[Devel] [PATCH hci-8.1 v4 0/3] selftests/bpf: run the xsk veth tests in a network namespace of their own

[Devel] [PATCH hci-8.1 v4 0/3] selftests/bpf: run the xsk veth tests in a network namespace of their own

[Konstantin Khorenko]

Patchset for task https://virtuozzo.atlassian.net/browse/VSTOR-150731

test_xsk.sh makes its veth pair in the host's namespace. The host's own
traffic on the pair reaches the AF_XDP sockets under test, so the tests
fail on a configured host and pass on a bare one. The series moves the
veth part into a namespace of its own, after two fixes it relies on:

- Patch 1: test_exit() exits 1 for a skip, so the runner counts every
  "skip all tests" as a failure.
- Patch 2: xskxceiver takes the fragment limit from the max_skb_frags
  sysctl, which AF_XDP does not check and which exists in the initial
  namespace only. Take MAX_SKB_FRAGS from the kernel's BTF.
- Patch 3: the namespace itself.

Assisted-by: LLM
Signed-off-by: Konstantin Khorenko <khorenko@virtuozzo.com>
---
Changes in v4:
- Split v3 into three patches. The test_exit() fix gets a Fixes: tag;
  mainline gets the same fix (VSTOR-151679), and a commit of its own
  here can later be replaced by the ms/ one.
- Patch 3: drop "mount --make-rprivate /", unshare --mount already
  makes the new mount namespace private.
- Messages: the sysctl fallback serves kernels without BTF only, it
  cannot work inside the namespace; the "as before" claim is gone.
- Link to v3: http://lore.virtuozzo.com/20261026012309.458048-1-eva.kurchatova@virtuozzo.com

Changes in v3:
- The namespace is unnamed, so it goes with its last process however
  the run ends, Ctrl-C, a timeout or a kill, and no outer run is left
  to delete host interfaces.
- Obtain max_skb_frags from the kernel BTF, removing -f option.
- Fix test_exit() exit status for skipped runs.
- With -d, the namespace is kept.
- Rebased.
- Link to v2: http://lore.virtuozzo.com/kernel/20261022024816.2435565-1-eva.kurchatova@virtuozzo.com/T/#t

Changes in v2:
- -f is checked as the sysctl is, 1 to 45 and digits only, so a value
  that would wrap the fragment count or the packet array is refused.
- -f goes into the namespace only when the host's sysctl file could be
  read; a bare -f made getopts complain.
- No network namespace skips the tests, as no veth support does, rather
  than failing them.
- The -f description says the option tells the tests the kernel's limit
  and changes nothing in the kernel.
- The -d hint names the namespace in a usable form.
- test_exit() exits 1 for a skip as it does for the veth one; as the
  script has it.
- Link to v1: http://lore.virtuozzo.com/kernel/20261020041218.1659392-1-eva.kurchatova@virtuozzo.com/T/#t

---
Eva Kurchatova (3):
      selftests/bpf: xsk: exit with the status test_exit() is given
      selftests/bpf: xsk: take the fragment limit from the kernel's BTF
      selftests/bpf: run the xsk veth tests in a network namespace of their own

 tools/testing/selftests/bpf/test_xsk.sh    | 30 +++++++++++++++++++++--
 tools/testing/selftests/bpf/xsk_prereqs.sh |  2 +-
 tools/testing/selftests/bpf/xskxceiver.c   | 39 +++++++++++++++++++++++++++++-
 3 files changed, 67 insertions(+), 4 deletions(-)
---
base-commit: 02401c5e378f282fbf11bdc65996253ff65dad9b
change-id: 20261027-b4-prep-khorenko-vstor-150731-xsk-veth-netns-d97e81caa801

Best regards,
--  
Konstantin Khorenko,
Virtuozzo Linux Kernel Team

_______________________________________________
Devel mailing list
Devel@openvz.org
https://lists.openvz.org/mailman/listinfo/devel

[Devel] [PATCH hci-8.1 v4 1/3] selftests/bpf: xsk: exit with the status test_exit() is given

[Konstantin Khorenko]

From: Eva Kurchatova <eva.kurchatova@virtuozzo.com>

test_exit() prints the status it is given, then exits 1 whatever it
was. Every "skip all tests" of test_xsk.sh, no veth support or no ip
utility, reaches the kselftest runner as a failure.

