/**
* drm_gem_cma_create_with_handle - allocate an object with the given size and
* return a GEM handle to it
* @file_priv: DRM file-private structure to register the handle for
* @drm: DRM device
* @size: size of the object to allocate
* @handle: return location for the GEM handle
*
* This function creates a CMA GEM object, allocating a physically contiguous
* chunk of memory as backing store. The GEM object is then added to the list
* of object associated with the given file and a handle to it is returned.
*
* Returns:
* A struct drm_gem_cma_object * on success or an ERR_PTR()-encoded negative
* error code on failure.
*/
static struct drm_gem_cma_object *
drm_gem_cma_create_with_handle(struct drm_file *file_priv,
struct drm_device *drm, size_t size,
uint32_t *handle)
{
struct drm_gem_cma_object *cma_obj;
struct drm_gem_object *gem_obj;
int ret;
cma_obj = drm_gem_cma_create(drm, size);
if (IS_ERR(cma_obj))
return cma_obj;
gem_obj = &cma_obj->base;
/*
* allocate a id of idr table where the obj is registered
* and handle has the id what user can see.
*/
ret = drm_gem_handle_create(file_priv, gem_obj, handle);
if (ret)
goto err_handle_create;
/* drop reference from allocate - handle holds it now. */
drm_gem_object_unreference_unlocked(gem_obj);
return cma_obj;
err_handle_create:
drm->driver->gem_free_object(gem_obj);
return ERR_PTR(ret);
}
struct vc4_bo *vc4_bo_create(struct drm_device *dev, size_t unaligned_size,
bool allow_unzeroed)
{
size_t size = roundup(unaligned_size, PAGE_SIZE);
struct vc4_dev *vc4 = to_vc4_dev(dev);
struct drm_gem_cma_object *cma_obj;
struct vc4_bo *bo;
if (size == 0)
return ERR_PTR(-EINVAL);
/* First, try to get a vc4_bo from the kernel BO cache. */
bo = vc4_bo_get_from_cache(dev, size);
if (bo) {
if (!allow_unzeroed)
memset(bo->base.vaddr, 0, bo->base.base.size);
return bo;
}
cma_obj = drm_gem_cma_create(dev, size);
if (IS_ERR(cma_obj)) {
/*
* If we've run out of CMA memory, kill the cache of
* CMA allocations we've got laying around and try again.
*/
vc4_bo_cache_purge(dev);
cma_obj = drm_gem_cma_create(dev, size);
if (IS_ERR(cma_obj)) {
DRM_ERROR("Failed to allocate from CMA:\n");
vc4_bo_stats_dump(vc4);
return ERR_PTR(-ENOMEM);
}
}
return to_vc4_bo(&cma_obj->base);
}
static int xylon_drm_gem_dumb_create(struct drm_file *file_priv,
struct drm_device *dev,
struct drm_mode_create_dumb *args)
{
struct drm_gem_cma_object *cma_obj;
struct drm_gem_object *gem_obj;
struct xylon_drm_device *xdev = dev->dev_private;
unsigned int buff_width;
int ret;
ret = xylon_drm_crtc_get_param(xdev->crtc, &buff_width,
XYLON_DRM_CRTC_BUFF_WIDTH);
if (ret)
return ret;
args->pitch = buff_width * DIV_ROUND_UP(args->bpp, 8);
args->size = (u64)(buff_width * DIV_ROUND_UP(args->bpp, 8) *
args->height);
cma_obj = drm_gem_cma_create(dev, (unsigned int)args->size);
if (IS_ERR(cma_obj))
return PTR_ERR(cma_obj);
gem_obj = &cma_obj->base;
ret = drm_gem_handle_create(file_priv, gem_obj, &args->handle);
if (ret)
goto err_handle_create;
