linux-kernel - Re: [PATCH 3/4] drm/mediatek: fix the wrong pixel clock when resolution is 4K

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]

Date:	Wed, 20 Jul 2016 11:41:58 +0200
From:	Philipp Zabel <p.zabel@...gutronix.de>
To:	Bibby Hsieh <bibby.hsieh@...iatek.com>
Cc:	David Airlie <airlied@...ux.ie>,
	Matthias Brugger <matthias.bgg@...il.com>,
	Daniel Vetter <daniel.vetter@...ll.ch>,
	dri-devel@...ts.freedesktop.org,
	linux-mediatek@...ts.infradead.org,
	Yingjoe Chen <yingjoe.chen@...iatek.com>,
	Cawa Cheng <cawa.cheng@...iatek.com>,
	Daniel Kurtz <djkurtz@...omium.org>,
	YT Shen <yt.shen@...iatek.com>,
	Thierry Reding <thierry.reding@...il.com>,
	CK Hu <ck.hu@...iatek.com>, Mao Huang <littlecvr@...omium.org>,
	linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
	Sascha Hauer <kernel@...gutronix.de>,
	Junzhi Zhao <junzhi.zhao@...iatek.com>
Subject: Re: [PATCH 3/4] drm/mediatek: fix the wrong pixel clock when
 resolution is 4K

Am Mittwoch, den 20.07.2016, 12:03 +0800 schrieb Bibby Hsieh:
> From: Junzhi Zhao <junzhi.zhao@...iatek.com>
> 
> Pixel clock should be 297MHz when resolution is 4K.
> 
> Signed-off-by: Junzhi Zhao <junzhi.zhao@...iatek.com>
> Signed-off-by: Bibby Hsieh <bibby.hsieh@...iatek.com>
> ---
>  drivers/gpu/drm/mediatek/mtk_dpi.c |  184 +++++++++++++++++++++++++-----------
>  1 file changed, 131 insertions(+), 53 deletions(-)
> 
> diff --git a/drivers/gpu/drm/mediatek/mtk_dpi.c b/drivers/gpu/drm/mediatek/mtk_dpi.c
> index d05ca79..c0f04d2 100644
> --- a/drivers/gpu/drm/mediatek/mtk_dpi.c
> +++ b/drivers/gpu/drm/mediatek/mtk_dpi.c
> @@ -60,14 +60,35 @@ enum mtk_dpi_out_color_format {
>  	MTK_DPI_COLOR_FORMAT_YCBCR_422_FULL
>  };
>  
> +enum mtk_dpi_clk_id {
> +	MTK_DPI_CLK_DPI_ENGINE,
> +	MTK_DPI_CLK_DPI_PIXEL,
> +	MTK_DPI_CLK_TVD_PLL,
> +	MTK_DPI_CLK_TVDPLL_MUX,
> +	MTK_DPI_CLK_TVDPLL_D2,
> +	MTK_DPI_CLK_TVDPLL_D4,
> +	MTK_DPI_CLK_TVDPLL_D8,
> +	MTK_DPI_CLK_TVDPLL_D16,
> +	MTK_DPI_CLK_COUNT,
> +};

I think this is going in the wrong direction. If the pixel clock output
isn't correct after a clk_set_rate(dpi->pixel_clk, rate), the clock
drivers should be fixed, not worked around in the dpi driver.

The TVDPLL_* mux and dividers are not direct inputs to the DPI module:

   tvdpll ("pll")
     |               ..|\
     v               ..| | mm_sel ----> mm_dpi_engine ("engine")
   tvdpll_594m(1/3)  ..|/
     |
     |`-> tvdpll_d2 -->|\
     |`-> tvdpll_d4 -->| | dpi0_sel --> mm_dpi_pixel ("pixel")
     |`-> tvdpll_d8 -->| |
     `--> tvdpll_d16 ->|/

Currently the code first sets the "pll" to the desired multiple of the
pixel clock manually (*3*4, *3*8) and than calls clk_set_rate on the
"pixel" clock which gets propagated by the clock framework up to
dpi0_sel. Since dpi0_sel doesn't have the CLK_SET_RATE_PARENT flag set,
it should just choose the tvdpll_d* input divider. I'd like not to give
the dpi driver direct access to all the intermediate clocks.

regards
Philipp