Re: [PATCH 1/4] drm/scheduler: make sure timer is restarted

Patch 1 and 2 are Reviewed-by: Nayan Deshmukh On Wed, Oct 17, 2018 at 12:30 AM Grodzovsky, Andrey wrote: > > Patches 1-3 Reviewed-by: Andrey Grodzovsky > > Patch 4 Acked-by: Andrey Grodzovsky > > Andrey > > > On 10/16/2018 07:55 AM, Christian König wrote: > >

Re: [PATCH 1/2] drm/scheduler: modify API to avoid redundancy

Hi Daniel, On Thu, Aug 9, 2018 at 2:27 PM Daniel Vetter wrote: > > On Fri, Jul 20, 2018 at 2:21 PM, Nayan Deshmukh > wrote: > > entity has a scheduler field and we don't need the sched argument > > in any of the functions where entity is provided. > > &

Re: [PATCH] drm/amdgpu: fix the null pointer to get timeline by scheduler fence

On Wed, Aug 8, 2018 at 4:58 PM Huang Rui wrote: > On Wed, Aug 08, 2018 at 03:10:07PM +0800, Koenig, Christian wrote: > > Yeah that is a known issue, but this solution is not correct either. > > > > See the scheduler where the job is execute on is simply not determined > > yet when we want to trac

Re: [PATCH] drm/scheduler: select the least loaded sched during entity init

Ah...you are correct. We will reschedule on the first job push. I didn't take that into account. Let's drop this patch then. Thanks, Nayan On Fri, Aug 3, 2018, 4:12 PM Christian König < ckoenig.leichtzumer...@gmail.com> wrote: > Am 03.08.2018 um 09:06 schrieb Nayan Deshm

[PATCH] drm/scheduler: select the least loaded sched during entity init

Instead of assigning entity to the first scheduler in the list assign it to the least loaded scheduler. Signed-off-by: Nayan Deshmukh --- drivers/gpu/drm/scheduler/gpu_scheduler.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/scheduler/gpu_scheduler.c b

Re: [PATCH] drm/scheduler: fix setting the priorty for entities

On Wed, Aug 1, 2018 at 7:58 PM Christian König < ckoenig.leichtzumer...@gmail.com> wrote: > Since we now deal with multiple rq we need to update all of them, not > just the current one. > > Signed-off-by: Christian König > Acked-by: Nayan Deshmukh > --- >

Re: [PATCH 2/4] drm/scheduler: add counter for total jobs in scheduler

tion. :-) > > Thanks, > Ray > > > Christian. > > > > Am 01.08.2018 um 15:15 schrieb Huang Rui: > > > On Wed, Aug 01, 2018 at 01:50:00PM +0530, Nayan Deshmukh wrote: > > > > > > This should need a commmit message. > > > > > > Thanks, &

Re: [PATCH 2/3] drm/scheduler: add counter for total jobs in scheduler

On Thu, Aug 2, 2018 at 11:29 AM Zhang, Jerry (Junwei) wrote: > On 08/02/2018 01:50 PM, Nayan Deshmukh wrote: > > > > > > On Thu, Aug 2, 2018 at 10:31 AM Zhang, Jerry (Junwei) < > jerry.zh...@amd.com <mailto:jerry.zh...@amd.com>> wrote: > > > >

Re: [PATCH 3/4] drm/scheduler: add new function to get least loaded sched v2

on question - if the run queues belong to different > schedulers they effectively point to different rings, > > it means we allow to move (reschedule) a drm_sched_entity from one ring > to another - i assume that the idea int the first place, that > > you have a set of HW rings

Re: [PATCH 2/3] drm/scheduler: add counter for total jobs in scheduler

On Thu, Aug 2, 2018 at 10:31 AM Zhang, Jerry (Junwei) wrote: > On 07/12/2018 02:36 PM, Nayan Deshmukh wrote: > > Signed-off-by: Nayan Deshmukh > > --- > > drivers/gpu/drm/scheduler/gpu_scheduler.c | 3 +++ > > include/drm/gpu_scheduler.h | 2

Re: [PATCH 3/4] drm/scheduler: add new function to get least loaded sched v2

> to another - i assume that the idea int the first place, that > > you have a set of HW rings and you can utilize any of them for your jobs > (like compute rings). Correct ? > > Andrey > > > On 08/01/2018 04:20 AM, Nayan Deshmukh wrote: > > The function selects the

