| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Verifier enforces that only certain program types can mutate sock{map,hash}
maps, that is update it or delete from it. Add test coverage for these
checks so we don't regress.
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20240527-sockmap-verify-deletes-v1-3-944b372f2101@cloudflare.com
|
|
This reverts commit ff91059932401894e6c86341915615c5eb0eca48.
This check is no longer needed. BPF programs attached to tracepoints are
now rejected by the verifier when they attempt to delete from a
sockmap/sockhash maps.
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20240527-sockmap-verify-deletes-v1-2-944b372f2101@cloudflare.com
|
|
We have seen an influx of syzkaller reports where a BPF program attached to
a tracepoint triggers a locking rule violation by performing a map_delete
on a sockmap/sockhash.
We don't intend to support this artificial use scenario. Extend the
existing verifier allowed-program-type check for updating sockmap/sockhash
to also cover deleting from a map.
From now on only BPF programs which were previously allowed to update
sockmap/sockhash can delete from these map types.
Fixes: ff9105993240 ("bpf, sockmap: Prevent lock inversion deadlock in map delete elem")
Reported-by: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>
Reported-by: syzbot+ec941d6e24f633a59172@syzkaller.appspotmail.com
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Tested-by: syzbot+ec941d6e24f633a59172@syzkaller.appspotmail.com
Acked-by: John Fastabend <john.fastabend@gmail.com>
Closes: https://syzkaller.appspot.com/bug?extid=ec941d6e24f633a59172
Link: https://lore.kernel.org/bpf/20240527-sockmap-verify-deletes-v1-1-944b372f2101@cloudflare.com
|
|
LED Select (LED_SEL) bit in the LED General Purpose IO Configuration
register is used to determine the functionality of external LED pins
(Speed Indicator, Link and Activity Indicator, Full Duplex Link
Indicator). The default value for this bit is 0 when no EEPROM is
present. If a EEPROM is present, the default value is the value of the
LED Select bit in the Configuration Flags of the EEPROM. A USB Reset or
Lite Reset (LRST) will cause this bit to be restored to the image value
last loaded from EEPROM, or to be set to 0 if no EEPROM is present.
While configuring the dual purpose GPIO/LED pins to LED outputs in the
LED General Purpose IO Configuration register, the LED_SEL bit is changed
as 0 and resulting the configured value from the EEPROM is cleared. The
issue is fixed by using read-modify-write approach.
Fixes: f293501c61c5 ("smsc95xx: configure LED outputs")
Signed-off-by: Parthiban Veerasooran <Parthiban.Veerasooran@microchip.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: Woojung Huh <woojung.huh@microchip.com>
Link: https://lore.kernel.org/r/20240523085314.167650-1-Parthiban.Veerasooran@microchip.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
There are two type of classes. "Leaf classes" that are the
bottom of the class hierarchy. "Inner classes" that are neither
the root class nor leaf classes. QoS rules can only specify leaf
classes as targets for traffic.
Root
/ \
/ \
1 2
/\
/ \
4 5
classes 1,4 and 5 are leaf classes.
class 2 is a inner class.
When a leaf class made as inner, or vice versa, resources associated
with send queue (send queue buffers and transmit schedulers) are not
getting freed.
Fixes: 5e6808b4c68d ("octeontx2-pf: Add support for HTB offload")
Signed-off-by: Hariprasad Kelam <hkelam@marvell.com>
Link: https://lore.kernel.org/r/20240523073626.4114-1-hkelam@marvell.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
syzkaller reported data-race of sk->sk_hash in unix_autobind() [0],
and the same ones exist in unix_bind_bsd() and unix_bind_abstract().
The three bind() functions prefetch sk->sk_hash locklessly and
use it later after validating that unix_sk(sk)->addr is NULL under
unix_sk(sk)->bindlock.
The prefetched sk->sk_hash is the hash value of unbound socket set
in unix_create1() and does not change until bind() completes.
There could be a chance that sk->sk_hash changes after the lockless
read. However, in such a case, non-NULL unix_sk(sk)->addr is visible
under unix_sk(sk)->bindlock, and bind() returns -EINVAL without using
the prefetched value.
The KCSAN splat is false-positive, but let's silence it by reading
sk->sk_hash under unix_sk(sk)->bindlock.
[0]:
BUG: KCSAN: data-race in unix_autobind / unix_autobind
write to 0xffff888034a9fb88 of 4 bytes by task 4468 on cpu 0:
__unix_set_addr_hash net/unix/af_unix.c:331 [inline]
unix_autobind+0x47a/0x7d0 net/unix/af_unix.c:1185
unix_dgram_connect+0x7e3/0x890 net/unix/af_unix.c:1373
__sys_connect_file+0xd7/0xe0 net/socket.c:2048
__sys_connect+0x114/0x140 net/socket.c:2065
__do_sys_connect net/socket.c:2075 [inline]
__se_sys_connect net/socket.c:2072 [inline]
__x64_sys_connect+0x40/0x50 net/socket.c:2072
do_syscall_x64 arch/x86/entry/common.c:52 [inline]
do_syscall_64+0x4f/0x110 arch/x86/entry/common.c:83
entry_SYSCALL_64_after_hwframe+0x46/0x4e
read to 0xffff888034a9fb88 of 4 bytes by task 4465 on cpu 1:
unix_autobind+0x28/0x7d0 net/unix/af_unix.c:1134
unix_dgram_connect+0x7e3/0x890 net/unix/af_unix.c:1373
__sys_connect_file+0xd7/0xe0 net/socket.c:2048
__sys_connect+0x114/0x140 net/socket.c:2065
__do_sys_connect net/socket.c:2075 [inline]
__se_sys_connect net/socket.c:2072 [inline]
__x64_sys_connect+0x40/0x50 net/socket.c:2072
do_syscall_x64 arch/x86/entry/common.c:52 [inline]
do_syscall_64+0x4f/0x110 arch/x86/entry/common.c:83
entry_SYSCALL_64_after_hwframe+0x46/0x4e
value changed: 0x000000e4 -> 0x000001e3
Reported by Kernel Concurrency Sanitizer on:
CPU: 1 PID: 4465 Comm: syz-executor.0 Not tainted 6.8.0-12822-gcd51db110a7e #12
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.16.0-0-gd239552ce722-prebuilt.qemu.org 04/01/2014
Fixes: afd20b9290e1 ("af_unix: Replace the big lock with small locks.")
Reported-by: syzkaller <syzkaller@googlegroups.com>
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Link: https://lore.kernel.org/r/20240522154218.78088-1-kuniyu@amazon.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Once unix_sk(sk)->addr is assigned under net->unx.table.locks and
unix_sk(sk)->bindlock, *(unix_sk(sk)->addr) and unix_sk(sk)->path are
fully set up, and unix_sk(sk)->addr is never changed.
unix_getname() and unix_copy_addr() access the two fields locklessly,
and commit ae3b564179bf ("missing barriers in some of unix_sock ->addr
and ->path accesses") added smp_store_release() and smp_load_acquire()
pairs.
In other functions, we still read unix_sk(sk)->addr locklessly to check
if the socket is bound, and KCSAN complains about it. [0]
Given these functions have no dependency for *(unix_sk(sk)->addr) and
unix_sk(sk)->path, READ_ONCE() is enough to annotate the data-race.
Note that it is safe to access unix_sk(sk)->addr locklessly if the socket
is found in the hash table. For example, the lockless read of otheru->addr
in unix_stream_connect() is safe.
Note also that newu->addr there is of the child socket that is still not
accessible from userspace, and smp_store_release() publishes the address
in case the socket is accept()ed and unix_getname() / unix_copy_addr()
is called.
