netfilter: nf_tables: cache device name in flowtable object
Devices going away have to grab the nfnl_lock from the netdev event path
to avoid races with control plane updates.
However, netlink dumps in netfilter do not hold nfnl_lock mutex. Cache
the device name into the objects to avoid an use-after-free situation
for a device that is going away.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Paolo Abeni [Thu, 22 Mar 2018 10:08:50 +0000 (11:08 +0100)]
netfilter: drop template ct when conntrack is skipped.
The ipv4 nf_ct code currently skips the nf_conntrak_in() call
for fragmented packets. As a results later matches/target can end
up manipulating template ct entry instead of 'real' ones.
Exploiting the above, syzbot found a way to trigger the following
splat:
WARNING: CPU: 1 PID: 4242 at net/netfilter/xt_cluster.c:55
xt_cluster_mt+0x6c1/0x840 net/netfilter/xt_cluster.c:127
Kernel panic - not syncing: panic_on_warn set ...
Florian Westphal [Sun, 18 Mar 2018 18:22:39 +0000 (19:22 +0100)]
netfilter: nf_tables: permit second nat hook if colliding hook is going away
Sergei Trofimovich reported that restoring an nft ruleset doesn't work
anymore unless old rule content is flushed first.
The problem stems from a recent change designed to prevent multiple nat
hooks at the same hook point locations and nftables transaction model.
A 'flush ruleset' won't take effect until the entire transaction has
completed.
So, if one has a nft.rules file that contains a 'flush ruleset',
followed by a nat hook register request, then 'nft -f file' will work,
but running 'nft -f file' again will fail with -EBUSY.
Reason is that nftables will place the flush/removal requests in the
transaction list, but it will not act on the removal until after all new
rules are in place.
The netfilter core will therefore get request to register a new nat
hook before the old one is removed -- this now fails as the netfilter
core can't know the existing hook is staged for removal.
To fix this, we can search the transaction log when a hook collision
is detected. The collision is okay if
1. there is a delete request pending for the nat hook that is already
registered.
2. there is no second add request for a matching nat hook.
This is required to only apply the exception once.
Fixes: f92b40a8b2645 ("netfilter: core: only allow one nat hook per hook point") Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Florian Westphal [Wed, 14 Mar 2018 12:37:58 +0000 (13:37 +0100)]
netfilter: nf_tables: meter: pick a set backend that supports updates
in nftables, 'meter' can be used to instantiate a hash-table at run
time:
rule add filter forward iif "internal" meter hostacct { ip saddr counter}
nft list meter ip filter hostacct
table ip filter {
meter hostacct {
type ipv4_addr
elements = { 192.168.0.1 : counter packets 8 bytes 2672, ..
because elemets get added on the fly, the kernel must chose a set
backend type that implements the ->update() function, otherwise
rule insertion fails with EOPNOTSUPP.
Therefore, skip set types that lack ->update, and also
make sure we do not discard a (bad) candidate when we did yet
find any candidate at all. This could happen when userspace prefers
low memory footprint -- the set implementation currently checked might
not be a fit at all. Make sure we pick it anyway (!bops). In
case next candidate is a better fix, it will be chosen instead.
But in case nothing else is found we at least have a non-ideal
match rather than no match at all.
David S. Miller [Tue, 20 Mar 2018 01:14:27 +0000 (21:14 -0400)]
Merge branch 'phy-relax-error-checking'
Grygorii Strashko says:
====================
net: phy: relax error checking when creating sysfs link netdev->phydev
Some ethernet drivers (like TI CPSW) may connect and manage >1 Net PHYs per
one netdevice, as result such drivers will produce warning during system
boot and fail to connect second phy to netdevice when PHYLIB framework
will try to create sysfs link netdev->phydev for second PHY
in phy_attach_direct(), because sysfs link with the same name has been
created already for the first PHY.
As result, second CPSW external port will became unusable.
This regression was introduced by commits: 5568363f0cb3 ("net: phy: Create sysfs reciprocal links for attached_dev/phydev" a3995460491d ("net: phy: Relax error checking on sysfs_create_link()"
Patch 1: exports sysfs_create_link_nowarn() function as preparation for Patch 2.
Patch 2: relaxes error checking when PHYLIB framework is creating sysfs
link netdev->phydev in phy_attach_direct(), suppresses warning by using
sysfs_create_link_nowarn() and adds error message instead, so links creation
failure is not fatal any more and system can continue working,
which fixes TI CPSW issue and makes boot logs accessible
in case of NFS boot, for example.
net: phy: relax error checking when creating sysfs link netdev->phydev
Some ethernet drivers (like TI CPSW) may connect and manage >1 Net PHYs per
one netdevice, as result such drivers will produce warning during system
boot and fail to connect second phy to netdevice when PHYLIB framework
will try to create sysfs link netdev->phydev for second PHY
in phy_attach_direct(), because sysfs link with the same name has been
created already for the first PHY. As result, second CPSW external
port will became unusable.
Fix it by relaxing error checking when PHYLIB framework is creating sysfs
link netdev->phydev in phy_attach_direct(), suppressing warning by using
sysfs_create_link_nowarn() and adding error message instead.
After this change links (phy->netdev and netdev->phy) creation failure is not
fatal any more and system can continue working, which fixes TI CPSW issue.
Cc: Florian Fainelli <f.fainelli@gmail.com> Cc: Andrew Lunn <andrew@lunn.ch> Fixes: a3995460491d ("net: phy: Relax error checking on sysfs_create_link()") Signed-off-by: Grygorii Strashko <grygorii.strashko@ti.com> Signed-off-by: David S. Miller <davem@davemloft.net>
The sysfs_create_link_nowarn() is going to be used in phylib framework in
subsequent patch which can be built as module. Hence, export
sysfs_create_link_nowarn() to avoid build errors.
Cc: Florian Fainelli <f.fainelli@gmail.com> Cc: Andrew Lunn <andrew@lunn.ch> Fixes: a3995460491d ("net: phy: Relax error checking on sysfs_create_link()") Signed-off-by: Grygorii Strashko <grygorii.strashko@ti.com> Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Ronak Doshi [Fri, 16 Mar 2018 21:49:19 +0000 (14:49 -0700)]
vmxnet3: use correct flag to indicate LRO feature
'Commit 45dac1d6ea04 ("vmxnet3: Changes for vmxnet3 adapter version 2
(fwd)")' introduced a flag "lro" in structure vmxnet3_adapter which is
used to indicate whether LRO is enabled or not. However, the patch
did not set the flag and hence it was never exercised.
So, when LRO is enabled, it resulted in poor TCP performance due to
delayed acks. This issue is seen with packets which are larger than
the mss getting a delayed ack rather than an immediate ack, thus
resulting in high latency.
This patch removes the lro flag and directly uses device features
against NETIF_F_LRO to check if lro is enabled.
Fixes: 45dac1d6ea04 ("vmxnet3: Changes for vmxnet3 adapter version 2 (fwd)") Reported-by: Rachel Lunnon <rachel_lunnon@stormagic.com> Signed-off-by: Ronak Doshi <doshir@vmware.com> Acked-by: Shrikrishna Khare <skhare@vmware.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Ronak Doshi [Fri, 16 Mar 2018 21:47:54 +0000 (14:47 -0700)]
vmxnet3: avoid xmit reset due to a race in vmxnet3
The field txNumDeferred is used by the driver to keep track of the number
of packets it has pushed to the emulation. The driver increments it on
pushing the packet to the emulation and the emulation resets it to 0 at
the end of the transmit.
There is a possibility of a race either when (a) ESX is under heavy load or
(b) workload inside VM is of low packet rate.
This race results in xmit hangs when network coalescing is disabled. This
change creates a local copy of txNumDeferred and uses it to perform ring
arithmetic.
Reported-by: Noriho Tanaka <ntanaka@vmware.com> Signed-off-by: Ronak Doshi <doshir@vmware.com> Acked-by: Shrikrishna Khare <skhare@vmware.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Sat, 17 Mar 2018 23:53:29 +0000 (19:53 -0400)]
Merge branch 'tcf_foo_init-NULL-deref'
Davide Caratti says:
====================
net/sched: fix NULL dereference in the error path of .init()
with several TC actions it's possible to see NULL pointer dereference,
when the .init() function calls tcf_idr_alloc(), fails at some point and
then calls tcf_idr_release(): this series fixes all them introducing
non-NULL tests in the .cleanup() function.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Davide Caratti [Thu, 15 Mar 2018 23:00:57 +0000 (00:00 +0100)]
net/sched: fix NULL dereference on the error path of tcf_skbmod_init()
when the following command
# tc action replace action skbmod swap mac index 100
is run for the first time, and tcf_skbmod_init() fails to allocate struct
tcf_skbmod_params, tcf_skbmod_cleanup() calls kfree_rcu(NULL), thus
causing the following error:
Davide Caratti [Thu, 15 Mar 2018 23:00:56 +0000 (00:00 +0100)]
net/sched: fix NULL dereference in the error path of tcf_sample_init()
when the following command
# tc action add action sample rate 100 group 100 index 100
is run for the first time, and psample_group_get(100) fails to create a
new group, tcf_sample_cleanup() calls psample_group_put(NULL), thus
causing the following error:
Davide Caratti [Thu, 15 Mar 2018 23:00:55 +0000 (00:00 +0100)]
net/sched: fix NULL dereference in the error path of tunnel_key_init()
when the following command
# tc action add action tunnel_key unset index 100
is run for the first time, and tunnel_key_init() fails to allocate struct
tcf_tunnel_key_params, tunnel_key_release() dereferences NULL pointers.
This causes the following error:
Davide Caratti [Thu, 15 Mar 2018 23:00:54 +0000 (00:00 +0100)]
net/sched: fix NULL dereference in the error path of tcf_csum_init()
when the following command
# tc action add action csum udp continue index 100
is run for the first time, and tcf_csum_init() fails allocating struct
tcf_csum, tcf_csum_cleanup() calls kfree_rcu(NULL,...). This causes the
following error:
fix this in tcf_csum_cleanup(), ensuring that kfree_rcu(param, ...) is
called only when param is not NULL.
Fixes: 9c5f69bbd75a ("net/sched: act_csum: don't use spinlock in the fast path") Signed-off-by: Davide Caratti <dcaratti@redhat.com> Acked-by: Jiri Pirko <jiri@mellanox.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Davide Caratti [Thu, 15 Mar 2018 23:00:53 +0000 (00:00 +0100)]
net/sched: fix NULL dereference in the error path of tcf_vlan_init()
when the following command
# tc actions replace action vlan pop index 100
is run for the first time, and tcf_vlan_init() fails allocating struct
tcf_vlan_params, tcf_vlan_cleanup() calls kfree_rcu(NULL, ...). This causes
the following error:
fix this in tcf_vlan_cleanup(), ensuring that kfree_rcu(p, ...) is called
only when p is not NULL.
Fixes: 4c5b9d9642c8 ("act_vlan: VLAN action rewrite to use RCU lock/unlock and update") Acked-by: Jiri Pirko <jiri@mellanox.com> Acked-by: Manish Kurup <manish.kurup@verizon.com> Signed-off-by: Davide Caratti <dcaratti@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
SZ Lin (林上智) [Thu, 15 Mar 2018 16:56:01 +0000 (00:56 +0800)]
net: ethernet: ti: cpsw: add check for in-band mode setting with RGMII PHY interface
According to AM335x TRM[1] 14.3.6.2, AM437x TRM[2] 15.3.6.2 and
DRA7 TRM[3] 24.11.4.8.7.3.3, in-band mode in EXT_EN(bit18) register is only
available when PHY is configured in RGMII mode with 10Mbps speed. It will
cause some networking issues without RGMII mode, such as carrier sense
errors and low throughput. TI also mentioned this issue in their forum[4].
This patch adds the check mechanism for PHY interface with RGMII interface
type, the in-band mode can only be set in RGMII mode with 10Mbps speed.
Matthias Brugger [Thu, 15 Mar 2018 16:54:20 +0000 (17:54 +0100)]
net: hns: Fix ethtool private flags
The driver implementation returns support for private flags, while
no private flags are present. When asked for the number of private
flags it returns the number of statistic flag names.
Fix this by returning EOPNOTSUPP for not implemented ethtool flags.
Signed-off-by: Matthias Brugger <mbrugger@suse.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Ido Schimmel [Thu, 15 Mar 2018 12:49:56 +0000 (14:49 +0200)]
mlxsw: spectrum_buffers: Set a minimum quota for CPU port traffic
In commit 9ffcc3725f09 ("mlxsw: spectrum: Allow packets to be trapped
from any PG") I fixed a problem where packets could not be trapped to
the CPU due to exceeded shared buffer quotas. The mentioned commit
explains the problem in detail.
The problem was fixed by assigning a minimum quota for the CPU port and
the traffic class used for scheduling traffic to the CPU.
However, commit 117b0dad2d54 ("mlxsw: Create a different trap group list
for each device") assigned different traffic classes to different
packet types and rendered the fix useless.
Fix the problem by assigning a minimum quota for the CPU port and all
the traffic classes that are currently in use.
Fixes: 117b0dad2d54 ("mlxsw: Create a different trap group list for each device") Signed-off-by: Ido Schimmel <idosch@mellanox.com> Reported-by: Eddie Shklaer <eddies@mellanox.com> Tested-by: Eddie Shklaer <eddies@mellanox.com> Acked-by: Jiri Pirko <jiri@mellanox.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Fixes: c5ad119fb6c0 ("net: sched: pfifo_fast use skb_array") Signed-off-by: Eric Dumazet <edumazet@google.com> Reported-by: syzbot+ed43b6903ab968b16f54@syzkaller.appspotmail.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> Acked-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Andrew Lunn [Fri, 16 Mar 2018 17:03:46 +0000 (18:03 +0100)]
net: dsa: mv88e6xxx: Fix binding documentation for MDIO busses
The MDIO busses are switch properties and so should be inside the
switch node. Fix the examples in the binding document.
Reported-by: 尤晓杰 <yxj790222@163.com> Signed-off-by: Andrew Lunn <andrew@lunn.ch> Fixes: a3c53be55c95 ("net: dsa: mv88e6xxx: Support multiple MDIO busses") Signed-off-by: David S. Miller <davem@davemloft.net>
skbuff: Fix not waking applications when errors are enqueued
When errors are enqueued to the error queue via sock_queue_err_skb()
function, it is possible that the waiting application is not notified.
Calling 'sk->sk_data_ready()' would not notify applications that
selected only POLLERR events in poll() (for example).
Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Reported-by: Randy E. Witt <randy.e.witt@intel.com> Reviewed-by: Eric Dumazet <edumazet@google.com> Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Nicolas Dichtel [Wed, 14 Mar 2018 20:10:23 +0000 (21:10 +0100)]
netlink: avoid a double skb free in genlmsg_mcast()
nlmsg_multicast() consumes always the skb, thus the original skb must be
freed only when this function is called with a clone.