Exit with the given status. The callers pass $ksft_fail or $ksft_skip,
so a failure still exits 1.

Fixes: 895b62eed2ab ("selftests: xsk: fix reporting of failed tests")
https://virtuozzo.atlassian.net/browse/VSTOR-150731

Feature: fix selftests
Signed-off-by: Eva Kurchatova <eva.kurchatova@virtuozzo.com>
Reviewed-by: Konstantin Khorenko <khorenko@virtuozzo.com>
---
 tools/testing/selftests/bpf/xsk_prereqs.sh | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/testing/selftests/bpf/xsk_prereqs.sh b/tools/testing/selftests/bpf/xsk_prereqs.sh
index 47c7b8064f38b..12176d85cd6cf 100755
--- a/tools/testing/selftests/bpf/xsk_prereqs.sh
+++ b/tools/testing/selftests/bpf/xsk_prereqs.sh
@@ -50,7 +50,7 @@ test_exit()
 	if [ $1 -ne 0 ]; then
 		test_status $1 $(basename $0)
 	fi
-	exit 1
+	exit $1
 }
 
 cleanup_iface()

-- 
2.43.0

_______________________________________________
Devel mailing list
Devel@openvz.org
https://lists.openvz.org/mailman/listinfo/devel

[Devel] [PATCH hci-8.1 v4 2/3] selftests/bpf: xsk: take the fragment limit from the kernel's BTF

[Konstantin Khorenko]

From: Eva Kurchatova <eva.kurchatova@virtuozzo.com>

too_many_frags sends a packet with as many fragments as the kernel
takes, then one that goes past the limit, and expects only the first to
arrive. xskxceiver reads the limit from the max_skb_frags sysctl, which
has two problems:

- AF_XDP does not use the sysctl: xsk_build_skb() checks the
  compile-time MAX_SKB_FRAGS. The sysctl only equals it by default, so
  on a host where it is lowered the test expects the wrong limit.
- The sysctl exists in the initial network namespace only. A run in any
  other namespace finds no file and falls back to 17, wrong for a kernel
  built with another CONFIG_MAX_SKB_FRAGS.

Take the limit from the kernel's BTF as the element count of
skb_shared_info's frags array: it is MAX_SKB_FRAGS itself and readable
from any namespace. The sysctl stays as a fallback for a kernel without
BTF.

https://virtuozzo.atlassian.net/browse/VSTOR-150731

Feature: fix selftests
Signed-off-by: Eva Kurchatova <eva.kurchatova@virtuozzo.com>
Reviewed-by: Konstantin Khorenko <khorenko@virtuozzo.com>
---
 tools/testing/selftests/bpf/xskxceiver.c | 39 +++++++++++++++++++++++++++++++-
 1 file changed, 38 insertions(+), 1 deletion(-)

diff --git a/tools/testing/selftests/bpf/xskxceiver.c b/tools/testing/selftests/bpf/xskxceiver.c
index 11f047b8af752..a4a0b4e2516ca 100644
--- a/tools/testing/selftests/bpf/xskxceiver.c
+++ b/tools/testing/selftests/bpf/xskxceiver.c
@@ -103,6 +103,7 @@
 #include "xsk.h"
 #include "xskxceiver.h"
 #include <bpf/bpf.h>
+#include <bpf/btf.h>
 #include <linux/filter.h>
 #include "../kselftest.h"
 #include "xsk_xdp_common.h"
@@ -325,12 +326,48 @@ static bool ifobj_zc_avail(struct ifobject *ifobject)
 	return zc_avail;
 }
 