[0]:
BUG: KCSAN: data-race in unix_bind / unix_listen
write (marked) to 0xffff88805f8d1840 of 8 bytes by task 13723 on cpu 0:
__unix_set_addr_hash net/unix/af_unix.c:329 [inline]
unix_bind_bsd net/unix/af_unix.c:1241 [inline]
unix_bind+0x881/0x1000 net/unix/af_unix.c:1319
__sys_bind+0x194/0x1e0 net/socket.c:1847
__do_sys_bind net/socket.c:1858 [inline]
__se_sys_bind net/socket.c:1856 [inline]
__x64_sys_bind+0x40/0x50 net/socket.c:1856
do_syscall_x64 arch/x86/entry/common.c:52 [inline]
do_syscall_64+0x4f/0x110 arch/x86/entry/common.c:83
entry_SYSCALL_64_after_hwframe+0x46/0x4e
read to 0xffff88805f8d1840 of 8 bytes by task 13724 on cpu 1:
unix_listen+0x72/0x180 net/unix/af_unix.c:734
__sys_listen+0xdc/0x160 net/socket.c:1881
__do_sys_listen net/socket.c:1890 [inline]
__se_sys_listen net/socket.c:1888 [inline]
__x64_sys_listen+0x2e/0x40 net/socket.c:1888
do_syscall_x64 arch/x86/entry/common.c:52 [inline]
do_syscall_64+0x4f/0x110 arch/x86/entry/common.c:83
entry_SYSCALL_64_after_hwframe+0x46/0x4e
value changed: 0x0000000000000000 -> 0xffff88807b5b1b40
Reported by Kernel Concurrency Sanitizer on:
CPU: 1 PID: 13724 Comm: syz-executor.4 Not tainted 6.8.0-12822-gcd51db110a7e #12
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.16.0-0-gd239552ce722-prebuilt.qemu.org 04/01/2014
Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Reported-by: syzkaller <syzkaller@googlegroups.com>
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Link: https://lore.kernel.org/r/20240522154002.77857-1-kuniyu@amazon.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
The hsr_ping test reports the following errors:
INFO: preparing interfaces for HSRv0.
INFO: Initial validation ping.
INFO: Longer ping test.
INFO: Cutting one link.
INFO: Delay the link and drop a few packages.
INFO: All good.
INFO: preparing interfaces for HSRv1.
RTNETLINK answers: File exists
RTNETLINK answers: File exists
RTNETLINK answers: File exists
RTNETLINK answers: File exists
RTNETLINK answers: File exists
RTNETLINK answers: File exists
Error: ipv4: Address already assigned.
Error: ipv6: address already assigned.
Error: ipv4: Address already assigned.
Error: ipv6: address already assigned.
Error: ipv4: Address already assigned.
Error: ipv6: address already assigned.
INFO: Initial validation ping.
That is because the cleanup code for the 2nd round test before
"setup_hsr_interfaces 1" is removed incorrectly in commit 680fda4f6714
("test: hsr: Remove script code already implemented in lib.sh").
This patch fixes it by re-setup the namespaces using
setup_ns ns1 ns2 ns3
command before "setup_hsr_interfaces 1". It deletes previous namespaces
and create new ones.
Fixes: 680fda4f6714 ("test: hsr: Remove script code already implemented in lib.sh")
Reviewed-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
Link: https://lore.kernel.org/r/6485d3005f467758d49f0f313c8c009759ba6b05.1716374462.git.tanggeliang@kylinos.cn
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
enic_set_vf_port assumes that the nl attribute IFLA_PORT_PROFILE
is of length PORT_PROFILE_MAX and that the nl attributes
IFLA_PORT_INSTANCE_UUID, IFLA_PORT_HOST_UUID are of length PORT_UUID_MAX.
These attributes are validated (in the function do_setlink in rtnetlink.c)
using the nla_policy ifla_port_policy. The policy defines IFLA_PORT_PROFILE
as NLA_STRING, IFLA_PORT_INSTANCE_UUID as NLA_BINARY and
IFLA_PORT_HOST_UUID as NLA_STRING. That means that the length validation
using the policy is for the max size of the attributes and not on exact
size so the length of these attributes might be less than the sizes that
enic_set_vf_port expects. This might cause an out of bands
read access in the memcpys of the data of these
attributes in enic_set_vf_port.
Fixes: f8bd909183ac ("net: Add ndo_{set|get}_vf_port support for enic dynamic vnics")
Signed-off-by: Roded Zats <rzats@paloaltonetworks.com>
Link: https://lore.kernel.org/r/20240522073044.33519-1-rzats@paloaltonetworks.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Add a test case to assert that the skb->pkt_type which was set from the BPF
program is retained from the netkit xmit side to the peer's device at tcx
ingress location.
# ./vmtest.sh -- ./test_progs -t netkit
[...]
./test_progs -t netkit
[ 1.140780] bpf_testmod: loading out-of-tree module taints kernel.
[ 1.141127] bpf_testmod: module verification failed: signature and/or required key missing - tainting kernel
[ 1.284601] tsc: Refined TSC clocksource calibration: 3408.006 MHz
[ 1.286672] clocksource: tsc: mask: 0xffffffffffffffff max_cycles: 0x311fd9b189d, max_idle_ns: 440795225691 ns
[ 1.290384] clocksource: Switched to clocksource tsc
#345 tc_netkit_basic:OK
#346 tc_netkit_device:OK
#347 tc_netkit_multi_links:OK
#348 tc_netkit_multi_opts:OK
#349 tc_netkit_neigh_links:OK
#350 tc_netkit_pkt_type:OK
Summary: 6/0 PASSED, 0 SKIPPED, 0 FAILED
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/r/20240524163619.26001-4-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
This adds simple tests around setting MAC addresses in the different
netkit modes.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/r/20240524163619.26001-3-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
When running Cilium connectivity test suite with netkit in L2 mode, we
found that compared to tcx a few tests were failing which pushed traffic
into an L7 proxy sitting in host namespace. The problem in particular is
around the invocation of eth_type_trans() in netkit.
In case of tcx, this is run before the tcx ingress is triggered inside
host namespace and thus if the BPF program uses the bpf_skb_change_type()
helper the newly set type is retained. However, in case of netkit, the
late eth_type_trans() invocation overrides the earlier decision from the
BPF program which eventually leads to the test failure.
Instead of eth_type_trans(), split out the relevant parts, meaning, reset
of mac header and call to eth_skb_pkt_type() before the BPF program is run
in order to have the same behavior as with tcx, and refactor a small helper
called eth_skb_pull_mac() which is run in case it's passed up the stack
where the mac header must be pulled. With this all connectivity tests pass.
Fixes: 35dfaad7188c ("netkit, bpf: Add bpf programmable net device")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Nikolay Aleksandrov <razor@blackwall.org>
Link: https://lore.kernel.org/r/20240524163619.26001-2-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
When running Cilium connectivity test suite with netkit in L2 mode, we
found that it is expected to be able to specify a custom MAC address for
the devices, in particular, cilium-cni obtains the specified MAC address
by querying the endpoint and sets the MAC address of the interface inside
the Pod. Thus, fix the missing support in netkit for L2 mode.
Fixes: 35dfaad7188c ("netkit, bpf: Add bpf programmable net device")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Nikolay Aleksandrov <razor@blackwall.org>
Acked-by: Stanislav Fomichev <sdf@google.com>
Link: https://lore.kernel.org/r/20240524163619.26001-1-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
Also updated couple of comments along the way.
One of the issues reported was indeed a bug in the code:
memset(ctx, 0, sizeof(ctx)) // original line
memset(ctx, 0, sizeof(*ctx)) // fixed line
That was a nice catch.
