Merge branch 'drm-next-3.19-wip' of git://people.freedesktop.org/~agd5f/linux into drm-next

- More cursor and hotspot handling fixes
- Fix some typos in the new smc fan control code and enable on CI
- VM and CS cleanups

* 'drm-next-3.19-wip' of git://people.freedesktop.org/~agd5f/linux:
  drm/radeon: enable smc fan control on CI
  drm/radeon: use pointers instead of indexes for CS chunks
  drm/radeon: remove duplicates check
  drm/ttm: optionally move duplicates to a separate list
  drm/radeon: check the right ring in radeon_evict_flags()
  drm/radeon: fix copy paste typos in fan control for si/ci
  drm/radeon: Hide cursor on CRTCs used by fbdev (v2)
  drm/radeon: add spinlock for BO_VA status protection (v2)
  drm/radeon: fence PT updates as shared
  drm/radeon: rename radeon_cs_reloc to radeon_bo_list
  drm/radeon: drop the handle from radeon_cs_reloc
  drm/radeon drop gobj from radeon_cs_reloc
  drm/radeon: fix typo in new fan control registers for SI/CI
  drm/radeon: sync all BOs involved in a CS
  drm/radeon: Move hotspot handling out of radeon_set_cursor
  drm/radeon: Re-show the cursor after a modeset

drivers/gpu/drm/qxl/qxl_release.c

@@ -264,7 +264,8 @@ int qxl_release_reserve_list(struct qxl_release *release, bool no_intr)
 	if (list_is_singular(&release->bos))
 		return 0;
 
-	ret = ttm_eu_reserve_buffers(&release->ticket, &release->bos, !no_intr);
+	ret = ttm_eu_reserve_buffers(&release->ticket, &release->bos,
+				     !no_intr, NULL);
 	if (ret)
 		return ret;
 

drivers/gpu/drm/radeon/atombios_crtc.c

@@ -2039,6 +2039,7 @@ int atombios_crtc_mode_set(struct drm_crtc *crtc,
 	atombios_crtc_set_base(crtc, x, y, old_fb);
 	atombios_overscan_setup(crtc, mode, adjusted_mode);
 	atombios_scaler_setup(crtc);
+	radeon_cursor_reset(crtc);
 	/* update the hw version fpr dpm */
 	radeon_crtc->hw_mode = *adjusted_mode;
 

drivers/gpu/drm/radeon/ci_dpm.c

@@ -937,7 +937,7 @@ static void ci_fan_ctrl_set_static_mode(struct radeon_device *rdev, u32 mode)
 	tmp |= TMIN(0);
 	WREG32_SMC(CG_FDO_CTRL2, tmp);
 
-	tmp = RREG32_SMC(CG_FDO_CTRL2) & FDO_PWM_MODE_MASK;
+	tmp = RREG32_SMC(CG_FDO_CTRL2) & ~FDO_PWM_MODE_MASK;
 	tmp |= FDO_PWM_MODE(mode);
 	WREG32_SMC(CG_FDO_CTRL2, tmp);
 }
@@ -1162,7 +1162,7 @@ static int ci_fan_ctrl_set_fan_speed_rpm(struct radeon_device *rdev,
 	tmp |= TARGET_PERIOD(tach_period);
 	WREG32_SMC(CG_TACH_CTRL, tmp);
 
-	ci_fan_ctrl_set_static_mode(rdev, FDO_PWM_MODE_STATIC);
+	ci_fan_ctrl_set_static_mode(rdev, FDO_PWM_MODE_STATIC_RPM);
 
 	return 0;
 }
@@ -1178,7 +1178,7 @@ static void ci_fan_ctrl_set_default_mode(struct radeon_device *rdev)
 		tmp |= FDO_PWM_MODE(pi->fan_ctrl_default_mode);
 		WREG32_SMC(CG_FDO_CTRL2, tmp);
 
-		tmp = RREG32_SMC(CG_FDO_CTRL2) & TMIN_MASK;
+		tmp = RREG32_SMC(CG_FDO_CTRL2) & ~TMIN_MASK;
 		tmp |= TMIN(pi->t_min);
 		WREG32_SMC(CG_FDO_CTRL2, tmp);
 		pi->fan_ctrl_is_in_default_mode = true;
@@ -5849,7 +5849,6 @@ int ci_dpm_init(struct radeon_device *rdev)
 			rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac;
 
 	pi->fan_ctrl_is_in_default_mode = true;
-	rdev->pm.dpm.fan.ucode_fan_control = false;
 
 	return 0;
 }

drivers/gpu/drm/radeon/cikd.h

@@ -213,18 +213,18 @@
 
 #define	CG_FDO_CTRL0					0xC0300064
 #define		FDO_STATIC_DUTY(x)			((x) << 0)
-#define		FDO_STATIC_DUTY_MASK			0x0000000F
+#define		FDO_STATIC_DUTY_MASK			0x000000FF
 #define		FDO_STATIC_DUTY_SHIFT			0
 #define	CG_FDO_CTRL1					0xC0300068
 #define		FMAX_DUTY100(x)				((x) << 0)
-#define		FMAX_DUTY100_MASK			0x0000000F
+#define		FMAX_DUTY100_MASK			0x000000FF
 #define		FMAX_DUTY100_SHIFT			0
 #define	CG_FDO_CTRL2					0xC030006C
 #define		TMIN(x)					((x) << 0)
-#define		TMIN_MASK				0x0000000F
+#define		TMIN_MASK				0x000000FF
 #define		TMIN_SHIFT				0
 #define		FDO_PWM_MODE(x)				((x) << 11)
-#define		FDO_PWM_MODE_MASK			(3 << 11)
+#define		FDO_PWM_MODE_MASK			(7 << 11)
 #define		FDO_PWM_MODE_SHIFT			11
 #define		TACH_PWM_RESP_RATE(x)			((x) << 25)
 #define		TACH_PWM_RESP_RATE_MASK			(0x7f << 25)

drivers/gpu/drm/radeon/evergreen_cs.c

@@ -35,7 +35,7 @@
 #define MIN(a,b)                   (((a)<(b))?(a):(b))
 
 int r600_dma_cs_next_reloc(struct radeon_cs_parser *p,
-			   struct radeon_cs_reloc **cs_reloc);
+			   struct radeon_bo_list **cs_reloc);
 struct evergreen_cs_track {
 	u32			group_size;
 	u32			nbanks;
@@ -1094,7 +1094,7 @@ static int evergreen_cs_parse_packet0(struct radeon_cs_parser *p,
 static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 {
 	struct evergreen_cs_track *track = (struct evergreen_cs_track *)p->track;
-	struct radeon_cs_reloc *reloc;
+	struct radeon_bo_list *reloc;
 	u32 last_reg;
 	u32 m, i, tmp, *ib;
 	int r;
@@ -1792,7 +1792,7 @@ static bool evergreen_is_safe_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 static int evergreen_packet3_check(struct radeon_cs_parser *p,
 				   struct radeon_cs_packet *pkt)
 {
-	struct radeon_cs_reloc *reloc;
+	struct radeon_bo_list *reloc;
 	struct evergreen_cs_track *track;
 	volatile u32 *ib;
 	unsigned idx;
@@ -2661,7 +2661,7 @@ int evergreen_cs_parse(struct radeon_cs_parser *p)
 			p->track = NULL;
 			return r;
 		}
-	} while (p->idx < p->chunks[p->chunk_ib_idx].length_dw);
+	} while (p->idx < p->chunk_ib->length_dw);
 #if 0
 	for (r = 0; r < p->ib.length_dw; r++) {
 		printk(KERN_INFO "%05d  0x%08X\n", r, p->ib.ptr[r]);
@@ -2684,8 +2684,8 @@ int evergreen_cs_parse(struct radeon_cs_parser *p)
  **/
 int evergreen_dma_cs_parse(struct radeon_cs_parser *p)
 {
-	struct radeon_cs_chunk *ib_chunk = &p->chunks[p->chunk_ib_idx];
-	struct radeon_cs_reloc *src_reloc, *dst_reloc, *dst2_reloc;
+	struct radeon_cs_chunk *ib_chunk = p->chunk_ib;
+	struct radeon_bo_list *src_reloc, *dst_reloc, *dst2_reloc;
 	u32 header, cmd, count, sub_cmd;
 	volatile u32 *ib = p->ib.ptr;
 	u32 idx;
@@ -3100,7 +3100,7 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p)
 			DRM_ERROR("Unknown packet type %d at %d !\n", cmd, idx);
 			return -EINVAL;
 		}
-	} while (p->idx < p->chunks[p->chunk_ib_idx].length_dw);
+	} while (p->idx < p->chunk_ib->length_dw);
 #if 0
 	for (r = 0; r < p->ib->length_dw; r++) {
 		printk(KERN_INFO "%05d  0x%08X\n", r, p->ib.ptr[r]);

drivers/gpu/drm/radeon/r100.c

@@ -1254,7 +1254,7 @@ int r100_reloc_pitch_offset(struct radeon_cs_parser *p,
 	int r;
 	u32 tile_flags = 0;
 	u32 tmp;
-	struct radeon_cs_reloc *reloc;
+	struct radeon_bo_list *reloc;
 	u32 value;
 
 	r = radeon_cs_packet_next_reloc(p, &reloc, 0);
@@ -1293,7 +1293,7 @@ int r100_packet3_load_vbpntr(struct radeon_cs_parser *p,
 			     int idx)
 {
 	unsigned c, i;
-	struct radeon_cs_reloc *reloc;
+	struct radeon_bo_list *reloc;
 	struct r100_cs_track *track;
 	int r = 0;
 	volatile uint32_t *ib;
@@ -1542,7 +1542,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
 			      struct radeon_cs_packet *pkt,
 			      unsigned idx, unsigned reg)
 {
-	struct radeon_cs_reloc *reloc;
+	struct radeon_bo_list *reloc;
 	struct r100_cs_track *track;
 	volatile uint32_t *ib;
 	uint32_t tmp;
@@ -1901,7 +1901,7 @@ int r100_cs_track_check_pkt3_indx_buffer(struct radeon_cs_parser *p,
 static int r100_packet3_check(struct radeon_cs_parser *p,
 			      struct radeon_cs_packet *pkt)
 {
-	struct radeon_cs_reloc *reloc;
+	struct radeon_bo_list *reloc;
 	struct r100_cs_track *track;
 	unsigned idx;
 	volatile uint32_t *ib;
@@ -2061,7 +2061,7 @@ int r100_cs_parse(struct radeon_cs_parser *p)
 		}
 		if (r)
 			return r;
-	} while (p->idx < p->chunks[p->chunk_ib_idx].length_dw);
+	} while (p->idx < p->chunk_ib->length_dw);
 	return 0;
 }
 

drivers/gpu/drm/radeon/r200.c

@@ -146,7 +146,7 @@ int r200_packet0_check(struct radeon_cs_parser *p,
 		       struct radeon_cs_packet *pkt,
 		       unsigned idx, unsigned reg)
 {
-	struct radeon_cs_reloc *reloc;
+	struct radeon_bo_list *reloc;
 	struct r100_cs_track *track;
 	volatile uint32_t *ib;
 	uint32_t tmp;

drivers/gpu/drm/radeon/r300.c

@@ -598,7 +598,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
 		struct radeon_cs_packet *pkt,
 		unsigned idx, unsigned reg)
 {
-	struct radeon_cs_reloc *reloc;
+	struct radeon_bo_list *reloc;
 	struct r100_cs_track *track;
 	volatile uint32_t *ib;
 	uint32_t tmp, tile_flags = 0;
@@ -1142,7 +1142,7 @@ fail:
 static int r300_packet3_check(struct radeon_cs_parser *p,
 			      struct radeon_cs_packet *pkt)
 {
-	struct radeon_cs_reloc *reloc;
+	struct radeon_bo_list *reloc;
 	struct r100_cs_track *track;
 	volatile uint32_t *ib;
 	unsigned idx;
@@ -1283,7 +1283,7 @@ int r300_cs_parse(struct radeon_cs_parser *p)
 		if (r) {
 			return r;
 		}
-	} while (p->idx < p->chunks[p->chunk_ib_idx].length_dw);
+	} while (p->idx < p->chunk_ib->length_dw);
 	return 0;
 }
 

drivers/gpu/drm/radeon/r600_cs.c

@@ -969,7 +969,7 @@ static int r600_cs_parse_packet0(struct radeon_cs_parser *p,
 static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 {
 	struct r600_cs_track *track = (struct r600_cs_track *)p->track;
-	struct radeon_cs_reloc *reloc;
+	struct radeon_bo_list *reloc;
 	u32 m, i, tmp, *ib;
 	int r;
 
@@ -1626,7 +1626,7 @@ static bool r600_is_safe_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 static int r600_packet3_check(struct radeon_cs_parser *p,
 				struct radeon_cs_packet *pkt)
 {
-	struct radeon_cs_reloc *reloc;
+	struct radeon_bo_list *reloc;
 	struct r600_cs_track *track;
 	volatile u32 *ib;
 	unsigned idx;
@@ -2316,7 +2316,7 @@ int r600_cs_parse(struct radeon_cs_parser *p)
 			p->track = NULL;
 			return r;
 		}
-	} while (p->idx < p->chunks[p->chunk_ib_idx].length_dw);
+	} while (p->idx < p->chunk_ib->length_dw);
 #if 0
 	for (r = 0; r < p->ib.length_dw; r++) {
 		printk(KERN_INFO "%05d  0x%08X\n", r, p->ib.ptr[r]);
@@ -2351,10 +2351,10 @@ static void r600_cs_parser_fini(struct radeon_cs_parser *parser, int error)
 
 static int r600_cs_parser_relocs_legacy(struct radeon_cs_parser *p)
 {
-	if (p->chunk_relocs_idx == -1) {
+	if (p->chunk_relocs == NULL) {
 		return 0;
 	}
-	p->relocs = kzalloc(sizeof(struct radeon_cs_reloc), GFP_KERNEL);
+	p->relocs = kzalloc(sizeof(struct radeon_bo_list), GFP_KERNEL);
 	if (p->relocs == NULL) {
 		return -ENOMEM;
 	}
@@ -2398,7 +2398,7 @@ int r600_cs_legacy(struct drm_device *dev, void *data, struct drm_file *filp,
 	/* Copy the packet into the IB, the parser will read from the
 	 * input memory (cached) and write to the IB (which can be
 	 * uncached). */
-	ib_chunk = &parser.chunks[parser.chunk_ib_idx];
+	ib_chunk = parser.chunk_ib;
 	parser.ib.length_dw = ib_chunk->length_dw;
 	*l = parser.ib.length_dw;
 	if (copy_from_user(ib, ib_chunk->user_ptr, ib_chunk->length_dw * 4)) {
@@ -2435,24 +2435,24 @@ void r600_cs_legacy_init(void)
  * GPU offset using the provided start.
  **/
 int r600_dma_cs_next_reloc(struct radeon_cs_parser *p,
-			   struct radeon_cs_reloc **cs_reloc)
+			   struct radeon_bo_list **cs_reloc)
 {
 	struct radeon_cs_chunk *relocs_chunk;
 	unsigned idx;
 
 	*cs_reloc = NULL;
-	if (p->chunk_relocs_idx == -1) {
+	if (p->chunk_relocs == NULL) {
 		DRM_ERROR("No relocation chunk !\n");
 		return -EINVAL;
 	}
-	relocs_chunk = &p->chunks[p->chunk_relocs_idx];
+	relocs_chunk = p->chunk_relocs;
 	idx = p->dma_reloc_idx;
 	if (idx >= p->nrelocs) {
 		DRM_ERROR("Relocs at %d after relocations chunk end %d !\n",
 			  idx, p->nrelocs);
 		return -EINVAL;
 	}
-	*cs_reloc = p->relocs_ptr[idx];
+	*cs_reloc = &p->relocs[idx];
 	p->dma_reloc_idx++;
 	return 0;
 }
@@ -2472,8 +2472,8 @@ int r600_dma_cs_next_reloc(struct radeon_cs_parser *p,
  **/
 int r600_dma_cs_parse(struct radeon_cs_parser *p)
 {
-	struct radeon_cs_chunk *ib_chunk = &p->chunks[p->chunk_ib_idx];
-	struct radeon_cs_reloc *src_reloc, *dst_reloc;
+	struct radeon_cs_chunk *ib_chunk = p->chunk_ib;
+	struct radeon_bo_list *src_reloc, *dst_reloc;
 	u32 header, cmd, count, tiled;
 	volatile u32 *ib = p->ib.ptr;
 	u32 idx, idx_value;
@@ -2619,7 +2619,7 @@ int r600_dma_cs_parse(struct radeon_cs_parser *p)
 			DRM_ERROR("Unknown packet type %d at %d !\n", cmd, idx);
 			return -EINVAL;
 		}
-	} while (p->idx < p->chunks[p->chunk_ib_idx].length_dw);
+	} while (p->idx < p->chunk_ib->length_dw);
 #if 0
 	for (r = 0; r < p->ib->length_dw; r++) {
 		printk(KERN_INFO "%05d  0x%08X\n", r, p->ib.ptr[r]);

drivers/gpu/drm/radeon/radeon.h

@@ -450,6 +450,15 @@ struct radeon_mman {
 #endif
 };
 
