Re: [RFC PATCH v2 1/3] drm/fourcc: Add 'bpp' field for formats with non-integer bytes-per-pixel

[Liviu Dudau]

On Mon, Sep 10, 2018 at 09:53:25PM +0200, Daniel Vetter wrote:
> On Mon, Sep 10, 2018 at 09:50:03AM +0100, Brian Starkey wrote:
> > Hi,
> > 
> > On Fri, Sep 07, 2018 at 09:28:44PM +0200, Daniel Vetter wrote:
> > > On Fri, Sep 07, 2018 at 01:45:36PM +0100, Brian Starkey wrote:
> > > > Hi Daniel,
> > > > 
> > > > On Fri, Aug 31, 2018 at 10:17:30AM +0200, Daniel Vetter wrote:
> > > > > On Thu, Aug 23, 2018 at 04:23:41PM +0100, Brian Starkey wrote:
> > > > > > Some formats have a non-integer number of bytes per pixel, which can't
> > > > > > be handled with the existing 'cpp' field in drm_format_info. To handle
> > > > > > these formats, add a 'bpp' field, which is only used if cpp[0] == 0.
> > > > > >
> > > > > > This updates all the users of format->cpp in the core DRM code,
> > > > > > converting them to use a new function to get the bits-per-pixel for any
> > > > > > format.
> > > > > >
> > > > > > It's assumed that drivers will use the 'bpp' field when they add support
> > > > > > for pixel formats with non-integer bytes-per-pixel.
> > > > > >
> > > > > > Signed-off-by: Brian Starkey <brian.starkey@xxxxxxx>
> > > > >
> > > > > I assume you still require that stuff is eventually aligned to bytes? In
> > > > > that case, can we subsume this into the tile work Alex is doing? It's
> > > > > essentially just another special case of having storage-size units
> > > > > measured in bytes which span more than 1x1 pixel. And I kinda don't want a
> > > > > metric pile of special cases here in the format code, because that just
> > > > > means every driver handles a different subset, with different bugs.
> > > > > -Daniel
> > > > 
> > > > Sorry for the delay, been struggling to free some cycles to think
> > > > about this.
> > > > 
> > > > I'm not sure how to pull this in with the tiling stuff. In the AFBC
> > > > case then our AFBC superblocks are always nice round numbers (256
> > > > pixels), and so it does end up being a multiple of bytes.
> > > > 
> > > > However, AFBC supports different superblock sizes, so picking just one
> > > > doesn't really work out, and putting AFBC in the core format table
> > > > which reflects AFBC doesn't seem good.
> > > > 
> > > > We could make something up (e.g. call these formats "tiled" with 2x4
> > > > tiles, which guarantees a multiple of 8), but it would be an
> > > > arbitrarily-selected lie, which often seems to spell trouble. If we
> > > > did do that, would you re-define cpp as "bytes-per-tile"? Otherwise
> > > > we still need to add a new field anyway.
> > > > 
> > > > What's the pile of special cases you're worried about? The helper I've
> > > > added here means that drivers which need to care can use one API and
> > > > not implement their own bugs.
> > > 
> > > I'm confused ... the new bits-per-pixel stuff you're adding here is for
> > > yuv formats, not afbc. I'm just suggesting we have only 1 way of
> > > describing such formats that need more descriptive power than cpp, whether
> > > they have some kind of pixel-groups or small tiles.
> > 
> > Well, not really. The three formats which have non-integer cpp are:
> > DRM_FORMAT_VUY101010, DRM_FORMAT_YUV420_8BIT and
> > DRM_FORMAT_YUV420_10BIT. These formats are only valid with non-linear
> > modifiers (no linear encoding is defined). Mali only supports them
> > with AFBC.
> > 
> > The formats themselves have no notion of tiling or grouping - the
> > modifier adds that. I'm not aware of any non-AFBC uses of these
> > formats, so I don't want to "make up" a small-tile layout restriction
> > for them.
> 
> Ah, I missed that.
> 
> > > For very special stuff like afbc you need to validate in the driver
> > > anyway, too complicated. So I have no idea why you bring this up here?
> > 
> > Sure, we can just let drivers provide their own format_info's for
> > these, if that's what you prefer. The core format checking code can
> > error out if it ever encounters them.
> 
> It's format_info we're talking about. What I mean is that you just set all
> these to 0 and let the format_info code ignore it. And then having a
> bespoke drm_format_check_afbc helper function or similar, which checks all
> the layout restrictions of afbc.
> 
> I still maintain that bpp and tile_size are equavalent, and we really
> don't need both. Both are defacto a form of numerator/denumerator. If you
> don't like that you have to introduce "fake" tiles for afbc, then we can
> rename tile_size to numerator and tile_h/w to denumerator_h/w. Doesn't
> change one bit of the math. bpp simply hardcodes a denumerator of 8, and I
> don't see why we need that special case. Except if you love to write
> redundant self tests for all the math :-)

