Changes asserts in sunrpc to use sock_owned_by_user() macro instead of
referencing sock_lock.owner directly.
Signed-off-by: John Heffner <jheffner@psc.edu>
Signed-off-by: David S. Miller <davem@davemloft.net>
Hopefully captured all single statement cases under net/. I'm
not too sure if there is some policy about #includes that are
"guaranteed" (ie., in the current tree) to be available through
some other #included header, so I just added linux/kernel.h to
each changed file that didn't #include it previously.
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
we upgraded the kernel of a nfs-server from 2.6.17.11 to 2.6.22.6. Since
then we get the message
lockd: too many open TCP sockets, consider increasing the number of nfsd threads
lockd: last TCP connect from ^\\236^\É^D
These random characters in the second line are caused by a bug in
svc_tcp_accept.
(Note: there are two previous __svc_print_addr(sin, buf, sizeof(buf))
calls in this function, either of which would initialize buf correctly;
but both are inside "if"'s and are not necessarily executed. This is
less obvious in the second case, which is inside a dprintk(), which is a
macro which expands to an if statement.)
Signed-off-by: Wolfgang Walter <wolfgang.walter@studentenwerk.mhn.de>
Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Commit aaf68cfbf2 added a bias
to sk_inuse, so this test for an unused socket now fails. So no
sockets get closed because they are old (they might get closed
if the client closed them).
This bug has existed since 2.6.21-rc1.
Thanks to Wolfgang Walter for finding and reporting the bug.
Cc: Wolfgang Walter <wolfgang.walter@studentenwerk.mhn.de>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Save the destination address of an incoming request over TCP like is
done already for UDP. It is necessary later for callbacks by the server.
Signed-off-by: Frank van Maarseveen <frankvm@frankvm.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
This while loop has an overly complex condition, which performs a couple of
assignments. This hurts readability.
We don't really need a loop at all. We can just return -EAGAIN and (providing
we set SK_DATA), the function will be called again.
So discard the loop, make the complex conditional become a few clear function
calls, and hopefully improve readability.
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Now that sk_defer_lock protects two different things, make the name more
generic.
Also don't bother with disabling _bh as the lock is only ever taken from
process context.
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
We currently use a special structure (struct skb_timeval) and plain
'struct timeval' to store packet timestamps in sk_buffs and struct
sock.
This has some drawbacks :
- Fixed resolution of micro second.
- Waste of space on 64bit platforms where sizeof(struct timeval)=16
I suggest using ktime_t that is a nice abstraction of high resolution
time services, currently capable of nanosecond resolution.
As sizeof(ktime_t) is 8 bytes, using ktime_t in 'struct sock' permits
a 8 byte shrink of this structure on 64bit architectures. Some other
structures also benefit from this size reduction (struct ipq in
ipv4/ip_fragment.c, struct frag_queue in ipv6/reassembly.c, ...)
Once this ktime infrastructure adopted, we can more easily provide
nanosecond resolution on top of it. (ioctl SIOCGSTAMPNS and/or
SO_TIMESTAMPNS/SCM_TIMESTAMPNS)
Note : this patch includes a bug correction in
compat_sock_get_timestamp() where a "err = 0;" was missing (so this
syscall returned -ENOENT instead of 0)
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
CC: Stephen Hemminger <shemminger@linux-foundation.org>
CC: John find <linux.kernel@free.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>
The return value of kernel_recvmsg() should be assigned to "err", not
compared with the random value of a never initialized "err" (and the "< 0"
check wrongly always returned false since == comparisons never have a
result < 0).
Spotted by the Coverity checker.
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Acked-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
When the last thread of nfsd exits, it shuts down all related sockets. It
currently uses svc_close_socket to do this, but that only is immediately
effective if the socket is not SK_BUSY.
If the socket is busy - i.e. if a request has arrived that has not yet been
processes - svc_close_socket is not effective and the shutdown process spins.
So create a new svc_force_close_socket which removes the SK_BUSY flag is set
and then calls svc_close_socket.
Also change some open-codes loops in svc_destroy to use
list_for_each_entry_safe.
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
They don't really save that much, and aren't worth the hassle.
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The sunrpc server code needs to know the source and destination address for
UDP packets so it can reply properly. It currently copies code out of the
network stack to pick the pieces out of the skb. This is ugly and causes
compile problems with the IPv6 stuff.
So, rip that out and use recv_msg instead. This is a much cleaner interface,
but has a slight cost in that the checksum is now checked before the copy, so
we don't benefit from doing both at the same time. This can probably be
fixed.
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Replace existing svc_create_socket() API to allow callers to pass addresses
larger than a sockaddr_in.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Add support for IPv6 addresses in the RPC server's UDP receive path.
[akpm@linux-foundation.org: cleanups]
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Modify svc_tcp_accept to support connecting on IPv6 sockets.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The only reason svcsock.c looks at a sockaddr's port is to check whether the
remote peer is connecting from a privileged port. Refactor this check to hide
processing that is specific to address format.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
CMSG_DATA comes in different sizes, depending on address family.
[akpm@linux-foundation.org: remove unneeded do/while (0)]
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The rq_daddr field must support larger addresses.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Expand the rq_addr field to allow it to contain larger addresses.
Specifically, we replace a 'sockaddr_in' with a 'sockaddr_storage', then
everywhere the 'sockaddr_in' was referenced, we use instead an accessor
function (svc_addr_in) which safely casts the _storage to _in.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Sockaddr_storage will allow us to store arbitrary socket addresses in the
svc_deferred_req struct.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
There are loads of places where the RPC server assumes that the rq_addr fields
contains an IPv4 address. Top among these are error and debugging messages
that display the server's IP address.
Let's refactor the address printing into a separate function that's smart
enough to figure out the difference between IPv4 and IPv6 addresses.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Clean-up: msg_name and msg_namelen are not used by sock_recvmsg, so don't
bother to set them in svc_recvfrom.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The remote peer's address won't change after the socket has been accepted. We
don't need to call ->getname on every incoming request.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Rather than calling svc_sock_enqueue at the end of svc_setup_socket, we now
call it (via svc_sock_recieved) after calling svc_setup_socket at each call
site.
We do this because a subsequent patch will insert some code between the two
calls at one call site.
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Sometimes we need to create an RPC service but not register it with the local
portmapper. NFSv4 delegation callback, for example.
Change the svc_makesock() API to allow optionally creating temporary or
permanent sockets, optionally registering with the local portmapper, and make
it return the ephemeral port of the new socket.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Currently in the RPC server, registering with the local portmapper and
creating "permanent" sockets are tied together. Expand the internal APIs to
allow these two socket characteristics to be separately specified.
This will be externalized in the next patch.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
If you lose this race, it can iput a socket inode twice and you get a BUG
in fs/inode.c
When I added the option for user-space to close a socket, I added some
cruft to svc_delete_socket so that I could call that function when closing
a socket per user-space request.
This was the wrong thing to do. I should have just set SK_CLOSE and let
normal mechanisms do the work.
Not only wrong, but buggy. The locking is all wrong and it openned up a
race where-by a socket could be closed twice.
So this patch:
Introduces svc_close_socket which sets SK_CLOSE then either leave
the close up to a thread, or calls svc_delete_socket if it can
get SK_BUSY.
Adds a bias to sk_busy which is removed when SK_DEAD is set,
This avoid races around shutting down the socket.
Changes several 'spin_lock' to 'spin_lock_bh' where the _bh
was missing.
Bugzilla-url: http://bugzilla.kernel.org/show_bug.cgi?id=7916
Signed-off-by: Neil Brown <neilb@suse.de>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Also remove {NFSD,RPC}_PARANOIA as having the defines doesn't really add
anything.
