BUG #6086: Segmentation fault

The following bug has been logged online:

Bug reference: 6086
Logged by: Dennis
Email address: dennis(dot)noordsij(at)helsinki(dot)fi
PostgreSQL version: 9.0.4
Operating system: FreeBSD 8.2 (64 bit)
Description: Segmentation fault
Details:

For some reason a 9.0.4 server which has always run completely stable, and
without any recent changes, crashed ("WARNING: terminating connection
because of crash of another server process" ..)

Since it is not a debug build the backtrace is not that useful, but
hopefully may give some idea of where to look, or ring a bell with someone.

#0 0x0000000806a0d5dc in pthread_mutex_lock () from /lib/libthr.so.3
#1 0x0000000800c1d965 in xmlRMutexLock () from /usr/local/lib/libxml2.so.5
#2 0x0000000800c6c67d in xmlDictFree () from /usr/local/lib/libxml2.so.5
#3 0x0000000800bb1ac5 in xmlFreeParserCtxt () from
/usr/local/lib/libxml2.so.5
#4 0x0000000800bc3f19 in xmlParseBalancedChunkMemoryRecover () from
/usr/local/lib/libxml2.so.5
#5 0x0000000800bc4090 in xmlParseBalancedChunkMemory () from
/usr/local/lib/libxml2.so.5
#6 0x000000000068c6a0 in cursor_to_xml ()
#7 0x000000000068c7f7 in xml_in ()
#8 0x00000000006a3d0a in InputFunctionCall ()
#9 0x00000000006a494a in OidInputFunctionCall ()
#10 0x00000000004db6f7 in coerce_type ()
#11 0x00000000004dbcd7 in coerce_to_target_type ()
#12 0x00000000004dbd8a in coerce_to_specific_type ()
#13 0x00000000004de092 in transformWithClause ()
#14 0x00000000004de84b in transformExpr ()
#15 0x00000000004ddfad in transformWithClause ()
#16 0x00000000004de84b in transformExpr ()
#17 0x00000000004ddfad in transformWithClause ()
#18 0x00000000004de84b in transformExpr ()
#19 0x00000000004e9cb7 in transformTargetEntry ()
#20 0x00000000004ea459 in transformTargetList ()
#21 0x00000000004bff84 in transformStmt ()
#22 0x00000000005f0624 in pg_analyze_and_rewrite_params ()
#23 0x000000000055d63f in SPI_cursor_open_with_paramlist ()
#24 0x000000000055e2da in SPI_prepare_params ()
#25 0x0000000805f0d31c in exec_get_datum_type () from
/usr/local/lib/postgresql/plpgsql.so
#26 0x0000000805f0df0f in exec_get_datum_type () from
/usr/local/lib/postgresql/plpgsql.so
#27 0x0000000805f0f8b3 in exec_get_datum_type () from
/usr/local/lib/postgresql/plpgsql.so
#28 0x0000000805f117ed in exec_get_datum_type () from
/usr/local/lib/postgresql/plpgsql.so
#29 0x0000000805f1109b in exec_get_datum_type () from
/usr/local/lib/postgresql/plpgsql.so
#30 0x0000000805f136f2 in plpgsql_exec_function () from
/usr/local/lib/postgresql/plpgsql.so
#31 0x0000000805f093a3 in plpgsql_call_handler () from
/usr/local/lib/postgresql/plpgsql.so
#32 0x0000000000544f00 in GetAttributeByNum ()
#33 0x0000000000541dfa in ExecProject ()
#34 0x0000000000556293 in ExecResult ()
#35 0x000000000054196d in ExecProcNode ()
#36 0x00000000005406bf in standard_ExecutorRun ()
#37 0x00000000005f2d98 in PostgresMain ()
#38 0x00000000005f4281 in PortalRun ()
#39 0x00000000005f0ebb in pg_parse_query ()
#40 0x00000000005f18c7 in PostgresMain ()
#41 0x00000000005c1a9d in ClosePostmasterPorts ()
#42 0x00000000005c2767 in PostmasterMain ()
#43 0x000000000056cc7e in main ()

Most of what it usually does is, as can be seen in this backtrace, running
plpgsql procedures which return XML data (so I'm hesitant to blame bad
hardware at this point). It happened during a normal query that is ran very
often, everything via psycopg2. Server uses WAL log shipping (crash did not
coincide with running an archive command, no base backup in progress)

I will switch to a debug build, but I haven't seen this crash before so it
might never produce a useful backtrace.

Thanks for any suggestions