Skip to content

B4/xsk #9601

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Closed
wants to merge 16 commits into from
Closed

B4/xsk #9601

wants to merge 16 commits into from

Conversation

bastien-curutchet
Copy link
Contributor

No description provided.

@bastien-curutchet bastien-curutchet force-pushed the b4/xsk branch 2 times, most recently from 002ca61 to d428179 Compare August 28, 2025 16:07
@bastien-curutchet bastien-curutchet force-pushed the b4/xsk branch 7 times, most recently from 36f8268 to b3bedae Compare September 2, 2025 12:55
@kernel-patches-daemon-bpf kernel-patches-daemon-bpf bot force-pushed the bpf-next_base branch 3 times, most recently from a9f9e99 to c20f844 Compare September 3, 2025 23:45
@bastien-curutchet
selftests/bpf: Integrate test_xsk.c to test_progs framework
a96ad0d 
Hi all,

This is a second version of a series I sent some time ago, it continues
the work of migrating the script tests into prog_tests.

The test_xsk.sh script covers many AF_XDP use cases. The tests it runs
are defined in xksxceiver.c. Since this script is used to test real
hardware, the goal here is to leave it as it is, and only integrate the
tests that run on veth peers into the test_progs framework.

Some tests are flaky so they can't be integrated in the CI as they are.
I think that fixing their flakyness would require a significant amount of
work. So, as first step, I've excluded them from the list of tests
migrated to the CI (see PATCH 13). If these tests get fixed at some
point, integrating them into the CI will be straightforward.

PATCH 1 extracts test_xsk[.c/.h] from xskxceiver[.c/.h] to make the
tests available to test_progs.
PATCH 2 to 5 fix small issues in the current test
PATCH 7 to 12 handle all errors to release resources instead of calling
exit() when any error occurs.
PATCH 13 isolates some flaky tests
PATCH 14 integrate the non-flaky tests to the test_progs framework

Maciej, I've fixed the bug you found in the initial series. I've
looked for any hardware able to run test_xsk.sh in my office, but I
couldn't find one ... So here again, only the veth part has been tested,
sorry about that.

To: Björn Töpel <bjorn@kernel.org>
To: Magnus Karlsson <magnus.karlsson@intel.com>
To: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
To: Jonathan Lemon <jonathan.lemon@gmail.com>
To: Alexei Starovoitov <ast@kernel.org>
To: Daniel Borkmann <daniel@iogearbox.net>
To: Andrii Nakryiko <andrii@kernel.org>
To: Martin KaFai Lau <martin.lau@linux.dev>
To: Eduard Zingerman <eddyz87@gmail.com>
To: Song Liu <song@kernel.org>
To: Yonghong Song <yonghong.song@linux.dev>
To: John Fastabend <john.fastabend@gmail.com>
To: KP Singh <kpsingh@kernel.org>
To: Stanislav Fomichev <sdf@fomichev.me>
To: Hao Luo <haoluo@google.com>
To: Jiri Olsa <jolsa@kernel.org>
To: Mykola Lysenko <mykolal@fb.com>
To: Shuah Khan <shuah@kernel.org>
To: David S. Miller <davem@davemloft.net>
To: Jakub Kicinski <kuba@kernel.org>
To: Jesper Dangaard Brouer <hawk@kernel.org>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Alexis Lothore <alexis.lothore@bootlin.com>
Cc: netdev@vger.kernel.org
Cc: bpf@vger.kernel.org
Cc: linux-kselftest@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Bastien Curutchet (eBPF Foundation) <bastien.curutchet@bootlin.com>

---
Changes in v3:
- Rebase on latest bpf-next_base to integrate commit c9110e6 ("selftests/bpf:
Fix count write in testapp_xdp_metadata_copy()").
- Move XDP_METADATA_COPY_* tests from flaky-tests to nominal tests
- Link to v2: https://lore.kernel.org/r/20250902-xsk-v2-0-17c6345d5215@bootlin.com

