In a similar effort to 413c18401, separate out the hot and cold paths in
GenerationAlloc() and SlabAlloc() to avoid having to setup the stack frame
for the hot path.
This additionally adjusts how we use the GenerationContext's freeblock.
Freeblock, when set, is now always empty and we only switch to using it
when the current allocation request finds the current block does not have
enough space and the freeblock is large enough to accomodate the
allocation.
This commit also adjusts GenerationFree() so that if we pfree the final
allocation in the current generation block, we now mark that block as
empty and keep it as the current block. Previously we free'd that block
and set the current block to NULL. Doing that meant we needed a special
case in GenerationAlloc to check if GenerationContext.block was NULL.
So this both reduces free/malloc calls and reduces the work done in
GenerationAlloc().
In passing, improve some comments in aset.c
Discussion: https://postgr.es/m/CAApHDvpHVSJqqb4B4OZLixr=CotKq-eKkbwZqvZVo_biYvUvQA@mail.gmail.com
Allocating from a free list or from a block which contains enough space
already, we deem to be common code paths and want to optimize for those.
Having to allocate a new block, either a normal block or a dedicated one
for a large allocation, we deem to be less common, therefore we class
that as "cold". Both cold paths require a malloc so are going to be
slower as a result of that regardless.
The main motivation here is to remove the calls to malloc() in the hot
path and because of this, the compiler is now free to not bother setting
up the stack frame in AllocSetAlloc(), thus making the hot path much
cheaper.
Author: Andres Freund
Reviewed-by: David Rowley
Discussion: https://postgr.es/m/20210719195950.gavgs6ujzmjfaiig@alap3.anarazel.de
Many modern compilers are able to optimize function calls to functions
where the parameters of the called function match a leading subset of
the calling function's parameters. If there are no instructions in the
calling function after the function is called, then the compiler is free
to avoid any stack frame setup and implement the function call as a
"jmp" rather than a "call". This is called sibling call optimization.
Here we adjust the memory allocation functions in mcxt.c to allow this
optimization. This requires moving some responsibility into the memory
context implementations themselves. It's now the responsibility of the
MemoryContext to check for malloc failures. This is good as it both
allows the sibling call optimization, but also because most small and
medium allocations won't call malloc and just allocate memory to an
existing block. That can't fail, so checking for NULLs in that case
isn't required.
Also, traditionally it's been the responsibility of palloc and the other
allocation functions in mcxt.c to check for invalid allocation size
requests. Here we also move the responsibility of checking that into the
MemoryContext. This isn't to allow the sibling call optimization, but
more because most of our allocators handle large allocations separately
and we can just add the size check when doing large allocations. We no
longer check this for non-large allocations at all.
To make checking the allocation request sizes and ERROR handling easier,
add some helper functions to mcxt.c for the allocators to use.
Author: Andres Freund
Reviewed-by: David Rowley
Discussion: https://postgr.es/m/20210719195950.gavgs6ujzmjfaiig@alap3.anarazel.de
Here we reduce the block size fields in AllocSetContext, GenerationContext
and SlabContext from Size down to uint32. Ever since c6e0fe1f2, blocks
for non-dedicated palloc chunks can no longer be larger than 1GB, so
there's no need to store the various block size fields as 64-bit values.
32 bits are enough to store 2^30.
Here we also further reduce the memory context struct sizes by getting rid
of the 'keeper' field which stores a pointer to the context's keeper
block. All the context types which have this field always allocate the
keeper block in the same allocation as the memory context itself, so the
keeper block always comes right at the end of the context struct. Add
some macros to calculate that address rather than storing it in the
context.
Overall, in AllocSetContext and GenerationContext, this saves 20 bytes on
64-bit builds which for ALLOCSET_SMALL_SIZES can sometimes mean the
difference between having to allocate a 2nd block and storing all the
required allocations on the keeper block alone. Such contexts are used
in relcache to store cache entries for indexes, of which there can be
a large number in a single backend.
