If the client is not using a delegation, the right thing to do is to return
it as soon as possible. This helps reduce the amount of state the server
has to track, as well as reducing the potential for conflicts with other
clients.
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Let the actual delegreturn stuff be run in the state manager thread rather
than allocating a separate kthread.
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
We really shouldn't be resetting the sequence ids when doing state
expiration recovery, since we don't know if the server still remembers our
previous state owners. There are servers out there that do attempt to
preserve client state even if the lease has expired. Such a server would
only release that state if a conflicting OPEN request occurs.
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Add a delegation cleanup phase to the state management loop, and do the
NFS4ERR_CB_PATH_DOWN recovery there.
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Add a flag to mark delegations as requiring return, then run a garbage
collector. In the future, this will allow for more flexible delegation
management, where delegations may be marked for return if it turns out
that they are not being referenced.
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
NFSv4 defines a number of state errors which the client does not currently
handle. Among those we should worry about are:
NFS4ERR_ADMIN_REVOKED - the server's administrator revoked our locks
and/or delegations.
NFS4ERR_BAD_STATEID - the client and server are out of sync, possibly
due to a delegation return racing with an OPEN
request.
NFS4ERR_OPENMODE - the client attempted to do something not sanctioned
by the open mode of the stateid. Should normally just
occur as a result of a delegation return race.
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Now that we're using the flags to indicate state that needs to be
recovered, as well as having implemented proper refcounting and spinlocking
on the state and open_owners, we can get rid of nfs_client->cl_sem. The
only remaining case that was dubious was the file locking, and that case is
now covered by the nfsi->rwsem.
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
The unlock path is currently failing to take the nfs_client->cl_sem read
lock, and hence the recovery path may see locks disappear from underneath
it.
Also ensure that it takes the nfs_inode->rwsem read lock so that it there
is no conflict with delegation recalls.
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
If the client for some reason is not able to recover all its state within
the time allotted for the grace period, and the server reboots again, the
client is not allowed to recover the state that was 'lost' using reboot
recovery.
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Without an extra lock, we cannot just assume that the delegation->inode is
valid when we're traversing the rcu-protected nfs_client lists. Use the
delegation->lock to ensure that it is truly valid.
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
When we can update_open_stateid(), we need to be certain that we don't
race with a delegation return. While we could do this by grabbing the
nfs_client->cl_lock, a dedicated spin lock in the delegation structure
will scale better.
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
If the admin has specified the "noresvport" option for an NFS mount
point, the kernel's NFS client uses an unprivileged source port for
the main NFS transport. The kernel's lockd client should use an
unprivileged port in this case as well.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
If the admin has specified the "noresvport" option for an NFS mount
point, the kernel's NFS client uses an unprivileged source port for
the main NFS transport. The kernel's mountd client should use an
unprivileged port in this case as well.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
The standard default security setting for NFS is AUTH_SYS. An NFS
client connects to NFS servers via a privileged source port and a
fixed standard destination port (2049). The client sends raw uid and
gid numbers to identify users making NFS requests, and the server
assumes an appropriate authority on the client has vetted these
values because the source port is privileged.
On Linux, by default in-kernel RPC services use a privileged port in
the range between 650 and 1023 to avoid using source ports of well-
known IP services. Using such a small range limits the number of NFS
mount points and the number of unique NFS servers to which a client
can connect concurrently.
An NFS client can use unprivileged source ports to expand the range of
source port numbers, allowing more concurrent server connections and
more NFS mount points. Servers must explicitly allow NFS connections
from unprivileged ports for this to work.
In the past, bumping the value of the sunrpc.max_resvport sysctl on
the client would permit the NFS client to use unprivileged ports.
Bumping this setting also changes the maximum port number used by
other in-kernel RPC services, some of which still required a port
number less than 1023.
This is exacerbated by the way source port numbers are chosen by the
Linux RPC client, which starts at the top of the range and works
downwards. It means that bumping the maximum means all RPC services
requesting a source port will likely get an unprivileged port instead
of a privileged one.
Changing this setting effects all NFS mount points on a client. A
sysadmin could not selectively choose which mount points would use
non-privileged ports and which could not.
