Merge branch 'drm-next-4.13' of git://people.freedesktop.org/~agd5f/linux into drm-next

New radeon and amdgpu features for 4.13:
- Lots of Vega10 bug fixes
- Preliminary Raven support
- KIQ support for compute rings
- MEC queue management rework from Andres
- Audio support for DCE6
- SR-IOV improvements
- Improved module parameters for controlling radeon vs amdgpu support
  for SI and CIK
- Bug fixes
- General code cleanups

[airlied: dropped drmP.h header from one file was needed and build broke]

* 'drm-next-4.13' of git://people.freedesktop.org/~agd5f/linux: (362 commits)
  drm/amdgpu: Fix compiler warnings
  drm/amdgpu: vm_update_ptes remove code duplication
  drm/amd/amdgpu: Port VCN over to new SOC15 macros
  drm/amd/amdgpu: Port PSP v10.0 over to new SOC15 macros
  drm/amd/amdgpu: Port PSP v3.1 over to new SOC15 macros
  drm/amd/amdgpu: Port NBIO v7.0 driver over to new SOC15 macros
  drm/amd/amdgpu: Port NBIO v6.1 driver over to new SOC15 macros
  drm/amd/amdgpu: Port UVD 7.0 over to new SOC15 macros
  drm/amd/amdgpu: Port MMHUB over to new SOC15 macros
  drm/amd/amdgpu: Cleanup gfxhub read-modify-write patterns
  drm/amd/amdgpu: Port GFXHUB over to new SOC15 macros
  drm/amd/amdgpu: Add offset variant to SOC15 macros
  drm/amd/powerplay: add avfs control for Vega10
  drm/amdgpu: add virtual display support for raven
  drm/amdgpu/gfx9: fix compute ring doorbell index
  drm/amd/amdgpu: Rename KIQ ring to avoid spaces
  drm/amd/amdgpu: gfx9 tidy ups (v2)
  drm/amdgpu: add contiguous flag in ucode bo create
  drm/amdgpu: fix missed gpu info firmware when cache firmware during S3
  drm/amdgpu: export test ib debugfs interface
  ...

drivers/gpu/drm/amd/amdgpu/Kconfig

@@ -5,15 +5,23 @@ config DRM_AMDGPU_SI
 	  Choose this option if you want to enable experimental support
 	  for SI asics.
 
+	  SI is already supported in radeon. Experimental support for SI
+	  in amdgpu will be disabled by default and is still provided by
+	  radeon. Use module options to override this:
+
+	  radeon.si_support=0 amdgpu.si_support=1
+
 config DRM_AMDGPU_CIK
 	bool "Enable amdgpu support for CIK parts"
 	depends on DRM_AMDGPU
 	help
-	  Choose this option if you want to enable experimental support
-	  for CIK asics.
+	  Choose this option if you want to enable support for CIK asics.
+
+	  CIK is already supported in radeon. Support for CIK in amdgpu
+	  will be disabled by default and is still provided by radeon.
+	  Use module options to override this:
 
-	  CIK is already supported in radeon.  CIK support in amdgpu
-	  is for experimentation and testing.
+	  radeon.cik_support=0 amdgpu.cik_support=1
 
 config DRM_AMDGPU_USERPTR
 	bool "Always enable userptr write support"

drivers/gpu/drm/amd/amdgpu/Makefile

@@ -24,7 +24,8 @@ amdgpu-y += amdgpu_device.o amdgpu_kms.o \
 	atombios_encoders.o amdgpu_sa.o atombios_i2c.o \
 	amdgpu_prime.o amdgpu_vm.o amdgpu_ib.o amdgpu_pll.o \
 	amdgpu_ucode.o amdgpu_bo_list.o amdgpu_ctx.o amdgpu_sync.o \
-	amdgpu_gtt_mgr.o amdgpu_vram_mgr.o amdgpu_virt.o amdgpu_atomfirmware.o
+	amdgpu_gtt_mgr.o amdgpu_vram_mgr.o amdgpu_virt.o amdgpu_atomfirmware.o \
+	amdgpu_queue_mgr.o
 
 # add asic specific block
 amdgpu-$(CONFIG_DRM_AMDGPU_CIK)+= cik.o cik_ih.o kv_smc.o kv_dpm.o \
@@ -34,7 +35,7 @@ amdgpu-$(CONFIG_DRM_AMDGPU_CIK)+= cik.o cik_ih.o kv_smc.o kv_dpm.o \
 amdgpu-$(CONFIG_DRM_AMDGPU_SI)+= si.o gmc_v6_0.o gfx_v6_0.o si_ih.o si_dma.o dce_v6_0.o si_dpm.o si_smc.o
 
 amdgpu-y += \
-	vi.o mxgpu_vi.o nbio_v6_1.o soc15.o mxgpu_ai.o
+	vi.o mxgpu_vi.o nbio_v6_1.o soc15.o mxgpu_ai.o nbio_v7_0.o
 
 # add GMC block
 amdgpu-y += \
@@ -54,7 +55,8 @@ amdgpu-y += \
 # add PSP block
 amdgpu-y += \
 	amdgpu_psp.o \
-	psp_v3_1.o
+	psp_v3_1.o \
+	psp_v10_0.o
 
 # add SMC block
 amdgpu-y += \
@@ -92,6 +94,11 @@ amdgpu-y += \
 	vce_v3_0.o \
 	vce_v4_0.o
 
+# add VCN block
+amdgpu-y += \
+	amdgpu_vcn.o \
+	vcn_v1_0.o
+
 # add amdkfd interfaces
 amdgpu-y += \
 	 amdgpu_amdkfd.o \

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

@@ -24,6 +24,7 @@
 #include "amd_shared.h"
 #include <drm/drmP.h>
 #include "amdgpu.h"
+#include "amdgpu_gfx.h"
 #include <linux/module.h>
 
 const struct kfd2kgd_calls *kfd2kgd;
@@ -60,9 +61,9 @@ int amdgpu_amdkfd_init(void)
 	return ret;
 }
 
-bool amdgpu_amdkfd_load_interface(struct amdgpu_device *rdev)
+bool amdgpu_amdkfd_load_interface(struct amdgpu_device *adev)
 {
-	switch (rdev->asic_type) {
+	switch (adev->asic_type) {
 #ifdef CONFIG_DRM_AMDGPU_CIK
 	case CHIP_KAVERI:
 		kfd2kgd = amdgpu_amdkfd_gfx_7_get_functions();
@@ -86,59 +87,83 @@ void amdgpu_amdkfd_fini(void)
 	}
 }
 
-void amdgpu_amdkfd_device_probe(struct amdgpu_device *rdev)
+void amdgpu_amdkfd_device_probe(struct amdgpu_device *adev)
 {
 	if (kgd2kfd)
-		rdev->kfd = kgd2kfd->probe((struct kgd_dev *)rdev,
-					rdev->pdev, kfd2kgd);
+		adev->kfd = kgd2kfd->probe((struct kgd_dev *)adev,
+					adev->pdev, kfd2kgd);
 }
 
-void amdgpu_amdkfd_device_init(struct amdgpu_device *rdev)
+void amdgpu_amdkfd_device_init(struct amdgpu_device *adev)
 {
-	if (rdev->kfd) {
+	int i;
+	int last_valid_bit;
+	if (adev->kfd) {
 		struct kgd2kfd_shared_resources gpu_resources = {
 			.compute_vmid_bitmap = 0xFF00,
-
-			.first_compute_pipe = 1,
-			.compute_pipe_count = 4 - 1,
+			.num_mec = adev->gfx.mec.num_mec,
+			.num_pipe_per_mec = adev->gfx.mec.num_pipe_per_mec,
+			.num_queue_per_pipe = adev->gfx.mec.num_queue_per_pipe
 		};
 
-		amdgpu_doorbell_get_kfd_info(rdev,
+		/* this is going to have a few of the MSBs set that we need to
+		 * clear */
+		bitmap_complement(gpu_resources.queue_bitmap,
+				  adev->gfx.mec.queue_bitmap,
+				  KGD_MAX_QUEUES);
+
+		/* remove the KIQ bit as well */
+		if (adev->gfx.kiq.ring.ready)
+			clear_bit(amdgpu_gfx_queue_to_bit(adev,
+							  adev->gfx.kiq.ring.me - 1,
+							  adev->gfx.kiq.ring.pipe,
+							  adev->gfx.kiq.ring.queue),
+				  gpu_resources.queue_bitmap);
+
+		/* According to linux/bitmap.h we shouldn't use bitmap_clear if
+		 * nbits is not compile time constant */
+		last_valid_bit = adev->gfx.mec.num_mec
+				* adev->gfx.mec.num_pipe_per_mec
+				* adev->gfx.mec.num_queue_per_pipe;
+		for (i = last_valid_bit; i < KGD_MAX_QUEUES; ++i)
+			clear_bit(i, gpu_resources.queue_bitmap);
+
+		amdgpu_doorbell_get_kfd_info(adev,
 				&gpu_resources.doorbell_physical_address,
 				&gpu_resources.doorbell_aperture_size,
 				&gpu_resources.doorbell_start_offset);
 
-		kgd2kfd->device_init(rdev->kfd, &gpu_resources);
+		kgd2kfd->device_init(adev->kfd, &gpu_resources);
 	}
 }
 
-void amdgpu_amdkfd_device_fini(struct amdgpu_device *rdev)
+void amdgpu_amdkfd_device_fini(struct amdgpu_device *adev)
 {
-	if (rdev->kfd) {
-		kgd2kfd->device_exit(rdev->kfd);
-		rdev->kfd = NULL;
+	if (adev->kfd) {
+		kgd2kfd->device_exit(adev->kfd);
+		adev->kfd = NULL;
 	}
 }
 
-void amdgpu_amdkfd_interrupt(struct amdgpu_device *rdev,
+void amdgpu_amdkfd_interrupt(struct amdgpu_device *adev,
 		const void *ih_ring_entry)
 {
-	if (rdev->kfd)
-		kgd2kfd->interrupt(rdev->kfd, ih_ring_entry);
+	if (adev->kfd)
+		kgd2kfd->interrupt(adev->kfd, ih_ring_entry);
 }
 
-void amdgpu_amdkfd_suspend(struct amdgpu_device *rdev)
+void amdgpu_amdkfd_suspend(struct amdgpu_device *adev)
 {
-	if (rdev->kfd)
-		kgd2kfd->suspend(rdev->kfd);
+	if (adev->kfd)
+		kgd2kfd->suspend(adev->kfd);
 }
 
-int amdgpu_amdkfd_resume(struct amdgpu_device *rdev)
+int amdgpu_amdkfd_resume(struct amdgpu_device *adev)
 {
 	int r = 0;
 
-	if (rdev->kfd)
-		r = kgd2kfd->resume(rdev->kfd);
+	if (adev->kfd)
+		r = kgd2kfd->resume(adev->kfd);
 
 	return r;
 }
@@ -147,7 +172,7 @@ int alloc_gtt_mem(struct kgd_dev *kgd, size_t size,
 			void **mem_obj, uint64_t *gpu_addr,
 			void **cpu_ptr)
 {
-	struct amdgpu_device *rdev = (struct amdgpu_device *)kgd;
+	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
 	struct kgd_mem **mem = (struct kgd_mem **) mem_obj;
 	int r;
 
@@ -159,10 +184,10 @@ int alloc_gtt_mem(struct kgd_dev *kgd, size_t size,
 	if ((*mem) == NULL)
 		return -ENOMEM;
 
-	r = amdgpu_bo_create(rdev, size, PAGE_SIZE, true, AMDGPU_GEM_DOMAIN_GTT,
+	r = amdgpu_bo_create(adev, size, PAGE_SIZE, true, AMDGPU_GEM_DOMAIN_GTT,
 			     AMDGPU_GEM_CREATE_CPU_GTT_USWC, NULL, NULL, &(*mem)->bo);
 	if (r) {
-		dev_err(rdev->dev,
+		dev_err(adev->dev,
 			"failed to allocate BO for amdkfd (%d)\n", r);
 		return r;
 	}
@@ -170,21 +195,21 @@ int alloc_gtt_mem(struct kgd_dev *kgd, size_t size,
 	/* map the buffer */
 	r = amdgpu_bo_reserve((*mem)->bo, true);
 	if (r) {
-		dev_err(rdev->dev, "(%d) failed to reserve bo for amdkfd\n", r);
+		dev_err(adev->dev, "(%d) failed to reserve bo for amdkfd\n", r);
 		goto allocate_mem_reserve_bo_failed;
 	}
 
 	r = amdgpu_bo_pin((*mem)->bo, AMDGPU_GEM_DOMAIN_GTT,
 				&(*mem)->gpu_addr);
 	if (r) {
-		dev_err(rdev->dev, "(%d) failed to pin bo for amdkfd\n", r);
+		dev_err(adev->dev, "(%d) failed to pin bo for amdkfd\n", r);
 		goto allocate_mem_pin_bo_failed;
 	}
 	*gpu_addr = (*mem)->gpu_addr;
 
 	r = amdgpu_bo_kmap((*mem)->bo, &(*mem)->cpu_ptr);
 	if (r) {
-		dev_err(rdev->dev,
+		dev_err(adev->dev,
 			"(%d) failed to map bo to kernel for amdkfd\n", r);
 		goto allocate_mem_kmap_bo_failed;
 	}
@@ -220,27 +245,27 @@ void free_gtt_mem(struct kgd_dev *kgd, void *mem_obj)
 
 uint64_t get_vmem_size(struct kgd_dev *kgd)
 {
-	struct amdgpu_device *rdev =
+	struct amdgpu_device *adev =
 		(struct amdgpu_device *)kgd;
 
 	BUG_ON(kgd == NULL);
 
-	return rdev->mc.real_vram_size;
+	return adev->mc.real_vram_size;
 }
 
 uint64_t get_gpu_clock_counter(struct kgd_dev *kgd)
 {
-	struct amdgpu_device *rdev = (struct amdgpu_device *)kgd;
+	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
 
-	if (rdev->gfx.funcs->get_gpu_clock_counter)
-		return rdev->gfx.funcs->get_gpu_clock_counter(rdev);
+	if (adev->gfx.funcs->get_gpu_clock_counter)
+		return adev->gfx.funcs->get_gpu_clock_counter(adev);
 	return 0;
 }
 
 uint32_t get_max_engine_clock_in_mhz(struct kgd_dev *kgd)
 {
-	struct amdgpu_device *rdev = (struct amdgpu_device *)kgd;
+	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
 
 	/* The sclk is in quantas of 10kHz */
-	return rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.sclk / 100;
+	return adev->pm.dpm.dyn_state.max_clock_voltage_on_ac.sclk / 100;
 }

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

@@ -39,15 +39,15 @@ struct kgd_mem {
 int amdgpu_amdkfd_init(void);
 void amdgpu_amdkfd_fini(void);
 
-bool amdgpu_amdkfd_load_interface(struct amdgpu_device *rdev);
+bool amdgpu_amdkfd_load_interface(struct amdgpu_device *adev);
 
-void amdgpu_amdkfd_suspend(struct amdgpu_device *rdev);
-int amdgpu_amdkfd_resume(struct amdgpu_device *rdev);
-void amdgpu_amdkfd_interrupt(struct amdgpu_device *rdev,
+void amdgpu_amdkfd_suspend(struct amdgpu_device *adev);
+int amdgpu_amdkfd_resume(struct amdgpu_device *adev);
+void amdgpu_amdkfd_interrupt(struct amdgpu_device *adev,
 			const void *ih_ring_entry);
-void amdgpu_amdkfd_device_probe(struct amdgpu_device *rdev);
-void amdgpu_amdkfd_device_init(struct amdgpu_device *rdev);
-void amdgpu_amdkfd_device_fini(struct amdgpu_device *rdev);
+void amdgpu_amdkfd_device_probe(struct amdgpu_device *adev);
+void amdgpu_amdkfd_device_init(struct amdgpu_device *adev);
+void amdgpu_amdkfd_device_fini(struct amdgpu_device *adev);
 
 struct kfd2kgd_calls *amdgpu_amdkfd_gfx_7_get_functions(void);
 struct kfd2kgd_calls *amdgpu_amdkfd_gfx_8_0_get_functions(void);

drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c

@@ -29,6 +29,7 @@
 #include "cikd.h"
 #include "cik_sdma.h"
 #include "amdgpu_ucode.h"
+#include "gfx_v7_0.h"
 #include "gca/gfx_7_2_d.h"
 #include "gca/gfx_7_2_enum.h"
 #include "gca/gfx_7_2_sh_mask.h"
@@ -38,8 +39,6 @@
 #include "gmc/gmc_7_1_sh_mask.h"
 #include "cik_structs.h"
 
-#define CIK_PIPE_PER_MEC	(4)
-
 enum {
 	MAX_TRAPID = 8,		/* 3 bits in the bitfield. */
 	MAX_WATCH_ADDRESSES = 4
@@ -185,8 +184,10 @@ static void unlock_srbm(struct kgd_dev *kgd)
 static void acquire_queue(struct kgd_dev *kgd, uint32_t pipe_id,
 				uint32_t queue_id)
 {
-	uint32_t mec = (++pipe_id / CIK_PIPE_PER_MEC) + 1;
-	uint32_t pipe = (pipe_id % CIK_PIPE_PER_MEC);
+	struct amdgpu_device *adev = get_amdgpu_device(kgd);
+
+	uint32_t mec = (++pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
+	uint32_t pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
 
 	lock_srbm(kgd, mec, pipe, queue_id, 0);
 }
@@ -243,18 +244,7 @@ static int kgd_set_pasid_vmid_mapping(struct kgd_dev *kgd, unsigned int pasid,
 static int kgd_init_pipeline(struct kgd_dev *kgd, uint32_t pipe_id,
 				uint32_t hpd_size, uint64_t hpd_gpu_addr)
 {
-	struct amdgpu_device *adev = get_amdgpu_device(kgd);
-
-	uint32_t mec = (++pipe_id / CIK_PIPE_PER_MEC) + 1;
-	uint32_t pipe = (pipe_id % CIK_PIPE_PER_MEC);
-
-	lock_srbm(kgd, mec, pipe, 0, 0);
-	WREG32(mmCP_HPD_EOP_BASE_ADDR, lower_32_bits(hpd_gpu_addr >> 8));
-	WREG32(mmCP_HPD_EOP_BASE_ADDR_HI, upper_32_bits(hpd_gpu_addr >> 8));
-	WREG32(mmCP_HPD_EOP_VMID, 0);
-	WREG32(mmCP_HPD_EOP_CONTROL, hpd_size);
-	unlock_srbm(kgd);
-
+	/* amdgpu owns the per-pipe state */
 	return 0;
 }
 
@@ -264,8 +254,8 @@ static int kgd_init_interrupts(struct kgd_dev *kgd, uint32_t pipe_id)
 	uint32_t mec;
 	uint32_t pipe;
 
-	mec = (pipe_id / CIK_PIPE_PER_MEC) + 1;
-	pipe = (pipe_id % CIK_PIPE_PER_MEC);
+	mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
+	pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
 
 	lock_srbm(kgd, mec, pipe, 0, 0);
 
@@ -309,55 +299,11 @@ static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
 	m = get_mqd(mqd);
 
 	is_wptr_shadow_valid = !get_user(wptr_shadow, wptr);
-
-	acquire_queue(kgd, pipe_id, queue_id);
-	WREG32(mmCP_MQD_BASE_ADDR, m->cp_mqd_base_addr_lo);
-	WREG32(mmCP_MQD_BASE_ADDR_HI, m->cp_mqd_base_addr_hi);
-	WREG32(mmCP_MQD_CONTROL, m->cp_mqd_control);
-
-	WREG32(mmCP_HQD_PQ_BASE, m->cp_hqd_pq_base_lo);
-	WREG32(mmCP_HQD_PQ_BASE_HI, m->cp_hqd_pq_base_hi);
-	WREG32(mmCP_HQD_PQ_CONTROL, m->cp_hqd_pq_control);
-
-	WREG32(mmCP_HQD_IB_CONTROL, m->cp_hqd_ib_control);
-	WREG32(mmCP_HQD_IB_BASE_ADDR, m->cp_hqd_ib_base_addr_lo);
-	WREG32(mmCP_HQD_IB_BASE_ADDR_HI, m->cp_hqd_ib_base_addr_hi);
-
-	WREG32(mmCP_HQD_IB_RPTR, m->cp_hqd_ib_rptr);
-
-	WREG32(mmCP_HQD_PERSISTENT_STATE, m->cp_hqd_persistent_state);
-	WREG32(mmCP_HQD_SEMA_CMD, m->cp_hqd_sema_cmd);
-	WREG32(mmCP_HQD_MSG_TYPE, m->cp_hqd_msg_type);
-
-	WREG32(mmCP_HQD_ATOMIC0_PREOP_LO, m->cp_hqd_atomic0_preop_lo);
-	WREG32(mmCP_HQD_ATOMIC0_PREOP_HI, m->cp_hqd_atomic0_preop_hi);
-	WREG32(mmCP_HQD_ATOMIC1_PREOP_LO, m->cp_hqd_atomic1_preop_lo);
-	WREG32(mmCP_HQD_ATOMIC1_PREOP_HI, m->cp_hqd_atomic1_preop_hi);
-
-	WREG32(mmCP_HQD_PQ_RPTR_REPORT_ADDR, m->cp_hqd_pq_rptr_report_addr_lo);
-	WREG32(mmCP_HQD_PQ_RPTR_REPORT_ADDR_HI,
-			m->cp_hqd_pq_rptr_report_addr_hi);
-
-	WREG32(mmCP_HQD_PQ_RPTR, m->cp_hqd_pq_rptr);
-
-	WREG32(mmCP_HQD_PQ_WPTR_POLL_ADDR, m->cp_hqd_pq_wptr_poll_addr_lo);
-	WREG32(mmCP_HQD_PQ_WPTR_POLL_ADDR_HI, m->cp_hqd_pq_wptr_poll_addr_hi);
-
-	WREG32(mmCP_HQD_PQ_DOORBELL_CONTROL, m->cp_hqd_pq_doorbell_control);
-
-	WREG32(mmCP_HQD_VMID, m->cp_hqd_vmid);
-
-	WREG32(mmCP_HQD_QUANTUM, m->cp_hqd_quantum);
-
-	WREG32(mmCP_HQD_PIPE_PRIORITY, m->cp_hqd_pipe_priority);
-	WREG32(mmCP_HQD_QUEUE_PRIORITY, m->cp_hqd_queue_priority);
-
-	WREG32(mmCP_HQD_IQ_RPTR, m->cp_hqd_iq_rptr);
-
 	if (is_wptr_shadow_valid)
-		WREG32(mmCP_HQD_PQ_WPTR, wptr_shadow);
+		m->cp_hqd_pq_wptr = wptr_shadow;
 
-	WREG32(mmCP_HQD_ACTIVE, m->cp_hqd_active);
+	acquire_queue(kgd, pipe_id, queue_id);
+	gfx_v7_0_mqd_commit(adev, m);
 	release_queue(kgd);
 
 	return 0;

drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c

@@ -28,6 +28,7 @@
 #include "amdgpu.h"
 #include "amdgpu_amdkfd.h"
 #include "amdgpu_ucode.h"
+#include "gfx_v8_0.h"
 #include "gca/gfx_8_0_sh_mask.h"
 #include "gca/gfx_8_0_d.h"
 #include "gca/gfx_8_0_enum.h"
@@ -38,8 +39,6 @@
 #include "vi_structs.h"
 #include "vid.h"
 
-#define VI_PIPE_PER_MEC	(4)
-
 struct cik_sdma_rlc_registers;
 
 /*
@@ -146,8 +145,10 @@ static void unlock_srbm(struct kgd_dev *kgd)
 static void acquire_queue(struct kgd_dev *kgd, uint32_t pipe_id,
 				uint32_t queue_id)
 {
-	uint32_t mec = (++pipe_id / VI_PIPE_PER_MEC) + 1;
-	uint32_t pipe = (pipe_id % VI_PIPE_PER_MEC);
+	struct amdgpu_device *adev = get_amdgpu_device(kgd);
+
+	uint32_t mec = (++pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
+	uint32_t pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
 
 	lock_srbm(kgd, mec, pipe, queue_id, 0);
 }
@@ -205,6 +206,7 @@ static int kgd_set_pasid_vmid_mapping(struct kgd_dev *kgd, unsigned int pasid,
 static int kgd_init_pipeline(struct kgd_dev *kgd, uint32_t pipe_id,
 				uint32_t hpd_size, uint64_t hpd_gpu_addr)
 {
+	/* amdgpu owns the per-pipe state */
 	return 0;
 }
 
@@ -214,8 +216,8 @@ static int kgd_init_interrupts(struct kgd_dev *kgd, uint32_t pipe_id)
 	uint32_t mec;
 	uint32_t pipe;
 
-	mec = (++pipe_id / VI_PIPE_PER_MEC) + 1;
-	pipe = (pipe_id % VI_PIPE_PER_MEC);
+	mec = (++pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
+	pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
 
 	lock_srbm(kgd, mec, pipe, 0, 0);
 
@@ -251,53 +253,11 @@ static int kgd_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
 	m = get_mqd(mqd);
 
 	valid_wptr = copy_from_user(&shadow_wptr, wptr, sizeof(shadow_wptr));
-	acquire_queue(kgd, pipe_id, queue_id);
-
-	WREG32(mmCP_MQD_CONTROL, m->cp_mqd_control);
-	WREG32(mmCP_MQD_BASE_ADDR, m->cp_mqd_base_addr_lo);
-	WREG32(mmCP_MQD_BASE_ADDR_HI, m->cp_mqd_base_addr_hi);
-
-	WREG32(mmCP_HQD_VMID, m->cp_hqd_vmid);
-	WREG32(mmCP_HQD_PERSISTENT_STATE, m->cp_hqd_persistent_state);
-	WREG32(mmCP_HQD_PIPE_PRIORITY, m->cp_hqd_pipe_priority);
-	WREG32(mmCP_HQD_QUEUE_PRIORITY, m->cp_hqd_queue_priority);
-	WREG32(mmCP_HQD_QUANTUM, m->cp_hqd_quantum);
-	WREG32(mmCP_HQD_PQ_BASE, m->cp_hqd_pq_base_lo);
-	WREG32(mmCP_HQD_PQ_BASE_HI, m->cp_hqd_pq_base_hi);
-	WREG32(mmCP_HQD_PQ_RPTR_REPORT_ADDR, m->cp_hqd_pq_rptr_report_addr_lo);
-	WREG32(mmCP_HQD_PQ_RPTR_REPORT_ADDR_HI,
-			m->cp_hqd_pq_rptr_report_addr_hi);
-
-	if (valid_wptr > 0)
-		WREG32(mmCP_HQD_PQ_WPTR, shadow_wptr);
-
-	WREG32(mmCP_HQD_PQ_CONTROL, m->cp_hqd_pq_control);
-	WREG32(mmCP_HQD_PQ_DOORBELL_CONTROL, m->cp_hqd_pq_doorbell_control);
-
-	WREG32(mmCP_HQD_EOP_BASE_ADDR, m->cp_hqd_eop_base_addr_lo);
-	WREG32(mmCP_HQD_EOP_BASE_ADDR_HI, m->cp_hqd_eop_base_addr_hi);
-	WREG32(mmCP_HQD_EOP_CONTROL, m->cp_hqd_eop_control);
-	WREG32(mmCP_HQD_EOP_RPTR, m->cp_hqd_eop_rptr);
-	WREG32(mmCP_HQD_EOP_WPTR, m->cp_hqd_eop_wptr);
-	WREG32(mmCP_HQD_EOP_EVENTS, m->cp_hqd_eop_done_events);
-
-	WREG32(mmCP_HQD_CTX_SAVE_BASE_ADDR_LO, m->cp_hqd_ctx_save_base_addr_lo);
-	WREG32(mmCP_HQD_CTX_SAVE_BASE_ADDR_HI, m->cp_hqd_ctx_save_base_addr_hi);
-	WREG32(mmCP_HQD_CTX_SAVE_CONTROL, m->cp_hqd_ctx_save_control);
-	WREG32(mmCP_HQD_CNTL_STACK_OFFSET, m->cp_hqd_cntl_stack_offset);
-	WREG32(mmCP_HQD_CNTL_STACK_SIZE, m->cp_hqd_cntl_stack_size);
-	WREG32(mmCP_HQD_WG_STATE_OFFSET, m->cp_hqd_wg_state_offset);
-	WREG32(mmCP_HQD_CTX_SAVE_SIZE, m->cp_hqd_ctx_save_size);
-
-	WREG32(mmCP_HQD_IB_CONTROL, m->cp_hqd_ib_control);
-
-	WREG32(mmCP_HQD_DEQUEUE_REQUEST, m->cp_hqd_dequeue_request);
-	WREG32(mmCP_HQD_ERROR, m->cp_hqd_error);
-	WREG32(mmCP_HQD_EOP_WPTR_MEM, m->cp_hqd_eop_wptr_mem);
-	WREG32(mmCP_HQD_EOP_DONES, m->cp_hqd_eop_dones);
-
-	WREG32(mmCP_HQD_ACTIVE, m->cp_hqd_active);
+	if (valid_wptr == 0)
+		m->cp_hqd_pq_wptr = shadow_wptr;
 
+	acquire_queue(kgd, pipe_id, queue_id);
+	gfx_v8_0_mqd_commit(adev, mqd);
 	release_queue(kgd);
 
 	return 0;

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

@@ -30,78 +30,6 @@
 #include "amdgpu.h"
 #include "amdgpu_trace.h"
 
-int amdgpu_cs_get_ring(struct amdgpu_device *adev, u32 ip_type,
-		       u32 ip_instance, u32 ring,
-		       struct amdgpu_ring **out_ring)
-{
-	/* Right now all IPs have only one instance - multiple rings. */
-	if (ip_instance != 0) {
-		DRM_ERROR("invalid ip instance: %d\n", ip_instance);
-		return -EINVAL;
-	}
-
-	switch (ip_type) {
-	default:
-		DRM_ERROR("unknown ip type: %d\n", ip_type);
-		return -EINVAL;
-	case AMDGPU_HW_IP_GFX:
-		if (ring < adev->gfx.num_gfx_rings) {
-			*out_ring = &adev->gfx.gfx_ring[ring];
-		} else {
-			DRM_ERROR("only %d gfx rings are supported now\n",
-				  adev->gfx.num_gfx_rings);
-			return -EINVAL;
-		}
-		break;
-	case AMDGPU_HW_IP_COMPUTE:
-		if (ring < adev->gfx.num_compute_rings) {
-			*out_ring = &adev->gfx.compute_ring[ring];
-		} else {
-			DRM_ERROR("only %d compute rings are supported now\n",
-				  adev->gfx.num_compute_rings);
-			return -EINVAL;
-		}
-		break;
-	case AMDGPU_HW_IP_DMA:
-		if (ring < adev->sdma.num_instances) {
-			*out_ring = &adev->sdma.instance[ring].ring;
-		} else {
-			DRM_ERROR("only %d SDMA rings are supported\n",
-				  adev->sdma.num_instances);
-			return -EINVAL;
-		}
-		break;
-	case AMDGPU_HW_IP_UVD:
-		*out_ring = &adev->uvd.ring;
-		break;
-	case AMDGPU_HW_IP_VCE:
-		if (ring < adev->vce.num_rings){
-			*out_ring = &adev->vce.ring[ring];
-		} else {
-			DRM_ERROR("only %d VCE rings are supported\n", adev->vce.num_rings);
-			return -EINVAL;
-		}
-		break;
-	case AMDGPU_HW_IP_UVD_ENC:
-		if (ring < adev->uvd.num_enc_rings){
-			*out_ring = &adev->uvd.ring_enc[ring];
-		} else {
-			DRM_ERROR("only %d UVD ENC rings are supported\n",
-				adev->uvd.num_enc_rings);
-			return -EINVAL;
-		}
-		break;
-	}
-
-	if (!(*out_ring && (*out_ring)->adev)) {
-		DRM_ERROR("Ring %d is not initialized on IP %d\n",
-			  ring, ip_type);
-		return -EINVAL;
-	}
-
-	return 0;
-}
-
 static int amdgpu_cs_user_fence_chunk(struct amdgpu_cs_parser *p,
 				      struct drm_amdgpu_cs_chunk_fence *data,
 				      uint32_t *offset)
@@ -597,7 +525,7 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 			goto error_free_pages;
 		}
 
