RE: [PATCH] drm/amdgpu: add rcu_barrier after entity fini

Looks good, Acked-by: Chunming Zhou -Original Message- From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf Of Emily Deng Sent: Thursday, May 17, 2018 11:05 AM To: amd-gfx@lists.freedesktop.org Cc: Deng, Emily Subject: [PATCH] drm/amdgpu: add rcu_barrier after entity

[PATCH] drm/amdgpu: add rcu_barrier after entity fini

To free the fence from the amdgpu_fence_slab, need twice call_rcu, to avoid the amdgpu_fence_slab_fini call kmem_cache_destroy(amdgpu_fence_slab) before kmem_cache_free(amdgpu_fence_slab, fence), add rcu_barrier after drm_sched_entity_fini. The kmem_cache_free(amdgpu_fence_slab, fence)'s call tra

Re: [PATCH v2] drm/scheduler: Remove obsolete spinlock.

Yes, every thing is clear, Acked-by: Chunming Zhou On 2018年05月16日 23:33, Andrey Grodzovsky wrote: This spinlock is superfluous, any call to drm_sched_entity_push_job should already be under a lock together with matching drm_sched_job_init to match the order of insertion into queue with job's f

Re: [PATCH] drm/amdgpu/display: remove VEGAM config option

On 2018-05-16 10:12 AM, Alex Deucher wrote: > Leftover from bringup. No need to keep it around for > upstream. > > Signed-off-by: Alex Deucher Reviewed-by: Harry Wentland Harry > --- > drivers/gpu/drm/amd/display/Kconfig | 7 --- > drivers/gpu/drm/amd/display/amd

Re: [PATCH 2/2] drm/amdgpu/display: fix vega12/20 handling in dal_asic_id.h

On 2018-05-16 04:37 PM, Alex Deucher wrote: > - Remove unused ASICREV_IS_VEGA12_p() macro > - Fix ASICREV_IS_VEGA12_P() macro to properly check against vega20 > > Signed-off-by: Alex Deucher Both patches are Reviewed-by: Harry Wentland Harry > --- > drivers/gpu/drm/amd/display/include/dal_as

Re: [PATCH] drm/amdgpu: conditionally compile amdgpu's amdkfd files

Hi Oded, Thanks for working on this! The Makefile changes look good. Instead of checking and calling function pointers in amdgpu_amdkfd_... functions at runtime, couldn't you just define empty stub functions in amdgpu_amdkfd.h if KFD is not enabled? I think that would make the code shorter and re

[PATCH 3/4] drm/amdgpu: add a df 1.7 implementation of enable_ecc_force_par_wr_rmw

Needed for proper memory setup depending on whether ECC is enabled on a particular board. Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/df_v1_7.c | 8 1 file changed, 8 insertions(+) diff --git a/drivers/gpu/drm/amd/amdgpu/df_v1_7.c b/drivers/gpu/drm/amd/amdgpu/df_v1_7.c

[PATCH 2/4] drm/amdgpu: add new DF callback for ECC setup

The ForceParWrRMW setting needs to be enabled for ECC, but disabled when ECC is not enabled. Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu.h | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index

[PATCH 4/4] drm/amdgpu/gmc9: disable partial wr rmw if ECC is not enabled

The vbios mistakenly sets this bit on some boards without ECC. This can lead to reduced performance in some workloads. Disable the bit if the board does not have ECC. Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c | 1 + 1 file changed, 1 insertion(+) diff --git a/driver

[PATCH 1/4] drm/amdgpu: add new DF 1.7 register defs

Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/include/asic_reg/df/df_1_7_offset.h | 4 drivers/gpu/drm/amd/include/asic_reg/df/df_1_7_sh_mask.h | 4 2 files changed, 8 insertions(+) diff --git a/drivers/gpu/drm/amd/include/asic_reg/df/df_1_7_offset.h b/drivers/gpu/drm/amd/inclu

[PATCH 1/2] drm/amdgpu/display: remove VEGA20 config option

Leftover from bringup. No need to keep it around for upstream. Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/display/Kconfig| 8 - drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c | 2 - drivers/gpu/drm/amd/display/dc/dce/dce_clocks.c| 6 - .../drm/amd/display/

[PATCH 2/2] drm/amdgpu/display: fix vega12/20 handling in dal_asic_id.h

- Remove unused ASICREV_IS_VEGA12_p() macro - Fix ASICREV_IS_VEGA12_P() macro to properly check against vega20 Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/display/include/dal_asic_id.h | 8 +++- 1 file changed, 3 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/amd/display

