[linux-sunxi] [PATCH] drm: sun4i: exclusively set HDMI-related clocks for dw-hdmi

Wed Aug 15 13:53:45 UTC 2018

Hi!

Dne sreda, 15. avgust 2018 ob 15:43:19 CEST je Icenowy Zheng napisal(a):
> 于 2018年8月15日 GMT+08:00 下午9:39:44, "Jernej Škrabec" 
<jernej.skrabec at gmail.com> 写到:
> >Hi!
> >
> >Dne sreda, 15. avgust 2018 ob 14:07:45 CEST je Icenowy Zheng
> >
> >napisal(a):
> >> The glue in sun4i-drm of dw-hdmi currently doesn't set the clocks of
> >> dw-hdmi exclusively, which will lead the display fails to initialize
> >
> >in
> >
> >> some situations.
> >> 
> >> Add the exclusivity to sun8i-dw-hdmi and sun8i-hdmi-phy.
> >> 
> >> Cc: stable at vger.kernel.org # v4.17+
> >> Signed-off-by: Icenowy Zheng <icenowy at aosc.io>
> >
> >Given that you want this patch to be merged in stable, I have to ask
> >what kind
> >of tests did you run to prove it works as expected?
> >
> >In the past, there were some issues with TCON exclusive rate setting.
> >It turns
> >out that as many scenarios as possible have to be tested before you can
> >say
> >it's working.
> >
> >I imagine that it has to be tested on at least A83T, H3 (or H5) and R40
> >(or
> >A64). They all have their own specifics which needs to be covered.
> >
> >Tests which I remember from top of my head:
> >- run with monitor connected at boot
> 
> BTW I think the bootloader will affect the state of displsy even
> without the patch.

I know it can be an issue. For this test I imagine you can just disable 
display driver in U-Boot so the PLL doesn't get set.

> 
> When could your U-Boot tweaks patchset to be done? I think they will
> be needed for a really reliable test.

I plan to do them after A64 driver is merged, so I can sync up dts files and 
switch driver to be enabled trough them. AFAIK, A64 patches are more or less 
ready and they will be merged after 4.19-rc1, right?

> 
> >- run with monitor disconnected at boot and connect it later
> >- playing with resolution switching (at least a few non-standard, e.g
> >something other than 1080p or 720p)
> 
> How to do it without starting a X server?

libdrm has in it's source superb testing tool called modetest. It is useful 
for resolution testing, plane testing (scaling, positioning, color formats, 
blending, ...) and setting atomic properties (like zpos). While it doesn't 
recognize sun4i-drm driver automatically, you can specify it through command 
line parameter.

Best regards,
Jernej

> 
> >Best regards,
> >Jernej
> >
> >> ---
> >> 
> >>  drivers/gpu/drm/sun4i/sun8i_dw_hdmi.c  | 11 ++++++++++-
> >>  drivers/gpu/drm/sun4i/sun8i_hdmi_phy.c |  7 +++++--
> >>  2 files changed, 15 insertions(+), 3 deletions(-)
> >> 
> >> diff --git a/drivers/gpu/drm/sun4i/sun8i_dw_hdmi.c
> >> b/drivers/gpu/drm/sun4i/sun8i_dw_hdmi.c index
> >
> >31875b636434..a10220518548
> >
> >> 100644
> >> --- a/drivers/gpu/drm/sun4i/sun8i_dw_hdmi.c
> >> +++ b/drivers/gpu/drm/sun4i/sun8i_dw_hdmi.c
> >> @@ -137,10 +137,16 @@ static int sun8i_dw_hdmi_bind(struct device
> >
> >*dev,
> >
> >> struct device *master, goto err_assert_ctrl_reset;
> >> 
> >>  	}
> >> 
> >> +	ret = clk_rate_exclusive_get(hdmi->clk_tmds);
> >> +	if (ret) {
> >> +		dev_err(dev, "Could not get exclusivity over the tmds clock\n");
> >> +		goto err_disable_clk_tmds;
> >> +	}
> >> +
> >> 
> >>  	phy_node = of_parse_phandle(dev->of_node, "phys", 0);
> >>  	if (!phy_node) {
> >>  	
> >>  		dev_err(dev, "Can't found PHY phandle\n");
> >> 
> >> -		goto err_disable_clk_tmds;
> >> +		goto err_put_clk_tmds_exclusivity;
> >> 
> >>  	}
> >>  	
> >>  	ret = sun8i_hdmi_phy_probe(hdmi, phy_node);
> >> 
> >> @@ -179,6 +185,8 @@ static int sun8i_dw_hdmi_bind(struct device *dev,
> >
> >struct
> >
> >> device *master, cleanup_encoder:
> >>  	drm_encoder_cleanup(encoder);
> >>  	sun8i_hdmi_phy_remove(hdmi);
> >> 
> >> +err_put_clk_tmds_exclusivity:
> >> +	clk_rate_exclusive_put(hdmi->clk_tmds);
> >> 
> >>  err_disable_clk_tmds:
> >>  	clk_disable_unprepare(hdmi->clk_tmds);
> >>  
> >>  err_assert_ctrl_reset:
> >> @@ -194,6 +202,7 @@ static void sun8i_dw_hdmi_unbind(struct device
> >
> >*dev,
> >
> >> struct device *master,
> >> 
> >>  	dw_hdmi_unbind(hdmi->hdmi);
> >>  	sun8i_hdmi_phy_remove(hdmi);
> >> 
> >> +	clk_rate_exclusive_put(hdmi->clk_tmds);
> >> 
> >>  	clk_disable_unprepare(hdmi->clk_tmds);
> >>  	reset_control_assert(hdmi->rst_ctrl);
> >>  
> >>  }
> >> 
> >> diff --git a/drivers/gpu/drm/sun4i/sun8i_hdmi_phy.c
> >> b/drivers/gpu/drm/sun4i/sun8i_hdmi_phy.c index
> >
> >82502b351aec..1e0b1d9bc0fb
> >
> >> 100644
> >> --- a/drivers/gpu/drm/sun4i/sun8i_hdmi_phy.c
> >> +++ b/drivers/gpu/drm/sun4i/sun8i_hdmi_phy.c
> >> @@ -511,13 +511,14 @@ int sun8i_hdmi_phy_probe(struct sun8i_dw_hdmi
> >
> >*hdmi,
> >
> >> struct device_node *node) }
> >> 
> >>  		clk_prepare_enable(phy->clk_phy);
> >> 
> >> +		clk_rate_exclusive_get(phy->clk_phy);
> >> 
> >>  	}
> >>  	
> >>  	phy->rst_phy = of_reset_control_get_shared(node, "phy");
> >>  	if (IS_ERR(phy->rst_phy)) {
> >>  	
> >>  		dev_err(dev, "Could not get phy reset control\n");
> >>  		ret = PTR_ERR(phy->rst_phy);
> >> 
> >> -		goto err_disable_clk_phy;
> >> +		goto err_put_clk_phy_exclusivity;
> >> 
> >>  	}
> >>  	
> >>  	ret = reset_control_deassert(phy->rst_phy);
> >> 
> >> @@ -548,7 +549,8 @@ int sun8i_hdmi_phy_probe(struct sun8i_dw_hdmi
> >
> >*hdmi,
> >
> >> struct device_node *node) reset_control_assert(phy->rst_phy);
> >> 
> >>  err_put_rst_phy:
> >>  	reset_control_put(phy->rst_phy);
> >> 
> >> -err_disable_clk_phy:
> >> +err_put_clk_phy_exclusivity:
> >> +	clk_rate_exclusive_put(phy->clk_phy);
> >> 
> >>  	clk_disable_unprepare(phy->clk_phy);
> >>  
> >>  err_put_clk_pll1:
> >>  	clk_put(phy->clk_pll1);
> >> 
> >> @@ -568,6 +570,7 @@ void sun8i_hdmi_phy_remove(struct sun8i_dw_hdmi
> >
> >*hdmi)
> >
> >>  	clk_disable_unprepare(phy->clk_mod);
> >>  	clk_disable_unprepare(phy->clk_bus);
> >> 
> >> +	clk_rate_exclusive_put(phy->clk_phy);
> >> 
> >>  	clk_disable_unprepare(phy->clk_phy);
> >>  	
> >>  	reset_control_assert(phy->rst_phy);