-		/* Fill the page arrays for all useptrs. */
+		/* Fill the page arrays for all userptrs. */
 		list_for_each_entry(e, &need_pages, tv.head) {
 			struct ttm_tt *ttm = e->robj->tbo.ttm;
 
@@ -917,9 +845,8 @@ static int amdgpu_cs_ib_fill(struct amdgpu_device *adev,
 				return -EINVAL;
 		}
 
-		r = amdgpu_cs_get_ring(adev, chunk_ib->ip_type,
-				       chunk_ib->ip_instance, chunk_ib->ring,
-				       &ring);
+		r = amdgpu_queue_mgr_map(adev, &parser->ctx->queue_mgr, chunk_ib->ip_type,
+					 chunk_ib->ip_instance, chunk_ib->ring, &ring);
 		if (r)
 			return r;
 
@@ -1021,16 +948,19 @@ static int amdgpu_cs_dependencies(struct amdgpu_device *adev,
 			struct amdgpu_ctx *ctx;
 			struct dma_fence *fence;
 
-			r = amdgpu_cs_get_ring(adev, deps[j].ip_type,
-					       deps[j].ip_instance,
-					       deps[j].ring, &ring);
-			if (r)
-				return r;
-
 			ctx = amdgpu_ctx_get(fpriv, deps[j].ctx_id);
 			if (ctx == NULL)
 				return -EINVAL;
 
+			r = amdgpu_queue_mgr_map(adev, &ctx->queue_mgr,
+						 deps[j].ip_type,
+						 deps[j].ip_instance,
+						 deps[j].ring, &ring);
+			if (r) {
+				amdgpu_ctx_put(ctx);
+				return r;
+			}
+
 			fence = amdgpu_ctx_get_fence(ctx, ring,
 						     deps[j].handle);
 			if (IS_ERR(fence)) {
@@ -1086,6 +1016,7 @@ 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;
@@ -1093,6 +1024,8 @@ 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;
@@ -1154,21 +1087,28 @@ 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;
 
-	r = amdgpu_cs_get_ring(adev, wait->in.ip_type, wait->in.ip_instance,
-			       wait->in.ring, &ring);
-	if (r)
-		return 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;
 
+	r = amdgpu_queue_mgr_map(adev, &ctx->queue_mgr,
+				 wait->in.ip_type, wait->in.ip_instance,
+				 wait->in.ring, &ring);
+	if (r) {
+		amdgpu_ctx_put(ctx);
+		return r;
+	}
+
 	fence = amdgpu_ctx_get_fence(ctx, ring, wait->in.handle);
 	if (IS_ERR(fence))
 		r = PTR_ERR(fence);
@@ -1204,15 +1144,17 @@ static struct dma_fence *amdgpu_cs_get_fence(struct amdgpu_device *adev,
 	struct dma_fence *fence;
 	int r;
 
-	r = amdgpu_cs_get_ring(adev, user->ip_type, user->ip_instance,
-			       user->ring, &ring);
-	if (r)
-		return ERR_PTR(r);
-
 	ctx = amdgpu_ctx_get(filp->driver_priv, user->ctx_id);
 	if (ctx == NULL)
 		return ERR_PTR(-EINVAL);
 
+	r = amdgpu_queue_mgr_map(adev, &ctx->queue_mgr, user->ip_type,
+				 user->ip_instance, user->ring, &ring);
+	if (r) {
+		amdgpu_ctx_put(ctx);
+		return ERR_PTR(r);
+	}
+
 	fence = amdgpu_ctx_get_fence(ctx, ring, user->seq_no);
 	amdgpu_ctx_put(ctx);
 
@@ -1333,12 +1275,15 @@ 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);

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

@@ -52,12 +52,20 @@ static int amdgpu_ctx_init(struct amdgpu_device *adev, struct amdgpu_ctx *ctx)
 		struct amd_sched_rq *rq;
 
 		rq = &ring->sched.sched_rq[AMD_SCHED_PRIORITY_NORMAL];
+
+		if (ring == &adev->gfx.kiq.ring)
+			continue;
+
 		r = amd_sched_entity_init(&ring->sched, &ctx->rings[i].entity,
 					  rq, amdgpu_sched_jobs);
 		if (r)
 			goto failed;
 	}
 
+	r = amdgpu_queue_mgr_init(adev, &ctx->queue_mgr);
+	if (r)
+		goto failed;
+
 	return 0;
 
 failed:
@@ -86,6 +94,8 @@ static void amdgpu_ctx_fini(struct amdgpu_ctx *ctx)
 	for (i = 0; i < adev->num_rings; i++)
 		amd_sched_entity_fini(&adev->rings[i]->sched,
 				      &ctx->rings[i].entity);
+
+	amdgpu_queue_mgr_fini(adev, &ctx->queue_mgr);
 }
 
 static int amdgpu_ctx_alloc(struct amdgpu_device *adev,

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

@@ -65,9 +65,11 @@
  * - 3.13.0 - Add PRT support
  * - 3.14.0 - Fix race in amdgpu_ctx_get_fence() and note new functionality
  * - 3.15.0 - Export more gpu info for gfx9
+ * - 3.16.0 - Add reserved vmid support
+ * - 3.17.0 - Add AMDGPU_NUM_VRAM_CPU_PAGE_FAULTS.
  */
 #define KMS_DRIVER_MAJOR	3
-#define KMS_DRIVER_MINOR	15
+#define KMS_DRIVER_MINOR	17
 #define KMS_DRIVER_PATCHLEVEL	0
 
 int amdgpu_vram_limit = 0;
@@ -92,7 +94,8 @@ int amdgpu_vm_size = -1;
 int amdgpu_vm_block_size = -1;
 int amdgpu_vm_fault_stop = 0;
 int amdgpu_vm_debug = 0;
-int amdgpu_vram_page_split = 1024;
+int amdgpu_vram_page_split = 512;
+int amdgpu_vm_update_mode = -1;
 int amdgpu_exp_hw_support = 0;
 int amdgpu_sched_jobs = 32;
 int amdgpu_sched_hw_submission = 2;
@@ -110,6 +113,8 @@ int amdgpu_prim_buf_per_se = 0;
 int amdgpu_pos_buf_per_se = 0;
 int amdgpu_cntl_sb_buf_per_se = 0;
 int amdgpu_param_buf_per_se = 0;
+int amdgpu_job_hang_limit = 0;
+int amdgpu_lbpw = -1;
 
 MODULE_PARM_DESC(vramlimit, "Restrict VRAM for testing, in megabytes");
 module_param_named(vramlimit, amdgpu_vram_limit, int, 0600);
@@ -177,6 +182,9 @@ module_param_named(vm_fault_stop, amdgpu_vm_fault_stop, int, 0444);
 MODULE_PARM_DESC(vm_debug, "Debug VM handling (0 = disabled (default), 1 = enabled)");
 module_param_named(vm_debug, amdgpu_vm_debug, int, 0644);
 
+MODULE_PARM_DESC(vm_update_mode, "VM update using CPU (0 = never (default except for large BAR(LB)), 1 = Graphics only, 2 = Compute only (default for LB), 3 = Both");
+module_param_named(vm_update_mode, amdgpu_vm_update_mode, int, 0444);
+
 MODULE_PARM_DESC(vram_page_split, "Number of pages after we split VRAM allocations (default 1024, -1 = disable)");
 module_param_named(vram_page_split, amdgpu_vram_page_split, int, 0444);
 
@@ -232,6 +240,24 @@ module_param_named(cntl_sb_buf_per_se, amdgpu_cntl_sb_buf_per_se, int, 0444);
 MODULE_PARM_DESC(param_buf_per_se, "the size of Off-Chip Pramater Cache per Shader Engine (default depending on gfx)");
 module_param_named(param_buf_per_se, amdgpu_param_buf_per_se, int, 0444);
 
+MODULE_PARM_DESC(job_hang_limit, "how much time allow a job hang and not drop it (default 0)");
+module_param_named(job_hang_limit, amdgpu_job_hang_limit, int ,0444);
+
+MODULE_PARM_DESC(lbpw, "Load Balancing Per Watt (LBPW) support (1 = enable, 0 = disable, -1 = auto)");
+module_param_named(lbpw, amdgpu_lbpw, int, 0444);
+
+#ifdef CONFIG_DRM_AMDGPU_SI
+int amdgpu_si_support = 0;
+MODULE_PARM_DESC(si_support, "SI support (1 = enabled, 0 = disabled (default))");
+module_param_named(si_support, amdgpu_si_support, int, 0444);
+#endif
+
+#ifdef CONFIG_DRM_AMDGPU_CIK
+int amdgpu_cik_support = 0;
+MODULE_PARM_DESC(cik_support, "CIK support (1 = enabled, 0 = disabled (default))");
+module_param_named(cik_support, amdgpu_cik_support, int, 0444);
+#endif
+
 
 static const struct pci_device_id pciidlist[] = {
 #ifdef  CONFIG_DRM_AMDGPU_SI
@@ -460,6 +486,9 @@ static const struct pci_device_id pciidlist[] = {
 	{0x1002, 0x6868, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VEGA10|AMD_EXP_HW_SUPPORT},
 	{0x1002, 0x686c, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VEGA10|AMD_EXP_HW_SUPPORT},
 	{0x1002, 0x687f, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VEGA10|AMD_EXP_HW_SUPPORT},
+	/* Raven */
+	{0x1002, 0x15dd, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_RAVEN|AMD_IS_APU|AMD_EXP_HW_SUPPORT},
+
 	{0, 0, 0}
 };
 
@@ -491,6 +520,7 @@ static int amdgpu_kick_out_firmware_fb(struct pci_dev *pdev)
 static int amdgpu_pci_probe(struct pci_dev *pdev,
 			    const struct pci_device_id *ent)
 {
+	struct drm_device *dev;
 	unsigned long flags = ent->driver_data;
 	int ret;
 
@@ -513,7 +543,29 @@ static int amdgpu_pci_probe(struct pci_dev *pdev,
 	if (ret)
 		return ret;
 
-	return drm_get_pci_dev(pdev, ent, &kms_driver);
+	dev = drm_dev_alloc(&kms_driver, &pdev->dev);
+	if (IS_ERR(dev))
+		return PTR_ERR(dev);
+
+	ret = pci_enable_device(pdev);
+	if (ret)
+		goto err_free;
+
+	dev->pdev = pdev;
+
+	pci_set_drvdata(pdev, dev);
+
+	ret = drm_dev_register(dev, ent->driver_data);
+	if (ret)
+		goto err_pci;
+
+	return 0;
+
+err_pci:
+	pci_disable_device(pdev);
+err_free:
+	drm_dev_unref(dev);
+	return ret;
 }
 
 static void
@@ -521,7 +573,8 @@ amdgpu_pci_remove(struct pci_dev *pdev)
 {
 	struct drm_device *dev = pci_get_drvdata(pdev);
 
-	drm_put_dev(dev);
+	drm_dev_unregister(dev);
+	drm_dev_unref(dev);
 }
 
 static void
@@ -817,7 +870,7 @@ static int __init amdgpu_init(void)
 	driver->num_ioctls = amdgpu_max_kms_ioctl;
 	amdgpu_register_atpx_handler();
 	/* let modprobe override vga console setting */
-	return drm_pci_init(driver, pdriver);
+	return pci_register_driver(pdriver);
 
 error_sched:
 	amdgpu_fence_slab_fini();
@@ -832,7 +885,7 @@ static int __init amdgpu_init(void)
 static void __exit amdgpu_exit(void)
 {
 	amdgpu_amdkfd_fini();
-	drm_pci_exit(driver, pdriver);
+	pci_unregister_driver(pdriver);
 	amdgpu_unregister_atpx_handler();
 	amdgpu_sync_fini();
 	amd_sched_fence_slab_fini();

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

@@ -541,6 +541,12 @@ void amdgpu_fence_driver_force_completion(struct amdgpu_device *adev)
 	}
 }
 
+void amdgpu_fence_driver_force_completion_ring(struct amdgpu_ring *ring)
+{
+	if (ring)
+		amdgpu_fence_write(ring, ring->fence_drv.sync_seq);
+}
+
 /*
  * Common fence implementation
  */
@@ -660,11 +666,17 @@ static const struct drm_info_list amdgpu_debugfs_fence_list[] = {
 	{"amdgpu_fence_info", &amdgpu_debugfs_fence_info, 0, NULL},
 	{"amdgpu_gpu_reset", &amdgpu_debugfs_gpu_reset, 0, NULL}
 };
+
+static const struct drm_info_list amdgpu_debugfs_fence_list_sriov[] = {
+	{"amdgpu_fence_info", &amdgpu_debugfs_fence_info, 0, NULL},
+};
 #endif
 
 int amdgpu_debugfs_fence_init(struct amdgpu_device *adev)
 {
 #if defined(CONFIG_DEBUG_FS)
+	if (amdgpu_sriov_vf(adev))
+		return amdgpu_debugfs_add_files(adev, amdgpu_debugfs_fence_list_sriov, 1);
 	return amdgpu_debugfs_add_files(adev, amdgpu_debugfs_fence_list, 2);
 #else
 	return 0;

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

@@ -224,8 +224,9 @@ void amdgpu_gart_table_vram_free(struct amdgpu_device *adev)
  *
  * Unbinds the requested pages from the gart page table and
  * replaces them with the dummy page (all asics).
+ * Returns 0 for success, -EINVAL for failure.
  */
-void amdgpu_gart_unbind(struct amdgpu_device *adev, uint64_t offset,
+int amdgpu_gart_unbind(struct amdgpu_device *adev, uint64_t offset,
 			int pages)
 {
 	unsigned t;
@@ -237,7 +238,7 @@ void amdgpu_gart_unbind(struct amdgpu_device *adev, uint64_t offset,
 
 	if (!adev->gart.ready) {
 		WARN(1, "trying to unbind memory from uninitialized GART !\n");
-		return;
+		return -EINVAL;
 	}
 
 	t = offset / AMDGPU_GPU_PAGE_SIZE;
@@ -258,6 +259,7 @@ void amdgpu_gart_unbind(struct amdgpu_device *adev, uint64_t offset,
 	}
 	mb();
 	amdgpu_gart_flush_gpu_tlb(adev, 0);
+	return 0;
 }
 
 /**

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

@@ -219,16 +219,6 @@ void amdgpu_gem_object_close(struct drm_gem_object *obj,
 	ttm_eu_backoff_reservation(&ticket, &list);
 }
 
-static int amdgpu_gem_handle_lockup(struct amdgpu_device *adev, int r)
-{
-	if (r == -EDEADLK) {
-		r = amdgpu_gpu_reset(adev);
-		if (!r)
-			r = -EAGAIN;
-	}
-	return r;
-}
-
 /*
  * GEM ioctls.
  */
@@ -249,20 +239,17 @@ int amdgpu_gem_create_ioctl(struct drm_device *dev, void *data,
 				      AMDGPU_GEM_CREATE_CPU_GTT_USWC |
 				      AMDGPU_GEM_CREATE_VRAM_CLEARED|
 				      AMDGPU_GEM_CREATE_SHADOW |
-				      AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS)) {
-		r = -EINVAL;
-		goto error_unlock;
-	}
+				      AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS))
+		return -EINVAL;
+
 	/* reject invalid gem domains */
 	if (args->in.domains & ~(AMDGPU_GEM_DOMAIN_CPU |
 				 AMDGPU_GEM_DOMAIN_GTT |
 				 AMDGPU_GEM_DOMAIN_VRAM |
 				 AMDGPU_GEM_DOMAIN_GDS |
 				 AMDGPU_GEM_DOMAIN_GWS |
-				 AMDGPU_GEM_DOMAIN_OA)) {
-		r = -EINVAL;
-		goto error_unlock;
-	}
+				 AMDGPU_GEM_DOMAIN_OA))
+		return -EINVAL;
 
 	/* create a gem object to contain this object in */
 	if (args->in.domains & (AMDGPU_GEM_DOMAIN_GDS |
@@ -274,10 +261,8 @@ int amdgpu_gem_create_ioctl(struct drm_device *dev, void *data,
 			size = size << AMDGPU_GWS_SHIFT;
 		else if (args->in.domains == AMDGPU_GEM_DOMAIN_OA)
 			size = size << AMDGPU_OA_SHIFT;
-		else {
-			r = -EINVAL;
-			goto error_unlock;
-		}
+		else
+			return -EINVAL;
 	}
 	size = roundup(size, PAGE_SIZE);
 
