linux.git/tools/testing/selftests/drivers/net/hw, branch master Linux kernel source tree selftests: drv-net: hide the devlink port_split test 2026-08-12T01:31:19+00:00 Jakub Kicinski kuba@kernel.org 2026-08-11T00:46:45+00:00 878b56de01e255172745775cd8afcec2bd80268a The devlink port_split test has limited applicability. NICs (as opposed to switches) require at least a re-probe to apply the split configuration. On top of that the test is not compatible with our driver env, it just splits all ports on the system, not only what NETIF points at. Long term we may want to add some indication in devlink whether the port splitting is runtime (cmode of sorts), and fix the test to follow driver env. But since no (known) NIC driver can support runtime anyway let's just hide the test from the selftest framework by moving it to extra files. Having this test randomly break unrelated NICs within the DUT makes people implement allow-lists for ksft, which then means their setups don't run new tests. It's very useful during test review to see whether the test works across all the runners. Reviewed-by: Petr Machata <petrm@nvidia.com> Link: https://patch.msgid.link/20260811004645.1072124-1-kuba@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org> 
The devlink port_split test has limited applicability.
NICs (as opposed to switches) require at least a re-probe
to apply the split configuration.

On top of that the test is not compatible with our driver env,
it just splits all ports on the system, not only what NETIF
points at.

Long term we may want to add some indication in devlink whether
the port splitting is runtime (cmode of sorts), and fix the
test to follow driver env. But since no (known) NIC driver can
support runtime anyway let's just hide the test from the selftest
framework by moving it to extra files.

Having this test randomly break unrelated NICs within the DUT
makes people implement allow-lists for ksft, which then means
their setups don't run new tests. It's very useful during test
review to see whether the test works across all the runners.

Reviewed-by: Petr Machata <petrm@nvidia.com>
Link: https://patch.msgid.link/20260811004645.1072124-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
selftests: drv-net: let ethtool stats settle before reading 2026-08-10T23:47:49+00:00 Jakub Kicinski kuba@kernel.org 2026-08-08T16:36:53+00:00 855631afe0f5a1fac96abdd8a30e3ccda966e15f Some devices refresh the statistics exposed via ethtool only periodically, every stats-block-usecs (as reported by ethtool -c). ethtool_std_stats and ethtool_rmon sample the counters immediately after generating traffic, so on such devices they can read stale values and fail with a delta short of the packets just sent. Add a hw_stats_settle() helper which sleeps for 1.25x the configured stats-block-usecs (defaulting to 20ms when the device reports no, or a zero, period). Use it for ethtool std stats and RMON. The 1.25x/20msec heuristic matches what the Python tests do. Reviewed-by: Petr Machata <petrm@nvidia.com> Link: https://patch.msgid.link/20260808163653.2460381-2-kuba@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org> 
Some devices refresh the statistics exposed via ethtool only
periodically, every stats-block-usecs (as reported by ethtool -c).
ethtool_std_stats and ethtool_rmon sample the counters immediately
after generating traffic, so on such devices they can read stale
values and fail with a delta short of the packets just sent.

Add a hw_stats_settle() helper which sleeps for 1.25x the configured
stats-block-usecs (defaulting to 20ms when the device reports no, or
a zero, period). Use it for ethtool std stats and RMON.
The 1.25x/20msec heuristic matches what the Python tests do.

Reviewed-by: Petr Machata <petrm@nvidia.com>
Link: https://patch.msgid.link/20260808163653.2460381-2-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
selftests: drv-net: pace ethtool_std_stats packet generation 2026-08-10T23:47:38+00:00 Jakub Kicinski kuba@kernel.org 2026-08-08T16:36:52+00:00 f4cf60568747edddb255106154684f8d56b1aa1b mausezahn defaults to sending packets back to back at the maximum rate, which can cause packet loss, especially if receiver is running a debug kernel. Space the generated packets out (-d 10usec), like ethtool_rmon already does. Reviewed-by: Petr Machata <petrm@nvidia.com> Link: https://patch.msgid.link/20260808163653.2460381-1-kuba@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org> 
mausezahn defaults to sending packets back to back at the maximum rate,
which can cause packet loss, especially if receiver is running a debug
kernel. Space the generated packets out (-d 10usec), like ethtool_rmon
already does.

Reviewed-by: Petr Machata <petrm@nvidia.com>
Link: https://patch.msgid.link/20260808163653.2460381-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
selftests/net: devmem.py: add check_rx_large_niov 2026-08-08T01:32:40+00:00 Bobby Eshleman bobbyeshleman@meta.com 2026-08-05T20:42:45+00:00 8ac4255c1e0c83d2e1559a18b8918673116fc8d6 Add a new devmem test case for binding the dmabuf with rx-page-size=16K. The test sweeps RX payload sizes straddling the niov boundary to cover the sub-niov, exact-niov, and multi-niov RX paths. Silence pylint invalid-name (`with open() as f`) and too-many-arguments (ncdevmem_rx grew to 6 args) at file scope. Acked-by: Stanislav Fomichev <sdf@fomichev.me> Reviewed-by: Nikolay Aleksandrov <razor@blackwall.org> Signed-off-by: Bobby Eshleman <bobbyeshleman@meta.com> Link: https://patch.msgid.link/20260805-tcpdm-large-niovs-v8-3-3e0225e2808c@meta.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> 
Add a new devmem test case for binding the dmabuf with rx-page-size=16K.
The test sweeps RX payload sizes straddling the niov boundary to cover
the sub-niov, exact-niov, and multi-niov RX paths.

