This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: Misaligned stack on 32-bit s390?
- From: Rich Felker <dalias at libc dot org>
- To: Andreas Krebbel <krebbel at linux dot vnet dot ibm dot com>
- Cc: Carlos O'Donell <carlos at redhat dot com>, Richard Henderson <rth at twiddle dot net>, Siddhesh Poyarekar <siddhesh at redhat dot com>, GNU C Library <libc-alpha at sourceware dot org>, stli at linux dot vnet dot ibm dot com
- Date: Wed, 12 Nov 2014 16:46:26 -0500
- Subject: Re: Misaligned stack on 32-bit s390?
- Authentication-results: sourceware.org; auth=none
- References: <54619F3E dot 8080306 at redhat dot com> <5461D6CA dot 9030902 at twiddle dot net> <5461DBC4 dot 9090508 at redhat dot com> <5462231E dot 3050502 at linux dot vnet dot ibm dot com> <5462DB6B dot 8060405 at redhat dot com> <546322E9 dot 3060501 at linux dot vnet dot ibm dot com>
On Wed, Nov 12, 2014 at 10:05:45AM +0100, Andreas Krebbel wrote:
> On 11/12/2014 05:00 AM, Carlos O'Donell wrote:
> > On 11/11/2014 09:54 AM, Andreas Krebbel wrote:
> >> Would you like to continue working on it or should we try to take
> >> over? (Stefan or myself)
> >
> > We can get it done. It would be great to have you review the final
> > patch.
> >
> > Please note that the patch I posted is incomplete, it fails to
> > readjust _dl_argv which is cached by the loader and needs to be
> > changed if argv is moved. Simple fix though.
> >
> > I also don't know how long s390 lasted without this breaking
> > something. I guess aligned stacks don't really matter all that
> > much ;-)
>
> The reason probably is that the broken alignment never reaches the executable. In _start.S we
> correct the alignment again. So it really is only a problem for code executed between the argv
> adjustments done by ld.so and the entry point of the executable.
How/why does any code get executed in that interval anyway?
Rich