@@ -286,21 +271,17 @@ int amdgpu_gem_create_ioctl(struct drm_device *dev, void *data,
 				     args->in.domain_flags,
 				     kernel, &gobj);
 	if (r)
-		goto error_unlock;
+		return r;
 
 	r = drm_gem_handle_create(filp, gobj, &handle);
 	/* drop reference from allocate - handle holds it now */
 	drm_gem_object_unreference_unlocked(gobj);
 	if (r)
-		goto error_unlock;
+		return r;
 
 	memset(args, 0, sizeof(*args));
 	args->out.handle = handle;
 	return 0;
-
-error_unlock:
-	r = amdgpu_gem_handle_lockup(adev, r);
-	return r;
 }
 
 int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data,
@@ -334,7 +315,7 @@ int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data,
 				     AMDGPU_GEM_DOMAIN_CPU, 0,
 				     0, &gobj);
 	if (r)
-		goto handle_lockup;
+		return r;
 
 	bo = gem_to_amdgpu_bo(gobj);
 	bo->prefered_domains = AMDGPU_GEM_DOMAIN_GTT;
@@ -374,7 +355,7 @@ int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data,
 	/* drop reference from allocate - handle holds it now */
 	drm_gem_object_unreference_unlocked(gobj);
 	if (r)
-		goto handle_lockup;
+		return r;
 
 	args->handle = handle;
 	return 0;
@@ -388,9 +369,6 @@ int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data,
 release_object:
 	drm_gem_object_unreference_unlocked(gobj);
 
-handle_lockup:
-	r = amdgpu_gem_handle_lockup(adev, r);
-
 	return r;
 }
 
@@ -456,7 +434,6 @@ unsigned long amdgpu_gem_timeout(uint64_t timeout_ns)
 int amdgpu_gem_wait_idle_ioctl(struct drm_device *dev, void *data,
 			      struct drm_file *filp)
 {
-	struct amdgpu_device *adev = dev->dev_private;
 	union drm_amdgpu_gem_wait_idle *args = data;
 	struct drm_gem_object *gobj;
 	struct amdgpu_bo *robj;
@@ -484,7 +461,6 @@ int amdgpu_gem_wait_idle_ioctl(struct drm_device *dev, void *data,
 		r = ret;
 
 	drm_gem_object_unreference_unlocked(gobj);
-	r = amdgpu_gem_handle_lockup(adev, r);
 	return r;
 }
 
@@ -593,9 +569,6 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
 	uint64_t va_flags;
 	int r = 0;
 
-	if (!adev->vm_manager.enabled)
-		return -ENOTTY;
-
 	if (args->va_address < AMDGPU_VA_RESERVED_SIZE) {
 		dev_err(&dev->pdev->dev,
 			"va_address 0x%lX is in reserved area 0x%X\n",
@@ -621,6 +594,11 @@ 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);
 	if ((args->operation != AMDGPU_VA_OP_CLEAR) &&

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

@@ -108,3 +108,209 @@ void amdgpu_gfx_parse_disable_cu(unsigned *mask, unsigned max_se, unsigned max_s
 		p = next + 1;
 	}
 }
+
+void amdgpu_gfx_compute_queue_acquire(struct amdgpu_device *adev)
+{
+	int i, queue, pipe, mec;
+
+	/* policy for amdgpu compute queue ownership */
+	for (i = 0; i < AMDGPU_MAX_COMPUTE_QUEUES; ++i) {
+		queue = i % adev->gfx.mec.num_queue_per_pipe;
+		pipe = (i / adev->gfx.mec.num_queue_per_pipe)
+			% adev->gfx.mec.num_pipe_per_mec;
+		mec = (i / adev->gfx.mec.num_queue_per_pipe)
+			/ adev->gfx.mec.num_pipe_per_mec;
+
+		/* we've run out of HW */
+		if (mec >= adev->gfx.mec.num_mec)
+			break;
+
+		if (adev->gfx.mec.num_mec > 1) {
+			/* policy: amdgpu owns the first two queues of the first MEC */
+			if (mec == 0 && queue < 2)
+				set_bit(i, adev->gfx.mec.queue_bitmap);
+		} else {
+			/* policy: amdgpu owns all queues in the first pipe */
+			if (mec == 0 && pipe == 0)
+				set_bit(i, adev->gfx.mec.queue_bitmap);
+		}
+	}
+
+	/* update the number of active compute rings */
+	adev->gfx.num_compute_rings =
+		bitmap_weight(adev->gfx.mec.queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES);
+
+	/* If you hit this case and edited the policy, you probably just
+	 * need to increase AMDGPU_MAX_COMPUTE_RINGS */
+	if (WARN_ON(adev->gfx.num_compute_rings > AMDGPU_MAX_COMPUTE_RINGS))
+		adev->gfx.num_compute_rings = AMDGPU_MAX_COMPUTE_RINGS;
+}
+
+static int amdgpu_gfx_kiq_acquire(struct amdgpu_device *adev,
+				  struct amdgpu_ring *ring)
+{
+	int queue_bit;
+	int mec, pipe, queue;
+
+	queue_bit = adev->gfx.mec.num_mec
+		    * adev->gfx.mec.num_pipe_per_mec
+		    * adev->gfx.mec.num_queue_per_pipe;
+
+	while (queue_bit-- >= 0) {
+		if (test_bit(queue_bit, adev->gfx.mec.queue_bitmap))
+			continue;
+
+		amdgpu_gfx_bit_to_queue(adev, queue_bit, &mec, &pipe, &queue);
+
+		/* Using pipes 2/3 from MEC 2 seems cause problems */
+		if (mec == 1 && pipe > 1)
+			continue;
+
+		ring->me = mec + 1;
+		ring->pipe = pipe;
+		ring->queue = queue;
+
+		return 0;
+	}
+
+	dev_err(adev->dev, "Failed to find a queue for KIQ\n");
+	return -EINVAL;
+}
+
+int amdgpu_gfx_kiq_init_ring(struct amdgpu_device *adev,
+			     struct amdgpu_ring *ring,
+			     struct amdgpu_irq_src *irq)
+{
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq;
+	int r = 0;
+
+	mutex_init(&kiq->ring_mutex);
+
+	r = amdgpu_wb_get(adev, &adev->virt.reg_val_offs);
+	if (r)
+		return r;
+
+	ring->adev = NULL;
+	ring->ring_obj = NULL;
+	ring->use_doorbell = true;
+	ring->doorbell_index = AMDGPU_DOORBELL_KIQ;
+
+	r = amdgpu_gfx_kiq_acquire(adev, ring);
+	if (r)
+		return r;
+
+	ring->eop_gpu_addr = kiq->eop_gpu_addr;
+	sprintf(ring->name, "kiq_%d.%d.%d", ring->me, ring->pipe, ring->queue);
+	r = amdgpu_ring_init(adev, ring, 1024,
+			     irq, AMDGPU_CP_KIQ_IRQ_DRIVER0);
+	if (r)
+		dev_warn(adev->dev, "(%d) failed to init kiq ring\n", r);
+
+	return r;
+}
+
+void amdgpu_gfx_kiq_free_ring(struct amdgpu_ring *ring,
+			      struct amdgpu_irq_src *irq)
+{
+	amdgpu_wb_free(ring->adev, ring->adev->virt.reg_val_offs);
+	amdgpu_ring_fini(ring);
+}
+
+void amdgpu_gfx_kiq_fini(struct amdgpu_device *adev)
+{
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq;
+
+	amdgpu_bo_free_kernel(&kiq->eop_obj, &kiq->eop_gpu_addr, NULL);
+}
+
+int amdgpu_gfx_kiq_init(struct amdgpu_device *adev,
+			unsigned hpd_size)
+{
+	int r;
+	u32 *hpd;
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq;
+
+	r = amdgpu_bo_create_kernel(adev, hpd_size, PAGE_SIZE,
+				    AMDGPU_GEM_DOMAIN_GTT, &kiq->eop_obj,
+				    &kiq->eop_gpu_addr, (void **)&hpd);
+	if (r) {
+		dev_warn(adev->dev, "failed to create KIQ bo (%d).\n", r);
+		return r;
+	}
+
+	memset(hpd, 0, hpd_size);
+
+	r = amdgpu_bo_reserve(kiq->eop_obj, true);
+	if (unlikely(r != 0))
+		dev_warn(adev->dev, "(%d) reserve kiq eop bo failed\n", r);
+	amdgpu_bo_kunmap(kiq->eop_obj);
+	amdgpu_bo_unreserve(kiq->eop_obj);
+
+	return 0;
+}
+
+/* create MQD for each compute queue */
+int amdgpu_gfx_compute_mqd_sw_init(struct amdgpu_device *adev,
+				   unsigned mqd_size)
+{
+	struct amdgpu_ring *ring = NULL;
+	int r, i;
+
+	/* create MQD for KIQ */
+	ring = &adev->gfx.kiq.ring;
+	if (!ring->mqd_obj) {
+		r = amdgpu_bo_create_kernel(adev, mqd_size, PAGE_SIZE,
+					    AMDGPU_GEM_DOMAIN_GTT, &ring->mqd_obj,
+					    &ring->mqd_gpu_addr, &ring->mqd_ptr);
+		if (r) {
+			dev_warn(adev->dev, "failed to create ring mqd ob (%d)", r);
+			return r;
+		}
+
+		/* prepare MQD backup */
+		adev->gfx.mec.mqd_backup[AMDGPU_MAX_COMPUTE_RINGS] = kmalloc(mqd_size, GFP_KERNEL);
+		if (!adev->gfx.mec.mqd_backup[AMDGPU_MAX_COMPUTE_RINGS])
+				dev_warn(adev->dev, "no memory to create MQD backup for ring %s\n", ring->name);
+	}
+
+	/* create MQD for each KCQ */
+	for (i = 0; i < adev->gfx.num_compute_rings; i++) {
+		ring = &adev->gfx.compute_ring[i];
+		if (!ring->mqd_obj) {
+			r = amdgpu_bo_create_kernel(adev, mqd_size, PAGE_SIZE,
+						    AMDGPU_GEM_DOMAIN_GTT, &ring->mqd_obj,
+						    &ring->mqd_gpu_addr, &ring->mqd_ptr);
+			if (r) {
+				dev_warn(adev->dev, "failed to create ring mqd ob (%d)", r);
+				return r;
+			}
+
+			/* prepare MQD backup */
+			adev->gfx.mec.mqd_backup[i] = kmalloc(mqd_size, GFP_KERNEL);
+			if (!adev->gfx.mec.mqd_backup[i])
+				dev_warn(adev->dev, "no memory to create MQD backup for ring %s\n", ring->name);
+		}
+	}
+
+	return 0;
+}
+
+void amdgpu_gfx_compute_mqd_sw_fini(struct amdgpu_device *adev)
+{
+	struct amdgpu_ring *ring = NULL;
+	int i;
+
+	for (i = 0; i < adev->gfx.num_compute_rings; i++) {
+		ring = &adev->gfx.compute_ring[i];
+		kfree(adev->gfx.mec.mqd_backup[i]);
+		amdgpu_bo_free_kernel(&ring->mqd_obj,
+				      &ring->mqd_gpu_addr,
+				      &ring->mqd_ptr);
+	}
+
+	ring = &adev->gfx.kiq.ring;
+	kfree(adev->gfx.mec.mqd_backup[AMDGPU_MAX_COMPUTE_RINGS]);
+	amdgpu_bo_free_kernel(&ring->mqd_obj,
+			      &ring->mqd_gpu_addr,
+			      &ring->mqd_ptr);
+}

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

@@ -30,4 +30,64 @@ void amdgpu_gfx_scratch_free(struct amdgpu_device *adev, uint32_t reg);
 void amdgpu_gfx_parse_disable_cu(unsigned *mask, unsigned max_se,
 		unsigned max_sh);
 
+void amdgpu_gfx_compute_queue_acquire(struct amdgpu_device *adev);
+
+int amdgpu_gfx_kiq_init_ring(struct amdgpu_device *adev,
+			     struct amdgpu_ring *ring,
+			     struct amdgpu_irq_src *irq);
+
+void amdgpu_gfx_kiq_free_ring(struct amdgpu_ring *ring,
+			      struct amdgpu_irq_src *irq);
+
+void amdgpu_gfx_kiq_fini(struct amdgpu_device *adev);
+int amdgpu_gfx_kiq_init(struct amdgpu_device *adev,
+			unsigned hpd_size);
+
+int amdgpu_gfx_compute_mqd_sw_init(struct amdgpu_device *adev,
+				   unsigned mqd_size);
+void amdgpu_gfx_compute_mqd_sw_fini(struct amdgpu_device *adev);
+
+/**
+ * amdgpu_gfx_create_bitmask - create a bitmask
+ *
+ * @bit_width: length of the mask
+ *
+ * create a variable length bit mask.
+ * Returns the bitmask.
+ */
+static inline u32 amdgpu_gfx_create_bitmask(u32 bit_width)
+{
+	return (u32)((1ULL << bit_width) - 1);
+}
+
+static inline int amdgpu_gfx_queue_to_bit(struct amdgpu_device *adev,
+					  int mec, int pipe, int queue)
+{
+	int bit = 0;
+
+	bit += mec * adev->gfx.mec.num_pipe_per_mec
+		* adev->gfx.mec.num_queue_per_pipe;
+	bit += pipe * adev->gfx.mec.num_queue_per_pipe;
+	bit += queue;
+
+	return bit;
+}
+
+static inline void amdgpu_gfx_bit_to_queue(struct amdgpu_device *adev, int bit,
+					   int *mec, int *pipe, int *queue)
+{
+	*queue = bit % adev->gfx.mec.num_queue_per_pipe;
+	*pipe = (bit / adev->gfx.mec.num_queue_per_pipe)
+		% adev->gfx.mec.num_pipe_per_mec;
+	*mec = (bit / adev->gfx.mec.num_queue_per_pipe)
+	       / adev->gfx.mec.num_pipe_per_mec;
+
+}
+static inline bool amdgpu_gfx_is_mec_queue_enabled(struct amdgpu_device *adev,
+						   int mec, int pipe, int queue)
+{
+	return test_bit(amdgpu_gfx_queue_to_bit(adev, mec, pipe, queue),
+			adev->gfx.mec.queue_bitmap);
+}
+
 #endif

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

@@ -121,6 +121,7 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 {
 	struct amdgpu_device *adev = ring->adev;
 	struct amdgpu_ib *ib = &ibs[0];
+	struct dma_fence *tmp = NULL;
 	bool skip_preamble, need_ctx_switch;
 	unsigned patch_offset = ~0;
 	struct amdgpu_vm *vm;
@@ -160,8 +161,16 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 		dev_err(adev->dev, "scheduling IB failed (%d).\n", r);
 		return r;
 	}
-	if (ring->funcs->emit_pipeline_sync && job && job->need_pipeline_sync)
+
+	if (ring->funcs->emit_pipeline_sync && job &&
+	    ((tmp = amdgpu_sync_get_fence(&job->sched_sync)) ||
+	     amdgpu_vm_need_pipeline_sync(ring, job))) {
 		amdgpu_ring_emit_pipeline_sync(ring);
+		dma_fence_put(tmp);
+	}
+
+	if (ring->funcs->insert_start)
+		ring->funcs->insert_start(ring);
 
 	if (vm) {
 		r = amdgpu_vm_flush(ring, job);
@@ -188,8 +197,6 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 			status |= AMDGPU_HAVE_CTX_SWITCH;
 		status |= job->preamble_status;
 
-		if (vm)
-			status |= AMDGPU_VM_DOMAIN;
 		amdgpu_ring_emit_cntxcntl(ring, status);
 	}
 
@@ -208,6 +215,9 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 		need_ctx_switch = false;
 	}
 
+	if (ring->funcs->emit_tmz)
+		amdgpu_ring_emit_tmz(ring, false);
+
 	if (ring->funcs->emit_hdp_invalidate
 #ifdef CONFIG_X86_64
 	    && !(adev->flags & AMD_IS_APU)

