On Tue, Apr 02, 2019 at 04:43:57PM -0700, Davidlohr Bueso wrote: > On Tue, 02 Apr 2019, Andrew Morton wrote: > > > Also, we didn't remove any down_write(mmap_sem)s from core code so I'm > > thinking that the benefit of removing a few mmap_sem-takings from a few > > obscure drivers (sorry ;)) is pretty small. > > afaik porting the remaining incorrect users of locked_vm to pinned_vm was > the next step before this one, which made converting locked_vm to atomic > hardly worth it. Daniel? Right, as you know I tried those incorrect users first, but there were concerns about user-visible changes regarding RLIMIT_MEMLOCK and pinned_vm/locked_vm without the accounting problem between all three being solved. To my knowledge no one has a solution for that, so in the meantime I'm taking the incremental step of getting rid of mmap_sem for locked_vm users. The locked_vm -> pinned_vm conversion can happen later.
