As of now all transmit queues transmit packets out of same scheduler
queue hierarchy. Due to this PFC frames sent by peer are not handled
properly, either all transmit queues are backpressured or none.
To fix this when user enables PFC for a given priority map relavant
transmit queue to a different scheduler queue hierarcy, so that
backpressure is applied only to the traffic egressing out of that TXQ.
Signed-off-by: Suman Ghosh <sumang@marvell.com>
Link: https://lore.kernel.org/r/20220830120304.158060-1-sumang@marvell.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Currently, the change function can be called by two ways. The one way is
that qdisc_change() will call it. Before calling change function,
qdisc_change() ensures tca[TCA_OPTIONS] is not empty. The other way is
that .init() will call it. The opt parameter is also checked before
calling change function in .init(). Therefore, it's no need to check the
input parameter opt in change function.
Signed-off-by: Zhengchao Shao <shaozhengchao@huawei.com>
Acked-by: Toke Høiland-Jørgensen <toke@toke.dk>
Link: https://lore.kernel.org/r/20220829071219.208646-1-shaozhengchao@huawei.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Eric Dumazet says:
====================
tcp: tcp challenge ack fixes
syzbot found a typical data-race addressed in the first patch.
While we are at it, second patch makes the global rate limit
per net-ns and disabled by default.
====================
Link: https://lore.kernel.org/r/20220830185656.268523-1-eric.dumazet@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Because per host rate limiting has been proven problematic (side channel
attacks can be based on it), per host rate limiting of challenge acks ideally
should be per netns and turned off by default.
This is a long due followup of following commits:
083ae30828 ("tcp: enable per-socket rate limiting of all 'challenge acks'")
f2b2c582e8 ("tcp: mitigate ACK loops for connections as tcp_sock")
75ff39ccc1 ("tcp: make challenge acks less predictable")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Jason Baron <jbaron@akamai.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
challenge_timestamp can be read an written by concurrent threads.
This was expected, but we need to annotate the race to avoid potential issues.
Following patch moves challenge_timestamp and challenge_count
to per-netns storage to provide better isolation.
Fixes: 354e4aa391 ("tcp: RFC 5961 5.2 Blind Data Injection Attack Mitigation")
Reported-by: syzbot <syzkaller@googlegroups.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The IP_UNICAST_IF socket option is used to set the outgoing interface
for outbound packets.
The IP_UNICAST_IF socket option was added as it was needed by the
Wine project, since no other existing option (SO_BINDTODEVICE socket
option, IP_PKTINFO socket option or the bind function) provided the
needed characteristics needed by the IP_UNICAST_IF socket option. [1]
The IP_UNICAST_IF socket option works well for unconnected sockets,
that is, the interface specified by the IP_UNICAST_IF socket option
is taken into consideration in the route lookup process when a packet
is being sent. However, for connected sockets, the outbound interface
is chosen when connecting the socket, and in the route lookup process
which is done when a packet is being sent, the interface specified by
the IP_UNICAST_IF socket option is being ignored.
This inconsistent behavior was reported and discussed in an issue
opened on systemd's GitHub project [2]. Also, a bug report was
submitted in the kernel's bugzilla [3].
To understand the problem in more detail, we can look at what happens
for UDP packets over IPv4 (The same analysis was done separately in
the referenced systemd issue).
When a UDP packet is sent the udp_sendmsg function gets called and
the following happens:
1. The oif member of the struct ipcm_cookie ipc (which stores the
output interface of the packet) is initialized by the ipcm_init_sk
function to inet->sk.sk_bound_dev_if (the device set by the
SO_BINDTODEVICE socket option).
2. If the IP_PKTINFO socket option was set, the oif member gets
overridden by the call to the ip_cmsg_send function.
3. If no output interface was selected yet, the interface specified
by the IP_UNICAST_IF socket option is used.
4. If the socket is connected and no destination address is
specified in the send function, the struct ipcm_cookie ipc is not
taken into consideration and the cached route, that was calculated in
the connect function is being used.
Thus, for a connected socket, the IP_UNICAST_IF sockopt isn't taken
into consideration.
This patch corrects the behavior of the IP_UNICAST_IF socket option
for connect()ed sockets by taking into consideration the
IP_UNICAST_IF sockopt when connecting the socket.
