- Linux-kselftest-mirror - lists.linaro.org

[PATCH net v2 2/2] selftests: drv-net: test XDP, HDS auto and the ioctl path

by Jakub Kicinski

Test XDP and HDS interaction. While at it add a test for using the IOCTL, as that turned out to be the real culprit. Testing bnxt: # NETIF=eth0 ./ksft-net-drv/drivers/net/hds.py KTAP version 1 1..12 ok 1 hds.get_hds ok 2 hds.get_hds_thresh ok 3 hds.set_hds_disable # SKIP disabling of HDS not supported by the device ok 4 hds.set_hds_enable ok 5 hds.set_hds_thresh_zero ok 6 hds.set_hds_thresh_max ok 7 hds.set_hds_thresh_gt ok 8 hds.set_xdp ok 9 hds.enabled_set_xdp ok 10 hds.ioctl ok 11 hds.ioctl_set_xdp ok 12 hds.ioctl_enabled_set_xdp # Totals: pass:11 fail:0 xfail:0 xpass:0 skip:1 error:0 and netdevsim: # ./ksft-net-drv/drivers/net/hds.py KTAP version 1 1..12 ok 1 hds.get_hds ok 2 hds.get_hds_thresh ok 3 hds.set_hds_disable ok 4 hds.set_hds_enable ok 5 hds.set_hds_thresh_zero ok 6 hds.set_hds_thresh_max ok 7 hds.set_hds_thresh_gt ok 8 hds.set_xdp ok 9 hds.enabled_set_xdp ok 10 hds.ioctl ok 11 hds.ioctl_set_xdp ok 12 hds.ioctl_enabled_set_xdp # Totals: pass:12 fail:0 xfail:0 xpass:0 skip:0 error:0 Netdevsim needs a sane default for tx/rx ring size. ethtool 6.11 is needed for the --disable-netlink option. Signed-off-by: Jakub Kicinski <kuba(a)kernel.org> --- Since this is targeting net there is no cfg.rpath(), I'll follow up once in net-next. v2: - add the ioctl tests - factor out some common logic CC: shuah(a)kernel.org CC: hawk(a)kernel.org CC: petrm(a)nvidia.com CC: willemb(a)google.com CC: jstancek(a)redhat.com CC: linux-kselftest(a)vger.kernel.org --- tools/testing/selftests/net/lib/Makefile | 3 + drivers/net/netdevsim/ethtool.c | 2 + .../testing/selftests/net/lib/xdp_dummy.bpf.c | 13 ++ tools/testing/selftests/drivers/net/hds.py | 145 +++++++++++++++++- 4 files changed, 160 insertions(+), 3 deletions(-) create mode 100644 tools/testing/selftests/net/lib/xdp_dummy.bpf.c diff --git a/tools/testing/selftests/net/lib/Makefile b/tools/testing/selftests/net/lib/Makefile index bc6b6762baf3..c22623b9a2a5 100644 --- a/tools/testing/selftests/net/lib/Makefile +++ b/tools/testing/selftests/net/lib/Makefile @@ -9,7 +9,10 @@ TEST_FILES := ../../../../../Documentation/netlink/specs TEST_FILES += ../../../../net/ynl TEST_GEN_FILES += csum +TEST_GEN_FILES += $(patsubst %.c,%.o,$(wildcard *.bpf.c)) TEST_INCLUDES := $(wildcard py/*.py sh/*.sh) include ../../lib.mk + +include ../bpf.mk diff --git a/drivers/net/netdevsim/ethtool.c b/drivers/net/netdevsim/ethtool.c index 5c80fbee7913..7ab358616e03 100644 --- a/drivers/net/netdevsim/ethtool.c +++ b/drivers/net/netdevsim/ethtool.c @@ -184,9 +184,11 @@ static const struct ethtool_ops nsim_ethtool_ops = { static void nsim_ethtool_ring_init(struct netdevsim *ns) { + ns->ethtool.ring.rx_pending = 512; ns->ethtool.ring.rx_max_pending = 4096; ns->ethtool.ring.rx_jumbo_max_pending = 4096; ns->ethtool.ring.rx_mini_max_pending = 4096; + ns->ethtool.ring.tx_pending = 512; ns->ethtool.ring.tx_max_pending = 4096; } diff --git a/tools/testing/selftests/net/lib/xdp_dummy.bpf.c b/tools/testing/selftests/net/lib/xdp_dummy.bpf.c new file mode 100644 index 000000000000..d988b2e0cee8 --- /dev/null +++ b/tools/testing/selftests/net/lib/xdp_dummy.bpf.c @@ -0,0 +1,13 @@ +// SPDX-License-Identifier: GPL-2.0 + +#define KBUILD_MODNAME "xdp_dummy" +#include <linux/bpf.h> +#include <bpf/bpf_helpers.h> + +SEC("xdp") +int xdp_dummy_prog(struct xdp_md *ctx) +{ + return XDP_PASS; +} + +char _license[] SEC("license") = "GPL"; diff --git a/tools/testing/selftests/drivers/net/hds.py b/tools/testing/selftests/drivers/net/hds.py index 394971b25c0b..90807b21a6eb 100755 --- a/tools/testing/selftests/drivers/net/hds.py +++ b/tools/testing/selftests/drivers/net/hds.py @@ -2,17 +2,54 @@ # SPDX-License-Identifier: GPL-2.0 import errno +import os from lib.py import ksft_run, ksft_exit, ksft_eq, ksft_raises, KsftSkipEx -from lib.py import EthtoolFamily, NlError +from lib.py import CmdExitFailure, EthtoolFamily, NlError from lib.py import NetDrvEnv +from lib.py import defer, ethtool, ip -def get_hds(cfg, netnl) -> None: + +def _get_hds_mode(cfg, netnl) -> str: try: rings = netnl.rings_get({'header': {'dev-index': cfg.ifindex}}) except NlError as e: raise KsftSkipEx('ring-get not supported by device') if 'tcp-data-split' not in rings: raise KsftSkipEx('tcp-data-split not supported by device') + return rings['tcp-data-split'] + + +def _xdp_onoff(cfg): + test_dir = os.path.dirname(os.path.realpath(__file__)) + prog = test_dir + "/../../net/lib/xdp_dummy.bpf.o" + ip(f"link set dev %s xdp obj %s sec xdp" % + (cfg.ifname, prog)) + ip(f"link set dev %s xdp off" % cfg.ifname) + + +def _ioctl_ringparam_modify(cfg, netnl) -> None: + """ + Helper for performing a hopefully unimportant IOCTL SET. + IOCTL does not support HDS, so it should not affect the HDS config. + """ + try: + rings = netnl.rings_get({'header': {'dev-index': cfg.ifindex}}) + except NlError as e: + raise KsftSkipEx('ring-get not supported by device') + + if 'tx' not in rings: + raise KsftSkipEx('setting Tx ring size not supported') + + try: + ethtool(f"--disable-netlink -G {cfg.ifname} tx {rings['tx'] // 2}") + except CmdExitFailure as e: + ethtool(f"--disable-netlink -G {cfg.ifname} tx {rings['tx'] * 2}") + defer(ethtool, f"-G {cfg.ifname} tx {rings['tx']}") + + +def get_hds(cfg, netnl) -> None: + _get_hds_mode(cfg, netnl) + def get_hds_thresh(cfg, netnl) -> None: try: @@ -104,6 +141,103 @@ from lib.py import NetDrvEnv netnl.rings_set({'header': {'dev-index': cfg.ifindex}, 'hds-thresh': hds_gt}) ksft_eq(e.exception.nl_msg.error, -errno.EINVAL) + +def set_xdp(cfg, netnl) -> None: + """ + Enable single-buffer XDP on the device. + When HDS is in "auto" / UNKNOWN mode, XDP installation should work. + """ + mode = _get_hds_mode(cfg, netnl) + if mode == 'enabled': + netnl.rings_set({'header': {'dev-index': cfg.ifindex}, + 'tcp-data-split': 'unknown'}) + + _xdp_onoff(cfg) + + +def enabled_set_xdp(cfg, netnl) -> None: + """ + Enable single-buffer XDP on the device. + When HDS is in "enabled" mode, XDP installation should not work. + """ + _get_hds_mode(cfg, netnl) + netnl.rings_set({'header': {'dev-index': cfg.ifindex}, + 'tcp-data-split': 'enabled'}) + + defer(netnl.rings_set, {'header': {'dev-index': cfg.ifindex}, + 'tcp-data-split': 'unknown'}) + + with ksft_raises(CmdExitFailure) as e: + _xdp_onoff(cfg) + + +def set_xdp(cfg, netnl) -> None: + """ + Enable single-buffer XDP on the device. + When HDS is in "auto" / UNKNOWN mode, XDP installation should work. + """ + mode = _get_hds_mode(cfg, netnl) + if mode == 'enabled': + netnl.rings_set({'header': {'dev-index': cfg.ifindex}, + 'tcp-data-split': 'unknown'}) + + _xdp_onoff(cfg) + + +def enabled_set_xdp(cfg, netnl) -> None: + """ + Enable single-buffer XDP on the device. + When HDS is in "enabled" mode, XDP installation should not work. + """ + _get_hds_mode(cfg, netnl) # Trigger skip if not supported + + netnl.rings_set({'header': {'dev-index': cfg.ifindex}, + 'tcp-data-split': 'enabled'}) + defer(netnl.rings_set, {'header': {'dev-index': cfg.ifindex}, + 'tcp-data-split': 'unknown'}) + + with ksft_raises(CmdExitFailure) as e: + _xdp_onoff(cfg) + + +def ioctl(cfg, netnl) -> None: + mode1 = _get_hds_mode(cfg, netnl) + _ioctl_ringparam_modify(cfg, netnl) + mode2 = _get_hds_mode(cfg, netnl) + + ksft_eq(mode1, mode2) + + +def ioctl_set_xdp(cfg, netnl) -> None: + """ + Like set_xdp(), but we perturb the settings via the legacy ioctl. + """ + mode = _get_hds_mode(cfg, netnl) + if mode == 'enabled': + netnl.rings_set({'header': {'dev-index': cfg.ifindex}, + 'tcp-data-split': 'unknown'}) + + _ioctl_ringparam_modify(cfg, netnl) + + _xdp_onoff(cfg) + + +def ioctl_enabled_set_xdp(cfg, netnl) -> None: + """ + Enable single-buffer XDP on the device. + When HDS is in "enabled" mode, XDP installation should not work. + """ + _get_hds_mode(cfg, netnl) # Trigger skip if not supported + + netnl.rings_set({'header': {'dev-index': cfg.ifindex}, + 'tcp-data-split': 'enabled'}) + defer(netnl.rings_set, {'header': {'dev-index': cfg.ifindex}, + 'tcp-data-split': 'unknown'}) + + with ksft_raises(CmdExitFailure) as e: + _xdp_onoff(cfg) + + def main() -> None: with NetDrvEnv(__file__, queue_count=3) as cfg: ksft_run([get_hds, @@ -112,7 +246,12 @@ from lib.py import NetDrvEnv set_hds_enable, set_hds_thresh_zero, set_hds_thresh_max, - set_hds_thresh_gt], + set_hds_thresh_gt, + set_xdp, + enabled_set_xdp, + ioctl, + ioctl_set_xdp, + ioctl_enabled_set_xdp], args=(cfg, EthtoolFamily())) ksft_exit() -- 2.48.1

