This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH 11/13] nptl: arm: Fix Race conditions in pthread cancellation (BZ#12683)
- From: Adhemerval Zanella <adhemerval dot zanella at linaro dot org>
- To: Phil Blundell <pb at pbcl dot net>, libc-alpha at sourceware dot org
- Date: Fri, 9 Oct 2015 10:27:54 -0300
- Subject: Re: [PATCH 11/13] nptl: arm: Fix Race conditions in pthread cancellation (BZ#12683)
- Authentication-results: sourceware.org; auth=none
- References: <1444234995-9542-1-git-send-email-adhemerval dot zanella at linaro dot com> <1444234995-9542-12-git-send-email-adhemerval dot zanella at linaro dot com> <1444318152 dot 5251 dot 6 dot camel at pbcl dot net>
On 08-10-2015 12:29, Phil Blundell wrote:
> On Wed, 2015-10-07 at 13:23 -0300, Adhemerval Zanella wrote:
>> From: Adhemerval Zanella <adhemerval.zanella@linaro.org>
>>
>> This patch adds the ARM modifications required for the BZ#12683 fix.
>> It basically removes the enable_asynccancel/disable_asynccancel
>> function
>> usage on code, provide a arch-specific symbol that contains global
>> markers to be used in SIGCANCEL handler.
>>
>> Checked on armhf.
>
> What architecture variant specifically? Have you verified that this at
> least compiles on older arches? (I think we still support everything
> back to ARMv4.)
I checked on arm7l and arm8l. I will try with a build for armv4.
>
>>
> + PSEUDO_CANCEL_BEFORE;
>> \
>> + ldr r0, =SYS_ify (syscall_name);
>> \
>> + PSEUDO_CANCEL_AFTER;
>>
>
> It's still not very clear to me that the PSEUDO_CANCEL_BEFORE and
> PSEUDO_CANCEL_AFTER macros are achieving anything other than making the
> code harder to read. Is there any reason these couldn't just be
> written in line?
Indeed, I will change it.
>
> p.
>