Eric Dumazet [Thu, 27 Apr 2017 04:10:55 +0000 (21:10 -0700)]
tcp: tcp_rack_reo_timeout() must update tp->tcp_mstamp
I wrongly assumed tp->tcp_mstamp was up to date at the time
tcp_rack_reo_timeout() was called.
It is not true, since we only update tcp->tcp_mstamp when receiving
a packet (as initially done in commit 69e996c58a35 ("tcp: add
tp->tcp_mstamp field")
tcp_rack_reo_timeout() being called by a timer and not an incoming
packet, we need to refresh tp->tcp_mstamp
Oliver Hartkopp [Wed, 26 Apr 2017 18:14:34 +0000 (20:14 +0200)]
can: fix CAN BCM build with CONFIG_PROC_FS disabled
The introduced namespace support moved the BCM variables for procfs into a
per-net data structure. This leads to a build failure with disabled procfs:
on x86_64:
when CONFIG_PROC_FS is not enabled:
../net/can/bcm.c:1541:14: error: 'struct netns_can' has no member named 'bcmproc_dir'
../net/can/bcm.c:1601:14: error: 'struct netns_can' has no member named 'bcmproc_dir'
../net/can/bcm.c:1696:11: error: 'struct netns_can' has no member named 'bcmproc_dir'
../net/can/bcm.c:1707:15: error: 'struct netns_can' has no member named 'bcmproc_dir'
David Howells [Wed, 26 Apr 2017 21:15:55 +0000 (22:15 +0100)]
statx: Kill fd-with-NULL-path support in favour of AT_EMPTY_PATH
With the new statx() syscall, the following both allow the attributes of
the file attached to a file descriptor to be retrieved:
statx(dfd, NULL, 0, ...);
and:
statx(dfd, "", AT_EMPTY_PATH, ...);
Change the code to reject the first option, though this means copying
the path and engaging pathwalk for the fstat() equivalent. dfd can be a
non-directory provided path is "".
[ The timing of this isn't wonderful, but applying this now before we
have statx() in any released kernel, before anybody starts using the
NULL special case. - Linus ]
Johannes Berg [Wed, 26 Apr 2017 07:58:53 +0000 (10:58 +0300)]
mac80211: disentangle iflist_mtx and chanctx_mtx
At least on iwlwifi, sometimes lockdep complains that we can
lock
chanctx_mtx -> mvm.mutex -> iflist_mtx
(due to iterate_interfaces)
and
iflist_mtx -> chanctx_mtx
Remove the latter dependency in mac80211 by using the RTNL
that we already hold in one case, and can relatively easily
achieve in the other case.
mac80211: don't parse encrypted management frames in ieee80211_frame_acked
ieee80211_frame_acked is called when a frame is acked by
the peer. In case this is a management frame, we check
if this an SMPS frame, in which case we can update our
antenna configuration.
When we parse the management frame we look at the category
in case it is an action frame. That byte sits after the IV
in case the frame was encrypted. This means that if the
frame was encrypted, we basically look at the IV instead
of looking at the category. It is then theorically
possible that we think that an SMPS action frame was acked
where really we had another frame that was encrypted.
Since the only management frame whose ack needs to be
tracked is the SMPS action frame, and that frame is not
a robust management frame, it will never be encrypted.
The easiest way to fix this problem is then to not look
at frames that were encrypted.
cfg80211: add request id parameter to .sched_scan_stop() signature
For multiple scheduled scan support the driver needs to know which
scheduled scan request is being stopped. Pass the request id in the
.sched_scan_stop() callback.
nl80211: allow multiple active scheduled scan requests
This patch implements the idea to have multiple scheduled scan requests
running concurrently. It mainly illustrates how to deal with the incoming
request from user-space in terms of backward compatibility. In order to
use multiple scheduled scans user-space needs to provide a flag attribute
NL80211_ATTR_SCHED_SCAN_MULTI to indicate support. If not the request is
treated as a legacy scan.
Drivers currently supporting scheduled scan are now indicating they support
a single scheduled scan request. This obsoletes WIPHY_FLAG_SUPPORTS_SCHED_SCAN.
There's no need to allocate a portid structure and then, for
each of those, walk the interfaces - we can just add a flag
to each interface and walk those directly. Due to padding in
the struct, we can even do it without any memory cost, and
it even simplifies the code.
