Use pskb_may_pull() to make sure the optional fields are in skb linear
parts, so we can safely read them later.
It's easy to reproduce the issue with a net driver that supports paged
skb data. Just create a L2TPv3 over IP tunnel and then generates some
network traffic.
Once reproduced, rx err in /sys/kernel/debug/l2tp/tunnels will increase.
Changes in v4:
1. s/l2tp_v3_pull_opt/l2tp_v3_ensure_opt_in_linear/
2. s/tunnel->version != L2TP_HDR_VER_2/tunnel->version == L2TP_HDR_VER_3/
3. Add 'Fixes' in commit messages.
Changes in v3:
1. To keep consistency, move the code out of l2tp_recv_common.
2. Use "net" instead of "net-next", since this is a bug fix.
Changes in v2:
1. Only fix L2TPv3 to make code simple.
To fix both L2TPv3 and L2TPv2, we'd better refactor l2tp_recv_common.
It's complicated to do so.
2. Reloading pointers after pskb_may_pull
Fixes: f7faffa3ff ("l2tp: Add L2TPv3 protocol support")
Fixes: 0d76751fad ("l2tp: Add L2TPv3 IP encapsulation (no UDP) support")
Fixes: a32e0eec70 ("l2tp: introduce L2TPv3 IP encapsulation support for IPv6")
Signed-off-by: Jacob Wen <jian.w.wen@oracle.com>
Acked-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Call tun_set_real_num_queues() after the increment of tun->numqueues
since the former depends on it. Otherwise, the number of queues is not
correctly accounted for, which results to warnings similar to:
"vnet0 selects TX queue 11, but real number of TX queues is 11".
Fixes: 0b7959b625 ("tun: publish tfile after it's fully initialized")
Reported-and-tested-by: George Amanakis <gamanakis@gmail.com>
Signed-off-by: George Amanakis <gamanakis@gmail.com>
Signed-off-by: Stanislav Fomichev <sdf@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
skb->cb may contain data from previous layers (in an observed case
IPv4 with L3 Master Device). In the observed scenario, the data in
IPCB(skb)->frags was misinterpreted as IP6CB(skb)->frag_max_size,
eventually caused an unexpected IPv6 fragmentation in ip6_fragment()
through ip6_finish_output().
This patch clears IP6CB(skb), which potentially contains garbage data,
on the SRH ip4ip6 encapsulation.
Fixes: 32d99d0b67 ("ipv6: sr: add support for ip4ip6 encapsulation")
Signed-off-by: Yohei Kanemaru <yohei.kanemaru@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Toshiaki Makita says:
====================
virtio_net: Fix problems around XDP tx and napi_tx
While I'm looking into how to account standard tx counters on XDP tx
processing, I found several bugs around XDP tx and napi_tx.
Patch1: Fix oops on error path. Patch2 depends on this.
Patch2: Fix memory corruption on freeing xdp_frames with napi_tx enabled.
Patch3: Minor fix patch5 depends on.
Patch4: Fix memory corruption on processing xdp_frames when XDP is disabled.
Also patch5 depends on this.
Patch5: Fix memory corruption on processing xdp_frames while XDP is being
disabled.
Patch6: Minor fix patch7 depends on.
Patch7: Fix memory corruption on freeing sk_buff or xdp_frames when a normal
queue is reused for XDP and vise versa.
v2:
- patch5: Make rcu_assign_pointer/synchronize_net conditional instead of
_virtnet_set_queues.
- patch7: Use napi_consume_skb() instead of dev_consume_skb_any()
====================
Signed-off-by: Toshiaki Makita <makita.toshiaki@lab.ntt.co.jp>
Signed-off-by: David S. Miller <davem@davemloft.net>
We do not reset or free up unused buffers when enabling/disabling XDP,
so it can happen that xdp_frames are freed after disabling XDP or
sk_buffs are freed after enabling XDP on xdp tx queues.
Thus we need to handle both forms (xdp_frames and sk_buffs) regardless
of XDP setting.
One way to trigger this problem is to disable XDP when napi_tx is
enabled. In that case, virtnet_xdp_set() calls virtnet_napi_enable()
which kicks NAPI. The NAPI handler will call virtnet_poll_cleantx()
which invokes free_old_xmit_skbs() for queues which have been used by
XDP.
Note that even with this change we need to keep skipping
free_old_xmit_skbs() from NAPI handlers when XDP is enabled, because XDP
tx queues do not aquire queue locks.