The printks covered by RPC_PARANOIA were triggered by badly formatted
packets and so should be ratelimited.
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
NFSd assumes that largest number of pages that will be needed for a
request+response is 2+N where N pages is the size of the largest permitted
read/write request. The '2' are 1 for the non-data part of the request, and 1
for the non-data part of the reply.
However, when a read request is not page-aligned, and we choose to use
->sendfile to send it directly from the page cache, we may need N+1 pages to
hold the whole reply. This can overflow and array and cause an Oops.
This patch increases size of the array for holding pages by one and makes sure
that entry is NULL when it is not in use.
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Stick NFS sockets in their own class to avoid some lockdep warnings. NFS
sockets are never exposed to user-space, and will hence not trigger certain
code paths that would otherwise pose deadlock scenarios.
[akpm@osdl.org: cleanups]
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Steven Dickson <SteveD@redhat.com>
Acked-by: Ingo Molnar <mingo@elte.hu>
Cc: Trond Myklebust <trond.myklebust@fys.uio.no>
Acked-by: Neil Brown <neilb@suse.de>
Cc: "David S. Miller" <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
[ Fixed patch corruption by quilt, pointed out by Peter Zijlstra ]
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Move process freezing functions from include/linux/sched.h to freezer.h, so
that modifications to the freezer or the kernel configuration don't require
recompiling just about everything.
[akpm@osdl.org: fix ueagle driver]
Signed-off-by: Nigel Cunningham <nigel@suspend2.net>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Cc: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
A recent patch fixed a problem which would occur when the refcount on an
auth_domain reached zero. This problem has not been reported in practice
despite existing in two major kernel releases because the refcount can
never reach zero.
This patch fixes the problems that stop the refcount reaching zero.
1/ We were adding to the refcount when inserting in the hash table,
but only removing from the hashtable when the refcount reached zero.
Obviously it never would. So don't count the implied reference of
being in the hash table.
2/ There are two paths on which a socket can be destroyed. One called
svcauth_unix_info_release(). The other didn't. So when the other was
taken, we can lose a reference to an ip_map which in-turn holds a
reference to an auth_domain
So unify the exit paths into svc_sock_put. This highlights the fact
that svc_delete_socket has slightly odd semantics - it does not drop
a reference but probably should. Fixing this need a bit more
thought and testing.
Signed-off-by: Neil Brown <neilb@suse.de>
Cc: Trond Myklebust <trond.myklebust@fys.uio.no>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This patch is suitable for just about any 2.6 kernel. It should go in
2.6.19 and 2.6.18.2 and possible even the .17 and .16 stable series.
This is a long standing bug that seems to have only recently become
apparent, presumably due to increasing use of NFS over TCP - many
distros seem to be making it the default.
The SK_CONN bit gets set when a listening socket may be ready
for an accept, just as SK_DATA is set when data may be available.
It is entirely possible for svc_tcp_accept to be called with neither
of these set. It doesn't happen often but there is a small race in
svc_sock_enqueue as SK_CONN and SK_DATA are tested outside the
spin_lock. They could be cleared immediately after the test and
before the lock is gained.
This normally shouldn't be a problem. The sockets are non-blocking so
trying to read() or accept() when ther is nothing to do is not a problem.
However: svc_tcp_recvfrom makes the decision "Should I accept() or
should I read()" based on whether SK_CONN is set or not. This usually
works but is not safe. The decision should be based on whether it is
a TCP_LISTEN socket or a TCP_CONNECTED socket.
Signed-off-by: Neil Brown <neilb@suse.de>
Cc: Adrian Bunk <bunk@stusta.de>
Cc: <stable@kernel.org>
Cc: Trond Myklebust <trond.myklebust@fys.uio.no>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
There is some confusion about the meaning of 'bufsz' for a sunrpc server.
In some cases it is the largest message that can be sent or received. In
other cases it is the largest 'payload' that can be included in a NFS
message.
In either case, it is not possible for both the request and the reply to be
this large. One of the request or reply may only be one page long, which
fits nicely with NFS.
So we remove 'bufsz' and replace it with two numbers: 'max_payload' and
'max_mesg'. Max_payload is the size that the server requests. It is used
by the server to check the max size allowed on a particular connection:
depending on the protocol a lower limit might be used.
max_mesg is the largest single message that can be sent or received. It is
calculated as the max_payload, rounded up to a multiple of PAGE_SIZE, and
with PAGE_SIZE added to overhead. Only one of the request and reply may be
this size. The other must be at most one page.
Cc: Greg Banks <gnb@sgi.com>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Speed up high call-rate workloads by caching the struct ip_map for the peer on
the connected struct svc_sock instead of looking it up in the ip_map cache
hashtable on every call. This helps workloads using AUTH_SYS authentication
over TCP.
Testing was on a 4 CPU 4 NIC Altix using 4 IRIX clients, each with 16
synthetic client threads simulating an rsync (i.e. recursive directory
listing) workload reading from an i386 RH9 install image (161480 regular files
in 10841 directories) on the server. That tree is small enough to fill in the
server's RAM so no disk traffic was involved. This setup gives a sustained
call rate in excess of 60000 calls/sec before being CPU-bound on the server.
Profiling showed strcmp(), called from ip_map_match(), was taking 4.8% of each
CPU, and ip_map_lookup() was taking 2.9%. This patch drops both contribution
into the profile noise.
Note that the above result overstates this value of this patch for most
workloads. The synthetic clients are all using separate IP addresses, so
there are 64 entries in the ip_map cache hash. Because the kernel measured
contained the bug fixed in commit
commit 1f1e030bf7
and was running on 64bit little-endian machine, probably all of those 64
entries were on a single chain, thus increasing the cost of ip_map_lookup().
With a modern kernel you would need more clients to see the same amount of
performance improvement. This patch has helped to scale knfsd to handle a
deployment with 2000 NFS clients.
Signed-off-by: Greg Banks <gnb@melbourne.sgi.com>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
.. by allocating the array of 'kvec' in 'struct svc_rqst'.
As we plan to increase RPCSVC_MAXPAGES from 8 upto 256, we can no longer
allocate an array of this size on the stack. So we allocate it in 'struct
svc_rqst'.
However svc_rqst contains (indirectly) an array of the same type and size
(actually several, but they are in a union). So rather than waste space, we
move those arrays out of the separately allocated union and into svc_rqst to
share with the kvec moved out of svc_tcp_recvfrom (various arrays are used at
different times, so there is no conflict).
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
We are planning to increase RPCSVC_MAXPAGES from about 8 to about 256. This
means we need to be a bit careful about arrays of size RPCSVC_MAXPAGES.
struct svc_rqst contains two such arrays. However the there are never more
that RPCSVC_MAXPAGES pages in the two arrays together, so only one array is
needed.
The two arrays are for the pages holding the request, and the pages holding
the reply. Instead of two arrays, we can simply keep an index into where the
first reply page is.
This patch also removes a number of small inline functions that probably
server to obscure what is going on rather than clarify it, and opencode the
needed functionality.
Also remove the 'rq_restailpage' variable as it is *always* 0. i.e. if the
response 'xdr' structure has a non-empty tail it is always in the same pages
as the head.
check counters are initilised and incr properly
check for consistant usage of ++ etc
maybe extra some inlines for common approach
general review
Signed-off-by: Neil Brown <neilb@suse.de>
Cc: Magnus Maatta <novell@kiruna.se>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Arrgg.. We cannot 'lockd_up' before 'svc_addsock' as we don't know the
protocol yet.... So switch it around again and save the name of the created
sockets so that it can be closed if lock_up fails.
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The refcount that nfsd holds on lockd is based on the number of open sockets.
So when we close a socket, we should decrement the ref (with lockd_down).
Currently when a socket is closed via writing to the portlist file, that
doesn't happen.
So: make sure we get an error return if the socket that was requested does is
not found, and call lockd_down if it was.
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Actually implement multiple pools. On NUMA machines, allocate a svc_pool per
NUMA node; on SMP a svc_pool per CPU; otherwise a single global pool. Enqueue
sockets on the svc_pool corresponding to the CPU on which the socket bh is run
(i.e. the NIC interrupt CPU). Threads have their cpu mask set to limit them
to the CPUs in the svc_pool that owns them.
This is the patch that allows an Altix to scale NFS traffic linearly
beyond 4 CPUs and 4 NICs.
Incorporates changes and feedback from Neil Brown, Trond Myklebust, and
Christoph Hellwig.
Signed-off-by: Greg Banks <gnb@melbourne.sgi.com>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Split out the list of idle threads and pending sockets from svc_serv into a
new svc_pool structure, and allocate a fixed number (in this patch, 1) of
pools per svc_serv. The new structure contains a lock which takes over
several of the duties of svc_serv->sv_lock, which is now relegated to
protecting only sv_tempsocks, sv_permsocks, and sv_tmpcnt in svc_serv.
The point is to move the hottest fields out of svc_serv and into svc_pool,
allowing a following patch to arrange for a svc_pool per NUMA node or per CPU.
This is a major step towards making the NFS server NUMA-friendly.
Signed-off-by: Greg Banks <gnb@melbourne.sgi.com>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The SK_BUSY bit in svc_sock->sk_flags ensures that we do not attempt to
enqueue a socket twice. Currently, setting and clearing the bit is protected
by svc_serv->sv_lock. As I intend to reduce the data that the lock protects
so it's not held when svc_sock_enqueue() tests and sets SK_BUSY, that test and
set needs to be atomic.
Signed-off-by: Greg Banks <gnb@melbourne.sgi.com>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Convert the svc_sock->sk_reserved variable from an int protected by
svc_serv->sv_lock, to an atomic. This reduces (by 1) the number of places we
need to take the (effectively global) svc_serv->sv_lock.
Signed-off-by: Greg Banks <gnb@melbourne.sgi.com>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Protect the svc_sock->sk_deferred list with a new lock svc_sock->sk_defer_lock
instead of svc_serv->sv_lock. Using the more fine-grained lock reduces the
number of places we need to take the svc_serv lock.
Signed-off-by: Greg Banks <gnb@melbourne.sgi.com>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Convert the svc_sock->sk_inuse counter from an int protected by
svc_serv->sv_lock, to an atomic. This reduces the number of places we need to
take the (effectively global) svc_serv->sv_lock.
Signed-off-by: Greg Banks <gnb@melbourne.sgi.com>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Following are 11 patches from Greg Banks which combine to make knfsd more
Numa-aware. They reduce hitting on 'global' data structures, and create some
data-structures that can be node-local.
knfsd threads are bound to a particular node, and the thread to handle a new
request is chosen from the threads that are attach to the node that received
the interrupt.
The distribution of threads across nodes can be controlled by a new file in
the 'nfsd' filesystem, though the default approach of an even spread is
probably fine for most sites.
Some (old) numbers that show the efficacy of these patches: N == number of
NICs == number of CPUs == nmber of clients. Number of NUMA nodes == N/2
N Throughput, MiB/s CPU usage, % (max=N*100)
Before After Before After
--- ------ ---- ----- -----
4 312 435 350 228
6 500 656 501 418
8 562 804 690 589
This patch:
Move the aging of RPC/TCP connection sockets from the main svc_recv() loop to
a timer which uses a mark-and-sweep algorithm every 6 minutes. This reduces
the amount of work that needs to be done in the main RPC loop and the length
of time we need to hold the (effectively global) svc_serv->sv_lock.
[akpm@osdl.org: cleanup]
Signed-off-by: Greg Banks <gnb@melbourne.sgi.com>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
John Heffner