+struct radeon_bo_list {
+	struct radeon_bo		*robj;
+	struct ttm_validate_buffer	tv;
+	uint64_t			gpu_offset;
+	unsigned			prefered_domains;
+	unsigned			allowed_domains;
+	uint32_t			tiling_flags;
+};
+
 /* bo virtual address in a specific vm */
 struct radeon_bo_va {
 	/* protected by bo being reserved */
@@ -920,6 +929,9 @@ struct radeon_vm {
 
 	struct rb_root		va;
 
+	/* protecting invalidated and freed */
+	spinlock_t		status_lock;
+
 	/* BOs moved, but not yet updated in the PT */
 	struct list_head	invalidated;
 
@@ -1044,19 +1056,7 @@ void cayman_dma_fini(struct radeon_device *rdev);
 /*
  * CS.
  */
-struct radeon_cs_reloc {
-	struct drm_gem_object		*gobj;
-	struct radeon_bo		*robj;
-	struct ttm_validate_buffer	tv;
-	uint64_t			gpu_offset;
-	unsigned			prefered_domains;
-	unsigned			allowed_domains;
-	uint32_t			tiling_flags;
-	uint32_t			handle;
-};
-
 struct radeon_cs_chunk {
-	uint32_t		chunk_id;
 	uint32_t		length_dw;
 	uint32_t		*kdata;
 	void __user		*user_ptr;
@@ -1074,16 +1074,15 @@ struct radeon_cs_parser {
 	unsigned		idx;
 	/* relocations */
 	unsigned		nrelocs;
-	struct radeon_cs_reloc	*relocs;
-	struct radeon_cs_reloc	**relocs_ptr;
-	struct radeon_cs_reloc	*vm_bos;
+	struct radeon_bo_list	*relocs;
+	struct radeon_bo_list	*vm_bos;
 	struct list_head	validated;
 	unsigned		dma_reloc_idx;
 	/* indices of various chunks */
-	int			chunk_ib_idx;
-	int			chunk_relocs_idx;
-	int			chunk_flags_idx;
-	int			chunk_const_ib_idx;
+	struct radeon_cs_chunk  *chunk_ib;
+	struct radeon_cs_chunk  *chunk_relocs;
+	struct radeon_cs_chunk  *chunk_flags;
+	struct radeon_cs_chunk  *chunk_const_ib;
 	struct radeon_ib	ib;
 	struct radeon_ib	const_ib;
 	void			*track;
@@ -1097,7 +1096,7 @@ struct radeon_cs_parser {
 
 static inline u32 radeon_get_ib_value(struct radeon_cs_parser *p, int idx)
 {
-	struct radeon_cs_chunk *ibc = &p->chunks[p->chunk_ib_idx];
+	struct radeon_cs_chunk *ibc = p->chunk_ib;
 
 	if (ibc->kdata)
 		return ibc->kdata[idx];
@@ -2975,7 +2974,7 @@ int radeon_vm_manager_init(struct radeon_device *rdev);
 void radeon_vm_manager_fini(struct radeon_device *rdev);
 int radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm);
 void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm);
-struct radeon_cs_reloc *radeon_vm_get_bos(struct radeon_device *rdev,
+struct radeon_bo_list *radeon_vm_get_bos(struct radeon_device *rdev,
 					  struct radeon_vm *vm,
                                           struct list_head *head);
 struct radeon_fence *radeon_vm_grab_id(struct radeon_device *rdev,
@@ -3089,7 +3088,7 @@ bool radeon_cs_packet_next_is_pkt3_nop(struct radeon_cs_parser *p);
 void radeon_cs_dump_packet(struct radeon_cs_parser *p,
 			   struct radeon_cs_packet *pkt);
 int radeon_cs_packet_next_reloc(struct radeon_cs_parser *p,
-				struct radeon_cs_reloc **cs_reloc,
+				struct radeon_bo_list **cs_reloc,
 				int nomm);
 int r600_cs_common_vline_parse(struct radeon_cs_parser *p,
 			       uint32_t *vline_start_end,

drivers/gpu/drm/radeon/radeon_cs.c

@@ -77,22 +77,18 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p)
 	struct drm_device *ddev = p->rdev->ddev;
 	struct radeon_cs_chunk *chunk;
 	struct radeon_cs_buckets buckets;
-	unsigned i, j;
-	bool duplicate, need_mmap_lock = false;
+	unsigned i;
+	bool need_mmap_lock = false;
 	int r;
 
-	if (p->chunk_relocs_idx == -1) {
+	if (p->chunk_relocs == NULL) {
 		return 0;
 	}
-	chunk = &p->chunks[p->chunk_relocs_idx];
+	chunk = p->chunk_relocs;
 	p->dma_reloc_idx = 0;
 	/* FIXME: we assume that each relocs use 4 dwords */
 	p->nrelocs = chunk->length_dw / 4;
-	p->relocs_ptr = kcalloc(p->nrelocs, sizeof(void *), GFP_KERNEL);
-	if (p->relocs_ptr == NULL) {
-		return -ENOMEM;
-	}
-	p->relocs = kcalloc(p->nrelocs, sizeof(struct radeon_cs_reloc), GFP_KERNEL);
+	p->relocs = kcalloc(p->nrelocs, sizeof(struct radeon_bo_list), GFP_KERNEL);
 	if (p->relocs == NULL) {
 		return -ENOMEM;
 	}
@@ -101,31 +97,17 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p)
 
 	for (i = 0; i < p->nrelocs; i++) {
 		struct drm_radeon_cs_reloc *r;
+		struct drm_gem_object *gobj;
 		unsigned priority;
 
-		duplicate = false;
 		r = (struct drm_radeon_cs_reloc *)&chunk->kdata[i*4];
-		for (j = 0; j < i; j++) {
-			if (r->handle == p->relocs[j].handle) {
-				p->relocs_ptr[i] = &p->relocs[j];
-				duplicate = true;
-				break;
-			}
-		}
-		if (duplicate) {
-			p->relocs[i].handle = 0;
-			continue;
-		}
-
-		p->relocs[i].gobj = drm_gem_object_lookup(ddev, p->filp,
-							  r->handle);
-		if (p->relocs[i].gobj == NULL) {
+		gobj = drm_gem_object_lookup(ddev, p->filp, r->handle);
+		if (gobj == NULL) {
 			DRM_ERROR("gem object lookup failed 0x%x\n",
 				  r->handle);
 			return -ENOENT;
 		}
-		p->relocs_ptr[i] = &p->relocs[i];
-		p->relocs[i].robj = gem_to_radeon_bo(p->relocs[i].gobj);
+		p->relocs[i].robj = gem_to_radeon_bo(gobj);
 
 		/* The userspace buffer priorities are from 0 to 15. A higher
 		 * number means the buffer is more important.
@@ -184,7 +166,6 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p)
 
 		p->relocs[i].tv.bo = &p->relocs[i].robj->tbo;
 		p->relocs[i].tv.shared = !r->write_domain;
-		p->relocs[i].handle = r->handle;
 
 		radeon_cs_buckets_add(&buckets, &p->relocs[i].tv.head,
 				      priority);
@@ -251,22 +232,20 @@ static int radeon_cs_get_ring(struct radeon_cs_parser *p, u32 ring, s32 priority
 
 static int radeon_cs_sync_rings(struct radeon_cs_parser *p)
 {
-	int i, r = 0;
+	struct radeon_bo_list *reloc;
+	int r;
 
-	for (i = 0; i < p->nrelocs; i++) {
+	list_for_each_entry(reloc, &p->validated, tv.head) {
 		struct reservation_object *resv;
 
-		if (!p->relocs[i].robj)
-			continue;
-
-		resv = p->relocs[i].robj->tbo.resv;
+		resv = reloc->robj->tbo.resv;
 		r = radeon_sync_resv(p->rdev, &p->ib.sync, resv,
-				     p->relocs[i].tv.shared);
+				     reloc->tv.shared);
 
 		if (r)
-			break;
+			return r;
 	}
-	return r;
+	return 0;
 }
 
 /* XXX: note that this is called from the legacy UMS CS ioctl as well */
@@ -286,10 +265,10 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
 	p->idx = 0;
 	p->ib.sa_bo = NULL;
 	p->const_ib.sa_bo = NULL;
-	p->chunk_ib_idx = -1;
-	p->chunk_relocs_idx = -1;
-	p->chunk_flags_idx = -1;
-	p->chunk_const_ib_idx = -1;
+	p->chunk_ib = NULL;
+	p->chunk_relocs = NULL;
+	p->chunk_flags = NULL;
+	p->chunk_const_ib = NULL;
 	p->chunks_array = kcalloc(cs->num_chunks, sizeof(uint64_t), GFP_KERNEL);
 	if (p->chunks_array == NULL) {
 		return -ENOMEM;
@@ -316,24 +295,23 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
 			return -EFAULT;
 		}
 		p->chunks[i].length_dw = user_chunk.length_dw;
-		p->chunks[i].chunk_id = user_chunk.chunk_id;
-		if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_RELOCS) {
-			p->chunk_relocs_idx = i;
+		if (user_chunk.chunk_id == RADEON_CHUNK_ID_RELOCS) {
+			p->chunk_relocs = &p->chunks[i];
 		}
-		if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_IB) {
-			p->chunk_ib_idx = i;
+		if (user_chunk.chunk_id == RADEON_CHUNK_ID_IB) {
+			p->chunk_ib = &p->chunks[i];
 			/* zero length IB isn't useful */
 			if (p->chunks[i].length_dw == 0)
 				return -EINVAL;
 		}
-		if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_CONST_IB) {
-			p->chunk_const_ib_idx = i;
+		if (user_chunk.chunk_id == RADEON_CHUNK_ID_CONST_IB) {
+			p->chunk_const_ib = &p->chunks[i];
 			/* zero length CONST IB isn't useful */
 			if (p->chunks[i].length_dw == 0)
 				return -EINVAL;
 		}
-		if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_FLAGS) {
-			p->chunk_flags_idx = i;
+		if (user_chunk.chunk_id == RADEON_CHUNK_ID_FLAGS) {
+			p->chunk_flags = &p->chunks[i];
 			/* zero length flags aren't useful */
 			if (p->chunks[i].length_dw == 0)
 				return -EINVAL;
@@ -342,10 +320,10 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
 		size = p->chunks[i].length_dw;
 		cdata = (void __user *)(unsigned long)user_chunk.chunk_data;
 		p->chunks[i].user_ptr = cdata;
-		if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_CONST_IB)
+		if (user_chunk.chunk_id == RADEON_CHUNK_ID_CONST_IB)
 			continue;
 
-		if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_IB) {
+		if (user_chunk.chunk_id == RADEON_CHUNK_ID_IB) {
 			if (!p->rdev || !(p->rdev->flags & RADEON_IS_AGP))
 				continue;
 		}
@@ -358,7 +336,7 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
 		if (copy_from_user(p->chunks[i].kdata, cdata, size)) {
 			return -EFAULT;
 		}
-		if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_FLAGS) {
+		if (user_chunk.chunk_id == RADEON_CHUNK_ID_FLAGS) {
 			p->cs_flags = p->chunks[i].kdata[0];
 			if (p->chunks[i].length_dw > 1)
 				ring = p->chunks[i].kdata[1];
@@ -399,8 +377,8 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
 static int cmp_size_smaller_first(void *priv, struct list_head *a,
 				  struct list_head *b)
 {
-	struct radeon_cs_reloc *la = list_entry(a, struct radeon_cs_reloc, tv.head);
-	struct radeon_cs_reloc *lb = list_entry(b, struct radeon_cs_reloc, tv.head);
+	struct radeon_bo_list *la = list_entry(a, struct radeon_bo_list, tv.head);
+	struct radeon_bo_list *lb = list_entry(b, struct radeon_bo_list, tv.head);
 
 	/* Sort A before B if A is smaller. */
 	return (int)la->robj->tbo.num_pages - (int)lb->robj->tbo.num_pages;
@@ -441,13 +419,15 @@ static void radeon_cs_parser_fini(struct radeon_cs_parser *parser, int error, bo
 
 	if (parser->relocs != NULL) {
 		for (i = 0; i < parser->nrelocs; i++) {
-			if (parser->relocs[i].gobj)
-				drm_gem_object_unreference_unlocked(parser->relocs[i].gobj);
+			struct radeon_bo *bo = parser->relocs[i].robj;
+			if (bo == NULL)
+				continue;
+
+			drm_gem_object_unreference_unlocked(&bo->gem_base);
 		}
 	}
 	kfree(parser->track);
 	kfree(parser->relocs);
-	kfree(parser->relocs_ptr);
 	drm_free_large(parser->vm_bos);
 	for (i = 0; i < parser->nchunks; i++)
 		drm_free_large(parser->chunks[i].kdata);
@@ -462,7 +442,7 @@ static int radeon_cs_ib_chunk(struct radeon_device *rdev,
 {
 	int r;
 
-	if (parser->chunk_ib_idx == -1)
+	if (parser->chunk_ib == NULL)
 		return 0;
 
 	if (parser->cs_flags & RADEON_CS_USE_VM)
@@ -505,9 +485,6 @@ static int radeon_bo_vm_update_pte(struct radeon_cs_parser *p,
 	if (r)
 		return r;
 
-	radeon_sync_resv(p->rdev, &p->ib.sync, vm->page_directory->tbo.resv,
-			 true);
-
 	r = radeon_vm_clear_freed(rdev, vm);
 	if (r)
 		return r;
@@ -525,10 +502,6 @@ static int radeon_bo_vm_update_pte(struct radeon_cs_parser *p,
 	for (i = 0; i < p->nrelocs; i++) {
 		struct radeon_bo *bo;
 
-		/* ignore duplicates */
-		if (p->relocs_ptr[i] != &p->relocs[i])
-			continue;
-
 		bo = p->relocs[i].robj;
 		bo_va = radeon_vm_bo_find(vm, bo);
 		if (bo_va == NULL) {
@@ -553,7 +526,7 @@ static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev,
 	struct radeon_vm *vm = &fpriv->vm;
 	int r;
 
-	if (parser->chunk_ib_idx == -1)
+	if (parser->chunk_ib == NULL)
 		return 0;
 	if ((parser->cs_flags & RADEON_CS_USE_VM) == 0)
 		return 0;
@@ -587,7 +560,7 @@ static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev,
 	}
 
 	if ((rdev->family >= CHIP_TAHITI) &&
-	    (parser->chunk_const_ib_idx != -1)) {
+	    (parser->chunk_const_ib != NULL)) {
 		r = radeon_ib_schedule(rdev, &parser->ib, &parser->const_ib, true);
 	} else {
 		r = radeon_ib_schedule(rdev, &parser->ib, NULL, true);
@@ -614,7 +587,7 @@ static int radeon_cs_ib_fill(struct radeon_device *rdev, struct radeon_cs_parser
 	struct radeon_vm *vm = NULL;
 	int r;
 
-	if (parser->chunk_ib_idx == -1)
+	if (parser->chunk_ib == NULL)
 		return 0;
 
 	if (parser->cs_flags & RADEON_CS_USE_VM) {
@@ -622,8 +595,8 @@ static int radeon_cs_ib_fill(struct radeon_device *rdev, struct radeon_cs_parser
 		vm = &fpriv->vm;
 
 		if ((rdev->family >= CHIP_TAHITI) &&
-		    (parser->chunk_const_ib_idx != -1)) {
-			ib_chunk = &parser->chunks[parser->chunk_const_ib_idx];
+		    (parser->chunk_const_ib != NULL)) {
+			ib_chunk = parser->chunk_const_ib;
 			if (ib_chunk->length_dw > RADEON_IB_VM_MAX_SIZE) {
 				DRM_ERROR("cs IB CONST too big: %d\n", ib_chunk->length_dw);
 				return -EINVAL;
@@ -642,13 +615,13 @@ static int radeon_cs_ib_fill(struct radeon_device *rdev, struct radeon_cs_parser
 				return -EFAULT;
 		}
 
-		ib_chunk = &parser->chunks[parser->chunk_ib_idx];
+		ib_chunk = parser->chunk_ib;
 		if (ib_chunk->length_dw > RADEON_IB_VM_MAX_SIZE) {
 			DRM_ERROR("cs IB too big: %d\n", ib_chunk->length_dw);
 			return -EINVAL;
 		}
 	}
-	ib_chunk = &parser->chunks[parser->chunk_ib_idx];
+	ib_chunk = parser->chunk_ib;
 
 	r =  radeon_ib_get(rdev, parser->ring, &parser->ib,
 			   vm, ib_chunk->length_dw * 4);
@@ -740,7 +713,7 @@ int radeon_cs_packet_parse(struct radeon_cs_parser *p,
 			   struct radeon_cs_packet *pkt,
 			   unsigned idx)
 {
-	struct radeon_cs_chunk *ib_chunk = &p->chunks[p->chunk_ib_idx];
+	struct radeon_cs_chunk *ib_chunk = p->chunk_ib;
 	struct radeon_device *rdev = p->rdev;
 	uint32_t header;
 
@@ -834,7 +807,7 @@ void radeon_cs_dump_packet(struct radeon_cs_parser *p,
  * GPU offset using the provided start.
  **/
 int radeon_cs_packet_next_reloc(struct radeon_cs_parser *p,
-				struct radeon_cs_reloc **cs_reloc,
+				struct radeon_bo_list **cs_reloc,
 				int nomm)
 {
 	struct radeon_cs_chunk *relocs_chunk;
@@ -842,12 +815,12 @@ int radeon_cs_packet_next_reloc(struct radeon_cs_parser *p,
 	unsigned idx;
 	int r;
 
-	if (p->chunk_relocs_idx == -1) {
+	if (p->chunk_relocs == NULL) {
 		DRM_ERROR("No relocation chunk !\n");
 		return -EINVAL;
 	}
 	*cs_reloc = NULL;
-	relocs_chunk = &p->chunks[p->chunk_relocs_idx];
+	relocs_chunk = p->chunk_relocs;
 	r = radeon_cs_packet_parse(p, &p3reloc, p->idx);
 	if (r)
 		return r;
@@ -873,6 +846,6 @@ int radeon_cs_packet_next_reloc(struct radeon_cs_parser *p,
 			(u64)relocs_chunk->kdata[idx + 3] << 32;
 		(*cs_reloc)->gpu_offset |= relocs_chunk->kdata[idx + 0];
 	} else
-		*cs_reloc = p->relocs_ptr[(idx / 4)];
+		*cs_reloc = &p->relocs[(idx / 4)];
 	return 0;
 }

drivers/gpu/drm/radeon/radeon_cursor.c

@@ -227,11 +227,24 @@ int radeon_crtc_cursor_move(struct drm_crtc *crtc,
 	return ret;
 }
 
-static void radeon_set_cursor(struct drm_crtc *crtc, struct drm_gem_object *obj,
-			      uint64_t gpu_addr, int hot_x, int hot_y)
+static int radeon_set_cursor(struct drm_crtc *crtc, struct drm_gem_object *obj)
 {
 	struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc);
 	struct radeon_device *rdev = crtc->dev->dev_private;
+	struct radeon_bo *robj = gem_to_radeon_bo(obj);
+	uint64_t gpu_addr;
+	int ret;
+
+	ret = radeon_bo_reserve(robj, false);
+	if (unlikely(ret != 0))
+		goto fail;
+	/* Only 27 bit offset for legacy cursor */
+	ret = radeon_bo_pin_restricted(robj, RADEON_GEM_DOMAIN_VRAM,
+				       ASIC_IS_AVIVO(rdev) ? 0 : 1 << 27,
+				       &gpu_addr);
+	radeon_bo_unreserve(robj);
+	if (ret)
+		goto fail;
 
 	if (ASIC_IS_DCE4(rdev)) {
 		WREG32(EVERGREEN_CUR_SURFACE_ADDRESS_HIGH + radeon_crtc->crtc_offset,
@@ -253,18 +266,12 @@ static void radeon_set_cursor(struct drm_crtc *crtc, struct drm_gem_object *obj,
 		WREG32(RADEON_CUR_OFFSET + radeon_crtc->crtc_offset, radeon_crtc->legacy_cursor_offset);
 	}
 
-	if (hot_x != radeon_crtc->cursor_hot_x ||
-	    hot_y != radeon_crtc->cursor_hot_y) {
-		int x, y;
-
-		x = radeon_crtc->cursor_x + radeon_crtc->cursor_hot_x - hot_x;
-		y = radeon_crtc->cursor_y + radeon_crtc->cursor_hot_y - hot_y;
+	return 0;
 
-		radeon_cursor_move_locked(crtc, x, y);
+fail:
+	drm_gem_object_unreference_unlocked(obj);
 
-		radeon_crtc->cursor_hot_x = hot_x;
-		radeon_crtc->cursor_hot_y = hot_y;
-	}
+	return ret;
 }
 
 int radeon_crtc_cursor_set2(struct drm_crtc *crtc,
@@ -276,10 +283,7 @@ int radeon_crtc_cursor_set2(struct drm_crtc *crtc,
 			    int32_t hot_y)
 {
 	struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc);
-	struct radeon_device *rdev = crtc->dev->dev_private;
 	struct drm_gem_object *obj;
-	struct radeon_bo *robj;
-	uint64_t gpu_addr;
 	int ret;
 
 	if (!handle) {
@@ -301,41 +305,76 @@ int radeon_crtc_cursor_set2(struct drm_crtc *crtc,
 		return -ENOENT;
 	}
 
-	robj = gem_to_radeon_bo(obj);
-	ret = radeon_bo_reserve(robj, false);
-	if (unlikely(ret != 0))
-		goto fail;
-	/* Only 27 bit offset for legacy cursor */
-	ret = radeon_bo_pin_restricted(robj, RADEON_GEM_DOMAIN_VRAM,
-				       ASIC_IS_AVIVO(rdev) ? 0 : 1 << 27,
-				       &gpu_addr);
-	radeon_bo_unreserve(robj);
-	if (ret)
-		goto fail;
-
 	radeon_crtc->cursor_width = width;
 	radeon_crtc->cursor_height = height;
 
 	radeon_lock_cursor(crtc, true);
-	radeon_set_cursor(crtc, obj, gpu_addr, hot_x, hot_y);
-	radeon_show_cursor(crtc);
+
+	if (hot_x != radeon_crtc->cursor_hot_x ||
+	    hot_y != radeon_crtc->cursor_hot_y) {
+		int x, y;
+
+		x = radeon_crtc->cursor_x + radeon_crtc->cursor_hot_x - hot_x;
+		y = radeon_crtc->cursor_y + radeon_crtc->cursor_hot_y - hot_y;
+
+		radeon_cursor_move_locked(crtc, x, y);
+
+		radeon_crtc->cursor_hot_x = hot_x;
+		radeon_crtc->cursor_hot_y = hot_y;
+	}
+
+	ret = radeon_set_cursor(crtc, obj);
+
+	if (ret)
+		DRM_ERROR("radeon_set_cursor returned %d, not changing cursor\n",
+			  ret);
+	else
+		radeon_show_cursor(crtc);
+
 	radeon_lock_cursor(crtc, false);
 
 unpin:
 	if (radeon_crtc->cursor_bo) {
-		robj = gem_to_radeon_bo(radeon_crtc->cursor_bo);
+		struct radeon_bo *robj = gem_to_radeon_bo(radeon_crtc->cursor_bo);
 		ret = radeon_bo_reserve(robj, false);
 		if (likely(ret == 0)) {
 			radeon_bo_unpin(robj);
 			radeon_bo_unreserve(robj);
 		}
-		drm_gem_object_unreference_unlocked(radeon_crtc->cursor_bo);
+		if (radeon_crtc->cursor_bo != obj)
+			drm_gem_object_unreference_unlocked(radeon_crtc->cursor_bo);
 	}
 
 	radeon_crtc->cursor_bo = obj;
 	return 0;
-fail:
-	drm_gem_object_unreference_unlocked(obj);
+}
 
-	return ret;
+/**
+ * radeon_cursor_reset - Re-set the current cursor, if any.
+ *
+ * @crtc: drm crtc
+ *
+ * If the CRTC passed in currently has a cursor assigned, this function
+ * makes sure it's visible.
+ */
+void radeon_cursor_reset(struct drm_crtc *crtc)
+{
+	struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc);
+	int ret;
+
+	if (radeon_crtc->cursor_bo) {
+		radeon_lock_cursor(crtc, true);
+
+		radeon_cursor_move_locked(crtc, radeon_crtc->cursor_x,
+					  radeon_crtc->cursor_y);
+
+		ret = radeon_set_cursor(crtc, radeon_crtc->cursor_bo);
+		if (ret)
+			DRM_ERROR("radeon_set_cursor returned %d, not showing "
+				  "cursor\n", ret);
+		else
+			radeon_show_cursor(crtc);
+
+		radeon_lock_cursor(crtc, false);
+	}
 }

drivers/gpu/drm/radeon/radeon_fb.c

@@ -48,10 +48,40 @@ struct radeon_fbdev {
 	struct radeon_device *rdev;
 };
 
