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

Pull drm updates from Dave Airlie:
 "This is a combo of -next and some -fixes that came in in the
  intervening time.

  Highlights:

  New drivers:
    ARM Armada driver for Marvell Armada 510 SOCs

  Intel:
    Broadwell initial support under a default off switch,
    Stereo/3D HDMI mode support
    Valleyview improvements
    Displayport improvements
    Haswell fixes
    initial mipi dsi panel support
    CRC support for debugging
    build with CONFIG_FB=n

  Radeon:
    enable DPM on a number of GPUs by default
    secondary GPU powerdown support
    enable HDMI audio by default
    Hawaii support

  Nouveau:
    dynamic pm code infrastructure reworked, does nothing major yet
    GK208 modesetting support
    MSI fixes, on by default again
    PMPEG improvements
    pageflipping fixes

  GMA500:
    minnowboard SDVO support

  VMware:
    misc fixes

  MSM:
    prime, plane and rendernodes support

  Tegra:
    rearchitected to put the drm driver into the drm subsystem.
    HDMI and gr2d support for tegra 114 SoC

  QXL:
    oops fix, and multi-head fixes

  DRM core:
    sysfs lifetime fixes
    client capability ioctl
    further cleanups to device midlayer
    more vblank timestamp fixes"

* 'drm-next' of git://people.freedesktop.org/~airlied/linux: (789 commits)
  drm/nouveau: do not map evicted vram buffers in nouveau_bo_vma_add
  drm/nvc0-/gr: shift wrapping bug in nvc0_grctx_generate_r406800
  drm/nouveau/pwr: fix missing mutex unlock in a failure path
  drm/nv40/therm: fix slowing down fan when pstate undefined
  drm/nv11-: synchronise flips to vblank, unless async flip requested
  drm/nvc0-: remove nasty fifo swmthd hack for flip completion method
  drm/nv10-: we no longer need to create nvsw object on user channels
  drm/nouveau: always queue flips relative to kernel channel activity
  drm/nouveau: there is no need to reserve/fence the new fb when flipping
  drm/nouveau: when bailing out of a pushbuf ioctl, do not remove previous fence
  drm/nouveau: allow nouveau_fence_ref() to be a noop
  drm/nvc8/mc: msi rearm is via the nvc0 method
  drm/ttm: Fix vma page_prot bit manipulation
  drm/vmwgfx: Fix a couple of compile / sparse warnings and errors
  drm/vmwgfx: Resource evict fixes
  drm/edid: compare actual vrefresh for all modes for quirks
  drm: shmob_drm: Convert to clk_prepare/unprepare
  drm/nouveau: fix 32-bit build
  drm/i915/opregion: fix build error on CONFIG_ACPI=n
  Revert "drm/radeon/audio: don't set speaker allocation on DCE4+"
  ...

5 files changed, 119 insertions, 58 deletions

diff --git a/drivers/gpu/drm/ttm/Makefile b/drivers/gpu/drm/ttm/Makefile
index b2b33dd..b433b9f 100644
--- a/drivers/gpu/drm/ttm/Makefile
+++ b/drivers/gpu/drm/ttm/Makefile
@@ -5,10 +5,6 @@ ccflags-y := -Iinclude/drm
 ttm-y := ttm_agp_backend.o ttm_memory.o ttm_tt.o ttm_bo.o \
 	ttm_bo_util.o ttm_bo_vm.o ttm_module.o \
 	ttm_object.o ttm_lock.o ttm_execbuf_util.o ttm_page_alloc.o \
-	ttm_bo_manager.o
-
-ifeq ($(CONFIG_SWIOTLB),y)
-ttm-y += ttm_page_alloc_dma.o
-endif
+	ttm_bo_manager.o ttm_page_alloc_dma.o
 
 obj-$(CONFIG_DRM_TTM) += ttm.o
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index f1a857e..8d5a646 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -429,8 +429,20 @@ static void ttm_bo_cleanup_refs_or_queue(struct ttm_buffer_object *bo)
 		sync_obj = driver->sync_obj_ref(bo->sync_obj);
 	spin_unlock(&bdev->fence_lock);
 
