[v2,00/10] drm: Add support for low-color frame buffer formats

Message

Geert Uytterhoeven March 7, 2022, 8:52 p.m. UTC

Hi all,

A long outstanding issue with the DRM subsystem has been the lack of
support for low-color displays, as used typically on older desktop
systems, and on small embedded displays.

This patch series adds support for color-indexed frame buffer formats
with 2, 4, and 16 colors.  It has been tested on ARAnyM using a
work-in-progress Atari DRM driver supporting 2, 4, 16, and 256 colors,
with text console operation, fbtest, and modetest.

Overview:
  - Patch 1 introduces a helper, to be used by later patches in the
    series,
  - Patch 2 introduces a flag to indicate color-indexed formats,
  - Patches 3 and 4 correct calculations of bits per pixel for sub-byte
    pixel formats,
  - Patches 5 and 6 introduce the new C[124] formats,
  - Patch 7 fixes an untested code path,
  - Patch 8 documents the use of "red" for light-on-dark displays,
  - Patches 9 and 10 add more fourcc codes for light-on-dark and
    dark-on-light frame buffer formats, which may be useful for e.g. the
    ssd130x and repaper drivers.

Changes compared to v1[1]:
  - Reshuffle patches,
  - New patch "[PATCH v2 02/10] drm/fourcc: Add
    drm_format_info.is_color_indexed flag",
  - Improve pixel descriptions,
  - Require depth to match bpp in drm_mode_legacy_fb_format(),
  - Set .is_color_indexed flag.
  - Use drm_format_info_bpp() helper instead of deprecated .depth field
    or format-dependent calculations,
  - Use new .is_color_indexed field instead of checking against a list
    of formats,
  - Add Acked-by,
  - Replace FIXME by TODO comment,
  - New patch "[PATCH v2 08/10] [RFC] drm/fourcc: Document that
    single-channel "red" can be any color",
  - Add rationale for adding new formats,
  - Add D[248] for completeness.

Notes:
  - Using modetest with the C4 formats requires [2].
  - Using modetest with the default XR24 format requires [3].
  - As this was used on emulated hardware only, and I do not have Atari
    hardware, I do not have performance figures to compare with fbdev.
    I hope to do proper measuring with an Amiga DRM driver, eventually.

Thanks for your comments!

[1] "[PATCH 0/8] drm: Add support for low-color frame buffer formats"
    https://lore.kernel.org/r/[e-mail]/
[2] "[PATCH libdrm 0/3] Add support for low-color frame buffer formats"
    https://lore.kernel.org/r/[e-mail]
[3] "[PATCH RFC libdrm 0/2] Big-endian fixes"
    https://lore.kernel.org/r/[e-mail]

Geert Uytterhoeven (10):
  drm/fourcc: Add drm_format_info_bpp() helper
  drm/fourcc: Add drm_format_info.is_color_indexed flag
  drm/client: Use actual bpp when allocating frame buffers
  drm/framebuffer: Use actual bpp for DRM_IOCTL_MODE_GETFB
  drm/fourcc: Add DRM_FORMAT_C[124]
  drm/fb-helper: Add support for DRM_FORMAT_C[124]
  [RFC] drm/gem-fb-helper: Use actual bpp for size calculations
  [RFC] drm/fourcc: Document that single-channel "red" can be any color
  [RFC] drm/fourcc: Add DRM_FORMAT_R[124]
  [RFC] drm/fourcc: Add DRM_FORMAT_D[1248]

 drivers/gpu/drm/drm_client.c                 |   4 +-
 drivers/gpu/drm/drm_fb_helper.c              | 101 ++++++++++++++-----
 drivers/gpu/drm/drm_fourcc.c                 |  55 +++++++++-
 drivers/gpu/drm/drm_framebuffer.c            |   2 +-
 drivers/gpu/drm/drm_gem_framebuffer_helper.c |  12 +--
 include/drm/drm_fourcc.h                     |   4 +
 include/uapi/drm/drm_fourcc.h                |  36 +++++--
 7 files changed, 169 insertions(+), 45 deletions(-)