>-----Original Message-----
>From: Roper, Matthew D
>Sent: Friday, December 21, 2018 6:54 AM
>To: Shankar, Uma <[email protected]>
>Cc: [email protected]; Lankhorst, Maarten
><[email protected]>; Syrjala, Ville <[email protected]>; 
>Sharma,
>Shashank <[email protected]>
>Subject: Re: [v4 1/4] drm/i915: Remove gamma_mode state variable
>
>On Fri, Dec 21, 2018 at 01:29:38AM +0530, Uma Shankar wrote:
>> Removed crtc state variable for gamma mode as it's redundant since
>> currently we have fixed modes on respective hardware platforms. This
>> was making this state variable irrelevant.
>>
>> Credits-to: Matt Roper <[email protected]>
>>
>> Signed-off-by: Uma Shankar <[email protected]>
>> ---
>>  drivers/gpu/drm/i915/intel_color.c   | 5 +----
>>  drivers/gpu/drm/i915/intel_display.c | 3 ---
>>  drivers/gpu/drm/i915/intel_drv.h     | 3 ---
>>  3 files changed, 1 insertion(+), 10 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/i915/intel_color.c
>> b/drivers/gpu/drm/i915/intel_color.c
>> index 37fd9dd..f32e4a7 100644
>> --- a/drivers/gpu/drm/i915/intel_color.c
>> +++ b/drivers/gpu/drm/i915/intel_color.c
>> @@ -370,12 +370,11 @@ static void haswell_load_luts(struct intel_crtc_state
>*crtc_state)
>>       * GAMMA_MODE is configured for split gamma and IPS_CTL has IPS
>enabled.
>>       */
>>      if (IS_HASWELL(dev_priv) && crtc_state->ips_enabled &&
>> -        (crtc_state->gamma_mode == GAMMA_MODE_MODE_SPLIT)) {
>> +            (I915_READ(GAMMA_MODE(crtc->pipe)) &
>GAMMA_MODE_MODE_SPLIT)) {
>
>We can probably avoid this read on each LUT load if we just handle it once 
>during
>hardware readout/sanitization.  Since our driver never actually sets MODE_SPLIT
>on this platform, we only have to worry about the BIOS leaving us in split 
>gamma
>mode; that means we could do
>
>        if (ips)
>                disable ips
>
>        intel_color_set_csc()
>        intel_color_load_luts()
>
>        if (ips)
>                re-enable ips
>
>once during intel_sanitize_crtc(), to immediately program the hardware to a
>known-good state at startup (linear CTM and disabled LUT's).  That would also 
>be
>a little bit more consistent overall; since we don't actually readout the 
>BIOS-set
>LUT's or CTM today, we wind up having them remain active for a while until they
>go away with the first fastset or modeset.
>
>Otherwise, this patch looks good, so
>
>Reviewed-by: Matt Roper <[email protected]>
>
>and you can deal with sanitizing the color management stuff at bootup as a
>separate patch later.

Thanks Matt for these useful suggestions. Will update the patch accordingly.

Regards,
Uma Shankar

>
>Matt
>
>>              hsw_disable_ips(crtc_state);
>>              reenable_ips = true;
>>      }
>>
>> -    crtc_state->gamma_mode = GAMMA_MODE_MODE_8BIT;
>>      I915_WRITE(GAMMA_MODE(crtc->pipe), GAMMA_MODE_MODE_8BIT);
>>
>>      i9xx_load_luts(crtc_state);
>> @@ -476,7 +475,6 @@ static void broadwell_load_luts(struct intel_crtc_state
>*crtc_state)
>>      bdw_load_gamma_lut(crtc_state,
>>                         INTEL_INFO(dev_priv)->color.degamma_lut_size);
>>
>> -    crtc_state->gamma_mode = GAMMA_MODE_MODE_SPLIT;
>>      I915_WRITE(GAMMA_MODE(pipe), GAMMA_MODE_MODE_SPLIT);
>>      POSTING_READ(GAMMA_MODE(pipe));
>>
>> @@ -532,7 +530,6 @@ static void glk_load_luts(struct intel_crtc_state
>> *crtc_state)
>>
>>      bdw_load_gamma_lut(crtc_state, 0);
>>
>> -    crtc_state->gamma_mode = GAMMA_MODE_MODE_10BIT;
>>      I915_WRITE(GAMMA_MODE(pipe), GAMMA_MODE_MODE_10BIT);
>>      POSTING_READ(GAMMA_MODE(pipe));
>>  }
>> diff --git a/drivers/gpu/drm/i915/intel_display.c
>> b/drivers/gpu/drm/i915/intel_display.c
>> index 3b70948..704d9d3 100644
>> --- a/drivers/gpu/drm/i915/intel_display.c
>> +++ b/drivers/gpu/drm/i915/intel_display.c
>> @@ -9679,9 +9679,6 @@ static bool haswell_get_pipe_config(struct intel_crtc
>*crtc,
>>      intel_get_pipe_src_size(crtc, pipe_config);
>>      intel_get_crtc_ycbcr_config(crtc, pipe_config);
>>
>> -    pipe_config->gamma_mode =
>> -            I915_READ(GAMMA_MODE(crtc->pipe)) &
>GAMMA_MODE_MODE_MASK;
>> -
>>      power_domain = POWER_DOMAIN_PIPE_PANEL_FITTER(crtc->pipe);
>>      if (intel_display_power_get_if_enabled(dev_priv, power_domain)) {
>>              power_domain_mask |= BIT_ULL(power_domain); diff --git
>> a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
>> index 1028af8..7427a36 100644
>> --- a/drivers/gpu/drm/i915/intel_drv.h
>> +++ b/drivers/gpu/drm/i915/intel_drv.h
>> @@ -921,9 +921,6 @@ struct intel_crtc_state {
>>
>>      struct intel_crtc_wm_state wm;
>>
>> -    /* Gamma mode programmed on the pipe */
>> -    uint32_t gamma_mode;
>> -
>>      /* bitmask of visible planes (enum plane_id) */
>>      u8 active_planes;
>>      u8 nv12_planes;
>> --
>> 1.9.1
>>
>
>--
>Matt Roper
>Graphics Software Engineer
>IoTG Platform Enabling & Development
>Intel Corporation
>(916) 356-2795
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Reply via email to