The calculation of RTTVAR involves the subtraction of two unsigned
numbers which
may causes rollover and results in very high values of RTTVAR when RTT > SRTT.
With this patch it is possible to set RTOmin = 1 to get the minimum of RTO at
4 times the clock granularity.
Change Notes:
v2)
*Replaced abs() by abs64() and long by __s64, changed patch
description.
Signed-off-by: Christian Schoch <e0326715@student.tuwien.ac.at>
CC: Vlad Yasevich <vyasevich@gmail.com>
CC: Sridhar Samudrala <sri@us.ibm.com>
CC: Neil Horman <nhorman@tuxdriver.com>
CC: linux-sctp@vger.kernel.org
Acked-by: Vlad Yasevich <vyasevich@gmail.com>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Consider the following program, that sets the second argument to the
sendto() syscall incorrectly:
#include <string.h>
#include <arpa/inet.h>
#include <sys/socket.h>
int main(void)
{
int fd;
struct sockaddr_in sa;
fd = socket(AF_INET, SOCK_STREAM, 132 /*IPPROTO_SCTP*/);
if (fd < 0)
return 1;
memset(&sa, 0, sizeof(sa));
sa.sin_family = AF_INET;
sa.sin_addr.s_addr = inet_addr("127.0.0.1");
sa.sin_port = htons(11111);
sendto(fd, NULL, 1, 0, (struct sockaddr *)&sa, sizeof(sa));
return 0;
}
We get -ENOMEM:
$ strace -e sendto ./demo
sendto(3, NULL, 1, 0, {sa_family=AF_INET, sin_port=htons(11111), sin_addr=inet_addr("127.0.0.1")}, 16) = -1 ENOMEM (Cannot allocate memory)
Propagate the error code from sctp_user_addto_chunk(), so that we will
tell user space what actually went wrong:
$ strace -e sendto ./demo
sendto(3, NULL, 1, 0, {sa_family=AF_INET, sin_port=htons(11111), sin_addr=inet_addr("127.0.0.1")}, 16) = -1 EFAULT (Bad address)
Noticed while running Trinity (the syscall fuzzer).
Signed-off-by: Tommi Rantala <tt.rantala@gmail.com>
Acked-by: Vlad Yasevich <vyasevich@gmail.com>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Trinity (the syscall fuzzer) discovered a memory leak in SCTP,
reproducible e.g. with the sendto() syscall by passing invalid
user space pointer in the second argument:
#include <string.h>
#include <arpa/inet.h>
#include <sys/socket.h>
int main(void)
{
int fd;
struct sockaddr_in sa;
fd = socket(AF_INET, SOCK_STREAM, 132 /*IPPROTO_SCTP*/);
if (fd < 0)
return 1;
memset(&sa, 0, sizeof(sa));
sa.sin_family = AF_INET;
sa.sin_addr.s_addr = inet_addr("127.0.0.1");
sa.sin_port = htons(11111);
sendto(fd, NULL, 1, 0, (struct sockaddr *)&sa, sizeof(sa));
return 0;
}
As far as I can tell, the leak has been around since ~2003.
Signed-off-by: Tommi Rantala <tt.rantala@gmail.com>
Acked-by: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit 13d782f ("sctp: Make the proc files per network namespace.")
changed the /proc/net/sctp/ struct file_operations opener functions to
use single_open_net() and seq_open_net().
Avoid leaking memory by using single_release_net() and seq_release_net()
as the release functions.
Discovered with Trinity (the syscall fuzzer).
Signed-off-by: Tommi Rantala <tt.rantala@gmail.com>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Correct spelling typo in net/sctp/socket.c
Signed-off-by: Masanari Iida <standby24x7@gmail.com>
Acked-by: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Bug introduced by commit edfee0339e
(sctp: check src addr when processing SACK to update transport state)
Signed-off-by: Zijie Pan <zijie.pan@6wind.com>
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Acked-by: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pull networking changes from David Miller:
"The most important bit in here is the fix for input route caching from
Eric Dumazet, it's a shame we couldn't fully analyze this in time for
3.6 as it's a 3.6 regression introduced by the routing cache removal.
Anyways, will send quickly to -stable after you pull this in.
Other changes of note:
1) Fix lockdep splats in team and bonding, from Eric Dumazet.
2) IPV6 adds link local route even when there is no link local
address, from Nicolas Dichtel.
