[PATCH v6 03/15] drm/bridge: tc358767: Simplify polling in tc_link_training()

[Tomi Valkeinen]

On 04/12/2019 20:27, Tomi Valkeinen wrote:
> Hi Andrey,
> 
> On 19/06/2019 08:27, Andrey Smirnov wrote:
> 
>> @@ -748,22 +748,19 @@ static int tc_set_video_mode(struct tc_data *tc,
>> Â  static int tc_wait_link_training(struct tc_data *tc)
>> Â  {
>> -Â Â Â  u32 timeout = 1000;
>> Â Â Â Â Â  u32 value;
>> Â Â Â Â Â  int ret;
>> -Â Â Â  do {
>> -Â Â Â Â Â Â Â  udelay(1);
>> -Â Â Â Â Â Â Â  tc_read(DP0_LTSTAT, &value);
>> -Â Â Â  } while ((!(value & LT_LOOPDONE)) && (--timeout));
>> -
>> -Â Â Â  if (timeout == 0) {
>> +Â Â Â  ret = tc_poll_timeout(tc, DP0_LTSTAT, LT_LOOPDONE,
>> +Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â  LT_LOOPDONE, 1, 1000);
> 
> This seems to break DP at least with some monitors for me. I think it's just a timeout problem, as 
> increasing the values helps.
> 
> Using ktime, I can see that during link training, the first call takes ~2ms, the second ~7ms. I 
> think this worked before, as udelay(1) takes much longer than 1 us.

Also the timeout in tc_aux_link_setup takes ~500us for me, and max is 1000us. So it works, but I 
think it's a bit tight.

  Tomi

-- 
Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki.
Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki