Merge branch 'drm-next-5.2' of git://people.freedesktop.org/~agd5f/li…

…nux into drm-next

amdgpu:
- Switch to HMM for userptr (reverted until HMM fixes land)
- New experimental SMU 11 replacement for powerplay for vega20 (not enabled by default)
- Initial RAS support for vega20
- BACO support for vega12
- BACO fixes for vega20
- Rework IH handling for page fault and retry interrupts
- Cleanly split CPU and GPU paths for GPUVM updates
- Powerplay fixes
- XGMI fixes
- Rework how DC interacts with atomic for planes
- Clean up and simplify DC/Powerplay interfaces
- Misc cleanups and bug fixes

amdkfd:
- Switch to HMM for userptr (reverted until HMM fixes land)
- Add initial RAS support
- MQD fixes

ttm:
- Unify DRM_FILE_PAGE_OFFSET handling
- Account for kernel allocations in kernel zone only
- Misc cleanups

Signed-off-by: Dave Airlie <[email protected]>

From: Alex Deucher <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]

drivers/gpu/drm/Kconfig

@@ -231,8 +231,6 @@ config DRM_AMDGPU
 
 source "drivers/gpu/drm/amd/amdgpu/Kconfig"
 
-source "drivers/gpu/drm/amd/lib/Kconfig"
-
 source "drivers/gpu/drm/nouveau/Kconfig"
 
 source "drivers/gpu/drm/i915/Kconfig"

drivers/gpu/drm/Makefile

@@ -57,7 +57,6 @@ obj-$(CONFIG_DRM_TTM)	+= ttm/
 obj-$(CONFIG_DRM_SCHED)	+= scheduler/
 obj-$(CONFIG_DRM_TDFX)	+= tdfx/
 obj-$(CONFIG_DRM_R128)	+= r128/
-obj-y			+= amd/lib/
 obj-$(CONFIG_HSA_AMD) += amd/amdkfd/
 obj-$(CONFIG_DRM_RADEON)+= radeon/
 obj-$(CONFIG_DRM_AMDGPU)+= amd/amdgpu/

drivers/gpu/drm/amd/amdgpu/Makefile

@@ -53,7 +53,8 @@ amdgpu-y += amdgpu_device.o amdgpu_kms.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_vf_error.o amdgpu_sched.o amdgpu_debugfs.o amdgpu_ids.o \
-	amdgpu_gmc.o amdgpu_xgmi.o amdgpu_csa.o
+	amdgpu_gmc.o amdgpu_xgmi.o amdgpu_csa.o amdgpu_ras.o amdgpu_vm_cpu.o \
+	amdgpu_vm_sdma.o
 
 # add asic specific block
 amdgpu-$(CONFIG_DRM_AMDGPU_CIK)+= cik.o cik_ih.o kv_smc.o kv_dpm.o \

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

@@ -83,6 +83,7 @@
 #include "amdgpu_gem.h"
 #include "amdgpu_doorbell.h"
 #include "amdgpu_amdkfd.h"
+#include "amdgpu_smu.h"
 
 #define MAX_GPU_INSTANCE		16
 
@@ -156,6 +157,8 @@ extern int amdgpu_emu_mode;
 extern uint amdgpu_smu_memory_pool_size;
 extern uint amdgpu_dc_feature_mask;
 extern struct amdgpu_mgpu_info mgpu_info;
+extern int amdgpu_ras_enable;
+extern uint amdgpu_ras_mask;
 
 #ifdef CONFIG_DRM_AMDGPU_SI
 extern int amdgpu_si_support;
@@ -702,7 +705,6 @@ enum amd_hw_ip_block_type {
 struct amd_powerplay {
 	void *pp_handle;
 	const struct amd_pm_funcs *pp_funcs;
-	uint32_t pp_feature;
 };
 
