[PATCH] x86-64: Move nptl/rseq-access.h to nptl/64 [BZ #32543]

[Mathieu Desnoyers]

On 11-Jan-2025 10:41:10 AM, Florian Weimer wrote:
> * H. J. Lu:
> 
> > commit 494d65129ed5ae1154b75cc189bbdde5e9ecf1df
> > Author: Michael Jeanson <mjeanson@efficios.com>
> > Date:   Thu Aug 1 10:35:34 2024 -0400
> >
> >     nptl: Introduce <rseq-access.h> for RSEQ_* accessors
> >
> > added things like
> >
> >        asm volatile ("movl %%fs:%P1(%q2),%0"                                  \
> >                      : "=r" (__value)                                         \
> >                      : "i" (offsetof (struct rseq_area, member)),             \
> >                        "r" (__rseq_offset));
> >
> > But this doesn't work for x32 when __rseq_offset is negative since the
> > address is computed as
> >
> > FS + 32-bit to 64-bit zero extension of __rseq_offset
> > + offsetof (struct rseq_area, member)
> >
> > Move x86_64/nptl/rseq-access.h to x86_64/nptl/64/rseq-access.h so that
> > sysdeps/nptl/rseq-access.h is used for x32.  This fixes BZ #32543.
> 
> Would this work on x86-64 and x86-64 x32?
> 
>        asm volatile ("movl %%fs:%P1(%q2),%0"                                  \
>                      : "=r" (__value)                                         \
>                      : "i" (offsetof (struct rseq_area, member)),             \
>                        "r" ((long long int) __rseq_offset));

__rseq_offset is a ptrdiff_t, which is indeed an issue specifically on
x32 because ptrdiff_t has size 4 and gets zero-extended to 64-bit.

The fix proposed by Florian is better than using the generic
sysdeps/nptl/rseq-access.h because it benefits from using the
%%fs segment selector prefix on x32 rather than compute the offset
based on __builtin_thread_pointer() with more instructions.

Thanks,

Mathieu

-- 
Mathieu Desnoyers
EfficiOS Inc.
http://www.efficios.com