syzbot reported an array-index-out-of-bounds when printing out bpf
insns. Further investigation shows the insn is illegal but
is printed out due to log level 1 or 2 before actual insn verification
in do_check().
This particular illegal insn is a MOVSX insn with offset value 2.
The legal offset value for MOVSX should be 8, 16 and 32.
The disasm sign-extension-size array index is calculated as
(insn->off / 8) - 1
and offset value 2 gives an out-of-bound index -1.
Tighten the checking for MOVSX insn in disasm.c to avoid
array-index-out-of-bounds issue.
Reported-by: syzbot+3758842a6c01012aa73b@syzkaller.appspotmail.com
Fixes: f835bb6222 ("bpf: Add kernel/bpftool asm support for new instructions")
Signed-off-by: Yonghong Song <yonghong.song@linux.dev>
Acked-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20230731204534.1975311-1-yonghong.song@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Daniel Xu says:
====================
Support defragmenting IPv(4|6) packets in BPF
=== Context ===
In the context of a middlebox, fragmented packets are tricky to handle.
The full 5-tuple of a packet is often only available in the first
fragment which makes enforcing consistent policy difficult. There are
really only two stateless options, neither of which are very nice:
1. Enforce policy on first fragment and accept all subsequent fragments.
This works but may let in certain attacks or allow data exfiltration.
2. Enforce policy on first fragment and drop all subsequent fragments.
This does not really work b/c some protocols may rely on
fragmentation. For example, DNS may rely on oversized UDP packets for
large responses.
So stateful tracking is the only sane option. RFC 8900 [0] calls this
out as well in section 6.3:
Middleboxes [...] should process IP fragments in a manner that is
consistent with [RFC0791] and [RFC8200]. In many cases, middleboxes
must maintain state in order to achieve this goal.
=== BPF related bits ===
Policy has traditionally been enforced from XDP/TC hooks. Both hooks
run before kernel reassembly facilities. However, with the new
BPF_PROG_TYPE_NETFILTER, we can rather easily hook into existing
netfilter reassembly infra.
The basic idea is we bump a refcnt on the netfilter defrag module and
then run the bpf prog after the defrag module runs. This allows bpf
progs to transparently see full, reassembled packets. The nice thing
about this is that progs don't have to carry around logic to detect
fragments.
=== Changelog ===
Changes from v5:
* Fix defrag disable codepaths
Changes from v4:
* Refactor module handling code to not sleep in rcu_read_lock()
* Also unify the v4 and v6 hook structs so they can share codepaths
* Fixed some checkpatch.pl formatting warnings
Changes from v3:
* Correctly initialize `addrlen` stack var for recvmsg()
Changes from v2:
* module_put() if ->enable() fails
* Fix CI build errors
Changes from v1:
* Drop bpf_program__attach_netfilter() patches
* static -> static const where appropriate
* Fix callback assignment order during registration
* Only request_module() if callbacks are missing
* Fix retval when modprobe fails in userspace
* Fix v6 defrag module name (nf_defrag_ipv6_hooks -> nf_defrag_ipv6)
* Simplify priority checking code
* Add warning if module doesn't assign callbacks in the future
* Take refcnt on module while defrag link is active
[0]: https://datatracker.ietf.org/doc/html/rfc8900
====================
Link: https://lore.kernel.org/r/cover.1689970773.git.dxu@dxuuu.xyz
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
These selftests tests 2 major scenarios: the BPF based defragmentation
can successfully be done and that packet pointers are invalidated after
calls to the kfunc. The logic is similar for both ipv4 and ipv6.
In the first scenario, we create a UDP client and UDP echo server. The
the server side is fairly straightforward: we attach the prog and simply
echo back the message.
The on the client side, we send fragmented packets to and expect the
reassembled message back from the server.
Signed-off-by: Daniel Xu <dxu@dxuuu.xyz>
Link: https://lore.kernel.org/r/33e40fdfddf43be93f2cb259303f132f46750953.1689970773.git.dxu@dxuuu.xyz
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
This commit adds support for enabling IP defrag using pre-existing
netfilter defrag support. Basically all the flag does is bump a refcnt
while the link the active. Checks are also added to ensure the prog
requesting defrag support is run _after_ netfilter defrag hooks.