+/**
+ * radeon_fb_helper_set_par - Hide cursor on CRTCs used by fbdev.
+ *
+ * @info: fbdev info
+ *
+ * This function hides the cursor on all CRTCs used by fbdev.
+ */
+static int radeon_fb_helper_set_par(struct fb_info *info)
+{
+	int ret;
+
+	ret = drm_fb_helper_set_par(info);
+
+	/* XXX: with universal plane support fbdev will automatically disable
+	 * all non-primary planes (including the cursor)
+	 */
+	if (ret == 0) {
+		struct drm_fb_helper *fb_helper = info->par;
+		int i;
+
+		for (i = 0; i < fb_helper->crtc_count; i++) {
+			struct drm_crtc *crtc = fb_helper->crtc_info[i].mode_set.crtc;
+
+			radeon_crtc_cursor_set2(crtc, NULL, 0, 0, 0, 0, 0);
+		}
+	}
+
+	return ret;
+}
+
 static struct fb_ops radeonfb_ops = {
 	.owner = THIS_MODULE,
 	.fb_check_var = drm_fb_helper_check_var,
-	.fb_set_par = drm_fb_helper_set_par,
+	.fb_set_par = radeon_fb_helper_set_par,
 	.fb_fillrect = cfb_fillrect,
 	.fb_copyarea = cfb_copyarea,
 	.fb_imageblit = cfb_imageblit,

drivers/gpu/drm/radeon/radeon_gem.c

@@ -548,7 +548,7 @@ static void radeon_gem_va_update_vm(struct radeon_device *rdev,
 				    struct radeon_bo_va *bo_va)
 {
 	struct ttm_validate_buffer tv, *entry;
-	struct radeon_cs_reloc *vm_bos;
+	struct radeon_bo_list *vm_bos;
 	struct ww_acquire_ctx ticket;
 	struct list_head list;
 	unsigned domain;
@@ -564,7 +564,7 @@ static void radeon_gem_va_update_vm(struct radeon_device *rdev,
 	if (!vm_bos)
 		return;
 
-	r = ttm_eu_reserve_buffers(&ticket, &list, true);
+	r = ttm_eu_reserve_buffers(&ticket, &list, true, NULL);
 	if (r)
 		goto error_free;
 

drivers/gpu/drm/radeon/radeon_legacy_crtc.c

@@ -1054,6 +1054,7 @@ static int radeon_crtc_mode_set(struct drm_crtc *crtc,
 			DRM_ERROR("Mode need scaling but only first crtc can do that.\n");
 		}
 	}
+	radeon_cursor_reset(crtc);
 	return 0;
 }
 

drivers/gpu/drm/radeon/radeon_mode.h

@@ -818,6 +818,7 @@ extern int radeon_crtc_cursor_set2(struct drm_crtc *crtc,
 				   int32_t hot_y);
 extern int radeon_crtc_cursor_move(struct drm_crtc *crtc,
 				   int x, int y);
+extern void radeon_cursor_reset(struct drm_crtc *crtc);
 
 extern int radeon_get_crtc_scanoutpos(struct drm_device *dev, int crtc,
 				      unsigned int flags,

drivers/gpu/drm/radeon/radeon_object.c

@@ -502,19 +502,20 @@ int radeon_bo_list_validate(struct radeon_device *rdev,
 			    struct ww_acquire_ctx *ticket,
 			    struct list_head *head, int ring)
 {
-	struct radeon_cs_reloc *lobj;
-	struct radeon_bo *bo;
+	struct radeon_bo_list *lobj;
+	struct list_head duplicates;
 	int r;
 	u64 bytes_moved = 0, initial_bytes_moved;
 	u64 bytes_moved_threshold = radeon_bo_get_threshold_for_moves(rdev);
 
-	r = ttm_eu_reserve_buffers(ticket, head, true);
+	INIT_LIST_HEAD(&duplicates);
+	r = ttm_eu_reserve_buffers(ticket, head, true, &duplicates);
 	if (unlikely(r != 0)) {
 		return r;
 	}
 
 	list_for_each_entry(lobj, head, tv.head) {
-		bo = lobj->robj;
+		struct radeon_bo *bo = lobj->robj;
 		if (!bo->pin_count) {
 			u32 domain = lobj->prefered_domains;
 			u32 allowed = lobj->allowed_domains;
@@ -562,6 +563,12 @@ int radeon_bo_list_validate(struct radeon_device *rdev,
 		lobj->gpu_offset = radeon_bo_gpu_offset(bo);
 		lobj->tiling_flags = bo->tiling_flags;
 	}
+
+	list_for_each_entry(lobj, &duplicates, tv.head) {
+		lobj->gpu_offset = radeon_bo_gpu_offset(lobj->robj);
+		lobj->tiling_flags = lobj->robj->tiling_flags;
+	}
+
 	return 0;
 }
 

drivers/gpu/drm/radeon/radeon_trace.h

@@ -38,7 +38,7 @@ TRACE_EVENT(radeon_cs,
 
 	    TP_fast_assign(
 			   __entry->ring = p->ring;
-			   __entry->dw = p->chunks[p->chunk_ib_idx].length_dw;
+			   __entry->dw = p->chunk_ib->length_dw;
 			   __entry->fences = radeon_fence_count_emitted(
 				p->rdev, p->ring);
 			   ),

drivers/gpu/drm/radeon/radeon_ttm.c

@@ -196,7 +196,7 @@ static void radeon_evict_flags(struct ttm_buffer_object *bo,
 	rbo = container_of(bo, struct radeon_bo, tbo);
 	switch (bo->mem.mem_type) {
 	case TTM_PL_VRAM:
-		if (rbo->rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready == false)
+		if (rbo->rdev->ring[radeon_copy_ring_index(rbo->rdev)].ready == false)
 			radeon_ttm_placement_from_domain(rbo, RADEON_GEM_DOMAIN_CPU);
 		else if (rbo->rdev->mc.visible_vram_size < rbo->rdev->mc.real_vram_size &&
 			 bo->mem.start < (rbo->rdev->mc.visible_vram_size >> PAGE_SHIFT)) {

drivers/gpu/drm/radeon/radeon_uvd.c

@@ -488,12 +488,12 @@ static int radeon_uvd_cs_reloc(struct radeon_cs_parser *p,
 			       unsigned buf_sizes[], bool *has_msg_cmd)
 {
 	struct radeon_cs_chunk *relocs_chunk;
-	struct radeon_cs_reloc *reloc;
+	struct radeon_bo_list *reloc;
 	unsigned idx, cmd, offset;
 	uint64_t start, end;
 	int r;
 
-	relocs_chunk = &p->chunks[p->chunk_relocs_idx];
+	relocs_chunk = p->chunk_relocs;
 	offset = radeon_get_ib_value(p, data0);
 	idx = radeon_get_ib_value(p, data1);
 	if (idx >= relocs_chunk->length_dw) {
@@ -502,7 +502,7 @@ static int radeon_uvd_cs_reloc(struct radeon_cs_parser *p,
 		return -EINVAL;
 	}
 
-	reloc = p->relocs_ptr[(idx / 4)];
+	reloc = &p->relocs[(idx / 4)];
 	start = reloc->gpu_offset;
 	end = start + radeon_bo_size(reloc->robj);
 	start += offset;
@@ -610,13 +610,13 @@ int radeon_uvd_cs_parse(struct radeon_cs_parser *p)
 		[0x00000003]	=	2048,
 	};
 
-	if (p->chunks[p->chunk_ib_idx].length_dw % 16) {
+	if (p->chunk_ib->length_dw % 16) {
 		DRM_ERROR("UVD IB length (%d) not 16 dwords aligned!\n",
-			  p->chunks[p->chunk_ib_idx].length_dw);
+			  p->chunk_ib->length_dw);
 		return -EINVAL;
 	}
 
-	if (p->chunk_relocs_idx == -1) {
+	if (p->chunk_relocs == NULL) {
 		DRM_ERROR("No relocation chunk !\n");
 		return -EINVAL;
 	}
@@ -640,7 +640,7 @@ int radeon_uvd_cs_parse(struct radeon_cs_parser *p)
 			DRM_ERROR("Unknown packet type %d !\n", pkt.type);
 			return -EINVAL;
 		}
-	} while (p->idx < p->chunks[p->chunk_ib_idx].length_dw);
+	} while (p->idx < p->chunk_ib->length_dw);
 
 	if (!has_msg_cmd) {
 		DRM_ERROR("UVD-IBs need a msg command!\n");

drivers/gpu/drm/radeon/radeon_vce.c

@@ -453,11 +453,11 @@ int radeon_vce_cs_reloc(struct radeon_cs_parser *p, int lo, int hi,
 			unsigned size)
 {
 	struct radeon_cs_chunk *relocs_chunk;
-	struct radeon_cs_reloc *reloc;
+	struct radeon_bo_list *reloc;
 	uint64_t start, end, offset;
 	unsigned idx;
 
-	relocs_chunk = &p->chunks[p->chunk_relocs_idx];
+	relocs_chunk = p->chunk_relocs;
 	offset = radeon_get_ib_value(p, lo);
 	idx = radeon_get_ib_value(p, hi);
 
@@ -467,7 +467,7 @@ int radeon_vce_cs_reloc(struct radeon_cs_parser *p, int lo, int hi,
 		return -EINVAL;
 	}
 