+/* AF_XDP checks MAX_SKB_FRAGS, which the sysctl only follows by default */
+static unsigned int btf_max_skb_frags(void)
+{
+	unsigned int max_skb_frags = 0;
+	const struct btf_member *m;
+	const struct btf_type *t;
+	struct btf *btf;
+	int id, i;
+
+	btf = btf__load_vmlinux_btf();
+	if (!btf)
+		return 0;
+
+	id = btf__find_by_name_kind(btf, "skb_shared_info", BTF_KIND_STRUCT);
+	if (id < 0)
+		goto out;
+
+	t = btf__type_by_id(btf, id);
+	m = btf_members(t);
+	for (i = 0; i < btf_vlen(t); i++, m++) {
+		if (strcmp(btf__name_by_offset(btf, m->name_off), "frags"))
+			continue;
+		t = btf__type_by_id(btf, m->type);
+		if (btf_is_array(t))
+			max_skb_frags = btf_array(t)->nelems;
+		break;
+	}
+out:
+	btf__free(btf);
+	return max_skb_frags;
+}
+
 #define MAX_SKB_FRAGS_PATH "/proc/sys/net/core/max_skb_frags"
 static unsigned int get_max_skb_frags(void)
 {
-	unsigned int max_skb_frags = 0;
+	unsigned int max_skb_frags = btf_max_skb_frags();
 	FILE *file;
 
+	if (max_skb_frags)
+		return max_skb_frags;
+
+	ksft_print_msg("No MAX_SKB_FRAGS in the kernel's BTF, reading %s\n", MAX_SKB_FRAGS_PATH);
 	file = fopen(MAX_SKB_FRAGS_PATH, "r");
 	if (!file) {
 		ksft_print_msg("Error opening %s\n", MAX_SKB_FRAGS_PATH);

-- 
2.43.0

_______________________________________________
Devel mailing list
Devel@openvz.org
https://lists.openvz.org/mailman/listinfo/devel

[Devel] [PATCH hci-8.1 v4 3/3] selftests/bpf: run the xsk veth tests in a network namespace of their own

[Konstantin Khorenko]

From: Eva Kurchatova <eva.kurchatova@virtuozzo.com>

test_xsk.sh makes its veth pair in the namespace it runs in, and the
host treats it as any new interface: address configuration, neighbour
discovery and the like happen on that pair. The AF_XDP sockets under
test receive the host's packets among their own. The tests count and
inspect what they receive, so they fail on a configured host and pass
on a bare one.

Re-run the veth part in an unnamed network namespace with a sysfs of its
own, which the busy poll setup writes to. The namespace goes away with
its last process however the run ends, and takes the pair with it.

With -d the pair has to outlive the run, so the namespace gets a name,
and the script prints how to enter and delete it.

If no network namespace can be created, the tests are skipped, as
without veth or ip. A failure after that, mounting sysfs in the
namespace included, fails the run.

https://virtuozzo.atlassian.net/browse/VSTOR-150731

Feature: fix selftests
Signed-off-by: Eva Kurchatova <eva.kurchatova@virtuozzo.com>
Signed-off-by: Konstantin Khorenko <khorenko@virtuozzo.com>
---
 tools/testing/selftests/bpf/test_xsk.sh | 30 ++++++++++++++++++++++++++++--
 1 file changed, 28 insertions(+), 2 deletions(-)