We also take care to avoid any issues w.r.t. module unloading -- while
defrag is active on a link, the module is prevented from unloading.
Signed-off-by: Daniel Xu <dxu@dxuuu.xyz>
Reviewed-by: Florian Westphal <fw@strlen.de>
Link: https://lore.kernel.org/r/5cff26f97e55161b7d56b09ddcf5f8888a5add1d.1689970773.git.dxu@dxuuu.xyz
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
In internal testing of test_maps, we sometimes observed failures like:
test_maps: test_maps.c:173: void test_hashmap_percpu(unsigned int, void *):
Assertion `bpf_map_update_elem(fd, &key, value, BPF_ANY) == 0' failed.
where the errno is ENOMEM. After some troubleshooting and enabling
the warnings, we saw:
[ 91.304708] percpu: allocation failed, size=8 align=8 atomic=1, atomic alloc failed, no space left
[ 91.304716] CPU: 51 PID: 24145 Comm: test_maps Kdump: loaded Tainted: G N 6.1.38-smp-DEV #7
[ 91.304719] Hardware name: Google Astoria/astoria, BIOS 0.20230627.0-0 06/27/2023
[ 91.304721] Call Trace:
[ 91.304724] <TASK>
[ 91.304730] [<ffffffffa7ef83b9>] dump_stack_lvl+0x59/0x88
[ 91.304737] [<ffffffffa7ef83f8>] dump_stack+0x10/0x18
[ 91.304738] [<ffffffffa75caa0c>] pcpu_alloc+0x6fc/0x870
[ 91.304741] [<ffffffffa75ca302>] __alloc_percpu_gfp+0x12/0x20
[ 91.304743] [<ffffffffa756785e>] alloc_bulk+0xde/0x1e0
[ 91.304746] [<ffffffffa7566c02>] bpf_mem_alloc_init+0xd2/0x2f0
[ 91.304747] [<ffffffffa7547c69>] htab_map_alloc+0x479/0x650
[ 91.304750] [<ffffffffa751d6e0>] map_create+0x140/0x2e0
[ 91.304752] [<ffffffffa751d413>] __sys_bpf+0x5a3/0x6c0
[ 91.304753] [<ffffffffa751c3ec>] __x64_sys_bpf+0x1c/0x30
[ 91.304754] [<ffffffffa7ef847a>] do_syscall_64+0x5a/0x80
[ 91.304756] [<ffffffffa800009b>] entry_SYSCALL_64_after_hwframe+0x63/0xcd
This makes sense, because in atomic context, percpu allocation would
not create new chunks; it would only create in non-atomic contexts.
And if during prefill all precpu chunks are full, -ENOMEM would
happen immediately upon next unit_alloc.
Prefill phase does not actually run in atomic context, so we can
use this fact to allocate non-atomically with GFP_KERNEL instead
of GFP_NOWAIT. This avoids the immediate -ENOMEM.
GFP_NOWAIT has to be used in unit_alloc when bpf program runs
in atomic context. Even if bpf program runs in non-atomic context,
in most cases, rcu read lock is enabled for the program so
GFP_NOWAIT is still needed. This is often also the case for
BPF_MAP_UPDATE_ELEM syscalls.
Signed-off-by: YiFei Zhu <zhuyifei@google.com>
Acked-by: Yonghong Song <yonghong.song@linux.dev>
Acked-by: Hou Tao <houtao1@huawei.com>
Link: https://lore.kernel.org/r/20230728043359.3324347-1-zhuyifei@google.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
The kernel test robot reported compilation warnings when -Wparentheses is
added to KBUILD_CFLAGS with gcc compiler. The following is the error message:
.../bpf-next/kernel/bpf/verifier.c: In function ‘coerce_reg_to_size_sx’:
.../bpf-next/kernel/bpf/verifier.c:5901:14:
error: suggest parentheses around comparison in operand of ‘==’ [-Werror=parentheses]
if (s64_max >= 0 == s64_min >= 0) {
~~~~~~~~^~~~
.../bpf-next/kernel/bpf/verifier.c: In function ‘coerce_subreg_to_size_sx’:
.../bpf-next/kernel/bpf/verifier.c:5965:14:
error: suggest parentheses around comparison in operand of ‘==’ [-Werror=parentheses]
if (s32_min >= 0 == s32_max >= 0) {
~~~~~~~~^~~~
To fix the issue, add proper parentheses for the above '>=' condition
to silence the warning/error.
I tried a few clang compilers like clang16 and clang18 and they do not emit
such warnings with -Wparentheses.
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202307281133.wi0c4SqG-lkp@intel.com/
Signed-off-by: Yonghong Song <yonghong.song@linux.dev>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Link: https://lore.kernel.org/r/20230728055740.2284534-1-yonghong.song@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Yonghong Song says:
====================
bpf: Support new insns from cpu v4
In previous discussion ([1]), it is agreed that we should introduce
cpu version 4 (llvm flag -mcpu=v4) which contains some instructions
which can simplify code, make code easier to understand, fix the
existing problem, or simply for feature completeness. More specifically,
the following new insns are proposed:
. sign extended load
. sign extended mov
. bswap
. signed div/mod
. ja with 32-bit offset
This patch set added kernel support for insns proposed in [1] except
BPF_ST which already has full kernel support. Beside the above proposed
insns, LLVM will generate BPF_ST insn as well under -mcpu=v4.
The llvm patch ([2]) has been merged into llvm-project 'main' branch.
The patchset implements interpreter, jit and verifier support for these new
insns.
For this patch set, I tested cpu v2/v3/v4 and the selftests are all passed.
I also tested selftests introduced in this patch set with additional changes
beside normal jit testing (bpf_jit_enable = 1 and bpf_jit_harden = 0)
- bpf_jit_enable = 0
- bpf_jit_enable = 1 and bpf_jit_harden = 1
and both testing passed.
[1] https://lore.kernel.org/bpf/4bfe98be-5333-1c7e-2f6d-42486c8ec039@meta.com/
[2] https://reviews.llvm.org/D144829
Changelogs:
v4 -> v5:
. for v4, patch 8/17 missed in mailing list and patchwork, so resend.
. rebase on top of master
v3 -> v4:
. some minor asm syntax adjustment based on llvm change.
. add clang version and target arch guard for new tests
so they can still compile with old llvm compilers.
. some changes to the bpf doc.
v2 -> v3:
. add missed disasm change from v2.
. handle signed load of ctx fields properly.
. fix some interpreter sdiv/smod error when bpf_jit_enable = 0.
. fix some verifier range bounding errors.
. add more C tests.
RFCv1 -> v2:
. add more verifier supports for signed extend load and mov insns.
. rename some insn names to be more consistent with intel practice.
. add cpuv4 test runner for test progs.
. add more unit and C tests.
. add documentation.
====================
Link: https://lore.kernel.org/r/20230728011143.3710005-1-yonghong.song@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
The following ldsx cases are tested:
- signed readonly map value
- read/write map value
- probed memory
- not-narrowed ctx field access
- narrowed ctx field access.
Without previous proper verifier/git handling, the test will fail.
If cpuv4 is not supported either by compiler or by jit,
the test will be skipped.
# ./test_progs -t ldsx_insn
#113/1 ldsx_insn/map_val and probed_memory:SKIP
#113/2 ldsx_insn/ctx_member_sign_ext:SKIP
#113/3 ldsx_insn/ctx_member_narrow_sign_ext:SKIP
#113 ldsx_insn:SKIP
Summary: 1/0 PASSED, 3 SKIPPED, 0 FAILED
Signed-off-by: Yonghong Song <yonghong.song@linux.dev>
Link: https://lore.kernel.org/r/20230728011336.3723434-1-yonghong.song@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Add unit tests for new ldsx insns. The test includes sign-extension
with a single value or with a value range.