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202405222314.UG5F2NHn-lkp@intel.com/
Closes: https://lore.kernel.org/oe-kbuild-all/202405232036.Xqoc3b0J-lkp@intel.com/
Signed-off-by: Shahab Vahedi <shahab@synopsys.com>
Link: https://lore.kernel.org/r/20240525035628.1026-1-list+bpf@vahedi.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
Validate libbpf's USDT-over-multi-uprobe logic by adding USDTs to
existing multi-uprobe tests. This checks correct libbpf fallback to
singular uprobes (when run on older kernels with buggy PID filtering).
We reuse already established child process and child thread testing
infrastructure, so additions are minimal. These test fail on either
older kernels or older version of libbpf that doesn't detect PID
filtering problems.
Acked-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/r/20240521163401.3005045-6-andrii@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
Extend existing multi-uprobe tests to test that PID filtering works
correctly. We already have child *process* tests, but we need also child
*thread* tests. This patch adds spawn_thread() helper to start child
thread, wait for it to be ready, and then instruct it to trigger desired
uprobes.
Additionally, we extend BPF-side code to track thread ID, not just
process ID. Also we detect whether extraneous triggerings with
unexpected process IDs happened, and validate that none of that happened
in practice.
These changes prove that fixed PID filtering logic for multi-uprobe
works as expected. These tests fail on old kernels.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Link: https://lore.kernel.org/r/20240521163401.3005045-5-andrii@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
Libbpf is automatically (and transparently to user) detecting
multi-uprobe support in the kernel, and, if supported, uses
multi-uprobes to improve USDT attachment speed.
USDTs can be attached system-wide or for the specific process by PID. In
the latter case, we rely on correct kernel logic of not triggering USDT
for unrelated processes.
As such, on older kernels that do support multi-uprobes, but still have
broken PID filtering logic, we need to fall back to singular uprobes.
Unfortunately, whether user is using PID filtering or not is known at
the attachment time, which happens after relevant BPF programs were
loaded into the kernel. Also unfortunately, we need to make a call
whether to use multi-uprobes or singular uprobe for SEC("usdt") programs
during BPF object load time, at which point we have no information about
possible PID filtering.
The distinction between single and multi-uprobes is small, but important
for the kernel. Multi-uprobes get BPF_TRACE_UPROBE_MULTI attach type,
and kernel internally substitiute different implementation of some of
BPF helpers (e.g., bpf_get_attach_cookie()) depending on whether uprobe
is multi or singular. So, multi-uprobes and singular uprobes cannot be
intermixed.
All the above implies that we have to make an early and conservative
call about the use of multi-uprobes. And so this patch modifies libbpf's
existing feature detector for multi-uprobe support to also check correct
PID filtering. If PID filtering is not yet fixed, we fall back to
singular uprobes for USDTs.
This extension to feature detection is simple thanks to kernel's -EINVAL
addition for pid < 0.
Acked-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/r/20240521163401.3005045-4-andrii@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
get_pid_task() internally already calls rcu_read_lock() and
rcu_read_unlock(), so there is no point to do this one extra time.
This is a drive-by improvement and has no correctness implications.
Acked-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/r/20240521163401.3005045-3-andrii@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
Current implementation of PID filtering logic for multi-uprobes in
uprobe_prog_run() is filtering down to exact *thread*, while the intent
for PID filtering it to filter by *process* instead. The check in
uprobe_prog_run() also differs from the analogous one in
uprobe_multi_link_filter() for some reason. The latter is correct,
checking task->mm, not the task itself.
Fix the check in uprobe_prog_run() to perform the same task->mm check.
While doing this, we also update get_pid_task() use to use PIDTYPE_TGID
type of lookup, given the intent is to get a representative task of an
entire process. This doesn't change behavior, but seems more logical. It
would hold task group leader task now, not any random thread task.
Last but not least, given multi-uprobe support is half-broken due to
this PID filtering logic (depending on whether PID filtering is
important or not), we need to make it easy for user space consumers
(including libbpf) to easily detect whether PID filtering logic was
already fixed.
We do it here by adding an early check on passed pid parameter. If it's
negative (and so has no chance of being a valid PID), we return -EINVAL.
Previous behavior would eventually return -ESRCH ("No process found"),
given there can't be any process with negative PID. This subtle change
won't make any practical change in behavior, but will allow applications
to detect PID filtering fixes easily. Libbpf fixes take advantage of
this in the next patch.
Cc: stable@vger.kernel.org
Acked-by: Jiri Olsa <jolsa@kernel.org>
Fixes: b733eeade420 ("bpf: Add pid filter support for uprobe_multi link")
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/r/20240521163401.3005045-2-andrii@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
err is a 32-bit integer, but elf_update returns an off_t, which is 64-bit
at least on 64-bit platforms. If symbols_patch is called on a binary between
2-4GB in size, the result will be negative when cast to a 32-bit integer,
which the code assumes means an error occurred. This can wrongly trigger
build failures when building very large kernel images.
Fixes: fbbb68de80a4 ("bpf: Add resolve_btfids tool to resolve BTF IDs in ELF object")
Signed-off-by: Friedrich Vock <friedrich.vock@gmx.de>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20240514070931.199694-1-friedrich.vock@gmx.de
|
|
When the skb is encapsulated, adjust the inner UDP header instead of the
outer one, and account for UDP header (instead of TCP) in the inline
header size calculation.
Fixes: 689adf0d4892 ("net/mlx5e: Add UDP GSO support")
Reported-by: Jason Baron <jbaron@akamai.com>
Closes: https://lore.kernel.org/netdev/c42961cb-50b9-4a9a-bd43-87fe48d88d29@akamai.com/
Signed-off-by: Gal Pressman <gal@nvidia.com>
Reviewed-by: Dragos Tatulea <dtatulea@nvidia.com>
Reviewed-by: Boris Pismenny <borisp@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Previously, the driver incorrectly used rx_dropped to report device
buffer exhaustion.
According to the documentation, rx_dropped should not be used to count
packets dropped due to buffer exhaustion, which is the purpose of
rx_missed_errors.
Use rx_missed_errors as intended for counting packets dropped due to
buffer exhaustion.
Fixes: 269e6b3af3bf ("net/mlx5e: Report additional error statistics in get stats ndo")
Signed-off-by: Carolina Jubran <cjubran@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
The ptp channel instance is only initialized when ptp traffic is first
processed by the driver. This means that there is a window in between when
port timestamping is enabled and ptp traffic is sent where the ptp channel
instance is not initialized. Accessing statistics during this window will
lead to an access violation (NULL + member offset). Check the validity of
the instance before attempting to query statistics.
BUG: unable to handle page fault for address: 0000000000003524
#PF: supervisor read access in kernel mode
#PF: error_code(0x0000) - not-present page
PGD 109dfc067 P4D 109dfc067 PUD 1064ef067 PMD 0
Oops: 0000 [#1] SMP
CPU: 0 PID: 420 Comm: ethtool Not tainted 6.9.0-rc2-rrameshbabu+ #245
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS Arch Linux 1.16.3-1-1 04/01/204
RIP: 0010:mlx5e_stats_ts_get+0x4c/0x130
<snip>
Call Trace:
<TASK>
? show_regs+0x60/0x70
? __die+0x24/0x70
? page_fault_oops+0x15f/0x430
? do_user_addr_fault+0x2c9/0x5c0
? exc_page_fault+0x63/0x110
? asm_exc_page_fault+0x27/0x30
? mlx5e_stats_ts_get+0x4c/0x130
? mlx5e_stats_ts_get+0x20/0x130
mlx5e_get_ts_stats+0x15/0x20
<snip>
Fixes: 3579032c08c1 ("net/mlx5e: Implement ethtool hardware timestamping statistics")
Signed-off-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Remove faulty check disabling checksum offload and GSO for offload of
simple IPsec tunnel L4 traffic. Comment previously describing the deleted
code incorrectly claimed the check prevented double tunnel (or three layers
of ip headers).
