Merge branch 'drm-mm-cleanup' into drm-next
* drm-mm-cleanup:
radeon: move blit functions to radeon_asic.h
radeon: kill decls for inline functions
radeon: consolidate asic-specific function decls for r600 & later
drm/radeon: kill radeon_bo->gobj pointer
drm/radeon: introduce gem_to_radeon_bo helper
drm/radeon: embed struct drm_gem_object
drm: mm: add helper to unwind scan state
drm: mm: add api for embedding struct drm_mm_node
drm: mm: extract node insert helper functions
drm: mm: track free areas implicitly
drm/nouveau: don't munge in drm_mm internals
diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
index a245d17..af60d9b 100644
--- a/drivers/gpu/drm/drm_edid.c
+++ b/drivers/gpu/drm/drm_edid.c
@@ -449,12 +449,11 @@
struct drm_display_mode *drm_mode_find_dmt(struct drm_device *dev,
int hsize, int vsize, int fresh)
{
+ struct drm_display_mode *mode = NULL;
int i;
- struct drm_display_mode *ptr, *mode;
- mode = NULL;
for (i = 0; i < drm_num_dmt_modes; i++) {
- ptr = &drm_dmt_modes[i];
+ const struct drm_display_mode *ptr = &drm_dmt_modes[i];
if (hsize == ptr->hdisplay &&
vsize == ptr->vdisplay &&
fresh == drm_mode_vrefresh(ptr)) {
@@ -885,7 +884,7 @@
}
static bool
-mode_is_rb(struct drm_display_mode *mode)
+mode_is_rb(const struct drm_display_mode *mode)
{
return (mode->htotal - mode->hdisplay == 160) &&
(mode->hsync_end - mode->hdisplay == 80) &&
@@ -894,7 +893,8 @@
}
static bool
-mode_in_hsync_range(struct drm_display_mode *mode, struct edid *edid, u8 *t)
+mode_in_hsync_range(const struct drm_display_mode *mode,
+ struct edid *edid, u8 *t)
{
int hsync, hmin, hmax;
@@ -910,7 +910,8 @@
}
static bool
-mode_in_vsync_range(struct drm_display_mode *mode, struct edid *edid, u8 *t)
+mode_in_vsync_range(const struct drm_display_mode *mode,
+ struct edid *edid, u8 *t)
{
int vsync, vmin, vmax;
@@ -941,7 +942,7 @@
}
static bool
-mode_in_range(struct drm_display_mode *mode, struct edid *edid,
+mode_in_range(const struct drm_display_mode *mode, struct edid *edid,
struct detailed_timing *timing)
{
u32 max_clock;
@@ -1472,7 +1473,7 @@
int hdisplay, int vdisplay)
{
int i, count, num_modes = 0;
- struct drm_display_mode *mode, *ptr;
+ struct drm_display_mode *mode;
struct drm_device *dev = connector->dev;
count = sizeof(drm_dmt_modes) / sizeof(struct drm_display_mode);
@@ -1482,7 +1483,7 @@
vdisplay = 0;
for (i = 0; i < count; i++) {
- ptr = &drm_dmt_modes[i];
+ const struct drm_display_mode *ptr = &drm_dmt_modes[i];
if (hdisplay && vdisplay) {
/*
* Only when two are valid, they will be used to check
diff --git a/drivers/gpu/drm/drm_edid_modes.h b/drivers/gpu/drm/drm_edid_modes.h
index 6eb7592..5f20644 100644
--- a/drivers/gpu/drm/drm_edid_modes.h
+++ b/drivers/gpu/drm/drm_edid_modes.h
@@ -32,7 +32,7 @@
* This table is copied from xfree86/modes/xf86EdidModes.c.
* But the mode with Reduced blank feature is deleted.
*/
-static struct drm_display_mode drm_dmt_modes[] = {
+static const struct drm_display_mode drm_dmt_modes[] = {
/* 640x350@85Hz */
{ DRM_MODE("640x350", DRM_MODE_TYPE_DRIVER, 31500, 640, 672,
736, 832, 0, 350, 382, 385, 445, 0,
@@ -266,7 +266,7 @@
static const int drm_num_dmt_modes =
sizeof(drm_dmt_modes) / sizeof(struct drm_display_mode);
-static struct drm_display_mode edid_est_modes[] = {
+static const struct drm_display_mode edid_est_modes[] = {
{ DRM_MODE("800x600", DRM_MODE_TYPE_DRIVER, 40000, 800, 840,
968, 1056, 0, 600, 601, 605, 628, 0,
DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, /* 800x600@60Hz */
diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
index 6977a1c..d421f9d 100644
--- a/drivers/gpu/drm/drm_fb_helper.c
+++ b/drivers/gpu/drm/drm_fb_helper.c
@@ -627,6 +627,11 @@
value = (red << info->var.red.offset) |
(green << info->var.green.offset) |
(blue << info->var.blue.offset);
+ if (info->var.transp.length > 0) {
+ u32 mask = (1 << info->var.transp.length) - 1;
+ mask <<= info->var.transp.offset;
+ value |= mask;
+ }
palette[regno] = value;
return 0;
}
diff --git a/drivers/gpu/drm/drm_gem.c b/drivers/gpu/drm/drm_gem.c
index aa8df25..57ce27c 100644
--- a/drivers/gpu/drm/drm_gem.c
+++ b/drivers/gpu/drm/drm_gem.c
@@ -101,7 +101,7 @@
dev->mm_private = mm;
- if (drm_ht_create(&mm->offset_hash, 19)) {
+ if (drm_ht_create(&mm->offset_hash, 12)) {
kfree(mm);
return -ENOMEM;
}
diff --git a/drivers/gpu/drm/drm_hashtab.c b/drivers/gpu/drm/drm_hashtab.c
index a93d7b4..e3a7568 100644
--- a/drivers/gpu/drm/drm_hashtab.c
+++ b/drivers/gpu/drm/drm_hashtab.c
@@ -39,27 +39,18 @@
int drm_ht_create(struct drm_open_hash *ht, unsigned int order)
{
- unsigned int i;
+ unsigned int size = 1 << order;
- ht->size = 1 << order;
ht->order = order;
- ht->fill = 0;
ht->table = NULL;
- ht->use_vmalloc = ((ht->size * sizeof(*ht->table)) > PAGE_SIZE);
- if (!ht->use_vmalloc) {
- ht->table = kcalloc(ht->size, sizeof(*ht->table), GFP_KERNEL);
- }
- if (!ht->table) {
- ht->use_vmalloc = 1;
- ht->table = vmalloc(ht->size*sizeof(*ht->table));
- }
+ if (size <= PAGE_SIZE / sizeof(*ht->table))
+ ht->table = kcalloc(size, sizeof(*ht->table), GFP_KERNEL);
+ else
+ ht->table = vzalloc(size*sizeof(*ht->table));
if (!ht->table) {
DRM_ERROR("Out of memory for hash table\n");
return -ENOMEM;
}
- for (i=0; i< ht->size; ++i) {
- INIT_HLIST_HEAD(&ht->table[i]);
- }
return 0;
}
EXPORT_SYMBOL(drm_ht_create);
@@ -180,7 +171,6 @@
list = drm_ht_find_key(ht, key);
if (list) {
hlist_del_init(list);
- ht->fill--;
return 0;
}
return -EINVAL;
@@ -189,7 +179,6 @@
int drm_ht_remove_item(struct drm_open_hash *ht, struct drm_hash_item *item)
{
hlist_del_init(&item->head);
- ht->fill--;
return 0;
}
EXPORT_SYMBOL(drm_ht_remove_item);
@@ -197,10 +186,10 @@
void drm_ht_remove(struct drm_open_hash *ht)
{
if (ht->table) {
- if (ht->use_vmalloc)
- vfree(ht->table);
- else
+ if ((PAGE_SIZE / sizeof(*ht->table)) >> ht->order)
kfree(ht->table);
+ else
+ vfree(ht->table);
ht->table = NULL;
}
}
diff --git a/drivers/gpu/drm/drm_modes.c b/drivers/gpu/drm/drm_modes.c
index 58e65f9..25bf873 100644
--- a/drivers/gpu/drm/drm_modes.c
+++ b/drivers/gpu/drm/drm_modes.c
@@ -593,7 +593,7 @@
*
* Return @modes's hsync rate in kHz, rounded to the nearest int.
