Return to BSD News archive
Received: by minnie.vk1xwt.ampr.org with NNTP
id AA6511 ; Sun, 10 Jan 93 11:19:09 EST
Path: sserve!manuel.anu.edu.au!munnari.oz.au!sgiblab!sdd.hp.com!spool.mu.edu!agate!doc.ic.ac.uk!uknet!cf-cm!myrddin.isl.cf.ac.uk!paul
From: paul@isl.cf.ac.uk (Paul Richards)
Newsgroups: comp.unix.bsd
Subject: gnu mathlib pow.s bug FIX
Message-ID: <1993Jan12.214029.26276@cm.cf.ac.uk>
Date: 12 Jan 93 21:40:28 GMT
Sender: news@cm.cf.ac.uk (Network News System)
Organization: Intelligent Systems Lab, ELSYM, Universiity of Wales, College of
Cardiff.
Lines: 32
I got around to looking at the pow.s bug I found in the mathlib library
today. The fix is included below. If you use the mathlib then you need
to apply this patch and recompile your apps.
The bug was that pow() was corrupting temporary values held in the
registers ebx,esi and edi by restoring their saved values from the
wrong place on the stack.
*** pow.s.orig Tue Jan 12 21:16:40 1993
--- pow.s Tue Jan 12 21:20:05 1993
***************
*** 155,161 ****
fchs
/*/NO_APP*/
.L1:
! leal -28(%ebp),%esp
popl %ebx
popl %esi
popl %edi
--- 155,161 ----
fchs
/*/NO_APP*/
.L1:
! leal -40(%ebp),%esp
popl %ebx
popl %esi
popl %edi
--
Paul Richards, University of Wales, College Cardiff
JANET:paul@uk.ac.cf.isl Internet:paul@isl.cf.ac.uk
UUCP: paul@cf-isl.UUCP or ...!uunet!mcsun!uknet!cf!isl!paul