Use try_cmpxchg instead of cmpxchg (*ptr, old, new) == old in
napi_if_scheduled_mark_missed. x86 CMPXCHG instruction returns
success in ZF flag, so this change saves a compare after cmpxchg
(and related move instruction in front of cmpxchg).
Also, try_cmpxchg implicitly assigns old *ptr value to "old" when cmpxchg
fails, enabling further code simplifications.
Cc: Eric Dumazet <edumazet@google.com>
Cc: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Uros Bizjak <ubizjak@gmail.com>
Link: https://lore.kernel.org/r/20220822143243.2798-1-ubizjak@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Current release - new code bugs:
- dsa: don't dereference NULL extack in dsa_slave_changeupper()
- dpaa: fix <1G ethernet on LS1046ARDB
- neigh: don't call kfree_skb() under spin_lock_irqsave()
Previous releases - regressions:
- r8152: fix the RX FIFO settings when suspending
- dsa: microchip: keep compatibility with device tree blobs with
no phy-mode
- Revert "net: macsec: update SCI upon MAC address change."
- Revert "xfrm: update SA curlft.use_time", comply with RFC 2367
Previous releases - always broken:
- netfilter: conntrack: work around exceeded TCP receive window
- ipsec: fix a null pointer dereference of dst->dev on a metadata
dst in xfrm_lookup_with_ifid
- moxa: get rid of asymmetry in DMA mapping/unmapping
- dsa: microchip: make learning configurable and keep it off
while standalone
- ice: xsk: prohibit usage of non-balanced queue id
- rxrpc: fix locking in rxrpc's sendmsg
Misc:
- another chunk of sysctl data race silencing
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
-----BEGIN PGP SIGNATURE-----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=azH5
-----END PGP SIGNATURE-----
Merge tag 'net-6.0-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Pull networking fixes from Jakub Kicinski:
"Including fixes from ipsec and netfilter (with one broken Fixes tag).
Current release - new code bugs:
- dsa: don't dereference NULL extack in dsa_slave_changeupper()
- dpaa: fix <1G ethernet on LS1046ARDB
- neigh: don't call kfree_skb() under spin_lock_irqsave()
Previous releases - regressions:
- r8152: fix the RX FIFO settings when suspending
- dsa: microchip: keep compatibility with device tree blobs with no
phy-mode
- Revert "net: macsec: update SCI upon MAC address change."
- Revert "xfrm: update SA curlft.use_time", comply with RFC 2367
Previous releases - always broken:
- netfilter: conntrack: work around exceeded TCP receive window
- ipsec: fix a null pointer dereference of dst->dev on a metadata dst
in xfrm_lookup_with_ifid
- moxa: get rid of asymmetry in DMA mapping/unmapping
- dsa: microchip: make learning configurable and keep it off while
standalone
- ice: xsk: prohibit usage of non-balanced queue id
- rxrpc: fix locking in rxrpc's sendmsg
Misc:
- another chunk of sysctl data race silencing"
* tag 'net-6.0-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (87 commits)
net: lantiq_xrx200: restore buffer if memory allocation failed
net: lantiq_xrx200: fix lock under memory pressure
net: lantiq_xrx200: confirm skb is allocated before using
net: stmmac: work around sporadic tx issue on link-up
ionic: VF initial random MAC address if no assigned mac
ionic: fix up issues with handling EAGAIN on FW cmds
ionic: clear broken state on generation change
rxrpc: Fix locking in rxrpc's sendmsg
net: ethernet: mtk_eth_soc: fix hw hash reporting for MTK_NETSYS_V2
MAINTAINERS: rectify file entry in BONDING DRIVER
i40e: Fix incorrect address type for IPv6 flow rules
ixgbe: stop resetting SYSTIME in ixgbe_ptp_start_cyclecounter
net: Fix a data-race around sysctl_somaxconn.
net: Fix a data-race around netdev_unregister_timeout_secs.
net: Fix a data-race around gro_normal_batch.
net: Fix data-races around sysctl_devconf_inherit_init_net.
net: Fix data-races around sysctl_fb_tunnels_only_for_init_net.
net: Fix a data-race around netdev_budget_usecs.
net: Fix data-races around sysctl_max_skb_frags.
net: Fix a data-race around netdev_budget.
...
Petr Machata says:
====================
mlxsw: Remove some unused code
This patchset removes code that is not used anymore after the following two
commits removed all users:
- commit b0d80c013b ("mlxsw: Remove Mellanox SwitchX-2 ASIC support")
- commit 9b43fbb8ce ("mlxsw: Remove Mellanox SwitchIB ASIC support")
====================
Link: https://lore.kernel.org/r/cover.1661350629.git.petrm@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Function mlxsw_core_port_type_get() is no longer used. So remove it.
Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
port_type_set devlink op is no longer used by any mlxsw driver,
so remove it.
Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
There are some IB leftovers that are no longer used in the code.
So remove them.
Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Jiri Pirko says:
====================
net: devlink: sync flash and dev info commands
Purpose of this patchset is to introduce consistency between two devlink
commands:
devlink dev info
Shows versions of running default flash target and components.
devlink dev flash
Flashes default flash target or component name (if specified
on cmdline).
Currently it is up to the driver what versions to expose and what flash
update component names to accept. This is inconsistent. Thankfully, only
netdevsim currently using components so it is still time
to sanitize this.
This patchset makes sure, that devlink.c calls into driver for
component flash update only in case the driver exposes the same version
name.
Example:
$ devlink dev info
netdevsim/netdevsim10:
driver netdevsim
versions:
running:
fw.mgmt 10.20.30
stored:
fw.mgmt 10.20.30
$ devlink dev flash netdevsim/netdevsim10 file somefile.bin
[fw.mgmt] Preparing to flash
[fw.mgmt] Flashing 100%
[fw.mgmt] Flash select
[fw.mgmt] Flashing done
$ devlink dev flash netdevsim/netdevsim10 file somefile.bin component fw.mgmt
[fw.mgmt] Preparing to flash
[fw.mgmt] Flashing 100%
[fw.mgmt] Flash select
[fw.mgmt] Flashing done
$ devlink dev flash netdevsim/netdevsim10 file somefile.bin component dummy
Error: selected component is not supported by this device.
====================
Link: https://lore.kernel.org/r/20220824122011.1204330-1-jiri@resnulli.us
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Limit the acceptance of component name passed to cmd_flash_update() to
match one of the versions returned by info_get(), marked by version type.
This makes things clearer and enforces 1:1 mapping between exposed
version and accepted flash component.
Check VERSION_TYPE_COMPONENT version type during cmd_flash_update()
execution by calling info_get() with different "req" context.
That causes info_get() to lookup the component name instead of
filling-up the netlink message.
Remove "UPDATE_COMPONENT" flag which becomes used.
Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Fix the only component user which is netdevsim. It uses component named
"fw.mgmt" in selftests. So add this version to info_get() output with
version type component.
Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Whenever the driver is called by his info_get() op, it may put multiple
version names and values to the netlink message. Extend by additional
helper devlink_info_version_running/stored_put_ext() that allows to
specify a version type that indicates when particular version name
represents a flash component.
This is going to be used in follow-up patch calling info_get() during
flash update command checking if version with this the version type
exists.
Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Aleksander Jan Bajkowski says:
====================
net: lantiq_xrx200: fix errors under memory pressure
This series fixes issues that can occur in the driver under memory pressure.
Situations when the system cannot allocate memory are rare, so the mentioned
bugs have been fixed recently. The patches have been tested on a BT Home
router with the Lantiq xRX200 chipset.
Changelog:
v3: - removed netdev_err() log from the first patch
v2:
- the second patch has been changed, so that under memory pressure situation
the driver will not receive packets indefinitely regardless of the NAPI budget,
- the third patch has been added.
====================
Link: https://lore.kernel.org/r/20220824215408.4695-1-olek2@wp.pl
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
In a situation where memory allocation fails, an invalid buffer address
is stored. When this descriptor is used again, the system panics in the
build_skb() function when accessing memory.
Fixes: 7ea6cd16f1 ("lantiq: net: fix duplicated skb in rx descriptor ring")
Signed-off-by: Aleksander Jan Bajkowski <olek2@wp.pl>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When the xrx200_hw_receive() function returns -ENOMEM, the NAPI poll
function immediately returns an error.
This is incorrect for two reasons:
* the function terminates without enabling interrupts or scheduling NAPI,
* the error code (-ENOMEM) is returned instead of the number of received
packets.
After the first memory allocation failure occurs, packet reception is
locked due to disabled interrupts from DMA..
Fixes: fe1a56420c ("net: lantiq: Add Lantiq / Intel VRX200 Ethernet driver")
Signed-off-by: Aleksander Jan Bajkowski <olek2@wp.pl>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
xrx200_hw_receive() assumes build_skb() always works and goes straight
to skb_reserve(). However, build_skb() can fail under memory pressure.
Add a check in case build_skb() failed to allocate and return NULL.