*/
-int drm_mode_hsync(struct drm_display_mode *mode)
+int drm_mode_hsync(const struct drm_display_mode *mode)
{
unsigned int calc_val;
@@ -627,7 +627,7 @@
* If it is 70.288, it will return 70Hz.
* If it is 59.6, it will return 60Hz.
*/
-int drm_mode_vrefresh(struct drm_display_mode *mode)
+int drm_mode_vrefresh(const struct drm_display_mode *mode)
{
int refresh = 0;
unsigned int calc_val;
@@ -725,7 +725,7 @@
* a pointer to it. Used to create new instances of established modes.
*/
struct drm_display_mode *drm_mode_duplicate(struct drm_device *dev,
- struct drm_display_mode *mode)
+ const struct drm_display_mode *mode)
{
struct drm_display_mode *nmode;
int new_id;
diff --git a/drivers/gpu/drm/i915/intel_sdvo.c b/drivers/gpu/drm/i915/intel_sdvo.c
index 6a09c14..318f398 100644
--- a/drivers/gpu/drm/i915/intel_sdvo.c
+++ b/drivers/gpu/drm/i915/intel_sdvo.c
@@ -1460,7 +1460,7 @@
* Note! This is in reply order (see loop in get_tv_modes).
* XXX: all 60Hz refresh?
*/
-struct drm_display_mode sdvo_tv_modes[] = {
+static const struct drm_display_mode sdvo_tv_modes[] = {
{ DRM_MODE("320x200", DRM_MODE_TYPE_DRIVER, 5815, 320, 321, 384,
416, 0, 200, 201, 232, 233, 0,
DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
diff --git a/drivers/gpu/drm/nouveau/nouveau_mem.c b/drivers/gpu/drm/nouveau/nouveau_mem.c
index 123969d..2b4e5e9 100644
--- a/drivers/gpu/drm/nouveau/nouveau_mem.c
+++ b/drivers/gpu/drm/nouveau/nouveau_mem.c
@@ -409,6 +409,7 @@
if (ret)
return ret;
+ dev_priv->ttm.bdev.dev = dev->dev;
ret = ttm_bo_device_init(&dev_priv->ttm.bdev,
dev_priv->ttm.bo_global_ref.ref.object,
&nouveau_bo_driver, DRM_FILE_PAGE_OFFSET,
diff --git a/drivers/gpu/drm/nouveau/nouveau_sgdma.c b/drivers/gpu/drm/nouveau/nouveau_sgdma.c
index 9a250eb..07b1151 100644
--- a/drivers/gpu/drm/nouveau/nouveau_sgdma.c
+++ b/drivers/gpu/drm/nouveau/nouveau_sgdma.c
@@ -12,6 +12,7 @@
struct drm_device *dev;
dma_addr_t *pages;
+ bool *ttm_alloced;
unsigned nr_pages;
u64 offset;
@@ -20,7 +21,8 @@
static int
nouveau_sgdma_populate(struct ttm_backend *be, unsigned long num_pages,
- struct page **pages, struct page *dummy_read_page)
+ struct page **pages, struct page *dummy_read_page,
+ dma_addr_t *dma_addrs)
{
struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)be;
struct drm_device *dev = nvbe->dev;
@@ -34,15 +36,25 @@
if (!nvbe->pages)
return -ENOMEM;
+ nvbe->ttm_alloced = kmalloc(sizeof(bool) * num_pages, GFP_KERNEL);
+ if (!nvbe->ttm_alloced)
+ return -ENOMEM;
+
nvbe->nr_pages = 0;
while (num_pages--) {
- nvbe->pages[nvbe->nr_pages] =
- pci_map_page(dev->pdev, pages[nvbe->nr_pages], 0,
+ if (dma_addrs[nvbe->nr_pages] != DMA_ERROR_CODE) {
+ nvbe->pages[nvbe->nr_pages] =
+ dma_addrs[nvbe->nr_pages];
+ nvbe->ttm_alloced[nvbe->nr_pages] = true;
+ } else {
+ nvbe->pages[nvbe->nr_pages] =
+ pci_map_page(dev->pdev, pages[nvbe->nr_pages], 0,
PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
- if (pci_dma_mapping_error(dev->pdev,
- nvbe->pages[nvbe->nr_pages])) {
- be->func->clear(be);
- return -EFAULT;
+ if (pci_dma_mapping_error(dev->pdev,
+ nvbe->pages[nvbe->nr_pages])) {
+ be->func->clear(be);
+ return -EFAULT;
+ }
}
nvbe->nr_pages++;
@@ -65,11 +77,14 @@
be->func->unbind(be);
while (nvbe->nr_pages--) {
- pci_unmap_page(dev->pdev, nvbe->pages[nvbe->nr_pages],
+ if (!nvbe->ttm_alloced[nvbe->nr_pages])
+ pci_unmap_page(dev->pdev, nvbe->pages[nvbe->nr_pages],
PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
}
kfree(nvbe->pages);
+ kfree(nvbe->ttm_alloced);
nvbe->pages = NULL;
+ nvbe->ttm_alloced = NULL;
nvbe->nr_pages = 0;
}
}
diff --git a/drivers/gpu/drm/nouveau/nv17_tv.c b/drivers/gpu/drm/nouveau/nv17_tv.c
index 28119fd..3900ceb 100644
--- a/drivers/gpu/drm/nouveau/nv17_tv.c
+++ b/drivers/gpu/drm/nouveau/nv17_tv.c
@@ -197,10 +197,12 @@
struct drm_connector *connector)
{
struct nv17_tv_norm_params *tv_norm = get_tv_norm(encoder);
- struct drm_display_mode *mode, *tv_mode;
+ const struct drm_display_mode *tv_mode;
int n = 0;
for (tv_mode = nv17_tv_modes; tv_mode->hdisplay; tv_mode++) {
+ struct drm_display_mode *mode;
+
mode = drm_mode_duplicate(encoder->dev, tv_mode);
mode->clock = tv_norm->tv_enc_mode.vrefresh *
diff --git a/drivers/gpu/drm/nouveau/nv17_tv.h b/drivers/gpu/drm/nouveau/nv17_tv.h
index 6bf0384..622e722 100644
--- a/drivers/gpu/drm/nouveau/nv17_tv.h
+++ b/drivers/gpu/drm/nouveau/nv17_tv.h
@@ -112,7 +112,7 @@
} nv17_tv_norms[NUM_TV_NORMS];
#define get_tv_norm(enc) (&nv17_tv_norms[to_tv_enc(enc)->tv_norm])
-extern struct drm_display_mode nv17_tv_modes[];
+extern const struct drm_display_mode nv17_tv_modes[];
static inline int interpolate(int y0, int y1, int y2, int x)
{
diff --git a/drivers/gpu/drm/nouveau/nv17_tv_modes.c b/drivers/gpu/drm/nouveau/nv17_tv_modes.c
index 9d3893c..4d1d29f 100644
--- a/drivers/gpu/drm/nouveau/nv17_tv_modes.c
+++ b/drivers/gpu/drm/nouveau/nv17_tv_modes.c
@@ -438,7 +438,7 @@
/* Timings similar to the ones the blob sets */
-struct drm_display_mode nv17_tv_modes[] = {
+const struct drm_display_mode nv17_tv_modes[] = {
{ DRM_MODE("320x200", DRM_MODE_TYPE_DRIVER, 0,
320, 344, 392, 560, 0, 200, 200, 202, 220, 0,
DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC
diff --git a/drivers/gpu/drm/radeon/r600_cs.c b/drivers/gpu/drm/radeon/r600_cs.c