1) MLX5 bug fixes from Saeed Mahameed et al:
- released wrong resources when firmware timeout happens
- fix wrong check for encapsulation size limits
- UAR memory leak
- ETHTOOL_GRXCLSRLALL failed to fill in info->data
2) Don't cache l3mdev on mis-matches local route, causes net devices to
leak refs. From Robert Shearman.
3) Handle fragmented SKBs properly in macsec driver, the problem is
that we were mis-sizing the sgvec table. From Jason A. Donenfeld.
4) We cannot have checksum offload enabled for inner UDP tunneled
packet during IPSEC, from Ansis Atteka.
5) Fix double SKB free in ravb driver, from Dan Carpenter.
6) Fix CPU port handling in b53 DSA driver, from Florian Dainelli.
7) Don't use on-stack buffers for usb_control_msg() in CAN usb driver,
from Maksim Salau.
8) Fix device leak in macvlan driver, from Herbert Xu. We have to purge
the broadcast queue properly on port destroy.
9) Fix tx ring entry limit on EF10 devices in sfc driver. From Bert
Kenward.
10) Fix memory leaks in team driver, from Pan Bian.
11) Don't setup ipv6_stub before it can be actually used, from Paolo
Abeni.
12) Fix tipc socket flow control accounting, from Parthasarathy
Bhuvaragan.
13) Fix crash on module unload in hso driver, from Andreas Kemnade.
14) Fix purging of bridge multicast entries, the problem is that if we
don't defer it to ndo_uninit it's possible for new entries to get
added after we purge. Fix from Xin Long.
15) Don't return garbage for PACKET_HDRLEN getsockopt, from Alexander
Potapenko.
16) Fix autoneg stall properly in PHY layer, and revert micrel driver
change that was papering over it. From Alexander Kochetkov.
17) Don't dereference an ipv4 route as an ipv6 one in the ip6_tunnnel
code, from Cong Wang.
18) Clear out the congestion control private of the TCP socket in all of
the right places, from Wei Wang.
19) rawv6_ioctl measures SKB length incorrectly, fix from Jamie
Bainbridge.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (41 commits)
ipv6: check raw payload size correctly in ioctl
tcp: memset ca_priv data to 0 properly
ipv6: check skb->protocol before lookup for nexthop
net: core: Prevent from dereferencing null pointer when releasing SKB
macsec: dynamically allocate space for sglist
Revert "phy: micrel: Disable auto negotiation on startup"
net: phy: fix auto-negotiation stall due to unavailable interrupt
net/packet: check length in getsockopt() called with PACKET_HDRLEN
net: ipv6: regenerate host route if moved to gc list
bridge: move bridge multicast cleanup to ndo_uninit
ipv6: fix source routing
qed: Fix error in the dcbx app meta data initialization.
netvsc: fix calculation of available send sections
net: hso: fix module unloading
tipc: fix socket flow control accounting error at tipc_recv_stream
tipc: fix socket flow control accounting error at tipc_send_stream
ipv6: move stub initialization after ipv6 setup completion
team: fix memory leaks
sfc: tx ring can only have 2048 entries for all EF10 NICs
macvlan: Fix device ref leak when purging bc_queue
...
In situations where an skb is paged, the transport header pointer and
tail pointer can be the same because the skb contents are in frags.
This results in ioctl(SIOCINQ/FIONREAD) incorrectly returning a
length of 0 when the length to receive is actually greater than zero.
skb->len is already correctly set in ip6_input_finish() with
pskb_pull(), so use skb->len as it always returns the correct result
for both linear and paged data.
Wei Wang [Wed, 26 Apr 2017 00:38:02 +0000 (17:38 -0700)]
tcp: memset ca_priv data to 0 properly
Always zero out ca_priv data in tcp_assign_congestion_control() so that
ca_priv data is cleared out during socket creation.
Also always zero out ca_priv data in tcp_reinit_congestion_control() so
that when cc algorithm is changed, ca_priv data is cleared out as well.
We should still zero out ca_priv data even in TCP_CLOSE state because
user could call connect() on AF_UNSPEC to disconnect the socket and
leave it in TCP_CLOSE state and later call setsockopt() to switch cc
algorithm on this socket.