Fixes: e015593573 ("net: lantiq_xrx200: convert to build_skb")
Reported-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Aleksander Jan Bajkowski <olek2@wp.pl>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
This is a follow-up to the discussion in [0]. It seems to me that
at least the IP version used on Amlogic SoC's sometimes has a problem
if register MAC_CTRL_REG is written whilst the chip is still processing
a previous write. But that's just a guess.
Adding a delay between two writes to this register helps, but we can
also simply omit the offending second write. This patch uses the second
approach and is based on a suggestion from Qi Duan.
Benefit of this approach is that we can save few register writes, also
on not affected chip versions.
[0] https://www.spinics.net/lists/netdev/msg831526.html
Fixes: bfab27a146 ("stmmac: add the experimental PCI support")
Suggested-by: Qi Duan <qi.duan@amlogic.com>
Suggested-by: Jerome Brunet <jbrunet@baylibre.com>
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Link: https://lore.kernel.org/r/e99857ce-bd90-5093-ca8c-8cd480b5a0a2@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Tony Nguyen says:
====================
Intel Wired LAN Driver Updates 2022-08-24 (ixgbe, i40e)
This series contains updates to ixgbe and i40e drivers.
Jake stops incorrect resetting of SYSTIME registers when starting
cyclecounter for ixgbe.
Sylwester corrects a check on source IP address when validating destination
for i40e.
* '10GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue:
i40e: Fix incorrect address type for IPv6 flow rules
ixgbe: stop resetting SYSTIME in ixgbe_ptp_start_cyclecounter
====================
Link: https://lore.kernel.org/r/20220824193748.874343-1-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Shannon Nelson says:
====================
ionic: bug fixes
These are a couple of maintenance bug fixes for the Pensando ionic
networking driver.
Mohamed takes care of a "plays well with others" issue where the
VF spec is a bit vague on VF mac addresses, but certain customers
have come to expect behavior based on other vendor drivers.
Shannon addresses a couple of corner cases seen in internal
stress testing.
====================
Link: https://lore.kernel.org/r/20220824165051.6185-1-snelson@pensando.io
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Assign a random mac address to the VF interface station
address if it boots with a zero mac address in order to match
similar behavior seen in other VF drivers. Handle the errors
where the older firmware does not allow the VF to set its own
station address.
Newer firmware will allow the VF to set the station mac address
if it hasn't already been set administratively through the PF.
Setting it will also be allowed if the VF has trust.
Fixes: fbb39807e9 ("ionic: support sr-iov operations")
Signed-off-by: R Mohamed Shah <mohamed@pensando.io>
Signed-off-by: Shannon Nelson <snelson@pensando.io>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
In looping on FW update tests we occasionally see the
FW_ACTIVATE_STATUS command fail while it is in its EAGAIN loop
waiting for the FW activate step to finsh inside the FW. The
firmware is complaining that the done bit is set when a new
dev_cmd is going to be processed.
Doing a clean on the cmd registers and doorbell before exiting
the wait-for-done and cleaning the done bit before the sleep
prevents this from occurring.
Fixes: fbfb803153 ("ionic: Add hardware init and device commands")
Signed-off-by: Shannon Nelson <snelson@pensando.io>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
There is a case found in heavy testing where a link flap happens just
before a firmware Recovery event and the driver gets stuck in the
BROKEN state. This comes from the driver getting interrupted by a FW
generation change when coming back up from the link flap, and the call
to ionic_start_queues() in ionic_link_status_check() fails. This can be
addressed by having the fw_up code clear the BROKEN bit if seen, rather
than waiting for a user to manually force the interface down and then
back up.
Fixes: 9e8eaf8427 ("ionic: stop watchdog when in broken state")
Signed-off-by: Shannon Nelson <snelson@pensando.io>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Fix three bugs in the rxrpc's sendmsg implementation:
(1) rxrpc_new_client_call() should release the socket lock when returning
an error from rxrpc_get_call_slot().
(2) rxrpc_wait_for_tx_window_intr() will return without the call mutex
held in the event that we're interrupted by a signal whilst waiting
for tx space on the socket or relocking the call mutex afterwards.
Fix this by: (a) moving the unlock/lock of the call mutex up to
rxrpc_send_data() such that the lock is not held around all of
rxrpc_wait_for_tx_window*() and (b) indicating to higher callers
whether we're return with the lock dropped. Note that this means
recvmsg() will not block on this call whilst we're waiting.
(3) After dropping and regaining the call mutex, rxrpc_send_data() needs
to go and recheck the state of the tx_pending buffer and the
tx_total_len check in case we raced with another sendmsg() on the same
call.