3) Fix ixgbe PTP implementation, from Jacob Keller.
4) Fix excessive stack usage in cxgb4 driver, from Vipul Pandya.
5) MAC length computed improperly in VLAN demux, from Antonio
Quartulli."
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (26 commits)
ipv6: release reference of ip6_null_entry's dst entry in __ip6_del_rt
Remove noisy printks from llcp_sock_connect
tipc: prevent dropped connections due to rcvbuf overflow
silence some noisy printks in irda
team: set qdisc_tx_busylock to avoid LOCKDEP splat
bonding: set qdisc_tx_busylock to avoid LOCKDEP splat
sctp: check src addr when processing SACK to update transport state
sctp: fix a typo in prototype of __sctp_rcv_lookup()
ipv4: add a fib_type to fib_info
can: mpc5xxx_can: fix section type conflict
can: peak_pcmcia: fix error return code
can: peak_pci: fix error return code
cxgb4: Fix build error due to missing linux/vmalloc.h include.
bnx2x: fix ring size for 10G functions
cxgb4: Dynamically allocate memory in t4_memory_rw() and get_vpd_params()
ixgbe: add support for X540-AT1
ixgbe: fix poll loop for FDIRCTRL.INIT_DONE bit
ixgbe: fix PTP ethtool timestamping function
ixgbe: (PTP) Fix PPS interrupt code
ixgbe: Fix PTP X540 SDP alignment code for PPS signal
...
Suppose we have an SCTP connection with two paths. After connection is
established, path1 is not available, thus this path is marked as inactive. Then
traffic goes through path2, but for some reasons packets are delayed (after
rto.max). Because packets are delayed, the retransmit mechanism will switch
again to path1. At this time, we receive a delayed SACK from path2. When we
update the state of the path in sctp_check_transmitted(), we do not take into
account the source address of the SACK, hence we update the wrong path.
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Acked-by: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Just to avoid confusion when people only reads this prototype.
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Acked-by: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pull vfs update from Al Viro:
- big one - consolidation of descriptor-related logics; almost all of
that is moved to fs/file.c
(BTW, I'm seriously tempted to rename the result to fd.c. As it is,
we have a situation when file_table.c is about handling of struct
file and file.c is about handling of descriptor tables; the reasons
are historical - file_table.c used to be about a static array of
struct file we used to have way back).
A lot of stray ends got cleaned up and converted to saner primitives,
disgusting mess in android/binder.c is still disgusting, but at least
doesn't poke so much in descriptor table guts anymore. A bunch of
relatively minor races got fixed in process, plus an ext4 struct file
leak.
- related thing - fget_light() partially unuglified; see fdget() in
there (and yes, it generates the code as good as we used to have).
- also related - bits of Cyrill's procfs stuff that got entangled into
that work; _not_ all of it, just the initial move to fs/proc/fd.c and
switch of fdinfo to seq_file.
- Alex's fs/coredump.c spiltoff - the same story, had been easier to
take that commit than mess with conflicts. The rest is a separate
pile, this was just a mechanical code movement.
- a few misc patches all over the place. Not all for this cycle,
there'll be more (and quite a few currently sit in akpm's tree)."
Fix up trivial conflicts in the android binder driver, and some fairly
simple conflicts due to two different changes to the sock_alloc_file()
interface ("take descriptor handling from sock_alloc_file() to callers"
vs "net: Providing protocol type via system.sockprotoname xattr of
/proc/PID/fd entries" adding a dentry name to the socket)
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs: (72 commits)
MAX_LFS_FILESIZE should be a loff_t
compat: fs: Generic compat_sys_sendfile implementation
fs: push rcu_barrier() from deactivate_locked_super() to filesystems
btrfs: reada_extent doesn't need kref for refcount
coredump: move core dump functionality into its own file
coredump: prevent double-free on an error path in core dumper
usb/gadget: fix misannotations
fcntl: fix misannotations
ceph: don't abuse d_delete() on failure exits
hypfs: ->d_parent is never NULL or negative
vfs: delete surplus inode NULL check
switch simple cases of fget_light to fdget
new helpers: fdget()/fdput()
switch o2hb_region_dev_write() to fget_light()
proc_map_files_readdir(): don't bother with grabbing files
make get_file() return its argument
vhost_set_vring(): turn pollstart/pollstop into bool
switch prctl_set_mm_exe_file() to fget_light()
switch xfs_find_handle() to fget_light()
switch xfs_swapext() to fget_light()
...
