This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH] nptl: support thread stacks that grow up [ping]
- From: "Carlos O'Donell" <carlos at redhat dot com>
- To: Roland McGrath <roland at hack dot frob dot com>, libc-alpha at sourceware dot org
- Date: Tue, 12 Jan 2016 19:07:50 -0500
- Subject: Re: [PATCH] nptl: support thread stacks that grow up [ping]
- Authentication-results: sourceware.org; auth=none
- References: <1439108237-31537-1-git-send-email-vapier at gentoo dot org> <20151229225736 dot GN25803 at vapier dot lan> <20160109083231 dot GA28268 at vapier dot lan> <20160112210341 dot BB09F2C3C2F at topped-with-meat dot com> <20160112220022 dot GJ4894 at vapier dot lan>
On 01/12/2016 05:00 PM, Mike Frysinger wrote:
> On 12 Jan 2016 13:03, Roland McGrath wrote:
>> It looks OK to me. If you can verify that it doesn't change the generated
>> code nontrivially on x86_64, then I think that is sufficient to declare it
>> safe enough for working machines. And grows-up machines are sufficiently
>> broken today in trunk that I'm not concerned about you making them worse.
>
> pthread_create.c & pthread_getattr_np.c have mostly the same disassembly.
> my gcc is being weird and changing a constant in one insn depending on
> whitespace in the file.
>
> allocatestack.c has some code shuffling, but it's the same size overall.
>
> tests still pass on x86_64.
>
> i'll push after the 2.23 branch.
Thanks. To be clear I have tested these patches on x86_64 and they passed
with no regressions.
Cheers,
Carlos.