[ Upstream commit 728356dede ]
To avoid use-after-free(s), use a refcount to keep track of the
usable references to any instantiated struct p9_req_t.
This commit adds p9_req_put(), p9_req_get() and p9_req_try_get() as
wrappers to kref_put(), kref_get() and kref_get_unless_zero().
These are used by the client and the transports to keep track of
valid requests' references.
p9_free_req() is added back and used as callback by kref_put().
Add SLAB_TYPESAFE_BY_RCU as it ensures that the memory freed by
kmem_cache_free() will not be reused for another type until the rcu
synchronisation period is over, so an address gotten under rcu read
lock is safe to inc_ref() without corrupting random memory while
the lock is held.
Link: http://lkml.kernel.org/r/1535626341-20693-1-git-send-email-asmadeus@codewreck.org
Co-developed-by: Dominique Martinet <dominique.martinet@cea.fr>
Signed-off-by: Tomas Bortoli <tomasbortoli@gmail.com>
Reported-by: syzbot+467050c1ce275af2a5b8@syzkaller.appspotmail.com
Signed-off-by: Dominique Martinet <dominique.martinet@cea.fr>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 91a76be37f ]
Having a specific cache for the fcall allocations helps speed up
end-to-end latency.
The caches will automatically be merged if there are multiple caches
of items with the same size so we do not need to try to share a cache
between different clients of the same size.
Since the msize is negotiated with the server, only allocate the cache
after that negotiation has happened - previous allocations or
allocations of different sizes (e.g. zero-copy fcall) are made with
kmalloc directly.
Some figures on two beefy VMs with Connect-IB (sriov) / trans=rdma,
with ior running 32 processes in parallel doing small 32 bytes IOs:
- no alloc (4.18-rc7 request cache): 65.4k req/s
- non-power of two alloc, no patch: 61.6k req/s
- power of two alloc, no patch: 62.2k req/s
- non-power of two alloc, with patch: 64.7k req/s
- power of two alloc, with patch: 65.1k req/s
Link: http://lkml.kernel.org/r/1532943263-24378-2-git-send-email-asmadeus@codewreck.org
Signed-off-by: Dominique Martinet <dominique.martinet@cea.fr>
Acked-by: Jun Piao <piaojun@huawei.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Greg Kurz <groug@kaod.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 523adb6cc1 ]
'msize' is often a power of two, or at least page-aligned, so avoiding
an overhead of two dozen bytes for each allocation will help the
allocator do its work and reduce memory fragmentation.
Link: http://lkml.kernel.org/r/1533825236-22896-1-git-send-email-asmadeus@codewreck.org
Suggested-by: Matthew Wilcox <willy@infradead.org>
Signed-off-by: Dominique Martinet <dominique.martinet@cea.fr>
Reviewed-by: Greg Kurz <groug@kaod.org>
Acked-by: Jun Piao <piaojun@huawei.com>
Cc: Matthew Wilcox <willy@infradead.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 996d5b4db4 ]
Replace the custom batch allocation with a slab. Use an IDR to store
pointers to the active requests instead of an array. We don't try to
handle P9_NOTAG specially; the IDR will happily shrink all the way back
once the TVERSION call has completed.
Link: http://lkml.kernel.org/r/20180711210225.19730-6-willy@infradead.org
Signed-off-by: Matthew Wilcox <willy@infradead.org>
Cc: Eric Van Hensbergen <ericvh@gmail.com>
Cc: Ron Minnich <rminnich@sandia.gov>
Cc: Latchesar Ionkov <lucho@ionkov.net>
Signed-off-by: Dominique Martinet <dominique.martinet@cea.fr>
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit 33d915d9e8 upstream.
As per the current design, in the case of sw crypto controlled devices,
it is the device which advertises the support for AP/VLAN iftype based
on it's ability to tranmsit packets encrypted in software
(In VLAN functionality, group traffic generated for a specific
VLAN group is always encrypted in software). Commit db3bdcb9c3
("mac80211: allow AP_VLAN operation on crypto controlled devices")
has introduced this change.
Since 4addr AP operation also uses AP/VLAN iftype, this conditional
way of advertising AP/VLAN support has broken 4addr AP mode operation on
crypto controlled devices which do not support VLAN functionality.
In the case of ath10k driver, not all firmwares have support for VLAN
functionality but all can support 4addr AP operation. Because AP/VLAN
support is not advertised for these devices, 4addr AP operations are
also blocked.
Fix this by allowing 4addr operation on devices which do not support
AP/VLAN iftype but can support 4addr AP operation (decision is based on
the wiphy flag WIPHY_FLAG_4ADDR_AP).
Cc: stable@vger.kernel.org
Fixes: db3bdcb9c3 ("mac80211: allow AP_VLAN operation on crypto controlled devices")
Signed-off-by: Manikanta Pubbisetty <mpubbise@codeaurora.org>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit d5bb334a8e upstream.
The minimum encryption key size for LE connections is 56 bits and to
align LE with BR/EDR, enforce 56 bits of minimum encryption key size for
BR/EDR connections as well.
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Cc: stable@vger.kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 5f3e2bf008 upstream.
Some TCP peers announce a very small MSS option in their SYN and/or
SYN/ACK messages.
This forces the stack to send packets with a very high network/cpu
overhead.
Linux has enforced a minimal value of 48. Since this value includes
the size of TCP options, and that the options can consume up to 40
bytes, this means that each segment can include only 8 bytes of payload.
In some cases, it can be useful to increase the minimal value
to a saner value.
We still let the default to 48 (TCP_MIN_SND_MSS), for compatibility
reasons.
Note that TCP_MAXSEG socket option enforces a minimal value
of (TCP_MIN_MSS). David Miller increased this minimal value
in commit c39508d6f1 ("tcp: Make TCP_MAXSEG minimum more correct.")
from 64 to 88.
We might in the future merge TCP_MIN_SND_MSS and TCP_MIN_MSS.
