There are no more external users of nfs4_state_mark_reclaim_nograce() or
nfs4_state_mark_reclaim_reboot(), so mark them as static.
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
We want SEQUENCE status bits to be handled by the state manager in order
to avoid threading issues.
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
nfs4_schedule_state_recovery() should only be used when we need to force
the state manager to check the lease. If we just want to start the
state manager in order to handle a state recovery situation, we should be
using nfs4_schedule_state_manager().
This patch fixes the abuses of nfs4_schedule_state_recovery() by replacing
its use with a set of helper functions that do the right thing.
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Fix bug where we currently retry the EXCHANGEID call again, eventhough
we already have a valid clientid. Instead, delay and retry the CREATE_SESSION
call.
Signed-off-by: Ricardo Labiaga <Ricardo.Labiaga@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
The problem was use of an int32, which when converted to a uint64
is sign extended resulting in a fileid that doesn't fit in 32 bits
even though the intent of the function is to fit the fileid into
32 bits.
Signed-off-by: Frank Filz <ffilzlnx@us.ibm.com>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
[Trond: Added an include for compat.h]
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
add kmalloc return value check in decode_and_add_ds
Signed-off-by: Stanislav Fomichev <kernel@fomichev.me>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
I've been adding in more artificial delays in the NFSv4 commit and close
codepaths to uncover races. The kernel I'm testing has the patch to
close the race in __rpc_wait_for_completion_task that's in Trond's
cthon2011 branch. The reproducer I've been using does this in a loop:
mkdir("DIR");
fd = open("DIR/FILE", O_WRONLY|O_CREAT|O_EXCL, 0644);
write(fd, "abcdefg", 7);
close(fd);
unlink("DIR/FILE");
rmdir("DIR");
The above reproducer shouldn't result in any silly-renaming. However,
when I add a "msleep(100)" just after the nfs_commit_clear_lock call in
nfs_commit_release, I can almost always force one to occur. If I can
force it to occur with that, then it can happen without that delay
given the right timing.
nfs_commit_inode waits for the NFS_INO_COMMIT bit to clear when called
with FLUSH_SYNC set. nfs_commit_rpcsetup on the other hand does not wait
for the task to complete before putting its reference to it, so the last
reference get put in rpc_release task and gets queued to a workqueue.
In this situation, the last open context reference may be put by the
COMMIT release instead of the close() syscall. The close() syscall
returns too quickly and the unlink runs while the d_count is still
high since the COMMIT release hasn't put its dentry reference yet.
Fix this by having rpc_commit_rpcsetup wait for the RPC call to complete
before putting the task reference when FLUSH_SYNC is set. With this, the
last reference is put by the process that's initiating the FLUSH_SYNC
commit and the race is closed.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Although they run as rpciod background tasks, under normal operation
(i.e. no SIGKILL), functions like nfs_sillyrename(), nfs4_proc_unlck()
and nfs4_do_close() want to be fully synchronous. This means that when we
exit, we want all references to the rpc_task to be gone, and we want
any dentry references etc. held by that task to be released.
For this reason these functions call __rpc_wait_for_completion_task(),
followed by rpc_put_task() in the expectation that the latter will be
releasing the last reference to the rpc_task, and thus ensuring that the
callback_ops->rpc_release() has been called synchronously.
This patch fixes a race which exists due to the fact that
rpciod calls rpc_complete_task() (in order to wake up the callers of
__rpc_wait_for_completion_task()) and then subsequently calls
rpc_put_task() without ensuring that these two steps are done atomically.
In order to avoid adding new spin locks, the patch uses the existing
waitqueue spin lock to order the rpc_task reference count releases between
the waiting process and rpciod.
The common case where nobody is waiting for completion is optimised for by
checking if the RPC_TASK_ASYNC flag is cleared and/or if the rpc_task
reference count is 1: in those cases we drop trying to grab the spin lock,
and immediately free up the rpc_task.
Those few processes that need to put the rpc_task from inside an
asynchronous context and that do not care about ordering are given a new
helper: rpc_put_task_async().
