Messages in this thread | | | From | Song Liu <> | Date | Tue, 27 Feb 2024 14:28:00 -0800 | Subject | Re: [PATCH v4] md/raid5: fix atomicity violation in raid5_cache_count |
| |
On Thu, Feb 1, 2024 at 11:11 PM Yu Kuai <yukuai1@huaweicloud.com> wrote: > > Hi, > > 在 2024/01/30 15:37, Song Liu 写道: > > On Thu, Jan 11, 2024 at 11:10 PM Gui-Dong Han <2045gemini@gmailcom> wrote: > >> > > [...] > >> > >> raid5_release_stripe(sh); > >> - conf->max_nr_stripes++; > >> + WRITE_ONCE(conf->max_nr_stripes, conf->max_nr_stripes + 1); > > > > This is weird. We are reading max_nr_stripes without READ_ONCE. > > We don't need READ_ONCE() here because writers are protected by > 'cache_size_mutex', there are no concurrent writers, it's safe to > read 'max_nr_stripes' directly.
OK, that makes sense. Applied to md-6.9.
Thanks, Song
| |