On Wed, 2008-07-23 at 09:35 +0000, Jarek Poplawski wrote:

On Wed, Jul 23, 2008 at 11:03:06AM +0200, Peter Zijlstra wrote:

quoted

On Wed, 2008-07-23 at 08:54 +0000, Jarek Poplawski wrote:

quoted

On Wed, Jul 23, 2008 at 12:59:21AM -0700, David Miller wrote:

quoted

From: Jarek Poplawski <redacted>
Date: Wed, 23 Jul 2008 06:20:36 +0000

quoted

PS: if there is nothing new in lockdep the classical method would
be to change this static array:

static struct lock_class_key
netdev_xmit_lock_key[ARRAY_SIZE(netdev_lock_type)];

to

static struct lock_class_key
netdev_xmit_lock_key[ARRAY_SIZE(netdev_lock_type)][MAX_NUM_TX_QUEUES];

and set lockdep classes per queue as well. (If we are sure we don't
need lockdep subclasses anywhere this could be optimized by using
one lock_class_key per 8 queues and spin_lock_nested()).

Unfortunately MAX_NUM_TX_QUEUES is USHORT_MAX, so this isn't really
a feasible approach.

Is it used by real devices already? Maybe for the beginning we could
start with something less?

quoted

spin_lock_nested() isn't all that viable either, as the subclass
limit is something like 8.

This method would need to do some additional counting: depending of
a queue number each 8 subsequent queues share (are set to) the same
class and their number mod 8 gives the subqueue number for
spin_lock_nested().

I'll try to find if there is something new around this in lockdep.
(lockdep people added to CC.)

There isn't.

Is there a static data structure that the driver needs to instantiate to
'create' a queue? Something like:

/* this imaginary e1000 hardware has 16 hardware queues */
static struct net_tx_queue e1000e_tx_queues[16]; 

I guess, not.

Then, IMHO, we could be practical and simply skip lockdep validation
for "some" range of locks: e.g. to set the table for the first 256
queues only, and to do e.g. __raw_spin_lock() for bigger numbers. (If
there are any bad locking patterns this should be enough for checking.)

definite NAK on using raw spinlock ops...

I'll go look at this multiqueue stuff to see if there is anything to be
done.. David, what would be a good commit to start reading?