Merge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux

Pull drm fixes from Dave Airlie:
 "Fixes for i915, nouveau and radeon:

   - i915: haswell stability, modeset rework fallout, ums fix
   - nouveau: misc fixes from code rework
   - radeon: pll rework fixes, more 2 level PTE cleanups.
   - core: warning fixes on 32-bit."

* 'drm-fixes' of git://people.freedesktop.org/~airlied/linux: (31 commits)
  nouveau: fix warning on 32-bit build.
  drm/nouveau/bios: fix typo in error message
  drm/nouveau: only call ttm_agp_tt_create when __OS_HAS_AGP
  drm/nv50/fb: fix double free of vram mm
  drm/nouveau/pm: do not stop reclocking if failing to set the fan speed
  drm/nouveau/pm: fix a typo related to the move to the therm subdev
  drm/nouveau/hwmon: fix the initialization condition
  drm: fix warning on 32-bit.
  drm: radeon: fix printk format warning
  drm/radeon: fix spelling typos in debugging output
  drm/radeon: Don't destroy I2C Bus Rec in radeon_ext_tmds_enc_destroy().
  drm/radeon: check if pcie gen 2 is already enabled (v2)
  drm/radeon/cayman: set VM max pfn at MC init
  drm/radeon: separate pt alloc from lru add
  drm/radeon: don't add the IB pool to all VMs v2
  drm/radeon: allocate page tables on demand v4
  drm/radeon: update comments to clarify VM setup (v2)
  drm/radeon: allocate PPLLs from low to high
  drm/radeon: fix compilation with backlight disabled
  drm/radeon: use %zu for formatting size_t
  ...

29 files changed, 443 insertions, 249 deletions

diff --git a/drivers/char/agp/intel-gtt.c b/drivers/char/agp/intel-gtt.c
index e01f5eaaec82..38390f7c6ab6 100644
--- a/drivers/char/agp/intel-gtt.c
+++ b/drivers/char/agp/intel-gtt.c
@@ -667,7 +667,7 @@ static int intel_gtt_init(void)
 	gtt_map_size = intel_private.base.gtt_total_entries * 4;
 
 	intel_private.gtt = NULL;
-	if (INTEL_GTT_GEN < 6)
+	if (INTEL_GTT_GEN < 6 && INTEL_GTT_GEN > 2)
 		intel_private.gtt = ioremap_wc(intel_private.gtt_bus_addr,
 					       gtt_map_size);
 	if (intel_private.gtt == NULL)
diff --git a/drivers/gpu/drm/drm_info.c b/drivers/gpu/drm/drm_info.c
index cdf8b1e7602d..441ebc1bdbef 100644
--- a/drivers/gpu/drm/drm_info.c
+++ b/drivers/gpu/drm/drm_info.c
@@ -239,7 +239,7 @@ int drm_vma_info(struct seq_file *m, void *data)
 	mutex_lock(&dev->struct_mutex);
 	seq_printf(m, "vma use count: %d, high_memory = %pK, 0x%pK\n",
 		   atomic_read(&dev->vma_count),
-		   high_memory, (void *)virt_to_phys(high_memory));
+		   high_memory, (void *)(unsigned long)virt_to_phys(high_memory));
 
 	list_for_each_entry(pt, &dev->vmalist, head) {
 		vma = pt->vma;
diff --git a/drivers/gpu/drm/i915/dvo_ch7xxx.c b/drivers/gpu/drm/i915/dvo_ch7xxx.c
index 38f3a6cb8c7d..3edd981e0770 100644
--- a/drivers/gpu/drm/i915/dvo_ch7xxx.c
+++ b/drivers/gpu/drm/i915/dvo_ch7xxx.c
@@ -303,10 +303,10 @@ static bool ch7xxx_get_hw_state(struct intel_dvo_device *dvo)
 
 	ch7xxx_readb(dvo, CH7xxx_PM, &val);
 
-	if (val & CH7xxx_PM_FPD)
-		return false;
-	else
+	if (val & (CH7xxx_PM_DVIL | CH7xxx_PM_DVIP))
 		return true;
+	else
+		return false;
 }
 
 static void ch7xxx_dump_regs(struct intel_dvo_device *dvo)
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 4f2831aa5fed..b84f7861e438 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -1341,9 +1341,14 @@ int __must_check i915_gem_object_get_pages(struct drm_i915_gem_object *obj);
 static inline struct page *i915_gem_object_get_page(struct drm_i915_gem_object *obj, int n)
 {
 	struct scatterlist *sg = obj->pages->sgl;
-	while (n >= SG_MAX_SINGLE_ALLOC) {
+	int nents = obj->pages->nents;
+	while (nents > SG_MAX_SINGLE_ALLOC) {
+		if (n < SG_MAX_SINGLE_ALLOC - 1)
+			break;
+
 		sg = sg_chain_ptr(sg + SG_MAX_SINGLE_ALLOC - 1);
 		n -= SG_MAX_SINGLE_ALLOC - 1;
+		nents -= SG_MAX_SINGLE_ALLOC - 1;
 	}
 	return sg_page(sg+n);
 }
@@ -1427,7 +1432,7 @@ int __must_check i915_gpu_idle(struct drm_device *dev);
 int __must_check i915_gem_idle(struct drm_device *dev);
 int i915_add_request(struct intel_ring_buffer *ring,
 		     struct drm_file *file,
-		     struct drm_i915_gem_request *request);
+		     u32 *seqno);
 int __must_check i915_wait_seqno(struct intel_ring_buffer *ring,
 				 uint32_t seqno);
 int i915_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf);
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 19dbdd7dd564..d33d02d13c96 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -1955,11 +1955,12 @@ i915_gem_next_request_seqno(struct intel_ring_buffer *ring)
 int
 i915_add_request(struct intel_ring_buffer *ring,
 		 struct drm_file *file,
-		 struct drm_i915_gem_request *request)
+		 u32 *out_seqno)
 {
 	drm_i915_private_t *dev_priv = ring->dev->dev_private;
-	uint32_t seqno;
+	struct drm_i915_gem_request *request;
 	u32 request_ring_position;
+	u32 seqno;
 	int was_empty;
 	int ret;
 
@@ -1974,11 +1975,9 @@ i915_add_request(struct intel_ring_buffer *ring,
 	if (ret)
 		return ret;
 
-	if (request == NULL) {
-		request = kmalloc(sizeof(*request), GFP_KERNEL);
-		if (request == NULL)
-			return -ENOMEM;
-	}
+	request = kmalloc(sizeof(*request), GFP_KERNEL);
+	if (request == NULL)
+		return -ENOMEM;
 
 	seqno = i915_gem_next_request_seqno(ring);
 
@@ -2030,6 +2029,8 @@ i915_add_request(struct intel_ring_buffer *ring,
 		}
 	}
 
+	if (out_seqno)
+		*out_seqno = seqno;
 	return 0;
 }
 
@@ -3959,6 +3960,9 @@ i915_gem_init_hw(struct drm_device *dev)
 	if (!intel_enable_gtt())
 		return -EIO;
 
+	if (IS_HASWELL(dev) && (I915_READ(0x120010) == 1))
+		I915_WRITE(0x9008, I915_READ(0x9008) | 0xf0000);
+
 	i915_gem_l3_remap(dev);
 
 	i915_gem_init_swizzling(dev);
@@ -4098,7 +4102,6 @@ i915_gem_entervt_ioctl(struct drm_device *dev, void *data,
 	}
 
 	BUG_ON(!list_empty(&dev_priv->mm.active_list));
-	BUG_ON(!list_empty(&dev_priv->mm.inactive_list));
 	mutex_unlock(&dev->struct_mutex);
 
 	ret = drm_irq_install(dev);
diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 64c1be0a9cfd..a4162ddff6c5 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -521,7 +521,7 @@
  */
 # define _3D_CHICKEN2_WM_READ_PIPELINED			(1 << 14)
 #define _3D_CHICKEN3	0x02090
-#define  _3D_CHICKEN_SF_DISABLE_FASTCLIP_CULL		(1 << 5)
+#define  _3D_CHICKEN3_SF_DISABLE_FASTCLIP_CULL		(1 << 5)
 