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

@@ -62,8 +62,9 @@ enum amdgpu_ih_clientid
     AMDGPU_IH_CLIENTID_MP0	    = 0x1e,
     AMDGPU_IH_CLIENTID_MP1	    = 0x1f,
 
-    AMDGPU_IH_CLIENTID_MAX
+    AMDGPU_IH_CLIENTID_MAX,
 
+    AMDGPU_IH_CLIENTID_VCN	    = AMDGPU_IH_CLIENTID_UVD
 };
 
 #define AMDGPU_IH_CLIENTID_LEGACY 0

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

@@ -83,7 +83,8 @@ static void amdgpu_irq_reset_work_func(struct work_struct *work)
 	struct amdgpu_device *adev = container_of(work, struct amdgpu_device,
 						  reset_work);
 
-	amdgpu_gpu_reset(adev);
+	if (!amdgpu_sriov_vf(adev))
+		amdgpu_gpu_reset(adev);
 }
 
 /* Disable *all* interrupts */

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

@@ -36,7 +36,11 @@ static void amdgpu_job_timedout(struct amd_sched_job *s_job)
 		  job->base.sched->name,
 		  atomic_read(&job->ring->fence_drv.last_seq),
 		  job->ring->fence_drv.sync_seq);
-	amdgpu_gpu_reset(job->adev);
+
+	if (amdgpu_sriov_vf(job->adev))
+		amdgpu_sriov_gpu_reset(job->adev, job);
+	else
+		amdgpu_gpu_reset(job->adev);
 }
 
 int amdgpu_job_alloc(struct amdgpu_device *adev, unsigned num_ibs,
@@ -57,9 +61,10 @@ 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)->need_pipeline_sync = false;
 
 	amdgpu_sync_create(&(*job)->sync);
+	amdgpu_sync_create(&(*job)->dep_sync);
+	amdgpu_sync_create(&(*job)->sched_sync);
 
 	return 0;
 }
@@ -98,6 +103,8 @@ static void amdgpu_job_free_cb(struct amd_sched_job *s_job)
 
 	dma_fence_put(job->fence);
 	amdgpu_sync_free(&job->sync);
+	amdgpu_sync_free(&job->dep_sync);
+	amdgpu_sync_free(&job->sched_sync);
 	kfree(job);
 }
 
@@ -107,6 +114,8 @@ void amdgpu_job_free(struct amdgpu_job *job)
 
 	dma_fence_put(job->fence);
 	amdgpu_sync_free(&job->sync);
+	amdgpu_sync_free(&job->dep_sync);
+	amdgpu_sync_free(&job->sched_sync);
 	kfree(job);
 }
 
@@ -138,11 +147,18 @@ static struct dma_fence *amdgpu_job_dependency(struct amd_sched_job *sched_job)
 	struct amdgpu_job *job = to_amdgpu_job(sched_job);
 	struct amdgpu_vm *vm = job->vm;
 
-	struct dma_fence *fence = amdgpu_sync_get_fence(&job->sync);
+	struct dma_fence *fence = amdgpu_sync_get_fence(&job->dep_sync);
+	int r;
 
+	if (amd_sched_dependency_optimized(fence, sched_job->s_entity)) {
+		r = amdgpu_sync_fence(job->adev, &job->sched_sync, fence);
+		if (r)
+			DRM_ERROR("Error adding fence to sync (%d)\n", r);
+	}
+	if (!fence)
+		fence = amdgpu_sync_get_fence(&job->sync);
 	while (fence == NULL && vm && !job->vm_id) {
 		struct amdgpu_ring *ring = job->ring;
-		int r;
 
 		r = amdgpu_vm_grab_id(vm, ring, &job->sync,
 				      &job->base.s_fence->finished,
@@ -153,9 +169,6 @@ static struct dma_fence *amdgpu_job_dependency(struct amd_sched_job *sched_job)
 		fence = amdgpu_sync_get_fence(&job->sync);
 	}
 
-	if (amd_sched_dependency_optimized(fence, sched_job->s_entity))
-		job->need_pipeline_sync = true;
-
 	return fence;
 }
 
@@ -163,6 +176,7 @@ static struct dma_fence *amdgpu_job_run(struct amd_sched_job *sched_job)
 {
 	struct dma_fence *fence = NULL;
 	struct amdgpu_job *job;
+	struct amdgpu_fpriv *fpriv = NULL;
 	int r;
 
 	if (!sched_job) {
@@ -174,10 +188,16 @@ static struct dma_fence *amdgpu_job_run(struct amd_sched_job *sched_job)
 	BUG_ON(amdgpu_sync_peek_fence(&job->sync, NULL));
 
 	trace_amdgpu_sched_run_job(job);
-	r = amdgpu_ib_schedule(job->ring, job->num_ibs, job->ibs, job, &fence);
-	if (r)
-		DRM_ERROR("Error scheduling IBs (%d)\n", r);
-
+	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))
+		DRM_ERROR("Skip scheduling IBs!\n");
+	else {
+		r = amdgpu_ib_schedule(job->ring, job->num_ibs, job->ibs, job, &fence);
+		if (r)
+			DRM_ERROR("Error scheduling IBs (%d)\n", r);
+	}
 	/* if gpu reset, hw fence will be replaced here */
 	dma_fence_put(job->fence);
 	job->fence = dma_fence_get(fence);

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

@@ -87,6 +87,41 @@ int amdgpu_driver_load_kms(struct drm_device *dev, unsigned long flags)
 	struct amdgpu_device *adev;
 	int r, acpi_status;
 
+#ifdef CONFIG_DRM_AMDGPU_SI
+	if (!amdgpu_si_support) {
+		switch (flags & AMD_ASIC_MASK) {
+		case CHIP_TAHITI:
+		case CHIP_PITCAIRN:
+		case CHIP_VERDE:
+		case CHIP_OLAND:
+		case CHIP_HAINAN:
+			dev_info(dev->dev,
+				 "SI support provided by radeon.\n");
+			dev_info(dev->dev,
+				 "Use radeon.si_support=0 amdgpu.si_support=1 to override.\n"
+				);
+			return -ENODEV;
+		}
+	}
+#endif
+#ifdef CONFIG_DRM_AMDGPU_CIK
+	if (!amdgpu_cik_support) {
+		switch (flags & AMD_ASIC_MASK) {
+		case CHIP_KAVERI:
+		case CHIP_BONAIRE:
+		case CHIP_HAWAII:
+		case CHIP_KABINI:
+		case CHIP_MULLINS:
+			dev_info(dev->dev,
+				 "CIK support provided by radeon.\n");
+			dev_info(dev->dev,
+				 "Use radeon.cik_support=0 amdgpu.cik_support=1 to override.\n"
+				);
+			return -ENODEV;
+		}
+	}
+#endif
+
 	adev = kzalloc(sizeof(struct amdgpu_device), GFP_KERNEL);
 	if (adev == NULL) {
 		return -ENOMEM;
@@ -235,6 +270,7 @@ 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;
@@ -247,6 +283,8 @@ 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:
@@ -319,6 +357,19 @@ static int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file
 			ib_start_alignment = AMDGPU_GPU_PAGE_SIZE;
 			ib_size_alignment = 1;
 			break;
+		case AMDGPU_HW_IP_VCN_DEC:
+			type = AMD_IP_BLOCK_TYPE_VCN;
+			ring_mask = adev->vcn.ring_dec.ready ? 1 : 0;
+			ib_start_alignment = AMDGPU_GPU_PAGE_SIZE;
+			ib_size_alignment = 16;
+			break;
+		case AMDGPU_HW_IP_VCN_ENC:
+			type = AMD_IP_BLOCK_TYPE_VCN;
+			for (i = 0; i < adev->vcn.num_enc_rings; i++)
+				ring_mask |= ((adev->vcn.ring_enc[i].ready ? 1 : 0) << i);
+			ib_start_alignment = AMDGPU_GPU_PAGE_SIZE;
+			ib_size_alignment = 1;
+			break;
 		default:
 			return -EINVAL;
 		}
@@ -361,6 +412,10 @@ static int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file
 		case AMDGPU_HW_IP_UVD_ENC:
 			type = AMD_IP_BLOCK_TYPE_UVD;
 			break;
+		case AMDGPU_HW_IP_VCN_DEC:
+		case AMDGPU_HW_IP_VCN_ENC:
+			type = AMD_IP_BLOCK_TYPE_VCN;
+			break;
 		default:
 			return -EINVAL;
 		}
@@ -397,6 +452,9 @@ static int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file
 	case AMDGPU_INFO_NUM_EVICTIONS:
 		ui64 = atomic64_read(&adev->num_evictions);
 		return copy_to_user(out, &ui64, min(size, 8u)) ? -EFAULT : 0;
+	case AMDGPU_INFO_NUM_VRAM_CPU_PAGE_FAULTS:
+		ui64 = atomic64_read(&adev->num_vram_cpu_page_faults);
+		return copy_to_user(out, &ui64, min(size, 8u)) ? -EFAULT : 0;
 	case AMDGPU_INFO_VRAM_USAGE:
 		ui64 = atomic64_read(&adev->vram_usage);
 		return copy_to_user(out, &ui64, min(size, 8u)) ? -EFAULT : 0;
@@ -730,6 +788,12 @@ void amdgpu_driver_lastclose_kms(struct drm_device *dev)
 	vga_switcheroo_process_delayed_switch();
 }
 
+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
  *
@@ -757,7 +821,8 @@ int amdgpu_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
 		goto out_suspend;
 	}
 
-	r = amdgpu_vm_init(adev, &fpriv->vm);
+	r = amdgpu_vm_init(adev, &fpriv->vm,
+			   AMDGPU_VM_CONTEXT_GFX);
 	if (r) {
 		kfree(fpriv);
 		goto out_suspend;
@@ -782,6 +847,7 @@ 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:
@@ -814,8 +880,10 @@ void amdgpu_driver_postclose_kms(struct drm_device *dev,
 
 	amdgpu_ctx_mgr_fini(&fpriv->ctx_mgr);
 
-	amdgpu_uvd_free_handles(adev, file_priv);
-	amdgpu_vce_free_handles(adev, file_priv);
+	if (adev->asic_type != CHIP_RAVEN) {
+		amdgpu_uvd_free_handles(adev, file_priv);
+		amdgpu_vce_free_handles(adev, file_priv);
+	}
 
 	amdgpu_vm_bo_rmv(adev, fpriv->prt_va);
 
@@ -948,6 +1016,7 @@ void amdgpu_disable_vblank_kms(struct drm_device *dev, unsigned int pipe)
 const struct drm_ioctl_desc amdgpu_ioctls_kms[] = {
 	DRM_IOCTL_DEF_DRV(AMDGPU_GEM_CREATE, amdgpu_gem_create_ioctl, DRM_AUTH|DRM_RENDER_ALLOW),
 	DRM_IOCTL_DEF_DRV(AMDGPU_CTX, amdgpu_ctx_ioctl, DRM_AUTH|DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(AMDGPU_VM, amdgpu_vm_ioctl, DRM_AUTH|DRM_RENDER_ALLOW),
 	DRM_IOCTL_DEF_DRV(AMDGPU_BO_LIST, amdgpu_bo_list_ioctl, DRM_AUTH|DRM_RENDER_ALLOW),
 	/* KMS */
 	DRM_IOCTL_DEF_DRV(AMDGPU_GEM_MMAP, amdgpu_gem_mmap_ioctl, DRM_AUTH|DRM_RENDER_ALLOW),

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

@@ -960,6 +960,7 @@ int amdgpu_bo_fault_reserve_notify(struct ttm_buffer_object *bo)
 		return -EINVAL;
 
 	/* hurrah the memory is not visible ! */
+	atomic64_inc(&adev->num_vram_cpu_page_faults);
 	amdgpu_ttm_placement_from_domain(abo, AMDGPU_GEM_DOMAIN_VRAM);
 	lpfn =	adev->mc.visible_vram_size >> PAGE_SHIFT;
 	for (i = 0; i < abo->placement.num_placement; i++) {

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

@@ -72,6 +72,7 @@ static int amdgpu_pp_early_init(void *handle)
 	case CHIP_CARRIZO:
 	case CHIP_STONEY:
 	case CHIP_VEGA10:
+	case CHIP_RAVEN:
 		adev->pp_enabled = true;
 		if (amdgpu_create_pp_handle(adev))
 			return -EINVAL;

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

@@ -30,6 +30,7 @@
 #include "amdgpu_ucode.h"
 #include "soc15_common.h"
 #include "psp_v3_1.h"
+#include "psp_v10_0.h"
 
 static void psp_set_funcs(struct amdgpu_device *adev);
 
@@ -61,6 +62,12 @@ static int psp_sw_init(void *handle)
 		psp->compare_sram_data = psp_v3_1_compare_sram_data;
 		psp->smu_reload_quirk = psp_v3_1_smu_reload_quirk;
 		break;
+	case CHIP_RAVEN:
+		psp->prep_cmd_buf = psp_v10_0_prep_cmd_buf;
+		psp->ring_init = psp_v10_0_ring_init;
+		psp->cmd_submit = psp_v10_0_cmd_submit;
+		psp->compare_sram_data = psp_v10_0_compare_sram_data;
+		break;
 	default:
 		return -EINVAL;
 	}
@@ -230,6 +237,13 @@ static int psp_asd_load(struct psp_context *psp)
 	int ret;
 	struct psp_gfx_cmd_resp *cmd;
 
+	/* If PSP version doesn't match ASD version, asd loading will be failed.
+	 * add workaround to bypass it for sriov now.
+	 * TODO: add version check to make it common
+	 */
+	if (amdgpu_sriov_vf(psp->adev))
+		return 0;
+
 	cmd = kzalloc(sizeof(struct psp_gfx_cmd_resp), GFP_KERNEL);
 	if (!cmd)
 		return -ENOMEM;
@@ -542,3 +556,12 @@ const struct amdgpu_ip_block_version psp_v3_1_ip_block =
 	.rev = 0,
 	.funcs = &psp_ip_funcs,
 };
+
+const struct amdgpu_ip_block_version psp_v10_0_ip_block =
+{
+	.type = AMD_IP_BLOCK_TYPE_PSP,
+	.major = 10,
+	.minor = 0,
+	.rev = 0,
+	.funcs = &psp_ip_funcs,
+};

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

@@ -138,4 +138,6 @@ extern const struct amdgpu_ip_block_version psp_v3_1_ip_block;
 extern int psp_wait_for(struct psp_context *psp, uint32_t reg_index,
 			uint32_t field_val, uint32_t mask, bool check_changed);
 
