[PATCH v2] drm/amdkfd: Fix circular locking dependency warning

Felix Kuehling felix.kuehling at amd.com
Wed Jun 24 16:56:05 UTC 2020


Am 2020-06-24 um 12:06 p.m. schrieb Mukul Joshi:
> Lockdep is spewing circular locking dependency warning when
> reading SDMA usage stats.
>
>   150.887733] ======================================================
> [  150.893903] WARNING: possible circular locking dependency detected
> [  150.905917] ------------------------------------------------------
> [  150.912129] kfdtest/4081 is trying to acquire lock:
> [  150.917002] ffff8f7f3762e118 (&mm->mmap_sem#2){++++}, at:
> 		__might_fault+0x3e/0x90
> [  150.924490]
>                but task is already holding lock:
> [  150.930320] ffff8f7f49d229e8 (&dqm->lock_hidden){+.+.}, at:
> 		destroy_queue_cpsch+0x29/0x210 [amdgpu]
> [  150.939432]
>                which lock already depends on the new lock.
> [  150.947603]
>                the existing dependency chain (in reverse order) is:
> [  150.955074]
>                -> #3 (&dqm->lock_hidden){+.+.}:
> [  150.960822]        __mutex_lock+0xa1/0x9f0
> [  150.964996]        evict_process_queues_cpsch+0x22/0x120 [amdgpu]
> [  150.971155]        kfd_process_evict_queues+0x3b/0xc0 [amdgpu]
> [  150.977054]        kgd2kfd_quiesce_mm+0x25/0x60 [amdgpu]
> [  150.982442]        amdgpu_amdkfd_evict_userptr+0x35/0x70 [amdgpu]
> [  150.988615]        amdgpu_mn_invalidate_hsa+0x41/0x60 [amdgpu]
> [  150.994448]        __mmu_notifier_invalidate_range_start+0xa4/0x240
> [  151.000714]        copy_page_range+0xd70/0xd80
> [  151.005159]        dup_mm+0x3ca/0x550
> [  151.008816]        copy_process+0x1bdc/0x1c70
> [  151.013183]        _do_fork+0x76/0x6c0
> [  151.016929]        __x64_sys_clone+0x8c/0xb0
> [  151.021201]        do_syscall_64+0x4a/0x1d0
> [  151.025404]        entry_SYSCALL_64_after_hwframe+0x49/0xbe
> [  151.030977]
>                -> #2 (&adev->notifier_lock){+.+.}:
> [  151.036993]        __mutex_lock+0xa1/0x9f0
> [  151.041168]        amdgpu_mn_invalidate_hsa+0x30/0x60 [amdgpu]
> [  151.047019]        __mmu_notifier_invalidate_range_start+0xa4/0x240
> [  151.053277]        copy_page_range+0xd70/0xd80
> [  151.057722]        dup_mm+0x3ca/0x550
> [  151.061388]        copy_process+0x1bdc/0x1c70
> [  151.065748]        _do_fork+0x76/0x6c0
> [  151.069499]        __x64_sys_clone+0x8c/0xb0
> [  151.073765]        do_syscall_64+0x4a/0x1d0
> [  151.077952]        entry_SYSCALL_64_after_hwframe+0x49/0xbe
> [  151.083523]
>                -> #1 (mmu_notifier_invalidate_range_start){+.+.}:
> [  151.090833]        change_protection+0x802/0xab0
> [  151.095448]        mprotect_fixup+0x187/0x2d0
> [  151.099801]        setup_arg_pages+0x124/0x250
> [  151.104251]        load_elf_binary+0x3a4/0x1464
> [  151.108781]        search_binary_handler+0x6c/0x210
> [  151.113656]        __do_execve_file.isra.40+0x7f7/0xa50
> [  151.118875]        do_execve+0x21/0x30
> [  151.122632]        call_usermodehelper_exec_async+0x17e/0x190
> [  151.128393]        ret_from_fork+0x24/0x30
> [  151.132489]
>                -> #0 (&mm->mmap_sem#2){++++}:
> [  151.138064]        __lock_acquire+0x11a1/0x1490
> [  151.142597]        lock_acquire+0x90/0x180
> [  151.146694]        __might_fault+0x68/0x90
> [  151.150879]        read_sdma_queue_counter+0x5f/0xb0 [amdgpu]
> [  151.156693]        update_sdma_queue_past_activity_stats+0x3b/0x90 [amdgpu]
> [  151.163725]        destroy_queue_cpsch+0x1ae/0x210 [amdgpu]
> [  151.169373]        pqm_destroy_queue+0xf0/0x250 [amdgpu]
> [  151.174762]        kfd_ioctl_destroy_queue+0x32/0x70 [amdgpu]
> [  151.180577]        kfd_ioctl+0x223/0x400 [amdgpu]
> [  151.185284]        ksys_ioctl+0x8f/0xb0
> [  151.189118]        __x64_sys_ioctl+0x16/0x20
> [  151.193389]        do_syscall_64+0x4a/0x1d0
> [  151.197569]        entry_SYSCALL_64_after_hwframe+0x49/0xbe
> [  151.203141]
>                other info that might help us debug this:
>
> [  151.211140] Chain exists of:
>                  &mm->mmap_sem#2 --> &adev->notifier_lock --> &dqm->lock_hidden
>
>   151.222535]  Possible unsafe locking scenario:
>
> [  151.228447]        CPU0                    CPU1
> [  151.232971]        ----                    ----
> [  151.237502]   lock(&dqm->lock_hidden);
> [  151.241254]                                lock(&adev->notifier_lock);
> [  151.247774]                                lock(&dqm->lock_hidden);
> [  151.254038]   lock(&mm->mmap_sem#2);
> [  151.257610]
>
> This commit fixes the warning by ensuring get_user() is not called
> while reading SDMA stats with dqm_lock held as get_user() could cause a
> page fault which leads to the circular locking scenario.
>
> Signed-off-by: Mukul Joshi <mukul.joshi at amd.com>

A few coding-style nit-picks inline. Please also run the patch through
checkpatch.pl. With those fixed, the patch is

Reviewed-by: Felix Kuehling <Felix.Kuehling at amd.com>


> ---
>  .../drm/amd/amdkfd/kfd_device_queue_manager.c |  75 +++++----
>  .../drm/amd/amdkfd/kfd_device_queue_manager.h |   3 +-
>  drivers/gpu/drm/amd/amdkfd/kfd_process.c      | 145 +++++++++++++++---
>  3 files changed, 162 insertions(+), 61 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
> index 21eb0998c4ae..57e579876037 100644
> --- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
> +++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
> @@ -153,24 +153,23 @@ static void decrement_queue_count(struct device_queue_manager *dqm,
>  		dqm->active_cp_queue_count--;
>  }
>  
> -int read_sdma_queue_counter(struct queue *q, uint64_t *val)
> +int read_sdma_queue_counter(uint64_t q_rptr, uint64_t *val)
>  {
>  	int ret;
>  	uint64_t tmp = 0;
>  
> -	if (!q || !val)
> +	if (!val)
>  		return -EINVAL;
>  	/*
>  	 * SDMA activity counter is stored at queue's RPTR + 0x8 location.
>  	 */
> -	if (!access_ok((const void __user *)((uint64_t)q->properties.read_ptr +
> +	if (!access_ok((const void __user *)(q_rptr +
>  					sizeof(uint64_t)), sizeof(uint64_t))) {
>  		pr_err("Can't access sdma queue activity counter\n");
>  		return -EFAULT;
>  	}
>  
> -	ret = get_user(tmp, (uint64_t *)((uint64_t)(q->properties.read_ptr) +
> -						    sizeof(uint64_t)));
> +	ret = get_user(tmp, (uint64_t *)(q_rptr + sizeof(uint64_t)));
>  	if (!ret) {
>  		*val = tmp;
>  	}
> @@ -178,27 +177,6 @@ int read_sdma_queue_counter(struct queue *q, uint64_t *val)
>  	return ret;
>  }
>  
> -static int update_sdma_queue_past_activity_stats(struct kfd_process_device *pdd,
> -						 struct queue *q)
> -{
> -	int ret;
> -	uint64_t val = 0;
> -
> -	if (!pdd)
> -		return -ENODEV;
> -
> -	ret = read_sdma_queue_counter(q, &val);
> -	if (ret) {
> -		pr_err("Failed to read SDMA queue counter for queue: %d\n",
> -				q->properties.queue_id);
> -		return ret;
> -	}
> -
> -	pdd->sdma_past_activity_counter += val;
> -
> -	return ret;
> -}
> -
>  static int allocate_doorbell(struct qcm_process_device *qpd, struct queue *q)
>  {
>  	struct kfd_dev *dev = qpd->dqm->dev;
> @@ -533,11 +511,6 @@ static int destroy_queue_nocpsch_locked(struct device_queue_manager *dqm,
>  	if (retval == -ETIME)
>  		qpd->reset_wavefronts = true;
>  
> -	/* Get the SDMA queue stats */
> -        if ((q->properties.type == KFD_QUEUE_TYPE_SDMA) ||
> -            (q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI)) {
> -                update_sdma_queue_past_activity_stats(qpd_to_pdd(qpd), q);
> -        }
>  
>  	mqd_mgr->free_mqd(mqd_mgr, q->mqd, q->mqd_mem_obj);
>  
> @@ -573,9 +546,24 @@ static int destroy_queue_nocpsch(struct device_queue_manager *dqm,
>  				struct queue *q)
>  {
>  	int retval;
> +	uint64_t sdma_val = 0;
> +	struct kfd_process_device *pdd = qpd_to_pdd(qpd);
> +
> +	/* Get the SDMA queue stats */
> +	if ((q->properties.type == KFD_QUEUE_TYPE_SDMA) ||
> +	    (q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI)) {
> +		retval = read_sdma_queue_counter((uint64_t)q->properties.read_ptr,
> +							&sdma_val);
> +		if (retval) {