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
The "bad_page()" page allocator sanity check was reported recently (call
chain as follows):
bad_page+0x69/0x91
free_hot_cold_page+0x81/0x144
skb_release_data+0x5f/0x98
__kfree_skb+0x11/0x1a
tcp_ack+0x6a3/0x1868
tcp_rcv_established+0x7a6/0x8b9
tcp_v4_do_rcv+0x2a/0x2fa
tcp_v4_rcv+0x9a2/0x9f6
do_timer+0x2df/0x52c
ip_local_deliver+0x19d/0x263
ip_rcv+0x539/0x57c
netif_receive_skb+0x470/0x49f
:virtio_net:virtnet_poll+0x46b/0x5c5
net_rx_action+0xac/0x1b3
__do_softirq+0x89/0x133
call_softirq+0x1c/0x28
do_softirq+0x2c/0x7d
do_IRQ+0xec/0xf5
default_idle+0x0/0x50
ret_from_intr+0x0/0xa
default_idle+0x29/0x50
cpu_idle+0x95/0xb8
start_kernel+0x220/0x225
_sinittext+0x22f/0x236
It occurs because an skb with a fraglist was freed from the tcp
retransmit queue when it was acked, but a page on that fraglist had
PG_Slab set (indicating it was allocated from the Slab allocator (which
means the free path above can't safely free it via put_page.
We tracked this back to an nfsv4 setacl operation, in which the nfs code
attempted to fill convert the passed in buffer to an array of pages in
__nfs4_proc_set_acl, which gets used by the skb->frags list in
xs_sendpages. __nfs4_proc_set_acl just converts each page in the buffer
to a page struct via virt_to_page, but the vfs allocates the buffer via
kmalloc, meaning the PG_slab bit is set. We can't create a buffer with
kmalloc and free it later in the tcp ack path with put_page, so we need
to either:
1) ensure that when we create the list of pages, no page struct has
PG_Slab set
or
2) not use a page list to send this data
Given that these buffers can be multiple pages and arbitrarily sized, I
think (1) is the right way to go. I've written the below patch to
allocate a page from the buddy allocator directly and copy the data over
to it. This ensures that we have a put_page free-able page for every
entry that winds up on an skb frag list, so it can be safely freed when
the frame is acked. We do a put page on each entry after the
rpc_call_sync call so as to drop our own reference count to the page,
leaving only the ref count taken by tcp_sendpages. This way the data
will be properly freed when the ack comes in
Successfully tested by myself to solve the above oops.
Note, as this is the result of a setacl operation that exceeded a page
of data, I think this amounts to a local DOS triggerable by an
uprivlidged user, so I'm CCing security on this as well.
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
CC: Trond Myklebust <Trond.Myklebust@netapp.com>
CC: security@kernel.org
CC: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
On recent 2.6.38-rc kernels, connectathon basic test 6 fails on
NFSv4 mounts of OpenSolaris with something like:
> ./test6: readdir
> ./test6: (/mnt/klimt/matisse.test) didn't read expected 'file.12' dir entry, pass 0
> ./test6: (/mnt/klimt/matisse.test) didn't read expected 'file.82' dir entry, pass 0
> ./test6: (/mnt/klimt/matisse.test) didn't read expected 'file.164' dir entry, pass 0
> ./test6: (/mnt/klimt/matisse.test) Test failed with 3 errors
> basic tests failed
> Tests failed, leaving /mnt/klimt mounted
> [cel@matisse cthon04]$
I narrowed the problem down to nfs4_decode_dirent() reporting that the
decode buffer had overflowed while decoding the entries for those
missing files.
verify_attr_len() assumes both it's pointer arguments reside on the
same page. When these arguments point to locations on two different
pages, verify_attr_len() can report false errors. This can happen now
that a large NFSv4 readdir result can span pages.
We have reasonably good checking in nfs4_decode_dirent() anyway, so
it should be safe to simply remove the extra checking.
At a guess, this was introduced by commit 6650239a, "NFS: Don't use
vm_map_ram() in readdir".
Cc: stable@kernel.org [2.6.37]
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Make the decoding of NFSv4 directory entries slightly more efficient
by:
1. Avoiding unnecessary byte swapping when checking XDR booleans,
and
2. Not bumping "p" when its value will be immediately replaced by
xdr_inline_decode()
This commit makes nfs4_decode_dirent() consistent with similar logic
in the other two decode_dirent() functions.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
There is no reason to be freeing the delegation cred in the rcu callback,
and doing so is resulting in a lockdep complaint that rpc_credcache_lock
is being called from both softirq and non-softirq contexts.
Reported-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Cc: stable@kernel.org
As stated in section 2.4 of RFC 5661, subsequent instances of the client need
to present the same co_ownerid. Concatinate the client's IP dot address,
host name, and the rpc_auth pseudoflavor to form the co_ownerid.
Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
If the call to nfs_wcc_update_inode() results in an attribute update, we
need to ensure that the inode's attr_gencount gets bumped too, otherwise
we are not protected against races with other GETATTR calls.
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
What we really want to know is the ref count.
Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Always assign the cb_process_state nfs_client pointer so a processing error
in cb_sequence after the nfs_client is found and referenced returns
a non-NULL cb_process_state nfs_client and the matching nfs_put_client in
nfs4_callback_compound dereferences the client.
Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
The information required to find the nfs_client cooresponding to the incoming
back channel request is contained in the NFS layer. Perform minimal checking
in the RPC layer pg_authenticate method, and push more detailed checking into
the NFS layer where the nfs_client can be found.
Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
nfsacl_encode() allocates memory in certain cases. This of course
is not guaranteed to work.
Since commit 9f06c719 "SUNRPC: New xdr_streams XDR encoder API", the
kernel's XDR encoders can't return a result indicating possibly a
failure, so a memory allocation failure in nfsacl_encode() has become
fatal (ie, the XDR code Oopses) in some cases.
However, the allocated memory is a tiny fixed amount, on the order
of 40-50 bytes. We can easily use a stack-allocated buffer for
this, with only a wee bit of nose-holding.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Nick Piggin reports:
> I'm getting use after frees in aio code in NFS
>
> [ 2703.396766] Call Trace:
> [ 2703.396858] [<ffffffff8100b057>] ? native_sched_clock+0x27/0x80
> [ 2703.396959] [<ffffffff8108509e>] ? put_lock_stats+0xe/0x40
> [ 2703.397058] [<ffffffff81088348>] ? lock_release_holdtime+0xa8/0x140
> [ 2703.397159] [<ffffffff8108a2a5>] lock_acquire+0x95/0x1b0
> [ 2703.397260] [<ffffffff811627db>] ? aio_put_req+0x2b/0x60
> [ 2703.397361] [<ffffffff81039701>] ? get_parent_ip+0x11/0x50
> [ 2703.397464] [<ffffffff81612a31>] _raw_spin_lock_irq+0x41/0x80
> [ 2703.397564] [<ffffffff811627db>] ? aio_put_req+0x2b/0x60
> [ 2703.397662] [<ffffffff811627db>] aio_put_req+0x2b/0x60
> [ 2703.397761] [<ffffffff811647fe>] do_io_submit+0x2be/0x7c0
> [ 2703.397895] [<ffffffff81164d0b>] sys_io_submit+0xb/0x10
> [ 2703.397995] [<ffffffff8100307b>] system_call_fastpath+0x16/0x1b
>
> Adding some tracing, it is due to nfs completing the request then
> returning something other than -EIOCBQUEUED, so aio.c
> also completes the request.
To address this, prevent the NFS direct I/O engine from completing
async iocbs when the forward path returns an error without starting
any I/O.
This fix appears to survive ^C during both "xfstest no. 208" and "fsx
-Z."
It's likely this bug has existed for a very long while, as we are seeing
very similar symptoms in OEL 5. Copying stable.
Cc: Stable <stable@kernel.org>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
On Mon, 17 Jan 2011, Mi Jinlong wrote:
>
>
> Jesper Juhl:
> > strrchr() can return NULL if nothing is found. If this happens we'll
> > dereference a NULL pointer in
> > fs/nfs/nfs4filelayoutdev.c::decode_and_add_ds().
> >
> > I tried to find some other code that guarantees that this can never
> > happen but I was unsuccessful. So, unless someone else can point to some
> > code that ensures this can never be a problem, I believe this patch is
> > needed.
> >
> > While I was changing this code I also noticed that all the dprintk()
> > statements, except one, start with "%s:". The one missing the ":" I added
> > it to.
>
> Maybe another one also should be changed at decode_and_add_ds() at line 243:
>
> 243 printk("%s Decoded address and port %s\n", __func__, buf);
>
Missed that one. Thanks.
Signed-off-by: Jesper Juhl <jj@chaosbits.net>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Cleanup of the allocated list entries should not call
put_nfs_open_context() on each entry, as the context will
always be NULL, causing an oops.
Signed-off-by: Fred Isaman <iisaman@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Unexport do_add_mount() and make ->d_automount() return the vfsmount to be
added rather than calling do_add_mount() itself. follow_automount() will then
do the addition.
This slightly complicates things as ->d_automount() normally wants to add the
new vfsmount to an expiration list and start an expiration timer. The problem
with that is that the vfsmount will be deleted if it has a refcount of 1 and
the timer will not repeat if the expiration list is empty.
To this end, we require the vfsmount to be returned from d_automount() with a
refcount of (at least) 2. One of these refs will be dropped unconditionally.
In addition, follow_automount() must get a 3rd ref around the call to
do_add_mount() lest it eat a ref and return an error, leaving the mount we
have open to being expired as we would otherwise have only 1 ref on it.
d_automount() should also add the the vfsmount to the expiration list (by
calling mnt_set_expiry()) and start the expiration timer before returning, if
this mechanism is to be used. The vfsmount will be unlinked from the
expiration list by follow_automount() if do_add_mount() fails.
This patch also fixes the call to do_add_mount() for AFS to propagate the mount
flags from the parent vfsmount.
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Make NFS use the new d_automount() dentry operation rather than abusing
follow_link() on directories.
Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Acked-by: Ian Kent <raven@themaw.net>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Add a dentry op (d_manage) to permit a filesystem to hold a process and make it
sleep when it tries to transit away from one of that filesystem's directories
during a pathwalk. The operation is keyed off a new dentry flag
(DCACHE_MANAGE_TRANSIT).
The filesystem is allowed to be selective about which processes it holds and
which it permits to continue on or prohibits from transiting from each flagged
directory. This will allow autofs to hold up client processes whilst letting
its userspace daemon through to maintain the directory or the stuff behind it
or mounted upon it.
The ->d_manage() dentry operation:
int (*d_manage)(struct path *path, bool mounting_here);
takes a pointer to the directory about to be transited away from and a flag
indicating whether the transit is undertaken by do_add_mount() or
do_move_mount() skipping through a pile of filesystems mounted on a mountpoint.
It should return 0 if successful and to let the process continue on its way;
-EISDIR to prohibit the caller from skipping to overmounted filesystems or
automounting, and to use this directory; or some other error code to return to
the user.
->d_manage() is called with namespace_sem writelocked if mounting_here is true
and no other locks held, so it may sleep. However, if mounting_here is true,
it may not initiate or wait for a mount or unmount upon the parameter
directory, even if the act is actually performed by userspace.
Within fs/namei.c, follow_managed() is extended to check with d_manage() first
on each managed directory, before transiting away from it or attempting to
automount upon it.
follow_down() is renamed follow_down_one() and should only be used where the
filesystem deliberately intends to avoid management steps (e.g. autofs).
A new follow_down() is added that incorporates the loop done by all other
callers of follow_down() (do_add/move_mount(), autofs and NFSD; whilst AFS, NFS
and CIFS do use it, their use is removed by converting them to use
d_automount()). The new follow_down() calls d_manage() as appropriate. It
also takes an extra parameter to indicate if it is being called from mount code
(with namespace_sem writelocked) which it passes to d_manage(). follow_down()
ignores automount points so that it can be used to mount on them.
__follow_mount_rcu() is made to abort rcu-walk mode if it hits a directory with
DCACHE_MANAGE_TRANSIT set on the basis that we're probably going to have to
sleep. It would be possible to enter d_manage() in rcu-walk mode too, and have
that determine whether to abort or not itself. That would allow the autofs
daemon to continue on in rcu-walk mode.
Note that DCACHE_MANAGE_TRANSIT on a directory should be cleared when it isn't
required as every tranist from that directory will cause d_manage() to be
invoked. It can always be set again when necessary.
==========================
WHAT THIS MEANS FOR AUTOFS
==========================
Autofs currently uses the lookup() inode op and the d_revalidate() dentry op to
trigger the automounting of indirect mounts, and both of these can be called
with i_mutex held.
autofs knows that the i_mutex will be held by the caller in lookup(), and so
can drop it before invoking the daemon - but this isn't so for d_revalidate(),
since the lock is only held on _some_ of the code paths that call it. This
means that autofs can't risk dropping i_mutex from its d_revalidate() function
before it calls the daemon.
