Messages in this thread | | | Date | Fri, 3 Jan 2014 20:55:19 +0100 | From | Oleg Nesterov <> | Subject | [PATCH v2 0/1] mm: fix the theoretical compound_lock() vs prep_new_page() race |
| |
On 12/23, Andrea Arcangeli wrote: > > Acked-by: Andrea Arcangeli <aarcange@redhat.com> > > Only one improvement possible, the smp_wmb() could have been put under > CONFIG_TRANSPARENT_HUGEPAGE somehow. No difference for x86-64 though.
I thought it would be better to serialize it with get_page_unless_zero() in general, but OK, As the changelog says I didn't find another problematic caller, so please see v2. I preserved your ack, thanks.
And I would like to know your opinion about other changes I sent ;)
Oleg.
| |