-	if (!ret)
+	if (!ret) {
+
+		/*
+		 * Make NO_EVICT bos immediately available to
+		 * shrinkers, now that they are queued for
+		 * destruction.
+		 */
+		if (bo->mem.placement & TTM_PL_FLAG_NO_EVICT) {
+			bo->mem.placement &= ~TTM_PL_FLAG_NO_EVICT;
+			ttm_bo_add_to_lru(bo);
+		}
+
 		ww_mutex_unlock(&bo->resv->lock);
+	}
 
 	kref_get(&bo->list_kref);
 	list_add_tail(&bo->ddestroy, &bdev->ddestroy);
@@ -986,24 +998,32 @@ out_unlock:
 	return ret;
 }
 
-static int ttm_bo_mem_compat(struct ttm_placement *placement,
-			     struct ttm_mem_reg *mem)
+static bool ttm_bo_mem_compat(struct ttm_placement *placement,
+			      struct ttm_mem_reg *mem,
+			      uint32_t *new_flags)
 {
 	int i;
 
 	if (mem->mm_node && placement->lpfn != 0 &&
 	    (mem->start < placement->fpfn ||
 	     mem->start + mem->num_pages > placement->lpfn))
-		return -1;
+		return false;
 
 	for (i = 0; i < placement->num_placement; i++) {
-		if ((placement->placement[i] & mem->placement &
-			TTM_PL_MASK_CACHING) &&
-			(placement->placement[i] & mem->placement &
-			TTM_PL_MASK_MEM))
-			return i;
+		*new_flags = placement->placement[i];
+		if ((*new_flags & mem->placement & TTM_PL_MASK_CACHING) &&
+		    (*new_flags & mem->placement & TTM_PL_MASK_MEM))
+			return true;
 	}
-	return -1;
+
+	for (i = 0; i < placement->num_busy_placement; i++) {
+		*new_flags = placement->busy_placement[i];
+		if ((*new_flags & mem->placement & TTM_PL_MASK_CACHING) &&
+		    (*new_flags & mem->placement & TTM_PL_MASK_MEM))
+			return true;
+	}
+
+	return false;
 }
 
 int ttm_bo_validate(struct ttm_buffer_object *bo,
@@ -1012,6 +1032,7 @@ int ttm_bo_validate(struct ttm_buffer_object *bo,
 			bool no_wait_gpu)
 {
 	int ret;
+	uint32_t new_flags;
 
 	lockdep_assert_held(&bo->resv->lock.base);
 	/* Check that range is valid */
@@ -1022,8 +1043,7 @@ int ttm_bo_validate(struct ttm_buffer_object *bo,
 	/*
 	 * Check whether we need to move buffer.
 	 */
-	ret = ttm_bo_mem_compat(placement, &bo->mem);
-	if (ret < 0) {
+	if (!ttm_bo_mem_compat(placement, &bo->mem, &new_flags)) {
 		ret = ttm_bo_move_buffer(bo, placement, interruptible,
 					 no_wait_gpu);
 		if (ret)
@@ -1033,7 +1053,7 @@ int ttm_bo_validate(struct ttm_buffer_object *bo,
 		 * Use the access and other non-mapping-related flag bits from
 		 * the compatible memory placement flags to the active flags
 		 */
-		ttm_flag_masked(&bo->mem.placement, placement->placement[ret],
+		ttm_flag_masked(&bo->mem.placement, new_flags,
 				~TTM_PL_MASK_MEMTYPE);
 	}
 	/*
diff --git a/drivers/gpu/drm/ttm/ttm_bo_util.c b/drivers/gpu/drm/ttm/ttm_bo_util.c
index 7cc904d..4834c46 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_util.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_util.c
@@ -343,19 +343,25 @@ int ttm_bo_move_memcpy(struct ttm_buffer_object *bo,
 	if (ret)
 		goto out;
 