If cpuv4 is not supported due to
(1) older compiler, e.g., less than clang version 18, or
(2) test runner test_progs and test_progs-no_alu32 which tests
cpu v2 and v3, or
(3) non-x86_64 arch not supporting new insns in jit yet,
a dummy program is added with below output:
#318/1 verifier_ldsx/cpuv4 is not supported by compiler or jit, use a dummy test:OK
#318 verifier_ldsx:OK
to indicate the test passed with a dummy test instead of actually
testing cpuv4. I am using a dummy prog to avoid changing the
verifier testing infrastructure. Once clang 18 is widely available
and other architectures support cpuv4, at least for CI run,
the dummy program can be removed.
Signed-off-by: Yonghong Song <yonghong.song@linux.dev>
Link: https://lore.kernel.org/r/20230728011304.3719139-1-yonghong.song@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Similar to no-alu32 runner, if clang compiler supports -mcpu=v4,
a cpuv4 runner is created to test bpf programs compiled with
-mcpu=v4.
The following are some num-of-insn statistics for each newer
instructions based on existing selftests, excluding subsequent
cpuv4 insn specific tests.
insn pattern # of instructions
reg = (s8)reg 4
reg = (s16)reg 4
reg = (s32)reg 144
reg = *(s8 *)(reg + off) 13
reg = *(s16 *)(reg + off) 14
reg = *(s32 *)(reg + off) 15215
reg = bswap16 reg 142
reg = bswap32 reg 38
reg = bswap64 reg 14
reg s/= reg 0
reg s%= reg 0
gotol <offset> 58
Note that in llvm -mcpu=v4 implementation, the compiler is a little
bit conservative about generating 'gotol' insn (32-bit branch offset)
as it didn't precise count the number of insns (e.g., some insns are
debug insns, etc.). Compared to old 'goto' insn, newer 'gotol' insn
should have comparable verification states to 'goto' insn.
With current patch set, all selftests passed with -mcpu=v4
when running test_progs-cpuv4 binary. The -mcpu=v3 and -mcpu=v2 run
are also successful.
Signed-off-by: Yonghong Song <yonghong.song@linux.dev>
Link: https://lore.kernel.org/r/20230728011250.3718252-1-yonghong.song@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Add asm support for new instructions so kernel verifier and bpftool
xlated insn dumps can have proper asm syntax for new instructions.
Acked-by: Eduard Zingerman <eddyz87@gmail.com>
Acked-by: Quentin Monnet <quentin@isovalent.com>
Signed-off-by: Yonghong Song <yonghong.song@linux.dev>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Currently, if user accesses a ctx member with signed types,
the compiler will generate an unsigned load followed by
necessary left and right shifts.
With the introduction of sign-extension load, compiler may
just emit a ldsx insn instead. Let us do a final movsx sign
extension to the final unsigned ctx load result to
satisfy original sign extension requirement.
Acked-by: Eduard Zingerman <eddyz87@gmail.com>
Signed-off-by: Yonghong Song <yonghong.song@linux.dev>
Link: https://lore.kernel.org/r/20230728011207.3712528-1-yonghong.song@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Add interpreter/jit support for new sign-extension mov insns.
The original 'MOV' insn is extended to support reg-to-reg
signed version for both ALU and ALU64 operations. For ALU mode,
the insn->off value of 8 or 16 indicates sign-extension
from 8- or 16-bit value to 32-bit value. For ALU64 mode,
the insn->off value of 8/16/32 indicates sign-extension
from 8-, 16- or 32-bit value to 64-bit value.
Acked-by: Eduard Zingerman <eddyz87@gmail.com>
Signed-off-by: Yonghong Song <yonghong.song@linux.dev>
Link: https://lore.kernel.org/r/20230728011202.3712300-1-yonghong.song@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Add interpreter/jit support for new sign-extension load insns
which adds a new mode (BPF_MEMSX).
Also add verifier support to recognize these insns and to
do proper verification with new insns. In verifier, besides
to deduce proper bounds for the dst_reg, probed memory access
is also properly handled.
