Re: [PATCH] drm/amdgpu: Fix extra call to amdgpu_ctx_put.

Am 11.10.2017 um 23:13 schrieb Andrey Grodzovsky: In amdgpu_cs_parser_init() in case of error handling amdgpu_ctx_put() is called without setting p->ctx to NULL after that, later amdgpu_cs_parser_fini() also calls amdgpu_ctx_put() again and mess up the reference count. Signed-off-by: Andrey Grod

Re: [PATCH] drm/amd/display: Add comment for NaN checks in DCN calcs

Hi Harry, couldn't a simple isnan() macro be used? Also a question below. Den 11 okt. 2017 17:01 skrev "Harry Wentland" : This is confusing as-is and really needs a comment. Signed-off-by: Harry Wentland --- drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_math.c | 4 1 file changed, 4 inser

TDR and VRAM lost handling in KMD (v2)

V2 summary Hi team please give your comments l When a job timed out (set from lockup_timeout kernel parameter), What KMD should do in TDR routine : 1.Update adev->gpu_reset_counter, and stop scheduler first 2.Set its fence error status to “ECANCELED”, 3.Find the co

Re: [PATCH 5/5] drm/amd/sched: signal and free remaining fences in amd_sched_entity_fini

Am 11.10.2017 um 18:30 schrieb Michel Dänzer: On 28/09/17 04:55 PM, Nicolai Hähnle wrote: From: Nicolai Hähnle Highly concurrent Piglit runs can trigger a race condition where a pending SDMA job on a buffer object is never executed because the corresponding process is killed (perhaps due to a

Re: [PATCH v2 2/2] drm/amdgpu: Add amdgpu_find_mm_node()

Am 11.10.2017 um 21:49 schrieb Harish Kasiviswanathan: v2: Use amdgpu_find_mm_node() in amdgpu_ttm_io_mem_pfn() Change-Id: I12231e18bb60152843cd0e0213ddd0d0e04e7497 Signed-off-by: Harish Kasiviswanathan Reviewed-by: Christian König --- drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 49 +++

Re: [PATCH] drm/amd/pp: add new sysfs pp_alloc_mem_for_smu

Am 12.10.2017 um 05:48 schrieb Alex Deucher: On Wed, Oct 11, 2017 at 7:28 AM, Rex Zhu wrote: Change-Id: Ie06f87445e7d6945472d88ac976693c98d96cd43 Signed-off-by: Rex Zhu Please add a better patch description. Something like: Add a sysfs interface to allocate a smu logging buffer on the fly.

Re: TDR and VRAM lost handling in KMD (v2)

Yeah, that sounds like a plan to me. Going to commit the patches I've already done in a minute, they still seem to implement quite a bit of what we want to do here. Regards, Christian. Am 12.10.2017 um 10:03 schrieb Liu, Monk: V2 summary Hi team *please give your comments* lWhen a job ti

Re: TDR and VRAM lost handling in KMD (v2)

I think we should stick to the plan where kernel contexts stay "stuck" after a GPU reset. This is the most robust behavior for the kernel. Even if the OpenGL spec says that an OpenGL context can be re-used without destroying and re-creating it, the UMD can take care of re-creating the kernel c

Re: TDR and VRAM lost handling in KMD (v2)

However, !guilty && ctx->reset_counter != adev->reset_counter does not imply that the context was lost. The way I understand it, we should return AMDGPU_CTX_INNOCENT_RESET if !guilty && ctx->vram_lost_counter != adev->vram_lost_counter. As far as I understand it, the case of !guilty && ctx->r

Re: TDR and VRAM lost handling in KMD (v2)

Hi David, Agreed. I'd also like to use the first variant (UMD compares the current vram_lost count with the share list's vram_lost count) for Mesa. Cheers, Nicolai On 12.10.2017 10:43, Mao, David wrote: Thanks Monk for the summary! Hi Nicolai, In order to block the usage of new context refe

Re: TDR and VRAM lost handling in KMD (v2)

Hi Monk, Thanks for the summary. Most of it looks good to me, though I can't speak to all the kernel internals. Just some comments: On 12.10.2017 10:03, Liu, Monk wrote: lFor cs_submit() IOCTL: 1.check if current ctx been marked “*guilty*”and return “*ECANCELED*”  if so. 2.set job->*vra

Re: TDR and VRAM lost handling in KMD (v2)