+	/*
+	 * Single TTM move. NOP.
+	 */
 	if (old_iomap == NULL && new_iomap == NULL)
 		goto out2;
+
+	/*
+	 * Move nonexistent data. NOP.
+	 */
 	if (old_iomap == NULL && ttm == NULL)
 		goto out2;
 
-	if (ttm->state == tt_unpopulated) {
+	/*
+	 * TTM might be null for moves within the same region.
+	 */
+	if (ttm && ttm->state == tt_unpopulated) {
 		ret = ttm->bdev->driver->ttm_tt_populate(ttm);
-		if (ret) {
-			/* if we fail here don't nuke the mm node
-			 * as the bo still owns it */
-			old_copy.mm_node = NULL;
+		if (ret)
 			goto out1;
-		}
 	}
 
 	add = 0;
@@ -381,11 +387,8 @@ int ttm_bo_move_memcpy(struct ttm_buffer_object *bo,
 						   prot);
 		} else
 			ret = ttm_copy_io_page(new_iomap, old_iomap, page);
-		if (ret) {
-			/* failing here, means keep old copy as-is */
-			old_copy.mm_node = NULL;
+		if (ret)
 			goto out1;
-		}
 	}
 	mb();
 out2:
@@ -403,7 +406,12 @@ out1:
 	ttm_mem_reg_iounmap(bdev, old_mem, new_iomap);
 out:
 	ttm_mem_reg_iounmap(bdev, &old_copy, old_iomap);
-	ttm_bo_mem_put(bo, &old_copy);
+
+	/*
+	 * On error, keep the mm node!
+	 */
+	if (!ret)
+		ttm_bo_mem_put(bo, &old_copy);
 	return ret;
 }
 EXPORT_SYMBOL(ttm_bo_move_memcpy);
diff --git a/drivers/gpu/drm/ttm/ttm_bo_vm.c b/drivers/gpu/drm/ttm/ttm_bo_vm.c
index 1006c15..ac617f3 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_vm.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_vm.c
@@ -41,6 +41,51 @@
 
 #define TTM_BO_VM_NUM_PREFAULT 16
 
+static int ttm_bo_vm_fault_idle(struct ttm_buffer_object *bo,
+				struct vm_area_struct *vma,
+				struct vm_fault *vmf)
+{
+	struct ttm_bo_device *bdev = bo->bdev;
+	int ret = 0;
+
+	spin_lock(&bdev->fence_lock);
+	if (likely(!test_bit(TTM_BO_PRIV_FLAG_MOVING, &bo->priv_flags)))
+		goto out_unlock;
+
+	/*
+	 * Quick non-stalling check for idle.
+	 */
+	ret = ttm_bo_wait(bo, false, false, true);
+	if (likely(ret == 0))
+		goto out_unlock;
+
+	/*
+	 * If possible, avoid waiting for GPU with mmap_sem
+	 * held.
+	 */
+	if (vmf->flags & FAULT_FLAG_ALLOW_RETRY) {
+		ret = VM_FAULT_RETRY;
+		if (vmf->flags & FAULT_FLAG_RETRY_NOWAIT)
+			goto out_unlock;
+
+		up_read(&vma->vm_mm->mmap_sem);
+		(void) ttm_bo_wait(bo, false, true, false);
+		goto out_unlock;
+	}
+
+	/*
+	 * Ordinary wait.
+	 */
+	ret = ttm_bo_wait(bo, false, true, false);
+	if (unlikely(ret != 0))
+		ret = (ret != -ERESTARTSYS) ? VM_FAULT_SIGBUS :
+			VM_FAULT_NOPAGE;
+
+out_unlock:
+	spin_unlock(&bdev->fence_lock);
+	return ret;
+}
+
 static int ttm_bo_vm_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 {
 	struct ttm_buffer_object *bo = (struct ttm_buffer_object *)
@@ -57,6 +102,7 @@ static int ttm_bo_vm_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 	int retval = VM_FAULT_NOPAGE;
 	struct ttm_mem_type_manager *man =
 		&bdev->man[bo->mem.mem_type];
+	struct vm_area_struct cvma;
 