 #define MI_MODE		0x0209c
 # define VS_TIMER_DISPATCH				(1 << 6)
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 2b6ce9b2674a..682bd3729baf 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -3253,6 +3253,16 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
 
 	if (HAS_PCH_CPT(dev))
 		intel_cpt_verify_modeset(dev, intel_crtc->pipe);
+
+	/*
+	 * There seems to be a race in PCH platform hw (at least on some
+	 * outputs) where an enabled pipe still completes any pageflip right
+	 * away (as if the pipe is off) instead of waiting for vblank. As soon
+	 * as the first vblank happend, everything works as expected. Hence just
+	 * wait for one vblank before returning to avoid strange things
+	 * happening.
+	 */
+	intel_wait_for_vblank(dev, intel_crtc->pipe);
 }
 
 static void ironlake_crtc_disable(struct drm_crtc *crtc)
@@ -7892,8 +7902,7 @@ static struct intel_quirk intel_quirks[] = {
 	/* ThinkPad T60 needs pipe A force quirk (bug #16494) */
 	{ 0x2782, 0x17aa, 0x201a, quirk_pipea_force },
 
-	/* 855 & before need to leave pipe A & dpll A up */
-	{ 0x3582, PCI_ANY_ID, PCI_ANY_ID, quirk_pipea_force },
+	/* 830/845 need to leave pipe A & dpll A up */
 	{ 0x2562, PCI_ANY_ID, PCI_ANY_ID, quirk_pipea_force },
 	{ 0x3577, PCI_ANY_ID, PCI_ANY_ID, quirk_pipea_force },
 
@@ -8049,29 +8058,42 @@ static void intel_enable_pipe_a(struct drm_device *dev)
 
 }
 
+static bool
+intel_check_plane_mapping(struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+	u32 reg, val;
+
+	if (dev_priv->num_pipe == 1)
+		return true;
+
+	reg = DSPCNTR(!crtc->plane);
+	val = I915_READ(reg);
+
+	if ((val & DISPLAY_PLANE_ENABLE) &&
+	    (!!(val & DISPPLANE_SEL_PIPE_MASK) == crtc->pipe))
+		return false;
+
+	return true;
+}
+
 static void intel_sanitize_crtc(struct intel_crtc *crtc)
 {
 	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 reg, val;
+	u32 reg;
 
 	/* Clear any frame start delays used for debugging left by the BIOS */
 	reg = PIPECONF(crtc->pipe);
 	I915_WRITE(reg, I915_READ(reg) & ~PIPECONF_FRAME_START_DELAY_MASK);
 
 	/* We need to sanitize the plane -> pipe mapping first because this will
-	 * disable the crtc (and hence change the state) if it is wrong. */
-	if (!HAS_PCH_SPLIT(dev)) {
+	 * disable the crtc (and hence change the state) if it is wrong. Note
+	 * that gen4+ has a fixed plane -> pipe mapping.  */
+	if (INTEL_INFO(dev)->gen < 4 && !intel_check_plane_mapping(crtc)) {
 		struct intel_connector *connector;
 		bool plane;
 
-		reg = DSPCNTR(crtc->plane);
-		val = I915_READ(reg);
-
-		if ((val & DISPLAY_PLANE_ENABLE) == 0 &&
-		    (!!(val & DISPPLANE_SEL_PIPE_MASK) == crtc->pipe))
-			goto ok;
-
 		DRM_DEBUG_KMS("[CRTC:%d] wrong plane connection detected!\n",
 			      crtc->base.base.id);
 
@@ -8095,7 +8117,6 @@ static void intel_sanitize_crtc(struct intel_crtc *crtc)
 		WARN_ON(crtc->active);
 		crtc->base.enabled = false;
 	}
-ok:
 
 	if (dev_priv->quirks & QUIRK_PIPEA_FORCE &&
 	    crtc->pipe == PIPE_A && !crtc->active) {
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index d1e8ddb2d6c0..1b727a5c9ee5 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -2369,8 +2369,9 @@ static void
 intel_dp_destroy(struct drm_connector *connector)
 {
 	struct drm_device *dev = connector->dev;
+	struct intel_dp *intel_dp = intel_attached_dp(connector);
 
-	if (intel_dpd_is_edp(dev))
+	if (is_edp(intel_dp))
 		intel_panel_destroy_backlight(dev);
 
 	drm_sysfs_connector_remove(connector);
diff --git a/drivers/gpu/drm/i915/intel_overlay.c b/drivers/gpu/drm/i915/intel_overlay.c
index ebff850a9ab6..495625914e4a 100644
--- a/drivers/gpu/drm/i915/intel_overlay.c
+++ b/drivers/gpu/drm/i915/intel_overlay.c
@@ -209,7 +209,6 @@ static void intel_overlay_unmap_regs(struct intel_overlay *overlay,
 }
 
 static int intel_overlay_do_wait_request(struct intel_overlay *overlay,
-					 struct drm_i915_gem_request *request,
 					 void (*tail)(struct intel_overlay *))
 {
 	struct drm_device *dev = overlay->dev;
@@ -218,12 +217,10 @@ static int intel_overlay_do_wait_request(struct intel_overlay *overlay,
 	int ret;
 
 	BUG_ON(overlay->last_flip_req);
-	ret = i915_add_request(ring, NULL, request);
-	if (ret) {
-	    kfree(request);
-	    return ret;
-	}
-	overlay->last_flip_req = request->seqno;
+	ret = i915_add_request(ring, NULL, &overlay->last_flip_req);
+	if (ret)
+		return ret;
+
 	overlay->flip_tail = tail;
 	ret = i915_wait_seqno(ring, overlay->last_flip_req);
 	if (ret)
@@ -240,7 +237,6 @@ static int intel_overlay_on(struct intel_overlay *overlay)
 	struct drm_device *dev = overlay->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_ring_buffer *ring = &dev_priv->ring[RCS];
-	struct drm_i915_gem_request *request;
 	int ret;
 
 	BUG_ON(overlay->active);
@@ -248,17 +244,9 @@ static int intel_overlay_on(struct intel_overlay *overlay)
 
 	WARN_ON(IS_I830(dev) && !(dev_priv->quirks & QUIRK_PIPEA_FORCE));
 
-	request = kzalloc(sizeof(*request), GFP_KERNEL);
-	if (request == NULL) {
-		ret = -ENOMEM;
-		goto out;
-	}
-
 	ret = intel_ring_begin(ring, 4);
-	if (ret) {
-		kfree(request);
-		goto out;
-	}
+	if (ret)
+		return ret;
 
 	intel_ring_emit(ring, MI_OVERLAY_FLIP | MI_OVERLAY_ON);
 	intel_ring_emit(ring, overlay->flip_addr | OFC_UPDATE);
@@ -266,9 +254,7 @@ static int intel_overlay_on(struct intel_overlay *overlay)
 	intel_ring_emit(ring, MI_NOOP);
 	intel_ring_advance(ring);
 
-	ret = intel_overlay_do_wait_request(overlay, request, NULL);
-out:
-	return ret;
+	return intel_overlay_do_wait_request(overlay, NULL);
 }
 
 /* overlay needs to be enabled in OCMD reg */
@@ -278,17 +264,12 @@ static int intel_overlay_continue(struct intel_overlay *overlay,
 	struct drm_device *dev = overlay->dev;
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	struct intel_ring_buffer *ring = &dev_priv->ring[RCS];
-	struct drm_i915_gem_request *request;
 	u32 flip_addr = overlay->flip_addr;
 	u32 tmp;
 	int ret;
 
 	BUG_ON(!overlay->active);
 