Both modular callers of sock_map_fd() had been buggy; sctp one leaks
descriptor and file if copy_to_user() fails, 9p one shouldn't be
exposing file in the descriptor table at all.
Switch both to sock_alloc_file(), export it, unexport sock_map_fd() and
make it static.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Conflicts:
net/netfilter/nfnetlink_log.c
net/netfilter/xt_LOG.c
Rather easy conflict resolution, the 'net' tree had bug fixes to make
sure we checked if a socket is a time-wait one or not and elide the
logging code if so.
Whereas on the 'net-next' side we are calculating the UID and GID from
the creds using different interfaces due to the user namespace changes
from Eric Biederman.
Signed-off-by: David S. Miller <davem@davemloft.net>
Using list_move_tail() instead of list_del() + list_add_tail().
spatch with a semantic match is used to found this problem.
(http://coccinelle.lip6.fr/)
Signed-off-by: Wei Yongjun <yongjun_wei@trendmicro.com.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
SCTP charges wmem_alloc via sctp_set_owner_w() in sctp_sendmsg() and via
skb_set_owner_w() in sctp_packet_transmit(). If a sender runs out of
sndbuf it will sleep in sctp_wait_for_sndbuf() and expects to be waken up
by __sctp_write_space().
Buffer space charged via sctp_set_owner_w() is released in sctp_wfree()
which calls __sctp_write_space() directly.
Buffer space charged via skb_set_owner_w() is released via sock_wfree()
which calls sk->sk_write_space() _if_ SOCK_USE_WRITE_QUEUE is not set.
sctp_endpoint_init() sets SOCK_USE_WRITE_QUEUE on all sockets.
Therefore if sctp_packet_transmit() manages to queue up more than sndbuf
bytes, sctp_wait_for_sndbuf() will never be woken up again unless it is
interrupted by a signal.
This could be fixed by clearing the SOCK_USE_WRITE_QUEUE flag but ...
Charging for the data twice does not make sense in the first place, it
leads to overcharging sndbuf by a factor 2. Therefore this patch only
charges a single byte in wmem_alloc when transmitting an SCTP packet to
ensure that the socket stays alive until the packet has been released.
This means that control chunks are no longer accounted for in wmem_alloc
which I believe is not a problem as skb->truesize will typically lead
to overcharging anyway and thus compensates for any control overhead.
Signed-off-by: Thomas Graf <tgraf@suug.ch>
CC: Vlad Yasevich <vyasevic@redhat.com>
CC: Neil Horman <nhorman@tuxdriver.com>
CC: David Miller <davem@davemloft.net>
Acked-by: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This is an initial merge in of Eric Biederman's work to start adding
user namespace support to the networking.
Signed-off-by: David S. Miller <davem@davemloft.net>
There is an extra semi-colon here, so we always return 0 instead of
calling __sctp_auth_cid().
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
struct seq_net_private has no struct net
if CONFIG_NET_NS is not enabled
Signed-off-by: Ulrich Weber <ulrich.weber@sophos.com>
Reviewed-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add struct net as a parameter to sctp_verify_param so it can be passed
to sctp_verify_ext_param where struct net will be needed when the sctp
tunables become per net tunables.
Add struct net as a parameter to sctp_verify_init so struct net can be
passed to sctp_verify_param.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There are a handle of state machine functions primarily those dealing
with processing INIT packets where there is neither a valid endpoint nor
a valid assoication from which to derive a struct net. Therefore add
struct net * to the parameter list of sctp_state_fn_t and update all of
the state machine functions.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
struct net will be needed shortly when the tunables are made per network
namespace.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This trickles up through sctp_sm_lookup_event up to sctp_do_sm
and up further into sctp_primitiv_NAME before the code reaches
places where struct net can be reliably found.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Start with an empty sctp_net_table that will be populated as the various
tunable sysctls are made per net.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Acked-by: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Acked-by: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
- Fix the sctp_af operations to work in all namespaces
- Enable sctp socket creation in all network namespaces.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Acked-by: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
- Convert all of the files under /proc/net/sctp to be per
network namespace.