 	/*
 	 * Work around locking order reversal in fault / nopfn
@@ -91,18 +137,11 @@ static int ttm_bo_vm_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 	 * Wait for buffer data in transit, due to a pipelined
 	 * move.
 	 */
-
-	spin_lock(&bdev->fence_lock);
-	if (test_bit(TTM_BO_PRIV_FLAG_MOVING, &bo->priv_flags)) {
-		ret = ttm_bo_wait(bo, false, true, false);
-		spin_unlock(&bdev->fence_lock);
-		if (unlikely(ret != 0)) {
-			retval = (ret != -ERESTARTSYS) ?
-			    VM_FAULT_SIGBUS : VM_FAULT_NOPAGE;
-			goto out_unlock;
-		}
-	} else
-		spin_unlock(&bdev->fence_lock);
+	ret = ttm_bo_vm_fault_idle(bo, vma, vmf);
+	if (unlikely(ret != 0)) {
+		retval = ret;
+		goto out_unlock;
+	}
 
 	ret = ttm_mem_io_lock(man, true);
 	if (unlikely(ret != 0)) {
@@ -126,26 +165,21 @@ static int ttm_bo_vm_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 	}
 
 	/*
-	 * Strictly, we're not allowed to modify vma->vm_page_prot here,
-	 * since the mmap_sem is only held in read mode. However, we
-	 * modify only the caching bits of vma->vm_page_prot and
-	 * consider those bits protected by
-	 * the bo->mutex, as we should be the only writers.
-	 * There shouldn't really be any readers of these bits except
-	 * within vm_insert_mixed()? fork?
-	 *
-	 * TODO: Add a list of vmas to the bo, and change the
-	 * vma->vm_page_prot when the object changes caching policy, with
-	 * the correct locks held.
+	 * Make a local vma copy to modify the page_prot member
+	 * and vm_flags if necessary. The vma parameter is protected
+	 * by mmap_sem in write mode.
 	 */
+	cvma = *vma;
+	cvma.vm_page_prot = vm_get_page_prot(cvma.vm_flags);
+
 	if (bo->mem.bus.is_iomem) {
-		vma->vm_page_prot = ttm_io_prot(bo->mem.placement,
-						vma->vm_page_prot);
+		cvma.vm_page_prot = ttm_io_prot(bo->mem.placement,
+						cvma.vm_page_prot);
 	} else {
 		ttm = bo->ttm;
-		vma->vm_page_prot = (bo->mem.placement & TTM_PL_FLAG_CACHED) ?
-		    vm_get_page_prot(vma->vm_flags) :
-		    ttm_io_prot(bo->mem.placement, vma->vm_page_prot);
+		if (!(bo->mem.placement & TTM_PL_FLAG_CACHED))
+			cvma.vm_page_prot = ttm_io_prot(bo->mem.placement,
+							cvma.vm_page_prot);
 
 		/* Allocate all page at once, most common usage */
 		if (ttm->bdev->driver->ttm_tt_populate(ttm)) {
@@ -172,7 +206,7 @@ static int ttm_bo_vm_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 			pfn = page_to_pfn(page);
 		}
 
-		ret = vm_insert_mixed(vma, address, pfn);
+		ret = vm_insert_mixed(&cvma, address, pfn);
 		/*
 		 * Somebody beat us to this PTE or prefaulting to
 		 * an already populated PTE, or prefaulting error.
diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c b/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c
index 7957bee..fb8259f 100644
--- a/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c
+++ b/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c
@@ -33,6 +33,7 @@
  *   when freed).
  */
 
+#if defined(CONFIG_SWIOTLB) || defined(CONFIG_INTEL_IOMMU)
 #define pr_fmt(fmt) "[TTM] " fmt
 
 #include <linux/dma-mapping.h>
@@ -1142,3 +1143,5 @@ int ttm_dma_page_alloc_debugfs(struct seq_file *m, void *data)
 	return 0;
 }
 EXPORT_SYMBOL_GPL(ttm_dma_page_alloc_debugfs);
+
+#endif