Fix another longstanding problem in copy_relation_data: it was blithely

assuming that a local char[] array would be aligned on at least a word
boundary.  There are architectures on which that is pretty much guaranteed to
NOT be the case ... and those arches also don't like non-aligned memory
accesses, meaning that log_newpage() would crash if it ever got invoked.
Even on Intel-ish machines there's a potential for a large performance penalty
from doing I/O to an inadequately aligned buffer.  So palloc it instead.

Backpatch to 8.0 --- 7.4 doesn't have this code.
This commit is contained in:
Tom Lane 2010-07-29 19:23:28 +00:00
parent 19600de638
commit c04b3f45b9

View File

@ -8,7 +8,7 @@
*
*
* IDENTIFICATION
* $PostgreSQL: pgsql/src/backend/commands/tablecmds.c,v 1.332.2.1 2010/07/23 20:04:09 petere Exp $
* $PostgreSQL: pgsql/src/backend/commands/tablecmds.c,v 1.332.2.2 2010/07/29 19:23:28 tgl Exp $
*
*-------------------------------------------------------------------------
*/
@ -7051,11 +7051,20 @@ static void
copy_relation_data(SMgrRelation src, SMgrRelation dst,
ForkNumber forkNum, bool istemp)
{
char *buf;
Page page;
bool use_wal;
BlockNumber nblocks;
BlockNumber blkno;
char buf[BLCKSZ];
Page page = (Page) buf;
/*
* palloc the buffer so that it's MAXALIGN'd. If it were just a local
* char[] array, the compiler might align it on any byte boundary, which
* can seriously hurt transfer speed to and from the kernel; not to
* mention possibly making log_newpage's accesses to the page header fail.
*/
buf = (char *) palloc(BLCKSZ);
page = (Page) buf;
/*
* We need to log the copied data in WAL iff WAL archiving/streaming is
@ -7084,6 +7093,8 @@ copy_relation_data(SMgrRelation src, SMgrRelation dst,
smgrextend(dst, forkNum, blkno, buf, true);
}
pfree(buf);
/*
* If the rel isn't temp, we must fsync it down to disk before it's safe
* to commit the transaction. (For a temp rel we don't care since the rel