 #define AMDGPU_RESET_MAGIC_NUM 64
@@ -842,6 +844,9 @@ struct amdgpu_device {
 	struct amd_powerplay		powerplay;
 	bool				pp_force_state_enabled;
 
+	/* smu */
+	struct smu_context		smu;
+
 	/* dpm */
 	struct amdgpu_pm		pm;
 	u32				cg_flags;
@@ -922,6 +927,8 @@ struct amdgpu_device {
 
 	int asic_reset_res;
 	struct work_struct		xgmi_reset_work;
+
+	bool                            in_baco_reset;
 };
 
 static inline struct amdgpu_device *amdgpu_ttm_adev(struct ttm_bo_device *bdev)

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

@@ -640,4 +640,8 @@ int kgd2kfd_post_reset(struct kfd_dev *kfd)
 void kgd2kfd_interrupt(struct kfd_dev *kfd, const void *ih_ring_entry)
 {
 }
+
+void kgd2kfd_set_sram_ecc_flag(struct kfd_dev *kfd)
+{
+}
 #endif

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

@@ -230,5 +230,6 @@ int kgd2kfd_quiesce_mm(struct mm_struct *mm);
 int kgd2kfd_resume_mm(struct mm_struct *mm);
 int kgd2kfd_schedule_evict_and_restore_process(struct mm_struct *mm,
 					       struct dma_fence *fence);
+void kgd2kfd_set_sram_ecc_flag(struct kfd_dev *kfd);
 
 #endif /* AMDGPU_AMDKFD_H_INCLUDED */

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

@@ -410,15 +410,7 @@ static int add_bo_to_vm(struct amdgpu_device *adev, struct kgd_mem *mem,
 	if (p_bo_va_entry)
 		*p_bo_va_entry = bo_va_entry;
 
-	/* Allocate new page tables if needed and validate
-	 * them.
-	 */
-	ret = amdgpu_vm_alloc_pts(adev, vm, va, amdgpu_bo_size(bo));
-	if (ret) {
-		pr_err("Failed to allocate pts, err=%d\n", ret);
-		goto err_alloc_pts;
-	}
-
+	/* Allocate validate page tables if needed */
 	ret = vm_validate_pt_pd_bos(vm);
 	if (ret) {
 		pr_err("validate_pt_pd_bos() failed\n");
@@ -741,13 +733,7 @@ static int update_gpuvm_pte(struct amdgpu_device *adev,
 		struct amdgpu_sync *sync)
 {
 	int ret;
-	struct amdgpu_vm *vm;
-	struct amdgpu_bo_va *bo_va;
-	struct amdgpu_bo *bo;
-
-	bo_va = entry->bo_va;
-	vm = bo_va->base.vm;
-	bo = bo_va->base.bo;
+	struct amdgpu_bo_va *bo_va = entry->bo_va;
 
 	/* Update the page tables  */
 	ret = amdgpu_vm_bo_update(adev, bo_va, false);
@@ -906,7 +892,8 @@ static int init_kfd_vm(struct amdgpu_vm *vm, void **process_info,
 		pr_err("validate_pt_pd_bos() failed\n");
 		goto validate_pd_fail;
 	}
-	amdgpu_bo_sync_wait(vm->root.base.bo, AMDGPU_FENCE_OWNER_KFD, false);
+	ret = amdgpu_bo_sync_wait(vm->root.base.bo,
+				  AMDGPU_FENCE_OWNER_KFD, false);
 	if (ret)
 		goto wait_pd_fail;
 	amdgpu_bo_fence(vm->root.base.bo,

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

@@ -28,8 +28,6 @@
 #include "atom.h"
 #include "atombios.h"
 
-#define get_index_into_master_table(master_table, table_name) (offsetof(struct master_table, table_name) / sizeof(uint16_t))
-
 bool amdgpu_atomfirmware_gpu_supports_virtualization(struct amdgpu_device *adev)
 {
 	int index = get_index_into_master_table(atom_master_list_of_data_tables_v2_1,
@@ -238,10 +236,71 @@ int amdgpu_atomfirmware_get_vram_type(struct amdgpu_device *adev)
 	return 0;
 }
 
+/*
+ * Return true if vbios enabled ecc by default, if umc info table is available
+ * or false if ecc is not enabled or umc info table is not available
+ */
+bool amdgpu_atomfirmware_mem_ecc_supported(struct amdgpu_device *adev)
+{
+	struct amdgpu_mode_info *mode_info = &adev->mode_info;
+	int index;
+	u16 data_offset, size;
+	union umc_info *umc_info;
+	u8 frev, crev;
+	bool ecc_default_enabled = false;
+
+	index = get_index_into_master_table(atom_master_list_of_data_tables_v2_1,
+			umc_info);
+
+	if (amdgpu_atom_parse_data_header(mode_info->atom_context,
+				index, &size, &frev, &crev, &data_offset)) {
+		/* support umc_info 3.1+ */
+		if ((frev == 3 && crev >= 1) || (frev > 3)) {
+			umc_info = (union umc_info *)
+				(mode_info->atom_context->bios + data_offset);
+			ecc_default_enabled =
+				(le32_to_cpu(umc_info->v31.umc_config) &
+				 UMC_CONFIG__DEFAULT_MEM_ECC_ENABLE) ? true : false;
+		}
+	}
+
+	return ecc_default_enabled;
+}
+
 union firmware_info {
 	struct atom_firmware_info_v3_1 v31;
 };
 
+/*
+ * Return true if vbios supports sram ecc or false if not
+ */
+bool amdgpu_atomfirmware_sram_ecc_supported(struct amdgpu_device *adev)
+{
+	struct amdgpu_mode_info *mode_info = &adev->mode_info;
+	int index;
+	u16 data_offset, size;
+	union firmware_info *firmware_info;
+	u8 frev, crev;
+	bool sram_ecc_supported = false;
+
+	index = get_index_into_master_table(atom_master_list_of_data_tables_v2_1,
+			firmwareinfo);
+
+	if (amdgpu_atom_parse_data_header(adev->mode_info.atom_context,
+				index, &size, &frev, &crev, &data_offset)) {
+		/* support firmware_info 3.1 + */
+		if ((frev == 3 && crev >=1) || (frev > 3)) {
+			firmware_info = (union firmware_info *)
+				(mode_info->atom_context->bios + data_offset);
+			sram_ecc_supported =
+				(le32_to_cpu(firmware_info->v31.firmware_capability) &
+				 ATOM_FIRMWARE_CAP_SRAM_ECC) ? true : false;
+		}
+	}
+
+	return sram_ecc_supported;
+}
+
 union smu_info {
 	struct atom_smu_info_v3_1 v31;
 };
@@ -346,11 +405,11 @@ int amdgpu_atomfirmware_get_gfx_info(struct amdgpu_device *adev)
 			(mode_info->atom_context->bios + data_offset);
 		switch (crev) {
 		case 4:
-			adev->gfx.config.max_shader_engines = gfx_info->v24.gc_num_se;
-			adev->gfx.config.max_cu_per_sh = gfx_info->v24.gc_num_cu_per_sh;
-			adev->gfx.config.max_sh_per_se = gfx_info->v24.gc_num_sh_per_se;
-			adev->gfx.config.max_backends_per_se = gfx_info->v24.gc_num_rb_per_se;
-			adev->gfx.config.max_texture_channel_caches = gfx_info->v24.gc_num_tccs;
+			adev->gfx.config.max_shader_engines = gfx_info->v24.max_shader_engines;
+			adev->gfx.config.max_cu_per_sh = gfx_info->v24.max_cu_per_sh;
+			adev->gfx.config.max_sh_per_se = gfx_info->v24.max_sh_per_se;
+			adev->gfx.config.max_backends_per_se = gfx_info->v24.max_backends_per_se;
+			adev->gfx.config.max_texture_channel_caches = gfx_info->v24.max_texture_channel_caches;
 			adev->gfx.config.max_gprs = le16_to_cpu(gfx_info->v24.gc_num_gprs);
 			adev->gfx.config.max_gs_threads = gfx_info->v24.gc_num_max_gs_thds;
 			adev->gfx.config.gs_vgt_table_depth = gfx_info->v24.gc_gs_table_depth;