Lastly, this mechanism of expanding the limit on the number of NFS
mount points was entirely undocumented.
To address the need for the NFS client to use a large range of source
ports without interfering with the activity of other in-kernel RPC
services, we introduce a new NFS mount option. This option explicitly
tells only the NFS client to use a non-privileged source port when
communicating with the NFS server for one specific mount point.
This new mount option is called "resvport," like the similar NFS mount
option on FreeBSD and Mac OS X. A sister patch for nfs-utils will be
submitted that documents this new option in nfs(5).
The default setting for this new mount option requires the NFS client
to use a privileged port, as before. Explicitly specifying the
"noresvport" mount option allows the NFS client to use an unprivileged
source port for this mount point when connecting to the NFS server
port.
This mount option is supported only for text-based NFS mounts.
[ Sidebar: it is widely known that security mechanisms based on the
use of privileged source ports are ineffective. However, the NFS
client can combine the use of unprivileged ports with the use of
secure authentication mechanisms, such as Kerberos. This allows a
large number of connections and mount points while ensuring a useful
level of security.
Eventually we may change the default setting for this option
depending on the security flavor used for the mount. For example,
if the mount is using only AUTH_SYS, then the default setting will
be "resvport;" if the mount is using a strong security flavor such
as krb5, the default setting will be "noresvport." ]
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
[Trond.Myklebust@netapp.com: Fixed a bug whereby nfs4_init_client()
was being called with incorrect arguments.]
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Make it possible for the NFSv4 mount set up logic to pass mount option
flags down the stack to nfs_create_rpc_client().
This is immediately useful if we want NFS mount options to modulate
settings of the underlying RPC transport, but it may be useful at some
later point if other parts of the NFSv4 mount initialization logic
want to know what the mount options are.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
The nfs_create_rpc_client() function sets up an RPC client for an NFS
mount point. Add an option that allows it to set up an RPC transport
from an unprivileged port.
Instead of having nfs_create_rpc_client()'s callers retain local
knowledge about how to set up an RPC client, create a couple of flag
arguments to control the use of RPC_CLNT_CREATE flags.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Clean up: convert nfs_mount() to take a single data structure argument to make
it simpler to add more arguments.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Clean up: The nfs_mount() function is not to be used outside of the
NFS client. Move its public declaration to fs/nfs/internal.h.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Clean up: I'm about to move the declaration of nfs_mount into
fs/nfs/internal.h and include it in fs/nfs/nfsroot.c. There's a
conflicting definition of nfs_path in fs/nfs/internal.h and
fs/nfs/nfsroot.c, so rename the private one.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
My understanding is that there is a push to turn the kernel_thread
interface into a non-exported symbol and move all kernel threads to use
the kthread API. This patch changes lockd to use kthread_run to spawn
the reclaimer thread.
I've made the assumption here that the extra module references taken
when we spawn this thread are unnecessary and removed them. I've also
added a KERN_ERR printk that pops if the thread can't be spawned to warn
the admin that the locks won't be reclaimed.
In the future, it would be nice to be able to notify userspace that
locks have been lost (probably by implementing SIGLOST), and adding some
good policies about how long we should reattempt to reclaim the locks.
Finally, I removed a comment about memory leaks that I believe is
obsolete and added a new one to clarify the result of sending a SIGKILL
to the reclaimer thread. As best I can tell, doing so doesn't actually
cause a memory leak.
I consider this patch 2.6.29 material.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
aops->readpages() and its NFS helper readpage_async_filler() will only
be called to do readahead I/O for newly allocated pages. So it's not
necessary to test for the always 0 dirty/uptodate page flags.
The removal of nfs_wb_page() call also fixes a readahead bug: the NFS
readahead has been synchronous since 2.6.23, because that call will
clear PG_readahead, which is the reminder for asynchronous readahead.
More background: the PG_readahead page flag is shared with PG_reclaim,
one for read path and the other for write path. clear_page_dirty_for_io()
unconditionally clears PG_readahead to prevent possible readahead residuals,
assuming itself to be always called in the write path. However, NFS is one
and the only exception in that it _always_ calls clear_page_dirty_for_io()
in the read path, i.e. for readpages()/readpage().