WANG Cong [Tue, 25 Apr 2017 21:37:15 +0000 (14:37 -0700)]
ipv6: check skb->protocol before lookup for nexthop
Andrey reported a out-of-bound access in ip6_tnl_xmit(), this
is because we use an ipv4 dst in ip6_tnl_xmit() and cast an IPv4
neigh key as an IPv6 address:
neigh = dst_neigh_lookup(skb_dst(skb),
&ipv6_hdr(skb)->daddr);
if (!neigh)
goto tx_err_link_failure;
if (addr_type == IPV6_ADDR_ANY)
addr6 = &ipv6_hdr(skb)->daddr;
memcpy(&fl6->daddr, addr6, sizeof(fl6->daddr));
Also the network header of the skb at this point should be still IPv4
for 4in6 tunnels, we shold not just use it as IPv6 header.
This patch fixes it by checking if skb->protocol is ETH_P_IPV6: if it
is, we are safe to do the nexthop lookup using skb_dst() and
ipv6_hdr(skb)->daddr; if not (aka IPv4), we have no clue about which
dest address we can pick here, we have to rely on callers to fill it
from tunnel config, so just fall to ip6_route_output() to make the
decision.
Willem de Bruijn [Tue, 25 Apr 2017 19:59:17 +0000 (15:59 -0400)]
virtio-net: on tx, only call napi_disable if tx napi is on
As of tx napi, device down (`ip link set dev $dev down`) hangs unless
tx napi is enabled. Else napi_enable is not called, so napi_disable
will spin on test_and_set_bit NAPI_STATE_SCHED.
====================
ibmvnic: Move sub crq init out of interrupt context
The sub crqs are currently intialized in interrupt context when
handling a crq response fromn the vios server. There is no reason
they must be initialized there.
Moving the initialization of the sub crqs to the ibmvnic_init routine
allows us to do the initialization outside of interrupt context and
make all of the allocations with GFP_KERNEL instead of GFP_ATOMIC.
====================
ibmvnic: Move initialization of sub crqs to ibmvnic_init
The sub crq structures are initialized in interrupt context while
handling the response to crqs when negotiating capabilities for
the driver. The sub crqs do not need to be initialized at this point
and can be moved to being done from ibmvnic_init. Moving the init
of the sub crqs to ibmvnic_init also allows use to allocate the
memory with GFP_KERNEL instead of GFP_ATOMIC.
ibmvnic: Split initialization of scrqs to its own routine
Split the sending of capability request crqs and the initialization
of sub crqs into their own routines. This is a first step to moving
the allocation of sub-crqs out of interrupt context.
The described GPIO reset property is applicable to *all* child PHYs. If
we have one reset line per PHY present on the MDIO bus, these
automatically become properties of the child PHY nodes.
Finally, indicate how the RESET pulse width must be defined, which is
the maximum value of all individual PHYs RESET pulse widths determined
by reading their datasheets.
Fixes: 69226896ad63 ("mdio_bus: Issue GPIO RESET to PHYs.") Signed-off-by: Florian Fainelli <[email protected]> Reviewed-by: Roger Quadros <[email protected]> Signed-off-by: David S. Miller <[email protected]>
====================
tcp: do not use tcp_time_stamp for rcv autotuning
Some devices or linux distributions use HZ=100 or HZ=250
TCP receive buffer autotuning has poor behavior caused by this choice.
Since autotuning happens after 4 ms or 10 ms, short distance flows
get their receive buffer tuned to a very high value, but after an initial
period where it was frozen to (too small) initial value.
With BBR (or other CC allowing to increase BDP), we are willing to
increase tcp_rmem[2], but this receive autotuning defect is a blocker
for hosts dealing with gazillions of TCP flows in the data centers,
since many of them have inflated RCVBUF. Risk of OOM is too high.
Note that TSO autodefer, tcp cubic, and TCP TS options (RFC 7323)
also suffer from our dependency to jiffies (via tcp_time_stamp).
We have ongoing efforts to improve all that in the future.
====================
Eric Dumazet [Tue, 25 Apr 2017 17:15:41 +0000 (10:15 -0700)]
tcp: switch rcv_rtt_est and rcvq_space to high resolution timestamps
Some devices or distributions use HZ=100 or HZ=250
TCP receive buffer autotuning has poor behavior caused by this choice.
Since autotuning happens after 4 ms or 10 ms, short distance flows
get their receive buffer tuned to a very high value, but after an initial
period where it was frozen to (too small) initial value.
With tp->tcp_mstamp introduction, we can switch to high resolution
timestamps almost for free (at the expense of 8 additional bytes per
TCP structure)
Note that some TCP stacks use usec TCP timestamps where this
patch makes even more sense : Many TCP flows have < 500 usec RTT.
Hopefully this finer TS option can be standardized soon.