Author: Melih Mutlu
Reviewed-by: David Rowley
Discussion: https://postgr.es/m/CAGPVpCSOW3uJ1QJmsMR9_oE3X7fG_z4q0AoU4R_w+2RzvroPFg@mail.gmail.com
If asked to decrease the size of a large (>8K) palloc chunk,
AllocSetRealloc could improperly change the Valgrind state of memory
beyond the new end of the chunk: it would mark data UNDEFINED as far
as the old end of the chunk after having done the realloc(3) call,
thus tromping on the state of memory that no longer belongs to it.
One would normally expect that memory to now be marked NOACCESS,
so that this mislabeling might prevent detection of later errors.
If realloc() had chosen to move the chunk someplace else (unlikely,
but well within its rights) we could also mismark perfectly-valid
DEFINED data as UNDEFINED, causing false-positive valgrind reports
later. Also, any malloc bookkeeping data placed within this area
might now be wrongly marked, causing additional problems.
Fix by replacing relevant uses of "oldsize" with "Min(size, oldsize)".
It's sufficient to mark as far as "size" when that's smaller, because
whatever remains in the new chunk size will be marked NOACCESS below,
and we expect realloc() to have taken care of marking the memory
beyond the new official end of the chunk.
While we're here, also rename the function's "oldsize" variable
to "oldchksize" to more clearly explain what it actually holds,
namely the distance to the end of the chunk (that is, requested size
plus trailing padding). This is more consistent with the use of
"size" and "chksize" to hold the new requested size and chunk size.
Add a new variable "oldsize" in the one stanza where we're actually
talking about the old requested size.
Oversight in commit c477f3e44. Back-patch to all supported branches,
as that was, just in case anybody wants to do valgrind testing on back
branches.
Karina Litskevich
Discussion: https://postgr.es/m/CACiT8iaAET-fmzjjZLjaJC4zwSJmrFyL7LAdHwaYyjjQOQ4hcg@mail.gmail.com
To allow testing for general support for fast bitscan intrinsics,
add symbols HAVE_BITSCAN_REVERSE and HAVE_BITSCAN_FORWARD.
Also do related cleanup in AllocSetFreeIndex(): Previously, we
tested for HAVE__BUILTIN_CLZ and copied the relevant internals of
pg_leftmost_one_pos32(), with a special fallback that does less
work than the general fallback for that function. Now that we have
a more general test, we just call pg_leftmost_one_pos32() directly
for platforms with intrinsic support. On gcc at least, there is no
difference in the binary for non-assert builds.
Discussion: https://www.postgresql.org/message-id/CAFBsxsEPc%2BFnX_0vmmQ5DHv60sk4rL_RZJ%2BMD6ei%3D76L0kFMvA%40mail.gmail.com
This fixes a couple of unused variable warnings that could be seen when
compiling with MEMORY_CONTEXT_CHECKING but not USE_ASSERT_CHECKING.
Defining MEMORY_CONTEXT_CHECKING without asserts is a little unusual,
however, we shouldn't be producing any warnings from such a build.
Author: Richard Guo
Discussion: https://postgr.es/m/CAMbWs4_D-vgLEh7eO47p=73u1jWO78NWf6Qfv1FndY1kG-Q-jA@mail.gmail.com
Because we added StaticAssertStmt() first before StaticAssertDecl(),
some uses as well as the instructions in c.h are now a bit backwards
from the "native" way static assertions are meant to be used in C.
This updates the guidance and moves some static assertions to better
places.
Specifically, since the addition of StaticAssertDecl(), we can put
static assertions at the file level. This moves a number of static
assertions out of function bodies, where they might have been stuck
out of necessity, to perhaps better places at the file level or in
header files.
Also, when the static assertion appears in a position where a
declaration is allowed, then using StaticAssertDecl() is more native
than StaticAssertStmt().
Reviewed-by: John Naylor <john.naylor@enterprisedb.com>
Discussion: https://www.postgresql.org/message-id/flat/941a04e7-dd6f-c0e4-8cdf-a33b3338cbda%40enterprisedb.com
Before commit c6e0fe1f2, functions such as AllocSetFree could pretty
safely presume that they were given a valid chunk pointer for their
own type of context, because the indirect call through a memory
context object and method struct would be very unlikely to work
otherwise. But now, if pfree() is mistakenly invoked on a pointer
to garbage, we have three chances in eight of ending up at one of
these functions. That means we need to take extra measures to
verify that we are looking at what we're supposed to be looking at,
especially in debug builds.
Hence, add code to verify that the chunk's back-link to a block header
leads to a memory context object that satisfies the right sort of
IsA() check. This is still a bit weaker than what we did before,
but for the moment assume that an IsA() check is sufficient.
As a compromise between speed and safety, implement these checks
as Asserts when dealing with small chunks but plain test-and-elogs
when dealing with large (external) chunks. The latter case should
not be too performance-critical, but the former case probably is.
In slab.c, all chunks are small; but nonetheless use a plain test
in SlabRealloc, because that is certainly not performance-critical,
indeed we should be suspicious that it's being called in error.
In aset.c, additionally add some assertions that the "value" field
of the chunk header is within the small range allowed for freelist
indexes. Without that, we might find ourselves trying to wipe
most of memory when CLOBBER_FREED_MEMORY is enabled, or scribbling
on a "freelist header" that's far away from the context object.
Eventually, field experience might show us that it's smarter for
these tests to be active always, but for now we'll try to get
away with just having them as assertions.
While at it, also be more uniform about asserting that context
objects passed as parameters are of the type we expect. Some
places missed that altogether, and slab.c was for no very good
reason doing it differently from the other allocators.
Discussion: https://postgr.es/m/3578387.1665244345@sss.pgh.pa.us
Traditionally, in MEMORY_CONTEXT_CHECKING builds, we only ever marked a
sentinel byte just beyond the requested size if there happened to be
enough space on the chunk to do so. For Slab and Generation context
types, we only rounded the size of the chunk up to the next maxalign
boundary, so it was often not that likely that those would ever have space
for the sentinel given that the majority of allocation requests are going
to be for sizes which are maxaligned. For AllocSet, it was a little
different as smaller allocations are rounded up to the next power-of-2
value rather than the next maxalign boundary, so we're a bit more likely
to have space for the sentinel byte, especially when we get away from tiny
sized allocations such as 8 or 16 bytes.
Here we make more of an effort to allow space so that there is enough room
for the sentinel byte in more cases. This makes it more likely that we'll
detect when buggy code accidentally writes beyond the end of any of its
memory allocations.
Each of the 3 MemoryContext types has been changed as follows:
The Slab allocator will now always set a sentinel byte. Both the current
usages of this MemoryContext type happen to use chunk sizes which were on
the maxalign boundary, so these never used sentinel bytes previously.
For the Generation allocator, we now always ensure there's enough space in
the allocation for a sentinel byte.
For AllocSet, this commit makes an adjustment for allocation sizes which
are greater than allocChunkLimit. We now ensure there is always space for
a sentinel byte. We don't alter the sentinel behavior for request sizes
<= allocChunkLimit. Making way for the sentinel byte for power-of-2
request sizes would require doubling up to the next power of 2. Some
analysis done on the request sizes made during installcheck shows that a
fairly large portion of allocation requests are for power-of-2 sizes. The
amount of additional memory for the sentinel there seems prohibitive, so
we do nothing for those here.
Author: David Rowley
Discussion: https://postgr.es/m/3478405.1661824539@sss.pgh.pa.us
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
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
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.
Nobody really uses this stuff, especially not since we created
valgrind-based infrastructure that does the same thing better.
It is thus unsurprising that the generation.c and slab.c versions
were actually broken. Rather than fix 'em, let's just remove 'em.
Alexander Lakhin
Discussion: https://postgr.es/m/8936216c-3492-3f6e-634b-d638fddc5f91@gmail.com
An AllocSet doubles the size of allocated blocks (up to maxBlockSize),
which means that the current block can represent half of the total
allocated space for the memory context. But the free space in the
current block may never have been touched, so don't count the
untouched memory as allocated for the purposes of
MemoryContextMemAllocated().
Discussion: https://postgr.es/m/ec63d70b668818255486a83ffadc3aec492c1f57.camel@j-davis.com
Use __builtin_clz() where available. Where it isn't, we can still win
a little by using the pg_leftmost_one_pos[] lookup table instead of
having a private table.
