rdelaage
/
postgresql
mirror of https://git.postgresql.org/git/postgresql.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
postgresql/src/backend/utils/mmgr/slab.c

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

769 lines
22 KiB
C
Raw Normal View History

Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
/*-------------------------------------------------------------------------
*
* slab.c
* SLAB allocator definitions.
*
* SLAB is a MemoryContext implementation designed for cases where large
* numbers of equally-sized objects are allocated (and freed).
*
*
Update copyright for 2022 Backpatch-through: 10
2022-01-08 01:04:57 +01:00
* Portions Copyright (c) 2017-2022, PostgreSQL Global Development Group
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
*
* IDENTIFICATION
* src/backend/utils/mmgr/slab.c
*
*
* NOTE:
* The constant allocation size allows significant simplification and various
* optimizations over more general purpose allocators. The blocks are carved
* into chunks of exactly the right size (plus alignment), not wasting any
* memory.
*
* The information about free chunks is maintained both at the block level and
* global (context) level. This is possible as the chunk size (and thus also
* the number of chunks per block) is fixed.
*
* On each block, free chunks are tracked in a simple linked list. Contents
* of free chunks is replaced with an index of the next free chunk, forming
* a very simple linked list. Each block also contains a counter of free
* chunks. Combined with the local block-level freelist, it makes it trivial
* to eventually free the whole block.
*
* At the context level, we use 'freelist' to track blocks ordered by number
* of free chunks, starting with blocks having a single allocated chunk, and
* with completely full blocks on the tail.
*
* This also allows various optimizations - for example when searching for
* free chunk, the allocator reuses space from the fullest blocks first, in
* the hope that some of the less full blocks will get completely empty (and
* returned back to the OS).
*
* For each block, we maintain pointer to the first free chunk - this is quite
* cheap and allows us to skip all the preceding used chunks, eliminating
Fix inconsistencies and typos in the tree, take 9 This addresses more issues with code comments, variable names and unreferenced variables. Author: Alexander Lakhin Discussion: https://postgr.es/m/7ab243e0-116d-3e44-d120-76b3df7abefd@gmail.com
2019-08-05 05:14:58 +02:00
* a significant number of lookups in many common usage patterns. In the worst
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
* case this performs as if the pointer was not maintained.
*
* We cache the freelist index for the blocks with the fewest free chunks
* (minFreeChunks), so that we don't have to search the freelist on every
* SlabAlloc() call, which is quite expensive.
*
*-------------------------------------------------------------------------
*/
#include "postgres.h"
Make the order of the header file includes consistent in backend modules. Similar to commits 7e735035f2 and dddf4cdc33, this commit makes the order of header file inclusion consistent for backend modules. In the passing, removed a couple of duplicate inclusions. Author: Vignesh C Reviewed-by: Kuntal Ghosh and Amit Kapila Discussion: https://postgr.es/m/CALDaNm2Sznv8RR6Ex-iJO6xAdsxgWhCoETkaYX=+9DW3q0QCfA@mail.gmail.com
2019-11-12 04:00:16 +01:00
#include "lib/ilist.h"
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
#include "utils/memdebug.h"
#include "utils/memutils.h"
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
#include "utils/memutils_memorychunk.h"
#include "utils/memutils_internal.h"
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
Use MAXALIGN() in calculations using sizeof(SlabBlock) c6e0fe1f2 added a new pointer field to SlabBlock to make it 4 bytes larger on 32-bit machines. Prior to that commit, the size of that struct was a multiple of 8, which meant that MAXALIGN(sizeof(SlabBlock)) was the same as sizeof(SlabBlock), however, after c6e0fe1f2, due to the addition of the new pointer field to store a pointer to the owning context, that was no longer true on builds with sizeof(void *) == 4. This problem was highlighted by an Assert failure which was checking that the pointer given to pfree() was MAXALIGNED. Various 32-bit ARM buildfarm animals were failing. These have MAXIMUM_ALIGNOF of 8. The only 32-bit testing I'd managed to do on c6e0fe1f2 had been on x86, which has a MAXIMUM_ALIGNOF of 4, therefore did not exhibit this issue. Here we define Slab_BLOCKHDRSZ and copy what is being done in aset.c and generation.c for doing calculations based on the size of the context's block type. This means that SlabAlloc() will now always return a MAXALIGNed pointer. This also fixes an incorrect sentinel_ok() check in SlabCheck() which was incorrectly checking the wrong sentinel byte. This must have previously not caused any issues due to the fullChunkSize never being large enough to store the sentinel byte. Diagnosed-by: Tomas Vondra, Tom Lane Author: Tomas Vondra, David Rowley Discussion: https://postgr.es/m/CAA4eK1%2B1JyW5TiL%3DyV-3Uq1CrfnTyn0Xrk5uArt31Z%3D8rgPhXQ%40mail.gmail.com
2022-08-30 04:36:04 +02:00
#define Slab_BLOCKHDRSZ MAXALIGN(sizeof(SlabBlock))
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
/*
* SlabContext is a specialized implementation of MemoryContext.
*/
typedef struct SlabContext
{
MemoryContextData header; /* Standard memory-context fields */
/* Allocation parameters for this context: */
Size chunkSize; /* chunk size */
Size fullChunkSize; /* chunk size including header and alignment */
Size blockSize; /* block size */
Rethink MemoryContext creation to improve performance. This patch makes a number of interrelated changes to reduce the overhead involved in creating/deleting memory contexts. The key ideas are: * Include the AllocSetContext header of an aset.c context in its first malloc request, rather than allocating it separately in TopMemoryContext. This means that we now always create an initial or "keeper" block in an aset, even if it never receives any allocation requests. * Create freelists in which we can save and recycle recently-destroyed asets (this idea is due to Robert Haas). * In the common case where the name of a context is a constant string, just store a pointer to it in the context header, rather than copying the string. The first change eliminates a palloc/pfree cycle per context, and also avoids bloat in TopMemoryContext, at the price that creating a context now involves a malloc/free cycle even if the context never receives any allocations. That would be a loser for some common usage patterns, but recycling short-lived contexts via the freelist eliminates that pain. Avoiding copying constant strings not only saves strlen() and strcpy() overhead, but is an essential part of the freelist optimization because it makes the context header size constant. Currently we make no attempt to use the freelist for contexts with non-constant names. (Perhaps someday we'll need to think harder about that, but in current usage, most contexts with custom names are long-lived anyway.) The freelist management in this initial commit is pretty simplistic, and we might want to refine it later --- but in common workloads that will never matter because the freelists will never get full anyway. To create a context with a non-constant name, one is now required to call AllocSetContextCreateExtended and specify the MEMCONTEXT_COPY_NAME option. AllocSetContextCreate becomes a wrapper macro, and it includes a test that will complain about non-string-literal context name parameters on gcc and similar compilers. An unfortunate side effect of making AllocSetContextCreate a macro is that one is now *required* to use the size parameter abstraction macros (ALLOCSET_DEFAULT_SIZES and friends) with it; the pre-9.6 habit of writing out individual size parameters no longer works unless you switch to AllocSetContextCreateExtended. Internally to the memory-context-related modules, the context creation APIs are simplified, removing the rather baroque original design whereby a context-type module called mcxt.c which then called back into the context-type module. That saved a bit of code duplication, but not much, and it prevented context-type modules from exercising control over the allocation of context headers. In passing, I converted the test-and-elog validation of aset size parameters into Asserts to save a few more cycles. The original thought was that callers might compute size parameters on the fly, but in practice nobody does that, so it's useless to expend cycles on checking those numbers in production builds. Also, mark the memory context method-pointer structs "const", just for cleanliness. Discussion: https://postgr.es/m/2264.1512870796@sss.pgh.pa.us
2017-12-13 19:55:12 +01:00
Size headerSize; /* allocated size of context header */
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
int chunksPerBlock; /* number of chunks per block */
int minFreeChunks; /* min number of free chunks in any block */
int nblocks; /* number of blocks allocated */
Allocate freechunks bitmap as part of SlabContext The bitmap used by SlabCheck to cross-check free chunks in a block used to be allocated for each SlabCheck call, and was never freed. The memory leak could be fixed by simply adding a pfree call, but it's actually a bad idea to do any allocations in SlabCheck at all as it assumes the state of the memory management as a whole is sane. So instead we allocate the bitmap as part of SlabContext, which means we don't need to do any allocations in SlabCheck and the bitmap goes away together with the SlabContext. Backpatch to 10, where the Slab context was introduced. Author: Tomas Vondra Reported-by: Andres Freund Reviewed-by: Tom Lane Backpatch-through: 10 Discussion: https://www.postgresql.org/message-id/20200116044119.g45f7pmgz4jmodxj%40alap3.anarazel.de
2020-01-17 14:06:28 +01:00
#ifdef MEMORY_CONTEXT_CHECKING
bool *freechunks; /* bitmap of free chunks in a block */
#endif
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
/* blocks with free space, grouped by number of free chunks: */
dlist_head freelist[FLEXIBLE_ARRAY_MEMBER];
} SlabContext;
/*
* SlabBlock
* Structure of a single block in SLAB allocator.
*
* node: doubly-linked list of blocks in global freelist
* nfree: number of free chunks in this block
* firstFreeChunk: index of the first free chunk
*/
typedef struct SlabBlock
{
dlist_node node; /* doubly-linked list */
int nfree; /* number of free chunks */
int firstFreeChunk; /* index of the first free chunk in the block */
Reduce size of common allocation header. The new slab allocator needs different per-allocation information than the classical aset.c. The definition in 58b25e981 wasn't sufficiently careful on 32 platforms with 8 byte alignment, leading to buildfarm failures. That's not entirely easy to fix by just adjusting the definition. As slab.c doesn't actually need the size part(s) of the common header, all chunks are equally sized after all, it seems better to instead reduce the header to the part needed by all allocators, namely which context an allocation belongs to. That has the advantage of reducing the overhead of slab allocations, and also allows for more flexibility in future allocators. To avoid spreading the logic about accessing a chunk's context around, centralize it in GetMemoryChunkContext(), which allows to delete a good number of lines. A followup commit will revise the mmgr/README portion about StandardChunkHeader, and more. Author: Andres Freund Discussion: https://postgr.es/m/20170228074420.aazv4iw6k562mnxg@alap3.anarazel.de
2017-02-28 08:32:22 +01:00
SlabContext *slab; /* owning context */
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
} SlabBlock;
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
#define Slab_CHUNKHDRSZ sizeof(MemoryChunk)
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
#define SlabPointerGetChunk(ptr) \
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
((MemoryChunk *)(((char *)(ptr)) - sizeof(MemoryChunk)))
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
#define SlabChunkGetPointer(chk) \
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
((void *)(((char *)(chk)) + sizeof(MemoryChunk)))
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
#define SlabBlockGetChunk(slab, block, idx) \
Use MAXALIGN() in calculations using sizeof(SlabBlock) c6e0fe1f2 added a new pointer field to SlabBlock to make it 4 bytes larger on 32-bit machines. Prior to that commit, the size of that struct was a multiple of 8, which meant that MAXALIGN(sizeof(SlabBlock)) was the same as sizeof(SlabBlock), however, after c6e0fe1f2, due to the addition of the new pointer field to store a pointer to the owning context, that was no longer true on builds with sizeof(void *) == 4. This problem was highlighted by an Assert failure which was checking that the pointer given to pfree() was MAXALIGNED. Various 32-bit ARM buildfarm animals were failing. These have MAXIMUM_ALIGNOF of 8. The only 32-bit testing I'd managed to do on c6e0fe1f2 had been on x86, which has a MAXIMUM_ALIGNOF of 4, therefore did not exhibit this issue. Here we define Slab_BLOCKHDRSZ and copy what is being done in aset.c and generation.c for doing calculations based on the size of the context's block type. This means that SlabAlloc() will now always return a MAXALIGNed pointer. This also fixes an incorrect sentinel_ok() check in SlabCheck() which was incorrectly checking the wrong sentinel byte. This must have previously not caused any issues due to the fullChunkSize never being large enough to store the sentinel byte. Diagnosed-by: Tomas Vondra, Tom Lane Author: Tomas Vondra, David Rowley Discussion: https://postgr.es/m/CAA4eK1%2B1JyW5TiL%3DyV-3Uq1CrfnTyn0Xrk5uArt31Z%3D8rgPhXQ%40mail.gmail.com
2022-08-30 04:36:04 +02:00
((MemoryChunk *) ((char *) (block) + Slab_BLOCKHDRSZ \
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
+ (idx * slab->fullChunkSize)))
#define SlabBlockStart(block) \
Use MAXALIGN() in calculations using sizeof(SlabBlock) c6e0fe1f2 added a new pointer field to SlabBlock to make it 4 bytes larger on 32-bit machines. Prior to that commit, the size of that struct was a multiple of 8, which meant that MAXALIGN(sizeof(SlabBlock)) was the same as sizeof(SlabBlock), however, after c6e0fe1f2, due to the addition of the new pointer field to store a pointer to the owning context, that was no longer true on builds with sizeof(void *) == 4. This problem was highlighted by an Assert failure which was checking that the pointer given to pfree() was MAXALIGNED. Various 32-bit ARM buildfarm animals were failing. These have MAXIMUM_ALIGNOF of 8. The only 32-bit testing I'd managed to do on c6e0fe1f2 had been on x86, which has a MAXIMUM_ALIGNOF of 4, therefore did not exhibit this issue. Here we define Slab_BLOCKHDRSZ and copy what is being done in aset.c and generation.c for doing calculations based on the size of the context's block type. This means that SlabAlloc() will now always return a MAXALIGNed pointer. This also fixes an incorrect sentinel_ok() check in SlabCheck() which was incorrectly checking the wrong sentinel byte. This must have previously not caused any issues due to the fullChunkSize never being large enough to store the sentinel byte. Diagnosed-by: Tomas Vondra, Tom Lane Author: Tomas Vondra, David Rowley Discussion: https://postgr.es/m/CAA4eK1%2B1JyW5TiL%3DyV-3Uq1CrfnTyn0Xrk5uArt31Z%3D8rgPhXQ%40mail.gmail.com
2022-08-30 04:36:04 +02:00
((char *) block + Slab_BLOCKHDRSZ)
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
#define SlabChunkIndex(slab, block, chunk) \
(((char *) chunk - SlabBlockStart(block)) / slab->fullChunkSize)
/*
* SlabContextCreate
* Create a new Slab context.
