All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH v2 1/2] drm/amdgpu: save the reset dump register value for devcoredump
@ 2022-05-26  9:48 Somalapuram Amaranath
  2022-05-26  9:48 ` [PATCH v2 2/2] drm/amdgpu: adding device coredump support Somalapuram Amaranath
  2022-05-26  9:54 ` [PATCH v2 1/2] drm/amdgpu: save the reset dump register value for devcoredump Sharma, Shashank
  0 siblings, 2 replies; 9+ messages in thread
From: Somalapuram Amaranath @ 2022-05-26  9:48 UTC (permalink / raw)
  To: amd-gfx
  Cc: alexander.deucher, Somalapuram Amaranath, christian.koenig,
	shashank.sharma

Allocate memory for register value and use the same values for devcoredump.
v1 -> V2: Change krealloc_array() to kmalloc_array()

Signed-off-by: Somalapuram Amaranath <Amaranath.Somalapuram@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h         | 1 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c | 7 +++++++
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c  | 6 +++---
 3 files changed, 11 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 76df583663c7..c79d9992b113 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -1042,6 +1042,7 @@ struct amdgpu_device {
 
 	/* reset dump register */
 	uint32_t                        *reset_dump_reg_list;
+	uint32_t			*reset_dump_reg_value;
 	int                             num_regs;
 
 	struct amdgpu_reset_domain	*reset_domain;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
index eedb12f6b8a3..f3ac7912c29c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
@@ -1709,17 +1709,24 @@ static ssize_t amdgpu_reset_dump_register_list_write(struct file *f,
 		i++;
 	} while (len < size);
 
+	new = kmalloc_array(i, sizeof(uint32_t), GFP_KERNEL);
+	if (!new) {
+		ret = -ENOMEM;
+		goto error_free;
+	}
 	ret = down_write_killable(&adev->reset_domain->sem);
 	if (ret)
 		goto error_free;
 
 	swap(adev->reset_dump_reg_list, tmp);
+	swap(adev->reset_dump_reg_value, new);
 	adev->num_regs = i;
 	up_write(&adev->reset_domain->sem);
 	ret = size;
 
 error_free:
 	kfree(tmp);
+	kfree(new);
 	return ret;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 4daa0e893965..866b4980a6fa 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -4720,15 +4720,15 @@ int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev,
 
 static int amdgpu_reset_reg_dumps(struct amdgpu_device *adev)
 {
-	uint32_t reg_value;
 	int i;
 
 	lockdep_assert_held(&adev->reset_domain->sem);
 	dump_stack();
 
 	for (i = 0; i < adev->num_regs; i++) {
-		reg_value = RREG32(adev->reset_dump_reg_list[i]);
-		trace_amdgpu_reset_reg_dumps(adev->reset_dump_reg_list[i], reg_value);
+		adev->reset_dump_reg_value[i] = RREG32(adev->reset_dump_reg_list[i]);
+		trace_amdgpu_reset_reg_dumps(adev->reset_dump_reg_list[i],
+				adev->reset_dump_reg_value[i]);
 	}
 
 	return 0;
-- 
2.32.0


^ permalink raw reply related	[flat|nested] 9+ messages in thread

* [PATCH v2 2/2] drm/amdgpu: adding device coredump support
  2022-05-26  9:48 [PATCH v2 1/2] drm/amdgpu: save the reset dump register value for devcoredump Somalapuram Amaranath
@ 2022-05-26  9:48 ` Somalapuram Amaranath
  2022-05-26  9:57   ` Sharma, Shashank
  2022-05-26 10:26   ` Wang, Yang(Kevin)
  2022-05-26  9:54 ` [PATCH v2 1/2] drm/amdgpu: save the reset dump register value for devcoredump Sharma, Shashank
  1 sibling, 2 replies; 9+ messages in thread
From: Somalapuram Amaranath @ 2022-05-26  9:48 UTC (permalink / raw)
  To: amd-gfx
  Cc: alexander.deucher, Somalapuram Amaranath, christian.koenig,
	shashank.sharma

Added device coredump information:
- Kernel version
- Module
- Time
- VRAM status
- Guilty process name and PID
- GPU register dumps
v1 -> v2: Variable name change
v1 -> v2: NULL check
v1 -> v2: Code alignment
v1 -> v2: Adding dummy amdgpu_devcoredump_free
v1 -> v2: memset reset_task_info to zero

Signed-off-by: Somalapuram Amaranath <Amaranath.Somalapuram@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h        |  3 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 67 ++++++++++++++++++++++
 2 files changed, 70 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index c79d9992b113..25a7b2c74928 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -1044,6 +1044,9 @@ struct amdgpu_device {
 	uint32_t                        *reset_dump_reg_list;
 	uint32_t			*reset_dump_reg_value;
 	int                             num_regs;
+	struct amdgpu_task_info         reset_task_info;
+	bool                            reset_vram_lost;
+	struct timespec64               reset_time;
 
 	struct amdgpu_reset_domain	*reset_domain;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 866b4980a6fa..ca97afe5be63 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -32,6 +32,8 @@
 #include <linux/slab.h>
 #include <linux/iommu.h>
 #include <linux/pci.h>
+#include <linux/devcoredump.h>
+#include <generated/utsrelease.h>
 
 #include <drm/drm_atomic_helper.h>
 #include <drm/drm_probe_helper.h>
@@ -4734,6 +4736,62 @@ static int amdgpu_reset_reg_dumps(struct amdgpu_device *adev)
 	return 0;
 }
 
+#ifdef CONFIG_DEV_COREDUMP
+static ssize_t amdgpu_devcoredump_read(char *buffer, loff_t offset,
+		size_t count, void *data, size_t datalen)
+{
+	struct drm_printer p;
+	struct amdgpu_device *adev = data;
+	struct drm_print_iterator iter;
+	int i;
+
+	if (adev == NULL)
+		return 0;
+
+	iter.data = buffer;
+	iter.offset = 0;
+	iter.start = offset;
+	iter.remain = count;
+
+	p = drm_coredump_printer(&iter);
+
+	drm_printf(&p, "**** AMDGPU Device Coredump ****\n");
+	drm_printf(&p, "kernel: " UTS_RELEASE "\n");
+	drm_printf(&p, "module: " KBUILD_MODNAME "\n");
+	drm_printf(&p, "time: %lld.%09ld\n", adev->reset_time.tv_sec, adev->reset_time.tv_nsec);
+	if (adev->reset_task_info.pid)
+		drm_printf(&p, "process_name: %s PID: %d\n",
+			   adev->reset_task_info.process_name,
+			   adev->reset_task_info.pid);
+
+	if (adev->reset_vram_lost)
+		drm_printf(&p, "VRAM is lost due to GPU reset!\n");
+	if (adev->num_regs) {
+		drm_printf(&p, "AMDGPU register dumps:\nOffset:     Value:\n");
+
+		for (i = 0; i < adev->num_regs; i++)
+			drm_printf(&p, "0x%08x: 0x%08x\n",
+				   adev->reset_dump_reg_list[i],
+				   adev->reset_dump_reg_value[i]);
+	}
+
+	return count - iter.remain;
+}
+
+static void amdgpu_devcoredump_free(void *data)
+{
+}
+
+static void amdgpu_reset_capture_coredumpm(struct amdgpu_device *adev)
+{
+	struct drm_device *dev = adev_to_drm(adev);
+
+	ktime_get_ts64(&adev->reset_time);
+	dev_coredumpm(dev->dev, THIS_MODULE, adev, 0, GFP_KERNEL,
+			amdgpu_devcoredump_read, amdgpu_devcoredump_free);
+}
+#endif
+
 int amdgpu_do_asic_reset(struct list_head *device_list_handle,
 			 struct amdgpu_reset_context *reset_context)
 {
@@ -4818,6 +4876,15 @@ int amdgpu_do_asic_reset(struct list_head *device_list_handle,
 					goto out;
 
 				vram_lost = amdgpu_device_check_vram_lost(tmp_adev);
+#ifdef CONFIG_DEV_COREDUMP
+				tmp_adev->reset_vram_lost = vram_lost;
+				memset(&tmp_adev->reset_task_info, 0,
+						sizeof(tmp_adev->reset_task_info));
+				if (reset_context->job && reset_context->job->vm)
+					tmp_adev->reset_task_info =
+						reset_context->job->vm->task_info;
+				amdgpu_reset_capture_coredumpm(tmp_adev);
+#endif
 				if (vram_lost) {
 					DRM_INFO("VRAM is lost due to GPU reset!\n");
 					amdgpu_inc_vram_lost(tmp_adev);
-- 
2.32.0


^ permalink raw reply related	[flat|nested] 9+ messages in thread

* Re: [PATCH v2 1/2] drm/amdgpu: save the reset dump register value for devcoredump
  2022-05-26  9:48 [PATCH v2 1/2] drm/amdgpu: save the reset dump register value for devcoredump Somalapuram Amaranath
  2022-05-26  9:48 ` [PATCH v2 2/2] drm/amdgpu: adding device coredump support Somalapuram Amaranath
@ 2022-05-26  9:54 ` Sharma, Shashank
  2022-05-31 13:08   ` Somalapuram, Amaranath
  1 sibling, 1 reply; 9+ messages in thread
From: Sharma, Shashank @ 2022-05-26  9:54 UTC (permalink / raw)
  To: Somalapuram Amaranath, amd-gfx; +Cc: alexander.deucher, christian.koenig

Hey Amar,

On 5/26/2022 11:48 AM, Somalapuram Amaranath wrote:
> Allocate memory for register value and use the same values for devcoredump.
> v1 -> V2: Change krealloc_array() to kmalloc_array()
> 
> Signed-off-by: Somalapuram Amaranath <Amaranath.Somalapuram@amd.com>
> ---
>   drivers/gpu/drm/amd/amdgpu/amdgpu.h         | 1 +
>   drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c | 7 +++++++
>   drivers/gpu/drm/amd/amdgpu/amdgpu_device.c  | 6 +++---
>   3 files changed, 11 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> index 76df583663c7..c79d9992b113 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> @@ -1042,6 +1042,7 @@ struct amdgpu_device {
>   
>   	/* reset dump register */
>   	uint32_t                        *reset_dump_reg_list;
> +	uint32_t			*reset_dump_reg_value;
>   	int                             num_regs;
>   
>   	struct amdgpu_reset_domain	*reset_domain;
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
> index eedb12f6b8a3..f3ac7912c29c 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
> @@ -1709,17 +1709,24 @@ static ssize_t amdgpu_reset_dump_register_list_write(struct file *f,
>   		i++;
>   	} while (len < size);
>   
> +	new = kmalloc_array(i, sizeof(uint32_t), GFP_KERNEL);
> +	if (!new) {
> +		ret = -ENOMEM;
> +		goto error_free;
> +	}
>   	ret = down_write_killable(&adev->reset_domain->sem);
>   	if (ret)
>   		goto error_free;
>   
>   	swap(adev->reset_dump_reg_list, tmp);
> +	swap(adev->reset_dump_reg_value, new);
>   	adev->num_regs = i;
>   	up_write(&adev->reset_domain->sem);
>   	ret = size;
>   
>   error_free:
>   	kfree(tmp);
> +	kfree(new);
>   	return ret;
>   }
>   
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> index 4daa0e893965..866b4980a6fa 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> @@ -4720,15 +4720,15 @@ int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev,
>   
>   static int amdgpu_reset_reg_dumps(struct amdgpu_device *adev)
>   {
> -	uint32_t reg_value;
>   	int i;
>   
>   	lockdep_assert_held(&adev->reset_domain->sem);
>   	dump_stack();
>   
>   	for (i = 0; i < adev->num_regs; i++) {
> -		reg_value = RREG32(adev->reset_dump_reg_list[i]);
> -		trace_amdgpu_reset_reg_dumps(adev->reset_dump_reg_list[i], reg_value);
> +		adev->reset_dump_reg_value[i] = RREG32(adev->reset_dump_reg_list[i]);
> +		trace_amdgpu_reset_reg_dumps(adev->reset_dump_reg_list[i],
> +				adev->reset_dump_reg_value[i]);
Please fix the alignment with the line above, after that, please feel 
free to use:
Reviewed-by: Shashank Sharma <shashank.sharma@amd.com>

