Skip to content

Commit

Permalink
Merge branch 'drm-intel-next' of git://git.kernel.org/pub/scm/linux/k…
Browse files Browse the repository at this point in the history 
…ernel/git/anholt/drm-intel

* 'drm-intel-next' of git://git.kernel.org/pub/scm/linux/kernel/git/anholt/drm-intel:
  drm/i915: fix scheduling while holding the new active list spinlock
  drm/i915: Allow tiling of objects with bit 17 swizzling by the CPU.
  drm/i915: Correctly set the write flag for get_user_pages in pread.
  drm/i915: Fix use of uninitialized var in 40a5f0d
  drm/i915: indicate framebuffer restore key in SysRq help message
  drm/i915: sync hdmi detection by hdmi identifier with 2D
  drm/i915: Fix a mismerge of the IGD patch (new .find_pll hooks missed)
  drm/i915: Implement batch and ring buffer dumping
  • Loading branch information
@torvalds
torvalds committed Apr 14, 2009
2 parents 6f66cbc + 68c8434 commit b897e6f
Show file tree
Hide file tree
Showing 9 changed files with 412 additions and 40 deletions.
7 changes: 7 additions & 0 deletions drivers/gpu/drm/i915/i915_drv.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -446,6 +446,9 @@ struct drm_i915_gem_object {
uint32_t tiling_mode;
uint32_t stride;

/** Record of address bit 17 of each page at last unbind. */
long *bit_17;

/** AGP mapping type (AGP_USER_MEMORY or AGP_USER_CACHED_MEMORY */
uint32_t agp_type;

Expand Down Expand Up @@ -635,9 +638,13 @@ int i915_gem_attach_phys_object(struct drm_device *dev,
void i915_gem_detach_phys_object(struct drm_device *dev,
struct drm_gem_object *obj);
void i915_gem_free_all_phys_object(struct drm_device *dev);
int i915_gem_object_get_pages(struct drm_gem_object *obj);
void i915_gem_object_put_pages(struct drm_gem_object *obj);

/* i915_gem_tiling.c */
void i915_gem_detect_bit_6_swizzle(struct drm_device *dev);
void i915_gem_object_do_bit_17_swizzle(struct drm_gem_object *obj);
void i915_gem_object_save_bit_17_swizzle(struct drm_gem_object *obj);

/* i915_gem_debug.c */
void i915_gem_dump_object(struct drm_gem_object *obj, int len,
Expand Down
187 changes: 156 additions & 31 deletions drivers/gpu/drm/i915/i915_gem.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -43,8 +43,6 @@ static int i915_gem_object_set_cpu_read_domain_range(struct drm_gem_object *obj,
uint64_t offset,
uint64_t size);
static void i915_gem_object_set_to_full_cpu_read_domain(struct drm_gem_object *obj);
static int i915_gem_object_get_pages(struct drm_gem_object *obj);
static void i915_gem_object_put_pages(struct drm_gem_object *obj);
static int i915_gem_object_wait_rendering(struct drm_gem_object *obj);
static int i915_gem_object_bind_to_gtt(struct drm_gem_object *obj,
unsigned alignment);
Expand Down Expand Up @@ -143,15 +141,27 @@ fast_shmem_read(struct page **pages,
int length)
{
char __iomem *vaddr;
int ret;
int unwritten;

vaddr = kmap_atomic(pages[page_base >> PAGE_SHIFT], KM_USER0);
if (vaddr == NULL)
return -ENOMEM;
ret = __copy_to_user_inatomic(data, vaddr + page_offset, length);
unwritten = __copy_to_user_inatomic(data, vaddr + page_offset, length);
kunmap_atomic(vaddr, KM_USER0);

return ret;
if (unwritten)
return -EFAULT;

return 0;
}

static int i915_gem_object_needs_bit17_swizzle(struct drm_gem_object *obj)
{
drm_i915_private_t *dev_priv = obj->dev->dev_private;
struct drm_i915_gem_object *obj_priv = obj->driver_private;

return dev_priv->mm.bit_6_swizzle_x == I915_BIT_6_SWIZZLE_9_10_17 &&
obj_priv->tiling_mode != I915_TILING_NONE;
}

static inline int
Expand Down Expand Up @@ -181,6 +191,64 @@ slow_shmem_copy(struct page *dst_page,
return 0;
}

static inline int
slow_shmem_bit17_copy(struct page *gpu_page,
int gpu_offset,
struct page *cpu_page,
int cpu_offset,
int length,
int is_read)
{
char *gpu_vaddr, *cpu_vaddr;

