rdelaage
/
postgresql
mirror of https://git.postgresql.org/git/postgresql.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Repair bugs in GiST page splitting code for multi-column indexes. 

When considering a non-last column in a multi-column GiST index,
gistsplit.c tries to improve on the split chosen by the opclass-specific
pickSplit function by considering penalties for the next column.  However,
there were two bugs in this code: it failed to recompute the union keys for
the leftmost index columns, even though these might well change after
reassigning tuples; and it included the old union keys in the recomputation
for the columns it did recompute, so that those keys couldn't get smaller
even if they should.  The first problem could result in an invalid index
in which searches wouldn't find index entries that are in fact present;
the second would make the index less efficient to search.

Both of these errors were caused by misuse of gistMakeUnionItVec, whose
API was designed in a way that just begged such errors to be made.  There
is no situation in which it's safe or useful to compute the union keys for
a subset of the index columns, and there is no caller that wants any
previous union keys to be included in the computation; so the undocumented
choice to treat the union keys as in/out rather than pure output parameters
is a waste of code as well as being dangerous.

Hence, rather than just making a minimal patch, I've changed the API of
gistMakeUnionItVec to remove the "startkey" parameter (it now always
processes all index columns) and treat the attr/isnull arrays as purely
output parameters.

In passing, also get rid of a couple of unnecessary and dangerous uses
of static variables in gistutil.c.  It's remarkable that the one in
gistMakeUnionKey hasn't given us portability troubles before now, because
in addition to posing a re-entrancy hazard, it was unsafely assuming that
a static char[] array would have at least Datum alignment.

Per investigation of a trouble report from Tomas Vondra.  (There are also
some bugs in contrib/btree_gist to be fixed, but that seems like material
for a separate patch.)  Back-patch to all supported branches.
This commit is contained in:
Tom Lane 2013-02-07 17:44:02 -05:00
parent c5aad8dc14
commit 166d534fcd
3 changed files with 46 additions and 50 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

39
src/backend/access/gist/gistsplit.c
View File

@ -19,21 +19,21 @@
typedef struct
{
Datum *attr;
int len;
OffsetNumber *entries;
int len;
Datum *attr;
bool *isnull;
bool *equiv;
} GistSplitUnion;
/*
* Forms unions of subkeys after page split, but
* uses only tuples that aren't in groups of equivalent tuples
* Form unions of subkeys after a page split, ignoring any tuples
* that are marked in gsvp->equiv[]
*/
static void
gistunionsubkeyvec(GISTSTATE *giststate, IndexTuple *itvec,
GistSplitUnion *gsvp, int startkey)
GistSplitUnion *gsvp)
{
IndexTuple *cleanedItVec;
int i,
@ -49,35 +49,36 @@ gistunionsubkeyvec(GISTSTATE *giststate, IndexTuple *itvec,
cleanedItVec[cleanedLen++] = itvec[gsvp->entries[i] - 1];
}
gistMakeUnionItVec(giststate, cleanedItVec, cleanedLen, startkey,
gistMakeUnionItVec(giststate, cleanedItVec, cleanedLen,
gsvp->attr, gsvp->isnull);
pfree(cleanedItVec);
}
/*
* unions subkeys for after user picksplit over attno-1 column
* Recompute unions of subkeys after a page split, ignoring any tuples
* that are marked in spl->spl_equiv[]
*/
static void
gistunionsubkey(GISTSTATE *giststate, IndexTuple *itvec, GistSplitVector *spl, int attno)
gistunionsubkey(GISTSTATE *giststate, IndexTuple *itvec, GistSplitVector *spl)
{
GistSplitUnion gsvp;
gsvp.equiv = spl->spl_equiv;
gsvp.attr = spl->spl_lattr;
gsvp.len = spl->splitVector.spl_nleft;
gsvp.entries = spl->splitVector.spl_left;
gsvp.len = spl->splitVector.spl_nleft;
gsvp.attr = spl->spl_lattr;
gsvp.isnull = spl->spl_lisnull;
gistunionsubkeyvec(giststate, itvec, &gsvp, attno);
gistunionsubkeyvec(giststate, itvec, &gsvp);
gsvp.attr = spl->spl_rattr;
gsvp.len = spl->splitVector.spl_nright;
gsvp.entries = spl->splitVector.spl_right;
gsvp.len = spl->splitVector.spl_nright;
gsvp.attr = spl->spl_rattr;
gsvp.isnull = spl->spl_risnull;
gistunionsubkeyvec(giststate, itvec, &gsvp, attno);
gistunionsubkeyvec(giststate, itvec, &gsvp);
}
/*
@ -443,14 +444,14 @@ gistUserPicksplit(Relation r, GistEntryVector *entryvec, int attno, GistSplitVec
*/
if (LenEquiv == 0)
{
gistunionsubkey(giststate, itup, v, attno + 1);
gistunionsubkey(giststate, itup, v);
}
else
{
cleanupOffsets(sv->spl_left, &sv->spl_nleft, v->spl_equiv, &LenEquiv);
cleanupOffsets(sv->spl_right, &sv->spl_nright, v->spl_equiv, &LenEquiv);
gistunionsubkey(giststate, itup, v, attno + 1);
gistunionsubkey(giststate, itup, v);
if (LenEquiv == 1)
{
/*
@ -471,7 +472,7 @@ gistUserPicksplit(Relation r, GistEntryVector *entryvec, int attno, GistSplitVec
* performance, because it's very rarely
*/
v->spl_equiv = NULL;
gistunionsubkey(giststate, itup, v, attno + 1);
gistunionsubkey(giststate, itup, v);
return false;
}
@ -562,7 +563,7 @@ gistSplitByKey(Relation r, Page page, IndexTuple *itup, int len, GISTSTATE *gist
v->splitVector.spl_left[v->splitVector.spl_nleft++] = i;
v->spl_equiv = NULL;
gistunionsubkey(giststate, itup, v, attno);
gistunionsubkey(giststate, itup, v);
}
else
{
@ -617,7 +618,7 @@ gistSplitByKey(Relation r, Page page, IndexTuple *itup, int len, GISTSTATE *gist
v->splitVector = backupSplit;
/* reunion left and right datums */
gistunionsubkey(giststate, itup, v, attno);
gistunionsubkey(giststate, itup, v);
}
}
}

55
src/backend/access/gist/gistutil.c
View File

@ -21,12 +21,6 @@
#include "storage/lmgr.h"
#include "utils/builtins.h"
/*
* static *S used for temporary storage (saves stack and palloc() call)
*/
static Datum attrS[INDEX_MAX_KEYS];
static bool isnullS[INDEX_MAX_KEYS];
/*
* Write itup vector to page, has no control of free space.
@ -148,12 +142,12 @@ gistfillitupvec(IndexTuple *vec, int veclen, int *memlen)
}
/*
* Make unions of keys in IndexTuple vector.
* Make unions of keys in IndexTuple vector (one union datum per index column).
* Union Datums are returned into the attr/isnull arrays.
* Resulting Datums aren't compressed.
*/
void
gistMakeUnionItVec(GISTSTATE *giststate, IndexTuple *itvec, int len, int startkey,
gistMakeUnionItVec(GISTSTATE *giststate, IndexTuple *itvec, int len,
Datum *attr, bool *isnull)
{
int i;
@ -162,19 +156,12 @@ gistMakeUnionItVec(GISTSTATE *giststate, IndexTuple *itvec, int len, int startke
evec = (GistEntryVector *) palloc((len + 2) * sizeof(GISTENTRY) + GEVHDRSZ);
for (i = startkey; i < giststate->tupdesc->natts; i++)
for (i = 0; i < giststate->tupdesc->natts; i++)
{
int j;
/* Collect non-null datums for this column */
evec->n = 0;
if (!isnull[i])
{
gistentryinit(evec->vector[evec->n], attr[i],
NULL, NULL, (OffsetNumber) 0,
FALSE);
evec->n++;
}
for (j = 0; j < len; j++)
{
Datum datum;
@ -192,7 +179,7 @@ gistMakeUnionItVec(GISTSTATE *giststate, IndexTuple *itvec, int len, int startke
evec->n++;
}
/* If this tuple vector was all NULLs, the union is NULL */
/* If this column was all NULLs, the union is NULL */
if (evec->n == 0)
{
attr[i] = (Datum) 0;
@ -202,6 +189,7 @@ gistMakeUnionItVec(GISTSTATE *giststate, IndexTuple *itvec, int len, int startke
{
if (evec->n == 1)
{
/* unionFn may expect at least two inputs */
evec->n = 2;
evec->vector[1] = evec->vector[0];
}
@ -224,11 +212,12 @@ gistMakeUnionItVec(GISTSTATE *giststate, IndexTuple *itvec, int len, int startke
IndexTuple
gistunion(Relation r, IndexTuple *itvec, int len, GISTSTATE *giststate)
{
memset(isnullS, TRUE, sizeof(bool) * giststate->tupdesc->natts);
Datum attr[INDEX_MAX_KEYS];
bool isnull[INDEX_MAX_KEYS];
gistMakeUnionItVec(giststate, itvec, len, 0, attrS, isnullS);
gistMakeUnionItVec(giststate, itvec, len, attr, isnull);
return gistFormTuple(giststate, r, attrS, isnullS, false);
return gistFormTuple(giststate, r, attr, isnull, false);
}
/*
@ -240,12 +229,15 @@ gistMakeUnionKey(GISTSTATE *giststate, int attno,
GISTENTRY *entry2, bool isnull2,
Datum *dst, bool *dstisnull)
{
/* we need a GistEntryVector with room for exactly 2 elements */
union
{
GistEntryVector gev;
char padding[2 * sizeof(GISTENTRY) + GEVHDRSZ];
} storage;
GistEntryVector *evec = &storage.gev;
int dstsize;
static char storage[2 * sizeof(GISTENTRY) + GEVHDRSZ];
GistEntryVector *evec = (GistEntryVector *) storage;
evec->n = 2;
if (isnull1 && isnull2)
@ -321,6 +313,8 @@ gistgetadjusted(Relation r, IndexTuple oldtup, IndexTuple addtup, GISTSTATE *gis
addentries[INDEX_MAX_KEYS];
bool oldisnull[INDEX_MAX_KEYS],
addisnull[INDEX_MAX_KEYS];
Datum attr[INDEX_MAX_KEYS];
bool isnull[INDEX_MAX_KEYS];
IndexTuple newtup = NULL;
int i;
@ -335,19 +329,20 @@ gistgetadjusted(Relation r, IndexTuple oldtup, IndexTuple addtup, GISTSTATE *gis
gistMakeUnionKey(giststate, i,
oldentries + i, oldisnull[i],
addentries + i, addisnull[i],
attrS + i, isnullS + i);
attr + i, isnull + i);
if (neednew)
/* we already need new key, so we can skip check */
continue;
if (isnullS[i])
if (isnull[i])
/* union of key may be NULL if and only if both keys are NULL */
continue;
if (!addisnull[i])
{
if (oldisnull[i] || gistKeyIsEQ(giststate, i, oldentries[i].key, attrS[i]) == false)
if (oldisnull[i] ||
!gistKeyIsEQ(giststate, i, oldentries[i].key, attr[i]))
neednew = true;
}
}
@ -355,7 +350,7 @@ gistgetadjusted(Relation r, IndexTuple oldtup, IndexTuple addtup, GISTSTATE *gis
if (neednew)
{
/* need to update key */
newtup = gistFormTuple(giststate, r, attrS, isnullS, false);
newtup = gistFormTuple(giststate, r, attr, isnull, false);
newtup->t_tid = oldtup->t_tid;
}

2
src/include/access/gist_private.h
View File

@ -500,7 +500,7 @@ extern void gistdentryinit(GISTSTATE *giststate, int nkey, GISTENTRY *e,
extern float gistpenalty(GISTSTATE *giststate, int attno,
GISTENTRY *key1, bool isNull1,
GISTENTRY *key2, bool isNull2);
extern void gistMakeUnionItVec(GISTSTATE *giststate, IndexTuple *itvec, int len, int startkey,
extern void gistMakeUnionItVec(GISTSTATE *giststate, IndexTuple *itvec, int len,
Datum *attr, bool *isnull);
extern bool gistKeyIsEQ(GISTSTATE *giststate, int attno, Datum a, Datum b);
extern void gistDeCompressAtt(GISTSTATE *giststate, Relation r, IndexTuple tuple, Page p,