-	request = kzalloc(sizeof(*request), GFP_KERNEL);
-	if (request == NULL)
-		return -ENOMEM;
-
 	if (load_polyphase_filter)
 		flip_addr |= OFC_UPDATE;
 
@@ -298,22 +279,14 @@ static int intel_overlay_continue(struct intel_overlay *overlay,
 		DRM_DEBUG("overlay underrun, DOVSTA: %x\n", tmp);
 
 	ret = intel_ring_begin(ring, 2);
-	if (ret) {
-		kfree(request);
+	if (ret)
 		return ret;
-	}
+
 	intel_ring_emit(ring, MI_OVERLAY_FLIP | MI_OVERLAY_CONTINUE);
 	intel_ring_emit(ring, flip_addr);
 	intel_ring_advance(ring);
 
-	ret = i915_add_request(ring, NULL, request);
-	if (ret) {
-		kfree(request);
-		return ret;
-	}
-
-	overlay->last_flip_req = request->seqno;
-	return 0;
+	return i915_add_request(ring, NULL, &overlay->last_flip_req);
 }
 
 static void intel_overlay_release_old_vid_tail(struct intel_overlay *overlay)
@@ -349,15 +322,10 @@ static int intel_overlay_off(struct intel_overlay *overlay)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_ring_buffer *ring = &dev_priv->ring[RCS];
 	u32 flip_addr = overlay->flip_addr;
-	struct drm_i915_gem_request *request;
 	int ret;
 
 	BUG_ON(!overlay->active);
 
-	request = kzalloc(sizeof(*request), GFP_KERNEL);
-	if (request == NULL)
-		return -ENOMEM;
-
 	/* According to intel docs the overlay hw may hang (when switching
 	 * off) without loading the filter coeffs. It is however unclear whether
 	 * this applies to the disabling of the overlay or to the switching off
@@ -365,10 +333,9 @@ static int intel_overlay_off(struct intel_overlay *overlay)
 	flip_addr |= OFC_UPDATE;
 
 	ret = intel_ring_begin(ring, 6);
-	if (ret) {
-		kfree(request);
+	if (ret)
 		return ret;
-	}
+
 	/* wait for overlay to go idle */
 	intel_ring_emit(ring, MI_OVERLAY_FLIP | MI_OVERLAY_CONTINUE);
 	intel_ring_emit(ring, flip_addr);
@@ -379,8 +346,7 @@ static int intel_overlay_off(struct intel_overlay *overlay)
 	intel_ring_emit(ring, MI_WAIT_FOR_EVENT | MI_WAIT_FOR_OVERLAY_FLIP);
 	intel_ring_advance(ring);
 
-	return intel_overlay_do_wait_request(overlay, request,
-					     intel_overlay_off_tail);
+	return intel_overlay_do_wait_request(overlay, intel_overlay_off_tail);
 }
 
 /* recover from an interruption due to a signal
@@ -425,24 +391,16 @@ static int intel_overlay_release_old_vid(struct intel_overlay *overlay)
 		return 0;
 
 	if (I915_READ(ISR) & I915_OVERLAY_PLANE_FLIP_PENDING_INTERRUPT) {
-		struct drm_i915_gem_request *request;
-
 		/* synchronous slowpath */
-		request = kzalloc(sizeof(*request), GFP_KERNEL);
-		if (request == NULL)
-			return -ENOMEM;
-
 		ret = intel_ring_begin(ring, 2);
-		if (ret) {
-			kfree(request);
+		if (ret)
 			return ret;
-		}
 
 		intel_ring_emit(ring, MI_WAIT_FOR_EVENT | MI_WAIT_FOR_OVERLAY_FLIP);
 		intel_ring_emit(ring, MI_NOOP);
 		intel_ring_advance(ring);
 
-		ret = intel_overlay_do_wait_request(overlay, request,
+		ret = intel_overlay_do_wait_request(overlay,
 						    intel_overlay_release_old_vid_tail);
 		if (ret)
 			return ret;
diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index b3b4b6cea8b0..72f41aaa71ff 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -3442,8 +3442,8 @@ static void gen6_init_clock_gating(struct drm_device *dev)
 		   GEN6_RCCUNIT_CLOCK_GATE_DISABLE);
 
 	/* Bspec says we need to always set all mask bits. */
-	I915_WRITE(_3D_CHICKEN, (0xFFFF << 16) |
-		   _3D_CHICKEN_SF_DISABLE_FASTCLIP_CULL);
+	I915_WRITE(_3D_CHICKEN3, (0xFFFF << 16) |
+		   _3D_CHICKEN3_SF_DISABLE_FASTCLIP_CULL);
 
 	/*
 	 * According to the spec the following bits should be
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/dcb.c b/drivers/gpu/drm/nouveau/core/subdev/bios/dcb.c
index 9ed6e728a94c..7d750382a833 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/dcb.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/dcb.c
@@ -43,7 +43,7 @@ dcb_table(struct nouveau_bios *bios, u8 *ver, u8 *hdr, u8 *cnt, u8 *len)
 	*ver = nv_ro08(bios, dcb);
 
 	if (*ver >= 0x41) {
-		nv_warn(bios, "DCB *ver 0x%02x unknown\n", *ver);
+		nv_warn(bios, "DCB version 0x%02x unknown\n", *ver);
 		return 0x0000;
 	} else
 	if (*ver >= 0x30) {
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv50.c
index 436e9efe7ef5..42d7539e6525 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv50.c
@@ -277,7 +277,6 @@ nv50_fb_dtor(struct nouveau_object *object)
 		__free_page(priv->r100c08_page);
 	}
 
-	nouveau_mm_fini(&priv->base.vram);
 	nouveau_fb_destroy(&priv->base);
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c b/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
index b29237970fa0..523178685180 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
@@ -134,7 +134,7 @@ nouveau_therm_fan_sense(struct nouveau_therm *therm)
 	end = ptimer->read(ptimer);
 
 	if (cycles == 5) {
-		tach = (u64)60000000000;
+		tach = (u64)60000000000ULL;
 		do_div(tach, (end - start));
 		return tach;
 	} else
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index 259e5f1adf47..35ac57f0aab6 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -456,6 +456,7 @@ static struct ttm_tt *
 nouveau_ttm_tt_create(struct ttm_bo_device *bdev, unsigned long size,
 		      uint32_t page_flags, struct page *dummy_read)
 {
+#if __OS_HAS_AGP
 	struct nouveau_drm *drm = nouveau_bdev(bdev);
 	struct drm_device *dev = drm->dev;
 
@@ -463,6 +464,7 @@ nouveau_ttm_tt_create(struct ttm_bo_device *bdev, unsigned long size,
 		return ttm_agp_tt_create(bdev, dev->agp->bridge, size,
 					 page_flags, dummy_read);
 	}
+#endif
 
 	return nouveau_sgdma_create_ttm(bdev, size, page_flags, dummy_read);
 }
diff --git a/drivers/gpu/drm/nouveau/nouveau_pm.c b/drivers/gpu/drm/nouveau/nouveau_pm.c
index 0bf64c90aa20..5566172774df 100644
--- a/drivers/gpu/drm/nouveau/nouveau_pm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_pm.c
@@ -52,7 +52,7 @@ nouveau_pm_perflvl_aux(struct drm_device *dev, struct nouveau_pm_level *perflvl,
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	struct nouveau_pm *pm = nouveau_pm(dev);
-	struct nouveau_therm *therm = nouveau_therm(drm);
+	struct nouveau_therm *therm = nouveau_therm(drm->device);
 	int ret;
 
 	/*XXX: not on all boards, we should control based on temperature
@@ -64,7 +64,6 @@ nouveau_pm_perflvl_aux(struct drm_device *dev, struct nouveau_pm_level *perflvl,
 		ret = therm->fan_set(therm, perflvl->fanspeed);
 		if (ret && ret != -ENODEV) {
 			NV_ERROR(drm, "fanspeed set failed: %d\n", ret);
-			return ret;
 		}
 	}
 