/* Use the unswizzled path if this page isn't affected. */
if ((page_to_phys(gpu_page) & (1 << 17)) == 0) {
if (is_read)
return slow_shmem_copy(cpu_page, cpu_offset,
gpu_page, gpu_offset, length);
else
return slow_shmem_copy(gpu_page, gpu_offset,
cpu_page, cpu_offset, length);
}

gpu_vaddr = kmap_atomic(gpu_page, KM_USER0);
if (gpu_vaddr == NULL)
return -ENOMEM;

cpu_vaddr = kmap_atomic(cpu_page, KM_USER1);
if (cpu_vaddr == NULL) {
kunmap_atomic(gpu_vaddr, KM_USER0);
return -ENOMEM;
}

/* Copy the data, XORing A6 with A17 (1). The user already knows he's
* XORing with the other bits (A9 for Y, A9 and A10 for X)
*/
while (length > 0) {
int cacheline_end = ALIGN(gpu_offset + 1, 64);
int this_length = min(cacheline_end - gpu_offset, length);
int swizzled_gpu_offset = gpu_offset ^ 64;

if (is_read) {
memcpy(cpu_vaddr + cpu_offset,
gpu_vaddr + swizzled_gpu_offset,
this_length);
} else {
memcpy(gpu_vaddr + swizzled_gpu_offset,
cpu_vaddr + cpu_offset,
this_length);
}
cpu_offset += this_length;
gpu_offset += this_length;
length -= this_length;
}

kunmap_atomic(cpu_vaddr, KM_USER1);
kunmap_atomic(gpu_vaddr, KM_USER0);

return 0;
}

/**
* This is the fast shmem pread path, which attempts to copy_from_user directly
* from the backing pages of the object to the user's address space. On a
Expand Down Expand Up @@ -269,6 +337,7 @@ i915_gem_shmem_pread_slow(struct drm_device *dev, struct drm_gem_object *obj,
int page_length;
int ret;
uint64_t data_ptr = args->data_ptr;
int do_bit17_swizzling;

remain = args->size;

Expand All @@ -286,13 +355,15 @@ i915_gem_shmem_pread_slow(struct drm_device *dev, struct drm_gem_object *obj,

down_read(&mm->mmap_sem);
pinned_pages = get_user_pages(current, mm, (uintptr_t)args->data_ptr,
num_pages, 0, 0, user_pages, NULL);
num_pages, 1, 0, user_pages, NULL);
up_read(&mm->mmap_sem);
if (pinned_pages < num_pages) {
ret = -EFAULT;
goto fail_put_user_pages;
}

do_bit17_swizzling = i915_gem_object_needs_bit17_swizzle(obj);

mutex_lock(&dev->struct_mutex);

ret = i915_gem_object_get_pages(obj);
Expand Down Expand Up @@ -327,11 +398,20 @@ i915_gem_shmem_pread_slow(struct drm_device *dev, struct drm_gem_object *obj,
if ((data_page_offset + page_length) > PAGE_SIZE)
page_length = PAGE_SIZE - data_page_offset;

ret = slow_shmem_copy(user_pages[data_page_index],
data_page_offset,
obj_priv->pages[shmem_page_index],
shmem_page_offset,
page_length);
if (do_bit17_swizzling) {
ret = slow_shmem_bit17_copy(obj_priv->pages[shmem_page_index],
shmem_page_offset,
user_pages[data_page_index],
data_page_offset,
page_length,
1);
} else {
ret = slow_shmem_copy(user_pages[data_page_index],
data_page_offset,
obj_priv->pages[shmem_page_index],
shmem_page_offset,
page_length);
}
if (ret)
goto fail_put_pages;

