On Thu, Jun 04, 2015 at 08:06:57AM +0800, Kinglong Mee wrote:
> On 6/3/2015 11:03 PM, J. Bruce Fields wrote:
> > On Tue, Jun 02, 2015 at 06:59:19PM +0800, Kinglong Mee wrote:
> >> nfsd enters a infinite loop and print message per 10 seconds,
> >>
> >> May 31 18:33:52 test-server kernel: Error sending entire callback!
> >> May 31 18:34:01 test-server kernel: Error sending entire callback!
> >>
> >> It is caused by a cb_layoutreturn got error -10008 (NFS4ERR_DELAY),
> >> and then, the client crash, nfsd enter the infinite loop.
> >>
> >> bc_sendto --> call_timeout --> nfsd4_cb_done --> nfsd4_cb_layout_done
> >> with error -10008 --> rpc_delay(task, HZ/100) --> bc_sendto ...
> >
> > How are you reproducing this?
>
> Yes,
>
> I test it by xfstests 074 with nfs client's kdump is on,
> set CONFIG_DEFAULT_HUNG_TASK_TIMEOUT, and client's blkmapd is down.
>
> 1. nfs client's write operation will get the layout of file,
> and then the getdeviceinfo,
> 2. but layout segment is not record by client for blkmapd is down,
> 3. client write data by sending WRITE to server,
> 4. nfs server will recall the layout of the file before WRITE,
> 5. network error cause the client reset the session and return NFS4ERR_DELAY,
> 6. so client's WRITE operation is waiting the reply,
> if the task hang 120s, client will crash.
> 7. so that, the next bc_sendto will fail with TIMEOUT,
> and cb_status is NFS4ERR_DELAY.
OK, that's complicated. Sounds like you're giving this code a
workout--thanks. I'll add the reproducer to the changelog....
--b.
>
> thanks,
> Kinglong Mee
>
> >
> > --b.
> >
> >>
> >> Signed-off-by: Kinglong Mee <kinglongmee@xxxxxxxxx>
> >> ---
> >> fs/nfsd/nfs4callback.c | 1 +
> >> 1 file changed, 1 insertion(+)
> >>
> >> diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c
> >> index 5694cfb..8b1ac8d 100644
> >> --- a/fs/nfsd/nfs4callback.c
> >> +++ b/fs/nfsd/nfs4callback.c
> >> @@ -875,6 +875,7 @@ static void nfsd4_cb_prepare(struct rpc_task *task, void *calldata)
> >> u32 minorversion = clp->cl_minorversion;
> >>
> >> cb->cb_minorversion = minorversion;
> >> + cb->cb_status = 0;
> >> if (minorversion) {
> >> if (!nfsd41_cb_get_slot(clp, task))
> >> return;
> >> --
> >> 2.4.2
> >
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
