7 files changed, 92 insertions, 22 deletions

diff --git a/drivers/char/drm/ati_pcigart.c b/drivers/char/drm/ati_pcigart.c
index 35d25d821c38..141f4dfa0a11 100644
--- a/drivers/char/drm/ati_pcigart.c
+++ b/drivers/char/drm/ati_pcigart.c
@@ -168,6 +168,12 @@ int drm_ati_pcigart_init(struct drm_device *dev, struct drm_ati_pcigart_info *ga
                 }
         }
 
+        if (gart_info->gart_table_location == DRM_ATI_GART_MAIN)
+                dma_sync_single_for_device(&dev->pdev->dev,
+                                           bus_address,
+                                           max_pages * sizeof(u32),
+                                           PCI_DMA_TODEVICE);
+
         ret = 1;
 
 #if defined(__i386__) || defined(__x86_64__)
diff --git a/drivers/char/drm/drm_scatter.c b/drivers/char/drm/drm_scatter.c
index 26d8f675ed5d..b2b0f3d41714 100644
--- a/drivers/char/drm/drm_scatter.c
+++ b/drivers/char/drm/drm_scatter.c
@@ -36,6 +36,15 @@
 
 #define DEBUG_SCATTER 0
 
+static inline void *drm_vmalloc_dma(unsigned long size)
+{
+#if defined(__powerpc__) && defined(CONFIG_NOT_COHERENT_CACHE)
+        return __vmalloc(size, GFP_KERNEL, PAGE_KERNEL | _PAGE_NO_CACHE);
+#else
+        return vmalloc_32(size);
+#endif
+}
+
 void drm_sg_cleanup(struct drm_sg_mem * entry)
 {
         struct page *page;
@@ -104,7 +113,7 @@ int drm_sg_alloc(struct drm_device *dev, struct drm_scatter_gather * request)
         }
         memset((void *)entry->busaddr, 0, pages * sizeof(*entry->busaddr));
 
-        entry->virtual = vmalloc_32(pages << PAGE_SHIFT);
+        entry->virtual = drm_vmalloc_dma(pages << PAGE_SHIFT);
         if (!entry->virtual) {
                 drm_free(entry->busaddr,
                          entry->pages * sizeof(*entry->busaddr), DRM_MEM_PAGES);
diff --git a/drivers/char/drm/drm_vm.c b/drivers/char/drm/drm_vm.c
index 3d65c4dcd0c6..945df72a51a9 100644
--- a/drivers/char/drm/drm_vm.c
+++ b/drivers/char/drm/drm_vm.c
@@ -54,13 +54,24 @@ static pgprot_t drm_io_prot(uint32_t map_type, struct vm_area_struct *vma)
         pgprot_val(tmp) |= _PAGE_NO_CACHE;
         if (map_type == _DRM_REGISTERS)
                 pgprot_val(tmp) |= _PAGE_GUARDED;
-#endif
-#if defined(__ia64__)
+#elif defined(__ia64__)
         if (efi_range_is_wc(vma->vm_start, vma->vm_end -
                                     vma->vm_start))
                 tmp = pgprot_writecombine(tmp);
         else
                 tmp = pgprot_noncached(tmp);
+#elif defined(__sparc__)
+        tmp = pgprot_noncached(tmp);
+#endif
+        return tmp;
+}
+
+static pgprot_t drm_dma_prot(uint32_t map_type, struct vm_area_struct *vma)
+{
+        pgprot_t tmp = vm_get_page_prot(vma->vm_flags);
+
+#if defined(__powerpc__) && defined(CONFIG_NOT_COHERENT_CACHE)
+        tmp |= _PAGE_NO_CACHE;
 #endif
         return tmp;
 }
@@ -603,9 +614,6 @@ static int drm_mmap_locked(struct file *filp, struct vm_area_struct *vma)
                 offset = dev->driver->get_reg_ofs(dev);
                 vma->vm_flags |= VM_IO; /* not in core dump */
                 vma->vm_page_prot = drm_io_prot(map->type, vma);
-#ifdef __sparc__
-                vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
-#endif
                 if (io_remap_pfn_range(vma, vma->vm_start,
                                        (map->offset + offset) >> PAGE_SHIFT,
                                        vma->vm_end - vma->vm_start,
@@ -624,6 +632,7 @@ static int drm_mmap_locked(struct file *filp, struct vm_area_struct *vma)
                     page_to_pfn(virt_to_page(map->handle)),
                     vma->vm_end - vma->vm_start, vma->vm_page_prot))
                         return -EAGAIN;
+                vma->vm_page_prot = drm_dma_prot(map->type, vma);
         /* fall through to _DRM_SHM */
         case _DRM_SHM:
                 vma->vm_ops = &drm_vm_shm_ops;
@@ -631,6 +640,7 @@ static int drm_mmap_locked(struct file *filp, struct vm_area_struct *vma)
                 /* Don't let this area swap.  Change when
                    DRM_KERNEL advisory is supported. */
                 vma->vm_flags |= VM_RESERVED;
+                vma->vm_page_prot = drm_dma_prot(map->type, vma);
                 break;
         case _DRM_SCATTER_GATHER:
                 vma->vm_ops = &drm_vm_sg_ops;
diff --git a/drivers/char/drm/i915_dma.c b/drivers/char/drm/i915_dma.c
index e9d6663bec73..a043bb12301a 100644
--- a/drivers/char/drm/i915_dma.c
+++ b/drivers/char/drm/i915_dma.c
@@ -804,6 +804,9 @@ void i915_driver_lastclose(struct drm_device * dev)
 {
         drm_i915_private_t *dev_priv = dev->dev_private;
 
+        if (!dev_priv)
+                return;
+
         if (dev_priv->agp_heap)
                 i915_mem_takedown(&(dev_priv->agp_heap));
 
diff --git a/drivers/char/drm/r300_cmdbuf.c b/drivers/char/drm/r300_cmdbuf.c
index 0f4afc44245c..f535812e4057 100644
--- a/drivers/char/drm/r300_cmdbuf.c
+++ b/drivers/char/drm/r300_cmdbuf.c
@@ -729,6 +729,47 @@ static void r300_discard_buffer(struct drm_device * dev, struct drm_buf * buf)
         buf->used = 0;
 }
 
+static void r300_cmd_wait(drm_radeon_private_t * dev_priv,
+                          drm_r300_cmd_header_t header)
+{
+        u32 wait_until;
+        RING_LOCALS;
+
+        if (!header.wait.flags)
+                return;
+
+        wait_until = 0;
+
+        switch(header.wait.flags) {
+        case R300_WAIT_2D:
+                wait_until = RADEON_WAIT_2D_IDLE;
+                break;
+        case R300_WAIT_3D:
+                wait_until = RADEON_WAIT_3D_IDLE;
+                break;
+        case R300_NEW_WAIT_2D_3D:
+                wait_until = RADEON_WAIT_2D_IDLE|RADEON_WAIT_3D_IDLE;
+                break;
+        case R300_NEW_WAIT_2D_2D_CLEAN:
+                wait_until = RADEON_WAIT_2D_IDLE|RADEON_WAIT_2D_IDLECLEAN;
+                break;
+        case R300_NEW_WAIT_3D_3D_CLEAN:
+                wait_until = RADEON_WAIT_3D_IDLE|RADEON_WAIT_3D_IDLECLEAN;
+                break;
+        case R300_NEW_WAIT_2D_2D_CLEAN_3D_3D_CLEAN:
+                wait_until = RADEON_WAIT_2D_IDLE|RADEON_WAIT_2D_IDLECLEAN;
+                wait_until |= RADEON_WAIT_3D_IDLE|RADEON_WAIT_3D_IDLECLEAN;
+                break;
+        default:
+                return;
+        }
+
+        BEGIN_RING(2);
+        OUT_RING(CP_PACKET0(RADEON_WAIT_UNTIL, 0));
+        OUT_RING(wait_until);
+        ADVANCE_RING();
+}
+
 static int r300_scratch(drm_radeon_private_t *dev_priv,
                         drm_radeon_kcmd_buffer_t *cmdbuf,
                         drm_r300_cmd_header_t header)
@@ -909,19 +950,8 @@ int r300_do_cp_cmdbuf(struct drm_device *dev,
                         break;
 