index 7831e08..4706294 100644
--- a/drivers/gpu/drm/radeon/r600_cs.c
+++ b/drivers/gpu/drm/radeon/r600_cs.c
@@ -71,75 +71,164 @@
u64 db_bo_mc;
};
+#define FMT_8_BIT(fmt, vc) [fmt] = { 1, 1, 1, vc }
+#define FMT_16_BIT(fmt, vc) [fmt] = { 1, 1, 2, vc }
+#define FMT_24_BIT(fmt) [fmt] = { 1, 1, 3, 0 }
+#define FMT_32_BIT(fmt, vc) [fmt] = { 1, 1, 4, vc }
+#define FMT_48_BIT(fmt) [fmt] = { 1, 1, 6, 0 }
+#define FMT_64_BIT(fmt, vc) [fmt] = { 1, 1, 8, vc }
+#define FMT_96_BIT(fmt) [fmt] = { 1, 1, 12, 0 }
+#define FMT_128_BIT(fmt, vc) [fmt] = { 1, 1, 16, vc }
+
+struct gpu_formats {
+ unsigned blockwidth;
+ unsigned blockheight;
+ unsigned blocksize;
+ unsigned valid_color;
+};
+
+static const struct gpu_formats color_formats_table[] = {
+ /* 8 bit */
+ FMT_8_BIT(V_038004_COLOR_8, 1),
+ FMT_8_BIT(V_038004_COLOR_4_4, 1),
+ FMT_8_BIT(V_038004_COLOR_3_3_2, 1),
+ FMT_8_BIT(V_038004_FMT_1, 0),
+
+ /* 16-bit */
+ FMT_16_BIT(V_038004_COLOR_16, 1),
+ FMT_16_BIT(V_038004_COLOR_16_FLOAT, 1),
+ FMT_16_BIT(V_038004_COLOR_8_8, 1),
+ FMT_16_BIT(V_038004_COLOR_5_6_5, 1),
+ FMT_16_BIT(V_038004_COLOR_6_5_5, 1),
+ FMT_16_BIT(V_038004_COLOR_1_5_5_5, 1),
+ FMT_16_BIT(V_038004_COLOR_4_4_4_4, 1),
+ FMT_16_BIT(V_038004_COLOR_5_5_5_1, 1),
+
+ /* 24-bit */
+ FMT_24_BIT(V_038004_FMT_8_8_8),
+
+ /* 32-bit */
+ FMT_32_BIT(V_038004_COLOR_32, 1),
+ FMT_32_BIT(V_038004_COLOR_32_FLOAT, 1),
+ FMT_32_BIT(V_038004_COLOR_16_16, 1),
+ FMT_32_BIT(V_038004_COLOR_16_16_FLOAT, 1),
+ FMT_32_BIT(V_038004_COLOR_8_24, 1),
+ FMT_32_BIT(V_038004_COLOR_8_24_FLOAT, 1),
+ FMT_32_BIT(V_038004_COLOR_24_8, 1),
+ FMT_32_BIT(V_038004_COLOR_24_8_FLOAT, 1),
+ FMT_32_BIT(V_038004_COLOR_10_11_11, 1),
+ FMT_32_BIT(V_038004_COLOR_10_11_11_FLOAT, 1),
+ FMT_32_BIT(V_038004_COLOR_11_11_10, 1),
+ FMT_32_BIT(V_038004_COLOR_11_11_10_FLOAT, 1),
+ FMT_32_BIT(V_038004_COLOR_2_10_10_10, 1),
+ FMT_32_BIT(V_038004_COLOR_8_8_8_8, 1),
+ FMT_32_BIT(V_038004_COLOR_10_10_10_2, 1),
+ FMT_32_BIT(V_038004_FMT_5_9_9_9_SHAREDEXP, 0),
+ FMT_32_BIT(V_038004_FMT_32_AS_8, 0),
+ FMT_32_BIT(V_038004_FMT_32_AS_8_8, 0),
+
+ /* 48-bit */
+ FMT_48_BIT(V_038004_FMT_16_16_16),
+ FMT_48_BIT(V_038004_FMT_16_16_16_FLOAT),
+
+ /* 64-bit */
+ FMT_64_BIT(V_038004_COLOR_X24_8_32_FLOAT, 1),
+ FMT_64_BIT(V_038004_COLOR_32_32, 1),
+ FMT_64_BIT(V_038004_COLOR_32_32_FLOAT, 1),
+ FMT_64_BIT(V_038004_COLOR_16_16_16_16, 1),
+ FMT_64_BIT(V_038004_COLOR_16_16_16_16_FLOAT, 1),
+
+ FMT_96_BIT(V_038004_FMT_32_32_32),
+ FMT_96_BIT(V_038004_FMT_32_32_32_FLOAT),
+
+ /* 128-bit */
+ FMT_128_BIT(V_038004_COLOR_32_32_32_32, 1),
+ FMT_128_BIT(V_038004_COLOR_32_32_32_32_FLOAT, 1),
+
+ [V_038004_FMT_GB_GR] = { 2, 1, 4, 0 },
+ [V_038004_FMT_BG_RG] = { 2, 1, 4, 0 },
+
+ /* block compressed formats */
+ [V_038004_FMT_BC1] = { 4, 4, 8, 0 },
+ [V_038004_FMT_BC2] = { 4, 4, 16, 0 },
+ [V_038004_FMT_BC3] = { 4, 4, 16, 0 },
+ [V_038004_FMT_BC4] = { 4, 4, 8, 0 },
+ [V_038004_FMT_BC5] = { 4, 4, 16, 0},
+
+};
+
+static inline bool fmt_is_valid_color(u32 format)
+{
+ if (format > ARRAY_SIZE(color_formats_table))
+ return false;
+
+ if (color_formats_table[format].valid_color)
+ return true;
+
+ return false;
+}
+
+static inline bool fmt_is_valid_texture(u32 format)
+{
+ if (format > ARRAY_SIZE(color_formats_table))
+ return false;
+
+ if (color_formats_table[format].blockwidth > 0)
+ return true;
+
+ return false;
+}
+
+static inline int fmt_get_blocksize(u32 format)
+{
+ if (format > ARRAY_SIZE(color_formats_table))
+ return 0;
+
+ return color_formats_table[format].blocksize;
+}
+
+static inline int fmt_get_nblocksx(u32 format, u32 w)
+{
+ unsigned bw;
+ if (format > ARRAY_SIZE(color_formats_table))
+ return 0;
+
+ bw = color_formats_table[format].blockwidth;
+ if (bw == 0)
+ return 0;
+
+ return (w + bw - 1) / bw;
+}
+
+static inline int fmt_get_nblocksy(u32 format, u32 h)
+{
+ unsigned bh;
+ if (format > ARRAY_SIZE(color_formats_table))
+ return 0;
+
+ bh = color_formats_table[format].blockheight;
+ if (bh == 0)
+ return 0;
+
+ return (h + bh - 1) / bh;
+}
+
static inline int r600_bpe_from_format(u32 *bpe, u32 format)
{
- switch (format) {
- case V_038004_COLOR_8:
- case V_038004_COLOR_4_4:
- case V_038004_COLOR_3_3_2:
- case V_038004_FMT_1:
- *bpe = 1;
- break;
- case V_038004_COLOR_16:
- case V_038004_COLOR_16_FLOAT:
- case V_038004_COLOR_8_8:
- case V_038004_COLOR_5_6_5:
- case V_038004_COLOR_6_5_5:
- case V_038004_COLOR_1_5_5_5:
- case V_038004_COLOR_4_4_4_4:
- case V_038004_COLOR_5_5_5_1:
- *bpe = 2;
- break;
- case V_038004_FMT_8_8_8:
- *bpe = 3;
- break;
- case V_038004_COLOR_32:
- case V_038004_COLOR_32_FLOAT:
- case V_038004_COLOR_16_16:
- case V_038004_COLOR_16_16_FLOAT:
- case V_038004_COLOR_8_24:
- case V_038004_COLOR_8_24_FLOAT:
- case V_038004_COLOR_24_8:
- case V_038004_COLOR_24_8_FLOAT:
- case V_038004_COLOR_10_11_11:
- case V_038004_COLOR_10_11_11_FLOAT:
- case V_038004_COLOR_11_11_10:
- case V_038004_COLOR_11_11_10_FLOAT:
- case V_038004_COLOR_2_10_10_10:
- case V_038004_COLOR_8_8_8_8:
- case V_038004_COLOR_10_10_10_2:
- case V_038004_FMT_5_9_9_9_SHAREDEXP:
- case V_038004_FMT_32_AS_8:
- case V_038004_FMT_32_AS_8_8:
- *bpe = 4;
- break;
- case V_038004_COLOR_X24_8_32_FLOAT:
- case V_038004_COLOR_32_32:
