[PATCH 04/15] drm/amd/display: Replace msleep with udelay while read edid return defer.

[Christian König]

Am 17.09.20 um 00:18 schrieb Alex Deucher:
> On Wed, Sep 16, 2020 at 6:16 PM Zhuo, Qingqing <Qingqing.Zhuo at amd.com> wrote:
>> [AMD Official Use Only - Internal Distribution Only]
>>
>> On Wed, Sep 16, 2020 at 3:42 PM Qingqing Zhuo <qingqing.zhuo at amd.com> wrote:
>>> From: jinlong zhang <jinlong.zhang at amd.com>
>>>
>>> [why]
>>> while read edid return defer, then it enter to msleep, but it actually
>>> took more time during msleep, this will cause remaining edid read
>>> fail.
>>>
>>> [how]
>>> Replacing msleep with udelay, it will not take any extra time, edid return pass finally.
>> How long of a delay are we talking about here?  Some platforms don't support long udelays and someone will send a patch to change this to msleep.
>>
>> Alex
>>
>> ---------------------
>>
>> Hi Alex,
>>
>> It's between 0-5ms for generic cases, though there exist some dongle workaround cases where we will do 70ms. Would this be a concern?
> I think ARM has a limit of 2ms for udelay.

Yeah, there is even a define somewhere for this.

If you need a delay which is longer than this but still more precise 
than msleep() then there is the high precision timer sleep as alternative.

I've forgotten the function name to use here, but there was a LWN 
article about this a few years ago. You just need to google a bit.

Regards,
Christian.

>
> Alex
>
>> Thank you,
>> Lillian
>>
>>
>>> Signed-off-by: jinlong zhang <jinlong.zhang at amd.com>
>>> Reviewed-by: Wenjing Liu <Wenjing.Liu at amd.com>
>>> Acked-by: Qingqing Zhuo <qingqing.zhuo at amd.com>
>>> ---
>>>   drivers/gpu/drm/amd/display/dc/dce/dce_aux.c | 2 +-
>>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_aux.c
>>> b/drivers/gpu/drm/amd/display/dc/dce/dce_aux.c
>>> index 743042d5905a..cdcad82765e0 100644
>>> --- a/drivers/gpu/drm/amd/display/dc/dce/dce_aux.c
>>> +++ b/drivers/gpu/drm/amd/display/dc/dce/dce_aux.c
>>> @@ -653,7 +653,7 @@ bool dce_aux_transfer_with_retries(struct ddc_service *ddc,
>>>                                          if ((*payload->reply == AUX_TRANSACTION_REPLY_AUX_DEFER) ||
>>>                                                  (*payload->reply == AUX_TRANSACTION_REPLY_I2C_OVER_AUX_DEFER)) {
>>>                                                  if (payload->defer_delay > 0)
>>> -                                                       msleep(payload->defer_delay);
>>> +
>>> + udelay(payload->defer_delay * 1000);
>>>                                          }
>>>                                  }
>>>                                  break;
>>> --
>>> 2.17.1
>>>
>>> _______________________________________________
>>> amd-gfx mailing list
>>> amd-gfx at lists.freedesktop.org
>>> https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Flist
>>> s.freedesktop.org%2Fmailman%2Flistinfo%2Famd-gfx&data=02%7C01%7Cqi
>>> ngqing.zhuo%40amd.com%7C36c3bee68c28448769fa08d85a884619%7C3dd8961fe48
>>> 84e608e11a82d994e183d%7C0%7C0%7C637358888627498307&sdata=mynpHpiup
>>> J%2FU2o5gZNW%2Bft%2Fg2beFY86%2BzMRWoTZCghQ%3D&reserved=0
> _______________________________________________
> amd-gfx mailing list
> amd-gfx at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/amd-gfx