[PATCH 1/4] drm/amdgpu: partial revert VRAM lost handling v2

[PATCH 1/4] drm/amdgpu: partial revert VRAM lost handling v2

[Christian König]

From: Christian König <christian.koenig@amd.com>

Keep blocking the CS, but revert everything else. Mapping BOs and info IOCTL
are harmless and can still happen even when VRAM content ist lost.

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c  | 11 -----------
 drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c |  5 -----
 drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c | 14 ++++++++++----
 3 files changed, 10 insertions(+), 20 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index ab83dfc..ba3e645 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -1257,16 +1257,12 @@ int amdgpu_cs_wait_ioctl(struct drm_device *dev, void *data,
 {
 	union drm_amdgpu_wait_cs *wait = data;
 	struct amdgpu_device *adev = dev->dev_private;
-	struct amdgpu_fpriv *fpriv = filp->driver_priv;
 	unsigned long timeout = amdgpu_gem_timeout(wait->in.timeout);
 	struct amdgpu_ring *ring = NULL;
 	struct amdgpu_ctx *ctx;
 	struct dma_fence *fence;
 	long r;
 
-	if (amdgpu_kms_vram_lost(adev, fpriv))
-		return -ENODEV;
-
 	ctx = amdgpu_ctx_get(filp->driver_priv, wait->in.ctx_id);
 	if (ctx == NULL)
 		return -EINVAL;
@@ -1335,16 +1331,12 @@ int amdgpu_cs_fence_to_handle_ioctl(struct drm_device *dev, void *data,
 				    struct drm_file *filp)
 {
 	struct amdgpu_device *adev = dev->dev_private;
-	struct amdgpu_fpriv *fpriv = filp->driver_priv;
 	union drm_amdgpu_fence_to_handle *info = data;
 	struct dma_fence *fence;
 	struct drm_syncobj *syncobj;
 	struct sync_file *sync_file;
 	int fd, r;
 
-	if (amdgpu_kms_vram_lost(adev, fpriv))
-		return -ENODEV;
-
 	fence = amdgpu_cs_get_fence(adev, filp, &info->in.fence);
 	if (IS_ERR(fence))
 		return PTR_ERR(fence);
@@ -1506,15 +1498,12 @@ int amdgpu_cs_wait_fences_ioctl(struct drm_device *dev, void *data,
 				struct drm_file *filp)
 {
 	struct amdgpu_device *adev = dev->dev_private;
-	struct amdgpu_fpriv *fpriv = filp->driver_priv;
 	union drm_amdgpu_wait_fences *wait = data;
 	uint32_t fence_count = wait->in.fence_count;
 	struct drm_amdgpu_fence *fences_user;
 	struct drm_amdgpu_fence *fences;
 	int r;
 
-	if (amdgpu_kms_vram_lost(adev, fpriv))
-		return -ENODEV;
 	/* Get the fences from userspace */
 	fences = kmalloc_array(fence_count, sizeof(struct drm_amdgpu_fence),
 			GFP_KERNEL);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index b0d45c8..c69048c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -577,11 +577,6 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
 			args->operation);
 		return -EINVAL;
 	}
-	if ((args->operation == AMDGPU_VA_OP_MAP) ||
-	    (args->operation == AMDGPU_VA_OP_REPLACE)) {
-		if (amdgpu_kms_vram_lost(adev, fpriv))
-			return -ENODEV;
-	}
 
 	INIT_LIST_HEAD(&list);
 	INIT_LIST_HEAD(&duplicates);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
index 8c6fd56..9fba6f9 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
@@ -269,7 +269,6 @@ static int amdgpu_firmware_info(struct drm_amdgpu_info_firmware *fw_info,
 static int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 {
 	struct amdgpu_device *adev = dev->dev_private;
-	struct amdgpu_fpriv *fpriv = filp->driver_priv;
 	struct drm_amdgpu_info *info = data;
 	struct amdgpu_mode_info *minfo = &adev->mode_info;
 	void __user *out = (void __user *)(uintptr_t)info->return_pointer;
@@ -282,8 +281,6 @@ static int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file
 
 	if (!info->return_size || !info->return_pointer)
 		return -EINVAL;
-	if (amdgpu_kms_vram_lost(adev, fpriv))
-		return -ENODEV;
 
 	switch (info->query) {
 	case AMDGPU_INFO_ACCEL_WORKING:
@@ -791,10 +788,19 @@ void amdgpu_driver_lastclose_kms(struct drm_device *dev)
 	vga_switcheroo_process_delayed_switch();
 }
 
+/**
+ * amdgpu_kms_vram_lost - check if VRAM was lost for this client
+ *
+ * @adev: amdgpu device
+ * @fpriv: client private
+ *
+ * Check if all CS is blocked for the client because of lost VRAM
+ */
 bool amdgpu_kms_vram_lost(struct amdgpu_device *adev,
 			  struct amdgpu_fpriv *fpriv)
 {
-	return fpriv->vram_lost_counter != atomic_read(&adev->vram_lost_counter);
+	return fpriv->vram_lost_counter !=
+		atomic_read(&adev->vram_lost_counter);
 }
 
 /**
-- 
2.7.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 2/4] drm/amdgpu: keep copy of VRAM lost counter in job

[Christian König]

From: Christian König <christian.koenig@amd.com>

Instead of reading the current counter from fpriv.

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h     |  1 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c  |  2 ++
 drivers/gpu/drm/amd/amdgpu/amdgpu_job.c | 13 +++++++------
 3 files changed, 10 insertions(+), 6 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 6a8f169..0b8b5c9 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -1116,6 +1116,7 @@ struct amdgpu_job {
 	uint32_t		gds_base, gds_size;
 	uint32_t		gws_base, gws_size;
 	uint32_t		oa_base, oa_size;
+	uint32_t		vram_lost_counter;
 
 	/* user fence handling */
 	uint64_t		uf_addr;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index ba3e645..fa8502f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -170,6 +170,8 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, void *data)
 	if (ret)
 		goto free_all_kdata;
 
+	p->job->vram_lost_counter = fpriv->vram_lost_counter;
+
 	if (p->uf_entry.robj)
 		p->job->uf_addr = uf_offset;
 	kfree(chunk_array);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
index 4510627..c76d17c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
@@ -61,6 +61,7 @@ int amdgpu_job_alloc(struct amdgpu_device *adev, unsigned num_ibs,
 	(*job)->vm = vm;
 	(*job)->ibs = (void *)&(*job)[1];
 	(*job)->num_ibs = num_ibs;
+	(*job)->vram_lost_counter = atomic_read(&adev->vram_lost_counter);
 
 	amdgpu_sync_create(&(*job)->sync);
 	amdgpu_sync_create(&(*job)->dep_sync);
@@ -177,8 +178,8 @@ static struct dma_fence *amdgpu_job_dependency(struct amd_sched_job *sched_job)
 static struct dma_fence *amdgpu_job_run(struct amd_sched_job *sched_job)
 {
 	struct dma_fence *fence = NULL;
+	struct amdgpu_device *adev;
 	struct amdgpu_job *job;
-	struct amdgpu_fpriv *fpriv = NULL;
 	int r;
 
 	if (!sched_job) {
@@ -186,17 +187,17 @@ static struct dma_fence *amdgpu_job_run(struct amd_sched_job *sched_job)
 		return NULL;
 	}
 	job = to_amdgpu_job(sched_job);
+	adev = job->adev;
 
 	BUG_ON(amdgpu_sync_peek_fence(&job->sync, NULL));
 
 	trace_amdgpu_sched_run_job(job);
-	if (job->vm)
-		fpriv = container_of(job->vm, struct amdgpu_fpriv, vm);
 	/* skip ib schedule when vram is lost */
-	if (fpriv && amdgpu_kms_vram_lost(job->adev, fpriv))
+	if (job->vram_lost_counter != atomic_read(&adev->vram_lost_counter)) {
 		DRM_ERROR("Skip scheduling IBs!\n");
-	else {
-		r = amdgpu_ib_schedule(job->ring, job->num_ibs, job->ibs, job, &fence);
+	} else {
+		r = amdgpu_ib_schedule(job->ring, job->num_ibs, job->ibs, job,
+				       &fence);
 		if (r)
 			DRM_ERROR("Error scheduling IBs (%d)\n", r);
 	}
-- 
2.7.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 3/4] drm/amdgpu: move the VRAM lost counter per context

[Christian König]

From: Christian König <christian.koenig@amd.com>

Instead of per device track the VRAM lost per context and return ECANCELED
instead of ENODEV.