Expand Down Expand Up @@ -383,9 +463,14 @@ i915_gem_pread_ioctl(struct drm_device *dev, void *data,
return -EINVAL;
}

ret = i915_gem_shmem_pread_fast(dev, obj, args, file_priv);
if (ret != 0)
if (i915_gem_object_needs_bit17_swizzle(obj)) {
ret = i915_gem_shmem_pread_slow(dev, obj, args, file_priv);
} else {
ret = i915_gem_shmem_pread_fast(dev, obj, args, file_priv);
if (ret != 0)
ret = i915_gem_shmem_pread_slow(dev, obj, args,
file_priv);
}

drm_gem_object_unreference(obj);

Expand Down Expand Up @@ -727,6 +812,7 @@ i915_gem_shmem_pwrite_slow(struct drm_device *dev, struct drm_gem_object *obj,
int page_length;
int ret;
uint64_t data_ptr = args->data_ptr;
int do_bit17_swizzling;

remain = args->size;

Expand All @@ -751,6 +837,8 @@ i915_gem_shmem_pwrite_slow(struct drm_device *dev, struct drm_gem_object *obj,
goto fail_put_user_pages;
}

do_bit17_swizzling = i915_gem_object_needs_bit17_swizzle(obj);

mutex_lock(&dev->struct_mutex);

ret = i915_gem_object_get_pages(obj);
Expand Down Expand Up @@ -785,11 +873,20 @@ i915_gem_shmem_pwrite_slow(struct drm_device *dev, struct drm_gem_object *obj,
if ((data_page_offset + page_length) > PAGE_SIZE)
page_length = PAGE_SIZE - data_page_offset;

ret = slow_shmem_copy(obj_priv->pages[shmem_page_index],
shmem_page_offset,
user_pages[data_page_index],
data_page_offset,
page_length);
if (do_bit17_swizzling) {
ret = slow_shmem_bit17_copy(obj_priv->pages[shmem_page_index],
shmem_page_offset,
user_pages[data_page_index],
data_page_offset,
page_length,
0);
} else {
ret = slow_shmem_copy(obj_priv->pages[shmem_page_index],
shmem_page_offset,
user_pages[data_page_index],
data_page_offset,
page_length);
}
if (ret)
goto fail_put_pages;

Expand Down Expand Up @@ -854,6 +951,8 @@ i915_gem_pwrite_ioctl(struct drm_device *dev, void *data,
ret = i915_gem_gtt_pwrite_slow(dev, obj, args,
file_priv);
}
} else if (i915_gem_object_needs_bit17_swizzle(obj)) {
ret = i915_gem_shmem_pwrite_slow(dev, obj, args, file_priv);
} else {
ret = i915_gem_shmem_pwrite_fast(dev, obj, args, file_priv);
if (ret == -EFAULT) {
Expand Down Expand Up @@ -1285,7 +1384,7 @@ i915_gem_mmap_gtt_ioctl(struct drm_device *dev, void *data,
return 0;
}

static void
void
i915_gem_object_put_pages(struct drm_gem_object *obj)
{
struct drm_i915_gem_object *obj_priv = obj->driver_private;
Expand All @@ -1297,6 +1396,9 @@ i915_gem_object_put_pages(struct drm_gem_object *obj)
if (--obj_priv->pages_refcount != 0)
return;

if (obj_priv->tiling_mode != I915_TILING_NONE)
i915_gem_object_save_bit_17_swizzle(obj);

for (i = 0; i < page_count; i++)
if (obj_priv->pages[i] != NULL) {
if (obj_priv->dirty)
Expand Down Expand Up @@ -1494,8 +1596,19 @@ i915_gem_retire_request(struct drm_device *dev,