No braces needed here.


> +			pr_err("Failed to read SDMA queue counter for queue: %d\n",
> +				q->properties.queue_id);
> +		}
> +	}
>  
>  	dqm_lock(dqm);
>  	retval = destroy_queue_nocpsch_locked(dqm, qpd, q);
> +	if (!retval)
> +		pdd->sdma_past_activity_counter += sdma_val;
>  	dqm_unlock(dqm);
>  
>  	return retval;
> @@ -1480,6 +1468,19 @@ static int destroy_queue_cpsch(struct device_queue_manager *dqm,
>  {
>  	int retval;
>  	struct mqd_manager *mqd_mgr;
> +	uint64_t sdma_val = 0;
> +	struct kfd_process_device *pdd = qpd_to_pdd(qpd);
> +
> +	/* Get the SDMA queue stats */
> +	if ((q->properties.type == KFD_QUEUE_TYPE_SDMA) ||
> +	    (q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI)) {
> +		retval = read_sdma_queue_counter((uint64_t)q->properties.read_ptr,
> +							&sdma_val);
> +		if (retval) {

No braces needed here.


> +			pr_err("Failed to read SDMA queue counter for queue: %d\n",
> +				q->properties.queue_id);
> +		}
> +	}
>  
>  	retval = 0;
>  
> @@ -1501,10 +1502,11 @@ static int destroy_queue_cpsch(struct device_queue_manager *dqm,
>  
>  	deallocate_doorbell(qpd, q);
>  
> -	if (q->properties.type == KFD_QUEUE_TYPE_SDMA)
> -		deallocate_sdma_queue(dqm, q);
> -	else if (q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI)
> +	if ((q->properties.type == KFD_QUEUE_TYPE_SDMA) ||
> +	    (q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI)) {
>  		deallocate_sdma_queue(dqm, q);
> +		pdd->sdma_past_activity_counter += sdma_val;
> +	}
>  
>  	list_del(&q->list);
>  	qpd->queue_count--;
> @@ -1520,11 +1522,6 @@ static int destroy_queue_cpsch(struct device_queue_manager *dqm,
>  		}
>  	}
>  
> -	/* Get the SDMA queue stats */
> -	if ((q->properties.type == KFD_QUEUE_TYPE_SDMA) ||
> -	    (q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI)) {
> -		update_sdma_queue_past_activity_stats(qpd_to_pdd(qpd), q);
> -	}
>  	/*
>  	 * Unconditionally decrement this counter, regardless of the queue's
>  	 * type
> diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h
> index 894bcf877f9e..49d8e324c636 100644
> --- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h
> +++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h
> @@ -251,6 +251,5 @@ static inline void dqm_unlock(struct device_queue_manager *dqm)
>  	mutex_unlock(&dqm->lock_hidden);
>  }
>  
> -int read_sdma_queue_counter(struct queue *q, uint64_t *val);
> -
> +int read_sdma_queue_counter(uint64_t q_rptr, uint64_t *val);
>  #endif /* KFD_DEVICE_QUEUE_MANAGER_H_ */
> diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process.c b/drivers/gpu/drm/amd/amdkfd/kfd_process.c
> index 8616a204e4c3..6dd1c861007b 100644
> --- a/drivers/gpu/drm/amd/amdkfd/kfd_process.c
> +++ b/drivers/gpu/drm/amd/amdkfd/kfd_process.c
> @@ -86,6 +86,13 @@ struct kfd_sdma_activity_handler_workarea {
>  	uint64_t sdma_activity_counter;
>  };
>  
> +struct temp_sdma_queue_list {
> +	uint64_t rptr;
> +	uint64_t sdma_val;
> +	unsigned int queue_id;
> +	struct list_head list;
> +};
> +
>  static void kfd_sdma_activity_worker(struct work_struct *work)
>  {
>  	struct kfd_sdma_activity_handler_workarea *workarea;
> @@ -96,6 +103,8 @@ static void kfd_sdma_activity_worker(struct work_struct *work)
>  	struct qcm_process_device *qpd;