- case V_038004_COLOR_32_32_FLOAT:
- case V_038004_COLOR_16_16_16_16:
- case V_038004_COLOR_16_16_16_16_FLOAT:
- *bpe = 8;
- break;
- case V_038004_FMT_16_16_16:
- case V_038004_FMT_16_16_16_FLOAT:
- *bpe = 6;
- break;
- case V_038004_FMT_32_32_32:
- case V_038004_FMT_32_32_32_FLOAT:
- *bpe = 12;
- break;
- case V_038004_COLOR_32_32_32_32:
- case V_038004_COLOR_32_32_32_32_FLOAT:
- *bpe = 16;
- break;
- case V_038004_FMT_GB_GR:
- case V_038004_FMT_BG_RG:
- case V_038004_COLOR_INVALID:
- default:
- *bpe = 16;
- return -EINVAL;
- }
+ unsigned res;
+ if (format > ARRAY_SIZE(color_formats_table))
+ goto fail;
+
+ res = color_formats_table[format].blocksize;
+ if (res == 0)
+ goto fail;
+
+ *bpe = res;
return 0;
+
+fail:
+ *bpe = 16;
+ return -EINVAL;
}
struct array_mode_checker {
@@ -148,7 +237,7 @@
u32 nbanks;
u32 npipes;
u32 nsamples;
- u32 bpe;
+ u32 blocksize;
};
/* returns alignment in pixels for pitch/height/depth and bytes for base */
@@ -162,7 +251,7 @@
u32 tile_height = 8;
u32 macro_tile_width = values->nbanks;
u32 macro_tile_height = values->npipes;
- u32 tile_bytes = tile_width * tile_height * values->bpe * values->nsamples;
+ u32 tile_bytes = tile_width * tile_height * values->blocksize * values->nsamples;
u32 macro_tile_bytes = macro_tile_width * macro_tile_height * tile_bytes;
switch (values->array_mode) {
@@ -174,7 +263,7 @@
*base_align = 1;
break;
case ARRAY_LINEAR_ALIGNED:
- *pitch_align = max((u32)64, (u32)(values->group_size / values->bpe));
+ *pitch_align = max((u32)64, (u32)(values->group_size / values->blocksize));
*height_align = tile_height;
*depth_align = 1;
*base_align = values->group_size;
@@ -182,7 +271,7 @@
case ARRAY_1D_TILED_THIN1:
*pitch_align = max((u32)tile_width,
(u32)(values->group_size /
- (tile_height * values->bpe * values->nsamples)));
+ (tile_height * values->blocksize * values->nsamples)));
*height_align = tile_height;
*depth_align = 1;
*base_align = values->group_size;
@@ -190,12 +279,12 @@
case ARRAY_2D_TILED_THIN1:
*pitch_align = max((u32)macro_tile_width,
(u32)(((values->group_size / tile_height) /
- (values->bpe * values->nsamples)) *
+ (values->blocksize * values->nsamples)) *
values->nbanks)) * tile_width;
*height_align = macro_tile_height * tile_height;
*depth_align = 1;
*base_align = max(macro_tile_bytes,
- (*pitch_align) * values->bpe * (*height_align) * values->nsamples);
+ (*pitch_align) * values->blocksize * (*height_align) * values->nsamples);
break;
default:
return -EINVAL;
@@ -234,21 +323,22 @@
static inline int r600_cs_track_validate_cb(struct radeon_cs_parser *p, int i)
{
struct r600_cs_track *track = p->track;
- u32 bpe = 0, slice_tile_max, size, tmp;
+ u32 slice_tile_max, size, tmp;
u32 height, height_align, pitch, pitch_align, depth_align;
u64 base_offset, base_align;
struct array_mode_checker array_check;
volatile u32 *ib = p->ib->ptr;
unsigned array_mode;
-
+ u32 format;
if (G_0280A0_TILE_MODE(track->cb_color_info[i])) {
dev_warn(p->dev, "FMASK or CMASK buffer are not supported by this kernel\n");
return -EINVAL;
}
size = radeon_bo_size(track->cb_color_bo[i]) - track->cb_color_bo_offset[i];
- if (r600_bpe_from_format(&bpe, G_0280A0_FORMAT(track->cb_color_info[i]))) {
+ format = G_0280A0_FORMAT(track->cb_color_info[i]);
+ if (!fmt_is_valid_color(format)) {
dev_warn(p->dev, "%s:%d cb invalid format %d for %d (0x%08X)\n",
- __func__, __LINE__, G_0280A0_FORMAT(track->cb_color_info[i]),
+ __func__, __LINE__, format,
i, track->cb_color_info[i]);
return -EINVAL;
}
@@ -267,7 +357,7 @@
array_check.nbanks = track->nbanks;
array_check.npipes = track->npipes;
array_check.nsamples = track->nsamples;
- array_check.bpe = bpe;
+ array_check.blocksize = fmt_get_blocksize(format);
if (r600_get_array_mode_alignment(&array_check,
&pitch_align, &height_align, &depth_align, &base_align)) {
dev_warn(p->dev, "%s invalid tiling %d for %d (0x%08X)\n", __func__,
@@ -310,7 +400,7 @@
}
/* check offset */
- tmp = height * pitch * bpe;
+ tmp = fmt_get_nblocksy(format, height) * fmt_get_nblocksx(format, pitch) * fmt_get_blocksize(format);
if ((tmp + track->cb_color_bo_offset[i]) > radeon_bo_size(track->cb_color_bo[i])) {
if (array_mode == V_0280A0_ARRAY_LINEAR_GENERAL) {
/* the initial DDX does bad things with the CB size occasionally */
@@ -432,7 +522,7 @@
array_check.nbanks = track->nbanks;
array_check.npipes = track->npipes;
array_check.nsamples = track->nsamples;
- array_check.bpe = bpe;
+ array_check.blocksize = bpe;
if (r600_get_array_mode_alignment(&array_check,
&pitch_align, &height_align, &depth_align, &base_align)) {
dev_warn(p->dev, "%s invalid tiling %d (0x%08X)\n", __func__,
@@ -1107,39 +1197,61 @@
return 0;
}
-static inline unsigned minify(unsigned size, unsigned levels)
+static inline unsigned mip_minify(unsigned size, unsigned level)
{
- size = size >> levels;
- if (size < 1)
- size = 1;
- return size;
+ unsigned val;
+
+ val = max(1U, size >> level);
+ if (level > 0)
+ val = roundup_pow_of_two(val);
+ return val;
}
-static void r600_texture_size(unsigned nfaces, unsigned blevel, unsigned nlevels,
- unsigned w0, unsigned h0, unsigned d0, unsigned bpe,
- unsigned pitch_align,
+static void r600_texture_size(unsigned nfaces, unsigned blevel, unsigned llevel,
