This reverts two changesets, ec3c0982a2
("[TCP]: TCP_DEFER_ACCEPT updates - process as established") and
the follow-on bug fix 9ae27e0adb
("tcp: Fix slab corruption with ipv6 and tcp6fuzz").
This change causes several problems, first reported by Ingo Molnar
as a distcc-over-loopback regression where connections were getting
stuck.
Ilpo Järvinen first spotted the locking problems. The new function
added by this code, tcp_defer_accept_check(), only has the
child socket locked, yet it is modifying state of the parent
listening socket.
Fixing that is non-trivial at best, because we can't simply just grab
the parent listening socket lock at this point, because it would
create an ABBA deadlock. The normal ordering is parent listening
socket --> child socket, but this code path would require the
reverse lock ordering.
Next is a problem noticed by Vitaliy Gusev, he noted:
----------------------------------------
>--- a/net/ipv4/tcp_timer.c
>+++ b/net/ipv4/tcp_timer.c
>@@ -481,6 +481,11 @@ static void tcp_keepalive_timer (unsigned long data)
> goto death;
> }
>
>+ if (tp->defer_tcp_accept.request && sk->sk_state == TCP_ESTABLISHED) {
>+ tcp_send_active_reset(sk, GFP_ATOMIC);
>+ goto death;
Here socket sk is not attached to listening socket's request queue. tcp_done()
will not call inet_csk_destroy_sock() (and tcp_v4_destroy_sock() which should
release this sk) as socket is not DEAD. Therefore socket sk will be lost for
freeing.
----------------------------------------
Finally, Alexey Kuznetsov argues that there might not even be any
real value or advantage to these new semantics even if we fix all
of the bugs:
----------------------------------------
Hiding from accept() sockets with only out-of-order data only
is the only thing which is impossible with old approach. Is this really
so valuable? My opinion: no, this is nothing but a new loophole
to consume memory without control.
----------------------------------------
So revert this thing for now.
Signed-off-by: David S. Miller <davem@davemloft.net>
In changeset 22dd485022
("raw: Raw socket leak.") code was added so that we
flush pending frames on raw sockets to avoid leaks.
The ipv4 part was fine, but the ipv6 part was not
done correctly. Unlike the ipv4 side, the ipv6 code
already has a .destroy method for rawv6_prot.
So now there were two assignments to this member, and
what the compiler does is use the last one, effectively
making the ipv6 parts of that changeset a NOP.
Fix this by removing the:
.destroy = inet6_destroy_sock,
line, and adding an inet6_destroy_sock() call to the
end of raw6_destroy().
Noticed by Al Viro.
Signed-off-by: David S. Miller <davem@davemloft.net>
Acked-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
When creation of a new conntrack entry in ctnetlink fails after having
set up the NAT mappings, the conntrack has an extension area allocated
that is not getting properly destroyed when freeing the conntrack again.
This means the NAT extension is still in the bysource hash, causing a
crash when walking over the hash chain the next time:
BUG: unable to handle kernel paging request at 00120fbd
IP: [<c03d394b>] nf_nat_setup_info+0x221/0x58a
*pde = 00000000
Oops: 0000 [#1] PREEMPT SMP
Pid: 2795, comm: conntrackd Not tainted (2.6.26-rc5 #1)
EIP: 0060:[<c03d394b>] EFLAGS: 00010206 CPU: 1
EIP is at nf_nat_setup_info+0x221/0x58a
EAX: 00120fbd EBX: 00120fbd ECX: 00000001 EDX: 00000000
ESI: 0000019e EDI: e853bbb4 EBP: e853bbc8 ESP: e853bb78
DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068
Process conntrackd (pid: 2795, ti=e853a000 task=f7de10f0 task.ti=e853a000)
Stack: 00000000 e853bc2c e85672ec 00000008 c0561084 63c1db4a 00000000 00000000
00000000 0002e109 61d2b1c3 00000000 00000000 00000000 01114e22 61d2b1c3
00000000 00000000 f7444674 e853bc04 00000008 c038e728 0000000a f7444674
Call Trace:
[<c038e728>] nla_parse+0x5c/0xb0
[<c0397c1b>] ctnetlink_change_status+0x190/0x1c6
[<c0397eec>] ctnetlink_new_conntrack+0x189/0x61f
[<c0119aee>] update_curr+0x3d/0x52
[<c03902d1>] nfnetlink_rcv_msg+0xc1/0xd8
[<c0390228>] nfnetlink_rcv_msg+0x18/0xd8
[<c0390210>] nfnetlink_rcv_msg+0x0/0xd8
[<c038d2ce>] netlink_rcv_skb+0x2d/0x71
[<c0390205>] nfnetlink_rcv+0x19/0x24
[<c038d0f5>] netlink_unicast+0x1b3/0x216
...