The bug could manifest itself as, for example, a process that's trying to
validate an automount dentry that gets made to wait because that dentry is
expired and needs cleaning up:
mkdir S ffffffff8014e05a 0 32580 24956
Call Trace:
[<ffffffff885371fd>] :autofs4:autofs4_wait+0x674/0x897
[<ffffffff80127f7d>] avc_has_perm+0x46/0x58
[<ffffffff8009fdcf>] autoremove_wake_function+0x0/0x2e
[<ffffffff88537be6>] :autofs4:autofs4_expire_wait+0x41/0x6b
[<ffffffff88535cfc>] :autofs4:autofs4_revalidate+0x91/0x149
[<ffffffff80036d96>] __lookup_hash+0xa0/0x12f
[<ffffffff80057a2f>] lookup_create+0x46/0x80
[<ffffffff800e6e31>] sys_mkdirat+0x56/0xe4
versus the automount daemon which wants to remove that dentry, but can't
because the normal process is holding the i_mutex lock:
automount D ffffffff8014e05a 0 32581 1 32561
Call Trace:
[<ffffffff80063c3f>] __mutex_lock_slowpath+0x60/0x9b
[<ffffffff8000ccf1>] do_path_lookup+0x2ca/0x2f1
[<ffffffff80063c89>] .text.lock.mutex+0xf/0x14
[<ffffffff800e6d55>] do_rmdir+0x77/0xde
[<ffffffff8005d229>] tracesys+0x71/0xe0
[<ffffffff8005d28d>] tracesys+0xd5/0xe0
which means that the system is deadlocked.
This patch allows autofs to hold up normal processes whilst the daemon goes
ahead and does things to the dentry tree behind the automouter point without
risking a deadlock as almost no locks are held in d_manage() and none in
d_automount().
Signed-off-by: David Howells <dhowells@redhat.com>
Was-Acked-by: Ian Kent <raven@themaw.net>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* 'vfs-scale-working' of git://git.kernel.org/pub/scm/linux/kernel/git/npiggin/linux-npiggin:
fs: fix do_last error case when need_reval_dot
nfs: add missing rcu-walk check
fs: hlist UP debug fixup
fs: fix dropping of rcu-walk from force_reval_path
fs: force_reval_path drop rcu-walk before d_invalidate
fs: small rcu-walk documentation fixes
Fixed up trivial conflicts in Documentation/filesystems/porting
Commit c0204fd2b8 (NFS: Clean up
nfs4_proc_create()) broke NFSv3 exclusive open by removing the code
that passes the O_EXCL flag down to nfs3_proc_create(). This patch
reverts that offending hunk from the original commit.
Reported-by: Nick Bowler <nbowler@elliptictech.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Cc: stable@kernel.org [2.6.37]
Tested-by: Nick Bowler <nbowler@elliptictech.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* 'nfs-for-2.6.38' of git://git.linux-nfs.org/projects/trondmy/nfs-2.6: (89 commits)
NFS fix the setting of exchange id flag
NFS: Don't use vm_map_ram() in readdir
NFSv4: Ensure continued open and lockowner name uniqueness
NFS: Move cl_delegations to the nfs_server struct
NFS: Introduce nfs_detach_delegations()
NFS: Move cl_state_owners and related fields to the nfs_server struct
NFS: Allow walking nfs_client.cl_superblocks list outside client.c
pnfs: layout roc code
pnfs: update nfs4_callback_recallany to handle layouts
pnfs: add CB_LAYOUTRECALL handling
pnfs: CB_LAYOUTRECALL xdr code
pnfs: change lo refcounting to atomic_t
pnfs: check that partial LAYOUTGET return is ignored
pnfs: add layout to client list before sending rpc
pnfs: serialize LAYOUTGET(openstateid)
pnfs: layoutget rpc code cleanup
pnfs: change how lsegs are removed from layout list
pnfs: change layout state seqlock to a spinlock
pnfs: add prefix to struct pnfs_layout_hdr fields
pnfs: add prefix to struct pnfs_layout_segment fields
...
Indicate support for referrals. Do not set any PNFS roles. Check the flags
returned by the server for validity. Do not use exchange flags from an old
client ID instance when recovering a client ID.
Update the EXCHID4_FLAG_XXX set to RFC 5661.
Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
vm_map_ram() is not available on NOMMU platforms, and causes trouble
on incoherrent architectures such as ARM when we access the page data
through both the direct and the virtual mapping.
The alternative is to use the direct mapping to access page data
for the case when we are not crossing a page boundary, but to copy
the data into a linear scratch buffer when we are accessing data
that spans page boundaries.
