All of lore.kernel.org
 help / color / mirror / Atom feed
From: Andrey Grodzovsky <Andrey.Grodzovsky@amd.com>
To: Nirmoy <nirmodas@amd.com>,
	20200311195759.49332-1-nirmoy.das@amd.com,
	amd-gfx@lists.freedesktop.org
Cc: Boyuan.Zhang@amd.com, nirmoy.das@amd.com,
	alexander.deucher@amd.com, James.Zhu@amd.com, Leo.Liu@amd.com,
	christian.koenig@amd.com
Subject: Re: [PATCH 1/1] drm/amdgpu: disable gpu_sched load balancer for vcn jobs
Date: Wed, 11 Mar 2020 16:35:49 -0400	[thread overview]
Message-ID: <ee63269d-8c95-cbb2-c64d-55a5061feb44@amd.com> (raw)
In-Reply-To: <4ee422df-5b57-ea1e-ae42-0ef702a657a8@amd.com>


On 3/11/20 4:32 PM, Nirmoy wrote:
>
> On 3/11/20 9:02 PM, Andrey Grodzovsky wrote:
>>
>> On 3/11/20 4:00 PM, Andrey Grodzovsky wrote:
>>>
>>> On 3/11/20 4:00 PM, Nirmoy Das wrote:
>>>> VCN HW  doesn't support dynamic load balance on multiple
>>>> instances for a context. This patch modifies entity's
>>>> sched_list to a sched_list consist of only one drm scheduler.
>>>>
>>>> Signed-off-by: Nirmoy Das <nirmoy.das@amd.com>
>>>> ---
>>>>   drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c   |  4 ++++
>>>>   drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c  | 13 +++++++++++++
>>>>   drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.h  |  1 +
>>>>   drivers/gpu/drm/amd/amdgpu/amdgpu_job.c  |  3 +++
>>>>   drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h |  1 +
>>>>   drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c    |  2 ++
>>>>   drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c    |  2 ++
>>>>   drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c    |  2 ++
>>>>   8 files changed, 28 insertions(+)
>>>>
>>>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c 
>>>> b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
>>>> index 8304d0c87899..db0eef19c636 100644
>>>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
>>>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
>>>> @@ -1203,6 +1203,7 @@ static int amdgpu_cs_submit(struct 
>>>> amdgpu_cs_parser *p,
>>>>                   union drm_amdgpu_cs *cs)
>>>>   {
>>>>       struct amdgpu_fpriv *fpriv = p->filp->driver_priv;
>>>> +    struct amdgpu_ring *ring = to_amdgpu_ring(p->entity->rq->sched);
>>>>       struct drm_sched_entity *entity = p->entity;
>>>>       enum drm_sched_priority priority;
>>>>       struct amdgpu_bo_list_entry *e;
>>>> @@ -1257,6 +1258,9 @@ static int amdgpu_cs_submit(struct 
>>>> amdgpu_cs_parser *p,
>>>>       priority = job->base.s_priority;
>>>>       drm_sched_entity_push_job(&job->base, entity);
>>>>   +    if (ring->funcs->no_gpu_sched_loadbalance)
>>>> +        amdgpu_ctx_disable_gpu_sched_load_balance(entity);
>>>> +
>>>
>>>
>>> Why this needs to be done each time a job is submitted and not once 
>>> in drm_sched_entity_init (same foramdgpu_job_submit bellow ?)
>>>
>>> Andrey
>>
>>
>> My bad - not in drm_sched_entity_init but in relevant amdgpu code.
>
>
> Hi Andrey,
>
> Do you mean drm_sched_job_init() or after creating VCN entities?
>
>
> Nirmoy


I guess after creating the VCN entities (has to be amdgpu specific code) 
- I just don't get why it needs to be done each time job is submitted, I 
mean - since you set .no_gpu_sched_loadbalance = true anyway this is 
always true and so shouldn't you just initialize the VCN entity with a 
schedulers list consisting of one scheduler and that it ?