Silence pylint invalid-name (`with open() as f`) and too-many-arguments
(ncdevmem_rx grew to 6 args) at file scope.

Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Reviewed-by: Nikolay Aleksandrov <razor@blackwall.org>
Signed-off-by: Bobby Eshleman <bobbyeshleman@meta.com>
Link: https://patch.msgid.link/20260805-tcpdm-large-niovs-v8-3-3e0225e2808c@meta.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
selftests/net: ncdevmem: add -b option to set rx-page-size on bind 2026-08-08T01:32:39+00:00 Bobby Eshleman bobbyeshleman@meta.com 2026-08-05T20:42:44+00:00 3e8c9ec4eb75b60cd03bcebb275553df09417e2e Add -b <bytes> to request a non-default niov size via NETDEV_A_DMABUF_RX_PAGE_SIZE. When the value exceeds PAGE_SIZE, udmabuf_alloc() switches to an MFD_HUGETLB-backed memfd so each 2 MB hugepage produces one naturally-aligned sg entry. Acked-by: Stanislav Fomichev <sdf@fomichev.me> Reviewed-by: Nikolay Aleksandrov <razor@blackwall.org> Signed-off-by: Bobby Eshleman <bobbyeshleman@meta.com> Link: https://patch.msgid.link/20260805-tcpdm-large-niovs-v8-2-3e0225e2808c@meta.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> 
Add -b <bytes> to request a non-default niov size via
NETDEV_A_DMABUF_RX_PAGE_SIZE. When the value exceeds PAGE_SIZE,
udmabuf_alloc() switches to an MFD_HUGETLB-backed memfd so each 2 MB
hugepage produces one naturally-aligned sg entry.

Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Reviewed-by: Nikolay Aleksandrov <razor@blackwall.org>
Signed-off-by: Bobby Eshleman <bobbyeshleman@meta.com>
Link: https://patch.msgid.link/20260805-tcpdm-large-niovs-v8-2-3e0225e2808c@meta.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
selftests: drv-net: hw: reset HDS mode after netkit devmem tests 2026-08-05T23:52:33+00:00 Jakub Kicinski kuba@kernel.org 2026-08-04T15:10:40+00:00 481e86a82219369ad19898013b11f47aeaee7ab9 HDS mode has confusing semantics. On GET kernel reports effective mode. On SET kernel expects explicit config. Effective mode on GET means that we know the current state, but we don't know if it's a driver default or user setting. This matter because driver default can change automatically when e.g. XDP is attached. Explicit user setting must not be lost. With that in mind, we can't restore the HDS setting like we restore other NIC config. We should always reset to default ("unknown"). This fixes an issue with tests running after the devmem test not being able to attach XDP, e.g. Exception| File "./xdp_metadata.py", line 105, in test_xdp_rss_hash [...] Exception| net.lib.py.utils.CmdExitFailure: Command failed Exception| CMD: ip link set dev ens9np0 xdpdrv pinned /sys/fs/bpf/xdp_metadata_test/xdp_rss_hash Exception| EXIT: 2 Exception| STDERR: Error: unable to install XDP to device using tcp-data-split. not ok 1 xdp_metadata.test_xdp_rss_hash.tcp Reviewed-by: Simon Horman <horms@kernel.org> Reviewed-by: Breno Leitao <leitao@debian.org> Reviewed-by: Bobby Eshleman <bobbyeshleman@meta.com> Link: https://patch.msgid.link/20260804151040.2755153-1-kuba@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org> 
HDS mode has confusing semantics. On GET kernel reports
effective mode. On SET kernel expects explicit config.
Effective mode on GET means that we know the current
state, but we don't know if it's a driver default
or user setting. This matter because driver default
can change automatically when e.g. XDP is attached.
Explicit user setting must not be lost.

With that in mind, we can't restore the HDS setting
like we restore other NIC config. We should always
reset to default ("unknown").

This fixes an issue with tests running after the devmem
test not being able to attach XDP, e.g.