In order to avoid reconnecting the socket, this option is still
ignored when applied on an already connected socket until connect()
is called again by the Richard Gobert.
Change the __ip4_datagram_connect function, which is called during
socket connection, to take into consideration the interface set by
the IP_UNICAST_IF socket option, in a similar way to what is done in
the udp_sendmsg function.
[1] https://lore.kernel.org/netdev/1328685717.4736.4.camel@edumazet-laptop/T/
[2] https://github.com/systemd/systemd/issues/11935#issuecomment-618691018
[3] https://bugzilla.kernel.org/show_bug.cgi?id=210255
Signed-off-by: Richard Gobert <richardbgobert@gmail.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Link: https://lore.kernel.org/r/20220829111554.GA1771@debian
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
__mkroute_input() uses fib_validate_source() to trigger an icmp redirect.
My understanding is that fib_validate_source() is used to know if the src
address and the gateway address are on the same link. For that,
fib_validate_source() returns 1 (same link) or 0 (not the same network).
__mkroute_input() is the only user of these positive values, all other
callers only look if the returned value is negative.
Since the below patch, fib_validate_source() didn't return anymore 1 when
both addresses are on the same network, because the route lookup returns
RT_SCOPE_LINK instead of RT_SCOPE_HOST. But this is, in fact, right.
Let's adapat the test to return 1 again when both addresses are on the same
link.
CC: stable@vger.kernel.org
Fixes: 747c143072 ("ip: fix dflt addr selection for connected nexthop")
Reported-by: kernel test robot <yujie.liu@intel.com>
Reported-by: Heng Qi <hengqi@linux.alibaba.com>
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Link: https://lore.kernel.org/r/20220829100121.3821-1-nicolas.dichtel@6wind.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The variable "other" in the struct red_stats is not used. Remove it.
Signed-off-by: Zhengchao Shao <shaozhengchao@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The variable "other" in the struct choke_sched_data is not used. Remove it.
Signed-off-by: Zhengchao Shao <shaozhengchao@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Functions that work on a pointer to virtual memory such as
virt_to_pfn() and users of that function such as
virt_to_page() are supposed to pass a pointer to virtual
memory, ideally a (void *) or other pointer. However since
many architectures implement virt_to_pfn() as a macro,
this function becomes polymorphic and accepts both a
(unsigned long) and a (void *).
If we instead implement a proper virt_to_pfn(void *addr)
function the following happens (occurred on arch/arm):
net/rds/message.c:357:56: warning: passing argument 1
of 'virt_to_pfn' makes pointer from integer without a
cast [-Wint-conversion]
Fix this with an explicit cast.
Cc: Santosh Shilimkar <santosh.shilimkar@oracle.com>
Cc: rds-devel@oss.oracle.com
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20220829132001.114858-1-linus.walleij@linaro.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
anon_vma->degree tracks the combined number of child anon_vmas and VMAs
that use the anon_vma as their ->anon_vma.
anon_vma_clone() then assumes that for any anon_vma attached to
src->anon_vma_chain other than src->anon_vma, it is impossible for it to
be a leaf node of the VMA tree, meaning that for such VMAs ->degree is
elevated by 1 because of a child anon_vma, meaning that if ->degree
equals 1 there are no VMAs that use the anon_vma as their ->anon_vma.
This assumption is wrong because the ->degree optimization leads to leaf
nodes being abandoned on anon_vma_clone() - an existing anon_vma is
reused and no new parent-child relationship is created. So it is
possible to reuse an anon_vma for one VMA while it is still tied to
another VMA.
This is an issue because is_mergeable_anon_vma() and its callers assume
that if two VMAs have the same ->anon_vma, the list of anon_vmas
attached to the VMAs is guaranteed to be the same. When this assumption
is violated, vma_merge() can merge pages into a VMA that is not attached
to the corresponding anon_vma, leading to dangling page->mapping
pointers that will be dereferenced during rmap walks.
Fix it by separately tracking the number of child anon_vmas and the
number of VMAs using the anon_vma as their ->anon_vma.
