20 Aug, 2011
1 commit
-
Use NUMA aware allocations to reduce latencies and increase throughput.
sunrpc kthreads can use kthread_create_on_node() if pool_mode is
"percpu" or "pernode", and svc_prepare_thread()/svc_init_buffer() can
also take into account NUMA node affinity for memory allocations.Signed-off-by: Eric Dumazet
CC: "J. Bruce Fields"
CC: Neil Brown
CC: David Miller
Reviewed-by: Greg Banks
[bfields@redhat.com: fix up caller nfs41_callback_up]
Signed-off-by: J. Bruce Fields
26 Jan, 2011
1 commit
-
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
Signed-off-by: Trond Myklebust
07 Jan, 2011
5 commits
-
Differentiate from server backchannel
Signed-off-by: Andy Adamson
Acked-by: Bruce Fields
Signed-off-by: Trond Myklebust -
Signed-off-by: Andy Adamson
Acked-by: Bruce Fields
Signed-off-by: Trond Myklebust -
Fixes a bug where the nfs_client could be freed during callback processing.
Refactor nfs_find_client to use minorversion specific means to locate the
correct nfs_client structure.In the NFS layer, V4.0 clients are found using the callback_ident field in the
CB_COMPOUND header. V4.1 clients are found using the sessionID in the
CB_SEQUENCE operation which is also compared against the sessionID associated
with the back channel thread after a successful CREATE_SESSION.Each of these methods finds the one an only nfs_client associated
with the incoming callback request - so nfs_find_client_next is not needed.In the RPC layer, the pg_authenticate call needs to find the nfs_client. For
the v4.0 callback service, the callback identifier has not been decoded so a
search by address, version, and minorversion is used. The sessionid for the
sessions based callback service has (usually) not been set for the
pg_authenticate on a CB_NULL call which can be sent prior to the return
of a CREATE_SESSION call, so the sessionid associated with the back channel
thread is not used to find the client in pg_authenticate for CB_NULL calls.Pass the referenced nfs_client to each CB_COMPOUND operation being proceesed
via the new cb_process_state structure. The reference is held across
cb_compound processing.Use the new cb_process_state struct to move the NFS4ERR_RETRY_UNCACHED_REP
processing from process_op into nfs4_callback_sequence where it belongs.Signed-off-by: Andy Adamson
Signed-off-by: Trond Myklebust -
The sessions based callback service is started prior to the CREATE_SESSION call
so that it can handle CB_NULL requests which can be sent before the
CREATE_SESSION call returns and the session ID is known.Set the callback sessionid after a sucessful CREATE_SESSION.
Signed-off-by: Andy Adamson
Signed-off-by: Trond Myklebust -
The new back channel transport means we call the normal creation routine as
well as svc_xprt_put.Signed-off-by: Andy Adamson
Signed-off-by: Trond Myklebust
18 Nov, 2010
1 commit
-
The big kernel lock has been removed from all these files at some point,
leaving only the #include.Remove this too as a cleanup.
Signed-off-by: Arnd Bergmann
Signed-off-by: Linus Torvalds
02 Oct, 2010
1 commit
-
Signed-off-by: Pavel Emelyanov
Signed-off-by: J. Bruce Fields
11 Aug, 2010
1 commit
-
This is more kernel-ish, saves some space, and also allows us to
expand the ops without breaking all the callers who are happy for the
new members to be NULL.The few places which defined their own param types are changed to the
new scheme (more which crept in recently fixed in following patches).Since we're touching them anyway, we change get() and set() to take a
const struct kernel_param (which they really are). This causes some
harmless warnings until we fix them (in following patches).To reduce churn, module_param_call creates the ops struct so the callers
don't have to change (and casts the functions to reduce warnings).
The modern version which takes an ops struct is called module_param_cb.Signed-off-by: Rusty Russell
Reviewed-by: Takashi Iwai
Tested-by: Phil Carmody
Cc: "David S. Miller"
Cc: Ville Syrjala
Cc: Dmitry Torokhov
Cc: Alessandro Rubini
Cc: Michal Januszewski
Cc: Trond Myklebust
Cc: "J. Bruce Fields"
Cc: Neil Brown
Cc: linux-kernel@vger.kernel.org
Cc: linux-input@vger.kernel.org
Cc: linux-fbdev-devel@lists.sourceforge.net
Cc: linux-nfs@vger.kernel.org
Cc: netdev@vger.kernel.org
27 Jan, 2010
1 commit
-
Clean up: Bruce observed we have more or less common logic in each of
svc_create_xprt()'s callers: the check to create an IPv6 RPC listener
socket only if CONFIG_IPV6 is set. I'm about to add another case
that does just the same.If we move the ifdefs into __svc_xpo_create(), then svc_create_xprt()
call sites can get rid of the "#ifdef" ugliness, and can use the same
logic with or without IPv6 support available in the kernel.Signed-off-by: Chuck Lever
Signed-off-by: J. Bruce Fields
06 Dec, 2009
1 commit
-
The NFSv4.1 spec-29 (18.36.3) says that the server MUST use an ONC RPC
(program) version number equal to 4 in callbacks sent to the client.