On 12.10.2017 10:49, Christian König wrote: However, !guilty && ctx->reset_counter != adev->reset_counter does not imply that the context was lost. The way I understand it, we should return AMDGPU_CTX_INNOCENT_RESET if !guilty && ctx->vram_lost_counter != adev->vram_lost_counter. As far as I

Re: TDR and VRAM lost handling in KMD (v2)

Am 12.10.2017 um 11:10 schrieb Nicolai Hähnle: On 12.10.2017 10:49, Christian König wrote: However, !guilty && ctx->reset_counter != adev->reset_counter does not imply that the context was lost. The way I understand it, we should return AMDGPU_CTX_INNOCENT_RESET if !guilty && ctx->vram_lost_c

Re: TDR and VRAM lost handling in KMD (v2)

On 12/10/17 11:23 AM, Christian König wrote: > Am 12.10.2017 um 11:10 schrieb Nicolai Hähnle: >> On 12.10.2017 10:49, Christian König wrote: However, !guilty && ctx->reset_counter != adev->reset_counter does not imply that the context was lost. The way I understand it, we should

Re: TDR and VRAM lost handling in KMD (v2)

On 12.10.2017 11:35, Michel Dänzer wrote: On 12/10/17 11:23 AM, Christian König wrote: Am 12.10.2017 um 11:10 schrieb Nicolai Hähnle: On 12.10.2017 10:49, Christian König wrote: However, !guilty && ctx->reset_counter != adev->reset_counter does not imply that the context was lost. The way I u

Re: TDR and VRAM lost handling in KMD (v2)

Am 12.10.2017 um 11:44 schrieb Nicolai Hähnle: On 12.10.2017 11:35, Michel Dänzer wrote: On 12/10/17 11:23 AM, Christian König wrote: Am 12.10.2017 um 11:10 schrieb Nicolai Hähnle: On 12.10.2017 10:49, Christian König wrote: However, !guilty && ctx->reset_counter != adev->reset_counter does n

Re: TDR and VRAM lost handling in KMD (v2)

Thanks Monk for the summary! Hi Nicolai, In order to block the usage of new context reference the old allocation, i think we need to do something in UMD so that KMD don’t need to monitor the resource list. I want to make sure we are on the same page. If you agree, then there might have two optio

[PATCH] drm/amdgpu: minor CS optimization

From: Christian König We only need to loop over all IBs for old UVD/VCE command stream patching. Signed-off-by: Christian König --- drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 37 +- 1 file changed, 19 insertions(+), 18 deletions(-) diff --git a/drivers/gpu/drm/am

Re: [PATCH 5/5] drm/amd/sched: signal and free remaining fences in amd_sched_entity_fini

On 12/10/17 10:05 AM, Christian König wrote: > Am 11.10.2017 um 18:30 schrieb Michel Dänzer: >> On 28/09/17 04:55 PM, Nicolai Hähnle wrote: >>> From: Nicolai Hähnle >>> >>> Highly concurrent Piglit runs can trigger a race condition where a >>> pending >>> SDMA job on a buffer object is never execu

[PATCH] drm/ttm: Fix unused variables with huge page support

Signed-off-by: Tom St Denis --- drivers/gpu/drm/ttm/ttm_page_alloc.c | 9 - 1 file changed, 8 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc.c b/drivers/gpu/drm/ttm/ttm_page_alloc.c index b6f16e73..95022473704b 100644 --- a/drivers/gpu/drm/ttm/ttm_page_

RE: TDR and VRAM lost handling in KMD (v2)

Hi team Very good, many policy and implement are agreed, looks we only have some arguments in amdgpu_ctx_query(), well I also confused with the current implement of it, ☹ First, I want to know if you guys agree that we don't update ctx->reset_counter in amdgpu_ctx_query() ? because I want to m

Re: [PATCH 5/5] drm/amd/sched: signal and free remaining fences in amd_sched_entity_fini

Am 12.10.2017 um 13:00 schrieb Michel Dänzer: On 12/10/17 10:05 AM, Christian König wrote: [SNIP] Is amd-staging-drm-next stable for you? It seemed stable before the changes you pushed this morning. :) As of cfb6dee86711 "drm/ttm: add transparent huge page support for cached allocations v2", I

Re: TDR and VRAM lost handling in KMD (v2)

Am 12.10.2017 um 13:37 schrieb Liu, Monk: Hi team Very good, many policy and implement are agreed, looks we only have some arguments in amdgpu_ctx_query(), well I also confused with the current implement of it, ☹ First, I want to know if you guys agree that we*don't update ctx->reset_counter i