Fixes: cb9f7a9a5c96 ("netlink: ensure to loop over all netns in genlmsg_multicast_allns()") Reported-by: Ben Hutchings <ben.hutchings@codethink.co.uk> Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Michal Kalderon [Wed, 14 Mar 2018 12:56:53 +0000 (14:56 +0200)]
qede: Fix qedr link update
Link updates were not reported to qedr correctly.
Leading to cases where a link could be down, but qedr
would see it as up.
In addition, once qede was loaded, link state would be up,
regardless of the actual link state.
Signed-off-by: Michal Kalderon <michal.kalderon@cavium.com> Signed-off-by: Ariel Elior <ariel.elior@cavium.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Michal Kalderon [Wed, 14 Mar 2018 12:49:28 +0000 (14:49 +0200)]
qed: Fix non TCP packets should be dropped on iWARP ll2 connection
FW workaround. The iWARP LL2 connection did not expect TCP packets
to arrive on it's connection. The fix drops any non-tcp packets
Fixes b5c29ca ("qed: iWARP CM - setup a ll2 connection for handling
SYN packets")
Signed-off-by: Michal Kalderon <Michal.Kalderon@cavium.com> Signed-off-by: Ariel Elior <Ariel.Elior@cavium.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Michal Kalderon [Wed, 14 Mar 2018 12:49:27 +0000 (14:49 +0200)]
qed: Fix MPA unalign flow in case header is split across two packets.
There is a corner case in the MPA unalign flow where a FPDU header is
split over two tcp segments. The length of the first fragment in this
case was not initialized properly and should be '1'
Fixes: c7d1d839 ("qed: Add support for MPA header being split over two tcp packets") Signed-off-by: Michal Kalderon <Michal.Kalderon@cavium.com> Signed-off-by: Ariel Elior <Ariel.Elior@cavium.com> Signed-off-by: David S. Miller <davem@davemloft.net>
There is no need for complex checking between the last consumed index
and current consumed index, a simple subtraction will do.
This also eliminates the possibility of a permanent transmit queue stall
under the following conditions:
- one CPU bursts ring->size worth of traffic (up to 256 buffers), to the
point where we run out of free descriptors, so we stop the transmit
queue at the end of bcm_sysport_xmit()
- because of our locking, we have the transmit process disable
interrupts which means we can be blocking the TX reclamation process
- when TX reclamation finally runs, we will be computing the difference
between ring->c_index (last consumed index by SW) and what the HW
reports through its register
- this register is masked with (ring->size - 1) = 0xff, which will lead
to stripping the upper bits of the index (register is 16-bits wide)
- we will be computing last_tx_cn as 0, which means there is no work to
be done, and we never wake-up the transmit queue, leaving it
permanently disabled
A practical example is e.g: ring->c_index aka last_c_index = 12, we
pushed 256 entries, HW consumer index = 268, we mask it with 0xff = 12,
so last_tx_cn == 0, nothing happens.
Fixes: 80105befdb4b ("net: systemport: add Broadcom SYSTEMPORT Ethernet MAC driver") Signed-off-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Tom Herbert [Tue, 13 Mar 2018 19:01:43 +0000 (12:01 -0700)]
kcm: lock lower socket in kcm_attach
Need to lock lower socket in order to provide mutual exclusion
with kcm_unattach.
v2: Add Reported-by for syzbot
Fixes: ab7ac4eb9832e32a09f4e804 ("kcm: Kernel Connection Multiplexor module") Reported-by: syzbot+ea75c0ffcd353d32515f064aaebefc5279e6161e@syzkaller.appspotmail.com Signed-off-by: Tom Herbert <tom@quantonium.net> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Fri, 16 Mar 2018 14:03:48 +0000 (10:03 -0400)]
Merge branch 'vlan-untag-and-insert-fixes'
Toshiaki Makita says:
====================
Fix vlan untag and insertion for bridge and vlan with reorder_hdr off
As Brandon Carpenter reported[1], sending non-vlan-offloaded packets from
bridge devices ends up with corrupted packets. He narrowed down this problem
and found that the root cause is in skb_reorder_vlan_header().
While I was working on fixing this problem, I found that the function does
not work properly for double tagged packets with reorder_hdr off as well.
Patch 1 fixes these 2 problems in skb_reorder_vlan_header().
And it turned out that fixing skb_reorder_vlan_header() is not sufficient
to receive double tagged packets with reorder_hdr off while I was testing the
fix. Vlan tags got out of order when vlan devices with reorder_hdr disabled
were stacked. Patch 2 fixes this problem.
Toshiaki Makita [Tue, 13 Mar 2018 05:51:28 +0000 (14:51 +0900)]
vlan: Fix out of order vlan headers with reorder header off
With reorder header off, received packets are untagged in skb_vlan_untag()
called from within __netif_receive_skb_core(), and later the tag will be
inserted back in vlan_do_receive().
This caused out of order vlan headers when we create a vlan device on top
of another vlan device, because vlan_do_receive() inserts a tag as the
outermost vlan tag. E.g. the outer tag is first removed in skb_vlan_untag()
and inserted back in vlan_do_receive(), then the inner tag is next removed
and inserted back as the outermost tag.
This patch fixes the behaviour by inserting the inner tag at the right
position.
Signed-off-by: Toshiaki Makita <makita.toshiaki@lab.ntt.co.jp> Signed-off-by: David S. Miller <davem@davemloft.net>
Toshiaki Makita [Tue, 13 Mar 2018 05:51:27 +0000 (14:51 +0900)]
net: Fix vlan untag for bridge and vlan_dev with reorder_hdr off
When we have a bridge with vlan_filtering on and a vlan device on top of
it, packets would be corrupted in skb_vlan_untag() called from
br_dev_xmit().
The problem sits in skb_reorder_vlan_header() used in skb_vlan_untag(),
which makes use of skb->mac_len. In this function mac_len is meant for
handling rx path with vlan devices with reorder_header disabled, but in
tx path mac_len is typically 0 and cannot be used, which is the problem
in this case.
The current code even does not properly handle rx path (skb_vlan_untag()
called from __netif_receive_skb_core()) with reorder_header off actually.
In rx path single tag case, it works as follows:
- Before skb_reorder_vlan_header()
mac_header data
v v
+-------------------+-------------+------+----
| ETH | VLAN | ETH |
| ADDRS | TPID | TCI | TYPE |
+-------------------+-------------+------+----
<-------- mac_len --------->
<------------->
to be removed
- After skb_reorder_vlan_header()
mac_header data
v v
+-------------------+------+----
| ETH | ETH |
| ADDRS | TYPE |
+-------------------+------+----
<-------- mac_len --------->
This is ok, but in rx double tag case, it corrupts packets:
- Before skb_reorder_vlan_header()
mac_header data
v v
+-------------------+-------------+-------------+------+----
| ETH | VLAN | VLAN | ETH |
| ADDRS | TPID | TCI | TPID | TCI | TYPE |
+-------------------+-------------+-------------+------+----
<--------------- mac_len ---------------->
<------------->
should be removed
<--------------------------->
actually will be removed
- After skb_reorder_vlan_header()
mac_header data
v v
+-------------------+------+----
| ETH | ETH |
| ADDRS | TYPE |
+-------------------+------+----
<--------------- mac_len ---------------->
So, two of vlan tags are both removed while only inner one should be
removed and mac_header (and mac_len) is broken.
skb_vlan_untag() is meant for removing the vlan header at (skb->data - 2),
so use skb->data and skb->mac_header to calculate the right offset.
Reported-by: Brandon Carpenter <brandon.carpenter@cypherpath.com> Fixes: a6e18ff11170 ("vlan: Fix untag operations of stacked vlans with REORDER_HEADER off") Signed-off-by: Toshiaki Makita <makita.toshiaki@lab.ntt.co.jp> Signed-off-by: David S. Miller <davem@davemloft.net>
Roman Mashak [Mon, 12 Mar 2018 20:20:58 +0000 (16:20 -0400)]
net sched actions: return explicit error when tunnel_key mode is not specified
If set/unset mode of the tunnel_key action is not provided, ->init() still
returns 0, and the caller proceeds with bogus 'struct tc_action *' object,
this results in crash:
% tc actions add action tunnel_key src_ip 1.1.1.1 dst_ip 2.2.2.1 id 7 index 1
Fixes: d0f6dd8a914f ("net/sched: Introduce act_tunnel_key") Signed-off-by: Roman Mashak <mrv@mojatatu.com> Acked-by: Cong Wang <xiyou.wangcong@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Ursula Braun [Tue, 13 Mar 2018 09:41:54 +0000 (10:41 +0100)]
net/smc: simplify wait when closing listen socket
Closing of a listen socket wakes up kernel_accept() of
smc_tcp_listen_worker(), and then has to wait till smc_tcp_listen_worker()
gives up the internal clcsock. The wait logic introduced with
commit 127f49705823 ("net/smc: release clcsock from tcp_listen_worker")
might wait longer than necessary. This patch implements the idea to
implement the wait just with flush_work(), and gets rid of the extra
smc_close_wait_listen_clcsock() function.
Fixes: 127f49705823 ("net/smc: release clcsock from tcp_listen_worker") Reported-by: Hans Wippel <hwippel@linux.vnet.ibm.com> Signed-off-by: Ursula Braun <ubraun@linux.vnet.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Cathy Zhou [Wed, 14 Mar 2018 17:56:07 +0000 (10:56 -0700)]
sunvnet: does not support GSO for sctp
The NETIF_F_GSO_SOFTWARE implies support for GSO on SCTP, but the
sunvnet driver does not support GSO for sctp. Here we remove the
NETIF_F_GSO_SOFTWARE feature flag and only report NETIF_F_ALL_TSO
instead.
Signed-off-by: Cathy Zhou <Cathy.Zhou@Oracle.COM> Signed-off-by: Shannon Nelson <shannon.nelson@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Wed, 14 Mar 2018 17:50:47 +0000 (13:50 -0400)]
Merge tag 'linux-can-fixes-for-4.16-20180314' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/mkl/linux-can
Marc Kleine-Budde says:
====================
pull-request: can 2018-03-14
this is a pull request of two patches for net/master.
Both patches are by Andri Yngvason and fix problems in the cc770 driver,
that show up quite fast on RT systems, but also on non RT setups.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Jonathan Toppins [Wed, 14 Mar 2018 16:36:25 +0000 (12:36 -0400)]
tg3: prevent scheduling while atomic splat
The problem was introduced in commit 506b0a395f26 ("[netdrv] tg3: APE heartbeat changes"). The bug occurs
because tp->lock spinlock is held which is obtained in tg3_start
by way of tg3_full_lock(), line 11571. The documentation for usleep_range()
specifically states it cannot be used inside a spinlock.
Fixes: 506b0a395f26 ("[netdrv] tg3: APE heartbeat changes") Signed-off-by: Jonathan Toppins <jtoppins@redhat.com> Acked-by: Michael Chan <michael.chan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Sabrina Dubroca [Wed, 14 Mar 2018 09:21:14 +0000 (10:21 +0100)]
ipv4: lock mtu in fnhe when received PMTU < net.ipv4.route.min_pmtu
Prior to the rework of PMTU information storage in commit 2c8cec5c10bc ("ipv4: Cache learned PMTU information in inetpeer."),
when a PMTU event advertising a PMTU smaller than
net.ipv4.route.min_pmtu was received, we would disable setting the DF
flag on packets by locking the MTU metric, and set the PMTU to
net.ipv4.route.min_pmtu.
Since then, we don't disable DF, and set PMTU to
net.ipv4.route.min_pmtu, so the intermediate router that has this link
with a small MTU will have to drop the packets.
This patch reestablishes pre-2.6.39 behavior by splitting
rtable->rt_pmtu into a bitfield with rt_mtu_locked and rt_pmtu.
rt_mtu_locked indicates that we shouldn't set the DF bit on that path,
and is checked in ip_dont_fragment().
One possible workaround is to set net.ipv4.route.min_pmtu to a value low
enough to accommodate the lowest MTU encountered.
Fixes: 2c8cec5c10bc ("ipv4: Cache learned PMTU information in inetpeer.") Signed-off-by: Sabrina Dubroca <sd@queasysnail.net> Reviewed-by: Stefano Brivio <sbrivio@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Wed, 14 Mar 2018 17:16:23 +0000 (13:16 -0400)]
Merge branch 'DPAA-Ethernet-fixes'
Madalin Bucur says:
====================
DPAA Ethernet fixes
This patch set is addressing several issues in the DPAA Ethernet
driver suite:
- module unload crash caused by wrong reference to device being left
in the cleanup code after the DSA related changes
- scheduling wile atomic bug in QMan code revealed during dpaa_eth
module unload
- a couple of error counter fixes, a duplicated init in dpaa_eth.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Camelia Groza [Wed, 14 Mar 2018 13:37:32 +0000 (08:37 -0500)]
dpaa_eth: remove duplicate increment of the tx_errors counter
The tx_errors counter is incremented by the dpaa_xmit caller.
Signed-off-by: Camelia Groza <camelia.groza@nxp.com> Signed-off-by: Madalin Bucur <madalin.bucur@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Madalin Bucur [Wed, 14 Mar 2018 13:37:29 +0000 (08:37 -0500)]
dpaa_eth: fix error in dpaa_remove()
The recent changes that make the driver probing compatible with DSA
were not propagated in the dpa_remove() function, breaking the
module unload function. Using the proper device to address the issue.
Signed-off-by: Madalin Bucur <madalin.bucur@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Madalin Bucur [Wed, 14 Mar 2018 13:37:28 +0000 (08:37 -0500)]
soc/fsl/qbman: fix issue in qman_delete_cgr_safe()
The wait_for_completion() call in qman_delete_cgr_safe()
was triggering a scheduling while atomic bug, replacing the
kthread with a smp_call_function_single() call to fix it.
Signed-off-by: Madalin Bucur <madalin.bucur@nxp.com> Signed-off-by: Roy Pledge <roy.pledge@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Wed, 14 Mar 2018 16:04:16 +0000 (09:04 -0700)]
net: use skb_to_full_sk() in skb_update_prio()
Andrei Vagin reported a KASAN: slab-out-of-bounds error in
skb_update_prio()
Since SYNACK might be attached to a request socket, we need to
get back to the listener socket.
Since this listener is manipulated without locks, add const
qualifiers to sock_cgroup_prioidx() so that the const can also
be used in skb_update_prio()
Also add the const qualifier to sock_cgroup_classid() for consistency.
Fixes: ca6fb0651883 ("tcp: attach SYNACK messages to request sockets instead of listener") Signed-off-by: Eric Dumazet <edumazet@google.com> Reported-by: Andrei Vagin <avagin@virtuozzo.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Andri Yngvason [Wed, 14 Mar 2018 11:52:57 +0000 (11:52 +0000)]
can: cc770: Fix queue stall & dropped RTR reply
While waiting for the TX object to send an RTR, an external message with a
matching id can overwrite the TX data. In this case we must call the rx
routine and then try transmitting the message that was overwritten again.
The queue was being stalled because the RX event did not generate an
interrupt to wake up the queue again and the TX event did not happen
because the TXRQST flag is reset by the chip when new data is received.
According to the CC770 datasheet the id of a message object should not be
changed while the MSGVAL bit is set. This has been fixed by resetting the
MSGVAL bit before modifying the object in the transmit function and setting
it after. It is not enough to set & reset CPUUPD.
It is important to keep the MSGVAL bit reset while the message object is
being modified. Otherwise, during RTR transmission, a frame with matching
id could trigger an rx-interrupt, which would cause a race condition
between the interrupt routine and the transmit function.
Signed-off-by: Andri Yngvason <andri.yngvason@marel.com> Tested-by: Richard Weinberger <richard@nod.at> Cc: linux-stable <stable@vger.kernel.org> Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Dan Carpenter [Tue, 13 Mar 2018 09:09:38 +0000 (12:09 +0300)]
qed: Use after free in qed_rdma_free()
We're dereferencing "p_hwfn->p_rdma_info" but that is freed on the line
before in qed_rdma_resc_free(p_hwfn).
Fixes: 9de506a547c0 ("qed: Free RoCE ILT Memory on rmmod qedr") Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Acked-by: Michal Kalderon <Michal.Kalderon@cavium.com> Signed-off-by: David S. Miller <davem@davemloft.net>
1) Refuse to insert 32 bit userspace socket policies on 64
bit systems like we do it for standard policies. We don't
have a compat layer, so inserting socket policies from
32 bit userspace will lead to a broken configuration.
2) Make the policy hold queue work without the flowcache.
Dummy bundles are not chached anymore, so we need to
generate a new one on each lookup as long as the SAs
are not yet in place.
3) Fix the validation of the esn replay attribute. The
The sanity check in verify_replay() is bypassed if
the XFRM_STATE_ESN flag is not set. Fix this by doing
the sanity check uncoditionally.
From Florian Westphal.
4) After most of the dst_entry garbage collection code
is removed, we may leak xfrm_dst entries as they are
neither cached nor tracked somewhere. Fix this by
reusing the 'uncached_list' to track xfrm_dst entries
too. From Xin Long.
5) Fix a rcu_read_lock/rcu_read_unlock imbalance in
xfrm_get_tos() From Xin Long.
6) Fix an infinite loop in xfrm_get_dst_nexthop. On
transport mode we fetch the child dst_entry after
we continue, so this pointer is never updated.
Fix this by fetching it before we continue.
7) Fix ESN sequence number gap after IPsec GSO packets.
We accidentally increment the sequence number counter
on the xfrm_state by one packet too much in the ESN
case. Fix this by setting the sequence number to the
correct value.
8) Reset the ethernet protocol after decapsulation only if a
mac header was set. Otherwise it breaks configurations
with TUN devices. From Yossi Kuperman.
9) Fix __this_cpu_read() usage in preemptible code. Use
this_cpu_read() instead in ipcomp_alloc_tfms().
From Greg Hackmann.
Please pull or let me know if there are problems.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Greg Hackmann [Wed, 7 Mar 2018 22:42:53 +0000 (14:42 -0800)]
net: xfrm: use preempt-safe this_cpu_read() in ipcomp_alloc_tfms()
f7c83bcbfaf5 ("net: xfrm: use __this_cpu_read per-cpu helper") added a
__this_cpu_read() call inside ipcomp_alloc_tfms().
At the time, __this_cpu_read() required the caller to either not care
about races or to handle preemption/interrupt issues. 3.15 tightened
the rules around some per-cpu operations, and now __this_cpu_read()
should never be used in a preemptible context. On 3.15 and later, we
need to use this_cpu_read() instead.
syzkaller reported this leading to the following kernel BUG while
fuzzing sendmsg:
Florian Fainelli [Mon, 12 Mar 2018 23:00:40 +0000 (16:00 -0700)]
net: dsa: Fix dsa_is_user_port() test inversion
During the conversion to dsa_is_user_port(), a condition ended up being
reversed, which would prevent the creation of any user port when using
the legacy binding and/or platform data, fix that.
Fixes: 4a5b85ffe2a0 ("net: dsa: use dsa_is_user_port everywhere") Signed-off-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Mon, 12 Mar 2018 19:21:08 +0000 (15:21 -0400)]
Merge branch '1GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-queue
Jeff Kirsher says:
====================
Intel Wired LAN Driver Updates 2018-03-12
This series contains fixes to e1000e only.
Benjamin Poirier provides two fixes, first reverts commits that changed
what happens to the link status when there is an error. These commits
were to resolve a race condition, but in the process of fixing the race
condition, they changed the behavior when an error occurred. Second fix
resolves a race condition by not setting "get_link_status" to false
after checking the link.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Paolo Abeni [Mon, 12 Mar 2018 13:54:24 +0000 (14:54 +0100)]
l2tp: fix races with ipv4-mapped ipv6 addresses
The l2tp_tunnel_create() function checks for v4mapped ipv6
sockets and cache that flag, so that l2tp core code can
reusing it at xmit time.
If the socket is provided by the userspace, the connection
status of the tunnel sockets can change between the tunnel
creation and the xmit call, so that syzbot is able to
trigger the following splat:
BUG: KASAN: use-after-free in ip6_dst_idev include/net/ip6_fib.h:192
[inline]
BUG: KASAN: use-after-free in ip6_xmit+0x1f76/0x2260
net/ipv6/ip6_output.c:264
Read of size 8 at addr ffff8801bd949318 by task syz-executor4/23448
This change addresses the issues:
* explicitly checking for TCP_ESTABLISHED for user space provided sockets
* dropping the v4mapped flag usage - it can become outdated - and
explicitly invoking ipv6_addr_v4mapped() instead
The issue is apparently there since ancient times.
v1 -> v2: (many thanks to Guillaume)
- with csum issue introduced in v1
- replace pr_err with pr_debug
- fix build issue with IPV6 disabled
- move l2tp_sk_is_v4mapped in l2tp_core.c
v2 -> v3:
- don't update inet_daddr for v4mapped address, unneeded
- drop rendundant check at creation time
Reported-and-tested-by: syzbot+92fa328176eb07e4ac1a@syzkaller.appspotmail.com Fixes: 3557baabf280 ("[L2TP]: PPP over L2TP driver core") Signed-off-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Paolo Abeni [Mon, 12 Mar 2018 13:54:23 +0000 (14:54 +0100)]
net: ipv6: keep sk status consistent after datagram connect failure
On unsuccesful ip6_datagram_connect(), if the failure is caused by
ip6_datagram_dst_update(), the sk peer information are cleared, but
the sk->sk_state is preserved.
If the socket was already in an established status, the overall sk
status is inconsistent and fouls later checks in datagram code.
Fix this saving the old peer information and restoring them in
case of failure. This also aligns ipv6 datagram connect() behavior
with ipv4.
v1 -> v2:
- added missing Fixes tag
Fixes: 85cb73ff9b74 ("net: ipv6: reset daddr and dport in sk if connect() fails") Signed-off-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
link_active = true
/* link_active is true, wrongly, and stays so because
* get_link_status is false */
Avoid this problem by making sure that we don't set get_link_status = false
after having checked the link.
It seems this problem has been present since the introduction of e1000e.
Link: https://lkml.org/lkml/2018/1/29/338 Reported-by: Alexander Duyck <alexander.duyck@gmail.com> Signed-off-by: Benjamin Poirier <bpoirier@suse.com> Acked-by: Alexander Duyck <alexander.h.duyck@intel.com> Tested-by: Aaron Brown <aaron.f.brown@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Commit 19110cfbb34d ("e1000e: Separate signaling for link check/link up")
changed what happens to the link status when there is an error which
happens after "get_link_status = false" in the copper check_for_link
callbacks. Previously, such an error would be ignored and the link
considered up. After that commit, any error implies that the link is down.