[PATCH 4/4] drm/scheduler: move idle entities to scheduler with less load v2

moving. v2: remove unused variable and an unecessary check Signed-off-by: Nayan Deshmukh --- drivers/gpu/drm/scheduler/gpu_scheduler.c | 22 ++ 1 file changed, 18 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/scheduler/gpu_scheduler.c b/drivers/gpu/drm/scheduler

[PATCH 2/4] drm/scheduler: add counter for total jobs in scheduler

Signed-off-by: Nayan Deshmukh --- drivers/gpu/drm/scheduler/gpu_scheduler.c | 3 +++ include/drm/gpu_scheduler.h | 2 ++ 2 files changed, 5 insertions(+) diff --git a/drivers/gpu/drm/scheduler/gpu_scheduler.c b/drivers/gpu/drm/scheduler/gpu_scheduler.c index a3eacc35cf98

[PATCH 3/4] drm/scheduler: add new function to get least loaded sched v2

The function selects the run queue from the rq_list with the least load. The load is decided by the number of jobs in a scheduler. v2: avoid using atomic read twice consecutively, instead store it locally Signed-off-by: Nayan Deshmukh --- drivers/gpu/drm/scheduler/gpu_scheduler.c | 25

[PATCH 1/4] drm/scheduler: add a list of run queues to the entity

These are the potential run queues on which the jobs from this entity can be scheduled. We will use this to do load balancing. Signed-off-by: Nayan Deshmukh --- drivers/gpu/drm/scheduler/gpu_scheduler.c | 8 include/drm/gpu_scheduler.h | 7 ++- 2 files changed, 14

Re: [PATCH 4/4] drm/scheduler: move idle entities to scheduler with less load

On Tue, Jul 31, 2018 at 5:02 PM Christian König < ckoenig.leichtzumer...@gmail.com> wrote: > Am 31.07.2018 um 12:37 schrieb Nayan Deshmukh: > > This is the first attempt to move entities between schedulers to > > have dynamic load balancing. We just move entities with n

Re: [PATCH 3/4] drm/scheduler: add new function to get least loaded sched

I will take care of that in v2. On Tue, Jul 31, 2018 at 4:57 PM Christian König < ckoenig.leichtzumer...@gmail.com> wrote: > Am 31.07.2018 um 12:37 schrieb Nayan Deshmukh: > > The function selects the run queue from the rq_list with the > > least load. The load is decided

[PATCH 3/4] drm/scheduler: add new function to get least loaded sched

The function selects the run queue from the rq_list with the least load. The load is decided by the number of jobs in a scheduler. Signed-off-by: Nayan Deshmukh --- drivers/gpu/drm/scheduler/gpu_scheduler.c | 26 ++ 1 file changed, 26 insertions(+) diff --git a/drivers

[PATCH 4/4] drm/scheduler: move idle entities to scheduler with less load

moving. Signed-off-by: Nayan Deshmukh --- drivers/gpu/drm/scheduler/gpu_scheduler.c | 25 - 1 file changed, 20 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/scheduler/gpu_scheduler.c b/drivers/gpu/drm/scheduler/gpu_scheduler.c index c67f65ad8f15..f665a84d48ef

[PATCH 2/4] drm/scheduler: add counter for total jobs in scheduler

Signed-off-by: Nayan Deshmukh --- drivers/gpu/drm/scheduler/gpu_scheduler.c | 3 +++ include/drm/gpu_scheduler.h | 2 ++ 2 files changed, 5 insertions(+) diff --git a/drivers/gpu/drm/scheduler/gpu_scheduler.c b/drivers/gpu/drm/scheduler/gpu_scheduler.c index a3eacc35cf98

[PATCH 1/4] drm/scheduler: add a list of run queues to the entity

These are the potential run queues on which the jobs from this entity can be scheduled. We will use this to do load balancing. Signed-off-by: Nayan Deshmukh --- drivers/gpu/drm/scheduler/gpu_scheduler.c | 8 include/drm/gpu_scheduler.h | 7 ++- 2 files changed, 14

[PATCH 0/4] drm/scheduler: add dynamic balancing

This is the my first attempt to include dynamic balancing as part of the scheduler. I have tried to handle the easy cases first to get the basic implementation running. Please share your thoughts!! *** BLURB HERE *** Nayan Deshmukh (4): drm/scheduler: add a list of run queues to the entity

