possible deadlock in shmem_fallocate (4)

Suren Baghdasaryan surenb at google.com
Thu Jul 16 02:49:52 UTC 2020


On Tue, Jul 14, 2020 at 11:36 PM Michal Hocko <mhocko at kernel.org> wrote:
>
> On Tue 14-07-20 10:32:20, Suren Baghdasaryan wrote:
> [...]
> > I'm not sure how we can annotate the fact that the inode_lock in
> > generic_file_write_iter and in shmem_fallocate always operate on
> > different inodes. Ideas?
>
> I believe that the standard way is to use lockdep_set_class on your
> backing inode. But asking lockdep experts would be better than relying
> on my vague recollection

Thanks Michal! I think https://lkml.org/lkml/2020/7/15/1390 should fix
it, however I was unable to reproduce the lockdep warning to confirm
that this warning gets fixed by the patch.

>
> --
> Michal Hocko
> SUSE Labs


More information about the devel mailing list