For now we allow both versions 1 and 4.Signed-off-by: Alexandros Batsakis
Signed-off-by: Trond Myklebust
04 Dec, 2009
1 commit
-
Signed-off-by: J. Bruce Fields
Signed-off-by: Trond Myklebust
10 Aug, 2009
1 commit
-
Tighten up the validity checking in param_set_port: check for NULL pointers.
Ensure that the option shows up on 'modinfo' output.Signed-off-by: Trond Myklebust
19 Jun, 2009
1 commit
-
Conflicts:
fs/nfs/client.c
fs/nfs/super.c
18 Jun, 2009
5 commits
-
Clear "ret" if the error return from svc_create_xprt(AF_INET6) was
-EAFNOSUPORT. Otherwise, callback start-up will succeed, but
nfs_callback_up() will return -EAFNOSUPPORT anyway, and the first
NFSv4 mount attempt after a reboot will fail.Bug introduced by commit f738f517 in 2.6.30-rc1.
Signed-off-by: Chuck Lever
Signed-off-by: Trond Myklebust -
Tracks the nfs_callback_info for both versions, enabling the callback
service for v4 and v4.1 to run concurrently and be stopped independently
of each other.Signed-off-by: Ricardo Labiaga
Signed-off-by: Benny Halevy -
Signed-off-by: Benny Halevy
-
nfs41_callback_up() initializes the necessary queues and creates the new
nfs41_callback_svc thread. This thread executes the callback service which
waits for requests to arrive on the svc_serv->sv_cb_list.NFS41_BC_MIN_CALLBACKS is set to 1 because we expect callbacks to not
cause substantial latency.The actual processing of the callback will be implemented as a separate patch.
There is only one NFSv4.1 callback service. The first caller of
nfs4_callback_up() creates the service, subsequent callers increment a
reference count on the service. The service is destroyed when the last
caller invokes nfs_callback_down().The transport needs to hold a reference to the callback service in order
to invoke it during callback processing. Currently this reference is only
obtained when the service is first created. This is incorrect, since
subsequent registrations for other transports will leave the xprt->serv
pointer uninitialized, leading to an oops when a callback arrives on
the "unreferenced" transport.This patch fixes the problem by ensuring that a reference to the service
is saved in xprt->serv, either because the service is created by this
invocation to nfs4_callback_up() or by a prior invocation.Signed-off-by: Ricardo Labiaga
Signed-off-by: Benny Halevy
[nfs41: Add a reference to svc_serv during callback service bring up]
Signed-off-by: Ricardo Labiaga
Signed-off-by: Benny Halevy
[Type check arguments of nfs_callback_up]
Signed-off-by: Ricardo Labiaga
Signed-off-by: Benny Halevy
[nfs41: save svc_serv in nfs_callback_info]
Signed-off-by: Benny Halevy
[Removal of ugly #ifdefs]
[nfs41: Update to removal of ugly #ifdefs]
Signed-off-by: Ricardo Labiaga
Signed-off-by: Benny Halevy -
move nfs4_init_callback into nfs4_init_client_minor_version
and nfs4_destroy_callback into nfs4_clear_client_minor_versionas these need to happen also when auto-negotiating the minorversion
once the callback service for nfs41 becomes different than for nfs4.0Signed-off-by: Benny Halevy
[nfs41: Fix checkpatch warning]
Signed-off-by: Ricardo Labiaga
Signed-off-by: Benny Halevy
[Type check arguments of nfs_callback_up]
Signed-off-by: Ricardo Labiaga
Signed-off-by: Benny Halevy
[nfs41: Backchannel: Remove FIXME comment]
Signed-off-by: Ricardo Labiaga
Signed-off-by: Benny Halevy
29 Mar, 2009
4 commits
-
Apparently a lot of people need to disable IPv6 completely on their
distributor-built systems, which have CONFIG_IPV6_MODULE enabled at
build time.They do this by blacklisting the ipv6.ko module. This causes the