- Don't print anything for /proc/net/sctp/snmp except in
the initial network namespaces as the snmp counters still
have to be converted to be per network namespace.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Acked-by: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The percpu sctp socket counter has nothing at all to do with the sctp
proc files, and having it in the wrong initialization is confusing,
and makes network namespace support a pain.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Acked-by: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
- Kill sctp_get_ctl_sock, it is useless now.
- Pass struct net where needed so net->sctp.ctl_sock is accessible.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Acked-by: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
- Move the address lists into struct net
- Add per network namespace initialization and cleanup
- Pass around struct net so it is everywhere I need it.
- Rename all of the global variable references into references
to the variables moved into struct net
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Acked-by: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
- Use struct net in the hash calculation
- Use sock_net(association.base.sk) in the association lookups.
- On receive calculate the network namespace from skb->dev.
- Pass struct net from receive down to the functions that actually
do the association lookup.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Acked-by: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
- Use struct net in the hash calculation
- Use sock_net(endpoint.base.sk) in the endpoint lookups.
- On receive calculate the network namespace from skb->dev.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Acked-by: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
- Add struct net into the port hash table hash calculation
- Add struct net inot the struct sctp_bind_bucket so there
is a memory of which network namespace a port is allocated in.
No need for a ref count because sctp_bind_bucket only exists
when there are sockets in the hash table and sockets can not
change their network namspace, and sockets already ref count
their network namespace.
- Add struct net into the key comparison when we are testing
to see if we have found the port hash table entry we are
looking for.
With these changes lookups in the port hash table becomes
safe to use in multiple network namespaces.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Acked-by: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch series is based on top of "Swap-over-NBD without deadlocking
v15" as it depends on the same reservation of PF_MEMALLOC reserves logic.
When a user or administrator requires swap for their application, they
create a swap partition and file, format it with mkswap and activate it
with swapon. In diskless systems this is not an option so if swap if
required then swapping over the network is considered. The two likely
scenarios are when blade servers are used as part of a cluster where the
form factor or maintenance costs do not allow the use of disks and thin
clients.
The Linux Terminal Server Project recommends the use of the Network Block
Device (NBD) for swap but this is not always an option. There is no
guarantee that the network attached storage (NAS) device is running Linux
or supports NBD. However, it is likely that it supports NFS so there are
users that want support for swapping over NFS despite any performance
concern. Some distributions currently carry patches that support swapping
over NFS but it would be preferable to support it in the mainline kernel.
Patch 1 avoids a stream-specific deadlock that potentially affects TCP.
Patch 2 is a small modification to SELinux to avoid using PFMEMALLOC
reserves.
Patch 3 adds three helpers for filesystems to handle swap cache pages.
For example, page_file_mapping() returns page->mapping for
file-backed pages and the address_space of the underlying
swap file for swap cache pages.
Patch 4 adds two address_space_operations to allow a filesystem
to pin all metadata relevant to a swapfile in memory. Upon
successful activation, the swapfile is marked SWP_FILE and
the address space operation ->direct_IO is used for writing
and ->readpage for reading in swap pages.
Patch 5 notes that patch 3 is bolting
filesystem-specific-swapfile-support onto the side and that
the default handlers have different information to what
is available to the filesystem. This patch refactors the
code so that there are generic handlers for each of the new
address_space operations.
Patch 6 adds an API to allow a vector of kernel addresses to be
translated to struct pages and pinned for IO.
Patch 7 adds support for using highmem pages for swap by kmapping
the pages before calling the direct_IO handler.
Patch 8 updates NFS to use the helpers from patch 3 where necessary.
Patch 9 avoids setting PF_private on PG_swapcache pages within NFS.
Patch 10 implements the new swapfile-related address_space operations
for NFS and teaches the direct IO handler how to manage
kernel addresses.
Patch 11 prevents page allocator recursions in NFS by using GFP_NOIO
where appropriate.
Patch 12 fixes a NULL pointer dereference that occurs when using
swap-over-NFS.
With the patches applied, it is possible to mount a swapfile that is on an
NFS filesystem. Swap performance is not great with a swap stress test
taking roughly twice as long to complete than if the swap device was
backed by NBD.
This patch: netvm: prevent a stream-specific deadlock
It could happen that all !SOCK_MEMALLOC sockets have buffered so much data
that we're over the global rmem limit. This will prevent SOCK_MEMALLOC
buffers from receiving data, which will prevent userspace from running,
which is needed to reduce the buffered data.
Fix this by exempting the SOCK_MEMALLOC sockets from the rmem limit. Once
this change it applied, it is important that sockets that set
SOCK_MEMALLOC do not clear the flag until the socket is being torn down.
If this happens, a warning is generated and the tokens reclaimed to avoid
accounting errors until the bug is fixed.
[davem@davemloft.net: Warning about clearing SOCK_MEMALLOC]
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Mel Gorman <mgorman@suse.de>
Acked-by: David S. Miller <davem@davemloft.net>
Acked-by: Rik van Riel <riel@redhat.com>
Cc: Trond Myklebust <Trond.Myklebust@netapp.com>
Cc: Neil Brown <neilb@suse.de>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Mike Christie <michaelc@cs.wisc.edu>
Cc: Eric B Munson <emunson@mgebm.net>
Cc: Sebastian Andrzej Siewior <sebastian@breakpoint.cc>
Cc: Mel Gorman <mgorman@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Use inet_iif() consistently, and for TCP record the input interface of
cached RX dst in inet sock.
rt->rt_iif is going to be encoded differently, so that we can
legitimately cache input routes in the FIB info more aggressively.
When the input interface is "use SKB device index" the rt->rt_iif will
be set to zero.
This forces us to move the TCP RX dst cache installation into the ipv4
specific code, and as well it should since doing the route caching for
ipv6 is pointless at the moment since it is not inspected in the ipv6
input paths yet.
Also, remove the unlikely on dst->obsolete, all ipv4 dsts have
obsolete set to a non-zero value to force invocation of the check
callback.
Signed-off-by: David S. Miller <davem@davemloft.net>
The ipv4 routing cache is non-deterministic, performance wise, and is
subject to reasonably easy to launch denial of service attacks.
The routing cache works great for well behaved traffic, and the world
was a much friendlier place when the tradeoffs that led to the routing
cache's design were considered.
What it boils down to is that the performance of the routing cache is
a product of the traffic patterns seen by a system rather than being a
product of the contents of the routing tables. The former of which is
controllable by external entitites.
Even for "well behaved" legitimate traffic, high volume sites can see
hit rates in the routing cache of only ~%10.
The general flow of this patch series is that first the routing cache
is removed. We build a completely new rtable entry every lookup
request.
Next we make some simplifications due to the fact that removing the
routing cache causes several members of struct rtable to become no
longer necessary.
Then we need to make some amends such that we can legally cache
pre-constructed routes in the FIB nexthops. Firstly, we need to
invalidate routes which are hit with nexthop exceptions. Secondly we
have to change the semantics of rt->rt_gateway such that zero means
that the destination is on-link and non-zero otherwise.
Now that the preparations are ready, we start caching precomputed
routes in the FIB nexthops. Output and input routes need different
kinds of care when determining if we can legally do such caching or
not. The details are in the commit log messages for those changes.
The patch series then winds down with some more struct rtable
simplifications and other tidy ups that remove unnecessary overhead.
On a SPARC-T3 output route lookups are ~876 cycles. Input route
lookups are ~1169 cycles with rpfilter disabled, and about ~1468
cycles with rpfilter enabled.
These measurements were taken with the kbench_mod test module in the
net_test_tools GIT tree:
git://git.kernel.org/pub/scm/linux/kernel/git/davem/net_test_tools.git
That GIT tree also includes a udpflood tester tool and stresses
route lookups on packet output.
For example, on the same SPARC-T3 system we can run:
time ./udpflood -l 10000000 10.2.2.11
with routing cache:
real 1m21.955s user 0m6.530s sys 1m15.390s
without routing cache:
real 1m31.678s user 0m6.520s sys 1m25.140s
Performance undoubtedly can easily be improved further.
For example fib_table_lookup() performs a lot of excessive
computations with all the masking and shifting, some of it
conditionalized to deal with edge cases.
Also, Eric's no-ref optimization for input route lookups can be
re-instated for the FIB nexthop caching code path. I would be really
pleased if someone would work on that.
In fact anyone suitable motivated can just fire up perf on the loading
of the test net_test_tools benchmark kernel module. I spend much of
my time going:
bash# perf record insmod ./kbench_mod.ko dst=172.30.42.22 src=74.128.0.1 iif=2
bash# perf report
Thanks to helpful feedback from Joe Perches, Eric Dumazet, Ben
Hutchings, and others.
