Messages in this thread | | | Date | Tue, 17 May 2016 11:31:10 -0700 | Subject | Re: [GIT PULL] killable rwsems for v4.7 | From | Linus Torvalds <> |
| |
On Tue, May 17, 2016 at 11:05 AM, Al Viro <viro@zeniv.linux.org.uk> wrote: > > I'm not sure - killable write is needed in a bunch of places there (and the > only reason it's not used in #work.lookups is to avoid even more merge > headache; as soon as both are merged, I'll post a trivial followup switching > half a dozen places to it), killable read... Do we really need it? > The only plausible user right now (->i_rwsem one, that is) is parallel readdir. > And I'm not convinced that we need to make that one killable. We can > (down_read_killable seems to be easy to put together), but is it worth > using it in that usecase?
It may not be a big deal.
Especially since I also expect to merge the "stop readdir in the middle if a signal is pending" patch that was buggy last time around but Ted has a fixed and extended version of it.
So together with the fact that most of the time we hopefully end up not blocking anyway (because taking the rwsem for writing is hopefully less common) I guess that the possible latency issue just isn't going to ever be noticeable.
So let's just see if we ever hit a situation where people end up even noticing the issue.
Linus
| |