Fixes: 7a3ef208e6 ("mm: prevent endless growth of anon_vma hierarchy")
Cc: stable@kernel.org
Acked-by: Michal Hocko <mhocko@suse.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Jann Horn <jannh@google.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The rtl8152 driver does not disable multicasting when userspace asks
it to. For example:
$ ifconfig eth0 -multicast -allmulti
$ tcpdump -p -i eth0 # will still capture multicast frames
Fix by clearing the device multicast filter table when multicast and
allmulti are both unset.
Tested as follows:
- Set multicast on eth0 network interface
- verify that multicast packets are coming in:
$ tcpdump -p -i eth0
- Clear multicast and allmulti on eth0 network interface
- verify that no more multicast packets are coming in:
$ tcpdump -p -i eth0
Signed-off-by: Sven van Ashbrook <svenva@chromium.org>
Acked-by: Hayes Wang <hayeswang@realtek.com>
Link: https://lore.kernel.org/r/20220830045923.net-next.v1.1.I4fee0ac057083d4f848caf0fa3a9fd466fc374a0@changeid
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When the GSO splitting feature of sch_cake is enabled, GSO superpackets
will be broken up and the resulting segments enqueued in place of the
original skb. In this case, CAKE calls consume_skb() on the original skb,
but still returns NET_XMIT_SUCCESS. This can confuse parent qdiscs into
assuming the original skb still exists, when it really has been freed. Fix
this by adding the __NET_XMIT_STOLEN flag to the return value in this case.
Fixes: 0c850344d3 ("sch_cake: Conditionally split GSO segments")
Signed-off-by: Toke Høiland-Jørgensen <toke@toke.dk>
Reported-by: zdi-disclosures@trendmicro.com # ZDI-CAN-18231
Link: https://lore.kernel.org/r/20220831092103.442868-1-toke@toke.dk
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
One test demonstrates the reentrancy of hash map update on the same
bucket should fail, and another one shows concureently updates of
the same hash map bucket should succeed and not fail due to
the reentrancy checking for bucket lock.
There is no trampoline support on s390x, so move htab_update to
denylist.
Signed-off-by: Hou Tao <houtao1@huawei.com>
Link: https://lore.kernel.org/r/20220831042629.130006-4-houtao@huaweicloud.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
In __htab_map_lookup_and_delete_batch() if htab_lock_bucket() returns
-EBUSY, it will go to next bucket. Going to next bucket may not only
skip the elements in current bucket silently, but also incur
out-of-bound memory access or expose kernel memory to userspace if
current bucket_cnt is greater than bucket_size or zero.
Fixing it by stopping batch operation and returning -EBUSY when
htab_lock_bucket() fails, and the application can retry or skip the busy
batch as needed.
Fixes: 20b6cc34ea ("bpf: Avoid hashtab deadlock with map_locked")
Reported-by: Hao Sun <sunhao.th@gmail.com>
Signed-off-by: Hou Tao <houtao1@huawei.com>
Link: https://lore.kernel.org/r/20220831042629.130006-3-houtao@huaweicloud.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Per-cpu htab->map_locked is used to prohibit the concurrent accesses
from both NMI and non-NMI contexts. But since commit 74d862b682
("sched: Make migrate_disable/enable() independent of RT"),
migrate_disable() is also preemptible under CONFIG_PREEMPT case, so now
map_locked also disallows concurrent updates from normal contexts
(e.g. userspace processes) unexpectedly as shown below:
process A process B
htab_map_update_elem()
htab_lock_bucket()
migrate_disable()
/* return 1 */
__this_cpu_inc_return()
/* preempted by B */
htab_map_update_elem()
/* the same bucket as A */
htab_lock_bucket()
migrate_disable()
/* return 2, so lock fails */
__this_cpu_inc_return()
return -EBUSY
A fix that seems feasible is using in_nmi() in htab_lock_bucket() and
only checking the value of map_locked for nmi context. But it will
re-introduce dead-lock on bucket lock if htab_lock_bucket() is re-entered
through non-tracing program (e.g. fentry program).
One cannot use preempt_disable() to fix this issue as htab_use_raw_lock
being false causes the bucket lock to be a spin lock which can sleep and
does not work with preempt_disable().
Therefore, use migrate_disable() when using the spinlock instead of
preempt_disable() and defer fixing concurrent updates to when the kernel
has its own BPF memory allocator.