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

@@ -24,12 +24,16 @@
 #ifndef __AMDGPU_ATOMFIRMWARE_H__
 #define __AMDGPU_ATOMFIRMWARE_H__
 
+#define get_index_into_master_table(master_table, table_name) (offsetof(struct master_table, table_name) / sizeof(uint16_t))
+
 bool amdgpu_atomfirmware_gpu_supports_virtualization(struct amdgpu_device *adev);
 void amdgpu_atomfirmware_scratch_regs_init(struct amdgpu_device *adev);
 int amdgpu_atomfirmware_allocate_fb_scratch(struct amdgpu_device *adev);
 int amdgpu_atomfirmware_get_vram_width(struct amdgpu_device *adev);
 int amdgpu_atomfirmware_get_vram_type(struct amdgpu_device *adev);
 int amdgpu_atomfirmware_get_clock_info(struct amdgpu_device *adev);
 int amdgpu_atomfirmware_get_gfx_info(struct amdgpu_device *adev);
+bool amdgpu_atomfirmware_mem_ecc_supported(struct amdgpu_device *adev);
+bool amdgpu_atomfirmware_sram_ecc_supported(struct amdgpu_device *adev);
 
 #endif

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

@@ -92,15 +92,6 @@ int amdgpu_map_static_csa(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 		return -ENOMEM;
 	}
 
-	r = amdgpu_vm_alloc_pts(adev, (*bo_va)->base.vm, csa_addr,
-				size);
-	if (r) {
-		DRM_ERROR("failed to allocate pts for static CSA, err=%d\n", r);
-		amdgpu_vm_bo_rmv(adev, *bo_va);
-		ttm_eu_backoff_reservation(&ticket, &list);
-		return r;
-	}
-
 	r = amdgpu_vm_bo_map(adev, *bo_va, csa_addr, 0, size,
 			     AMDGPU_PTE_READABLE | AMDGPU_PTE_WRITEABLE |
 			     AMDGPU_PTE_EXECUTABLE);

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

@@ -26,6 +26,7 @@
 #include <drm/drm_auth.h>
 #include "amdgpu.h"
 #include "amdgpu_sched.h"
+#include "amdgpu_ras.h"
 
 #define to_amdgpu_ctx_entity(e)	\
 	container_of((e), struct amdgpu_ctx_entity, entity)
@@ -344,6 +345,7 @@ static int amdgpu_ctx_query2(struct amdgpu_device *adev,
 {
 	struct amdgpu_ctx *ctx;
 	struct amdgpu_ctx_mgr *mgr;
+	uint32_t ras_counter;
 
 	if (!fpriv)
 		return -EINVAL;
@@ -368,6 +370,21 @@ static int amdgpu_ctx_query2(struct amdgpu_device *adev,
 	if (atomic_read(&ctx->guilty))
 		out->state.flags |= AMDGPU_CTX_QUERY2_FLAGS_GUILTY;
 
+	/*query ue count*/
+	ras_counter = amdgpu_ras_query_error_count(adev, false);
+	/*ras counter is monotonic increasing*/
+	if (ras_counter != ctx->ras_counter_ue) {
+		out->state.flags |= AMDGPU_CTX_QUERY2_FLAGS_RAS_UE;
+		ctx->ras_counter_ue = ras_counter;
+	}
+
+	/*query ce count*/
+	ras_counter = amdgpu_ras_query_error_count(adev, true);
+	if (ras_counter != ctx->ras_counter_ce) {
+		out->state.flags |= AMDGPU_CTX_QUERY2_FLAGS_RAS_CE;
+		ctx->ras_counter_ce = ras_counter;
+	}
+
 	mutex_unlock(&mgr->lock);
 	return 0;
 }
@@ -541,32 +558,26 @@ void amdgpu_ctx_mgr_init(struct amdgpu_ctx_mgr *mgr)
 	idr_init(&mgr->ctx_handles);
 }
 
