tests load/attach bpf prog with maps, perfbuf and ringbuf, pinning
them. Then effective caps are dropped and we verify we can
- pick up the pin
- create ringbuf/perfbuf
- get ringbuf/perfbuf events, carry out map update, lookup and delete
- create a link
Negative testing also ensures
- BPF prog load fails
- BPF map create fails
- get fd by id fails
- get next id fails
- query fails
- BTF load fails
Signed-off-by: Alan Maguire <alan.maguire@oracle.com>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/r/1652970334-30510-3-git-send-email-alan.maguire@oracle.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
With unprivileged BPF disabled, all cmds associated with the BPF syscall
are blocked to users without CAP_BPF/CAP_SYS_ADMIN. However there are
use cases where we may wish to allow interactions with BPF programs
without being able to load and attach them. So for example, a process
with required capabilities loads/attaches a BPF program, and a process
with less capabilities interacts with it; retrieving perf/ring buffer
events, modifying map-specified config etc. With all BPF syscall
commands blocked as a result of unprivileged BPF being disabled,
this mode of interaction becomes impossible for processes without
CAP_BPF.
As Alexei notes
"The bpf ACL model is the same as traditional file's ACL.
The creds and ACLs are checked at open(). Then during file's write/read
additional checks might be performed. BPF has such functionality already.
Different map_creates have capability checks while map_lookup has:
map_get_sys_perms(map, f) & FMODE_CAN_READ.
In other words it's enough to gate FD-receiving parts of bpf
with unprivileged_bpf_disabled sysctl.
The rest is handled by availability of FD and access to files in bpffs."
So key fd creation syscall commands BPF_PROG_LOAD and BPF_MAP_CREATE
are blocked with unprivileged BPF disabled and no CAP_BPF.
And as Alexei notes, map creation with unprivileged BPF disabled off
blocks creation of maps aside from array, hash and ringbuf maps.
Programs responsible for loading and attaching the BPF program
can still control access to its pinned representation by restricting
permissions on the pin path, as with normal files.
Signed-off-by: Alan Maguire <alan.maguire@oracle.com>
Acked-by: Yonghong Song <yhs@fb.com>
Acked-by: Shung-Hsi Yu <shung-hsi.yu@suse.com>
Acked-by: KP Singh <kpsingh@kernel.org>
Link: https://lore.kernel.org/r/1652970334-30510-2-git-send-email-alan.maguire@oracle.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Tracing and syscall BPF program types are very convenient to add BPF
capabilities to subsystem otherwise not BPF capable.
When we add kfuncs capabilities to those program types, we can add
BPF features to subsystems without having to touch BPF core.
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Link: https://lore.kernel.org/r/20220518205924.399291-2-benjamin.tissoires@redhat.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Joanne Koong says:
====================
Add a bhash2 table hashed by port + address
This patchset proposes adding a bhash2 table that hashes by port and address.
The motivation behind bhash2 is to expedite bind requests in situations where
the port has many sockets in its bhash table entry, 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
The patches are as follows:
1/2 - Adds a second bhash table (bhash2) hashed by port and address
2/2 - Adds a test for timing how long an additional bind request takes when
the bhash entry is populated
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 - 0.000018 seconds
ipv6:
before - 0.002431 seconds
with bhash2 - 0.000021 seconds
====================
Link: https://lore.kernel.org/r/20220520001834.2247810-1-kuba@kernel.org
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 test locally, I did:
* ulimit -n 65535000
* ip addr add 2001:0db8:0:f101::1 dev eth0
* ./bind_bhash_test 443
Signed-off-by: Joanne Koong <joannelkoong@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
We currently have one tcp bind table (bhash) which hashes by port
number only. In the socket bind path, we check for bind conflicts by
traversing the specified port's inet_bind2_bucket while holding the
bucket'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, checking for a bind conflict 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 proposes adding a second bind table, bhash2, that hashes by
port and ip address. Searching the bhash2 table leads to significantly
faster conflict resolution and less time holding the spinlock.
Signed-off-by: Joanne Koong <joannelkoong@gmail.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Acked-by: Kuniyuki Iwashima <kuniyu@amazon.co.jp>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
GCC array-bounds warns that ipc_coredump_get_list() under-allocates
the size of struct iosm_cd_table *cd_table.