>  	struct device_queue_manager *dqm;
>  	int ret = 0;
> +	struct temp_sdma_queue_list sdma_q_list;
> +	struct temp_sdma_queue_list *sdma_q, *next;
>  
>  	workarea = container_of(work, struct kfd_sdma_activity_handler_workarea,
>  				sdma_activity_work);
> @@ -109,41 +118,137 @@ static void kfd_sdma_activity_worker(struct work_struct *work)
>  	qpd = &pdd->qpd;
>  	if (!dqm || !qpd)
>  		return;
> +	/*
> +	 * Total SDMA activity is current SDMA activity + past SDMA activity
> +	 * Past SDMA count is stored in pdd.
> +	 * To get the current activity counters for all active SDMA queues,
> +	 * we loop over all SDMA queues and get their counts from user-space.
> +	 *
> +	 * We cannot call get_user() with dqm_lock held as it can cause
> +	 * a circular lock dependency situation. To read the SDMA stats,
> +	 * we need to do the following:
> +	 *
> +	 * 1. Create a temporary list of SDMA queue nodes from the qpd->queues_list,
> +	 *    with dqm_lock/dqm_unlock().
> +	 * 2. Call get_user() for each node in temporary list without dqm_lock.
> +	 *    Save the SDMA count for each node and also add the count to the total
> +	 *    SDMA count counter.
> +	 *    Its possible, during this step, a few SDMA queue nodes got deleted
> +	 *    from the qpd->queues_list.
> +	 * 3. Do a second pass over qpd->queues_list to check if any nodes got deleted.
> +	 *    If any node got deleted, its SDMA count would be captured in the sdma
> +	 *    past activity counter. So subtract the SDMA counter stored in step 2
> +	 *    for this node from the total SDMA count.
> +	 */
> +	INIT_LIST_HEAD(&sdma_q_list.list);
>  
> -	mm = get_task_mm(pdd->process->lead_thread);
> -	if (!mm) {
> -		return;
> +	/*
> +	 * Create the temp list of all SDMA queues
> +	 */
> +	dqm_lock(dqm);
> +
> +	list_for_each_entry(q, &qpd->queues_list, list) {
> +		if ((q->properties.type != KFD_QUEUE_TYPE_SDMA) &&
> +		    (q->properties.type != KFD_QUEUE_TYPE_SDMA_XGMI))
> +			continue;
> +
> +		sdma_q = kzalloc(sizeof(struct temp_sdma_queue_list), GFP_KERNEL);
> +		if (!sdma_q) {
> +			dqm_unlock(dqm);
> +			goto cleanup;
> +		}
> +
> +		INIT_LIST_HEAD(&sdma_q->list);
> +		sdma_q->rptr = (uint64_t)q->properties.read_ptr;
> +		sdma_q->queue_id = q->properties.queue_id;
> +		list_add_tail(&sdma_q->list, &sdma_q_list.list);
>  	}
>  
> -	use_mm(mm);
> +	/*
> +	 * If the temp list is empty, then no SDMA queues nodes were found in
> +	 * qpd->queues_list. Return the past activity count as the total sdma
> +	 * count
> +	 */
> +	if (list_empty(&sdma_q_list.list)) {
> +		workarea->sdma_activity_counter = pdd->sdma_past_activity_counter;
> +		dqm_unlock(dqm);
> +		return;
> +	}
>  
> -	dqm_lock(dqm);
> +	dqm_unlock(dqm);
>  
>  	/*
> -	 * Total SDMA activity is current SDMA activity + past SDMA activity
> +	 * Get the usage count for each SDMA queue in temp_list.
>  	 */
> -	workarea->sdma_activity_counter = pdd->sdma_past_activity_counter;
> +	mm = get_task_mm(pdd->process->lead_thread);
> +	if (!mm)
> +		goto cleanup;
> +
> +	use_mm(mm);
> +
> +	list_for_each_entry(sdma_q, &sdma_q_list.list, list) {
> +		val = 0;
> +		ret = read_sdma_queue_counter(sdma_q->rptr, &val);
> +		if (ret)
> +			pr_debug("Failed to read SDMA queue active "
> +				 "counter for queue id: %d",