Acked-by: Eduard Zingerman <eddyz87@gmail.com>
Signed-off-by: Yonghong Song <yonghong.song@linux.dev>
Link: https://lore.kernel.org/r/20230728011156.3711870-1-yonghong.song@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Splitting these out into separate helper functions means that we
actually pass an uninitialized variable into another function call
if dec_active() happens to not be inlined, and CONFIG_PREEMPT_RT
is disabled:
kernel/bpf/memalloc.c: In function 'add_obj_to_free_list':
kernel/bpf/memalloc.c:200:9: error: 'flags' is used uninitialized [-Werror=uninitialized]
200 | dec_active(c, flags);
Avoid this by passing the flags by reference, so they either get
initialized and dereferenced through a pointer, or the pointer never
gets accessed at all.
Fixes: 18e027b1c7 ("bpf: Factor out inc/dec of active flag into helpers.")
Suggested-by: Alexei Starovoitov <alexei.starovoitov@gmail.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20230725202653.2905259-1-arnd@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Lorenz Bauer says:
====================
We want to replace iptables TPROXY with a BPF program at TC ingress.
To make this work in all cases we need to assign a SO_REUSEPORT socket
to an skb, which is currently prohibited. This series adds support for
such sockets to bpf_sk_assing.
I did some refactoring to cut down on the amount of duplicate code. The
key to this is to use INDIRECT_CALL in the reuseport helpers. To show
that this approach is not just beneficial to TC sk_assign I removed
duplicate code for bpf_sk_lookup as well.
Joint work with Daniel Borkmann.
Signed-off-by: Lorenz Bauer <lmb@isovalent.com>
---
Changes in v6:
- Reject unhashed UDP sockets in bpf_sk_assign to avoid ref leak
- Link to v5: https://lore.kernel.org/r/20230613-so-reuseport-v5-0-f6686a0dbce0@isovalent.com
Changes in v5:
- Drop reuse_sk == sk check in inet[6]_steal_stock (Kuniyuki)
- Link to v4: https://lore.kernel.org/r/20230613-so-reuseport-v4-0-4ece76708bba@isovalent.com
Changes in v4:
- WARN_ON_ONCE if reuseport socket is refcounted (Kuniyuki)
- Use inet[6]_ehashfn_t to shorten function declarations (Kuniyuki)
- Shuffle documentation patch around (Kuniyuki)
- Update commit message to explain why IPv6 needs EXPORT_SYMBOL
- Link to v3: https://lore.kernel.org/r/20230613-so-reuseport-v3-0-907b4cbb7b99@isovalent.com
Changes in v3:
- Fix warning re udp_ehashfn and udp6_ehashfn (Simon)
- Return higher scoring connected UDP reuseport sockets (Kuniyuki)
- Fix ipv6 module builds
- Link to v2: https://lore.kernel.org/r/20230613-so-reuseport-v2-0-b7c69a342613@isovalent.com
Changes in v2:
- Correct commit abbrev length (Kuniyuki)
- Reduce duplication (Kuniyuki)
- Add checks on sk_state (Martin)
- Split exporting inet[6]_lookup_reuseport into separate patch (Eric)
---
Daniel Borkmann (1):
selftests/bpf: Test that SO_REUSEPORT can be used with sk_assign helper
====================
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
We use two programs to check that the new reuseport logic is executed
appropriately.
The first is a TC clsact program which bpf_sk_assigns
the skb to a UDP or TCP socket created by user space. Since the test
communicates via lo we see both directions of packets in the eBPF.
Traffic ingressing to the reuseport socket is identified by looking
at the destination port. For TCP, we additionally need to make sure
that we only assign the initial SYN packets towards our listening
socket. The network stack then creates a request socket which
transitions to ESTABLISHED after the 3WHS.
The second is a reuseport program which shares the fact that
it has been executed with user space. This tells us that the delayed
lookup mechanism is working.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Co-developed-by: Lorenz Bauer <lmb@isovalent.com>
Signed-off-by: Lorenz Bauer <lmb@isovalent.com>
Cc: Joe Stringer <joe@cilium.io>
Link: https://lore.kernel.org/r/20230720-so-reuseport-v6-8-7021b683cdae@isovalent.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Currently the bpf_sk_assign helper in tc BPF context refuses SO_REUSEPORT
sockets. This means we can't use the helper to steer traffic to Envoy,
which configures SO_REUSEPORT on its sockets. In turn, we're blocked
from removing TPROXY from our setup.