+ unsigned w0, unsigned h0, unsigned d0, unsigned format,
+ unsigned block_align, unsigned height_align, unsigned base_align,
unsigned *l0_size, unsigned *mipmap_size)
{
- unsigned offset, i, level, face;
- unsigned width, height, depth, rowstride, size;
+ unsigned offset, i, level;
+ unsigned width, height, depth, size;
+ unsigned blocksize;
+ unsigned nbx, nby;
+ unsigned nlevels = llevel - blevel + 1;
- w0 = minify(w0, 0);
- h0 = minify(h0, 0);
- d0 = minify(d0, 0);
+ *l0_size = -1;
+ blocksize = fmt_get_blocksize(format);
+
+ w0 = mip_minify(w0, 0);
+ h0 = mip_minify(h0, 0);
+ d0 = mip_minify(d0, 0);
for(i = 0, offset = 0, level = blevel; i < nlevels; i++, level++) {
- width = minify(w0, i);
- height = minify(h0, i);
- depth = minify(d0, i);
- for(face = 0; face < nfaces; face++) {
- rowstride = ALIGN((width * bpe), pitch_align);
- size = height * rowstride * depth;
- offset += size;
- offset = (offset + 0x1f) & ~0x1f;
- }
+ width = mip_minify(w0, i);
+ nbx = fmt_get_nblocksx(format, width);
+
+ nbx = round_up(nbx, block_align);
+
+ height = mip_minify(h0, i);
+ nby = fmt_get_nblocksy(format, height);
+ nby = round_up(nby, height_align);
+
+ depth = mip_minify(d0, i);
+
+ size = nbx * nby * blocksize;
+ if (nfaces)
+ size *= nfaces;
+ else
+ size *= depth;
+
+ if (i == 0)
+ *l0_size = size;
+
+ if (i == 0 || i == 1)
+ offset = round_up(offset, base_align);
+
+ offset += size;
}
- *l0_size = ALIGN((w0 * bpe), pitch_align) * h0 * d0;
*mipmap_size = offset;
- if (!nlevels)
+ if (llevel == 0)
*mipmap_size = *l0_size;
if (!blevel)
*mipmap_size -= *l0_size;
@@ -1163,11 +1275,13 @@
u32 tiling_flags)
{
struct r600_cs_track *track = p->track;
- u32 nfaces, nlevels, blevel, w0, h0, d0, bpe = 0;
+ u32 nfaces, llevel, blevel, w0, h0, d0;
u32 word0, word1, l0_size, mipmap_size;
u32 height_align, pitch, pitch_align, depth_align;
+ u32 array, barray, larray;
u64 base_align;
struct array_mode_checker array_check;
+ u32 format;
/* on legacy kernel we don't perform advanced check */
if (p->rdev == NULL)
@@ -1193,19 +1307,25 @@
case V_038000_SQ_TEX_DIM_3D:
break;
case V_038000_SQ_TEX_DIM_CUBEMAP:
- nfaces = 6;
+ if (p->family >= CHIP_RV770)
+ nfaces = 8;
+ else
+ nfaces = 6;
break;
case V_038000_SQ_TEX_DIM_1D_ARRAY:
case V_038000_SQ_TEX_DIM_2D_ARRAY:
+ array = 1;
+ break;
case V_038000_SQ_TEX_DIM_2D_MSAA:
case V_038000_SQ_TEX_DIM_2D_ARRAY_MSAA:
default:
dev_warn(p->dev, "this kernel doesn't support %d texture dim\n", G_038000_DIM(word0));
return -EINVAL;
}
- if (r600_bpe_from_format(&bpe, G_038004_DATA_FORMAT(word1))) {
+ format = G_038004_DATA_FORMAT(word1);
+ if (!fmt_is_valid_texture(format)) {
dev_warn(p->dev, "%s:%d texture invalid format %d\n",
- __func__, __LINE__, G_038004_DATA_FORMAT(word1));
+ __func__, __LINE__, format);
return -EINVAL;
}
@@ -1216,7 +1336,7 @@
array_check.nbanks = track->nbanks;
array_check.npipes = track->npipes;
array_check.nsamples = 1;
- array_check.bpe = bpe;
+ array_check.blocksize = fmt_get_blocksize(format);
if (r600_get_array_mode_alignment(&array_check,
&pitch_align, &height_align, &depth_align, &base_align)) {
dev_warn(p->dev, "%s:%d tex array mode (%d) invalid\n",
@@ -1245,22 +1365,29 @@
word0 = radeon_get_ib_value(p, idx + 4);
word1 = radeon_get_ib_value(p, idx + 5);
blevel = G_038010_BASE_LEVEL(word0);
- nlevels = G_038014_LAST_LEVEL(word1);
- r600_texture_size(nfaces, blevel, nlevels, w0, h0, d0, bpe,
- (pitch_align * bpe),
+ llevel = G_038014_LAST_LEVEL(word1);
+ if (array == 1) {
+ barray = G_038014_BASE_ARRAY(word1);
+ larray = G_038014_LAST_ARRAY(word1);
+
+ nfaces = larray - barray + 1;
+ }
+ r600_texture_size(nfaces, blevel, llevel, w0, h0, d0, format,
+ pitch_align, height_align, base_align,
&l0_size, &mipmap_size);
/* using get ib will give us the offset into the texture bo */
word0 = radeon_get_ib_value(p, idx + 2) << 8;
if ((l0_size + word0) > radeon_bo_size(texture)) {
dev_warn(p->dev, "texture bo too small (%d %d %d %d -> %d have %ld)\n",
- w0, h0, bpe, word0, l0_size, radeon_bo_size(texture));
+ w0, h0, format, word0, l0_size, radeon_bo_size(texture));
+ dev_warn(p->dev, "alignments %d %d %d %lld\n", pitch, pitch_align, height_align, base_align);
return -EINVAL;
}
/* using get ib will give us the offset into the mipmap bo */
word0 = radeon_get_ib_value(p, idx + 3) << 8;
if ((mipmap_size + word0) > radeon_bo_size(mipmap)) {
/*dev_warn(p->dev, "mipmap bo too small (%d %d %d %d %d %d -> %d have %ld)\n",
- w0, h0, bpe, blevel, nlevels, word0, mipmap_size, radeon_bo_size(texture));*/
+ w0, h0, format, blevel, nlevels, word0, mipmap_size, radeon_bo_size(texture));*/
}
return 0;
}
diff --git a/drivers/gpu/drm/radeon/r600d.h b/drivers/gpu/drm/radeon/r600d.h
index a5d898b..d1f5986 100644
--- a/drivers/gpu/drm/radeon/r600d.h
+++ b/drivers/gpu/drm/radeon/r600d.h
@@ -1303,6 +1303,11 @@
#define V_038004_FMT_16_16_16_FLOAT 0x0000002E
#define V_038004_FMT_32_32_32 0x0000002F
#define V_038004_FMT_32_32_32_FLOAT 0x00000030
+#define V_038004_FMT_BC1 0x00000031
+#define V_038004_FMT_BC2 0x00000032
+#define V_038004_FMT_BC3 0x00000033
+#define V_038004_FMT_BC4 0x00000034
+#define V_038004_FMT_BC5 0x00000035
#define R_038010_SQ_TEX_RESOURCE_WORD4_0 0x038010
#define S_038010_FORMAT_COMP_X(x) (((x) & 0x3) << 0)