*
* parent: parent context, or NULL if top-level context
Allow memory contexts to have both fixed and variable ident strings. Originally, we treated memory context names as potentially variable in all cases, and therefore always copied them into the context header. Commit 9fa6f00b1 rethought this a little bit and invented a distinction between fixed and variable names, skipping the copy step for the former. But we can make things both simpler and more useful by instead allowing there to be two parts to a context's identification, a fixed "name" and an optional, variable "ident". The name supplied in the context create call is now required to be a compile-time-constant string in all cases, as it is never copied but just pointed to. The "ident" string, if wanted, is supplied later. This is needed because typically we want the ident to be stored inside the context so that it's cleaned up automatically on context deletion; that means it has to be copied into the context before we can set the pointer. The cost of this approach is basically just an additional pointer field in struct MemoryContextData, which isn't much overhead, and is bought back entirely in the AllocSet case by not needing a headerSize field anymore, since we no longer have to cope with variable header length. In addition, we can simplify the internal interfaces for memory context creation still further, saving a few cycles there. And it's no longer true that a custom identifier disqualifies a context from participating in aset.c's freelist scheme, so possibly there's some win on that end. All the places that were using non-compile-time-constant context names are adjusted to put the variable info into the "ident" instead. This allows more effective identification of those contexts in many cases; for example, subsidary contexts of relcache entries are now identified by both type (e.g. "index info") and relname, where before you got only one or the other. Contexts associated with PL function cache entries are now identified more fully and uniformly, too. I also arranged for plancache contexts to use the query source string as their identifier. This is basically free for CachedPlanSources, as they contained a copy of that string already. We pay an extra pstrdup to do it for CachedPlans. That could perhaps be avoided, but it would make things more fragile (since the CachedPlanSource is sometimes destroyed first). I suspect future improvements in error reporting will require CachedPlans to have a copy of that string anyway, so it's not clear that it's worth moving mountains to avoid it now. This also changes the APIs for context statistics routines so that the context-specific routines no longer assume that output goes straight to stderr, nor do they know all details of the output format. This is useful immediately to reduce code duplication, and it also allows for external code to do something with stats output that's different from printing to stderr. The reason for pushing this now rather than waiting for v12 is that it rethinks some of the API changes made by commit 9fa6f00b1. Seems better for extension authors to endure just one round of API changes not two. Discussion: https://postgr.es/m/CAB=Je-FdtmFZ9y9REHD7VsSrnCkiBhsA4mdsLKSPauwXtQBeNA@mail.gmail.com
2018-03-27 22:46:47 +02:00
* name: name of context (must be statically allocated)
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
* blockSize: allocation block size
* chunkSize: allocation chunk size
*
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
* The MAXALIGN(chunkSize) may not exceed MEMORYCHUNK_MAX_VALUE
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
*/
MemoryContext
SlabContextCreate(MemoryContext parent,
const char *name,
Size blockSize,
Size chunkSize)
{
int chunksPerBlock;
Size fullChunkSize;
Size freelistSize;
Rethink MemoryContext creation to improve performance. This patch makes a number of interrelated changes to reduce the overhead involved in creating/deleting memory contexts. The key ideas are: * Include the AllocSetContext header of an aset.c context in its first malloc request, rather than allocating it separately in TopMemoryContext. This means that we now always create an initial or "keeper" block in an aset, even if it never receives any allocation requests. * Create freelists in which we can save and recycle recently-destroyed asets (this idea is due to Robert Haas). * In the common case where the name of a context is a constant string, just store a pointer to it in the context header, rather than copying the string. The first change eliminates a palloc/pfree cycle per context, and also avoids bloat in TopMemoryContext, at the price that creating a context now involves a malloc/free cycle even if the context never receives any allocations. That would be a loser for some common usage patterns, but recycling short-lived contexts via the freelist eliminates that pain. Avoiding copying constant strings not only saves strlen() and strcpy() overhead, but is an essential part of the freelist optimization because it makes the context header size constant. Currently we make no attempt to use the freelist for contexts with non-constant names. (Perhaps someday we'll need to think harder about that, but in current usage, most contexts with custom names are long-lived anyway.) The freelist management in this initial commit is pretty simplistic, and we might want to refine it later --- but in common workloads that will never matter because the freelists will never get full anyway. To create a context with a non-constant name, one is now required to call AllocSetContextCreateExtended and specify the MEMCONTEXT_COPY_NAME option. AllocSetContextCreate becomes a wrapper macro, and it includes a test that will complain about non-string-literal context name parameters on gcc and similar compilers. An unfortunate side effect of making AllocSetContextCreate a macro is that one is now *required* to use the size parameter abstraction macros (ALLOCSET_DEFAULT_SIZES and friends) with it; the pre-9.6 habit of writing out individual size parameters no longer works unless you switch to AllocSetContextCreateExtended. Internally to the memory-context-related modules, the context creation APIs are simplified, removing the rather baroque original design whereby a context-type module called mcxt.c which then called back into the context-type module. That saved a bit of code duplication, but not much, and it prevented context-type modules from exercising control over the allocation of context headers. In passing, I converted the test-and-elog validation of aset size parameters into Asserts to save a few more cycles. The original thought was that callers might compute size parameters on the fly, but in practice nobody does that, so it's useless to expend cycles on checking those numbers in production builds. Also, mark the memory context method-pointer structs "const", just for cleanliness. Discussion: https://postgr.es/m/2264.1512870796@sss.pgh.pa.us
2017-12-13 19:55:12 +01:00
Size headerSize;
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
SlabContext *slab;
Rethink MemoryContext creation to improve performance. This patch makes a number of interrelated changes to reduce the overhead involved in creating/deleting memory contexts. The key ideas are: * Include the AllocSetContext header of an aset.c context in its first malloc request, rather than allocating it separately in TopMemoryContext. This means that we now always create an initial or "keeper" block in an aset, even if it never receives any allocation requests. * Create freelists in which we can save and recycle recently-destroyed asets (this idea is due to Robert Haas). * In the common case where the name of a context is a constant string, just store a pointer to it in the context header, rather than copying the string. The first change eliminates a palloc/pfree cycle per context, and also avoids bloat in TopMemoryContext, at the price that creating a context now involves a malloc/free cycle even if the context never receives any allocations. That would be a loser for some common usage patterns, but recycling short-lived contexts via the freelist eliminates that pain. Avoiding copying constant strings not only saves strlen() and strcpy() overhead, but is an essential part of the freelist optimization because it makes the context header size constant. Currently we make no attempt to use the freelist for contexts with non-constant names. (Perhaps someday we'll need to think harder about that, but in current usage, most contexts with custom names are long-lived anyway.) The freelist management in this initial commit is pretty simplistic, and we might want to refine it later --- but in common workloads that will never matter because the freelists will never get full anyway. To create a context with a non-constant name, one is now required to call AllocSetContextCreateExtended and specify the MEMCONTEXT_COPY_NAME option. AllocSetContextCreate becomes a wrapper macro, and it includes a test that will complain about non-string-literal context name parameters on gcc and similar compilers. An unfortunate side effect of making AllocSetContextCreate a macro is that one is now *required* to use the size parameter abstraction macros (ALLOCSET_DEFAULT_SIZES and friends) with it; the pre-9.6 habit of writing out individual size parameters no longer works unless you switch to AllocSetContextCreateExtended. Internally to the memory-context-related modules, the context creation APIs are simplified, removing the rather baroque original design whereby a context-type module called mcxt.c which then called back into the context-type module. That saved a bit of code duplication, but not much, and it prevented context-type modules from exercising control over the allocation of context headers. In passing, I converted the test-and-elog validation of aset size parameters into Asserts to save a few more cycles. The original thought was that callers might compute size parameters on the fly, but in practice nobody does that, so it's useless to expend cycles on checking those numbers in production builds. Also, mark the memory context method-pointer structs "const", just for cleanliness. Discussion: https://postgr.es/m/2264.1512870796@sss.pgh.pa.us
2017-12-13 19:55:12 +01:00
int i;
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
/* ensure MemoryChunk's size is properly maxaligned */
StaticAssertStmt(Slab_CHUNKHDRSZ == MAXALIGN(Slab_CHUNKHDRSZ),
"sizeof(MemoryChunk) is not maxaligned");
Assert(MAXALIGN(chunkSize) <= MEMORYCHUNK_MAX_VALUE);
Reduce size of common allocation header. The new slab allocator needs different per-allocation information than the classical aset.c. The definition in 58b25e981 wasn't sufficiently careful on 32 platforms with 8 byte alignment, leading to buildfarm failures. That's not entirely easy to fix by just adjusting the definition. As slab.c doesn't actually need the size part(s) of the common header, all chunks are equally sized after all, it seems better to instead reduce the header to the part needed by all allocators, namely which context an allocation belongs to. That has the advantage of reducing the overhead of slab allocations, and also allows for more flexibility in future allocators. To avoid spreading the logic about accessing a chunk's context around, centralize it in GetMemoryChunkContext(), which allows to delete a good number of lines. A followup commit will revise the mmgr/README portion about StandardChunkHeader, and more. Author: Andres Freund Discussion: https://postgr.es/m/20170228074420.aazv4iw6k562mnxg@alap3.anarazel.de
2017-02-28 08:32:22 +01:00
Make slab allocator work on platforms with MAXIMUM_ALIGNOF < sizeof(int). Notably, m68k only needs 2-byte alignment. Per report from Christoph Berg. Discussion: https://www.postgresql.org/message-id/20170517193957.fwntkgi6epuso5l2@msg.df7cb.de
2017-05-18 21:22:13 +02:00
/* Make sure the linked list node fits inside a freed chunk */
if (chunkSize < sizeof(int))
chunkSize = sizeof(int);
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
/* chunk, including SLAB header (both addresses nicely aligned) */
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
fullChunkSize = Slab_CHUNKHDRSZ + MAXALIGN(chunkSize);
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
/* Make sure the block can store at least one chunk. */
Use MAXALIGN() in calculations using sizeof(SlabBlock) c6e0fe1f2 added a new pointer field to SlabBlock to make it 4 bytes larger on 32-bit machines. Prior to that commit, the size of that struct was a multiple of 8, which meant that MAXALIGN(sizeof(SlabBlock)) was the same as sizeof(SlabBlock), however, after c6e0fe1f2, due to the addition of the new pointer field to store a pointer to the owning context, that was no longer true on builds with sizeof(void *) == 4. This problem was highlighted by an Assert failure which was checking that the pointer given to pfree() was MAXALIGNED. Various 32-bit ARM buildfarm animals were failing. These have MAXIMUM_ALIGNOF of 8. The only 32-bit testing I'd managed to do on c6e0fe1f2 had been on x86, which has a MAXIMUM_ALIGNOF of 4, therefore did not exhibit this issue. Here we define Slab_BLOCKHDRSZ and copy what is being done in aset.c and generation.c for doing calculations based on the size of the context's block type. This means that SlabAlloc() will now always return a MAXALIGNed pointer. This also fixes an incorrect sentinel_ok() check in SlabCheck() which was incorrectly checking the wrong sentinel byte. This must have previously not caused any issues due to the fullChunkSize never being large enough to store the sentinel byte. Diagnosed-by: Tomas Vondra, Tom Lane Author: Tomas Vondra, David Rowley Discussion: https://postgr.es/m/CAA4eK1%2B1JyW5TiL%3DyV-3Uq1CrfnTyn0Xrk5uArt31Z%3D8rgPhXQ%40mail.gmail.com
2022-08-30 04:36:04 +02:00
if (blockSize < fullChunkSize + Slab_BLOCKHDRSZ)
Reduce size of common allocation header. The new slab allocator needs different per-allocation information than the classical aset.c. The definition in 58b25e981 wasn't sufficiently careful on 32 platforms with 8 byte alignment, leading to buildfarm failures. That's not entirely easy to fix by just adjusting the definition. As slab.c doesn't actually need the size part(s) of the common header, all chunks are equally sized after all, it seems better to instead reduce the header to the part needed by all allocators, namely which context an allocation belongs to. That has the advantage of reducing the overhead of slab allocations, and also allows for more flexibility in future allocators. To avoid spreading the logic about accessing a chunk's context around, centralize it in GetMemoryChunkContext(), which allows to delete a good number of lines. A followup commit will revise the mmgr/README portion about StandardChunkHeader, and more. Author: Andres Freund Discussion: https://postgr.es/m/20170228074420.aazv4iw6k562mnxg@alap3.anarazel.de
2017-02-28 08:32:22 +01:00
elog(ERROR, "block size %zu for slab is too small for %zu chunks",
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
blockSize, chunkSize);
/* Compute maximum number of chunks per block */
Use MAXALIGN() in calculations using sizeof(SlabBlock) c6e0fe1f2 added a new pointer field to SlabBlock to make it 4 bytes larger on 32-bit machines. Prior to that commit, the size of that struct was a multiple of 8, which meant that MAXALIGN(sizeof(SlabBlock)) was the same as sizeof(SlabBlock), however, after c6e0fe1f2, due to the addition of the new pointer field to store a pointer to the owning context, that was no longer true on builds with sizeof(void *) == 4. This problem was highlighted by an Assert failure which was checking that the pointer given to pfree() was MAXALIGNED. Various 32-bit ARM buildfarm animals were failing. These have MAXIMUM_ALIGNOF of 8. The only 32-bit testing I'd managed to do on c6e0fe1f2 had been on x86, which has a MAXIMUM_ALIGNOF of 4, therefore did not exhibit this issue. Here we define Slab_BLOCKHDRSZ and copy what is being done in aset.c and generation.c for doing calculations based on the size of the context's block type. This means that SlabAlloc() will now always return a MAXALIGNed pointer. This also fixes an incorrect sentinel_ok() check in SlabCheck() which was incorrectly checking the wrong sentinel byte. This must have previously not caused any issues due to the fullChunkSize never being large enough to store the sentinel byte. Diagnosed-by: Tomas Vondra, Tom Lane Author: Tomas Vondra, David Rowley Discussion: https://postgr.es/m/CAA4eK1%2B1JyW5TiL%3DyV-3Uq1CrfnTyn0Xrk5uArt31Z%3D8rgPhXQ%40mail.gmail.com
2022-08-30 04:36:04 +02:00
chunksPerBlock = (blockSize - Slab_BLOCKHDRSZ) / fullChunkSize;
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
/* The freelist starts with 0, ends with chunksPerBlock. */
freelistSize = sizeof(dlist_head) * (chunksPerBlock + 1);
Rethink MemoryContext creation to improve performance. This patch makes a number of interrelated changes to reduce the overhead involved in creating/deleting memory contexts. The key ideas are: * Include the AllocSetContext header of an aset.c context in its first malloc request, rather than allocating it separately in TopMemoryContext. This means that we now always create an initial or "keeper" block in an aset, even if it never receives any allocation requests. * Create freelists in which we can save and recycle recently-destroyed asets (this idea is due to Robert Haas). * In the common case where the name of a context is a constant string, just store a pointer to it in the context header, rather than copying the string. The first change eliminates a palloc/pfree cycle per context, and also avoids bloat in TopMemoryContext, at the price that creating a context now involves a malloc/free cycle even if the context never receives any allocations. That would be a loser for some common usage patterns, but recycling short-lived contexts via the freelist eliminates that pain. Avoiding copying constant strings not only saves strlen() and strcpy() overhead, but is an essential part of the freelist optimization because it makes the context header size constant. Currently we make no attempt to use the freelist for contexts with non-constant names. (Perhaps someday we'll need to think harder about that, but in current usage, most contexts with custom names are long-lived anyway.) The freelist management in this initial commit is pretty simplistic, and we might want to refine it later --- but in common workloads that will never matter because the freelists will never get full anyway. To create a context with a non-constant name, one is now required to call AllocSetContextCreateExtended and specify the MEMCONTEXT_COPY_NAME option. AllocSetContextCreate becomes a wrapper macro, and it includes a test that will complain about non-string-literal context name parameters on gcc and similar compilers. An unfortunate side effect of making AllocSetContextCreate a macro is that one is now *required* to use the size parameter abstraction macros (ALLOCSET_DEFAULT_SIZES and friends) with it; the pre-9.6 habit of writing out individual size parameters no longer works unless you switch to AllocSetContextCreateExtended. Internally to the memory-context-related modules, the context creation APIs are simplified, removing the rather baroque original design whereby a context-type module called mcxt.c which then called back into the context-type module. That saved a bit of code duplication, but not much, and it prevented context-type modules from exercising control over the allocation of context headers. In passing, I converted the test-and-elog validation of aset size parameters into Asserts to save a few more cycles. The original thought was that callers might compute size parameters on the fly, but in practice nobody does that, so it's useless to expend cycles on checking those numbers in production builds. Also, mark the memory context method-pointer structs "const", just for cleanliness. Discussion: https://postgr.es/m/2264.1512870796@sss.pgh.pa.us
2017-12-13 19:55:12 +01:00
/*
* Allocate the context header. Unlike aset.c, we never try to combine
* this with the first regular block; not worth the extra complication.
*/
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
Allow memory contexts to have both fixed and variable ident strings. Originally, we treated memory context names as potentially variable in all cases, and therefore always copied them into the context header. Commit 9fa6f00b1 rethought this a little bit and invented a distinction between fixed and variable names, skipping the copy step for the former. But we can make things both simpler and more useful by instead allowing there to be two parts to a context's identification, a fixed "name" and an optional, variable "ident". The name supplied in the context create call is now required to be a compile-time-constant string in all cases, as it is never copied but just pointed to. The "ident" string, if wanted, is supplied later. This is needed because typically we want the ident to be stored inside the context so that it's cleaned up automatically on context deletion; that means it has to be copied into the context before we can set the pointer. The cost of this approach is basically just an additional pointer field in struct MemoryContextData, which isn't much overhead, and is bought back entirely in the AllocSet case by not needing a headerSize field anymore, since we no longer have to cope with variable header length. In addition, we can simplify the internal interfaces for memory context creation still further, saving a few cycles there. And it's no longer true that a custom identifier disqualifies a context from participating in aset.c's freelist scheme, so possibly there's some win on that end. All the places that were using non-compile-time-constant context names are adjusted to put the variable info into the "ident" instead. This allows more effective identification of those contexts in many cases; for example, subsidary contexts of relcache entries are now identified by both type (e.g. "index info") and relname, where before you got only one or the other. Contexts associated with PL function cache entries are now identified more fully and uniformly, too. I also arranged for plancache contexts to use the query source string as their identifier. This is basically free for CachedPlanSources, as they contained a copy of that string already. We pay an extra pstrdup to do it for CachedPlans. That could perhaps be avoided, but it would make things more fragile (since the CachedPlanSource is sometimes destroyed first). I suspect future improvements in error reporting will require CachedPlans to have a copy of that string anyway, so it's not clear that it's worth moving mountains to avoid it now. This also changes the APIs for context statistics routines so that the context-specific routines no longer assume that output goes straight to stderr, nor do they know all details of the output format. This is useful immediately to reduce code duplication, and it also allows for external code to do something with stats output that's different from printing to stderr. The reason for pushing this now rather than waiting for v12 is that it rethinks some of the API changes made by commit 9fa6f00b1. Seems better for extension authors to endure just one round of API changes not two. Discussion: https://postgr.es/m/CAB=Je-FdtmFZ9y9REHD7VsSrnCkiBhsA4mdsLKSPauwXtQBeNA@mail.gmail.com
2018-03-27 22:46:47 +02:00
/* Size of the memory context header */
headerSize = offsetof(SlabContext, freelist) + freelistSize;
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
Allocate freechunks bitmap as part of SlabContext The bitmap used by SlabCheck to cross-check free chunks in a block used to be allocated for each SlabCheck call, and was never freed. The memory leak could be fixed by simply adding a pfree call, but it's actually a bad idea to do any allocations in SlabCheck at all as it assumes the state of the memory management as a whole is sane. So instead we allocate the bitmap as part of SlabContext, which means we don't need to do any allocations in SlabCheck and the bitmap goes away together with the SlabContext. Backpatch to 10, where the Slab context was introduced. Author: Tomas Vondra Reported-by: Andres Freund Reviewed-by: Tom Lane Backpatch-through: 10 Discussion: https://www.postgresql.org/message-id/20200116044119.g45f7pmgz4jmodxj%40alap3.anarazel.de
2020-01-17 14:06:28 +01:00
#ifdef MEMORY_CONTEXT_CHECKING
Initial pgindent and pgperltidy run for v13. Includes some manual cleanup of places that pgindent messed up, most of which weren't per project style anyway. Notably, it seems some people didn't absorb the style rules of commit c9d297751, because there were a bunch of new occurrences of function calls with a newline just after the left paren, all with faulty expectations about how the rest of the call would get indented.