Also drop the initial right shift by ALLOC_MINBITS in favor of
subtracting ALLOC_MINBITS from the leftmost-one-pos result. This
is a win because the compiler can fold that adjustment into other
constants it'd have to add anyway, making the shift-removal free.
Also, we can explain this coding as an unrolled form of
pg_leftmost_one_pos32(), even though that's a bit ahistorical
since it long predates pg_bitutils.h.
John Naylor, with some cosmetic adjustments by me
Discussion: https://postgr.es/m/CACPNZCuNUGMxjK7WTn_=WZnRbfASDdBxmjsVf2+m9MdmeNw_sg@mail.gmail.com
Commit 5dd7fc1519 added block-level memory accounting, but used int64 variable to
track the amount of allocated memory. That is incorrect, because we have Size for
exactly these purposes, but it was mostly harmless until c477f3e449 which changed
how we handle with repalloc() when downsizing the chunk. Previously we've ignored
these cases and just kept using the original chunk, but now we need to update the
accounting, and the code was doing this:
context->mem_allocated += blksize - oldblksize;
Both blksize and oldblksize are Size (so unsigned) which means the subtraction
underflows, producing a very high positive value. On 64-bit platforms (where Size
has the same size as mem_alllocated) this happens to work because the result wraps
to the right value, but on (some) 32-bit platforms this fails.
This fixes two things - it changes mem_allocated (and related variables) to Size,
and it splits the update to two separate steps, to prevent any underflows.
Discussion: https://www.postgresql.org/message-id/15151.1570163761%40sss.pgh.pa.us
Up to now, if you resized a large (>8K) palloc chunk down to a smaller
size, aset.c made no attempt to return any space to the malloc pool.
That's unpleasant if a really large allocation is resized to a
significantly smaller size. I think no such cases existed when this
code was designed, and I'm not sure whether they're common even yet,
but an upcoming fix to encoding conversion will certainly create such
cases. Therefore, fix AllocSetRealloc so that it gives realloc()
a chance to do something with the block. This doesn't noticeably
increase complexity, we mostly just have to change the order in which
the cases are considered.
Back-patch to all supported branches.
Discussion: https://postgr.es/m/20190816181418.GA898@alvherre.pgsql
Discussion: https://postgr.es/m/3614.1569359690@sss.pgh.pa.us
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
We can allow this macro to accept either abbreviated or non-abbreviated
allocation parameters by making use of __VA_ARGS__. As noted by Andres
Freund, it's unlikely that any compiler would have __builtin_constant_p
but not __VA_ARGS__, so this gives up little or no error checking, and
it avoids a minor but annoying API break for extensions.
With this change, there is no reason for anybody to call
AllocSetContextCreateExtended directly, so in HEAD I renamed it to
AllocSetContextCreateInternal. It's probably too late for an ABI
break like that in 11, though.
Discussion: https://postgr.es/m/20181012170355.bhxi273skjt6sag4@alap3.anarazel.de
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
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
Revise aset.c so that all the "private" fields of chunk headers are
marked NOACCESS when outside the module, improving on the previous
coding which protected only requested_size. Fix a couple of corner
case bugs, such as failing to re-protect the header during a failure
exit from AllocSetRealloc, and wrong padding-size calculation for an
oversize allocation request.
Apply the same design to generation.c, and also fix several bugs therein
that I found by dint of hacking the code to use generation.c as the
standard allocator and then running the core regression tests with it.
Notably, we have to track the actual size of each block, else the
wipe_mem call in GenerationReset clears the wrong amount of memory for
an oversize-chunk block; and GenerationCheck needs a way of identifying
freed chunks that isn't fooled by palloc(0). I chose to fix the latter
by resetting the context pointer to NULL in a freed chunk, roughly like
what happens in a freed aset.c chunk.
Discussion: https://postgr.es/m/E1eHa4J-0006hI-Q8@gemulon.postgresql.org
Add commentary about what we're doing and why. Apply the method used for
padding in GenerationChunk to AllocChunkData, replacing the rather ad-hoc
solution used in commit 7e3aa03b4. Reorder fields in GenerationChunk so
that the padding calculation will work even if sizeof(size_t) is different
from sizeof(void *) --- likely that will never happen, but we don't need
the assumption if we do it like this. Improve static assertions about
alignment.