Cc: Trond Myklebust <Trond.Myklebust@netapp.com>
Signed-off-by: Wu Fengguang <wfg@linux.intel.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Since v9ses->uid is unsigned, it would seem better to use simple_strtoul that
simple_strtol.
A simplified version of the semantic patch that makes this change is as
follows: (http://www.emn.fr/x-info/coccinelle/)
// <smpl>
@r2@
long e;
position p;
@@
e = simple_strtol@p(...)
@@
position p != r2.p;
type T;
T e;
@@
e =
- simple_strtol@p
+ simple_strtoul
(...)
// </smpl>
Signed-off-by: Julia Lawall <julia@diku.dk>
Acked-by: Eric Van Hensbergen <ericvh@gmail.com>
d_iname is rubbish for long file names.
Use d_name.name in printks instead.
Signed-off-by: Wu Fengguang <wfg@linux.intel.com>
Acked-by: Eric Van Hensbergen <ericvh@gmail.com>
While testing a kernel with memory poisoning enabled, I saw some warnings
about the redzone getting clobbered when chasing DFS referrals. The
buffer allocation for the unicode converted version of the searchName is
too small and needs to take null termination into account.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Acked-by: Steve French <sfrench@us.ibm.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Define the OCFS2_FEATURE_COMPAT_JBD2 bit in the filesystem header.
Signed-off-by: Joel Becker <joel.becker@oracle.com>
Signed-off-by: Mark Fasheh <mfasheh@suse.com>
When we create xattr bucket during the process of xattr set, we always
need to update the ocfs2_xattr_search since even if the bucket size is
the same as block size, the offset will change because of the removal
of the ocfs2_xattr_block header.
Signed-off-by: Tao Ma <tao.ma@oracle.com>
Signed-off-by: Mark Fasheh <mfasheh@suse.com>
Miles Lane tailing /sys files hit a BUG which Pekka Enberg has tracked
to my 966c8c12dc sprint_symbol(): use
less stack exposing a bug in slub's list_locations() -
kallsyms_lookup() writes a 0 to namebuf[KSYM_NAME_LEN-1], but that was
beyond the end of page provided.
The 100 slop which list_locations() allows at end of page looks roughly
enough for all the other stuff it might print after the symbol before
it checks again: break out KSYM_SYMBOL_LEN earlier than before.
Latencytop and ftrace and are using KSYM_NAME_LEN buffers where they
need KSYM_SYMBOL_LEN buffers, and vmallocinfo a 2*KSYM_NAME_LEN buffer
where it wants a KSYM_SYMBOL_LEN buffer: fix those before anyone copies
them.
[akpm@linux-foundation.org: ftrace.h needs module.h]
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Cc: Christoph Lameter <cl@linux-foundation.org>
Cc Miles Lane <miles.lane@gmail.com>
Acked-by: Pekka Enberg <penberg@cs.helsinki.fi>
Acked-by: Steven Rostedt <srostedt@redhat.com>
Acked-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
On umount two event will be dispatched to watcher:
1: inotify_dev_queue_event(.., IN_UNMOUNT,..)
2: remove_watch(watch, dev)
->inotify_dev_queue_event(.., IN_IGNORED, ..)
But if watcher has IN_ONESHOT bit set then the watcher will be released
inside first event. Which result in accessing invalid object later. IMHO
it is not pure regression. This bug wasn't triggered while initial
inotify interface testing phase because of another bug in IN_ONESHOT
handling logic :)
commit ac74c00e49
Author: Ulisses Furquim <ulissesf@gmail.com>
Date: Fri Feb 8 04:18:16 2008 -0800
inotify: fix check for one-shot watches before destroying them
As the IN_ONESHOT bit is never set when an event is sent we must check it
in the watch's mask and not in the event's mask.