The reason that bpf_sk_assign refuses such sockets is that the
bpf_sk_lookup helpers don't execute SK_REUSEPORT programs. Instead,
one of the reuseport sockets is selected by hash. This could cause
dispatch to the "wrong" socket:
sk = bpf_sk_lookup_tcp(...) // select SO_REUSEPORT by hash
bpf_sk_assign(skb, sk) // SK_REUSEPORT wasn't executed
Fixing this isn't as simple as invoking SK_REUSEPORT from the lookup
helpers unfortunately. In the tc context, L2 headers are at the start
of the skb, while SK_REUSEPORT expects L3 headers instead.
Instead, we execute the SK_REUSEPORT program when the assigned socket
is pulled out of the skb, further up the stack. This creates some
trickiness with regards to refcounting as bpf_sk_assign will put both
refcounted and RCU freed sockets in skb->sk. reuseport sockets are RCU
freed. We can infer that the sk_assigned socket is RCU freed if the
reuseport lookup succeeds, but convincing yourself of this fact isn't
straight forward. Therefore we defensively check refcounting on the
sk_assign sock even though it's probably not required in practice.
Fixes: 8e368dc72e ("bpf: Fix use of sk->sk_reuseport from sk_assign")
Fixes: cf7fbe660f ("bpf: Add socket assign support")
Co-developed-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Cc: Joe Stringer <joe@cilium.io>
Link: https://lore.kernel.org/bpf/CACAyw98+qycmpQzKupquhkxbvWK4OFyDuuLMBNROnfWMZxUWeA@mail.gmail.com/
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: Lorenz Bauer <lmb@isovalent.com>
Link: https://lore.kernel.org/r/20230720-so-reuseport-v6-7-7021b683cdae@isovalent.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
The current implementation was extracted from inet[6]_lhash2_lookup
in commit 80b373f74f ("inet: Extract helper for selecting socket
from reuseport group") and commit 5df6531292 ("inet6: Extract helper
for selecting socket from reuseport group"). In the original context,
sk is always in TCP_LISTEN state and so did not have a separate check.
Add documentation that specifies which sk_state are valid to pass to
the function.
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: Lorenz Bauer <lmb@isovalent.com>
Link: https://lore.kernel.org/r/20230720-so-reuseport-v6-5-7021b683cdae@isovalent.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
There are currently four copies of reuseport_lookup: one each for
(TCP, UDP)x(IPv4, IPv6). This forces us to duplicate all callers of
those functions as well. This is already the case for sk_lookup
helpers (inet,inet6,udp4,udp6)_lookup_run_bpf.
There are two differences between the reuseport_lookup helpers:
1. They call different hash functions depending on protocol
2. UDP reuseport_lookup checks that sk_state != TCP_ESTABLISHED
Move the check for sk_state into the caller and use the INDIRECT_CALL
infrastructure to cut down the helpers to one per IP version.
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: Lorenz Bauer <lmb@isovalent.com>
Link: https://lore.kernel.org/r/20230720-so-reuseport-v6-4-7021b683cdae@isovalent.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
The semantics for bpf_sk_assign are as follows:
sk = some_lookup_func()
bpf_sk_assign(skb, sk)
bpf_sk_release(sk)
That is, the sk is not consumed by bpf_sk_assign. The function
therefore needs to make sure that sk lives long enough to be
consumed from __inet_lookup_skb. The path through the stack for a
TCPv4 packet is roughly:
netif_receive_skb_core: takes RCU read lock
__netif_receive_skb_core:
sch_handle_ingress:
tcf_classify:
bpf_sk_assign()
deliver_ptype_list_skb:
deliver_skb:
ip_packet_type->func == ip_rcv:
ip_rcv_core:
ip_rcv_finish_core:
dst_input:
ip_local_deliver:
ip_local_deliver_finish:
ip_protocol_deliver_rcu:
tcp_v4_rcv:
__inet_lookup_skb:
skb_steal_sock
The existing helper takes advantage of the fact that everything
happens in the same RCU critical section: for sockets with
SOCK_RCU_FREE set bpf_sk_assign never takes a reference.
skb_steal_sock then checks SOCK_RCU_FREE again and does sock_put
if necessary.
