Messages in this thread | | | Subject | Re: One (possible) x86 get_user_pages bug | From | Peter Zijlstra <> | Date | Thu, 27 Jan 2011 17:56:43 +0100 |
| |
On Thu, 2011-01-27 at 16:07 +0000, Jan Beulich wrote: > > Nick, based on your doing of the initial implementation, would > you be able to estimate whether disabling get_user_pages_fast() > altogether for Xen would be performing measurably worse than > adding the locks (but continuing to avoid acquiring mm->mmap_sem) > as suggested by Xiaowei? That's of course only if the latter is correct > at all, of which I haven't fully convinced myself yet.
Also, I don't think taking the pte_lock is sufficient in your case.. since x86 mmu_gather frees the page tables themselves too, your dereference of the pgd/pud/pmd will be subject to the same deref after free problems.
| |