Thinking on this some more, it might make sense to have different locks for
sendmsg() and recvmsg(). There's probably no need to make recvmsg() wait
for sendmsg(). It does mean that recvmsg() can return MSG_EOR indicating
that a call is dead before a sendmsg() to that call returns - but that can
currently happen anyway.
Without fix (2), something like the following can be induced:
WARNING: bad unlock balance detected!
5.16.0-rc6-syzkaller #0 Not tainted
-------------------------------------
syz-executor011/3597 is trying to release lock (&call->user_mutex) at:
[<ffffffff885163a3>] rxrpc_do_sendmsg+0xc13/0x1350 net/rxrpc/sendmsg.c:748
but there are no more locks to release!
other info that might help us debug this:
no locks held by syz-executor011/3597.
...
Call Trace:
<TASK>
__dump_stack lib/dump_stack.c:88 [inline]
dump_stack_lvl+0xcd/0x134 lib/dump_stack.c:106
print_unlock_imbalance_bug include/trace/events/lock.h:58 [inline]
__lock_release kernel/locking/lockdep.c:5306 [inline]
lock_release.cold+0x49/0x4e kernel/locking/lockdep.c:5657
__mutex_unlock_slowpath+0x99/0x5e0 kernel/locking/mutex.c:900
rxrpc_do_sendmsg+0xc13/0x1350 net/rxrpc/sendmsg.c:748
rxrpc_sendmsg+0x420/0x630 net/rxrpc/af_rxrpc.c:561
sock_sendmsg_nosec net/socket.c:704 [inline]
sock_sendmsg+0xcf/0x120 net/socket.c:724
____sys_sendmsg+0x6e8/0x810 net/socket.c:2409
___sys_sendmsg+0xf3/0x170 net/socket.c:2463
__sys_sendmsg+0xe5/0x1b0 net/socket.c:2492
do_syscall_x64 arch/x86/entry/common.c:50 [inline]
do_syscall_64+0x35/0xb0 arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x44/0xae
[Thanks to Hawkins Jiawei and Khalid Masum for their attempts to fix this]
Fixes: bc5e3a546d ("rxrpc: Use MSG_WAITALL to tell sendmsg() to temporarily ignore signals")
Reported-by: syzbot+7f0483225d0c94cb3441@syzkaller.appspotmail.com
Signed-off-by: David Howells <dhowells@redhat.com>
Reviewed-by: Marc Dionne <marc.dionne@auristor.com>
Tested-by: syzbot+7f0483225d0c94cb3441@syzkaller.appspotmail.com
cc: Hawkins Jiawei <yin31149@gmail.com>
cc: Khalid Masum <khalid.masum.92@gmail.com>
cc: Dan Carpenter <dan.carpenter@oracle.com>
cc: linux-afs@lists.infradead.org
Link: https://lore.kernel.org/r/166135894583.600315.7170979436768124075.stgit@warthog.procyon.org.uk
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
deadlock") in the previous fix pull required cgroup core to grab
cpus_read_lock() before invoking ->attach(). Unfortunately, it missed adding
cpus_read_lock() in cgroup_attach_task_all(). Fix it.
-----BEGIN PGP SIGNATURE-----
iIQEABYIACwWIQTfIjM1kS57o3GsC/uxYfJx3gVYGQUCYwe0GA4cdGpAa2VybmVs
Lm9yZwAKCRCxYfJx3gVYGee0AP9jrsUgnmis/PzqyyPlkD95rRSDyyUNjMjfHnJe
HW+YbgD/XcEo1eJvijqP1g/ZJhRKQl6vA1JSMgnL9obc3wNpGg8=
=7LzT
-----END PGP SIGNATURE-----
Merge tag 'cgroup-for-6.0-rc2-fixes-2' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup
Pull another cgroup fix from Tejun Heo:
"Commit 4f7e723643 ("cgroup: Fix threadgroup_rwsem <->
cpus_read_lock() deadlock") required the cgroup
core to grab cpus_read_lock() before invoking ->attach().
Unfortunately, it missed adding cpus_read_lock() in
cgroup_attach_task_all(). Fix it"
* tag 'cgroup-for-6.0-rc2-fixes-2' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup:
cgroup: Add missing cpus_read_lock() to cgroup_attach_task_all()
qdisc_reset() is clearing qdisc->q.qlen and qdisc->qstats.backlog
_after_ calling qdisc->ops->reset. There is no need to clear them
again in the specific reset function.
Signed-off-by: Zhengchao Shao <shaozhengchao@huawei.com>
Link: https://lore.kernel.org/r/20220824005231.345727-1-shaozhengchao@huawei.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
is_post_ct_flow() function will process only ct_state ESTABLISHED,
then offload_pre_check() function will check FLOW_DISSECTOR_KEY_CT flag.
