__attribute__ (__cleanup__) vs. _Unwind_Resume

[Thomas Schwinge]

Hello!

On Mon, Apr 20, 2009 at 12:29:02PM -0700, Ulrich Drepper wrote:
> Thomas Schwinge wrote:
> > One possibility to fix this would be to add -lgcc_eh to libc.so-gnulib in
> > Makeconfig (I don't know yet if that is legitimate), or perhaps, instead
> > of that, we should introduce / use something like
> > nptl/sysdeps/pthread/unwind-resume.c?
> 
> The later, probably.  This is how this work on Linux.

In fact, as it is enough to have a sysdeps/mach/hurd/unwind-resume.c à la
``#include "nptl/sysdeps/pthread/unwind-resume.c"'' (plus copying the
three unwind-resume-referring lines from nptl/sysdeps/Makefile to
sysdeps/mach/hurd/Makefile), and there being nothing NPTL-specific in
unwind-resume.c (apart from the comment ``libgcc_s.so.1 must be installed
for *pthread_cancel* to work''), I wonder if it wouldn't be fine to move
that file to a generic location so that it can be picked up by every
glibc port?  Or are there cases, where this unwinding support wouldn't be
appropriate?

Additionally, couldn't this be unified with both
nptl/sysdeps/pthread/unwind-forcedunwind.c and the libgcc_s handling in
several backtrace.c files?


Regards,
 Thomas
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 191 bytes
Desc: Digital signature
URL: <http://sourceware.org/pipermail/libc-alpha/attachments/20090421/ea21073a/attachment.sig>