CVE-2019-11479 -- tcp mss hardcoded to 48
Signed-off-by: Eric Dumazet <edumazet@google.com>
Suggested-by: Jonathan Looney <jtl@netflix.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Tyler Hicks <tyhicks@canonical.com>
Cc: Bruce Curtis <brucec@netflix.com>
Cc: Jonathan Lemon <jonathan.lemon@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 3b4929f65b upstream.
Jonathan Looney reported that TCP can trigger the following crash
in tcp_shifted_skb() :
BUG_ON(tcp_skb_pcount(skb) < pcount);
This can happen if the remote peer has advertized the smallest
MSS that linux TCP accepts : 48
An skb can hold 17 fragments, and each fragment can hold 32KB
on x86, or 64KB on PowerPC.
This means that the 16bit witdh of TCP_SKB_CB(skb)->tcp_gso_segs
can overflow.
Note that tcp_sendmsg() builds skbs with less than 64KB
of payload, so this problem needs SACK to be enabled.
SACK blocks allow TCP to coalesce multiple skbs in the retransmit
queue, thus filling the 17 fragments to maximal capacity.
CVE-2019-11477 -- u16 overflow of TCP_SKB_CB(skb)->tcp_gso_segs
Fixes: 832d11c5cd ("tcp: Try to restore large SKBs while SACK processing")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Jonathan Looney <jtl@netflix.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Reviewed-by: Tyler Hicks <tyhicks@canonical.com>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Bruce Curtis <brucec@netflix.com>
Cc: Jonathan Lemon <jonathan.lemon@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This reverts commit 38f092c41c which is
commit d5bb334a8e upstream.
Lots of people have reported issues with this patch, and as there does
not seem to be a fix going into Linus's kernel tree any time soon,
revert the commit in the stable trees so as to get people's machines
working properly again.
Reported-by: Vasily Khoruzhick <anarsoul@gmail.com>
Reported-by: Hans de Goede <hdegoede@redhat.com>
Cc: Jeremy Cline <jeremy@jcline.org>
Cc: Marcel Holtmann <marcel@holtmann.org>
Cc: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 9b3040a6aa upstream.
Define __ipv4_neigh_lookup_noref to return NULL when CONFIG_INET is disabled.
Fixes: 4b2a2bfeb3 ("neighbor: Call __ipv4_neigh_lookup_noref in neigh_xmit")
Reported-by: kbuild test robot <lkp@intel.com>
Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Cc: Nobuhiro Iwamatsu <nobuhiro1.iwamatsu@toshiba.co.jp>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit e52972c11d ]
Commit 38030d7cb7 ("net/tls: avoid NULL-deref on resync during device removal")
tried to fix a potential NULL-dereference by taking the
context rwsem. Unfortunately the RX resync may get called
from soft IRQ, so we can't use the rwsem to protect from
the device disappearing. Because we are guaranteed there
can be only one resync at a time (it's called from strparser)
use a bit to indicate resync is busy and make device
removal wait for the bit to get cleared.
Note that there is a leftover "flags" field in struct
tls_context already.
Fixes: 4799ac81e5 ("tls: Add rx inline crypto offload")
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit b7999b0772 ]
In Jianlin's testing, netperf was broken with 'Connection reset by peer',
as the cookie check failed in rt6_check() and ip6_dst_check() always
returned NULL.
It's caused by Commit 93531c6743 ("net/ipv6: separate handling of FIB
entries from dst based routes"), where the cookie can be got only when
'c1'(see below) for setting dst_cookie whereas rt6_check() is called
when !'c1' for checking dst_cookie, as we can see in ip6_dst_check().
Since in ip6_dst_check() both rt6_dst_from_check() (c1) and rt6_check()
(!c1) will check the 'from' cookie, this patch is to remove the c1 check
in rt6_get_cookie(), so that the dst_cookie can always be set properly.
c1:
(rt->rt6i_flags & RTF_PCPU || unlikely(!list_empty(&rt->rt6i_uncached)))
Fixes: 93531c6743 ("net/ipv6: separate handling of FIB entries from dst based routes")
Reported-by: Jianlin Shi <jishi@redhat.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit df453700e8 ]
According to Amit Klein and Benny Pinkas, IP ID generation is too weak
and might be used by attackers.
Even with recent net_hash_mix() fix (netns: provide pure entropy for net_hash_mix())
having 64bit key and Jenkins hash is risky.
It is time to switch to siphash and its 128bit keys.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Amit Klein <aksecurity@gmail.com>
Reported-by: Benny Pinkas <benny@pinkas.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit f80c5dad7b ]
This commit makes the kernel not send the next queued HCI command until
a command complete arrives for the last HCI command sent to the
controller. This change avoids a problem with some buggy controllers
(seen on two SKUs of QCA9377) that send an extra command complete event
for the previous command after the kernel had already sent a new HCI
command to the controller.
The problem was reproduced when starting an active scanning procedure,
where an extra command complete event arrives for the LE_SET_RANDOM_ADDR
command. When this happends the kernel ends up not processing the
command complete for the following commmand, LE_SET_SCAN_PARAM, and
ultimately behaving as if a passive scanning procedure was being
performed, when in fact controller is performing an active scanning
procedure. This makes it impossible to discover BLE devices as no device
found events are sent to userspace.
This problem is reproducible on 100% of the attempts on the affected
controllers. The extra command complete event can be seen at timestamp
27.420131 on the btmon logs bellow.
