- 09 Apr, 2013 10 commits
-
-
J. Bruce Fields authored
The logic here is better expressed with a switch statement. While we're here, CLOSED stateids (or stateids of an unkown type--which would indicate a server bug) should probably return nfserr_bad_stateid, though this behavior shouldn't affect any non-buggy client. Signed-off-by:J. Bruce Fields <bfields@redhat.com>
-
J. Bruce Fields authored
Make sure the client gives us an adequate backchannel. Signed-off-by: -
J. Bruce Fields authored
Negotiation of the 4.1 session forechannel attributes is a mess. Fix: - Move it all into check_forechannel_attrs instead of spreading it between that, alloc_session, and init_forechannel_attrs. - set a minimum "slotsize" so that our drc memory limits apply even for small maxresponsesize_cached. This also fixes some bugs when slotsize becomes <= 0. Signed-off-by: -
J. Bruce Fields authored
Pass this struct by reference, not by value, and return an error instead of a boolean to allow for future additions. Signed-off-by: -
J. Bruce Fields authored
Don't actually close any opens until we don't need them at all. This means being left with write access when it's not really necessary, but that's better than putting a file that might still have posix locks held on it, as we have been. Reported-by:
Toralf Förster <toralf.foerster@gmx.de> Cc: stable@kernel.org Signed-off-by:
-
J. Bruce Fields authored
In the 4.1 case we're supposed to release lockowners as soon as they're no longer used. It would probably be more efficient to reference count them, but that's slightly fiddly due to the need to have callbacks from locks.c to take into account lock merging and splitting. For most cases just scanning the inode's lock list on unlock for matching locks will be sufficient. Signed-off-by: -
J. Bruce Fields authored
More logic that's unnecessary in the 4.1 case. Signed-off-by: -
J. Bruce Fields authored
The replay_owner will never be used in the sessions case. Signed-off-by: -
J. Bruce Fields authored
Signed-off-by: -
Wei Yongjun authored
memory allocated by kmem_cache_alloc() should be freed using kmem_cache_free(), not kfree(). Signed-off-by:
Wei Yongjun <yongjun_wei@trendmicro.com.cn> Cc: stable@kernel.org Signed-off-by:
-
- 08 Apr, 2013 1 commit
-
-
J. Bruce Fields authored
Closed stateid's are kept around a little while to handle close replays in the 4.0 case. So we stash them in the last-used stateid in the oo_last_closed_stateid field of the open owner. We can free that in encode_seqid_op_tail once the seqid on the open owner is next incremented. But we don't want to do that on the close itself; so we set NFS4_OO_PURGE_CLOSE flag set on the open owner, skip freeing it the first time through encode_seqid_op_tail, then when we see that flag set next time we free it. This is unnecessarily baroque. Instead, just move the logic that increments the seqid out of the xdr code and into the operation code itself. The justification given for the current placement is that we need to wait till the last minute to be sure we know whether the status is a sequence-id-mutating error or not, but examination of the code shows that can't actually happen. Reported-by:
Yanchuan Nian <ycnian@gmail.com> Tested-by:
-
- 04 Apr, 2013 4 commits
-
-
J. Bruce Fields authored
Signed-off-by: -
J. Bruce Fields authored
If nothing else this simplifies the nfs4_state_shutdown_net logic a tad. Signed-off-by: -
J. Bruce Fields authored
Once we've unhashed the delegation, it's only hanging around for the benefit of an oustanding recall, which only needs the encoded filehandle, stateid, and dl_retries counter. No point keeping the file around any longer, or keeping it hashed. This also fixes a race: calls to idr_remove should really be serialized by the caller, but the nfs4_put_delegation call from the callback code isn't taking the state lock. (Better might be to cancel the callback before destroying the delegation, and remove any need for reference counting--but I don't see an easy way to cancel an rpc call.) Signed-off-by: -
J. Bruce Fields authored
Signed-off-by:
-
- 03 Apr, 2013 25 commits
-
-
Alexey Khoroshilov authored
If kmalloc() fails in cache_open(), module cd->owner left locked. The patch adds module_put(cd->owner) on this path. Found by Linux Driver Verification project (linuxtesting.org). Signed-off-by:
Alexey Khoroshilov <khoroshilov@ispras.ru> Signed-off-by:
-
fanchaoting authored
when create /proc/fs/nfs/exports error, we should remove /proc/fs/nfs, if don't do it, it maybe cause Memory leak. Signed-off-by:
fanchaoting <fanchaoting@cn.fujitsu.com> Reviewed-by:
chendt.fnst <chendt.fnst@cn.fujitsu.com> Signed-off-by:
-
fanchaoting authored
we should return error status directly when nfs4_preprocess_stateid_op return error. Signed-off-by:
-
Jeff Layton authored
We only ever traverse the hash chains in the forward direction, so a double pointer list head isn't really necessary. Signed-off-by:
Jeff Layton <jlayton@redhat.com> Signed-off-by:
-
J. Bruce Fields authored
This changes session destruction to be similar to client destruction in that attempts to destroy a session while in use (which should be rare corner cases) result in DELAY. This simplifies things somewhat and helps meet a coming 4.2 requirement. Signed-off-by: -
J. Bruce Fields authored
When a setclientid_confirm or create_session confirms a client after a client reboot, it also destroys any previous state held by that client. The shutdown of that previous state must be careful not to free the client out from under threads processing other requests that refer to the client. This is a particular problem in the NFSv4.1 case when we hold a reference to a session (hence a client) throughout compound processing. The server attempts to handle this by unhashing the client at the time it's destroyed, then delaying the final free to the end. But this still leaves some races in the current code. I believe it's simpler just to fail the attempt to destroy the client by returning NFS4ERR_DELAY. This is a case that should never happen anyway. Signed-off-by: -
J. Bruce Fields authored
The locking here is very fiddly, and there's no reason for us to be setting cstate->session, since this is the only op in the compound. Let's just take the state lock and drop the reference counting. Signed-off-by: -
J. Bruce Fields authored
destroy_session uses the session and client without continuously holding any reference or locks. Put the whole thing under the state lock for now. Signed-off-by: -
J. Bruce Fields authored
Signed-off-by: -
J. Bruce Fields authored
I'm not sure what the check for clientid expiry was meant to do here. The check for a matching session is redundant given the previous check for state: a client without state is, in particular, a client without sessions. Signed-off-by: -
J. Bruce Fields authored
E.g. printk's that just report the return value from an op are uninteresting as we already do that in the main proc_compound loop. Signed-off-by: -
J. Bruce Fields authored
Signed-off-by: -
J. Bruce Fields authored
This should never happen. (Note: the comparable case in setclientid_confirm *can* happen, since updating a client record can result in both confirmed and unconfirmed records with the same clientid.) Signed-off-by: -
ycnian@gmail.com authored
NFS4_OO_PURGE_CLOSE is not handled properly. To avoid memory leak, nfs4 stateid which is pointed by oo_last_closed_stid is freed in nfsd4_close(), but NFS4_OO_PURGE_CLOSE isn't cleared meanwhile. So the stateid released in THIS close procedure may be freed immediately in the coming encoding function. Sorry that Signed-off-by was forgotten in last version. Signed-off-by:
-
Yanchuan Nian authored
lk_rflags is never used anywhere, and rflags is not defined in struct nfsd4_lock. Signed-off-by:
-
J. Bruce Fields authored
Once we drop the lock here there's nothing keeping the client around: the only lock still held is the xpt_lock on this socket, but this socket no longer has any connection with the client so there's no way for other code to know we're still using the client. The solution is simple: all nfsd4_probe_callback does is set a few variables and queue some work, so there's no reason we can't just keep it under the lock. Signed-off-by: -
J. Bruce Fields authored
Dropping the session's reference count after the client's means we leave a window where the session's se_client pointer is NULL. An xpt_user callback that encounters such a session may then crash: [ 303.956011] BUG: unable to handle kernel NULL pointer dereference at 0000000000000318 [ 303.959061] IP: [<ffffffff81481a8e>] _raw_spin_lock+0x1e/0x40 [ 303.959061] PGD 37811067 PUD 3d498067 PMD 0 [ 303.959061] Oops: 0002 [#8] PREEMPT SMP [ 303.959061] Modules linked in: md5 nfsd auth_rpcgss nfs_acl snd_hda_intel snd_hda_codec snd_hwdep snd_pcm snd_page_alloc microcode psmouse snd_timer serio_raw pcspkr evdev snd soundcore i2c_piix4 i2c_core intel_agp intel_gtt processor button nfs lockd sunrpc fscache ata_generic pata_acpi ata_piix uhci_hcd libata btrfs usbcore usb_common crc32c scsi_mod libcrc32c zlib_deflate floppy virtio_balloon virtio_net virtio_pci virtio_blk virtio_ring virtio [ 303.959061] CPU 0 [ 303.959061] Pid: 264, comm: nfsd Tainted: G D 3.8.0-ARCH+ #156 Bochs Bochs [ 303.959061] RIP: 0010:[<ffffffff81481a8e>] [<ffffffff81481a8e>] _raw_spin_lock+0x1e/0x40 [ 303.959061] RSP: 0018:ffff880037877dd8 EFLAGS: 00010202 [ 303.959061] RAX: 0000000000000100 RBX: ffff880037a2b698 RCX: ffff88003d879278 [ 303.959061] RDX: ffff88003d879278 RSI: dead000000100100 RDI: 0000000000000318 [ 303.959061] RBP: ffff880037877dd8 R08: ffff88003c5a0f00 R09: 0000000000000002 [ 303.959061] R10: 0000000000000001 R11: 0000000000000000 R12: 0000000000000000 [ 303.959061] R13: 0000000000000318 R14: ffff880037a2b680 R15: ffff88003c1cbe00 [ 303.959061] FS: 0000000000000000(0000) GS:ffff88003fc00000(0000) knlGS:0000000000000000 [ 303.959061] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b [ 303.959061] CR2: 0000000000000318 CR3: 000000003d49c000 CR4: 00000000000006f0 [ 303.959061] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 [ 303.959061] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 [ 303.959061] Process nfsd (pid: 264, threadinfo ffff880037876000, task ffff88003c1fd0a0) [ 303.959061] Stack: [ 303.959061] ffff880037877e08 ffffffffa03772ec ffff88003d879000 ffff88003d879278 [ 303.959061] ffff88003d879080 0000000000000000 ffff880037877e38 ffffffffa0222a1f [ 303.959061] 0000000000107ac0 ffff88003c22e000 ffff88003d879000 ffff88003c1cbe00 [ 303.959061] Call Trace: [ 303.959061] [<ffffffffa03772ec>] nfsd4_conn_lost+0x3c/0xa0 [nfsd] [ 303.959061] [<ffffffffa0222a1f>] svc_delete_xprt+0x10f/0x180 [sunrpc] [ 303.959061] [<ffffffffa0223d96>] svc_recv+0xe6/0x580 [sunrpc] [ 303.959061] [<ffffffffa03587c5>] nfsd+0xb5/0x140 [nfsd] [ 303.959061] [<ffffffffa0358710>] ? nfsd_destroy+0x90/0x90 [nfsd] [ 303.959061] [<ffffffff8107ae00>] kthread+0xc0/0xd0 [ 303.959061] [<ffffffff81010000>] ? perf_trace_xen_mmu_set_pte_at+0x50/0x100 [ 303.959061] [<ffffffff8107ad40>] ? kthread_freezable_should_stop+0x70/0x70 [ 303.959061] [<ffffffff814898ec>] ret_from_fork+0x7c/0xb0 [ 303.959061] [<ffffffff8107ad40>] ? kthread_freezable_should_stop+0x70/0x70 [ 303.959061] Code: ff ff 5d c3 90 90 90 90 90 90 90 90 90 0f 1f 44 00 00 55 65 48 8b 04 25 f0 c6 00 00 48 89 e5 83 80 44 e0 ff ff 01 b8 00 01 00 00 <3e> 66 0f c1 07 0f b6 d4 38 c2 74 0f 66 0f 1f 44 00 00 f3 90 0f [ 303.959061] RIP [<ffffffff81481a8e>] _raw_spin_lock+0x1e/0x40 [ 303.959061] RSP <ffff880037877dd8> [ 303.959061] CR2: 0000000000000318 [ 304.001218] ---[ end trace 2d809cd4a7931f5a ]--- [ 304.001903] note: nfsd[264] exited with preempt_count 2 Reported-by:
Bryan Schumaker <bjschuma@netapp.com> Signed-off-by:
-
J. Bruce Fields authored
If a client sets an owner (or group_owner or acl) attribute on open for create, and the mapping of that owner to an id fails, then we return BAD_OWNER. But BAD_OWNER is a seqid-mutating error, so we can't shortcut the open processing that case: we have to at least look up the owner so we can find the seqid to bump. Signed-off-by: -
J. Bruce Fields authored
This BUG_ON just crashes the thread a little earlier than it would otherwise--it doesn't seem useful. Signed-off-by: -
Jeff Layton authored
We've now increased the size of the duplicate reply cache by quite a bit, but the number of hash buckets has not changed. So, we've gone from an average hash chain length of 16 in the old code to 4096 when the cache is its largest. Change the code to scale out the number of buckets with the max size of the cache. At the same time, we also need to fix the hash function since the existing one isn't really suitable when there are more than 256 buckets. Move instead to use the stock hash_32 function for this. Testing on a machine that had 2048 buckets showed that this gave a smaller longest:average ratio than the existing hash function: The formula here is longest hash bucket searched divided by average number of entries per bucket at the time that we saw that longest bucket: old hash: 68/(39258/2048) == 3.547404 hash_32: 45/(33773/2048) == 2.728807 Signed-off-by: -
Jeff Layton authored
The typical case with the DRC is a cache miss, so if we keep track of the max number of entries that we've ever walked over in a search, then we should have a reasonable estimate of the longest hash chain that we've ever seen. With that, we'll also keep track of the total size of the cache when we see the longest chain. In the case of a tie, we prefer to track the smallest total cache size in order to properly gauge the worst-case ratio of max vs. avg chain length. Signed-off-by:
-
Jeff Layton authored
For presenting statistics relating to duplicate reply cache. Signed-off-by:
-
Jeff Layton authored
Signed-off-by:
-
Jeff Layton authored
Break out the function that compares the rqstp and checksum against a reply cache entry. While we're at it, track the efficacy of the checksum over the NFS data by tracking the cases where we would have incorrectly matched a DRC entry if we had not tracked it or the length. Signed-off-by:
-
Jeff Layton authored
The most common case is to do a search of the cache, followed by an insert. In the case where we have to allocate an entry off the slab, then we end up having to redo the search, which is wasteful. Better optimize the code for the common case by eliminating the initial search of the cache and always preallocating an entry. In the case of a cache hit, we'll end up just freeing that entry but that's preferable to an extra search. Signed-off-by:
-