In passing, fix a couple of oversights in the "large chunk" path in
GenerationAlloc().
Discussion: https://postgr.es/m/E1eHa4J-0006hI-Q8@gemulon.postgresql.org
Don't move parenthesized lines to the left, even if that means they
flow past the right margin.
By default, BSD indent lines up statement continuation lines that are
within parentheses so that they start just to the right of the preceding
left parenthesis. However, traditionally, if that resulted in the
continuation line extending to the right of the desired right margin,
then indent would push it left just far enough to not overrun the margin,
if it could do so without making the continuation line start to the left of
the current statement indent. That makes for a weird mix of indentations
unless one has been completely rigid about never violating the 80-column
limit.
This behavior has been pretty universally panned by Postgres developers.
Hence, disable it with indent's new -lpl switch, so that parenthesized
lines are always lined up with the preceding left paren.
This patch is much less interesting than the first round of indent
changes, but also bulkier, so I thought it best to separate the effects.
Discussion: https://postgr.es/m/E1dAmxK-0006EE-1r@gemulon.postgresql.org
Discussion: https://postgr.es/m/30527.1495162840@sss.pgh.pa.us
Change pg_bsd_indent to follow upstream rules for placement of comments
to the right of code, and remove pgindent hack that caused comments
following #endif to not obey the general rule.
Commit e3860ffa4d wasn't actually using
the published version of pg_bsd_indent, but a hacked-up version that
tried to minimize the amount of movement of comments to the right of
code. The situation of interest is where such a comment has to be
moved to the right of its default placement at column 33 because there's
code there. BSD indent has always moved right in units of tab stops
in such cases --- but in the previous incarnation, indent was working
in 8-space tab stops, while now it knows we use 4-space tabs. So the
net result is that in about half the cases, such comments are placed
one tab stop left of before. This is better all around: it leaves
more room on the line for comment text, and it means that in such
cases the comment uniformly starts at the next 4-space tab stop after
the code, rather than sometimes one and sometimes two tabs after.
Also, ensure that comments following #endif are indented the same
as comments following other preprocessor commands such as #else.
That inconsistency turns out to have been self-inflicted damage
from a poorly-thought-through post-indent "fixup" in pgindent.
This patch is much less interesting than the first round of indent
changes, but also bulkier, so I thought it best to separate the effects.
Discussion: https://postgr.es/m/E1dAmxK-0006EE-1r@gemulon.postgresql.org
Discussion: https://postgr.es/m/30527.1495162840@sss.pgh.pa.us
The new indent version includes numerous fixes thanks to Piotr Stefaniak.
The main changes visible in this commit are:
* Nicer formatting of function-pointer declarations.
* No longer unexpectedly removes spaces in expressions using casts,
sizeof, or offsetof.
* No longer wants to add a space in "struct structname *varname", as
well as some similar cases for const- or volatile-qualified pointers.
* Declarations using PG_USED_FOR_ASSERTS_ONLY are formatted more nicely.
* Fixes bug where comments following declarations were sometimes placed
with no space separating them from the code.
* Fixes some odd decisions for comments following case labels.
* Fixes some cases where comments following code were indented to less
than the expected column 33.
On the less good side, it now tends to put more whitespace around typedef
names that are not listed in typedefs.list. This might encourage us to
put more effort into typedef name collection; it's not really a bug in
indent itself.
There are more changes coming after this round, having to do with comment
indentation and alignment of lines appearing within parentheses. I wanted
to limit the size of the diffs to something that could be reviewed without
one's eyes completely glazing over, so it seemed better to split up the
changes as much as practical.
Discussion: https://postgr.es/m/E1dAmxK-0006EE-1r@gemulon.postgresql.org
Discussion: https://postgr.es/m/30527.1495162840@sss.pgh.pa.us
Large chunks (those too large for any palloc freelist) are managed as
separate blocks. Formerly, realloc'ing or pfree'ing such a chunk required
O(N) time in a context with N blocks, since we had to traipse down the
singly-linked block list to locate the block's predecessor before we could
fix the list links. This can result in O(N^2) runtime in situations where
large numbers of such chunks are manipulated within one context. Cases
like that were not foreseen in the original design of aset.c, and indeed
didn't arise until fairly recently. But such problems can now occur in
reorderbuffer.c and in hash joining, both of which make repeated large
requests without scaling up their request size as they do so, and which
will free their requests in not-necessarily-LIFO order.
