AMD-GFX Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: "Khatri, Sunil" <sukhatri@amd.com>
To: "Christian König" <christian.koenig@amd.com>,
	"Sunil Khatri" <sunil.khatri@amd.com>,
	"Alex Deucher" <alexander.deucher@amd.com>
Cc: amd-gfx@lists.freedesktop.org, Liang@rtg-sunil-navi33.amd.com,
	Prike <Prike.Liang@amd.com>
Subject: Re: [PATCH] drm/amdgpu/userq: add mutex lock/unlocked version of amdgpu_userq_put
Date: Fri, 20 Mar 2026 17:40:55 +0530	[thread overview]
Message-ID: <8251843c-db9b-4841-a6d8-807bbd4e5fe6@amd.com> (raw)
In-Reply-To: <df13f705-78c5-4e3a-86bf-8fc969b30d5c@amd.com>

Sent a different patch to fix the irrelevant get/put.
drm/amdgpu/userq: cleanup amdgpu_userq_get/put where not needed

regards
Sunil khatri

On 20-03-2026 04:44 pm, Khatri, Sunil wrote:
>
> On 20-03-2026 04:27 pm, Christian König wrote:
>> On 3/20/26 11:53, Khatri, Sunil wrote:
>>> On 20-03-2026 03:43 pm, Christian König wrote:
>>>> On 3/20/26 11:02, Khatri, Sunil wrote:
>>>>> On 20-03-2026 03:16 pm, Christian König wrote:
>>>>>> On 3/20/26 10:41, Sunil Khatri wrote:
>>>>>>> There is a possibility of deadlock when last reference to a 
>>>>>>> queue is
>>>>>>> put in certain situations where mutex is already help when calling
>>>>>>> the amdgpu_userq_destroy.
>>>>>>    From functions
>>>>>> As far as I can see that is illegal to begin with. Why are we 
>>>>>> doing that?
>>>>> This is to fix the deadlock that prike shared and many other 
>>>>> places where deadlock could still be caused.
>>>>> There is a possibility of amdgpu_userq_put being called for last 
>>>>> reference from amdgpu_userq_restore_worker or 
>>>>> amdgpu_eviction_fence_suspend_worker via amdgpu_evf_mgr_shutdown
>>>>> and all these functions already hold the userq_mutex and on last 
>>>>> reference when destroy is called it again takes userq_mutex and 
>>>>> causing deadlock.
>>>>>
>>>>> Thats why when we are dropping the reference we pass the 
>>>>> information of the handled could be called with lock already taken 
>>>>> and hence the handling.
>>>> Well that sounds like completely broken handling.
>>>>
>>>> Why are dropping an userqueu reference while holding the lock in 
>>>> the first place?
>>> we are doing at withing the locked state in most of the place in the 
>>> code.
>>> few examples:
>>> static void amdgpu_userq_restore_worker(struct work_struct *work)
>>> {
>>>      struct amdgpu_userq_mgr *uq_mgr = work_to_uq_mgr(work, 
>>> resume_work.work);
>>>      struct amdgpu_fpriv *fpriv = uq_mgr_to_fpriv(uq_mgr);
>>>      struct dma_fence *ev_fence;
>>>      int ret;
>>>
>>>      mutex_lock(&uq_mgr->userq_mutex);
>>>      ev_fence = amdgpu_evf_mgr_get_fence(&fpriv->evf_mgr);
>>>      if (!dma_fence_is_signaled(ev_fence))
>>>          goto unlock;
>>>
>>>      ret = amdgpu_userq_vm_validate(uq_mgr);
>>>      if (ret) {
>>>          drm_file_err(uq_mgr->file, "Failed to validate BOs to 
>>> restore\n");
>>>          goto unlock;
>>>      }
>>>
>>>      // Here the restore all is going through all the queues one bye 
>>> one by doing get and put
>> That is ok, the question is why the heck is doing that get/put? See 
>> we need get/put because we are *not* holding the lock.