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Tested-by: Marc Kleine-Budde <mkl@pengutronix.de>
Cc: stable@kernel.org [2.6.37]
dcache_inode_lock can be replaced with per-inode locking. Use existing
inode->i_lock for this. This is slightly non-trivial because we sometimes
need to find the inode from the dentry, which requires d_inode to be
stabilised (either with refcount or d_lock).
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
Require filesystems be aware of .d_revalidate being called in rcu-walk
mode (nd->flags & LOOKUP_RCU). For now do a simple push down, returning
-ECHILD from all implementations.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
Reduce some branches and memory accesses in dcache lookup by adding dentry
flags to indicate common d_ops are set, rather than having to check them.
This saves a pointer memory access (dentry->d_op) in common path lookup
situations, and saves another pointer load and branch in cases where we
have d_op but not the particular operation.
Patched with:
git grep -E '[.>]([[:space:]])*d_op([[:space:]])*=' | xargs sed -e 's/\([^\t ]*\)->d_op = \(.*\);/d_set_d_op(\1, \2);/' -e 's/\([^\t ]*\)\.d_op = \(.*\);/d_set_d_op(\&\1, \2);/' -i
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
RCU free the struct inode. This will allow:
- Subsequent store-free path walking patch. The inode must be consulted for
permissions when walking, so an RCU inode reference is a must.
- sb_inode_list_lock to be moved inside i_lock because sb list walkers who want
to take i_lock no longer need to take sb_inode_list_lock to walk the list in
the first place. This will simplify and optimize locking.
- Could remove some nested trylock loops in dcache code
- Could potentially simplify things a bit in VM land. Do not need to take the
page lock to follow page->mapping.
The downsides of this is the performance cost of using RCU. In a simple
creat/unlink microbenchmark, performance drops by about 10% due to inability to
reuse cache-hot slab objects. As iterations increase and RCU freeing starts
kicking over, this increases to about 20%.
In cases where inode lifetimes are longer (ie. many inodes may be allocated
during the average life span of a single inode), a lot of this cache reuse is
not applicable, so the regression caused by this patch is smaller.
The cache-hot regression could largely be avoided by using SLAB_DESTROY_BY_RCU,
however this adds some complexity to list walking and store-free path walking,
so I prefer to implement this at a later date, if it is shown to be a win in
real situations. I haven't found a regression in any non-micro benchmark so I
doubt it will be a problem.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
The remaining usages for dcache_lock is to allow atomic, multi-step read-side
operations over the directory tree by excluding modifications to the tree.
Also, to walk in the leaf->root direction in the tree where we don't have
a natural d_lock ordering.
This could be accomplished by taking every d_lock, but this would mean a
huge number of locks and actually gets very tricky.
Solve this instead by using the rename seqlock for multi-step read-side
operations, retry in case of a rename so we don't walk up the wrong parent.
Concurrent dentry insertions are not serialised against. Concurrent deletes
are tricky when walking up the directory: our parent might have been deleted
when dropping locks so also need to check and retry for that.
We can also use the rename lock in cases where livelock is a worry (and it
is introduced in subsequent patch).
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
Add a new lock, dcache_inode_lock, to protect the inode's i_dentry list
from concurrent modification. d_alias is also protected by d_lock.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
Make d_count non-atomic and protect it with d_lock. This allows us to ensure a
0 refcount dentry remains 0 without dcache_lock. It is also fairly natural when
we start protecting many other dentry members with d_lock.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
Change d_delete from a dentry deletion notification to a dentry caching
advise, more like ->drop_inode. Require it to be constant and idempotent,
and not take d_lock. This is how all existing filesystems use the callback
anyway.
This makes fine grained dentry locking of dput and dentry lru scanning
much simpler.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
In order to enable migration support, we will want to move some of the
structures that are subject to migration into the struct nfs_server.
In particular, if we are to move the state_owner and state_owner_id to
being a per-filesystem structure, then we should label the resulting
open/lock owners with a per-filesytem label to ensure global uniqueness.
This patch does so by adding the super block s_dev to the open/lock owner
name.
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Delegations are per-inode, not per-nfs_client. When a server file
system is migrated, delegations on the client must be moved from the
source to the destination nfs_server. Make it easier to manage a
mount point's delegation list across a migration event by moving the
list to the nfs_server struct.
Clean up: I added documenting comments to public functions I changed
in this patch. For consistency I added comments to all the other
public functions in fs/nfs/delegation.c.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