Signed-off-by: David S. Miller <davem@davemloft.net>
I've seen several attempts recently made to do quick failover of sctp transports
by reducing various retransmit timers and counters. While its possible to
implement a faster failover on multihomed sctp associations, its not
particularly robust, in that it can lead to unneeded retransmits, as well as
false connection failures due to intermittent latency on a network.
Instead, lets implement the new ietf quick failover draft found here:
http://tools.ietf.org/html/draft-nishida-tsvwg-sctp-failover-05
This will let the sctp stack identify transports that have had a small number of
errors, and avoid using them quickly until their reliability can be
re-established. I've tested this out on two virt guests connected via multiple
isolated virt networks and believe its in compliance with the above draft and
works well.
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
CC: Vlad Yasevich <vyasevich@gmail.com>
CC: Sridhar Samudrala <sri@us.ibm.com>
CC: "David S. Miller" <davem@davemloft.net>
CC: linux-sctp@vger.kernel.org
CC: joe@perches.com
Acked-by: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add a big comment explaining how the field works, and use defines
instead of magic constants for the values assigned to it.
Suggested by Joe Perches.
Signed-off-by: David S. Miller <davem@davemloft.net>
These patches implement the final mechanism necessary to really allow
us to go without the route cache in ipv4.
We need a place to have long-term storage of PMTU/redirect information
which is independent of the routes themselves, yet does not get us
back into a situation where we have to write to metrics or anything
like that.
For this we use an "next-hop exception" table in the FIB nexthops.
The one thing I desperately want to avoid is having to create clone
routes in the FIB trie for this purpose, because that is very
expensive. However, I'm willing to entertain such an idea later
if this current scheme proves to have downsides that the FIB trie
variant would not have.
In order to accomodate this any such scheme, we need to be able to
produce a full flow key at PMTU/redirect time. That required an
adjustment of the interface call-sites used to propagate these events.
For a PMTU/redirect with a fully specified socket, we pass that socket
and use it to produce the flow key.
Otherwise we use a passed in SKB to formulate the key. There are two
cases that need to be distinguished, ICMP message processing (in which
case the IP header is at skb->data) and output packet processing
(mostly tunnels, and in all such cases the IP header is at ip_hdr(skb)).
We also have to make the code able to handle the case where the dst
itself passed into the dst_ops->{update_pmtu,redirect} method is
invalidated. This matters for calls from sockets that have cached
that route. We provide a inet{,6} helper function for this purpose,
and edit SCTP specially since it caches routes at the transport rather
than socket level.
Signed-off-by: David S. Miller <davem@davemloft.net>
This will be used so that we can compose a full flow key.
Even though we have a route in this context, we need more. In the
future the routes will be without destination address, source address,
etc. keying. One ipv4 route will cover entire subnets, etc.
In this environment we have to have a way to possess persistent storage
for redirects and PMTU information. This persistent storage will exist
in the FIB tables, and that's why we'll need to be able to rebuild a
full lookup flow key here. Using that flow key will do a fib_lookup()
and create/update the persistent entry.
Signed-off-by: David S. Miller <davem@davemloft.net>
Fix the following sparse warning:
* symbol 'sctp_init_cause_fixed' was not declared. Should it be
static?
Signed-off-by: Ioan Orghici <ioanorghici@gmail.com>
Acked-by: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
A few days ago Dave Jones reported this oops:
[22766.294255] general protection fault: 0000 [#1] PREEMPT SMP
[22766.295376] CPU 0
[22766.295384] Modules linked in:
[22766.387137] ffffffffa169f292 6b6b6b6b6b6b6b6b ffff880147c03a90
ffff880147c03a74
[22766.387135] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 00000000000
[22766.387136] Process trinity-watchdo (pid: 10896, threadinfo ffff88013e7d2000,
[22766.387137] Stack:
[22766.387140] ffff880147c03a10
[22766.387140] ffffffffa169f2b6
[22766.387140] ffff88013ed95728
[22766.387143] 0000000000000002
[22766.387143] 0000000000000000
[22766.387143] ffff880003fad062
[22766.387144] ffff88013c120000
[22766.387144]
[22766.387145] Call Trace:
[22766.387145] <IRQ>
[22766.387150] [<ffffffffa169f292>] ? __sctp_lookup_association+0x62/0xd0
[sctp]
[22766.387154] [<ffffffffa169f2b6>] __sctp_lookup_association+0x86/0xd0 [sctp]
[22766.387157] [<ffffffffa169f597>] sctp_rcv+0x207/0xbb0 [sctp]
[22766.387161] [<ffffffff810d4da8>] ? trace_hardirqs_off_caller+0x28/0xd0
[22766.387163] [<ffffffff815827e3>] ? nf_hook_slow+0x133/0x210
[22766.387166] [<ffffffff815902fc>] ? ip_local_deliver_finish+0x4c/0x4c0
[22766.387168] [<ffffffff8159043d>] ip_local_deliver_finish+0x18d/0x4c0
[22766.387169] [<ffffffff815902fc>] ? ip_local_deliver_finish+0x4c/0x4c0
[22766.387171] [<ffffffff81590a07>] ip_local_deliver+0x47/0x80
[22766.387172] [<ffffffff8158fd80>] ip_rcv_finish+0x150/0x680
[22766.387174] [<ffffffff81590c54>] ip_rcv+0x214/0x320
[22766.387176] [<ffffffff81558c07>] __netif_receive_skb+0x7b7/0x910
[22766.387178] [<ffffffff8155856c>] ? __netif_receive_skb+0x11c/0x910
[22766.387180] [<ffffffff810d423e>] ? put_lock_stats.isra.25+0xe/0x40
[22766.387182] [<ffffffff81558f83>] netif_receive_skb+0x23/0x1f0
[22766.387183] [<ffffffff815596a9>] ? dev_gro_receive+0x139/0x440
[22766.387185] [<ffffffff81559280>] napi_skb_finish+0x70/0xa0
[22766.387187] [<ffffffff81559cb5>] napi_gro_receive+0xf5/0x130
[22766.387218] [<ffffffffa01c4679>] e1000_receive_skb+0x59/0x70 [e1000e]
[22766.387242] [<ffffffffa01c5aab>] e1000_clean_rx_irq+0x28b/0x460 [e1000e]
[22766.387266] [<ffffffffa01c9c18>] e1000e_poll+0x78/0x430 [e1000e]
[22766.387268] [<ffffffff81559fea>] net_rx_action+0x1aa/0x3d0
[22766.387270] [<ffffffff810a495f>] ? account_system_vtime+0x10f/0x130
[22766.387273] [<ffffffff810734d0>] __do_softirq+0xe0/0x420
[22766.387275] [<ffffffff8169826c>] call_softirq+0x1c/0x30
[22766.387278] [<ffffffff8101db15>] do_softirq+0xd5/0x110
[22766.387279] [<ffffffff81073bc5>] irq_exit+0xd5/0xe0
[22766.387281] [<ffffffff81698b03>] do_IRQ+0x63/0xd0
[22766.387283] [<ffffffff8168ee2f>] common_interrupt+0x6f/0x6f
[22766.387283] <EOI>
[22766.387284]
[22766.387285] [<ffffffff8168eed9>] ? retint_swapgs+0x13/0x1b
[22766.387285] Code: c0 90 5d c3 66 0f 1f 44 00 00 4c 89 c8 5d c3 0f 1f 00 55 48
89 e5 48 83
ec 20 48 89 5d e8 4c 89 65 f0 4c 89 6d f8 66 66 66 66 90 <0f> b7 87 98 00 00 00
48 89 fb
49 89 f5 66 c1 c0 08 66 39 46 02
[22766.387307]
[22766.387307] RIP
[22766.387311] [<ffffffffa168a2c9>] sctp_assoc_is_match+0x19/0x90 [sctp]
[22766.387311] RSP <ffff880147c039b0>
[22766.387142] ffffffffa16ab120
[22766.599537] ---[ end trace 3f6dae82e37b17f5 ]---
[22766.601221] Kernel panic - not syncing: Fatal exception in interrupt
It appears from his analysis and some staring at the code that this is likely
occuring because an association is getting freed while still on the
sctp_assoc_hashtable. As a result, we get a gpf when traversing the hashtable
while a freed node corrupts part of the list.
Nominally I would think that an mibalanced refcount was responsible for this,
but I can't seem to find any obvious imbalance. What I did note however was
that the two places where we create an association using
sctp_primitive_ASSOCIATE (__sctp_connect and sctp_sendmsg), have failure paths
which free a newly created association after calling sctp_primitive_ASSOCIATE.
sctp_primitive_ASSOCIATE brings us into the sctp_sf_do_prm_asoc path, which
issues a SCTP_CMD_NEW_ASOC side effect, which in turn adds a new association to
the aforementioned hash table. the sctp command interpreter that process side
effects has not way to unwind previously processed commands, so freeing the
association from the __sctp_connect or sctp_sendmsg error path would lead to a
freed association remaining on this hash table.
I've fixed this but modifying sctp_[un]hash_established to use hlist_del_init,
which allows us to proerly use hlist_unhashed to check if the node is on a
hashlist safely during a delete. That in turn alows us to safely call
sctp_unhash_established in the __sctp_connect and sctp_sendmsg error paths
before freeing them, regardles of what the associations state is on the hash
list.
I noted, while I was doing this, that the __sctp_unhash_endpoint was using
hlist_unhsashed in a simmilar fashion, but never nullified any removed nodes
pointers to make that function work properly, so I fixed that up in a simmilar
fashion.
I attempted to test this using a virtual guest running the SCTP_RR test from
netperf in a loop while running the trinity fuzzer, both in a loop. I wasn't
able to recreate the problem prior to this fix, nor was I able to trigger the
failure after (neither of which I suppose is suprising). Given the trace above
however, I think its likely that this is what we hit.
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Reported-by: davej@redhat.com
CC: davej@redhat.com
CC: "David S. Miller" <davem@davemloft.net>
CC: Vlad Yasevich <vyasevich@gmail.com>
CC: Sridhar Samudrala <sri@us.ibm.com>
CC: linux-sctp@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
This adjusts the call to dst_ops->update_pmtu() so that we can
transparently handle the fact that, in the future, the dst itself can
be invalidated by the PMTU update (when we have non-host routes cached
in sockets).
Signed-off-by: David S. Miller <davem@davemloft.net>
While doing some recent work on sctp sack bundling I noted that
sctp_packet_append_chunk was pretty inefficient. Specifially, it was called
recursively while trying to bundle auth and sack chunks. Because of that we
call sctp_packet_bundle_sack and sctp_packet_bundle_auth a total of 4 times for
every call to sctp_packet_append_chunk, knowing that at least 3 of those calls
will do nothing.
So lets refactor sctp_packet_bundle_auth to have an outer part that does the
attempted bundling, and an inner part that just does the chunk appends. This
saves us several calls per iteration that we just don't need.
Also, noticed that the auth and sack bundling fail to free the chunks they
allocate if the append fails, so make sure we add that in
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
CC: Vlad Yasevich <vyasevich@gmail.com>
CC: "David S. Miller" <davem@davemloft.net>
CC: linux-sctp@vger.kernel.org
Acked-by: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
It was noticed recently that when we send data on a transport, its possible that
we might bundle a sack that arrived on a different transport. While this isn't
a major problem, it does go against the SHOULD requirement in section 6.4 of RFC
2960:
An endpoint SHOULD transmit reply chunks (e.g., SACK, HEARTBEAT ACK,
etc.) to the same destination transport address from which it
received the DATA or control chunk to which it is replying. This
rule should also be followed if the endpoint is bundling DATA chunks
together with the reply chunk.
This patch seeks to correct that. It restricts the bundling of sack operations
to only those transports which have moved the ctsn of the association forward
since the last sack. By doing this we guarantee that we only bundle outbound
saks on a transport that has received a chunk since the last sack. This brings
us into stricter compliance with the RFC.
Vlad had initially suggested that we strictly allow only sack bundling on the
transport that last moved the ctsn forward. While this makes sense, I was
concerned that doing so prevented us from bundling in the case where we had
received chunks that moved the ctsn on multiple transports. In those cases, the
RFC allows us to select any of the transports having received chunks to bundle
the sack on. so I've modified the approach to allow for that, by adding a state
variable to each transport that tracks weather it has moved the ctsn since the
last sack. This I think keeps our behavior (and performance), close enough to
our current profile that I think we can do this without a sysctl knob to
enable/disable it.
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
CC: Vlad Yaseivch <vyasevich@gmail.com>
CC: David S. Miller <davem@davemloft.net>
CC: linux-sctp@vger.kernel.org
Reported-by: Michele Baldessari <michele@redhat.com>
Reported-by: sorin serban <sserban@redhat.com>
Acked-by: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
