Messages in this thread | | | Subject | Re: [PATCH 4.19 082/131] nvme: fix possible deadlock when I/O is blocked | From | Sagi Grimberg <> | Date | Thu, 2 Jul 2020 15:15:48 -0700 |
| |
> Hi! > >> From: Sagi Grimberg <sagi@grimberg.me> >> >> [ Upstream commit 3b4b19721ec652ad2c4fe51dfbe5124212b5f581 ] >> >> Revert fab7772bfbcf ("nvme-multipath: revalidate nvme_ns_head gendisk >> in nvme_validate_ns") >> >> When adding a new namespace to the head disk (via nvme_mpath_set_live) >> we will see partition scan which triggers I/O on the mpath device node. >> This process will usually be triggered from the scan_work which holds >> the scan_lock. If I/O blocks (if we got ana change currently have only >> available paths but none are accessible) this can deadlock on the head >> disk bd_mutex as both partition scan I/O takes it, and head disk revalidation >> takes it to check for resize (also triggered from scan_work on a different >> path). See trace [1]. >> >> The mpath disk revalidation was originally added to detect online disk >> size change, but this is no longer needed since commit cb224c3af4df >> ("nvme: Convert to use set_capacity_revalidate_and_notify") which already >> updates resize info without unnecessarily revalidating the disk (the > > Unfortunately, v4.19-stable does not contain cb224c3af4df. According > to changelog, it seems it should be cherry-picked?
You are absolutely right,
The reference commit is a part of the series: 78317c5d58e6 ("scsi: Convert to use set_capacity_revalidate_and_notify") cb224c3af4df ("nvme: Convert to use set_capacity_revalidate_and_notify") 3cbc28bb902b ("xen-blkfront.c: Convert to use set_capacity_revalidate_and_notify") 662155e2898d ("virtio_blk.c: Convert to use set_capacity_revalidate_and_notify") e598a72faeb5 ("block/genhd: Notify udev about capacity change")
It would be cool if they are cherry picked, although they don't qualify as stable patches per se...
| |