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h     |  6 ++----
 drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c  |  9 +++++----
 drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c |  1 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c | 16 ----------------
 4 files changed, 8 insertions(+), 24 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 0b8b5c9..591d6b2 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -735,10 +735,11 @@ struct amdgpu_ctx {
 	struct amdgpu_device    *adev;
 	struct amdgpu_queue_mgr queue_mgr;
 	unsigned		reset_counter;
+	uint32_t		vram_lost_counter;
 	spinlock_t		ring_lock;
 	struct dma_fence	**fences;
 	struct amdgpu_ctx_ring	rings[AMDGPU_MAX_RINGS];
-	bool preamble_presented;
+	bool			preamble_presented;
 };
 
 struct amdgpu_ctx_mgr {
@@ -773,7 +774,6 @@ struct amdgpu_fpriv {
 	struct mutex		bo_list_lock;
 	struct idr		bo_list_handles;
 	struct amdgpu_ctx_mgr	ctx_mgr;
-	u32			vram_lost_counter;
 };
 
 /*
@@ -1868,8 +1868,6 @@ static inline bool amdgpu_has_atpx(void) { return false; }
 extern const struct drm_ioctl_desc amdgpu_ioctls_kms[];
 extern const int amdgpu_max_kms_ioctl;
 
-bool amdgpu_kms_vram_lost(struct amdgpu_device *adev,
-			  struct amdgpu_fpriv *fpriv);
 int amdgpu_driver_load_kms(struct drm_device *dev, unsigned long flags);
 void amdgpu_driver_unload_kms(struct drm_device *dev);
 void amdgpu_driver_lastclose_kms(struct drm_device *dev);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index fa8502f..359c89c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -170,7 +170,11 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, void *data)
 	if (ret)
 		goto free_all_kdata;
 
-	p->job->vram_lost_counter = fpriv->vram_lost_counter;
+	p->job->vram_lost_counter = atomic_read(&p->adev->vram_lost_counter);
+	if (p->ctx->vram_lost_counter != p->job->vram_lost_counter) {
+		ret = -ECANCELED;
+		goto free_all_kdata;
+	}
 
 	if (p->uf_entry.robj)
 		p->job->uf_addr = uf_offset;
@@ -1191,7 +1195,6 @@ static int amdgpu_cs_submit(struct amdgpu_cs_parser *p,
 int amdgpu_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 {
 	struct amdgpu_device *adev = dev->dev_private;
-	struct amdgpu_fpriv *fpriv = filp->driver_priv;
 	union drm_amdgpu_cs *cs = data;
 	struct amdgpu_cs_parser parser = {};
 	bool reserved_buffers = false;
@@ -1199,8 +1202,6 @@ int amdgpu_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 
 	if (!adev->accel_working)
 		return -EBUSY;
-	if (amdgpu_kms_vram_lost(adev, fpriv))
-		return -ENODEV;
 
 	parser.adev = adev;
 	parser.filp = filp;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
index 75c933b..6718a41 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
@@ -45,6 +45,7 @@ static int amdgpu_ctx_init(struct amdgpu_device *adev, struct amdgpu_ctx *ctx)
 	}
 
 	ctx->reset_counter = atomic_read(&adev->gpu_reset_counter);
+	ctx->vram_lost_counter = atomic_read(&adev->vram_lost_counter);
 
 	/* create context entity for each ring */
 	for (i = 0; i < adev->num_rings; i++) {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
index 9fba6f9..0fc36b2 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
@@ -789,21 +789,6 @@ void amdgpu_driver_lastclose_kms(struct drm_device *dev)
 }
 
 /**
- * amdgpu_kms_vram_lost - check if VRAM was lost for this client
- *
- * @adev: amdgpu device
- * @fpriv: client private
- *
- * Check if all CS is blocked for the client because of lost VRAM
- */
-bool amdgpu_kms_vram_lost(struct amdgpu_device *adev,
-			  struct amdgpu_fpriv *fpriv)
-{
-	return fpriv->vram_lost_counter !=
-		atomic_read(&adev->vram_lost_counter);
-}
-
-/**
  * amdgpu_driver_open_kms - drm callback for open
  *
  * @dev: drm dev pointer
@@ -859,7 +844,6 @@ int amdgpu_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
 
 	amdgpu_ctx_mgr_init(&fpriv->ctx_mgr);
 
-	fpriv->vram_lost_counter = atomic_read(&adev->vram_lost_counter);
 	file_priv->driver_priv = fpriv;
 
 out_suspend:
-- 
2.7.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 4/4] drm/amdgpu: set -ECANCELED when dropping jobs

[Christian König]

From: Christian König <christian.koenig@amd.com>

And return from the wait functions the fence error code.