#define G_038010_FORMAT_COMP_X(x) (((x) >> 0) & 0x3)
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index 15e86b4..55fefe7 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -329,6 +329,7 @@
union radeon_gart_table table;
struct page **pages;
dma_addr_t *pages_addr;
+ bool *ttm_alloced;
bool ready;
};
@@ -341,7 +342,8 @@
void radeon_gart_unbind(struct radeon_device *rdev, unsigned offset,
int pages);
int radeon_gart_bind(struct radeon_device *rdev, unsigned offset,
- int pages, struct page **pagelist);
+ int pages, struct page **pagelist,
+ dma_addr_t *dma_addr);
/*
diff --git a/drivers/gpu/drm/radeon/radeon_gart.c b/drivers/gpu/drm/radeon/radeon_gart.c
index 9ac1bf0..f0534ef 100644
--- a/drivers/gpu/drm/radeon/radeon_gart.c
+++ b/drivers/gpu/drm/radeon/radeon_gart.c
@@ -149,8 +149,9 @@
p = t / (PAGE_SIZE / RADEON_GPU_PAGE_SIZE);
for (i = 0; i < pages; i++, p++) {
if (rdev->gart.pages[p]) {
- pci_unmap_page(rdev->pdev, rdev->gart.pages_addr[p],
- PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
+ if (!rdev->gart.ttm_alloced[p])
+ pci_unmap_page(rdev->pdev, rdev->gart.pages_addr[p],
+ PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
rdev->gart.pages[p] = NULL;
rdev->gart.pages_addr[p] = rdev->dummy_page.addr;
page_base = rdev->gart.pages_addr[p];
@@ -165,7 +166,7 @@
}
int radeon_gart_bind(struct radeon_device *rdev, unsigned offset,
- int pages, struct page **pagelist)
+ int pages, struct page **pagelist, dma_addr_t *dma_addr)
{
unsigned t;
unsigned p;
@@ -180,15 +181,22 @@
p = t / (PAGE_SIZE / RADEON_GPU_PAGE_SIZE);
for (i = 0; i < pages; i++, p++) {
- /* we need to support large memory configurations */
- /* assume that unbind have already been call on the range */
- rdev->gart.pages_addr[p] = pci_map_page(rdev->pdev, pagelist[i],
+ /* On TTM path, we only use the DMA API if TTM_PAGE_FLAG_DMA32
+ * is requested. */
+ if (dma_addr[i] != DMA_ERROR_CODE) {
+ rdev->gart.ttm_alloced[p] = true;
+ rdev->gart.pages_addr[p] = dma_addr[i];
+ } else {
+ /* we need to support large memory configurations */
+ /* assume that unbind have already been call on the range */
+ rdev->gart.pages_addr[p] = pci_map_page(rdev->pdev, pagelist[i],
0, PAGE_SIZE,
PCI_DMA_BIDIRECTIONAL);
- if (pci_dma_mapping_error(rdev->pdev, rdev->gart.pages_addr[p])) {
- /* FIXME: failed to map page (return -ENOMEM?) */
- radeon_gart_unbind(rdev, offset, pages);
- return -ENOMEM;
+ if (pci_dma_mapping_error(rdev->pdev, rdev->gart.pages_addr[p])) {
+ /* FIXME: failed to map page (return -ENOMEM?) */
+ radeon_gart_unbind(rdev, offset, pages);
+ return -ENOMEM;
+ }
}
rdev->gart.pages[p] = pagelist[i];
page_base = rdev->gart.pages_addr[p];
@@ -251,6 +259,12 @@
radeon_gart_fini(rdev);
return -ENOMEM;
}
+ rdev->gart.ttm_alloced = kzalloc(sizeof(bool) *
+ rdev->gart.num_cpu_pages, GFP_KERNEL);
+ if (rdev->gart.ttm_alloced == NULL) {
+ radeon_gart_fini(rdev);
+ return -ENOMEM;
+ }
/* set GART entry to point to the dummy page by default */
for (i = 0; i < rdev->gart.num_cpu_pages; i++) {
rdev->gart.pages_addr[i] = rdev->dummy_page.addr;
@@ -267,6 +281,8 @@
rdev->gart.ready = false;
kfree(rdev->gart.pages);
kfree(rdev->gart.pages_addr);
+ kfree(rdev->gart.ttm_alloced);
rdev->gart.pages = NULL;
rdev->gart.pages_addr = NULL;
+ rdev->gart.ttm_alloced = NULL;
}
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index e17e595..177adc8 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -513,6 +513,7 @@
if (r) {
return r;
}
+ rdev->mman.bdev.dev = rdev->dev;
/* No others user of address space so set it to 0 */
r = ttm_bo_device_init(&rdev->mman.bdev,
rdev->mman.bo_global_ref.ref.object,
@@ -647,6 +648,7 @@
unsigned long num_pages;
struct page **pages;
struct page *dummy_read_page;
+ dma_addr_t *dma_addrs;
bool populated;
bool bound;
unsigned offset;
@@ -655,12 +657,14 @@
static int radeon_ttm_backend_populate(struct ttm_backend *backend,
unsigned long num_pages,
struct page **pages,
- struct page *dummy_read_page)
+ struct page *dummy_read_page,
+ dma_addr_t *dma_addrs)
{
struct radeon_ttm_backend *gtt;
gtt = container_of(backend, struct radeon_ttm_backend, backend);
gtt->pages = pages;
+ gtt->dma_addrs = dma_addrs;
gtt->num_pages = num_pages;
gtt->dummy_read_page = dummy_read_page;
gtt->populated = true;
@@ -673,6 +677,7 @@
gtt = container_of(backend, struct radeon_ttm_backend, backend);
gtt->pages = NULL;
+ gtt->dma_addrs = NULL;
gtt->num_pages = 0;
gtt->dummy_read_page = NULL;
gtt->populated = false;
@@ -693,7 +698,7 @@
gtt->num_pages, bo_mem, backend);
}
r = radeon_gart_bind(gtt->rdev, gtt->offset,
- gtt->num_pages, gtt->pages);
+ gtt->num_pages, gtt->pages, gtt->dma_addrs);
if (r) {
DRM_ERROR("failed to bind %lu pages at 0x%08X\n",
gtt->num_pages, gtt->offset);
diff --git a/drivers/gpu/drm/ttm/ttm_agp_backend.c b/drivers/gpu/drm/ttm/ttm_agp_backend.c
index f999e36..1c4a72f 100644
--- a/drivers/gpu/drm/ttm/ttm_agp_backend.c
+++ b/drivers/gpu/drm/ttm/ttm_agp_backend.c
@@ -47,7 +47,8 @@
static int ttm_agp_populate(struct ttm_backend *backend,
unsigned long num_pages, struct page **pages,
- struct page *dummy_read_page)
+ struct page *dummy_read_page,
+ dma_addr_t *dma_addrs)
{
struct ttm_agp_backend *agp_be =
container_of(backend, struct ttm_agp_backend, backend);
