Re: [PATCH -rt] catch put_task_struct RCU handling up to mainline

[Paul E. McKenney]

On Mon, Jul 10, 2006 at 07:10:49PM +0100, Esben Nielsen wrote:
> On Mon, 10 Jul 2006, Paul E. McKenney wrote:
> >On Sat, Jul 08, 2006 at 02:59:37PM +0100, Esben Nielsen wrote:

[ . . . ]

> >>The work should be defered to a low priority task. Using rcu is
> >>probably overkill because it also introduces other delays. A tasklet
> >>or a dedicated task would be better.
> >
> >Agreed -- if there is in fact a legitimate non-error code path, then
> >a patch that used some deferral mechanism would be good.  But RCU is
> >overkill, and misleading overkill at that!
> >
> 
> I think this is a legitimate situation. lock 1 is owned by B which is
> blocked on lock 2 which is owned by C
> 
>  CPU1:                                      CPU2
>     RT task A locks lock 1                C runs something
>     A boosts B to RT
>     A does get_task_struct B
>     A enables interrupts                  C unlocks lock 2
>     An very long interrupt is running     B unlocks lock 2
>                                           B unlocks lock 1
>                                           B is deboosted
>                                           B exits
>     A gets CPU1 again
>     A does put_task_struct B
> 
> I don't know if the timing is realistic, but theoretically it is possible.
> It might also be possible the B exits on another CPU even without the long
> interrupt handler. If A has cpu affinity to CPU1 it is enough if a higher 
> priority task preempts it on CPU1.

For this to happen, either A has to be at a lower priority than the irq
tasks or the interrupt has to be a hard irq (e.g., scheduling clock
interrupt).  In the first case, the added cleanup processing seems
inconsequential compared to (say) an interrupt doing network protocol
processing.  In the second case, B does not do its put_task_struct()
until after the hard irq returns (because the put_task_struct() is invoked
from a call_rcu() callback), which makes the above scenario unlikely,
though perhaps not impossible.

If the second scenario is in fact possible, would you be willing to
supply the appropriate deferral code?  I believe we both agree that RCU
is not really the right deferral mechanism in this situation.

							Thanx, Paul
-
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/