Re: [patch] sched, tty: fix scheduling latencies in tty_io.c

[Alan Cox]

On Maw, 2004-09-14 at 13:27, Ingo Molnar wrote:
> would it be a sufficient fix to grab some sort of tty_sem mutex in the
> places where the patch drops the BKL - or are there other entry paths to
> this code?

There are a whole load of entry points. I've been trying to document all
the ldisc layer stuff as I fix the basic bugs in it. A semaphore isnt
sufficient because some of the entry points have to be multi-entry so
you'd need to go modify all the ldisc internals (I think that will have
to happen eventually). It also relies on it for read v write locking
still.

So far all I've fixed is the vfs/ldisc locking to ensure that open/close
and other events are properly sequenced. I've not yet finished tackling
the ldisc/driver cases where driver->close() completes and the ldisc
calls the driver still.

At that point its at least then down to the drivers to get fixed.

-
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/