diff --git a/tools/testing/selftests/bpf/test_xsk.sh b/tools/testing/selftests/bpf/test_xsk.sh
index 65aafe0003db0..1554d1ec02b12 100755
--- a/tools/testing/selftests/bpf/test_xsk.sh
+++ b/tools/testing/selftests/bpf/test_xsk.sh
@@ -68,7 +68,8 @@
 # Run with verbose output:
 #   sudo ./test_xsk.sh -v
 #
-# Set up veth interfaces and leave them up so xskxceiver can be launched in a debugger:
+# Set up veth interfaces in a network namespace and leave them up so
+# xskxceiver can be launched in a debugger there:
 #   sudo ./test_xsk.sh -d
 #
 # Run test suite for physical device in loopback mode
@@ -162,8 +163,33 @@ if [ ! -z $ETH ]; then
 	VETH1=${ETH}
 else
 	validate_root_exec
-	validate_veth_support ${VETH0}
 	validate_ip_utility
+	# The host configures and talks on any new interface in its namespace
+	if [ -z "$XSK_NETNS" ]; then
+		if ! unshare --net true; then
+			echo "skip all tests: cannot create a network namespace" >&2
+			test_exit $ksft_skip
+		fi
+		if [[ $debug -ne 1 ]]; then
+			# Unnamed, it goes with its last process; its own sysfs shows its interfaces
+			XSK_NETNS=1 exec unshare --net --mount sh -c \
+				'mount -t sysfs sysfs /sys && exec bash "$0" "$@"' \
+				"$0" "$@"
+		fi
+		# The pair of a run by hand outlives it, in a namespace with a name
+		XSK_NETNS=xsk${VETH0_POSTFIX}
+		ip netns add ${XSK_NETNS} || exit 1
+		trap "ip netns del ${XSK_NETNS}" EXIT
+		# This run's veth names were never created; the pair is in the namespace
+		trap "exit 1" INT
+		XSK_NETNS=${XSK_NETNS} ip netns exec ${XSK_NETNS} bash "$0" "$@" || exit
+		trap - EXIT
+		echo "The pair is up in network namespace ${XSK_NETNS}," \
+			"use ip netns exec ${XSK_NETNS};" \
+			"ip netns del ${XSK_NETNS} removes it"
+		exit 0
+	fi
+	validate_veth_support ${VETH0}
 	setup_vethPairs
 
 	retval=$?

-- 
2.43.0

_______________________________________________
Devel mailing list
Devel@openvz.org
https://lists.openvz.org/mailman/listinfo/devel

Re: [Devel] [PATCH hci-8.1 v4 0/3] selftests/bpf: run the xsk veth tests in a network namespace of their own

[Eva Kurchatova (Virtuozzo)]

On 10/27/26 20:15, Konstantin Khorenko wrote:
> Patchset for task https://virtuozzo.atlassian.net/browse/VSTOR-150731
>
> test_xsk.sh makes its veth pair in the host's namespace. The host's own
> traffic on the pair reaches the AF_XDP sockets under test, so the tests
> fail on a configured host and pass on a bare one. The series moves the
> veth part into a namespace of its own, after two fixes it relies on:
>
> - Patch 1: test_exit() exits 1 for a skip, so the runner counts every
>    "skip all tests" as a failure.
> - Patch 2: xskxceiver takes the fragment limit from the max_skb_frags
>    sysctl, which AF_XDP does not check and which exists in the initial
>    namespace only. Take MAX_SKB_FRAGS from the kernel's BTF.
> - Patch 3: the namespace itself.
>
> Assisted-by: LLM
> Signed-off-by: Konstantin Khorenko <khorenko@virtuozzo.com>
> ---
> Changes in v4:
> - Split v3 into three patches. The test_exit() fix gets a Fixes: tag;
>    mainline gets the same fix (VSTOR-151679), and a commit of its own
>    here can later be replaced by the ms/ one.
> - Patch 3: drop "mount --make-rprivate /", unshare --mount already
>    makes the new mount namespace private.
> - Messages: the sysctl fallback serves kernels without BTF only, it
>    cannot work inside the namespace; the "as before" claim is gone.
> - Link to v3: http://lore.virtuozzo.com/20261026012309.458048-1-eva.kurchatova@virtuozzo.com
>
> Changes in v3:
> - The namespace is unnamed, so it goes with its last process however
>    the run ends, Ctrl-C, a timeout or a kill, and no outer run is left
>    to delete host interfaces.
> - Obtain max_skb_frags from the kernel BTF, removing -f option.
> - Fix test_exit() exit status for skipped runs.
> - With -d, the namespace is kept.
> - Rebased.
> - Link to v2: http://lore.virtuozzo.com/kernel/20261022024816.2435565-1-eva.kurchatova@virtuozzo.com/T/#t
>
> Changes in v2:
> - -f is checked as the sysctl is, 1 to 45 and digits only, so a value
>    that would wrap the fragment count or the packet array is refused.
> - -f goes into the namespace only when the host's sysctl file could be
>    read; a bare -f made getopts complain.
> - No network namespace skips the tests, as no veth support does, rather
>    than failing them.
> - The -f description says the option tells the tests the kernel's limit
>    and changes nothing in the kernel.
> - The -d hint names the namespace in a usable form.
> - test_exit() exits 1 for a skip as it does for the veth one; as the
>    script has it.
> - Link to v1: http://lore.virtuozzo.com/kernel/20261020041218.1659392-1-eva.kurchatova@virtuozzo.com/T/#t
>
> ---
> Eva Kurchatova (3):
>        selftests/bpf: xsk: exit with the status test_exit() is given
>        selftests/bpf: xsk: take the fragment limit from the kernel's BTF
>        selftests/bpf: run the xsk veth tests in a network namespace of their own
>
>   tools/testing/selftests/bpf/test_xsk.sh    | 30 +++++++++++++++++++++--
>   tools/testing/selftests/bpf/xsk_prereqs.sh |  2 +-
>   tools/testing/selftests/bpf/xskxceiver.c   | 39 +++++++++++++++++++++++++++++-
>   3 files changed, 67 insertions(+), 4 deletions(-)
> ---
> base-commit: 02401c5e378f282fbf11bdc65996253ff65dad9b
> change-id: 20261027-b4-prep-khorenko-vstor-150731-xsk-veth-netns-d97e81caa801
>
> Best regards,
> --
> Konstantin Khorenko,
> Virtuozzo Linux Kernel Team


