| Message ID | 20200928122613.434820-1-hch@lst.de (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | writeback: don't warn on an unregistered BDI in __mark_inode_dirty | expand |
On Mon 28-09-20 14:26:13, Christoph Hellwig wrote: > BDIs get unregistered during device removal, and this WARN can be > trivially triggered by hot-removing a NVMe device while running fsx > It is otherwise harmless as we still hold a BDI reference, and the > writeback has been shut down already. > > Signed-off-by: Christoph Hellwig <hch@lst.de> > --- > > I have a vague memory someone else sent this patch alredy, but couldn't > find it in my mailing list folder. But given that my current NVMe > tests trigger it easily I'd rather get it fixed ASAP. Did this patch get lost? I don't see it upstream or in Jens' tree. FWIW I agree the warning may result in false positive so I'm OK with removing it. So feel free to add: Reviewed-by: Jan Kara <jack@suse.cz> Honza > > fs/fs-writeback.c | 4 ---- > 1 file changed, 4 deletions(-) > > diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c > index e6005c78bfa93e..acfb55834af23c 100644 > --- a/fs/fs-writeback.c > +++ b/fs/fs-writeback.c > @@ -2321,10 +2321,6 @@ void __mark_inode_dirty(struct inode *inode, int flags) > > wb = locked_inode_to_wb_and_lock_list(inode); > > - WARN((wb->bdi->capabilities & BDI_CAP_WRITEBACK) && > - !test_bit(WB_registered, &wb->state), > - "bdi-%s not registered\n", bdi_dev_name(wb->bdi)); > - > inode->dirtied_when = jiffies; > if (dirtytime) > inode->dirtied_time_when = jiffies; > -- > 2.28.0 >
On Wed, Dec 09, 2020 at 05:35:36PM +0100, Jan Kara wrote: > > I have a vague memory someone else sent this patch alredy, but couldn't > > find it in my mailing list folder. But given that my current NVMe > > tests trigger it easily I'd rather get it fixed ASAP. > > Did this patch get lost? I don't see it upstream or in Jens' tree. FWIW I > agree the warning may result in false positive so I'm OK with removing it. > So feel free to add: > > Reviewed-by: Jan Kara <jack@suse.cz> Yes, it looks like this patch got lost somewhere again.
On Wed 09-12-20 18:47:50, Christoph Hellwig wrote: > On Wed, Dec 09, 2020 at 05:35:36PM +0100, Jan Kara wrote: > > > I have a vague memory someone else sent this patch alredy, but couldn't > > > find it in my mailing list folder. But given that my current NVMe > > > tests trigger it easily I'd rather get it fixed ASAP. > > > > Did this patch get lost? I don't see it upstream or in Jens' tree. FWIW I > > agree the warning may result in false positive so I'm OK with removing it. > > So feel free to add: > > > > Reviewed-by: Jan Kara <jack@suse.cz> > > Yes, it looks like this patch got lost somewhere again. Since Jens didn't reply, I've just picked up the patch to my tree and will push it to Linus tomorrow. Honza
On 12/16/20 3:57 AM, Jan Kara wrote: > On Wed 09-12-20 18:47:50, Christoph Hellwig wrote: >> On Wed, Dec 09, 2020 at 05:35:36PM +0100, Jan Kara wrote: >>>> I have a vague memory someone else sent this patch alredy, but couldn't >>>> find it in my mailing list folder. But given that my current NVMe >>>> tests trigger it easily I'd rather get it fixed ASAP. >>> >>> Did this patch get lost? I don't see it upstream or in Jens' tree. FWIW I >>> agree the warning may result in false positive so I'm OK with removing it. >>> So feel free to add: >>> >>> Reviewed-by: Jan Kara <jack@suse.cz> >> >> Yes, it looks like this patch got lost somewhere again. > > Since Jens didn't reply, I've just picked up the patch to my tree and will > push it to Linus tomorrow. Sorry about that... That sounds fine to me, thanks Jan.
diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c index e6005c78bfa93e..acfb55834af23c 100644 --- a/fs/fs-writeback.c +++ b/fs/fs-writeback.c @@ -2321,10 +2321,6 @@ void __mark_inode_dirty(struct inode *inode, int flags) wb = locked_inode_to_wb_and_lock_list(inode); - WARN((wb->bdi->capabilities & BDI_CAP_WRITEBACK) && - !test_bit(WB_registered, &wb->state), - "bdi-%s not registered\n", bdi_dev_name(wb->bdi)); - inode->dirtied_when = jiffies; if (dirtytime) inode->dirtied_time_when = jiffies;
BDIs get unregistered during device removal, and this WARN can be trivially triggered by hot-removing a NVMe device while running fsx It is otherwise harmless as we still hold a BDI reference, and the writeback has been shut down already. Signed-off-by: Christoph Hellwig <hch@lst.de> --- I have a vague memory someone else sent this patch alredy, but couldn't find it in my mailing list folder. But given that my current NVMe tests trigger it easily I'd rather get it fixed ASAP. fs/fs-writeback.c | 4 ---- 1 file changed, 4 deletions(-)