@@ -706,8 +705,7 @@ nouveau_hwmon_init(struct drm_device *dev)
 	struct device *hwmon_dev;
 	int ret = 0;
 
-	if (!therm || !therm->temp_get || !therm->attr_get ||
-		!therm->attr_set || therm->temp_get(therm) < 0)
+	if (!therm || !therm->temp_get || !therm->attr_get || !therm->attr_set)
 		return -ENODEV;
 
 	hwmon_dev = hwmon_device_register(&dev->pdev->dev);
diff --git a/drivers/gpu/drm/radeon/atombios_crtc.c b/drivers/gpu/drm/radeon/atombios_crtc.c
index 96184d02c8d9..2e566e123e9e 100644
--- a/drivers/gpu/drm/radeon/atombios_crtc.c
+++ b/drivers/gpu/drm/radeon/atombios_crtc.c
@@ -1690,10 +1690,10 @@ static int radeon_atom_pick_pll(struct drm_crtc *crtc)
 		}
 		/* all other cases */
 		pll_in_use = radeon_get_pll_use_mask(crtc);
-		if (!(pll_in_use & (1 << ATOM_PPLL2)))
-			return ATOM_PPLL2;
 		if (!(pll_in_use & (1 << ATOM_PPLL1)))
 			return ATOM_PPLL1;
+		if (!(pll_in_use & (1 << ATOM_PPLL2)))
+			return ATOM_PPLL2;
 		DRM_ERROR("unable to allocate a PPLL\n");
 		return ATOM_PPLL_INVALID;
 	} else {
@@ -1715,10 +1715,10 @@ static int radeon_atom_pick_pll(struct drm_crtc *crtc)
 			}
 			/* all other cases */
 			pll_in_use = radeon_get_pll_use_mask(crtc);
-			if (!(pll_in_use & (1 << ATOM_PPLL2)))
-				return ATOM_PPLL2;
 			if (!(pll_in_use & (1 << ATOM_PPLL1)))
 				return ATOM_PPLL1;
+			if (!(pll_in_use & (1 << ATOM_PPLL2)))
+				return ATOM_PPLL2;
 			DRM_ERROR("unable to allocate a PPLL\n");
 			return ATOM_PPLL_INVALID;
 		} else {
diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c
index a1f49c5fd74b..14313ad43b76 100644
--- a/drivers/gpu/drm/radeon/evergreen.c
+++ b/drivers/gpu/drm/radeon/evergreen.c
@@ -3431,9 +3431,14 @@ void evergreen_pcie_gen2_enable(struct radeon_device *rdev)
 	if (!(mask & DRM_PCIE_SPEED_50))
 		return;
 
+	speed_cntl = RREG32_PCIE_P(PCIE_LC_SPEED_CNTL);
+	if (speed_cntl & LC_CURRENT_DATA_RATE) {
+		DRM_INFO("PCIE gen 2 link speeds already enabled\n");
+		return;
+	}
+
 	DRM_INFO("enabling PCIE gen 2 link speeds, disable with radeon.pcie_gen2=0\n");
 
-	speed_cntl = RREG32_PCIE_P(PCIE_LC_SPEED_CNTL);
 	if ((speed_cntl & LC_OTHER_SIDE_EVER_SENT_GEN2) ||
 	    (speed_cntl & LC_OTHER_SIDE_SUPPORTS_GEN2)) {
 
diff --git a/drivers/gpu/drm/radeon/ni.c b/drivers/gpu/drm/radeon/ni.c
index 8bcb554ea0c5..8c74c729586d 100644
--- a/drivers/gpu/drm/radeon/ni.c
+++ b/drivers/gpu/drm/radeon/ni.c
@@ -770,9 +770,13 @@ static int cayman_pcie_gart_enable(struct radeon_device *rdev)
 	WREG32(0x15DC, 0);
 
 	/* empty context1-7 */
+	/* Assign the pt base to something valid for now; the pts used for
+	 * the VMs are determined by the application and setup and assigned
+	 * on the fly in the vm part of radeon_gart.c
+	 */
 	for (i = 1; i < 8; i++) {
 		WREG32(VM_CONTEXT0_PAGE_TABLE_START_ADDR + (i << 2), 0);
-		WREG32(VM_CONTEXT0_PAGE_TABLE_END_ADDR + (i << 2), 0);
+		WREG32(VM_CONTEXT0_PAGE_TABLE_END_ADDR + (i << 2), rdev->vm_manager.max_pfn);
 		WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (i << 2),
 			rdev->gart.table_addr >> 12);
 	}
@@ -1572,12 +1576,6 @@ void cayman_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
 	if (vm == NULL)
 		return;
 
-	radeon_ring_write(ring, PACKET0(VM_CONTEXT0_PAGE_TABLE_START_ADDR + (vm->id << 2), 0));
-	radeon_ring_write(ring, 0);
-
-	radeon_ring_write(ring, PACKET0(VM_CONTEXT0_PAGE_TABLE_END_ADDR + (vm->id << 2), 0));
-	radeon_ring_write(ring, vm->last_pfn);
-
 	radeon_ring_write(ring, PACKET0(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm->id << 2), 0));
 	radeon_ring_write(ring, vm->pd_gpu_addr >> 12);
 
diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c
index 70c800ff6190..cda280d157da 100644
--- a/drivers/gpu/drm/radeon/r600.c
+++ b/drivers/gpu/drm/radeon/r600.c
@@ -3703,6 +3703,12 @@ static void r600_pcie_gen2_enable(struct radeon_device *rdev)
 	if (!(mask & DRM_PCIE_SPEED_50))
 		return;
 
+	speed_cntl = RREG32_PCIE_P(PCIE_LC_SPEED_CNTL);
+	if (speed_cntl & LC_CURRENT_DATA_RATE) {
+		DRM_INFO("PCIE gen 2 link speeds already enabled\n");
+		return;
+	}
+
 	DRM_INFO("enabling PCIE gen 2 link speeds, disable with radeon.pcie_gen2=0\n");
 
 	/* 55 nm r6xx asics */
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index b04c06444d8b..8c42d54c2e26 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -663,9 +663,14 @@ struct radeon_vm {
 	struct list_head		list;
 	struct list_head		va;
 	unsigned			id;
-	unsigned			last_pfn;
-	u64				pd_gpu_addr;
-	struct radeon_sa_bo		*sa_bo;
+
+	/* contains the page directory */
+	struct radeon_sa_bo		*page_directory;
+	uint64_t			pd_gpu_addr;
+
+	/* array of page tables, one for each page directory entry */
+	struct radeon_sa_bo		**page_tables;
+
 	struct mutex			mutex;
 	/* last fence for cs using this vm */
 	struct radeon_fence		*fence;
@@ -1843,9 +1848,10 @@ extern void radeon_ttm_set_active_vram_size(struct radeon_device *rdev, u64 size
  */
 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_init(struct radeon_device *rdev, struct radeon_vm *vm);
 void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm);
 int radeon_vm_alloc_pt(struct radeon_device *rdev, struct radeon_vm *vm);
+void radeon_vm_add_to_lru(struct radeon_device *rdev, struct radeon_vm *vm);
 struct radeon_fence *radeon_vm_grab_id(struct radeon_device *rdev,
 				       struct radeon_vm *vm, int ring);
 void radeon_vm_fence(struct radeon_device *rdev,
diff --git a/drivers/gpu/drm/radeon/radeon_acpi.c b/drivers/gpu/drm/radeon/radeon_acpi.c
index b0a5688c67f8..196d28d99570 100644
--- a/drivers/gpu/drm/radeon/radeon_acpi.c
+++ b/drivers/gpu/drm/radeon/radeon_acpi.c
@@ -201,7 +201,7 @@ static int radeon_atif_verify_interface(acpi_handle handle,
 
 	size = *(u16 *) info->buffer.pointer;
 	if (size < 12) {
-		DRM_INFO("ATIF buffer is too small: %lu\n", size);
+		DRM_INFO("ATIF buffer is too small: %zu\n", size);
 		err = -EINVAL;
 		goto out;
 	}
@@ -370,6 +370,7 @@ int radeon_atif_handler(struct radeon_device *rdev,
 
 			radeon_set_backlight_level(rdev, enc, req.backlight_level);
 
+#if defined(CONFIG_BACKLIGHT_CLASS_DEVICE) || defined(CONFIG_BACKLIGHT_CLASS_DEVICE_MODULE)
 			if (rdev->is_atom_bios) {
 				struct radeon_encoder_atom_dig *dig = enc->enc_priv;
 				backlight_force_update(dig->bl_dev,
@@ -379,6 +380,7 @@ int radeon_atif_handler(struct radeon_device *rdev,
 				backlight_force_update(dig->bl_dev,
 						       BACKLIGHT_UPDATE_HOTKEY);
 			}
+#endif
 		}
 	}
 	/* TODO: check other events */
@@ -485,7 +487,7 @@ static int radeon_atcs_verify_interface(acpi_handle handle,
 
 	size = *(u16 *) info->buffer.pointer;
 	if (size < 8) {
-		DRM_INFO("ATCS buffer is too small: %lu\n", size);
+		DRM_INFO("ATCS buffer is too small: %zu\n", size);
 		err = -EINVAL;
 		goto out;
 	}
diff --git a/drivers/gpu/drm/radeon/radeon_atpx_handler.c b/drivers/gpu/drm/radeon/radeon_atpx_handler.c
index 582e99449c12..1aa3f910b993 100644
--- a/drivers/gpu/drm/radeon/radeon_atpx_handler.c
+++ b/drivers/gpu/drm/radeon/radeon_atpx_handler.c
@@ -148,7 +148,7 @@ static int radeon_atpx_verify_interface(struct radeon_atpx *atpx)
 
 	size = *(u16 *) info->buffer.pointer;
 	if (size < 8) {
-		printk("ATPX buffer is too small: %lu\n", size);
+		printk("ATPX buffer is too small: %zu\n", size);
 		err = -EINVAL;
 		goto out;
 	}
diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c
index cb7b7c062fef..41672cc563fb 100644
--- a/drivers/gpu/drm/radeon/radeon_cs.c
+++ b/drivers/gpu/drm/radeon/radeon_cs.c
@@ -478,6 +478,7 @@ static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev,
 	}
 
 out:
+	radeon_vm_add_to_lru(rdev, vm);
 	mutex_unlock(&vm->mutex);
 	mutex_unlock(&rdev->vm_manager.lock);
 	return r;
diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c
index 64a42647f08a..bd13ca09eb62 100644
--- a/drivers/gpu/drm/radeon/radeon_device.c
+++ b/drivers/gpu/drm/radeon/radeon_device.c
@@ -1018,6 +1018,10 @@ int radeon_device_init(struct radeon_device *rdev,
 		return r;
 	/* initialize vm here */
 	mutex_init(&rdev->vm_manager.lock);
+	/* Adjust VM size here.
+	 * Currently set to 4GB ((1 << 20) 4k pages).
+	 * Max GPUVM size for cayman and SI is 40 bits.
+	 */
 	rdev->vm_manager.max_pfn = 1 << 20;
 	INIT_LIST_HEAD(&rdev->vm_manager.lru_vm);
 
diff --git a/drivers/gpu/drm/radeon/radeon_gart.c b/drivers/gpu/drm/radeon/radeon_gart.c
index f0c06d196b75..a7677dd1ce98 100644
--- a/drivers/gpu/drm/radeon/radeon_gart.c
+++ b/drivers/gpu/drm/radeon/radeon_gart.c
@@ -423,6 +423,18 @@ void radeon_gart_fini(struct radeon_device *rdev)
  */
 
 /**
+ * radeon_vm_num_pde - return the number of page directory entries
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Calculate the number of page directory entries (cayman+).
+ */
+static unsigned radeon_vm_num_pdes(struct radeon_device *rdev)
+{
+	return rdev->vm_manager.max_pfn >> RADEON_VM_BLOCK_SIZE;
+}
+
+/**
  * radeon_vm_directory_size - returns the size of the page directory in bytes
  *
  * @rdev: radeon_device pointer
@@ -431,7 +443,7 @@ void radeon_gart_fini(struct radeon_device *rdev)
  */
 static unsigned radeon_vm_directory_size(struct radeon_device *rdev)
 {
-	return (rdev->vm_manager.max_pfn >> RADEON_VM_BLOCK_SIZE) * 8;
+	return RADEON_GPU_PAGE_ALIGN(radeon_vm_num_pdes(rdev) * 8);
 }
 
 /**
@@ -451,11 +463,11 @@ int radeon_vm_manager_init(struct radeon_device *rdev)
 
 	if (!rdev->vm_manager.enabled) {
 		/* allocate enough for 2 full VM pts */
-		size = RADEON_GPU_PAGE_ALIGN(radeon_vm_directory_size(rdev));
-		size += RADEON_GPU_PAGE_ALIGN(rdev->vm_manager.max_pfn * 8);
+		size = radeon_vm_directory_size(rdev);
+		size += rdev->vm_manager.max_pfn * 8;
 		size *= 2;
 		r = radeon_sa_bo_manager_init(rdev, &rdev->vm_manager.sa_manager,
-					      size,
+					      RADEON_GPU_PAGE_ALIGN(size),
 					      RADEON_GEM_DOMAIN_VRAM);
 		if (r) {
 			dev_err(rdev->dev, "failed to allocate vm bo (%dKB)\n",
@@ -476,7 +488,7 @@ int radeon_vm_manager_init(struct radeon_device *rdev)
 
 	/* restore page table */
 	list_for_each_entry(vm, &rdev->vm_manager.lru_vm, list) {
-		if (vm->sa_bo == NULL)
+		if (vm->page_directory == NULL)
 			continue;
 
 		list_for_each_entry(bo_va, &vm->va, vm_list) {
@@ -500,16 +512,25 @@ static void radeon_vm_free_pt(struct radeon_device *rdev,
 				    struct radeon_vm *vm)
 {
 	struct radeon_bo_va *bo_va;
+	int i;
 
-	if (!vm->sa_bo)
+	if (!vm->page_directory)
 		return;
 
 	list_del_init(&vm->list);
-	radeon_sa_bo_free(rdev, &vm->sa_bo, vm->fence);
+	radeon_sa_bo_free(rdev, &vm->page_directory, vm->fence);
 
 	list_for_each_entry(bo_va, &vm->va, vm_list) {
 		bo_va->valid = false;
 	}
+
+	if (vm->page_tables == NULL)
+		return;
+
+	for (i = 0; i < radeon_vm_num_pdes(rdev); i++)
+		radeon_sa_bo_free(rdev, &vm->page_tables[i], vm->fence);
+
+	kfree(vm->page_tables);
 }
 
 /**
@@ -546,63 +567,106 @@ void radeon_vm_manager_fini(struct radeon_device *rdev)
 }
 
 /**
+ * radeon_vm_evict - evict page table to make room for new one
+ *
+ * @rdev: radeon_device pointer
+ * @vm: VM we want to allocate something for
+ *
+ * Evict a VM from the lru, making sure that it isn't @vm. (cayman+).
+ * Returns 0 for success, -ENOMEM for failure.
+ *
+ * Global and local mutex must be locked!
+ */
+int radeon_vm_evict(struct radeon_device *rdev, struct radeon_vm *vm)
+{
+	struct radeon_vm *vm_evict;
+
+	if (list_empty(&rdev->vm_manager.lru_vm))
+		return -ENOMEM;
+
+	vm_evict = list_first_entry(&rdev->vm_manager.lru_vm,
+				    struct radeon_vm, list);
+	if (vm_evict == vm)
+		return -ENOMEM;
+
+	mutex_lock(&vm_evict->mutex);
+	radeon_vm_free_pt(rdev, vm_evict);
+	mutex_unlock(&vm_evict->mutex);
+	return 0;
+}
+
+/**
  * radeon_vm_alloc_pt - allocates a page table for a VM
  *
  * @rdev: radeon_device pointer
  * @vm: vm to bind
  *
  * Allocate a page table for the requested vm (cayman+).
- * Also starts to populate the page table.
  * Returns 0 for success, error for failure.
  *
  * Global and local mutex must be locked!
  */
 int radeon_vm_alloc_pt(struct radeon_device *rdev, struct radeon_vm *vm)
 {
-	struct radeon_vm *vm_evict;
-	int r;
+	unsigned pd_size, pts_size;
 	u64 *pd_addr;
-	int tables_size;
+	int r;
 
 	if (vm == NULL) {
 		return -EINVAL;
 	}
 
-	/* allocate enough to cover the current VM size */
-	tables_size = RADEON_GPU_PAGE_ALIGN(radeon_vm_directory_size(rdev));
-	tables_size += RADEON_GPU_PAGE_ALIGN(vm->last_pfn * 8);
-
-	if (vm->sa_bo != NULL) {
-		/* update lru */
-		list_del_init(&vm->list);
-		list_add_tail(&vm->list, &rdev->vm_manager.lru_vm);
+	if (vm->page_directory != NULL) {
 		return 0;
 	}
 
 retry:
-	r = radeon_sa_bo_new(rdev, &rdev->vm_manager.sa_manager, &vm->sa_bo,
-			     tables_size, RADEON_GPU_PAGE_SIZE, false);
+	pd_size = RADEON_GPU_PAGE_ALIGN(radeon_vm_directory_size(rdev));
+	r = radeon_sa_bo_new(rdev, &rdev->vm_manager.sa_manager,
+			     &vm->page_directory, pd_size,
+			     RADEON_GPU_PAGE_SIZE, false);
 	if (r == -ENOMEM) {
-		if (list_empty(&rdev->vm_manager.lru_vm)) {
+		r = radeon_vm_evict(rdev, vm);
+		if (r)
 			return r;
-		}
-		vm_evict = list_first_entry(&rdev->vm_manager.lru_vm, struct radeon_vm, list);
-		mutex_lock(&vm_evict->mutex);
-		radeon_vm_free_pt(rdev, vm_evict);
-		mutex_unlock(&vm_evict->mutex);
 		goto retry;
 
 	} else if (r) {
 		return r;
 	}
 
-	pd_addr = radeon_sa_bo_cpu_addr(vm->sa_bo);
-	vm->pd_gpu_addr = radeon_sa_bo_gpu_addr(vm->sa_bo);
-	memset(pd_addr, 0, tables_size);
+	vm->pd_gpu_addr = radeon_sa_bo_gpu_addr(vm->page_directory);
+
+	/* Initially clear the page directory */
+	pd_addr = radeon_sa_bo_cpu_addr(vm->page_directory);
+	memset(pd_addr, 0, pd_size);
+
+	pts_size = radeon_vm_num_pdes(rdev) * sizeof(struct radeon_sa_bo *);
+	vm->page_tables = kzalloc(pts_size, GFP_KERNEL);
+
+	if (vm->page_tables == NULL) {
+		DRM_ERROR("Cannot allocate memory for page table array\n");
+		radeon_sa_bo_free(rdev, &vm->page_directory, vm->fence);
+		return -ENOMEM;
+	}
+
+	return 0;
+}
 
+/**
+ * radeon_vm_add_to_lru - add VMs page table to LRU list
+ *
+ * @rdev: radeon_device pointer
+ * @vm: vm to add to LRU
+ *
+ * Add the allocated page table to the LRU list (cayman+).
+ *
+ * Global mutex must be locked!
+ */
+void radeon_vm_add_to_lru(struct radeon_device *rdev, struct radeon_vm *vm)
+{
+	list_del_init(&vm->list);
 	list_add_tail(&vm->list, &rdev->vm_manager.lru_vm);
-	return radeon_vm_bo_update_pte(rdev, vm, rdev->ring_tmp_bo.bo,
-				       &rdev->ring_tmp_bo.bo->tbo.mem);
 }
 
 /**
@@ -793,20 +857,6 @@ int radeon_vm_bo_set_addr(struct radeon_device *rdev,
 	}
 
 	mutex_lock(&vm->mutex);
-	if (last_pfn > vm->last_pfn) {
-		/* release mutex and lock in right order */
-		mutex_unlock(&vm->mutex);
-		mutex_lock(&rdev->vm_manager.lock);
-		mutex_lock(&vm->mutex);
-		/* and check again */
-		if (last_pfn > vm->last_pfn) {
-			/* grow va space 32M by 32M */
-			unsigned align = ((32 << 20) >> 12) - 1;
-			radeon_vm_free_pt(rdev, vm);
-			vm->last_pfn = (last_pfn + align) & ~align;
-		}
-		mutex_unlock(&rdev->vm_manager.lock);
-	}
 	head = &vm->va;
 	last_offset = 0;
 	list_for_each_entry(tmp, &vm->va, vm_list) {
@@ -865,6 +915,155 @@ uint64_t radeon_vm_map_gart(struct radeon_device *rdev, uint64_t addr)
 }
 