This is avoidable - we just need a flexible array. Nothing calls
sizeof() on struct iosm_cd_list or anything that contains it.
Reviewed-by: M Chetan Kumar <m.chetan.kumar@intel.com>
Link: https://lore.kernel.org/r/20220520060013.2309497-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
GCC array bounds checking complains that ulp_id is validated
only against upper bound. Make it unsigned.
Reviewed-by: Michael Chan <michael.chan@broadcom.com>
Link: https://lore.kernel.org/r/20220520061955.2312968-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Commit 49bb39bdda ("selftests: fib_nexthops: Make the test more robust")
increased the timeout of ping commands to 5 seconds, to make the test
more robust. Make the timeout configurable using '-w' argument to allow
user to change it depending on the system that runs the test. Some systems
suffer from slow forwarding performance, so they may need to change the
timeout.
Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Shuah Khan <skhan@linuxfoundation.org>
Link: https://lore.kernel.org/r/20220519070921.3559701-1-amcohen@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Sometimes t7xx_cldma_gpd_set_next_ptr() is called under spin lock,
so add 'gfp_mask' parameter in t7xx_cldma_gpd_set_next_ptr() to pass
the flag.
Fixes: 39d439047f ("net: wwan: t7xx: Add control DMA interface")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Reviewed-by: Loic Poulain <loic.poulain@linaro.org>
Link: https://lore.kernel.org/r/20220519032108.2996400-1-yangyingliang@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When user sets skb_defer_max to 1 the kick threshold is 0
(half of 1). If we increment queue length before the check
the kick will never happen, and the skb may get stranded.
This is likely harmless but can be avoided by moving the
increment after the check. This way skb_defer_max == 1
will always kick. Still a silly config to have, but
somehow that feels more correct.
While at it drop a comment which seems to be outdated
or confusing, and wrap the defer_count write with
a WRITE_ONCE() since it's read on the fast path
that avoids taking the lock.
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20220518185522.2038683-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Siena only supports software TSO. This means more code can be deleted,
as pointed out by the Smatch static checker warning:
drivers/net/ethernet/sfc/siena/tx.c:184 __efx_siena_enqueue_skb()
warn: duplicate check 'segments' (previous on line 158)
Fixes: 956f2d86cb ("sfc/siena: Remove build references to missing functionality")
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Link: https://lore.kernel.org/kernel-janitors/YoH5tJMnwuGTrn1Z@kili/
Signed-off-by: Martin Habets <habetsm.xilinx@gmail.com>
Link: https://lore.kernel.org/r/165294463549.23865.4557617334650441347.stgit@palantir17.mph.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Currently the trampoline_count test doesn't include any fmod_ret bpf
programs, fix it to make the test cover all possible trampoline program
types.
Since fmod_ret bpf programs can't be attached to __set_task_comm function,
as it's neither whitelisted for error injection nor a security hook, change
it to bpf_modify_return_test.
This patch also does some other cleanups such as removing duplicate code,
dropping inconsistent comments, etc.
Signed-off-by: Yuntao Wang <ytcoode@gmail.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20220519150610.601313-1-ytcoode@gmail.com
Mat Martineau says:
====================
This patch set adds BPF access to mptcp_sock structures, along with
associated self tests. You may recognize some of the code from earlier
(https://lore.kernel.org/bpf/20200918121046.190240-6-nicolas.rybowski@tessares.net/)
but it has been reworked quite a bit.
v1 -> v2: Emit BTF type, add func_id checks in verifier.c and bpf_trace.c,
remove build check for CONFIG_BPF_JIT, add selftest check for CONFIG_MPTCP,
and add a patch to include CONFIG_IKCONFIG/CONFIG_IKCONFIG_PROC for the
BPF self tests.
v2 -> v3: Access sysctl through the filesystem to work around CI use of
the more limited busybox sysctl command.
v3 -> v4: Dropped special case kernel code for tcp_sock is_mptcp, use
existing bpf_tcp_helpers.h, and add check for 'ip mptcp monitor' support.
v4 -> v5: Use BPF test skeleton, more consistent use of ASSERT macros,
drop some unnecessary parameters / checks, and use tracing to acquire
MPTCP token.
Geliang Tang (6):
bpf: add bpf_skc_to_mptcp_sock_proto
selftests/bpf: Enable CONFIG_IKCONFIG_PROC in config
selftests/bpf: test bpf_skc_to_mptcp_sock
selftests/bpf: verify token of struct mptcp_sock
selftests/bpf: verify ca_name of struct mptcp_sock
selftests/bpf: verify first of struct mptcp_sock
====================
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
This patch verifies the 'first' struct member of struct mptcp_sock, which
points to the first subflow of msk. Save 'sk' in mptcp_storage, and verify
it with 'first' in verify_msk().
v5:
- Use ASSERT_EQ() instead of a manual comparison + log (Andrii).
Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Link: https://lore.kernel.org/bpf/20220519233016.105670-8-mathew.j.martineau@linux.intel.com
This patch verifies another member of struct mptcp_sock, ca_name. Add a
new function get_msk_ca_name() to read the sysctl tcp_congestion_control
and verify it in verify_msk().
v3: Access the sysctl through the filesystem to avoid compatibility
issues with the busybox sysctl command.
v4: use ASSERT_* instead of CHECK_FAIL (Andrii)
v5: use ASSERT_STRNEQ() instead of strncmp() (Andrii)
Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Link: https://lore.kernel.org/bpf/20220519233016.105670-7-mathew.j.martineau@linux.intel.com
This patch verifies the struct member token of struct mptcp_sock. Add a
new member token in struct mptcp_storage to store the token value of the
msk socket got by bpf_skc_to_mptcp_sock(). Trace the kernel function
mptcp_pm_new_connection() by using bpf fentry prog to obtain the msk token
and save it in a global bpf variable. Pass the variable to verify_msk() to
verify it with the token saved in socket_storage_map.
v4:
- use ASSERT_* instead of CHECK_FAIL (Andrii)
- skip the test if 'ip mptcp monitor' is not supported (Mat)
v5:
- Drop 'ip mptcp monitor', trace mptcp_pm_new_connection instead (Martin)
- Use ASSERT_EQ (Andrii)
Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Link: https://lore.kernel.org/bpf/20220519233016.105670-6-mathew.j.martineau@linux.intel.com
This patch extends the MPTCP test base, to test the new helper
bpf_skc_to_mptcp_sock().
Define struct mptcp_sock in bpf_tcp_helpers.h, use bpf_skc_to_mptcp_sock
to get the msk socket in progs/mptcp_sock.c and store the infos in
socket_storage_map.
Get the infos from socket_storage_map in prog_tests/mptcp.c. Add a new
function verify_msk() to verify the infos of MPTCP socket, and rename
verify_sk() to verify_tsk() to verify TCP socket only.
v2: Add CONFIG_MPTCP check for clearer error messages
v4:
- use ASSERT_* instead of CHECK_FAIL (Andrii)
- drop bpf_mptcp_helpers.h (Andrii)
v5:
- some 'ASSERT_*' were replaced in the next commit by mistake.
- Drop CONFIG_MPTCP (Martin)
- Use ASSERT_EQ (Andrii)
Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Link: https://lore.kernel.org/bpf/20220519233016.105670-5-mathew.j.martineau@linux.intel.com
This patch adds a base for MPTCP specific tests.
It is currently limited to the is_mptcp field in case of plain TCP
connection because there is no easy way to get the subflow sk from a msk
in userspace. This implies that we cannot lookup the sk_storage attached
to the subflow sk in the sockops program.
v4:
- add copyright 2022 (Andrii)
- use ASSERT_* instead of CHECK_FAIL (Andrii)
- drop SEC("version") (Andrii)
- use is_mptcp in tcp_sock, instead of bpf_tcp_sock (Martin & Andrii)
v5:
- Drop connect_to_mptcp_fd (Martin)
- Use BPF test skeleton (Andrii)
- Use ASSERT_EQ (Andrii)
- Drop the 'msg' parameter of verify_sk
Co-developed-by: Geliang Tang <geliang.tang@suse.com>
Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Signed-off-by: Nicolas Rybowski <nicolas.rybowski@tessares.net>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Link: https://lore.kernel.org/bpf/20220519233016.105670-4-mathew.j.martineau@linux.intel.com
CONFIG_IKCONFIG_PROC is required by BPF selftests, otherwise we get
errors like this:
libbpf: failed to open system Kconfig
libbpf: failed to load object 'kprobe_multi'
libbpf: failed to load BPF skeleton 'kprobe_multi': -22
It's because /proc/config.gz is opened in bpf_object__read_kconfig_file()
in tools/lib/bpf/libbpf.c:
file = gzopen("/proc/config.gz", "r");
So this patch enables CONFIG_IKCONFIG and CONFIG_IKCONFIG_PROC in
tools/testing/selftests/bpf/config.