Fixes: f1267798c980 ("net/mlx5: Fix checksum issue of VXLAN and IPsec crypto offload")
Signed-off-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
rx_create no longer allocates a modify_hdr instance that needs to be
cleaned up. The mlx5_modify_header_dealloc call will lead to a NULL pointer
dereference. A leak in the rules also previously occurred since there are
now two rules populated related to status.
BUG: kernel NULL pointer dereference, address: 0000000000000000
#PF: supervisor read access in kernel mode
#PF: error_code(0x0000) - not-present page
PGD 109907067 P4D 109907067 PUD 116890067 PMD 0
Oops: 0000 [#1] SMP
CPU: 1 PID: 484 Comm: ip Not tainted 6.9.0-rc2-rrameshbabu+ #254
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS Arch Linux 1.16.3-1-1 04/01/2014
RIP: 0010:mlx5_modify_header_dealloc+0xd/0x70
<snip>
Call Trace:
<TASK>
? show_regs+0x60/0x70
? __die+0x24/0x70
? page_fault_oops+0x15f/0x430
? free_to_partial_list.constprop.0+0x79/0x150
? do_user_addr_fault+0x2c9/0x5c0
? exc_page_fault+0x63/0x110
? asm_exc_page_fault+0x27/0x30
? mlx5_modify_header_dealloc+0xd/0x70
rx_create+0x374/0x590
rx_add_rule+0x3ad/0x500
? rx_add_rule+0x3ad/0x500
? mlx5_cmd_exec+0x2c/0x40
? mlx5_create_ipsec_obj+0xd6/0x200
mlx5e_accel_ipsec_fs_add_rule+0x31/0xf0
mlx5e_xfrm_add_state+0x426/0xc00
<snip>
Fixes: 94af50c0a9bb ("net/mlx5e: Unify esw and normal IPsec status table creation/destruction")
Signed-off-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
The MTMP register (0x900a) capability offset is off-by-one, move it to
the right place.
Fixes: 1f507e80c700 ("net/mlx5: Expose NIC temperature via hardware monitoring kernel API")
Signed-off-by: Gal Pressman <gal@nvidia.com>
Reviewed-by: Cosmin Ratiu <cratiu@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
A proper query to MPIR needs to set the correct value in the depth field.
On embedded CPU this value is not necessarily zero. As there is no real
use case for multi-PF netdev on the embedded CPU of the smart NIC, block
this option.
This fixes the following failure:
ACCESS_REG(0x805) op_mod(0x1) failed, status bad system state(0x4), syndrome (0x685f19), err(-5)
Fixes: 678eb448055a ("net/mlx5: SD, Implement basic query and instantiation")
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Currently, the driver does not enforce that lag bond slaves must have
matching roce capabilities. Yet, in mlx5_do_bond(), the driver attempts
to enable roce on all vports of the bond slaves, causing the following
syndrome when one slave has no roce fw support:
mlx5_cmd_out_err:809:(pid 25427): MODIFY_NIC_VPORT_CONTEXT(0×755) op_mod(0×0)
failed, status bad parameter(0×3), syndrome (0xc1f678), err(-22)
Thus, create HW lag only if bond's slaves agree on roce state,
either all slaves have roce support resulting in a roce lag bond,
or none do, resulting in a raw eth bond.
Fixes: 7907f23adc18 ("net/mlx5: Implement RoCE LAG feature")
Signed-off-by: Maher Sanalla <msanalla@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Following a similar reinstate for the KSZ8081 and KSZ9031.
Older kernels would use the genphy_soft_reset if the PHY did not implement
a .soft_reset.
The KSZ8061 errata described here:
https://ww1.microchip.com/downloads/en/DeviceDoc/KSZ8061-Errata-DS80000688B.pdf
and worked around with 232ba3a51c ("net: phy: Micrel KSZ8061: link failure after cable connect")
is back again without this soft reset.
Fixes: 6e2d85ec0559 ("net: phy: Stop with excessive soft reset")
Tested-by: Karim Ben Houcine <karim.benhoucine@landisgyr.com>
Signed-off-by: Mathieu Othacehe <othacehe@gnu.org>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
The implicit conversion from unsigned int to enum
proc_cn_event is invalid, so explicitly cast it
for compilation in a C++ compiler.
/usr/include/linux/cn_proc.h: In function 'proc_cn_event valid_event(proc_cn_event)':
/usr/include/linux/cn_proc.h:72:17: error: invalid conversion from 'unsigned int' to 'proc_cn_event' [-fpermissive]
72 | ev_type &= PROC_EVENT_ALL;
| ^
| |
| unsigned int
Signed-off-by: Matt Jan <zoo868e@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Fix the following -Wformat-security compile warnings adding missing
format arguments:
latency-collector.c: In function ‘show_available’:
latency-collector.c:938:17: warning: format not a string literal and
no format arguments [-Wformat-security]
938 | warnx(no_tracer_msg);
| ^~~~~
latency-collector.c:943:17: warning: format not a string literal and
no format arguments [-Wformat-security]
943 | warnx(no_latency_tr_msg);
| ^~~~~
latency-collector.c: In function ‘find_default_tracer’:
latency-collector.c:986:25: warning: format not a string literal and
no format arguments [-Wformat-security]
986 | errx(EXIT_FAILURE, no_tracer_msg);
|
^~~~
latency-collector.c: In function ‘scan_arguments’:
latency-collector.c:1881:33: warning: format not a string literal and
no format arguments [-Wformat-security]
1881 | errx(EXIT_FAILURE, no_tracer_msg);
| ^~~~
Link: https://lore.kernel.org/linux-trace-kernel/20240404011009.32945-1-skhan@linuxfoundation.org
Cc: stable@vger.kernel.org
Fixes: e23db805da2df ("tracing/tools: Add the latency-collector to tools directory")
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
|
|
An issue was found on the RTL8125b when transmitting small fragmented
packets, whereby invalid entries were inserted into the transmit ring
buffer, subsequently leading to calls to dma_unmap_single() with a null
address.
This was caused by rtl8169_start_xmit() not noticing changes to nr_frags
which may occur when small packets are padded (to work around hardware
quirks) in rtl8169_tso_csum_v2().
To fix this, postpone inspecting nr_frags until after any padding has been
applied.
Fixes: 9020845fb5d6 ("r8169: improve rtl8169_start_xmit")
Cc: stable@vger.kernel.org
Signed-off-by: Ken Milmore <ken.milmore@gmail.com>
Reviewed-by: Heiner Kallweit <hkallweit1@gmail.com>
Link: https://lore.kernel.org/r/27ead18b-c23d-4f49-a020-1fc482c5ac95@gmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Unlike ice, idpf does not check, if user has requested at least 1 combined
channel. Instead, it relies on a check in the core code. Unfortunately, the
check does not trigger for us because of the hacky .set_channels()
interpretation logic that is not consistent with the core code.
This naturally leads to user being able to trigger a crash with an invalid
input. This is how:
1. ethtool -l <IFNAME> -> combined: 40
2. ethtool -L <IFNAME> rx 0 tx 0
combined number is not specified, so command becomes {rx_count = 0,
tx_count = 0, combined_count = 40}.
3. ethnl_set_channels checks, if there is at least 1 RX and 1 TX channel,
comparing (combined_count + rx_count) and (combined_count + tx_count)
to zero. Obviously, (40 + 0) is greater than zero, so the core code
deems the input OK.
4. idpf interprets `rx 0 tx 0` as 0 channels and tries to proceed with such
configuration.
The issue has to be solved fundamentally, as current logic is also known to
cause AF_XDP problems in ice [0].
Interpret the command in a way that is more consistent with ethtool
manual [1] (--show-channels and --set-channels) and new ice logic.
Considering that in the idpf driver only the difference between RX and TX
queues forms dedicated channels, change the correct way to set number of
channels to:
ethtool -L <IFNAME> combined 10 /* For symmetric queues */
ethtool -L <IFNAME> combined 8 tx 2 rx 0 /* For asymmetric queues */
[0] https://lore.kernel.org/netdev/20240418095857.2827-1-larysa.zaremba@intel.com/
[1] https://man7.org/linux/man-pages/man8/ethtool.8.html
Fixes: 02cbfba1add5 ("idpf: add ethtool callbacks")
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Reviewed-by: Igor Bagnucki <igor.bagnucki@intel.com>
Signed-off-by: Larysa Zaremba <larysa.zaremba@intel.com>
Tested-by: Krishneil Singh <krishneil.k.singh@intel.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
A bug occurs because a safety check guarding AF_XDP-related queues in
ethnl_set_channels(), does not trigger. This happens, because kernel and
ice driver interpret the ethtool command differently.
How the bug occurs:
1. ethtool -l <IFNAME> -> combined: 40
2. Attach AF_XDP to queue 30
3. ethtool -L <IFNAME> rx 15 tx 15
combined number is not specified, so command becomes {rx_count = 15,
tx_count = 15, combined_count = 40}.
4. ethnl_set_channels checks, if there are any AF_XDP of queues from the
new (combined_count + rx_count) to the old one, so from 55 to 40, check
does not trigger.
5. ice interprets `rx 15 tx 15` as 15 combined channels and deletes the
queue that AF_XDP is attached to.
Interpret the command in a way that is more consistent with ethtool
manual [0] (--show-channels and --set-channels).
Considering that in the ice driver only the difference between RX and TX
queues forms dedicated channels, change the correct way to set number of
channels to:
ethtool -L <IFNAME> combined 10 /* For symmetric queues */
ethtool -L <IFNAME> combined 8 tx 2 rx 0 /* For asymmetric queues */
[0] https://man7.org/linux/man-pages/man8/ethtool.8.html
Fixes: 87324e747fde ("ice: Implement ethtool ops for channels")
Reviewed-by: Michal Swiatkowski <michal.swiatkowski@linux.intel.com>
Signed-off-by: Larysa Zaremba <larysa.zaremba@intel.com>
Tested-by: Chandan Kumar Rout <chandanx.rout@intel.com>
Tested-by: Pucha Himasekhar Reddy <himasekharx.reddy.pucha@intel.com>
Acked-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
When nci_rx_work() receives a zero-length payload packet, it should not
discard the packet and exit the loop. Instead, it should continue
processing subsequent packets.
Fixes: d24b03535e5e ("nfc: nci: Fix uninit-value in nci_dev_up and nci_ntf_packet")
Signed-off-by: Ryosuke Yasuoka <ryasuoka@redhat.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20240521153444.535399-1-ryasuoka@redhat.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Christoph reported the following splat:
WARNING: CPU: 1 PID: 772 at net/ipv4/af_inet.c:761 __inet_accept+0x1f4/0x4a0
Modules linked in:
CPU: 1 PID: 772 Comm: syz-executor510 Not tainted 6.9.0-rc7-g7da7119fe22b #56
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.11.0-2.el7 04/01/2014
RIP: 0010:__inet_accept+0x1f4/0x4a0 net/ipv4/af_inet.c:759
Code: 04 38 84 c0 0f 85 87 00 00 00 41 c7 04 24 03 00 00 00 48 83 c4 10 5b 41 5c 41 5d 41 5e 41 5f 5d c3 cc cc cc cc e8 ec b7 da fd <0f> 0b e9 7f fe ff ff e8 e0 b7 da fd 0f 0b e9 fe fe ff ff 89 d9 80
RSP: 0018:ffffc90000c2fc58 EFLAGS: 00010293
RAX: ffffffff836bdd14 RBX: 0000000000000000 RCX: ffff888104668000
RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000000
RBP: dffffc0000000000 R08: ffffffff836bdb89 R09: fffff52000185f64
R10: dffffc0000000000 R11: fffff52000185f64 R12: dffffc0000000000
R13: 1ffff92000185f98 R14: ffff88810754d880 R15: ffff8881007b7800
FS: 000000001c772880(0000) GS:ffff88811b280000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00007fb9fcf2e178 CR3: 00000001045d2002 CR4: 0000000000770ef0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
PKRU: 55555554
Call Trace:
<TASK>
inet_accept+0x138/0x1d0 net/ipv4/af_inet.c:786
do_accept+0x435/0x620 net/socket.c:1929
__sys_accept4_file net/socket.c:1969 [inline]
__sys_accept4+0x9b/0x110 net/socket.c:1999
__do_sys_accept net/socket.c:2016 [inline]
__se_sys_accept net/socket.c:2013 [inline]
__x64_sys_accept+0x7d/0x90 net/socket.c:2013
do_syscall_x64 arch/x86/entry/common.c:52 [inline]
do_syscall_64+0x58/0x100 arch/x86/entry/common.c:83
entry_SYSCALL_64_after_hwframe+0x76/0x7e
RIP: 0033:0x4315f9
Code: fd ff 48 81 c4 80 00 00 00 e9 f1 fe ff ff 0f 1f 00 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 0f 83 ab b4 fd ff c3 66 2e 0f 1f 84 00 00 00 00
RSP: 002b:00007ffdb26d9c78 EFLAGS: 00000246 ORIG_RAX: 000000000000002b
RAX: ffffffffffffffda RBX: 0000000000400300 RCX: 00000000004315f9
RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000004
RBP: 00000000006e1018 R08: 0000000000400300 R09: 0000000000400300
R10: 0000000000400300 R11: 0000000000000246 R12: 0000000000000000
R13: 000000000040cdf0 R14: 000000000040ce80 R15: 0000000000000055
</TASK>
The reproducer invokes shutdown() before entering the listener status.
After commit 94062790aedb ("tcp: defer shutdown(SEND_SHUTDOWN) for
TCP_SYN_RECV sockets"), the above causes the child to reach the accept
syscall in FIN_WAIT1 status.
Eric noted we can relax the existing assertion in __inet_accept()
Reported-by: Christoph Paasch <cpaasch@apple.com>
Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/490
Suggested-by: Eric Dumazet <edumazet@google.com>
Fixes: 94062790aedb ("tcp: defer shutdown(SEND_SHUTDOWN) for TCP_SYN_RECV sockets")
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/23ab880a44d8cfd967e84de8b93dbf48848e3d8c.1716299669.git.pabeni@redhat.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Recently, we had some servers upgraded to the latest kernel and noticed
the indicator from the user side showed worse results than before. It is
caused by the limitation of tp->rcv_wnd.
In 2018 commit a337531b942b ("tcp: up initial rmem to 128KB and SYN rwin
to around 64KB") limited the initial value of tp->rcv_wnd to 65535, most
CDN teams would not benefit from this change because they cannot have a
large window to receive a big packet, which will be slowed down especially
in long RTT. Small rcv_wnd means slow transfer speed, to some extent. It's
the side effect for the latency/time-sensitive users.
To avoid future confusion, current change doesn't affect the initial
receive window on the wire in a SYN or SYN+ACK packet which are set within
65535 bytes according to RFC 7323 also due to the limit in
__tcp_transmit_skb():
th->window = htons(min(tp->rcv_wnd, 65535U));
In one word, __tcp_transmit_skb() already ensures that constraint is
respected, no matter how large tp->rcv_wnd is. The change doesn't violate
RFC.
Let me provide one example if with or without the patch:
Before:
client --- SYN: rwindow=65535 ---> server
client <--- SYN+ACK: rwindow=65535 ---- server
client --- ACK: rwindow=65536 ---> server
Note: for the last ACK, the calculation is 512 << 7.
After:
client --- SYN: rwindow=65535 ---> server
client <--- SYN+ACK: rwindow=65535 ---- server
client --- ACK: rwindow=175232 ---> server
Note: I use the following command to make it work:
ip route change default via [ip] dev eth0 metric 100 initrwnd 120
For the last ACK, the calculation is 1369 << 7.
When we apply such a patch, having a large rcv_wnd if the user tweak this
knob can help transfer data more rapidly and save some rtts.
Fixes: a337531b942b ("tcp: up initial rmem to 128KB and SYN rwin to around 64KB")
Signed-off-by: Jason Xing <kernelxing@tencent.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Link: https://lore.kernel.org/r/20240521134220.12510-1-kerneljasonxing@gmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
In the prueth_probe() function, if one of the calls to emac_phy_connect()
fails due to of_phy_connect() returning NULL, then the subsequent call to
phy_attached_info() will dereference a NULL pointer.
Check the return code of emac_phy_connect and fail cleanly if there is an
error.
Fixes: 128d5874c082 ("net: ti: icssg-prueth: Add ICSSG ethernet driver")
Cc: stable@vger.kernel.org
Signed-off-by: Romain Gantois <romain.gantois@bootlin.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: MD Danish Anwar <danishanwar@ti.com>
Link: https://lore.kernel.org/r/20240521-icssg-prueth-fix-v1-1-b4b17b1433e9@bootlin.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
In tls_init(), a write memory barrier is missing, and store-store
reordering may cause NULL dereference in tls_{setsockopt,getsockopt}.
CPU0 CPU1
----- -----
// In tls_init()
// In tls_ctx_create()
ctx = kzalloc()
ctx->sk_proto = READ_ONCE(sk->sk_prot) -(1)
// In update_sk_prot()
WRITE_ONCE(sk->sk_prot, tls_prots) -(2)
// In sock_common_setsockopt()
READ_ONCE(sk->sk_prot)->setsockopt()
// In tls_{setsockopt,getsockopt}()
ctx->sk_proto->setsockopt() -(3)
In the above scenario, when (1) and (2) are reordered, (3) can observe
the NULL value of ctx->sk_proto, causing NULL dereference.
To fix it, we rely on rcu_assign_pointer() which implies the release
barrier semantic. By moving rcu_assign_pointer() after ctx->sk_proto is
initialized, we can ensure that ctx->sk_proto are visible when
changing sk->sk_prot.
Fixes: d5bee7374b68 ("net/tls: Annotate access to sk_prot with READ_ONCE/WRITE_ONCE")
Signed-off-by: Yewon Choi <woni9911@gmail.com>
Signed-off-by: Dae R. Jeong <threeearcat@gmail.com>
Link: https://lore.kernel.org/netdev/ZU4OJG56g2V9z_H7@dragonet/T/
Link: https://lore.kernel.org/r/Zkx4vjSFp0mfpjQ2@libra05
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
The assignment of pps_enable is protected by tmreg_lock, but the read
operation of pps_enable is not. So the Coverity tool reports a lock
evasion warning which may cause data race to occur when running in a
multithread environment. Although this issue is almost impossible to
occur, we'd better fix it, at least it seems more logically reasonable,
and it also prevents Coverity from continuing to issue warnings.
Fixes: 278d24047891 ("net: fec: ptp: Enable PPS output based on ptp clock")
Signed-off-by: Wei Fang <wei.fang@nxp.com>
Link: https://lore.kernel.org/r/20240521023800.17102-1-wei.fang@nxp.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
This reverts commit 565736048bd5f9888990569993c6b6bfdf6dcb6d.
According to the commit, it implements a manual AN-37 for some
"troublesome" Juniper MX5 switches. This appears to be a workaround for a
particular switch.
It has been reported that this causes a severe breakage for other switches,
including a Cisco 3560CX-12PD-S.
The code appears to be a workaround for a specific switch which fails to
link in SFI mode. It expects to see AN-37 auto negotiation in order to
link. The Cisco switch is not expecting AN-37 auto negotiation. When the
device starts the manual AN-37, the Cisco switch decides that the port is
confused and stops attempting to link with it. This persists until a power
cycle. A simple driver unload and reload does not resolve the issue, even
if loading with a version of the driver which lacks this workaround.
The authors of the workaround commit have not responded with
clarifications, and the result of the workaround is complete failure to
connect with other switches.
This appears to be a case where the driver can either "correctly" link with
the Juniper MX5 switch, at the cost of bricking the link with the Cisco
switch, or it can behave properly for the Cisco switch, but fail to link
with the Junipir MX5 switch. I do not know enough about the standards
involved to clearly determine whether either switch is at fault or behaving
incorrectly. Nor do I know whether there exists some alternative fix which
corrects behavior with both switches.
Revert the workaround for the Juniper switch.
Fixes: 565736048bd5 ("ixgbe: Manual AN-37 for troublesome link partners for X550 SFI")
Link: https://lore.kernel.org/netdev/cbe874db-9ac9-42b8-afa0-88ea910e1e99@intel.com/T/
Link: https://forum.proxmox.com/threads/intel-x553-sfp-ixgbe-no-go-on-pve8.135129/#post-612291
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Cc: Jeff Daly <jeffd@silicom-usa.com>
Cc: kernel.org-fo5k2w@ycharbi.fr
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://lore.kernel.org/r/20240520-net-2024-05-20-revert-silicom-switch-workaround-v1-1-50f80f261c94@intel.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Testing a network device that has large numbers of bytes/packets may
overflow. Using stats64 when comparing fixes this problem.
I tripped on this while iterating on a qstats patch for mlx5. See below
for confirmation without my added code that this is a bug.
Before this patch (with added debugging output):
$ NETIF=eth0 tools/testing/selftests/drivers/net/stats.py
KTAP version 1
1..4
ok 1 stats.check_pause
ok 2 stats.check_fec
rstat: 481708634 qstat: 666201639514 key: tx-bytes
not ok 3 stats.pkt_byte_sum
ok 4 stats.qstat_by_ifindex
Note the huge delta above ^^^ in the rtnl vs qstats.
After this patch:
$ NETIF=eth0 tools/testing/selftests/drivers/net/stats.py
KTAP version 1
1..4
ok 1 stats.check_pause
ok 2 stats.check_fec
ok 3 stats.pkt_byte_sum
ok 4 stats.qstat_by_ifindex
It looks like rtnl_fill_stats in net/core/rtnetlink.c will attempt to
copy the 64bit stats into a 32bit structure which is probably why this
behavior is occurring.
To show this is happening, you can get the underlying stats that the
stats.py test uses like this:
$ ./cli.py --spec ../../../Documentation/netlink/specs/rt_link.yaml \
--do getlink --json '{"ifi-index": 7}'
And examine the output (heavily snipped to show relevant fields):
'stats': {
'multicast': 3739197,
'rx-bytes': 1201525399,
'rx-packets': 56807158,
'tx-bytes': 492404458,
'tx-packets': 1200285371,
'stats64': {
'multicast': 3739197,
'rx-bytes': 35561263767,
'rx-packets': 56807158,
'tx-bytes': 666212335338,
'tx-packets': 1200285371,
The stats.py test prior to this patch was using the 'stats' structure
above, which matches the failure output on my system.
