This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH] powerpc: Add missing insn in swapcontext [BZ #20004]
- From: Carlos O'Donell <carlos at redhat dot com>
- To: "Paul E. Murphy" <murphyp at linux dot vnet dot ibm dot com>, "libc-alpha at sourceware dot org" <libc-alpha at sourceware dot org>, Tulio Magno Quites Machado Filho <tuliom at linux dot vnet dot ibm dot com>
- Date: Thu, 28 Apr 2016 13:58:19 -0400
- Subject: Re: [PATCH] powerpc: Add missing insn in swapcontext [BZ #20004]
- Authentication-results: sourceware.org; auth=none
- References: <57221FF1 dot 7030902 at linux dot vnet dot ibm dot com> <57222E1F dot 1050506 at redhat dot com> <57223CAD dot 8040903 at linux dot vnet dot ibm dot com>
On 04/28/2016 12:39 PM, Paul E. Murphy wrote:
> On 04/28/2016 10:37 AM, Carlos O'Donell wrote:
>> On 04/28/2016 10:36 AM, Paul E. Murphy wrote:
>>> A missing instruction was discovered in the compat version of
>>> swapcontext while running the GCC test suite.
>>>
>>> 2016-04-28 Paul E. Murphy <murphyp@linux.vnet.ibm.com>
>>> Segher Boessenkool <segher@gcc.gnu.org>
>>>
>>> [BZ #20004]
>>> * sysdeps/unix/sysv/linux/powerpc/powerpc64/swapcontext.S:
>>> (__novec_swapcontext): Add missing load.
>>> ---
>>> sysdeps/unix/sysv/linux/powerpc/powerpc64/swapcontext.S | 1 +
>>> 1 file changed, 1 insertion(+)
>>>
>>> diff --git a/sysdeps/unix/sysv/linux/powerpc/powerpc64/swapcontext.S b/sysdeps/unix/sysv/linux/powerpc/powerpc64/swapcontext.S
>>> index a166dcc..a05a9cc 100644
>>> --- a/sysdeps/unix/sysv/linux/powerpc/powerpc64/swapcontext.S
>>> +++ b/sysdeps/unix/sysv/linux/powerpc/powerpc64/swapcontext.S
>>> @@ -163,6 +163,7 @@ ENTRY(__novec_swapcontext)
>>> cmpdi r0,0
>>> bne L(nv_do_sigret)
>>>
>>> + ld r8,.LC__dl_hwcap@toc(r2)
>>> # ifdef SHARED
>>> /* Load _rtld-global._dl_hwcap. */
>>> ld r8,RTLD_GLOBAL_RO_DL_HWCAP_OFFSET(r8)
>>>
>>
>> How did you test this?
>
> A slightly hacked up, debate-ably correct, example provided by the
> makecontext man page, which is modified to explicitly call the old
> version. I directly linked the generated object files.
>
> Without this patch it segfaults. With it, seemingly expected output.
Looks good to me then :-)
--
Cheers,
Carlos.