SUN6I a31 OPP table clock issue

SUN6I a31 OPP table clock issue

[Sergey Suloev]

Hi,

I was able to catch oops coming from OPP table on BananaPi M2.

https://pastebin.com/ZRaXTtEx

Let me know if need more detail.

Thank you

SUN6I a31 OPP table clock issue

[Maxime Ripard]

On Sat, Apr 07, 2018 at 10:53:10PM +0300, Sergey Suloev wrote:
> I was able to catch oops coming from OPP table on BananaPi M2.
> 
> https://pastebin.com/ZRaXTtEx
> 
> Let me know if need more detail.

Of course we do.

Which kernel version was this for? With which configuration? What
makes you say it's related to the OPP table, is it a regression or
something you just noticed?

Maxime

-- 
Maxime Ripard, Bootlin (formerly Free Electrons)
Embedded Linux and Kernel engineering
https://bootlin.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20180409/cd58ac63/attachment.sig>

SUN6I a31 OPP table clock issue

[Sergey Suloev]

On 04/09/2018 11:53 AM, Maxime Ripard wrote:
> On Sat, Apr 07, 2018 at 10:53:10PM +0300, Sergey Suloev wrote:
>> I was able to catch oops coming from OPP table on BananaPi M2.
>>
>> https://pastebin.com/ZRaXTtEx
>>
>> Let me know if need more detail.
> Of course we do.
>
> Which kernel version was this for? With which configuration? What
> makes you say it's related to the OPP table, is it a regression or
> something you just noticed?
Ok, first of all, I am having the problem using linus tree, 4.16 as it 
is, with no extra patches.
My config file is here [1].
The problem is easily avoided by just commenting all freqs out of the 
OPP table except for 1008 MHz, i.e. things work well when there is no 
freq switch at all.
It makes me think that there might be an issue in A31 clocks driver as 
it is the only important component used by cpufreq-dt.
As of cpufreq-dt driver it seems ok because works fine on other boards.

>
> Maxime
>
>
>
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel


[1] https://pastebin.com/eZxbra2f

SUN6I a31 OPP table clock issue

[Maxime Ripard]

On Mon, Apr 09, 2018 at 01:35:42PM +0300, Sergey Suloev wrote:
> On 04/09/2018 11:53 AM, Maxime Ripard wrote:
> > On Sat, Apr 07, 2018 at 10:53:10PM +0300, Sergey Suloev wrote:
> > > I was able to catch oops coming from OPP table on BananaPi M2.
> > > 
> > > https://pastebin.com/ZRaXTtEx
> > > 
> > > Let me know if need more detail.
> > Of course we do.
> > 
> > Which kernel version was this for? With which configuration? What
> > makes you say it's related to the OPP table, is it a regression or
> > something you just noticed?
>
> Ok, first of all, I am having the problem using linus tree, 4.16 as it is,
> with no extra patches.

Your kernel version is marked as dirty in your logs, which means you
have local changes. So it doesn't really qualify as "as is".

> My config file is here [1].
> The problem is easily avoided by just commenting all freqs out of the OPP
> table except for 1008 MHz, i.e. things work well when there is no freq
> switch at all.
> It makes me think that there might be an issue in A31 clocks driver as it is
> the only important component used by cpufreq-dt.
> As of cpufreq-dt driver it seems ok because works fine on other boards.

Could you add some printk's to ccu_mux_notifier_nb to see which
pointer is NULL?

Maxime

-- 
Maxime Ripard, Bootlin (formerly Free Electrons)
Embedded Linux and Kernel engineering
https://bootlin.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20180409/8f93bff4/attachment.sig>

SUN6I a31 OPP table clock issue

[Sergey Suloev]

On 04/09/2018 02:28 PM, Maxime Ripard wrote:
> On Mon, Apr 09, 2018 at 01:35:42PM +0300, Sergey Suloev wrote:
>> On 04/09/2018 11:53 AM, Maxime Ripard wrote:
>>> On Sat, Apr 07, 2018 at 10:53:10PM +0300, Sergey Suloev wrote:
>>>> I was able to catch oops coming from OPP table on BananaPi M2.
>>>>
>>>> https://pastebin.com/ZRaXTtEx
>>>>
>>>> Let me know if need more detail.
>>> Of course we do.
>>>
>>> Which kernel version was this for? With which configuration? What
>>> makes you say it's related to the OPP table, is it a regression or
>>> something you just noticed?
>> Ok, first of all, I am having the problem using linus tree, 4.16 as it is,
>> with no extra patches.
> Your kernel version is marked as dirty in your logs, which means you
> have local changes. So it doesn't really qualify as "as is".
well, yeah, there is a small change but in? absolutely different area
>
>> My config file is here [1].
>> The problem is easily avoided by just commenting all freqs out of the OPP
>> table except for 1008 MHz, i.e. things work well when there is no freq
>> switch at all.
>> It makes me think that there might be an issue in A31 clocks driver as it is
>> the only important component used by cpufreq-dt.
>> As of cpufreq-dt driver it seems ok because works fine on other boards.
> Could you add some printk's to ccu_mux_notifier_nb to see which
> pointer is NULL?
ok

> Maxime
>
>
>
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel