Merge branch 'drm-next-4.14' of git://people.freedesktop.org/~agd5f/l…

…inux into drm-next

- Stop reprogramming the MC, the vbios already does this in asic_init
- Reduce internal gart to 256M (this does not affect the ttm GTT pool size)
- Initial support for huge pages
- Rework bo migration logic
- Lots of improvements for vega10
- Powerplay fixes
- Additional Raven enablement
- SR-IOV improvements
- Bug fixes
- Code cleanup

* 'drm-next-4.14' of git://people.freedesktop.org/~agd5f/linux: (138 commits)
  drm/amdgpu: fix header on gfx9 clear state
  drm/amdgpu: reduce the time of reading VBIOS
  drm/amdgpu/virtual_dce: Remove the rmmod error message
  drm/amdgpu/gmc9: disable legacy vga features in gmc init
  drm/amdgpu/gmc8: disable legacy vga features in gmc init
  drm/amdgpu/gmc7: disable legacy vga features in gmc init
  drm/amdgpu/gmc6: disable legacy vga features in gmc init (v2)
  drm/radeon: Set depth on low mem to 16 bpp instead of 8 bpp
  drm/amdgpu: fix the incorrect scratch reg number on gfx v6
  drm/amdgpu: fix the incorrect scratch reg number on gfx v7
  drm/amdgpu: fix the incorrect scratch reg number on gfx v8
  drm/amdgpu: fix the incorrect scratch reg number on gfx v9
  drm/amd/powerplay: add support for 3DP 4K@120Hz on vega10.
  drm/amdgpu: enable huge page handling in the VM v5
  drm/amdgpu: increase fragmentation size for Vega10 v2
  drm/amdgpu: ttm_bind only when user needs gpu_addr in bo pin
  drm/amdgpu: correct clock info for SRIOV
  drm/amdgpu/gmc8: SRIOV need to program fb location
  drm/amdgpu: disable firmware loading for psp v10
  drm/amdgpu:fix gfx fence allocate size
  ...

drivers/gpu/drm/amd/amdgpu/Makefile

@@ -25,7 +25,7 @@ amdgpu-y += amdgpu_device.o amdgpu_kms.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_queue_mgr.o
+	amdgpu_queue_mgr.o amdgpu_vf_error.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

@@ -68,13 +68,16 @@
 
 #include "gpu_scheduler.h"
 #include "amdgpu_virt.h"
+#include "amdgpu_gart.h"
 
 /*
  * Modules parameters.
  */
 extern int amdgpu_modeset;
 extern int amdgpu_vram_limit;
-extern int amdgpu_gart_size;
+extern int amdgpu_vis_vram_limit;
+extern unsigned amdgpu_gart_size;
+extern int amdgpu_gtt_size;
 extern int amdgpu_moverate;
 extern int amdgpu_benchmarking;
 extern int amdgpu_testing;
@@ -104,6 +107,7 @@ extern unsigned amdgpu_pcie_gen_cap;
 extern unsigned amdgpu_pcie_lane_cap;
 extern unsigned amdgpu_cg_mask;
 extern unsigned amdgpu_pg_mask;
+extern unsigned amdgpu_sdma_phase_quantum;
 extern char *amdgpu_disable_cu;
 extern char *amdgpu_virtual_display;
 extern unsigned amdgpu_pp_feature_mask;
@@ -531,49 +535,6 @@ int amdgpu_mode_dumb_mmap(struct drm_file *filp,
 int amdgpu_fence_slab_init(void);
 void amdgpu_fence_slab_fini(void);
 