Bluetooth monitor ver 5.50
= Note: Linux version 5.0.0+ (x86_64) 0.352340
= Note: Bluetooth subsystem version 2.22 0.352343
= New Index: 80:C5:F2:8F:87:84 (Primary,USB,hci0) [hci0] 0.352344
= Open Index: 80:C5:F2:8F:87:84 [hci0] 0.352345
= Index Info: 80:C5:F2:8F:87:84 (Qualcomm) [hci0] 0.352346
@ MGMT Open: bluetoothd (privileged) version 1.14 {0x0001} 0.352347
@ MGMT Open: btmon (privileged) version 1.14 {0x0002} 0.352366
@ MGMT Open: btmgmt (privileged) version 1.14 {0x0003} 27.302164
@ MGMT Command: Start Discovery (0x0023) plen 1 {0x0003} [hci0] 27.302310
Address type: 0x06
LE Public
LE Random
< HCI Command: LE Set Random Address (0x08|0x0005) plen 6 #1 [hci0] 27.302496
Address: 15:60:F2:91:B2:24 (Non-Resolvable)
> HCI Event: Command Complete (0x0e) plen 4 #2 [hci0] 27.419117
LE Set Random Address (0x08|0x0005) ncmd 1
Status: Success (0x00)
< HCI Command: LE Set Scan Parameters (0x08|0x000b) plen 7 #3 [hci0] 27.419244
Type: Active (0x01)
Interval: 11.250 msec (0x0012)
Window: 11.250 msec (0x0012)
Own address type: Random (0x01)
Filter policy: Accept all advertisement (0x00)
> HCI Event: Command Complete (0x0e) plen 4 #4 [hci0] 27.420131
LE Set Random Address (0x08|0x0005) ncmd 1
Status: Success (0x00)
< HCI Command: LE Set Scan Enable (0x08|0x000c) plen 2 #5 [hci0] 27.420259
Scanning: Enabled (0x01)
Filter duplicates: Enabled (0x01)
> HCI Event: Command Complete (0x0e) plen 4 #6 [hci0] 27.420969
LE Set Scan Parameters (0x08|0x000b) ncmd 1
Status: Success (0x00)
> HCI Event: Command Complete (0x0e) plen 4 #7 [hci0] 27.421983
LE Set Scan Enable (0x08|0x000c) ncmd 1
Status: Success (0x00)
@ MGMT Event: Command Complete (0x0001) plen 4 {0x0003} [hci0] 27.422059
Start Discovery (0x0023) plen 1
Status: Success (0x00)
Address type: 0x06
LE Public
LE Random
@ MGMT Event: Discovering (0x0013) plen 2 {0x0003} [hci0] 27.422067
Address type: 0x06
LE Public
LE Random
Discovery: Enabled (0x01)
@ MGMT Event: Discovering (0x0013) plen 2 {0x0002} [hci0] 27.422067
Address type: 0x06
LE Public
LE Random
Discovery: Enabled (0x01)
@ MGMT Event: Discovering (0x0013) plen 2 {0x0001} [hci0] 27.422067
Address type: 0x06
LE Public
LE Random
Discovery: Enabled (0x01)
Signed-off-by: João Paulo Rechi Vita <jprvita@endlessm.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 025c65e119 ]
If an xfrmi is associated to a vrf layer 3 master device,
xfrm_policy_check() fails after traffic decapsulation. The input
interface is replaced by the layer 3 master device, and hence
xfrmi_decode_session() can't match the xfrmi anymore to satisfy
policy checking.
Extend ingress xfrmi lookup to honor the original layer 3 slave
device, allowing xfrm interfaces to operate within a vrf domain.
Fixes: f203b76d78 ("xfrm: Add virtual xfrm interfaces")
Signed-off-by: Martin Willi <martin@strongswan.org>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit dbb2483b2a ]
In commit 6a53b75932 ("xfrm: check id proto in validate_tmpl()")
I introduced a check for xfrm protocol, but according to Herbert
IPSEC_PROTO_ANY should only be used as a wildcard for lookup, so
it should be removed from validate_tmpl().
And, IPSEC_PROTO_ANY is expected to only match 3 IPSec-specific
protocols, this is why xfrm_state_flush() could still miss
IPPROTO_ROUTING, which leads that those entries are left in
net->xfrm.state_all before exit net. Fix this by replacing
IPSEC_PROTO_ANY with zero.
This patch also extracts the check from validate_tmpl() to
xfrm_id_proto_valid() and uses it in parse_ipsecrequest().
With this, no other protocols should be added into xfrm.
Fixes: 6a53b75932 ("xfrm: check id proto in validate_tmpl()")
Reported-by: syzbot+0bf0519d6e0de15914fe@syzkaller.appspotmail.com
Cc: Steffen Klassert <steffen.klassert@secunet.com>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Acked-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 61fb0d0168 ]
At ipv6 route dismantle, fib6_drop_pcpu_from() is responsible
for finding all percpu routes and set their ->from pointer
to NULL, so that fib6_ref can reach its expected value (1).
The problem right now is that other cpus can still catch the
route being deleted, since there is no rcu grace period
between the route deletion and call to fib6_drop_pcpu_from()
This can leak the fib6 and associated resources, since no
notifier will take care of removing the last reference(s).
I decided to add another boolean (fib6_destroying) instead
of reusing/renaming exception_bucket_flushed to ease stable backports,
and properly document the memory barriers used to implement this fix.
This patch has been co-developped with Wei Wang.
Fixes: 93531c6743 ("net/ipv6: separate handling of FIB entries from dst based routes")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: syzbot <syzkaller@googlegroups.com>
Cc: Wei Wang <weiwan@google.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Martin Lau <kafai@fb.com>
Acked-by: Wei Wang <weiwan@google.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 6491d69839 ]
This is similar to commit e285d5bfb7 ("NFC: Fix the number of pipes")
where we changed NFC_HCI_MAX_PIPES from 127 to 128.
As the comment next to the define explains, the pipe identifier is 7
bits long. The highest possible pipe is 127, but the number of possible
pipes is 128. As the code is now, then there is potential for an
out of bounds array access:
net/nfc/nci/hci.c:297 nci_hci_cmd_received() warn: array off by one?