Revert commit 19110cfbb34d ("e1000e: Separate signaling for link check/link
up") and its followups. After reverting, the race condition described in
the log of commit 19110cfbb34d is reintroduced. It may still be triggered
by LSC events but this should keep the link down in case the link is
electrically unstable, as discussed. The race may no longer be
triggered by RXO events because commit 4aea7a5c5e94 ("e1000e: Avoid
receiver overrun interrupt bursts") restored reading icr in the Other
handler.
Link: https://lkml.org/lkml/2018/3/1/789 Signed-off-by: Benjamin Poirier <bpoirier@suse.com> Acked-by: Alexander Duyck <alexander.h.duyck@intel.com> Tested-by: Aaron Brown <aaron.f.brown@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The following patchset contains Netfilter fixes for your net tree, they are:
1) Fixed hashtable representation doesn't support timeout flag, skip it
otherwise rules to add elements from the packet fail bogusly fail with
EOPNOTSUPP.
2) Fix bogus error with 32-bits ebtables userspace and 64-bits kernel,
patch from Florian Westphal.
3) Sanitize proc names in several x_tables extensions, also from Florian.
4) Add sanitization to ebt_among wormhash logic, from Florian.
5) Missing release of hook array in flowtable.
====================
David S. Miller [Mon, 12 Mar 2018 15:16:40 +0000 (11:16 -0400)]
Merge tag 'linux-can-fixes-for-4.16-20180312' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/mkl/linux-can
Marc Kleine-Budde says:
====================
pull-request: can 2018-03-12
this is a pull reqeust of 6 patches for net/master.
The first patch is by Wolfram Sang and fixes a bitshift vs. comparison mistake
in the m_can driver. Two patches of Marek Vasut repair the error handling in
the ifi driver. The two patches by Stephane Grosjean fix a "echo_skb is
occupied!" bug in the peak/pcie_fd driver. Bich HEMON's patch adds pinctrl
select state calls to the m_can's driver to further improve power saving during
suspend.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Xin Long [Sat, 10 Mar 2018 10:57:50 +0000 (18:57 +0800)]
sock_diag: request _diag module only when the family or proto has been registered
Now when using 'ss' in iproute, kernel would try to load all _diag
modules, which also causes corresponding family and proto modules
to be loaded as well due to module dependencies.
Like after running 'ss', sctp, dccp, af_packet (if it works as a module)
would be loaded.
As these family and proto modules are loaded unintentionally, it
could cause some problems, like:
- Some debug tools use 'ss' to collect the socket info, which loads all
those diag and family and protocol modules. It's noisy for identifying
issues.
- Users usually expect to drop sctp init packet silently when they
have no sense of sctp protocol instead of sending abort back.
- It wastes resources (especially with multiple netns), and SCTP module
can't be unloaded once it's loaded.
...
In short, it's really inappropriate to have these family and proto
modules loaded unexpectedly when just doing debugging with inet_diag.
This patch is to introduce sock_load_diag_module() where it loads
the _diag module only when it's corresponding family or proto has
been already registered.
Note that we can't just load _diag module without the family or
proto loaded, as some symbols used in _diag module are from the
family or proto module.
v1->v2:
- move inet proto check to inet_diag to avoid a compiling err.
v2->v3:
- define sock_load_diag_module in sock.c and export one symbol
only.
- improve the changelog.
Reported-by: Sabrina Dubroca <sd@queasysnail.net> Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com> Acked-by: Phil Sutter <phil@nwl.cc> Acked-by: Sabrina Dubroca <sd@queasysnail.net> Signed-off-by: Xin Long <lucien.xin@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Mon, 12 Mar 2018 14:58:28 +0000 (10:58 -0400)]
Merge branch 'bnxt_en-Bug-fixes'
Michael Chan says:
====================
bnxt_en: Bug fixes.
There are 3 bug fixes in this series to fix regressions recently
introduced when adding the new ring reservations scheme. 2 minor
fixes in the TC Flower code to return standard errno values and
to elide some unnecessary warning dmesg. One Fixes the VLAN TCI
value passed to the stack by including the entire 16-bit VLAN TCI,
and the last fix is to check for valid VNIC ID before setting up or
shutting down LRO/GRO.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Michael Chan [Sat, 10 Mar 2018 04:46:10 +0000 (23:46 -0500)]
bnxt_en: Check valid VNIC ID in bnxt_hwrm_vnic_set_tpa().
During initialization, if we encounter errors, there is a code path that
calls bnxt_hwrm_vnic_set_tpa() with invalid VNIC ID. This may cause a
warning in firmware logs.
Fixes: c0c050c58d84 ("bnxt_en: New Broadcom ethernet driver.") Signed-off-by: Michael Chan <michael.chan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Venkat Duvvuru [Sat, 10 Mar 2018 04:46:09 +0000 (23:46 -0500)]
bnxt_en: close & open NIC, only when the interface is in running state.
bnxt_restore_pf_fw_resources routine frees PF resources by calling
close_nic and allocates the resources back, by doing open_nic. However,
this is not needed, if the PF is already in closed state.
This bug causes the driver to call open the device and call request_irq()
when it is not needed. Ultimately, pci_disable_msix() will crash
when bnxt_en is unloaded.
This patch fixes the problem by skipping __bnxt_close_nic and
__bnxt_open_nic inside bnxt_restore_pf_fw_resources routine, if the
interface is not running.
Fixes: 80fcaf46c092 ("bnxt_en: Restore MSIX after disabling SRIOV.") Signed-off-by: Venkat Duvvuru <venkatkumar.duvvuru@broadcom.com> Signed-off-by: Michael Chan <michael.chan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Venkat Duvvuru [Sat, 10 Mar 2018 04:46:08 +0000 (23:46 -0500)]
bnxt_en: Return standard Linux error codes for hwrm flow cmds.
Currently, internal error value is returned by the driver, when
hwrm_cfa_flow_alloc() fails due lack of resources. We should be returning
Linux errno value -ENOSPC instead.
This patch also converts other similar command errors to standard Linux errno
code (-EIO) in bnxt_tc.c
Fixes: db1d36a27324 ("bnxt_en: add TC flower offload flow_alloc/free FW cmds") Signed-off-by: Venkat Duvvuru <venkatkumar.duvvuru@broadcom.com> Signed-off-by: Michael Chan <michael.chan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Michael Chan [Sat, 10 Mar 2018 04:46:07 +0000 (23:46 -0500)]
bnxt_en: Fix regressions when setting up MQPRIO TX rings.
Recent changes added the bnxt_init_int_mode() call in the driver's open
path whenever ring reservations are changed. This call was previously
only called in the probe path. In the open path, if MQPRIO TC has been
setup, the bnxt_init_int_mode() call would reset and mess up the MQPRIO
per TC rings.
Fix it by not re-initilizing bp->tx_nr_rings_per_tc in
bnxt_init_int_mode(). Instead, initialize it in the probe path only
after the bnxt_init_int_mode() call.
Fixes: 674f50a5b026 ("bnxt_en: Implement new method to reserve rings.") Signed-off-by: Michael Chan <michael.chan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Michael Chan [Sat, 10 Mar 2018 04:46:06 +0000 (23:46 -0500)]
bnxt_en: Pass complete VLAN TCI to the stack.
When receiving a packet with VLAN tag, pass the entire 16-bit TCI to the
stack when calling __vlan_hwaccel_put_tag(). The current code is only
passing the 12-bit tag and it is missing the priority bits.