4 months, 1 week

3
2
0 0

[PATCH] selftests/mount: Close 'fd' when write fails

by ritvikfoss＠gmail.com

From: Ritvik Gupta <ritvikfoss(a)gmail.com> 1. Close the file descriptor when write fails. 2. Introduce 'close_or_die' helper function to reduce repetition. Signed-off-by: Ritvik Gupta <ritvikfoss(a)gmail.com> --- .../selftests/mount/unprivileged-remount-test.c | 13 +++++++++---- 1 file changed, 9 insertions(+), 4 deletions(-) diff --git a/tools/testing/selftests/mount/unprivileged-remount-test.c b/tools/testing/selftests/mount/unprivileged-remount-test.c index d2917054fe3a..3dd9df58725b 100644 --- a/tools/testing/selftests/mount/unprivileged-remount-test.c +++ b/tools/testing/selftests/mount/unprivileged-remount-test.c @@ -54,6 +54,13 @@ static void die(char *fmt, ...) exit(EXIT_FAILURE); } +static void close_or_die(char *filename, int fd) { + if (close(fd) != 0) { + die("close of %s failed: %s\n", + filename, strerror(errno)); + } +} + static void vmaybe_write_file(bool enoent_ok, char *filename, char *fmt, va_list ap) { char buf[4096]; @@ -79,6 +86,7 @@ static void vmaybe_write_file(bool enoent_ok, char *filename, char *fmt, va_list } written = write(fd, buf, buf_len); if (written != buf_len) { + close_or_die(filename, fd); if (written >= 0) { die("short write to %s\n", filename); } else { @@ -86,10 +94,7 @@ static void vmaybe_write_file(bool enoent_ok, char *filename, char *fmt, va_list filename, strerror(errno)); } } - if (close(fd) != 0) { - die("close of %s failed: %s\n", - filename, strerror(errno)); - } + close_or_die(filename, fd); } static void maybe_write_file(char *filename, char *fmt, ...) -- 2.48.1

4 months, 1 week

3
5
0 0

[PATCH] selftets: lib: remove reference to prime_numbers

by Tamir Duberstein

Remove a leftover shell script reference from commit 313b38a6ecb4 ("lib/prime_numbers: convert self-test to KUnit"). Reported-by: kernel test robot <oliver.sang(a)intel.com> Closes: https://lore.kernel.org/oe-lkp/202502171110.708d965a-lkp@intel.com Fixes: 313b38a6ecb4 ("lib/prime_numbers: convert self-test to KUnit") Signed-off-by: Tamir Duberstein <tamird(a)gmail.com> --- tools/testing/selftests/lib/Makefile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tools/testing/selftests/lib/Makefile b/tools/testing/selftests/lib/Makefile index c52fe3ad8e98..66dcbe2e39fa 100644 --- a/tools/testing/selftests/lib/Makefile +++ b/tools/testing/selftests/lib/Makefile @@ -4,5 +4,5 @@ # No binaries, but make sure arg-less "make" doesn't trigger "run_tests" all: -TEST_PROGS := printf.sh bitmap.sh prime_numbers.sh scanf.sh +TEST_PROGS := printf.sh bitmap.sh scanf.sh include ../lib.mk --- base-commit: 0ae0fa3bf0b44c8611d114a9f69985bf451010c3 change-id: 20250217-fix-prime-numbers-f5202155b226 Best regards, -- Tamir Duberstein <tamird(a)gmail.com>