Re: [PATCH] drm/amd/pp: Workaround flickering issue on RV

On 2018-05-14 09:17 AM, Alex Deucher wrote: > On Mon, May 14, 2018 at 4:06 AM, Rex Zhu wrote: >> Screen flickering observed while running 1080p video using >> MPV_VAAPI/VDPAU with 4x4K@60 monitors >> >> Need to set higher mclk in this configuration. >> >> Signed-off-by: Rex Zhu > > Probably long

Re: [PATCH xf86-video-amdgpu 10/13] Push staged color properties when DPMS state toggles On

On 2018-05-03 08:31 PM, sunpeng...@amd.com wrote: > From: "Leo (Sunpeng) Li" > > This will persist color management properties on a CRTC across DPMS > state changes. > > Signed-off-by: Leo (Sunpeng) Li > --- > src/drmmode_display.c | 6 ++ > 1 file changed, 6 insertions(+) > > diff --git

Re: [PATCH xf86-video-amdgpu 08/13] Set driver-private CRTC's dpms mode on disable

On 2018-05-03 08:31 PM, sunpeng...@amd.com wrote: > From: "Leo (Sunpeng) Li" > > The dpms_mode flag on the driver-private CRTC was not being set when > it's DPMS state is set to off. This causes some problems when toggling > it back on, as some conditionals check this flag. > > Signed-off-by: Le

Re: [PATCH xf86-video-amdgpu 03/13] List disabled color properties on RandR outputs without a CRTC

On 2018-05-03 08:31 PM, sunpeng...@amd.com wrote: > From: "Leo (Sunpeng) Li" > > The properties on an RandR output needs to stay consistent throughout > it's lifecycle. However, we cannot list color properties on an output if > there is no CRTC attached. > > Therefore, create a fake CRTC, and in

Re: [PATCH xf86-video-amdgpu 02/13] Push color properties to kernel DRM on CRTC init

On 2018-05-03 08:31 PM, sunpeng...@amd.com wrote: > From: "Leo (Sunpeng) Li" > > Push staged values on the driver-private CRTC, to kernel DRM when it's > initialized. This is to flush out any previous state that hardware was > in, and set them to their default values. > > Signed-off-by: Leo (Sun

Re: [PATCH xf86-video-amdgpu 01/13] Add color management properties to driver-private CRTC object

On 2018-05-03 08:31 PM, sunpeng...@amd.com wrote: > From: "Leo (Sunpeng) Li" > > Non-legacy color management consists of 3 properties on the CRTC: > Degamma LUT, Color Transformation Matrix (CTM), and Gamma LUT. > > Add these properties to the driver-private CRTC, and initialize them > when the

Re: [PATCH xf86-video-amdgpu 00/13] Enabling Color Management - Round 2

On 2018-05-03 08:31 PM, sunpeng...@amd.com wrote: > > 3. The three color management properties (Degamma LUT, Color Transform Matrix >(CTM), and Gamma LUT) are hard-coded into the DDX driver, to be listed (as >disabled) regardless of whether a CRTC is attached on the output, or > whether >

Re: [PATCH 3/6] drm/amdgpu: Add CG/PG flags for VCN

On Wed, May 16, 2018 at 8:52 AM, Rex Zhu wrote: > Signed-off-by: Rex Zhu Reviewed-by: Alex Deucher > --- > drivers/gpu/drm/amd/include/amd_shared.h | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/drivers/gpu/drm/amd/include/amd_shared.h > b/drivers/gpu/drm/amd/incl

Re: [PATCH 2/6] drm/amd/pp: Add smu support for vcn cg/pg on RV

On Wed, May 16, 2018 at 8:52 AM, Rex Zhu wrote: > Signed-off-by: Rex Zhu Reviewed-by: Alex Deucher > --- > drivers/gpu/drm/amd/powerplay/hwmgr/smu10_hwmgr.c | 25 > ++- > 1 file changed, 24 insertions(+), 1 deletion(-) > > diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/

Re: [PATCH 1/6] drm/amdgpu: skip CG for VCN when late_init/fini

On Wed, May 16, 2018 at 8:52 AM, Rex Zhu wrote: > Signed-off-by: Rex Zhu Please add a patch description. E.g., something like: VCN clockgating is handled manually like VCE and UVD. With that fixed: Reviewed-by: Alex Deucher > --- > drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 2 ++ > 1 file