We call skb_cow_data, which is good anyway to ensure we can actually
modify the skb as such (another error from prior). Now that we have the
number of fragments required, we can safely allocate exactly that amount
of memory.
Fixes: c09440f7dcb3 ("macsec: introduce IEEE 802.1AE driver") Signed-off-by: Jason A. Donenfeld <[email protected]> Acked-by: Sabrina Dubroca <[email protected]> Signed-off-by: David S. Miller <[email protected]>
net: phy: fix auto-negotiation stall due to unavailable interrupt
The Ethernet link on an interrupt driven PHY was not coming up if the Ethernet
cable was plugged before the Ethernet interface was brought up.
The patch trigger PHY state machine to update link state if PHY was requested to
do auto-negotiation and auto-negotiation complete flag already set.
During power-up cycle the PHY do auto-negotiation, generate interrupt and set
auto-negotiation complete flag. Interrupt is handled by PHY state machine but
doesn't update link state because PHY is in PHY_READY state. After some time
MAC bring up, start and request PHY to do auto-negotiation. If there are no new
settings to advertise genphy_config_aneg() doesn't start PHY auto-negotiation.
PHY continue to stay in auto-negotiation complete state and doesn't fire
interrupt. At the same time PHY state machine expect that PHY started
auto-negotiation and is waiting for interrupt from PHY and it won't get it.
Merge tag 'sound-4.11' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound
Pull sound fixes from Takashi Iwai:
"Since we got a bonus week, let me try to screw a few pending fixes.
A slightly large fix is the locking fix in ASoC STI driver, but it's
pretty board-specific, and the risk is fairly low.
All the rest are small / trivial fixes, mostly marked as stable, for
ALSA sequencer core, ASoC topology, ASoC Intel bytcr and Firewire
drivers"
* tag 'sound-4.11' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
ASoC: intel: Fix PM and non-atomic crash in bytcr drivers
ALSA: firewire-lib: fix inappropriate assignment between signed/unsigned type
ALSA: seq: Don't break snd_use_lock_sync() loop by timeout
ASoC: topology: Fix to store enum text values
ASoC: STI: Fix null ptr deference in IRQ handler
ALSA: oxfw: fix regression to handle Stanton SCS.1m/1d
Kalle Valo [Wed, 26 Apr 2017 11:21:00 +0000 (14:21 +0300)]
Merge tag 'iwlwifi-next-for-kalle-2017-04-26' of git://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi/iwlwifi-next
One more pull-request intended for 4.12. These are the changes:
* The firmware for 7265D and 3168 NICs is frozen at version 29;
* Sari continues working heavily on support for A000 series;
* A bunch of fixes;
* Some cleanups here and there;
James Hughes [Tue, 25 Apr 2017 09:15:06 +0000 (10:15 +0100)]
brcmfmac: Make skb header writable before use
The driver was making changes to the skb_header without
ensuring it was writable (i.e. uncloned).
This patch also removes some boiler plate header size
checking/adjustment code as that is also handled by the
skb_cow_header function used to make header writable.
James Hughes [Mon, 24 Apr 2017 11:40:50 +0000 (12:40 +0100)]
brcmfmac: Ensure pointer correctly set if skb data location changes
The incoming skb header may be resized if header space is
insufficient, which might change the data adddress in the skb.
Ensure that a cached pointer to that data is correctly set by
moving assignment to after any possible changes.
Pan Bian [Mon, 24 Apr 2017 00:40:28 +0000 (08:40 +0800)]
rndis_wlan: add return value validation
Function create_singlethread_workqueue() will return a NULL pointer if
there is no enough memory, and its return value should be validated
before using. However, in function rndis_wlan_bind(), its return value
is not checked. This may cause NULL dereference bugs. This patch fixes
it.
Pan Bian [Sun, 23 Apr 2017 13:19:38 +0000 (21:19 +0800)]
libertas: check return value of alloc_workqueue
Function alloc_workqueue() will return a NULL pointer if there is no
enough memory, and its return value should be validated before using.
However, in function if_spi_probe(), its return value is not checked.
This may result in a NULL dereference bug. This patch fixes the bug.
Pan Bian [Sun, 23 Apr 2017 07:00:23 +0000 (15:00 +0800)]
mt7601u: check return value of alloc_skb
Function alloc_skb() will return a NULL pointer if there is no enough
memory. However, in function mt7601u_mcu_msg_alloc(), its return value
is not validated before it is used. This patch fixes it.