My $.02 worth of thoughts:

I get the fact that Daniel doesn't like us to add 3 new variables into
format_info (bpp, tile_w, tile_h) and that adding a "bits_per_unit"
variable should be able to take care of linear (where unit = 1 pixel)
and tiled (where unit = tile_w * tile_h pixels) formats. And I also see
Daniel's option 2 below, where he says it is reasonable to check AFBC
without using format_info. 

However, the problem we are trying to solve is 2 fold: we are trying to
calculate the size of the framebuffer (and the "bits_per_unit" or
Brian's bpp is useful for that), but we also try to validate the sizes
passed by userspace based on the drm_fourcc.h+modifier info. In that
case, the driver still needs to store somewhere the tile_w/tile_h for
that given format in order to check that the stride is a whole multiple
of tile sizes, so we thought that putting it in format_info is not
entirely pointless, because others might use those variables in order to
do their driver specific validation, without creating new structures.

Did I capture the discussion correctly? If so, can we agree that it is
not just the framebuffer size calculation that matters and that tiled
formats validation requires a tile_w/tile_h info, therefore Alex's
patches and Brian's need to be discussed separately (so that we can
bikeshed on whether format_info is the right place or not)?

Best regards,
Liviu

> 
> So two options that I think are reasonable:
> - one common numerator/denumerator. I don't care how you call that
>   bikeshed.
> - don't check afbc using format_info, have your own helper that does that
>   using custom code.
> 
> Cheers, Daniel
> 
> > Cheers,
> > -Brian
> > 
> > > -Daniel
> > > 
> > > > 
> > > > Cheers,
> > > > -Brian
> > > > 
> > > > >
> > > > > > ---
> > > > > >  drivers/gpu/drm/drm_fb_cma_helper.c          |  6 +++-
> > > > > >  drivers/gpu/drm/drm_fb_helper.c              |  8 +++--
> > > > > >  drivers/gpu/drm/drm_fourcc.c                 | 50 ++++++++++++++++++++++++++++
> > > > > >  drivers/gpu/drm/drm_framebuffer.c            |  8 ++---
> > > > > >  drivers/gpu/drm/drm_gem_framebuffer_helper.c |  3 +-
> > > > > >  include/drm/drm_fourcc.h                     |  4 +++
> > > > > >  6 files changed, 70 insertions(+), 9 deletions(-)
> > > > > >
> > > > > > diff --git a/drivers/gpu/drm/drm_fb_cma_helper.c b/drivers/gpu/drm/drm_fb_cma_helper.c
> > > > > > index 186d00adfb5f..e279d70d3e60 100644
> > > > > > --- a/drivers/gpu/drm/drm_fb_cma_helper.c
> > > > > > +++ b/drivers/gpu/drm/drm_fb_cma_helper.c
> > > > > > @@ -118,13 +118,17 @@ dma_addr_t drm_fb_cma_get_gem_addr(struct drm_framebuffer *fb,
> > > > > >  {
> > > > > >  	struct drm_gem_cma_object *obj;
> > > > > >  	dma_addr_t paddr;
> > > > > > +	u8 bpp = drm_format_info_plane_bpp(fb->format, plane);
> > > > > > +
> > > > > > +	/* This can't work for non-integer bytes-per-pixel */
> > > > > > +	WARN_ON(bpp % 8);
> > > > > >
> > > > > >  	obj = drm_fb_cma_get_gem_obj(fb, plane);
> > > > > >  	if (!obj)
> > > > > >  		return 0;
> > > > > >
> > > > > >  	paddr = obj->paddr + fb->offsets[plane];
> > > > > > -	paddr += fb->format->cpp[plane] * (state->src_x >> 16);
> > > > > > +	paddr += (bpp / 8) * (state->src_x >> 16);
> > > > > >  	paddr += fb->pitches[plane] * (state->src_y >> 16);
> > > > > >
> > > > > >  	return paddr;
> > > > > > diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
> > > > > > index 0646b108030b..ab369f250af4 100644
> > > > > > --- a/drivers/gpu/drm/drm_fb_helper.c
> > > > > > +++ b/drivers/gpu/drm/drm_fb_helper.c
> > > > > > @@ -1572,6 +1572,7 @@ int drm_fb_helper_check_var(struct fb_var_screeninfo *var,
> > > > > >  	struct drm_fb_helper *fb_helper = info->par;
> > > > > >  	struct drm_framebuffer *fb = fb_helper->fb;
> > > > > >  	int depth;
> > > > > > +	u8 bpp = drm_format_info_plane_bpp(fb->format, 0);
> > > > > >
> > > > > >  	if (var->pixclock != 0 || in_dbg_master())
> > > > > >  		return -EINVAL;
> > > > > > @@ -1580,14 +1581,14 @@ int drm_fb_helper_check_var(struct fb_var_screeninfo *var,
> > > > > >  	 * Changes struct fb_var_screeninfo are currently not pushed back
> > > > > >  	 * to KMS, hence fail if different settings are requested.
> > > > > >  	 */
> > > > > > -	if (var->bits_per_pixel != fb->format->cpp[0] * 8 ||
> > > > > > +	if (var->bits_per_pixel != bpp ||
> > > > > >  	    var->xres > fb->width || var->yres > fb->height ||
> > > > > >  	    var->xres_virtual > fb->width || var->yres_virtual > fb->height) {
> > > > > >  		DRM_DEBUG("fb requested width/height/bpp can't fit in current fb "
> > > > > >  			  "request %dx%d-%d (virtual %dx%d) > %dx%d-%d\n",
> > > > > >  			  var->xres, var->yres, var->bits_per_pixel,
> > > > > >  			  var->xres_virtual, var->yres_virtual,
> > > > > > -			  fb->width, fb->height, fb->format->cpp[0] * 8);
> > > > > > +			  fb->width, fb->height, bpp);
> > > > > >  		return -EINVAL;
> > > > > >  	}
> > > > > >
> > > > > > @@ -1949,11 +1950,12 @@ void drm_fb_helper_fill_var(struct fb_info *info, struct drm_fb_helper *fb_helpe
> > > > > >  			    uint32_t fb_width, uint32_t fb_height)
> > > > > >  {
> > > > > >  	struct drm_framebuffer *fb = fb_helper->fb;
> > > > > > +	u8 bpp = drm_format_info_plane_bpp(fb->format, 0);
> > > > > >
> > > > > >  	info->pseudo_palette = fb_helper->pseudo_palette;
> > > > > >  	info->var.xres_virtual = fb->width;
> > > > > >  	info->var.yres_virtual = fb->height;
> > > > > > -	info->var.bits_per_pixel = fb->format->cpp[0] * 8;
> > > > > > +	info->var.bits_per_pixel = bpp;
> > > > > >  	info->var.accel_flags = FB_ACCELF_TEXT;
> > > > > >  	info->var.xoffset = 0;
> > > > > >  	info->var.yoffset = 0;
> > > > > > diff --git a/drivers/gpu/drm/drm_fourcc.c b/drivers/gpu/drm/drm_fourcc.c
> > > > > > index 3b42c25bd58d..bb28919c32f3 100644
> > > > > > --- a/drivers/gpu/drm/drm_fourcc.c
> > > > > > +++ b/drivers/gpu/drm/drm_fourcc.c
> > > > > > @@ -272,10 +272,60 @@ int drm_format_plane_cpp(uint32_t format, int plane)
> > > > > >  	if (!info || plane >= info->num_planes)
> > > > > >  		return 0;
> > > > > >
> > > > > > +	/*
> > > > > > +	 * Not valid for formats with non-integer cpp,
> > > > > > +	 * use drm_format{_info}_plane_bpp instead
> > > > > > +	 */
> > > > > > +	WARN_ON(!info->cpp[0]);
> > > > > > +
> > > > > >  	return info->cpp[plane];
> > > > > >  }
> > > > > >  EXPORT_SYMBOL(drm_format_plane_cpp);
> > > > > >
> > > > > > +/**
> > > > > > + * drm_format_plane_bpp - determine the bits per pixel value
> > > > > > + * @format: pixel format (DRM_FORMAT_*)
> > > > > > + * @plane: plane index
> > > > > > + *
> > > > > > + * Returns:
> > > > > > + * The bits per pixel value for the specified plane.
> > > > > > + */
