Re: [PATCHv2 1/8] wait_event_interruptible_locked() interface

[Michal Nazarewicz <mina86@mina86.com>]

> On Fri, 9 Apr 2010, Michal Nazarewicz wrote:
>> New wait_event_interruptible{,_exclusive}_locked{,_irq,_irqsave}
>> macros added.  They work just like versions without _locked* suffix

Thomas Gleixner <tglx@linutronix.de> writes:
> The _irqsave variant is really not necessary. It's actively wrong.
>
> If you go to wait then the state _before_ acquiring the waitqueue head
> lock must be irqs enabled. Otherwise you would schedule with
> interrupts disabled after the unlock_irqrestore which is a BUG.
>
> So if there is code which uses spin_lock_irqsave() in the wait path
> then this code is wrong and needs to be fixed to spin_(un)lock_irq()
> first instead of adding a bogus interface.

I haven't seen the big picture here.  Thanks for pointing that out.

>> +
>> +#define __wait_event_interruptible_locked(wq, condition, ret, exclusive, lock, unlock, lock_args) \
>
> That will also simplify this to (wq, condition, exclusive, lockmode)
>
>> +do {									\
>> +	DEFINE_WAIT(__wait);						\
>> +									\
>> +	if (exclusive)							\
>> +		__wait.flags |= WQ_FLAG_EXCLUSIVE;			\
>> +	else								\
>> +		__wait.flags &= ~WQ_FLAG_EXCLUSIVE;			\
>> +	__add_wait_queue_tail(&(wq), &__wait);				\
>> +									\
>> +	do {								\
>> +		set_current_state(TASK_INTERRUPTIBLE);			\
>> +		if (signal_pending(current)) {				\
>> +			ret = -ERESTARTSYS;				\
>> +			break;						\
>> +		}							\
>> +		spin_unlock ##unlock lock_args;				\
>> +		schedule();						\
>> +		spin_lock ##lock lock_args;				\
>> +	} while (!(condition));						\
>> +	__remove_wait_queue(&(wq), &__wait);				\
>> +	__set_current_state(TASK_RUNNING);				\
>> +} while (0)
>> +
>> +
>> +/**
>> + * wait_event_interruptible_locked - sleep until a condition gets true
>> + * @wq: the waitqueue to wait on
>> + * @condition: a C expression for the event to wait for
>> + *
>> + * The process is put to sleep (TASK_INTERRUPTIBLE) until the
>> + * @condition evaluates to true or a signal is received.
>> + * The @condition is checked each time the waitqueue @wq is woken up.
>> + *
>> + * It must be called with wq.lock being held.  This spinlock is
>> + * unlocked while sleeping but @condition testing is done while lock
>> + * is held and when this macro exits the lock is held.
>> + *
>> + * The lock is locked/unlocked using spin_lock()/spin_unlock()
>> + * functions which must match the way they are locked/unlocked outside
>> + * of this macro.
>> + *
>> + * wake_up_locked() has to be called after changing any variable that could
>> + * change the result of the wait condition.
>> + *
>> + * The function will return -ERESTARTSYS if it was interrupted by a
>> + * signal and 0 if @condition evaluated to true.
>> + */
>> +#define wait_event_interruptible_locked(wq, condition)			\
>> +({									\
>> +	int __ret = 0;							\
>> +	if (!(condition))						\
>> +		__wait_event_interruptible_locked(wq, condition, __ret, 0, , , (&(wq).lock)); \
>
>   I had to look more than once to figure out how that code might
>   return anything else than 0. Can we please change that to
>
>   if (!(condition))
>      __ret = __wait_.....();
>
>   to make that less confusing ?   

That's really how the rest of the wait_event*() macros are done so I'd
prefer to stack with the rest of the code.

>> +	__ret;								\
>> +})

Again, I'll resend the patches by the end of the week.

-- 
Best regards,                                         _     _
 .o. | Liege of Serenly Enlightened Majesty of      o' \,=./ `o
 ..o | Computer Science,  Michal "mina86" Nazarewicz   (o o)
 ooo +--<mina86-tlen.pl>--<jid:mina86-jabber.org>--ooO--(_)--Ooo--