Move invocation of the extension destructors to nf_conntrack_free()
to fix this problem.
Fixes http://bugzilla.kernel.org/show_bug.cgi?id=10875
Reported-and-Tested-by: Krzysztof Piotr Oledzki <ole@ans.pl>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
The message "nf_log_packet: can't log since no backend logging module loaded
in! Please either load one, or disable logging explicitly" was displayed for
each logged packet when no userspace application is listening to nflog events.
The message seems to warn for a problem with a kernel module missing but as
said before this is not the case. I thus propose to suppress the message (I
don't see any reason to flood the log because a user application has crashed.)
Signed-off-by: Eric Leblond <eric@inl.fr>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
IPV6_MULTICAST_HOPS, for example, is not valid for stream sockets.
Since they are virtually unavailable for stream sockets,
we should return ENOPROTOOPT instead of EINVAL.
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Only 0 and 1 are valid for IPV6_MULTICAST_LOOP socket option,
and we should return an error of EINVAL otherwise, per RFC3493.
Based on patch from Shan Wei <shanwei@cn.fujitsu.com>.
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
When specifing the outgoing hop limit as ancillary data for sendmsg(),
the kernel doesn't check the integer hop limit value as specified in
[RFC-3542] section 6.3.
Signed-off-by: Shan Wei <shanwei@cn.fujitsu.com>
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
1) We may have route lifetime larger than INT_MAX.
In that case we had wired value in lifetime.
Use INT_MAX if lifetime does not fit in s32.
2) Lifetime is valid iif RTF_EXPIRES is set.
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Step 8.5 in RFC 4340 says for the newly cloned socket
Initialize S.GAR := S.ISS,
but what in fact the code (minisocks.c) does is
Initialize S.GAR := S.ISR,
which is wrong (typo?) -- fixed by the patch.
Signed-off-by: Gerrit Renker <gerrit@erg.abdn.ac.uk>
This fixes a bug in computing the inter-packet-interval t_ipi = s/X:
scaled_div32(a, b) uses u32 for b, but in "scaled_div32(s, X)" the type of the
sending rate `X' is u64. Since X is scaled by 2^6, this truncates rates greater
than 2^26 Bps (~537 Mbps).
Using full 64-bit division now.
Signed-off-by: Gerrit Renker <gerrit@erg.abdn.ac.uk>
This fixes a bug in the reverse lookup of p: given a value f(p), instead of p,
the function returned the smallest tabulated value f(p).
The smallest tabulated value of
10^6 * f(p) = sqrt(2*p/3) + 12 * sqrt(3*p/8) * (32 * p^3 + p)
for p=0.0001 is 8172.
Since this value is scaled by 10^6, the outcome of this bug is that a loss
of 8172/10^6 = 0.8172% was reported whenever the input was below the table
resolution of 0.01%.
This means that the value was over 80 times too high, resulting in large spikes
of the initial loss interval, thus unnecessarily reducing the throughput.
Also corrected the printk format (%u for u32).
Signed-off-by: Gerrit Renker <gerrit@erg.abdn.ac.uk>
This fixes an oversight from an earlier patch, ensuring that Ack Vectors
are not processed on request sockets.
The issue is that Ack Vectors must not be parsed on request sockets, since
the Ack Vector feature depends on the selection of the (TX) CCID. During the
initial handshake the CCIDs are undefined, and so RFC 4340, 10.3 applies:
"Using CCID-specific options and feature options during a negotiation
for the corresponding CCID feature is NOT RECOMMENDED [...]"
And it is not even possible: when the server receives the Request from the
client, the CCID and Ack vector features are undefined; when the Ack finalising
the 3-way hanshake arrives, the request socket has not been cloned yet into a
full socket. (This order is necessary, since otherwise the newly created socket
would have to be destroyed whenever an option error occurred - a malicious
hacker could simply send garbage options and exploit this.)