-	reloc = p->relocs_ptr[(idx / 4)];
+	reloc = &p->relocs[(idx / 4)];
 	start = reloc->gpu_offset;
 	end = start + radeon_bo_size(reloc->robj);
 	start += offset;
@@ -534,7 +534,7 @@ int radeon_vce_cs_parse(struct radeon_cs_parser *p)
 	uint32_t *size = &tmp;
 	int i, r;
 
-	while (p->idx < p->chunks[p->chunk_ib_idx].length_dw) {
+	while (p->idx < p->chunk_ib->length_dw) {
 		uint32_t len = radeon_get_ib_value(p, p->idx);
 		uint32_t cmd = radeon_get_ib_value(p, p->idx + 1);
 

drivers/gpu/drm/radeon/radeon_vm.c

@@ -125,41 +125,37 @@ void radeon_vm_manager_fini(struct radeon_device *rdev)
  * Add the page directory to the list of BOs to
  * validate for command submission (cayman+).
  */
-struct radeon_cs_reloc *radeon_vm_get_bos(struct radeon_device *rdev,
+struct radeon_bo_list *radeon_vm_get_bos(struct radeon_device *rdev,
 					  struct radeon_vm *vm,
 					  struct list_head *head)
 {
-	struct radeon_cs_reloc *list;
+	struct radeon_bo_list *list;
 	unsigned i, idx;
 
 	list = drm_malloc_ab(vm->max_pde_used + 2,
-			     sizeof(struct radeon_cs_reloc));
+			     sizeof(struct radeon_bo_list));
 	if (!list)
 		return NULL;
 
 	/* add the vm page table to the list */
-	list[0].gobj = NULL;
 	list[0].robj = vm->page_directory;
 	list[0].prefered_domains = RADEON_GEM_DOMAIN_VRAM;
 	list[0].allowed_domains = RADEON_GEM_DOMAIN_VRAM;
 	list[0].tv.bo = &vm->page_directory->tbo;
 	list[0].tv.shared = true;
 	list[0].tiling_flags = 0;
-	list[0].handle = 0;
 	list_add(&list[0].tv.head, head);
 
 	for (i = 0, idx = 1; i <= vm->max_pde_used; i++) {
 		if (!vm->page_tables[i].bo)
 			continue;
 
-		list[idx].gobj = NULL;
 		list[idx].robj = vm->page_tables[i].bo;
 		list[idx].prefered_domains = RADEON_GEM_DOMAIN_VRAM;
 		list[idx].allowed_domains = RADEON_GEM_DOMAIN_VRAM;
 		list[idx].tv.bo = &list[idx].robj->tbo;
 		list[idx].tv.shared = true;
 		list[idx].tiling_flags = 0;
-		list[idx].handle = 0;
 		list_add(&list[idx++].tv.head, head);
 	}
 
@@ -491,7 +487,9 @@ int radeon_vm_bo_set_addr(struct radeon_device *rdev,
 			tmp->vm = vm;
 			tmp->addr = bo_va->addr;
 			tmp->bo = radeon_bo_ref(bo_va->bo);
+			spin_lock(&vm->status_lock);
 			list_add(&tmp->vm_status, &vm->freed);
+			spin_unlock(&vm->status_lock);
 		}
 
 		interval_tree_remove(&bo_va->it, &vm->va);
@@ -802,11 +800,11 @@ static void radeon_vm_frag_ptes(struct radeon_device *rdev,
  *
  * Global and local mutex must be locked!
  */
-static void radeon_vm_update_ptes(struct radeon_device *rdev,
-				  struct radeon_vm *vm,
-				  struct radeon_ib *ib,
-				  uint64_t start, uint64_t end,
-				  uint64_t dst, uint32_t flags)
+static int radeon_vm_update_ptes(struct radeon_device *rdev,
+				 struct radeon_vm *vm,
+				 struct radeon_ib *ib,
+				 uint64_t start, uint64_t end,
+				 uint64_t dst, uint32_t flags)
 {
 	uint64_t mask = RADEON_VM_PTE_COUNT - 1;
 	uint64_t last_pte = ~0, last_dst = ~0;
@@ -819,8 +817,12 @@ static void radeon_vm_update_ptes(struct radeon_device *rdev,
 		struct radeon_bo *pt = vm->page_tables[pt_idx].bo;
 		unsigned nptes;
 		uint64_t pte;
+		int r;
 
 		radeon_sync_resv(rdev, &ib->sync, pt->tbo.resv, true);
+		r = reservation_object_reserve_shared(pt->tbo.resv);
+		if (r)
+			return r;
 
 		if ((addr & ~mask) == (end & ~mask))
 			nptes = end - addr;
@@ -854,6 +856,8 @@ static void radeon_vm_update_ptes(struct radeon_device *rdev,
 				    last_pte + 8 * count,
 				    last_dst, flags);
 	}
+
+	return 0;
 }
 
 /**
@@ -878,7 +882,7 @@ static void radeon_vm_fence_pts(struct radeon_vm *vm,
 	end >>= radeon_vm_block_size;
 
 	for (i = start; i <= end; ++i)
-		radeon_bo_fence(vm->page_tables[i].bo, fence, false);
+		radeon_bo_fence(vm->page_tables[i].bo, fence, true);
 }
 
 /**
@@ -911,7 +915,9 @@ int radeon_vm_bo_update(struct radeon_device *rdev,
 		return -EINVAL;
 	}
 
+	spin_lock(&vm->status_lock);
 	list_del_init(&bo_va->vm_status);
+	spin_unlock(&vm->status_lock);
 
 	bo_va->flags &= ~RADEON_VM_PAGE_VALID;
 	bo_va->flags &= ~RADEON_VM_PAGE_SYSTEM;
@@ -987,9 +993,13 @@ int radeon_vm_bo_update(struct radeon_device *rdev,
 			radeon_sync_fence(&ib.sync, vm->ids[i].last_id_use);
 	}
 
-	radeon_vm_update_ptes(rdev, vm, &ib, bo_va->it.start,
-			      bo_va->it.last + 1, addr,
-			      radeon_vm_page_flags(bo_va->flags));
+	r = radeon_vm_update_ptes(rdev, vm, &ib, bo_va->it.start,
+				  bo_va->it.last + 1, addr,
+				  radeon_vm_page_flags(bo_va->flags));
+	if (r) {
+		radeon_ib_free(rdev, &ib);
+		return r;
+	}
 
 	radeon_asic_vm_pad_ib(rdev, &ib);
 	WARN_ON(ib.length_dw > ndw);
@@ -1022,17 +1032,25 @@ int radeon_vm_bo_update(struct radeon_device *rdev,
 int radeon_vm_clear_freed(struct radeon_device *rdev,
 			  struct radeon_vm *vm)
 {
-	struct radeon_bo_va *bo_va, *tmp;
+	struct radeon_bo_va *bo_va;
 	int r;
 
-	list_for_each_entry_safe(bo_va, tmp, &vm->freed, vm_status) {
+	spin_lock(&vm->status_lock);
+	while (!list_empty(&vm->freed)) {
+		bo_va = list_first_entry(&vm->freed,
+			struct radeon_bo_va, vm_status);
+		spin_unlock(&vm->status_lock);
+
 		r = radeon_vm_bo_update(rdev, bo_va, NULL);
 		radeon_bo_unref(&bo_va->bo);
 		radeon_fence_unref(&bo_va->last_pt_update);
 		kfree(bo_va);
 		if (r)
 			return r;
+
+		spin_lock(&vm->status_lock);
 	}
+	spin_unlock(&vm->status_lock);
 	return 0;
 
 }
@@ -1051,14 +1069,23 @@ int radeon_vm_clear_freed(struct radeon_device *rdev,
 int radeon_vm_clear_invalids(struct radeon_device *rdev,
 			     struct radeon_vm *vm)
 {
-	struct radeon_bo_va *bo_va, *tmp;
+	struct radeon_bo_va *bo_va;
 	int r;
 
-	list_for_each_entry_safe(bo_va, tmp, &vm->invalidated, vm_status) {
+	spin_lock(&vm->status_lock);
+	while (!list_empty(&vm->invalidated)) {
+		bo_va = list_first_entry(&vm->invalidated,
+			struct radeon_bo_va, vm_status);
+		spin_unlock(&vm->status_lock);
+
 		r = radeon_vm_bo_update(rdev, bo_va, NULL);
 		if (r)
 			return r;
+
+		spin_lock(&vm->status_lock);
 	}
+	spin_unlock(&vm->status_lock);
+
 	return 0;
 }
 