                 case R300_CMD_WAIT:
-                        /* simple enough, we can do it here */
                         DRM_DEBUG("R300_CMD_WAIT\n");
-                        if (header.wait.flags == 0)
-                                break;  /* nothing to do */
-
-                        {
-                                RING_LOCALS;
-
-                                BEGIN_RING(2);
-                                OUT_RING(CP_PACKET0(RADEON_WAIT_UNTIL, 0));
-                                OUT_RING((header.wait.flags & 0xf) << 14);
-                                ADVANCE_RING();
-                        }
+                        r300_cmd_wait(dev_priv, header);
                         break;
 
                 case R300_CMD_SCRATCH:
diff --git a/drivers/char/drm/radeon_drm.h b/drivers/char/drm/radeon_drm.h
index 71e5b21fad2c..aab82e121e07 100644
--- a/drivers/char/drm/radeon_drm.h
+++ b/drivers/char/drm/radeon_drm.h
@@ -225,8 +225,20 @@ typedef union {
 #define R300_CMD_WAIT                   7
 #       define R300_WAIT_2D             0x1
 #       define R300_WAIT_3D             0x2
+/* these two defines are DOING IT WRONG - however
+ * we have userspace which relies on using these.
+ * The wait interface is backwards compat new 
+ * code should use the NEW_WAIT defines below
+ * THESE ARE NOT BIT FIELDS
+ */
 #       define R300_WAIT_2D_CLEAN       0x3
 #       define R300_WAIT_3D_CLEAN       0x4
+
+#       define R300_NEW_WAIT_2D_3D      0x3
+#       define R300_NEW_WAIT_2D_2D_CLEAN        0x4
+#       define R300_NEW_WAIT_3D_3D_CLEAN        0x6
+#       define R300_NEW_WAIT_2D_2D_CLEAN_3D_3D_CLEAN    0x8
+
 #define R300_CMD_SCRATCH                8
 
 typedef union {
diff --git a/drivers/char/drm/radeon_mem.c b/drivers/char/drm/radeon_mem.c
index 78b34fa7c89a..4af5286a36fb 100644
--- a/drivers/char/drm/radeon_mem.c
+++ b/drivers/char/drm/radeon_mem.c
@@ -88,7 +88,7 @@ static struct mem_block *alloc_block(struct mem_block *heap, int size,
 
         list_for_each(p, heap) {
                 int start = (p->start + mask) & ~mask;
-                if (p->file_priv == 0 && start + size <= p->start + p->size)
+                if (p->file_priv == NULL && start + size <= p->start + p->size)
                         return split_block(p, start, size, file_priv);
         }
 
@@ -113,7 +113,7 @@ static void free_block(struct mem_block *p)
         /* Assumes a single contiguous range.  Needs a special file_priv in
          * 'heap' to stop it being subsumed.
          */
-        if (p->next->file_priv == 0) {
+        if (p->next->file_priv == NULL) {
                 struct mem_block *q = p->next;
                 p->size += q->size;
                 p->next = q->next;
@@ -121,7 +121,7 @@ static void free_block(struct mem_block *p)
                 drm_free(q, sizeof(*q), DRM_MEM_BUFS);
         }
 
-        if (p->prev->file_priv == 0) {
+        if (p->prev->file_priv == NULL) {
                 struct mem_block *q = p->prev;
                 q->size += p->size;
                 q->next = p->next;
@@ -174,7 +174,7 @@ void radeon_mem_release(struct drm_file *file_priv, struct mem_block *heap)
          * 'heap' to stop it being subsumed.
          */
         list_for_each(p, heap) {
-                while (p->file_priv == 0 && p->next->file_priv == 0) {
+                while (p->file_priv == NULL && p->next->file_priv == NULL) {
                         struct mem_block *q = p->next;
                         p->size += q->size;
                         p->next = q->next;