When config tc filter match ct_state(0/0x3f), dissector->used_keys
with FLOW_DISSECTOR_KEY_CT bit, function offload_pre_check() will
return false, so not offload. This is a special case that can be handled
safely.
Therefore, modify to let initial packet which won't go through conntrack
can be offloaded, as long as the cared ct fields are all zero.
Signed-off-by: Wenjuan Geng <wenjuan.geng@corigine.com>
Reviewed-by: Louis Peens <louis.peens@corigine.com>
Signed-off-by: Simon Horman <simon.horman@corigine.com>
Link: https://lore.kernel.org/r/20220823090122.403631-1-simon.horman@corigine.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Introduce a simple helper function to replace a common pattern.
When accessing the GRO header, we fetch the pointer from frag0,
then test its validity and fetch it from the skb when necessary.
This leads to the pattern
skb_gro_header_fast -> skb_gro_header_hard -> skb_gro_header_slow
recurring many times throughout GRO code.
This patch replaces these patterns with a single inlined function
call, improving code readability.
Signed-off-by: Richard Gobert <richardbgobert@gmail.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20220823071034.GA56142@debian
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
As all callbacks are converted now, fix the text reflecting that change.
Suggested-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Link: https://lore.kernel.org/r/20220823070213.1008956-1-jiri@resnulli.us
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Joanne Koong says:
====================
Add a second bind table hashed by port and address
Currently, there is one bind hashtable (bhash) that hashes by port only.
This patchset adds a second bind table (bhash2) that hashes by port and
address.
The motivation for adding bhash2 is to expedite bind requests in situations
where the port has many sockets in its bhash table entry (eg a large number
of sockets bound to different addresses on the same port), which makes checking
bind conflicts costly especially given that we acquire the table entry spinlock
while doing so, which can cause softirq cpu lockups and can prevent new tcp
connections.
We ran into this problem at Meta where the traffic team binds a large number
of IPs to port 443 and the bind() call took a significant amount of time
which led to cpu softirq lockups, which caused packet drops and other failures
on the machine.
When experimentally testing this on a local server for ~24k sockets bound to
the port, the results seen were:
ipv4:
before - 0.002317 seconds
with bhash2 - 0.000020 seconds
ipv6:
before - 0.002431 seconds
with bhash2 - 0.000021 seconds
The additions to the initial bhash2 submission [0] are:
* Updating bhash2 in the cases where a socket's rcv saddr changes after it has
* been bound
* Adding locks for bhash2 hashbuckets
[0] https://lore.kernel.org/netdev/20220520001834.2247810-1-kuba@kernel.org/
v3: https://lore.kernel.org/netdev/20220722195406.1304948-2-joannelkoong@gmail.com/
v2: https://lore.kernel.org/netdev/20220712235310.1935121-1-joannelkoong@gmail.com/
v1: https://lore.kernel.org/netdev/20220623234242.2083895-2-joannelkoong@gmail.com/
====================
Link: https://lore.kernel.org/r/20220822181023.3979645-1-joannelkoong@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
This patch adds 2 new tests: sk_bind_sendto_listen and
sk_connect_zero_addr.
The sk_bind_sendto_listen test exercises the path where a socket's
rcv saddr changes after it has been added to the binding tables,
and then a listen() on the socket is invoked. The listen() should
succeed.
The sk_bind_sendto_listen test is copied over from one of syzbot's
tests: https://syzkaller.appspot.com/x/repro.c?x=1673a38df00000
The sk_connect_zero_addr test exercises the path where the socket was
never previously added to the binding tables and it gets assigned a
saddr upon a connect() to address 0.
Signed-off-by: Joanne Koong <joannelkoong@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
This test populates the bhash table for a given port with
MAX_THREADS * MAX_CONNECTIONS sockets, and then times how long
a bind request on the port takes.
When populating the bhash table, we create the sockets and then bind
the sockets to the same address and port (SO_REUSEADDR and SO_REUSEPORT
are set). When timing how long a bind on the port takes, we bind on a
different address without SO_REUSEPORT set. We do not set SO_REUSEPORT
because we are interested in the case where the bind request does not
go through the tb->fastreuseport path, which is fragile (eg
tb->fastreuseport path does not work if binding with a different uid).
To run the script:
Usage: ./bind_bhash.sh [-6 | -4] [-p port] [-a address]
6: use ipv6
4: use ipv4
port: Port number
address: ip address
Without any arguments, ./bind_bhash.sh defaults to ipv6 using ip address
"2001:0db8:0:f101::1" on port 443.