Re: [PATCH 4/6] drm/amdgpu: Add runtime vcn cg/pg enablement

On Wed, May 16, 2018 at 8:53 AM, Rex Zhu wrote: > Signed-off-by: Rex Zhu Please add a better patch description. E.g., drm/amdgpu: Add runtime vcn cg/pg enablement Enable support for dynamically powering up/down VCN on demand. Signed-off-by: Rex Zhu With that fixed: Reviewed-by: Alex Deuch

Re: [PATCH 6/6] drm/amdgpu: Enable static pg feature on RV

On Wed, May 16, 2018 at 8:53 AM, Rex Zhu wrote: > Signed-off-by: Rex Zhu > --- > drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h | 11 ++ > drivers/gpu/drm/amd/amdgpu/soc15.c | 3 +- > drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c | 183 > ++-- > 3 files changed, 187 inser

Re: [PATCH 5/6] drm/amdgpu: Enable static cg for vcn on RV

On Wed, May 16, 2018 at 8:53 AM, Rex Zhu wrote: > Signed-off-by: Rex Zhu > --- > drivers/gpu/drm/amd/amdgpu/soc15.c| 3 ++- > drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c | 16 > 2 files changed, 10 insertions(+), 9 deletions(-) > > diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c

Re: [PATCH xf86-video-amdgpu] Replace 'foo == NULL' with '!foo'

On Wed, May 16, 2018 at 10:51 AM, Michel Dänzer wrote: > From: Michel Dänzer > > Shorter and sweeter. :) > > Signed-off-by: Michel Dänzer Reviewed-by: Alex Deucher > --- > src/amdgpu_bo_helper.c | 10 -- > src/amdgpu_dri2.c| 10 +- > src/amdgpu_glamor.c

[PATCH v2] drm/scheduler: Remove obsolete spinlock.

This spinlock is superfluous, any call to drm_sched_entity_push_job should already be under a lock together with matching drm_sched_job_init to match the order of insertion into queue with job's fence seqence number. v2: Improve patch description. Add functions documentation describing the locking

[PATCH xf86-video-amdgpu] Replace 'foo == NULL' with '!foo'

From: Michel Dänzer Shorter and sweeter. :) Signed-off-by: Michel Dänzer --- src/amdgpu_bo_helper.c | 10 -- src/amdgpu_dri2.c| 10 +- src/amdgpu_glamor.c | 2 +- src/amdgpu_glamor_wrappers.c | 2 +- src/amdgpu_kms.c | 8 src/a

Re: [PATCH 2/2] drm/scheduler: Remove obsolete spinlock.

On 05/16/2018 09:16 AM, Lucas Stach wrote: Am Mittwoch, den 16.05.2018, 15:10 +0200 schrieb Christian König: Am 16.05.2018 um 15:00 schrieb Lucas Stach: Am Mittwoch, den 16.05.2018, 14:32 +0200 schrieb Christian König: Am 16.05.2018 um 14:28 schrieb Lucas Stach: Am Mittwoch, den 16.05.2018,

Re: [PATCH] drm/amdgpu/display: remove VEGAM config option

On 2018-05-16 04:12 PM, Alex Deucher wrote: > Leftover from bringup. No need to keep it around for > upstream. > > Signed-off-by: Alex Deucher Nice, thanks. Reviewed-by: Michel Dänzer -- Earthling Michel Dänzer | http://www.amd.com Libre software enthusiast

[PATCH] drm/amdgpu/display: remove VEGAM config option

Leftover from bringup. No need to keep it around for upstream. Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/display/Kconfig | 7 --- drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 4 drivers/gpu/drm/amd/display/dc/bios/command_table_helper.

Re: [PATCH 2/2] drm/scheduler: Remove obsolete spinlock.

Am Mittwoch, den 16.05.2018, 15:10 +0200 schrieb Christian König: > Am 16.05.2018 um 15:00 schrieb Lucas Stach: > > Am Mittwoch, den 16.05.2018, 14:32 +0200 schrieb Christian König: > > > Am 16.05.2018 um 14:28 schrieb Lucas Stach: > > > > Am Mittwoch, den 16.05.2018, 14:08 +0200 schrieb Christian

Re: [PATCH 2/2] drm/scheduler: Remove obsolete spinlock.