-/*
- * GART structures, functions & helpers
- */
-struct amdgpu_mc;
-
-#define AMDGPU_GPU_PAGE_SIZE 4096
-#define AMDGPU_GPU_PAGE_MASK (AMDGPU_GPU_PAGE_SIZE - 1)
-#define AMDGPU_GPU_PAGE_SHIFT 12
-#define AMDGPU_GPU_PAGE_ALIGN(a) (((a) + AMDGPU_GPU_PAGE_MASK) & ~AMDGPU_GPU_PAGE_MASK)
-
-struct amdgpu_gart {
-	dma_addr_t			table_addr;
-	struct amdgpu_bo		*robj;
-	void				*ptr;
-	unsigned			num_gpu_pages;
-	unsigned			num_cpu_pages;
-	unsigned			table_size;
-#ifdef CONFIG_DRM_AMDGPU_GART_DEBUGFS
-	struct page			**pages;
-#endif
-	bool				ready;
-
-	/* Asic default pte flags */
-	uint64_t			gart_pte_flags;
-
-	const struct amdgpu_gart_funcs *gart_funcs;
-};
-
-int amdgpu_gart_table_ram_alloc(struct amdgpu_device *adev);
-void amdgpu_gart_table_ram_free(struct amdgpu_device *adev);
-int amdgpu_gart_table_vram_alloc(struct amdgpu_device *adev);
-void amdgpu_gart_table_vram_free(struct amdgpu_device *adev);
-int amdgpu_gart_table_vram_pin(struct amdgpu_device *adev);
-void amdgpu_gart_table_vram_unpin(struct amdgpu_device *adev);
-int amdgpu_gart_init(struct amdgpu_device *adev);
-void amdgpu_gart_fini(struct amdgpu_device *adev);
-int amdgpu_gart_unbind(struct amdgpu_device *adev, uint64_t offset,
-			int pages);
-int amdgpu_gart_bind(struct amdgpu_device *adev, uint64_t offset,
-		     int pages, struct page **pagelist,
-		     dma_addr_t *dma_addr, uint64_t flags);
-int amdgpu_ttm_recover_gart(struct amdgpu_device *adev);
-
 /*
  * VMHUB structures, functions & helpers
  */
@@ -598,22 +559,20 @@ struct amdgpu_mc {
 	 * about vram size near mc fb location */
 	u64			mc_vram_size;
 	u64			visible_vram_size;
-	u64			gtt_size;
-	u64			gtt_start;
-	u64			gtt_end;
+	u64			gart_size;
+	u64			gart_start;
+	u64			gart_end;
 	u64			vram_start;
 	u64			vram_end;
 	unsigned		vram_width;
 	u64			real_vram_size;
 	int			vram_mtrr;
-	u64                     gtt_base_align;
 	u64                     mc_mask;
 	const struct firmware   *fw;	/* MC firmware */
 	uint32_t                fw_version;
 	struct amdgpu_irq_src	vm_fault;
 	uint32_t		vram_type;
 	uint32_t                srbm_soft_reset;
-	struct amdgpu_mode_mc_save save;
 	bool			prt_warning;
 	uint64_t		stolen_size;
 	/* apertures */
@@ -1159,7 +1118,9 @@ struct amdgpu_cs_parser {
 	struct list_head		validated;
 	struct dma_fence		*fence;
 	uint64_t			bytes_moved_threshold;
+	uint64_t			bytes_moved_vis_threshold;
 	uint64_t			bytes_moved;
+	uint64_t			bytes_moved_vis;
 	struct amdgpu_bo_list_entry	*evictable;
 
 	/* user fence */
@@ -1231,7 +1192,9 @@ struct amdgpu_wb {
 int amdgpu_wb_get(struct amdgpu_device *adev, u32 *wb);
 void amdgpu_wb_free(struct amdgpu_device *adev, u32 wb);
 int amdgpu_wb_get_64bit(struct amdgpu_device *adev, u32 *wb);
+int amdgpu_wb_get_256Bit(struct amdgpu_device *adev, u32 *wb);
 void amdgpu_wb_free_64bit(struct amdgpu_device *adev, u32 wb);
+void amdgpu_wb_free_256bit(struct amdgpu_device *adev, u32 wb);
 
 void amdgpu_get_pcie_info(struct amdgpu_device *adev);
 
@@ -1557,6 +1520,10 @@ struct amdgpu_device {
 	spinlock_t gc_cac_idx_lock;
 	amdgpu_rreg_t			gc_cac_rreg;
 	amdgpu_wreg_t			gc_cac_wreg;
+	/* protects concurrent se_cac register access */
+	spinlock_t se_cac_idx_lock;
+	amdgpu_rreg_t			se_cac_rreg;
+	amdgpu_wreg_t			se_cac_wreg;
 	/* protects concurrent ENDPOINT (audio) register access */
 	spinlock_t audio_endpt_idx_lock;
 	amdgpu_block_rreg_t		audio_endpt_rreg;
@@ -1593,6 +1560,7 @@ struct amdgpu_device {
 		spinlock_t		lock;
 		s64			last_update_us;
 		s64			accum_us; /* accumulated microseconds */
+		s64			accum_us_vis; /* for visible VRAM */
 		u32			log2_max_MBps;
 	} mm_stats;
 