4 months, 1 week

2
3
0 0

[PATCH] selftests/mount: Explicitly define buffer size

by ritvikfoss＠gmail.com

From: Ritvik Gupta <ritvikfoss(a)gmail.com> Define macro ('MAX_BUF_SIZE') for buffer size instead of hardcoded value '4096', to improve readability. Signed-off-by: Ritvik Gupta <ritvikfoss(a)gmail.com> --- tools/testing/selftests/mount/unprivileged-remount-test.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/tools/testing/selftests/mount/unprivileged-remount-test.c b/tools/testing/selftests/mount/unprivileged-remount-test.c index d2917054fe3a..67a550b7c69b 100644 --- a/tools/testing/selftests/mount/unprivileged-remount-test.c +++ b/tools/testing/selftests/mount/unprivileged-remount-test.c @@ -45,6 +45,8 @@ # define MS_STRICTATIME (1 << 24) #endif +#define MAX_BUF_SIZE 4096 + static void die(char *fmt, ...) { va_list ap; @@ -56,7 +58,7 @@ static void die(char *fmt, ...) static void vmaybe_write_file(bool enoent_ok, char *filename, char *fmt, va_list ap) { - char buf[4096]; + char buf[MAX_BUF_SIZE]; int fd; ssize_t written; int buf_len; -- 2.48.1

4 months, 1 week

3
2
0 0

[PATCH net-next v4 0/9] Device memory TCP TX

by Mina Almasry

v4: https://lore.kernel.org/netdev/20250203223916.1064540-1-almasrymina@google.… === v4 mainly addresses the critical driver support issue surfaced in v3 by Paolo and Stan. Drivers aiming to support netmem_tx should make sure not to pass the netmem dma-addrs to the dma-mapping APIs, as these dma-addrs may come from dma-bufs. Additionally other feedback from v3 is addressed. Major changes: - Add helpers to handle netmem dma-addrs. Add GVE support for netmem_tx. - Fix binding->tx_vec not being freed on error paths during the tx binding. - Add a minimal devmem_tx test to devmem.py. - Clean up everything obsolete from the cover letter (Paolo). v3: https://patchwork.kernel.org/project/netdevbpf/list/?series=929401&state=* === Address minor comments from RFCv2 and fix a few build warnings and ynl-regen issues. No major changes. RFC v2: https://patchwork.kernel.org/project/netdevbpf/list/?series=920056&state=* ======= RFC v2 addresses much of the feedback from RFC v1. I plan on sending something close to this as net-next reopens, sending it slightly early to get feedback if any. Major changes: -------------- - much improved UAPI as suggested by Stan. We now interpret the iov_base of the passed in iov from userspace as the offset into the dmabuf to send from. This removes the need to set iov.iov_base = NULL which may be confusing to users, and enables us to send multiple iovs in the same sendmsg() call. ncdevmem and the docs show a sample use of that. - Removed the duplicate dmabuf iov_iter in binding->iov_iter. I think this is good improvment as it was confusing to keep track of 2 iterators for the same sendmsg, and mistracking both iterators caused a couple of bugs reported in the last iteration that are now resolved with this streamlining. - Improved test coverage in ncdevmem. Now multiple sendmsg() are tested, and sending multiple iovs in the same sendmsg() is tested. - Fixed issue where dmabuf unmapping was happening in invalid context (Stan). ==================================================================== The TX path had been dropped from the Device Memory TCP patch series post RFCv1 [1], to make that series slightly easier to review. This series rebases the implementation of the TX path on top of the net_iov/netmem framework agreed upon and merged. The motivation for the feature is thoroughly described in the docs & cover letter of the original proposal, so I don't repeat the lengthy descriptions here, but they are available in [1]. Full outline on usage of the TX path is detailed in the documentation included with this series. Test example is available via the kselftest included in the series as well. The series is relatively small, as the TX path for this feature largely piggybacks on the existing MSG_ZEROCOPY implementation. Patch Overview: --------------- 1. Documentation & tests to give high level overview of the feature being added. 1. Add netmem refcounting needed for the TX path. 2. Devmem TX netlink API. 3. Devmem TX net stack implementation. 4. Make dma-buf unbinding scheduled work to handle TX cases where it gets freed from contexts where we can't sleep. 5. Add devmem TX documentation. 6. Add scaffolding enabling driver support for netmem_tx. Add helpers, driver feature flag, and docs to enable drivers to declare netmem_tx support. 7. Guard netmem_tx against being enabled against drivers that don't support it. 8. Add devmem_tx selftests. Add TX path to ncdevmem and add a test to devmem.py. Testing: -------- Testing is very similar to devmem TCP RX path. The ncdevmem test used for the RX path is now augemented with client functionality to test TX path. * Test Setup: Kernel: net-next with this RFC and memory provider API cherry-picked locally. Hardware: Google Cloud A3 VMs. NIC: GVE with header split & RSS & flow steering support. Performance results are not included with this version, unfortunately. I'm having issues running the dma-buf exporter driver against the upstream kernel on my test setup. The issues are specific to that dma-buf exporter and do not affect this patch series. I plan to follow up this series with perf fixes if the tests point to issues once they're up and running. Special thanks to Stan who took a stab at rebasing the TX implementation on top of the netmem/net_iov framework merged. Parts of his proposal [2] that are reused as-is are forked off into their own patches to give full credit. [1] https://lore.kernel.org/netdev/20240909054318.1809580-1-almasrymina@google.… [2] https://lore.kernel.org/netdev/20240913150913.1280238-2-sdf@fomichev.me/T/#… Cc: sdf(a)fomichev.me Cc: asml.silence(a)gmail.com Cc: dw(a)davidwei.uk Cc: Jamal Hadi Salim <jhs(a)mojatatu.com> Cc: Victor Nogueira <victor(a)mojatatu.com> Cc: Pedro Tammela <pctammela(a)mojatatu.com> Cc: Samiullah Khawaja <skhawaja(a)google.com> Mina Almasry (8): net: add get_netmem/put_netmem support net: devmem: Implement TX path net: devmem: make dmabuf unbinding scheduled work net: add devmem TCP TX documentation net: enable driver support for netmem TX gve: add netmem TX support to GVE DQO-RDA mode net: check for driver support in netmem TX selftests: ncdevmem: Implement devmem TCP TX Stanislav Fomichev (1): net: devmem: TCP tx netlink api Documentation/netlink/specs/netdev.yaml | 12 + Documentation/networking/devmem.rst | 150 ++++++++- .../networking/net_cachelines/net_device.rst | 1 + Documentation/networking/netdev-features.rst | 5 + Documentation/networking/netmem.rst | 14 +- drivers/net/ethernet/google/gve/gve_main.c | 4 + drivers/net/ethernet/google/gve/gve_tx_dqo.c | 8 +- include/linux/netdevice.h | 2 + include/linux/skbuff.h | 17 +- include/linux/skbuff_ref.h | 4 +- include/net/netmem.h | 23 ++ include/net/sock.h | 1 + include/uapi/linux/netdev.h | 1 + net/core/datagram.c | 48 ++- net/core/dev.c | 3 + net/core/devmem.c | 114 ++++++- net/core/devmem.h | 69 +++- net/core/netdev-genl-gen.c | 13 + net/core/netdev-genl-gen.h | 1 + net/core/netdev-genl.c | 73 ++++- net/core/skbuff.c | 48 ++- net/core/sock.c | 6 + net/ipv4/ip_output.c | 3 +- net/ipv4/tcp.c | 46 ++- net/ipv6/ip6_output.c | 3 +- net/vmw_vsock/virtio_transport_common.c | 5 +- tools/include/uapi/linux/netdev.h | 1 + .../selftests/drivers/net/hw/devmem.py | 28 +- .../selftests/drivers/net/hw/ncdevmem.c | 300 +++++++++++++++++- 29 files changed, 931 insertions(+), 72 deletions(-) -- 2.48.1.601.g30ceb7b040-goog

