Re: [PATCH] Add schedule_timeout_{interruptible,uninterruptible}{,_msecs}() interfaces

[Nishanth Aravamudan]

On 23.07.2005 [19:01:57 +0200], Roman Zippel wrote:
> Hi,
> 
> On Sat, 23 Jul 2005, Nishanth Aravamudan wrote:
> 
> > 	set_current_state(TASK_{,UN}INTERRUPTIBLE);
> > 	schedule_timeout(msecs_to_jiffies(some_constant_msecs));
> > 
> > just have an interface that allows
> > 
> > 	schedule_timeout_msecs_{,un}interruptible(some_constant_msecs);
> > 
> > and push the jiffies conversion to common code?
> 
> What's wrong with just:
> 
> 	schedule_timeout_{,un}interruptible(msecs_to_jiffies(some_constant_msecs));

Nothing, I suppose. I just prefer directly using msecs. I understand
your point more now, I think. You are worried about those people that
actually use the return value of schedule_timeout().

> The majority of users use a constant, which can already be converted at 
> compile tile.
> Additionally such an interface also had to return a ms value and instead 
> of that constant conversion, the user is better off to work with jiffies 
> directly.

So, I just spent a good hour looking at every caller of
schedule_timeout() which actually stores the return value. Beyond the
other wrappers for it (wait_event(), wait_for_completion(),
sys_nanosleep(), etc., which I will leave alone using schedule_timeout()
until I can change *their* parameters ;) ), I found two cases.

1) Sleep, see if you actually slept the whole time:

	remainder = schedule_timeout(some_value_in_jiffies);
	if (!remaining)
		report_timeout();

2) Sleep in a loop, keeping track of remaining timeout each iteration:

	while (timeout) {
		do_some_stuff();
		timeout = schedule_timeout(timeout);
		if (some_condition)
			break;
	}

Clearly, neither needs to use jiffies. The former only wants to know if
the full timeout elapsed. I didn't find anyone returning that stored
value (again, excepting wrapper interfaces) to the caller. They just
want to know if they should return -ETIME{,DOUT}. The latter just is a
means to guarantee the entire time is slept, but doesn't care about the
units.

Now, some of these might depend on structures which have members with
jiffy-unit values. But I will be more than happy to try and either leave
them alone or convert those structures. We'll see about that on a
case-by-case basis?

Thanks,
Nish
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/