'ndev->hci_dev->pipes[pipe]' '0-127 == 127'
Fixes: 11f54f2286 ("NFC: nci: Add HCI over NCI protocol support")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <alexander.levin@microsoft.com>
[ Upstream commit 3c79107631 ]
else, we leak the addresses to userspace via ctnetlink events
and dumps.
Compute an ID on demand based on the immutable parts of nf_conn struct.
Another advantage compared to using an address is that there is no
immediate re-use of the same ID in case the conntrack entry is freed and
reallocated again immediately.
Fixes: 3583240249 ("[NETFILTER]: nf_conntrack_expect: kill unique ID")
Fixes: 7f85f91472 ("[NETFILTER]: nf_conntrack: kill unique ID")
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit d5bb334a8e upstream.
The minimum encryption key size for LE connections is 56 bits and to
align LE with BR/EDR, enforce 56 bits of minimum encryption key size for
BR/EDR connections as well.
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Cc: stable@vger.kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit fbd019737d ]
Ying triggered a call trace when doing an asconf testing:
BUG: scheduling while atomic: swapper/12/0/0x10000100
Call Trace:
<IRQ> [<ffffffffa4375904>] dump_stack+0x19/0x1b
[<ffffffffa436fcaf>] __schedule_bug+0x64/0x72
[<ffffffffa437b93a>] __schedule+0x9ba/0xa00
[<ffffffffa3cd5326>] __cond_resched+0x26/0x30
[<ffffffffa437bc4a>] _cond_resched+0x3a/0x50
[<ffffffffa3e22be8>] kmem_cache_alloc_node+0x38/0x200
[<ffffffffa423512d>] __alloc_skb+0x5d/0x2d0
[<ffffffffc0995320>] sctp_packet_transmit+0x610/0xa20 [sctp]
[<ffffffffc098510e>] sctp_outq_flush+0x2ce/0xc00 [sctp]
[<ffffffffc098646c>] sctp_outq_uncork+0x1c/0x20 [sctp]
[<ffffffffc0977338>] sctp_cmd_interpreter.isra.22+0xc8/0x1460 [sctp]
[<ffffffffc0976ad1>] sctp_do_sm+0xe1/0x350 [sctp]
[<ffffffffc099443d>] sctp_primitive_ASCONF+0x3d/0x50 [sctp]
[<ffffffffc0977384>] sctp_cmd_interpreter.isra.22+0x114/0x1460 [sctp]
[<ffffffffc0976ad1>] sctp_do_sm+0xe1/0x350 [sctp]
[<ffffffffc097b3a4>] sctp_assoc_bh_rcv+0xf4/0x1b0 [sctp]
[<ffffffffc09840f1>] sctp_inq_push+0x51/0x70 [sctp]
[<ffffffffc099732b>] sctp_rcv+0xa8b/0xbd0 [sctp]
As it shows, the first sctp_do_sm() running under atomic context (NET_RX
softirq) invoked sctp_primitive_ASCONF() that uses GFP_KERNEL flag later,
and this flag is supposed to be used in non-atomic context only. Besides,
sctp_do_sm() was called recursively, which is not expected.
Vlad tried to fix this recursive call in Commit c078669340 ("sctp: Fix
oops when sending queued ASCONF chunks") by introducing a new command
SCTP_CMD_SEND_NEXT_ASCONF. But it didn't work as this command is still
used in the first sctp_do_sm() call, and sctp_primitive_ASCONF() will
be called in this command again.
To avoid calling sctp_do_sm() recursively, we send the next queued ASCONF
not by sctp_primitive_ASCONF(), but by sctp_sf_do_prm_asconf() in the 1st
sctp_do_sm() directly.
Reported-by: Ying Xu <yinxu@redhat.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit fe384e2fa3 ]
callers of tcf_gact_goto_chain_index() can potentially read an old value
of the chain index, or even dereference a NULL 'goto_chain' pointer,
because 'goto_chain' and 'tcfa_action' are read in the traffic path
without caring of concurrent write in the control path. The most recent
value of chain index can be read also from a->tcfa_action (it's encoded
there together with TC_ACT_GOTO_CHAIN bits), so we don't really need to
dereference 'goto_chain': just read the chain id from the control action.
Fixes: e457d86ada ("net: sched: add couple of goto_chain helpers")
Signed-off-by: Davide Caratti <dcaratti@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin (Microsoft) <sashal@kernel.org>
[ Upstream commit 044175a067 ]
When the umem is cleaned up, the task that created it might already be
gone. If the task was gone, the xdp_umem_release function did not free
the pages member of struct xdp_umem.
It turned out that the task lookup was not needed at all; The code was
a left-over when we moved from task accounting to user accounting [1].
This patch fixes the memory leak by removing the task lookup logic
completely.
[1] https://lore.kernel.org/netdev/20180131135356.19134-3-bjorn.topel@gmail.com/
Link: https://lore.kernel.org/netdev/c1cb2ca8-6a14-3980-8672-f3de0bb38dfd@suse.cz/
Fixes: c0c77d8fb7 ("xsk: add user memory registration support sockopt")
Reported-by: Jiri Slaby <jslaby@suse.cz>
Signed-off-by: Björn Töpel <bjorn.topel@intel.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Sasha Levin (Microsoft) <sashal@kernel.org>
[ Upstream commit 273fe3f100 ]
Set deletion after flush coming in the same batch results in EBUSY. Add
set use counter to track the number of references to this set from
rules. We cannot rely on the list of bindings for this since such list
is still populated from the preparation phase.
Reported-by: Václav Zindulka <vaclav.zindulka@tlapnet.cz>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 40ba1d9b4d ]
The abort path can cause a double-free of an anonymous set.
Added-and-to-be-aborted rule looks like this:
udp dport { 137, 138 } drop
The to-be-aborted transaction list looks like this:
newset
newsetelem
newsetelem
rule
This gets walked in reverse order, so first pass disables the rule, the
set elements, then the set.