Thanks, I've had a look. This is a clean split of my v3 series into 3 
commits, with some wording tightened. The dropped `mount --make-rprivate 
/` is correct, as `unshare --mount` already makes the new mount 
namespace private.

This is good to merge.


Signed-off-by: Eva Kurchatova <eva.kurchatova@virtuozzo.com>

_______________________________________________
Devel mailing list
Devel@openvz.org
https://lists.openvz.org/mailman/listinfo/devel

Re: [Devel] [PATCH hci-8.1 v4 0/3] selftests/bpf: run the xsk veth tests in a network namespace of their own

[Konstantin Khorenko]

On Tue, 27 Oct 2026 19:15:18 +0100, Konstantin Khorenko wrote:
> selftests/bpf: run the xsk veth tests in a network namespace of their own
> 
> Patchset for task https://virtuozzo.atlassian.net/browse/VSTOR-150731
> 
> test_xsk.sh makes its veth pair in the host's namespace. The host's own
> traffic on the pair reaches the AF_XDP sockets under test, so the tests
> fail on a configured host and pass on a bare one. The series moves the
> veth part into a namespace of its own, after two fixes it relies on:
> 
> [...]

Applied to branch-rh10-6.12.0-211.52.1.22.x.vz10-ovz, will be released in rh10-6.12.0-211.52.1.22.12.vz10, thanks!

[1/3] selftests/bpf: xsk: exit with the status test_exit() is given
      https://bitbucket.org/virtuozzocore/vzkernel.vzs/commits/db5e04c0d837
[2/3] selftests/bpf: xsk: take the fragment limit from the kernel's BTF
      https://bitbucket.org/virtuozzocore/vzkernel.vzs/commits/af464697646e
[3/3] selftests/bpf: run the xsk veth tests in a network namespace of their own
      https://bitbucket.org/virtuozzocore/vzkernel.vzs/commits/d9f6317f6c14

-- 
Konstantin Khorenko,
Virtuozzo Linux Kernel Team


_______________________________________________
Devel mailing list
Devel@openvz.org
https://lists.openvz.org/mailman/listinfo/devel