Re: [PATCH] apm-emulation: use wait_event_freezable() instead offreezer_[do_not_]count()

[Jiri Kosina]

On Thu, 18 Aug 2011, Tejun Heo wrote:

> vfork is moving away from freezer_[do_not_]count() one way or the
> other leaving apm_ioctl() as the only user.  apm_ioctl() just wants to
> wait for suspend/resume cycle to complete without hindering the
> freezer.  Use wait_event_freezable() instead.
> 
> The only annoyance is that wait_event_freezable() wakes up with
> -ERESTART if there are pending signals while apm_ioctl() wants to
> ignore all signals until suspend is complete.  We can play with
> @current->[real_]blocked but this is hardly a performance or latency
> critical path - simply chill a bit on each iteration until
> SUSPEND_DONE for unlikely cases where there are pending signals.
> 
> Signed-off-by: Tejun Heo <tj@xxxxxxxxxx>
> Cc: Oleg Nesterov <oleg@xxxxxxxxxx>
> ---
> Compile tested only.  It would be great if someone w/ affected
> configuraiton can test this.
> 
> Thank you.
> 
>  drivers/char/apm-emulation.c |   16 ++++++----------
>  1 file changed, 6 insertions(+), 10 deletions(-)
> 
> Index: work/drivers/char/apm-emulation.c
> ===================================================================
> --- work.orig/drivers/char/apm-emulation.c
> +++ work/drivers/char/apm-emulation.c
> @@ -300,17 +300,13 @@ apm_ioctl(struct file *filp, u_int cmd, 
>  			/*
>  			 * Wait for the suspend/resume to complete.  If there
>  			 * are pending acknowledges, we wait here for them.
> +			 * wait_event_freezable() is interruptible and pending
> +			 * signal can cause busy looping.  We aren't doing
> +			 * anything critical, chill a bit on each iteration.
>  			 */
> -			freezer_do_not_count();
> -
> -			wait_event(apm_suspend_waitqueue,
> -				   as->suspend_state == SUSPEND_DONE);
> -
> -			/*
> -			 * Since we are waiting until the suspend is done, the
> -			 * try_to_freeze() in freezer_count() will not trigger
> -			 */
> -			freezer_count();
> +			while (wait_event_freezable(apm_suspend_waitqueue,
> +					as->suspend_state == SUSPEND_DONE))
> +				msleep(10);
>  			break;
>  		case SUSPEND_ACKTO:
>  			as->suspend_result = -ETIMEDOUT;

Applied, thanks Tejun.

-- 
Jiri Kosina
SUSE Labs
--
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/