+extern const struct amdgpu_ip_block_version psp_v10_0_ip_block;
+
 #endif

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

+/*
+ * Copyright 2017 Valve Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Andres Rodriguez
+ */
+
+#include "amdgpu.h"
+#include "amdgpu_ring.h"
+
+static int amdgpu_queue_mapper_init(struct amdgpu_queue_mapper *mapper,
+				    int hw_ip)
+{
+	if (!mapper)
+		return -EINVAL;
+
+	if (hw_ip > AMDGPU_MAX_IP_NUM)
+		return -EINVAL;
+
+	mapper->hw_ip = hw_ip;
+	mutex_init(&mapper->lock);
+
+	memset(mapper->queue_map, 0, sizeof(mapper->queue_map));
+
+	return 0;
+}
+
+static struct amdgpu_ring *amdgpu_get_cached_map(struct amdgpu_queue_mapper *mapper,
+					  int ring)
+{
+	return mapper->queue_map[ring];
+}
+
+static int amdgpu_update_cached_map(struct amdgpu_queue_mapper *mapper,
+			     int ring, struct amdgpu_ring *pring)
+{
+	if (WARN_ON(mapper->queue_map[ring])) {
+		DRM_ERROR("Un-expected ring re-map\n");
+		return -EINVAL;
+	}
+
+	mapper->queue_map[ring] = pring;
+
+	return 0;
+}
+
+static int amdgpu_identity_map(struct amdgpu_device *adev,
+			       struct amdgpu_queue_mapper *mapper,
+			       int ring,
+			       struct amdgpu_ring **out_ring)
+{
+	switch (mapper->hw_ip) {
+	case AMDGPU_HW_IP_GFX:
+		*out_ring = &adev->gfx.gfx_ring[ring];
+		break;
+	case AMDGPU_HW_IP_COMPUTE:
+		*out_ring = &adev->gfx.compute_ring[ring];
+		break;
+	case AMDGPU_HW_IP_DMA:
+		*out_ring = &adev->sdma.instance[ring].ring;
+		break;
+	case AMDGPU_HW_IP_UVD:
+		*out_ring = &adev->uvd.ring;
+		break;
+	case AMDGPU_HW_IP_VCE:
+		*out_ring = &adev->vce.ring[ring];
+		break;
+	case AMDGPU_HW_IP_UVD_ENC:
+		*out_ring = &adev->uvd.ring_enc[ring];
+		break;
+	case AMDGPU_HW_IP_VCN_DEC:
+		*out_ring = &adev->vcn.ring_dec;
+		break;
+	case AMDGPU_HW_IP_VCN_ENC:
+		*out_ring = &adev->vcn.ring_enc[ring];
+		break;
+	default:
+		*out_ring = NULL;
+		DRM_ERROR("unknown HW IP type: %d\n", mapper->hw_ip);
+		return -EINVAL;
+	}
+
+	return amdgpu_update_cached_map(mapper, ring, *out_ring);
+}
+
+static enum amdgpu_ring_type amdgpu_hw_ip_to_ring_type(int hw_ip)
+{
+	switch (hw_ip) {
+	case AMDGPU_HW_IP_GFX:
+		return AMDGPU_RING_TYPE_GFX;
+	case AMDGPU_HW_IP_COMPUTE:
+		return AMDGPU_RING_TYPE_COMPUTE;
+	case AMDGPU_HW_IP_DMA:
+		return AMDGPU_RING_TYPE_SDMA;
+	case AMDGPU_HW_IP_UVD:
+		return AMDGPU_RING_TYPE_UVD;
+	case AMDGPU_HW_IP_VCE:
+		return AMDGPU_RING_TYPE_VCE;
+	default:
+		DRM_ERROR("Invalid HW IP specified %d\n", hw_ip);
+		return -1;
+	}
+}
+
+static int amdgpu_lru_map(struct amdgpu_device *adev,
+			  struct amdgpu_queue_mapper *mapper,
+			  int user_ring,
+			  struct amdgpu_ring **out_ring)
+{
+	int r, i, j;
+	int ring_type = amdgpu_hw_ip_to_ring_type(mapper->hw_ip);
+	int ring_blacklist[AMDGPU_MAX_RINGS];
+	struct amdgpu_ring *ring;
+
+	/* 0 is a valid ring index, so initialize to -1 */
+	memset(ring_blacklist, 0xff, sizeof(ring_blacklist));
+
+	for (i = 0, j = 0; i < AMDGPU_MAX_RINGS; i++) {
+		ring = mapper->queue_map[i];
+		if (ring)
+			ring_blacklist[j++] = ring->idx;
+	}
+
+	r = amdgpu_ring_lru_get(adev, ring_type, ring_blacklist,
+				j, out_ring);
+	if (r)
+		return r;
+
+	return amdgpu_update_cached_map(mapper, user_ring, *out_ring);
+}
+
+/**
+ * amdgpu_queue_mgr_init - init an amdgpu_queue_mgr struct
+ *
+ * @adev: amdgpu_device pointer
+ * @mgr: amdgpu_queue_mgr structure holding queue information
+ *
+ * Initialize the the selected @mgr (all asics).
+ *
+ * Returns 0 on success, error on failure.
+ */
+int amdgpu_queue_mgr_init(struct amdgpu_device *adev,
+			  struct amdgpu_queue_mgr *mgr)
+{
+	int i, r;
+
+	if (!adev || !mgr)
+		return -EINVAL;
+
+	memset(mgr, 0, sizeof(*mgr));
+
+	for (i = 0; i < AMDGPU_MAX_IP_NUM; ++i) {
+		r = amdgpu_queue_mapper_init(&mgr->mapper[i], i);
+		if (r)
+			return r;
+	}
+
+	return 0;
+}
+
+/**
+ * amdgpu_queue_mgr_fini - de-initialize an amdgpu_queue_mgr struct
+ *
+ * @adev: amdgpu_device pointer
+ * @mgr: amdgpu_queue_mgr structure holding queue information
+ *
+ * De-initialize the the selected @mgr (all asics).
+ *
+ * Returns 0 on success, error on failure.
+ */
+int amdgpu_queue_mgr_fini(struct amdgpu_device *adev,
+			  struct amdgpu_queue_mgr *mgr)
+{
+	return 0;
+}
+
+/**
+ * amdgpu_queue_mgr_map - Map a userspace ring id to an amdgpu_ring
+ *
+ * @adev: amdgpu_device pointer
+ * @mgr: amdgpu_queue_mgr structure holding queue information
+ * @hw_ip: HW IP enum
+ * @instance: HW instance
+ * @ring: user ring id
+ * @our_ring: pointer to mapped amdgpu_ring
+ *
+ * Map a userspace ring id to an appropriate kernel ring. Different
+ * policies are configurable at a HW IP level.
+ *
+ * Returns 0 on success, error on failure.
+ */
+int amdgpu_queue_mgr_map(struct amdgpu_device *adev,
+			 struct amdgpu_queue_mgr *mgr,
+			 int hw_ip, int instance, int ring,
+			 struct amdgpu_ring **out_ring)
+{
+	int r, ip_num_rings;
+	struct amdgpu_queue_mapper *mapper = &mgr->mapper[hw_ip];
+
+	if (!adev || !mgr || !out_ring)
+		return -EINVAL;
+
+	if (hw_ip >= AMDGPU_MAX_IP_NUM)
+		return -EINVAL;
+
+	if (ring >= AMDGPU_MAX_RINGS)
+		return -EINVAL;
+
+	/* Right now all IPs have only one instance - multiple rings. */
+	if (instance != 0) {
+		DRM_ERROR("invalid ip instance: %d\n", instance);
+		return -EINVAL;
+	}
+
+	switch (hw_ip) {
+	case AMDGPU_HW_IP_GFX:
+		ip_num_rings = adev->gfx.num_gfx_rings;
+		break;
+	case AMDGPU_HW_IP_COMPUTE:
+		ip_num_rings = adev->gfx.num_compute_rings;
+		break;
+	case AMDGPU_HW_IP_DMA:
+		ip_num_rings = adev->sdma.num_instances;
+		break;
+	case AMDGPU_HW_IP_UVD:
+		ip_num_rings = 1;
+		break;
+	case AMDGPU_HW_IP_VCE:
+		ip_num_rings = adev->vce.num_rings;
+		break;
+	case AMDGPU_HW_IP_UVD_ENC:
+		ip_num_rings = adev->uvd.num_enc_rings;
+		break;
+	case AMDGPU_HW_IP_VCN_DEC:
+		ip_num_rings = 1;
+		break;
+	case AMDGPU_HW_IP_VCN_ENC:
+		ip_num_rings = adev->vcn.num_enc_rings;
+		break;
+	default:
+		DRM_ERROR("unknown ip type: %d\n", hw_ip);
+		return -EINVAL;
+	}
+
+	if (ring >= ip_num_rings) {
+		DRM_ERROR("Ring index:%d exceeds maximum:%d for ip:%d\n",
+				ring, ip_num_rings, hw_ip);
+		return -EINVAL;
+	}
+
+	mutex_lock(&mapper->lock);
+
+	*out_ring = amdgpu_get_cached_map(mapper, ring);
+	if (*out_ring) {
+		/* cache hit */
+		r = 0;
+		goto out_unlock;
+	}
+
+	switch (mapper->hw_ip) {
+	case AMDGPU_HW_IP_GFX:
+	case AMDGPU_HW_IP_UVD:
+	case AMDGPU_HW_IP_VCE:
+	case AMDGPU_HW_IP_UVD_ENC:
+	case AMDGPU_HW_IP_VCN_DEC:
+	case AMDGPU_HW_IP_VCN_ENC:
+		r = amdgpu_identity_map(adev, mapper, ring, out_ring);
+		break;
+	case AMDGPU_HW_IP_DMA:
+	case AMDGPU_HW_IP_COMPUTE:
+		r = amdgpu_lru_map(adev, mapper, ring, out_ring);
+		break;
+	default:
+		*out_ring = NULL;
+		r = -EINVAL;
+		DRM_ERROR("unknown HW IP type: %d\n", mapper->hw_ip);
+	}
+
+out_unlock:
+	mutex_unlock(&mapper->lock);
+	return r;
+}

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

@@ -135,6 +135,8 @@ void amdgpu_ring_commit(struct amdgpu_ring *ring)
 
 	if (ring->funcs->end_use)
 		ring->funcs->end_use(ring);
+
+	amdgpu_ring_lru_touch(ring->adev, ring);
 }
 
 /**
@@ -253,10 +255,13 @@ int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring,
 	}
 
 	ring->max_dw = max_dw;
+	INIT_LIST_HEAD(&ring->lru_list);
+	amdgpu_ring_lru_touch(adev, ring);
 
 	if (amdgpu_debugfs_ring_init(adev, ring)) {
 		DRM_ERROR("Failed to register debugfs file for rings !\n");
 	}
+
 	return 0;
 }
 
@@ -294,6 +299,84 @@ void amdgpu_ring_fini(struct amdgpu_ring *ring)
 	ring->adev->rings[ring->idx] = NULL;
 }
 
+static void amdgpu_ring_lru_touch_locked(struct amdgpu_device *adev,
+					 struct amdgpu_ring *ring)
+{
+	/* list_move_tail handles the case where ring isn't part of the list */
+	list_move_tail(&ring->lru_list, &adev->ring_lru_list);
+}
+
+static bool amdgpu_ring_is_blacklisted(struct amdgpu_ring *ring,
+				       int *blacklist, int num_blacklist)
+{
+	int i;
+
+	for (i = 0; i < num_blacklist; i++) {
+		if (ring->idx == blacklist[i])
+			return true;
+	}
+
+	return false;
+}
+
+/**
+ * amdgpu_ring_lru_get - get the least recently used ring for a HW IP block
+ *
+ * @adev: amdgpu_device pointer
+ * @type: amdgpu_ring_type enum
+ * @blacklist: blacklisted ring ids array
+ * @num_blacklist: number of entries in @blacklist
+ * @ring: output ring
+ *
+ * Retrieve the amdgpu_ring structure for the least recently used ring of
+ * a specific IP block (all asics).
+ * Returns 0 on success, error on failure.
+ */
+int amdgpu_ring_lru_get(struct amdgpu_device *adev, int type, int *blacklist,
+			int num_blacklist, struct amdgpu_ring **ring)
+{
+	struct amdgpu_ring *entry;
+
+	/* List is sorted in LRU order, find first entry corresponding
+	 * to the desired HW IP */
+	*ring = NULL;
+	spin_lock(&adev->ring_lru_list_lock);
+	list_for_each_entry(entry, &adev->ring_lru_list, lru_list) {
+		if (entry->funcs->type != type)
+			continue;
+
+		if (amdgpu_ring_is_blacklisted(entry, blacklist, num_blacklist))
+			continue;
+
+		*ring = entry;
+		amdgpu_ring_lru_touch_locked(adev, *ring);
+		break;
+	}
+	spin_unlock(&adev->ring_lru_list_lock);
+
+	if (!*ring) {
+		DRM_ERROR("Ring LRU contains no entries for ring type:%d\n", type);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+/**
+ * amdgpu_ring_lru_touch - mark a ring as recently being used
+ *
+ * @adev: amdgpu_device pointer
+ * @ring: ring to touch
+ *
+ * Move @ring to the tail of the lru list
+ */
+void amdgpu_ring_lru_touch(struct amdgpu_device *adev, struct amdgpu_ring *ring)
+{
+	spin_lock(&adev->ring_lru_list_lock);
+	amdgpu_ring_lru_touch_locked(adev, ring);
+	spin_unlock(&adev->ring_lru_list_lock);
+}
+
 /*
  * Debugfs info
  */

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

@@ -47,7 +47,9 @@ enum amdgpu_ring_type {
 	AMDGPU_RING_TYPE_UVD,
 	AMDGPU_RING_TYPE_VCE,
 	AMDGPU_RING_TYPE_KIQ,
-	AMDGPU_RING_TYPE_UVD_ENC
+	AMDGPU_RING_TYPE_UVD_ENC,
+	AMDGPU_RING_TYPE_VCN_DEC,
+	AMDGPU_RING_TYPE_VCN_ENC
 };
 
 struct amdgpu_device;
@@ -76,6 +78,7 @@ struct amdgpu_fence_driver {
 int amdgpu_fence_driver_init(struct amdgpu_device *adev);
 void amdgpu_fence_driver_fini(struct amdgpu_device *adev);
 void amdgpu_fence_driver_force_completion(struct amdgpu_device *adev);
+void amdgpu_fence_driver_force_completion_ring(struct amdgpu_ring *ring);
 
 int amdgpu_fence_driver_init_ring(struct amdgpu_ring *ring,
 				  unsigned num_hw_submission);
@@ -130,6 +133,7 @@ struct amdgpu_ring_funcs {
 	int (*test_ib)(struct amdgpu_ring *ring, long timeout);
 	/* insert NOP packets */
 	void (*insert_nop)(struct amdgpu_ring *ring, uint32_t count);
+	void (*insert_start)(struct amdgpu_ring *ring);
 	void (*insert_end)(struct amdgpu_ring *ring);
 	/* pad the indirect buffer to the necessary number of dw */
 	void (*pad_ib)(struct amdgpu_ring *ring, struct amdgpu_ib *ib);
@@ -142,6 +146,7 @@ struct amdgpu_ring_funcs {
 	void (*emit_cntxcntl) (struct amdgpu_ring *ring, uint32_t flags);
 	void (*emit_rreg)(struct amdgpu_ring *ring, uint32_t reg);
 	void (*emit_wreg)(struct amdgpu_ring *ring, uint32_t reg, uint32_t val);
+	void (*emit_tmz)(struct amdgpu_ring *ring, bool start);
 };
 
 struct amdgpu_ring {
@@ -149,6 +154,7 @@ struct amdgpu_ring {
 	const struct amdgpu_ring_funcs	*funcs;
 	struct amdgpu_fence_driver	fence_drv;
 	struct amd_gpu_scheduler	sched;
+	struct list_head		lru_list;
 
 	struct amdgpu_bo	*ring_obj;
 	volatile uint32_t	*ring;
@@ -180,6 +186,7 @@ struct amdgpu_ring {
 	u64			cond_exe_gpu_addr;
 	volatile u32		*cond_exe_cpu_addr;
 	unsigned		vm_inv_eng;
+	bool			has_compute_vm_bug;
 #if defined(CONFIG_DEBUG_FS)
 	struct dentry *ent;
 #endif
@@ -194,6 +201,9 @@ int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring,
 		     unsigned ring_size, struct amdgpu_irq_src *irq_src,
 		     unsigned irq_type);
 void amdgpu_ring_fini(struct amdgpu_ring *ring);
+int amdgpu_ring_lru_get(struct amdgpu_device *adev, int type, int *blacklist,
+			int num_blacklist, struct amdgpu_ring **ring);
+void amdgpu_ring_lru_touch(struct amdgpu_device *adev, struct amdgpu_ring *ring);
 static inline void amdgpu_ring_clear_ring(struct amdgpu_ring *ring)
 {
 	int i = 0;

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

@@ -298,6 +298,25 @@ struct dma_fence *amdgpu_sync_get_fence(struct amdgpu_sync *sync)
 	return NULL;
 }
 
+int amdgpu_sync_wait(struct amdgpu_sync *sync, bool intr)
+{
+	struct amdgpu_sync_entry *e;
+	struct hlist_node *tmp;
+	int i, r;
+
+	hash_for_each_safe(sync->fences, i, tmp, e, node) {
+		r = dma_fence_wait(e->fence, intr);
+		if (r)
+			return r;
+
+		hash_del(&e->node);
+		dma_fence_put(e->fence);
+		kmem_cache_free(amdgpu_sync_slab, e);
+	}
+
+	return 0;
+}
+
 /**
  * amdgpu_sync_free - free the sync object
  *

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

@@ -49,6 +49,7 @@ int amdgpu_sync_resv(struct amdgpu_device *adev,
 struct dma_fence *amdgpu_sync_peek_fence(struct amdgpu_sync *sync,
 				     struct amdgpu_ring *ring);
 struct dma_fence *amdgpu_sync_get_fence(struct amdgpu_sync *sync);
+int amdgpu_sync_wait(struct amdgpu_sync *sync, bool intr);
 void amdgpu_sync_free(struct amdgpu_sync *sync);
 int amdgpu_sync_init(void);
 void amdgpu_sync_fini(void);

