| Message ID | 20210203132022.92406-9-jinpu.wang@cloud.ionos.com |
|---|---|
| State | Superseded |
| Headers | show |
| Series | Some misc fixes | expand |
On Wed, Feb 03, 2021 at 02:20:22PM +0100, Jack Wang wrote: > From: Xiao Ni <xni@redhat.com> > > One customer reports a crash problem which causes by flush request. It > triggers a warning before crash. > > /* new request after previous flush is completed */ > if (ktime_after(req_start, mddev->prev_flush_start)) { > WARN_ON(mddev->flush_bio); > mddev->flush_bio = bio; > bio = NULL; > } > > The WARN_ON is triggered. We use spin lock to protect prev_flush_start and > flush_bio in md_flush_request. But there is no lock protection in > md_submit_flush_data. It can set flush_bio to NULL first because of > compiler reordering write instructions. > > For example, flush bio1 sets flush bio to NULL first in > md_submit_flush_data. An interrupt or vmware causing an extended stall > happen between updating flush_bio and prev_flush_start. Because flush_bio > is NULL, flush bio2 can get the lock and submit to underlayer disks. Then > flush bio1 updates prev_flush_start after the interrupt or extended stall. > > Then flush bio3 enters in md_flush_request. The start time req_start is > behind prev_flush_start. The flush_bio is not NULL(flush bio2 hasn't > finished). So it can trigger the WARN_ON now. Then it calls INIT_WORK > again. INIT_WORK() will re-initialize the list pointers in the > work_struct, which then can result in a corrupted work list and the > work_struct queued a second time. With the work list corrupted, it can > lead in invalid work items being used and cause a crash in > process_one_work. > > We need to make sure only one flush bio can be handled at one same time. > So add spin lock in md_submit_flush_data to protect prev_flush_start and > flush_bio in an atomic way. > > Reviewed-by: David Jeffery <djeffery@redhat.com> > Signed-off-by: Xiao Ni <xni@redhat.com> > Signed-off-by: Song Liu <songliubraving@fb.com> > [jwang: backport dc5d17a3c39b06aef866afca19245a9cfb533a79 to 4.19] I can not take patches backported to older kernels that "skip" kernel releases. For example, if I take this into 4.19.y, and then someone moves to 5.4 or 5.10, they will hit the same issue. So please provide a backported series for all affected releases, back as far as you want, but never skip releases. I can't take this series, I'll drop it for now and wait for an updated set of patches. thanks, greg k-h
On Thu, Feb 4, 2021 at 4:12 PM Greg KH <gregkh@linuxfoundation.org> wrote: > > On Wed, Feb 03, 2021 at 02:20:22PM +0100, Jack Wang wrote: > > From: Xiao Ni <xni@redhat.com> > > > > One customer reports a crash problem which causes by flush request. It > > triggers a warning before crash. > > > > /* new request after previous flush is completed */ > > if (ktime_after(req_start, mddev->prev_flush_start)) { > > WARN_ON(mddev->flush_bio); > > mddev->flush_bio = bio; > > bio = NULL; > > } > > > > The WARN_ON is triggered. We use spin lock to protect prev_flush_start and > > flush_bio in md_flush_request. But there is no lock protection in > > md_submit_flush_data. It can set flush_bio to NULL first because of > > compiler reordering write instructions. > > > > For example, flush bio1 sets flush bio to NULL first in > > md_submit_flush_data. An interrupt or vmware causing an extended stall > > happen between updating flush_bio and prev_flush_start. Because flush_bio > > is NULL, flush bio2 can get the lock and submit to underlayer disks. Then > > flush bio1 updates prev_flush_start after the interrupt or extended stall. > > > > Then flush bio3 enters in md_flush_request. The start time req_start is > > behind prev_flush_start. The flush_bio is not NULL(flush bio2 hasn't > > finished). So it can trigger the WARN_ON now. Then it calls INIT_WORK > > again. INIT_WORK() will re-initialize the list pointers in the > > work_struct, which then can result in a corrupted work list and the > > work_struct queued a second time. With the work list corrupted, it can > > lead in invalid work items being used and cause a crash in > > process_one_work. > > > > We need to make sure only one flush bio can be handled at one same time. > > So add spin lock in md_submit_flush_data to protect prev_flush_start and > > flush_bio in an atomic way. > > > > Reviewed-by: David Jeffery <djeffery@redhat.com> > > Signed-off-by: Xiao Ni <xni@redhat.com> > > Signed-off-by: Song Liu <songliubraving@fb.com> > > [jwang: backport dc5d17a3c39b06aef866afca19245a9cfb533a79 to 4.19] > > I can not take patches backported to older kernels that "skip" kernel > releases. > > For example, if I take this into 4.19.y, and then someone moves to 5.4 > or 5.10, they will hit the same issue. > > So please provide a backported series for all affected releases, back as > far as you want, but never skip releases. > > I can't take this series, I'll drop it for now and wait for an updated > set of patches. > > thanks, > > greg k-h Hi Greg, Thanks for reply, only this patch should be backported also to 5.4/5.10, this backport can be applied cleanly to stable/linux-5.4.y and stable/linux-5.10.y, I will send the backport for them later today! Thanks! J
diff --git a/drivers/md/md.c b/drivers/md/md.c index 80ca13594c18..09f0d8e70b70 100644 --- a/drivers/md/md.c +++ b/drivers/md/md.c @@ -474,8 +474,10 @@ static void md_submit_flush_data(struct work_struct *ws) * could wait for this and below md_handle_request could wait for those * bios because of suspend check */ + spin_lock_irq(&mddev->lock); mddev->last_flush = mddev->start_flush; mddev->flush_bio = NULL; + spin_unlock_irq(&mddev->lock); wake_up(&mddev->sb_wait); if (bio->bi_iter.bi_size == 0) {