4 months, 1 week

3
23
0 0

[PATCH net-next] selftests: fib_nexthops: do not mark skipped tests as failed

by Hangbin Liu

The current test marks all unexpected return values as failed and sets ret to 1. If a test is skipped, the entire test also returns 1, incorrectly indicating failure. To fix this, add a skipped variable and set ret to 4 if it was previously 0. Otherwise, keep ret set to 1. Signed-off-by: Hangbin Liu <liuhangbin(a)gmail.com> --- tools/testing/selftests/net/fib_nexthops.sh | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/tools/testing/selftests/net/fib_nexthops.sh b/tools/testing/selftests/net/fib_nexthops.sh index 77c83d9508d3..6a58e23e1588 100755 --- a/tools/testing/selftests/net/fib_nexthops.sh +++ b/tools/testing/selftests/net/fib_nexthops.sh @@ -76,11 +76,13 @@ log_test() printf "TEST: %-60s [ OK ]\n" "${msg}" nsuccess=$((nsuccess+1)) else - ret=1 - nfail=$((nfail+1)) if [[ $rc -eq $ksft_skip ]]; then + [[ $ret -eq 0 ]] && ret=$ksft_skip + nskip=$((nskip+1)) printf "TEST: %-60s [SKIP]\n" "${msg}" else + ret=1 + nfail=$((nfail+1)) printf "TEST: %-60s [FAIL]\n" "${msg}" fi @@ -2528,6 +2530,7 @@ done if [ "$TESTS" != "none" ]; then printf "\nTests passed: %3d\n" ${nsuccess} printf "Tests failed: %3d\n" ${nfail} + printf "Tests skipped: %2d\n" ${nskip} fi exit $ret -- 2.46.0

4 months, 1 week

2
1
0 0

[PATCH] kunit: tool: Implement listing of available architectures

by Thomas Weißschuh

To implement custom scripting around kunit.py it is useful to get a list of available architectures. While it is possible to manually inspect tools/testing/kunit/qemu_configs/, this is annoying to implement and introduces a dependency on a kunit.py implementation detail. Introduce 'kunit.py run --arch help' which lists all known architectures in an easy to parse list. This is equivalent on how QEMU implements listing of possible argument values. Signed-off-by: Thomas Weißschuh <thomas.weissschuh(a)linutronix.de> --- Documentation/dev-tools/kunit/run_wrapper.rst | 2 ++ tools/testing/kunit/kunit_kernel.py | 8 ++++++++ 2 files changed, 10 insertions(+) diff --git a/Documentation/dev-tools/kunit/run_wrapper.rst b/Documentation/dev-tools/kunit/run_wrapper.rst index 19ddf5e07013314c608b570e297a8ff79a8efe7f..6697c71ee8ca020b8ac7e91b46e29ab082d9dea0 100644 --- a/Documentation/dev-tools/kunit/run_wrapper.rst +++ b/Documentation/dev-tools/kunit/run_wrapper.rst @@ -182,6 +182,8 @@ via UML. To run tests on qemu, by default it requires two flags: is ignored), the tests will run via UML. Non-UML architectures, for example: i386, x86_64, arm and so on; run on qemu. + ``--arch help`` lists all valid ``--arch`` values. + - ``--cross_compile``: Specifies the Kbuild toolchain. It passes the same argument as passed to the ``CROSS_COMPILE`` variable used by Kbuild. As a reminder, this will be the prefix for the toolchain diff --git a/tools/testing/kunit/kunit_kernel.py b/tools/testing/kunit/kunit_kernel.py index d30f90eae9a4237e85910fd36f7f1c731d952319..e04195b135edc8f1aabe21d094b276e47c4f6848 100644 --- a/tools/testing/kunit/kunit_kernel.py +++ b/tools/testing/kunit/kunit_kernel.py @@ -14,6 +14,7 @@ import os import shlex import shutil import signal +import sys import threading from typing import Iterator, List, Optional, Tuple from types import FrameType @@ -201,6 +202,13 @@ def _default_qemu_config_path(arch: str) -> str: return config_path options = [f[:-3] for f in os.listdir(QEMU_CONFIGS_DIR) if f.endswith('.py')] + + if arch == 'help': + print('um') + for option in options: + print(option) + sys.exit() + raise ConfigError(arch + ' is not a valid arch, options are ' + str(sorted(options))) def _get_qemu_ops(config_path: str, --- base-commit: 2014c95afecee3e76ca4a56956a936e23283f05b change-id: 20250220-kunit-list-552a8cdc011e Best regards, -- Thomas Weißschuh <thomas.weissschuh(a)linutronix.de>

