Re: [PATCH] i2c: pnx: fix warnings caused by enabling unprepared clock

[Vladimir Zapolskiy <vz@mleia.com>]

On 20.10.2015 11:54, Shubhrajyoti Datta wrote:
> On Mon, Oct 19, 2015 at 11:49 PM, Vladimir Zapolskiy <vz@mleia.com> wrote:
>> On 19.10.2015 19:21, Shubhrajyoti Datta wrote:
>>> On Sun, Oct 18, 2015 at 12:22 AM, Vladimir Zapolskiy <vz@mleia.com> wrote:
>>>> If common clock framework is configured, the driver generates a warning,
>>>> which is fixed by this change:
>>> Maybe just describe the issue and the fix.
>>> Is it just a warning or  the clk enable doesn't work ?
>>>
>>> I feel the crash log in the commit msg is not very informative.
>>
>> It is not a crash, it is a WARN_ON(1) backtrace.
> 
> ok
> 
>>
>> The backtrace is informative enough IMHO, because if you check the code
>> around given drivers/clk/clk.c:727 you may find the rootcause, the
>> WARN_ON() and that the clock is not enabled as a result.
>>
>> CLK_COMMON selects HAVE_CLK_PREPARE and all drivers used on platforms
>> with common clock framework must have clk_prepare/clk_unprepare, this
>> information is omitted as well-known.
>>
>> But if you insist, I will improve the commit message, I'm interested in
>> applying this trivial fix as soon as possible, then concentrate on doing
>> something more fascinating.
> 
> My request will be to describe the issue.
> Like the probe fails or you access registers with clocks not enabled
> 
> Or it is a warning fix and the controller works etc.

If I change initial description to the one below, would you be satisfied
with it?

The driver can not be used on a platform with common clock framework
until clk_prepare/clk_unprepare calls are added, otherwise clk_enable
calls will fail and a warning is generated.

--
With best wishes,
Vladimir