This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: libgomp nvptx plugin: rework initialisation and support the proposed load/unload hooks (was: Merge current set of OpenACC changes from gomp-4_0-branch)
- From: Ilya Verbin <iverbin at gmail dot com>
- To: Thomas Schwinge <thomas at codesourcery dot com>
- Cc: Julian Brown <julian at codesourcery dot com>, Jakub Jelinek <jakub at redhat dot com>, gcc-patches at gcc dot gnu dot org, Kirill Yukhin <kirill dot yukhin at gmail dot com>
- Date: Wed, 25 Feb 2015 15:02:37 +0300
- Subject: Re: libgomp nvptx plugin: rework initialisation and support the proposed load/unload hooks (was: Merge current set of OpenACC changes from gomp-4_0-branch)
- Authentication-results: sourceware.org; auth=none
- References: <20150123172053 dot GA22341 at msticlxl57 dot ims dot intel dot com> <87lhkpr70s dot fsf at schwinge dot name> <20150126143426 dot GA8810 at msticlxl57 dot ims dot intel dot com> <20150127140741 dot 3f6d650b at octopus> <20150203112844 dot GA54416 at msticlxl57 dot ims dot intel dot com> <20150203130000 dot 7be35f5e at octopus> <20150203200104 dot GB54416 at msticlxl57 dot ims dot intel dot com> <20150204150545 dot 2628ad35 at octopus> <20150224112951 dot 363b71fd at octopus> <87a902ib93 dot fsf at schwinge dot name>
On Wed, Feb 25, 2015 at 10:36:08 +0100, Thomas Schwinge wrote:
> > Julian Brown <julian@codesourcery.com> wrote:
> > OK for gomp4 branch? I could commit Ilya's patch there too if so.
>
> I'll leave the decision to Jakub, but, what about trunk? As Ilya
> indicated in
> <http://news.gmane.org/find-root.php?message_id=%3C20150116231632.GB48380%40msticlxl57.ims.intel.com%3E>,
> (at least part of) these patches are fixing a regression with offloading
> From shared libraries. (And maybe the rest qualifies as fixes and
> extensions to new code (offloading), so no danger to cause any
> regressions compared to the last GCC release?)
BTW, when I removed calls to gomp_init_tables in <https://gcc.gnu.org/ml/gcc-patches/2015-01/msg02275.html>,
I could accidentally remove some necessary gomp_mutex_lock/unlock.
Also GOMP_offload_[un]register require some mutexes, as noted by Jakub.
I'm going to fix this. So, I think we should commit all dependent patches to
gomp4 branch, and I will post a fix for mutexes on top of them.
Thanks,
-- Ilya