Suggested-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20220519233016.105670-3-mathew.j.martineau@linux.intel.com
This patch implements a new struct bpf_func_proto, named
bpf_skc_to_mptcp_sock_proto. Define a new bpf_id BTF_SOCK_TYPE_MPTCP,
and a new helper bpf_skc_to_mptcp_sock(), which invokes another new
helper bpf_mptcp_sock_from_subflow() in net/mptcp/bpf.c to get struct
mptcp_sock from a given subflow socket.
v2: Emit BTF type, add func_id checks in verifier.c and bpf_trace.c,
remove build check for CONFIG_BPF_JIT
v5: Drop EXPORT_SYMBOL (Martin)
Co-developed-by: Nicolas Rybowski <nicolas.rybowski@tessares.net>
Co-developed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Nicolas Rybowski <nicolas.rybowski@tessares.net>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20220519233016.105670-2-mathew.j.martineau@linux.intel.com
comments from Andrii Nakryiko, details in here:
https://lore.kernel.org/lkml/20220511093854.411-1-zhoufeng.zf@bytedance.com/T/
use /* */ instead of //
use libbpf_num_possible_cpus() instead of sysconf(_SC_NPROCESSORS_ONLN)
use 8 bytes for value size
fix memory leak
use ASSERT_EQ instead of ASSERT_OK
add bpf_loop to fetch values on each possible CPU
Fixes: ed7c13776e ("selftests/bpf: add test case for bpf_map_lookup_percpu_elem")
Signed-off-by: Feng Zhou <zhoufeng.zf@bytedance.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20220518025053.20492-1-zhoufeng.zf@bytedance.com
Looks like the IPv6 version of the patch under Fixes was
a copy/paste of the IPv4 but hit the wrong spot.
It is tcp_v6_rcv() which uses drop_reason as a boolean, and
needs to be protected against reason == 0 before calling free.
tcp_v6_do_rcv() has a pretty straightforward flow.
The resulting warning looks like this:
WARNING: CPU: 1 PID: 0 at net/core/skbuff.c:775
Call Trace:
tcp_v6_rcv (net/ipv6/tcp_ipv6.c:1767)
ip6_protocol_deliver_rcu (net/ipv6/ip6_input.c:438)
ip6_input_finish (include/linux/rcupdate.h:726)
ip6_input (include/linux/netfilter.h:307)
Fixes: f8319dfd1b ("net: tcp: reset 'drop_reason' to NOT_SPCIFIED in tcp_v{4,6}_rcv()")
Tested-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Link: https://lore.kernel.org/r/20220520021347.2270207-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Alex Elder says:
====================
net: ipa: a mix of patches
This series includes a mix of things things that are generally
minor. The first four are sort of unrelated fixes, and summarizing
them here wouldn't be that helpful.
The last three together make it so only the "configuration data" we
need after initialization is saved for later use. Most such data is
used only during driver initialization. But endpoint configuration
is needed later, so the last patch saves a copy of that. Eventually
we'll want to support reconfiguring endpoints at runtime as well,
and this will facilitate that.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
All elements of the default endpoint configuration are used in the
code when programming an endpoint for use. But none of the other
configuration data is ever needed once things are initialized.
So rather than saving a pointer to *all* of the configuration data,
save a copy of only the endpoint configuration portion.