>>
>> When we are holding the lock no get/put is needed at all.
> Got your point. In that case there is a bug in code and this could be 
> resolved with clean up or doing get/put where queue mutex is not locked.
> Will send the update here.
>
> Regards
> Sunil Khatri
>>
>> Regards,
>> Christian.
>>
>>> and doing the restore of each queue. Now during put if its last 
>>> reference due to race with another thread in putting we will
>>> call the destroy with locks taken which causes deadlock.
>>>      ret = amdgpu_userq_restore_all(uq_mgr);
>>>      if (ret) {
>>>          drm_file_err(uq_mgr->file, "Failed to restore all queues\n");
>>>          goto unlock;
>>>      }
>>>
>>> unlock:
>>>      mutex_unlock(&uq_mgr->userq_mutex);
>>>      dma_fence_put(ev_fence);
>>> }
>>>
>>> Another example:
>>> static void
>>> amdgpu_eviction_fence_suspend_worker(struct work_struct *work)
>>> {
>>>      struct amdgpu_eviction_fence_mgr *evf_mgr =
>>>          container_of(work, struct amdgpu_eviction_fence_mgr,
>>>                   suspend_work);
>>>      struct amdgpu_fpriv *fpriv =
>>>          container_of(evf_mgr, struct amdgpu_fpriv, evf_mgr);
>>>      struct amdgpu_userq_mgr *uq_mgr = &fpriv->userq_mgr;
>>>      struct dma_fence *ev_fence;
>>>      bool cookie;
>>>
>>>      mutex_lock(&uq_mgr->userq_mutex);
>>>
>>>      /*
>>>       * This is intentionally after taking the userq_mutex since we do
>>>       * allocate memory while holding this lock, but only after 
>>> ensuring that
>>>       * the eviction fence is signaled.
>>>       */
>>>      cookie = dma_fence_begin_signalling();
>>>
>>>      ev_fence = amdgpu_evf_mgr_get_fence(evf_mgr);
>>>
>>>
>>> /* Here in userq evict we do a ref get and put a various places 
>>> while the mutex is already taken */
>>>
>>>      amdgpu_userq_evict(uq_mgr, !evf_mgr->shutdown);
>>>
>>>      /*
>>>       * Signaling the eviction fence must be done while holding the
>>>       * userq_mutex. Otherwise we won't resume the queues before 
>>> issuing the
>>>       * next fence.
>>>       */
>>>      dma_fence_signal(ev_fence);
>>>      dma_fence_end_signalling(cookie);
>>>      dma_fence_put(ev_fence);
>>>      mutex_unlock(&uq_mgr->userq_mutex);
>>>
>>> }
>>>
>>> Regards
>>> Sunil.
>>>
>>>
>>>> Regards,
>>>> Christian.
>>>>
>>>>> Regards
>>>>> Sunil khatri
>>>>>
>>>>>> Regards,
>>>>>> Christian.
>>>>>>
>>>>>>> So based on the thread where it could be
>>>>>>> locked we pass the locked information in the destroy functionality
>>>>>>> to avoid taking the lock again.
