Hello,

I was reading GiST core codes when I found an XLogInsert()
call that can produce a corrupted WAL record.

== Summary ==
There is an execution path that produces a WAL record whose
xl_info indicates XLOG_GIST_PAGE_UPDATE while the record
actually contains a gistxlogPageSplit structure.

== Details ==
(Line numbers are for HEAD as of Wed Dec 23 19:42:15 2009 +0000.)

The problematic XLogInsert() call is on gistxlog.c, line 770:
 recptr = XLogInsert(RM_GIST_ID, XLOG_GIST_PAGE_UPDATE, rdata);
where the last argument rdata has a pointer assigned either
on line 741 or on line 752.

When rdata comes from formSplitRdata() at line 741,
rdata contains a reference to a gistxlogPageSplit structure.
This is inconsistent with the second argument XLOG_GIST_PAGE_UPDATE.

== Importance ==
I think this poses possible data loss under multiple consecutive crashes.

== Fix ==
I include a simple patch (for HEAD as of the datetime above)
that, I suppose, prevents the corrupt WAL production.
I would be glad if you liked it.
/* I'm including the patch instead of attaching it because
   it seems the spam filter on pgsql-hackers is hard on me, sigh. */

Please note that the execution path exists at least in current HEAD,
REL8_2_STABLE and the branches in between.

Sincerely,

Yoichi Hirai
  Dept. of Computer Science, The University of Tokyo


diff --git a/src/backend/access/gist/gistxlog.c 
b/src/backend/access/gist/gistxlog.c
index adf27ff..74e72f0 100644
--- a/src/backend/access/gist/gistxlog.c
+++ b/src/backend/access/gist/gistxlog.c
@@ -654,6 +654,7 @@ gistContinueInsert(gistIncompleteInsert *insert)
                        XLogRecPtr      recptr;
                        Buffer          tempbuffer = InvalidBuffer;
                        int                     ntodelete = 0;
+                       uint8           info;
 
                        numbuffer = 1;
                        buffers[0] = ReadBuffer(index, insert->path[i]);
@@ -738,6 +739,7 @@ gistContinueInsert(gistIncompleteInsert *insert)
                                for (j = 0; j < ntodelete; j++)
                                        PageIndexTupleDelete(pages[0], 
todelete[j]);
 
+                               info = XLOG_GIST_PAGE_SPLIT;
                                rdata = formSplitRdata(index->rd_node, 
insert->path[i],
                                                                           
false, &(insert->key),
                                                                         
gistMakePageLayout(buffers, numbuffer));
@@ -751,6 +753,7 @@ gistContinueInsert(gistIncompleteInsert *insert)
                                        PageIndexTupleDelete(pages[0], 
todelete[j]);
                                gistfillbuffer(pages[0], itup, lenitup, 
InvalidOffsetNumber);
 
+                               info = XLOG_GIST_PAGE_UPDATE;
                                rdata = formUpdateRdata(index->rd_node, 
buffers[0],
                                                                                
todelete, ntodelete,
                                                                                
itup, lenitup, &(insert->key));
@@ -767,7 +770,7 @@ gistContinueInsert(gistIncompleteInsert *insert)
                                GistPageGetOpaque(pages[j])->rightlink = 
InvalidBlockNumber;
                                MarkBufferDirty(buffers[j]);
                        }
-                       recptr = XLogInsert(RM_GIST_ID, XLOG_GIST_PAGE_UPDATE, 
rdata);
+                       recptr = XLogInsert(RM_GIST_ID, info, rdata);
                        for (j = 0; j < numbuffer; j++)
                        {
                                PageSetLSN(pages[j], recptr);

-- 
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

Reply via email to