This will eventually allow endpoint configuration to be modifiable
at runtime. But even before that it means we won't keep a pointer
to configuration data after when no longer needed.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Rename the just-moved data structure types to drop the "_data"
suffix, to make it more obvious they are no longer meant to be used
just as read-only initialization data. Rename the fields and
variables of these types to use "config" instead of "data" in the
name. This is another small step meant to facilitate review.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Move the definitions of the structures defining endpoint-specific
configuration data out of "ipa_data.h" and into "ipa_endpoint.h".
This is a trivial movement of code without any other change, to
prepare for the next few patches.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
About half of the fields set by the call in ipa_modem_netdev_setup()
are overwritten after the call. Instead, just skip the call, and
open-code the (other) assignments it makes to the net_device
structure fields.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
If we program an RX endpoint to have no header (header length is 0),
header-related endpoint configuration values are meaningless and are
ignored.
The only case we support that defines a header is QMAP endpoints.
In ipa_endpoint_init_hdr_ext() we set the endianness mask value
unconditionally, but it should not be done if there is no header
(meaning it is not configured for QMAP).
Set the endianness conditionally, and rearrange the logic in that
function slightly to avoid testing the qmap flag twice.
Delete an incorrect comment in ipa_endpoint_init_aggr().
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
The CHANNEL_NOT_RUNNING error condition has been generalized, so
rename it to be INCORRECT_CHANNEL_STATE.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
In gsi_channel_update(), a reference count is taken on the last
completed transaction "to keep it from completing" before we give
the event back to the hardware. Completion processing for that
transaction (and any other "new" ones) will not occur until after
this function returns, so there's no risk it completing early. So
there's no need to take and drop the additional transaction
reference.
Use local variables in the call to gsi_evt_ring_doorbell().
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pablo Neira Ayuso says:
====================
Netfilter updates for net-next
The following patchset contains Netfilter updates for net-next, misc
updates and fallout fixes from recent Florian's code rewritting (from
last pull request):
1) Use new flowi4_l3mdev field in ip_route_me_harder(), from Martin Willi.
2) Avoid unnecessary GC with a timestamp in conncount, from William Tu
and Yifeng Sun.
3) Remove TCP conntrack debugging, from Florian Westphal.
4) Fix compilation warning in ctnetlink, from Florian.
* git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf-next:
netfilter: ctnetlink: fix up for "netfilter: conntrack: remove unconfirmed list"
netfilter: conntrack: remove pr_debug callsites from tcp tracker
netfilter: nf_conncount: reduce unnecessary GC
netfilter: Use l3mdev flow key when re-routing mangled packets
====================
Link: https://lore.kernel.org/r/20220519220206.722153-1-pablo@netfilter.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
I missed this in the barrage of GCC 12 warnings. Commit cf2df74e20
("net: fix dev_fill_forward_path with pppoe + bridge") changed
the pointer into an array.
Fixes: d7e6f58360 ("Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net")
Link: https://lore.kernel.org/r/20220520012555.2262461-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Mat Martineau says:
====================
mptcp: Miscellaneous fixes and a new test case
Patches 1 and 3 remove helpers that were iterating over the subflow
connection list without proper locking. Iteration was not needed in
either case.
Patch 2 fixes handling of MP_FAIL timeout, checking for orphaned
subflows instead of using the MPTCP socket data lock and connection
state.
Patch 4 adds a test for MP_FAIL timeout using tc pedit to induce checksum
failures.
====================
Link: https://lore.kernel.org/r/20220518220446.209750-1-mathew.j.martineau@linux.intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Add the multiple subflows test case for MP_FAIL, to test the MP_FAIL
reset case. Use the test_linkfail value to make 1024KB test files.
Invoke reset_with_fail() to use 'iptables' and 'tc action pedit' rules
to produce the bit flips to trigger the checksum failures on ns2eth2.
Add delays on ns2eth1 to make sure more data can translate on ns2eth2.
The check_invert flag is enabled in reset_with_fail(), so this test
prints out the inverted bytes, instead of the file mismatch errors.
Invoke pedit_action_pkts() to get the numbers of the packets edited
by the tc pedit actions, and print this numbers to the output.
Co-developed-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The MPTCP socket's conn_list (list of subflows) requires the socket lock
to access. The MP_FAIL timeout code added such an access, where it would
check the list of subflows both in timer context and (later) in workqueue
context where the socket lock is held.