Re: [PATCH] drm/scheduler: only kill entity if last user is killed

On Thu, Jul 26, 2018 at 5:16 PM Christian König < ckoenig.leichtzumer...@gmail.com> wrote: > Note which task is using the entity and only kill it if the last user of > the entity is killed. This should prevent problems when entities are > leaked to > child processes. > > Signed-off-by: Christian K

[PATCH 2/2] drm/scheduler: remove sched field from the entity

The scheduler of the entity is decided by the run queue on which it is queued. This patch avoids us the effort required to maintain a sync between rq and sched field when we start shifting entites among different rqs. Signed-off-by: Nayan Deshmukh --- drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c

[PATCH 1/2] drm/scheduler: modify API to avoid redundancy

entity has a scheduler field and we don't need the sched argument in any of the functions where entity is provided. Signed-off-by: Nayan Deshmukh --- drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c| 2 +- drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c | 13 + drivers/gpu/drm/amd/a

Re: [PATCH v2 1/2] drm/scheduler: add a pointer to scheduler in the rq

I forgot to add these: On Fri, Jul 13, 2018 at 3:21 PM Nayan Deshmukh wrote: > > This patch is in preparation for a better load balancing in > scheduler. It allows us to associate entities with the > run queues instead of binding them to a scheduler. > > Signed-off-by: Nayan Des

[PATCH v2 2/2] drm/scheduler: modify args of drm_sched_entity_init

replace run queue by a list of run queues and remove the sched arg as that is part of run queue itself Signed-off-by: Nayan Deshmukh Reviewed-by: Christian König --- drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c | 4 ++-- drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 3 +-- drivers/gpu/drm/amd

[PATCH v2 1/2] drm/scheduler: add a pointer to scheduler in the rq

This patch is in preparation for a better load balancing in scheduler. It allows us to associate entities with the run queues instead of binding them to a scheduler. Signed-off-by: Nayan Deshmukh --- drivers/gpu/drm/scheduler/gpu_scheduler.c | 6 -- include/drm/gpu_scheduler.h

Re: [PATCH 3/3] drm/scheduler: modify args of drm_sched_entity_init

On Thu, Jul 12, 2018 at 11:21 PM Eric Anholt wrote: > > Nayan Deshmukh writes: > > diff --git a/drivers/gpu/drm/scheduler/gpu_scheduler.c > > b/drivers/gpu/drm/scheduler/gpu_scheduler.c > > index 3dc1a4f07e3f..b2dbd1c1ba69 100644 > > --- a/drivers/gpu/drm/sched

Re: [PATCH 3/3] drm/scheduler: modify args of drm_sched_entity_init

On Thu, Jul 12, 2018 at 12:07 PM Nayan Deshmukh wrote: > > replace run queue by a list of run queues and remove the > sched arg as that is part of run queue itself > > Signed-off-by: Nayan Deshmukh > --- > drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c | 4 ++-- > dr

[PATCH 3/3] drm/scheduler: modify args of drm_sched_entity_init

replace run queue by a list of run queues and remove the sched arg as that is part of run queue itself Signed-off-by: Nayan Deshmukh --- drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c | 4 ++-- drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 4 ++-- drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c | 4

[PATCH 2/3] drm/scheduler: add counter for total jobs in scheduler

Signed-off-by: Nayan Deshmukh --- drivers/gpu/drm/scheduler/gpu_scheduler.c | 3 +++ include/drm/gpu_scheduler.h | 2 ++ 2 files changed, 5 insertions(+) diff --git a/drivers/gpu/drm/scheduler/gpu_scheduler.c b/drivers/gpu/drm/scheduler/gpu_scheduler.c index 429b1328653a

[PATCH 1/3] drm/scheduler: add a pointer to scheduler in the rq

Signed-off-by: Nayan Deshmukh --- drivers/gpu/drm/scheduler/gpu_scheduler.c | 6 -- include/drm/gpu_scheduler.h | 2 ++ 2 files changed, 6 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/scheduler/gpu_scheduler.c b/drivers/gpu/drm/scheduler/gpu_scheduler.c index

[PATCH 0/3] drm/scheduler: preparation for load balancing

, Nayan Deshmukh Nayan Deshmukh (3): drm/scheduler: add a pointer to scheduler in the rq drm/scheduler: add counter for total jobs in scheduler drm/scheduler: modify args of drm_sched_entity_init drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c | 4 ++-- drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c