 /**
+ * radeon_vm_update_pdes - make sure that page directory is valid
+ *
+ * @rdev: radeon_device pointer
+ * @vm: requested vm
+ * @start: start of GPU address range
+ * @end: end of GPU address range
+ *
+ * Allocates new page tables if necessary
+ * and updates the page directory (cayman+).
+ * Returns 0 for success, error for failure.
+ *
+ * Global and local mutex must be locked!
+ */
+static int radeon_vm_update_pdes(struct radeon_device *rdev,
+				 struct radeon_vm *vm,
+				 uint64_t start, uint64_t end)
+{
+	static const uint32_t incr = RADEON_VM_PTE_COUNT * 8;
+
+	uint64_t last_pde = ~0, last_pt = ~0;
+	unsigned count = 0;
+	uint64_t pt_idx;
+	int r;
+
+	start = (start / RADEON_GPU_PAGE_SIZE) >> RADEON_VM_BLOCK_SIZE;
+	end = (end / RADEON_GPU_PAGE_SIZE) >> RADEON_VM_BLOCK_SIZE;
+
+	/* walk over the address space and update the page directory */
+	for (pt_idx = start; pt_idx <= end; ++pt_idx) {
+		uint64_t pde, pt;
+
+		if (vm->page_tables[pt_idx])
+			continue;
+
+retry:
+		r = radeon_sa_bo_new(rdev, &rdev->vm_manager.sa_manager,
+				     &vm->page_tables[pt_idx],
+				     RADEON_VM_PTE_COUNT * 8,
+				     RADEON_GPU_PAGE_SIZE, false);
+
+		if (r == -ENOMEM) {
+			r = radeon_vm_evict(rdev, vm);
+			if (r)
+				return r;
+			goto retry;
+		} else if (r) {
+			return r;
+		}
+
+		pde = vm->pd_gpu_addr + pt_idx * 8;
+
+		pt = radeon_sa_bo_gpu_addr(vm->page_tables[pt_idx]);
+
+		if (((last_pde + 8 * count) != pde) ||
+		    ((last_pt + incr * count) != pt)) {
+
+			if (count) {
+				radeon_asic_vm_set_page(rdev, last_pde,
+							last_pt, count, incr,
+							RADEON_VM_PAGE_VALID);
+			}
+
+			count = 1;
+			last_pde = pde;
+			last_pt = pt;
+		} else {
+			++count;
+		}
+	}
+
+	if (count) {
+		radeon_asic_vm_set_page(rdev, last_pde, last_pt, count,
+					incr, RADEON_VM_PAGE_VALID);
+
+	}
+
+	return 0;
+}
+
+/**
+ * radeon_vm_update_ptes - make sure that page tables are valid
+ *
+ * @rdev: radeon_device pointer
+ * @vm: requested vm
+ * @start: start of GPU address range
+ * @end: end of GPU address range
+ * @dst: destination address to map to
+ * @flags: mapping flags
+ *
+ * Update the page tables in the range @start - @end (cayman+).
+ *
+ * Global and local mutex must be locked!
+ */
+static void radeon_vm_update_ptes(struct radeon_device *rdev,
+				  struct radeon_vm *vm,
+				  uint64_t start, uint64_t end,
+				  uint64_t dst, uint32_t flags)
+{
+	static const uint64_t mask = RADEON_VM_PTE_COUNT - 1;
+
+	uint64_t last_pte = ~0, last_dst = ~0;
+	unsigned count = 0;
+	uint64_t addr;
+
+	start = start / RADEON_GPU_PAGE_SIZE;
+	end = end / RADEON_GPU_PAGE_SIZE;
+
+	/* walk over the address space and update the page tables */
+	for (addr = start; addr < end; ) {
+		uint64_t pt_idx = addr >> RADEON_VM_BLOCK_SIZE;
+		unsigned nptes;
+		uint64_t pte;
+
+		if ((addr & ~mask) == (end & ~mask))
+			nptes = end - addr;
+		else
+			nptes = RADEON_VM_PTE_COUNT - (addr & mask);
+
+		pte = radeon_sa_bo_gpu_addr(vm->page_tables[pt_idx]);
+		pte += (addr & mask) * 8;
+
+		if (((last_pte + 8 * count) != pte) ||
+		    ((count + nptes) > 1 << 11)) {
+
+			if (count) {
+				radeon_asic_vm_set_page(rdev, last_pte,
+							last_dst, count,
+							RADEON_GPU_PAGE_SIZE,
+							flags);
+			}
+
+			count = nptes;
+			last_pte = pte;
+			last_dst = dst;
+		} else {
+			count += nptes;
+		}
+
+		addr += nptes;
+		dst += nptes * RADEON_GPU_PAGE_SIZE;
+	}
+
+	if (count) {
+		radeon_asic_vm_set_page(rdev, last_pte,	last_dst, count,
+					RADEON_GPU_PAGE_SIZE, flags);
+	}
+}
+
+/**
  * radeon_vm_bo_update_pte - map a bo into the vm page table
  *
  * @rdev: radeon_device pointer
@@ -887,12 +1086,11 @@ int radeon_vm_bo_update_pte(struct radeon_device *rdev,
 	struct radeon_semaphore *sem = NULL;
 	struct radeon_bo_va *bo_va;
 	unsigned nptes, npdes, ndw;
-	uint64_t pe, addr;
-	uint64_t pfn;
+	uint64_t addr;
 	int r;
 
 	/* nothing to do if vm isn't bound */
-	if (vm->sa_bo == NULL)
+	if (vm->page_directory == NULL)
 		return 0;
 
 	bo_va = radeon_vm_bo_find(vm, bo);
@@ -939,25 +1137,29 @@ int radeon_vm_bo_update_pte(struct radeon_device *rdev,
 		}
 	}
 
-	/* estimate number of dw needed */
-	/* reserve space for 32-bit padding */
-	ndw = 32;
-
 	nptes = radeon_bo_ngpu_pages(bo);
 
-	pfn = (bo_va->soffset / RADEON_GPU_PAGE_SIZE);
+	/* assume two extra pdes in case the mapping overlaps the borders */
+	npdes = (nptes >> RADEON_VM_BLOCK_SIZE) + 2;
 
-	/* handle cases where a bo spans several pdes  */
-	npdes = (ALIGN(pfn + nptes, RADEON_VM_PTE_COUNT) -
-		 (pfn & ~(RADEON_VM_PTE_COUNT - 1))) >> RADEON_VM_BLOCK_SIZE;
+	/* estimate number of dw needed */
+	/* semaphore, fence and padding */
+	ndw = 32;
+
+	if (RADEON_VM_BLOCK_SIZE > 11)
+		/* reserve space for one header for every 2k dwords */
+		ndw += (nptes >> 11) * 3;
+	else
+		/* reserve space for one header for
+		    every (1 << BLOCK_SIZE) entries */
+		ndw += (nptes >> RADEON_VM_BLOCK_SIZE) * 3;
 
-	/* reserve space for one header for every 2k dwords */
-	ndw += (nptes >> 11) * 3;
 	/* reserve space for pte addresses */
 	ndw += nptes * 2;
 
 	/* reserve space for one header for every 2k dwords */
 	ndw += (npdes >> 11) * 3;
+
 	/* reserve space for pde addresses */
 	ndw += npdes * 2;
 
@@ -971,22 +1173,14 @@ int radeon_vm_bo_update_pte(struct radeon_device *rdev,
 		radeon_fence_note_sync(vm->fence, ridx);
 	}
 
-	/* update page table entries */
-	pe = vm->pd_gpu_addr;
-	pe += radeon_vm_directory_size(rdev);
-	pe += (bo_va->soffset / RADEON_GPU_PAGE_SIZE) * 8;
-
-	radeon_asic_vm_set_page(rdev, pe, addr, nptes,
-				RADEON_GPU_PAGE_SIZE, bo_va->flags);
-
-	/* update page directory entries */
-	addr = pe;
-
-	pe = vm->pd_gpu_addr;
-	pe += ((bo_va->soffset / RADEON_GPU_PAGE_SIZE) >> RADEON_VM_BLOCK_SIZE) * 8;
+	r = radeon_vm_update_pdes(rdev, vm, bo_va->soffset, bo_va->eoffset);
+	if (r) {
+		radeon_ring_unlock_undo(rdev, ring);
+		return r;
+	}
 
-	radeon_asic_vm_set_page(rdev, pe, addr, npdes,
-				RADEON_VM_PTE_COUNT * 8, RADEON_VM_PAGE_VALID);
+	radeon_vm_update_ptes(rdev, vm, bo_va->soffset, bo_va->eoffset,
+			      addr, bo_va->flags);
 
 	radeon_fence_unref(&vm->fence);
 	r = radeon_fence_emit(rdev, &vm->fence, ridx);
@@ -997,6 +1191,7 @@ int radeon_vm_bo_update_pte(struct radeon_device *rdev,
 	radeon_ring_unlock_commit(rdev, ring);
 	radeon_semaphore_free(rdev, &sem, vm->fence);
 	radeon_fence_unref(&vm->last_flush);
+
 	return 0;
 }
 
@@ -1056,31 +1251,15 @@ void radeon_vm_bo_invalidate(struct radeon_device *rdev,
  * @rdev: radeon_device pointer
  * @vm: requested vm
  *
- * Init @vm (cayman+).
- * Map the IB pool and any other shared objects into the VM
- * by default as it's used by all VMs.
- * Returns 0 for success, error for failure.
+ * Init @vm fields (cayman+).
  */
-int radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm)
+void radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm)
 {
-	struct radeon_bo_va *bo_va;
-	int r;
-
 	vm->id = 0;
 	vm->fence = NULL;
-	vm->last_pfn = 0;
 	mutex_init(&vm->mutex);
 	INIT_LIST_HEAD(&vm->list);
 	INIT_LIST_HEAD(&vm->va);
-
-	/* map the ib pool buffer at 0 in virtual address space, set
-	 * read only
-	 */
-	bo_va = radeon_vm_bo_add(rdev, vm, rdev->ring_tmp_bo.bo);
-	r = radeon_vm_bo_set_addr(rdev, bo_va, RADEON_VA_IB_OFFSET,
-				  RADEON_VM_PAGE_READABLE |
-				  RADEON_VM_PAGE_SNOOPED);
-	return r;
 }
 
 /**
@@ -1102,17 +1281,6 @@ void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm)
 	radeon_vm_free_pt(rdev, vm);
 	mutex_unlock(&rdev->vm_manager.lock);
 