Comparing side by side, rx-bytes and tx-bytes, and getting ethtool -S
output:
rx-bytes stats: 1201525399
rx-bytes stats64: 35561263767
rx-bytes ethtool: 36203402638
tx-bytes stats: 492404458
tx-bytes stats64: 666212335338
tx-bytes ethtool: 666215360113
Note that the above was taken from a system with an mlx5 NIC, which only
exposes ndo_get_stats64.
Based on the ethtool output and qstat output, it appears that stats.py
should be updated to use the 'stats64' structure for accurate
comparisons when packet/byte counters get very large.
To confirm that this was not related to the qstats code I was iterating
on, I booted a kernel without my driver changes and re-ran the test
which shows the qstats are skipped (as they don't exist for mlx5):
NETIF=eth0 tools/testing/selftests/drivers/net/stats.py
KTAP version 1
1..4
ok 1 stats.check_pause
ok 2 stats.check_fec
ok 3 stats.pkt_byte_sum # SKIP qstats not supported by the device
ok 4 stats.qstat_by_ifindex # SKIP No ifindex supports qstats
But, fetching the stats using the CLI
$ ./cli.py --spec ../../../Documentation/netlink/specs/rt_link.yaml \
--do getlink --json '{"ifi-index": 7}'
Shows the same issue (heavily snipped for relevant fields only):
'stats': {
'multicast': 105489,
'rx-bytes': 530879526,
'rx-packets': 751415,
'tx-bytes': 2510191396,
'tx-packets': 27700323,
'stats64': {
'multicast': 105489,
'rx-bytes': 530879526,
'rx-packets': 751415,
'tx-bytes': 15395093284,
'tx-packets': 27700323,
Comparing side by side with ethtool -S on the unmodified mlx5 driver:
tx-bytes stats: 2510191396
tx-bytes stats64: 15395093284
tx-bytes ethtool: 17718435810
Fixes: f0e6c86e4bab ("testing: net-drv: add a driver test for stats reporting")
Signed-off-by: Joe Damato <jdamato@fastly.com>
Link: https://lore.kernel.org/r/20240520235850.190041-1-jdamato@fastly.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
With the rework of how the __string() handles dynamic strings where it
saves off the source string in field in the helper structure[1], the
assignment of that value to the trace event field is stored in the helper
value and does not need to be passed in again.
This means that with:
__string(field, mystring)
Which use to be assigned with __assign_str(field, mystring), no longer
needs the second parameter and it is unused. With this, __assign_str()
will now only get a single parameter.
There's over 700 users of __assign_str() and because coccinelle does not
handle the TRACE_EVENT() macro I ended up using the following sed script:
git grep -l __assign_str | while read a ; do
sed -e 's/\(__assign_str([^,]*[^ ,]\) *,[^;]*/\1)/' $a > /tmp/test-file;
mv /tmp/test-file $a;
done
I then searched for __assign_str() that did not end with ';' as those
were multi line assignments that the sed script above would fail to catch.
Note, the same updates will need to be done for:
__assign_str_len()
__assign_rel_str()
__assign_rel_str_len()
I tested this with both an allmodconfig and an allyesconfig (build only for both).
[1] https://lore.kernel.org/linux-trace-kernel/20240222211442.634192653@goodmis.org/
Link: https://lore.kernel.org/linux-trace-kernel/20240516133454.681ba6a0@rorschach.local.home
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Julia Lawall <Julia.Lawall@inria.fr>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Acked-by: Jani Nikula <jani.nikula@intel.com>
Acked-by: Christian König <christian.koenig@amd.com> for the amdgpu parts.
Acked-by: Thomas Hellström <thomas.hellstrom@linux.intel.com> #for
Acked-by: Rafael J. Wysocki <rafael@kernel.org> # for thermal
Acked-by: Takashi Iwai <tiwai@suse.de>
Acked-by: Darrick J. Wong <djwong@kernel.org> # xfs
Tested-by: Guenter Roeck <linux@roeck-us.net>
|
|
Use '%pD' to print out the filename, and print out the actual offset
within the file too, rather than just what the virtual address of the
mapping is (which doesn't tell you anything about any mapping offsets).
Also, use the exact vma_lookup() instead of find_vma() - the latter
looks up any vma _after_ the address, which is of questionable value
(yes, maybe you fell off the beginning, but you'd be more likely to fall
off the end).
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
This uses the new ASM_INPUT_RM macro to avoid the bad code generation
issue that clang has with more generic asm inputs.
This ends up avoiding generating code like this:
mov %r10,(%rsp)
tzcnt (%rsp),%rcx
which now becomes just
tzcnt %r10,%rcx
and in the process ends up also removing a few unnecessary stack frames
when the only use was that pointless "asm uses memory location off stack".
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Don't force the inputs to be 'unsigned long', when the comparison can
easily be done in 32-bit if that's more appropriate.
Note that while we can look at the inputs to choose an appropriate size
for the compare instruction, the output is fixed at 'unsigned long'.
That's not technically optimal either, since a 32-bit 'sbbl' would often
be sufficient.
But for the outgoing mask we don't know how the mask ends up being used
(ie we have uses that have an incoming 32-bit array index, but end up
using the mask for other things). That said, it only costs the extra
REX prefix to always generate the 64-bit mask.
[ A 'sbbl' also always technically generates a 64-bit mask, but with the
upper 32 bits clear: that's fine for when the incoming index that will
be masked is already 32-bit, but not if you use the mask to mask a
pointer afterwards, like the file table lookup does ]
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Work around clang problems with asm constraints that have multiple
possibilities, particularly "g" and "rm".
Clang seems to turn inputs like that into the most generic form, which
is the memory input - but to make matters worse, clang won't even use a
possible original memory location, but will spill the value to stack,
and use the stack for the asm input.
See
https://github.com/llvm/llvm-project/issues/20571#issuecomment-980933442
for some explanation of why clang has this strange behavior, but the end
result is that "g" and "rm" really end up generating horrid code.
Link: https://github.com/llvm/llvm-project/issues/20571
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Our applications, built on Elasticsearch[0], frequently create and
delete files. These applications operate within containers, some with a
memory limit exceeding 100GB. Over prolonged periods, the accumulation
of negative dentries within these containers can amount to tens of
gigabytes.
Upon container exit, directories are deleted. However, due to the
numerous associated dentries, this process can be time-consuming. Our
users have expressed frustration with this prolonged exit duration,
which constitutes our first issue.
Simultaneously, other processes may attempt to access the parent
directory of the Elasticsearch directories. Since the task responsible
for deleting the dentries holds the inode lock, processes attempting
directory lookup experience significant delays. This issue, our second
problem, is easily demonstrated:
- Task 1 generates negative dentries:
$ pwd
~/test
$ mkdir es && cd es/ && ./create_and_delete_files.sh
[ After generating tens of GB dentries ]
$ cd ~/test && rm -rf es
[ It will take a long duration to finish ]
- Task 2 attempts to lookup the 'test/' directory
$ pwd
~/test
$ ls
The 'ls' command in Task 2 experiences prolonged execution as Task 1
is deleting the dentries.
We've devised a solution to address both issues by deleting associated
dentry when removing a file. Interestingly, we've noted that a similar
patch was proposed years ago[1], although it was rejected citing the
absence of tangible issues caused by negative dentries. Given our
current challenges, we're resubmitting the proposal. All relevant
stakeholders from previous discussions have been included for reference.