>>>>>>>
>>>>>>> Cc: Liang, Prike <Prike.Liang@amd.com>
>>>>>>> Suggested-by: Liang, Prike <Prike.Liang@amd.com>
>>>>>>> Signed-off-by: Sunil Khatri <sunil.khatri@amd.com>
>>>>>>> ---
>>>>>>>     drivers/gpu/drm/amd/amdgpu/amdgpu_userq.c     | 52 
>>>>>>> +++++++++++++------
>>>>>>>     drivers/gpu/drm/amd/amdgpu/amdgpu_userq.h     |  2 +-
>>>>>>>     .../gpu/drm/amd/amdgpu/amdgpu_userq_fence.c   |  4 +-
>>>>>>>     3 files changed, 40 insertions(+), 18 deletions(-)
>>>>>>>
>>>>>>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.c 
>>>>>>> b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.c
>>>>>>> index ced9ade44be4..9482664e9c2c 100644
>>>>>>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.c
>>>>>>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.c
>>>>>>> @@ -617,13 +617,17 @@ amdgpu_userq_get_doorbell_index(struct 
>>>>>>> amdgpu_userq_mgr *uq_mgr,
>>>>>>>     }
>>>>>>>       static int
>>>>>>> -amdgpu_userq_destroy(struct amdgpu_userq_mgr *uq_mgr, struct 
>>>>>>> amdgpu_usermode_queue *queue)
>>>>>>> +amdgpu_userq_destroy(struct amdgpu_userq_mgr *uq_mgr, struct 
>>>>>>> amdgpu_usermode_queue *queue,
>>>>>>> +             bool locked)
>>>>>>>     {
>>>>>>>         struct amdgpu_device *adev = uq_mgr->adev;
>>>>>>>         int r = 0;
>>>>>>>     - cancel_delayed_work_sync(&uq_mgr->resume_work);
>>>>>>> +    /* It safe to unlock since we are in destroy and the queue 
>>>>>>> ref is only this */
>>>>>>> +    if (locked)
>>>>>>> +        mutex_unlock(&uq_mgr->userq_mutex);
>>>>>>>     + cancel_delayed_work_sync(&uq_mgr->resume_work);
>>>>>>>         /* Cancel any pending hang detection work and cleanup */
>>>>>>> cancel_delayed_work_sync(&queue->hang_detect_work);
>>>>>>>     @@ -657,13 +661,27 @@ amdgpu_userq_destroy(struct 
>>>>>>> amdgpu_userq_mgr *uq_mgr, struct amdgpu_usermode_que
>>>>>>>             queue->state = AMDGPU_USERQ_STATE_HUNG;
>>>>>>>         }
>>>>>>>         amdgpu_userq_cleanup(queue);
>>>>>>> -    mutex_unlock(&uq_mgr->userq_mutex);
>>>>>>> +
>>>>>>> +    if (!locked)
>>>>>>> +        mutex_unlock(&uq_mgr->userq_mutex);
>>>>>>> pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
>>>>>>>           return r;
>>>>>>>     }
>>>>>>>     +static void amdgpu_userq_kref_destroy_locked(struct kref 
>>>>>>> *kref)
>>>>>>> +{
>>>>>>> +    int r;
>>>>>>> +    struct amdgpu_usermode_queue *queue =
>>>>>>> +        container_of(kref, struct amdgpu_usermode_queue, 
>>>>>>> refcount);
>>>>>>> +    struct amdgpu_userq_mgr *uq_mgr = queue->userq_mgr;
>>>>>>> +
>>>>>>> +    r = amdgpu_userq_destroy(uq_mgr, queue, true);
>>>>>>> +    if (r)
>>>>>>> +        drm_file_err(uq_mgr->file, "Failed to destroy usermode 
>>>>>>> queue %d\n", r);
>>>>>>> +}
>>>>>>> +
>>>>>>>     static void amdgpu_userq_kref_destroy(struct kref *kref)
>>>>>>>     {
>>>>>>>         int r;
>>>>>>> @@ -671,7 +689,7 @@ static void amdgpu_userq_kref_destroy(struct 
>>>>>>> kref *kref)
>>>>>>>             container_of(kref, struct amdgpu_usermode_queue, 
>>>>>>> refcount);
>>>>>>>         struct amdgpu_userq_mgr *uq_mgr = queue->userq_mgr;
>>>>>>>     -    r = amdgpu_userq_destroy(uq_mgr, queue);
>>>>>>> +    r = amdgpu_userq_destroy(uq_mgr, queue, false);