Re: [PATCH] drm/ttm: Fix unused variables with huge page support

Am 12.10.2017 um 13:26 schrieb Tom St Denis: Signed-off-by: Tom St Denis Reviewed-by: Christian König --- drivers/gpu/drm/ttm/ttm_page_alloc.c | 9 - 1 file changed, 8 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc.c b/drivers/gpu/drm/ttm/ttm_page_

Re: TDR and VRAM lost handling in KMD (v2)

On 12.10.2017 13:49, Christian König wrote: Am 12.10.2017 um 13:37 schrieb Liu, Monk: Hi team Very good, many policy and implement are agreed, looks we only have some arguments in amdgpu_ctx_query(), well I also confused with the current implement of it, ☹ First, I want to know if you guys agr

Re: [PATCH] drm/amd/display: Add comment for NaN checks in DCN calcs

On 2017-10-12 03:32 AM, Nils Wallménius wrote: > Hi Harry, couldn't a simple isnan() macro be used? Also a question below. > Both very good points. I'll send a new patch. Harry > Den 11 okt. 2017 17:01 skrev "Harry Wentland" >: > > This is confusing as-is an

Re: [PATCH 5/5] drm/amd/sched: signal and free remaining fences in amd_sched_entity_fini

On 12/10/17 01:44 PM, Christian König wrote: > Am 12.10.2017 um 13:00 schrieb Michel Dänzer: > >> Anyway, unless anyone knows which commits from amd-staging-drm-next are >> needed to make 1d00402b4da2 stable in 4.14, the safe course of action >> seems to be reverting it (and ac7afe6b3cf3, which de

Re: [PATCH 5/5] drm/amd/sched: signal and free remaining fences in amd_sched_entity_fini

Am 12.10.2017 um 15:42 schrieb Michel Dänzer: On 12/10/17 01:44 PM, Christian König wrote: Am 12.10.2017 um 13:00 schrieb Michel Dänzer: Anyway, unless anyone knows which commits from amd-staging-drm-next are needed to make 1d00402b4da2 stable in 4.14, the safe course of action seems to be rev

Re: [PATCH] amdgpu/dc: Use DRM new-style object iterators.

On 2017-10-12 02:00 AM, Maarten Lankhorst wrote: Op 11-10-17 om 22:40 schreef Harry Wentland: On 2017-10-11 03:46 PM, Maarten Lankhorst wrote: Op 11-10-17 om 20:55 schreef Leo: On 2017-10-11 10:30 AM, Maarten Lankhorst wrote: Op 11-10-17 om 16:24 schreef sunpeng...@amd.com: From: "Leo (Su

drm/amdgpu: LVDS-1 output only turns on with sleep 1s workaround

Hi, I have a situation which might be a bug in drm/amd/amdgpu: I've got a TFT screen unit with an integrated AMD G-T56N mainboard used in a signage application. I build kernel and xorg myself. Kernels 4.9.29, 4.9.52 and 4.13.4 behave the same. git.freedesktop.org/~airlied/linux.git drm-next co

Re: [PATCH 5/5] drm/amd/sched: signal and free remaining fences in amd_sched_entity_fini

On 12/10/17 03:50 PM, Christian König wrote: > Am 12.10.2017 um 15:42 schrieb Michel Dänzer: >> On 12/10/17 01:44 PM, Christian König wrote: >>> Am 12.10.2017 um 13:00 schrieb Michel Dänzer: >>> Anyway, unless anyone knows which commits from amd-staging-drm-next are needed to make 1d00402

[PATCH] drm/amd/display: Explicit casting for grph object ids

C++ compilers don't like the implicit conversion Signed-off-by: Harry Wentland --- drivers/gpu/drm/amd/display/include/grph_object_id.h | 12 ++-- 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/amd/display/include/grph_object_id.h b/drivers/gpu/drm/amd/dis

RE: [PATCH] drm/amdgpu: minor CS optimization

> -Original Message- > From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf > Of Christian König > Sent: Thursday, October 12, 2017 6:20 AM > To: Grodzovsky, Andrey; amd-gfx@lists.freedesktop.org > Subject: [PATCH] drm/amdgpu: minor CS optimization > > From: Christian Kön