Having a line longer than 80 characters is better than breaking a string
like that. Breaking strings makes it harder to grep for them. A long
line is a warning in checkpatch.pl. A broken string may actually be
considered an error.


> +				 sdma_q->queue_id);
> +			else {

Inconsistent indentation. The "else" should be at the same level as the
"if". Also, if you need braces in one if-else branch, it's better to use
braces on all of them for consistency and readability.


> +				sdma_q->sdma_val = val;
> +				workarea->sdma_activity_counter += val;
> +			}
> +	}
> +
> +	unuse_mm(mm);
> +	mmput(mm);
>  
>  	/*
> -	 * Get the current activity counters for all active SDMA queues
> +	 * Do a second iteration over qpd_queues_list to check if any SDMA
> +	 * nodes got deleted while fetching SDMA counter.
>  	 */
> +	dqm_lock(dqm);
> +
> +	workarea->sdma_activity_counter += pdd->sdma_past_activity_counter;
> +
>  	list_for_each_entry(q, &qpd->queues_list, list) {
> -		if ((q->properties.type == KFD_QUEUE_TYPE_SDMA) ||
> -		    (q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI)) {
> -			val = 0;
> -			ret = read_sdma_queue_counter(q, &val);
> -			if (ret)
> -				pr_debug("Failed to read SDMA queue active "
> -					 "counter for queue id: %d",
> -					 q->properties.queue_id);
> -			else
> -				workarea->sdma_activity_counter += val;
> +		if (list_empty(&sdma_q_list.list)) {

No braces needed here.

Regards,
  Felix


> +			break;
> +		}
> +
> +		if ((q->properties.type != KFD_QUEUE_TYPE_SDMA) &&
> +		    (q->properties.type != KFD_QUEUE_TYPE_SDMA_XGMI))
> +			continue;
> +
> +		list_for_each_entry_safe(sdma_q, next, &sdma_q_list.list, list) {
> +			if (((uint64_t)q->properties.read_ptr == sdma_q->rptr) &&
> +			     (sdma_q->queue_id == q->properties.queue_id)) {
> +				list_del(&sdma_q->list);
> +				kfree(sdma_q);
> +				break;
> +			}
>  		}
>  	}
>  
>  	dqm_unlock(dqm);
> -	unuse_mm(mm);
> -	mmput(mm);
> +
> +	/*
> +	 * If temp list is not empty, it implies some queues got deleted
> +	 * from qpd->queues_list during SDMA usage read. Subtract the SDMA
> +	 * count for each node from the total SDMA count.
> +	 */
> +	list_for_each_entry_safe(sdma_q, next, &sdma_q_list.list, list) {
> +		workarea->sdma_activity_counter -= sdma_q->sdma_val;
> +		list_del(&sdma_q->list);
> +		kfree(sdma_q);
> +	}
> +
> +	return;
> +
> +cleanup:
> +	list_for_each_entry_safe(sdma_q, next, &sdma_q_list.list, list) {
> +		list_del(&sdma_q->list);
> +		kfree(sdma_q);
> +	}
>  }
>  
>  static ssize_t kfd_procfs_show(struct kobject *kobj, struct attribute *attr,


More information about the amd-gfx mailing list