Re: [PATCH v4] blk-mq: Fix race conditions in request timeout handling

["tj@kernel.org" <tj@kernel.org>]

Hello, Israel.

On Wed, Apr 11, 2018 at 07:16:14PM +0300, Israel Rukshin wrote:
> >Just noticed this one, this looks interesting to me as well. Israel,
> >can you run your test with this patch?
> 
> Yes, I just did and it looks good.

Awesome.

> >>+++ b/include/linux/blkdev.h
> >>@@ -227,6 +227,8 @@ struct request {
> >>� ����� unsigned int extra_len;��� /* length of alignment and padding */
> >>� +��� bool missed_completion;
> >>+
> >
> >Would be nicer if we can flag this somewhere instead of adding a hole to
> >struct request...

I missed it before.  It's actually being put in an existing hole, so
the struct size stays the same before and after.  It's a bit of
cheating cuz this is one of the two holes which can be removed by
swapping two fields.

Re. making it a flag, regardless of whether this is a flag or a
separate field, we need to add a new field because there currently is
no field which can be modified by the party who doesn't own the
request, so if we make it a flag, we need to add sth like unsigned
long atom_flags.

Thanks.

-- 
tejun