Message ID | 1b7bb28b-939e-c111-9bb0-5091ab1cdcf1@oracle.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
On Wed, Jun 22, 2016 at 6:18 AM, Anand Jain <anand.jain@oracle.com> wrote: > > > Thanks for the review Chris. > > On 06/21/2016 09:00 PM, Chris Mason wrote: >> On 06/21/2016 06:24 AM, Anand Jain wrote: >>> From: Anand Jain <Anand.Jain@oracle.com> >>> >>> Further to the commit >>> bc178622d40d87e75abc131007342429c9b03351 >>> btrfs: use rcu_barrier() to wait for bdev puts at unmount >>> >>> This patch implements a method to time wait on the __free_device() >>> which actually does the bdev put. This is needed as the user space >>> running 'btrfs fi show -d' immediately after the replace and >>> unmount, is still reading older information from the device. >> >> Thanks for working on this Anand. Since it looks like blkdev_put can >> deadlock against us, can we please switch to making sure we fully >> flush >> the outstanding IO? It's probably enough to do a sync_blockdev() >> call >> before we allow the unmount to finish, but we can toss in an >> invalidate_bdev for good measure. > > > ------------ > # git diff > diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c > index 604daf315669..e0ad29d6fe9a 100644 > --- a/fs/btrfs/volumes.c > +++ b/fs/btrfs/volumes.c > @@ -870,6 +870,11 @@ static void btrfs_close_one_device(struct > btrfs_device *device) > if (device->missing) > fs_devices->missing_devices--; > > + if (device->bdev && device->writeable) { > + sync_blockdev(device->bdev); > + invalidate_bdev(device->bdev); > + } > + > new_device = btrfs_alloc_device(NULL, &device->devid, > device->uuid); > BUG_ON(IS_ERR(new_device)); /* -ENOMEM */ > ----------- > > > However, theoretically still there might be a problem - at the end of > unmount, if the device exclusive open is not actually closed, then > there might be a race with another program which is trying to open > the device in exclusive mode. Like for eg: > unmount /btrfs; fsck /dev/X > and here fsck might fail to open the device if it wins the race. This true, but at least we know he'll have up to date buffers if he does manage to open the device. With the generic code, the blkdev_put happens after the super is gone, so I'm not sure we can completely fix this from inside our callback. -chris -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 06/23/2016 05:47 AM, Chris Mason wrote: > > > On Wed, Jun 22, 2016 at 6:18 AM, Anand Jain <anand.jain@oracle.com> wrote: >> >> >> Thanks for the review Chris. >> >> On 06/21/2016 09:00 PM, Chris Mason wrote: >>> On 06/21/2016 06:24 AM, Anand Jain wrote: >>>> From: Anand Jain <Anand.Jain@oracle.com> >>>> >>>> Further to the commit >>>> bc178622d40d87e75abc131007342429c9b03351 >>>> btrfs: use rcu_barrier() to wait for bdev puts at unmount >>>> >>>> This patch implements a method to time wait on the __free_device() >>>> which actually does the bdev put. This is needed as the user space >>>> running 'btrfs fi show -d' immediately after the replace and >>>> unmount, is still reading older information from the device. >>> >>> Thanks for working on this Anand. Since it looks like blkdev_put can >>> deadlock against us, can we please switch to making sure we fully flush >>> the outstanding IO? It's probably enough to do a sync_blockdev() call >>> before we allow the unmount to finish, but we can toss in an >>> invalidate_bdev for good measure. >> >> >> ------------ >> # git diff >> diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c >> index 604daf315669..e0ad29d6fe9a 100644 >> --- a/fs/btrfs/volumes.c >> +++ b/fs/btrfs/volumes.c >> @@ -870,6 +870,11 @@ static void btrfs_close_one_device(struct >> btrfs_device *device) >> if (device->missing) >> fs_devices->missing_devices--; >> >> + if (device->bdev && device->writeable) { >> + sync_blockdev(device->bdev); >> + invalidate_bdev(device->bdev); >> + } >> + >> new_device = btrfs_alloc_device(NULL, &device->devid, >> device->uuid); >> BUG_ON(IS_ERR(new_device)); /* -ENOMEM */ >> ----------- >> >> >> However, theoretically still there might be a problem - at the end of >> unmount, if the device exclusive open is not actually closed, then >> there might be a race with another program which is trying to open >> the device in exclusive mode. Like for eg: >> unmount /btrfs; fsck /dev/X >> and here fsck might fail to open the device if it wins the race. > > This true, but at least we know he'll have up to date buffers if he does > manage to open the device. > > With the generic code, the blkdev_put happens after the super is gone, > so I'm not sure we can completely fix this from inside our callback. Makes sense, sent out v3 with title (btrfs: make sure device is synced before return) Also sent out RFC patch btrfs: make sure device is synced before return where I have tried not to background blkdev_put(), which seems to be better, it works fine per fstests. Thanks, Anand > -chris > > > -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 604daf315669..e0ad29d6fe9a 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -870,6 +870,11 @@ static void btrfs_close_one_device(struct btrfs_device *device) if (device->missing) fs_devices->missing_devices--; + if (device->bdev && device->writeable) { + sync_blockdev(device->bdev); + invalidate_bdev(device->bdev); + } + new_device = btrfs_alloc_device(NULL, &device->devid, device->uuid);