- Shashank
>   	}
>   
>   	return 0;

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: [PATCH v2 2/2] drm/amdgpu: adding device coredump support
  2022-05-26  9:48 ` [PATCH v2 2/2] drm/amdgpu: adding device coredump support Somalapuram Amaranath
@ 2022-05-26  9:57   ` Sharma, Shashank
  2022-05-26 10:26   ` Wang, Yang(Kevin)
  1 sibling, 0 replies; 9+ messages in thread
From: Sharma, Shashank @ 2022-05-26  9:57 UTC (permalink / raw)
  To: Somalapuram Amaranath, amd-gfx; +Cc: alexander.deucher, christian.koenig



On 5/26/2022 11:48 AM, Somalapuram Amaranath wrote:
> Added device coredump information:
> - Kernel version
> - Module
> - Time
> - VRAM status
> - Guilty process name and PID
> - GPU register dumps
> v1 -> v2: Variable name change
> v1 -> v2: NULL check
> v1 -> v2: Code alignment
> v1 -> v2: Adding dummy amdgpu_devcoredump_free
> v1 -> v2: memset reset_task_info to zero
> 
> Signed-off-by: Somalapuram Amaranath <Amaranath.Somalapuram@amd.com>
> ---
>   drivers/gpu/drm/amd/amdgpu/amdgpu.h        |  3 +
>   drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 67 ++++++++++++++++++++++
>   2 files changed, 70 insertions(+)
> 
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> index c79d9992b113..25a7b2c74928 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> @@ -1044,6 +1044,9 @@ struct amdgpu_device {
>   	uint32_t                        *reset_dump_reg_list;
>   	uint32_t			*reset_dump_reg_value;
>   	int                             num_regs;
> +	struct amdgpu_task_info         reset_task_info;
> +	bool                            reset_vram_lost;
> +	struct timespec64               reset_time;
>   
>   	struct amdgpu_reset_domain	*reset_domain;
>   
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> index 866b4980a6fa..ca97afe5be63 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> @@ -32,6 +32,8 @@
>   #include <linux/slab.h>
>   #include <linux/iommu.h>
>   #include <linux/pci.h>
> +#include <linux/devcoredump.h>
> +#include <generated/utsrelease.h>
>   
>   #include <drm/drm_atomic_helper.h>
>   #include <drm/drm_probe_helper.h>
> @@ -4734,6 +4736,62 @@ static int amdgpu_reset_reg_dumps(struct amdgpu_device *adev)
>   	return 0;
>   }
>   
> +#ifdef CONFIG_DEV_COREDUMP
> +static ssize_t amdgpu_devcoredump_read(char *buffer, loff_t offset,
> +		size_t count, void *data, size_t datalen)
> +{
> +	struct drm_printer p;
> +	struct amdgpu_device *adev = data;
> +	struct drm_print_iterator iter;
> +	int i;
> +
> +	if (adev == NULL)
> +		return 0;
> +
> +	iter.data = buffer;
> +	iter.offset = 0;
> +	iter.start = offset;
> +	iter.remain = count;
> +
> +	p = drm_coredump_printer(&iter);
> +
> +	drm_printf(&p, "**** AMDGPU Device Coredump ****\n");
> +	drm_printf(&p, "kernel: " UTS_RELEASE "\n");
> +	drm_printf(&p, "module: " KBUILD_MODNAME "\n");
> +	drm_printf(&p, "time: %lld.%09ld\n", adev->reset_time.tv_sec, adev->reset_time.tv_nsec);
> +	if (adev->reset_task_info.pid)
> +		drm_printf(&p, "process_name: %s PID: %d\n",
> +			   adev->reset_task_info.process_name,
> +			   adev->reset_task_info.pid);
> +
> +	if (adev->reset_vram_lost)
> +		drm_printf(&p, "VRAM is lost due to GPU reset!\n");
> +	if (adev->num_regs) {
> +		drm_printf(&p, "AMDGPU register dumps:\nOffset:     Value:\n");
> +
> +		for (i = 0; i < adev->num_regs; i++)
> +			drm_printf(&p, "0x%08x: 0x%08x\n",
> +				   adev->reset_dump_reg_list[i],
> +				   adev->reset_dump_reg_value[i]);
> +	}
> +
> +	return count - iter.remain;
> +}
> +
> +static void amdgpu_devcoredump_free(void *data)
> +{
> +}
> +
> +static void amdgpu_reset_capture_coredumpm(struct amdgpu_device *adev)
> +{
> +	struct drm_device *dev = adev_to_drm(adev);
> +
> +	ktime_get_ts64(&adev->reset_time);
> +	dev_coredumpm(dev->dev, THIS_MODULE, adev, 0, GFP_KERNEL,
> +			amdgpu_devcoredump_read, amdgpu_devcoredump_free);
Alignment with line above.
> +}
> +#endif
> +
>   int amdgpu_do_asic_reset(struct list_head *device_list_handle,
>   			 struct amdgpu_reset_context *reset_context)
>   {
> @@ -4818,6 +4876,15 @@ int amdgpu_do_asic_reset(struct list_head *device_list_handle,
>   					goto out;
>   
>   				vram_lost = amdgpu_device_check_vram_lost(tmp_adev);
> +#ifdef CONFIG_DEV_COREDUMP
> +				tmp_adev->reset_vram_lost = vram_lost;
> +				memset(&tmp_adev->reset_task_info, 0,
> +						sizeof(tmp_adev->reset_task_info));
Alignment with the line above.
> +				if (reset_context->job && reset_context->job->vm)
> +					tmp_adev->reset_task_info =
> +						reset_context->job->vm->task_info;
> +				amdgpu_reset_capture_coredumpm(tmp_adev);
> +#endif
>   				if (vram_lost) {
>   					DRM_INFO("VRAM is lost due to GPU reset!\n");
>   					amdgpu_inc_vram_lost(tmp_adev);

With above fixed, feel free to use:
Reviewed-by: Shashank Sharma <shashank.sharma@amd.com>

- Shashank

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: [PATCH v2 2/2] drm/amdgpu: adding device coredump support
  2022-05-26  9:48 ` [PATCH v2 2/2] drm/amdgpu: adding device coredump support Somalapuram Amaranath
  2022-05-26  9:57   ` Sharma, Shashank
@ 2022-05-26 10:26   ` Wang, Yang(Kevin)
  2022-05-31 13:14     ` Somalapuram, Amaranath
  1 sibling, 1 reply; 9+ messages in thread
From: Wang, Yang(Kevin) @ 2022-05-26 10:26 UTC (permalink / raw)
  To: Somalapuram, Amaranath, amd-gfx
  Cc: Deucher, Alexander, Koenig, Christian, Sharma, Shashank

[-- Attachment #1: Type: text/plain, Size: 5874 bytes --]

[AMD Official Use Only - General]



________________________________
From: amd-gfx <amd-gfx-bounces@lists.freedesktop.org> on behalf of Somalapuram Amaranath <Amaranath.Somalapuram@amd.com>
Sent: Thursday, May 26, 2022 5:48 PM
To: amd-gfx@lists.freedesktop.org <amd-gfx@lists.freedesktop.org>
Cc: Deucher, Alexander <Alexander.Deucher@amd.com>; Somalapuram, Amaranath <Amaranath.Somalapuram@amd.com>; Koenig, Christian <Christian.Koenig@amd.com>; Sharma, Shashank <Shashank.Sharma@amd.com>
Subject: [PATCH v2 2/2] drm/amdgpu: adding device coredump support

Added device coredump information:
- Kernel version
- Module
- Time
- VRAM status
- Guilty process name and PID
- GPU register dumps
v1 -> v2: Variable name change
v1 -> v2: NULL check
v1 -> v2: Code alignment
v1 -> v2: Adding dummy amdgpu_devcoredump_free
v1 -> v2: memset reset_task_info to zero

Signed-off-by: Somalapuram Amaranath <Amaranath.Somalapuram@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h        |  3 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 67 ++++++++++++++++++++++
 2 files changed, 70 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index c79d9992b113..25a7b2c74928 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -1044,6 +1044,9 @@ struct amdgpu_device {
         uint32_t                        *reset_dump_reg_list;
         uint32_t                        *reset_dump_reg_value;
         int                             num_regs;
+       struct amdgpu_task_info         reset_task_info;
+       bool                            reset_vram_lost;
+       struct timespec64               reset_time;

[kevin]:
the CONFIG_DEV_COREDUMP check is needed for above variable to avoid compiler warning when coredump feautre is not enabled.

         struct amdgpu_reset_domain      *reset_domain;

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 866b4980a6fa..ca97afe5be63 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -32,6 +32,8 @@
 #include <linux/slab.h>
 #include <linux/iommu.h>
 #include <linux/pci.h>
+#include <linux/devcoredump.h>
+#include <generated/utsrelease.h>

 #include <drm/drm_atomic_helper.h>
 #include <drm/drm_probe_helper.h>
@@ -4734,6 +4736,62 @@ static int amdgpu_reset_reg_dumps(struct amdgpu_device *adev)
         return 0;
 }

+#ifdef CONFIG_DEV_COREDUMP
+static ssize_t amdgpu_devcoredump_read(char *buffer, loff_t offset,
+               size_t count, void *data, size_t datalen)
+{
+       struct drm_printer p;
+       struct amdgpu_device *adev = data;
+       struct drm_print_iterator iter;
+       int i;
+
+       if (adev == NULL)
+               return 0;
[kevin]:
 this check is not needed, because this private data is passed by our driver as below:

       dev_coredumpm(dev->dev, THIS_MODULE, adev, 0, GFP_KERNEL,
                       amdgpu_devcoredump_read, amdgpu_devcoredump_free);
+
+       iter.data = buffer;
+       iter.offset = 0;
+       iter.start = offset;
+       iter.remain = count;
+
+       p = drm_coredump_printer(&iter);
+
+       drm_printf(&p, "**** AMDGPU Device Coredump ****\n");
+       drm_printf(&p, "kernel: " UTS_RELEASE "\n");
+       drm_printf(&p, "module: " KBUILD_MODNAME "\n");
+       drm_printf(&p, "time: %lld.%09ld\n", adev->reset_time.tv_sec, adev->reset_time.tv_nsec);
+       if (adev->reset_task_info.pid)
+               drm_printf(&p, "process_name: %s PID: %d\n",
+                          adev->reset_task_info.process_name,
+                          adev->reset_task_info.pid);
+
+       if (adev->reset_vram_lost)
+               drm_printf(&p, "VRAM is lost due to GPU reset!\n");
+       if (adev->num_regs) {
+               drm_printf(&p, "AMDGPU register dumps:\nOffset:     Value:\n");
+
+               for (i = 0; i < adev->num_regs; i++)
+                       drm_printf(&p, "0x%08x: 0x%08x\n",
+                                  adev->reset_dump_reg_list[i],
+                                  adev->reset_dump_reg_value[i]);
+       }
+
+       return count - iter.remain;
+}
+
+static void amdgpu_devcoredump_free(void *data)
+{
+}
+
+static void amdgpu_reset_capture_coredumpm(struct amdgpu_device *adev)
+{
+       struct drm_device *dev = adev_to_drm(adev);
+
+       ktime_get_ts64(&adev->reset_time);
+       dev_coredumpm(dev->dev, THIS_MODULE, adev, 0, GFP_KERNEL,
+                       amdgpu_devcoredump_read, amdgpu_devcoredump_free);
+}
+#endif
+
 int amdgpu_do_asic_reset(struct list_head *device_list_handle,
                          struct amdgpu_reset_context *reset_context)
 {
@@ -4818,6 +4876,15 @@ int amdgpu_do_asic_reset(struct list_head *device_list_handle,
                                         goto out;

                                 vram_lost = amdgpu_device_check_vram_lost(tmp_adev);
+#ifdef CONFIG_DEV_COREDUMP
+                               tmp_adev->reset_vram_lost = vram_lost;
+                               memset(&tmp_adev->reset_task_info, 0,
+                                               sizeof(tmp_adev->reset_task_info));
+                               if (reset_context->job && reset_context->job->vm)
+                                       tmp_adev->reset_task_info =
+                                               reset_context->job->vm->task_info;
+                               amdgpu_reset_capture_coredumpm(tmp_adev);
+#endif
                                 if (vram_lost) {
                                         DRM_INFO("VRAM is lost due to GPU reset!\n");
                                         amdgpu_inc_vram_lost(tmp_adev);
--
2.32.0


[-- Attachment #2: Type: text/html, Size: 13768 bytes --]

^ permalink raw reply related	[flat|nested] 9+ messages in thread

* Re: [PATCH v2 1/2] drm/amdgpu: save the reset dump register value for devcoredump
  2022-05-26  9:54 ` [PATCH v2 1/2] drm/amdgpu: save the reset dump register value for devcoredump Sharma, Shashank
@ 2022-05-31 13:08   ` Somalapuram, Amaranath
  2022-05-31 13:17     ` Sharma, Shashank
  0 siblings, 1 reply; 9+ messages in thread
From: Somalapuram, Amaranath @ 2022-05-31 13:08 UTC (permalink / raw)
  To: Sharma, Shashank, Somalapuram Amaranath, amd-gfx
  Cc: alexander.deucher, christian.koenig

[-- Attachment #1: Type: text/plain, Size: 3537 bytes --]


On 5/26/2022 3:24 PM, Sharma, Shashank wrote:
> Hey Amar,
>
> On 5/26/2022 11:48 AM, Somalapuram Amaranath wrote:
>> Allocate memory for register value and use the same values for 
>> devcoredump.
>> v1 -> V2: Change krealloc_array() to kmalloc_array()
>>
>> Signed-off-by: Somalapuram Amaranath <Amaranath.Somalapuram@amd.com>
>> ---
>>   drivers/gpu/drm/amd/amdgpu/amdgpu.h         | 1 +
>>   drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c | 7 +++++++
>>   drivers/gpu/drm/amd/amdgpu/amdgpu_device.c  | 6 +++---
>>   3 files changed, 11 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h 
>> b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
>> index 76df583663c7..c79d9992b113 100644
>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
>> @@ -1042,6 +1042,7 @@ struct amdgpu_device {
>>         /* reset dump register */
>>       uint32_t                        *reset_dump_reg_list;
>> +    uint32_t            *reset_dump_reg_value;
>>       int                             num_regs;
>>         struct amdgpu_reset_domain    *reset_domain;
>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c 
>> b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
>> index eedb12f6b8a3..f3ac7912c29c 100644
>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
>> @@ -1709,17 +1709,24 @@ static ssize_t 
>> amdgpu_reset_dump_register_list_write(struct file *f,
>>           i++;
>>       } while (len < size);
>>   +    new = kmalloc_array(i, sizeof(uint32_t), GFP_KERNEL);
>> +    if (!new) {
>> +        ret = -ENOMEM;
>> +        goto error_free;
>> +    }
>>       ret = down_write_killable(&adev->reset_domain->sem);
>>       if (ret)
>>           goto error_free;
>>         swap(adev->reset_dump_reg_list, tmp);
>> +    swap(adev->reset_dump_reg_value, new);
>>       adev->num_regs = i;
>>       up_write(&adev->reset_domain->sem);
>>       ret = size;
>>     error_free:
>>       kfree(tmp);
>> +    kfree(new);
>>       return ret;
>>   }
>>   diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c 
>> b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
>> index 4daa0e893965..866b4980a6fa 100644
>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
>> @@ -4720,15 +4720,15 @@ int amdgpu_device_pre_asic_reset(struct 
>> amdgpu_device *adev,
>>     static int amdgpu_reset_reg_dumps(struct amdgpu_device *adev)
>>   {
>> -    uint32_t reg_value;
>>       int i;
>>         lockdep_assert_held(&adev->reset_domain->sem);
>>       dump_stack();
>>         for (i = 0; i < adev->num_regs; i++) {
>> -        reg_value = RREG32(adev->reset_dump_reg_list[i]);
>> - trace_amdgpu_reset_reg_dumps(adev->reset_dump_reg_list[i], reg_value);
>> +        adev->reset_dump_reg_value[i] = 
>> RREG32(adev->reset_dump_reg_list[i]);
>> + trace_amdgpu_reset_reg_dumps(adev->reset_dump_reg_list[i],
>> +                adev->reset_dump_reg_value[i]);

Alignment is showing spaces in email. But its 2 tabs in code.


Regards,
S.Amarnath


> Please fix the alignment with the line above, after that, please feel 
> free to use:
> Reviewed-by: Shashank Sharma <shashank.sharma@amd.com>
>
> - Shashank
>>       }
>>         return 0;

[-- Attachment #2.1: Type: text/html, Size: 6818 bytes --]

[-- Attachment #2.2: imNmJuhYCmZqwnZU.png --]
[-- Type: image/png, Size: 10366 bytes --]

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: [PATCH v2 2/2] drm/amdgpu: adding device coredump support
  2022-05-26 10:26   ` Wang, Yang(Kevin)
@ 2022-05-31 13:14     ` Somalapuram, Amaranath
  0 siblings, 0 replies; 9+ messages in thread
From: Somalapuram, Amaranath @ 2022-05-31 13:14 UTC (permalink / raw)
  To: Wang, Yang(Kevin), Somalapuram, Amaranath, amd-gfx
  Cc: Deucher, Alexander, Koenig, Christian, Sharma, Shashank

[-- Attachment #1: Type: text/plain, Size: 6474 bytes --]


On 5/26/2022 3:56 PM, Wang, Yang(Kevin) wrote:
>
> [AMD Official Use Only - General]
>
>
>
>
> ------------------------------------------------------------------------
> *From:* amd-gfx <amd-gfx-bounces@lists.freedesktop.org> on behalf of 
> Somalapuram Amaranath <Amaranath.Somalapuram@amd.com>
> *Sent:* Thursday, May 26, 2022 5:48 PM
> *To:* amd-gfx@lists.freedesktop.org <amd-gfx@lists.freedesktop.org>
> *Cc:* Deucher, Alexander <Alexander.Deucher@amd.com>; Somalapuram, 
> Amaranath <Amaranath.Somalapuram@amd.com>; Koenig, Christian 
> <Christian.Koenig@amd.com>; Sharma, Shashank <Shashank.Sharma@amd.com>
> *Subject:* [PATCH v2 2/2] drm/amdgpu: adding device coredump support
> Added device coredump information:
> - Kernel version
> - Module
> - Time
> - VRAM status
> - Guilty process name and PID
> - GPU register dumps
> v1 -> v2: Variable name change
> v1 -> v2: NULL check
> v1 -> v2: Code alignment
> v1 -> v2: Adding dummy amdgpu_devcoredump_free
> v1 -> v2: memset reset_task_info to zero
>
> Signed-off-by: Somalapuram Amaranath <Amaranath.Somalapuram@amd.com>
> ---
>  drivers/gpu/drm/amd/amdgpu/amdgpu.h        |  3 +
>  drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 67 ++++++++++++++++++++++
>  2 files changed, 70 insertions(+)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h 
> b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> index c79d9992b113..25a7b2c74928 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> @@ -1044,6 +1044,9 @@ struct amdgpu_device {
>          uint32_t *reset_dump_reg_list;
>          uint32_t *reset_dump_reg_value;
>          int                             num_regs;
> +       struct amdgpu_task_info         reset_task_info;
> +       bool                            reset_vram_lost;
> +       struct timespec64               reset_time;
>
> [kevin]:
> the CONFIG_DEV_COREDUMP check is needed for above variable to avoid 
> compiler warning when coredump feautre is not enabled.
>
Agreed.
>          struct amdgpu_reset_domain      *reset_domain;
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c 
> b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> index 866b4980a6fa..ca97afe5be63 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> @@ -32,6 +32,8 @@
>  #include <linux/slab.h>
>  #include <linux/iommu.h>
>  #include <linux/pci.h>
> +#include <linux/devcoredump.h>
> +#include <generated/utsrelease.h>
>
>  #include <drm/drm_atomic_helper.h>
>  #include <drm/drm_probe_helper.h>
> @@ -4734,6 +4736,62 @@ static int amdgpu_reset_reg_dumps(struct 
> amdgpu_device *adev)
>          return 0;
>  }
>
> +#ifdef CONFIG_DEV_COREDUMP
> +static ssize_t amdgpu_devcoredump_read(char *buffer, loff_t offset,
> +               size_t count, void *data, size_t datalen)
> +{
> +       struct drm_printer p;
> +       struct amdgpu_device *adev = data;
> +       struct drm_print_iterator iter;
> +       int i;
> +
> +       if (adev == NULL)
> +               return 0;
> [kevin]:
>  this check is not needed, because this private data is passed by our 
> driver as below:
>
In my testing if the reset is unsuccessful amdgpu_devcoredump_read will 
not be called.

Shashank: Any inputs on this.


Regards,

S.Amarnath

>  dev_coredumpm(dev->dev, THIS_MODULE, adev, 0, GFP_KERNEL,
>  amdgpu_devcoredump_read, amdgpu_devcoredump_free);
> +
> +       iter.data = buffer;
> +       iter.offset = 0;
> +       iter.start = offset;
> +       iter.remain = count;
> +
> +       p = drm_coredump_printer(&iter);
> +
> +       drm_printf(&p, "**** AMDGPU Device Coredump ****\n");
> +       drm_printf(&p, "kernel: " UTS_RELEASE "\n");
> +       drm_printf(&p, "module: " KBUILD_MODNAME "\n");
> +       drm_printf(&p, "time: %lld.%09ld\n", adev->reset_time.tv_sec, 
> adev->reset_time.tv_nsec);
> +       if (adev->reset_task_info.pid)
> +               drm_printf(&p, "process_name: %s PID: %d\n",
> + adev->reset_task_info.process_name,
> + adev->reset_task_info.pid);
> +
> +       if (adev->reset_vram_lost)
> +               drm_printf(&p, "VRAM is lost due to GPU reset!\n");
> +       if (adev->num_regs) {
> +               drm_printf(&p, "AMDGPU register dumps:\nOffset:     
> Value:\n");
> +
> +               for (i = 0; i < adev->num_regs; i++)
> +                       drm_printf(&p, "0x%08x: 0x%08x\n",
> + adev->reset_dump_reg_list[i],
> + adev->reset_dump_reg_value[i]);
> +       }
> +
> +       return count - iter.remain;
> +}
> +
> +static void amdgpu_devcoredump_free(void *data)
> +{
> +}
> +
> +static void amdgpu_reset_capture_coredumpm(struct amdgpu_device *adev)
> +{
> +       struct drm_device *dev = adev_to_drm(adev);
> +
> +       ktime_get_ts64(&adev->reset_time);
> +       dev_coredumpm(dev->dev, THIS_MODULE, adev, 0, GFP_KERNEL,
> +                       amdgpu_devcoredump_read, amdgpu_devcoredump_free);
> +}
> +#endif
> +
>  int amdgpu_do_asic_reset(struct list_head *device_list_handle,
>                           struct amdgpu_reset_context *reset_context)
>  {
> @@ -4818,6 +4876,15 @@ int amdgpu_do_asic_reset(struct list_head 
> *device_list_handle,
>                                          goto out;
>
>                                  vram_lost = 
> amdgpu_device_check_vram_lost(tmp_adev);
> +#ifdef CONFIG_DEV_COREDUMP
> + tmp_adev->reset_vram_lost = vram_lost;
> + memset(&tmp_adev->reset_task_info, 0,
> + sizeof(tmp_adev->reset_task_info));
> +                               if (reset_context->job && 
> reset_context->job->vm)
> + tmp_adev->reset_task_info =
> + reset_context->job->vm->task_info;
> + amdgpu_reset_capture_coredumpm(tmp_adev);
> +#endif
>                                  if (vram_lost) {
>                                          DRM_INFO("VRAM is lost due to 
> GPU reset!\n");
> amdgpu_inc_vram_lost(tmp_adev);
> -- 
> 2.32.0
>