To fix, change the block list from singly-linked to doubly-linked.
This adds another 4 or 8 bytes to ALLOC_BLOCKHDRSZ, but that doesn't
seem like unacceptable overhead, since aset.c's blocks are normally
8K or more, and never less than 1K in current practice.
In passing, get rid of some redundant AllocChunkGetPointer() calls in
AllocSetRealloc (the compiler might be smart enough to optimize these
away anyway, but no need to assume that) and improve AllocSetCheck's
checking of block header fields.
Back-patch to 9.4 where reorderbuffer.c appeared. We could take this
further back, but currently there's no evidence that it would be useful.
Discussion: https://postgr.es/m/CAMkU=1x1hvue1XYrZoWk_omG0Ja5nBvTdvgrOeVkkeqs71CV8g@mail.gmail.com
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
An upcoming patch introduces a new type of memory context. To avoid
duplicating debugging infrastructure within aset.c, move useful pieces
to memdebug.[ch].
While touching aset.c, fix printf format code in AllocFree* debug
macros.
Author: Tomas Vondra
Reviewed-By: Andres Freund
Discussion: https://postgr.es/m/b3b2245c-b37a-e1e5-ebc4-857c914bc747@2ndquadrant.com
The previous behavior was to silently change them to something valid.
That obscured the bugs fixed in commit ea268cdc9, and generally seems
less useful than complaining. Unlike the previous commit, though,
we'll do this in HEAD only --- it's a bit too late to be possibly
breaking third-party code in 9.6.
Discussion: <CA+TgmobNcELVd3QmLD3tx=w7+CokRQiC4_U0txjz=WHpfdkU=w@mail.gmail.com>
I found that half a dozen (nearly 5%) of our AllocSetContextCreate calls
had typos in the context-sizing parameters. While none of these led to
especially significant problems, they did create minor inefficiencies,
and it's now clear that expecting people to copy-and-paste those calls
accurately is not a great idea. Let's reduce the risk of future errors
by introducing single macros that encapsulate the common use-cases.
Three such macros are enough to cover all but two special-purpose contexts;
those two calls can be left as-is, I think.
While this patch doesn't in itself improve matters for third-party
extensions, it doesn't break anything for them either, and they can
gradually adopt the simplified notation over time.
In passing, change TopMemoryContext to use the default allocation
parameters. Formerly it could only be extended 8K at a time. That was
probably reasonable when this code was written; but nowadays we create
many more contexts than we did then, so that it's not unusual to have a
couple hundred K in TopMemoryContext, even without considering various
dubious code that sticks other things there. There seems no good reason
not to let it use growing blocks like most other contexts.
Back-patch to 9.6, mostly because that's still close enough to HEAD that
it's easy to do so, and keeping the branches in sync can be expected to
avoid some future back-patching pain. The bugs fixed by these changes
don't seem to be significant enough to justify fixing them further back.
Discussion: <21072.1472321324@sss.pgh.pa.us>
We had a report from Stefan Kaltenbrunner of a case in which postmaster
log files overran available disk space because multiple backends spewed
enormous context stats dumps upon hitting an out-of-memory condition.
Given the lack of similar reports, this isn't a common problem, but it
still seems worth doing something about. However, we don't want to just
blindly truncate the output, because that might prevent diagnosis of OOM
problems. What seems like a workable compromise is to limit the dump to
100 child contexts per parent, and summarize the space used within any
additional child contexts. That should help because practical cases where
the dump 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, we hope.
Anyway it doesn't take much code or memory space to do this, so let's try
it like this and see how things go.
Since the summarization mechanism requires passing totals back up anyway,
I took the opportunity to add a "grand total" line to the end of the
printout.
Peter Eisentraut
David Rowley
Bruce Momjian
Tom Lane
John Naylor
Fujii Masao
Jeff Davis
Tomas Vondra
Michael Paquier
Andres Freund