Re: [PATCH v2] mmc: sdhci-of-arasan: fix set_clock when a phy is supported

[Ulf Hansson]

On 4 May 2016 at 11:24, Shawn Lin <shawn.lin@xxxxxxxxxxxxxx> wrote:
> å 2016/5/4 16:35, Heiko Stuebner åé:
>>
>> Am Mittwoch, 4. Mai 2016, 09:48:55 schrieb Shawn Lin:
>>>
>>> å 2016/4/28 18:38, Ulf Hansson åé:
>>>>
>>>> On 28 April 2016 at 10:38, Shawn Lin <shawn.lin@xxxxxxxxxxxxxx> wrote:
>>>>>
>>>>> commit 61b914eb81f8 ("mmc: sdhci-of-arasan: add phy support for
>>>>> sdhci-of-arasan") introduce phy support for arasan. According to
>>>>> the vendor's databook, we should make sure the phy is in poweroff
>>>>> status before we configure the clk stuff. Otherwise it may cause
>>>>> some IO sample timing issues from the test. And we don't need this
>>>>> extra operation while running in low performance mode since phy
>>>>> doesn't trigger sampling block.
>>>>>
>>>>> Signed-off-by: Shawn Lin <shawn.lin@xxxxxxxxxxxxxx>
>>>>
>>>>
>>>> Thanks, applied for next!
>>>
>>>
>>> Thanks for applying v2 but, could you please drop it and applied v3
>>> I just send :) .
>>> Because v2 introduces a bug. Set_clock callback will be
>>> called under the protection of spinlock in sdhci_do_set_ios. However
>>> PHY APIs need mutex which fails the kernel's debug check
>>>
>>> BUG: sleeping function called from invalid context at
>>> kernel/locking/mutex.c
>>
>>
>> I think in general you will need to provide a follow-up patch instead of
>> replacing the old version, as Ulf might not want to restructure an already
>> published branch (depends on the maintainer), especially as already other
>> stuff is on top of your patch [0].
>>
>
> Aha.. yes I see other patches on top of mine now. Thanks for reminding
> it.
>
> Ulf, would you mind that I send a new patch to fix it, or you are prone
> to restructure the next?

I re-base my next branch occasionally, as I prefer it to be as clean
as possible without obvious bisect issues.

In this case there were no conflict by just removing the patch, so I
did that, thanks!

Kind regards
Uffe