Messages in this thread | | | Date | Thu, 8 Mar 2012 14:45:26 -0800 | From | Andrew Morton <> | Subject | Re: [PATCH] hugetlbfs: lockdep annotate root inode properly |
| |
On Thu, 8 Mar 2012 17:33:34 -0500 Dave Jones <davej@redhat.com> wrote:
> On Thu, Mar 08, 2012 at 02:19:38PM -0800, Andrew Morton wrote: > > On Thu, 8 Mar 2012 21:49:52 +0000 > > Al Viro <viro@ZenIV.linux.org.uk> wrote: > > > > > > So we need to pull the i_mutex out of hugetlbfs_file_mmap(). > > > > > > IIRC, you have a patch in your tree doing just that... > > > > Nope. > > > > But it seems that you've recently seen such a patch - can you recall > > where? > > this ? https://lkml.org/lkml/2012/2/23/64 >
Thanks, yes, probably that. Needs the i_size_read()/write() changes.
I worry a bit about the region handling code in mm/hugetlb.c.
* The region data structures are protected by a combination of the mmap_sem * and the hugetlb_instantion_mutex. To access or modify a region the caller * must either hold the mmap_sem for write, or the mmap_sem for read and * the hugetlb_instantiation mutex:
I hope that's true - it would be nice to have some debug assertions in the various region_foo() functions to verify that the required locks are held.
But if that code is all nice and tight, I guess that removing that i_mutex acquisition will be pretty simple.
| |