Re: [PATCH] PM / QoS: Fix device resume latency PM QoS

[Andy Shevchenko]

On Thu, Oct 26, 2017 at 11:38 AM, Rafael J. Wysocki <rafael@xxxxxxxxxx> wrote:
> On Wed, Oct 25, 2017 at 10:06 PM, Andy Shevchenko
> <andy.shevchenko@xxxxxxxxx> wrote:
>> On Tue, Oct 24, 2017 at 11:49 AM, Rafael J. Wysocki <rjw@xxxxxxxxxxxxx> wrote:
>>> On Tuesday, October 24, 2017 7:54:09 AM CEST Ramesh Thomas wrote:
>>>> On 2017-10-20 at 13:27:34 +0200, Rafael J. Wysocki wrote:
>>
>>>> >  static ssize_t pm_qos_resume_latency_store(struct device *dev,
>>>> > @@ -228,11 +235,19 @@ static ssize_t pm_qos_resume_latency_sto
>>>> >     s32 value;
>>>> >     int ret;
>>
>>>> > +   if (!kstrtos32(buf, 0, &value)) {
>>>> > +           /*
>>>> > +            * Prevent users from writing negative or "no constraint" values
>>>> > +            * directly.
>>>> > +            */
>>>> > +           if (value < 0 || value == PM_QOS_RESUME_LATENCY_NO_CONSTRAINT)
>>>> > +                   return -EINVAL;
>>
>>>> > +           if (value == 0)
>>>> > +                   value = PM_QOS_RESUME_LATENCY_NO_CONSTRAINT;
>>>> > +   } else if (!strcmp(buf, "n/a") || !strcmp(buf, "n/a\n")) {
>>>>
>>>> Can the 2 checks for "n/a" be combined by checking first 3 characters?
>>>
>>> No, because "n/asomething" would then match too.
>>
>> If I don't missed anything, kernfs is aware of \n which means the
>> first check is enough.
>> Am I correct?
>
> I'm not sure, honestly. :-)

Okay, just a summary:
1. kernfs guarantees that buffer is NULL terminated
2. sysfs guarantees that the buffer is not empty
3. kstrto* are aware of '\n'
4. sysfs_streq() and __sysfs_match_string() are aware of '\n'

Thus, we just may use sysfs_streq() for that.

I will prepare a clean up patch on top of this fix if you are okay with it.

> Anyway, that can be fixed up later and the bug in question is rather urgent.

Sure.

-- 
With Best Regards,
Andy Shevchenko