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c  | 7 ++++++-
 drivers/gpu/drm/amd/amdgpu/amdgpu_job.c | 1 +
 2 files changed, 7 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index 359c89c..0185d35 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -1283,6 +1283,8 @@ int amdgpu_cs_wait_ioctl(struct drm_device *dev, void *data,
 		r = PTR_ERR(fence);
 	else if (fence) {
 		r = dma_fence_wait_timeout(fence, true, timeout);
+		if (r > 0 && fence->error)
+			r = fence->error;
 		dma_fence_put(fence);
 	} else
 		r = 1;
@@ -1420,6 +1422,9 @@ static int amdgpu_cs_wait_all_fences(struct amdgpu_device *adev,
 
 		if (r == 0)
 			break;
+
+		if (fence->error)
+			return fence->error;
 	}
 
 	memset(wait, 0, sizeof(*wait));
@@ -1480,7 +1485,7 @@ static int amdgpu_cs_wait_any_fence(struct amdgpu_device *adev,
 	wait->out.status = (r > 0);
 	wait->out.first_signaled = first;
 	/* set return value 0 to indicate success */
-	r = 0;
+	r = array[first]->error;
 
 err_free_fence_array:
 	for (i = 0; i < fence_count; i++)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
index c76d17c..7067edf 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
@@ -194,6 +194,7 @@ static struct dma_fence *amdgpu_job_run(struct amd_sched_job *sched_job)
 	trace_amdgpu_sched_run_job(job);
 	/* skip ib schedule when vram is lost */
 	if (job->vram_lost_counter != atomic_read(&adev->vram_lost_counter)) {
+		dma_fence_set_error(&job->base.s_fence->finished, -ECANCELED);
 		DRM_ERROR("Skip scheduling IBs!\n");
 	} else {
 		r = amdgpu_ib_schedule(job->ring, job->num_ibs, job->ibs, job,
-- 
2.7.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Re: [PATCH 4/4] drm/amdgpu: set -ECANCELED when dropping jobs

[Nicolai Hähnle]

On 09.10.2017 17:34, Christian König wrote:
> From: Christian König <christian.koenig@amd.com>
> 
> And return from the wait functions the fence error code.
> 
> Signed-off-by: Christian König <christian.koenig@amd.com>

For the series:

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>

> ---
>   drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c  | 7 ++++++-
>   drivers/gpu/drm/amd/amdgpu/amdgpu_job.c | 1 +
>   2 files changed, 7 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> index 359c89c..0185d35 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> @@ -1283,6 +1283,8 @@ int amdgpu_cs_wait_ioctl(struct drm_device *dev, void *data,
>   		r = PTR_ERR(fence);
>   	else if (fence) {
>   		r = dma_fence_wait_timeout(fence, true, timeout);
> +		if (r > 0 && fence->error)
> +			r = fence->error;
>   		dma_fence_put(fence);
>   	} else
>   		r = 1;
> @@ -1420,6 +1422,9 @@ static int amdgpu_cs_wait_all_fences(struct amdgpu_device *adev,
>   
>   		if (r == 0)
>   			break;
> +
> +		if (fence->error)
> +			return fence->error;
>   	}
>   
>   	memset(wait, 0, sizeof(*wait));
> @@ -1480,7 +1485,7 @@ static int amdgpu_cs_wait_any_fence(struct amdgpu_device *adev,
>   	wait->out.status = (r > 0);
>   	wait->out.first_signaled = first;
>   	/* set return value 0 to indicate success */
> -	r = 0;
> +	r = array[first]->error;
>   
>   err_free_fence_array:
>   	for (i = 0; i < fence_count; i++)
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
> index c76d17c..7067edf 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
> @@ -194,6 +194,7 @@ static struct dma_fence *amdgpu_job_run(struct amd_sched_job *sched_job)
>   	trace_amdgpu_sched_run_job(job);
>   	/* skip ib schedule when vram is lost */
>   	if (job->vram_lost_counter != atomic_read(&adev->vram_lost_counter)) {
> +		dma_fence_set_error(&job->base.s_fence->finished, -ECANCELED);
>   		DRM_ERROR("Skip scheduling IBs!\n");
>   	} else {
>   		r = amdgpu_ib_schedule(job->ring, job->num_ibs, job->ibs, job,
> 