Signed-off-by: Michael Chan <michael.chan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
bnxt_en: Remove unwanted ovs-offload messages in some conditions
In some conditions when the driver fails to add a flow in HW and returns
an error back to the stack, the stack continues to invoke get_flow_stats()
and/or del_flow() on it. The driver fails these APIs with an error message
"no flow_node for cookie". The message gets logged repeatedly as long as
the stack keeps invoking these functions.
Fix this by removing the corresponding netdev_info() calls from these
functions.
Fixes: d7bc73053024 ("bnxt_en: add code to query TC flower offload stats") Signed-off-by: Sriharsha Basavapatna <sriharsha.basavapatna@broadcom.com> Signed-off-by: Michael Chan <michael.chan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Eddie Wai [Sat, 10 Mar 2018 04:46:04 +0000 (23:46 -0500)]
bnxt_en: Fix vnic accounting in the bnxt_check_rings() path.
The number of vnics to check must be determined ahead of time because
only standard RX rings require vnics to support RFS. The logic is
similar to the ring reservation logic and we can now use the
refactored common functions to do most of the work in setting up
the firmware message.
Fixes: 8f23d638b36b ("bnxt_en: Expand bnxt_check_rings() to check all resources.") Signed-off-by: Eddie Wai <eddie.wai@broadcom.com> Signed-off-by: Michael Chan <michael.chan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Michael Chan [Sat, 10 Mar 2018 04:46:03 +0000 (23:46 -0500)]
bnxt_en: Refactor the functions to reserve hardware rings.
The bnxt_hwrm_reserve_{pf|vf}_rings() functions are very similar to
the bnxt_hwrm_check_{pf|vf}_rings() functions. Refactor the former
so that the latter can make use of common code in the next patch.
Signed-off-by: Michael Chan <michael.chan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Brad Mouring [Thu, 8 Mar 2018 22:23:03 +0000 (16:23 -0600)]
net: phy: Tell caller result of phy_change()
In 664fcf123a30e (net: phy: Threaded interrupts allow some simplification)
the phy_interrupt system was changed to use a traditional threaded
interrupt scheme instead of a workqueue approach.
With this change, the phy status check moved into phy_change, which
did not report back to the caller whether or not the interrupt was
handled. This means that, in the case of a shared phy interrupt,
only the first phydev's interrupt registers are checked (since
phy_interrupt() would always return IRQ_HANDLED). This leads to
interrupt storms when it is a secondary device that's actually the
interrupt source.
Signed-off-by: Brad Mouring <brad.mouring@ni.com> Signed-off-by: David S. Miller <davem@davemloft.net>
can: peak/pcie_fd: remove useless code when interface starts
When an interface starts, the echo_skb array is empty and the network
queue should be started only. This patch replaces useless code and locks
when the internal RX_BARRIER message is received from the IP core, telling
the driver that tx may start.
This patch makes atomic the handling of the linux-can echo_skb array and
the network tx queue. This prevents from the "BUG! echo_skb is occupied!"
message to be printed by the linux-can core, in SMP environments.
Marek Vasut [Thu, 1 Mar 2018 18:34:00 +0000 (19:34 +0100)]
can: ifi: Repair the error handling
The new version of the IFI CANFD core has significantly less complex
error state indication logic. In particular, the warning/error state
bits are no longer all over the place, but are all present in the
STATUS register. Moreover, there is a new IRQ register bit indicating
transition between error states (active/warning/passive/busoff).
This patch makes use of this bit to weed out the obscure selective
INTERRUPT register clearing, which was used to carry over the error
state indication into the poll function. While at it, this patch
fixes the handling of the ACTIVE state, since the hardware provides
indication of the core being in ACTIVE state and that in turn fixes
the state transition indication toward userspace. Finally, register
reads in the poll function are moved to the matching subfunctions
since those are also no longer needed in the poll function.
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Heiko Schocher <hs@denx.de> Cc: Markus Marb <markus@marb.org> Cc: Marc Kleine-Budde <mkl@pengutronix.de> Cc: linux-stable <stable@vger.kernel.org> Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Marek Vasut [Mon, 5 Mar 2018 20:29:52 +0000 (21:29 +0100)]
can: ifi: Check core revision upon probe
Older versions of the core are not compatible with the driver due
to various intrusive fixes of the core. Read out the VER register,
check the core revision bitfield and verify if the core in use is
new enough (rev 2.1 or newer) to work correctly with this driver.
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Heiko Schocher <hs@denx.de> Cc: Markus Marb <markus@marb.org> Cc: Marc Kleine-Budde <mkl@pengutronix.de> Cc: linux-stable <stable@vger.kernel.org> Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
zhangliping [Fri, 9 Mar 2018 02:08:50 +0000 (10:08 +0800)]
openvswitch: meter: fix the incorrect calculation of max delta_t
Max delat_t should be the full_bucket/rate instead of the full_bucket.
Also report EINVAL if the rate is zero.
Fixes: 96fbc13d7e77 ("openvswitch: Add meter infrastructure") Cc: Andy Zhou <azhou@ovn.org> Signed-off-by: zhangliping <zhangliping02@baidu.com> Acked-by: Pravin B Shelar <pshelar@ovn.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Shannon Nelson [Fri, 9 Mar 2018 00:17:23 +0000 (16:17 -0800)]
macvlan: filter out unsupported feature flags
Adding a macvlan device on top of a lowerdev that supports
the xfrm offloads fails with a new regression:
# ip link add link ens1f0 mv0 type macvlan
RTNETLINK answers: Operation not permitted
Tracing down the failure shows that the macvlan device inherits
the NETIF_F_HW_ESP and NETIF_F_HW_ESP_TX_CSUM feature flags
from the lowerdev, but with no dev->xfrmdev_ops API filled
in, it doesn't actually support xfrm. When the request is
made to add the new macvlan device, the XFRM listener for
NETDEV_REGISTER calls xfrm_api_check() which fails the new
registration because dev->xfrmdev_ops is NULL.
The macvlan creation succeeds when we filter out the ESP
feature flags in macvlan_fix_features(), so let's filter them
out like we're already filtering out ~NETIF_F_NETNS_LOCAL.
When XFRM support is added in the future, we can add the flags
into MACVLAN_FEATURES.
This same problem could crop up in the future with any other
new feature flags, so let's filter out any flags that aren't
defined as supported in macvlan.
Fixes: d77e38e612a0 ("xfrm: Add an IPsec hardware offloading API") Reported-by: Alexey Kodanev <alexey.kodanev@oracle.com> Signed-off-by: Shannon Nelson <shannon.nelson@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net>
netfilter: bridge: ebt_among: add more missing match size checks
ebt_among is special, it has a dynamic match size and is exempt
from the central size checks.
commit c4585a2823edf ("bridge: ebt_among: add missing match size checks")
added validation for pool size, but missed fact that the macros
ebt_among_wh_src/dst can already return out-of-bound result because
they do not check value of wh_src/dst_ofs (an offset) vs. the size
of the match that userspace gave to us.
v2:
check that offset has correct alignment.
Paolo Abeni points out that we should also check that src/dst
wormhash arrays do not overlap, and src + length lines up with
start of dst (or vice versa).
v3: compact wormhash_sizes_valid() part
NB: Fixes tag is intentionally wrong, this bug exists from day
one when match was added for 2.6 kernel. Tag is there so stable
maintainers will notice this one too.
Tested with same rules from the earlier patch.
Fixes: c4585a2823edf ("bridge: ebt_among: add missing match size checks") Reported-by: <syzbot+bdabab6f1983a03fc009@syzkaller.appspotmail.com> Signed-off-by: Florian Westphal <fw@strlen.de> Reviewed-by: Eric Dumazet <edumazet@google.com> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
netfilter: ebtables: fix erroneous reject of last rule
The last rule in the blob has next_entry offset that is same as total size.