Merge tag 'arc-4.11-final' of git://git.kernel.org/pub/scm/linux/kernel/git/vgupta/arc
Pull ARC fix from Vineet Gupta:
"Last minute fixes for ARC:
- build error in Mellanox nps platform
- addressing lack of saving FPU regs in releavnt configs"
* tag 'arc-4.11-final' of git://git.kernel.org/pub/scm/linux/kernel/git/vgupta/arc:
ARCv2: entry: save Accumulator register pair (r58:59) if present
ARC: [plat-eznps] Fix build error
Eric Dumazet [Tue, 25 Apr 2017 18:36:52 +0000 (11:36 -0700)]
net: move xdp_prog field in RX cache lines
(struct net_device, xdp_prog) field should be moved in RX cache lines,
reducing latencies when a single packet is received on idle host,
since netif_elide_gro() needs it.
Sara Sharon [Sun, 8 Jan 2017 14:45:46 +0000 (16:45 +0200)]
iwlwifi: adjust NVM parsing APIs for new a000 method
In a000 devices we will get all nvm data from the firmware,
and can save most of the parsing.
Export two APIs that op mode will still use.
Adjust API of init_sbands to be independent of NVM file structure
so it can be used by op mode as well.
Johannes Berg [Mon, 6 Mar 2017 14:49:00 +0000 (15:49 +0100)]
iwlwifi: pcie: apply no-reclaim logic only to group 0
When applying no-reclaim logic to commands other than the group
zero for legacy commands, commands such as 0x1c (TX_CMD in group
0) can't be used in any other group. Fix that by applying this
logic only for group 0 - it's not and should never be needed for
any other groups.
Sara Sharon [Sun, 5 Mar 2017 09:38:58 +0000 (11:38 +0200)]
iwlwifi: rename wait_for_tx_queues_empty
Rename current wait_tx_queue_empty to wait_tx_queues_empty since
it waits for multiple queues (up to 32).
Next patch will add a wait for single TX queue which is needed for
gen2 to be scalable for 512.
Sara Sharon [Mon, 6 Feb 2017 17:09:32 +0000 (19:09 +0200)]
iwlwifi: mvm: support station type API
Support change to ADD_STA API to support station types.
Each station is assigned its type.
This simplifies FW handling of the broadcast and multicast
stations:
* broadcast station is identified by its type and not the mac
address.
* multicast queue is no longer treated differently. The opening
and closing of it is done by referring to its station.
There is no need to specify it in the MAC command.
* When disabling TX to all station driver can disable the traffic
on multicast station, so FW doesn't have to do it.
Change is backward compatible.
Change the order of adding and removing the stations according to
FW requirements.
Sara Sharon [Thu, 23 Feb 2017 11:15:07 +0000 (13:15 +0200)]
iwlwifi: mvm: remove references to queue_info in new TX path
Most of the fields aren't needed in new TX path.
Enlarging the struct to 512 queues will consume a lot of memory.
Remove all references to the struct in the new TX path.
Move mac80211 queue mapping outside, since it will be needed per
queue for TVQM mode.
Add warning in paths that shouldn't be hit.
Johannes Berg [Tue, 28 Feb 2017 21:45:25 +0000 (22:45 +0100)]
iwlwifi: remove module loading failure message
When CONFIG_DEBUG_TEST_DRIVER_REMOVE is set, iwlwifi crashes
when the opmode module cannot be loaded, due to completing
the completion before using drv->dev, which can then already
be freed.
Fix this by removing the (fairly useless) message. Moving the
completion later causes a deadlock instead, so that's not an
option.
Sara Sharon [Wed, 22 Feb 2017 10:08:19 +0000 (12:08 +0200)]
iwlwifi: mvm: use defines instead of variables for shared dwell times
Most of the dwells are constant across different scan types.
Use defines instead of depending on scan type.
This is needed as preparation to having different scan type per
band.
Sara Sharon [Thu, 23 Feb 2017 20:42:01 +0000 (22:42 +0200)]
iwlwifi: mvm: remove color definition
We use the full station field as sta_id, and never
use the color.
FW are about to clean the color out, so those defines
are incorrect now (and were redundant before).
Sara Sharon [Wed, 22 Feb 2017 17:34:17 +0000 (19:34 +0200)]
iwlwifi: mvm: move internally to use bigger INVALID_TXQ
We can't use IEEE80211_INVAL_HW_QUEUE to mark a queue as
invalid since 255 will be a valid value for a TVQM queue
index.
Use IWL_MVM_INVALID_QUEUE instead for accessing txq_id.
reserved_queue can stay a u8 since reserved_queue is not
used when TVQM is enabled.