-- 
Lerne, wie die Welt wirklich ist,
Aber vergiss niemals, wie sie sein sollte.
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

RE: [PATCH 4/4] drm/amdgpu: set -ECANCELED when dropping jobs

[Liu, Monk]

You should use "dma_fence_wait_timeout()" instead of fence->error, that's more make sense 



-----Original Message-----
From: amd-gfx [mailto:amd-gfx-bounces@lists.freedesktop.org] On Behalf Of Christian K?nig
Sent: 2017年10月9日 23:35
To: amd-gfx@lists.freedesktop.org
Subject: [PATCH 4/4] drm/amdgpu: set -ECANCELED when dropping jobs

From: Christian König <christian.koenig@amd.com>

And return from the wait functions the fence error code.

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c  | 7 ++++++-  drivers/gpu/drm/amd/amdgpu/amdgpu_job.c | 1 +
 2 files changed, 7 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index 359c89c..0185d35 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -1283,6 +1283,8 @@ int amdgpu_cs_wait_ioctl(struct drm_device *dev, void *data,
 		r = PTR_ERR(fence);
 	else if (fence) {
 		r = dma_fence_wait_timeout(fence, true, timeout);
+		if (r > 0 && fence->error)
+			r = fence->error;
 		dma_fence_put(fence);
 	} else
 		r = 1;
@@ -1420,6 +1422,9 @@ static int amdgpu_cs_wait_all_fences(struct amdgpu_device *adev,
 
 		if (r == 0)
 			break;
+
+		if (fence->error)
+			return fence->error;
 	}
 
 	memset(wait, 0, sizeof(*wait));
@@ -1480,7 +1485,7 @@ static int amdgpu_cs_wait_any_fence(struct amdgpu_device *adev,
 	wait->out.status = (r > 0);
 	wait->out.first_signaled = first;
 	/* set return value 0 to indicate success */
-	r = 0;
+	r = array[first]->error;
 
 err_free_fence_array:
 	for (i = 0; i < fence_count; i++)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
index c76d17c..7067edf 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
@@ -194,6 +194,7 @@ static struct dma_fence *amdgpu_job_run(struct amd_sched_job *sched_job)
 	trace_amdgpu_sched_run_job(job);
 	/* skip ib schedule when vram is lost */
 	if (job->vram_lost_counter != atomic_read(&adev->vram_lost_counter)) {
+		dma_fence_set_error(&job->base.s_fence->finished, -ECANCELED);
 		DRM_ERROR("Skip scheduling IBs!\n");
 	} else {
 		r = amdgpu_ib_schedule(job->ring, job->num_ibs, job->ibs, job,
--
2.7.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

RE: [PATCH 4/4] drm/amdgpu: set -ECANCELED when dropping jobs

[Liu, Monk]

Sorry, should be "dma_fence_get_error"

-----Original Message-----
From: amd-gfx [mailto:amd-gfx-bounces@lists.freedesktop.org] On Behalf Of Liu, Monk
Sent: 2017年10月10日 12:37
To: Christian König <ckoenig.leichtzumerken@gmail.com>; amd-gfx@lists.freedesktop.org
Subject: RE: [PATCH 4/4] drm/amdgpu: set -ECANCELED when dropping jobs

You should use "dma_fence_wait_timeout()" instead of fence->error, that's more make sense 



-----Original Message-----
From: amd-gfx [mailto:amd-gfx-bounces@lists.freedesktop.org] On Behalf Of Christian K?nig
Sent: 2017年10月9日 23:35
To: amd-gfx@lists.freedesktop.org
Subject: [PATCH 4/4] drm/amdgpu: set -ECANCELED when dropping jobs

From: Christian König <christian.koenig@amd.com>

And return from the wait functions the fence error code.

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c  | 7 ++++++-  drivers/gpu/drm/amd/amdgpu/amdgpu_job.c | 1 +
 2 files changed, 7 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index 359c89c..0185d35 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -1283,6 +1283,8 @@ int amdgpu_cs_wait_ioctl(struct drm_device *dev, void *data,
 		r = PTR_ERR(fence);
 	else if (fence) {
 		r = dma_fence_wait_timeout(fence, true, timeout);
+		if (r > 0 && fence->error)
+			r = fence->error;
 		dma_fence_put(fence);
 	} else
 		r = 1;
@@ -1420,6 +1422,9 @@ static int amdgpu_cs_wait_all_fences(struct amdgpu_device *adev,
 
 		if (r == 0)
 			break;
+
+		if (fence->error)
+			return fence->error;
 	}
 
 	memset(wait, 0, sizeof(*wait));
@@ -1480,7 +1485,7 @@ static int amdgpu_cs_wait_any_fence(struct amdgpu_device *adev,
 	wait->out.status = (r > 0);
 	wait->out.first_signaled = first;
 	/* set return value 0 to indicate success */
-	r = 0;
+	r = array[first]->error;
 
 err_free_fence_array:
 	for (i = 0; i < fence_count; i++)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
index c76d17c..7067edf 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
@@ -194,6 +194,7 @@ static struct dma_fence *amdgpu_job_run(struct amd_sched_job *sched_job)
 	trace_amdgpu_sched_run_job(job);
 	/* skip ib schedule when vram is lost */
 	if (job->vram_lost_counter != atomic_read(&adev->vram_lost_counter)) {
+		dma_fence_set_error(&job->base.s_fence->finished, -ECANCELED);
 		DRM_ERROR("Skip scheduling IBs!\n");
 	} else {
 		r = amdgpu_ib_schedule(job->ring, job->num_ibs, job->ibs, job,
--
2.7.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

RE: [PATCH 4/4] drm/amdgpu: set -ECANCELED when dropping jobs

[Liu, Monk]

Should be "dma_fence_get_status()"

-----Original Message-----
From: amd-gfx [mailto:amd-gfx-bounces@lists.freedesktop.org] On Behalf Of Liu, Monk
Sent: 2017年10月10日 12:37
To: Christian König <ckoenig.leichtzumerken@gmail.com>; amd-gfx@lists.freedesktop.org
Subject: RE: [PATCH 4/4] drm/amdgpu: set -ECANCELED when dropping jobs

You should use "dma_fence_wait_timeout()" instead of fence->error, that's more make sense 



-----Original Message-----
From: amd-gfx [mailto:amd-gfx-bounces@lists.freedesktop.org] On Behalf Of Christian K?nig
Sent: 2017年10月9日 23:35
To: amd-gfx@lists.freedesktop.org
Subject: [PATCH 4/4] drm/amdgpu: set -ECANCELED when dropping jobs

From: Christian König <christian.koenig@amd.com>

And return from the wait functions the fence error code.

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c  | 7 ++++++-  drivers/gpu/drm/amd/amdgpu/amdgpu_job.c | 1 +
 2 files changed, 7 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index 359c89c..0185d35 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -1283,6 +1283,8 @@ int amdgpu_cs_wait_ioctl(struct drm_device *dev, void *data,
 		r = PTR_ERR(fence);
 	else if (fence) {
 		r = dma_fence_wait_timeout(fence, true, timeout);
+		if (r > 0 && fence->error)
+			r = fence->error;
 		dma_fence_put(fence);
 	} else
 		r = 1;
@@ -1420,6 +1422,9 @@ static int amdgpu_cs_wait_all_fences(struct amdgpu_device *adev,
 
 		if (r == 0)
 			break;
+
+		if (fence->error)
+			return fence->error;
 	}
 
 	memset(wait, 0, sizeof(*wait));
@@ -1480,7 +1485,7 @@ static int amdgpu_cs_wait_any_fence(struct amdgpu_device *adev,
 	wait->out.status = (r > 0);
 	wait->out.first_signaled = first;
 	/* set return value 0 to indicate success */
-	r = 0;
+	r = array[first]->error;
 
 err_free_fence_array:
 	for (i = 0; i < fence_count; i++)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
index c76d17c..7067edf 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
@@ -194,6 +194,7 @@ static struct dma_fence *amdgpu_job_run(struct amd_sched_job *sched_job)
 	trace_amdgpu_sched_run_job(job);
 	/* skip ib schedule when vram is lost */
 	if (job->vram_lost_counter != atomic_read(&adev->vram_lost_counter)) {
+		dma_fence_set_error(&job->base.s_fence->finished, -ECANCELED);
 		DRM_ERROR("Skip scheduling IBs!\n");
 	} else {
 		r = amdgpu_ib_schedule(job->ring, job->num_ibs, job->ibs, job,
--
2.7.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

RE: [PATCH 2/4] drm/amdgpu: keep copy of VRAM lost counter in job

[Liu, Monk]

Christian, 

Can you elaborate what benefit it brings that keep a counter copy in JOB ? 
Because you know when vram lost happens, the thing real matter is : if the BO in BO_LIST of this job modified/created before this vram lost,
Not that if the job creating timestamp ...

Since kernel side now doesn't implement a way to compare counter per BO (which is too low efficient considering OCL app that have 5000+ BO in BO_LIST), 

if we always consider BO is created/modified since this job's context created (you know many job can refer to one specific BO like texture, which is long time ago modified before JOB creation), than we just need to compare the counter in ctx, 

BR Monk

-----Original Message-----
From: amd-gfx [mailto:amd-gfx-bounces@lists.freedesktop.org] On Behalf Of Christian K?nig
Sent: Monday, October 9, 2017 11:35 PM
To: amd-gfx@lists.freedesktop.org
Subject: [PATCH 2/4] drm/amdgpu: keep copy of VRAM lost counter in job

From: Christian König <christian.koenig@amd.com>

Instead of reading the current counter from fpriv.

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h     |  1 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c  |  2 ++  drivers/gpu/drm/amd/amdgpu/amdgpu_job.c | 13 +++++++------
 3 files changed, 10 insertions(+), 6 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 6a8f169..0b8b5c9 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -1116,6 +1116,7 @@ struct amdgpu_job {
 	uint32_t		gds_base, gds_size;
 	uint32_t		gws_base, gws_size;
 	uint32_t		oa_base, oa_size;
+	uint32_t		vram_lost_counter;
 
 	/* user fence handling */
 	uint64_t		uf_addr;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index ba3e645..fa8502f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -170,6 +170,8 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, void *data)
 	if (ret)
 		goto free_all_kdata;
 
+	p->job->vram_lost_counter = fpriv->vram_lost_counter;
+
 	if (p->uf_entry.robj)
 		p->job->uf_addr = uf_offset;
 	kfree(chunk_array);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
index 4510627..c76d17c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
@@ -61,6 +61,7 @@ int amdgpu_job_alloc(struct amdgpu_device *adev, unsigned num_ibs,
 	(*job)->vm = vm;
 	(*job)->ibs = (void *)&(*job)[1];
 	(*job)->num_ibs = num_ibs;
+	(*job)->vram_lost_counter = atomic_read(&adev->vram_lost_counter);
 
 	amdgpu_sync_create(&(*job)->sync);
 	amdgpu_sync_create(&(*job)->dep_sync);
@@ -177,8 +178,8 @@ static struct dma_fence *amdgpu_job_dependency(struct amd_sched_job *sched_job)  static struct dma_fence *amdgpu_job_run(struct amd_sched_job *sched_job)  {
 	struct dma_fence *fence = NULL;
+	struct amdgpu_device *adev;
 	struct amdgpu_job *job;
-	struct amdgpu_fpriv *fpriv = NULL;
 	int r;
 
 	if (!sched_job) {
@@ -186,17 +187,17 @@ static struct dma_fence *amdgpu_job_run(struct amd_sched_job *sched_job)
 		return NULL;
 	}
 	job = to_amdgpu_job(sched_job);
+	adev = job->adev;
 
 	BUG_ON(amdgpu_sync_peek_fence(&job->sync, NULL));
 
 	trace_amdgpu_sched_run_job(job);
-	if (job->vm)
-		fpriv = container_of(job->vm, struct amdgpu_fpriv, vm);
 	/* skip ib schedule when vram is lost */
-	if (fpriv && amdgpu_kms_vram_lost(job->adev, fpriv))
+	if (job->vram_lost_counter != atomic_read(&adev->vram_lost_counter)) {
 		DRM_ERROR("Skip scheduling IBs!\n");
-	else {
-		r = amdgpu_ib_schedule(job->ring, job->num_ibs, job->ibs, job, &fence);
+	} else {
+		r = amdgpu_ib_schedule(job->ring, job->num_ibs, job->ibs, job,
+				       &fence);
 		if (r)
 			DRM_ERROR("Error scheduling IBs (%d)\n", r);
 	}
--
2.7.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx