[PATCH 1/3] drm: exynos: hdmi: add exynos5 support to mixer driver

RAHUL SHARMA rahul.sharma at samsung.com
Wed Sep 12 22:20:26 PDT 2012



------- Original Message -------
Sender : In-Ki Dae<inki.dae at samsung.com> S5/Senior Engineer/System S/W Lab./Samsung Electronics
Date : Sep 13, 2012 10:07 (GMT+05:30)
Title : Re: [PATCH 1/3] drm: exynos: hdmi: add exynos5 support to mixer driver



2012? 9? 13? ???? Joonyoung Shim<jy0922.shim at samsung.com>?? ??:
> On 09/13/2012 11:53 AM, Inki Dae wrote:
>
> -----Original Message-----
> From: Joonyoung Shim [mailto:jy0922.shim at samsung.com]
> Sent: Thursday, September 13, 2012 10:44 AM
> To: Rahul Sharma
> Cc: dri-devel at lists.freedesktop.org; sw0312.kim at samsung.com;
> inki.dae at samsung.com; kyungmin.park at samsung.com; prashanth.g at samsung.com;
> joshi at samsung.com; s.shirish at samsung.com; fahad.k at samsung.com;
> l.krishna at samsung.com; r.sh.open at gmail.com
> Subject: Re: [PATCH 1/3] drm: exynos: hdmi: add exynos5 support to mixer
> driver
>
> Hi, Rahul.
>
> On 09/12/2012 09:08 PM, Rahul Sharma wrote:
>
> Added support for exynos5 to drm mixer driver. Exynos5 works
> with dt enabled while in exynos4 mixer device information can
> be passed either way (dt or plf data). This situation is taken
> cared.
>
> Signed-off-by: Rahul Sharma <rahul.sharma at samsung.com>
> Signed-off-by: Shirish S <s.shirish at samsung.com>
> Signed-off-by: Fahad Kunnathadi <fahad.k at samsung.com>
> ---
>    drivers/gpu/drm/exynos/exynos_mixer.c |  153
>
> ++++++++++++++++++++++++++++++---
>
>    drivers/gpu/drm/exynos/regs-mixer.h   |    2 +
>    2 files changed, 142 insertions(+), 13 deletions(-)
>
> diff --git a/drivers/gpu/drm/exynos/exynos_mixer.c
>
> b/drivers/gpu/drm/exynos/exynos_mixer.c
>
> index 8a43ee1..7d04a40 100644
> --- a/drivers/gpu/drm/exynos/exynos_mixer.c
> +++ b/drivers/gpu/drm/exynos/exynos_mixer.c
> @@ -71,6 +71,7 @@ struct mixer_resources {
>         struct clk              *sclk_mixer;
>         struct clk              *sclk_hdmi;
>         struct clk              *sclk_dac;
> +       bool                    is_soc_exynos5;
>    };
>
>    struct mixer_context {
> @@ -251,7 +252,8 @@ static void mixer_vsync_set_update(struct
>
> mixer_context *ctx, bool enable)
>
>         mixer_reg_writemask(res, MXR_STATUS, enable ?
>                         MXR_STATUS_SYNC_ENABLE : 0, MXR_STATUS_SYNC_ENABLE);
>
> -       vp_reg_write(res, VP_SHADOW_UPDATE, enable ?
> +       if (!res->is_soc_exynos5)
> +               vp_reg_write(res, VP_SHADOW_UPDATE, enable ?
>                         VP_SHADOW_UPDATE_ENABLE : 0);
>    }
>
> @@ -615,15 +617,21 @@ static void mixer_win_reset(struct mixer_context
>
> *ctx)
>
>         val = MXR_GRP_CFG_ALPHA_VAL(0);
>         mixer_reg_write(res, MXR_VIDEO_CFG, val);
>
> -       /* configuration of Video Processor Registers */
> -       vp_win_reset(ctx);
> -       vp_default_filter(res);
> +       if (!res->is_soc_exynos5) {
> +               /* configuration of Video Processor Registers */
> +               vp_win_reset(ctx);
> +               vp_default_filter(res);
> +       }
>
>         /* disable all layers */
>         mixer_reg_writemask(res, MXR_CFG, 0, MXR_CFG_GRP0_ENABLE);
>         mixer_reg_writemask(res, MXR_CFG, 0, MXR_CFG_GRP1_ENABLE);
>         mixer_reg_writemask(res, MXR_CFG, 0, MXR_CFG_VP_ENABLE);
>
> +       /* enable vsync interrupt after mixer reset*/
> +       mixer_reg_writemask(res, MXR_INT_EN, MXR_INT_EN_VSYNC,
> +                       MXR_INT_EN_VSYNC);
> +
>         mixer_vsync_set_update(ctx, true);
>         spin_unlock_irqrestore(&res->reg_slock, flags);
>    }
> @@ -645,7 +653,8 @@ static void mixer_poweron(struct mixer_context *ctx)
>         pm_runtime_get_sync(ctx->dev);
>
>         clk_enable(res->mixer);
> -       clk_enable(res->vp);
> +       if (!res->is_soc_exynos5)
> +               clk_enable(res->vp);
>         clk_enable(res->sclk_mixer);
>
>         mixer_reg_write(res, MXR_INT_EN, ctx->int_en);
> @@ -666,7 +675,8 @@ static void mixer_poweroff(struct mixer_context
>
> *ctx)
>
>         ctx->int_en = mixer_reg_read(res, MXR_INT_EN);
>
>         clk_disable(res->mixer
>
> Let's think to disassociate hdmi and mixer. I have plan to unify to one
> many things of exynos hdmi. Above problem occurs because exynos5 doesn't
> have video processor ip. Even if we use a field such is_soc_exynos5, the
> is_soc_exynos5 is unsuitable naming if other exynos SoC also doesn't
> have video processor ip.
> 

one more thing, exynos5 uses GScaler instead of Video processor. the GScaler can be also used as post processor but exynos5 spec has no any descriptions to this. so we should check that first and next let's update things related to hdmi.

Thanks Joonyoung, Inki,

I agree with you. mixer_resources_init_exynnos4,5 has code repetition here but Since Exy4
is non-DT and and Exy5 is DT enbled, method for getting resource are different. I need to
replace the platform_get_resource_byname to platform_get_resource by index.
Then I will be able to group the two functions into one. It will also help us when we add DT support
for Exynos4 hdmi driver. How does it look to you?

is_soc_exynos5 variable is used for 2 scenarios
1) Resource initialization. I will eliminate its usage by unifying the resource init function.
2) Video Processor. I can add a property in mixer node in device tree stating that VP support is
required or not. Based on that I can Disable VP support. By default it will be enabled.
With this method Exy4, Exy5, or any other SoC in future can define the property accordingly.

For Exynos-4 I will revert name to "s5p-mixer".

I have access to exy5 GScaler specs. It is providing very basic VPP fucntionality, can only see
CSC which again looks optional to me for hdmi. We may need to add GSC for scaling but can be
optioanlly accessed as v4l driver from application as doing in the current project. If you foresee
any impact on gscalar addition due to my changes please tell me.

regards.
Rahul Sharma
_______________________________________________
> dri-devel mailing list
> dri-devel at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/dri-devel
> 


More information about the dri-devel mailing list