   Exception|   File "./xdp_metadata.py", line 105, in test_xdp_rss_hash
   [...]
   Exception| net.lib.py.utils.CmdExitFailure: Command failed
   Exception| CMD: ip link set dev ens9np0 xdpdrv pinned /sys/fs/bpf/xdp_metadata_test/xdp_rss_hash
   Exception|   EXIT: 2
   Exception|   STDERR: Error: unable to install XDP to device using tcp-data-split.
   not ok 1 xdp_metadata.test_xdp_rss_hash.tcp

Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: Breno Leitao <leitao@debian.org>
Reviewed-by: Bobby Eshleman <bobbyeshleman@meta.com>
Link: https://patch.msgid.link/20260804151040.2755153-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
selftests: drv-net: enable devmem TCP in the test config 2026-08-03T20:25:13+00:00 Bobby Eshleman bobbyeshleman@meta.com 2026-07-31T23:36:23+00:00 8c7badd19e13ceeb7ae34ea0b25fc5e4e13c43d3 The config fragment already sets CONFIG_UDMABUF=y, but kconfig silently drops it. UDMABUF/NET_DEVMEM both depend on DMA_SHARED_BUFFER, which we can't enable directly, so we need to enable a config that selects it. We use SYNC_FILE for that purpose here. Additionally, we flip on CONFIG_NET_DEVMEM as well. Suggested-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Bobby Eshleman <bobbyeshleman@meta.com> Reviewed-by: Mina Almasry <almasrymina@google.com> Link: https://patch.msgid.link/20260731-selftests-devmem-config-v1-1-098014348d9d@meta.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> 
The config fragment already sets CONFIG_UDMABUF=y, but kconfig silently
drops it. UDMABUF/NET_DEVMEM both depend on DMA_SHARED_BUFFER, which we
can't enable directly, so we need to enable a config that selects it. We
use SYNC_FILE for that purpose here.

Additionally, we flip on CONFIG_NET_DEVMEM as well.

Suggested-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Bobby Eshleman <bobbyeshleman@meta.com>
Reviewed-by: Mina Almasry <almasrymina@google.com>
Link: https://patch.msgid.link/20260731-selftests-devmem-config-v1-1-098014348d9d@meta.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
selftests: drv-net: ncdevmem: Open /dev/udmabuf O_RDONLY 2026-07-24T21:53:16+00:00 T.J. Mercier tjmercier@google.com 2026-07-22T20:54:42+00:00 be6f0d0bae229ebd03e9bfe736f78e2d6b35885f Write permissions on the /dev/udmabuf device file are not required to issue ioctls and allocate udmabufs. Applications should be opening this file as O_RDONLY. The BPF dmabuf_iter selftest already does this. [1] Users are pointing to these selftests as examples of how use udmabuf, and encountering permission errors on systems where write permissions are not available on /dev/udmabuf. Apply the principle of least privilege to selftests which use udmabuf by removing the write access mode from drivers/net/hw/ncdevmem.c. [1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/tools/testing/selftests/bpf/prog_tests/dmabuf_iter.c?h=v7.1#n49 Signed-off-by: T.J. Mercier <tjmercier@google.com> Reviewed-by: Bobby Eshleman <bobbyeshleman@meta.com> Link: https://patch.msgid.link/20260722205442.1894665-1-tjmercier@google.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> 
Write permissions on the /dev/udmabuf device file are not required to
issue ioctls and allocate udmabufs. Applications should be opening this
file as O_RDONLY. The BPF dmabuf_iter selftest already does this. [1]

Users are pointing to these selftests as examples of how use udmabuf,
and encountering permission errors on systems where write permissions
are not available on /dev/udmabuf. Apply the principle of least
privilege to selftests which use udmabuf by removing the write access
mode from drivers/net/hw/ncdevmem.c.

[1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/tools/testing/selftests/bpf/prog_tests/dmabuf_iter.c?h=v7.1#n49

Signed-off-by: T.J. Mercier <tjmercier@google.com>
Reviewed-by: Bobby Eshleman <bobbyeshleman@meta.com>
Link: https://patch.msgid.link/20260722205442.1894665-1-tjmercier@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
selftests: drv-net: Fix TSO test doc 2026-07-23T17:12:51+00:00 Petr Vorel pvorel@suse.cz 2026-07-20T21:51:49+00:00 1306cf6dc1dfd34b97ccff98fcb08168c352dbef Replace copy paste from csum.py with a real test purpose. Signed-off-by: Petr Vorel <pvorel@suse.cz> Link: https://patch.msgid.link/20260720215149.631145-2-pvorel@suse.cz Signed-off-by: Jakub Kicinski <kuba@kernel.org> 
Replace copy paste from csum.py with a real test purpose.

Signed-off-by: Petr Vorel <pvorel@suse.cz>
Link: https://patch.msgid.link/20260720215149.631145-2-pvorel@suse.cz
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
selftests: drv-net: Fix csum path in doc 2026-07-23T17:12:51+00:00 Petr Vorel petr.vorel@gmail.com 2026-07-20T21:51:48+00:00 b8363908ea1ea08c726c4804211a38dd353219fe C source was moved in 1d0dc857b5d87. Signed-off-by: Petr Vorel <pvorel@suse.cz> Link: https://patch.msgid.link/20260720215149.631145-1-pvorel@suse.cz Signed-off-by: Jakub Kicinski <kuba@kernel.org> 
C source was moved in 1d0dc857b5d87.

Signed-off-by: Petr Vorel <pvorel@suse.cz>
Link: https://patch.msgid.link/20260720215149.631145-1-pvorel@suse.cz
Signed-off-by: Jakub Kicinski <kuba@kernel.org>