Andrey


>
>>
>> Andrey
>>
>>
>>>
>>>
>>>> amdgpu_vm_move_to_lru_tail(p->adev, &fpriv->vm);
>>>>         ttm_eu_fence_buffer_objects(&p->ticket, &p->validated, 
>>>> p->fence);
>>>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c 
>>>> b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
>>>> index fa575bdc03c8..1127e8f77721 100644
>>>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
>>>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
>>>> @@ -559,6 +559,19 @@ void amdgpu_ctx_priority_override(struct 
>>>> amdgpu_ctx *ctx,
>>>>       }
>>>>   }
>>>>   +/**
>>>> + * amdgpu_ctx_disable_gpu_sched_load_balance - disable gpu_sched's 
>>>> load balancer
>>>> + * @entity: entity on which load balancer will be disabled
>>>> + */
>>>> +void amdgpu_ctx_disable_gpu_sched_load_balance(struct 
>>>> drm_sched_entity *entity)
>>>> +{
>>>> +    struct drm_gpu_scheduler **scheds = &entity->rq->sched;
>>>> +
>>>> +    /* disable gpu_sched's job load balancer by assigning only one */
>>>> +    /* drm scheduler to the entity */
>>>> +    drm_sched_entity_modify_sched(entity, scheds, 1);
>>>> +}
>>>> +
>>>>   int amdgpu_ctx_wait_prev_fence(struct amdgpu_ctx *ctx,
>>>>                      struct drm_sched_entity *entity)
>>>>   {
>>>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.h 
>>>> b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.h
>>>> index de490f183af2..3a2f900b8000 100644
>>>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.h
>>>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.h
>>>> @@ -90,5 +90,6 @@ void amdgpu_ctx_mgr_fini(struct amdgpu_ctx_mgr 
>>>> *mgr);
>>>>     void amdgpu_ctx_init_sched(struct amdgpu_device *adev);
>>>>   +void amdgpu_ctx_disable_gpu_sched_load_balance(struct 
>>>> drm_sched_entity *entity);
>>>>     #endif
>>>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c 
>>>> b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
>>>> index 4981e443a884..64dad7ba74da 100644
>>>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
>>>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
>>>> @@ -140,6 +140,7 @@ void amdgpu_job_free(struct amdgpu_job *job)
>>>>   int amdgpu_job_submit(struct amdgpu_job *job, struct 
>>>> drm_sched_entity *entity,
>>>>                 void *owner, struct dma_fence **f)
>>>>   {
>>>> +    struct amdgpu_ring *ring = to_amdgpu_ring(entity->rq->sched);
>>>>       enum drm_sched_priority priority;
>>>>       int r;
>>>>   @@ -154,6 +155,8 @@ int amdgpu_job_submit(struct amdgpu_job *job, 
>>>> struct drm_sched_entity *entity,
>>>>       amdgpu_job_free_resources(job);
>>>>       priority = job->base.s_priority;
>>>>       drm_sched_entity_push_job(&job->base, entity);
>>>> +    if (ring->funcs->no_gpu_sched_loadbalance)
>>>> +        amdgpu_ctx_disable_gpu_sched_load_balance(entity);
>>>>         return 0;
>>>>   }
>>>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h 
>>>> b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
>>>> index 448c76cbf3ed..f78fe1a6912b 100644
>>>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
>>>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
>>>> @@ -115,6 +115,7 @@ struct amdgpu_ring_funcs {
>>>>       u32            nop;
>>>>       bool            support_64bit_ptrs;
>>>>       bool            no_user_fence;
>>>> +    bool            no_gpu_sched_loadbalance;
>>>>       unsigned        vmhub;
>>>>       unsigned        extra_dw;
>>>>   diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c 
>>>> b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
>>>> index 71f61afdc655..749ccdb5fbfb 100644
>>>> --- a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
>>>> +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
>>>> @@ -1871,6 +1871,7 @@ static const struct amdgpu_ring_funcs 
>>>> vcn_v1_0_dec_ring_vm_funcs = {
>>>>       .align_mask = 0xf,
>>>>       .support_64bit_ptrs = false,
>>>>       .no_user_fence = true,
>>>> +    .no_gpu_sched_loadbalance = true,
>>>>       .vmhub = AMDGPU_MMHUB_0,
>>>>       .get_rptr = vcn_v1_0_dec_ring_get_rptr,
>>>>       .get_wptr = vcn_v1_0_dec_ring_get_wptr,
>>>> @@ -1905,6 +1906,7 @@ static const struct amdgpu_ring_funcs 
>>>> vcn_v1_0_enc_ring_vm_funcs = {
>>>>       .nop = VCN_ENC_CMD_NO_OP,
>>>>       .support_64bit_ptrs = false,
>>>>       .no_user_fence = true,
>>>> +    .no_gpu_sched_loadbalance = true,
>>>>       .vmhub = AMDGPU_MMHUB_0,
>>>>       .get_rptr = vcn_v1_0_enc_ring_get_rptr,
>>>>       .get_wptr = vcn_v1_0_enc_ring_get_wptr,
>>>> diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c 
>>>> b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c
>>>> index f2745fd1ddb3..c48423b54bc5 100644
>>>> --- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c
>>>> +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c
>>>> @@ -1954,6 +1954,7 @@ static const struct amd_ip_funcs 
>>>> vcn_v2_0_ip_funcs = {
>>>>   static const struct amdgpu_ring_funcs vcn_v2_0_dec_ring_vm_funcs = {
>>>>       .type = AMDGPU_RING_TYPE_VCN_DEC,
>>>>       .align_mask = 0xf,
>>>> +    .no_gpu_sched_loadbalance = true,
>>>>       .vmhub = AMDGPU_MMHUB_0,
>>>>       .get_rptr = vcn_v2_0_dec_ring_get_rptr,
>>>>       .get_wptr = vcn_v2_0_dec_ring_get_wptr,
>>>> @@ -1984,6 +1985,7 @@ static const struct amdgpu_ring_funcs 
>>>> vcn_v2_0_dec_ring_vm_funcs = {
>>>>   static const struct amdgpu_ring_funcs vcn_v2_0_enc_ring_vm_funcs = {
>>>>       .type = AMDGPU_RING_TYPE_VCN_ENC,
>>>>       .align_mask = 0x3f,
>>>> +    .no_gpu_sched_loadbalance = true,
>>>>       .nop = VCN_ENC_CMD_NO_OP,
>>>>       .vmhub = AMDGPU_MMHUB_0,
>>>>       .get_rptr = vcn_v2_0_enc_ring_get_rptr,
>>>> diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c 
>>>> b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c
>>>> index 9b22e2b55132..1cc8e1420fc8 100644
>>>> --- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c
>>>> +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c
>>>> @@ -1478,6 +1478,7 @@ static void vcn_v2_5_dec_ring_set_wptr(struct 
>>>> amdgpu_ring *ring)
>>>>   static const struct amdgpu_ring_funcs vcn_v2_5_dec_ring_vm_funcs = {
>>>>       .type = AMDGPU_RING_TYPE_VCN_DEC,
>>>>       .align_mask = 0xf,
>>>> +    .no_gpu_sched_loadbalance = true,
>>>>       .vmhub = AMDGPU_MMHUB_1,
>>>>       .get_rptr = vcn_v2_5_dec_ring_get_rptr,
>>>>       .get_wptr = vcn_v2_5_dec_ring_get_wptr,
>>>> @@ -1577,6 +1578,7 @@ static void vcn_v2_5_enc_ring_set_wptr(struct 
>>>> amdgpu_ring *ring)
>>>>   static const struct amdgpu_ring_funcs vcn_v2_5_enc_ring_vm_funcs = {
>>>>       .type = AMDGPU_RING_TYPE_VCN_ENC,
>>>>       .align_mask = 0x3f,
>>>> +    .no_gpu_sched_loadbalance = true,
>>>>       .nop = VCN_ENC_CMD_NO_OP,
>>>>       .vmhub = AMDGPU_MMHUB_1,
>>>>       .get_rptr = vcn_v2_5_enc_ring_get_rptr,
>>> _______________________________________________
>>> amd-gfx mailing list
>>> amd-gfx@lists.freedesktop.org
>>> https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.freedesktop.org%2Fmailman%2Flistinfo%2Famd-gfx&amp;data=02%7C01%7Candrey.grodzovsky%40amd.com%7Cc6958b3a536448ecd9bd08d7c5f6cfc9%7C3dd8961fe4884e608e11a82d994e183d%7C0%7C0%7C637195536148567145&amp;sdata=OCuOocxzDKPVCUKDUzqqbL3lKj4lYFyDR8Ly%2FtG3Gi0%3D&amp;reserved=0 
>>>
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

  reply	other threads:[~2020-03-11 20:35 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-03-11 20:00 [PATCH 1/1] drm/amdgpu: disable gpu_sched load balancer for vcn jobs Nirmoy Das
2020-03-11 20:00 ` Andrey Grodzovsky
2020-03-11 20:02   ` Andrey Grodzovsky
2020-03-11 20:32     ` Nirmoy
2020-03-11 20:35       ` Andrey Grodzovsky [this message]
2020-03-11 20:55         ` Nirmoy
2020-03-12  8:50           ` Christian König
2020-03-12 10:56             ` Nirmoy
2020-03-17 20:46               ` Luben Tuikov
2020-03-11 20:14 ` James Zhu
2020-03-11 20:36   ` Nirmoy
  -- strict thread matches above, loose matches on Subject: below --
2020-03-11 19:57 Nirmoy Das

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=ee63269d-8c95-cbb2-c64d-55a5061feb44@amd.com \
    --to=andrey.grodzovsky@amd.com \
    --cc=20200311195759.49332-1-nirmoy.das@amd.com \
    --cc=Boyuan.Zhang@amd.com \
    --cc=James.Zhu@amd.com \
    --cc=Leo.Liu@amd.com \
    --cc=alexander.deucher@amd.com \
    --cc=amd-gfx@lists.freedesktop.org \
    --cc=christian.koenig@amd.com \
    --cc=nirmodas@amd.com \
    --cc=nirmoy.das@amd.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.