Rather than check the list twice, remove the check in the timeout
handler and only depend on the check in the workqueue. Also remove the
MPTCP_FAIL_NO_RESPONSE flag, since mptcp_mp_fail_no_response() has
insignificant overhead and can be checked on each worker run.
Fixes: 49fa1919d6 ("mptcp: reset subflow when MP_FAIL doesn't respond")
Reported-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
MP_FAIL timeout (waiting for a peer to respond to an MP_FAIL with
another MP_FAIL) is implemented using the MPTCP socket's sk_timer. That
timer is also used at MPTCP socket close, so it's important to not have
the two timer users interfere with each other.
At MPTCP socket close, all subflows are orphaned before sk_timer is
manipulated. By checking the SOCK_DEAD flag on the subflows, each
subflow can determine if the timer is safe to alter without acquiring
any MPTCP-level lock. This replaces code that was using the
mptcp_data_lock and MPTCP-level socket state checks that did not
correctly protect the timer.
Fixes: 49fa1919d6 ("mptcp: reset subflow when MP_FAIL doesn't respond")
Reviewed-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The mentioned helper requires the msk socket lock, and the
current callers don't own it nor can't acquire it, so the
access is racy.
All the current callers are really checking for infinite mapping
fallback, and the latter condition is explicitly tracked by
the relevant msk variable: we can safely remove the caller usage
- and the caller itself.
The issue is present since MP_FAIL implementation, but the
fix only applies since the infinite fallback support, ence the
somewhat unexpected fixes tag.
Fixes: 0530020a7c ("mptcp: track and update contiguous data status")
Acked-and-tested-by: Geliang Tang <geliang.tang@suse.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
This patch improves TCP PRR loss recovery behavior for a corner
case. Previously during PRR conservation-bound mode, it strictly
sends the amount equals to the amount newly acked or s/acked.
The patch changes s.t. PRR may send additional amount that was banked
previously (e.g. application-limited) in the conservation-bound
mode, similar to the slow-start mode. This unifies and simplifies the
algorithm further and may improve the recovery latency. This change
still follow the general packet conservation design principle and
always keep inflight/cwnd below the slow start threshold set
by the congestion control module.
PRR is described in RFC 6937. We'll include this change in the
latest revision rfc6937-bis as well.
Reported-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20220519003410.2531936-1-ycheng@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The current title of our section of the documentation is
Linux Networking Documentation. Since we're describing
a section of Linux Documentation repeating those two
words seems redundant.
Link: https://lore.kernel.org/r/20220518234346.2088436-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
GCC 12 warns:
drivers/net/wwan/iosm/iosm_ipc_protocol_ops.c: In function ‘ipc_protocol_dl_td_process’:
drivers/net/wwan/iosm/iosm_ipc_protocol_ops.c:406:13: warning: the comparison will always evaluate as ‘true’ for the address of ‘cb’ will never be NULL [-Waddress]
406 | if (!IPC_CB(skb)) {
| ^
Indeed the check seems entirely pointless. Hopefully the other
validation checks will catch if the cb is bad, but it can't be
NULL.
Reviewed-by: M Chetan Kumar <m.chetan.kumar@intel.com>
Link: https://lore.kernel.org/r/20220519004342.2109832-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Martin Blumenstingl says:
====================
lantiq_gswip: Two small fixes
While updating the Lantiq target in OpenWrt to Linux 5.15 I came across
an FDB related error message. While that still needs to be solved I
found two other small issues on the way.
This series fixes the two minor issues found while revisiting the FDB
code in the lantiq_gswip driver:
- The first patch fixes the start index used in gswip_port_fdb() to
find the entry with the matching bridge. The updated logic is now
consistent with the rest of the driver.
- The second patch fixes a typo in a dev_err() message.
[0] https://lore.kernel.org/netdev/20220517194015.1081632-1-martin.blumenstingl@googlemail.com/
====================
Link: https://lore.kernel.org/r/20220518220051.1520023-1-martin.blumenstingl@googlemail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
gswip_port_fdb_dump() reads the MAC bridge entries. The error message
should say "failed to read mac bridge entry". While here, also add the
index to the error print so humans can get to the cause of the problem
easier.
Acked-by: Hauke Mehrtens <hauke@hauke-m.de>
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