4 months, 1 week

2
1
0 0

[PATCH net-next v10 00/13] net: Improve netns handling in rtnetlink

by Xiao Liang

This patch series includes some netns-related improvements and fixes for rtnetlink, to make link creation more intuitive: 1) Creating link in another net namespace doesn't conflict with link names in current one. 2) Refector rtnetlink link creation. Create link in target namespace directly. So that # ip link add netns ns1 link-netns ns2 tun0 type gre ... will create tun0 in ns1, rather than create it in ns2 and move to ns1. And don't conflict with another interface named "tun0" in current netns. Patch 01 avoids link name conflict in different netns. To achieve 2), there're mainly 3 steps: - Patch 02 packs newlink() parameters into a struct, including the original "src_net" along with more netns context. No semantic changes are introduced. - Patch 03 ~ 09 converts device drivers to use the explicit netns extracted from params. - Patch 10 ~ 11 removes the old netns parameter, and converts rtnetlink to create device in target netns directly. Patch 12 ~ 13 adds some tests for link name and link netns. --- BTW please note there're some issues found in current code: - In amt_newlink() drivers/net/amt.c: amt->net = net; ... amt->stream_dev = dev_get_by_index(net, ... Uses net, but amt_lookup_upper_dev() only searches in dev_net. So the AMT device may not be properly deleted if it's in a different netns from lower dev. - In lowpan_newlink() in net/ieee802154/6lowpan/core.c: wdev = dev_get_by_index(dev_net(ldev), nla_get_u32(tb[IFLA_LINK])); Looks for IFLA_LINK in dev_net, but in theory the ifindex is defined in link netns. And thanks to Kuniyuki for fixing related issues in gtp and pfcp: https://lore.kernel.org/netdev/20250110014754.33847-1-kuniyu@amazon.com/ --- v10: - Move link/peer net helper functions to from patch 02 to 03. - Remove redundant tunnel->net assignment for IPv4 tunnels (patch 05). - Initialize tunnel->net before calling register_netdevice() for IPv6 tunnels (patch 07). - Coding style fixes. v9: link: https://lore.kernel.org/all/20250210133002.883422-1-shaw.leon@gmail.com/ - Change the prototype of macvlan_common_newlink(). - Minor fixes of coding style and local variables. v8: link: https://lore.kernel.org/all/20250113143719.7948-1-shaw.leon@gmail.com/ - Move dev and ext_ack out from param struct. - Validate link_net and dev_net are identical for 6lowpan. v7: link: https://lore.kernel.org/all/20250104125732.17335-1-shaw.leon@gmail.com/ - Add selftest kconfig. - Remove a duplicated test of ip6gre. v6: link: https://lore.kernel.org/all/20241218130909.2173-1-shaw.leon@gmail.com/ - Split prototype, driver and rtnetlink changes. - Add more tests for link netns. - Fix IPv6 tunnel net overwriten in ndo_init(). - Reorder variable declarations. - Exclude a ip_tunnel-specific patch. v5: link: https://lore.kernel.org/all/20241209140151.231257-1-shaw.leon@gmail.com/ - Fix function doc in batman-adv. - Include peer_net in rtnl newlink parameters. v4: link: https://lore.kernel.org/all/20241118143244.1773-1-shaw.leon@gmail.com/ - Pack newlink() parameters to a single struct. - Use ynl async_msg_queue.empty() in selftest. v3: link: https://lore.kernel.org/all/20241113125715.150201-1-shaw.leon@gmail.com/ - Drop "netns_atomic" flag and module parameter. Add netns parameter to newlink() instead, and convert drivers accordingly. - Move python NetNSEnter helper to net selftest lib. v2: link: https://lore.kernel.org/all/20241107133004.7469-1-shaw.leon@gmail.com/ - Check NLM_F_EXCL to ensure only link creation is affected. - Add self tests for link name/ifindex conflict and notifications in different netns. - Changes in dummy driver and ynl in order to add the test case. v1: link: https://lore.kernel.org/all/20241023023146.372653-1-shaw.leon@gmail.com/ Xiao Liang (13): rtnetlink: Lookup device in target netns when creating link rtnetlink: Pack newlink() params into struct net: Use link/peer netns in newlink() of rtnl_link_ops ieee802154: 6lowpan: Validate link netns in newlink() of rtnl_link_ops net: ip_tunnel: Don't set tunnel->net in ip_tunnel_init() net: ip_tunnel: Use link netns in newlink() of rtnl_link_ops net: ipv6: Init tunnel link-netns before registering dev net: ipv6: Use link netns in newlink() of rtnl_link_ops net: xfrm: Use link netns in newlink() of rtnl_link_ops rtnetlink: Remove "net" from newlink params rtnetlink: Create link directly in target net namespace selftests: net: Add python context manager for netns entering selftests: net: Add test cases for link and peer netns drivers/infiniband/ulp/ipoib/ipoib_netlink.c | 9 +- drivers/net/amt.c | 11 +- drivers/net/bareudp.c | 9 +- drivers/net/bonding/bond_netlink.c | 6 +- drivers/net/can/dev/netlink.c | 4 +- drivers/net/can/vxcan.c | 7 +- .../ethernet/qualcomm/rmnet/rmnet_config.c | 9 +- drivers/net/geneve.c | 9 +- drivers/net/gtp.c | 10 +- drivers/net/ipvlan/ipvlan.h | 3 +- drivers/net/ipvlan/ipvlan_main.c | 8 +- drivers/net/ipvlan/ipvtap.c | 6 +- drivers/net/macsec.c | 9 +- drivers/net/macvlan.c | 21 +-- drivers/net/macvtap.c | 6 +- drivers/net/netkit.c | 14 +- drivers/net/pfcp.c | 9 +- drivers/net/ppp/ppp_generic.c | 9 +- drivers/net/team/team_core.c | 6 +- drivers/net/veth.c | 7 +- drivers/net/vrf.c | 5 +- drivers/net/vxlan/vxlan_core.c | 9 +- drivers/net/wireguard/device.c | 7 +- drivers/net/wireless/virtual/virt_wifi.c | 8 +- drivers/net/wwan/wwan_core.c | 16 +- include/linux/if_macvlan.h | 6 +- include/net/ip_tunnels.h | 5 +- include/net/rtnetlink.h | 40 ++++- net/8021q/vlan_netlink.c | 9 +- net/batman-adv/soft-interface.c | 9 +- net/bridge/br_netlink.c | 6 +- net/caif/chnl_net.c | 5 +- net/core/rtnetlink.c | 34 +++-- net/hsr/hsr_netlink.c | 12 +- net/ieee802154/6lowpan/core.c | 7 +- net/ipv4/ip_gre.c | 22 ++- net/ipv4/ip_tunnel.c | 7 +- net/ipv4/ip_vti.c | 9 +- net/ipv4/ipip.c | 9 +- net/ipv6/ip6_gre.c | 26 ++-- net/ipv6/ip6_tunnel.c | 18 ++- net/ipv6/ip6_vti.c | 14 +- net/ipv6/sit.c | 20 ++- net/xfrm/xfrm_interface_core.c | 15 +- tools/testing/selftests/net/Makefile | 1 + tools/testing/selftests/net/config | 5 + .../testing/selftests/net/lib/py/__init__.py | 2 +- tools/testing/selftests/net/lib/py/netns.py | 18 +++ tools/testing/selftests/net/link_netns.py | 141 ++++++++++++++++++ tools/testing/selftests/net/netns-name.sh | 10 ++ 50 files changed, 486 insertions(+), 181 deletions(-) create mode 100755 tools/testing/selftests/net/link_netns.py -- 2.48.1

4 months, 1 week

4
24
0 0

[PATCH v2] ww_mutex: convert self-test to KUnit

by Tamir Duberstein

Convert this unit test to a KUnit test. This allows the test to benefit from the KUnit tooling. Note that care is taken to avoid test-ending assertions in worker threads, which is unsafe in KUnit (and wasn't done before this change either). Signed-off-by: Tamir Duberstein <tamird(a)gmail.com> --- I tested this using: $ tools/testing/kunit/kunit.py run --arch arm64 --make_options LLVM=1 ww_mutex ; [12:48:16] ================== ww_mutex (5 subtests) =================== ; [12:48:16] ======================= test_mutex ======================== ; [12:48:16] [PASSED] flags=0 ; [12:48:16] [PASSED] flags=1 ; [12:48:16] [PASSED] flags=2 ; [12:48:16] [PASSED] flags=3 ; [12:48:16] [PASSED] flags=4 ; [12:48:17] [PASSED] flags=5 ; [12:48:17] [PASSED] flags=6 ; [12:48:17] [PASSED] flags=7 ; [12:48:17] =================== [PASSED] test_mutex ==================== ; [12:48:17] ========================= test_aa ========================= ; [12:48:17] [PASSED] lock ; [12:48:17] [PASSED] trylock ; [12:48:17] ===================== [PASSED] test_aa ===================== ; [12:48:17] ======================== test_abba ======================== ; [12:48:17] [PASSED] trylock=0,resolve=0 ; [12:48:17] [PASSED] trylock=1,resolve=1 ; [12:48:17] [PASSED] trylock=0,resolve=0 ; [12:48:17] [PASSED] trylock=1,resolve=1 ; [12:48:17] ==================== [PASSED] test_abba ==================== ; [12:48:17] ======================= test_cycle ======================== ; [12:48:17] [PASSED] nthreads=2 ; [12:48:17] =================== [PASSED] test_cycle ==================== ; [12:48:21] ========================= stress ========================== ; [12:48:21] [PASSED] nlocks=16,nthreads_per_cpu=2,flags=1 ; [12:48:23] [PASSED] nlocks=16,nthreads_per_cpu=2,flags=2 ; [12:48:23] [PASSED] nlocks=2046,nthreads_per_cpu=3,flags=7 ; [12:48:23] ===================== [PASSED] stress ====================== ; [12:48:23] ==================== [PASSED] ww_mutex ===================== ; [12:48:23] ============================================================ ; [12:48:23] Testing complete. Ran 18 tests: passed: 18 --- Changes in v2: - Avoid KUNIT_ASSERT_* in non-main thread. (David Gow) - Include rationale and details in the commit message. (Boqun Feng) - Introduce test_abba_param to avoid odd-looking bit shifting. (Dan Carpenter) - Rebase on linux-next. - Move the test to tests/ per KUnit style guide. - Link to v1: https://lore.kernel.org/r/20250210-ww_mutex-kunit-convert-v1-1-972f0201f71e… --- kernel/locking/Makefile | 3 +- kernel/locking/tests/Makefile | 3 + .../{test-ww_mutex.c => tests/ww_mutex_kunit.c} | 319 +++++++++++---------- lib/Kconfig.debug | 12 +- tools/testing/selftests/locking/ww_mutex.sh | 19 -- 5 files changed, 176 insertions(+), 180 deletions(-) diff --git a/kernel/locking/Makefile b/kernel/locking/Makefile index 0db4093d17b8..b5fa68d18823 100644 --- a/kernel/locking/Makefile +++ b/kernel/locking/Makefile @@ -30,5 +30,6 @@ obj-$(CONFIG_DEBUG_SPINLOCK) += spinlock.o obj-$(CONFIG_DEBUG_SPINLOCK) += spinlock_debug.o obj-$(CONFIG_QUEUED_RWLOCKS) += qrwlock.o obj-$(CONFIG_LOCK_TORTURE_TEST) += locktorture.o -obj-$(CONFIG_WW_MUTEX_SELFTEST) += test-ww_mutex.o obj-$(CONFIG_LOCK_EVENT_COUNTS) += lock_events.o + +obj-y += tests/ diff --git a/kernel/locking/tests/Makefile b/kernel/locking/tests/Makefile new file mode 100644 index 000000000000..8ddde822c793 --- /dev/null +++ b/kernel/locking/tests/Makefile @@ -0,0 +1,3 @@ +# SPDX-License-Identifier: GPL-2.0-only + +obj-$(CONFIG_WW_MUTEX_KUNIT_TEST) += ww_mutex_kunit.o diff --git a/kernel/locking/test-ww_mutex.c b/kernel/locking/tests/ww_mutex_kunit.c similarity index 66% rename from kernel/locking/test-ww_mutex.c rename to kernel/locking/tests/ww_mutex_kunit.c index bcb1b9fea588..a8d412231735 100644 --- a/kernel/locking/test-ww_mutex.c +++ b/kernel/locking/tests/ww_mutex_kunit.c @@ -3,10 +3,10 @@ * Module-based API test facility for ww_mutexes */ -#include <linux/kernel.h> - +#include <kunit/test.h> #include <linux/completion.h> #include <linux/delay.h> +#include <linux/kernel.h> #include <linux/kthread.h> #include <linux/module.h> #include <linux/prandom.h> @@ -54,12 +54,39 @@ static void test_mutex_work(struct work_struct *work) ww_mutex_unlock(&mtx->mutex); } -static int __test_mutex(unsigned int flags) +static const unsigned int *gen_range( + unsigned int *storage, + const unsigned int min, + const unsigned int max, + const int *prev) +{ + if (prev != NULL) { + if (*prev >= max) + return NULL; + *storage = *prev + 1; + } else { + *storage = min; + } + return storage; +} + +static const void *test_mutex_gen_params(const void *prev, char *desc) +{ + static unsigned int storage; + const unsigned int *next = gen_range(&storage, 0, __TEST_MTX_LAST - 1, prev); + + if (next != NULL) + snprintf(desc, KUNIT_PARAM_DESC_SIZE, "flags=%x", *next); + return next; +} + +static void test_mutex(struct kunit *test) { #define TIMEOUT (HZ / 16) + const unsigned int *param = test->param_value; + const unsigned int flags = *param; struct test_mutex mtx; struct ww_acquire_ctx ctx; - int ret; ww_mutex_init(&mtx.mutex, &ww_class); if (flags & TEST_MTX_CTX) @@ -79,53 +106,42 @@ static int __test_mutex(unsigned int flags) if (flags & TEST_MTX_SPIN) { unsigned long timeout = jiffies + TIMEOUT; - ret = 0; do { if (completion_done(&mtx.done)) { - ret = -EINVAL; + KUNIT_FAIL(test, "mutual exclusion failure"); break; } cond_resched(); } while (time_before(jiffies, timeout)); } else { - ret = wait_for_completion_timeout(&mtx.done, TIMEOUT); + KUNIT_EXPECT_EQ(test, wait_for_completion_timeout(&mtx.done, TIMEOUT), 0); } ww_mutex_unlock(&mtx.mutex); if (flags & TEST_MTX_CTX) ww_acquire_fini(&ctx); - if (ret) { - pr_err("%s(flags=%x): mutual exclusion failure\n", - __func__, flags); - ret = -EINVAL; - } - flush_work(&mtx.work); destroy_work_on_stack(&mtx.work); - return ret; #undef TIMEOUT } -static int test_mutex(void) +static const void *test_aa_gen_params(const void *prev, char *desc) { - int ret; - int i; - - for (i = 0; i < __TEST_MTX_LAST; i++) { - ret = __test_mutex(i); - if (ret) - return ret; - } + static unsigned int storage; + const unsigned int *next = gen_range(&storage, 0, 1, prev); - return 0; + if (next != NULL) + snprintf(desc, KUNIT_PARAM_DESC_SIZE, *next ? "trylock" : "lock"); + return next; } -static int test_aa(bool trylock) +static void test_aa(struct kunit *test) { + const unsigned int *param = test->param_value; + const bool trylock = *param; struct ww_mutex mutex; struct ww_acquire_ctx ctx; int ret; - const char *from = trylock ? "trylock" : "lock"; ww_mutex_init(&mutex, &ww_class); ww_acquire_init(&ctx, &ww_class); @@ -133,46 +149,42 @@ static int test_aa(bool trylock) if (!trylock) { ret = ww_mutex_lock(&mutex, &ctx); if (ret) { - pr_err("%s: initial lock failed!