Johannes Berg [Thu, 13 Apr 2017 13:50:27 +0000 (15:50 +0200)]
mac80211: rewrite monitor mode delivery logic
The monitor mode delivery logic makes it hard to add any
kind of filtering in an efficient way, because the monitor
SKB is created first and then passed to all interfaces.
Rewrite the logic to create the monitor SKB the first time
it's actually needed, and then keep delivering it.
cfg80211: Fix dfs state propagation for non-DFS center channel
When part of a bigger bandwidth (160 MHz) channel falls in DFS
channel range it is possible that the center frequency may not
necessarily be a radar channel. Remove the sanity check on channel
flag for IEEE80211_CHAN_RADAR in regulatory_propagate_dfs_state(),
this should fix the dfs state propagation for non-DFS center freq
which has DFS channels in it's bandwidth, should also fix unnecessary
WARN_ON() spam in regulatory_propagate_dfs_state().
Fixes: 8976672736d6 ("cfg80211: Share Channel DFS state across wiphys of same DFS domain") Reported-by: Johannes Berg <[email protected]> Signed-off-by: Vasanthakumar Thiagarajan <[email protected]> Signed-off-by: Johannes Berg <[email protected]>
Sara Sharon [Wed, 22 Feb 2017 17:35:10 +0000 (19:35 +0200)]
iwlwifi: mvm: map cab_queue to different txq_id
cab_queue can now get bigger than u8, since in TVQM we will support
512 queues..
Support it by maintaining internal mapping between the actual number
and mac80211 queue (IWL_MVM_DQA_GCAST_QUEUE).
For pre-a000 the internal queue will be the same as the mac80211
queue.
Haim Dreyfuss [Thu, 2 Feb 2017 12:49:50 +0000 (14:49 +0200)]
iwlwifi: mvm: Ignore wifi mcc update in the driver while associated
Wifi mcc (mobile country code) update is forbidden while associated.
Currently, FW prevents these updates and the driver is unaware to
this logic. From now on, the FW sends every wifi mcc update to the
driver. The driver in his turn needs to decide whether to
ignore it or not, depends on the association state.
Johannes Berg [Fri, 24 Feb 2017 11:02:22 +0000 (12:02 +0100)]
iwlwifi: mvm: avoid variable shadowing
Remove an extra variable 'queue' that already exists.
Also, since there are no code paths that use 'queue'
without intializing it, remove the unnecessary zero
initialization.
net/packet: check length in getsockopt() called with PACKET_HDRLEN
In the case getsockopt() is called with PACKET_HDRLEN and optlen < 4
|val| remains uninitialized and the syscall may behave differently
depending on its value, and even copy garbage to userspace on certain
architectures. To fix this we now return -EINVAL if optlen is too small.
David Ahern [Tue, 25 Apr 2017 16:17:29 +0000 (09:17 -0700)]
net: ipv6: regenerate host route if moved to gc list
Taking down the loopback device wreaks havoc on IPv6 routing. By
extension, taking down a VRF device wreaks havoc on its table.
Dmitry and Andrey both reported heap out-of-bounds reports in the IPv6
FIB code while running syzkaller fuzzer. The root cause is a dead dst
that is on the garbage list gets reinserted into the IPv6 FIB. While on
the gc (or perhaps when it gets added to the gc list) the dst->next is
set to an IPv4 dst. A subsequent walk of the ipv6 tables causes the
out-of-bounds access.
Andrey's reproducer was the key to getting to the bottom of this.
With IPv6, host routes for an address have the dst->dev set to the
loopback device. When the 'lo' device is taken down, rt6_ifdown initiates
a walk of the fib evicting routes with the 'lo' device which means all
host routes are removed. That process moves the dst which is attached to
an inet6_ifaddr to the gc list and marks it as dead.
The recent change to keep global IPv6 addresses added a new function,
fixup_permanent_addr, that is called on admin up. That function restarts
dad for an inet6_ifaddr and when it completes the host route attached
to it is inserted into the fib. Since the route was marked dead and
moved to the gc list, re-inserting the route causes the reported
out-of-bounds accesses. If the device with the address is taken down
or the address is removed, the WARN_ON in fib6_del is triggered.
All of those faults are fixed by regenerating the host route if the
existing one has been moved to the gc list, something that can be
determined by checking if the rt6i_ref counter is 0.