-	/* remove all bo at this point non are busy any more because unbind
-	 * waited for the last vm fence to signal
-	 */
-	r = radeon_bo_reserve(rdev->ring_tmp_bo.bo, false);
-	if (!r) {
-		bo_va = radeon_vm_bo_find(vm, rdev->ring_tmp_bo.bo);
-		list_del_init(&bo_va->bo_list);
-		list_del_init(&bo_va->vm_list);
-		radeon_bo_unreserve(rdev->ring_tmp_bo.bo);
-		kfree(bo_va);
-	}
 	if (!list_empty(&vm->va)) {
 		dev_err(rdev->dev, "still active bo inside vm\n");
 	}
diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
index 83b8d8aa71c0..dc781c49b96b 100644
--- a/drivers/gpu/drm/radeon/radeon_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_kms.c
@@ -419,6 +419,7 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
 	/* new gpu have virtual address space support */
 	if (rdev->family >= CHIP_CAYMAN) {
 		struct radeon_fpriv *fpriv;
+		struct radeon_bo_va *bo_va;
 		int r;
 
 		fpriv = kzalloc(sizeof(*fpriv), GFP_KERNEL);
@@ -426,7 +427,15 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
 			return -ENOMEM;
 		}
 
-		r = radeon_vm_init(rdev, &fpriv->vm);
+		radeon_vm_init(rdev, &fpriv->vm);
+
+		/* map the ib pool buffer read only into
+		 * virtual address space */
+		bo_va = radeon_vm_bo_add(rdev, &fpriv->vm,
+					 rdev->ring_tmp_bo.bo);
+		r = radeon_vm_bo_set_addr(rdev, bo_va, RADEON_VA_IB_OFFSET,
+					  RADEON_VM_PAGE_READABLE |
+					  RADEON_VM_PAGE_SNOOPED);
 		if (r) {
 			radeon_vm_fini(rdev, &fpriv->vm);
 			kfree(fpriv);
@@ -454,6 +463,17 @@ void radeon_driver_postclose_kms(struct drm_device *dev,
 	/* new gpu have virtual address space support */
 	if (rdev->family >= CHIP_CAYMAN && file_priv->driver_priv) {
 		struct radeon_fpriv *fpriv = file_priv->driver_priv;
+		struct radeon_bo_va *bo_va;
+		int r;
+
+		r = radeon_bo_reserve(rdev->ring_tmp_bo.bo, false);
+		if (!r) {
+			bo_va = radeon_vm_bo_find(&fpriv->vm,
+						  rdev->ring_tmp_bo.bo);
+			if (bo_va)
+				radeon_vm_bo_rmv(rdev, bo_va);
+			radeon_bo_unreserve(rdev->ring_tmp_bo.bo);
+		}
 
 		radeon_vm_fini(rdev, &fpriv->vm);
 		kfree(fpriv);
diff --git a/drivers/gpu/drm/radeon/radeon_legacy_encoders.c b/drivers/gpu/drm/radeon/radeon_legacy_encoders.c
index 92487e614778..a13ad9d707cf 100644
--- a/drivers/gpu/drm/radeon/radeon_legacy_encoders.c
+++ b/drivers/gpu/drm/radeon/radeon_legacy_encoders.c
@@ -269,27 +269,6 @@ static const struct drm_encoder_helper_funcs radeon_legacy_lvds_helper_funcs = {
 	.disable = radeon_legacy_encoder_disable,
 };
 
-#if defined(CONFIG_BACKLIGHT_CLASS_DEVICE) || defined(CONFIG_BACKLIGHT_CLASS_DEVICE_MODULE)
-
-static uint8_t radeon_legacy_lvds_level(struct backlight_device *bd)
-{
-	struct radeon_backlight_privdata *pdata = bl_get_data(bd);
-	uint8_t level;
-
-	/* Convert brightness to hardware level */
-	if (bd->props.brightness < 0)
-		level = 0;
-	else if (bd->props.brightness > RADEON_MAX_BL_LEVEL)
-		level = RADEON_MAX_BL_LEVEL;
-	else
-		level = bd->props.brightness;
-
-	if (pdata->negative)
-		level = RADEON_MAX_BL_LEVEL - level;
-
-	return level;
-}
-
 u8
 radeon_legacy_get_backlight_level(struct radeon_encoder *radeon_encoder)
 {
@@ -331,6 +310,27 @@ radeon_legacy_set_backlight_level(struct radeon_encoder *radeon_encoder, u8 leve
 	radeon_legacy_lvds_update(&radeon_encoder->base, dpms_mode);
 }
 
+#if defined(CONFIG_BACKLIGHT_CLASS_DEVICE) || defined(CONFIG_BACKLIGHT_CLASS_DEVICE_MODULE)
+
+static uint8_t radeon_legacy_lvds_level(struct backlight_device *bd)
+{
+	struct radeon_backlight_privdata *pdata = bl_get_data(bd);
+	uint8_t level;
+
+	/* Convert brightness to hardware level */
+	if (bd->props.brightness < 0)
+		level = 0;
+	else if (bd->props.brightness > RADEON_MAX_BL_LEVEL)
+		level = RADEON_MAX_BL_LEVEL;
+	else
+		level = bd->props.brightness;
+
+	if (pdata->negative)
+		level = RADEON_MAX_BL_LEVEL - level;
+
+	return level;
+}
+
 static int radeon_legacy_backlight_update_status(struct backlight_device *bd)
 {
 	struct radeon_backlight_privdata *pdata = bl_get_data(bd);
@@ -991,11 +991,7 @@ static void radeon_legacy_tmds_ext_mode_set(struct drm_encoder *encoder,
 static void radeon_ext_tmds_enc_destroy(struct drm_encoder *encoder)
 {
 	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	struct radeon_encoder_ext_tmds *tmds = radeon_encoder->enc_priv;
-	if (tmds) {
-		if (tmds->i2c_bus)
-			radeon_i2c_destroy(tmds->i2c_bus);
-	}
+	/* don't destroy the i2c bus record here, this will be done in radeon_i2c_fini */
 	kfree(radeon_encoder->enc_priv);
 	drm_encoder_cleanup(encoder);
 	kfree(radeon_encoder);
diff --git a/drivers/gpu/drm/radeon/radeon_ring.c b/drivers/gpu/drm/radeon/radeon_ring.c
index bba66902c83b..47634f27f2e5 100644
--- a/drivers/gpu/drm/radeon/radeon_ring.c
+++ b/drivers/gpu/drm/radeon/radeon_ring.c
@@ -305,7 +305,7 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v)
 {
 #if DRM_DEBUG_CODE
 	if (ring->count_dw <= 0) {
-		DRM_ERROR("radeon: writting more dword to ring than expected !\n");
+		DRM_ERROR("radeon: writing more dwords to the ring than expected!\n");
 	}
 #endif
 	ring->ring[ring->wptr++] = v;
diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c
index f79633a036c3..df8dd7701643 100644
--- a/drivers/gpu/drm/radeon/si.c
+++ b/drivers/gpu/drm/radeon/si.c
@@ -2407,12 +2407,13 @@ static int si_pcie_gart_enable(struct radeon_device *rdev)
 	WREG32(0x15DC, 0);
 
 	/* empty context1-15 */
-	/* FIXME start with 4G, once using 2 level pt switch to full
-	 * vm size space
-	 */
 	/* set vm size, must be a multiple of 4 */
 	WREG32(VM_CONTEXT1_PAGE_TABLE_START_ADDR, 0);
 	WREG32(VM_CONTEXT1_PAGE_TABLE_END_ADDR, rdev->vm_manager.max_pfn);
+	/* Assign the pt base to something valid for now; the pts used for
+	 * the VMs are determined by the application and setup and assigned
+	 * on the fly in the vm part of radeon_gart.c
+	 */
 	for (i = 1; i < 16; i++) {
 		if (i < 8)
 			WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (i << 2),