Fixes: 74d862b682 ("sched: Make migrate_disable/enable() independent of RT")
Reviewed-by: Hao Luo <haoluo@google.com>
Signed-off-by: Hou Tao <houtao1@huawei.com>
Link: https://lore.kernel.org/r/20220831042629.130006-2-houtao@huaweicloud.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
This patch adds a test to ensure bpf_setsockopt(TCP_CONGESTION, "not_exist")
will not trigger the kernel module autoload.
Before the fix:
[ 40.535829] BUG: sleeping function called from invalid context at include/linux/sched/mm.h:274
[...]
[ 40.552134] tcp_ca_find_autoload.constprop.0+0xcb/0x200
[ 40.552689] tcp_set_congestion_control+0x99/0x7b0
[ 40.553203] do_tcp_setsockopt+0x3ed/0x2240
[...]
[ 40.556041] __bpf_setsockopt+0x124/0x640
Signed-off-by: Martin KaFai Lau <martin.lau@linux.dev>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20220830231953.792412-1-martin.lau@linux.dev
This patch adds logic to compute the MDIO period based on
the i1clk, and thereafter write the MDIO period into the YU
MDIO config register. The i1clk resource from the ACPI table
is used to provide addressing to YU bootrecord PLL registers.
The values in these registers are used to compute MDIO period.
If the i1clk resource is not present in the ACPI table, then
the current default hardcorded value of 430Mhz is used.
The i1clk clock value of 430MHz is only accurate for boards
with BF2 mid bin and main bin SoCs. The BF2 high bin SoCs
have i1clk = 500MHz, but can support a slower MDIO period.
Fixes: f92e1869d7 ("Add Mellanox BlueField Gigabit Ethernet driver")
Reviewed-by: Asmaa Mnebhi <asmaa@nvidia.com>
Signed-off-by: David Thompson <davthompson@nvidia.com>
Link: https://lore.kernel.org/r/20220826155916.12491-1-davthompson@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The bpf_tail_call_static function is currently not defined unless
using clang >= 8.
To support bpf_tail_call_static on GCC we can check if __clang__ is
not defined to enable bpf_tail_call_static.
We need to use GCC assembly syntax when the compiler does not define
__clang__ as LLVM inline assembly is not fully compatible with GCC.
Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20220829210546.755377-1-james.hilliard1@gmail.com
Pull fscache/cachefiles fixes from David Howells:
- Fix kdoc on fscache_use/unuse_cookie().
- Fix the error returned by cachefiles_ondemand_copen() from an upcall
result.
- Fix the distribution of requests in on-demand mode in cachefiles to
be fairer by cycling through them rather than picking the one with
the lowest ID each time (IDs being reused).
* tag 'fscache-fixes-20220831' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs:
cachefiles: make on-demand request distribution fairer
cachefiles: fix error return code in cachefiles_ondemand_copen()
fscache: fix misdocumented parameter
Pull crypto fix from Herbert Xu:
"Fix a boot performance regression due to an unnecessary dependency on
XOR_BLOCKS"
* tag 'v6.0-p2' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6:
crypto: lib - remove unneeded selection of XOR_BLOCKS
Pull LSM support for IORING_OP_URING_CMD from Paul Moore:
"Add SELinux and Smack controls to the io_uring IORING_OP_URING_CMD.
These are necessary as without them the IORING_OP_URING_CMD remains
outside the purview of the LSMs (Luis' LSM patch, Casey's Smack patch,
and my SELinux patch). They have been discussed at length with the
io_uring folks, and Jens has given his thumbs-up on the relevant
patches (see the commit descriptions).
There is one patch that is not strictly necessary, but it makes
testing much easier and is very trivial: the /dev/null
IORING_OP_URING_CMD patch."
* tag 'lsm-pr-20220829' of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/lsm:
Smack: Provide read control for io_uring_cmd
/dev/null: add IORING_OP_URING_CMD support
selinux: implement the security_uring_cmd() LSM hook
lsm,io_uring: add LSM hooks for the new uring_cmd file op
Guangbin Huang says:
====================
net: hns3: updates for -next
This series includes some updates for the HNS3 ethernet driver.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
For some new devices, the FEC mode can not be set to OFF in speed 200G.
In order to flexibly adapt to all types of devices, driver queries
fec ability from firmware to decide whether OFF mode can be supported.
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch supports llrs fec mode in speed 200G for some new devices, and
suppoprts querying llrs fec ability from firmware.
Signed-off-by: Hao Lan <lanhao@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
For some new devices, driver can queries fec ability from firmware to
decide which FEC mode can be supported.
If devices of old version which not support querying fec ability, driver
sets fixed ability according to current speed.
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As some new devices may not support GRO offload and flow table director,
to support these devices, driver needs to querying capabilities of GRO
offload and flow table director from firmware. Whether the driver
supports these two features depends on capabilities.
For old device of version HNAE3_DEVICE_VERSION_V2, driver sets their
capabilities of these two features to fixed value.
Setting default features of netdev and debugfs also need to identify
whether support these two features.
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Mika Westerberg says:
====================
thunderbolt: net: Enable full end-to-end flow control
Thunderbolt/USB4 host controllers support full end-to-end flow control
that prevents dropping packets if there are not enough hardware receive
buffers. So far it has not been enabled for the networking driver yet
but this series changes that. There is one snag though: the second
generation (Intel Falcon Ridge) had a bug that needs special quirk to
get it working. We had that in the early stages of the Thunderbolt/USB4
driver but it got dropped because it was not needed at the time. Now we
add it back as a quirk for the host controller (NHI).
The first patch of this series is a bugfix that I'm planning to push for
v6.0-rc. Rest are v6.1 material. This also includes a patch that shows
the XDomain link type in sysfs the same way we do for USB4 routers and
updates the networking driver module description.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
It is Thunderbolt/USB4 now so reflect that in the module description too
to avoid any confusion. No functional changes.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
USB4NET protocol allows the networking drivers to take advantage of
end-to-end flow control supported by the USB4 host interface. This
should prevent the receiving side from dropping network packets.
In adddition add a module parameter that can be used to turn this off
just in case it causes problems.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As we are now enabling full end-to-end flow control to the Thunderbolt
networking driver, in order for it to work properly on second generation
Thunderbolt hardware (Falcon Ridge), we need to add back the workaround
that was removed with commit 53f13319d1 ("thunderbolt: Get rid of E2E
workaround"). However, this time we only apply it for Falcon Ridge
controllers as a form of an additional quirk. For non-Falcon Ridge this
does nothing.
While there fix a typo 'reqister' -> 'register' in the comment.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Following what we do for routers already, extend this to XDomain
connections as well. This will show in sysfs whether the link is in USB4
or Thunderbolt mode.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If the other host starts sending packets early on it is possible that we
are still in the middle of populating the initial Rx ring packets to the
ring. This causes the tbnet_poll() to mess over the queue and causes
list corruption. This happens specifically when connected with macOS as
it seems start sending various IP discovery packets as soon as its side
of the paths are configured.
To prevent this we move the DMA path enabling to happen after we have
primed the Rx ring. This makes sure no incoming packets can arrive
before we are ready to handle them.
Fixes: e69b6c02b4 ("net: Add support for networking over Thunderbolt cable")
Cc: stable@vger.kernel.org
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The function neigh_timer_handler() is a timer handler that runs in an
atomic context. When used by rocker, neigh_timer_handler() calls
"kzalloc(.., GFP_KERNEL)" that may sleep. As a result, the sleep in
atomic context bug will happen. One of the processes is shown below:
ofdpa_fib4_add()
...
neigh_add_timer()
(wait a timer)
neigh_timer_handler()
neigh_release()
neigh_destroy()
rocker_port_neigh_destroy()
rocker_world_port_neigh_destroy()
ofdpa_port_neigh_destroy()
ofdpa_port_ipv4_neigh()
kzalloc(sizeof(.., GFP_KERNEL) //may sleep
This patch changes the gfp_t parameter of kzalloc() from GFP_KERNEL to
GFP_ATOMIC in order to mitigate the bug.
Fixes: 00fc0c51e3 ("rocker: Change world_ops API and implementation to be switchdev independant")
Signed-off-by: Duoming Zhou <duoming@zju.edu.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