Am 16.05.2018 um 15:00 schrieb Lucas Stach: Am Mittwoch, den 16.05.2018, 14:32 +0200 schrieb Christian König: Am 16.05.2018 um 14:28 schrieb Lucas Stach: Am Mittwoch, den 16.05.2018, 14:08 +0200 schrieb Christian König: Yes, exactly. For normal user space command submission we should have ton

[PATCH umr] Introduction of Vega20.

Signed-off-by: Tom St Denis --- scripts/soc15_asic.sh | 3 +- src/lib/asic/CMakeLists.txt | 1 + src/lib/asic/vega20.c | 51 src/lib/asic/vega20.i | 189 src/lib/discover_by_did.c | 6 ++ src/lib/discover_by_name

Re: [PATCH 2/2] drm/scheduler: Remove obsolete spinlock.

Am Mittwoch, den 16.05.2018, 14:32 +0200 schrieb Christian König: > Am 16.05.2018 um 14:28 schrieb Lucas Stach: > > Am Mittwoch, den 16.05.2018, 14:08 +0200 schrieb Christian König: > > > Yes, exactly. > > > > > > For normal user space command submission we should have tons of > > > locks > > > gu

[PATCH 6/6] drm/amdgpu: Enable static pg feature on RV

Signed-off-by: Rex Zhu --- drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h | 11 ++ drivers/gpu/drm/amd/amdgpu/soc15.c | 3 +- drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c | 183 ++-- 3 files changed, 187 insertions(+), 10 deletions(-) diff --git a/drivers/gpu/drm/amd/amdg

[PATCH 4/6] drm/amdgpu: Add runtime vcn cg/pg enablement

Signed-off-by: Rex Zhu --- drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c | 19 ++- 1 file changed, 14 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c index 58e4953..8b0d491 100644 --- a/drivers/gpu/drm/amd/am

[PATCH 3/6] drm/amdgpu: Add CG/PG flags for VCN

Signed-off-by: Rex Zhu --- drivers/gpu/drm/amd/include/amd_shared.h | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/include/amd_shared.h b/drivers/gpu/drm/amd/include/amd_shared.h index 33de330..b178176 100644 --- a/drivers/gpu/drm/amd/include/amd_shared

[PATCH 5/6] drm/amdgpu: Enable static cg for vcn on RV

Signed-off-by: Rex Zhu --- drivers/gpu/drm/amd/amdgpu/soc15.c| 3 ++- drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c | 16 2 files changed, 10 insertions(+), 9 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c b/drivers/gpu/drm/amd/amdgpu/soc15.c index 8ccbcf9..485cb43 1

[PATCH 2/6] drm/amd/pp: Add smu support for vcn cg/pg on RV

Signed-off-by: Rex Zhu --- drivers/gpu/drm/amd/powerplay/hwmgr/smu10_hwmgr.c | 25 ++- 1 file changed, 24 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/smu10_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/smu10_hwmgr.c index 32b1524..436326b 10

[PATCH 1/6] drm/amdgpu: skip CG for VCN when late_init/fini

Signed-off-by: Rex Zhu --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index fcd4bb2..25bee46 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b

Re: [PATCH 2/2] drm/scheduler: Remove obsolete spinlock.

Am 16.05.2018 um 14:28 schrieb Lucas Stach: Am Mittwoch, den 16.05.2018, 14:08 +0200 schrieb Christian König: Yes, exactly. For normal user space command submission we should have tons of locks guaranteeing that (e.g. just the VM lock should do). For kernel moves we have the mutex for the GTT

Re: [PATCH 2/2] drm/scheduler: Remove obsolete spinlock.

Am Mittwoch, den 16.05.2018, 14:08 +0200 schrieb Christian König: > Yes, exactly. > > For normal user space command submission we should have tons of > locks  > guaranteeing that (e.g. just the VM lock should do). > > For kernel moves we have the mutex for the GTT windows which protects > it. >

Re: [PATCH 27/57] drm/amdgpu: Add vega20 soc init sequence on emulator (v3)

On Wed, May 16, 2018 at 2:11 PM, Grazvydas Ignotas wrote: > On Tue, May 15, 2018 at 5:59 PM, Alex Deucher wrote: >> From: Shaoyun Liu >> >> v2: cleanups (Alex) >> v3: make it vega20 only (Alex) >> >> Signed-off-by: Shaoyun Liu >> Acked-by: Alex Deucher >> Signed-off-by: Alex Deucher >> --- >>

Re: [PATCH 2/2] drm/scheduler: Remove obsolete spinlock.