Changes in v2:
- Rebase on the latest bpf-next_base and integrate the newly added tests
  to the work (adjust_tail* and tx_queue_consumer tests)
- Re-order patches to split xkxceiver sooner.
- Fix the bug reported by Maciej.
- Fix verbose mode in test_xsk.sh by keeping kselftest (remove PATCH 1,
  7 and 8)
- Link to v1: https://lore.kernel.org/r/20250313-xsk-v1-0-7374729a93b9@bootlin.com

--- b4-submit-tracking ---
# This section is used internally by b4 prep for tracking purposes.
{
  "series": {
    "revision": 3,
    "change-id": "20250218-xsk-0cf90e975d14",
    "prefixes": [
      "bpf-next"
    ],
    "history": {
      "v1": [
        "20250313-xsk-v1-0-7374729a93b9@bootlin.com"
      ],
      "v2": [
        "20250902-xsk-v2-0-17c6345d5215@bootlin.com"
      ]
    }
  }
}
@bastien-curutchet
selftests/bpf: test_xsk: Split xskxceiver
583c025 
AF_XDP features are tested by the test_xsk.sh script but not by the
test_progs framework. The tests used by the script are defined in
xksxceiver.c which can't be integrated in the test_progs framework as is.

Extract these test definitions from xskxceiver{.c/.h} to put them in new
test_xsk{.c/.h} files.
Keep the main() function and its unshared dependencies in xksxceiver to
avoid impacting the test_xsk.sh script which is often used to test real
hardware.

Signed-off-by: Bastien Curutchet (eBPF Foundation) <bastien.curutchet@bootlin.com>
@bastien-curutchet
selftests/bpf: test_xsk: Initialize bitmap before use
ea10b12 
bitmap is used before being initialized.

Initialize it to zero before using it.

Signed-off-by: Bastien Curutchet (eBPF Foundation) <bastien.curutchet@bootlin.com>
@bastien-curutchet
selftests/bpf: test_xsk: Fix memory leaks
a204f2f 
Some tests introduce memory leaks by not freeing all the pkt_stream
objects they're creating.

Fix these memory leaks.

Signed-off-by: Bastien Curutchet (eBPF Foundation) <bastien.curutchet@bootlin.com>
@bastien-curutchet
selftests/bpf: test_xsk: Wrap test clean-up in functions
1a9e8b6 
The clean-up done at the end of a test in __testapp_validate_traffic()
isn't wrapped in a function. It isn't convenient if we want to use it
somewhere else in the code.

Wrap the clean-up in two new functions : the first deletes the sockets,
the second releases the umem.

Signed-off-by: Bastien Curutchet (eBPF Foundation) <bastien.curutchet@bootlin.com>
@bastien-curutchet
selftests/bpf: test_xsk: Release resources when swap fails
52829fd 
testapp_validate_traffic() doesn't release the sockets and the umem
created by the threads if the test isn't currently in its last step.
Thus, if the swap_xsk_resources() fails before the last step, the
created resources aren't cleaned up.

Clean the sockets and the umem in case of swap_xsk_resources() failure.

Signed-off-by: Bastien Curutchet (eBPF Foundation) <bastien.curutchet@bootlin.com>
@bastien-curutchet
selftests/bpf: test_xsk: Add return value to init_iface()
6e319fd 
init_iface() doesn't have any return value while it can fail. In case of
failure it calls exit_on_error() which exits the application
immediately. This prevents the following tests from being run and isn't
compliant with the CI

Add a return value to init_iface() so errors can be handled more
smoothly.

Signed-off-by: Bastien Curutchet (eBPF Foundation) <bastien.curutchet@bootlin.com>
@bastien-curutchet
selftests/bpf: test_xsk: Don't exit immediately when xsk_attach fails
2236e67 
xsk_reattach_xdp calls exit_with_error() on failures. This exits the
program immediately. It prevents the following tests from being run and
isn't compliant with the CI.

Add a return value to the functions handling XDP attachments to handle
errors more smoothly.

