All of lore.kernel.org
 help / color / mirror / Atom feed
From: Jesse Zhang <zhexi.zhang-5C7GfCeVMHo@public.gmane.org>
To: amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org
Cc: Jesse Zhang <zhexi.zhang-5C7GfCeVMHo@public.gmane.org>
Subject: [PATCH] drm/amd/amdgpu: Add mutex to protect IB resources
Date: Wed, 6 Nov 2019 14:56:56 +0800	[thread overview]
Message-ID: <1573023416-10462-1-git-send-email-zhexi.zhang@amd.com> (raw)

Unloading driver has call trace when unloading happens
2s after loading driver.

Since ring test are delayed after initializing driver,
it is possible that driver has been unloaded before or
while doing ring test.

Add mutex to prevent ring test and driver finalization
occurs at the same time and check before doing ring test
if required resources still exist.

Change-Id: I27b52c2c630ad3853c6384e7b6906f0fae590ead
Signed-off-by: Jesse Zhang <zhexi.zhang@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h        | 1 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 1 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c     | 9 +++++++++
 3 files changed, 11 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 0469cc5..b825ad0 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -923,6 +923,7 @@ struct amdgpu_device {
 	struct amdgpu_ring		*rings[AMDGPU_MAX_RINGS];
 	bool				ib_pool_ready;
 	struct amdgpu_sa_manager	ring_tmp_bo;
+	struct mutex			ib_lock;
 
 	/* interrupts */
 	struct amdgpu_irq		irq;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index b2f38b0..8adbb25 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -2806,6 +2806,7 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	mutex_init(&adev->lock_reset);
 	mutex_init(&adev->virt.dpm_mutex);
 	mutex_init(&adev->psp.mutex);
+	mutex_init(&adev->ib_lock);
 
 	r = amdgpu_device_check_arguments(adev);
 	if (r)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
index d121bbd..ef5339e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
@@ -317,10 +317,12 @@ int amdgpu_ib_pool_init(struct amdgpu_device *adev)
  */
 void amdgpu_ib_pool_fini(struct amdgpu_device *adev)
 {
+	mutex_lock(&adev->ib_lock);
 	if (adev->ib_pool_ready) {
 		amdgpu_sa_bo_manager_fini(adev, &adev->ring_tmp_bo);
 		adev->ib_pool_ready = false;
 	}
+	mutex_unlock(&adev->ib_lock);
 }
 
 /**
@@ -364,6 +366,11 @@ int amdgpu_ib_ring_tests(struct amdgpu_device *adev)
 		struct amdgpu_ring *ring = adev->rings[i];
 		long tmo;
 
+		if (adev->shutdown) {
+			DRM_INFO("Device finalized, skip unfinished ring test\n");
+			return 0;
+		}
+
 		/* KIQ rings don't have an IB test because we never submit IBs
 		 * to them and they have no interrupt support.
 		 */
@@ -381,7 +388,9 @@ int amdgpu_ib_ring_tests(struct amdgpu_device *adev)
 		else
 			tmo = tmo_gfx;
 
+		mutex_lock(&adev->ib_lock);
 		r = amdgpu_ring_test_ib(ring, tmo);
+		mutex_unlock(&adev->ib_lock);
 		if (!r) {
 			DRM_DEV_DEBUG(adev->dev, "ib test on %s succeeded\n",
 				      ring->name);
-- 
2.7.4

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

WARNING: multiple messages have this Message-ID (diff)
From: Jesse Zhang <zhexi.zhang@amd.com>
To: <amd-gfx@lists.freedesktop.org>
Cc: Jesse Zhang <zhexi.zhang@amd.com>
Subject: [PATCH] drm/amd/amdgpu: Add mutex to protect IB resources
Date: Wed, 6 Nov 2019 14:56:56 +0800	[thread overview]
Message-ID: <1573023416-10462-1-git-send-email-zhexi.zhang@amd.com> (raw)
Message-ID: <20191106065656.y0OOsLxPXm5aRRp9TQ_Q9OzMovqkOgUSsWmwJDJubms@z> (raw)

Unloading driver has call trace when unloading happens
2s after loading driver.

Since ring test are delayed after initializing driver,
it is possible that driver has been unloaded before or
while doing ring test.

Add mutex to prevent ring test and driver finalization
occurs at the same time and check before doing ring test
if required resources still exist.

Change-Id: I27b52c2c630ad3853c6384e7b6906f0fae590ead
Signed-off-by: Jesse Zhang <zhexi.zhang@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h        | 1 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 1 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c     | 9 +++++++++
 3 files changed, 11 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 0469cc5..b825ad0 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -923,6 +923,7 @@ struct amdgpu_device {
 	struct amdgpu_ring		*rings[AMDGPU_MAX_RINGS];
 	bool				ib_pool_ready;
 	struct amdgpu_sa_manager	ring_tmp_bo;
+	struct mutex			ib_lock;
 
 	/* interrupts */
 	struct amdgpu_irq		irq;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index b2f38b0..8adbb25 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -2806,6 +2806,7 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	mutex_init(&adev->lock_reset);
 	mutex_init(&adev->virt.dpm_mutex);
 	mutex_init(&adev->psp.mutex);
+	mutex_init(&adev->ib_lock);
 
 	r = amdgpu_device_check_arguments(adev);
 	if (r)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
index d121bbd..ef5339e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
@@ -317,10 +317,12 @@ int amdgpu_ib_pool_init(struct amdgpu_device *adev)
  */
 void amdgpu_ib_pool_fini(struct amdgpu_device *adev)
 {
+	mutex_lock(&adev->ib_lock);
 	if (adev->ib_pool_ready) {
 		amdgpu_sa_bo_manager_fini(adev, &adev->ring_tmp_bo);
 		adev->ib_pool_ready = false;
 	}
+	mutex_unlock(&adev->ib_lock);
 }
 
 /**
@@ -364,6 +366,11 @@ int amdgpu_ib_ring_tests(struct amdgpu_device *adev)
 		struct amdgpu_ring *ring = adev->rings[i];
 		long tmo;
 
+		if (adev->shutdown) {
+			DRM_INFO("Device finalized, skip unfinished ring test\n");
+			return 0;
+		}
+
 		/* KIQ rings don't have an IB test because we never submit IBs
 		 * to them and they have no interrupt support.
 		 */
@@ -381,7 +388,9 @@ int amdgpu_ib_ring_tests(struct amdgpu_device *adev)
 		else
 			tmo = tmo_gfx;
 
+		mutex_lock(&adev->ib_lock);
 		r = amdgpu_ring_test_ib(ring, tmo);
+		mutex_unlock(&adev->ib_lock);
 		if (!r) {
 			DRM_DEV_DEBUG(adev->dev, "ib test on %s succeeded\n",
 				      ring->name);
-- 
2.7.4

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

             reply	other threads:[~2019-11-06  6:56 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-11-06  6:56 Jesse Zhang [this message]
2019-11-06  6:56 ` [PATCH] drm/amd/amdgpu: Add mutex to protect IB resources Jesse Zhang
     [not found] ` <1573023416-10462-1-git-send-email-zhexi.zhang-5C7GfCeVMHo@public.gmane.org>
2019-11-06  9:35   ` Christian König
2019-11-06  9:35     ` Christian König

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=1573023416-10462-1-git-send-email-zhexi.zhang@amd.com \
    --to=zhexi.zhang-5c7gfcevmho@public.gmane.org \
    --cc=amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org \
    /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.