Re: [PATCHv2 7/7] drm/omap: hdmi4: fix use of uninitialized var

Re: [PATCHv2 7/7] drm/omap: hdmi4: fix use of uninitialized var

[Tony Lindgren]

* Tomi Valkeinen <tomi.valkeinen@ti.com> [190930 10:38]:
> If use_mclk is false, mclk_mode is written to a register without
> initialization. This doesn't cause any ill effects as the written value
> is not used when use_mclk is false.
> 
> To fix this, write use_mclk only when use_mclk is true.

Hey nice catch. Based on a quick test looks like this fixes an
issue where power consumption stays higher after using HDMI.

Would be nice to have merged in the v5.4-rc series:

Tested-by: Tony Lindgren <tony@atomide.com>
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCHv2 7/7] drm/omap: hdmi4: fix use of uninitialized var

[Tomi Valkeinen]

On 08/10/2019 17:13, Tony Lindgren wrote:
> * Tomi Valkeinen <tomi.valkeinen@ti.com> [190930 10:38]:
>> If use_mclk is false, mclk_mode is written to a register without
>> initialization. This doesn't cause any ill effects as the written value
>> is not used when use_mclk is false.
>>
>> To fix this, write use_mclk only when use_mclk is true.
> 
> Hey nice catch. Based on a quick test looks like this fixes an
> issue where power consumption stays higher after using HDMI.
> 
> Would be nice to have merged in the v5.4-rc series:
> 
> Tested-by: Tony Lindgren <tony@atomide.com>

Really? Ok, well, then it was a good random find =).

I did already push this to drm-misc-next, as I thought it does not have 
any real effect. I'll check if it's ok to push to drm-misc-fixes too, 
with Cc stable.

  Tomi

-- 
Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki.
Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCHv2 7/7] drm/omap: hdmi4: fix use of uninitialized var

[Tony Lindgren]

* Tomi Valkeinen <tomi.valkeinen@ti.com> [191008 14:17]:
> On 08/10/2019 17:13, Tony Lindgren wrote:
> > * Tomi Valkeinen <tomi.valkeinen@ti.com> [190930 10:38]:
> > > If use_mclk is false, mclk_mode is written to a register without
> > > initialization. This doesn't cause any ill effects as the written value
> > > is not used when use_mclk is false.
> > > 
> > > To fix this, write use_mclk only when use_mclk is true.
> > 
> > Hey nice catch. Based on a quick test looks like this fixes an
> > issue where power consumption stays higher after using HDMI.
> > 
> > Would be nice to have merged in the v5.4-rc series:
> > 
> > Tested-by: Tony Lindgren <tony@atomide.com>
> 
> Really? Ok, well, then it was a good random find =).

Yeah so it seems :) Earlier I thought there's still some
clkctrl setting wrong after using HDMI, but did not see
anything diffing the clkctrl registers before and after
and gave up.

> I did already push this to drm-misc-next, as I thought it does not have any
> real effect. I'll check if it's ok to push to drm-misc-fixes too, with Cc
> stable.

OK great thanks.

Tony
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCHv2 7/7] drm/omap: hdmi4: fix use of uninitialized var

[Tomi Valkeinen]

On 08/10/2019 17:21, Tony Lindgren wrote:
> * Tomi Valkeinen <tomi.valkeinen@ti.com> [191008 14:17]:
>> On 08/10/2019 17:13, Tony Lindgren wrote:
>>> * Tomi Valkeinen <tomi.valkeinen@ti.com> [190930 10:38]:
>>>> If use_mclk is false, mclk_mode is written to a register without
>>>> initialization. This doesn't cause any ill effects as the written value
>>>> is not used when use_mclk is false.
>>>>
>>>> To fix this, write use_mclk only when use_mclk is true.
>>>
>>> Hey nice catch. Based on a quick test looks like this fixes an
>>> issue where power consumption stays higher after using HDMI.
>>>
>>> Would be nice to have merged in the v5.4-rc series:
>>>
>>> Tested-by: Tony Lindgren <tony@atomide.com>
>>
>> Really? Ok, well, then it was a good random find =).
> 
> Yeah so it seems :) Earlier I thought there's still some
> clkctrl setting wrong after using HDMI, but did not see
> anything diffing the clkctrl registers before and after
> and gave up.
> 
>> I did already push this to drm-misc-next, as I thought it does not have any
>> real effect. I'll check if it's ok to push to drm-misc-fixes too, with Cc
>> stable.
> 
> OK great thanks.

Pushing this to fixes too would cause conflicts, so we shouldn't push 
without good reason. How much power saving you see?

I think this can still be sent to stable later, after it has been merged 
to mainline.

  Tomi

-- 
Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki.
Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCHv2 7/7] drm/omap: hdmi4: fix use of uninitialized var

[Tony Lindgren]

* Tomi Valkeinen <tomi.valkeinen@ti.com> [191010 06:48]:
> On 08/10/2019 17:21, Tony Lindgren wrote:
> > * Tomi Valkeinen <tomi.valkeinen@ti.com> [191008 14:17]:
> > > On 08/10/2019 17:13, Tony Lindgren wrote:
> > > > * Tomi Valkeinen <tomi.valkeinen@ti.com> [190930 10:38]:
> > > > > If use_mclk is false, mclk_mode is written to a register without
> > > > > initialization. This doesn't cause any ill effects as the written value
> > > > > is not used when use_mclk is false.
> > > > > 
> > > > > To fix this, write use_mclk only when use_mclk is true.
> > > > 
> > > > Hey nice catch. Based on a quick test looks like this fixes an
> > > > issue where power consumption stays higher after using HDMI.
> > > > 
> > > > Would be nice to have merged in the v5.4-rc series:
> > > > 
> > > > Tested-by: Tony Lindgren <tony@atomide.com>
> > > 
> > > Really? Ok, well, then it was a good random find =).
> > 
> > Yeah so it seems :) Earlier I thought there's still some
> > clkctrl setting wrong after using HDMI, but did not see
> > anything diffing the clkctrl registers before and after
> > and gave up.
> > 
> > > I did already push this to drm-misc-next, as I thought it does not have any
> > > real effect. I'll check if it's ok to push to drm-misc-fixes too, with Cc
> > > stable.
> > 
> > OK great thanks.
> 
> Pushing this to fixes too would cause conflicts, so we shouldn't push
> without good reason. How much power saving you see?

Sure no rush with this one. I should also test again that it
really fixes the issue I'm seeing.

Hmm so what register does this clock actually change?

I'm seeing an increase of few tens of extra mW, which means at
least one day of standby time less for me :) It does not happen
always, maybe half of the time.

> I think this can still be sent to stable later, after it has been merged to
> mainline.

Yes sure.

Thanks,

Tony
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCHv2 7/7] drm/omap: hdmi4: fix use of uninitialized var

[Tomi Valkeinen]

On 10/10/2019 16:24, Tony Lindgren wrote:

> Hmm so what register does this clock actually change?
> 
> I'm seeing an increase of few tens of extra mW, which means at
> least one day of standby time less for me :) It does not happen
> always, maybe half of the time.

I have no idea why this would affect power consumption. As far as I can 
understand, the bits written here are a clk divider MCLK. I don't see 
why that would affect.

Maybe Jyri or Peter has an idea, I have never looked at the HDMI audio side.

  Tomi

-- 
Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki.
Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCHv2 7/7] drm/omap: hdmi4: fix use of uninitialized var

[Tony Lindgren]

* Tomi Valkeinen <tomi.valkeinen@ti.com> [191011 10:25]:
> On 10/10/2019 16:24, Tony Lindgren wrote:
> 
> > Hmm so what register does this clock actually change?
> > 
> > I'm seeing an increase of few tens of extra mW, which means at
> > least one day of standby time less for me :) It does not happen
> > always, maybe half of the time.
> 
> I have no idea why this would affect power consumption. As far as I can
> understand, the bits written here are a clk divider MCLK. I don't see why
> that would affect.

Yeah you're right, and I just got lucky initially.

I have seen the power consumption stay higher already with
the patch applied. The clocks seem just fine.

> Maybe Jyri or Peter has an idea, I have never looked at the HDMI audio side.

I'll try dumping out the hdmi registers before and after
when I get a chance.

Regards,

Tony
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel