In a follow up to these patches,
- commit 0bdf399342c5("net: Avoid address overwrite in kernel_connect")
- commit 86a7e0b69bd5("net: prevent rewrite of msg_name in sock_sendmsg()")
- commit c889a99a21bf("net: prevent address rewrite in kernel_bind()")
- commit 01b2885d9415("net: Save and restore msg_namelen in sock_sendmsg")
this patch series introduces BPF selftests that test the interaction
between BPF sockaddr hooks and socket operations in kernel space. It
focuses on regression test coverage to ensure that these operations do not
overwrite their address parameter and also provides some sanity checks
around kernel_getpeername() and kernel_getsockname().
It introduces two new components: a kernel module called sock_addr_testmod
and a new test program called sock_addr_kern which is loosely modeled after
and adapted from the old-style bpf/test_sock_addr.c selftest. When loaded,
the kernel module will perform some socket operation in kernel space. The
kernel module accepts five parameters controlling which socket operation
will be performed and its inputs:
MODULE_PARM_DESC(ip, "IPv4/IPv6/Unix address to use for socket operation");
MODULE_PARM_DESC(port, "Port number to use for socket operation");
MODULE_PARM_DESC(af, "Address family (AF_INET, AF_INET6, or AF_UNIX)");
MODULE_PARM_DESC(type, "Socket type (SOCK_STREAM or SOCK_DGRAM)");
MODULE_PARM_DESC(op, "Socket operation (BIND=0, CONNECT=1, SENDMSG=2)");
On module init, the socket operation is performed and results of are
exposed through debugfs.
- /sys/kernel/debug/sock_addr_testmod/success
Indicates success or failure of the operation.
- /sys/kernel/debug/sock_addr_testmod/addr
The value of the address parameter after the operation.
- /sys/kernel/debug/sock_addr_testmod/sock_name
The value of kernel_getsockname() after the socket operation (if relevant).
- /sys/kernel/debug/sock_addr_testmod/peer_name
The value of kernel_getpeername(() after the socket operation (if relevant).
The sock_addr_kern test program loads and unloads the kernel module to
drive kernel socket operations, reads the results from debugfs, makes sure
that the operation did not overwrite the address, and any result from
kernel_getpeername() or kernel_getsockname() were as expected.
== Patches ==
- Patch 1 introduces sock_addr_testmod and functions necessary for the test
program to load and unload the module.
- Patches 2-6 transform existing test helpers and introduce new test helpers
to enable the sock_addr_kern test program.
- Patch 7 implements the sock_addr_kern test program.
- Patch 8 fixes the sock_addr bind test program to work for big endian
architectures such as s390x.
Jordan Rife (8):
selftests/bpf: Introduce sock_addr_testmod
selftests/bpf: Add module load helpers
selftests/bpf: Factor out cmp_addr
selftests/bpf: Add recv_msg_from_client to network helpers
selftests/bpf: Factor out load_path and defines from test_sock_addr
selftests/bpf: Add setup/cleanup subcommands
selftests/bpf: Add sock_addr_kern prog_test
selftests/bpf: Fix bind program for big endian systems
tools/testing/selftests/bpf/Makefile | 46 +-
tools/testing/selftests/bpf/network_helpers.c | 65 ++
tools/testing/selftests/bpf/network_helpers.h | 5 +
.../selftests/bpf/prog_tests/sock_addr.c | 34 -
.../selftests/bpf/prog_tests/sock_addr_kern.c | 631 ++++++++++++++++++
.../testing/selftests/bpf/progs/bind4_prog.c | 18 +-
.../testing/selftests/bpf/progs/bind6_prog.c | 18 +-
tools/testing/selftests/bpf/progs/bind_prog.h | 19 +
.../testing/selftests/bpf/sock_addr_helpers.c | 46 ++
.../testing/selftests/bpf/sock_addr_helpers.h | 44 ++
.../bpf/sock_addr_testmod/.gitignore | 6 +
.../selftests/bpf/sock_addr_testmod/Makefile | 20 +
.../bpf/sock_addr_testmod/sock_addr_testmod.c | 256 +++++++
tools/testing/selftests/bpf/test_sock_addr.c | 76 +--
tools/testing/selftests/bpf/test_sock_addr.sh | 10 +-
tools/testing/selftests/bpf/testing_helpers.c | 44 +-
tools/testing/selftests/bpf/testing_helpers.h | 2 +
17 files changed, 1196 insertions(+), 144 deletions(-)
create mode 100644 tools/testing/selftests/bpf/prog_tests/sock_addr_kern.c
create mode 100644 tools/testing/selftests/bpf/progs/bind_prog.h
create mode 100644 tools/testing/selftests/bpf/sock_addr_helpers.c
create mode 100644 tools/testing/selftests/bpf/sock_addr_helpers.h
create mode 100644 tools/testing/selftests/bpf/sock_addr_testmod/.gitignore
create mode 100644 tools/testing/selftests/bpf/sock_addr_testmod/Makefile
create mode 100644 tools/testing/selftests/bpf/sock_addr_testmod/sock_addr_testmod.c
--
2.44.0.478.gd926399ef9-goog
There is a spelling mistake in a ksft_test_result_skip message. Fix it.
Signed-off-by: Colin Ian King <colin.i.king(a)gmail.com>
---
tools/testing/selftests/mm/ksm_functional_tests.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/tools/testing/selftests/mm/ksm_functional_tests.c b/tools/testing/selftests/mm/ksm_functional_tests.c
index 2d277620fad2..db845dca8d19 100644
--- a/tools/testing/selftests/mm/ksm_functional_tests.c
+++ b/tools/testing/selftests/mm/ksm_functional_tests.c
@@ -502,7 +502,7 @@ static void test_child_ksm_err(int status)
else if (status == -2)
ksft_test_result_fail("Merge in child failed\n");
else if (status == -3)
- ksft_test_result_skip("Merge in child skiped\n");
+ ksft_test_result_skip("Merge in child skipped\n");
}
/* Verify that prctl ksm flag is inherited. */
--
2.39.2
Here are two fixes related to MPTCP.
The first patch fixes when the MPTcpExtMPCapableFallbackACK MIB counter
is modified: it should only be incremented when a connection was using
MPTCP options, but then a fallback to TCP has been done. This patch also
checks the counter is not incremented by mistake during the connect
selftests. This counter was wrongly incremented since its introduction
in v5.7.
The second patch fixes a wrong parsing of the 'dev' endpoint options in
the selftests: the wrong variable was used. This option was not used
before, but it is going to be soon. This issue is visible since v5.18.
Signed-off-by: Matthieu Baerts (NGI0) <matttbe(a)kernel.org>
---
Davide Caratti (1):
mptcp: don't account accept() of non-MPC client as fallback to TCP
Geliang Tang (1):
selftests: mptcp: join: fix dev in check_endpoint
net/mptcp/protocol.c | 2 --
net/mptcp/subflow.c | 2 ++
tools/testing/selftests/net/mptcp/mptcp_connect.sh | 9 +++++++++
tools/testing/selftests/net/mptcp/mptcp_join.sh | 4 +++-
4 files changed, 14 insertions(+), 3 deletions(-)
---
base-commit: 0ba80d96585662299d4ea4624043759ce9015421
change-id: 20240329-upstream-net-20240329-fallback-mib-b0fec9c6189b
Best regards,
--
Matthieu Baerts (NGI0) <matttbe(a)kernel.org>
The netdev CI runs in a VM and captures serial, so stdout and
stderr get combined. Because there's a missing new line in
stderr the test ends up corrupting KTAP:
# Successok 1 selftests: net: reuseaddr_conflict
which should have been:
# Success
ok 1 selftests: net: reuseaddr_conflict
Fixes: 422d8dc6fd3a ("selftest: add a reuseaddr test")
Signed-off-by: Jakub Kicinski <kuba(a)kernel.org>
---
Low risk and seems worth backporting to stable, hence the fixes tag.
CC: shuah(a)kernel.org
CC: jbacik(a)fb.com
CC: linux-kselftest(a)vger.kernel.org
---
tools/testing/selftests/net/reuseaddr_conflict.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/tools/testing/selftests/net/reuseaddr_conflict.c b/tools/testing/selftests/net/reuseaddr_conflict.c
index 7c5b12664b03..bfb07dc49518 100644
--- a/tools/testing/selftests/net/reuseaddr_conflict.c
+++ b/tools/testing/selftests/net/reuseaddr_conflict.c
@@ -109,6 +109,6 @@ int main(void)
fd1 = open_port(0, 1);
if (fd1 >= 0)
error(1, 0, "Was allowed to create an ipv4 reuseport on an already bound non-reuseport socket with no ipv6");
- fprintf(stderr, "Success");
+ fprintf(stderr, "Success\n");
return 0;
}
--
2.44.0