Xin Long [Tue, 25 Apr 2017 14:58:37 +0000 (22:58 +0800)]
bridge: move bridge multicast cleanup to ndo_uninit
During removing a bridge device, if the bridge is still up, a new mdb entry
still can be added in br_multicast_add_group() after all mdb entries are
removed in br_multicast_dev_del(). Like the path:
This could happen when ip link remove a bridge or destroy a netns with a
bridge device inside.
With Nikolay's suggestion, this patch is to clean up bridge multicast in
ndo_uninit after bridge dev is shutdown, instead of br_dev_delete, so
that netif_running check in br_multicast_add_group can avoid this issue.
v1->v2:
- fix this issue by moving br_multicast_dev_del to ndo_uninit, instead
of calling dev_close in br_dev_delete.
(NOTE: Depends upon b6fe0440c637 ("bridge: implement missing ndo_uninit()"))
Commit a149e7c7ce81 ("ipv6: sr: add support for SRH injection through
setsockopt") introduced handling of IPV6_SRCRT_TYPE_4, but at the same
time restricted it to only IPV6_SRCRT_TYPE_0 and
IPV6_SRCRT_TYPE_4. Previously, ipv6_push_exthdr() and fl6_update_dst()
would also handle other values (ie STRICT and TYPE_2).
Restore previous source routing behavior, by handling IPV6_SRCRT_STRICT
and IPV6_SRCRT_TYPE_2 the same way as IPV6_SRCRT_TYPE_0 in
ipv6_push_exthdr() and fl6_update_dst().
Fixes: a149e7c7ce81 ("ipv6: sr: add support for SRH injection through setsockopt") Signed-off-by: Sabrina Dubroca <[email protected]> Reviewed-by: Hannes Frederic Sowa <[email protected]> Signed-off-by: David S. Miller <[email protected]>
David S. Miller [Tue, 18 Apr 2017 19:36:58 +0000 (15:36 -0400)]
net: Generic XDP
This provides a generic SKB based non-optimized XDP path which is used
if either the driver lacks a specific XDP implementation, or the user
requests it via a new IFLA_XDP_FLAGS value named XDP_FLAGS_SKB_MODE.
It is arguable that perhaps I should have required something like
this as part of the initial XDP feature merge.
I believe this is critical for two reasons:
1) Accessibility. More people can play with XDP with less
dependencies. Yes I know we have XDP support in virtio_net, but
that just creates another depedency for learning how to use this
facility.
I wrote this to make life easier for the XDP newbies.
2) As a model for what the expected semantics are. If there is a pure
generic core implementation, it serves as a semantic example for
driver folks adding XDP support.
One thing I have not tried to address here is the issue of
XDP_PACKET_HEADROOM, thanks to Daniel for spotting that. It seems
incredibly expensive to do a skb_cow(skb, XDP_PACKET_HEADROOM) or
whatever even if the XDP program doesn't try to push headers at all.
I think we really need the verifier to somehow propagate whether
certain XDP helpers are used or not.
v5:
- Handle both negative and positive offset after running prog
- Fix mac length in XDP_TX case (Alexei)
- Use rcu_dereference_protected() in free_netdev (kbuild test robot)
v4:
- Fix MAC header adjustmnet before calling prog (David Ahern)
- Disable LRO when generic XDP is installed (Michael Chan)
- Bypass qdisc et al. on XDP_TX and record the event (Alexei)
- Do not perform generic XDP on reinjected packets (DaveM)
v3:
- Make sure XDP program sees packet at MAC header, push back MAC
header if we do XDP_TX. (Alexei)
- Elide GRO when generic XDP is in use. (Alexei)
- Add XDP_FLAG_SKB_MODE flag which the user can use to request generic
XDP even if the driver has an XDP implementation. (Alexei)
- Report whether SKB mode is in use in rtnl_xdp_fill() via XDP_FLAGS
attribute. (Daniel)
v2:
- Add some "fall through" comments in switch statements based
upon feedback from Andrew Lunn
- Use RCU for generic xdp_prog, thanks to Johannes Berg.
Teng Qin [Tue, 25 Apr 2017 02:00:37 +0000 (19:00 -0700)]
bpf: map_get_next_key to return first key on NULL
When iterating through a map, we need to find a key that does not exist
in the map so map_get_next_key will give us the first key of the map.
This often requires a lot of guessing in production systems.
This patch makes map_get_next_key return the first key when the key
pointer in the parameter is NULL.
netvsc: fix calculation of available send sections
My change (introduced in 4.11) to use find_first_clear_bit
incorrectly assumed that the size argument was words, not bits.
The effect was only a small limited number of the available send
sections were being actually used. This can cause performance loss
with some workloads.
Since map_words is now used only during initialization, it can
be on stack instead of in per-device data.
Fixes: b58a185801da ("netvsc: simplify get next send section") Signed-off-by: Stephen Hemminger <[email protected]> Signed-off-by: David S. Miller <[email protected]>
This patch adds hardware channel APIs support between
VF and PF for tunnelling configuration for the VFs.
According to that configuration VFs can run VXLAN/GENEVE/GRE
tunnels over it with tunnel features offloaded.
Using these APIs VF can also request for UDP ports configuration
to the PF, although PF and it's child VFs share the same port.
This patch configures UDP ports locally instead of
configuring them in deferred context which would be
helpful in synchronizing UDP ports configuration for VFs
which will be enabled in further patches.
qede: Disable tunnel offloads for non offloaded UDP ports
This patch disables tunnel offloads via ndo_features_check()
if given UDP port is not offloaded to hardware. This in turn
allows to run multiple tunnel interfaces using different UDP ports.
This patch changes the tunnel APIs to use per tunnel
info instead of using bitmasks for all tunnels and also
uses single struct to hold the data to prepare multiple
variant of tunnel configuration ramrods to be sent to the hardware.
tipc: fix socket flow control accounting error at tipc_recv_stream
Until now in tipc_recv_stream(), we update the received
unacknowledged bytes based on a stack variable and not based on the
actual message size.
If the user buffer passed at tipc_recv_stream() is smaller than the
received skb, the size variable in stack differs from the actual
message size in the skb. This leads to a flow control accounting
error causing permanent congestion.
In this commit, we fix this accounting error by always using the
size of the incoming message.
Fixes: 10724cc7bb78 ("tipc: redesign connection-level flow control") Signed-off-by: Parthasarathy Bhuvaragan <[email protected]> Reviewed-by: Jon Maloy <[email protected]> Signed-off-by: David S. Miller <[email protected]>
tipc: fix socket flow control accounting error at tipc_send_stream
Until now in tipc_send_stream(), we return -1 when the socket
encounters link congestion even if the socket had successfully
sent partial data. This is incorrect as the application resends
the same the partial data leading to data corruption at
receiver's end.
In this commit, we return the partially sent bytes as the return
value at link congestion.
Fixes: 10724cc7bb78 ("tipc: redesign connection-level flow control") Signed-off-by: Parthasarathy Bhuvaragan <[email protected]> Reviewed-by: Jon Maloy <[email protected]> Signed-off-by: David S. Miller <[email protected]>
Merge tag 'asoc-fix-v4.11-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound into for-linus
ASoC: Fixes for v4.11
A few last minute fixes for v4.11, the STI fix is relatively large but
driver specific and has been cooking in -next for a little while now:
- A fix from Takashi for some suspend/resume related crashes in the
Intel drivers.
- A fix from Mousumi Jana for issues with incorrectly created
enumeration controls generated from topology files which could cause
problems for userspace.
- Fixes from Arnaud Pouliquen for some crashes due to races with the
interrupt handler in the STI driver.
Paolo Abeni [Mon, 24 Apr 2017 12:18:28 +0000 (14:18 +0200)]
ipv6: move stub initialization after ipv6 setup completion
The ipv6 stub pointer is currently initialized before the ipv6
routing subsystem: a 3rd party can access and use such stub
before the routing data is ready.
Moreover, such pointer is not cleared in case of initialization
error, possibly leading to dangling pointers usage.
This change addresses the above moving the stub initialization
at the end of ipv6 init code.
Fixes: 5f81bd2e5d80 ("ipv6: export a stub for IPv6 symbols used by vxlan") Signed-off-by: Paolo Abeni <[email protected]> Acked-by: Cong Wang <[email protected]> Signed-off-by: David S. Miller <[email protected]>
David S. Miller [Tue, 25 Apr 2017 15:41:57 +0000 (11:41 -0400)]
Merge branch 'l2tpeth-info'
Guillaume Nault says:
====================
l2tp: add informations about l2tpeth interfaces in /sys
Patch #1 lets userspace retrieve the naming scheme of an l2tpeth
interface, using /sys/class/net/<iface>/name_assign_type.
Patch #2 adds the DEVTYPE field in /sys/class/net/<iface>/uevent so
that userspace can reliably know if a device is an l2tpeth interface.
====================