- v2: Use napi_consume_skb() instead of dev_consume_skb_any()
Fixes: 4941d472bf ("virtio-net: do not reset during XDP set")
Signed-off-by: Toshiaki Makita <makita.toshiaki@lab.ntt.co.jp>
Acked-by: Jason Wang <jasowang@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
put_page() can work as a fallback for freeing xdp_frames, but the
appropriate way is to use xdp_return_frame().
Fixes: cac320c850 ("virtio_net: convert to use generic xdp_frame and xdp_return_frame API")
Signed-off-by: Toshiaki Makita <makita.toshiaki@lab.ntt.co.jp>
Acked-by: Jason Wang <jasowang@redhat.com>
Acked-by: Jesper Dangaard Brouer <brouer@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit 8dcc5b0ab0 ("virtio_net: fix ndo_xdp_xmit crash towards dev not
ready for XDP") tried to avoid access to unexpected sq while XDP is
disabled, but was not complete.
There was a small window which causes out of bounds sq access in
virtnet_xdp_xmit() while disabling XDP.
An example case of
- curr_queue_pairs = 6 (2 for SKB and 4 for XDP)
- online_cpu_num = xdp_queue_paris = 4
when XDP is enabled:
CPU 0 CPU 1
(Disabling XDP) (Processing redirected XDP frames)
virtnet_xdp_xmit()
virtnet_xdp_set()
_virtnet_set_queues()
set curr_queue_pairs (2)
check if rq->xdp_prog is not NULL
virtnet_xdp_sq(vi)
qp = curr_queue_pairs -
xdp_queue_pairs +
smp_processor_id()
= 2 - 4 + 1 = -1
sq = &vi->sq[qp] // out of bounds access
set xdp_queue_pairs (0)
rq->xdp_prog = NULL
Basically we should not change curr_queue_pairs and xdp_queue_pairs
while someone can read the values. Thus, when disabling XDP, assign NULL
to rq->xdp_prog first, and wait for RCU grace period, then change
xxx_queue_pairs.
Note that we need to keep the current order when enabling XDP though.
- v2: Make rcu_assign_pointer/synchronize_net conditional instead of
_virtnet_set_queues.
Fixes: 186b3c998c ("virtio-net: support XDP_REDIRECT")
Signed-off-by: Toshiaki Makita <makita.toshiaki@lab.ntt.co.jp>
Acked-by: Jason Wang <jasowang@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When XDP is disabled, curr_queue_pairs + smp_processor_id() can be
larger than max_queue_pairs.
There is no guarantee that we have enough XDP send queues dedicated for
each cpu when XDP is disabled, so do not count drops on sq in that case.
Fixes: 5b8f3c8d30 ("virtio_net: Add XDP related stats")
Signed-off-by: Toshiaki Makita <makita.toshiaki@lab.ntt.co.jp>
Acked-by: Jason Wang <jasowang@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When _virtnet_set_queues() failed we did not restore real_num_rx_queues.
Fix this by placing the change of real_num_rx_queues after
_virtnet_set_queues().
This order is also in line with virtnet_set_channels().
Fixes: 4941d472bf ("virtio-net: do not reset during XDP set")
Signed-off-by: Toshiaki Makita <makita.toshiaki@lab.ntt.co.jp>
Acked-by: Jason Wang <jasowang@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When napi_tx is enabled, virtnet_poll_cleantx() called
free_old_xmit_skbs() even for xdp send queue.
This is bogus since the queue has xdp_frames, not sk_buffs, thus mangled
device tx bytes counters because skb->len is meaningless value, and even
triggered oops due to general protection fault on freeing them.
Since xdp send queues do not aquire locks, old xdp_frames should be
freed only in virtnet_xdp_xmit(), so just skip free_old_xmit_skbs() for
xdp send queues.
Similarly virtnet_poll_tx() called free_old_xmit_skbs(). This NAPI
handler is called even without calling start_xmit() because cb for tx is
by default enabled. Once the handler is called, it enabled the cb again,
and then the handler would be called again. We don't need this handler
for XDP, so don't enable cb as well as not calling free_old_xmit_skbs().
Also, we need to disable tx NAPI when disabling XDP, so
virtnet_poll_tx() can safely access curr_queue_pairs and
xdp_queue_pairs, which are not atomically updated while disabling XDP.