Yes, exactly. For normal user space command submission we should have tons of locks guaranteeing that (e.g. just the VM lock should do). For kernel moves we have the mutex for the GTT windows which protects it. The could be problems with the UVD/VCE queues to cleanup the handles when an appl

Re: [PATCH 2/2] drm/scheduler: Remove obsolete spinlock.

So are you saying that you expect this to  be already in code for any usage of drm_sched_fence_create and drm_sched_entity_push_job ? lock() drm_sched_fence_create() ... (some code) drm_sched_entity_push_job() unlock() Andrey On 05/16/2018 07:23 AM, Christian König wrote: drm_sched_fence_

Re: [PATCH 2/2] drm/scheduler: Remove obsolete spinlock.

drm_sched_fence_create() assigns a sequence number to the fence it creates. Now drm_sched_fence_create() is called by drm_sched_job_init() to initialize the jobs we want to push on the scheduler queue. When you now call drm_sched_entity_push_job() without a protecting lock it can happen that

Re: [PATCH 2/2] drm/scheduler: Remove obsolete spinlock.

Can you please elaborate more, maybe give an example - I don't understand yet the problematic scenario. Andrey On 05/16/2018 02:50 AM, Christian König wrote: No, that spinlock is indeed incorrect. I See even when we protect the spsc queue with a spinlock that doesn't make it correct. It can

Re: [PATCH 27/57] drm/amdgpu: Add vega20 soc init sequence on emulator (v3)

On Tue, May 15, 2018 at 5:59 PM, Alex Deucher wrote: > From: Shaoyun Liu > > v2: cleanups (Alex) > v3: make it vega20 only (Alex) > > Signed-off-by: Shaoyun Liu > Acked-by: Alex Deucher > Signed-off-by: Alex Deucher > --- > drivers/gpu/drm/amd/amdgpu/emu_soc.c | 10091 > +

Re: [PATCH 29/57] drm/amdgpu: Add vega20 to dc support check

On Tue, May 15, 2018 at 5:59 PM, Alex Deucher wrote: > From: Feifei Xu > > Signed-off-by: Feifei Xu > Reviewed-by: Alex Deucher > Reviewed-by: Hawking Zhang > Signed-off-by: Alex Deucher > --- > drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 1 + > 1 file changed, 1 insertion(+) > > diff --git

Re: [PATCH 30/57] drm/amd: Add dce-12.1 gpio aux registers

On Tue, May 15, 2018 at 5:59 PM, Alex Deucher wrote: > From: Roman Li > > Updating dce12 register headers by adding dc registers > required for potential DP LTTPR support. > > Signed-off-by: Roman Li > Acked-by: Alex Deucher > Signed-off-by: Alex Deucher > --- > .../drm/amd/include/asic_reg/d

Re: [PATCH 04/21] drm/amdgpu: Add GFXv9 kfd2kgd interface functions

On Wed, May 16, 2018 at 1:08 AM, Felix Kuehling wrote: > On 2018-05-15 05:41 AM, Dave Airlie wrote: >>> +static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id, >>> + uint32_t queue_id, uint32_t __user *wptr, >>> + uint32_t wptr_shift,

[PATCH] drm/amdgpu: conditionally compile amdgpu's amdkfd files

In case CONFIG_HSA_AMD is not chosen, there is no need to compile amdkfd files that reside inside amdgpu dirver. In addition, because amdkfd depends on x86_64 architecture and amdgpu is not, compiling amdkfd files under i386 architecture can cause compiler errors and warnings. This patch modifies

Re: [PATCH 00/57] Vega20 support

Patches #1, #15, #18, #30, #33-#36, #39, #53-#54, #56 are Acked-by: Christian König . Patches #2-#14, #16-#17, #20-#25, #28, #31, #40, #43, #55, #57 are Reviewed-by: Christian König . Patch #19 still has a "//TODO: Need to update this for vega20" comment in the middle of the code. Should p

Re: [PATCH 1/2] drm/amdgpu: Skip drm_sched_entity realted ops for KIQ ring.

Am 15.05.2018 um 21:42 schrieb Alex Deucher: On Tue, May 15, 2018 at 3:31 PM, Andrey Grodzovsky wrote: Follwoing change 75fbed2 we need to skip KIQ ring when iterating amdgpu_ctx's scheduler entites. Signed-off-by: Andrey Grodzovsky Typo in the title: realted -> related Typo in the descripti