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index af61fc2..0b6a55a 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -406,11 +406,12 @@
}
if (bo->mem.mem_type == TTM_PL_SYSTEM) {
+ if (bdev->driver->move_notify)
+ bdev->driver->move_notify(bo, mem);
bo->mem = *mem;
mem->mm_node = NULL;
goto moved;
}
-
}
if (bdev->driver->move_notify)
diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc.c b/drivers/gpu/drm/ttm/ttm_page_alloc.c
index b1e02ff..35849db 100644
--- a/drivers/gpu/drm/ttm/ttm_page_alloc.c
+++ b/drivers/gpu/drm/ttm/ttm_page_alloc.c
@@ -38,6 +38,7 @@
#include <linux/mm.h>
#include <linux/seq_file.h> /* for seq_printf */
#include <linux/slab.h>
+#include <linux/dma-mapping.h>
#include <asm/atomic.h>
@@ -662,7 +663,8 @@
* cached pages.
*/
int ttm_get_pages(struct list_head *pages, int flags,
- enum ttm_caching_state cstate, unsigned count)
+ enum ttm_caching_state cstate, unsigned count,
+ dma_addr_t *dma_address, struct device *dev)
{
struct ttm_page_pool *pool = ttm_get_pool(flags, cstate);
struct page *p = NULL;
@@ -681,14 +683,22 @@
gfp_flags |= GFP_HIGHUSER;
for (r = 0; r < count; ++r) {
- p = alloc_page(gfp_flags);
+ if ((flags & TTM_PAGE_FLAG_DMA32) && dma_address) {
+ void *addr;
+ addr = dma_alloc_coherent(dev, PAGE_SIZE,
+ &dma_address[r],
+ gfp_flags);
+ if (addr == NULL)
+ return -ENOMEM;
+ p = virt_to_page(addr);
+ } else
+ p = alloc_page(gfp_flags);
if (!p) {
printk(KERN_ERR TTM_PFX
"Unable to allocate page.");
return -ENOMEM;
}
-
list_add(&p->lru, pages);
}
return 0;
@@ -720,7 +730,7 @@
printk(KERN_ERR TTM_PFX
"Failed to allocate extra pages "
"for large request.");
- ttm_put_pages(pages, 0, flags, cstate);
+ ttm_put_pages(pages, 0, flags, cstate, NULL, NULL);
return r;
}
}
@@ -731,17 +741,30 @@
/* Put all pages in pages list to correct pool to wait for reuse */
void ttm_put_pages(struct list_head *pages, unsigned page_count, int flags,
- enum ttm_caching_state cstate)
+ enum ttm_caching_state cstate, dma_addr_t *dma_address,
+ struct device *dev)
{
unsigned long irq_flags;
struct ttm_page_pool *pool = ttm_get_pool(flags, cstate);
struct page *p, *tmp;
+ unsigned r;
if (pool == NULL) {
/* No pool for this memory type so free the pages */
+ r = page_count-1;
list_for_each_entry_safe(p, tmp, pages, lru) {
- __free_page(p);
+ if ((flags & TTM_PAGE_FLAG_DMA32) && dma_address) {
+ void *addr = page_address(p);
+ WARN_ON(!addr || !dma_address[r]);
+ if (addr)
+ dma_free_coherent(dev, PAGE_SIZE,
+ addr,
+ dma_address[r]);
+ dma_address[r] = 0;
+ } else
+ __free_page(p);
+ r--;
}
/* Make the pages list empty */
INIT_LIST_HEAD(pages);
diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c
index af789dc..0f8fc9f 100644
--- a/drivers/gpu/drm/ttm/ttm_tt.c
+++ b/drivers/gpu/drm/ttm/ttm_tt.c
@@ -49,12 +49,16 @@
static void ttm_tt_alloc_page_directory(struct ttm_tt *ttm)
{
ttm->pages = drm_calloc_large(ttm->num_pages, sizeof(*ttm->pages));
+ ttm->dma_address = drm_calloc_large(ttm->num_pages,
+ sizeof(*ttm->dma_address));
}
static void ttm_tt_free_page_directory(struct ttm_tt *ttm)
{
drm_free_large(ttm->pages);
ttm->pages = NULL;
+ drm_free_large(ttm->dma_address);
+ ttm->dma_address = NULL;
}
static void ttm_tt_free_user_pages(struct ttm_tt *ttm)
@@ -105,7 +109,8 @@
INIT_LIST_HEAD(&h);
- ret = ttm_get_pages(&h, ttm->page_flags, ttm->caching_state, 1);
+ ret = ttm_get_pages(&h, ttm->page_flags, ttm->caching_state, 1,
+ &ttm->dma_address[index], ttm->be->bdev->dev);
if (ret != 0)
return NULL;
@@ -164,7 +169,7 @@
}
be->func->populate(be, ttm->num_pages, ttm->pages,
- ttm->dummy_read_page);
+ ttm->dummy_read_page, ttm->dma_address);
ttm->state = tt_unbound;
return 0;
}
@@ -298,7 +303,8 @@
count++;
}
}
- ttm_put_pages(&h, count, ttm->page_flags, ttm->caching_state);
+ ttm_put_pages(&h, count, ttm->page_flags, ttm->caching_state,
+ ttm->dma_address, ttm->be->bdev->dev);
ttm->state = tt_unpopulated;
ttm->first_himem_page = ttm->num_pages;
ttm->last_lomem_page = -1;
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c b/drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c
index 80bc37b..87e43e0 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c
@@ -102,7 +102,8 @@
static int vmw_ttm_populate(struct ttm_backend *backend,
unsigned long num_pages, struct page **pages,
- struct page *dummy_read_page)
+ struct page *dummy_read_page,
+ dma_addr_t *dma_addrs)
{
struct vmw_ttm_backend *vmw_be =
container_of(backend, struct vmw_ttm_backend, backend);
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
index 96949b9..df04661 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
@@ -322,7 +322,7 @@
ttm_lock_set_kill(&dev_priv->fbdev_master.lock, false, SIGTERM);
dev_priv->active_master = &dev_priv->fbdev_master;
-
+ dev_priv->bdev.dev = dev->dev;
ret = ttm_bo_device_init(&dev_priv->bdev,
dev_priv->bo_global_ref.ref.object,
&vmw_bo_driver, VMWGFX_FILE_PAGE_OFFSET,
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c
index 29113c9..b3a2cd5 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c
@@ -345,7 +345,7 @@
return connector_status_disconnected;
}
-static struct drm_display_mode vmw_ldu_connector_builtin[] = {
+static const struct drm_display_mode vmw_ldu_connector_builtin[] = {
/* 640x480@60Hz */
{ DRM_MODE("640x480", DRM_MODE_TYPE_DRIVER, 25175, 640, 656,
752, 800, 0, 480, 489, 492, 525, 0,
@@ -429,7 +429,6 @@
struct drm_device *dev = connector->dev;
struct vmw_private *dev_priv = vmw_priv(dev);
struct drm_display_mode *mode = NULL;
- struct drm_display_mode *bmode;
struct drm_display_mode prefmode = { DRM_MODE("preferred",
DRM_MODE_TYPE_DRIVER | DRM_MODE_TYPE_PREFERRED,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
@@ -459,6 +458,8 @@
}
for (i = 0; vmw_ldu_connector_builtin[i].type != 0; i++) {
+ const struct drm_display_mode *bmode;
+
bmode = &vmw_ldu_connector_builtin[i];
if (bmode->hdisplay > max_width ||
bmode->vdisplay > max_height)
diff --git a/include/drm/drm_crtc.h b/include/drm/drm_crtc.h
index 080a6e3..60edf9b 100644
--- a/include/drm/drm_crtc.h
+++ b/include/drm/drm_crtc.h
@@ -659,7 +659,7 @@
extern void drm_mode_probed_add(struct drm_connector *connector, struct drm_display_mode *mode);
extern void drm_mode_remove(struct drm_connector *connector, struct drm_display_mode *mode);
extern struct drm_display_mode *drm_mode_duplicate(struct drm_device *dev,
- struct drm_display_mode *mode);
+ const struct drm_display_mode *mode);
extern void drm_mode_debug_printmodeline(struct drm_display_mode *mode);
extern void drm_mode_config_init(struct drm_device *dev);
extern void drm_mode_config_reset(struct drm_device *dev);
@@ -685,8 +685,8 @@
extern void drm_mode_prune_invalid(struct drm_device *dev,
struct list_head *mode_list, bool verbose);
extern void drm_mode_sort(struct list_head *mode_list);
-extern int drm_mode_hsync(struct drm_display_mode *mode);
-extern int drm_mode_vrefresh(struct drm_display_mode *mode);
+extern int drm_mode_hsync(const struct drm_display_mode *mode);
+extern int drm_mode_vrefresh(const struct drm_display_mode *mode);
extern void drm_mode_set_crtcinfo(struct drm_display_mode *p,
int adjust_flags);
extern void drm_mode_connector_list_update(struct drm_connector *connector);
diff --git a/include/drm/drm_hashtab.h b/include/drm/drm_hashtab.h
index 0af087a..3650d5d 100644
--- a/include/drm/drm_hashtab.h
+++ b/include/drm/drm_hashtab.h
@@ -45,14 +45,10 @@
};
struct drm_open_hash {
- unsigned int size;
- unsigned int order;
- unsigned int fill;
struct hlist_head *table;
- int use_vmalloc;
+ u8 order;
};
-
extern int drm_ht_create(struct drm_open_hash *ht, unsigned int order);
extern int drm_ht_insert_item(struct drm_open_hash *ht, struct drm_hash_item *item);
extern int drm_ht_just_insert_please(struct drm_open_hash *ht, struct drm_hash_item *item,
diff --git a/include/drm/ttm/ttm_bo_driver.h b/include/drm/ttm/ttm_bo_driver.h
index 1da8af6..38ff068 100644
--- a/include/drm/ttm/ttm_bo_driver.h
+++ b/include/drm/ttm/ttm_bo_driver.h
@@ -50,13 +50,15 @@
* @pages: Array of pointers to ttm pages.
* @dummy_read_page: Page to be used instead of NULL pages in the
* array @pages.
+ * @dma_addrs: Array of DMA (bus) address of the ttm pages.
*
* Populate the backend with ttm pages. Depending on the backend,
* it may or may not copy the @pages array.
*/
int (*populate) (struct ttm_backend *backend,
unsigned long num_pages, struct page **pages,
- struct page *dummy_read_page);
+ struct page *dummy_read_page,
+ dma_addr_t *dma_addrs);
/**
* struct ttm_backend_func member clear
*
@@ -149,6 +151,7 @@
* @swap_storage: Pointer to shmem struct file for swap storage.
* @caching_state: The current caching state of the pages.
* @state: The current binding state of the pages.
+ * @dma_address: The DMA (bus) addresses of the pages (if TTM_PAGE_FLAG_DMA32)
*
* This is a structure holding the pages, caching- and aperture binding
* status for a buffer object that isn't backed by fixed (VRAM / AGP)
@@ -173,6 +176,7 @@
tt_unbound,
tt_unpopulated,
} state;
+ dma_addr_t *dma_address;
};
#define TTM_MEMTYPE_FLAG_FIXED (1 << 0) /* Fixed (on-card) PCI memory */
@@ -547,6 +551,7 @@
struct list_head device_list;
struct ttm_bo_global *glob;
struct ttm_bo_driver *driver;
+ struct device *dev;
rwlock_t vm_lock;
struct ttm_mem_type_manager man[TTM_NUM_MEM_TYPES];
spinlock_t fence_lock;
diff --git a/include/drm/ttm/ttm_page_alloc.h b/include/drm/ttm/ttm_page_alloc.h
index 1168214..ccb6b7a 100644
--- a/include/drm/ttm/ttm_page_alloc.h
+++ b/include/drm/ttm/ttm_page_alloc.h
@@ -36,11 +36,15 @@
* @flags: ttm flags for page allocation.
* @cstate: ttm caching state for the page.
* @count: number of pages to allocate.
+ * @dma_address: The DMA (bus) address of pages (if TTM_PAGE_FLAG_DMA32 set).
+ * @dev: struct device for appropiate DMA accounting.
*/
int ttm_get_pages(struct list_head *pages,
int flags,
enum ttm_caching_state cstate,
- unsigned count);
+ unsigned count,
+ dma_addr_t *dma_address,
+ struct device *dev);
/**
* Put linked list of pages to pool.
*
@@ -49,11 +53,15 @@
* count.
* @flags: ttm flags for page allocation.
* @cstate: ttm caching state.
+ * @dma_address: The DMA (bus) address of pages (if TTM_PAGE_FLAG_DMA32 set).
+ * @dev: struct device for appropiate DMA accounting.
*/
void ttm_put_pages(struct list_head *pages,
unsigned page_count,
int flags,
- enum ttm_caching_state cstate);
+ enum ttm_caching_state cstate,
+ dma_addr_t *dma_address,
+ struct device *dev);
/**
* Initialize pool allocator.
*/