Fixes: b92f1e6751 ("virtio-net: transmit napi")
Fixes: 7b0411ef4a ("virtio-net: clean tx descriptors from rx napi")
Signed-off-by: Toshiaki Makita <makita.toshiaki@lab.ntt.co.jp>
Acked-by: Jason Wang <jasowang@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit 4e09ff5362 ("virtio-net: disable NAPI only when enabled during
XDP set") tried to fix inappropriate NAPI enabling/disabling when
!netif_running(), but was not complete.
On error path virtio_net could enable NAPI even when !netif_running().
This can cause enabling NAPI twice on virtnet_open(), which would
trigger BUG_ON() in napi_enable().
Fixes: 4941d472bf ("virtio-net: do not reset during XDP set")
Signed-off-by: Toshiaki Makita <makita.toshiaki@lab.ntt.co.jp>
Acked-by: Jason Wang <jasowang@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Lorenzo Bianconi says:
====================
erspan: always reports output key to userspace
Erspan protocol relies on output key to set session id header field.
However TUNNEL_KEY bit is cleared in order to not add key field to
the external GRE header and so the configured o_key is not reported
to userspace.
Fix the issue adding TUNNEL_KEY bit to the o_flags parameter dumping
device info
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
As Erspan_v4, Erspan_v6 protocol relies on o_key to configure
session id header field. However TUNNEL_KEY bit is cleared in
ip6erspan_tunnel_xmit since ERSPAN protocol does not set the key field
of the external GRE header and so the configured o_key is not reported
to userspace. The issue can be triggered with the following reproducer:
$ip link add ip6erspan1 type ip6erspan local 2000::1 remote 2000::2 \
key 1 seq erspan_ver 1
$ip link set ip6erspan1 up
ip -d link sh ip6erspan1
ip6erspan1@NONE: <BROADCAST,MULTICAST> mtu 1422 qdisc noop state DOWN mode DEFAULT
link/ether ba:ff:09:24:c3:0e brd ff:ff:ff:ff:ff:ff promiscuity 0 minmtu 68 maxmtu 1500
ip6erspan remote 2000::2 local 2000::1 encaplimit 4 flowlabel 0x00000 ikey 0.0.0.1 iseq oseq
Fix the issue adding TUNNEL_KEY bit to the o_flags parameter in
ip6gre_fill_info
Fixes: 5a963eb61b ("ip6_gre: Add ERSPAN native tunnel support")
Signed-off-by: Lorenzo Bianconi <lorenzo.bianconi@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Erspan protocol (version 1 and 2) relies on o_key to configure
session id header field. However TUNNEL_KEY bit is cleared in
erspan_xmit since ERSPAN protocol does not set the key field
of the external GRE header and so the configured o_key is not reported
to userspace. The issue can be triggered with the following reproducer:
$ip link add erspan1 type erspan local 192.168.0.1 remote 192.168.0.2 \
key 1 seq erspan_ver 1
$ip link set erspan1 up
$ip -d link sh erspan1
erspan1@NONE: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1450 qdisc pfifo_fast state UNKNOWN mode DEFAULT
link/ether 52:aa:99:95:9a:b5 brd ff:ff:ff:ff:ff:ff promiscuity 0 minmtu 68 maxmtu 1500
erspan remote 192.168.0.2 local 192.168.0.1 ttl inherit ikey 0.0.0.1 iseq oseq erspan_index 0
Fix the issue adding TUNNEL_KEY bit to the o_flags parameter in
ipgre_fill_info
Fixes: 84e54fe0a5 ("gre: introduce native tunnel support for ERSPAN")
Signed-off-by: Lorenzo Bianconi <lorenzo.bianconi@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
After a timeout event caused by for example a broadcast storm, when
the MAC and PHY are reset, the BQL TX queue needs to be reset as
well. Otherwise, the device will exhibit severe performance issues
even after the storm has ended.
Co-authored-by: David Gounaris <david.gounaris@infinera.com>
Signed-off-by: Mathias Thore <mathias.thore@infinera.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Johannes Berg says:
====================
various compat ioctl fixes
Back a long time ago, I already fixed a few of these by passing
the size of the struct ifreq to do_sock_ioctl(). However, Robert
found more cases, and now it won't be as simple because we'd have
to pass that down all the way to e.g. bond_do_ioctl() which isn't
really feasible.
Therefore, restore the old code.
While looking at why SIOCGIFNAME was broken, I realized that Al
had removed that case - which had been handled in an explicit
separate function - as well, and looking through his work at the
time I saw that bond ioctls were also affected by the erroneous
removal.
I've restored SIOCGIFNAME and bond ioctls by going through the
(now renamed) dev_ifsioc() instead of reintroducing their own
helper functions, which I hope is correct but have only tested
with SIOCGIFNAME.
====================
Acked-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
Same story as before, these use struct ifreq and thus need
to be read with the shorter version to not cause faults.
Cc: stable@vger.kernel.org
Fixes: f92d4fc953 ("kill bond_ioctl()")
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As reported by Robert O'Callahan in
https://bugzilla.kernel.org/show_bug.cgi?id=202273
reverting the previous changes in this area broke
the SIOCGIFNAME ioctl in compat again (I'd previously
fixed it after his previous report of breakage in
https://bugzilla.kernel.org/show_bug.cgi?id=199469).
This is obviously because I fixed SIOCGIFNAME more or
less by accident.
Fix it explicitly now by making it pass through the
restored compat translation code.
Cc: stable@vger.kernel.org
Fixes: 4cf808e7ac ("kill dev_ifname32()")
Reported-by: Robert O'Callahan <robert@ocallahan.org>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This reverts commit bf4405737f ("kill dev_ifsioc()").
This wasn't really unused as implied by the original commit,
it still handles the copy to/from user differently, and the
commit thus caused issues such as
https://bugzilla.kernel.org/show_bug.cgi?id=199469
and
https://bugzilla.kernel.org/show_bug.cgi?id=202273
However, deviating from a strict revert, rename dev_ifsioc()
to compat_ifreq_ioctl() to be clearer as to its purpose and
add a comment.
Cc: stable@vger.kernel.org
Fixes: bf4405737f ("kill dev_ifsioc()")
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This reverts commit 1cebf8f143 ("socket: fix struct ifreq
size in compat ioctl"), it's a bugfix for another commit that
I'll revert next.
This is not a 'perfect' revert, I'm keeping some coding style
intact rather than revert to the state with indentation errors.
Cc: stable@vger.kernel.org
Fixes: 1cebf8f143 ("socket: fix struct ifreq size in compat ioctl")
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pull networking fixes from David Miller:
1) Need to save away the IV across tls async operations, from Dave
Watson.
2) Upon successful packet processing, we should liberate the SKB with
dev_consume_skb{_irq}(). From Yang Wei.
3) Only apply RX hang workaround on effected macb chips, from Harini
Katakam.
4) Dummy netdev need a proper namespace assigned to them, from Josh
Elsasser.
5) Some paths of nft_compat run lockless now, and thus we need to use a
proper refcnt_t. From Florian Westphal.
6) Avoid deadlock in mlx5 by doing IRQ locking, from Moni Shoua.
7) netrom does not refcount sockets properly wrt. timers, fix that by
using the sock timer API. From Cong Wang.
8) Fix locking of inexact inserts of xfrm policies, from Florian
Westphal.
9) Missing xfrm hash generation bump, also from Florian.
10) Missing of_node_put() in hns driver, from Yonglong Liu.
11) Fix DN_IFREQ_SIZE, from Johannes Berg.
12) ip6mr notifier is invoked during traversal of wrong table, from Nir
Dotan.
13) TX promisc settings not performed correctly in qed, from Manish
Chopra.
14) Fix OOB access in vhost, from Jason Wang.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (52 commits)
MAINTAINERS: Add entry for XDP (eXpress Data Path)
net: set default network namespace in init_dummy_netdev()
net: b44: replace dev_kfree_skb_xxx by dev_consume_skb_xxx for drop profiles
net: caif: call dev_consume_skb_any when skb xmit done
net: 8139cp: replace dev_kfree_skb_irq by dev_consume_skb_irq for drop profiles
net: macb: Apply RXUBR workaround only to versions with errata
net: ti: replace dev_kfree_skb_irq by dev_consume_skb_irq for drop profiles
net: apple: replace dev_kfree_skb_irq by dev_consume_skb_irq for drop profiles
net: amd8111e: replace dev_kfree_skb_irq by dev_consume_skb_irq
net: alteon: replace dev_kfree_skb_irq by dev_consume_skb_irq
net: tls: Fix deadlock in free_resources tx
net: tls: Save iv in tls_rec for async crypto requests
vhost: fix OOB in get_rx_bufs()
qed: Fix stack out of bounds bug
qed: Fix system crash in ll2 xmit
qed: Fix VF probe failure while FLR
qed: Fix LACP pdu drops for VFs
qed: Fix bug in tx promiscuous mode settings
net: i825xx: replace dev_kfree_skb_irq by dev_consume_skb_irq for drop profiles
netfilter: ipt_CLUSTERIP: fix warning unused variable cn
...