creation of the NFSv4 callback service listener to fail if
CONFIG_IPV6_MODULE is set, but the module cannot be loaded.Now that the kernel's PF_INET6 RPC listeners are completely separate
from PF_INET listeners, we can always start PF_INET. Then the NFS
client can try to start a PF_INET6 listener, but it isn't required
to be available.Signed-off-by: Chuck Lever
Signed-off-by: Trond Myklebust -
We're about to convert over to using separate PF_INET and PF_INET6
listeners, instead of a single PF_INET6 listener that also receives
AF_INET requests and maps them to AF_INET6.Clear the way by removing the logic in lockd and the NFSv4 callback
server that creates an AF_INET6 service listener.Signed-off-by: Chuck Lever
Signed-off-by: Trond Myklebust -
Since an RPC service listener's protocol family is specified now via
svc_create_xprt(), it no longer needs to be passed to svc_create() or
svc_create_pooled(). Remove that argument from the synopsis of those
functions, and remove the sv_family field from the svc_serv struct.Signed-off-by: Chuck Lever
Signed-off-by: Trond Myklebust -
The sv_family field is going away. Pass a protocol family argument to
svc_create_xprt() instead of extracting the family from the passed-in
svc_serv struct.Again, as this is a listener socket and not an address, we make this
new argument an "int" protocol family, instead of an "sa_family_t."Signed-off-by: Chuck Lever
Signed-off-by: Trond Myklebust
24 Dec, 2008
1 commit
-
This patch adds client-side support to allow for callbacks other than
AUTH_SYS.Signed-off-by: Olga Kornievskaia
Signed-off-by: J. Bruce Fields
Signed-off-by: Trond Myklebust
18 Oct, 2008
1 commit
-
Allow the NFS callback server to listen for requests via an AF_INET6 or
AF_INET socket when IPv6 support is present in the kernel.Signed-off-by: Chuck Lever
Signed-off-by: Trond Myklebust
30 Sep, 2008
1 commit
-
Introduce and initialize an address family field in the svc_serv structure.
This field will determine what family to use for the service's listener
sockets and what families are advertised via the local rpcbind daemon.Signed-off-by: Chuck Lever
Signed-off-by: J. Bruce Fields
10 Jul, 2008
2 commits
-
If the nfsv4 callback thread is rapidly brought up and down, it's
possible that nfs_callback_svc might never get a chance to run. If
this happens, the cleanup at thread exit might never occur, throwing
the refcounting off and nfs_callback_info in an incorrect state.Move the clean functions into nfs_callback_down. Also change the
nfs_callback_info struct to track the svc_rqst rather than svc_serv
since we need to know that to call svc_exit_thread.Signed-off-by: Jeff Layton
Signed-off-by: Trond Myklebust -
The nfs_callback_mutex is sufficient protection.
Signed-off-by: Jeff Layton
Signed-off-by: Trond Myklebust
17 May, 2008
1 commit
-
__FUNCTION__ is gcc-specific, use __func__
Signed-off-by: Harvey Harrison
Cc: Trond Myklebust
Cc: "J. Bruce Fields"
Signed-off-by: Andrew Morton
Signed-off-by: Trond Myklebust
24 Apr, 2008
3 commits
-
When svc_recv returns an unexpected error, nfs_callback_svc will print a
warning and exit. This problematic for several reasons. In particular,
it will cause the reference counts for the thread to be wrong, and no
new thread will be started until all nfs4 mounts are unmounted.Rather than exiting on error from svc_recv, have the thread do a 1s
sleep and then retry the loop. This is unlikely to cause any harm, and
if the error turns out to be something temporary then it may be able to
recover.Signed-off-by: Jeff Layton
Signed-off-by: J. Bruce Fields -
Obvious comment nit.