Signed-off-by: Bastien Curutchet (eBPF Foundation) <bastien.curutchet@bootlin.com>
@bastien-curutchet
selftests/bpf: test_xsk: Don't exit immediately when gettimeofday fails
dd23b2a 
exit_with_error() is called when gettimeofday() fails. This exits the
program immediately. It prevents the following tests from being run and
isn't compliant with the CI.

Return TEST_FAILURE instead of calling exit_on_error().

Signed-off-by: Bastien Curutchet (eBPF Foundation) <bastien.curutchet@bootlin.com>
@bastien-curutchet
selftests/bpf: test_xsk: Don't exit immediately when workers fail
37b355e 
TX and RX workers can fail in many places. These failures trigger a call
to exit_with_error() which exits the program immediately. It prevents the
following tests from running and isn't compliant with the CI.

Add return value to functions that can fail.
Handle failures more smoothly through report_failure().

Signed-off-by: Bastien Curutchet (eBPF Foundation) <bastien.curutchet@bootlin.com>
@bastien-curutchet
selftests/bpf: test_xsk: Don't exit immediately if validate_traffic f…
ea85a42 
…ails

__testapp_validate_traffic() calls exit_with_error() on failures. This
exits the program immediately. It prevents the following tests from
running and isn't compliant with the CI.

Return TEST_FAILURE instead of calling exit_with_error().
Release the resource of the 1st thread if a failure happens between its
creation and the creation of the second thread.

Signed-off-by: Bastien Curutchet (eBPF Foundation) <bastien.curutchet@bootlin.com>
@bastien-curutchet
selftests/bpf: test_xsk: Don't exit immediately on allocation failures
f755eee 
If any allocation in the pkt_stream_*() helpers fail, exit_with_error() is
called. This terminates the program immediately. It prevents the following
tests from running and isn't compliant with the CI.

Return NULL in case of allocation failure.
Return TEST_FAILURE when something goes wrong in the packet generation.
Clean up the resources if a failure happens between two steps of a test.

Signed-off-by: Bastien Curutchet (eBPF Foundation) <bastien.curutchet@bootlin.com>
@bastien-curutchet
selftests/bpf: test_xsk: Move exit_with_error to xskxceiver.c
d74b8cf 
exit_with_error() isn't called by test_xsk.c. It shouldn't be called
in the future either, since some resources need to be cleaned before
exiting when an error occurs.

Move the definition of exit_with_error() to xskxceiver.c, where it's
still used.

Signed-off-by: Bastien Curutchet (eBPF Foundation) <bastien.curutchet@bootlin.com>
@bastien-curutchet
selftests/bpf: test_xsk: Isolate flaky tests
af6abac 
Some tests are flaky and fail from time to time on virtual interfaces.
Adding them to the CI would trigger lots of 'false' errors.

Remove the flaky tests from the nominal tests table so they won't be
run by the CI in upcoming patch.
Create a flaky_tests table to hold them.
Use this flaky table in xskxceiver.c to keep all the tests available
from the test_xsk.sh script.

Signed-off-by: Bastien Curutchet (eBPF Foundation) <bastien.curutchet@bootlin.com>
@bastien-curutchet
selftests/bpf: test_xsk: Integrate test_xsk.c to test_progs framework
7a40724 
test_xsk.c isn't part of the test_progs framework.

Integrate the tests defined by test_xsk.c into the test_progs framework
through a new file : prog_tests/xsk.c. ZeroCopy mode isn't tested in it
as veth peers don't support it.

Move test_xsk{.c/.h} to prog_tests/.

Add the find_bit library to test_progs sources in the Makefile as it is
is used by test_xsk.c

Signed-off-by: Bastien Curutchet (eBPF Foundation) <bastien.curutchet@bootlin.com>
@kernel-patches-daemon-bpf kernel-patches-daemon-bpf bot force-pushed the bpf-next_base branch 8 times, most recently from bf64b12 to 01c3a00 Compare September 9, 2025 16:33
@kernel-patches-daemon-bpf
Copy link

Automatically cleaning up stale PR; feel free to reopen if needed

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
1 participant
@bastien-curutchet