2020-05-14 19:06:38 +02:00
Allocate freechunks bitmap as part of SlabContext The bitmap used by SlabCheck to cross-check free chunks in a block used to be allocated for each SlabCheck call, and was never freed. The memory leak could be fixed by simply adding a pfree call, but it's actually a bad idea to do any allocations in SlabCheck at all as it assumes the state of the memory management as a whole is sane. So instead we allocate the bitmap as part of SlabContext, which means we don't need to do any allocations in SlabCheck and the bitmap goes away together with the SlabContext. Backpatch to 10, where the Slab context was introduced. Author: Tomas Vondra Reported-by: Andres Freund Reviewed-by: Tom Lane Backpatch-through: 10 Discussion: https://www.postgresql.org/message-id/20200116044119.g45f7pmgz4jmodxj%40alap3.anarazel.de
2020-01-17 14:06:28 +01:00
/*
* With memory checking, we need to allocate extra space for the bitmap of
* free chunks. The bitmap is an array of bools, so we don't need to worry
* about alignment.
*/
headerSize += chunksPerBlock * sizeof(bool);
#endif
Rethink MemoryContext creation to improve performance. This patch makes a number of interrelated changes to reduce the overhead involved in creating/deleting memory contexts. The key ideas are: * Include the AllocSetContext header of an aset.c context in its first malloc request, rather than allocating it separately in TopMemoryContext. This means that we now always create an initial or "keeper" block in an aset, even if it never receives any allocation requests. * Create freelists in which we can save and recycle recently-destroyed asets (this idea is due to Robert Haas). * In the common case where the name of a context is a constant string, just store a pointer to it in the context header, rather than copying the string. The first change eliminates a palloc/pfree cycle per context, and also avoids bloat in TopMemoryContext, at the price that creating a context now involves a malloc/free cycle even if the context never receives any allocations. That would be a loser for some common usage patterns, but recycling short-lived contexts via the freelist eliminates that pain. Avoiding copying constant strings not only saves strlen() and strcpy() overhead, but is an essential part of the freelist optimization because it makes the context header size constant. Currently we make no attempt to use the freelist for contexts with non-constant names. (Perhaps someday we'll need to think harder about that, but in current usage, most contexts with custom names are long-lived anyway.) The freelist management in this initial commit is pretty simplistic, and we might want to refine it later --- but in common workloads that will never matter because the freelists will never get full anyway. To create a context with a non-constant name, one is now required to call AllocSetContextCreateExtended and specify the MEMCONTEXT_COPY_NAME option. AllocSetContextCreate becomes a wrapper macro, and it includes a test that will complain about non-string-literal context name parameters on gcc and similar compilers. An unfortunate side effect of making AllocSetContextCreate a macro is that one is now *required* to use the size parameter abstraction macros (ALLOCSET_DEFAULT_SIZES and friends) with it; the pre-9.6 habit of writing out individual size parameters no longer works unless you switch to AllocSetContextCreateExtended. Internally to the memory-context-related modules, the context creation APIs are simplified, removing the rather baroque original design whereby a context-type module called mcxt.c which then called back into the context-type module. That saved a bit of code duplication, but not much, and it prevented context-type modules from exercising control over the allocation of context headers. In passing, I converted the test-and-elog validation of aset size parameters into Asserts to save a few more cycles. The original thought was that callers might compute size parameters on the fly, but in practice nobody does that, so it's useless to expend cycles on checking those numbers in production builds. Also, mark the memory context method-pointer structs "const", just for cleanliness. Discussion: https://postgr.es/m/2264.1512870796@sss.pgh.pa.us
2017-12-13 19:55:12 +01:00
slab = (SlabContext *) malloc(headerSize);
if (slab == NULL)
{
MemoryContextStats(TopMemoryContext);
ereport(ERROR,
(errcode(ERRCODE_OUT_OF_MEMORY),
errmsg("out of memory"),
errdetail("Failed while creating memory context \"%s\".",
name)));
}
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
Rethink MemoryContext creation to improve performance. This patch makes a number of interrelated changes to reduce the overhead involved in creating/deleting memory contexts. The key ideas are: * Include the AllocSetContext header of an aset.c context in its first malloc request, rather than allocating it separately in TopMemoryContext. This means that we now always create an initial or "keeper" block in an aset, even if it never receives any allocation requests. * Create freelists in which we can save and recycle recently-destroyed asets (this idea is due to Robert Haas). * In the common case where the name of a context is a constant string, just store a pointer to it in the context header, rather than copying the string. The first change eliminates a palloc/pfree cycle per context, and also avoids bloat in TopMemoryContext, at the price that creating a context now involves a malloc/free cycle even if the context never receives any allocations. That would be a loser for some common usage patterns, but recycling short-lived contexts via the freelist eliminates that pain. Avoiding copying constant strings not only saves strlen() and strcpy() overhead, but is an essential part of the freelist optimization because it makes the context header size constant. Currently we make no attempt to use the freelist for contexts with non-constant names. (Perhaps someday we'll need to think harder about that, but in current usage, most contexts with custom names are long-lived anyway.) The freelist management in this initial commit is pretty simplistic, and we might want to refine it later --- but in common workloads that will never matter because the freelists will never get full anyway. To create a context with a non-constant name, one is now required to call AllocSetContextCreateExtended and specify the MEMCONTEXT_COPY_NAME option. AllocSetContextCreate becomes a wrapper macro, and it includes a test that will complain about non-string-literal context name parameters on gcc and similar compilers. An unfortunate side effect of making AllocSetContextCreate a macro is that one is now *required* to use the size parameter abstraction macros (ALLOCSET_DEFAULT_SIZES and friends) with it; the pre-9.6 habit of writing out individual size parameters no longer works unless you switch to AllocSetContextCreateExtended. Internally to the memory-context-related modules, the context creation APIs are simplified, removing the rather baroque original design whereby a context-type module called mcxt.c which then called back into the context-type module. That saved a bit of code duplication, but not much, and it prevented context-type modules from exercising control over the allocation of context headers. In passing, I converted the test-and-elog validation of aset size parameters into Asserts to save a few more cycles. The original thought was that callers might compute size parameters on the fly, but in practice nobody does that, so it's useless to expend cycles on checking those numbers in production builds. Also, mark the memory context method-pointer structs "const", just for cleanliness. Discussion: https://postgr.es/m/2264.1512870796@sss.pgh.pa.us
2017-12-13 19:55:12 +01:00
/*
* Avoid writing code that can fail between here and MemoryContextCreate;
* we'd leak the header if we ereport in this stretch.
*/
/* Fill in SlabContext-specific header fields */
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
slab->chunkSize = chunkSize;
slab->fullChunkSize = fullChunkSize;
Rethink MemoryContext creation to improve performance. This patch makes a number of interrelated changes to reduce the overhead involved in creating/deleting memory contexts. The key ideas are: * Include the AllocSetContext header of an aset.c context in its first malloc request, rather than allocating it separately in TopMemoryContext. This means that we now always create an initial or "keeper" block in an aset, even if it never receives any allocation requests. * Create freelists in which we can save and recycle recently-destroyed asets (this idea is due to Robert Haas). * In the common case where the name of a context is a constant string, just store a pointer to it in the context header, rather than copying the string. The first change eliminates a palloc/pfree cycle per context, and also avoids bloat in TopMemoryContext, at the price that creating a context now involves a malloc/free cycle even if the context never receives any allocations. That would be a loser for some common usage patterns, but recycling short-lived contexts via the freelist eliminates that pain. Avoiding copying constant strings not only saves strlen() and strcpy() overhead, but is an essential part of the freelist optimization because it makes the context header size constant. Currently we make no attempt to use the freelist for contexts with non-constant names. (Perhaps someday we'll need to think harder about that, but in current usage, most contexts with custom names are long-lived anyway.) The freelist management in this initial commit is pretty simplistic, and we might want to refine it later --- but in common workloads that will never matter because the freelists will never get full anyway. To create a context with a non-constant name, one is now required to call AllocSetContextCreateExtended and specify the MEMCONTEXT_COPY_NAME option. AllocSetContextCreate becomes a wrapper macro, and it includes a test that will complain about non-string-literal context name parameters on gcc and similar compilers. An unfortunate side effect of making AllocSetContextCreate a macro is that one is now *required* to use the size parameter abstraction macros (ALLOCSET_DEFAULT_SIZES and friends) with it; the pre-9.6 habit of writing out individual size parameters no longer works unless you switch to AllocSetContextCreateExtended. Internally to the memory-context-related modules, the context creation APIs are simplified, removing the rather baroque original design whereby a context-type module called mcxt.c which then called back into the context-type module. That saved a bit of code duplication, but not much, and it prevented context-type modules from exercising control over the allocation of context headers. In passing, I converted the test-and-elog validation of aset size parameters into Asserts to save a few more cycles. The original thought was that callers might compute size parameters on the fly, but in practice nobody does that, so it's useless to expend cycles on checking those numbers in production builds. Also, mark the memory context method-pointer structs "const", just for cleanliness. Discussion: https://postgr.es/m/2264.1512870796@sss.pgh.pa.us
2017-12-13 19:55:12 +01:00
slab->blockSize = blockSize;
slab->headerSize = headerSize;
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
slab->chunksPerBlock = chunksPerBlock;
slab->minFreeChunks = 0;
Rethink MemoryContext creation to improve performance. This patch makes a number of interrelated changes to reduce the overhead involved in creating/deleting memory contexts. The key ideas are: * Include the AllocSetContext header of an aset.c context in its first malloc request, rather than allocating it separately in TopMemoryContext. This means that we now always create an initial or "keeper" block in an aset, even if it never receives any allocation requests. * Create freelists in which we can save and recycle recently-destroyed asets (this idea is due to Robert Haas). * In the common case where the name of a context is a constant string, just store a pointer to it in the context header, rather than copying the string. The first change eliminates a palloc/pfree cycle per context, and also avoids bloat in TopMemoryContext, at the price that creating a context now involves a malloc/free cycle even if the context never receives any allocations. That would be a loser for some common usage patterns, but recycling short-lived contexts via the freelist eliminates that pain. Avoiding copying constant strings not only saves strlen() and strcpy() overhead, but is an essential part of the freelist optimization because it makes the context header size constant. Currently we make no attempt to use the freelist for contexts with non-constant names. (Perhaps someday we'll need to think harder about that, but in current usage, most contexts with custom names are long-lived anyway.) The freelist management in this initial commit is pretty simplistic, and we might want to refine it later --- but in common workloads that will never matter because the freelists will never get full anyway. To create a context with a non-constant name, one is now required to call AllocSetContextCreateExtended and specify the MEMCONTEXT_COPY_NAME option. AllocSetContextCreate becomes a wrapper macro, and it includes a test that will complain about non-string-literal context name parameters on gcc and similar compilers. An unfortunate side effect of making AllocSetContextCreate a macro is that one is now *required* to use the size parameter abstraction macros (ALLOCSET_DEFAULT_SIZES and friends) with it; the pre-9.6 habit of writing out individual size parameters no longer works unless you switch to AllocSetContextCreateExtended. Internally to the memory-context-related modules, the context creation APIs are simplified, removing the rather baroque original design whereby a context-type module called mcxt.c which then called back into the context-type module. That saved a bit of code duplication, but not much, and it prevented context-type modules from exercising control over the allocation of context headers. In passing, I converted the test-and-elog validation of aset size parameters into Asserts to save a few more cycles. The original thought was that callers might compute size parameters on the fly, but in practice nobody does that, so it's useless to expend cycles on checking those numbers in production builds. Also, mark the memory context method-pointer structs "const", just for cleanliness. Discussion: https://postgr.es/m/2264.1512870796@sss.pgh.pa.us
2017-12-13 19:55:12 +01:00
slab->nblocks = 0;
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
/* initialize the freelist slots */
for (i = 0; i < (slab->chunksPerBlock + 1); i++)
dlist_init(&slab->freelist[i]);
Rethink MemoryContext creation to improve performance. This patch makes a number of interrelated changes to reduce the overhead involved in creating/deleting memory contexts. The key ideas are: * Include the AllocSetContext header of an aset.c context in its first malloc request, rather than allocating it separately in TopMemoryContext. This means that we now always create an initial or "keeper" block in an aset, even if it never receives any allocation requests. * Create freelists in which we can save and recycle recently-destroyed asets (this idea is due to Robert Haas). * In the common case where the name of a context is a constant string, just store a pointer to it in the context header, rather than copying the string. The first change eliminates a palloc/pfree cycle per context, and also avoids bloat in TopMemoryContext, at the price that creating a context now involves a malloc/free cycle even if the context never receives any allocations. That would be a loser for some common usage patterns, but recycling short-lived contexts via the freelist eliminates that pain. Avoiding copying constant strings not only saves strlen() and strcpy() overhead, but is an essential part of the freelist optimization because it makes the context header size constant. Currently we make no attempt to use the freelist for contexts with non-constant names. (Perhaps someday we'll need to think harder about that, but in current usage, most contexts with custom names are long-lived anyway.) The freelist management in this initial commit is pretty simplistic, and we might want to refine it later --- but in common workloads that will never matter because the freelists will never get full anyway. To create a context with a non-constant name, one is now required to call AllocSetContextCreateExtended and specify the MEMCONTEXT_COPY_NAME option. AllocSetContextCreate becomes a wrapper macro, and it includes a test that will complain about non-string-literal context name parameters on gcc and similar compilers. An unfortunate side effect of making AllocSetContextCreate a macro is that one is now *required* to use the size parameter abstraction macros (ALLOCSET_DEFAULT_SIZES and friends) with it; the pre-9.6 habit of writing out individual size parameters no longer works unless you switch to AllocSetContextCreateExtended. Internally to the memory-context-related modules, the context creation APIs are simplified, removing the rather baroque original design whereby a context-type module called mcxt.c which then called back into the context-type module. That saved a bit of code duplication, but not much, and it prevented context-type modules from exercising control over the allocation of context headers. In passing, I converted the test-and-elog validation of aset size parameters into Asserts to save a few more cycles. The original thought was that callers might compute size parameters on the fly, but in practice nobody does that, so it's useless to expend cycles on checking those numbers in production builds. Also, mark the memory context method-pointer structs "const", just for cleanliness. Discussion: https://postgr.es/m/2264.1512870796@sss.pgh.pa.us
2017-12-13 19:55:12 +01:00
Allocate freechunks bitmap as part of SlabContext The bitmap used by SlabCheck to cross-check free chunks in a block used to be allocated for each SlabCheck call, and was never freed. The memory leak could be fixed by simply adding a pfree call, but it's actually a bad idea to do any allocations in SlabCheck at all as it assumes the state of the memory management as a whole is sane. So instead we allocate the bitmap as part of SlabContext, which means we don't need to do any allocations in SlabCheck and the bitmap goes away together with the SlabContext. Backpatch to 10, where the Slab context was introduced. Author: Tomas Vondra Reported-by: Andres Freund Reviewed-by: Tom Lane Backpatch-through: 10 Discussion: https://www.postgresql.org/message-id/20200116044119.g45f7pmgz4jmodxj%40alap3.anarazel.de
2020-01-17 14:06:28 +01:00
#ifdef MEMORY_CONTEXT_CHECKING
/* set the freechunks pointer right after the freelists array */
slab->freechunks
= (bool *) slab + offsetof(SlabContext, freelist) + freelistSize;
#endif
Rethink MemoryContext creation to improve performance. This patch makes a number of interrelated changes to reduce the overhead involved in creating/deleting memory contexts. The key ideas are: * Include the AllocSetContext header of an aset.c context in its first malloc request, rather than allocating it separately in TopMemoryContext. This means that we now always create an initial or "keeper" block in an aset, even if it never receives any allocation requests. * Create freelists in which we can save and recycle recently-destroyed asets (this idea is due to Robert Haas). * In the common case where the name of a context is a constant string, just store a pointer to it in the context header, rather than copying the string. The first change eliminates a palloc/pfree cycle per context, and also avoids bloat in TopMemoryContext, at the price that creating a context now involves a malloc/free cycle even if the context never receives any allocations. That would be a loser for some common usage patterns, but recycling short-lived contexts via the freelist eliminates that pain. Avoiding copying constant strings not only saves strlen() and strcpy() overhead, but is an essential part of the freelist optimization because it makes the context header size constant. Currently we make no attempt to use the freelist for contexts with non-constant names. (Perhaps someday we'll need to think harder about that, but in current usage, most contexts with custom names are long-lived anyway.) The freelist management in this initial commit is pretty simplistic, and we might want to refine it later --- but in common workloads that will never matter because the freelists will never get full anyway. To create a context with a non-constant name, one is now required to call AllocSetContextCreateExtended and specify the MEMCONTEXT_COPY_NAME option. AllocSetContextCreate becomes a wrapper macro, and it includes a test that will complain about non-string-literal context name parameters on gcc and similar compilers. An unfortunate side effect of making AllocSetContextCreate a macro is that one is now *required* to use the size parameter abstraction macros (ALLOCSET_DEFAULT_SIZES and friends) with it; the pre-9.6 habit of writing out individual size parameters no longer works unless you switch to AllocSetContextCreateExtended. Internally to the memory-context-related modules, the context creation APIs are simplified, removing the rather baroque original design whereby a context-type module called mcxt.c which then called back into the context-type module. That saved a bit of code duplication, but not much, and it prevented context-type modules from exercising control over the allocation of context headers. In passing, I converted the test-and-elog validation of aset size parameters into Asserts to save a few more cycles. The original thought was that callers might compute size parameters on the fly, but in practice nobody does that, so it's useless to expend cycles on checking those numbers in production builds. Also, mark the memory context method-pointer structs "const", just for cleanliness. Discussion: https://postgr.es/m/2264.1512870796@sss.pgh.pa.us
