kiq->eop_obj won't allowed to be freed until GFX finish, so we still better
unmap this object as long as we don't need to access it via CPU
发件人: amd-gfx 代表 zhoucm1
发送时间: 2016年12月28日 10:31:52
收件人: Yu, Xiangliang; amd-gfx@lists.freedesktop.org
抄送: Liu, Monk
主题: R
From: "Hongbo.He"
rollback to ttm default callback since OCL perf benchmark huge performance drop
Change-Id: If4cfc5a52872a7e00a7cbcad6efa46c57ee533de
Signed-off-by: Roger.He
---
drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/dr
> -Original Message-
> From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf
> Of zhoucm1
> Sent: Wednesday, December 28, 2016 10:32 AM
> To: Yu, Xiangliang ; amd-
> g...@lists.freedesktop.org
> Cc: Liu, Monk
> Subject: Re: [V2 1/1] drm/amdgpu/gfx8: add support kernel int
On 2016年12月28日 10:27, Yu, Xiangliang wrote:
+
>+ amdgpu_bo_kunmap(kiq->eop_obj);
>+
I see it is in many places, which is don't need I think,
amdgpu_bo_free_kernel will handle it if I'm correct.
Regards,
David zhou
___
amd-gfx mailing list
amd-gfx
Does anyone can give some comments?
This is urgent case.
Thanks!
Xiangliang Yu
> -Original Message-
> From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf
> Of Xiangliang Yu
> Sent: Friday, December 23, 2016 3:08 PM
> To: amd-gfx@lists.freedesktop.org
> Cc: Yu, Xiangli
It's a well known problem that the completion interrupts are notorious
unreliable.
That's why we have a fallback timer in amdgpu_fence.c which kicks an
extra hardware probe after a certain timeout. Please double check that
this one is working as expected.
Another possibility is that the memo