[00/11] adding state checker for gamma lut values
mbox series

Message ID 1554816771-13819-1-git-send-email-swati2.sharma@intel.com
Headers show
  • adding state checker for gamma lut values
Related show


Sharma, Swati2 April 9, 2019, 1:32 p.m. UTC
Thanks to Jani N, Matt and Ville for the review comments. Hopefully
I have addressed all the current review comments and ready to receive more :)

In this patch series, added state checker to validate gamma_lut values. This
reads hardware state, and compares the originally requested
state to the state read from hardware.

v1: -Implementation done for legacy platforms (removed all the placeholders) (Jani)
    -Added inverse function of drm_color_lut_extract to convert hardware
     read values back to user values (code written by Jani)
    -Renamed get_config() to color_config() (Jani)
    -Placed all platform specific shifts and masks in i915_reg.h (Jani)
    -Renamed i9xx_get_config to i9xx_color_config and all related
     functions (Jani)
    -Removed debug logs from compare function (Jani)
    -Renamed intel_compare_blob to intel_compare_lut and added platform specific
     bit precision of the readout into the function (Jani)
    -Added check if blobs can be NULL (Jani)
    -Added function in intel_color.c that returns the bit precision (Jani),
     didn't add in device info since its gonna die soon (Ville)

v2: -Moved intel_compare_lut() from intel_display.c to intel_color.c (Jani)
    -Changed color_config() func names to get_color_config() and same
     for gamma func too (Jani)
    -Removed blank line in i915_reg.h and aligned MASK with their
     register name (Jani)
    -Mask definition code indented and defined using REG_GENMASK() and
     used using REG_FIELD_GET() (Jani)
    -Made bit_precision func inline (Jani/Matt)
    -Assigned bit_precision according to GAMMA_MODE (Matt/Ville)
    -Changed IS_ERR(blob) condition to (!blob) (Jani)
    -Rearranged blob check and lut_size conditions (Jani)
    -Used inline function for the comparison (Jani)
    -Separated the get config part from the state checker part to
     another patch (Jani)
    -Retained "internal" i9xx_internal_gamma_config for consistency
    -Removed crtc_state_is_legacy_gamma check and replaced with
     GAMMA_MODE (Matt)
    -Created whole platform specific patch series as submitted by Ville for
     clean up intel_color_check()
    -Rebased on top of Ville's changes

TODO:  -Matt suggested to rename get_gamma_config to get_gamma_lut or
        i9xx_read_out_gamma_lut, Ville also named functions like i9xx_read_lut_8(),
        i9xx_read_lut_10(). Should we follow this?
       -Answering Matt's query regarding intension to read and compare degamma/ctm".
        Degamma will be my first priority after this and later ctm.
        Should we have combined func or separate func for gamma/degamma val?
       -Add separate func to log errors. I am not sure, what needs to be printed here?
        Mismatched RGB value? 
       -Will make changes in next rev for changes made by Ville for cherryview_load_luts()

Swati Sharma (11):
  [v2] drm/i915: Introduce vfunc intel_get_color_config to create hw lut
  [v2] drm/i915: Extract i9xx_get_color_config()
  [v2] drm/i915: Extract cherryview_get_color_config()
  [v2] drm/i915: Extract i965_get_color_config()
  [v2] drm/i915: Extract icl_get_color_config()
  [v2] drm/i915: Extract glk_get_color_config()
  [v2] drm/i915: Extract bdw_get_color_config()
  [v2] drm/i915: Extract ivb_get_color_config()
  [v2] drm/i915: Extract ilk_get_color_config()
  [v2] drm/i915: Enable intel_get_color_config()
  [v2] drm/i915: Add intel_compare_color_lut() to compare hw and sw
    gamma lut values

 drivers/gpu/drm/i915/i915_drv.h      |   1 +
 drivers/gpu/drm/i915/i915_reg.h      |  15 ++
 drivers/gpu/drm/i915/intel_color.c   | 347 ++++++++++++++++++++++++++++++++++-
 drivers/gpu/drm/i915/intel_display.c |  13 ++
 drivers/gpu/drm/i915/intel_drv.h     |   4 +
 5 files changed, 375 insertions(+), 5 deletions(-)