After synchronize_rcu(), we then destroy those in same order: rule, set
element, set element, newset.
Problem is that the anonymous set has already been bound to the rule, so
the rule (lookup expression destructor) already frees the set, when then
cause use-after-free when trying to delete the elements from this set,
then try to free the set again when handling the newset expression.
Rule releases the bound set in first place from the abort path, this
causes the use-after-free on set element removal when undoing the new
element transactions. To handle this, skip new element transaction if
set is bound from the abort path.
This is still causes the use-after-free on set element removal. To
handle this, remove transaction from the list when the set is already
bound.
Joint work with Florian Westphal.
Fixes: f6ac858589 ("netfilter: nf_tables: unbind set in rule from commit path")
Bugzilla: https://bugzilla.netfilter.org/show_bug.cgi?id=1325
Acked-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit b8e2040063 ]
Add .release_ops, that is called in case of error at a later stage in
the expression initialization path, ie. .select_ops() has been already
set up operations and that needs to be undone. This allows us to unwind
.select_ops from the error path, ie. release the dynamic operations for
this extension.
Moreover, allocate one single operation instead of recycling them, this
comes at the cost of consuming a bit more memory per rule, but it
simplifies the infrastructure.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
Anonymous sets that are bound to rules from the same transaction trigger
a kernel splat from the abort path due to double set list removal and
double free.
This patch updates the logic to search for the transaction that is
responsible for creating the set and disable the set list removal and
release, given the rule is now responsible for this. Lookup is reverse
since the transaction that adds the set is likely to be at the tail of
the list.
Moreover, this patch adds the unbind step to deliver the event from the
commit path. This should not be done from the worker thread, since we
have no guarantees of in-order delivery to the listener.
This patch removes the assumption that both activate and deactivate
callbacks need to be provided.
Fixes: cd5125d8f5 ("netfilter: nf_tables: split set destruction in deactivate and destroy phase")
Reported-by: Mikhail Morfikov <mmorfikov@gmail.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
[ Upstream commit cd5125d8f5 ]
Splits unbind_set into destroy_set and unbinding operation.
Unbinding removes set from lists (so new transaction would not
find it anymore) but keeps memory allocated (so packet path continues
to work).
Rebind function is added to allow unrolling in case transaction
that wants to remove set is aborted.
Destroy function is added to free the memory, but this could occur
outside of transaction in the future.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit d4289fcc9b ]
Currently, IPv6 defragmentation code drops non-last fragments that
are smaller than 1280 bytes: see
commit 0ed4229b08 ("ipv6: defrag: drop non-last frags smaller than min mtu")
This behavior is not specified in IPv6 RFCs and appears to break
compatibility with some IPv6 implemenations, as reported here:
https://www.spinics.net/lists/netdev/msg543846.html
This patch re-uses common IP defragmentation queueing and reassembly
code in IPv6, removing the 1280 byte restriction.
v2: change handling of overlaps to match that of upstream.
Signed-off-by: Peter Oskolkov <posk@google.com>
Reported-by: Tom Herbert <tom@herbertland.com>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit c23f35d19d ]
This is a refactoring patch: without changing runtime behavior,
it moves rbtree-related code from IPv4-specific files/functions
into .h/.c defrag files shared with IPv6 defragmentation code.
v2: make handling of overlapping packets match upstream.
Signed-off-by: Peter Oskolkov <posk@google.com>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Florian Westphal <fw@strlen.de>
Cc: Tom Herbert <tom@herbertland.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit b4f47f3848 ]
Unlike '&&' operator, the '&' does not have short-circuit
evaluation semantics. IOW both sides of the operator always
get evaluated. Fix the wrong operator in
tls_is_sk_tx_device_offloaded(), which would lead to
out-of-bounds access for for non-full sockets.
Fixes: 4799ac81e5 ("tls: Add rx inline crypto offload")
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Dirk van der Merwe <dirk.vandermerwe@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit f75a2804da ]
xfrm_state_put() moves struct xfrm_state to the GC list
and schedules the GC work to clean it up. On net exit call
path, xfrm_state_flush() is called to clean up and
xfrm_flush_gc() is called to wait for the GC work to complete
before exit.
However, this doesn't work because one of the ->destructor(),
ipcomp_destroy(), schedules the same GC work again inside
the GC work. It is hard to wait for such a nested async
callback. This is also why syzbot still reports the following
warning:
WARNING: CPU: 1 PID: 33 at net/ipv6/xfrm6_tunnel.c:351 xfrm6_tunnel_net_exit+0x2cb/0x500 net/ipv6/xfrm6_tunnel.c:351
...
ops_exit_list.isra.0+0xb0/0x160 net/core/net_namespace.c:153
cleanup_net+0x51d/0xb10 net/core/net_namespace.c:551
process_one_work+0xd0c/0x1ce0 kernel/workqueue.c:2153
worker_thread+0x143/0x14a0 kernel/workqueue.c:2296
kthread+0x357/0x430 kernel/kthread.c:246
ret_from_fork+0x3a/0x50 arch/x86/entry/entry_64.S:352
In fact, it is perfectly fine to bypass GC and destroy xfrm_state
synchronously on net exit call path, because it is in process context
and doesn't need a work struct to do any blocking work.
This patch introduces xfrm_state_put_sync() which simply bypasses
GC, and lets its callers to decide whether to use this synchronous
version. On net exit path, xfrm_state_fini() and
xfrm6_tunnel_net_exit() use it. And, as ipcomp_destroy() itself is
blocking, it can use xfrm_state_put_sync() directly too.
Also rename xfrm_state_gc_destroy() to ___xfrm_state_destroy() to
reflect this change.
Fixes: b48c05ab5d ("xfrm: Fix warning in xfrm6_tunnel_net_exit.")
Reported-and-tested-by: syzbot+e9aebef558e3ed673934@syzkaller.appspotmail.com
Cc: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 8c83f2df9c ]
Configuration check to accept source route IP options should be made on
the incoming netdevice when the skb->dev is an l3mdev master. The route
lookup for the source route next hop also needs the incoming netdev.
v2->v3:
- Simplify by passing the original netdevice down the stack (per David
Ahern).
Signed-off-by: Stephen Suryaputra <ssuryaextr@gmail.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 355b985537 ]
net_hash_mix() currently uses kernel address of a struct net,
and is used in many places that could be used to reveal this
address to a patient attacker, thus defeating KASLR, for
the typical case (initial net namespace, &init_net is
not dynamically allocated)
I believe the original implementation tried to avoid spending
too many cycles in this function, but security comes first.
Also provide entropy regardless of CONFIG_NET_NS.
Fixes: 0b4419162a ("netns: introduce the net_hash_mix "salt" for hashes")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Amit Klein <aksecurity@gmail.com>
Reported-by: Benny Pinkas <benny@pinkas.net>
Cc: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 8e2f311a68 ]
Following command:
iptables -D FORWARD -m physdev ...
causes connectivity loss in some setups.
Reason is that iptables userspace will probe kernel for the module revision
of the physdev patch, and physdev has an artificial dependency on
br_netfilter (xt_physdev use makes no sense unless a br_netfilter module
is loaded).
This causes the "phydev" module to be loaded, which in turn enables the
"call-iptables" infrastructure.
bridged packets might then get dropped by the iptables ruleset.
The better fix would be to change the "call-iptables" defaults to 0 and
enforce explicit setting to 1, but that breaks backwards compatibility.
This does the next best thing: add a request_module call to checkentry.
This was a stray '-D ... -m physdev' won't activate br_netfilter
anymore.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 273160ffc6 ]
sctp_hdr(skb) only works when skb->transport_header is set properly.
But in Netfilter, skb->transport_header for ipv6 is not guaranteed
to be right value for sctphdr. It would cause to fail to check the
checksum for sctp packets.
So fix it by using offset, which is always right in all places.
v1->v2:
- Fix the changelog.
Fixes: e6d8b64b34 ("net: sctp: fix and consolidate SCTP checksumming code")
Reported-by: Li Shuang <shuali@redhat.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit a4dc6a4915 ]
When using fanouts with AF_PACKET, the demux functions such as
fanout_demux_cpu will return an index in the fanout socket array, which
corresponds to the selected socket.
The ordering of this array depends on the order the sockets were added
to a given fanout group, so for FANOUT_CPU this means sockets are bound
to cpus in the order they are configured, which is OK.
However, when stopping then restarting the interface these sockets are
bound to, the sockets are reassigned to the fanout group in the reverse
order, due to the fact that they were inserted at the head of the
interface's AF_PACKET socket list.
This means that traffic that was directed to the first socket in the
fanout group is now directed to the last one after an interface restart.
In the case of FANOUT_CPU, traffic from CPU0 will be directed to the
socket that used to receive traffic from the last CPU after an interface
restart.
This commit introduces a helper to add a socket at the tail of a list,
then uses it to register AF_PACKET sockets.
Note that this changes the order in which sockets are listed in /proc and
with sock_diag.
Fixes: dc99f60069 ("packet: Add fanout support")
Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 6321aa1975 ]
clang warns about overflowing the data[] member in the struct pnpipehdr:
net/phonet/pep.c:295:8: warning: array index 4 is past the end of the array (which contains 1 element) [-Warray-bounds]
if (hdr->data[4] == PEP_IND_READY)
^ ~
include/net/phonet/pep.h:66:3: note: array 'data' declared here
u8 data[1];
Using a flexible array member at the end of the struct avoids the
warning, but since we cannot have a flexible array member inside
of the union, each index now has to be moved back by one, which
makes it a little uglier.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Rémi Denis-Courmont <remi@remlab.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit c4f5627f7e upstream.
With commit e163376220 ("Bluetooth: Handle bt_accept_enqueue() socket
atomically") lock_sock[_nested]() is used to acquire the socket lock
before manipulating the socket. lock_sock[_nested]() may block, which
is problematic since bt_accept_enqueue() can be called in bottom half
context (e.g. from rfcomm_connect_ind()):
[<ffffff80080d81ec>] __might_sleep+0x4c/0x80
[<ffffff800876c7b0>] lock_sock_nested+0x24/0x58
[<ffffff8000d7c27c>] bt_accept_enqueue+0x48/0xd4 [bluetooth]
[<ffffff8000e67d8c>] rfcomm_connect_ind+0x190/0x218 [rfcomm]
Add a parameter to bt_accept_enqueue() to indicate whether the
function is called from BH context, and acquire the socket lock
with bh_lock_sock_nested() if that's the case.
Also adapt all callers of bt_accept_enqueue() to pass the new
parameter:
- l2cap_sock_new_connection_cb()
- uses lock_sock() to lock the parent socket => process context
- rfcomm_connect_ind()
- acquires the parent socket lock with bh_lock_sock() => BH
context
- __sco_chan_add()
- called from sco_chan_add(), which is called from sco_connect().
parent is NULL, hence bt_accept_enqueue() isn't called in this
code path and we can ignore it
- also called from sco_conn_ready(). uses bh_lock_sock() to acquire
the parent lock => BH context
Fixes: e163376220 ("Bluetooth: Handle bt_accept_enqueue() socket atomically")
Signed-off-by: Matthias Kaehlcke <mka@chromium.org>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Cc: stable@vger.kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 3da1ed7ac3 ]
Extract IP options in cipso_v4_error and use __icmp_send.
Signed-off-by: Sergey Nazarov <s-nazarov@yandex.ru>
Acked-by: Paul Moore <paul@paul-moore.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 9ef6b42ad6 ]
Add __icmp_send function having ip_options struct parameter
Signed-off-by: Sergey Nazarov <s-nazarov@yandex.ru>
Reviewed-by: Paul Moore <paul@paul-moore.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 5e1a99eae8 ]
For ip rules, we need to use 'ipproto ipv6-icmp' to match ICMPv6 headers.
But for ip -6 route, currently we only support tcp, udp and icmp.
Add ICMPv6 support so we can match ipv6-icmp rules for route lookup.
v2: As David Ahern and Sabrina Dubroca suggested, Add an argument to
rtm_getroute_parse_ip_proto() to handle ICMP/ICMPv6 with different family.
Reported-by: Jianlin Shi <jishi@redhat.com>
Fixes: eacb9384a3 ("ipv6: support sport, dport and ip_proto in RTM_GETROUTE")
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 46b1c18f9d ]
In the series fc8b81a598 ("Merge branch 'lockless-qdisc-series'")
John made the assumption that the data path had no need to read
the qdisc qlen (number of packets in the qdisc).
It is true when pfifo_fast is used as the root qdisc, or as direct MQ/MQPRIO
children.
But pfifo_fast can be used as leaf in class full qdiscs, and existing
logic needs to access the child qlen in an efficient way.
HTB breaks badly, since it uses cl->leaf.q->q.qlen in :
htb_activate() -> WARN_ON()
htb_dequeue_tree() to decide if a class can be htb_deactivated
when it has no more packets.
HFSC, DRR, CBQ, QFQ have similar issues, and some calls to
qdisc_tree_reduce_backlog() also read q.qlen directly.
Using qdisc_qlen_sum() (which iterates over all possible cpus)
in the data path is a non starter.
It seems we have to put back qlen in a central location,
at least for stable kernels.
For all qdisc but pfifo_fast, qlen is guarded by the qdisc lock,
so the existing q.qlen{++|--} are correct.
For 'lockless' qdisc (pfifo_fast so far), we need to use atomic_{inc|dec}()
because the spinlock might be not held (for example from
pfifo_fast_enqueue() and pfifo_fast_dequeue())
This patch adds atomic_qlen (in the same location than qlen)
and renames the following helpers, since we want to express
they can be used without qdisc lock, and that qlen is no longer percpu.
- qdisc_qstats_cpu_qlen_dec -> qdisc_qstats_atomic_qlen_dec()
- qdisc_qstats_cpu_qlen_inc -> qdisc_qstats_atomic_qlen_inc()
Later (net-next) we might revert this patch by tracking all these
qlen uses and replace them by a more efficient method (not having
to access a precise qlen, but an empty/non_empty status that might
be less expensive to maintain/track).
Another possibility is to have a legacy pfifo_fast version that would
be used when used a a child qdisc, since the parent qdisc needs
a spinlock anyway. But then, future lockless qdiscs would also
have the same problem.
Fixes: 7e66016f2c ("net: sched: helpers to sum qlen and qlen for per cpu logic")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: John Fastabend <john.fastabend@gmail.com>
Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Cc: Cong Wang <xiyou.wangcong@gmail.com>
Cc: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 10f4e76587 ]
In the forward chain, the iif is changed from slave device to master vrf
device. Thus, flow offload does not find a match on the lower slave
device.
This patch uses the cached route, ie. dst->dev, to update the iif and
oif fields in the flow entry.
After this patch, the following example works fine:
# ip addr add dev eth0 1.1.1.1/24
# ip addr add dev eth1 10.0.0.1/24
# ip link add user1 type vrf table 1
# ip l set user1 up
# ip l set dev eth0 master user1
# ip l set dev eth1 master user1
# nft add table firewall
# nft add flowtable f fb1 { hook ingress priority 0 \; devices = { eth0, eth1 } \; }
# nft add chain f ftb-all {type filter hook forward priority 0 \; policy accept \; }
# nft add rule f ftb-all ct zone 1 ip protocol tcp flow offload @fb1
# nft add rule f ftb-all ct zone 1 ip protocol udp flow offload @fb1
Signed-off-by: wenxu <wenxu@ucloud.cn>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit c09551c6ff ]
According to the algorithm described in the comment block at the
beginning of ip_rt_send_redirect, the host should try to send
'ip_rt_redirect_number' ICMP redirect packets with an exponential
backoff and then stop sending them at all assuming that the destination
ignores redirects.
If the device has previously sent some ICMP error packets that are
rate-limited (e.g TTL expired) and continues to receive traffic,
the redirect packets will never be transmitted. This happens since
peer->rate_tokens will be typically greater than 'ip_rt_redirect_number'
and so it will never be reset even if the redirect silence timeout
(ip_rt_redirect_silence) has elapsed without receiving any packet
requiring redirects.
Fix it by using a dedicated counter for the number of ICMP redirect
packets that has been sent by the host
I have not been able to identify a given commit that introduced the
issue since ip_rt_send_redirect implements the same rate-limiting
algorithm from commit 1da177e4c3 ("Linux-2.6.12-rc2")
Signed-off-by: Lorenzo Bianconi <lorenzo.bianconi@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit d5256083f6 ]
While implementing ipvlan l3 and l3s mode for kubernetes CNI plugin,
I ran into the issue that while l3 mode is working fine, l3s mode
does not have any connectivity to kube-apiserver and hence all pods
end up in Error state as well. The ipvlan master device sits on
top of a bond device and hostns traffic to kube-apiserver (also running
in hostns) is DNATed from 10.152.183.1:443 to 139.178.29.207:37573
where the latter is the address of the bond0. While in l3 mode, a
curl to https://10.152.183.1:443 or to https://139.178.29.207:37573
works fine from hostns, neither of them do in case of l3s. In the
latter only a curl to https://127.0.0.1:37573 appeared to work where
for local addresses of bond0 I saw kernel suddenly starting to emit
ARP requests to query HW address of bond0 which remained unanswered
and neighbor entries in INCOMPLETE state. These ARP requests only
happen while in l3s.
Debugging this further, I found the issue is that l3s mode is piggy-
backing on l3 master device, and in this case local routes are using
l3mdev_master_dev_rcu(dev) instead of net->loopback_dev as per commit
f5a0aab84b ("net: ipv4: dst for local input routes should use l3mdev
if relevant") and 5f02ce24c2 ("net: l3mdev: Allow the l3mdev to be
a loopback"). I found that reverting them back into using the
net->loopback_dev fixed ipvlan l3s connectivity and got everything
working for the CNI.
Now judging from 4fbae7d83c ("ipvlan: Introduce l3s mode") and the
l3mdev paper in [0] the only sole reason why ipvlan l3s is relying
on l3 master device is to get the l3mdev_ip_rcv() receive hook for
setting the dst entry of the input route without adding its own
ipvlan specific hacks into the receive path, however, any l3 domain
semantics beyond just that are breaking l3s operation. Note that
ipvlan also has the ability to dynamically switch its internal
operation from l3 to l3s for all ports via ipvlan_set_port_mode()
at runtime. In any case, l3 vs l3s soley distinguishes itself by
'de-confusing' netfilter through switching skb->dev to ipvlan slave
device late in NF_INET_LOCAL_IN before handing the skb to L4.
Minimal fix taken here is to add a IFF_L3MDEV_RX_HANDLER flag which,
if set from ipvlan setup, gets us only the wanted l3mdev_l3_rcv() hook
without any additional l3mdev semantics on top. This should also have
minimal impact since dev->priv_flags is already hot in cache. With
this set, l3s mode is working fine and I also get things like
masquerading pod traffic on the ipvlan master properly working.
[0] https://netdevconf.org/1.2/papers/ahern-what-is-l3mdev-paper.pdf
Fixes: f5a0aab84b ("net: ipv4: dst for local input routes should use l3mdev if relevant")
Fixes: 5f02ce24c2 ("net: l3mdev: Allow the l3mdev to be a loopback")
Fixes: 4fbae7d83c ("ipvlan: Introduce l3s mode")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Cc: Mahesh Bandewar <maheshb@google.com>
Cc: David Ahern <dsa@cumulusnetworks.com>
Cc: Florian Westphal <fw@strlen.de>
Cc: Martynas Pumputis <m@lambda.lt>
Acked-by: David Ahern <dsa@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit f97f4dd8b3 ]
IPv4 routing tables are flushed in two cases:
1. In response to events in the netdev and inetaddr notification chains
2. When a network namespace is being dismantled
In both cases only routes associated with a dead nexthop group are
flushed. However, a nexthop group will only be marked as dead in case it
is populated with actual nexthops using a nexthop device. This is not
the case when the route in question is an error route (e.g.,
'blackhole', 'unreachable').
Therefore, when a network namespace is being dismantled such routes are
not flushed and leaked [1].
To reproduce:
# ip netns add blue
# ip -n blue route add unreachable 192.0.2.0/24
# ip netns del blue
Fix this by not skipping error routes that are not marked with
RTNH_F_DEAD when flushing the routing tables.
To prevent the flushing of such routes in case #1, add a parameter to
fib_table_flush() that indicates if the table is flushed as part of
namespace dismantle or not.
Note that this problem does not exist in IPv6 since error routes are
associated with the loopback device.
[1]
unreferenced object 0xffff888066650338 (size 56):
comm "ip", pid 1206, jiffies 4294786063 (age 26.235s)
hex dump (first 32 bytes):
00 00 00 00 00 00 00 00 b0 1c 62 61 80 88 ff ff ..........ba....
e8 8b a1 64 80 88 ff ff 00 07 00 08 fe 00 00 00 ...d............
backtrace:
[<00000000856ed27d>] inet_rtm_newroute+0x129/0x220
[<00000000fcdfc00a>] rtnetlink_rcv_msg+0x397/0xa20
[<00000000cb85801a>] netlink_rcv_skb+0x132/0x380
[<00000000ebc991d2>] netlink_unicast+0x4c0/0x690
[<0000000014f62875>] netlink_sendmsg+0x929/0xe10
[<00000000bac9d967>] sock_sendmsg+0xc8/0x110
[<00000000223e6485>] ___sys_sendmsg+0x77a/0x8f0
[<000000002e94f880>] __sys_sendmsg+0xf7/0x250
[<00000000ccb1fa72>] do_syscall_64+0x14d/0x610
[<00000000ffbe3dae>] entry_SYSCALL_64_after_hwframe+0x49/0xbe
[<000000003a8b605b>] 0xffffffffffffffff
unreferenced object 0xffff888061621c88 (size 48):
comm "ip", pid 1206, jiffies 4294786063 (age 26.235s)
hex dump (first 32 bytes):
6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b kkkkkkkkkkkkkkkk
6b 6b 6b 6b 6b 6b 6b 6b d8 8e 26 5f 80 88 ff ff kkkkkkkk..&_....
backtrace:
[<00000000733609e3>] fib_table_insert+0x978/0x1500
[<00000000856ed27d>] inet_rtm_newroute+0x129/0x220
[<00000000fcdfc00a>] rtnetlink_rcv_msg+0x397/0xa20
[<00000000cb85801a>] netlink_rcv_skb+0x132/0x380
[<00000000ebc991d2>] netlink_unicast+0x4c0/0x690
[<0000000014f62875>] netlink_sendmsg+0x929/0xe10
[<00000000bac9d967>] sock_sendmsg+0xc8/0x110
[<00000000223e6485>] ___sys_sendmsg+0x77a/0x8f0
[<000000002e94f880>] __sys_sendmsg+0xf7/0x250
[<00000000ccb1fa72>] do_syscall_64+0x14d/0x610
[<00000000ffbe3dae>] entry_SYSCALL_64_after_hwframe+0x49/0xbe
[<000000003a8b605b>] 0xffffffffffffffff
Fixes: 8cced9eff1 ("[NETNS]: Enable routing configuration in non-initial namespace.")
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