> > > > > > +int drm_format_plane_bpp(uint32_t format, int plane)
> > > > > > +{
> > > > > > +	const struct drm_format_info *info;
> > > > > > +
> > > > > > +	info = drm_format_info(format);
> > > > > > +	if (!info)
> > > > > > +		return 0;
> > > > > > +
> > > > > > +	return drm_format_info_plane_bpp(info, plane);
> > > > > > +}
> > > > > > +EXPORT_SYMBOL(drm_format_plane_bpp);
> > > > > > +
> > > > > > +/**
> > > > > > + * drm_format_info_plane_bpp - determine the bits per pixel value
> > > > > > + *
> > > > > > + * Convenience function which handles formats with both integer
> > > > > > + * and non-integer bytes-per-pixel.
> > > > > > + *
> > > > > > + * @format: pixel format info structure
> > > > > > + * @plane: plane index
> > > > > > + *
> > > > > > + * Returns:
> > > > > > + * The bits per pixel value for the specified plane.
> > > > > > + */
> > > > > > +int drm_format_info_plane_bpp(const struct drm_format_info *info, int plane)
> > > > > > +{
> > > > > > +	if (plane >= info->num_planes)
> > > > > > +		return 0;
> > > > > > +
> > > > > > +	if (info->cpp[0])
> > > > > > +		return info->cpp[plane] * 8;
> > > > > > +
> > > > > > +	return info->bpp[plane];
> > > > > > +}
> > > > > > +EXPORT_SYMBOL(drm_format_info_plane_bpp);
> > > > > > +
> > > > > >  /**
> > > > > >   * drm_format_horz_chroma_subsampling - get the horizontal chroma subsampling factor
> > > > > >   * @format: pixel format (DRM_FORMAT_*)
> > > > > > diff --git a/drivers/gpu/drm/drm_framebuffer.c b/drivers/gpu/drm/drm_framebuffer.c
> > > > > > index 8c4d32adcc17..7e00360ff70d 100644
> > > > > > --- a/drivers/gpu/drm/drm_framebuffer.c
> > > > > > +++ b/drivers/gpu/drm/drm_framebuffer.c
> > > > > > @@ -185,20 +185,20 @@ static int framebuffer_check(struct drm_device *dev,
> > > > > >  	for (i = 0; i < info->num_planes; i++) {
> > > > > >  		unsigned int width = fb_plane_width(r->width, info, i);
> > > > > >  		unsigned int height = fb_plane_height(r->height, info, i);
> > > > > > -		unsigned int cpp = info->cpp[i];
> > > > > > +		unsigned int bpp = drm_format_info_plane_bpp(info, i);
> > > > > >
> > > > > >  		if (!r->handles[i]) {
> > > > > >  			DRM_DEBUG_KMS("no buffer object handle for plane %d\n", i);
> > > > > >  			return -EINVAL;
> > > > > >  		}
> > > > > >
> > > > > > -		if ((uint64_t) width * cpp > UINT_MAX)
> > > > > > +		if ((uint64_t) DIV_ROUND_UP(width * bpp, 8) > UINT_MAX)
> > > > > >  			return -ERANGE;
> > > > > >
> > > > > >  		if ((uint64_t) height * r->pitches[i] + r->offsets[i] > UINT_MAX)
> > > > > >  			return -ERANGE;
> > > > > >
> > > > > > -		if (r->pitches[i] < width * cpp) {
> > > > > > +		if ((uint64_t) r->pitches[i] * 8 < (uint64_t) width * bpp) {
> > > > > >  			DRM_DEBUG_KMS("bad pitch %u for plane %d\n", r->pitches[i], i);
> > > > > >  			return -EINVAL;
> > > > > >  		}
> > > > > > @@ -476,7 +476,7 @@ int drm_mode_getfb(struct drm_device *dev,
> > > > > >  	r->height = fb->height;
> > > > > >  	r->width = fb->width;
> > > > > >  	r->depth = fb->format->depth;
> > > > > > -	r->bpp = fb->format->cpp[0] * 8;
> > > > > > +	r->bpp = drm_format_info_plane_bpp(fb->format, 0);
> > > > > >  	r->pitch = fb->pitches[0];
> > > > > >
> > > > > >  	/* GET_FB() is an unprivileged ioctl so we must not return a
> > > > > > diff --git a/drivers/gpu/drm/drm_gem_framebuffer_helper.c b/drivers/gpu/drm/drm_gem_framebuffer_helper.c
