Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.kernel > #1223505 > unrolled thread

[PATCH 2.6.32 42/62] fixing infinite OPEN loop in 4.0 stateid recovery

Started byWilly Tarreau <w@1wt.eu>
First post2015-09-13 01:30 +0200
Last post2015-09-16 07:40 +0200
Articles 4 — 2 participants

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.


Contents

  [PATCH 2.6.32 42/62] fixing infinite OPEN loop in 4.0 stateid recovery Willy Tarreau <w@1wt.eu> - 2015-09-13 01:30 +0200
    Re: [PATCH 2.6.32 42/62] fixing infinite OPEN loop in 4.0 stateid  recovery "Kornievskaia, Olga" <Olga.Kornievskaia@netapp.com> - 2015-09-15 02:00 +0200
      Re: [PATCH 2.6.32 42/62] fixing infinite OPEN loop in 4.0 stateid recovery Willy Tarreau <w@1wt.eu> - 2015-09-15 07:40 +0200
        Re: [PATCH 2.6.32 42/62] fixing infinite OPEN loop in 4.0 stateid recovery Willy Tarreau <w@1wt.eu> - 2015-09-16 07:40 +0200

#1223505 — [PATCH 2.6.32 42/62] fixing infinite OPEN loop in 4.0 stateid recovery

FromWilly Tarreau <w@1wt.eu>
Date2015-09-13 01:30 +0200
Subject[PATCH 2.6.32 42/62] fixing infinite OPEN loop in 4.0 stateid recovery
Message-ID<q82p3-6WW-5@gated-at.bofh.it>
2.6.32-longterm review patch.  If anyone has any objections, please let me know.

------------------

From: Olga Kornievskaia <kolga@netapp.com>

commit e8d975e73e5fa05f983fbf2723120edcf68e0b38 upstream.

Problem: When an operation like WRITE receives a BAD_STATEID, even though
recovery code clears the RECLAIM_NOGRACE recovery flag before recovering
the open state, because of clearing delegation state for the associated
inode, nfs_inode_find_state_and_recover() gets called and it makes the
same state with RECLAIM_NOGRACE flag again. As a results, when we restart
looking over the open states, we end up in the infinite loop instead of
breaking out in the next test of state flags.

Solution: unset the RECLAIM_NOGRACE set because of
calling of nfs_inode_find_state_and_recover() after returning from calling
recover_open() function.

Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
[bwh: Backported to 3.2: adjust context]
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit ef8500b18fc4bb03286a93b6032d56ec7bcbfd15)

Signed-off-by: Willy Tarreau <w@1wt.eu>
---
 fs/nfs/nfs4state.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c
index 2a7f163..71ee6f6 100644
--- a/fs/nfs/nfs4state.c
+++ b/fs/nfs/nfs4state.c
@@ -929,6 +929,8 @@ restart:
 							__func__);
 				}
 				nfs4_put_open_state(state);
+				clear_bit(NFS4CLNT_RECLAIM_NOGRACE,
+					&state->flags);
 				goto restart;
 			}
 		}
-- 
1.7.12.2.21.g234cd45.dirty



--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [next] | [standalone]


#1224549 — Re: [PATCH 2.6.32 42/62] fixing infinite OPEN loop in 4.0 stateid recovery

From"Kornievskaia, Olga" <Olga.Kornievskaia@netapp.com>
Date2015-09-15 02:00 +0200
SubjectRe: [PATCH 2.6.32 42/62] fixing infinite OPEN loop in 4.0 stateid recovery
Message-ID<q8LPc-4Cp-7@gated-at.bofh.it>
In reply to#1223505
Hi Willy,

I believe the patch introduced another problem and needs to be corrected. 

> On Sep 12, 2015, at 6:56 PM, Willy Tarreau <w@1wt.eu> wrote:
> 
> 2.6.32-longterm review patch.  If anyone has any objections, please let me know.
> 
> ------------------
> 
> From: Olga Kornievskaia <kolga@netapp.com>
> 
> commit e8d975e73e5fa05f983fbf2723120edcf68e0b38 upstream.
> 
> Problem: When an operation like WRITE receives a BAD_STATEID, even though
> recovery code clears the RECLAIM_NOGRACE recovery flag before recovering
> the open state, because of clearing delegation state for the associated
> inode, nfs_inode_find_state_and_recover() gets called and it makes the
> same state with RECLAIM_NOGRACE flag again. As a results, when we restart
> looking over the open states, we end up in the infinite loop instead of
> breaking out in the next test of state flags.
> 
> Solution: unset the RECLAIM_NOGRACE set because of
> calling of nfs_inode_find_state_and_recover() after returning from calling
> recover_open() function.
> 
> Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
> [bwh: Backported to 3.2: adjust context]
> Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
> (cherry picked from commit ef8500b18fc4bb03286a93b6032d56ec7bcbfd15)
> 
> Signed-off-by: Willy Tarreau <w@1wt.eu>
> ---
> fs/nfs/nfs4state.c | 2 ++
> 1 file changed, 2 insertions(+)
> 
> diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c
> index 2a7f163..71ee6f6 100644
> --- a/fs/nfs/nfs4state.c
> +++ b/fs/nfs/nfs4state.c
> @@ -929,6 +929,8 @@ restart:
> 							__func__);
> 				}
> 				nfs4_put_open_state(state);
> +				clear_bit(NFS4CLNT_RECLAIM_NOGRACE,
> +					&state->flags);
> 				goto restart;
> 			}
> 		}
> -- 
> 1.7.12.2.21.g234cd45.dirty
> 
> 
> 

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]


#1224648

FromWilly Tarreau <w@1wt.eu>
Date2015-09-15 07:40 +0200
Message-ID<q8R8d-3V1-13@gated-at.bofh.it>
In reply to#1224549
Hi Olga,

On Mon, Sep 14, 2015 at 11:54:34PM +0000, Kornievskaia, Olga wrote:
> Hi Willy,
> 
> I believe the patch introduced another problem and needs to be corrected. 

Can you be more specific ? What problem, or how to verify which one ?
Should the patch simply be reverted ?

Thanks,
Willy

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]


#1225725

FromWilly Tarreau <w@1wt.eu>
Date2015-09-16 07:40 +0200
Message-ID<q9dBM-3zy-7@gated-at.bofh.it>
In reply to#1224648
Hi Olga,

On Tue, Sep 15, 2015 at 02:36:06PM +0000, Kornievskaia, Olga wrote:
> 
> Hi Willy,
> 
> After checking with the list, I believe the course of action will be to
> correct the patch with the patch below instead of reverting it.

OK but as far as I can tell, mainline is still not fixed regarding this
issue. I can't introduce in a stable branch a fix which is not yet in
mainline. Thus I'll simply remove the patch from this series and will
merge both patches in a future series once your fix reaches mainline.

Note that I picked this fix from 3.2 (commit ef8500b18fc4bb) so my
understanding is that this patch needs to be reverted from 3.2 as well
for the time being ?

Thanks very much for the detailed investigations!

Willy

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [standalone]


Back to top | Article view | linux.kernel


csiph-web