[-- Attachment #2: Type: text/html, Size: 19292 bytes --]

^ permalink raw reply	[flat|nested] 9+ messages in thread

* RE: [PATCH v2 1/2] drm/amdgpu: save the reset dump register value for devcoredump
  2022-05-31 13:08   ` Somalapuram, Amaranath
@ 2022-05-31 13:17     ` Sharma, Shashank
  2022-05-31 13:24       ` Christian König
  0 siblings, 1 reply; 9+ messages in thread
From: Sharma, Shashank @ 2022-05-31 13:17 UTC (permalink / raw)
  To: Somalapuram, Amaranath, amd-gfx; +Cc: Deucher, Alexander, Koenig, Christian


[-- Attachment #1.1: Type: text/plain, Size: 3945 bytes --]

It’s not correct even in the snapshot 😃.

[cid:image001.png@01D87501.96BF5BD0]

The *a*dev->reset_dump line should start just below the open brace in the line above “dumps*(*adev->reset”.

Check the Linux kernel coding guidelines for extending the line below the open brace.


Regards
Shashank
From: Somalapuram, Amaranath <Amaranath.Somalapuram@amd.com>
Sent: Tuesday, May 31, 2022 3:09 PM
To: Sharma, Shashank <Shashank.Sharma@amd.com>; Somalapuram, Amaranath <Amaranath.Somalapuram@amd.com>; amd-gfx@lists.freedesktop.org
Cc: Koenig, Christian <Christian.Koenig@amd.com>; Deucher, Alexander <Alexander.Deucher@amd.com>
Subject: Re: [PATCH v2 1/2] drm/amdgpu: save the reset dump register value for devcoredump



On 5/26/2022 3:24 PM, Sharma, Shashank wrote:
Hey Amar,

On 5/26/2022 11:48 AM, Somalapuram Amaranath wrote:

Allocate memory for register value and use the same values for devcoredump.
v1 -> V2: Change krealloc_array() to kmalloc_array()

Signed-off-by: Somalapuram Amaranath <Amaranath.Somalapuram@amd.com><mailto:Amaranath.Somalapuram@amd.com>
---
  drivers/gpu/drm/amd/amdgpu/amdgpu.h         | 1 +
  drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c | 7 +++++++
  drivers/gpu/drm/amd/amdgpu/amdgpu_device.c  | 6 +++---
  3 files changed, 11 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 76df583663c7..c79d9992b113 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -1042,6 +1042,7 @@ struct amdgpu_device {
        /* reset dump register */
      uint32_t                        *reset_dump_reg_list;
+    uint32_t            *reset_dump_reg_value;
      int                             num_regs;
        struct amdgpu_reset_domain    *reset_domain;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
index eedb12f6b8a3..f3ac7912c29c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
@@ -1709,17 +1709,24 @@ static ssize_t amdgpu_reset_dump_register_list_write(struct file *f,
          i++;
      } while (len < size);
  +    new = kmalloc_array(i, sizeof(uint32_t), GFP_KERNEL);
+    if (!new) {
+        ret = -ENOMEM;
+        goto error_free;
+    }
      ret = down_write_killable(&adev->reset_domain->sem);
      if (ret)
          goto error_free;
        swap(adev->reset_dump_reg_list, tmp);
+    swap(adev->reset_dump_reg_value, new);
      adev->num_regs = i;
      up_write(&adev->reset_domain->sem);
      ret = size;
    error_free:
      kfree(tmp);
+    kfree(new);
      return ret;
  }
  diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 4daa0e893965..866b4980a6fa 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -4720,15 +4720,15 @@ int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev,
    static int amdgpu_reset_reg_dumps(struct amdgpu_device *adev)
  {
-    uint32_t reg_value;
      int i;
        lockdep_assert_held(&adev->reset_domain->sem);
      dump_stack();
        for (i = 0; i < adev->num_regs; i++) {
-        reg_value = RREG32(adev->reset_dump_reg_list[i]);
-        trace_amdgpu_reset_reg_dumps(adev->reset_dump_reg_list[i], reg_value);
+        adev->reset_dump_reg_value[i] = RREG32(adev->reset_dump_reg_list[i]);
+        trace_amdgpu_reset_reg_dumps(adev->reset_dump_reg_list[i],
+                adev->reset_dump_reg_value[i]);

Alignment is showing spaces in email. But its 2 tabs in code.

Regards,
S.Amarnath


Please fix the alignment with the line above, after that, please feel free to use:
Reviewed-by: Shashank Sharma <shashank.sharma@amd.com><mailto:shashank.sharma@amd.com>

- Shashank

      }
        return 0;

[-- Attachment #1.2: Type: text/html, Size: 9080 bytes --]

[-- Attachment #2: image001.png --]
[-- Type: image/png, Size: 10366 bytes --]

^ permalink raw reply related	[flat|nested] 9+ messages in thread

* Re: [PATCH v2 1/2] drm/amdgpu: save the reset dump register value for devcoredump
  2022-05-31 13:17     ` Sharma, Shashank
@ 2022-05-31 13:24       ` Christian König
  0 siblings, 0 replies; 9+ messages in thread
From: Christian König @ 2022-05-31 13:24 UTC (permalink / raw)
  To: Sharma, Shashank, Somalapuram, Amaranath, amd-gfx; +Cc: Deucher, Alexander

[-- Attachment #1: Type: text/plain, Size: 5071 bytes --]

I strongly recommend installing coding style setting for your favorite 
editor.

