[PATCH v12 16/18] drm: bridge: analogix/dp: expand the wait time for looking AUX CH reply flag

[Jingoo Han]

On Wednesday, December 23, 2015 9:51 PM, Yakir Yang wrote:
> 
> On Rockchip platform, sometimes driver would failed at reading EDID
> message, and it's caused by the AUX reply flag wouldn't received under
> the 100*10us wait time.

The problem is specific for Rockchip platform.
Also, 1 ms is long time.
The best way is that your hardware engineers find the root cause.
I cannot understand why your engineers cannot find the root cause. :-(

> 
> But after expand the wait time a little, the AUX reply flag would be
> set, so maybe the wait time is a little critical. Besides the analogix
> dp book haven't reminded the standard wait for looking AUX reply flag,
> so I thought it's okay to expand the wait time.
> 
> And the external wait time won't hurt Exynos DP too much, cause they
> wouldn't meet this problem, then driver would received the reply command
> very soon, so no more additional wait time would bring to Exynos platform.

Then, when loop time happens on Exynos platform, it will take long time
to return error.

> 
> Signed-off-by: Yakir Yang <ykk at rock-chips.com>
> ---
> Changes in v12:
> - Using another way to expand the AUX reply wait time (Jingoo)
> 
> Changes in v11: None
> Changes in v10: None
> Changes in v9: None
> Changes in v8: None
> Changes in v7: None
> Changes in v6: None
> Changes in v5: None
> Changes in v4: None
> Changes in v3: None
> Changes in v2: None
> 
>  drivers/gpu/drm/bridge/analogix/analogix_dp_reg.c | 10 ++++------
>  1 file changed, 4 insertions(+), 6 deletions(-)
> 
> diff --git a/drivers/gpu/drm/bridge/analogix/analogix_dp_reg.c
> b/drivers/gpu/drm/bridge/analogix/analogix_dp_reg.c
> index cba3ffd..8687eea 100644
> --- a/drivers/gpu/drm/bridge/analogix/analogix_dp_reg.c
> +++ b/drivers/gpu/drm/bridge/analogix/analogix_dp_reg.c
> @@ -471,7 +471,7 @@ int analogix_dp_start_aux_transaction(struct analogix_dp_device *dp)
>  {
>  	int reg;
>  	int retval = 0;
> -	int timeout_loop = 0;
> +	unsigned long timeout;
> 
>  	/* Enable AUX CH operation */
>  	reg = readl(dp->reg_base + ANALOGIX_DP_AUX_CH_CTL_2);
> @@ -479,14 +479,12 @@ int analogix_dp_start_aux_transaction(struct analogix_dp_device *dp)
>  	writel(reg, dp->reg_base + ANALOGIX_DP_AUX_CH_CTL_2);
> 
>  	/* Is AUX CH command reply received? */
> -	reg = readl(dp->reg_base + ANALOGIX_DP_INT_STA);
> -	while (!(reg & RPLY_RECEIV)) {
> -		timeout_loop++;
> -		if (DP_TIMEOUT_LOOP_COUNT < timeout_loop) {
> +	timeout = jiffies + msecs_to_jiffies(5);
> +	while ((readl(dp->reg_base + ANALOGIX_DP_INT_STA) & RPLY_RECEIV) == 0) {
> +		if (time_after(jiffies, timeout)) {
>  			dev_err(dp->dev, "AUX CH command reply failed!\n");
>  			return -ETIMEDOUT;
>  		}
> -		reg = readl(dp->reg_base + ANALOGIX_DP_INT_STA);

Sorry, I don't like your patch.

The problem happens because of Rockchip platform.
So, you need to add workaround for only Rockchip platform.

Just add new DT property and new variable for the value for wait time.
When, the probe is called, new wait time value is read from Rockchip DT file.
Then, the new wait time value can be written to the new variable.

    new DT property: wait-time-aux
    new variable: wait_time_aux


If ( ) // New DT 
    wait_time_aux = New DT;
else 
    wait_time_aux = 10;


>  		usleep_range(10, 11);

If there is NO  new wait time value from DT file, the default value '10' is
used for sleep.

But, if there is new wait time value from DT file, new wait time value
can be used for sleep.

                 usleep_range(dp->wait_time_aux, dp->wait_time_aux + 1);

What I want to say is that there should be NO effect on Exynos platform,
because of the hardware bug of Rockchip platform.

Best regards,
Jingoo Han

>  	}
> 
> --
> 1.9.1