[Intel-gfx] [PATCH] drm/intel: Set 8-bit gamma mode for the palette

[Adam Jackson]

On Mon, 2010-04-26 at 22:19 +0100, Peter Clifton wrote:

> @@ -3439,11 +3442,16 @@ void intel_crtc_load_lut(struct drm_crtc *crtc)
>  
>  	/* use legacy palette for Ironlake */
>  	if (IS_IRONLAKE(dev))
> -		palreg = (intel_crtc->pipe == 0) ? LGC_PALETTE_A :
> -						   LGC_PALETTE_B;
> +		pal_reg = (intel_crtc->pipe == 0) ? LGC_PALETTE_A :
> +						    LGC_PALETTE_B;
> +
> +	/* Switch to 8-bit gamma mode */
> +	pipeconf &= ~PIPEACONF_GAMMA;
> +	I915_WRITE(pipeconf_reg, pipeconf);
> +	I915_READ(pipeconf_reg);

This is wrong for gen5, where bits 25 and 24 control gamma:

Value Name     Description                    Project
00b   8 bit    8-bit Legacy Palette Mode      All
01b   10 bt    10-bit Precision Palette Mode  All
10b   12 bit   12-bit Interpolated Gamma Mode All
11b   Reserved Reserved                       All

Where 10-bit is a real 1024-stop gamma, and 12-bit is 512 lerp'd stops
(like gen4's 129 lerp'd stops).  I don't think it'd matter in the real
world since I doubt any OEM is insane enough to set up 12-bit in the
BIOS, but.

On gen4 and earlier though, bit 25 is border color enable.  Pretty sure
we only use that in VGA load detection.  Although, we use it in VGA load
detection even on gen5, which means load detection on gen5 almost
certainly does not work.

Patch looks good otherwise.

- ajax
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part
URL: <http://lists.freedesktop.org/archives/intel-gfx/attachments/20100427/abe8901b/attachment.sig>