RE: Subject: [PATCH V7 2/4] mm: frontswap: core code

[Dan Magenheimer <dan.magenheimer@oracle.com>]

> > My concern was race in counters. Even you allow race in frontswap_succ_puts++,
> >
> > Don't you need some lock for
> > 	sis->frontswap_pages++
> > 	sis->frontswap_pages--
> 
> Hmmm... OK, you've convinced me.  If this counter should be one and
> a race leaves it as zero, I think data corruption could result on
> a swapoff or partial swapoff.  And after thinking about it, I
> think I also need to check for locking on frontswap_set/clear
> as I don't think these bitfield modifiers are atomic.
> 
> Thanks for pointing this out.  Good catch!  I will need to
> play with this and test it so probably will not submit V8 until
> next week as today is a vacation day for me.

Silly me: Of course set_bit and clear_bit ARE atomic.  I will
post V8 later today with the only changes being frontswap_pages
is now a type atomic_t.

Thanks again for catching this, Kame!

Thanks,
Dan

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>