>>>>>>>         if (r)
>>>>>>>             drm_file_err(uq_mgr->file, "Failed to destroy 
>>>>>>> usermode queue %d\n", r);
>>>>>>>     }
>>>>>>> @@ -689,10 +707,14 @@ struct amdgpu_usermode_queue 
>>>>>>> *amdgpu_userq_get(struct amdgpu_userq_mgr *uq_mgr,
>>>>>>>         return queue;
>>>>>>>     }
>>>>>>>     -void amdgpu_userq_put(struct amdgpu_usermode_queue *queue)
>>>>>>> +void amdgpu_userq_put(struct amdgpu_usermode_queue *queue, bool 
>>>>>>> locked)
>>>>>>>     {
>>>>>>> -    if (queue)
>>>>>>> -        kref_put(&queue->refcount, amdgpu_userq_kref_destroy);
>>>>>>> +    if (queue) {
>>>>>>> +        if (locked)
>>>>>>> +            kref_put(&queue->refcount, 
>>>>>>> amdgpu_userq_kref_destroy_locked);
>>>>>>> +        else
>>>>>>> +            kref_put(&queue->refcount, amdgpu_userq_kref_destroy);
>>>>>>> +    }
>>>>>>>     }
>>>>>>>       static int amdgpu_userq_priority_permit(struct drm_file 
>>>>>>> *filp,
>>>>>>> @@ -978,7 +1000,7 @@ int amdgpu_userq_ioctl(struct drm_device 
>>>>>>> *dev, void *data,
>>>>>>>             if (!queue)
>>>>>>>                 return -ENOENT;
>>>>>>>     -        amdgpu_userq_put(queue);
>>>>>>> +        amdgpu_userq_put(queue, false);
>>>>>>>             break;
>>>>>>>         }
>>>>>>>     @@ -1007,7 +1029,7 @@ amdgpu_userq_restore_all(struct 
>>>>>>> amdgpu_userq_mgr *uq_mgr)
>>>>>>>                 drm_file_err(uq_mgr->file,
>>>>>>>                          "trying restore queue without va 
>>>>>>> mapping\n");
>>>>>>>                 queue->state = AMDGPU_USERQ_STATE_INVALID_VA;
>>>>>>> -            amdgpu_userq_put(queue);
>>>>>>> +            amdgpu_userq_put(queue, true);
>>>>>>>                 continue;
>>>>>>>             }
>>>>>>>     @@ -1015,7 +1037,7 @@ amdgpu_userq_restore_all(struct 
>>>>>>> amdgpu_userq_mgr *uq_mgr)
>>>>>>>             if (r)
>>>>>>>                 ret = r;
>>>>>>>     -        amdgpu_userq_put(queue);
>>>>>>> +        amdgpu_userq_put(queue, true);
>>>>>>>         }
>>>>>>>           if (ret)
>>>>>>> @@ -1258,7 +1280,7 @@ amdgpu_userq_evict_all(struct 
>>>>>>> amdgpu_userq_mgr *uq_mgr)
>>>>>>>             r = amdgpu_userq_preempt_helper(queue);
>>>>>>>             if (r)
>>>>>>>                 ret = r;
>>>>>>> -        amdgpu_userq_put(queue);
>>>>>>> +        amdgpu_userq_put(queue, true);
>>>>>>>         }
>>>>>>>           if (ret)
>>>>>>> @@ -1298,17 +1320,17 @@ amdgpu_userq_wait_for_signal(struct 
>>>>>>> amdgpu_userq_mgr *uq_mgr)
>>>>>>>             struct dma_fence *f = queue->last_fence;
>>>>>>>               if (!f || dma_fence_is_signaled(f)) {
>>>>>>> -            amdgpu_userq_put(queue);
>>>>>>> +            amdgpu_userq_put(queue, true);
>>>>>>>                 continue;
>>>>>>>             }
>>>>>>>             ret = dma_fence_wait_timeout(f, true, 
>>>>>>> msecs_to_jiffies(100));
>>>>>>>             if (ret <= 0) {
>>>>>>>                 drm_file_err(uq_mgr->file, "Timed out waiting 
>>>>>>> for fence=%llu:%llu\n",
>>>>>>>                          f->context, f->seqno);
>>>>>>> -            amdgpu_userq_put(queue);
>>>>>>> +            amdgpu_userq_put(queue, true);
>>>>>>>                 return -ETIMEDOUT;