[pull] amdgpu drm-fixes-4.14

Hi Dave, One memory management regression fix. The following changes since commit 545036a9944e9d6e50fed4ca03117147c880ff71: Merge tag 'drm-misc-fixes-2017-10-11' of git://anongit.freedesktop.org/drm/drm-misc into drm-fixes (2017-10-12 10:38:09 +1000) are available in the git repository at:

Re: [pull] amdgpu drm-fixes-4.14

On 12/10/17 05:58 PM, Alex Deucher wrote: > Hi Dave, > > One memory management regression fix. > > The following changes since commit 545036a9944e9d6e50fed4ca03117147c880ff71: > > Merge tag 'drm-misc-fixes-2017-10-11' of > git://anongit.freedesktop.org/drm/drm-misc into drm-fixes (2017-10-12

Re: [PATCH 5/5] drm/amd/sched: signal and free remaining fences in amd_sched_entity_fini

On 12/10/17 01:00 PM, Michel Dänzer wrote: > > [0] I also got this, but I don't know yet if it's related: No, that seems to be a separate issue; I can still reproduce it with the huge page related changes reverted. Unfortunately, it doesn't seem to happen reliably on every piglit run. Even befor

RE: [pull] amdgpu drm-fixes-4.14

> -Original Message- > From: Michel Dänzer [mailto:mic...@daenzer.net] > Sent: Thursday, October 12, 2017 12:21 PM > To: Alex Deucher > Cc: amd-gfx@lists.freedesktop.org; dri-de...@lists.freedesktop.org; > airl...@gmail.com; Deucher, Alexander > Subject: Re: [pull] amdgpu drm-fixes-4.14 >

Re: [pull] amdgpu drm-fixes-4.14

Am 12.10.2017 um 18:20 schrieb Michel Dänzer: On 12/10/17 05:58 PM, Alex Deucher wrote: Hi Dave, One memory management regression fix. The following changes since commit 545036a9944e9d6e50fed4ca03117147c880ff71: Merge tag 'drm-misc-fixes-2017-10-11' of git://anongit.freedesktop.org/drm/dr

Re: [PATCH 5/5] drm/amd/sched: signal and free remaining fences in amd_sched_entity_fini

Am 12.10.2017 um 18:49 schrieb Michel Dänzer: On 12/10/17 01:00 PM, Michel Dänzer wrote: [0] I also got this, but I don't know yet if it's related: No, that seems to be a separate issue; I can still reproduce it with the huge page related changes reverted. Unfortunately, it doesn't seem to happ

[PATCH] drm/ttm: fix the fix for huge compound pages

From: Christian König We don't use compound pages at the moment. Take this into account when freeing them. Signed-off-by: Christian König --- drivers/gpu/drm/ttm/ttm_page_alloc.c | 21 - 1 file changed, 16 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/ttm/ttm_

Re: [pull] amdgpu drm-fixes-4.14

On Thu, Oct 12, 2017 at 1:02 PM, Christian König wrote: > Am 12.10.2017 um 18:20 schrieb Michel Dänzer: >> >> On 12/10/17 05:58 PM, Alex Deucher wrote: >>> >>> Hi Dave, >>> >>> One memory management regression fix. >>> >>> The following changes since commit >>> 545036a9944e9d6e50fed4ca03117147c880

[PATCH] drm/amd/display: Fix warning about overflow

We're overflowing the last bit. Cast it explicitly Signed-off-by: Harry Wentland --- drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c b/drivers/gpu/drm/amd/display/dc/bios/bi

Re: [PATCH] drm/amdgpu: introduce AMDGPU_GEM_CREATE_EXPLICIT_SYNC v2

On Thu, Sep 21, 2017 at 4:38 PM, Andres Rodriguez wrote: > Hi Christian, > > The reference radv patches are on the list. The basic idea is to only set > the explicit sync flag for buffers allocated for dri usage. Did you mean "only set the explicit sync flag for buffers NOT allocated for dri usag

Re: [PATCH] drm/amdgpu: introduce AMDGPU_GEM_CREATE_EXPLICIT_SYNC v2

Am 12.10.2017 um 20:27 schrieb Marek Olšák: On Thu, Sep 21, 2017 at 4:38 PM, Andres Rodriguez wrote: Hi Christian, The reference radv patches are on the list. The basic idea is to only set the explicit sync flag for buffers allocated for dri usage. Did you mean "only set the explicit sync fla

Re: [PATCH] drm/amdgpu: introduce AMDGPU_GEM_CREATE_EXPLICIT_SYNC v2

On 2017-10-12 02:27 PM, Marek Olšák wrote: On Thu, Sep 21, 2017 at 4:38 PM, Andres Rodriguez wrote: Hi Christian, The reference radv patches are on the list. The basic idea is to only set the explicit sync flag for buffers allocated for dri usage. Did you mean "only set the explicit sync f

[ANNOUNCE] libdrm 2.4.84

-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 libdrm 2.4.84 has been released. Alex Deucher (1): tests/amdgpu: add missing header to SOURCES Andrey Grodzovsky (1): amdgpu: Add deadlock detection test suit. Anuj Phogat (1): intel: Change a KBL pci id to GT2 from GT1.5 Chris

Re: [PATCH 1/3] drm/syncobj: extract two helpers from drm_syncobj_create

Sorry, this codepath is not tested by radeonsi. Marek On Sun, Oct 1, 2017 at 1:20 AM, Zhou, David(ChunMing) wrote: > Could you test and review it? On hand, I have no env. > > Regards, > David Zhou > > 发自坚果 Pro > > Marek Ol?醟 于 2017年9月30日 下午11:56写道： > > The idea sounds good. > > Marek > > On Sat

[PATCH] drm/amdgpu: bump version for vram lost counter query

Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c index b9a3258..38cd8bd 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_dr

[PATCH v2 1/6] drm/amd/display: Use DRM new-style object iterators.

From: "Leo (Sunpeng) Li" Use the correct for_each_new/old_* iterators instead of for_each_* List of affected functions: amdgpu_dm_find_first_crtc_matching_connector: use for_each_new - Old from_state_var flag was always choosing the new state amdgpu_dm_display_resume: use for_each_new

[PATCH 4/6] drm/amd/display: Unify amdgpu_dm state variable namings.

From: "Leo (Sunpeng) Li" Use dm_new_*_state and dm_old_*_state for their respective amdgpu_dm new and old object states. Helps with readability, and enforces use of new DRM api (choose either new, or old). Signed-off-by: Leo (Sunpeng) Li --- drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c |

[PATCH 2/6] drm/amd/display: Use new DRM API where possible

From: "Leo (Sunpeng) Li" To conform to DRM's new API, we should not be accessing a DRM object's internal state directly. Rather, the DRM for_each_old/new_* iterators, and drm_atomic_get_old/new_* interface should be used. This is an ongoing process. For now, update the DRM-facing atomic function

[PATCH 5/6] drm/amd/display: Fix typo

From: "Leo (Sunpeng) Li" undersacn -> underscan Signed-off-by: Leo (Sunpeng) Li --- drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_

[PATCH 0/6] Use new DRM API where possible, and cleanups.

From: "Leo (Sunpeng) Li" Hi Dave, This series reworks the previous patch. Patch 1 is a v2 of the previous, and additional patches are from the feedback received. They apply on top of your drm-next-amd-dc-staging branch. Thanks, Leo Leo (Sunpeng) Li (6): drm/amd/display: Use DRM new-style obj

[PATCH 6/6] drm/amd/display: Remove useless pcrtc pointer

From: "Leo (Sunpeng) Li" in amdgpu_dm_atomic_commit_tail. Just use crtc instead. Signed-off-by: Leo (Sunpeng) Li --- drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/

[PATCH 3/6] drm/amd/display: Unify DRM state variable namings.

From: "Leo (Sunpeng) Li" Use new_*_state and old_*_state for their respective new/old DRM object states. Signed-off-by: Leo (Sunpeng) Li --- drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 80 +++ 1 file changed, 40 insertions(+), 40 deletions(-) diff --git a/drivers/g

[amd-staging-drm-next] compilation error with latest commit #1b006d838f78

Hello, next (regression) compilation error: drivers/gpu/drm/amd/amdgpu/../display/dc/core/dc_resource.c: In function ‘resource_map_pool_resources’: drivers/gpu/drm/amd/amdgpu/../display/dc/core/dc_resource.c:1688:14: error: implicit declaration of function ‘acquire_first_split_pipe’; did you

RE: TDR and VRAM lost handling in KMD (v2)

>That sounds sane, but unfortunately might not be possible with the existing >IOCTL. Keep in mind that we need to keep backward compatibility here. unfortunately the current scheme on amdgpu_ctx_query() won’t work with TDR feature, which is aim to support vulkan/mesa/close-ogl/radv … It’s enumer