2017-12-13 19:55:12 +01:00
/* Finally, do the type-independent part of context creation */
MemoryContextCreate((MemoryContext) slab,
T_SlabContext,
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
MCTX_SLAB_ID,
Rethink MemoryContext creation to improve performance. This patch makes a number of interrelated changes to reduce the overhead involved in creating/deleting memory contexts. The key ideas are: * Include the AllocSetContext header of an aset.c context in its first malloc request, rather than allocating it separately in TopMemoryContext. This means that we now always create an initial or "keeper" block in an aset, even if it never receives any allocation requests. * Create freelists in which we can save and recycle recently-destroyed asets (this idea is due to Robert Haas). * In the common case where the name of a context is a constant string, just store a pointer to it in the context header, rather than copying the string. The first change eliminates a palloc/pfree cycle per context, and also avoids bloat in TopMemoryContext, at the price that creating a context now involves a malloc/free cycle even if the context never receives any allocations. That would be a loser for some common usage patterns, but recycling short-lived contexts via the freelist eliminates that pain. Avoiding copying constant strings not only saves strlen() and strcpy() overhead, but is an essential part of the freelist optimization because it makes the context header size constant. Currently we make no attempt to use the freelist for contexts with non-constant names. (Perhaps someday we'll need to think harder about that, but in current usage, most contexts with custom names are long-lived anyway.) The freelist management in this initial commit is pretty simplistic, and we might want to refine it later --- but in common workloads that will never matter because the freelists will never get full anyway. To create a context with a non-constant name, one is now required to call AllocSetContextCreateExtended and specify the MEMCONTEXT_COPY_NAME option. AllocSetContextCreate becomes a wrapper macro, and it includes a test that will complain about non-string-literal context name parameters on gcc and similar compilers. An unfortunate side effect of making AllocSetContextCreate a macro is that one is now *required* to use the size parameter abstraction macros (ALLOCSET_DEFAULT_SIZES and friends) with it; the pre-9.6 habit of writing out individual size parameters no longer works unless you switch to AllocSetContextCreateExtended. Internally to the memory-context-related modules, the context creation APIs are simplified, removing the rather baroque original design whereby a context-type module called mcxt.c which then called back into the context-type module. That saved a bit of code duplication, but not much, and it prevented context-type modules from exercising control over the allocation of context headers. In passing, I converted the test-and-elog validation of aset size parameters into Asserts to save a few more cycles. The original thought was that callers might compute size parameters on the fly, but in practice nobody does that, so it's useless to expend cycles on checking those numbers in production builds. Also, mark the memory context method-pointer structs "const", just for cleanliness. Discussion: https://postgr.es/m/2264.1512870796@sss.pgh.pa.us
2017-12-13 19:55:12 +01:00
parent,
Allow memory contexts to have both fixed and variable ident strings. Originally, we treated memory context names as potentially variable in all cases, and therefore always copied them into the context header. Commit 9fa6f00b1 rethought this a little bit and invented a distinction between fixed and variable names, skipping the copy step for the former. But we can make things both simpler and more useful by instead allowing there to be two parts to a context's identification, a fixed "name" and an optional, variable "ident". The name supplied in the context create call is now required to be a compile-time-constant string in all cases, as it is never copied but just pointed to. The "ident" string, if wanted, is supplied later. This is needed because typically we want the ident to be stored inside the context so that it's cleaned up automatically on context deletion; that means it has to be copied into the context before we can set the pointer. The cost of this approach is basically just an additional pointer field in struct MemoryContextData, which isn't much overhead, and is bought back entirely in the AllocSet case by not needing a headerSize field anymore, since we no longer have to cope with variable header length. In addition, we can simplify the internal interfaces for memory context creation still further, saving a few cycles there. And it's no longer true that a custom identifier disqualifies a context from participating in aset.c's freelist scheme, so possibly there's some win on that end. All the places that were using non-compile-time-constant context names are adjusted to put the variable info into the "ident" instead. This allows more effective identification of those contexts in many cases; for example, subsidary contexts of relcache entries are now identified by both type (e.g. "index info") and relname, where before you got only one or the other. Contexts associated with PL function cache entries are now identified more fully and uniformly, too. I also arranged for plancache contexts to use the query source string as their identifier. This is basically free for CachedPlanSources, as they contained a copy of that string already. We pay an extra pstrdup to do it for CachedPlans. That could perhaps be avoided, but it would make things more fragile (since the CachedPlanSource is sometimes destroyed first). I suspect future improvements in error reporting will require CachedPlans to have a copy of that string anyway, so it's not clear that it's worth moving mountains to avoid it now. This also changes the APIs for context statistics routines so that the context-specific routines no longer assume that output goes straight to stderr, nor do they know all details of the output format. This is useful immediately to reduce code duplication, and it also allows for external code to do something with stats output that's different from printing to stderr. The reason for pushing this now rather than waiting for v12 is that it rethinks some of the API changes made by commit 9fa6f00b1. Seems better for extension authors to endure just one round of API changes not two. Discussion: https://postgr.es/m/CAB=Je-FdtmFZ9y9REHD7VsSrnCkiBhsA4mdsLKSPauwXtQBeNA@mail.gmail.com
2018-03-27 22:46:47 +02:00
name);
Rethink MemoryContext creation to improve performance. This patch makes a number of interrelated changes to reduce the overhead involved in creating/deleting memory contexts. The key ideas are: * Include the AllocSetContext header of an aset.c context in its first malloc request, rather than allocating it separately in TopMemoryContext. This means that we now always create an initial or "keeper" block in an aset, even if it never receives any allocation requests. * Create freelists in which we can save and recycle recently-destroyed asets (this idea is due to Robert Haas). * In the common case where the name of a context is a constant string, just store a pointer to it in the context header, rather than copying the string. The first change eliminates a palloc/pfree cycle per context, and also avoids bloat in TopMemoryContext, at the price that creating a context now involves a malloc/free cycle even if the context never receives any allocations. That would be a loser for some common usage patterns, but recycling short-lived contexts via the freelist eliminates that pain. Avoiding copying constant strings not only saves strlen() and strcpy() overhead, but is an essential part of the freelist optimization because it makes the context header size constant. Currently we make no attempt to use the freelist for contexts with non-constant names. (Perhaps someday we'll need to think harder about that, but in current usage, most contexts with custom names are long-lived anyway.) The freelist management in this initial commit is pretty simplistic, and we might want to refine it later --- but in common workloads that will never matter because the freelists will never get full anyway. To create a context with a non-constant name, one is now required to call AllocSetContextCreateExtended and specify the MEMCONTEXT_COPY_NAME option. AllocSetContextCreate becomes a wrapper macro, and it includes a test that will complain about non-string-literal context name parameters on gcc and similar compilers. An unfortunate side effect of making AllocSetContextCreate a macro is that one is now *required* to use the size parameter abstraction macros (ALLOCSET_DEFAULT_SIZES and friends) with it; the pre-9.6 habit of writing out individual size parameters no longer works unless you switch to AllocSetContextCreateExtended. Internally to the memory-context-related modules, the context creation APIs are simplified, removing the rather baroque original design whereby a context-type module called mcxt.c which then called back into the context-type module. That saved a bit of code duplication, but not much, and it prevented context-type modules from exercising control over the allocation of context headers. In passing, I converted the test-and-elog validation of aset size parameters into Asserts to save a few more cycles. The original thought was that callers might compute size parameters on the fly, but in practice nobody does that, so it's useless to expend cycles on checking those numbers in production builds. Also, mark the memory context method-pointer structs "const", just for cleanliness. Discussion: https://postgr.es/m/2264.1512870796@sss.pgh.pa.us
2017-12-13 19:55:12 +01:00
return (MemoryContext) slab;
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
}
/*
* SlabReset
* Frees all memory which is allocated in the given set.
*
* The code simply frees all the blocks in the context - we don't keep any
* keeper blocks or anything like that.
*/
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
void
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
SlabReset(MemoryContext context)
{
int i;
SlabContext *slab = castNode(SlabContext, context);
Assert(slab);
#ifdef MEMORY_CONTEXT_CHECKING
/* Check for corruption and leaks before freeing */
SlabCheck(context);
#endif
/* walk over freelists and free the blocks */
for (i = 0; i <= slab->chunksPerBlock; i++)
{
dlist_mutable_iter miter;
dlist_foreach_modify(miter, &slab->freelist[i])
{
SlabBlock *block = dlist_container(SlabBlock, node, miter.cur);
dlist_delete(miter.cur);
#ifdef CLOBBER_FREED_MEMORY
wipe_mem(block, slab->blockSize);
#endif
free(block);
slab->nblocks--;
Revert "Specialize MemoryContextMemAllocated()." This reverts commit e00912e11a9ec2d29274ed8a6465e81385906dc2.
2020-03-19 19:20:39 +01:00
context->mem_allocated -= slab->blockSize;
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
}
}
slab->minFreeChunks = 0;
Assert(slab->nblocks == 0);
Revert "Specialize MemoryContextMemAllocated()." This reverts commit e00912e11a9ec2d29274ed8a6465e81385906dc2.
2020-03-19 19:20:39 +01:00
Assert(context->mem_allocated == 0);
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
}
/*
* SlabDelete
Rethink MemoryContext creation to improve performance. This patch makes a number of interrelated changes to reduce the overhead involved in creating/deleting memory contexts. The key ideas are: * Include the AllocSetContext header of an aset.c context in its first malloc request, rather than allocating it separately in TopMemoryContext. This means that we now always create an initial or "keeper" block in an aset, even if it never receives any allocation requests. * Create freelists in which we can save and recycle recently-destroyed asets (this idea is due to Robert Haas). * In the common case where the name of a context is a constant string, just store a pointer to it in the context header, rather than copying the string. The first change eliminates a palloc/pfree cycle per context, and also avoids bloat in TopMemoryContext, at the price that creating a context now involves a malloc/free cycle even if the context never receives any allocations. That would be a loser for some common usage patterns, but recycling short-lived contexts via the freelist eliminates that pain. Avoiding copying constant strings not only saves strlen() and strcpy() overhead, but is an essential part of the freelist optimization because it makes the context header size constant. Currently we make no attempt to use the freelist for contexts with non-constant names. (Perhaps someday we'll need to think harder about that, but in current usage, most contexts with custom names are long-lived anyway.) The freelist management in this initial commit is pretty simplistic, and we might want to refine it later --- but in common workloads that will never matter because the freelists will never get full anyway. To create a context with a non-constant name, one is now required to call AllocSetContextCreateExtended and specify the MEMCONTEXT_COPY_NAME option. AllocSetContextCreate becomes a wrapper macro, and it includes a test that will complain about non-string-literal context name parameters on gcc and similar compilers. An unfortunate side effect of making AllocSetContextCreate a macro is that one is now *required* to use the size parameter abstraction macros (ALLOCSET_DEFAULT_SIZES and friends) with it; the pre-9.6 habit of writing out individual size parameters no longer works unless you switch to AllocSetContextCreateExtended. Internally to the memory-context-related modules, the context creation APIs are simplified, removing the rather baroque original design whereby a context-type module called mcxt.c which then called back into the context-type module. That saved a bit of code duplication, but not much, and it prevented context-type modules from exercising control over the allocation of context headers. In passing, I converted the test-and-elog validation of aset size parameters into Asserts to save a few more cycles. The original thought was that callers might compute size parameters on the fly, but in practice nobody does that, so it's useless to expend cycles on checking those numbers in production builds. Also, mark the memory context method-pointer structs "const", just for cleanliness. Discussion: https://postgr.es/m/2264.1512870796@sss.pgh.pa.us
2017-12-13 19:55:12 +01:00
* Free all memory which is allocated in the given context.
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
*/
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
void
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
SlabDelete(MemoryContext context)
{
Rethink MemoryContext creation to improve performance. This patch makes a number of interrelated changes to reduce the overhead involved in creating/deleting memory contexts. The key ideas are: * Include the AllocSetContext header of an aset.c context in its first malloc request, rather than allocating it separately in TopMemoryContext. This means that we now always create an initial or "keeper" block in an aset, even if it never receives any allocation requests. * Create freelists in which we can save and recycle recently-destroyed asets (this idea is due to Robert Haas). * In the common case where the name of a context is a constant string, just store a pointer to it in the context header, rather than copying the string. The first change eliminates a palloc/pfree cycle per context, and also avoids bloat in TopMemoryContext, at the price that creating a context now involves a malloc/free cycle even if the context never receives any allocations. That would be a loser for some common usage patterns, but recycling short-lived contexts via the freelist eliminates that pain. Avoiding copying constant strings not only saves strlen() and strcpy() overhead, but is an essential part of the freelist optimization because it makes the context header size constant. Currently we make no attempt to use the freelist for contexts with non-constant names. (Perhaps someday we'll need to think harder about that, but in current usage, most contexts with custom names are long-lived anyway.) The freelist management in this initial commit is pretty simplistic, and we might want to refine it later --- but in common workloads that will never matter because the freelists will never get full anyway. To create a context with a non-constant name, one is now required to call AllocSetContextCreateExtended and specify the MEMCONTEXT_COPY_NAME option. AllocSetContextCreate becomes a wrapper macro, and it includes a test that will complain about non-string-literal context name parameters on gcc and similar compilers. An unfortunate side effect of making AllocSetContextCreate a macro is that one is now *required* to use the size parameter abstraction macros (ALLOCSET_DEFAULT_SIZES and friends) with it; the pre-9.6 habit of writing out individual size parameters no longer works unless you switch to AllocSetContextCreateExtended. Internally to the memory-context-related modules, the context creation APIs are simplified, removing the rather baroque original design whereby a context-type module called mcxt.c which then called back into the context-type module. That saved a bit of code duplication, but not much, and it prevented context-type modules from exercising control over the allocation of context headers. In passing, I converted the test-and-elog validation of aset size parameters into Asserts to save a few more cycles. The original thought was that callers might compute size parameters on the fly, but in practice nobody does that, so it's useless to expend cycles on checking those numbers in production builds. Also, mark the memory context method-pointer structs "const", just for cleanliness. Discussion: https://postgr.es/m/2264.1512870796@sss.pgh.pa.us
2017-12-13 19:55:12 +01:00
/* Reset to release all the SlabBlocks */
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
SlabReset(context);
Rethink MemoryContext creation to improve performance. This patch makes a number of interrelated changes to reduce the overhead involved in creating/deleting memory contexts. The key ideas are: * Include the AllocSetContext header of an aset.c context in its first malloc request, rather than allocating it separately in TopMemoryContext. This means that we now always create an initial or "keeper" block in an aset, even if it never receives any allocation requests. * Create freelists in which we can save and recycle recently-destroyed asets (this idea is due to Robert Haas). * In the common case where the name of a context is a constant string, just store a pointer to it in the context header, rather than copying the string. The first change eliminates a palloc/pfree cycle per context, and also avoids bloat in TopMemoryContext, at the price that creating a context now involves a malloc/free cycle even if the context never receives any allocations. That would be a loser for some common usage patterns, but recycling short-lived contexts via the freelist eliminates that pain. Avoiding copying constant strings not only saves strlen() and strcpy() overhead, but is an essential part of the freelist optimization because it makes the context header size constant. Currently we make no attempt to use the freelist for contexts with non-constant names. (Perhaps someday we'll need to think harder about that, but in current usage, most contexts with custom names are long-lived anyway.) The freelist management in this initial commit is pretty simplistic, and we might want to refine it later --- but in common workloads that will never matter because the freelists will never get full anyway. To create a context with a non-constant name, one is now required to call AllocSetContextCreateExtended and specify the MEMCONTEXT_COPY_NAME option. AllocSetContextCreate becomes a wrapper macro, and it includes a test that will complain about non-string-literal context name parameters on gcc and similar compilers. An unfortunate side effect of making AllocSetContextCreate a macro is that one is now *required* to use the size parameter abstraction macros (ALLOCSET_DEFAULT_SIZES and friends) with it; the pre-9.6 habit of writing out individual size parameters no longer works unless you switch to AllocSetContextCreateExtended. Internally to the memory-context-related modules, the context creation APIs are simplified, removing the rather baroque original design whereby a context-type module called mcxt.c which then called back into the context-type module. That saved a bit of code duplication, but not much, and it prevented context-type modules from exercising control over the allocation of context headers. In passing, I converted the test-and-elog validation of aset size parameters into Asserts to save a few more cycles. The original thought was that callers might compute size parameters on the fly, but in practice nobody does that, so it's useless to expend cycles on checking those numbers in production builds. Also, mark the memory context method-pointer structs "const", just for cleanliness. Discussion: https://postgr.es/m/2264.1512870796@sss.pgh.pa.us
2017-12-13 19:55:12 +01:00
/* And free the context header */
free(context);
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
}
/*
* SlabAlloc
* Returns pointer to allocated memory of given size or NULL if
* request could not be completed; memory is added to the slab.