Add multiple people as maintainers for XDP, sorted alphabetically.
XDP is also tied to driver level support and code, but we cannot add all
drivers to the list. Instead K: and N: match on 'xdp' in hope to catch some
of those changes in drivers.
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Acked-by: David S. Miller <davem@davemloft.net>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Assign a default net namespace to netdevs created by init_dummy_netdev().
Fixes a NULL pointer dereference caused by busy-polling a socket bound to
an iwlwifi wireless device, which bumps the per-net BUSYPOLLRXPACKETS stat
if napi_poll() received packets:
BUG: unable to handle kernel NULL pointer dereference at 0000000000000190
IP: napi_busy_loop+0xd6/0x200
Call Trace:
sock_poll+0x5e/0x80
do_sys_poll+0x324/0x5a0
SyS_poll+0x6c/0xf0
do_syscall_64+0x6b/0x1f0
entry_SYSCALL_64_after_hwframe+0x3d/0xa2
Fixes: 7db6b048da ("net: Commonize busy polling code to focus on napi_id instead of socket")
Signed-off-by: Josh Elsasser <jelsasser@appneta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The skb should be freed by dev_consume_skb_any() in b44_start_xmit()
when bounce_skb is used. The skb is be replaced by bounce_skb, so the
original skb should be consumed(not drop).
dev_consume_skb_irq() should be called in b44_tx() when skb xmit
done. It makes drop profiles(dropwatch, perf) more friendly.
Signed-off-by: Yang Wei <yang.wei9@zte.com.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
The skb shouled be consumed when xmit done, it makes drop profiles
(dropwatch, perf) more friendly.
dev_kfree_skb_irq()/kfree_skb() shouled be replaced by
dev_consume_skb_any(), it makes code cleaner.
Signed-off-by: Yang Wei <yang.wei9@zte.com.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
dev_consume_skb_irq() should be called in cp_tx() when skb xmit
done. It makes drop profiles(dropwatch, perf) more friendly.
Signed-off-by: Yang Wei <yang.wei9@zte.com.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
The interrupt handler contains a workaround for RX hang applicable
to Zynq and AT91RM9200 only. Subsequent versions do not need this
workaround. This workaround unnecessarily resets RX whenever RX used
bit read is observed, which can be often under heavy traffic. There
is no other action performed on RX UBR interrupt. Hence introduce a
CAPS mask; enable this interrupt and workaround only on affected
versions.
Signed-off-by: Harini Katakam <harini.katakam@xilinx.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
dev_consume_skb_irq() should be called in cpmac_end_xmit() when
xmit done. It makes drop profiles more friendly.
Signed-off-by: Yang Wei <yang.wei9@zte.com.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
dev_consume_skb_irq() should be called in bmac_txdma_intr() when
xmit done. It makes drop profiles more friendly.
Signed-off-by: Yang Wei <yang.wei9@zte.com.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
dev_consume_skb_irq() should be called in amd8111e_tx() when xmit
done. It makes drop profiles more friendly.
Signed-off-by: Yang Wei <yang.wei9@zte.com.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
dev_consume_skb_irq() should be called in ace_tx_int() when xmit
done. It makes drop profiles more friendly.
Signed-off-by: Yang Wei <yang.wei9@zte.com.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
If there are outstanding async tx requests (when crypto returns EINPROGRESS),
there is a potential deadlock: the tx work acquires the lock, while we
cancel_delayed_work_sync() while holding the lock. Drop the lock while waiting
for the work to complete.
Fixes: a42055e8d2 ("Add support for async encryption of records...")
Signed-off-by: Dave Watson <davejwatson@fb.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
aead_request_set_crypt takes an iv pointer, and we change the iv
soon after setting it. Some async crypto algorithms don't save the iv,
so we need to save it in the tls_rec for async requests.
Found by hardcoding x64 aesni to use async crypto manager (to test the async
codepath), however I don't think this combination can happen in the wild.
Presumably other hardware offloads will need this fix, but there have been
no user reports.
Fixes: a42055e8d2 ("Add support for async encryption of records...")
Signed-off-by: Dave Watson <davejwatson@fb.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
After batched used ring updating was introduced in commit e2b3b35eb9
("vhost_net: batch used ring update in rx"). We tend to batch heads in
vq->heads for more than one packet. But the quota passed to
get_rx_bufs() was not correctly limited, which can result a OOB write
in vq->heads.
headcount = get_rx_bufs(vq, vq->heads + nvq->done_idx,
vhost_len, &in, vq_log, &log,
likely(mergeable) ? UIO_MAXIOV : 1);
UIO_MAXIOV was still used which is wrong since we could have batched
used in vq->heads, this will cause OOB if the next buffer needs more
than 960 (1024 (UIO_MAXIOV) - 64 (VHOST_NET_BATCH)) heads after we've
batched 64 (VHOST_NET_BATCH) heads:
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
=============================================================================
BUG kmalloc-8k (Tainted: G B ): Redzone overwritten
-----------------------------------------------------------------------------
INFO: 0x00000000fd93b7a2-0x00000000f0713384. First byte 0xa9 instead of 0xcc
INFO: Allocated in alloc_pd+0x22/0x60 age=3933677 cpu=2 pid=2674
kmem_cache_alloc_trace+0xbb/0x140
alloc_pd+0x22/0x60
gen8_ppgtt_create+0x11d/0x5f0
i915_ppgtt_create+0x16/0x80
i915_gem_create_context+0x248/0x390
i915_gem_context_create_ioctl+0x4b/0xe0
drm_ioctl_kernel+0xa5/0xf0
drm_ioctl+0x2ed/0x3a0
do_vfs_ioctl+0x9f/0x620
ksys_ioctl+0x6b/0x80
__x64_sys_ioctl+0x11/0x20
do_syscall_64+0x43/0xf0
entry_SYSCALL_64_after_hwframe+0x44/0xa9
INFO: Slab 0x00000000d13e87af objects=3 used=3 fp=0x (null) flags=0x200000000010201
INFO: Object 0x0000000003278802 @offset=17064 fp=0x00000000e2e6652b
Fixing this by allocating UIO_MAXIOV + VHOST_NET_BATCH iovs for
vhost-net. This is done through set the limitation through
vhost_dev_init(), then set_owner can allocate the number of iov in a
per device manner.
This fixes CVE-2018-16880.
Fixes: e2b3b35eb9 ("vhost_net: batch used ring update in rx")
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Manish Chopra says:
====================
qed: Bug fixes
This series have SR-IOV and some general fixes.
Please consider applying it to "net"
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Cache number of fragments in the skb locally as in case
of linear skb (with zero fragments), tx completion
(or freeing of skb) may happen before driver tries
to get number of frgaments from the skb which could
lead to stale access to an already freed skb.
Signed-off-by: Manish Chopra <manishc@marvell.com>
Signed-off-by: Ariel Elior <aelior@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
VFs may hit VF-PF channel timeout while probing, as in some
cases it was observed that VF FLR and VF "acquire" message
transaction (i.e first message from VF to PF in VF's probe flow)
could occur simultaneously which could lead VF to fail sending
"acquire" message to PF as VF is marked disabled from HW perspective
due to FLR, which will result into channel timeout and VF probe failure.
In such cases, try retrying VF "acquire" message so that in later
attempts it could be successful to pass message to PF after the VF
FLR is completed and can be probed successfully.
Signed-off-by: Manish Chopra <manishc@marvell.com>
Signed-off-by: Ariel Elior <aelior@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
VF is always configured to drop control frames
(with reserved mac addresses) but to work LACP
on the VFs, it would require LACP control frames
to be forwarded or transmitted successfully.
This patch fixes this in such a way that trusted VFs
(marked through ndo_set_vf_trust) would be allowed to
pass the control frames such as LACP pdus.
Signed-off-by: Manish Chopra <manishc@marvell.com>
Signed-off-by: Ariel Elior <aelior@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When running tx switched traffic between VNICs
created via a bridge(to which VFs are added),
adapter drops the unicast packets in tx flow due to
VNIC's ucast mac being unknown to it. But VF interfaces
being in promiscuous mode should have caused adapter
to accept all the unknown ucast packets. Later, it
was found that driver doesn't really configure tx
promiscuous mode settings to accept all unknown unicast macs.
This patch fixes tx promiscuous mode settings to accept all
unknown/unmatched unicast macs and works out the scenario.
Signed-off-by: Manish Chopra <manishc@marvell.com>
Signed-off-by: Ariel Elior <aelior@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
dev_consume_skb_irq() should be called in i596_interrupt() when skb
xmit done. It makes drop profiles(dropwatch, perf) more friendly.
Signed-off-by: Yang Wei <albin_yang@163.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pablo Neira Ayuso says:
====================
Netfilter/IPVS fixes for net
The following patchset contains Netfilter/IPVS fixes for your net tree:
1) The nftnl mutex is now per-netns, therefore use reference counter
for matches and targets to deal with concurrent updates from netns.
Moreover, place extensions in a pernet list. Patches from Florian Westphal.
2) Bail out with EINVAL in case of negative timeouts via setsockopt()
through ip_vs_set_timeout(), from ZhangXiaoxu.
3) Spurious EINVAL on ebtables 32bit binary with 64bit kernel, also
from Florian.
4) Reset TCP option header parser in case of fingerprint mismatch,
otherwise follow up overlapping fingerprint definitions including
TCP options do not work, from Fernando Fernandez Mancera.
5) Compilation warning in ipt_CLUSTER with CONFIG_PROC_FS unset.
From Anders Roxell.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
This reverts commit 2830bf6f05.
The underlying assumption that one sparse section belongs into a single
numa node doesn't hold really. Robert Shteynfeld has reported a boot
failure. The boot log was not captured but his memory layout is as
follows:
Early memory node ranges
node 1: [mem 0x0000000000001000-0x0000000000090fff]
node 1: [mem 0x0000000000100000-0x00000000dbdf8fff]
node 1: [mem 0x0000000100000000-0x0000001423ffffff]
node 0: [mem 0x0000001424000000-0x0000002023ffffff]
This means that node0 starts in the middle of a memory section which is
also in node1. memmap_init_zone tries to initialize padding of a
section even when it is outside of the given pfn range because there are
code paths (e.g. memory hotplug) which assume that the full worth of
memory section is always initialized.
In this particular case, though, such a range is already intialized and
most likely already managed by the page allocator. Scribbling over
those pages corrupts the internal state and likely blows up when any of
those pages gets used.
Reported-by: Robert Shteynfeld <robert.shteynfeld@gmail.com>
Fixes: 2830bf6f05 ("mm, memory_hotplug: initialize struct pages for the full memory section")
Cc: stable@kernel.org
Signed-off-by: Michal Hocko <mhocko@suse.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
When CONFIG_PROC_FS isn't set the variable cn isn't used.
net/ipv4/netfilter/ipt_CLUSTERIP.c: In function ‘clusterip_net_exit’:
net/ipv4/netfilter/ipt_CLUSTERIP.c:849:24: warning: unused variable ‘cn’ [-Wunused-variable]
struct clusterip_net *cn = clusterip_pernet(net);
^~
Rework so the variable 'cn' is declared inside "#ifdef CONFIG_PROC_FS".
Fixes: b12f7bad5a ("netfilter: ipt_CLUSTERIP: remove wrong WARN_ON_ONCE in netns exit routine")
Signed-off-by: Anders Roxell <anders.roxell@linaro.org>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
When we check the tcp options of a packet and it doesn't match the current
fingerprint, the tcp packet option pointer must be restored to its initial
value in order to do the proper tcp options check for the next fingerprint.
Here we can see an example.
Assumming the following fingerprint base with two lines:
S10:64:1:60:M*,S,T,N,W6: Linux:3.0::Linux 3.0
S20:64:1:60:M*,S,T,N,W7: Linux:4.19:arch:Linux 4.1
Where TCP options are the last field in the OS signature, all of them overlap
except by the last one, ie. 'W6' versus 'W7'.
In case a packet for Linux 4.19 kicks in, the osf finds no matching because the
TCP options pointer is updated after checking for the TCP options in the first
line.
Therefore, reset pointer back to where it should be.
Fixes: 11eeef41d5 ("netfilter: passive OS fingerprint xtables match")
Signed-off-by: Fernando Fernandez Mancera <ffmancera@riseup.net>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Unlike ip(6)tables ebtables only counts user-defined chains.
The effect is that a 32bit ebtables binary on a 64bit kernel can do
'ebtables -N FOO' only after adding at least one rule, else the request
fails with -EINVAL.
This is a similar fix as done in
3f1e53abff ("netfilter: ebtables: don't attempt to allocate 0-sized compat array").
Fixes: 7d7d7e0211 ("netfilter: compat: reject huge allocation requests")
Reported-by: Francesco Ruggeri <fruggeri@arista.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Duec to a typo, mv88e6390_serdes_irq_setup() calls itself, rather than
mv88e6390x_serdes_irq_setup(). It then blows the stack, and shortly
after the machine blows up.
Fixes: 2defda1f4b ("net: dsa: mv88e6xxx: Add support for SERDES on ports 2-8 for 6390X")
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
When the MC route socket is closed, mroute_clean_tables() is called to
cleanup existing routes. Mistakenly notifiers call was put on the cleanup
of the unresolved MC route entries cache.
