Oracle blows up in malloc(3c) on NCR machines

[Steinar Overbeck Cook]

We have an application which uses dynamic SQL against an Oracle
database.  The machine is a NCR Tower 32/600 with UNIX V.2 (NCR
1.03.02).

After extracting approx 480 records, the application core dumps
because of SIGSEGV (Segment violation).  This happens on whatever
table we are extracting from.

The bug is only there when we are using Oracle Release 5.1.17.4
which we have purchased from NCR.  The bug is NOT there on the
versions we have bought directly from Oracle.

The stack trace in sdb comes out thus:

	malloc: address 0x18a54
	*malloc(1032,199,147564)
	_findbuf(147564,147564,14679000)
	_filbuf(147564,0,102)
	fgets(14679000,200,147564)
	upiref(157555,1403,1403)   [upiexn.c]
	upigem(157280,157544,157280)
	sqlgem(1403,1403,14679284)
	sqlret(0,14679332,3818)
	sqlfch(151460,151464,171892)
	FETCHKol()   [DYNSEL.c:852]
	main(argc=4,argv=0xdffd58,14679404)   [DYNSEL.c:595]
	*

Does anybody have a clue as to what is wrong ? My theory is that Oracle
overwrites the internal tables of malloc(3c).

Please answer by e-mail.
-- 
Steinar Overbeck Cook, Fellesdata a.s, P.O. Box 248, 0212 OSLO 2, NORWAY
Phone : +47 2 52 80 80                            Fax   : +47 2 52 85 10
E-mail : ...!mcvax!ndosl!fdmetd!steinar  or       steinar at fdmetd.uucp
<The opinions expressed, if any, do not represent Fellesdata a.s>