*/
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
void *
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
SlabAlloc(MemoryContext context, Size size)
{
SlabContext *slab = castNode(SlabContext, context);
SlabBlock *block;
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
MemoryChunk *chunk;
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
int idx;
Assert(slab);
Assert((slab->minFreeChunks >= 0) &&
(slab->minFreeChunks < slab->chunksPerBlock));
/* make sure we only allow correct request size */
if (size != slab->chunkSize)
Reduce size of common allocation header. The new slab allocator needs different per-allocation information than the classical aset.c. The definition in 58b25e981 wasn't sufficiently careful on 32 platforms with 8 byte alignment, leading to buildfarm failures. That's not entirely easy to fix by just adjusting the definition. As slab.c doesn't actually need the size part(s) of the common header, all chunks are equally sized after all, it seems better to instead reduce the header to the part needed by all allocators, namely which context an allocation belongs to. That has the advantage of reducing the overhead of slab allocations, and also allows for more flexibility in future allocators. To avoid spreading the logic about accessing a chunk's context around, centralize it in GetMemoryChunkContext(), which allows to delete a good number of lines. A followup commit will revise the mmgr/README portion about StandardChunkHeader, and more. Author: Andres Freund Discussion: https://postgr.es/m/20170228074420.aazv4iw6k562mnxg@alap3.anarazel.de
2017-02-28 08:32:22 +01:00
elog(ERROR, "unexpected alloc chunk size %zu (expected %zu)",
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
size, slab->chunkSize);
/*
* If there are no free chunks in any existing block, create a new block
* and put it to the last freelist bucket.
*
* slab->minFreeChunks == 0 means there are no blocks with free chunks,
* thanks to how minFreeChunks is updated at the end of SlabAlloc().
*/
if (slab->minFreeChunks == 0)
{
block = (SlabBlock *) malloc(slab->blockSize);
if (block == NULL)
return NULL;
block->nfree = slab->chunksPerBlock;
block->firstFreeChunk = 0;
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
block->slab = slab;
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
/*
* Put all the chunks on a freelist. Walk the chunks and point each
* one to the next one.
*/
for (idx = 0; idx < slab->chunksPerBlock; idx++)
{
chunk = SlabBlockGetChunk(slab, block, idx);
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
*(int32 *) MemoryChunkGetPointer(chunk) = (idx + 1);
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
}
/*
* And add it to the last freelist with all chunks empty.
*
* We know there are no blocks in the freelist, otherwise we wouldn't
* need a new block.
*/
Assert(dlist_is_empty(&slab->freelist[slab->chunksPerBlock]));
dlist_push_head(&slab->freelist[slab->chunksPerBlock], &block->node);
slab->minFreeChunks = slab->chunksPerBlock;
slab->nblocks += 1;
Revert "Specialize MemoryContextMemAllocated()." This reverts commit e00912e11a9ec2d29274ed8a6465e81385906dc2.
2020-03-19 19:20:39 +01:00
context->mem_allocated += slab->blockSize;
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
}
/* grab the block from the freelist (even the new block is there) */
block = dlist_head_element(SlabBlock, node,
&slab->freelist[slab->minFreeChunks]);
/* make sure we actually got a valid block, with matching nfree */
Assert(block != NULL);
Assert(slab->minFreeChunks == block->nfree);
Assert(block->nfree > 0);
/* we know index of the first free chunk in the block */
idx = block->firstFreeChunk;
/* make sure the chunk index is valid, and that it's marked as empty */
Assert((idx >= 0) && (idx < slab->chunksPerBlock));
/* compute the chunk location block start (after the block header) */
chunk = SlabBlockGetChunk(slab, block, idx);
/*
* Update the block nfree count, and also the minFreeChunks as we've
* decreased nfree for a block with the minimum number of free chunks
* (because that's how we chose the block).
*/
block->nfree--;
slab->minFreeChunks = block->nfree;
/*
* Remove the chunk from the freelist head. The index of the next free
* chunk is stored in the chunk itself.
*/
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
VALGRIND_MAKE_MEM_DEFINED(MemoryChunkGetPointer(chunk), sizeof(int32));
block->firstFreeChunk = *(int32 *) MemoryChunkGetPointer(chunk);
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
Assert(block->firstFreeChunk >= 0);
Assert(block->firstFreeChunk <= slab->chunksPerBlock);
Assert((block->nfree != 0 &&
block->firstFreeChunk < slab->chunksPerBlock) ||
(block->nfree == 0 &&
block->firstFreeChunk == slab->chunksPerBlock));
/* move the whole block to the right place in the freelist */
dlist_delete(&block->node);
dlist_push_head(&slab->freelist[block->nfree], &block->node);
/*
* And finally update minFreeChunks, i.e. the index to the block with the
* lowest number of free chunks. We only need to do that when the block
* got full (otherwise we know the current block is the right one). We'll
* simply walk the freelist until we find a non-empty entry.
*/
if (slab->minFreeChunks == 0)
{
for (idx = 1; idx <= slab->chunksPerBlock; idx++)
{
if (dlist_is_empty(&slab->freelist[idx]))
continue;
/* found a non-empty freelist */
slab->minFreeChunks = idx;
break;
}
}
if (slab->minFreeChunks == slab->chunksPerBlock)
slab->minFreeChunks = 0;
/* Prepare to initialize the chunk header. */
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
VALGRIND_MAKE_MEM_UNDEFINED(chunk, Slab_CHUNKHDRSZ);
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
MemoryChunkSetHdrMask(chunk, block, MAXALIGN(slab->chunkSize),
MCTX_SLAB_ID);
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
#ifdef MEMORY_CONTEXT_CHECKING
/* slab mark to catch clobber of "unused" space */
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
if (slab->chunkSize < (slab->fullChunkSize - Slab_CHUNKHDRSZ))
Reduce size of common allocation header. The new slab allocator needs different per-allocation information than the classical aset.c. The definition in 58b25e981 wasn't sufficiently careful on 32 platforms with 8 byte alignment, leading to buildfarm failures. That's not entirely easy to fix by just adjusting the definition. As slab.c doesn't actually need the size part(s) of the common header, all chunks are equally sized after all, it seems better to instead reduce the header to the part needed by all allocators, namely which context an allocation belongs to. That has the advantage of reducing the overhead of slab allocations, and also allows for more flexibility in future allocators. To avoid spreading the logic about accessing a chunk's context around, centralize it in GetMemoryChunkContext(), which allows to delete a good number of lines. A followup commit will revise the mmgr/README portion about StandardChunkHeader, and more. Author: Andres Freund Discussion: https://postgr.es/m/20170228074420.aazv4iw6k562mnxg@alap3.anarazel.de
2017-02-28 08:32:22 +01:00
{
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
set_sentinel(MemoryChunkGetPointer(chunk), size);
Reduce size of common allocation header. The new slab allocator needs different per-allocation information than the classical aset.c. The definition in 58b25e981 wasn't sufficiently careful on 32 platforms with 8 byte alignment, leading to buildfarm failures. That's not entirely easy to fix by just adjusting the definition. As slab.c doesn't actually need the size part(s) of the common header, all chunks are equally sized after all, it seems better to instead reduce the header to the part needed by all allocators, namely which context an allocation belongs to. That has the advantage of reducing the overhead of slab allocations, and also allows for more flexibility in future allocators. To avoid spreading the logic about accessing a chunk's context around, centralize it in GetMemoryChunkContext(), which allows to delete a good number of lines. A followup commit will revise the mmgr/README portion about StandardChunkHeader, and more. Author: Andres Freund Discussion: https://postgr.es/m/20170228074420.aazv4iw6k562mnxg@alap3.anarazel.de
2017-02-28 08:32:22 +01:00
VALGRIND_MAKE_MEM_NOACCESS(((char *) chunk) +
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
Slab_CHUNKHDRSZ + slab->chunkSize,
Reduce size of common allocation header. The new slab allocator needs different per-allocation information than the classical aset.c. The definition in 58b25e981 wasn't sufficiently careful on 32 platforms with 8 byte alignment, leading to buildfarm failures. That's not entirely easy to fix by just adjusting the definition. As slab.c doesn't actually need the size part(s) of the common header, all chunks are equally sized after all, it seems better to instead reduce the header to the part needed by all allocators, namely which context an allocation belongs to. That has the advantage of reducing the overhead of slab allocations, and also allows for more flexibility in future allocators. To avoid spreading the logic about accessing a chunk's context around, centralize it in GetMemoryChunkContext(), which allows to delete a good number of lines. A followup commit will revise the mmgr/README portion about StandardChunkHeader, and more. Author: Andres Freund Discussion: https://postgr.es/m/20170228074420.aazv4iw6k562mnxg@alap3.anarazel.de
2017-02-28 08:32:22 +01:00
slab->fullChunkSize -
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
(slab->chunkSize + Slab_CHUNKHDRSZ));
Reduce size of common allocation header. The new slab allocator needs different per-allocation information than the classical aset.c. The definition in 58b25e981 wasn't sufficiently careful on 32 platforms with 8 byte alignment, leading to buildfarm failures. That's not entirely easy to fix by just adjusting the definition. As slab.c doesn't actually need the size part(s) of the common header, all chunks are equally sized after all, it seems better to instead reduce the header to the part needed by all allocators, namely which context an allocation belongs to. That has the advantage of reducing the overhead of slab allocations, and also allows for more flexibility in future allocators. To avoid spreading the logic about accessing a chunk's context around, centralize it in GetMemoryChunkContext(), which allows to delete a good number of lines. A followup commit will revise the mmgr/README portion about StandardChunkHeader, and more. Author: Andres Freund Discussion: https://postgr.es/m/20170228074420.aazv4iw6k562mnxg@alap3.anarazel.de
2017-02-28 08:32:22 +01:00
}
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
#endif
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
#ifdef RANDOMIZE_ALLOCATED_MEMORY
/* fill the allocated space with junk */
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
randomize_mem((char *) MemoryChunkGetPointer(chunk), size);
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
#endif
Revert "Specialize MemoryContextMemAllocated()." This reverts commit e00912e11a9ec2d29274ed8a6465e81385906dc2.
2020-03-19 19:20:39 +01:00
Assert(slab->nblocks * slab->blockSize == context->mem_allocated);
Add transparent block-level memory accounting Adds accounting of memory allocated in a memory context. Compared to various ad hoc solutions, the main advantage is that the accounting is transparent and does not require direct control over allocations (this matters for use cases where the allocations happen in user code, like for example aggregate states allocated in a transition functions). To reduce overhead, the accounting happens at the block level (not for individual chunks) and only the context immediately owning the block is updated. When inquiring about amount of memory allocated in a context, we have to recursively walk all children contexts. This "lazy" accounting works well for cases with relatively small number of contexts in the relevant subtree and/or with infrequent inquiries. Author: Jeff Davis Reivewed-by: Tomas Vondra, Melanie Plageman, Soumyadeep Chakraborty Discussion: https://www.postgresql.org/message-id/flat/027a129b8525601c6a680d27ce3a7172dab61aab.camel@j-davis.com
2019-10-01 03:13:39 +02:00
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
return MemoryChunkGetPointer(chunk);
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
}
/*
* SlabFree
* Frees allocated memory; memory is removed from the slab.
*/
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
void
SlabFree(void *pointer)
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
{
int idx;
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
MemoryChunk *chunk = PointerGetMemoryChunk(pointer);
SlabBlock *block = MemoryChunkGetBlock(chunk);
SlabContext *slab = block->slab;
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
#ifdef MEMORY_CONTEXT_CHECKING
/* Test for someone scribbling on unused space in chunk */
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
if (slab->chunkSize < (slab->fullChunkSize - Slab_CHUNKHDRSZ))
Reduce size of common allocation header. The new slab allocator needs different per-allocation information than the classical aset.c. The definition in 58b25e981 wasn't sufficiently careful on 32 platforms with 8 byte alignment, leading to buildfarm failures. That's not entirely easy to fix by just adjusting the definition. As slab.c doesn't actually need the size part(s) of the common header, all chunks are equally sized after all, it seems better to instead reduce the header to the part needed by all allocators, namely which context an allocation belongs to. That has the advantage of reducing the overhead of slab allocations, and also allows for more flexibility in future allocators. To avoid spreading the logic about accessing a chunk's context around, centralize it in GetMemoryChunkContext(), which allows to delete a good number of lines. A followup commit will revise the mmgr/README portion about StandardChunkHeader, and more. Author: Andres Freund Discussion: https://postgr.es/m/20170228074420.aazv4iw6k562mnxg@alap3.anarazel.de
2017-02-28 08:32:22 +01:00
if (!sentinel_ok(pointer, slab->chunkSize))
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
elog(WARNING, "detected write past chunk end in %s %p",
slab->header.name, chunk);
#endif
/* compute index of the chunk with respect to block start */
idx = SlabChunkIndex(slab, block, chunk);
/* add chunk to freelist, and update block nfree count */
*(int32 *) pointer = block->firstFreeChunk;
block->firstFreeChunk = idx;
block->nfree++;
Assert(block->nfree > 0);
Assert(block->nfree <= slab->chunksPerBlock);
#ifdef CLOBBER_FREED_MEMORY
/* XXX don't wipe the int32 index, used for block-level freelist */
wipe_mem((char *) pointer + sizeof(int32),
Reduce size of common allocation header. The new slab allocator needs different per-allocation information than the classical aset.c. The definition in 58b25e981 wasn't sufficiently careful on 32 platforms with 8 byte alignment, leading to buildfarm failures. That's not entirely easy to fix by just adjusting the definition. As slab.c doesn't actually need the size part(s) of the common header, all chunks are equally sized after all, it seems better to instead reduce the header to the part needed by all allocators, namely which context an allocation belongs to. That has the advantage of reducing the overhead of slab allocations, and also allows for more flexibility in future allocators. To avoid spreading the logic about accessing a chunk's context around, centralize it in GetMemoryChunkContext(), which allows to delete a good number of lines. A followup commit will revise the mmgr/README portion about StandardChunkHeader, and more. Author: Andres Freund Discussion: https://postgr.es/m/20170228074420.aazv4iw6k562mnxg@alap3.anarazel.de
2017-02-28 08:32:22 +01:00
slab->chunkSize - sizeof(int32));
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
#endif
/* remove the block from a freelist */
dlist_delete(&block->node);
/*
* See if we need to update the minFreeChunks field for the slab - we only
* need to do that if there the block had that number of free chunks
* before we freed one. In that case, we check if there still are blocks
* in the original freelist and we either keep the current value (if there
* still are blocks) or increment it by one (the new block is still the
* one with minimum free chunks).
*
* The one exception is when the block will get completely free - in that
* case we will free it, se we can't use it for minFreeChunks. It however
* means there are no more blocks with free chunks.
*/
if (slab->minFreeChunks == (block->nfree - 1))
{
/* Have we removed the last chunk from the freelist? */
if (dlist_is_empty(&slab->freelist[slab->minFreeChunks]))
{
/* but if we made the block entirely free, we'll free it */
if (block->nfree == slab->chunksPerBlock)
slab->minFreeChunks = 0;
else
slab->minFreeChunks++;
}
}
/* If the block is now completely empty, free it. */
if (block->nfree == slab->chunksPerBlock)
{
free(block);
slab->nblocks--;
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
slab->header.mem_allocated -= slab->blockSize;
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
}
else
dlist_push_head(&slab->freelist[block->nfree], &block->node);
Assert(slab->nblocks >= 0);
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
Assert(slab->nblocks * slab->blockSize == slab->header.mem_allocated);
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
}
/*
* SlabRealloc
Suppress compiler warning in slab.c. Compilers that don't realize that elog(ERROR) doesn't return complained that SlabRealloc() failed to return a value. While at it, fix the rather muddled header comment for the function. Per buildfarm.