In a case where the MC socket closes before an unresolved route expires,
the notifier call leads to a crash, caused by the driver trying to
increment a non initialized refcount_t object [1] and then when handling
is done, to decrement it [2]. This was detected by a test recently added in
commit 6d4efada3b ("selftests: forwarding: Add multicast routing test").
Fix that by putting notifiers call on the resolved entries traversal,
instead of on the unresolved entries traversal.
[1]
[ 245.748967] refcount_t: increment on 0; use-after-free.
[ 245.754829] WARNING: CPU: 3 PID: 3223 at lib/refcount.c:153 refcount_inc_checked+0x2b/0x30
...
[ 245.802357] Hardware name: Mellanox Technologies Ltd. MSN2740/SA001237, BIOS 5.6.5 06/07/2016
[ 245.811873] RIP: 0010:refcount_inc_checked+0x2b/0x30
...
[ 245.907487] Call Trace:
[ 245.910231] mlxsw_sp_router_fib_event.cold.181+0x42/0x47 [mlxsw_spectrum]
[ 245.917913] notifier_call_chain+0x45/0x7
[ 245.922484] atomic_notifier_call_chain+0x15/0x20
[ 245.927729] call_fib_notifiers+0x15/0x30
[ 245.932205] mroute_clean_tables+0x372/0x3f
[ 245.936971] ip6mr_sk_done+0xb1/0xc0
[ 245.940960] ip6_mroute_setsockopt+0x1da/0x5f0
...
[2]
[ 246.128487] refcount_t: underflow; use-after-free.
[ 246.133859] WARNING: CPU: 0 PID: 7 at lib/refcount.c:187 refcount_sub_and_test_checked+0x4c/0x60
[ 246.183521] Hardware name: Mellanox Technologies Ltd. MSN2740/SA001237, BIOS 5.6.5 06/07/2016
...
[ 246.193062] Workqueue: mlxsw_core_ordered mlxsw_sp_router_fibmr_event_work [mlxsw_spectrum]
[ 246.202394] RIP: 0010:refcount_sub_and_test_checked+0x4c/0x60
...
[ 246.298889] Call Trace:
[ 246.301617] refcount_dec_and_test_checked+0x11/0x20
[ 246.307170] mlxsw_sp_router_fibmr_event_work.cold.196+0x47/0x78 [mlxsw_spectrum]
[ 246.315531] process_one_work+0x1fa/0x3f0
[ 246.320005] worker_thread+0x2f/0x3e0
[ 246.324083] kthread+0x118/0x130
[ 246.327683] ? wq_update_unbound_numa+0x1b0/0x1b0
[ 246.332926] ? kthread_park+0x80/0x80
[ 246.337013] ret_from_fork+0x1f/0x30
Fixes: 088aa3eec2 ("ip6mr: Support fib notifications")
Signed-off-by: Nir Dotan <nird@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Digging through the ioctls with Al because of the previous
patches, we found that on 64-bit decnet's dn_dev_ioctl()
is wrong, because struct ifreq::ifr_ifru is actually 24
bytes (not 16 as expected from struct sockaddr) due to the
ifru_map and ifru_settings members.
Clearly, decnet expects the ioctl to be called with a struct
like
struct ifreq_dn {
char ifr_name[IFNAMSIZ];
struct sockaddr_dn ifr_addr;
};
since it does
struct ifreq *ifr = ...;
struct sockaddr_dn *sdn = (struct sockaddr_dn *)&ifr->ifr_addr;
This means that DN_IFREQ_SIZE is too big for what it wants on
64-bit, as it is
sizeof(struct ifreq) - sizeof(struct sockaddr) +
sizeof(struct sockaddr_dn)
This assumes that sizeof(struct sockaddr) is the size of ifr_ifru
but that isn't true.
Fix this to use offsetof(struct ifreq, ifr_ifru).
This indeed doesn't really matter much - the result is that we
copy in/out 8 bytes more than we should on 64-bit platforms. In
case the "struct ifreq_dn" lands just on the end of a page though
it might lead to faults.
As far as I can tell, it has been like this forever, so it seems
very likely that nobody cares.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If phy_power_on() fails in rk_gmac_powerup(), clocks are left enabled.
Found by Linux Driver Verification project (linuxtesting.org).
Signed-off-by: Alexey Khoroshilov <khoroshilov@ispras.ru>
Signed-off-by: David S. Miller <davem@davemloft.net>