@@ -1687,6 +1655,8 @@ struct amdgpu_device {
 	bool has_hw_reset;
 	u8				reset_magic[AMDGPU_RESET_MAGIC_NUM];
 
+	/* record last mm index being written through WREG32*/
+	unsigned long last_mm_index;
 };
 
 static inline struct amdgpu_device *amdgpu_ttm_adev(struct ttm_bo_device *bdev)
@@ -1742,6 +1712,8 @@ void amdgpu_mm_wdoorbell64(struct amdgpu_device *adev, u32 index, u64 v);
 #define WREG32_DIDT(reg, v) adev->didt_wreg(adev, (reg), (v))
 #define RREG32_GC_CAC(reg) adev->gc_cac_rreg(adev, (reg))
 #define WREG32_GC_CAC(reg, v) adev->gc_cac_wreg(adev, (reg), (v))
+#define RREG32_SE_CAC(reg) adev->se_cac_rreg(adev, (reg))
+#define WREG32_SE_CAC(reg, v) adev->se_cac_wreg(adev, (reg), (v))
 #define RREG32_AUDIO_ENDPT(block, reg) adev->audio_endpt_rreg(adev, (block), (reg))
 #define WREG32_AUDIO_ENDPT(block, reg, v) adev->audio_endpt_wreg(adev, (block), (reg), (v))
 #define WREG32_P(reg, val, mask)				\
@@ -1792,50 +1764,6 @@ void amdgpu_mm_wdoorbell64(struct amdgpu_device *adev, u32 index, u64 v);
 #define RBIOS16(i) (RBIOS8(i) | (RBIOS8((i)+1) << 8))
 #define RBIOS32(i) ((RBIOS16(i)) | (RBIOS16((i)+2) << 16))
 
-/*
- * RING helpers.
- */
-static inline void amdgpu_ring_write(struct amdgpu_ring *ring, uint32_t v)
-{
-	if (ring->count_dw <= 0)
-		DRM_ERROR("amdgpu: writing more dwords to the ring than expected!\n");
-	ring->ring[ring->wptr++ & ring->buf_mask] = v;
-	ring->wptr &= ring->ptr_mask;
-	ring->count_dw--;
-}
-
-static inline void amdgpu_ring_write_multiple(struct amdgpu_ring *ring, void *src, int count_dw)
-{
-	unsigned occupied, chunk1, chunk2;
-	void *dst;
-
-	if (unlikely(ring->count_dw < count_dw)) {
-		DRM_ERROR("amdgpu: writing more dwords to the ring than expected!\n");
-		return;
-	}
-
-	occupied = ring->wptr & ring->buf_mask;
-	dst = (void *)&ring->ring[occupied];
-	chunk1 = ring->buf_mask + 1 - occupied;
-	chunk1 = (chunk1 >= count_dw) ? count_dw: chunk1;
-	chunk2 = count_dw - chunk1;
-	chunk1 <<= 2;
-	chunk2 <<= 2;
-
-	if (chunk1)
-		memcpy(dst, src, chunk1);
-
-	if (chunk2) {
-		src += chunk1;
-		dst = (void *)ring->ring;
-		memcpy(dst, src, chunk2);
-	}
-
-	ring->wptr += count_dw;
-	ring->wptr &= ring->ptr_mask;
-	ring->count_dw -= count_dw;
-}
-
 static inline struct amdgpu_sdma_instance *
 amdgpu_get_sdma_instance(struct amdgpu_ring *ring)
 {
@@ -1898,7 +1826,6 @@ amdgpu_get_sdma_instance(struct amdgpu_ring *ring)
 #define amdgpu_ih_get_wptr(adev) (adev)->irq.ih_funcs->get_wptr((adev))
 #define amdgpu_ih_decode_iv(adev, iv) (adev)->irq.ih_funcs->decode_iv((adev), (iv))
 #define amdgpu_ih_set_rptr(adev) (adev)->irq.ih_funcs->set_rptr((adev))
-#define amdgpu_display_set_vga_render_state(adev, r) (adev)->mode_info.funcs->set_vga_render_state((adev), (r))
 #define amdgpu_display_vblank_get_counter(adev, crtc) (adev)->mode_info.funcs->vblank_get_counter((adev), (crtc))
 #define amdgpu_display_vblank_wait(adev, crtc) (adev)->mode_info.funcs->vblank_wait((adev), (crtc))
 #define amdgpu_display_backlight_set_level(adev, e, l) (adev)->mode_info.funcs->backlight_set_level((e), (l))
@@ -1911,8 +1838,6 @@ amdgpu_get_sdma_instance(struct amdgpu_ring *ring)
 #define amdgpu_display_page_flip_get_scanoutpos(adev, crtc, vbl, pos) (adev)->mode_info.funcs->page_flip_get_scanoutpos((adev), (crtc), (vbl), (pos))
 #define amdgpu_display_add_encoder(adev, e, s, c) (adev)->mode_info.funcs->add_encoder((adev), (e), (s), (c))
 #define amdgpu_display_add_connector(adev, ci, sd, ct, ib, coi, h, r) (adev)->mode_info.funcs->add_connector((adev), (ci), (sd), (ct), (ib), (coi), (h), (r))
-#define amdgpu_display_stop_mc_access(adev, s) (adev)->mode_info.funcs->stop_mc_access((adev), (s))
-#define amdgpu_display_resume_mc_access(adev, s) (adev)->mode_info.funcs->resume_mc_access((adev), (s))
 #define amdgpu_emit_copy_buffer(adev, ib, s, d, b) (adev)->mman.buffer_funcs->emit_copy_buffer((ib),  (s), (d), (b))
 #define amdgpu_emit_fill_buffer(adev, ib, s, d, b) (adev)->mman.buffer_funcs->emit_fill_buffer((ib), (s), (d), (b))
 #define amdgpu_gfx_get_gpu_clock_counter(adev) (adev)->gfx.funcs->get_gpu_clock_counter((adev))
@@ -1927,7 +1852,8 @@ void amdgpu_pci_config_reset(struct amdgpu_device *adev);
 bool amdgpu_need_post(struct amdgpu_device *adev);
 void amdgpu_update_display_priority(struct amdgpu_device *adev);
 
-void amdgpu_cs_report_moved_bytes(struct amdgpu_device *adev, u64 num_bytes);
+void amdgpu_cs_report_moved_bytes(struct amdgpu_device *adev, u64 num_bytes,
+				  u64 num_vis_bytes);
 void amdgpu_ttm_placement_from_domain(struct amdgpu_bo *abo, u32 domain);
 bool amdgpu_ttm_bo_is_amdgpu_bo(struct ttm_buffer_object *bo);
 int amdgpu_ttm_tt_get_user_pages(struct ttm_tt *ttm, struct page **pages);
@@ -1943,7 +1869,7 @@ bool amdgpu_ttm_tt_is_readonly(struct ttm_tt *ttm);
 uint64_t amdgpu_ttm_tt_pte_flags(struct amdgpu_device *adev, struct ttm_tt *ttm,
 				 struct ttm_mem_reg *mem);
 void amdgpu_vram_location(struct amdgpu_device *adev, struct amdgpu_mc *mc, u64 base);
-void amdgpu_gtt_location(struct amdgpu_device *adev, struct amdgpu_mc *mc);
+void amdgpu_gart_location(struct amdgpu_device *adev, struct amdgpu_mc *mc);
 void amdgpu_ttm_set_active_vram_size(struct amdgpu_device *adev, u64 size);
 int amdgpu_ttm_init(struct amdgpu_device *adev);
 void amdgpu_ttm_fini(struct amdgpu_device *adev);

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

@@ -1686,7 +1686,7 @@ void amdgpu_atombios_scratch_regs_lock(struct amdgpu_device *adev, bool lock)
 {
 	uint32_t bios_6_scratch;
 
-	bios_6_scratch = RREG32(mmBIOS_SCRATCH_6);
+	bios_6_scratch = RREG32(adev->bios_scratch_reg_offset + 6);
 
 	if (lock) {
 		bios_6_scratch |= ATOM_S6_CRITICAL_STATE;
@@ -1696,15 +1696,17 @@ void amdgpu_atombios_scratch_regs_lock(struct amdgpu_device *adev, bool lock)
 		bios_6_scratch |= ATOM_S6_ACC_MODE;
 	}
 
-	WREG32(mmBIOS_SCRATCH_6, bios_6_scratch);
+	WREG32(adev->bios_scratch_reg_offset + 6, bios_6_scratch);
 }
 
 void amdgpu_atombios_scratch_regs_init(struct amdgpu_device *adev)
 {
 	uint32_t bios_2_scratch, bios_6_scratch;
 
-	bios_2_scratch = RREG32(mmBIOS_SCRATCH_2);
-	bios_6_scratch = RREG32(mmBIOS_SCRATCH_6);
+	adev->bios_scratch_reg_offset = mmBIOS_SCRATCH_0;
+
+	bios_2_scratch = RREG32(adev->bios_scratch_reg_offset + 2);
+	bios_6_scratch = RREG32(adev->bios_scratch_reg_offset + 6);
 
 	/* let the bios control the backlight */
 	bios_2_scratch &= ~ATOM_S2_VRI_BRIGHT_ENABLE;
@@ -1715,16 +1717,16 @@ void amdgpu_atombios_scratch_regs_init(struct amdgpu_device *adev)
 	/* clear the vbios dpms state */
 	bios_2_scratch &= ~ATOM_S2_DEVICE_DPMS_STATE;
 
-	WREG32(mmBIOS_SCRATCH_2, bios_2_scratch);
-	WREG32(mmBIOS_SCRATCH_6, bios_6_scratch);
+	WREG32(adev->bios_scratch_reg_offset + 2, bios_2_scratch);
+	WREG32(adev->bios_scratch_reg_offset + 6, bios_6_scratch);
 }
 
 void amdgpu_atombios_scratch_regs_save(struct amdgpu_device *adev)
 {
 	int i;
 
 	for (i = 0; i < AMDGPU_BIOS_NUM_SCRATCH; i++)
-		adev->bios_scratch[i] = RREG32(mmBIOS_SCRATCH_0 + i);
+		adev->bios_scratch[i] = RREG32(adev->bios_scratch_reg_offset + i);
 }
 
 void amdgpu_atombios_scratch_regs_restore(struct amdgpu_device *adev)
@@ -1738,20 +1740,30 @@ void amdgpu_atombios_scratch_regs_restore(struct amdgpu_device *adev)
 	adev->bios_scratch[7] &= ~ATOM_S7_ASIC_INIT_COMPLETE_MASK;
 
 	for (i = 0; i < AMDGPU_BIOS_NUM_SCRATCH; i++)
-		WREG32(mmBIOS_SCRATCH_0 + i, adev->bios_scratch[i]);
+		WREG32(adev->bios_scratch_reg_offset + i, adev->bios_scratch[i]);
 }
 
 void amdgpu_atombios_scratch_regs_engine_hung(struct amdgpu_device *adev,
 					      bool hung)
 {
-	u32 tmp = RREG32(mmBIOS_SCRATCH_3);
+	u32 tmp = RREG32(adev->bios_scratch_reg_offset + 3);
 
 	if (hung)
 		tmp |= ATOM_S3_ASIC_GUI_ENGINE_HUNG;
 	else
 		tmp &= ~ATOM_S3_ASIC_GUI_ENGINE_HUNG;
 
-	WREG32(mmBIOS_SCRATCH_3, tmp);
+	WREG32(adev->bios_scratch_reg_offset + 3, tmp);
+}
+
+bool amdgpu_atombios_scratch_need_asic_init(struct amdgpu_device *adev)
+{
+	u32 tmp = RREG32(adev->bios_scratch_reg_offset + 7);
+
+	if (tmp & ATOM_S7_ASIC_INIT_COMPLETE_MASK)
+		return false;
+	else
+		return true;
 }
 
 /* Atom needs data in little endian format

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

@@ -200,6 +200,7 @@ void amdgpu_atombios_scratch_regs_save(struct amdgpu_device *adev);
 void amdgpu_atombios_scratch_regs_restore(struct amdgpu_device *adev);
 void amdgpu_atombios_scratch_regs_engine_hung(struct amdgpu_device *adev,
 					      bool hung);
+bool amdgpu_atombios_scratch_need_asic_init(struct amdgpu_device *adev);
 
 void amdgpu_atombios_copy_swap(u8 *dst, u8 *src, u8 num_bytes, bool to_le);
 int amdgpu_atombios_get_max_vddc(struct amdgpu_device *adev, u8 voltage_type,