drm_gem_object_unreference_unlocked(gem_obj);
return PTR_ERR_OR_ZERO(cma_obj);
err_handle_create:
drm_gem_cma_free_object(gem_obj);
return ret;
}
static int vc4_create_rcl_bo(struct drm_device *dev, struct vc4_exec_info *exec,
struct vc4_rcl_setup *setup)
{
struct drm_vc4_submit_cl *args = exec->args;
bool has_bin = args->bin_cl_size != 0;
uint8_t min_x_tile = args->min_x_tile;
uint8_t min_y_tile = args->min_y_tile;
uint8_t max_x_tile = args->max_x_tile;
uint8_t max_y_tile = args->max_y_tile;
uint8_t xtiles = max_x_tile - min_x_tile + 1;
uint8_t ytiles = max_y_tile - min_y_tile + 1;
uint8_t x, y;
uint32_t size, loop_body_size;
size = VC4_PACKET_TILE_RENDERING_MODE_CONFIG_SIZE;
loop_body_size = VC4_PACKET_TILE_COORDINATES_SIZE;
if (args->flags & VC4_SUBMIT_CL_USE_CLEAR_COLOR) {
size += VC4_PACKET_CLEAR_COLORS_SIZE +
VC4_PACKET_TILE_COORDINATES_SIZE +
VC4_PACKET_STORE_TILE_BUFFER_GENERAL_SIZE;
}
if (setup->color_read) {
if (args->color_read.flags &
VC4_SUBMIT_RCL_SURFACE_READ_IS_FULL_RES) {
loop_body_size += VC4_PACKET_LOAD_FULL_RES_TILE_BUFFER_SIZE;
} else {
loop_body_size += VC4_PACKET_LOAD_TILE_BUFFER_GENERAL_SIZE;
}
}
if (setup->zs_read) {
if (args->zs_read.flags &
VC4_SUBMIT_RCL_SURFACE_READ_IS_FULL_RES) {
loop_body_size += VC4_PACKET_LOAD_FULL_RES_TILE_BUFFER_SIZE;
} else {
if (setup->color_read &&
!(args->color_read.flags &
VC4_SUBMIT_RCL_SURFACE_READ_IS_FULL_RES)) {
loop_body_size += VC4_PACKET_TILE_COORDINATES_SIZE;
loop_body_size += VC4_PACKET_STORE_TILE_BUFFER_GENERAL_SIZE;
}
loop_body_size += VC4_PACKET_LOAD_TILE_BUFFER_GENERAL_SIZE;
}
}
if (has_bin) {
size += VC4_PACKET_WAIT_ON_SEMAPHORE_SIZE;
loop_body_size += VC4_PACKET_BRANCH_TO_SUB_LIST_SIZE;
}
if (setup->msaa_color_write)
loop_body_size += VC4_PACKET_STORE_FULL_RES_TILE_BUFFER_SIZE;
if (setup->msaa_zs_write)
loop_body_size += VC4_PACKET_STORE_FULL_RES_TILE_BUFFER_SIZE;
if (setup->zs_write)
loop_body_size += VC4_PACKET_STORE_TILE_BUFFER_GENERAL_SIZE;
if (setup->color_write)
loop_body_size += VC4_PACKET_STORE_MS_TILE_BUFFER_SIZE;
/* We need a VC4_PACKET_TILE_COORDINATES in between each store. */
loop_body_size += VC4_PACKET_TILE_COORDINATES_SIZE *
((setup->msaa_color_write != NULL) +
(setup->msaa_zs_write != NULL) +
(setup->color_write != NULL) +
(setup->zs_write != NULL) - 1);
size += xtiles * ytiles * loop_body_size;
setup->rcl = drm_gem_cma_create(dev, size);
if (!setup->rcl)
return -ENOMEM;
list_addtail(&to_vc4_bo(&setup->rcl->base)->unref_head,
&exec->unref_list);
rcl_u8(setup, VC4_PACKET_TILE_RENDERING_MODE_CONFIG);
rcl_u32(setup,
(setup->color_write ? (setup->color_write->paddr +
args->color_write.offset) :
0));
rcl_u16(setup, args->width);
rcl_u16(setup, args->height);
rcl_u16(setup, args->color_write.bits);
/* The tile buffer gets cleared when the previous tile is stored. If
* the clear values changed between frames, then the tile buffer has
* stale clear values in it, so we have to do a store in None mode (no
* writes) so that we trigger the tile buffer clear.
*/
if (args->flags & VC4_SUBMIT_CL_USE_CLEAR_COLOR) {
rcl_u8(setup, VC4_PACKET_CLEAR_COLORS);
rcl_u32(setup, args->clear_color[0]);
rcl_u32(setup, args->clear_color[1]);
rcl_u32(setup, args->clear_z);
rcl_u8(setup, args->clear_s);
vc4_tile_coordinates(setup, 0, 0);
rcl_u8(setup, VC4_PACKET_STORE_TILE_BUFFER_GENERAL);
rcl_u16(setup, VC4_LOADSTORE_TILE_BUFFER_NONE);
rcl_u32(setup, 0); /* no address, since we're in None mode */
}
for (y = min_y_tile; y <= max_y_tile; y++) {
for (x = min_x_tile; x <= max_x_tile; x++) {
bool first = (x == min_x_tile && y == min_y_tile);
bool last = (x == max_x_tile && y == max_y_tile);
emit_tile(exec, setup, x, y, first, last);
}
}
BUG_ON(setup->next_offset != size);
exec->ct1ca = setup->rcl->paddr;
exec->ct1ea = setup->rcl->paddr + setup->next_offset;
return 0;
}
/*
* Copies in the user's binning command list and generates the validated bin
* CL, along with associated data (shader records, uniforms).
*/
static int
vc4_get_bcl(struct drm_device *dev, struct vc4_exec_info *exec)
{
struct drm_vc4_submit_cl *args = exec->args;
void *temp = NULL;
void *bin;
int ret = 0;
uint32_t bin_offset = 0;
uint32_t shader_rec_offset = roundup(bin_offset + args->bin_cl_size,
16);
uint32_t uniforms_offset = shader_rec_offset + args->shader_rec_size;
uint32_t exec_size = uniforms_offset + args->uniforms_size;
uint32_t temp_size = exec_size + (sizeof(struct vc4_shader_state) *
args->shader_rec_count);
if (uniforms_offset < shader_rec_offset ||
exec_size < uniforms_offset ||
args->shader_rec_count >= (UINT_MAX /
sizeof(struct vc4_shader_state)) ||
temp_size < exec_size) {
DRM_ERROR("overflow in exec arguments\n");
goto fail;
}
/* Allocate space where we'll store the copied in user command lists
* and shader records.
*
* We don't just copy directly into the BOs because we need to
* read the contents back for validation, and I think the
* bo->vaddr is uncached access.
*/
temp = kmalloc(temp_size, GFP_KERNEL);
if (!temp) {
DRM_ERROR("Failed to allocate storage for copying "
"in bin/render CLs.\n");
ret = -ENOMEM;
goto fail;
}
bin = temp + bin_offset;
exec->shader_rec_u = temp + shader_rec_offset;
exec->uniforms_u = temp + uniforms_offset;
exec->shader_state = temp + exec_size;
exec->shader_state_size = args->shader_rec_count;
ret = copy_from_user(bin,
(void __user *)(uintptr_t)args->bin_cl,
args->bin_cl_size);
if (ret) {
DRM_ERROR("Failed to copy in bin cl\n");
goto fail;
}
ret = copy_from_user(exec->shader_rec_u,
(void __user *)(uintptr_t)args->shader_rec,
args->shader_rec_size);
if (ret) {
DRM_ERROR("Failed to copy in shader recs\n");
goto fail;
}
ret = copy_from_user(exec->uniforms_u,
(void __user *)(uintptr_t)args->uniforms,
args->uniforms_size);
if (ret) {
DRM_ERROR("Failed to copy in uniforms cl\n");
goto fail;
}
exec->exec_bo = drm_gem_cma_create(dev, exec_size);
#if 0
if (IS_ERR(exec->exec_bo)) {
DRM_ERROR("Couldn't allocate BO for exec\n");
ret = PTR_ERR(exec->exec_bo);
exec->exec_bo = NULL;
goto fail;
}
#endif
list_addtail(&to_vc4_bo(&exec->exec_bo->base)->unref_head,
&exec->unref_list);
exec->ct0ca = exec->exec_bo->paddr + bin_offset;
exec->bin_u = bin;
exec->shader_rec_v = exec->exec_bo->vaddr + shader_rec_offset;
exec->shader_rec_p = exec->exec_bo->paddr + shader_rec_offset;
exec->shader_rec_size = args->shader_rec_size;
exec->uniforms_v = exec->exec_bo->vaddr + uniforms_offset;
exec->uniforms_p = exec->exec_bo->paddr + uniforms_offset;
exec->uniforms_size = args->uniforms_size;
ret = vc4_validate_bin_cl(dev,
exec->exec_bo->vaddr + bin_offset,
bin,
exec);
if (ret)
goto fail;
ret = vc4_validate_shader_recs(dev, exec);
fail:
kfree(temp);
return ret;
}
static int hisi_drm_fbdev_probe(struct drm_fb_helper *helper,
struct drm_fb_helper_surface_size *sizes)
{
struct hisi_drm_fbdev *fbdev = to_hisi_drm_fbdev(helper);
struct drm_mode_fb_cmd2 mode_cmd = { 0 };
struct drm_device *dev = helper->dev;
struct drm_gem_cma_object *obj;
struct drm_framebuffer *fb;
unsigned int bytes_per_pixel;
unsigned long offset;
struct fb_info *fbi;
size_t size;
int ret;
/* TODO: Need to use ion heaps to create frame buffer?? */
DRM_DEBUG_DRIVER("surface width(%d), height(%d) and bpp(%d)\n",
sizes->surface_width, sizes->surface_height,
sizes->surface_bpp);
bytes_per_pixel = DIV_ROUND_UP(sizes->surface_bpp, 8);
sizes->surface_depth = PREFERRED_BPP;
mode_cmd.width = sizes->surface_width;
mode_cmd.height = sizes->surface_height * HISI_NUM_FRAMEBUFFERS;
mode_cmd.pitches[0] = sizes->surface_width * bytes_per_pixel;
mode_cmd.pixel_format = drm_mode_legacy_fb_format(sizes->surface_bpp,
sizes->surface_depth);
size = roundup(mode_cmd.pitches[0] * mode_cmd.height, PAGE_SIZE);
obj = drm_gem_cma_create(dev, size);
if (IS_ERR(obj))
return -ENOMEM;
fbi = framebuffer_alloc(0, dev->dev);
if (!fbi) {
dev_err(dev->dev, "Failed to allocate framebuffer info.\n");
ret = -ENOMEM;
goto err_drm_gem_cma_free_object;
}
fbdev->fb = hisi_drm_fb_alloc(dev, &mode_cmd, &obj, 1, true);
if (IS_ERR(fbdev->fb)) {
dev_err(dev->dev, "Failed to allocate DRM framebuffer.\n");
ret = PTR_ERR(fbdev->fb);
goto err_framebuffer_release;
}
fb = &fbdev->fb->fb;
helper->fb = fb;
helper->fbdev = fbi;
fbi->par = helper;
fbi->flags = FBINFO_FLAG_DEFAULT;
fbi->fbops = &hisi_drm_fbdev_ops;
ret = fb_alloc_cmap(&fbi->cmap, 256, 0);
if (ret) {
dev_err(dev->dev, "Failed to allocate color map.\n");
goto err_hisi_drm_fb_destroy;
}
drm_fb_helper_fill_fix(fbi, fb->pitches[0], fb->depth);
drm_fb_helper_fill_var(fbi, helper, fb->width,
fb->height/HISI_NUM_FRAMEBUFFERS);
offset = fbi->var.xoffset * bytes_per_pixel;
offset += fbi->var.yoffset * fb->pitches[0];
dev->mode_config.fb_base = (resource_size_t)obj->paddr;
fbi->screen_base = obj->vaddr + offset;
fbi->fix.smem_start = (unsigned long)(obj->paddr + offset);
fbi->screen_size = size;
fbi->fix.smem_len = size;
DRM_DEBUG_DRIVER("exit successfully.\n");
return 0;
err_hisi_drm_fb_destroy:
drm_framebuffer_unregister_private(fb);
hisi_drm_fb_destroy(fb);
err_framebuffer_release:
framebuffer_release(fbi);
err_drm_gem_cma_free_object:
drm_gem_cma_free_object(&obj->base);
return ret;
}
