Date: Wed, 11 Dec 96 18:51:02 +0000 From: dwmalone@maths.tcd.ie To: FreeBSD-gnats-submit@freebsd.org Subject: misc/2195: HAVE_FPU = yes produces buged libm Message-ID: <9612111851.aa08441@synge.maths.tcd.ie> Resent-Message-ID: <199612111900.LAA29038@freefall.freebsd.org>
next in thread | raw e-mail | index | archive | help
>Number: 2195 >Category: misc >Synopsis: HAVE_FPU = yes produces buged libm >Confidential: no >Severity: non-critical >Priority: low >Responsible: freebsd-bugs >State: open >Class: sw-bug >Submitter-Id: current-users >Arrival-Date: Wed Dec 11 11:00:01 PST 1996 >Last-Modified: >Originator: David Malone >Organization: Trinity College Dublin. >Release: FreeBSD 2.2 >Environment: Running FreeBSD-2.2 from ctm'ed sources on a intel pentium. >Description: If you compile libm with HAVE_FPU=yes then the libm code seems messed up. >How-To-Repeat: Some code I have for a ray-tracer core dumps on certain input when libm is compiled with HAVE_FPU=yes. The stack seems pretty badly trashed, as gdb produces nothing sensible. If without recompiling you drop in a non-fpu libm then everything works fine. If anyone wants the code and input which causes the problem let me know - but I wasn't able to track it down and there is a fiar bit of code in use. >Fix: Compile without HAVE_FPU=yes - maybe the option should be removed from /etc/make.conf >Audit-Trail: >Unformatted:
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?9612111851.aa08441>