@@ -1081,6 +1108,7 @@ void radeon_vm_bo_rmv(struct radeon_device *rdev,
 
 	mutex_lock(&vm->mutex);
 	interval_tree_remove(&bo_va->it, &vm->va);
+	spin_lock(&vm->status_lock);
 	list_del(&bo_va->vm_status);
 
 	if (bo_va->addr) {
@@ -1090,6 +1118,7 @@ void radeon_vm_bo_rmv(struct radeon_device *rdev,
 		radeon_fence_unref(&bo_va->last_pt_update);
 		kfree(bo_va);
 	}
+	spin_unlock(&vm->status_lock);
 
 	mutex_unlock(&vm->mutex);
 }
@@ -1110,10 +1139,10 @@ void radeon_vm_bo_invalidate(struct radeon_device *rdev,
 
 	list_for_each_entry(bo_va, &bo->va, bo_list) {
 		if (bo_va->addr) {
-			mutex_lock(&bo_va->vm->mutex);
+			spin_lock(&bo_va->vm->status_lock);
 			list_del(&bo_va->vm_status);
 			list_add(&bo_va->vm_status, &bo_va->vm->invalidated);
-			mutex_unlock(&bo_va->vm->mutex);
+			spin_unlock(&bo_va->vm->status_lock);
 		}
 	}
 }
@@ -1141,6 +1170,7 @@ int radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm)
 	}
 	mutex_init(&vm->mutex);
 	vm->va = RB_ROOT;
+	spin_lock_init(&vm->status_lock);
 	INIT_LIST_HEAD(&vm->invalidated);
 	INIT_LIST_HEAD(&vm->freed);
 

drivers/gpu/drm/radeon/si_dpm.c

@@ -5893,7 +5893,7 @@ static void si_fan_ctrl_set_static_mode(struct radeon_device *rdev, u32 mode)
 	tmp |= TMIN(0);
 	WREG32(CG_FDO_CTRL2, tmp);
 
-	tmp = RREG32(CG_FDO_CTRL2) & FDO_PWM_MODE_MASK;
+	tmp = RREG32(CG_FDO_CTRL2) & ~FDO_PWM_MODE_MASK;
 	tmp |= FDO_PWM_MODE(mode);
 	WREG32(CG_FDO_CTRL2, tmp);
 }
@@ -6098,7 +6098,7 @@ static int si_fan_ctrl_set_fan_speed_rpm(struct radeon_device *rdev,
 	tmp |= TARGET_PERIOD(tach_period);
 	WREG32(CG_TACH_CTRL, tmp);
 
-	si_fan_ctrl_set_static_mode(rdev, FDO_PWM_MODE_STATIC);
+	si_fan_ctrl_set_static_mode(rdev, FDO_PWM_MODE_STATIC_RPM);
 
 	return 0;
 }
@@ -6114,7 +6114,7 @@ static void si_fan_ctrl_set_default_mode(struct radeon_device *rdev)
 		tmp |= FDO_PWM_MODE(si_pi->fan_ctrl_default_mode);
 		WREG32(CG_FDO_CTRL2, tmp);
 
-		tmp = RREG32(CG_FDO_CTRL2) & TMIN_MASK;
+		tmp = RREG32(CG_FDO_CTRL2) & ~TMIN_MASK;
 		tmp |= TMIN(si_pi->t_min);
 		WREG32(CG_FDO_CTRL2, tmp);
 		si_pi->fan_ctrl_is_in_default_mode = true;

drivers/gpu/drm/radeon/sid.h

@@ -208,18 +208,18 @@
 
 #define	CG_FDO_CTRL0					0x754
 #define		FDO_STATIC_DUTY(x)			((x) << 0)
-#define		FDO_STATIC_DUTY_MASK			0x0000000F
+#define		FDO_STATIC_DUTY_MASK			0x000000FF
 #define		FDO_STATIC_DUTY_SHIFT			0
 #define	CG_FDO_CTRL1					0x758
 #define		FMAX_DUTY100(x)				((x) << 0)
-#define		FMAX_DUTY100_MASK			0x0000000F
+#define		FMAX_DUTY100_MASK			0x000000FF
 #define		FMAX_DUTY100_SHIFT			0
 #define	CG_FDO_CTRL2					0x75C
 #define		TMIN(x)					((x) << 0)
-#define		TMIN_MASK				0x0000000F
+#define		TMIN_MASK				0x000000FF
 #define		TMIN_SHIFT				0
 #define		FDO_PWM_MODE(x)				((x) << 11)
-#define		FDO_PWM_MODE_MASK			(3 << 11)
+#define		FDO_PWM_MODE_MASK			(7 << 11)
 #define		FDO_PWM_MODE_SHIFT			11
 #define		TACH_PWM_RESP_RATE(x)			((x) << 25)
 #define		TACH_PWM_RESP_RATE_MASK			(0x7f << 25)

drivers/gpu/drm/ttm/ttm_execbuf_util.c

@@ -93,7 +93,8 @@ EXPORT_SYMBOL(ttm_eu_backoff_reservation);
  */
 
 int ttm_eu_reserve_buffers(struct ww_acquire_ctx *ticket,
-			   struct list_head *list, bool intr)
+			   struct list_head *list, bool intr,
+			   struct list_head *dups)
 {
 	struct ttm_bo_global *glob;
 	struct ttm_validate_buffer *entry;
@@ -117,6 +118,13 @@ int ttm_eu_reserve_buffers(struct ww_acquire_ctx *ticket,
 			__ttm_bo_unreserve(bo);
 
 			ret = -EBUSY;
+
+		} else if (ret == -EALREADY && dups) {
+			struct ttm_validate_buffer *safe = entry;
+			entry = list_prev_entry(entry, head);
+			list_del(&safe->head);
+			list_add(&safe->head, dups);
+			continue;
 		}
 
 		if (!ret) {

drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c

@@ -2487,7 +2487,8 @@ int vmw_execbuf_process(struct drm_file *file_priv,
 	if (unlikely(ret != 0))
 		goto out_err_nores;
 
-	ret = ttm_eu_reserve_buffers(&ticket, &sw_context->validate_nodes, true);
+	ret = ttm_eu_reserve_buffers(&ticket, &sw_context->validate_nodes,
+				     true, NULL);
 	if (unlikely(ret != 0))
 		goto out_err;
 
@@ -2677,7 +2678,8 @@ void __vmw_execbuf_release_pinned_bo(struct vmw_private *dev_priv,
 	query_val.shared = false;
 	list_add_tail(&query_val.head, &validate_list);
 
-	ret = ttm_eu_reserve_buffers(&ticket, &validate_list, false);
+	ret = ttm_eu_reserve_buffers(&ticket, &validate_list,
+				     false, NULL);
 	if (unlikely(ret != 0)) {
 		vmw_execbuf_unpin_panic(dev_priv);
 		goto out_no_reserve;

drivers/gpu/drm/vmwgfx/vmwgfx_resource.c

@@ -1222,7 +1222,7 @@ vmw_resource_check_buffer(struct vmw_resource *res,
 	val_buf->bo = ttm_bo_reference(&res->backup->base);
 	val_buf->shared = false;
 	list_add_tail(&val_buf->head, &val_list);
-	ret = ttm_eu_reserve_buffers(NULL, &val_list, interruptible);
+	ret = ttm_eu_reserve_buffers(NULL, &val_list, interruptible, NULL);
 	if (unlikely(ret != 0))
 		goto out_no_reserve;
 

include/drm/ttm/ttm_execbuf_util.h

@@ -68,6 +68,7 @@ extern void ttm_eu_backoff_reservation(struct ww_acquire_ctx *ticket,
  *           non-blocking reserves should be tried.
  * @list:    thread private list of ttm_validate_buffer structs.
  * @intr:    should the wait be interruptible
+ * @dups:    [out] optional list of duplicates.
  *
  * Tries to reserve bos pointed to by the list entries for validation.
  * If the function returns 0, all buffers are marked as "unfenced",
@@ -83,6 +84,11 @@ extern void ttm_eu_backoff_reservation(struct ww_acquire_ctx *ticket,
  * calling process receives a signal while waiting. In that case, no
  * buffers on the list will be reserved upon return.
  *
+ * If dups is non NULL all buffers already reserved by the current thread
+ * (e.g. duplicates) are added to this list, otherwise -EALREADY is returned
+ * on the first already reserved buffer and all buffers from the list are
+ * unreserved again.
+ *
  * Buffers reserved by this function should be unreserved by
  * a call to either ttm_eu_backoff_reservation() or
  * ttm_eu_fence_buffer_objects() when command submission is complete or
@@ -90,7 +96,8 @@ extern void ttm_eu_backoff_reservation(struct ww_acquire_ctx *ticket,
  */
 
 extern int ttm_eu_reserve_buffers(struct ww_acquire_ctx *ticket,
-				  struct list_head *list, bool intr);
+				  struct list_head *list, bool intr,
+				  struct list_head *dups);
 
 /**
  * function ttm_eu_fence_buffer_objects.