Some alternative solutions are also under discussion[2][3], such as
shrinking child dentries outside of the parent inode lock or even
asynchronously shrinking child dentries. However, given the
straightforward nature of the current solution, I believe this approach
is still necessary.
[ NOTE! This is a pretty fundamental change in how we deal with
unlinking dentries, and it doesn't change the fact that you can have
lots of negative dentries from just doing negative lookups.
But the kernel test robot is at least initially happy with this from a
performance angle, so I'm applying this ASAP just to get more testing
and as a "known fix for an issue people hit in real life".
Put another way: we should still look at the alternatives, and this
patch may get reverted if somebody finds a performance regression on
some other load. - Linus ]
Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Yafang Shao <laoar.shao@gmail.com>
Link: https://github.com/elastic/elasticsearch [0]
Link: https://patchwork.kernel.org/project/linux-fsdevel/patch/1502099673-31620-1-git-send-email-wangkai86@huawei.com [1]
Link: https://lore.kernel.org/linux-fsdevel/20240511200240.6354-2-torvalds@linux-foundation.org/ [2]
Link: https://lore.kernel.org/linux-fsdevel/CAHk-=wjEMf8Du4UFzxuToGDnF3yLaMcrYeyNAaH1NJWa6fwcNQ@mail.gmail.com/ [3]
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Christian Brauner <brauner@kernel.org>
Cc: Jan Kara <jack@suse.cz>
Cc: Waiman Long <longman@redhat.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Wangkai <wangkai86@huawei.com>
Cc: Colin Walters <walters@verbum.org>
Tested-by: kernel test robot <oliver.sang@intel.com>
Link: https://lore.kernel.org/all/202405221518.ecea2810-oliver.sang@intel.com/
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
[bug]
In the virtio_pci_common.c function vp_del_vqs, vp_dev->is_avq is involved
to determine whether it is admin virtqueue, but this function vp_dev->is_avq
may be empty. For installations, virtio_pci_legacy does not assign a value
to vp_dev->is_avq.
[fix]
Check whether it is vp_dev->is_avq before use.
[test]
Test with virsh Attach device
Before this patch, the following command would crash the guest system
After applying the patch, everything seems to be working fine.
Signed-off-by: Li Zhang <zhanglikernel@gmail.com>
Message-Id: <1710566754-3532-1-git-send-email-zhanglikernel@gmail.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
|
|
When request_irq() fails, error path calls vp_del_vqs(). There, as vq is
present in the list, free_irq() is called for the same vector. That
causes following splat:
[ 0.414355] Trying to free already-free IRQ 27
[ 0.414403] WARNING: CPU: 1 PID: 1 at kernel/irq/manage.c:1899 free_irq+0x1a1/0x2d0
[ 0.414510] Modules linked in:
[ 0.414540] CPU: 1 PID: 1 Comm: swapper/0 Not tainted 6.9.0-rc4+ #27
[ 0.414540] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.3-1.fc39 04/01/2014
[ 0.414540] RIP: 0010:free_irq+0x1a1/0x2d0
[ 0.414540] Code: 1e 00 48 83 c4 08 48 89 e8 5b 5d 41 5c 41 5d 41 5e 41 5f c3 cc cc cc cc 90 8b 74 24 04 48 c7 c7 98 80 6c b1 e8 00 c9 f7 ff 90 <0f> 0b 90 90 48 89 ee 4c 89 ef e8 e0 20 b8 00 49 8b 47 40 48 8b 40
[ 0.414540] RSP: 0000:ffffb71480013ae0 EFLAGS: 00010086
[ 0.414540] RAX: 0000000000000000 RBX: ffffa099c2722000 RCX: 0000000000000000
[ 0.414540] RDX: 0000000000000000 RSI: ffffb71480013998 RDI: 0000000000000001
[ 0.414540] RBP: 0000000000000246 R08: 00000000ffffdfff R09: 0000000000000001
[ 0.414540] R10: 00000000ffffdfff R11: ffffffffb18729c0 R12: ffffa099c1c91760
[ 0.414540] R13: ffffa099c1c916a4 R14: ffffa099c1d2f200 R15: ffffa099c1c91600
[ 0.414540] FS: 0000000000000000(0000) GS:ffffa099fec40000(0000) knlGS:0000000000000000
[ 0.414540] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 0.414540] CR2: 0000000000000000 CR3: 0000000008e3e001 CR4: 0000000000370ef0
[ 0.414540] Call Trace:
[ 0.414540] <TASK>
[ 0.414540] ? __warn+0x80/0x120
[ 0.414540] ? free_irq+0x1a1/0x2d0
[ 0.414540] ? report_bug+0x164/0x190
[ 0.414540] ? handle_bug+0x3b/0x70
[ 0.414540] ? exc_invalid_op+0x17/0x70
[ 0.414540] ? asm_exc_invalid_op+0x1a/0x20
[ 0.414540] ? free_irq+0x1a1/0x2d0
[ 0.414540] vp_del_vqs+0xc1/0x220
[ 0.414540] vp_find_vqs_msix+0x305/0x470
[ 0.414540] vp_find_vqs+0x3e/0x1a0
[ 0.414540] vp_modern_find_vqs+0x1b/0x70
[ 0.414540] init_vqs+0x387/0x600
[ 0.414540] virtnet_probe+0x50a/0xc80
[ 0.414540] virtio_dev_probe+0x1e0/0x2b0
[ 0.414540] really_probe+0xc0/0x2c0
[ 0.414540] ? __pfx___driver_attach+0x10/0x10
[ 0.414540] __driver_probe_device+0x73/0x120
[ 0.414540] driver_probe_device+0x1f/0xe0
[ 0.414540] __driver_attach+0x88/0x180
[ 0.414540] bus_for_each_dev+0x85/0xd0
[ 0.414540] bus_add_driver+0xec/0x1f0
[ 0.414540] driver_register+0x59/0x100
[ 0.414540] ? __pfx_virtio_net_driver_init+0x10/0x10
[ 0.414540] virtio_net_driver_init+0x90/0xb0
[ 0.414540] do_one_initcall+0x58/0x230
[ 0.414540] kernel_init_freeable+0x1a3/0x2d0
[ 0.414540] ? __pfx_kernel_init+0x10/0x10
[ 0.414540] kernel_init+0x1a/0x1c0
[ 0.414540] ret_from_fork+0x31/0x50
[ 0.414540] ? __pfx_kernel_init+0x10/0x10
[ 0.414540] ret_from_fork_asm+0x1a/0x30
[ 0.414540] </TASK>
Fix this by calling deleting the current vq when request_irq() fails.
Fixes: 0b0f9dc52ed0 ("Revert "virtio_pci: use shared interrupts for virtqueues"")
Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Message-Id: <20240426150845.3999481-1-jiri@resnulli.us>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
|
Jakub Sitnicki
Parthiban Veerasooran
Hariprasad Kelam
Kuniyuki Iwashima
Geliang Tang
Roded Zats
Daniel Borkmann
Shahab Vahedi
Andrii Nakryiko
Friedrich Vock
Gal Pressman
Carolina Jubran
Rahul Rameshbabu
Tariq Toukan
Maher Sanalla
Mathieu Othacehe
Matt Jan
Shuah Khan
Ken Milmore
Larysa Zaremba
Ryosuke Yasuoka
Paolo Abeni
Jason Xing
Romain Gantois
Dae R. Jeong
Wei Fang
Jacob Keller
Joe Damato
Steven Rostedt (Google)
Linus Torvalds
Yafang Shao
Li Zhang
Jiri Pirko