2017-03-08 22:19:37 +01:00
* Change the allocated size of a chunk.
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
*
Suppress compiler warning in slab.c. Compilers that don't realize that elog(ERROR) doesn't return complained that SlabRealloc() failed to return a value. While at it, fix the rather muddled header comment for the function. Per buildfarm.
2017-03-08 22:19:37 +01:00
* As Slab is designed for allocating equally-sized chunks of memory, it can't
* do an actual chunk size change. We try to be gentle and allow calls with
* exactly the same size, as in that case we can simply return the same
* chunk. When the size differs, we throw an error.
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
*
Suppress compiler warning in slab.c. Compilers that don't realize that elog(ERROR) doesn't return complained that SlabRealloc() failed to return a value. While at it, fix the rather muddled header comment for the function. Per buildfarm.
2017-03-08 22:19:37 +01:00
* We could also allow requests with size < chunkSize. That however seems
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
* rather pointless - Slab is meant for chunks of constant size, and moreover
* realloc is usually used to enlarge the chunk.
*/
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
void *
SlabRealloc(void *pointer, Size size)
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
{
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
MemoryChunk *chunk = PointerGetMemoryChunk(pointer);
SlabBlock *block = MemoryChunkGetBlock(chunk);
SlabContext *slab = block->slab;
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
Assert(slab);
/* can't do actual realloc with slab, but let's try to be gentle */
if (size == slab->chunkSize)
return pointer;
elog(ERROR, "slab allocator does not support realloc()");
Suppress compiler warning in slab.c. Compilers that don't realize that elog(ERROR) doesn't return complained that SlabRealloc() failed to return a value. While at it, fix the rather muddled header comment for the function. Per buildfarm.
2017-03-08 22:19:37 +01:00
return NULL; /* keep compiler quiet */
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
}
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
/*
* SlabGetChunkContext
* Return the MemoryContext that 'pointer' belongs to.
*/
MemoryContext
SlabGetChunkContext(void *pointer)
{
MemoryChunk *chunk = PointerGetMemoryChunk(pointer);
SlabBlock *block = MemoryChunkGetBlock(chunk);
SlabContext *slab = block->slab;
Assert(slab != NULL);
return &slab->header;
}
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
/*
* SlabGetChunkSpace
* Given a currently-allocated chunk, determine the total space
* it occupies (including all memory-allocation overhead).
*/
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
Size
SlabGetChunkSpace(void *pointer)
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
{
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
MemoryChunk *chunk = PointerGetMemoryChunk(pointer);
SlabBlock *block = MemoryChunkGetBlock(chunk);
SlabContext *slab = block->slab;
Reduce size of common allocation header. The new slab allocator needs different per-allocation information than the classical aset.c. The definition in 58b25e981 wasn't sufficiently careful on 32 platforms with 8 byte alignment, leading to buildfarm failures. That's not entirely easy to fix by just adjusting the definition. As slab.c doesn't actually need the size part(s) of the common header, all chunks are equally sized after all, it seems better to instead reduce the header to the part needed by all allocators, namely which context an allocation belongs to. That has the advantage of reducing the overhead of slab allocations, and also allows for more flexibility in future allocators. To avoid spreading the logic about accessing a chunk's context around, centralize it in GetMemoryChunkContext(), which allows to delete a good number of lines. A followup commit will revise the mmgr/README portion about StandardChunkHeader, and more. Author: Andres Freund Discussion: https://postgr.es/m/20170228074420.aazv4iw6k562mnxg@alap3.anarazel.de
2017-02-28 08:32:22 +01:00
Assert(slab);
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
Reduce size of common allocation header. The new slab allocator needs different per-allocation information than the classical aset.c. The definition in 58b25e981 wasn't sufficiently careful on 32 platforms with 8 byte alignment, leading to buildfarm failures. That's not entirely easy to fix by just adjusting the definition. As slab.c doesn't actually need the size part(s) of the common header, all chunks are equally sized after all, it seems better to instead reduce the header to the part needed by all allocators, namely which context an allocation belongs to. That has the advantage of reducing the overhead of slab allocations, and also allows for more flexibility in future allocators. To avoid spreading the logic about accessing a chunk's context around, centralize it in GetMemoryChunkContext(), which allows to delete a good number of lines. A followup commit will revise the mmgr/README portion about StandardChunkHeader, and more. Author: Andres Freund Discussion: https://postgr.es/m/20170228074420.aazv4iw6k562mnxg@alap3.anarazel.de
2017-02-28 08:32:22 +01:00
return slab->fullChunkSize;
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
}
/*
* SlabIsEmpty
* Is an Slab empty of any allocated space?
*/
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
bool
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
SlabIsEmpty(MemoryContext context)
{
SlabContext *slab = castNode(SlabContext, context);
Assert(slab);
return (slab->nblocks == 0);
}
/*
* SlabStats
Rethink MemoryContext creation to improve performance. This patch makes a number of interrelated changes to reduce the overhead involved in creating/deleting memory contexts. The key ideas are: * Include the AllocSetContext header of an aset.c context in its first malloc request, rather than allocating it separately in TopMemoryContext. This means that we now always create an initial or "keeper" block in an aset, even if it never receives any allocation requests. * Create freelists in which we can save and recycle recently-destroyed asets (this idea is due to Robert Haas). * In the common case where the name of a context is a constant string, just store a pointer to it in the context header, rather than copying the string. The first change eliminates a palloc/pfree cycle per context, and also avoids bloat in TopMemoryContext, at the price that creating a context now involves a malloc/free cycle even if the context never receives any allocations. That would be a loser for some common usage patterns, but recycling short-lived contexts via the freelist eliminates that pain. Avoiding copying constant strings not only saves strlen() and strcpy() overhead, but is an essential part of the freelist optimization because it makes the context header size constant. Currently we make no attempt to use the freelist for contexts with non-constant names. (Perhaps someday we'll need to think harder about that, but in current usage, most contexts with custom names are long-lived anyway.) The freelist management in this initial commit is pretty simplistic, and we might want to refine it later --- but in common workloads that will never matter because the freelists will never get full anyway. To create a context with a non-constant name, one is now required to call AllocSetContextCreateExtended and specify the MEMCONTEXT_COPY_NAME option. AllocSetContextCreate becomes a wrapper macro, and it includes a test that will complain about non-string-literal context name parameters on gcc and similar compilers. An unfortunate side effect of making AllocSetContextCreate a macro is that one is now *required* to use the size parameter abstraction macros (ALLOCSET_DEFAULT_SIZES and friends) with it; the pre-9.6 habit of writing out individual size parameters no longer works unless you switch to AllocSetContextCreateExtended. Internally to the memory-context-related modules, the context creation APIs are simplified, removing the rather baroque original design whereby a context-type module called mcxt.c which then called back into the context-type module. That saved a bit of code duplication, but not much, and it prevented context-type modules from exercising control over the allocation of context headers. In passing, I converted the test-and-elog validation of aset size parameters into Asserts to save a few more cycles. The original thought was that callers might compute size parameters on the fly, but in practice nobody does that, so it's useless to expend cycles on checking those numbers in production builds. Also, mark the memory context method-pointer structs "const", just for cleanliness. Discussion: https://postgr.es/m/2264.1512870796@sss.pgh.pa.us
2017-12-13 19:55:12 +01:00
* Compute stats about memory consumption of a Slab context.
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
*
Allow memory contexts to have both fixed and variable ident strings. Originally, we treated memory context names as potentially variable in all cases, and therefore always copied them into the context header. Commit 9fa6f00b1 rethought this a little bit and invented a distinction between fixed and variable names, skipping the copy step for the former. But we can make things both simpler and more useful by instead allowing there to be two parts to a context's identification, a fixed "name" and an optional, variable "ident". The name supplied in the context create call is now required to be a compile-time-constant string in all cases, as it is never copied but just pointed to. The "ident" string, if wanted, is supplied later. This is needed because typically we want the ident to be stored inside the context so that it's cleaned up automatically on context deletion; that means it has to be copied into the context before we can set the pointer. The cost of this approach is basically just an additional pointer field in struct MemoryContextData, which isn't much overhead, and is bought back entirely in the AllocSet case by not needing a headerSize field anymore, since we no longer have to cope with variable header length. In addition, we can simplify the internal interfaces for memory context creation still further, saving a few cycles there. And it's no longer true that a custom identifier disqualifies a context from participating in aset.c's freelist scheme, so possibly there's some win on that end. All the places that were using non-compile-time-constant context names are adjusted to put the variable info into the "ident" instead. This allows more effective identification of those contexts in many cases; for example, subsidary contexts of relcache entries are now identified by both type (e.g. "index info") and relname, where before you got only one or the other. Contexts associated with PL function cache entries are now identified more fully and uniformly, too. I also arranged for plancache contexts to use the query source string as their identifier. This is basically free for CachedPlanSources, as they contained a copy of that string already. We pay an extra pstrdup to do it for CachedPlans. That could perhaps be avoided, but it would make things more fragile (since the CachedPlanSource is sometimes destroyed first). I suspect future improvements in error reporting will require CachedPlans to have a copy of that string anyway, so it's not clear that it's worth moving mountains to avoid it now. This also changes the APIs for context statistics routines so that the context-specific routines no longer assume that output goes straight to stderr, nor do they know all details of the output format. This is useful immediately to reduce code duplication, and it also allows for external code to do something with stats output that's different from printing to stderr. The reason for pushing this now rather than waiting for v12 is that it rethinks some of the API changes made by commit 9fa6f00b1. Seems better for extension authors to endure just one round of API changes not two. Discussion: https://postgr.es/m/CAB=Je-FdtmFZ9y9REHD7VsSrnCkiBhsA4mdsLKSPauwXtQBeNA@mail.gmail.com
2018-03-27 22:46:47 +02:00
* printfunc: if not NULL, pass a human-readable stats string to this.
* passthru: pass this pointer through to printfunc.
* totals: if not NULL, add stats about this context into *totals.
Add function to log the memory contexts of specified backend process. Commit 3e98c0bafb added pg_backend_memory_contexts view to display the memory contexts of the backend process. However its target process is limited to the backend that is accessing to the view. So this is not so convenient when investigating the local memory bloat of other backend process. To improve this situation, this commit adds pg_log_backend_memory_contexts() function that requests to log the memory contexts of the specified backend process. This information can be also collected by calling MemoryContextStats(TopMemoryContext) via a debugger. But this technique cannot be used in some environments because no debugger is available there. So, pg_log_backend_memory_contexts() allows us to see the memory contexts of specified backend more easily. Only superusers are allowed to request to log the memory contexts because allowing any users to issue this request at an unbounded rate would cause lots of log messages and which can lead to denial of service. On receipt of the request, at the next CHECK_FOR_INTERRUPTS(), the target backend logs its memory contexts at LOG_SERVER_ONLY level, so that these memory contexts will appear in the server log but not be sent to the client. It logs one message per memory context. Because if it buffers all memory contexts into StringInfo to log them as one message, which may require the buffer to be enlarged very much and lead to OOM error since there can be a large number of memory contexts in a backend. When a backend process is consuming huge memory, logging all its memory contexts might overrun available disk space. To prevent this, now this patch limits the number of child contexts to log per parent to 100. As with MemoryContextStats(), it supposes that practical cases where the log gets long will typically be huge numbers of siblings under the same parent context; while the additional debugging value from seeing details about individual siblings beyond 100 will not be large. There was another proposed patch to add the function to return the memory contexts of specified backend as the result sets, instead of logging them, in the discussion. However that patch is not included in this commit because it had several issues to address. Thanks to Tatsuhito Kasahara, Andres Freund, Tom Lane, Tomas Vondra, Michael Paquier, Kyotaro Horiguchi and Zhihong Yu for the discussion. Bump catalog version. Author: Atsushi Torikoshi Reviewed-by: Kyotaro Horiguchi, Zhihong Yu, Fujii Masao Discussion: https://postgr.es/m/0271f440ac77f2a4180e0e56ebd944d1@oss.nttdata.com
2021-04-06 06:44:15 +02:00
* print_to_stderr: print stats to stderr if true, elog otherwise.
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
*/
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
void
Allow memory contexts to have both fixed and variable ident strings. Originally, we treated memory context names as potentially variable in all cases, and therefore always copied them into the context header. Commit 9fa6f00b1 rethought this a little bit and invented a distinction between fixed and variable names, skipping the copy step for the former. But we can make things both simpler and more useful by instead allowing there to be two parts to a context's identification, a fixed "name" and an optional, variable "ident". The name supplied in the context create call is now required to be a compile-time-constant string in all cases, as it is never copied but just pointed to. The "ident" string, if wanted, is supplied later. This is needed because typically we want the ident to be stored inside the context so that it's cleaned up automatically on context deletion; that means it has to be copied into the context before we can set the pointer. The cost of this approach is basically just an additional pointer field in struct MemoryContextData, which isn't much overhead, and is bought back entirely in the AllocSet case by not needing a headerSize field anymore, since we no longer have to cope with variable header length. In addition, we can simplify the internal interfaces for memory context creation still further, saving a few cycles there. And it's no longer true that a custom identifier disqualifies a context from participating in aset.c's freelist scheme, so possibly there's some win on that end. All the places that were using non-compile-time-constant context names are adjusted to put the variable info into the "ident" instead. This allows more effective identification of those contexts in many cases; for example, subsidary contexts of relcache entries are now identified by both type (e.g. "index info") and relname, where before you got only one or the other. Contexts associated with PL function cache entries are now identified more fully and uniformly, too. I also arranged for plancache contexts to use the query source string as their identifier. This is basically free for CachedPlanSources, as they contained a copy of that string already. We pay an extra pstrdup to do it for CachedPlans. That could perhaps be avoided, but it would make things more fragile (since the CachedPlanSource is sometimes destroyed first). I suspect future improvements in error reporting will require CachedPlans to have a copy of that string anyway, so it's not clear that it's worth moving mountains to avoid it now. This also changes the APIs for context statistics routines so that the context-specific routines no longer assume that output goes straight to stderr, nor do they know all details of the output format. This is useful immediately to reduce code duplication, and it also allows for external code to do something with stats output that's different from printing to stderr. The reason for pushing this now rather than waiting for v12 is that it rethinks some of the API changes made by commit 9fa6f00b1. Seems better for extension authors to endure just one round of API changes not two. Discussion: https://postgr.es/m/CAB=Je-FdtmFZ9y9REHD7VsSrnCkiBhsA4mdsLKSPauwXtQBeNA@mail.gmail.com
2018-03-27 22:46:47 +02:00
SlabStats(MemoryContext context,
MemoryStatsPrintFunc printfunc, void *passthru,
Add function to log the memory contexts of specified backend process. Commit 3e98c0bafb added pg_backend_memory_contexts view to display the memory contexts of the backend process. However its target process is limited to the backend that is accessing to the view. So this is not so convenient when investigating the local memory bloat of other backend process. To improve this situation, this commit adds pg_log_backend_memory_contexts() function that requests to log the memory contexts of the specified backend process. This information can be also collected by calling MemoryContextStats(TopMemoryContext) via a debugger. But this technique cannot be used in some environments because no debugger is available there. So, pg_log_backend_memory_contexts() allows us to see the memory contexts of specified backend more easily. Only superusers are allowed to request to log the memory contexts because allowing any users to issue this request at an unbounded rate would cause lots of log messages and which can lead to denial of service. On receipt of the request, at the next CHECK_FOR_INTERRUPTS(), the target backend logs its memory contexts at LOG_SERVER_ONLY level, so that these memory contexts will appear in the server log but not be sent to the client. It logs one message per memory context. Because if it buffers all memory contexts into StringInfo to log them as one message, which may require the buffer to be enlarged very much and lead to OOM error since there can be a large number of memory contexts in a backend. When a backend process is consuming huge memory, logging all its memory contexts might overrun available disk space. To prevent this, now this patch limits the number of child contexts to log per parent to 100. As with MemoryContextStats(), it supposes that practical cases where the log gets long will typically be huge numbers of siblings under the same parent context; while the additional debugging value from seeing details about individual siblings beyond 100 will not be large. There was another proposed patch to add the function to return the memory contexts of specified backend as the result sets, instead of logging them, in the discussion. However that patch is not included in this commit because it had several issues to address. Thanks to Tatsuhito Kasahara, Andres Freund, Tom Lane, Tomas Vondra, Michael Paquier, Kyotaro Horiguchi and Zhihong Yu for the discussion. Bump catalog version. Author: Atsushi Torikoshi Reviewed-by: Kyotaro Horiguchi, Zhihong Yu, Fujii Masao Discussion: https://postgr.es/m/0271f440ac77f2a4180e0e56ebd944d1@oss.nttdata.com
2021-04-06 06:44:15 +02:00
MemoryContextCounters *totals,
bool print_to_stderr)
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
{
SlabContext *slab = castNode(SlabContext, context);
Size nblocks = 0;
Size freechunks = 0;
Rethink MemoryContext creation to improve performance. This patch makes a number of interrelated changes to reduce the overhead involved in creating/deleting memory contexts. The key ideas are: * Include the AllocSetContext header of an aset.c context in its first malloc request, rather than allocating it separately in TopMemoryContext. This means that we now always create an initial or "keeper" block in an aset, even if it never receives any allocation requests. * Create freelists in which we can save and recycle recently-destroyed asets (this idea is due to Robert Haas). * In the common case where the name of a context is a constant string, just store a pointer to it in the context header, rather than copying the string. The first change eliminates a palloc/pfree cycle per context, and also avoids bloat in TopMemoryContext, at the price that creating a context now involves a malloc/free cycle even if the context never receives any allocations. That would be a loser for some common usage patterns, but recycling short-lived contexts via the freelist eliminates that pain. Avoiding copying constant strings not only saves strlen() and strcpy() overhead, but is an essential part of the freelist optimization because it makes the context header size constant. Currently we make no attempt to use the freelist for contexts with non-constant names. (Perhaps someday we'll need to think harder about that, but in current usage, most contexts with custom names are long-lived anyway.) The freelist management in this initial commit is pretty simplistic, and we might want to refine it later --- but in common workloads that will never matter because the freelists will never get full anyway. To create a context with a non-constant name, one is now required to call AllocSetContextCreateExtended and specify the MEMCONTEXT_COPY_NAME option. AllocSetContextCreate becomes a wrapper macro, and it includes a test that will complain about non-string-literal context name parameters on gcc and similar compilers. An unfortunate side effect of making AllocSetContextCreate a macro is that one is now *required* to use the size parameter abstraction macros (ALLOCSET_DEFAULT_SIZES and friends) with it; the pre-9.6 habit of writing out individual size parameters no longer works unless you switch to AllocSetContextCreateExtended. Internally to the memory-context-related modules, the context creation APIs are simplified, removing the rather baroque original design whereby a context-type module called mcxt.c which then called back into the context-type module. That saved a bit of code duplication, but not much, and it prevented context-type modules from exercising control over the allocation of context headers. In passing, I converted the test-and-elog validation of aset size parameters into Asserts to save a few more cycles. The original thought was that callers might compute size parameters on the fly, but in practice nobody does that, so it's useless to expend cycles on checking those numbers in production builds. Also, mark the memory context method-pointer structs "const", just for cleanliness. Discussion: https://postgr.es/m/2264.1512870796@sss.pgh.pa.us
2017-12-13 19:55:12 +01:00
Size totalspace;
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
Size freespace = 0;
int i;
Rethink MemoryContext creation to improve performance. This patch makes a number of interrelated changes to reduce the overhead involved in creating/deleting memory contexts. The key ideas are: * Include the AllocSetContext header of an aset.c context in its first malloc request, rather than allocating it separately in TopMemoryContext. This means that we now always create an initial or "keeper" block in an aset, even if it never receives any allocation requests. * Create freelists in which we can save and recycle recently-destroyed asets (this idea is due to Robert Haas). * In the common case where the name of a context is a constant string, just store a pointer to it in the context header, rather than copying the string. The first change eliminates a palloc/pfree cycle per context, and also avoids bloat in TopMemoryContext, at the price that creating a context now involves a malloc/free cycle even if the context never receives any allocations. That would be a loser for some common usage patterns, but recycling short-lived contexts via the freelist eliminates that pain. Avoiding copying constant strings not only saves strlen() and strcpy() overhead, but is an essential part of the freelist optimization because it makes the context header size constant. Currently we make no attempt to use the freelist for contexts with non-constant names. (Perhaps someday we'll need to think harder about that, but in current usage, most contexts with custom names are long-lived anyway.) The freelist management in this initial commit is pretty simplistic, and we might want to refine it later --- but in common workloads that will never matter because the freelists will never get full anyway. To create a context with a non-constant name, one is now required to call AllocSetContextCreateExtended and specify the MEMCONTEXT_COPY_NAME option. AllocSetContextCreate becomes a wrapper macro, and it includes a test that will complain about non-string-literal context name parameters on gcc and similar compilers. An unfortunate side effect of making AllocSetContextCreate a macro is that one is now *required* to use the size parameter abstraction macros (ALLOCSET_DEFAULT_SIZES and friends) with it; the pre-9.6 habit of writing out individual size parameters no longer works unless you switch to AllocSetContextCreateExtended. Internally to the memory-context-related modules, the context creation APIs are simplified, removing the rather baroque original design whereby a context-type module called mcxt.c which then called back into the context-type module. That saved a bit of code duplication, but not much, and it prevented context-type modules from exercising control over the allocation of context headers. In passing, I converted the test-and-elog validation of aset size parameters into Asserts to save a few more cycles. The original thought was that callers might compute size parameters on the fly, but in practice nobody does that, so it's useless to expend cycles on checking those numbers in production builds. Also, mark the memory context method-pointer structs "const", just for cleanliness. Discussion: https://postgr.es/m/2264.1512870796@sss.pgh.pa.us
2017-12-13 19:55:12 +01:00
/* Include context header in totalspace */
totalspace = slab->headerSize;
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
for (i = 0; i <= slab->chunksPerBlock; i++)
{
dlist_iter iter;
dlist_foreach(iter, &slab->freelist[i])
{
SlabBlock *block = dlist_container(SlabBlock, node, iter.cur);
nblocks++;
totalspace += slab->blockSize;
freespace += slab->fullChunkSize * block->nfree;
freechunks += block->nfree;
}
}
Allow memory contexts to have both fixed and variable ident strings. Originally, we treated memory context names as potentially variable in all cases, and therefore always copied them into the context header. Commit 9fa6f00b1 rethought this a little bit and invented a distinction between fixed and variable names, skipping the copy step for the former. But we can make things both simpler and more useful by instead allowing there to be two parts to a context's identification, a fixed "name" and an optional, variable "ident". The name supplied in the context create call is now required to be a compile-time-constant string in all cases, as it is never copied but just pointed to. The "ident" string, if wanted, is supplied later. This is needed because typically we want the ident to be stored inside the context so that it's cleaned up automatically on context deletion; that means it has to be copied into the context before we can set the pointer. The cost of this approach is basically just an additional pointer field in struct MemoryContextData, which isn't much overhead, and is bought back entirely in the AllocSet case by not needing a headerSize field anymore, since we no longer have to cope with variable header length. In addition, we can simplify the internal interfaces for memory context creation still further, saving a few cycles there. And it's no longer true that a custom identifier disqualifies a context from participating in aset.c's freelist scheme, so possibly there's some win on that end. All the places that were using non-compile-time-constant context names are adjusted to put the variable info into the "ident" instead. This allows more effective identification of those contexts in many cases; for example, subsidary contexts of relcache entries are now identified by both type (e.g. "index info") and relname, where before you got only one or the other. Contexts associated with PL function cache entries are now identified more fully and uniformly, too. I also arranged for plancache contexts to use the query source string as their identifier. This is basically free for CachedPlanSources, as they contained a copy of that string already. We pay an extra pstrdup to do it for CachedPlans. That could perhaps be avoided, but it would make things more fragile (since the CachedPlanSource is sometimes destroyed first). I suspect future improvements in error reporting will require CachedPlans to have a copy of that string anyway, so it's not clear that it's worth moving mountains to avoid it now. This also changes the APIs for context statistics routines so that the context-specific routines no longer assume that output goes straight to stderr, nor do they know all details of the output format. This is useful immediately to reduce code duplication, and it also allows for external code to do something with stats output that's different from printing to stderr. The reason for pushing this now rather than waiting for v12 is that it rethinks some of the API changes made by commit 9fa6f00b1. Seems better for extension authors to endure just one round of API changes not two. Discussion: https://postgr.es/m/CAB=Je-FdtmFZ9y9REHD7VsSrnCkiBhsA4mdsLKSPauwXtQBeNA@mail.gmail.com
2018-03-27 22:46:47 +02:00
if (printfunc)
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
{
Allow memory contexts to have both fixed and variable ident strings. Originally, we treated memory context names as potentially variable in all cases, and therefore always copied them into the context header. Commit 9fa6f00b1 rethought this a little bit and invented a distinction between fixed and variable names, skipping the copy step for the former. But we can make things both simpler and more useful by instead allowing there to be two parts to a context's identification, a fixed "name" and an optional, variable "ident". The name supplied in the context create call is now required to be a compile-time-constant string in all cases, as it is never copied but just pointed to. The "ident" string, if wanted, is supplied later. This is needed because typically we want the ident to be stored inside the context so that it's cleaned up automatically on context deletion; that means it has to be copied into the context before we can set the pointer. The cost of this approach is basically just an additional pointer field in struct MemoryContextData, which isn't much overhead, and is bought back entirely in the AllocSet case by not needing a headerSize field anymore, since we no longer have to cope with variable header length. In addition, we can simplify the internal interfaces for memory context creation still further, saving a few cycles there. And it's no longer true that a custom identifier disqualifies a context from participating in aset.c's freelist scheme, so possibly there's some win on that end. All the places that were using non-compile-time-constant context names are adjusted to put the variable info into the "ident" instead. This allows more effective identification of those contexts in many cases; for example, subsidary contexts of relcache entries are now identified by both type (e.g. "index info") and relname, where before you got only one or the other. Contexts associated with PL function cache entries are now identified more fully and uniformly, too. I also arranged for plancache contexts to use the query source string as their identifier. This is basically free for CachedPlanSources, as they contained a copy of that string already. We pay an extra pstrdup to do it for CachedPlans. That could perhaps be avoided, but it would make things more fragile (since the CachedPlanSource is sometimes destroyed first). I suspect future improvements in error reporting will require CachedPlans to have a copy of that string anyway, so it's not clear that it's worth moving mountains to avoid it now. This also changes the APIs for context statistics routines so that the context-specific routines no longer assume that output goes straight to stderr, nor do they know all details of the output format. This is useful immediately to reduce code duplication, and it also allows for external code to do something with stats output that's different from printing to stderr. The reason for pushing this now rather than waiting for v12 is that it rethinks some of the API changes made by commit 9fa6f00b1. Seems better for extension authors to endure just one round of API changes not two. Discussion: https://postgr.es/m/CAB=Je-FdtmFZ9y9REHD7VsSrnCkiBhsA4mdsLKSPauwXtQBeNA@mail.gmail.com
2018-03-27 22:46:47 +02:00
char stats_string[200];
snprintf(stats_string, sizeof(stats_string),
Fix incorrect format placeholders
2021-12-22 08:34:10 +01:00
"%zu total in %zu blocks; %zu free (%zu chunks); %zu used",
Allow memory contexts to have both fixed and variable ident strings. Originally, we treated memory context names as potentially variable in all cases, and therefore always copied them into the context header. Commit 9fa6f00b1 rethought this a little bit and invented a distinction between fixed and variable names, skipping the copy step for the former. But we can make things both simpler and more useful by instead allowing there to be two parts to a context's identification, a fixed "name" and an optional, variable "ident". The name supplied in the context create call is now required to be a compile-time-constant string in all cases, as it is never copied but just pointed to. The "ident" string, if wanted, is supplied later. This is needed because typically we want the ident to be stored inside the context so that it's cleaned up automatically on context deletion; that means it has to be copied into the context before we can set the pointer. The cost of this approach is basically just an additional pointer field in struct MemoryContextData, which isn't much overhead, and is bought back entirely in the AllocSet case by not needing a headerSize field anymore, since we no longer have to cope with variable header length. In addition, we can simplify the internal interfaces for memory context creation still further, saving a few cycles there. And it's no longer true that a custom identifier disqualifies a context from participating in aset.c's freelist scheme, so possibly there's some win on that end. All the places that were using non-compile-time-constant context names are adjusted to put the variable info into the "ident" instead. This allows more effective identification of those contexts in many cases; for example, subsidary contexts of relcache entries are now identified by both type (e.g. "index info") and relname, where before you got only one or the other. Contexts associated with PL function cache entries are now identified more fully and uniformly, too. I also arranged for plancache contexts to use the query source string as their identifier. This is basically free for CachedPlanSources, as they contained a copy of that string already. We pay an extra pstrdup to do it for CachedPlans. That could perhaps be avoided, but it would make things more fragile (since the CachedPlanSource is sometimes destroyed first). I suspect future improvements in error reporting will require CachedPlans to have a copy of that string anyway, so it's not clear that it's worth moving mountains to avoid it now. This also changes the APIs for context statistics routines so that the context-specific routines no longer assume that output goes straight to stderr, nor do they know all details of the output format. This is useful immediately to reduce code duplication, and it also allows for external code to do something with stats output that's different from printing to stderr. The reason for pushing this now rather than waiting for v12 is that it rethinks some of the API changes made by commit 9fa6f00b1. Seems better for extension authors to endure just one round of API changes not two. Discussion: https://postgr.es/m/CAB=Je-FdtmFZ9y9REHD7VsSrnCkiBhsA4mdsLKSPauwXtQBeNA@mail.gmail.com
2018-03-27 22:46:47 +02:00
totalspace, nblocks, freespace, freechunks,
totalspace - freespace);
Add function to log the memory contexts of specified backend process. Commit 3e98c0bafb added pg_backend_memory_contexts view to display the memory contexts of the backend process. However its target process is limited to the backend that is accessing to the view. So this is not so convenient when investigating the local memory bloat of other backend process. To improve this situation, this commit adds pg_log_backend_memory_contexts() function that requests to log the memory contexts of the specified backend process. This information can be also collected by calling MemoryContextStats(TopMemoryContext) via a debugger. But this technique cannot be used in some environments because no debugger is available there. So, pg_log_backend_memory_contexts() allows us to see the memory contexts of specified backend more easily. Only superusers are allowed to request to log the memory contexts because allowing any users to issue this request at an unbounded rate would cause lots of log messages and which can lead to denial of service. On receipt of the request, at the next CHECK_FOR_INTERRUPTS(), the target backend logs its memory contexts at LOG_SERVER_ONLY level, so that these memory contexts will appear in the server log but not be sent to the client. It logs one message per memory context. Because if it buffers all memory contexts into StringInfo to log them as one message, which may require the buffer to be enlarged very much and lead to OOM error since there can be a large number of memory contexts in a backend. When a backend process is consuming huge memory, logging all its memory contexts might overrun available disk space. To prevent this, now this patch limits the number of child contexts to log per parent to 100. As with MemoryContextStats(), it supposes that practical cases where the log gets long will typically be huge numbers of siblings under the same parent context; while the additional debugging value from seeing details about individual siblings beyond 100 will not be large. There was another proposed patch to add the function to return the memory contexts of specified backend as the result sets, instead of logging them, in the discussion. However that patch is not included in this commit because it had several issues to address. Thanks to Tatsuhito Kasahara, Andres Freund, Tom Lane, Tomas Vondra, Michael Paquier, Kyotaro Horiguchi and Zhihong Yu for the discussion. Bump catalog version. Author: Atsushi Torikoshi Reviewed-by: Kyotaro Horiguchi, Zhihong Yu, Fujii Masao Discussion: https://postgr.es/m/0271f440ac77f2a4180e0e56ebd944d1@oss.nttdata.com
2021-04-06 06:44:15 +02:00
printfunc(context, passthru, stats_string, print_to_stderr);
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
}
if (totals)
{
totals->nblocks += nblocks;
totals->freechunks += freechunks;
totals->totalspace += totalspace;
totals->freespace += freespace;
}
}
#ifdef MEMORY_CONTEXT_CHECKING
/*
* SlabCheck
* Walk through chunks and check consistency of memory.
*
* NOTE: report errors as WARNING, *not* ERROR or FATAL. Otherwise you'll
* find yourself in an infinite loop when trouble occurs, because this
* routine will be entered again when elog cleanup tries to release memory!
*/
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
void
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
SlabCheck(MemoryContext context)
{
int i;
SlabContext *slab = castNode(SlabContext, context);
Rethink MemoryContext creation to improve performance. This patch makes a number of interrelated changes to reduce the overhead involved in creating/deleting memory contexts. The key ideas are: * Include the AllocSetContext header of an aset.c context in its first malloc request, rather than allocating it separately in TopMemoryContext. This means that we now always create an initial or "keeper" block in an aset, even if it never receives any allocation requests. * Create freelists in which we can save and recycle recently-destroyed asets (this idea is due to Robert Haas). * In the common case where the name of a context is a constant string, just store a pointer to it in the context header, rather than copying the string. The first change eliminates a palloc/pfree cycle per context, and also avoids bloat in TopMemoryContext, at the price that creating a context now involves a malloc/free cycle even if the context never receives any allocations. That would be a loser for some common usage patterns, but recycling short-lived contexts via the freelist eliminates that pain. Avoiding copying constant strings not only saves strlen() and strcpy() overhead, but is an essential part of the freelist optimization because it makes the context header size constant. Currently we make no attempt to use the freelist for contexts with non-constant names. (Perhaps someday we'll need to think harder about that, but in current usage, most contexts with custom names are long-lived anyway.) The freelist management in this initial commit is pretty simplistic, and we might want to refine it later --- but in common workloads that will never matter because the freelists will never get full anyway. To create a context with a non-constant name, one is now required to call AllocSetContextCreateExtended and specify the MEMCONTEXT_COPY_NAME option. AllocSetContextCreate becomes a wrapper macro, and it includes a test that will complain about non-string-literal context name parameters on gcc and similar compilers. An unfortunate side effect of making AllocSetContextCreate a macro is that one is now *required* to use the size parameter abstraction macros (ALLOCSET_DEFAULT_SIZES and friends) with it; the pre-9.6 habit of writing out individual size parameters no longer works unless you switch to AllocSetContextCreateExtended. Internally to the memory-context-related modules, the context creation APIs are simplified, removing the rather baroque original design whereby a context-type module called mcxt.c which then called back into the context-type module. That saved a bit of code duplication, but not much, and it prevented context-type modules from exercising control over the allocation of context headers. In passing, I converted the test-and-elog validation of aset size parameters into Asserts to save a few more cycles. The original thought was that callers might compute size parameters on the fly, but in practice nobody does that, so it's useless to expend cycles on checking those numbers in production builds. Also, mark the memory context method-pointer structs "const", just for cleanliness. Discussion: https://postgr.es/m/2264.1512870796@sss.pgh.pa.us
2017-12-13 19:55:12 +01:00
const char *name = slab->header.name;
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
Assert(slab);
Assert(slab->chunksPerBlock > 0);
/* walk all the freelists */
for (i = 0; i <= slab->chunksPerBlock; i++)
{
int j,
nfree;
dlist_iter iter;
/* walk all blocks on this freelist */
dlist_foreach(iter, &slab->freelist[i])
{
int idx;
SlabBlock *block = dlist_container(SlabBlock, node, iter.cur);
/*
* Make sure the number of free chunks (in the block header)
* matches position in the freelist.
*/
if (block->nfree != i)
elog(WARNING, "problem in slab %s: number of free chunks %d in block %p does not match freelist %d",
name, block->nfree, block, i);
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
/* make sure the slab pointer correctly points to this context */
if (block->slab != slab)
elog(WARNING, "problem in slab %s: bogus slab link in block %p",
name, block);
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
/* reset the bitmap of free chunks for this block */
Allocate freechunks bitmap as part of SlabContext The bitmap used by SlabCheck to cross-check free chunks in a block used to be allocated for each SlabCheck call, and was never freed. The memory leak could be fixed by simply adding a pfree call, but it's actually a bad idea to do any allocations in SlabCheck at all as it assumes the state of the memory management as a whole is sane. So instead we allocate the bitmap as part of SlabContext, which means we don't need to do any allocations in SlabCheck and the bitmap goes away together with the SlabContext. Backpatch to 10, where the Slab context was introduced. Author: Tomas Vondra Reported-by: Andres Freund Reviewed-by: Tom Lane Backpatch-through: 10 Discussion: https://www.postgresql.org/message-id/20200116044119.g45f7pmgz4jmodxj%40alap3.anarazel.de
2020-01-17 14:06:28 +01:00
memset(slab->freechunks, 0, (slab->chunksPerBlock * sizeof(bool)));
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
idx = block->firstFreeChunk;
/*
* Now walk through the chunks, count the free ones and also
* perform some additional checks for the used ones. As the chunk
* freelist is stored within the chunks themselves, we have to
* walk through the chunks and construct our own bitmap.
*/
nfree = 0;
while (idx < slab->chunksPerBlock)
{
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
MemoryChunk *chunk;
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
/* count the chunk as free, add it to the bitmap */
nfree++;
Allocate freechunks bitmap as part of SlabContext The bitmap used by SlabCheck to cross-check free chunks in a block used to be allocated for each SlabCheck call, and was never freed. The memory leak could be fixed by simply adding a pfree call, but it's actually a bad idea to do any allocations in SlabCheck at all as it assumes the state of the memory management as a whole is sane. So instead we allocate the bitmap as part of SlabContext, which means we don't need to do any allocations in SlabCheck and the bitmap goes away together with the SlabContext. Backpatch to 10, where the Slab context was introduced. Author: Tomas Vondra Reported-by: Andres Freund Reviewed-by: Tom Lane Backpatch-through: 10 Discussion: https://www.postgresql.org/message-id/20200116044119.g45f7pmgz4jmodxj%40alap3.anarazel.de
2020-01-17 14:06:28 +01:00
slab->freechunks[idx] = true;
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
/* read index of the next free chunk */
chunk = SlabBlockGetChunk(slab, block, idx);
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
VALGRIND_MAKE_MEM_DEFINED(MemoryChunkGetPointer(chunk), sizeof(int32));
idx = *(int32 *) MemoryChunkGetPointer(chunk);
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
}
for (j = 0; j < slab->chunksPerBlock; j++)
{
/* non-zero bit in the bitmap means chunk the chunk is used */
Allocate freechunks bitmap as part of SlabContext The bitmap used by SlabCheck to cross-check free chunks in a block used to be allocated for each SlabCheck call, and was never freed. The memory leak could be fixed by simply adding a pfree call, but it's actually a bad idea to do any allocations in SlabCheck at all as it assumes the state of the memory management as a whole is sane. So instead we allocate the bitmap as part of SlabContext, which means we don't need to do any allocations in SlabCheck and the bitmap goes away together with the SlabContext. Backpatch to 10, where the Slab context was introduced. Author: Tomas Vondra Reported-by: Andres Freund Reviewed-by: Tom Lane Backpatch-through: 10 Discussion: https://www.postgresql.org/message-id/20200116044119.g45f7pmgz4jmodxj%40alap3.anarazel.de
2020-01-17 14:06:28 +01:00
if (!slab->freechunks[j])
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
{
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
MemoryChunk *chunk = SlabBlockGetChunk(slab, block, j);
SlabBlock *chunkblock = (SlabBlock *) MemoryChunkGetBlock(chunk);
/*
* check the chunk's blockoffset correctly points back to
* the block
*/
if (chunkblock != block)
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
elog(WARNING, "problem in slab %s: bogus block link in block %p, chunk %p",
name, block, chunk);
/* there might be sentinel (thanks to alignment) */
Improve performance of and reduce overheads of memory management Whenever we palloc a chunk of memory, traditionally, we prefix the returned pointer with a pointer to the memory context to which the chunk belongs. This is required so that we're able to easily determine the owning context when performing operations such as pfree() and repalloc(). For the AllocSet context, prior to this commit we additionally prefixed the pointer to the owning context with the size of the chunk. This made the header 16 bytes in size. This 16-byte overhead was required for all AllocSet allocations regardless of the allocation size. For the generation context, the problem was worse; in addition to the pointer to the owning context and chunk size, we also stored a pointer to the owning block so that we could track the number of freed chunks on a block. The slab allocator had a 16-byte chunk header. The changes being made here reduce the chunk header size down to just 8 bytes for all 3 of our memory context types. For small to medium sized allocations, this significantly increases the number of chunks that we can fit on a given block which results in much more efficient use of memory. Additionally, this commit completely changes the rule that pointers to palloc'd memory must be directly prefixed by a pointer to the owning memory context and instead, we now insist that they're directly prefixed by an 8-byte value where the least significant 3-bits are set to a value to indicate which type of memory context the pointer belongs to. Using those 3 bits as an index (known as MemoryContextMethodID) to a new array which stores the methods for each memory context type, we're now able to pass the pointer given to functions such as pfree() and repalloc() to the function specific to that context implementation to allow them to devise their own methods of finding the memory context which owns the given allocated chunk of memory. The reason we're able to reduce the chunk header down to just 8 bytes is because of the way we make use of the remaining 61 bits of the required 8-byte chunk header. Here we also implement a general-purpose MemoryChunk struct which makes use of those 61 remaining bits to allow the storage of a 30-bit value which the MemoryContext is free to use as it pleases, and also the number of bytes which must be subtracted from the chunk to get a reference to the block that the chunk is stored on (also 30 bits). The 1 additional remaining bit is to denote if the chunk is an "external" chunk or not. External here means that the chunk header does not store the 30-bit value or the block offset. The MemoryContext can use these external chunks at any time, but must use them if any of the two 30-bit fields are not large enough for the value(s) that need to be stored in them. When the chunk is marked as external, it is up to the MemoryContext to devise its own means to determine the block offset. Using 3-bits for the MemoryContextMethodID does mean we're limiting ourselves to only having a maximum of 8 different memory context types. We could reduce the bit space for the 30-bit value a little to make way for more than 3 bits, but it seems like it might be better to do that only if we ever need more than 8 context types. This would only be a problem if some future memory context type which does not use MemoryChunk really couldn't give up any of the 61 remaining bits in the chunk header. With this MemoryChunk, each of our 3 memory context types can quickly obtain a reference to the block any given chunk is located on. AllocSet is able to find the context to which the chunk is owned, by first obtaining a reference to the block by subtracting the block offset as is stored in the 'hdrmask' field and then referencing the block's 'aset' field. The Generation context uses the same method, but GenerationBlock did not have a field pointing back to the owning context, so one is added by this commit. In aset.c and generation.c, all allocations larger than allocChunkLimit are stored on dedicated blocks. When there's just a single chunk on a block like this, it's easy to find the block from the chunk, we just subtract the size of the block header from the chunk pointer. The size of these chunks is also known as we store the endptr on the block, so we can just subtract the pointer to the allocated memory from that. Because we can easily find the owning block and the size of the chunk for these dedicated blocks, we just always use external chunks for allocation sizes larger than allocChunkLimit. For generation.c, this sidesteps the problem of non-external MemoryChunks being unable to represent chunk sizes >= 1GB. This is less of a problem for aset.c as we store the free list index in the MemoryChunk's spare 30-bit field (the value of which will never be close to using all 30-bits). We can easily reverse engineer the chunk size from this when needed. Storing this saves AllocSetFree() from having to make a call to AllocSetFreeIndex() to determine which free list to put the newly freed chunk on. For the slab allocator, this commit adds a new restriction that slab chunks cannot be >= 1GB in size. If there happened to be any users of slab.c which used chunk sizes this large, they really should be using AllocSet instead. Here we also add a restriction that normal non-dedicated blocks cannot be 1GB or larger. It's now not possible to pass a 'maxBlockSize' >= 1GB during the creation of an AllocSet or Generation context. Allocations can still be larger than 1GB, it's just these will always be on dedicated blocks (which do not have the 1GB restriction). Author: Andres Freund, David Rowley Discussion: https://postgr.es/m/CAApHDvpjauCRXcgcaL6+e3eqecEHoeRm9D-kcbuvBitgPnW=vw@mail.gmail.com
2022-08-29 07:15:00 +02:00
if (slab->chunkSize < (slab->fullChunkSize - Slab_CHUNKHDRSZ))
Use MAXALIGN() in calculations using sizeof(SlabBlock) c6e0fe1f2 added a new pointer field to SlabBlock to make it 4 bytes larger on 32-bit machines. Prior to that commit, the size of that struct was a multiple of 8, which meant that MAXALIGN(sizeof(SlabBlock)) was the same as sizeof(SlabBlock), however, after c6e0fe1f2, due to the addition of the new pointer field to store a pointer to the owning context, that was no longer true on builds with sizeof(void *) == 4. This problem was highlighted by an Assert failure which was checking that the pointer given to pfree() was MAXALIGNED. Various 32-bit ARM buildfarm animals were failing. These have MAXIMUM_ALIGNOF of 8. The only 32-bit testing I'd managed to do on c6e0fe1f2 had been on x86, which has a MAXIMUM_ALIGNOF of 4, therefore did not exhibit this issue. Here we define Slab_BLOCKHDRSZ and copy what is being done in aset.c and generation.c for doing calculations based on the size of the context's block type. This means that SlabAlloc() will now always return a MAXALIGNed pointer. This also fixes an incorrect sentinel_ok() check in SlabCheck() which was incorrectly checking the wrong sentinel byte. This must have previously not caused any issues due to the fullChunkSize never being large enough to store the sentinel byte. Diagnosed-by: Tomas Vondra, Tom Lane Author: Tomas Vondra, David Rowley Discussion: https://postgr.es/m/CAA4eK1%2B1JyW5TiL%3DyV-3Uq1CrfnTyn0Xrk5uArt31Z%3D8rgPhXQ%40mail.gmail.com
2022-08-30 04:36:04 +02:00
if (!sentinel_ok(chunk, Slab_CHUNKHDRSZ + slab->chunkSize))
Reduce size of common allocation header. The new slab allocator needs different per-allocation information than the classical aset.c. The definition in 58b25e981 wasn't sufficiently careful on 32 platforms with 8 byte alignment, leading to buildfarm failures. That's not entirely easy to fix by just adjusting the definition. As slab.c doesn't actually need the size part(s) of the common header, all chunks are equally sized after all, it seems better to instead reduce the header to the part needed by all allocators, namely which context an allocation belongs to. That has the advantage of reducing the overhead of slab allocations, and also allows for more flexibility in future allocators. To avoid spreading the logic about accessing a chunk's context around, centralize it in GetMemoryChunkContext(), which allows to delete a good number of lines. A followup commit will revise the mmgr/README portion about StandardChunkHeader, and more. Author: Andres Freund Discussion: https://postgr.es/m/20170228074420.aazv4iw6k562mnxg@alap3.anarazel.de
2017-02-28 08:32:22 +01:00
elog(WARNING, "problem in slab %s: detected write past chunk end in block %p, chunk %p",
name, block, chunk);
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
}
}
/*
* Make sure we got the expected number of free chunks (as tracked
* in the block header).
*/
if (nfree != block->nfree)
elog(WARNING, "problem in slab %s: number of free chunks %d in block %p does not match bitmap %d",
name, block->nfree, block, nfree);
}
}
Add transparent block-level memory accounting Adds accounting of memory allocated in a memory context. Compared to various ad hoc solutions, the main advantage is that the accounting is transparent and does not require direct control over allocations (this matters for use cases where the allocations happen in user code, like for example aggregate states allocated in a transition functions). To reduce overhead, the accounting happens at the block level (not for individual chunks) and only the context immediately owning the block is updated. When inquiring about amount of memory allocated in a context, we have to recursively walk all children contexts. This "lazy" accounting works well for cases with relatively small number of contexts in the relevant subtree and/or with infrequent inquiries. Author: Jeff Davis Reivewed-by: Tomas Vondra, Melanie Plageman, Soumyadeep Chakraborty Discussion: https://www.postgresql.org/message-id/flat/027a129b8525601c6a680d27ce3a7172dab61aab.camel@j-davis.com
2019-10-01 03:13:39 +02:00
Revert "Specialize MemoryContextMemAllocated()." This reverts commit e00912e11a9ec2d29274ed8a6465e81385906dc2.
2020-03-19 19:20:39 +01:00
Assert(slab->nblocks * slab->blockSize == context->mem_allocated);
Add "Slab" MemoryContext implementation for efficient equal-sized allocations. The default general purpose aset.c style memory context is not a great choice for allocations that are all going to be evenly sized, especially when those objects aren't small, and have varying lifetimes. There tends to be a lot of fragmentation, larger allocations always directly go to libc rather than have their cost amortized over several pallocs. These problems lead to the introduction of ad-hoc slab allocators in reorderbuffer.c. But it turns out that the simplistic implementation leads to problems when a lot of objects are allocated and freed, as aset.c is still the underlying implementation. Especially freeing can easily run into O(n^2) behavior in aset.c. While the O(n^2) behavior in aset.c can, and probably will, be addressed, custom allocators for this behavior are more efficient both in space and time. This allocator is for evenly sized allocations, and supports both cheap allocations and freeing, without fragmenting significantly. It does so by allocating evenly sized blocks via malloc(), and carves them into chunks that can be used for allocations. In order to release blocks to the OS as early as possible, chunks are allocated from the fullest block that still has free objects, increasing the likelihood of a block being entirely unused. A subsequent commit uses this in reorderbuffer.c, but a further allocator is needed to resolve the performance problems triggering this work. There likely are further potentialy uses of this allocator besides reorderbuffer.c. There's potential further optimizations of the new slab.c, in particular the array of freelists could be replaced by a more intelligent structure - but for now this looks more than good enough. Author: Tomas Vondra, editorialized by Andres Freund Reviewed-By: Andres Freund, Petr Jelinek, Robert Haas, Jim Nasby Discussion: https://postgr.es/m/d15dff83-0b37-28ed-0809-95a5cc7292ad@2ndquadrant.com
2017-02-27 12:41:44 +01:00
}
#endif /* MEMORY_CONTEXT_CHECKING */