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

@@ -745,6 +745,7 @@ int amdgpu_ttm_bind(struct ttm_buffer_object *bo, struct ttm_mem_reg *bo_mem)
 		return r;
 	}
 
+	spin_lock(&gtt->adev->gtt_list_lock);
 	flags = amdgpu_ttm_tt_pte_flags(gtt->adev, ttm, bo_mem);
 	gtt->offset = (u64)bo_mem->start << PAGE_SHIFT;
 	r = amdgpu_gart_bind(gtt->adev, gtt->offset, ttm->num_pages,
@@ -753,12 +754,13 @@ int amdgpu_ttm_bind(struct ttm_buffer_object *bo, struct ttm_mem_reg *bo_mem)
 	if (r) {
 		DRM_ERROR("failed to bind %lu pages at 0x%08llX\n",
 			  ttm->num_pages, gtt->offset);
-		return r;
+		goto error_gart_bind;
 	}
-	spin_lock(&gtt->adev->gtt_list_lock);
+
 	list_add_tail(&gtt->list, &gtt->adev->gtt_list);
+error_gart_bind:
 	spin_unlock(&gtt->adev->gtt_list_lock);
-	return 0;
+	return r;
 }
 
 int amdgpu_ttm_recover_gart(struct amdgpu_device *adev)
@@ -789,6 +791,7 @@ int amdgpu_ttm_recover_gart(struct amdgpu_device *adev)
 static int amdgpu_ttm_backend_unbind(struct ttm_tt *ttm)
 {
 	struct amdgpu_ttm_tt *gtt = (void *)ttm;
+	int r;
 
 	if (gtt->userptr)
 		amdgpu_ttm_tt_unpin_userptr(ttm);
@@ -797,14 +800,17 @@ static int amdgpu_ttm_backend_unbind(struct ttm_tt *ttm)
 		return 0;
 
 	/* unbind shouldn't be done for GDS/GWS/OA in ttm_bo_clean_mm */
-	if (gtt->adev->gart.ready)
-		amdgpu_gart_unbind(gtt->adev, gtt->offset, ttm->num_pages);
-
 	spin_lock(&gtt->adev->gtt_list_lock);
+	r = amdgpu_gart_unbind(gtt->adev, gtt->offset, ttm->num_pages);
+	if (r) {
+		DRM_ERROR("failed to unbind %lu pages at 0x%08llX\n",
+			  gtt->ttm.ttm.num_pages, gtt->offset);
+		goto error_unbind;
+	}
 	list_del_init(&gtt->list);
+error_unbind:
 	spin_unlock(&gtt->adev->gtt_list_lock);
-
-	return 0;
+	return r;
 }
 
 static void amdgpu_ttm_backend_destroy(struct ttm_tt *ttm)
@@ -1115,7 +1121,7 @@ int amdgpu_ttm_init(struct amdgpu_device *adev)
 	/* Change the size here instead of the init above so only lpfn is affected */
 	amdgpu_ttm_set_active_vram_size(adev, adev->mc.visible_vram_size);
 
-	r = amdgpu_bo_create(adev, 256 * 1024, PAGE_SIZE, true,
+	r = amdgpu_bo_create(adev, adev->mc.stolen_size, PAGE_SIZE, true,
 			     AMDGPU_GEM_DOMAIN_VRAM,
 			     AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED |
 			     AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS,
@@ -1462,6 +1468,9 @@ static ssize_t amdgpu_ttm_vram_read(struct file *f, char __user *buf,
 	if (size & 0x3 || *pos & 0x3)
 		return -EINVAL;
 
+	if (*pos >= adev->mc.mc_vram_size)
+		return -ENXIO;
+
 	while (size) {
 		unsigned long flags;
 		uint32_t value;

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

@@ -197,6 +197,27 @@ void amdgpu_ucode_print_sdma_hdr(const struct common_firmware_header *hdr)
 	}
 }
 
+void amdgpu_ucode_print_gpu_info_hdr(const struct common_firmware_header *hdr)
+{
+	uint16_t version_major = le16_to_cpu(hdr->header_version_major);
+	uint16_t version_minor = le16_to_cpu(hdr->header_version_minor);
+
+	DRM_DEBUG("GPU_INFO\n");
+	amdgpu_ucode_print_common_hdr(hdr);
+
+	if (version_major == 1) {
+		const struct gpu_info_firmware_header_v1_0 *gpu_info_hdr =
+			container_of(hdr, struct gpu_info_firmware_header_v1_0, header);
+
+		DRM_DEBUG("version_major: %u\n",
+			  le16_to_cpu(gpu_info_hdr->version_major));
+		DRM_DEBUG("version_minor: %u\n",
+			  le16_to_cpu(gpu_info_hdr->version_minor));
+	} else {
+		DRM_ERROR("Unknown gpu_info ucode version: %u.%u\n", version_major, version_minor);
+	}
+}
+
 int amdgpu_ucode_validate(const struct firmware *fw)
 {
 	const struct common_firmware_header *hdr =
@@ -253,6 +274,15 @@ amdgpu_ucode_get_load_type(struct amdgpu_device *adev, int load_type)
 			return AMDGPU_FW_LOAD_DIRECT;
 		else
 			return AMDGPU_FW_LOAD_PSP;
+	case CHIP_RAVEN:
+#if 0
+		if (!load_type)
+			return AMDGPU_FW_LOAD_DIRECT;
+		else
+			return AMDGPU_FW_LOAD_PSP;
+#else
+		return AMDGPU_FW_LOAD_DIRECT;
+#endif
 	default:
 		DRM_ERROR("Unknow firmware load type\n");
 	}
@@ -349,7 +379,8 @@ int amdgpu_ucode_init_bo(struct amdgpu_device *adev)
 
 	err = amdgpu_bo_create(adev, adev->firmware.fw_size, PAGE_SIZE, true,
 				amdgpu_sriov_vf(adev) ? AMDGPU_GEM_DOMAIN_VRAM : AMDGPU_GEM_DOMAIN_GTT,
-				0, NULL, NULL, bo);
+				AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS,
+				NULL, NULL, bo);
 	if (err) {
 		dev_err(adev->dev, "(%d) Firmware buffer allocate failed\n", err);
 		goto failed;

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

@@ -113,6 +113,32 @@ struct sdma_firmware_header_v1_1 {
 	uint32_t digest_size;
 };
 
+/* gpu info payload */
+struct gpu_info_firmware_v1_0 {
+	uint32_t gc_num_se;
+	uint32_t gc_num_cu_per_sh;
+	uint32_t gc_num_sh_per_se;
+	uint32_t gc_num_rb_per_se;
+	uint32_t gc_num_tccs;
+	uint32_t gc_num_gprs;
+	uint32_t gc_num_max_gs_thds;
+	uint32_t gc_gs_table_depth;
+	uint32_t gc_gsprim_buff_depth;
+	uint32_t gc_parameter_cache_depth;
+	uint32_t gc_double_offchip_lds_buffer;
+	uint32_t gc_wave_size;
+	uint32_t gc_max_waves_per_simd;
+	uint32_t gc_max_scratch_slots_per_cu;
+	uint32_t gc_lds_size;
+};
+
+/* version_major=1, version_minor=0 */
+struct gpu_info_firmware_header_v1_0 {
+	struct common_firmware_header header;
+	uint16_t version_major; /* version */
+	uint16_t version_minor; /* version */
+};
+
 /* header is fixed size */
 union amdgpu_firmware_header {
 	struct common_firmware_header common;
@@ -124,6 +150,7 @@ union amdgpu_firmware_header {
 	struct rlc_firmware_header_v2_0 rlc_v2_0;
 	struct sdma_firmware_header_v1_0 sdma;
 	struct sdma_firmware_header_v1_1 sdma_v1_1;
+	struct gpu_info_firmware_header_v1_0 gpu_info;
 	uint8_t raw[0x100];
 };
 
@@ -184,6 +211,7 @@ void amdgpu_ucode_print_smc_hdr(const struct common_firmware_header *hdr);
 void amdgpu_ucode_print_gfx_hdr(const struct common_firmware_header *hdr);
 void amdgpu_ucode_print_rlc_hdr(const struct common_firmware_header *hdr);
 void amdgpu_ucode_print_sdma_hdr(const struct common_firmware_header *hdr);
+void amdgpu_ucode_print_gpu_info_hdr(const struct common_firmware_header *hdr);
 int amdgpu_ucode_validate(const struct firmware *fw);
 bool amdgpu_ucode_hdr_version(union amdgpu_firmware_header *hdr,
 				uint16_t hdr_major, uint16_t hdr_minor);

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

@@ -165,35 +165,14 @@ int amdgpu_vce_sw_init(struct amdgpu_device *adev, unsigned long size)
 	adev->vce.fw_version = ((version_major << 24) | (version_minor << 16) |
 				(binary_id << 8));
 
-	/* allocate firmware, stack and heap BO */
-
-	r = amdgpu_bo_create(adev, size, PAGE_SIZE, true,
-			     AMDGPU_GEM_DOMAIN_VRAM,
-			     AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED |
-			     AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS,
-			     NULL, NULL, &adev->vce.vcpu_bo);
+	r = amdgpu_bo_create_kernel(adev, size, PAGE_SIZE,
+				    AMDGPU_GEM_DOMAIN_VRAM, &adev->vce.vcpu_bo,
+				    &adev->vce.gpu_addr, &adev->vce.cpu_addr);
 	if (r) {
 		dev_err(adev->dev, "(%d) failed to allocate VCE bo\n", r);
 		return r;
 	}
 
-	r = amdgpu_bo_reserve(adev->vce.vcpu_bo, false);
-	if (r) {
-		amdgpu_bo_unref(&adev->vce.vcpu_bo);
-		dev_err(adev->dev, "(%d) failed to reserve VCE bo\n", r);
-		return r;
-	}
-
-	r = amdgpu_bo_pin(adev->vce.vcpu_bo, AMDGPU_GEM_DOMAIN_VRAM,
-			  &adev->vce.gpu_addr);
-	amdgpu_bo_unreserve(adev->vce.vcpu_bo);
-	if (r) {
-		amdgpu_bo_unref(&adev->vce.vcpu_bo);
-		dev_err(adev->dev, "(%d) VCE bo pin failed\n", r);
-		return r;
-	}
-
-
 	ring = &adev->vce.ring[0];
 	rq = &ring->sched.sched_rq[AMD_SCHED_PRIORITY_NORMAL];
 	r = amd_sched_entity_init(&ring->sched, &adev->vce.entity,
@@ -230,7 +209,8 @@ int amdgpu_vce_sw_fini(struct amdgpu_device *adev)
 
 	amd_sched_entity_fini(&adev->vce.ring[0].sched, &adev->vce.entity);
 
-	amdgpu_bo_unref(&adev->vce.vcpu_bo);
+	amdgpu_bo_free_kernel(&adev->vce.vcpu_bo, &adev->vce.gpu_addr,
+		(void **)&adev->vce.cpu_addr);
 
 	for (i = 0; i < adev->vce.num_rings; i++)
 		amdgpu_ring_fini(&adev->vce.ring[i]);

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

@@ -33,6 +33,8 @@
 struct amdgpu_vce {
 	struct amdgpu_bo	*vcpu_bo;
 	uint64_t		gpu_addr;
+	void			*cpu_addr;
+	void			*saved_bo;
 	unsigned		fw_version;
 	unsigned		fb_version;
 	atomic_t		handles[AMDGPU_MAX_VCE_HANDLES];

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

+/*
+ * Copyright 2016 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __AMDGPU_VCN_H__
+#define __AMDGPU_VCN_H__
+
+#define AMDGPU_VCN_STACK_SIZE		(200*1024)
+#define AMDGPU_VCN_HEAP_SIZE		(256*1024)
+#define AMDGPU_VCN_SESSION_SIZE		(50*1024)
+#define AMDGPU_VCN_FIRMWARE_OFFSET	256
+#define AMDGPU_VCN_MAX_ENC_RINGS	3
+
+#define VCN_DEC_CMD_FENCE		0x00000000
+#define VCN_DEC_CMD_TRAP		0x00000001
+#define VCN_DEC_CMD_WRITE_REG		0x00000004
+#define VCN_DEC_CMD_REG_READ_COND_WAIT	0x00000006
+#define VCN_DEC_CMD_PACKET_START	0x0000000a
+#define VCN_DEC_CMD_PACKET_END		0x0000000b
+
+#define VCN_ENC_CMD_NO_OP		0x00000000
+#define VCN_ENC_CMD_END 		0x00000001
+#define VCN_ENC_CMD_IB			0x00000002
+#define VCN_ENC_CMD_FENCE		0x00000003
+#define VCN_ENC_CMD_TRAP		0x00000004
+#define VCN_ENC_CMD_REG_WRITE		0x0000000b
+#define VCN_ENC_CMD_REG_WAIT		0x0000000c
+
+struct amdgpu_vcn {
+	struct amdgpu_bo	*vcpu_bo;
+	void			*cpu_addr;
+	uint64_t		gpu_addr;
+	unsigned		fw_version;
+	void			*saved_bo;
+	struct delayed_work	idle_work;
+	const struct firmware	*fw;	/* VCN firmware */
+	struct amdgpu_ring	ring_dec;
+	struct amdgpu_ring	ring_enc[AMDGPU_VCN_MAX_ENC_RINGS];
+	struct amdgpu_irq_src	irq;
+	struct amd_sched_entity entity_dec;
+	struct amd_sched_entity entity_enc;
+	unsigned		num_enc_rings;
+};
+
+int amdgpu_vcn_sw_init(struct amdgpu_device *adev);
+int amdgpu_vcn_sw_fini(struct amdgpu_device *adev);
+int amdgpu_vcn_suspend(struct amdgpu_device *adev);
+int amdgpu_vcn_resume(struct amdgpu_device *adev);
+void amdgpu_vcn_ring_begin_use(struct amdgpu_ring *ring);
+void amdgpu_vcn_ring_end_use(struct amdgpu_ring *ring);
+
+int amdgpu_vcn_dec_ring_test_ring(struct amdgpu_ring *ring);
+int amdgpu_vcn_dec_ring_test_ib(struct amdgpu_ring *ring, long timeout);
+
+int amdgpu_vcn_enc_ring_test_ring(struct amdgpu_ring *ring);
+int amdgpu_vcn_enc_ring_test_ib(struct amdgpu_ring *ring, long timeout);
+
+#endif

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

@@ -22,6 +22,7 @@
  */
 
 #include "amdgpu.h"
+#define MAX_KIQ_REG_WAIT	100000
 
 int amdgpu_allocate_static_csa(struct amdgpu_device *adev)
 {
@@ -105,8 +106,9 @@ void amdgpu_virt_init_setting(struct amdgpu_device *adev)
 	/* enable virtual display */
 	adev->mode_info.num_crtc = 1;
 	adev->enable_virtual_display = true;
+	adev->cg_flags = 0;
+	adev->pg_flags = 0;
 
-	mutex_init(&adev->virt.lock_kiq);
 	mutex_init(&adev->virt.lock_reset);
 }
 
@@ -120,17 +122,19 @@ uint32_t amdgpu_virt_kiq_rreg(struct amdgpu_device *adev, uint32_t reg)
 
 	BUG_ON(!ring->funcs->emit_rreg);
 
-	mutex_lock(&adev->virt.lock_kiq);
+	mutex_lock(&kiq->ring_mutex);
 	amdgpu_ring_alloc(ring, 32);
 	amdgpu_ring_emit_rreg(ring, reg);
 	amdgpu_fence_emit(ring, &f);
 	amdgpu_ring_commit(ring);
-	mutex_unlock(&adev->virt.lock_kiq);
+	mutex_unlock(&kiq->ring_mutex);
 
-	r = dma_fence_wait(f, false);
-	if (r)
-		DRM_ERROR("wait for kiq fence error: %ld.\n", r);
+	r = dma_fence_wait_timeout(f, false, msecs_to_jiffies(MAX_KIQ_REG_WAIT));
 	dma_fence_put(f);
+	if (r < 1) {
+		DRM_ERROR("wait for kiq fence error: %ld.\n", r);
+		return ~0;
+	}
 
 	val = adev->wb.wb[adev->virt.reg_val_offs];
 
