cpp /usr/include/threads.h fails; modfl segfaults

[Corinna Vinschen]

On Sep  1 11:28, Brian Inglis wrote:
> On 2020-08-31 13:41, Corinna Vinschen wrote:
> > On Aug 31 13:24, Brian Inglis wrote:
> >>>> The upstream patch changed only amd64/x86_64 code sequences for multiple modules
> >>>> including modfl, and left i386/x86 untouched for those modules.
> >>
> >> Just pointing out that they only modify their amd64/x86_64 code which doesn't
> >> push/pop rax/eax:
> > 
> > Where are you looking at?  As you could see from my output, I was
> > looking at the master branch of the upstream repo.
> 
> Sorry I didn't see your point there as I wasn't aware there were SF repos.
> 
> > This lengthy discussion for a minor asm snippet doesn't make any sense.
> > If you think this is wrong, send patches to cygwin-patches and explain
> > where you got it from, preferrably as a git patch from the upstream
> > repo.
> 
> Sorry for wasting your time.
> I was looking at the bug/patch content and didn't realize someone later added a
> bogus clobber on their x86 code path.

No, really, patches are super great and *I* know *you* know how it works.

Given this code is at least questionable, what about sending a patch to
upstream and a matching patch to cygwin-patches?  The upstream mailing
list is at mingw-w64-public AT lists DOT sourceforge DOT net.


Thanks,
Corinna

-- 
Corinna Vinschen
Cygwin Maintainer