Those should be available for pretty much every more widely used code 
editor.

Regards,
Christian.

Am 31.05.22 um 15:17 schrieb Sharma, Shashank:
>
> It’s not correct even in the snapshot 😃.
>
> The **a**dev->reset_dump line should start just below the open brace 
> in the line above “dumps*(*adev->reset”.
>
> Check the Linux kernel coding guidelines for extending the line below 
> the open brace.
>
> Regards
>
> Shashank
>
> *From:* Somalapuram, Amaranath <Amaranath.Somalapuram@amd.com>
> *Sent:* Tuesday, May 31, 2022 3:09 PM
> *To:* Sharma, Shashank <Shashank.Sharma@amd.com>; Somalapuram, 
> Amaranath <Amaranath.Somalapuram@amd.com>; amd-gfx@lists.freedesktop.org
> *Cc:* Koenig, Christian <Christian.Koenig@amd.com>; Deucher, Alexander 
> <Alexander.Deucher@amd.com>
> *Subject:* Re: [PATCH v2 1/2] drm/amdgpu: save the reset dump register 
> value for devcoredump
>
> On 5/26/2022 3:24 PM, Sharma, Shashank wrote:
>
>     Hey Amar,
>
>     On 5/26/2022 11:48 AM, Somalapuram Amaranath wrote:
>
>         Allocate memory for register value and use the same values for
>         devcoredump.
>         v1 -> V2: Change krealloc_array() to kmalloc_array()
>
>         Signed-off-by: Somalapuram Amaranath
>         <Amaranath.Somalapuram@amd.com>
>         <mailto:Amaranath.Somalapuram@amd.com>
>         ---
>           drivers/gpu/drm/amd/amdgpu/amdgpu.h         | 1 +
>           drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c | 7 +++++++
>           drivers/gpu/drm/amd/amdgpu/amdgpu_device.c  | 6 +++---
>           3 files changed, 11 insertions(+), 3 deletions(-)
>
>         diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
>         b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
>         index 76df583663c7..c79d9992b113 100644
>         --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
>         +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
>         @@ -1042,6 +1042,7 @@ struct amdgpu_device {
>                 /* reset dump register */
>               uint32_t *reset_dump_reg_list;
>         +    uint32_t            *reset_dump_reg_value;
>               int                             num_regs;
>                 struct amdgpu_reset_domain    *reset_domain;
>         diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
>         b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
>         index eedb12f6b8a3..f3ac7912c29c 100644
>         --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
>         +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
>         @@ -1709,17 +1709,24 @@ static ssize_t
>         amdgpu_reset_dump_register_list_write(struct file *f,
>                   i++;
>               } while (len < size);
>           +    new = kmalloc_array(i, sizeof(uint32_t), GFP_KERNEL);
>         +    if (!new) {
>         +        ret = -ENOMEM;
>         +        goto error_free;
>         +    }
>               ret = down_write_killable(&adev->reset_domain->sem);
>               if (ret)
>                   goto error_free;
>                 swap(adev->reset_dump_reg_list, tmp);
>         +    swap(adev->reset_dump_reg_value, new);
>               adev->num_regs = i;
>               up_write(&adev->reset_domain->sem);
>               ret = size;
>             error_free:
>               kfree(tmp);
>         +    kfree(new);
>               return ret;
>           }
>           diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
>         b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
>         index 4daa0e893965..866b4980a6fa 100644
>         --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
>         +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
>         @@ -4720,15 +4720,15 @@ int
>         amdgpu_device_pre_asic_reset(struct amdgpu_device *adev,
>             static int amdgpu_reset_reg_dumps(struct amdgpu_device *adev)
>           {
>         -    uint32_t reg_value;
>               int i;
>         lockdep_assert_held(&adev->reset_domain->sem);
>               dump_stack();
>                 for (i = 0; i < adev->num_regs; i++) {
>         -        reg_value = RREG32(adev->reset_dump_reg_list[i]);
>         - trace_amdgpu_reset_reg_dumps(adev->reset_dump_reg_list[i],
>         reg_value);
>         +        adev->reset_dump_reg_value[i] =
>         RREG32(adev->reset_dump_reg_list[i]);
>         + trace_amdgpu_reset_reg_dumps(adev->reset_dump_reg_list[i],
>         +                adev->reset_dump_reg_value[i]);
>
> Alignment is showing spaces in email. But its 2 tabs in code.
>
>
> Regards,
> S.Amarnath
>
>     Please fix the alignment with the line above, after that, please
>     feel free to use:
>     Reviewed-by: Shashank Sharma <shashank.sharma@amd.com>
>     <mailto:shashank.sharma@amd.com>
>
>     - Shashank
>
>               }
>                 return 0;
>

[-- Attachment #2.1: Type: text/html, Size: 11383 bytes --]

[-- Attachment #2.2: image001.png --]
[-- Type: image/png, Size: 10366 bytes --]

^ permalink raw reply	[flat|nested] 9+ messages in thread

end of thread, other threads:[~2022-05-31 13:24 UTC | newest]

Thread overview: 9+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-05-26  9:48 [PATCH v2 1/2] drm/amdgpu: save the reset dump register value for devcoredump Somalapuram Amaranath
2022-05-26  9:48 ` [PATCH v2 2/2] drm/amdgpu: adding device coredump support Somalapuram Amaranath
2022-05-26  9:57   ` Sharma, Shashank
2022-05-26 10:26   ` Wang, Yang(Kevin)
2022-05-31 13:14     ` Somalapuram, Amaranath
2022-05-26  9:54 ` [PATCH v2 1/2] drm/amdgpu: save the reset dump register value for devcoredump Sharma, Shashank
2022-05-31 13:08   ` Somalapuram, Amaranath
2022-05-31 13:17     ` Sharma, Shashank
2022-05-31 13:24       ` Christian König

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.