Signed-off-by: Gerrit Renker <gerrit@erg.abdn.ac.uk>
This patch fixes the following sparse warnings:
* nested min(max()) expression:
net/dccp/ccids/ccid3.c:91:21: warning: symbol '__x' shadows an earlier one
net/dccp/ccids/ccid3.c:91:21: warning: symbol '__y' shadows an earlier one
* Declaration of function prototypes in .c instead of .h file, resulting in
"should it be static?" warnings.
* Declared "struct dccpw" static (local to dccp_probe).
* Disabled dccp_delayed_ack() - not fully removed due to RFC 4340, 11.3
("Receivers SHOULD implement delayed acknowledgement timers ...").
* Used a different local variable name to avoid
net/dccp/ackvec.c:293:13: warning: symbol 'state' shadows an earlier one
net/dccp/ackvec.c:238:33: originally declared here
* Removed unused functions `dccp_ackvector_print' and `dccp_ackvec_print'.
Signed-off-by: Gerrit Renker <gerrit@erg.abdn.ac.uk>
In commit $(825de27d9e) (from 27th May, commit
message `dccp ccid-3: Fix "t_ipi explosion" bug'), the CCID-3 window counter
computation was fixed to cope with RTTs < 4 microseconds.
Such RTTs can be found e.g. when running CCID-3 over loopback. The fix removed
a check against RTT < 4, but introduced a divide-by-zero bug.
All steady-state RTTs in DCCP are filtered using dccp_sample_rtt(), which
ensures non-zero samples. However, a zero RTT is possible on initialisation,
when there is no RTT sample from the Request/Response exchange.
The fix is to use the fallback-RTT from RFC 4340, 3.4.
This is also better than just fixing update_win_count() since it allows other
parts of the code to always assume that the RTT is non-zero during the time
that the CCID is used.
Signed-off-by: Gerrit Renker <gerrit@erg.abdn.ac.uk>
Most legacy software do not like tables > 255 as rtm_table is u8
so tb_id is sent &0xff and it is possible to mismatch for example
table 510 with table 254 (main).
This patch introduces RT_TABLE_COMPAT=252 so the code uses it if
tb_id > 255. It makes such old applications happy, new
ones are still able to use RTA_TABLE to get a proper table id.
Signed-off-by: Krzysztof Piotr Oledzki <ole@ans.pl>
Acked-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
When pfkey has no km listeners, it still does a lot of work
before finding out there aint nobody out there.
If a tree falls in a forest and no one is around to hear it, does it make
a sound? In this case it makes a lot of noise:
With this short-circuit adding 10s of thousands of SAs using
netlink improves performance by ~10%.
Signed-off-by: Jamal Hadi Salim <hadi@cyberus.ca>
Signed-off-by: David S. Miller <davem@davemloft.net>
Wei Yongjun noticed that we may call reqsk_free on request sock objects where
the opt fields may not be initialized, fix it by introducing inet_reqsk_alloc
where we initialize ->opt to NULL and set ->pktopts to NULL in
inet6_reqsk_alloc.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The bindv6only is tuned via sysctl. It is already on a struct net
and per-net sysctls allow for its modification (ipv6_sysctl_net_init).
Despite this the value configured in the init net is used for the
rest of them.
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch adds a check to the set_channel flow. When attempting to change
the channel while in IBSS mode, and the new channel does not support IBSS
mode, the flow return with an error value with no consequences on the
mac80211 and driver state.
Signed-off-by: Assaf Krauss <assaf.krauss@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Sufficient scans (at least 2 or 3) should have been done within 7
seconds to find an existing IBSS to join. This should improve IBSS
creation latency; and since IBSS merging is still in effect, shouldn't
have detrimental effects on eventual IBSS convergence.
Signed-off-by: Dan Williams <dcbw@redhat.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
This patch fixes the issue of slow reconnection to an IBSS cell after
disconnection from it. Now the interface's bssid is reset upon ifdown.
ieee80211_sta_find_ibss:
if (found && memcmp(ifsta->bssid, bssid, ETH_ALEN) != 0 &&
(bss = ieee80211_rx_bss_get(dev, bssid,
local->hw.conf.channel->center_freq,
ifsta->ssid, ifsta->ssid_len)))
Note:
In general disconnection is still not handled properly in mac80211
Signed-off-by: Assaf Krauss <assaf.krauss@intel.com>
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Otherwise userspace has no idea the IBSS creation succeeded.
Signed-off-by: Dan Williams <dcbw@redhat.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
skb_splice_bits temporary drops the socket lock while iterating over
the socket queue in order to break a reverse locking condition which
happens with sendfile. This, however, opens a window of opportunity
for tcp_collapse() to aggregate skbs and thus potentially free the
current skb used in skb_splice_bits and tcp_read_sock.
This patch fixes the problem by (re-)getting the same "logical skb"
after the lock has been temporary dropped.
Based on idea and initial patch from Evgeniy Polyakov.
Signed-off-by: Octavian Purdila <opurdila@ixiacom.com>
Acked-by: Evgeniy Polyakov <johnpol@2ka.mipt.ru>
Signed-off-by: David S. Miller <davem@davemloft.net>
TCP "resets sent" counter is not incremented when a TCP Reset is
sent via tcp_send_active_reset().
Signed-off-by: Sridhar Samudrala <sri@us.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The program below just leaks the raw kernel socket
int main() {
int fd = socket(PF_INET, SOCK_RAW, IPPROTO_UDP);
struct sockaddr_in addr;
memset(&addr, 0, sizeof(addr));
inet_aton("127.0.0.1", &addr.sin_addr);
addr.sin_family = AF_INET;
addr.sin_port = htons(2048);
sendto(fd, "a", 1, MSG_MORE, &addr, sizeof(addr));
return 0;
}
Corked packet is allocated via sock_wmalloc which holds the owner socket,
so one should uncork it and flush all pending data on close. Do this in the
same way as in UDP.
Signed-off-by: Denis V. Lunev <den@openvz.org>
Acked-by: Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit e9df2e8fd8 ("[IPV6]: Use
appropriate sock tclass setting for routing lookup.") also changed the
way that ECN capable transports mark this capability in IPv6. As a
result, SCTP was not marking ECN capablity because the traffic class
was never set. This patch brings back the markings for IPv6 traffic.
Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When fast retransmit is triggered by a sack, we should flush the queue
only once so that only 1 retransmit happens. Also, since we could
potentially have non-fast-rtx chunks on the retransmit queue, we need
make sure any chunks eligable for fast retransmit are sent first
during fast retransmission.
Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com>
Tested-by: Wei Yongjun <yjwei@cn.fujitsu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When we are trying to fast retransmit the lowest outstanding TSN, we
need to restart the T3-RTX timer, so that subsequent timeouts will
correctly tag all the packets necessary for retransmissions.
Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com>
Tested-by: Wei Yongjun <yjwei@cn.fujitsu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Correctly keep track of Fast Recovery state and do not reduce
congestion window multiple times during sucht state.
Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com>
Tested-by: Wei Yongjun <yjwei@cn.fujitsu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There's no need to execute sctp_v4_dst_saddr() for each
iteration, just move it out of loop.
Signed-off-by: Gui Jianfeng <guijianfeng@cn.fujitsu.com>
Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If the current retran_path is the only active one, it should
update it to the the next inactive one.
Signed-off-by: Gui Jianfeng <guijianfeng@cn.fujitsu.com>
Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This bug is able to corrupt fackets_out in very rare cases.
In order for this to cause corruption:
1) DSACK in the middle of previous SACK block must be generated.
2) In order to take that particular branch, part or all of the
DSACKed segment must already be SACKed so that we have that
in cache in the first place.
3) The new info must be top enough so that fackets_out will be
updated on this iteration.
...then fack_count is updated while skb wasn't, then we walk again
that particular segment thus updating fack_count twice for
a single skb and finally that value is assigned to fackets_out
by tcp_sacktag_one.
It is safe to call tcp_sacktag_one just once for a segment (at
DSACK), no need to call again for plain SACK.
Potential problem of the miscount are limited to premature entry
to recovery and to inflated reordering metric (which could even
cancel each other out in the most the luckiest scenarios :-)).
Both are quite insignificant in worst case too and there exists
also code to reset them (fackets_out once sacked_out becomes zero
and reordering metric on RTO).
This has been reported by a number of people, because it occurred
quite rarely, it has been very evasive. Andy Furniss was able to
get it to occur couple of times so that a bit more info was
collected about the problem using a debug patch, though it still
required lot of checking around. Thanks also to others who have
tried to help here.
This is listed as Bugzilla #10346. The bug was introduced by
me in commit 68f8353b48 ([TCP]: Rewrite SACK block processing &
sack_recv_cache use), I probably thought back then that there's
need to scan that entry twice or didn't dare to make it go
through it just once there. Going through twice would have
required restoring fack_count after the walk but as noted above,
I chose to drop the additional walk step altogether here.
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch fixes the usage of RIPEMD-160 in xfrm_algo which in turn
allows hmac(rmd160) to be used as authentication mechanism in IPsec
ESP and AH (see RFC 2857).
Signed-off-by: Adrian-Ken Rueegsegger <rueegsegger@swiss-it.ch>
Acked-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
IPv6 UDP sockets wth IPv4 mapped address use udp_sendmsg to send the data
actually. In this case ip_flush_pending_frames should be called instead
of ip6_flush_pending_frames.
Signed-off-by: Denis V. Lunev <den@openvz.org>
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
It is not allowed to change underlying protocol for
int fd = socket(PF_INET6, SOCK_RAW, IPPROTO_UDP);
Signed-off-by: Denis V. Lunev <den@openvz.org>
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
The outgoing interface index (ipi6_ifindex) in IPV6_PKTINFO
ancillary data, is not checked if the source address (ipi6_addr)
is unspecified. If the ipi6_ifindex is the not-exist interface,
it should be fail.
Based on patch from Shan Wei <shanwei@cn.fujitsu.com> and
Brian Haley <brian.haley@hp.com>.
Signed-off-by: Shan Wei <shanwei@cn.fujitsu.com>
Signed-off-by: Brian Haley <brian.haley@hp.com>
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
If get destination options with length which is not enough for that
option,getsockopt() will still return the real length of the option,
which is larger then the buffer space.
This is because ipv6_getsockopt_sticky() returns the real length of
the option.
This patch fix this problem.
Signed-off-by: Yang Hongyang <yanghy@cn.fujitsu.com>
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
If we pass NULL data buffer to getsockopt(), it will return 0,
and the option length is set to -EFAULT:
getsockopt(sk, IPPROTO_IPV6, IPV6_DSTOPTS, NULL, &len);
This is because ipv6_getsockopt_sticky() will return -EFAULT or
-EINVAL if some error occur.
This patch fix this problem.
Signed-off-by: Yang Hongyang <yanghy@cn.fujitsu.com>
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
- Allow longer lifetimes (>= 0x7fffffff/HZ) on 64bit archs
by using unsigned long.
- Shadow this arithmetic overflow workaround by introducing
helper functions: addrconf_timeout_fixup() and
addrconf_finite_timeout().
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
I discover a strange behavior in [ipv4 in ipv6] tunnel. When IPv6 tunnel
payload is less than 40(0x28), packet can be sent to network, received in
physical interface, but not seen in IP tunnel interface. No counter increase
in tunnel interface.
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
As of now, the prefix length is not vaildated when adding or deleting
addresses. The value is passed directly into the inet6_ifaddr structure
and later passed on to memcmp() as length indicator which relies on
the value never to exceed 128 (bits).
Due to the missing check, the currently code allows for any 8 bit
value to be passed on as prefix length while using the netlink
interface, and any 32 bit value while using the ioctl interface.
[Use unsigned int instead to generate better code - yoshfuji]
Signed-off-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
ip6_sk_dst_lookup returns held dst entry. It should be released
on all paths beyond this point. Add missed release when up->pending
is set.
Bug report and initial patch by Denis V. Lunev <den@openvz.org>.
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Acked-by: Denis V. Lunev <den@openvz.org>
Commit 7cbca67c07 ("[IPV6]: Support
Source Address Selection API (RFC5014)") introduced NULL dereference
of asoc to sctp_v6_get_saddr in net/sctp/ipv6.c.
Pointed out by Johann Felix Soden <johfel@users.sourceforge.net>.
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