This made "ebtables32 -A OUTPUT -d de:ad:be:ef:01:02" fail on 64 bit kernel.
David S. Miller [Fri, 9 Mar 2018 18:03:57 +0000 (13:03 -0500)]
Merge branch 'erspan-fixes'
William Tu says:
====================
a couple of erspan fixes
The series fixes a couple of erspan issues.
The first patch adds the erspan v2 proto type to the ip6 tunnel lookup.
The second patch improves the error handling when users screws the
version number in metadata. The final patch makes sure the skb has
enough headroom for pushing erspan header when xmit.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
William Tu [Fri, 9 Mar 2018 15:34:41 +0000 (07:34 -0800)]
ip6erspan: improve error handling for erspan version number.
When users fill in incorrect erspan version number through
the struct erspan_metadata uapi, current code skips pushing
the erspan header but continue pushing the gre header, which
is incorrect. The patch fixes it by returning error.
Signed-off-by: William Tu <u9012063@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Fri, 9 Mar 2018 18:02:14 +0000 (13:02 -0500)]
Merge branch 'mlxsw-ACL-and-mirroring-fixes'
Ido Schimmel says:
====================
mlxsw: ACL and mirroring fixes
The first patch fixes offload of rules using the 'pass' action. Instead
of continuing to evaluate lower priority rules, the binding is
terminated and the packet proceeds to the bridge and router blocks on
ingress, or goes out of the port on egress.
Second patch prevents the user from mirroring more than once from a
given {Port, Direction} as this is not supported by the device.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Petr Machata [Fri, 9 Mar 2018 13:33:53 +0000 (15:33 +0200)]
mlxsw: spectrum: Prevent duplicate mirrors
The Spectrum ASIC doesn't support mirroring more than once from a single
binding point (which is a port-direction pair). Therefore detect that a
second binding of a given binding point is attempted.
To that end, extend struct mlxsw_sp_span_inspected_port to track whether
a given binding point is bound or not. Extend
mlxsw_sp_span_entry_port_find() to look for ports based on the full
unique key: port number, direction, and boundness.
Besides fixing the overt bug where configured mirrors are not offloaded,
this also fixes a more subtle bug: mlxsw_sp_span_inspected_port_del()
just defers to mlxsw_sp_span_entry_bound_port_find(), and that used to
find the first port with the right number (disregarding the type). Thus
by adding and removing egress and ingress mirrors in the right order,
one could trick the system into believing it has no egress mirrors when
in fact it did have some. That then caused that
mlxsw_sp_span_port_mtu_update() didn't update mirroring buffer when MTU
was changed.
Fixes: 763b4b70afcd ("mlxsw: spectrum: Add support in matchall mirror TC offloading") Signed-off-by: Petr Machata <petrm@mellanox.com> Signed-off-by: Ido Schimmel <idosch@mellanox.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Jason Wang [Fri, 9 Mar 2018 06:50:34 +0000 (14:50 +0800)]
vhost_net: examine pointer types during un-producing
After commit fc72d1d54dd9 ("tuntap: XDP transmission"), we can
actually queueing XDP pointers in the pointer ring, so we should
examine the pointer type before freeing the pointer.
Fixes: fc72d1d54dd9 ("tuntap: XDP transmission") Reported-by: Michael S. Tsirkin <mst@redhat.com> Acked-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Jason Wang <jasowang@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Jason Wang [Fri, 9 Mar 2018 06:50:33 +0000 (14:50 +0800)]
vhost_net: keep private_data and rx_ring synced
We get pointer ring from the exported sock, this means we should keep
rx_ring and vq->private synced during both vq stop and backend set,
otherwise we may see stale rx_ring.
Fixes: c67df11f6e480 ("vhost_net: try batch dequing from skb array") Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Jason Wang <jasowang@redhat.com> Acked-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Fixes: c67df11f6e480 ("vhost_net: try batch dequing from skb array") Signed-off-by: Alexander Potapenko <glider@google.com> Signed-off-by: Jason Wang <jasowang@redhat.com> Acked-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Daniel Axtens [Fri, 9 Mar 2018 03:06:09 +0000 (14:06 +1100)]
net: use skb_is_gso_sctp() instead of open-coding
As well as the basic conversion, I noticed that a lot of the
SCTP code checks gso_type without first checking skb_is_gso()
so I have added that where appropriate.
Also, document the helper.
Cc: Daniel Borkmann <daniel@iogearbox.net> Cc: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com> Signed-off-by: Daniel Axtens <dja@axtens.net> Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Mon, 5 Mar 2018 16:51:03 +0000 (08:51 -0800)]
ieee802154: 6lowpan: fix possible NULL deref in lowpan_device_event()
A tun device type can trivially be set to arbitrary value using
TUNSETLINK ioctl().
Therefore, lowpan_device_event() must really check that ieee802154_ptr
is not NULL.
Fixes: 2c88b5283f60d ("ieee802154: 6lowpan: remove check on null") Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Alexander Aring <alex.aring@gmail.com> Cc: Stefan Schmidt <stefan@osg.samsung.com> Reported-by: syzbot <syzkaller@googlegroups.com> Acked-by: Stefan Schmidt <stefan@osg.samsung.com> Signed-off-by: David S. Miller <davem@davemloft.net>
ipv6: fix access to non-linear packet in ndisc_fill_redirect_hdr_option()
Fix the following slab-out-of-bounds kasan report in
ndisc_fill_redirect_hdr_option when the incoming ipv6 packet is not
linear and the accessed data are not in the linear data region of orig_skb.
[ 1503.122508] ==================================================================
[ 1503.122832] BUG: KASAN: slab-out-of-bounds in ndisc_send_redirect+0x94e/0x990
[ 1503.123036] Read of size 1184 at addr ffff8800298ab6b0 by task netperf/1932
[ 1503.158698] Memory state around the buggy address:
[ 1503.158816] ffff8800298ab900: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
[ 1503.158988] ffff8800298ab980: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
[ 1503.159165] >ffff8800298aba00: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
[ 1503.159338] ^
[ 1503.159436] ffff8800298aba80: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
[ 1503.159610] ffff8800298abb00: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
[ 1503.159785] ==================================================================
[ 1503.159964] Disabling lock debugging due to kernel taint
The test scenario to trigger the issue consists of 4 devices:
- H0: data sender, connected to LAN0
- H1: data receiver, connected to LAN1
- GW0 and GW1: routers between LAN0 and LAN1. Both of them have an
ethernet connection on LAN0 and LAN1
On H{0,1} set GW0 as default gateway while on GW0 set GW1 as next hop for
data from LAN0 to LAN1.
Moreover create an ip6ip6 tunnel between H0 and H1 and send 3 concurrent
data streams (TCP/UDP/SCTP) from H0 to H1 through ip6ip6 tunnel (send
buffer size is set to 16K). While data streams are active flush the route
cache on HA multiple times.
I have not been able to identify a given commit that introduced the issue
since, using the reproducer described above, the kasan report has been
triggered from 4.14 and I have not gone back further.
Reported-by: Jianlin Shi <jishi@redhat.com> Reviewed-by: Stefano Brivio <sbrivio@redhat.com> Reviewed-by: Eric Dumazet <edumazet@google.com> Signed-off-by: Lorenzo Bianconi <lorenzo.bianconi@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
====================
hv_netvsc: fix multicast flags and sync
This set of patches deals with the handling of multicast flags
and addresses in transparent VF mode. The recent set of patches
(in linux-net) had a couple of bugs.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>