Re: [PATCH RESEND] poll: allow f_op->poll to sleep, take #2

[Tejun Heo <tj@kernel.org>]

Miklos Szeredi wrote:
> On Thu, 20 Nov 2008, Tejun Heo wrote:
> 
> [snip]
> 
>> +int poll_schedule_timeout(struct poll_wqueues *pwq, int state,
>> +			  ktime_t *expires, unsigned long slack)
> 
> The 'state' parameter is unused, and is always called with the
> TASK_INTERRUPTIBLE value.  Shouldn't it be removed?
> 
>> +{
>> +	int rc = -EINTR;
>> +
>> +	set_current_state(TASK_INTERRUPTIBLE);

Aieee... this should have been set_current_state(state).  We can also
remove @state but this being a schedule() function I think it's better
to pass @state explicitly.

>> +	if (!pwq->triggered)
>> +		rc = schedule_hrtimeout_range(expires, slack, HRTIMER_MODE_ABS);
>> +	__set_current_state(TASK_RUNNING);
>> +
>> +	/* clear triggered for the next iteration */
>> +	pwq->triggered = 0;
>> +
>> +	return rc;
>> +}
>> +
>> +EXPORT_SYMBOL(poll_schedule_timeout);
> 
> Checkpatch warning:
> 
> WARNING: EXPORT_SYMBOL(foo); should immediately follow its function/variable
> #118: FILE: fs/select.c:216:
> +EXPORT_SYMBOL(poll_schedule_timeout);
> 
> [snip]
> 
>> Index: work/include/linux/poll.h
>> ===================================================================
>> --- work.orig/include/linux/poll.h
>> +++ work/include/linux/poll.h
>> @@ -57,6 +57,8 @@ struct poll_table_entry {
>>  struct poll_wqueues {
>>  	poll_table pt;
>>  	struct poll_table_page * table;
>> +	struct task_struct * polling_task;
>> +	int triggered;
> 
> Checkpatch error:
> 
> ERROR: "foo * bar" should be "foo *bar"
> #173: FILE: include/linux/poll.h:60:
> +       struct task_struct * polling_task;

For both, I was trying to stay consistent with the environment.  I
find mixed styles in close proximity much uglier than slightly
different but consistent style.  Eh... Is the consensus checkpatch or
die?

Thanks.

-- 
tejun