This provides a speedup of about 4X when NBuffers is large enough.
There is also a useful reduction in sinval traffic, since we
only do CacheInvalidateSmgr() once not once per fork.
Simon Riggs, reviewed and somewhat revised by Tom Lane
for (i = 0; i < ndelrels; i++)
{
SMgrRelation srel = smgropen(delrels[i], InvalidBackendId);
- ForkNumber fork;
- for (fork = 0; fork <= MAX_FORKNUM; fork++)
- {
- smgrdounlink(srel, fork, false);
- }
+ smgrdounlink(srel, false);
smgrclose(srel);
}
ForkNumber fork;
for (fork = 0; fork <= MAX_FORKNUM; fork++)
- {
XLogDropRelation(xnodes[i], fork);
- smgrdounlink(srel, fork, true);
- }
+ smgrdounlink(srel, true);
smgrclose(srel);
}
ForkNumber fork;
for (fork = 0; fork <= MAX_FORKNUM; fork++)
- {
XLogDropRelation(xlrec->xnodes[i], fork);
- smgrdounlink(srel, fork, true);
- }
+ smgrdounlink(srel, true);
smgrclose(srel);
}
}
if (pending->atCommit == isCommit)
{
SMgrRelation srel;
- int i;
srel = smgropen(pending->relnode, pending->backend);
- for (i = 0; i <= MAX_FORKNUM; i++)
- {
- smgrdounlink(srel, i, false);
- }
+ smgrdounlink(srel, false);
smgrclose(srel);
}
/* must explicitly free the list entry */
* DropRelFileNodeBuffers
*
* This function removes from the buffer pool all the pages of the
- * specified relation that have block numbers >= firstDelBlock.
+ * specified relation fork that have block numbers >= firstDelBlock.
* (In particular, with firstDelBlock = 0, all pages are removed.)
* Dirty pages are simply dropped, without bothering to write them
* out first. Therefore, this is NOT rollback-able, and so should be
}
}
+/* ---------------------------------------------------------------------
+ * DropRelFileNodeAllBuffers
+ *
+ * This function removes from the buffer pool all the pages of all
+ * forks of the specified relation. It's equivalent to calling
+ * DropRelFileNodeBuffers once per fork with firstDelBlock = 0.
+ * --------------------------------------------------------------------
+ */
+void
+DropRelFileNodeAllBuffers(RelFileNodeBackend rnode)
+{
+ int i;
+
+ /* If it's a local relation, it's localbuf.c's problem. */
+ if (rnode.backend != InvalidBackendId)
+ {
+ if (rnode.backend == MyBackendId)
+ DropRelFileNodeAllLocalBuffers(rnode.node);
+ return;
+ }
+
+ for (i = 0; i < NBuffers; i++)
+ {
+ volatile BufferDesc *bufHdr = &BufferDescriptors[i];
+
+ /*
+ * As in DropRelFileNodeBuffers, an unlocked precheck should be safe
+ * and saves some cycles.
+ */
+ if (!RelFileNodeEquals(bufHdr->tag.rnode, rnode.node))
+ continue;
+
+ LockBufHdr(bufHdr);
+ if (RelFileNodeEquals(bufHdr->tag.rnode, rnode.node))
+ InvalidateBuffer(bufHdr); /* releases spinlock */
+ else
+ UnlockBufHdr(bufHdr);
+ }
+}
+
/* ---------------------------------------------------------------------
* DropDatabaseBuffers
*
}
}
+/*
+ * DropRelFileNodeAllLocalBuffers
+ * This function removes from the buffer pool all pages of all forks
+ * of the specified relation.
+ *
+ * See DropRelFileNodeAllBuffers in bufmgr.c for more notes.
+ */
+void
+DropRelFileNodeAllLocalBuffers(RelFileNode rnode)
+{
+ int i;
+
+ for (i = 0; i < NLocBuffer; i++)
+ {
+ BufferDesc *bufHdr = &LocalBufferDescriptors[i];
+ LocalBufferLookupEnt *hresult;
+
+ if ((bufHdr->flags & BM_TAG_VALID) &&
+ RelFileNodeEquals(bufHdr->tag.rnode, rnode))
+ {
+ if (LocalRefCount[i] != 0)
+ elog(ERROR, "block %u of %s is still referenced (local %u)",
+ bufHdr->tag.blockNum,
+ relpathbackend(bufHdr->tag.rnode, MyBackendId,
+ bufHdr->tag.forkNum),
+ LocalRefCount[i]);
+ /* Remove entry from hashtable */
+ hresult = (LocalBufferLookupEnt *)
+ hash_search(LocalBufHash, (void *) &bufHdr->tag,
+ HASH_REMOVE, NULL);
+ if (!hresult) /* shouldn't happen */
+ elog(ERROR, "local buffer hash table corrupted");
+ /* Mark buffer invalid */
+ CLEAR_BUFFERTAG(bufHdr->tag);
+ bufHdr->flags = 0;
+ bufHdr->usage_count = 0;
+ }
+ }
+}
+
/*
* InitLocalBuffers -
* init the local buffer cache. Since most queries (esp. multi-user ones)
}
/*
- * smgrdounlink() -- Immediately unlink a relation.
+ * smgrdounlink() -- Immediately unlink all forks of a relation.
+ *
+ * All forks of the relation are removed from the store. This should
+ * not be used during transactional operations, since it can't be undone.
+ *
+ * If isRedo is true, it is okay for the underlying file(s) to be gone
+ * already.
+ *
+ * This is equivalent to calling smgrdounlinkfork for each fork, but
+ * it's significantly quicker so should be preferred when possible.
+ */
+void
+smgrdounlink(SMgrRelation reln, bool isRedo)
+{
+ RelFileNodeBackend rnode = reln->smgr_rnode;
+ int which = reln->smgr_which;
+ ForkNumber forknum;
+
+ /* Close the forks at smgr level */
+ for (forknum = 0; forknum <= MAX_FORKNUM; forknum++)
+ (*(smgrsw[which].smgr_close)) (reln, forknum);
+
+ /*
+ * Get rid of any remaining buffers for the relation. bufmgr will just
+ * drop them without bothering to write the contents.
+ */
+ DropRelFileNodeAllBuffers(rnode);
+
+ /*
+ * It'd be nice to tell the stats collector to forget it immediately, too.
+ * But we can't because we don't know the OID (and in cases involving
+ * relfilenode swaps, it's not always clear which table OID to forget,
+ * anyway).
+ */
+
+ /*
+ * Send a shared-inval message to force other backends to close any
+ * dangling smgr references they may have for this rel. We should do this
+ * before starting the actual unlinking, in case we fail partway through
+ * that step. Note that the sinval message will eventually come back to
+ * this backend, too, and thereby provide a backstop that we closed our
+ * own smgr rel.
+ */
+ CacheInvalidateSmgr(rnode);
+
+ /*
+ * Delete the physical file(s).
+ *
+ * Note: smgr_unlink must treat deletion failure as a WARNING, not an
+ * ERROR, because we've already decided to commit or abort the current
+ * xact.
+ */
+ for (forknum = 0; forknum <= MAX_FORKNUM; forknum++)
+ (*(smgrsw[which].smgr_unlink)) (rnode, forknum, isRedo);
+}
+
+/*
+ * smgrdounlinkfork() -- Immediately unlink one fork of a relation.
*
* The specified fork of the relation is removed from the store. This
* should not be used during transactional operations, since it can't be
* already.
*/
void
-smgrdounlink(SMgrRelation reln, ForkNumber forknum, bool isRedo)
+smgrdounlinkfork(SMgrRelation reln, ForkNumber forknum, bool isRedo)
{
RelFileNodeBackend rnode = reln->smgr_rnode;
int which = reln->smgr_which;
- /* Close the fork */
+ /* Close the fork at smgr level */
(*(smgrsw[which].smgr_close)) (reln, forknum);
/*
- * Get rid of any remaining buffers for the relation. bufmgr will just
+ * Get rid of any remaining buffers for the fork. bufmgr will just
* drop them without bothering to write the contents.
*/
DropRelFileNodeBuffers(rnode, forknum, 0);
extern void MarkLocalBufferDirty(Buffer buffer);
extern void DropRelFileNodeLocalBuffers(RelFileNode rnode, ForkNumber forkNum,
BlockNumber firstDelBlock);
+extern void DropRelFileNodeAllLocalBuffers(RelFileNode rnode);
extern void AtEOXact_LocalBuffers(bool isCommit);
#endif /* BUFMGR_INTERNALS_H */
extern void FlushDatabaseBuffers(Oid dbid);
extern void DropRelFileNodeBuffers(RelFileNodeBackend rnode,
ForkNumber forkNum, BlockNumber firstDelBlock);
+extern void DropRelFileNodeAllBuffers(RelFileNodeBackend rnode);
extern void DropDatabaseBuffers(Oid dbid);
#define RelationGetNumberOfBlocks(reln) \
extern void smgrcloseall(void);
extern void smgrclosenode(RelFileNodeBackend rnode);
extern void smgrcreate(SMgrRelation reln, ForkNumber forknum, bool isRedo);
-extern void smgrdounlink(SMgrRelation reln, ForkNumber forknum,
- bool isRedo);
+extern void smgrdounlink(SMgrRelation reln, bool isRedo);
+extern void smgrdounlinkfork(SMgrRelation reln, ForkNumber forknum, bool isRedo);
extern void smgrextend(SMgrRelation reln, ForkNumber forknum,
BlockNumber blocknum, char *buffer, bool skipFsync);
extern void smgrprefetch(SMgrRelation reln, ForkNumber forknum,