@@ -146,15 +150,15 @@ void amdgpu_virt_kiq_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v)
 
 	BUG_ON(!ring->funcs->emit_wreg);
 
-	mutex_lock(&adev->virt.lock_kiq);
+	mutex_lock(&kiq->ring_mutex);
 	amdgpu_ring_alloc(ring, 32);
 	amdgpu_ring_emit_wreg(ring, reg, v);
 	amdgpu_fence_emit(ring, &f);
 	amdgpu_ring_commit(ring);
-	mutex_unlock(&adev->virt.lock_kiq);
+	mutex_unlock(&kiq->ring_mutex);
 
-	r = dma_fence_wait(f, false);
-	if (r)
+	r = dma_fence_wait_timeout(f, false, msecs_to_jiffies(MAX_KIQ_REG_WAIT));
+	if (r < 1)
 		DRM_ERROR("wait for kiq fence error: %ld.\n", r);
 	dma_fence_put(f);
 }

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

@@ -52,7 +52,6 @@ struct amdgpu_virt {
 	uint64_t			csa_vmid0_addr;
 	bool chained_ib_support;
 	uint32_t			reg_val_offs;
-	struct mutex			lock_kiq;
 	struct mutex                    lock_reset;
 	struct amdgpu_irq_src		ack_irq;
 	struct amdgpu_irq_src		rcv_irq;
@@ -97,7 +96,7 @@ void amdgpu_virt_kiq_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v);
 int amdgpu_virt_request_full_gpu(struct amdgpu_device *adev, bool init);
 int amdgpu_virt_release_full_gpu(struct amdgpu_device *adev, bool init);
 int amdgpu_virt_reset_gpu(struct amdgpu_device *adev);
-int amdgpu_sriov_gpu_reset(struct amdgpu_device *adev, bool voluntary);
+int amdgpu_sriov_gpu_reset(struct amdgpu_device *adev, struct amdgpu_job *job);
 int amdgpu_virt_alloc_mm_table(struct amdgpu_device *adev);
 void amdgpu_virt_free_mm_table(struct amdgpu_device *adev);
 

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

@@ -84,6 +84,16 @@ struct amdgpu_bo_list_entry;
 
 /* hardcode that limit for now */
 #define AMDGPU_VA_RESERVED_SIZE			(8 << 20)
+/* max vmids dedicated for process */
+#define AMDGPU_VM_MAX_RESERVED_VMID	1
+
+#define AMDGPU_VM_CONTEXT_GFX 0
+#define AMDGPU_VM_CONTEXT_COMPUTE 1
+
+/* See vm_update_mode */
+#define AMDGPU_VM_USE_CPU_FOR_GFX (1 << 0)
+#define AMDGPU_VM_USE_CPU_FOR_COMPUTE (1 << 1)
+
 
 struct amdgpu_vm_pt {
 	struct amdgpu_bo	*bo;
@@ -123,8 +133,13 @@ struct amdgpu_vm {
 
 	/* client id */
 	u64                     client_id;
+	/* dedicated to vm */
+	struct amdgpu_vm_id	*reserved_vmid[AMDGPU_MAX_VMHUBS];
 	/* each VM will map on CSA */
 	struct amdgpu_bo_va *csa_bo_va;
+
+	/* Flag to indicate if VM tables are updated by CPU or GPU (SDMA) */
+	bool                    use_cpu_for_update;
 };
 
 struct amdgpu_vm_id {
@@ -152,6 +167,7 @@ struct amdgpu_vm_id_manager {
 	unsigned		num_ids;
 	struct list_head	ids_lru;
 	struct amdgpu_vm_id	ids[AMDGPU_NUM_VM];
+	atomic_t		reserved_vmid_num;
 };
 
 struct amdgpu_vm_manager {
@@ -168,8 +184,6 @@ struct amdgpu_vm_manager {
 	uint32_t				block_size;
 	/* vram base address for page table entry  */
 	u64					vram_base_offset;
-	/* is vm enabled? */
-	bool					enabled;
 	/* vm pte handling */
 	const struct amdgpu_vm_pte_funcs        *vm_pte_funcs;
 	struct amdgpu_ring                      *vm_pte_rings[AMDGPU_MAX_RINGS];
@@ -181,11 +195,18 @@ struct amdgpu_vm_manager {
 	/* partial resident texture handling */
 	spinlock_t				prt_lock;
 	atomic_t				num_prt_users;
+
+	/* controls how VM page tables are updated for Graphics and Compute.
+	 * BIT0[= 0] Graphics updated by SDMA [= 1] by CPU
+	 * BIT1[= 0] Compute updated by SDMA [= 1] by CPU
+	 */
+	int					vm_update_mode;
 };
 
 void amdgpu_vm_manager_init(struct amdgpu_device *adev);
 void amdgpu_vm_manager_fini(struct amdgpu_device *adev);
-int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm);
+int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm,
+		   int vm_context);
 void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm);
 void amdgpu_vm_get_pd_bo(struct amdgpu_vm *vm,
 			 struct list_head *validated,
@@ -239,5 +260,9 @@ int amdgpu_vm_bo_clear_mappings(struct amdgpu_device *adev,
 void amdgpu_vm_bo_rmv(struct amdgpu_device *adev,
 		      struct amdgpu_bo_va *bo_va);
 void amdgpu_vm_adjust_size(struct amdgpu_device *adev, uint64_t vm_size);
+int amdgpu_vm_ioctl(struct drm_device *dev, void *data, struct drm_file *filp);
+bool amdgpu_vm_need_pipeline_sync(struct amdgpu_ring *ring,
+				  struct amdgpu_job *job);
+void amdgpu_vm_check_compute_bug(struct amdgpu_device *adev);
 
 #endif

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

@@ -964,62 +964,62 @@ static bool cik_read_bios_from_rom(struct amdgpu_device *adev,
 }
 
 static const struct amdgpu_allowed_register_entry cik_allowed_read_registers[] = {
-	{mmGRBM_STATUS, false},
-	{mmGB_ADDR_CONFIG, false},
-	{mmMC_ARB_RAMCFG, false},
-	{mmGB_TILE_MODE0, false},
-	{mmGB_TILE_MODE1, false},
-	{mmGB_TILE_MODE2, false},
-	{mmGB_TILE_MODE3, false},
-	{mmGB_TILE_MODE4, false},
-	{mmGB_TILE_MODE5, false},
-	{mmGB_TILE_MODE6, false},
-	{mmGB_TILE_MODE7, false},
-	{mmGB_TILE_MODE8, false},
-	{mmGB_TILE_MODE9, false},
-	{mmGB_TILE_MODE10, false},
-	{mmGB_TILE_MODE11, false},
-	{mmGB_TILE_MODE12, false},
-	{mmGB_TILE_MODE13, false},
-	{mmGB_TILE_MODE14, false},
-	{mmGB_TILE_MODE15, false},
-	{mmGB_TILE_MODE16, false},
-	{mmGB_TILE_MODE17, false},
-	{mmGB_TILE_MODE18, false},
-	{mmGB_TILE_MODE19, false},
-	{mmGB_TILE_MODE20, false},
-	{mmGB_TILE_MODE21, false},
-	{mmGB_TILE_MODE22, false},
-	{mmGB_TILE_MODE23, false},
-	{mmGB_TILE_MODE24, false},
-	{mmGB_TILE_MODE25, false},
-	{mmGB_TILE_MODE26, false},
-	{mmGB_TILE_MODE27, false},
-	{mmGB_TILE_MODE28, false},
-	{mmGB_TILE_MODE29, false},
-	{mmGB_TILE_MODE30, false},
-	{mmGB_TILE_MODE31, false},
-	{mmGB_MACROTILE_MODE0, false},
-	{mmGB_MACROTILE_MODE1, false},
-	{mmGB_MACROTILE_MODE2, false},
-	{mmGB_MACROTILE_MODE3, false},
-	{mmGB_MACROTILE_MODE4, false},
-	{mmGB_MACROTILE_MODE5, false},
-	{mmGB_MACROTILE_MODE6, false},
-	{mmGB_MACROTILE_MODE7, false},
-	{mmGB_MACROTILE_MODE8, false},
-	{mmGB_MACROTILE_MODE9, false},
-	{mmGB_MACROTILE_MODE10, false},
-	{mmGB_MACROTILE_MODE11, false},
-	{mmGB_MACROTILE_MODE12, false},
-	{mmGB_MACROTILE_MODE13, false},
-	{mmGB_MACROTILE_MODE14, false},
-	{mmGB_MACROTILE_MODE15, false},
-	{mmCC_RB_BACKEND_DISABLE, false, true},
-	{mmGC_USER_RB_BACKEND_DISABLE, false, true},
-	{mmGB_BACKEND_MAP, false, false},
-	{mmPA_SC_RASTER_CONFIG, false, true},
-	{mmPA_SC_RASTER_CONFIG_1, false, true},
+	{mmGRBM_STATUS},
+	{mmGB_ADDR_CONFIG},
+	{mmMC_ARB_RAMCFG},
+	{mmGB_TILE_MODE0},
+	{mmGB_TILE_MODE1},
+	{mmGB_TILE_MODE2},
+	{mmGB_TILE_MODE3},
+	{mmGB_TILE_MODE4},
+	{mmGB_TILE_MODE5},
+	{mmGB_TILE_MODE6},
+	{mmGB_TILE_MODE7},
+	{mmGB_TILE_MODE8},
+	{mmGB_TILE_MODE9},
+	{mmGB_TILE_MODE10},
+	{mmGB_TILE_MODE11},
+	{mmGB_TILE_MODE12},
+	{mmGB_TILE_MODE13},
+	{mmGB_TILE_MODE14},
+	{mmGB_TILE_MODE15},
+	{mmGB_TILE_MODE16},
+	{mmGB_TILE_MODE17},
+	{mmGB_TILE_MODE18},
+	{mmGB_TILE_MODE19},
+	{mmGB_TILE_MODE20},
+	{mmGB_TILE_MODE21},
+	{mmGB_TILE_MODE22},
+	{mmGB_TILE_MODE23},
+	{mmGB_TILE_MODE24},
+	{mmGB_TILE_MODE25},
+	{mmGB_TILE_MODE26},
+	{mmGB_TILE_MODE27},
+	{mmGB_TILE_MODE28},
+	{mmGB_TILE_MODE29},
+	{mmGB_TILE_MODE30},
+	{mmGB_TILE_MODE31},
+	{mmGB_MACROTILE_MODE0},
+	{mmGB_MACROTILE_MODE1},
+	{mmGB_MACROTILE_MODE2},
+	{mmGB_MACROTILE_MODE3},
+	{mmGB_MACROTILE_MODE4},
+	{mmGB_MACROTILE_MODE5},
+	{mmGB_MACROTILE_MODE6},
+	{mmGB_MACROTILE_MODE7},
+	{mmGB_MACROTILE_MODE8},
+	{mmGB_MACROTILE_MODE9},
+	{mmGB_MACROTILE_MODE10},
+	{mmGB_MACROTILE_MODE11},
+	{mmGB_MACROTILE_MODE12},
+	{mmGB_MACROTILE_MODE13},
+	{mmGB_MACROTILE_MODE14},
+	{mmGB_MACROTILE_MODE15},
+	{mmCC_RB_BACKEND_DISABLE, true},
+	{mmGC_USER_RB_BACKEND_DISABLE, true},
+	{mmGB_BACKEND_MAP, false},
+	{mmPA_SC_RASTER_CONFIG, true},
+	{mmPA_SC_RASTER_CONFIG_1, true},
 };
 
 static uint32_t cik_read_indexed_register(struct amdgpu_device *adev,
@@ -1050,11 +1050,10 @@ static int cik_read_register(struct amdgpu_device *adev, u32 se_num,
 		if (reg_offset != cik_allowed_read_registers[i].reg_offset)
 			continue;
 
-		if (!cik_allowed_read_registers[i].untouched)
-			*value = cik_allowed_read_registers[i].grbm_indexed ?
-				 cik_read_indexed_register(adev, se_num,
-							   sh_num, reg_offset) :
-				 RREG32(reg_offset);
+		*value = cik_allowed_read_registers[i].grbm_indexed ?
+			 cik_read_indexed_register(adev, se_num,
+						   sh_num, reg_offset) :
+			 RREG32(reg_offset);
 		return 0;
 	}
 	return -EINVAL;

drivers/gpu/drm/amd/amdgpu/gfx_v7_0.h

@@ -29,4 +29,9 @@ extern const struct amdgpu_ip_block_version gfx_v7_1_ip_block;
 extern const struct amdgpu_ip_block_version gfx_v7_2_ip_block;
 extern const struct amdgpu_ip_block_version gfx_v7_3_ip_block;
 
+struct amdgpu_device;
+struct cik_mqd;
+
+int gfx_v7_0_mqd_commit(struct amdgpu_device *adev, struct cik_mqd *mqd);
+
 #endif

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.h

@@ -27,4 +27,9 @@
 extern const struct amdgpu_ip_block_version gfx_v8_0_ip_block;
 extern const struct amdgpu_ip_block_version gfx_v8_1_ip_block;
 
+struct amdgpu_device;
+struct vi_mqd;
+
+int gfx_v8_0_mqd_commit(struct amdgpu_device *adev, struct vi_mqd *mqd);
+
 #endif

drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.h

@@ -28,7 +28,8 @@ int gfxhub_v1_0_gart_enable(struct amdgpu_device *adev);
 void gfxhub_v1_0_gart_disable(struct amdgpu_device *adev);
 void gfxhub_v1_0_set_fault_enable_default(struct amdgpu_device *adev,
 					  bool value);
-
+void gfxhub_v1_0_init(struct amdgpu_device *adev);
+u64 gfxhub_v1_0_get_mc_fb_offset(struct amdgpu_device *adev);
 extern const struct amd_ip_funcs gfxhub_v1_0_ip_funcs;
 extern const struct amdgpu_ip_block_version gfxhub_v1_0_ip_block;
 

drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.h

@@ -28,6 +28,10 @@ int mmhub_v1_0_gart_enable(struct amdgpu_device *adev);
 void mmhub_v1_0_gart_disable(struct amdgpu_device *adev);
 void mmhub_v1_0_set_fault_enable_default(struct amdgpu_device *adev,
 					 bool value);
+void mmhub_v1_0_init(struct amdgpu_device *adev);
+int mmhub_v1_0_set_clockgating(struct amdgpu_device *adev,
+			       enum amd_clockgating_state state);
+void mmhub_v1_0_get_clockgating(struct amdgpu_device *adev, u32 *flags);
 
 extern const struct amd_ip_funcs mmhub_v1_0_ip_funcs;
 extern const struct amdgpu_ip_block_version mmhub_v1_0_ip_block;

drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c

@@ -643,8 +643,9 @@ static int sdma_v3_0_gfx_resume(struct amdgpu_device *adev)
 		WREG32(mmSDMA0_GFX_RB_CNTL + sdma_offsets[i], rb_cntl);
 
 		/* Initialize the ring buffer's read and write pointers */
+		ring->wptr = 0;
 		WREG32(mmSDMA0_GFX_RB_RPTR + sdma_offsets[i], 0);
-		WREG32(mmSDMA0_GFX_RB_WPTR + sdma_offsets[i], 0);
+		sdma_v3_0_ring_set_wptr(ring);
 		WREG32(mmSDMA0_GFX_IB_RPTR + sdma_offsets[i], 0);
 		WREG32(mmSDMA0_GFX_IB_OFFSET + sdma_offsets[i], 0);
 
@@ -659,9 +660,6 @@ static int sdma_v3_0_gfx_resume(struct amdgpu_device *adev)
 		WREG32(mmSDMA0_GFX_RB_BASE + sdma_offsets[i], ring->gpu_addr >> 8);
 		WREG32(mmSDMA0_GFX_RB_BASE_HI + sdma_offsets[i], ring->gpu_addr >> 40);
 
-		ring->wptr = 0;
-		WREG32(mmSDMA0_GFX_RB_WPTR + sdma_offsets[i], lower_32_bits(ring->wptr) << 2);
-
 		doorbell = RREG32(mmSDMA0_GFX_DOORBELL + sdma_offsets[i]);
 
 		if (ring->use_doorbell) {