On my local machine, I see:
ipv4:
before - 0.002317 seconds
with bhash2 - 0.000020 seconds
ipv6:
before - 0.002431 seconds
with bhash2 - 0.000021 seconds
Signed-off-by: Joanne Koong <joannelkoong@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The current bind hashtable (bhash) is hashed by port only.
In the socket bind path, we have to check for bind conflicts by
traversing the specified port's inet_bind_bucket while holding the
hashbucket's spinlock (see inet_csk_get_port() and
inet_csk_bind_conflict()). In instances where there are tons of
sockets hashed to the same port at different addresses, the bind
conflict check is time-intensive and can cause softirq cpu lockups,
as well as stops new tcp connections since __inet_inherit_port()
also contests for the spinlock.
This patch adds a second bind table, bhash2, that hashes by
port and sk->sk_rcv_saddr (ipv4) and sk->sk_v6_rcv_saddr (ipv6).
Searching the bhash2 table leads to significantly faster conflict
resolution and less time holding the hashbucket spinlock.
Please note a few things:
* There can be the case where the a socket's address changes after it
has been bound. There are two cases where this happens:
1) The case where there is a bind() call on INADDR_ANY (ipv4) or
IPV6_ADDR_ANY (ipv6) and then a connect() call. The kernel will
assign the socket an address when it handles the connect()
2) In inet_sk_reselect_saddr(), which is called when rebuilding the
sk header and a few pre-conditions are met (eg rerouting fails).
In these two cases, we need to update the bhash2 table by removing the
entry for the old address, and add a new entry reflecting the updated
address.
* The bhash2 table must have its own lock, even though concurrent
accesses on the same port are protected by the bhash lock. Bhash2 must
have its own lock to protect against cases where sockets on different
ports hash to different bhash hashbuckets but to the same bhash2
hashbucket.
This brings up a few stipulations:
1) When acquiring both the bhash and the bhash2 lock, the bhash2 lock
will always be acquired after the bhash lock and released before the
bhash lock is released.
2) There are no nested bhash2 hashbucket locks. A bhash2 lock is always
acquired+released before another bhash2 lock is acquired+released.
* The bhash table cannot be superseded by the bhash2 table because for
bind requests on INADDR_ANY (ipv4) or IPV6_ADDR_ANY (ipv6), every socket
bound to that port must be checked for a potential conflict. The bhash
table is the only source of port->socket associations.
Signed-off-by: Joanne Koong <joannelkoong@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Pablo Neira Ayuso says:
====================
Netfilter fixes for net
1) Fix crash with malformed ebtables blob which do not provide all
entry points, from Florian Westphal.
2) Fix possible TCP connection clogging up with default 5-days
timeout in conntrack, from Florian.
3) Fix crash in nf_tables tproxy with unsupported chains, also from Florian.
4) Do not allow to update implicit chains.
5) Make table handle allocation per-netns to fix data race.
6) Do not truncated payload length and offset, and checksum offset.
Instead report EINVAl.
7) Enable chain stats update via static key iff no error occurs.
8) Restrict osf expression to ip, ip6 and inet families.
9) Restrict tunnel expression to netdev family.
10) Fix crash when trying to bind again an already bound chain.
11) Flowtable garbage collector might leave behind pending work to
delete entries. This patch comes with a previous preparation patch
as dependency.
12) Allow net.netfilter.nf_conntrack_frag6_high_thresh to be lowered,
from Eric Dumazet.
* git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf:
netfilter: nf_defrag_ipv6: allow nf_conntrack_frag6_high_thresh increases
netfilter: flowtable: fix stuck flows on cleanup due to pending work
netfilter: flowtable: add function to invoke garbage collection immediately
netfilter: nf_tables: disallow binding to already bound chain
netfilter: nft_tunnel: restrict it to netdev family
netfilter: nft_osf: restrict osf to ipv4, ipv6 and inet families
netfilter: nf_tables: do not leave chain stats enabled on error
netfilter: nft_payload: do not truncate csum_offset and csum_type
netfilter: nft_payload: report ERANGE for too long offset and length
netfilter: nf_tables: make table handle allocation per-netns friendly
netfilter: nf_tables: disallow updates of implicit chain
netfilter: nft_tproxy: restrict to prerouting hook
netfilter: conntrack: work around exceeded receive window
netfilter: ebtables: reject blobs that don't provide all entry points
====================
Link: https://lore.kernel.org/r/20220824220330.64283-1-pablo@netfilter.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Commit c078290a2b ("selftests: include bonding tests into the kselftest
infra") adds the bonding tests in the directory:
tools/testing/selftests/drivers/net/bonding/
The file entry in MAINTAINERS for the BONDING DRIVER however refers to:
tools/testing/selftests/net/bonding/
Hence, ./scripts/get_maintainer.pl --self-test=patterns complains about a
broken file pattern.
Repair this file entry in BONDING DRIVER.
Signed-off-by: Lukas Bulwahn <lukas.bulwahn@gmail.com>
Acked-by: Jonathan Toppins <jtoppins@redhat.com>
Link: https://lore.kernel.org/r/20220824072945.28606-1-lukas.bulwahn@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
davinci_mdio.c uses mdio bitbang APIs, so it should select
MDIO_BITBANG to prevent build errors.
arm-linux-gnueabi-ld: drivers/net/ethernet/ti/davinci_mdio.o: in function `davinci_mdio_remove':
drivers/net/ethernet/ti/davinci_mdio.c:649: undefined reference to `free_mdio_bitbang'
arm-linux-gnueabi-ld: drivers/net/ethernet/ti/davinci_mdio.o: in function `davinci_mdio_probe':
drivers/net/ethernet/ti/davinci_mdio.c:545: undefined reference to `alloc_mdio_bitbang'
arm-linux-gnueabi-ld: drivers/net/ethernet/ti/davinci_mdio.o: in function `davinci_mdiobb_read':
drivers/net/ethernet/ti/davinci_mdio.c:236: undefined reference to `mdiobb_read'
arm-linux-gnueabi-ld: drivers/net/ethernet/ti/davinci_mdio.o: in function `davinci_mdiobb_write':
drivers/net/ethernet/ti/davinci_mdio.c:253: undefined reference to `mdiobb_write'
Fixes: d04807b806 ("net: ethernet: ti: davinci_mdio: Add workaround for errata i2329")
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Cc: Grygorii Strashko <grygorii.strashko@ti.com>
Cc: Ravi Gunasekaran <r-gunasekaran@ti.com>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Paolo Abeni <pabeni@redhat.com>
Cc: Naresh Kamboju <naresh.kamboju@linaro.org>
Cc: Sudip Mukherjee (Codethink) <sudipm.mukherjee@gmail.com>
Link: https://lore.kernel.org/r/20220824024216.4939-1-rdunlap@infradead.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Stephen Rothwell reported htmldocs warning when merging net-next tree:
Documentation/admin-guide/sysctl/net.rst:37: WARNING: Malformed table.
Text in column margin in table line 4.
========= =================== = ========== ==================
Directory Content Directory Content
========= =================== = ========== ==================
802 E802 protocol mptcp Multipath TCP
appletalk Appletalk protocol netfilter Network Filter
ax25 AX25 netrom NET/ROM
bridge Bridging rose X.25 PLP layer
core General parameter tipc TIPC
ethernet Ethernet protocol unix Unix domain sockets
ipv4 IP version 4 x25 X.25 protocol
ipv6 IP version 6
========= =================== = ========== ==================
The warning above is caused by cells in second "Content" column of
/proc/sys/net subdirectory table which are in column margin.
Align these cells against the column header to fix the warning.
Link: https://lore.kernel.org/linux-next/20220823134905.57ed08d5@canb.auug.org.au/
Fixes: 1202cdd665 ("Remove DECnet support from kernel")
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Bagas Sanjaya <bagasdotme@gmail.com>
Link: https://lore.kernel.org/r/20220824035804.204322-1-bagasdotme@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
It was not possible to create 1-tuple flow director
rule for IPv6 flow type. It was caused by incorrectly
checking for source IP address when validating user provided
destination IP address.
Fix this by changing ip6src to correct ip6dst address
in destination IP address validation for IPv6 flow type.
Fixes: efca91e89b ("i40e: Add flow director support for IPv6")
Signed-off-by: Sylwester Dziedziuch <sylwesterx.dziedziuch@intel.com>
Tested-by: Gurucharan <gurucharanx.g@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
The ixgbe_ptp_start_cyclecounter is intended to be called whenever the
cyclecounter parameters need to be changed.
Since commit a9763f3cb5 ("ixgbe: Update PTP to support X550EM_x
devices"), this function has cleared the SYSTIME registers and reset the
TSAUXC DISABLE_SYSTIME bit.
While these need to be cleared during ixgbe_ptp_reset, it is wrong to clear
them during ixgbe_ptp_start_cyclecounter. This function may be called
during both reset and link status change. When link changes, the SYSTIME
counter is still operating normally, but the cyclecounter should be updated
to account for the possibly changed parameters.
Clearing SYSTIME when link changes causes the timecounter to jump because
the cycle counter now reads zero.
Extract the SYSTIME initialization out to a new function and call this
during ixgbe_ptp_reset. This prevents the timecounter adjustment and avoids
an unnecessary reset of the current time.
This also restores the original SYSTIME clearing that occurred during
ixgbe_ptp_reset before the commit above.
Reported-by: Steve Payne <spayne@aurora.tech>
Reported-by: Ilya Evenbach <ievenbach@aurora.tech>
Fixes: a9763f3cb5 ("ixgbe: Update PTP to support X550EM_x devices")
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Gurucharan <gurucharanx.g@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
- Fix build warning for when MODULES and FTRACE_WITH_DIRECT_CALLS are not
set. A warning happens with ops_references_rec() defined but not used.
-----BEGIN PGP SIGNATURE-----
iIoEABYIADIWIQRRSw7ePDh/lE+zeZMp5XQQmuv6qgUCYwTkGhQccm9zdGVkdEBn
b29kbWlzLm9yZwAKCRAp5XQQmuv6qtUTAP4tOmf0I0c+GsWTzpecvv7fa+9rxmZa
SfBuoPqzC/TBqAEArqaf91+57aehCrJC3X5HaE7OJisW9nd2Epnvrpxk4QY=
=0yZV
-----END PGP SIGNATURE-----
Merge tag 'trace-v6.0-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace
Pull tracing fix from Steven Rostedt:
- Fix build warning for when MODULES and FTRACE_WITH_DIRECT_CALLS are
not set. A warning happens with ops_references_rec() defined but not
used.
* tag 'trace-v6.0-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace:
ftrace: Fix build warning for ops_references_rec() not used
Pull dmi update from Jean Delvare.
Tiny cleanup.
* 'dmi-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jdelvare/staging:
firmware: dmi: Use the proper accessor for the version field
Kuniyuki Iwashima says:
====================
net: sysctl: Fix data-races around net.core.XXX
This series fixes data-races around all knobs in net_core_table and
netns_core_table except for bpf stuff.
These knobs are skipped:
- 4 bpf knobs
- netdev_rss_key: Written only once by net_get_random_once() and
read-only knob
- rps_sock_flow_entries: Protected with sock_flow_mutex
- flow_limit_cpu_bitmap: Protected with flow_limit_update_mutex
- flow_limit_table_len: Protected with flow_limit_update_mutex
- default_qdisc: Protected with qdisc_mod_lock
- warnings: Unused
- high_order_alloc_disable: Protected with static_key_mutex
- skb_defer_max: Already using READ_ONCE()
- sysctl_txrehash: Already using READ_ONCE()
Note 5th patch fixes net.core.message_cost and net.core.message_burst,
and lib/ratelimit.c does not have an explicit maintainer.
Changes:
v3:
* Fix build failures of CONFIG_SYSCTL=n case in 13th & 14th patches
v2: https://lore.kernel.org/netdev/20220818035227.81567-1-kuniyu@amazon.com/
* Remove 4 bpf knobs and added 6 knobs
v1: https://lore.kernel.org/netdev/20220816052347.70042-1-kuniyu@amazon.com/
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
While reading sysctl_somaxconn, it can be changed concurrently.
Thus, we need to add READ_ONCE() to its reader.
Fixes: 1da177e4c3 ("Linux-2.6.12-rc2")
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
While reading netdev_unregister_timeout_secs, it can be changed
concurrently. Thus, we need to add READ_ONCE() to its reader.
Fixes: 5aa3afe107 ("net: make unregister netdev warning timeout configurable")
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Acked-by: Dmitry Vyukov <dvyukov@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
While reading gro_normal_batch, it can be changed concurrently.
Thus, we need to add READ_ONCE() to its reader.
Fixes: 323ebb61e3 ("net: use listified RX for handling GRO_NORMAL skbs")
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Acked-by: Edward Cree <ecree.xilinx@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
While reading sysctl_devconf_inherit_init_net, it can be changed
concurrently. Thus, we need to add READ_ONCE() to its readers.
Fixes: 856c395cfa ("net: introduce a knob to control whether to inherit devconf config")
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
While reading sysctl_fb_tunnels_only_for_init_net, it can be changed
concurrently. Thus, we need to add READ_ONCE() to its readers.
Fixes: 79134e6ce2 ("net: do not create fallback tunnels for non-default namespaces")
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
While reading netdev_budget_usecs, it can be changed concurrently.
Thus, we need to add READ_ONCE() to its reader.
Fixes: 7acf8a1e8a ("Replace 2 jiffies with sysctl netdev_budget_usecs to enable softirq tuning")
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
