Re: [PATCH] dio: lock refcount operations

[Jeff Moyer]

==> Regarding Re: [PATCH] dio: lock refcount operations; Zach Brown <zach.brown@xxxxxxxxxx> adds:

>> I don't believe that this can happen.

zach.brown> Yeah, I think my brain made the leap to spurious wake-ups from
zach.brown> hashed wait queues.  Which aren't being used :).  As long as
zach.brown> it's a private wait queue and sleeps and sleeps with
zach.brown> UNINTERRUPTIBLE it seems ok.

zach.brown> Do you think the cleanup shouldn't be done?  It seems easier to
zach.brown> understand after the patch, and makes dio_await_one() pretty
zach.brown> darn straight forward.

The patch looks sane to me, and I appreciate all of your comments in the
code.

zach.brown> The addition of the interrupt masking spin lock acquiry in
zach.brown> dio_bio_submit() looks alarming.  This lock acquiry existed in
zach.brown> that path before the recent dio completion patch set.  We
zach.brown> shouldn't expect significant performance regression from
zach.brown> returning to the behaviour that existed before the completion
zach.brown> clean up work. 

Are you going to quantify this at all?  I think we should.

-Jeff
-
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/