> > > > > > index acfbc0641a06..dfe224ccaeba 100644
> > > > > > --- a/drivers/gpu/drm/drm_gem_framebuffer_helper.c
> > > > > > +++ b/drivers/gpu/drm/drm_gem_framebuffer_helper.c
> > > > > > @@ -161,6 +161,7 @@ drm_gem_fb_create_with_funcs(struct drm_device *dev, struct drm_file *file,
> > > > > >  		unsigned int width = mode_cmd->width / (i ? info->hsub : 1);
> > > > > >  		unsigned int height = mode_cmd->height / (i ? info->vsub : 1);
> > > > > >  		unsigned int min_size;
> > > > > > +		u8 bpp = drm_format_info_plane_bpp(fb->format, i);
> > > > > >
> > > > > >  		objs[i] = drm_gem_object_lookup(file, mode_cmd->handles[i]);
> > > > > >  		if (!objs[i]) {
> > > > > > @@ -170,7 +171,7 @@ drm_gem_fb_create_with_funcs(struct drm_device *dev, struct drm_file *file,
> > > > > >  		}
> > > > > >
> > > > > >  		min_size = (height - 1) * mode_cmd->pitches[i]
> > > > > > -			 + width * info->cpp[i]
> > > > > > +			 + DIV_ROUND_UP(width * bpp, 8)
> > > > > >  			 + mode_cmd->offsets[i];
> > > > > >
> > > > > >  		if (objs[i]->size < min_size) {
> > > > > > diff --git a/include/drm/drm_fourcc.h b/include/drm/drm_fourcc.h
> > > > > > index 3e86408dac9f..d4af4dab1623 100644
> > > > > > --- a/include/drm/drm_fourcc.h
> > > > > > +++ b/include/drm/drm_fourcc.h
> > > > > > @@ -36,6 +36,7 @@ struct drm_mode_fb_cmd2;
> > > > > >   *	use in new code and set to 0 for new formats.
> > > > > >   * @num_planes: Number of color planes (1 to 3)
> > > > > >   * @cpp: Number of bytes per pixel (per plane)
> > > > > > + * @bpp: Number of bits per pixel (per plane), only valid if cpp[0] == 0.
> > > > > >   * @hsub: Horizontal chroma subsampling factor
> > > > > >   * @vsub: Vertical chroma subsampling factor
> > > > > >   * @has_alpha: Does the format embeds an alpha component?
> > > > > > @@ -45,6 +46,7 @@ struct drm_format_info {
> > > > > >  	u8 depth;
> > > > > >  	u8 num_planes;
> > > > > >  	u8 cpp[3];
> > > > > > +	u8 bpp[3];
> > > > > >  	u8 hsub;
> > > > > >  	u8 vsub;
> > > > > >  	bool has_alpha;
> > > > > > @@ -66,6 +68,8 @@ drm_get_format_info(struct drm_device *dev,
> > > > > >  uint32_t drm_mode_legacy_fb_format(uint32_t bpp, uint32_t depth);
> > > > > >  int drm_format_num_planes(uint32_t format);
> > > > > >  int drm_format_plane_cpp(uint32_t format, int plane);
> > > > > > +int drm_format_plane_bpp(uint32_t format, int plane);
> > > > > > +int drm_format_info_plane_bpp(const struct drm_format_info *format, int plane);
> > > > > >  int drm_format_horz_chroma_subsampling(uint32_t format);
> > > > > >  int drm_format_vert_chroma_subsampling(uint32_t format);
> > > > > >  int drm_format_plane_width(int width, uint32_t format, int plane);
> > > > > > --
> > > > > > 2.16.1
> > > > > >
> > > > >
> > > > > --
> > > > > Daniel Vetter
> > > > > Software Engineer, Intel Corporation
> > > > > http://blog.ffwll.ch
> > > 
> > > -- 
> > > Daniel Vetter
> > > Software Engineer, Intel Corporation
> > > http://blog.ffwll.ch
> > _______________________________________________
> > dri-devel mailing list
> > dri-devel@xxxxxxxxxxxxxxxxxxxxx
> > https://lists.freedesktop.org/mailman/listinfo/dri-devel
> 
> -- 
> Daniel Vetter
> Software Engineer, Intel Corporation
> http://blog.ffwll.ch

-- 
====================
| I would like to |
| fix the world,  |
| but they're not |
| giving me the   |
 \ source code!  /
  ---------------
    Â\_(ã)_/Â