linux-kernel - Re: [Freedreno] [PATCH 1/5] drm/msm/dsi: Remove useless math in DSC calculation

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20221004222317.6or3w6vwgyd3yy6z@SoMainline.org>
Date:   Wed, 5 Oct 2022 00:23:17 +0200
From:   Marijn Suijten <marijn.suijten@...ainline.org>
To:     Abhinav Kumar <quic_abhinavk@...cinc.com>
Cc:     phone-devel@...r.kernel.org, Rob Clark <robdclark@...il.com>,
        Dmitry Baryshkov <dmitry.baryshkov@...aro.org>,
        Vinod Koul <vkoul@...nel.org>, Marek Vasut <marex@...x.de>,
        freedreno@...ts.freedesktop.org,
        Douglas Anderson <dianders@...omium.org>,
        Thomas Zimmermann <tzimmermann@...e.de>,
        Jami Kettunen <jami.kettunen@...ainline.org>,
        Vladimir Lypak <vladimir.lypak@...il.com>,
        linux-arm-msm@...r.kernel.org,
        Konrad Dybcio <konrad.dybcio@...ainline.org>,
        dri-devel@...ts.freedesktop.org,
        Javier Martinez Canillas <javierm@...hat.com>,
        David Airlie <airlied@...ux.ie>,
        Martin Botka <martin.botka@...ainline.org>,
        ~postmarketos/upstreaming@...ts.sr.ht,
        Daniel Vetter <daniel@...ll.ch>,
        AngeloGioacchino Del Regno 
        <angelogioacchino.delregno@...ainline.org>,
        Alex Deucher <alexander.deucher@....com>,
        Sean Paul <sean@...rly.run>, linux-kernel@...r.kernel.org
Subject: Re: [Freedreno] [PATCH 1/5] drm/msm/dsi: Remove useless math in DSC
 calculation

On 2022-10-04 07:33:49, Abhinav Kumar wrote:
> 
> 
> On 10/1/2022 12:08 PM, Marijn Suijten wrote:
> > Multiplying a value by 2 and adding 1 to it always results in a value
> > that is uneven, and that 1 gets truncated immediately when performing
> > integer division by 2 again.  There is no "rounding" possible here.
> > 
> > Fixes: b9080324d6ca ("drm/msm/dsi: add support for dsc data")
> > Signed-off-by: Marijn Suijten <marijn.suijten@...ainline.org>
> > ---
> >   drivers/gpu/drm/msm/dsi/dsi_host.c | 7 +------
> >   1 file changed, 1 insertion(+), 6 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/msm/dsi/dsi_host.c b/drivers/gpu/drm/msm/dsi/dsi_host.c
> > index 8e4bc586c262..e05bae647431 100644
> > --- a/drivers/gpu/drm/msm/dsi/dsi_host.c
> > +++ b/drivers/gpu/drm/msm/dsi/dsi_host.c
> > @@ -1864,12 +1864,7 @@ static int dsi_populate_dsc_params(struct drm_dsc_config *dsc)
> >   	data = 2048 * (dsc->rc_model_size - dsc->initial_offset + num_extra_mux_bits);
> >   	dsc->slice_bpg_offset = DIV_ROUND_UP(data, groups_total);
> >   
> > -	/* bpp * 16 + 0.5 */
> > -	data = dsc->bits_per_pixel * 16;
> > -	data *= 2;
> > -	data++;
> > -	data /= 2;
> > -	target_bpp_x16 = data;
> > +	target_bpp_x16 = dsc->bits_per_pixel * 16;
> >   
> Since this patch is titled, "remove useless math", even the 
> target_bpp_x16 math looks redundant to me,
> 
> first we do
> 
> target_bpp_x16 = dsc->bits_per_pixel * 16;
> 
> then in the next line we do
> 
> data = (dsc->initial_xmit_delay * target_bpp_x16) / 16;
> 
> the *16 and /16 will cancel out here.
> 
> Instead we can just do
> 
> data = (dsc->initial_xmit_delay * dsc->drm->bits_per_pixel) ?

Thanks, good catch!  I might have been so focused on explaining the
effect of this patch and uselessness of the proposed `+ 0.5` rounding
here that I missed this intermediate variable now becoming redundant as
well.

Corrected for v2!

- Marijn

> >   	data = (dsc->initial_xmit_delay * target_bpp_x16) / 16;
> >   	final_value =  dsc->rc_model_size - data + num_extra_mux_bits;