[Intel-gfx] [PATCH] [v2] drm/i915: Clean up the ring scaling calculations

[Jesse Barnes]

On Tue,  1 Oct 2013 15:32:38 -0700
Ben Widawsky <benjamin.widawsky at intel.com> wrote:

> This patch attempts to clean up the ring/IA scaling programming in the
> following ways.
> 1. Fix the comment about the DDR frequency. The math is 266MHz, not
> 133MHz. Formula was right, docs are wrong.
> 
> 2. Mask the DCLK register since I don't know how it is defined on future
> platforms.
> 
> 3. use mult_frac instead of magic math.
> 
> This helps for future platform enabling.
> 
> v2: Actually use the right patch. The v1 was a mix of things, none of
> which was right. Note that due to rounding, we actually get different
> values (slightly higher) for the effective ring frequency.
> 
> CC: Jesse Barnes <jbarnes at virtuousgeek.org>
> CC: Chris Wilson <chris at chris-wilson.co.uk>
> Signed-off-by: Ben Widawsky <ben at bwidawsk.net>
> ---
>  drivers/gpu/drm/i915/intel_pm.c | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
> index 698257c..cb0876b 100644
> --- a/drivers/gpu/drm/i915/intel_pm.c
> +++ b/drivers/gpu/drm/i915/intel_pm.c
> @@ -3663,9 +3663,9 @@ void gen6_update_ring_freq(struct drm_device *dev)
>  	/* Convert from kHz to MHz */
>  	max_ia_freq /= 1000;
>  
> -	min_ring_freq = I915_READ(MCHBAR_MIRROR_BASE_SNB + DCLK);
> -	/* convert DDR frequency from units of 133.3MHz to bandwidth */
> -	min_ring_freq = (2 * 4 * min_ring_freq + 2) / 3;
> +	min_ring_freq = I915_READ(MCHBAR_MIRROR_BASE_SNB + DCLK) & 0xf;
> +	/* convert DDR frequency from units of 266.6MHz to bandwidth */
> +	min_ring_freq = mult_frac(min_ring_freq, 8, 3);
>  
>  	/*
>  	 * For each potential GPU frequency, load a ring frequency we'd like
> @@ -3678,7 +3678,7 @@ void gen6_update_ring_freq(struct drm_device *dev)
>  		unsigned int ia_freq = 0, ring_freq = 0;
>  
>  		if (IS_HASWELL(dev)) {
> -			ring_freq = (gpu_freq * 5 + 3) / 4;
> +			ring_freq = mult_frac(gpu_freq, 4, 3);
>  			ring_freq = max(min_ring_freq, ring_freq);
>  			/* leave ia_freq as the default, chosen by cpufreq */
>  		} else {

Seems like the second mult_frac should be 5,4 rather than 4,3, to
match the original calculation?