TESTCASE:
mkdir mnt
mount -ttmpfs none mnt
mkdir mnt/d
./inotify mnt/d&
umount mnt ## << lockup or crash here
TESTSOURCE:
/* gcc -oinotify inotify.c */
#include <stdio.h>
#include <stdlib.h>
#include <sys/inotify.h>
int main(int argc, char **argv)
{
char buf[1024];
struct inotify_event *ie;
char *p;
int i;
ssize_t l;
p = argv[1];
i = inotify_init();
inotify_add_watch(i, p, ~0);
l = read(i, buf, sizeof(buf));
printf("read %d bytes\n", l);
ie = (struct inotify_event *) buf;
printf("event mask: %d\n", ie->mask);
return 0;
}
Signed-off-by: Dmitri Monakhov <dmonakhov@openvz.org>
Cc: John McCutchan <ttb@tentacle.dhs.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Robert Love <rlove@google.com>
Cc: Ulisses Furquim <ulissesf@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The large pages fix from bcf8039ed4 broke 32-bit pagemap by pulling the
pagemap entry code out into a function with the wrong return type.
Pagemap entries are 64 bits on all systems and unsigned long is only 32
bits on 32-bit systems.
Signed-off-by: Matt Mackall <mpm@selenic.com>
Reported-by: Doug Graham <dgraham@nortel.com>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Dave Hansen <dave@linux.vnet.ibm.com>
Cc: <stable@kernel.org> [2.6.26.x, 2.6.27.x]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Revert
commit e8ced39d5e
Author: Mingming Cao <cmm@us.ibm.com>
Date: Fri Jul 11 19:27:31 2008 -0400
percpu_counter: new function percpu_counter_sum_and_set
As described in
revert "percpu counter: clean up percpu_counter_sum_and_set()"
the new percpu_counter_sum_and_set() is racy against updates to the
cpu-local accumulators on other CPUs. Revert that change.
This means that ext4 will be slow again. But correct.
Reported-by: Eric Dumazet <dada1@cosmosbay.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mingming Cao <cmm@us.ibm.com>
Cc: <linux-ext4@vger.kernel.org>
Cc: <stable@kernel.org> [2.6.27.x]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Revert
commit 1f7c14c62c
Author: Mingming Cao <cmm@us.ibm.com>
Date: Thu Oct 9 12:50:59 2008 -0400
percpu counter: clean up percpu_counter_sum_and_set()
Before this patch we had the following:
percpu_counter_sum(): return the percpu_counter's value
percpu_counter_sum_and_set(): return the percpu_counter's value, copying
that value into the central value and zeroing the per-cpu counters before
returning.
After this patch, percpu_counter_sum_and_set() has gone, and
percpu_counter_sum() gets the old percpu_counter_sum_and_set()
functionality.
Problem is, as Eric points out, the old percpu_counter_sum_and_set()
functionality was racy and wrong. It zeroes out counters on "other" cpus,
without holding any locks which will prevent races agaist updates from
those other CPUS.
This patch reverts 1f7c14c62c. This means
that percpu_counter_sum_and_set() still has the race, but
percpu_counter_sum() does not.
Note that this is not a simple revert - ext4 has since started using
percpu_counter_sum() for its dirty_blocks counter as well.
Note that this revert patch changes percpu_counter_sum() semantics.
Before the patch, a call to percpu_counter_sum() will bring the counter's
central counter mostly up-to-date, so a following percpu_counter_read()
will return a close value.
After this patch, a call to percpu_counter_sum() will leave the counter's
central accumulator unaltered, so a subsequent call to
percpu_counter_read() can now return a significantly inaccurate result.
If there is any code in the tree which was introduced after
e8ced39d5e was merged, and which depends
upon the new percpu_counter_sum() semantics, that code will break.
Reported-by: Eric Dumazet <dada1@cosmosbay.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mingming Cao <cmm@us.ibm.com>
Cc: <linux-ext4@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The patch 6341c39 "tracehook: exec" introduced a small regression in
2.6.27 regarding binfmt_misc exec event reporting. Since the reporting
is now done in the common search_binary_handler() function, an exec
of a misc binary will result in two (or possibly multiple) exec events
being reported, instead of just a single one, because the misc handler
contains a recursive call to search_binary_handler.
To add to the confusion, if PTRACE_O_TRACEEXEC is not active, the multiple
SIGTRAP signals will in fact cause only a single ptrace intercept, as the
signals are not queued. However, if PTRACE_O_TRACEEXEC is on, the debugger
will actually see multiple ptrace intercepts (PTRACE_EVENT_EXEC).
The test program included below demonstrates the problem.
This change fixes the bug by calling tracehook_report_exec() only in the
outermost search_binary_handler() call (bprm->recursion_depth == 0).
The additional change to restore bprm->recursion_depth after each binfmt
load_binary call is actually superfluous for this bug, since we test the
value saved on entry to search_binary_handler(). But it keeps the use of
of the depth count to its most obvious expected meaning. Depending on what
binfmt handlers do in certain cases, there could have been false-positive
tests for recursion limits before this change.
/* Test program using PTRACE_O_TRACEEXEC.
This forks and exec's the first argument with the rest of the arguments,
while ptrace'ing. It expects to see one PTRACE_EVENT_EXEC stop and
then a successful exit, with no other signals or events in between.
Test for kernel doing two PTRACE_EVENT_EXEC stops for a binfmt_misc exec:
$ gcc -g traceexec.c -o traceexec
$ sudo sh -c 'echo :test:M::foobar::/bin/cat: > /proc/sys/fs/binfmt_misc/register'
$ echo 'foobar test' > ./foobar
$ chmod +x ./foobar
$ ./traceexec ./foobar; echo $?
==> good <==
foobar test
0
$
==> bad <==
foobar test
unexpected status 0x4057f != 0
3
$
*/
#include <stdio.h>
#include <sys/types.h>
#include <sys/wait.h>
#include <sys/ptrace.h>
#include <unistd.h>
#include <signal.h>
#include <stdlib.h>
static void
wait_for (pid_t child, int expect)
{
int status;
pid_t p = wait (&status);
if (p != child)
{
perror ("wait");
exit (2);
}
if (status != expect)
{
fprintf (stderr, "unexpected status %#x != %#x\n", status, expect);
exit (3);
}
}
int
main (int argc, char **argv)
{
pid_t child = fork ();
if (child < 0)
{
perror ("fork");
return 127;
}
else if (child == 0)
{
ptrace (PTRACE_TRACEME);
raise (SIGUSR1);
execv (argv[1], &argv[1]);
perror ("execve");
_exit (127);
}
wait_for (child, W_STOPCODE (SIGUSR1));
if (ptrace (PTRACE_SETOPTIONS, child,
0L, (void *) (long) PTRACE_O_TRACEEXEC) != 0)
{
perror ("PTRACE_SETOPTIONS");
return 4;
}
if (ptrace (PTRACE_CONT, child, 0L, 0L) != 0)
{
perror ("PTRACE_CONT");
return 5;
}
wait_for (child, W_STOPCODE (SIGTRAP | (PTRACE_EVENT_EXEC << 8)));
if (ptrace (PTRACE_CONT, child, 0L, 0L) != 0)
{
perror ("PTRACE_CONT");
return 6;
}
wait_for (child, W_EXITCODE (0, 0));
return 0;
}
Reported-by: Arnd Bergmann <arnd@arndb.de>
CC: Ulrich Weigand <ulrich.weigand@de.ibm.com>
Signed-off-by: Roland McGrath <roland@redhat.com>
While 440037287c "[PATCH] switch all filesystems over to
d_obtain_alias" removed some cases where fh_to_dentry() and
fh_to_parent() could return NULL, there are still a few NULL returns
left in individual filesystems. Thus it was a mistake for that commit
to remove the handling of NULL returns in the callers.
Revert those parts of 440037287c which removed the NULL handling.
(We could, alternatively, modify all implementations to return -ESTALE
instead of NULL, but that proves to require fixing a number of
filesystems, and in some cases it's arguably more natural to return
NULL.)
Thanks to David for original patch and Linus, Christoph, and Hugh for
review.
Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Cc: David Howells <dhowells@redhat.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Changeset a238b790d5 (Call fasync()
functions without the BKL) introduced a race which could leave
file->f_flags in a state inconsistent with what the underlying
driver/filesystem believes. Revert that change, and also fix the same
races in ioctl_fioasync() and ioctl_fionbio().
This is a minimal, short-term fix; the real fix will not involve the
BKL.
Reported-by: Oleg Nesterov <oleg@redhat.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: stable@kernel.org
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>