-void amdgpu_ctx_mgr_entity_flush(struct amdgpu_ctx_mgr *mgr)
+long amdgpu_ctx_mgr_entity_flush(struct amdgpu_ctx_mgr *mgr, long timeout)
 {
 	unsigned num_entities = amdgput_ctx_total_num_entities();
 	struct amdgpu_ctx *ctx;
 	struct idr *idp;
 	uint32_t id, i;
-	long max_wait = MAX_WAIT_SCHED_ENTITY_Q_EMPTY;
 
 	idp = &mgr->ctx_handles;
 
 	mutex_lock(&mgr->lock);
 	idr_for_each_entry(idp, ctx, id) {
-
-		if (!ctx->adev) {
-			mutex_unlock(&mgr->lock);
-			return;
-		}
-
 		for (i = 0; i < num_entities; i++) {
 			struct drm_sched_entity *entity;
 
 			entity = &ctx->entities[0][i].entity;
-			max_wait = drm_sched_entity_flush(entity, max_wait);
+			timeout = drm_sched_entity_flush(entity, timeout);
 		}
 	}
 	mutex_unlock(&mgr->lock);
+	return timeout;
 }
 
 void amdgpu_ctx_mgr_entity_fini(struct amdgpu_ctx_mgr *mgr)
@@ -579,10 +590,6 @@ void amdgpu_ctx_mgr_entity_fini(struct amdgpu_ctx_mgr *mgr)
 	idp = &mgr->ctx_handles;
 
 	idr_for_each_entry(idp, ctx, id) {
-
-		if (!ctx->adev)
-			return;
-
 		if (kref_read(&ctx->refcount) != 1) {
 			DRM_ERROR("ctx %p is still alive\n", ctx);
 			continue;

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

@@ -49,6 +49,8 @@ struct amdgpu_ctx {
 	enum drm_sched_priority		override_priority;
 	struct mutex			lock;
 	atomic_t			guilty;
+	uint32_t			ras_counter_ce;
+	uint32_t			ras_counter_ue;
 };
 
 struct amdgpu_ctx_mgr {
@@ -82,7 +84,7 @@ int amdgpu_ctx_wait_prev_fence(struct amdgpu_ctx *ctx,
 
 void amdgpu_ctx_mgr_init(struct amdgpu_ctx_mgr *mgr);
 void amdgpu_ctx_mgr_entity_fini(struct amdgpu_ctx_mgr *mgr);
-void amdgpu_ctx_mgr_entity_flush(struct amdgpu_ctx_mgr *mgr);
+long amdgpu_ctx_mgr_entity_flush(struct amdgpu_ctx_mgr *mgr, long timeout);
 void amdgpu_ctx_mgr_fini(struct amdgpu_ctx_mgr *mgr);
 
 #endif

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

@@ -568,10 +568,9 @@ static ssize_t amdgpu_debugfs_sensor_read(struct file *f, char __user *buf,
 	idx = *pos >> 2;
 
 	valuesize = sizeof(values);
-	if (adev->powerplay.pp_funcs && adev->powerplay.pp_funcs->read_sensor)
-		r = amdgpu_dpm_read_sensor(adev, idx, &values[0], &valuesize);
-	else
-		return -EINVAL;
+	r = amdgpu_dpm_read_sensor(adev, idx, &values[0], &valuesize);
+	if (r)
+		return r;
 
 	if (size > valuesize)
 		return -EINVAL;