This approach assumes that SOCK_RCU_FREE is never set on a sk
between bpf_sk_assign and skb_steal_sock, but this invariant is
violated by unhashed UDP sockets. A new UDP socket is created
in TCP_CLOSE state but without SOCK_RCU_FREE set. That flag is only
added in udp_lib_get_port() which happens when a socket is bound.
When bpf_sk_assign was added it wasn't possible to access unhashed
UDP sockets from BPF, so this wasn't a problem. This changed
in commit 0c48eefae7 ("sock_map: Lift socket state restriction
for datagram sockets"), but the helper wasn't adjusted accordingly.
The following sequence of events will therefore lead to a refcount
leak:
1. Add socket(AF_INET, SOCK_DGRAM) to a sockmap.
2. Pull socket out of sockmap and bpf_sk_assign it. Since
SOCK_RCU_FREE is not set we increment the refcount.
3. bind() or connect() the socket, setting SOCK_RCU_FREE.
4. skb_steal_sock will now set refcounted = false due to
SOCK_RCU_FREE.
5. tcp_v4_rcv() skips sock_put().
Fix the problem by rejecting unhashed sockets in bpf_sk_assign().
This matches the behaviour of __inet_lookup_skb which is ultimately
the goal of bpf_sk_assign().
Fixes: cf7fbe660f ("bpf: Add socket assign support")
Cc: Joe Stringer <joe@cilium.io>
Signed-off-by: Lorenz Bauer <lmb@isovalent.com>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Link: https://lore.kernel.org/r/20230720-so-reuseport-v6-2-7021b683cdae@isovalent.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Contrary to TCP, UDP reuseport groups can contain TCP_ESTABLISHED
sockets. To support these properly we remember whether a group has
a connected socket and skip the fast reuseport early-return. In
effect we continue scoring all reuseport sockets and then choose the
one with the highest score.
The current code fails to re-calculate the score for the result of
lookup_reuseport. According to Kuniyuki Iwashima:
1) SO_INCOMING_CPU is set
-> selected sk might have +1 score
2) BPF prog returns ESTABLISHED and/or SO_INCOMING_CPU sk
-> selected sk will have more than 8
Using the old score could trigger more lookups depending on the
order that sockets are created.
sk -> sk (SO_INCOMING_CPU) -> sk (ESTABLISHED)
| |
`-> select the next SO_INCOMING_CPU sk
|
`-> select itself (We should save this lookup)
Fixes: efc6b6f6c3 ("udp: Improve load balancing for SO_REUSEPORT.")
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: Lorenz Bauer <lmb@isovalent.com>
Link: https://lore.kernel.org/r/20230720-so-reuseport-v6-1-7021b683cdae@isovalent.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Shannon Nelson says:
====================
ionic: add FLR support
Add support for handing and recovering from a PCI FLR event.
This patchset first moves some code around to make it usable
from multiple paths, then adds the PCI error handler callbacks
for reset_prepare and reset_done.
Example test:
echo 1 > /sys/bus/pci/devices/0000:2a:00.0/reset
v4:
- don't remove ionic_dev_teardown() in ionic_probe() in patch 2/4
- remove clear_bit() change from patch 3/4
v3:
Link: https://lore.kernel.org/netdev/20230717170001.30539-1-shannon.nelson@amd.com/
- removed first patch, it is already merged into net
v2:
Link: https://lore.kernel.org/netdev/20230713192936.45152-1-shannon.nelson@amd.com/
- removed redundant pci_save/restore_state() calls
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Add support for the PCI reset handlers in order to manage an FLR event.
Signed-off-by: Shannon Nelson <shannon.nelson@amd.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