>>>>>>>             }
>>>>>>> -        amdgpu_userq_put(queue);
>>>>>>> +        amdgpu_userq_put(queue, true);
>>>>>>>         }
>>>>>>>           return 0;
>>>>>>> @@ -1366,7 +1388,7 @@ void amdgpu_userq_mgr_fini(struct 
>>>>>>> amdgpu_userq_mgr *userq_mgr)
>>>>>>>             if (!queue)
>>>>>>>                 break;
>>>>>>>     -        amdgpu_userq_put(queue);
>>>>>>> +        amdgpu_userq_put(queue, false);
>>>>>>>         }
>>>>>>>           xa_destroy(&userq_mgr->userq_xa);
>>>>>>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.h 
>>>>>>> b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.h
>>>>>>> index f0abc16d02cc..2a496e74ec6a 100644
>>>>>>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.h
>>>>>>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.h
>>>>>>> @@ -116,7 +116,7 @@ struct amdgpu_db_info {
>>>>>>>     };
>>>>>>>       struct amdgpu_usermode_queue *amdgpu_userq_get(struct 
>>>>>>> amdgpu_userq_mgr *uq_mgr, u32 qid);
>>>>>>> -void amdgpu_userq_put(struct amdgpu_usermode_queue *queue);
>>>>>>> +void amdgpu_userq_put(struct amdgpu_usermode_queue *queue, bool 
>>>>>>> locked);
>>>>>>>       int amdgpu_userq_ioctl(struct drm_device *dev, void *data, 
>>>>>>> struct drm_file *filp);
>>>>>>>     diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.c 
>>>>>>> b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.c
>>>>>>> index 18390d37a7e0..10e08cb6bd13 100644
>>>>>>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.c
>>>>>>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.c
>>>>>>> @@ -612,7 +612,7 @@ int amdgpu_userq_signal_ioctl(struct 
>>>>>>> drm_device *dev, void *data,
>>>>>>>         kfree(syncobj_handles);
>>>>>>>           if (queue)
>>>>>>> -        amdgpu_userq_put(queue);
>>>>>>> +        amdgpu_userq_put(queue, false);
>>>>>>>           return r;
>>>>>>>     }
>>>>>>> @@ -914,7 +914,7 @@ amdgpu_userq_wait_return_fence_info(struct 
>>>>>>> drm_file *filp,
>>>>>>>             r = 0;
>>>>>>>       put_waitq:
>>>>>>> -    amdgpu_userq_put(waitq);
>>>>>>> +    amdgpu_userq_put(waitq, false);
>>>>>>>       free_fences:
>>>>>>>         while (num_fences--)

  reply	other threads:[~2026-03-20 12:11 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-03-20  9:41 [PATCH] drm/amdgpu/userq: add mutex lock/unlocked version of amdgpu_userq_put Sunil Khatri
2026-03-20  9:46 ` Christian König
2026-03-20 10:02   ` Khatri, Sunil
2026-03-20 10:13     ` Christian König
2026-03-20 10:53       ` Khatri, Sunil
2026-03-20 10:57         ` Christian König
2026-03-20 11:14           ` Khatri, Sunil
2026-03-20 12:10             ` Khatri, Sunil [this message]
  -- strict thread matches above, loose matches on Subject: below --
2026-03-20  9:43 Sunil Khatri

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=8251843c-db9b-4841-a6d8-807bbd4e5fe6@amd.com \
    --to=sukhatri@amd.com \
    --cc=Liang@rtg-sunil-navi33.amd.com \
    --cc=Prike.Liang@amd.com \
    --cc=alexander.deucher@amd.com \
    --cc=amd-gfx@lists.freedesktop.org \
    --cc=christian.koenig@amd.com \
    --cc=sunil.khatri@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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox