Commit f1549c09 authored by Likun Gao's avatar Likun Gao Committed by Alex Deucher
Browse files

drm/amdgpu: support reset flag set for gpu reset 



Move reset_context out of gpu recover function to make it configurable
for different reset purpose.
For the reset way of call gpu_recovery sysfs, force to use full reset
method. Otherwise, try soft reset by default if the related ASIC
supportted, if soft reset failed, will use full reset.

Signed-off-by: default avatarLikun Gao <Likun.Gao@amd.com>
Reviewed-by: default avatarHawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: default avatarAlex Deucher <alexander.deucher@amd.com>
parent 58e969b6

drivers/gpu/drm/amd/amdgpu/amdgpu.h

+2 −3
Original line number Diff line number Diff line
@@ -1253,9 +1253,8 @@ int emu_soc_asic_init(struct amdgpu_device *adev); 
bool amdgpu_device_has_job_running(struct amdgpu_device *adev);

bool amdgpu_device_should_recover_gpu(struct amdgpu_device *adev);

int amdgpu_device_gpu_recover(struct amdgpu_device *adev,

			      struct amdgpu_job* job);

int amdgpu_device_gpu_recover(struct amdgpu_device *adev,

			      struct amdgpu_job *job);

			      struct amdgpu_job *job,

			      struct amdgpu_reset_context *reset_context);

void amdgpu_device_pci_config_reset(struct amdgpu_device *adev);

int amdgpu_device_pci_reset(struct amdgpu_device *adev);

bool amdgpu_device_need_post(struct amdgpu_device *adev);

drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c

+8 −1
Original line number Diff line number Diff line
@@ -129,7 +129,14 @@ static void amdgpu_amdkfd_reset_work(struct work_struct *work) 
	struct amdgpu_device *adev = container_of(work, struct amdgpu_device,

						  kfd.reset_work);



	amdgpu_device_gpu_recover(adev, NULL);

	struct amdgpu_reset_context reset_context;

	memset(&reset_context, 0, sizeof(reset_context));



	reset_context.method = AMD_RESET_METHOD_NONE;

	reset_context.reset_req_dev = adev;

	clear_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags);



	amdgpu_device_gpu_recover(adev, NULL, &reset_context);

}



void amdgpu_amdkfd_device_init(struct amdgpu_device *adev)

drivers/gpu/drm/amd/amdgpu/amdgpu_device.c

+7 −13
Original line number Diff line number Diff line
@@ -5109,7 +5109,8 @@ static inline void amdgpu_device_stop_pending_resets(struct amdgpu_device *adev) 
 */



int amdgpu_device_gpu_recover(struct amdgpu_device *adev,

			      struct amdgpu_job *job)

			      struct amdgpu_job *job,

			      struct amdgpu_reset_context *reset_context)

{

	struct list_head device_list, *device_list_handle =  NULL;

	bool job_signaled = false;
@@ -5119,9 +5120,6 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev, 
	bool need_emergency_restart = false;

	bool audio_suspended = false;

	int tmp_vram_lost_counter;

	struct amdgpu_reset_context reset_context;



	memset(&reset_context, 0, sizeof(reset_context));



	/*

	 * Special case: RAS triggered and full reset isn't supported
@@ -5147,12 +5145,8 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev, 
	if (hive)

		mutex_lock(&hive->hive_lock);



	reset_context.method = AMD_RESET_METHOD_NONE;

	reset_context.reset_req_dev = adev;

	reset_context.job = job;

	reset_context.hive = hive;

	clear_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags);



	reset_context->job = job;

	reset_context->hive = hive;

	/*

	 * Build list of devices to reset.

	 * In case we are in XGMI hive mode, resort the device list
@@ -5245,7 +5239,7 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev, 


retry:	/* Rest of adevs pre asic reset from XGMI hive. */

	list_for_each_entry(tmp_adev, device_list_handle, reset_list) {

		r = amdgpu_device_pre_asic_reset(tmp_adev, &reset_context);

		r = amdgpu_device_pre_asic_reset(tmp_adev, reset_context);

		/*TODO Should we stop ?*/

		if (r) {

			dev_err(tmp_adev->dev, "GPU pre asic reset failed with err, %d for drm dev, %s ",
@@ -5272,7 +5266,7 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev, 
		if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 2))

			amdgpu_ras_resume(adev);

	} else {

		r = amdgpu_do_asic_reset(device_list_handle, &reset_context);

		r = amdgpu_do_asic_reset(device_list_handle, reset_context);

		if (r && r == -EAGAIN)

			goto retry;

	}
@@ -5292,7 +5286,7 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev, 
		if (amdgpu_gpu_recovery == 2 &&

			!(tmp_vram_lost_counter < atomic_read(&adev->vram_lost_counter)))

			amdgpu_device_recheck_guilty_jobs(

				tmp_adev, device_list_handle, &reset_context);

				tmp_adev, device_list_handle, reset_context);



		for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {

			struct amdgpu_ring *ring = tmp_adev->rings[i];

drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c

+8 −1
Original line number Diff line number Diff line
@@ -844,7 +844,14 @@ static void amdgpu_debugfs_reset_work(struct work_struct *work) 
	struct amdgpu_device *adev = container_of(work, struct amdgpu_device,

						  reset_work);



	amdgpu_device_gpu_recover(adev, NULL);

	struct amdgpu_reset_context reset_context;

	memset(&reset_context, 0, sizeof(reset_context));



	reset_context.method = AMD_RESET_METHOD_NONE;

	reset_context.reset_req_dev = adev;

	set_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags);



	amdgpu_device_gpu_recover(adev, NULL, &reset_context);

}



#endif

drivers/gpu/drm/amd/amdgpu/amdgpu_job.c

+9 −1
Original line number Diff line number Diff line
@@ -29,6 +29,7 @@ 


#include "amdgpu.h"

#include "amdgpu_trace.h"

#include "amdgpu_reset.h"



static enum drm_gpu_sched_stat amdgpu_job_timedout(struct drm_sched_job *s_job)

{
@@ -64,7 +65,14 @@ static enum drm_gpu_sched_stat amdgpu_job_timedout(struct drm_sched_job *s_job) 
		  ti.process_name, ti.tgid, ti.task_name, ti.pid);



	if (amdgpu_device_should_recover_gpu(ring->adev)) {

		r = amdgpu_device_gpu_recover(ring->adev, job);

		struct amdgpu_reset_context reset_context;

		memset(&reset_context, 0, sizeof(reset_context));



		reset_context.method = AMD_RESET_METHOD_NONE;

		reset_context.reset_req_dev = adev;

		clear_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags);



		r = amdgpu_device_gpu_recover(ring->adev, job, &reset_context);

		if (r)

			DRM_ERROR("GPU Recovery Failed: %d\n", r);

	} else {