if (obj->write_domain != 0)
i915_gem_object_move_to_flushing(obj);
else
else {
/* Take a reference on the object so it won't be
* freed while the spinlock is held. The list
* protection for this spinlock is safe when breaking
* the lock like this since the next thing we do
* is just get the head of the list again.
*/
drm_gem_object_reference(obj);
i915_gem_object_move_to_inactive(obj);
spin_unlock(&dev_priv->mm.active_list_lock);
drm_gem_object_unreference(obj);
spin_lock(&dev_priv->mm.active_list_lock);
}
}
out:
spin_unlock(&dev_priv->mm.active_list_lock);
Expand Down Expand Up @@ -1884,7 +1997,7 @@ i915_gem_evict_everything(struct drm_device *dev)
return ret;
}

static int
int
i915_gem_object_get_pages(struct drm_gem_object *obj)
{
struct drm_i915_gem_object *obj_priv = obj->driver_private;
Expand Down Expand Up @@ -1922,6 +2035,10 @@ i915_gem_object_get_pages(struct drm_gem_object *obj)
}
obj_priv->pages[i] = page;
}

if (obj_priv->tiling_mode != I915_TILING_NONE)
i915_gem_object_do_bit_17_swizzle(obj);

return 0;
}

Expand Down Expand Up @@ -3002,13 +3119,13 @@ i915_gem_get_relocs_from_user(struct drm_i915_gem_exec_object *exec_list,
drm_free(*relocs, reloc_count * sizeof(**relocs),
DRM_MEM_DRIVER);
*relocs = NULL;
return ret;
return -EFAULT;
}

reloc_index += exec_list[i].relocation_count;
}

return ret;
return 0;
}

static int
Expand All @@ -3017,23 +3134,28 @@ i915_gem_put_relocs_to_user(struct drm_i915_gem_exec_object *exec_list,
struct drm_i915_gem_relocation_entry *relocs)
{
uint32_t reloc_count = 0, i;
int ret;
int ret = 0;

for (i = 0; i < buffer_count; i++) {
struct drm_i915_gem_relocation_entry __user *user_relocs;
int unwritten;

user_relocs = (void __user *)(uintptr_t)exec_list[i].relocs_ptr;

if (ret == 0) {
ret = copy_to_user(user_relocs,
&relocs[reloc_count],
exec_list[i].relocation_count *
sizeof(*relocs));
unwritten = copy_to_user(user_relocs,
&relocs[reloc_count],
exec_list[i].relocation_count *
sizeof(*relocs));

if (unwritten) {
ret = -EFAULT;
goto err;
}

reloc_count += exec_list[i].relocation_count;
}

err:
drm_free(relocs, reloc_count * sizeof(*relocs), DRM_MEM_DRIVER);

return ret;
Expand Down Expand Up @@ -3243,7 +3365,7 @@ i915_gem_execbuffer(struct drm_device *dev, void *data,
exec_offset = exec_list[args->buffer_count - 1].offset;

#if WATCH_EXEC
i915_gem_dump_object(object_list[args->buffer_count - 1],
i915_gem_dump_object(batch_obj,
args->batch_len,
__func__,
~0);
Expand Down Expand Up @@ -3308,10 +3430,12 @@ i915_gem_execbuffer(struct drm_device *dev, void *data,
(uintptr_t) args->buffers_ptr,
exec_list,
sizeof(*exec_list) * args->buffer_count);
if (ret)
if (ret) {
ret = -EFAULT;
DRM_ERROR("failed to copy %d exec entries "
"back to user (%d)\n",
args->buffer_count, ret);
}
}

/* Copy the updated relocations out regardless of current error
Expand Down Expand Up @@ -3593,6 +3717,7 @@ void i915_gem_free_object(struct drm_gem_object *obj)
i915_gem_free_mmap_offset(obj);

drm_free(obj_priv->page_cpu_valid, 1, DRM_MEM_DRIVER);
kfree(obj_priv->bit_17);
drm_free(obj->driver_private, 1, DRM_MEM_DRIVER);
}

Expand Down
Loading

0 comments on commit b897e6f

Please sign in to comment.