#12873: Clear MMX state in i686 longjmp()
----------------------------------+-----------------------------------------
Reporter: jdemeyer | Owner: jdemeyer
Type: defect | Status: new
Priority: major | Milestone: sage-5.0
Component: c_lib | Resolution:
Keywords: | Work issues:
Report Upstream: N/A | Reviewers:
Authors: Jeroen Demeyer | Merged in:
Dependencies: | Stopgaps:
----------------------------------+-----------------------------------------
Description changed by jdemeyer:
Old description:
> The libc `setjmp`/`longjmp` functions on i686 hardware do not fully
> restore the x87 FPU state, in particular when MMX instructions have been
> used. The proposed solution is to execute the `EMMS` instruction before
> `longjmp()`. For more context, see #12777 where this bug was discovered.
>
> To see the problem in action, compile [attachment:emms.c] on i686
> hardware with
> {{{
> gcc -lm -march=native -m32 -mfpmath=387 emms.c -o emms
> }}}
New description:
The libc `setjmp`/`longjmp` functions on i686 hardware do not restore the
x87 FPU state, in particular when MMX instructions have been used. On
Linux, the FPU state is restored when an interrupt occurs, but on Solaris
this is not the case.
The proposed solution is to clear the FPU state (using the `EMMS`
instruction) in the signal handler. For more context, see #12777 where
this bug was discovered.
To see the problem in action, compile [attachment:emms.c] on i686 hardware
with
{{{
gcc -lm -march=native -m32 -mfpmath=387 emms.c -o emms
}}}
--
--
Ticket URL: <http://trac.sagemath.org/sage_trac/ticket/12873#comment:5>
Sage <http://www.sagemath.org>
Sage: Creating a Viable Open Source Alternative to Magma, Maple, Mathematica,
and MATLAB
--
You received this message because you are subscribed to the Google Groups
"sage-trac" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/sage-trac?hl=en.
