| Message ID | 1419321131-6121-1-git-send-email-xiong.y.zhang@intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Tue, Dec 23, 2014 at 03:52:11PM +0800, Xiong Zhang wrote: > After i915 commit: > commit bd008e5b2953186fc0c6633a885ade95e7043800 > Author: Chris Wilson <chris@chris-wilson.co.uk> > Date: Tue Oct 7 14:13:51 2014 +0100 > > drm: Implement O_NONBLOCK support on /dev/dri/cardN > > the return value for drm short_buffer read is -1 and errno is > EAGAIN. No, it is not. -Chris
> -----Original Message----- > From: Chris Wilson [mailto:chris@chris-wilson.co.uk] > Sent: Tuesday, December 23, 2014 5:53 PM > To: Zhang, Xiong Y > Cc: intel-gfx@lists.freedesktop.org > Subject: Re: [Intel-gfx] [PATCH] igt: Correct the return value for drm > short_buffer read > > On Tue, Dec 23, 2014 at 03:52:11PM +0800, Xiong Zhang wrote: > > After i915 commit: > > commit bd008e5b2953186fc0c6633a885ade95e7043800 > > Author: Chris Wilson <chris@chris-wilson.co.uk> > > Date: Tue Oct 7 14:13:51 2014 +0100 > > > > drm: Implement O_NONBLOCK support on /dev/dri/cardN > > > > the return value for drm short_buffer read is -1 and errno is EAGAIN. > > No, it is not. > -Chris Without this patch, system fail in short-buffer-block and short-buffer-nonblock subtest. With this patch, these two subtest could pass. > > -- > Chris Wilson, Intel Open Source Technology Centre
On Tue, Dec 23, 2014 at 10:14:15AM +0000, Zhang, Xiong Y wrote: > > -----Original Message----- > > From: Chris Wilson [mailto:chris@chris-wilson.co.uk] > > Sent: Tuesday, December 23, 2014 5:53 PM > > To: Zhang, Xiong Y > > Cc: intel-gfx@lists.freedesktop.org > > Subject: Re: [Intel-gfx] [PATCH] igt: Correct the return value for drm > > short_buffer read > > > > On Tue, Dec 23, 2014 at 03:52:11PM +0800, Xiong Zhang wrote: > > > After i915 commit: > > > commit bd008e5b2953186fc0c6633a885ade95e7043800 > > > Author: Chris Wilson <chris@chris-wilson.co.uk> > > > Date: Tue Oct 7 14:13:51 2014 +0100 > > > > > > drm: Implement O_NONBLOCK support on /dev/dri/cardN > > > > > > the return value for drm short_buffer read is -1 and errno is EAGAIN. > > > > No, it is not. > > -Chris > Without this patch, system fail in short-buffer-block and short-buffer-nonblock subtest. > With this patch, these two subtest could pass. That's the point of the test, the kernel behaviour is wrong. There is a patch to fix the kernel. -Chris
> -----Original Message----- > From: Chris Wilson [mailto:chris@chris-wilson.co.uk] > Sent: Tuesday, December 23, 2014 7:31 PM > To: Zhang, Xiong Y > Cc: intel-gfx@lists.freedesktop.org > Subject: Re: [Intel-gfx] [PATCH] igt: Correct the return value for drm > short_buffer read > > On Tue, Dec 23, 2014 at 10:14:15AM +0000, Zhang, Xiong Y wrote: > > > -----Original Message----- > > > From: Chris Wilson [mailto:chris@chris-wilson.co.uk] > > > Sent: Tuesday, December 23, 2014 5:53 PM > > > To: Zhang, Xiong Y > > > Cc: intel-gfx@lists.freedesktop.org > > > Subject: Re: [Intel-gfx] [PATCH] igt: Correct the return value for > > > drm short_buffer read > > > > > > On Tue, Dec 23, 2014 at 03:52:11PM +0800, Xiong Zhang wrote: > > > > After i915 commit: > > > > commit bd008e5b2953186fc0c6633a885ade95e7043800 > > > > Author: Chris Wilson <chris@chris-wilson.co.uk> > > > > Date: Tue Oct 7 14:13:51 2014 +0100 > > > > > > > > drm: Implement O_NONBLOCK support on /dev/dri/cardN > > > > > > > > the return value for drm short_buffer read is -1 and errno is EAGAIN. > > > > > > No, it is not. > > > -Chris > > Without this patch, system fail in short-buffer-block and > short-buffer-nonblock subtest. > > With this patch, these two subtest could pass. > > That's the point of the test, the kernel behaviour is wrong. There is a patch to fix > the kernel. > -Chris [Zhang, Xiong Y] Oh, I know it. Thanks. So could you send this patch to fix it ? > > -- > Chris Wilson, Intel Open Source Technology Centre
On Fri, Dec 26, 2014 at 01:16:06AM +0000, Zhang, Xiong Y wrote: > > -----Original Message----- > > From: Chris Wilson [mailto:chris@chris-wilson.co.uk] > > Sent: Tuesday, December 23, 2014 7:31 PM > > To: Zhang, Xiong Y > > Cc: intel-gfx@lists.freedesktop.org > > Subject: Re: [Intel-gfx] [PATCH] igt: Correct the return value for drm > > short_buffer read > > > > On Tue, Dec 23, 2014 at 10:14:15AM +0000, Zhang, Xiong Y wrote: > > > > -----Original Message----- > > > > From: Chris Wilson [mailto:chris@chris-wilson.co.uk] > > > > Sent: Tuesday, December 23, 2014 5:53 PM > > > > To: Zhang, Xiong Y > > > > Cc: intel-gfx@lists.freedesktop.org > > > > Subject: Re: [Intel-gfx] [PATCH] igt: Correct the return value for > > > > drm short_buffer read > > > > > > > > On Tue, Dec 23, 2014 at 03:52:11PM +0800, Xiong Zhang wrote: > > > > > After i915 commit: > > > > > commit bd008e5b2953186fc0c6633a885ade95e7043800 > > > > > Author: Chris Wilson <chris@chris-wilson.co.uk> > > > > > Date: Tue Oct 7 14:13:51 2014 +0100 > > > > > > > > > > drm: Implement O_NONBLOCK support on /dev/dri/cardN > > > > > > > > > > the return value for drm short_buffer read is -1 and errno is EAGAIN. > > > > > > > > No, it is not. > > > > -Chris > > > Without this patch, system fail in short-buffer-block and > > short-buffer-nonblock subtest. > > > With this patch, these two subtest could pass. > > > > That's the point of the test, the kernel behaviour is wrong. There is a patch to fix > > the kernel. > > -Chris > [Zhang, Xiong Y] Oh, I know it. Thanks. > So could you send this patch to fix it ? http://patchwork.freedesktop.org/patch/38174/ Daniel wanted the igt test first... -Chris
diff --git a/tests/drm_read.c b/tests/drm_read.c index 334f26a..a17d981 100644 --- a/tests/drm_read.c +++ b/tests/drm_read.c @@ -145,7 +145,8 @@ static void test_short_buffer(int in, int nonblock) alarm(3); - igt_assert_eq(read(fd, buffer, 4), 0); + igt_assert_eq(read(fd, buffer, 4), -1); + igt_assert_eq(errno, EAGAIN); igt_assert(read(fd, buffer, 40) > 0); igt_assert(read(fd, buffer, 40) > 0);
After i915 commit: commit bd008e5b2953186fc0c6633a885ade95e7043800 Author: Chris Wilson <chris@chris-wilson.co.uk> Date: Tue Oct 7 14:13:51 2014 +0100 drm: Implement O_NONBLOCK support on /dev/dri/cardN the return value for drm short_buffer read is -1 and errno is EAGAIN. Signed-off-by: Xiong Zhang <xiong.y.zhang@intel.com> --- tests/drm_read.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)