Signed-off-by: J. Bruce Fields
-
There's a general push to convert kernel threads to use the (much
cleaner) kthread API. This patch converts the NFSv4 callback kernel
thread to the kthread API. In addition to being generally cleaner this
also removes the dependency on signals when shutting down the thread.Note that this patch depends on the recent patches to svc_recv() to
make it check kthread_should_stop() periodically. Those patches are
in Bruce's tree at the moment and are slated for 2.6.26 along with
the lockd conversion, so this conversion is probably also appropriate
for 2.6.26.Signed-off-by: Jeff Layton
Acked-by: Trond Myklebust
Signed-off-by: J. Bruce Fields
22 Feb, 2008
1 commit
-
Sorry for the noise, but here's the v3 of this compilation fix :)
There are some places, which declare the char buf[...] on the stack
to push it later into dprintk(). Since the dprintk sometimes (if the
CONFIG_SYSCTL=n) becomes an empty do { } while (0) stub, these buffers
cause gcc to produce appropriate warnings.Wrap these buffers with RPC_IFDEBUG macro, as Trond proposed, to
compile them out when not needed.Signed-off-by: Pavel Emelyanov
Acked-by: J. Bruce Fields
Signed-off-by: Trond Myklebust
21 Feb, 2008
1 commit
-
Now that the reference counting on the callback thread is working as
expected, it uncovers another problem. Peter Staubach noticed while
testing that patch on an older kernel that he would occasionally see
this printk in rpc_register fire:"RPC: failed to contact portmap (errno -512).
The NFSv4 callback thread is signaled by nfs_callback_down(), but never
flushes that signal. All of the shutdown processing is done with that
signal pending. This makes it fail the call to unregister the port with
the portmapper.In actuality, this rpc_register call isn't necessary at all since the
port isn't actually registered with the portmapper anymore. Regardless,
there doesn't seem to be any reason to leave the signal pending while
the thread is being shut down and flushing it should generally silence
that printk.Signed-off-by: Jeff Layton
Signed-off-by: Trond Myklebust
14 Feb, 2008
1 commit
-
The reference counting for the NFSv4 callback thread stays artificially
high. When this thread comes down, it doesn't properly tear down the
svc_serv, causing a memory leak. In my testing on an older kernel on
x86_64, memory would leak out of the 8k kmalloc slab. So, we're leaking
at least a page of memory every time the thread comes down.svc_create() creates the svc_serv with a sv_nrthreads count of 1, and
then svc_create_thread() increments that count. Whenever the callback
thread is started it has a sv_nrthreads count of 2. When coming down, it
calls svc_exit_thread() which decrements that count and if it hits 0, it
tears everything down. That never happens here since the count is always
at 2 when the thread exits.The problem is that nfs_callback_up() should be calling svc_destroy() on
the svc_serv on both success and failure. This is how lockd_up_proto()
handles the reference counting, and doing that here fixes the leak.Signed-off-by: Jeff Layton
Signed-off-by: Trond Myklebust
02 Feb, 2008
1 commit
-
Modify the various kernel RPC svcs to use the svc_create_xprt service.
Signed-off-by: Tom Tucker
Acked-by: Neil Brown
Reviewed-by: Chuck Lever
Reviewed-by: Greg Banks
Signed-off-by: J. Bruce Fields
30 Jan, 2008
2 commits
-
Adjust arguments and callers of nfs_find_client() to pass a
"struct sockaddr *" instead of "struct sockaddr_in *" to support non-IPv4
addresses.Signed-off-by: Chuck Lever
Cc: Aurelien CharbonTrond: Also fix up protocol version number argument in nfs_find_client() to
use the correct u32 type.Signed-off-by: Trond Myklebust
-
Clean up: The client side peer address is available in callback_proc.c,
so move a dprintk out of fs/nfs/callback.c and into
fs/nfs/callback_proc.c.This is more consistent with other debugging messages, and the proc
routines have more information about each request to display.Signed-off-by: Chuck Lever
Cc: Aurelien Charbon
Signed-off-by: Trond Myklebust
18 Jul, 2007
1 commit
-
Currently, the freezer treats all tasks as freezable, except for the kernel
threads that explicitly set the PF_NOFREEZE flag for themselves. This
approach is problematic, since it requires every kernel thread to either
set PF_NOFREEZE explicitly, or call try_to_freeze(), even if it doesn't
care for the freezing of tasks at all.It seems better to only require the kernel threads that want to or need to
be frozen to use some freezer-related code and to remove any
freezer-related code from the other (nonfreezable) kernel threads, which is
done in this patch.The patch causes all kernel threads to be nonfreezable by default (ie. to
have PF_NOFREEZE set by default) and introduces the set_freezable()
function that should be called by the freezable kernel threads in order to
unset PF_NOFREEZE. It also makes all of the currently freezable kernel
threads call set_freezable(), so it shouldn't cause any (intentional)
change of behaviour to appear. Additionally, it updates documentation to
describe the freezing of tasks more accurately.[akpm@linux-foundation.org: build fixes]
Signed-off-by: Rafael J. Wysocki
Acked-by: Nigel Cunningham
Cc: Pavel Machek
Cc: Oleg Nesterov
Cc: Gautham R Shenoy
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
