Hi Liviu,

On Fri, Jul 26, 2019 at 01:08:23PM +0000, Liviu Dudau wrote:

Hi Lowry,

On Fri, Jul 26, 2019 at 07:51:02AM +0000, Lowry Li (Arm Technology China) wrote:

...

The drm_format_info doesn't have any cpp or block_size (both are zero) information for arm only afbc format YU08/YU10. we need to compute it by ourselves.

Signed-off-by: Lowry Li (Arm Technology China) lowry.li@arm.com

.../drm/arm/display/komeda/komeda_format_caps.c | 23 ++++++++++++++++++++++ .../drm/arm/display/komeda/komeda_format_caps.h | 3 +++ .../drm/arm/display/komeda/komeda_framebuffer.c | 6 ++++-- 3 files changed, 30 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/arm/display/komeda/komeda_format_caps.c b/drivers/gpu/drm/arm/display/komeda/komeda_format_caps.c index cd4d9f5..3c9e060 100644 --- a/drivers/gpu/drm/arm/display/komeda/komeda_format_caps.c +++ b/drivers/gpu/drm/arm/display/komeda/komeda_format_caps.c @@ -35,6 +35,29 @@ return NULL; }

+u32 komeda_get_afbc_format_bpp(const struct drm_format_info *info, u64 modifier) +{

• u32 bpp;
• WARN_ON(modifier == DRM_FORMAT_MOD_LINEAR);

Is it not better to return the value from info->char_per_block if modifier is DRM_FORMAT_MOD_LINEAR? Or return 0?

This function is only working for the afbc format and will remove this WARN_ON check.

...

• switch (info->format) {
• case DRM_FORMAT_YUV420_8BIT:

• bpp = 12;
  

• break;
  

• case DRM_FORMAT_YUV420_10BIT:

• bpp = 15;
  

• break;
  

• default:

• bpp = info->cpp[0] * 8;
  

• break;
  

• }
• WARN_ON(bpp == 0);
• return bpp;

+}

/* Two assumptions

• 1. RGB always has YTR
• 2. Tiled RGB always has SC

diff --git a/drivers/gpu/drm/arm/display/komeda/komeda_format_caps.h b/drivers/gpu/drm/arm/display/komeda/komeda_format_caps.h index 3631910..32273cf 100644 --- a/drivers/gpu/drm/arm/display/komeda/komeda_format_caps.h +++ b/drivers/gpu/drm/arm/display/komeda/komeda_format_caps.h @@ -97,6 +97,9 @@ static inline const char *komeda_get_format_name(u32 fourcc, u64 modifier) komeda_get_format_caps(struct komeda_format_caps_table *table, u32 fourcc, u64 modifier);

+u32 komeda_get_afbc_format_bpp(const struct drm_format_info *info,

• 	       u64 modifier);
  

u32 *komeda_get_layer_fourcc_list(struct komeda_format_caps_table *table, u32 layer_type, u32 *n_fmts);

diff --git a/drivers/gpu/drm/arm/display/komeda/komeda_framebuffer.c b/drivers/gpu/drm/arm/display/komeda/komeda_framebuffer.c index 10bf63e..966d0c7 100644 --- a/drivers/gpu/drm/arm/display/komeda/komeda_framebuffer.c +++ b/drivers/gpu/drm/arm/display/komeda/komeda_framebuffer.c @@ -44,7 +44,7 @@ static int komeda_fb_create_handle(struct drm_framebuffer *fb, const struct drm_format_info *info = fb->format; struct drm_gem_object *obj; u32 alignment_w = 0, alignment_h = 0, alignment_header;

• u32 n_blocks = 0, min_size = 0;

• u32 n_blocks = 0, min_size = 0, bpp;

  obj = drm_gem_object_lookup(file, mode_cmd->handles[0]); if (!obj) {

@@ -86,10 +86,12 @@ static int komeda_fb_create_handle(struct drm_framebuffer *fb, kfb->offset_payload = ALIGN(n_blocks * AFBC_HEADER_SIZE, alignment_header);

• bpp = komeda_get_afbc_format_bpp(info, fb->modifier); kfb->afbc_size = kfb->offset_payload + n_blocks *

• 	 ALIGN(info->cpp[0] * AFBC_SUPERBLK_PIXELS,
  

• 	 ALIGN(bpp * AFBC_SUPERBLK_PIXELS / 8,
         AFBC_SUPERBLK_ALIGNMENT);
  

  min_size = kfb->afbc_size + fb->offsets[0];
• if (min_size > obj->size) { DRM_DEBUG_KMS("afbc size check failed, obj_size: 0x%lx. min_size 0x%x.\n", obj->size, min_size);

Patch doesn't apply to tip of drm-misc-next. What is this patch depending on?

Best regards, Liviu

Will rebase and send again. Thanks for the review.

Best regards, Lowry

...

-- 1.9.1

--

| I would like to | | fix the world, | | but they're not | | giving me the | \ source code! /

---

¯\_(ツ)_/¯