\n", __func__); + KUNIT_FAIL(test, "initial lock failed, ret=%d", ret); goto out; } } else { ret = !ww_mutex_trylock(&mutex, &ctx); if (ret) { - pr_err("%s: initial trylock failed!\n", __func__); + KUNIT_FAIL(test, "initial trylock failed, ret=%d", ret); goto out; } } - if (ww_mutex_trylock(&mutex, NULL)) { - pr_err("%s: trylocked itself without context from %s!\n", __func__, from); + ret = ww_mutex_trylock(&mutex, NULL); + if (ret) { + KUNIT_FAIL(test, "trylocked itself without context, ret=%d", ret); ww_mutex_unlock(&mutex); - ret = -EINVAL; goto out; } - if (ww_mutex_trylock(&mutex, &ctx)) { - pr_err("%s: trylocked itself with context from %s!\n", __func__, from); + ret = ww_mutex_trylock(&mutex, &ctx); + if (ret) { + KUNIT_FAIL(test, "trylocked itself with context, ret=%d", ret); ww_mutex_unlock(&mutex); - ret = -EINVAL; goto out; } ret = ww_mutex_lock(&mutex, &ctx); if (ret != -EALREADY) { - pr_err("%s: missed deadlock for recursing, ret=%d from %s\n", - __func__, ret, from); + KUNIT_FAIL(test, "missed deadlock for recursing, ret=%d", ret); if (!ret) ww_mutex_unlock(&mutex); - ret = -EINVAL; goto out; } ww_mutex_unlock(&mutex); - ret = 0; out: ww_acquire_fini(&ctx); - return ret; } struct test_abba { @@ -217,11 +229,36 @@ static void test_abba_work(struct work_struct *work) abba->result = err; } -static int test_abba(bool trylock, bool resolve) +union test_abba_param { + unsigned int value; + struct { + unsigned int trylock : 1; + unsigned int resolve : 1; + }; +}; + +static const void *test_abba_gen_params(const void *prev, char *desc) { + static unsigned int storage; + const unsigned int *next = gen_range(&storage, 0b00, 0b11, prev); + + if (next != NULL) { + const union test_abba_param param = { .value = *next }; + + snprintf(desc, KUNIT_PARAM_DESC_SIZE, "trylock=%d,resolve=%d", + param.trylock, param.resolve); + } + return next; +} + +static void test_abba(struct kunit *test) +{ + const union test_abba_param *param = test->param_value; + const bool trylock = param->trylock; + const bool resolve = param->resolve; struct test_abba abba; struct ww_acquire_ctx ctx; - int err, ret; + int err; ww_mutex_init(&abba.a_mutex, &ww_class); ww_mutex_init(&abba.b_mutex, &ww_class); @@ -259,21 +296,17 @@ static int test_abba(bool trylock, bool resolve) flush_work(&abba.work); destroy_work_on_stack(&abba.work); - ret = 0; if (resolve) { if (err || abba.result) { - pr_err("%s: failed to resolve ABBA deadlock, A err=%d, B err=%d\n", - __func__, err, abba.result); - ret = -EINVAL; + KUNIT_FAIL(test, "failed to resolve ABBA deadlock, A err=%d, B err=%d", + err, abba.result); } } else { if (err != -EDEADLK && abba.result != -EDEADLK) { - pr_err("%s: missed ABBA deadlock, A err=%d, B err=%d\n", - __func__, err, abba.result); - ret = -EINVAL; + KUNIT_FAIL(test, "missed ABBA deadlock, A err=%d, B err=%d", + err, abba.result); } } - return ret; } struct test_cycle { @@ -314,15 +347,25 @@ static void test_cycle_work(struct work_struct *work) cycle->result = err ?: erra; } -static int __test_cycle(unsigned int nthreads) +static const void *test_cycle_gen_params(const void *prev, char *desc) { + static unsigned int storage; + const unsigned int *next = gen_range(&storage, 2, num_online_cpus(), prev); + + if (next != NULL) + snprintf(desc, KUNIT_PARAM_DESC_SIZE, "nthreads=%d", *next); + return next; +} + +static void test_cycle(struct kunit *test) +{ + const unsigned int *param = test->param_value; + const unsigned int nthreads = *param; struct test_cycle *cycles; unsigned int n, last = nthreads - 1; - int ret; - cycles = kmalloc_array(nthreads, sizeof(*cycles), GFP_KERNEL); - if (!cycles) - return -ENOMEM; + cycles = kunit_kmalloc_array(test, nthreads, sizeof(*cycles), GFP_KERNEL); + KUNIT_ASSERT_NOT_NULL(test, cycles); for (n = 0; n < nthreads; n++) { struct test_cycle *cycle = &cycles[n]; @@ -348,41 +391,24 @@ static int __test_cycle(unsigned int nthreads) flush_workqueue(wq); - ret = 0; for (n = 0; n < nthreads; n++) { struct test_cycle *cycle = &cycles[n]; if (!cycle->result) continue; - pr_err("cyclic deadlock not resolved, ret[%d/%d] = %d\n", - n, nthreads, cycle->result); - ret = -EINVAL; + KUNIT_FAIL(test, "cyclic deadlock not resolved, ret[%d/%d] = %d", + n, nthreads, cycle->result); break; } for (n = 0; n < nthreads; n++) ww_mutex_destroy(&cycles[n].a_mutex); - kfree(cycles); - return ret; -} - -static int test_cycle(unsigned int ncpus) -{ - unsigned int n; - int ret; - - for (n = 2; n <= ncpus + 1; n++) { - ret = __test_cycle(n); - if (ret) - return ret; - } - - return 0; } struct stress { struct work_struct work; + struct kunit *test; struct ww_mutex *locks; unsigned long timeout; int nlocks; @@ -401,12 +427,12 @@ static inline u32 prandom_u32_below(u32 ceil) return ret; } -static int *get_random_order(int count) +static int *get_random_order(struct kunit *test, int count) { int *order; int n, r; - order = kmalloc_array(count, sizeof(*order), GFP_KERNEL); + order = kunit_kmalloc_array(test, count, sizeof(*order), GFP_KERNEL); if (!order) return order; @@ -435,7 +461,8 @@ static void stress_inorder_work(struct work_struct *work) struct ww_acquire_ctx ctx; int *order; - order = get_random_order(nlocks); + order = get_random_order(stress->test, nlocks); + KUNIT_EXPECT_NOT_NULL(stress->test, order); if (!order) return; @@ -472,13 +499,10 @@ static void stress_inorder_work(struct work_struct *work) ww_acquire_fini(&ctx); if (err) { - pr_err_once("stress (%s) failed with %d\n", - __func__, err); + KUNIT_FAIL(stress->test, "lock[%d] failed, err=%d", n, err); break; } } while (!time_after(jiffies, stress->timeout)); - - kfree(order); } struct reorder_lock { @@ -495,19 +519,19 @@ static void stress_reorder_work(struct work_struct *work) int *order; int n, err; - order = get_random_order(stress->nlocks); + order = get_random_order(stress->test, stress->nlocks); + KUNIT_EXPECT_NOT_NULL(stress->test, order); if (!order) return; for (n = 0; n < stress->nlocks; n++) { - ll = kmalloc(sizeof(*ll), GFP_KERNEL); + ll = kunit_kmalloc(stress->test, sizeof(*ll), GFP_KERNEL); + KUNIT_EXPECT_NOT_NULL(stress->test, ll); if (!ll) - goto out; - + return; ll->lock = &stress->locks[order[n]]; list_add(&ll->link, &locks); } - kfree(order); order = NULL; do { @@ -523,8 +547,7 @@ static void stress_reorder_work(struct work_struct *work) ww_mutex_unlock(ln->lock); if (err != -EDEADLK) { - pr_err_once("stress (%s) failed with %d\n", - __func__, err); + KUNIT_FAIL(stress->test, "lock failed, err=%d", err); break; } @@ -538,11 +561,6 @@ static void stress_reorder_work(struct work_struct *work) ww_acquire_fini(&ctx); } while (!time_after(jiffies, stress->timeout)); - -out: - list_for_each_entry_safe(ll, ln, &locks, link) - kfree(ll); - kfree(order); } static void stress_one_work(struct work_struct *work) @@ -558,8 +576,7 @@ static void stress_one_work(struct work_struct *work) dummy_load(stress); ww_mutex_unlock(lock); } else { - pr_err_once("stress (%s) failed with %d\n", - __func__, err); + KUNIT_FAIL(stress->test, "lock failed, err=%d", err); break; } } while (!time_after(jiffies, stress->timeout)); @@ -570,22 +587,41 @@ static void stress_one_work(struct work_struct *work) #define STRESS_ONE BIT(2) #define STRESS_ALL (STRESS_INORDER | STRESS_REORDER | STRESS_ONE) -static int stress(int nlocks, int nthreads, unsigned int flags) +struct stress_case { + int nlocks; + int nthreads_per_cpu; + unsigned int flags; +}; + +static const struct stress_case stress_cases[] = { + { 16, 2, STRESS_INORDER }, + { 16, 2, STRESS_REORDER }, + { 2046, hweight32(STRESS_ALL), STRESS_ALL }, +}; + +static void stress_case_to_desc(const struct stress_case *param, char *desc) { + snprintf(desc, KUNIT_PARAM_DESC_SIZE, "nlocks=%d,nthreads_per_cpu=%d,flags=%x", + param->nlocks, param->nthreads_per_cpu, param->flags); +} + +KUNIT_ARRAY_PARAM(stress_cases, stress_cases, stress_case_to_desc); + +static void stress(struct kunit *test) +{ + const struct stress_case *param = test->param_value; + const int nlocks = param->nlocks; + int nthreads = param->nthreads_per_cpu * num_online_cpus(); + const unsigned int flags = param->flags; struct ww_mutex *locks; struct stress *stress_array; int n, count; - locks = kmalloc_array(nlocks, sizeof(*locks), GFP_KERNEL); - if (!locks) - return -ENOMEM; + locks = kunit_kmalloc_array(test, nlocks, sizeof(*locks), GFP_KERNEL); + KUNIT_ASSERT_NOT_NULL(test, locks); - stress_array = kmalloc_array(nthreads, sizeof(*stress_array), - GFP_KERNEL); - if (!stress_array) { - kfree(locks); - return -ENOMEM; - } + stress_array = kunit_kmalloc_array(test, nthreads, sizeof(*stress_array), GFP_KERNEL); + KUNIT_ASSERT_NOT_NULL(test, stress_array); for (n = 0; n < nlocks; n++) ww_mutex_init(&locks[n], &ww_class); @@ -617,6 +653,7 @@ static int stress(int nlocks, int nthreads, unsigned int flags) stress = &stress_array[count++]; INIT_WORK(&stress->work, fn); + stress->test = test; stress->locks = locks; stress->nlocks = nlocks; stress->timeout = jiffies + 2*HZ; @@ -629,70 +666,42 @@ static int stress(int nlocks, int nthreads, unsigned int flags) for (n = 0; n < nlocks; n++) ww_mutex_destroy(&locks[n]); - kfree(stress_array); - kfree(locks); - - return 0; } -static int __init test_ww_mutex_init(void) +static int ww_mutex_suite_init(struct kunit_suite *suite) { - int ncpus = num_online_cpus(); - int ret, i; - - printk(KERN_INFO "Beginning ww mutex selftests\n"); - - prandom_seed_state(&rng, get_random_u64()); - wq = alloc_workqueue("test-ww_mutex", WQ_UNBOUND, 0); if (!wq) return -ENOMEM; - ret = test_mutex(); - if (ret) - return ret; - - ret = test_aa(false); - if (ret) - return ret; - - ret = test_aa(true); - if (ret) - return ret; - - for (i = 0; i < 4; i++) { - ret = test_abba(i & 1, i & 2); - if (ret) - return ret; - } - - ret = test_cycle(ncpus); - if (ret) - return ret; - - ret = stress(16, 2*ncpus, STRESS_INORDER); - if (ret) - return ret; - - ret = stress(16, 2*ncpus, STRESS_REORDER); - if (ret) - return ret; - - ret = stress(2046, hweight32(STRESS_ALL)*ncpus, STRESS_ALL); - if (ret) - return ret; + prandom_seed_state(&rng, get_random_u64()); - printk(KERN_INFO "All ww mutex selftests passed\n"); return 0; } -static void __exit test_ww_mutex_exit(void) +static void ww_mutex_suite_exit(struct kunit_suite *suite) { - destroy_workqueue(wq); + if (wq) + destroy_workqueue(wq); } -module_init(test_ww_mutex_init); -module_exit(test_ww_mutex_exit); +static struct kunit_case ww_mutex_cases[] = { + KUNIT_CASE_PARAM(test_mutex, test_mutex_gen_params), + KUNIT_CASE_PARAM(test_aa, test_aa_gen_params), + KUNIT_CASE_PARAM(test_abba, test_abba_gen_params), + KUNIT_CASE_PARAM(test_cycle, test_cycle_gen_params), + KUNIT_CASE_PARAM_ATTR(stress, stress_cases_gen_params, {.speed = KUNIT_SPEED_SLOW}), + {}, +}; + +static struct kunit_suite ww_mutex_suite = { + .name = "ww_mutex", + .suite_init = ww_mutex_suite_init, + .suite_exit = ww_mutex_suite_exit, + .test_cases = ww_mutex_cases, +}; + +kunit_test_suite(ww_mutex_suite); MODULE_LICENSE("GPL"); MODULE_AUTHOR("Intel Corporation"); diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index 85b95d645b10..2c9da8647eaf 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -1595,16 +1595,18 @@ config LOCK_TORTURE_TEST Say M if you want these torture tests to build as a module. Say N if you are unsure. -config WW_MUTEX_SELFTEST - tristate "Wait/wound mutex selftests" +config WW_MUTEX_KUNIT_TEST + tristate "KUnit test for wait/wound mutex" if !KUNIT_ALL_TESTS + depends on KUNIT + default KUNIT_ALL_TESTS help - This option provides a kernel module that runs tests on the - on the struct ww_mutex locking API. + This option provides a KUnit test that exercises the struct + ww_mutex locking API. It is recommended to enable DEBUG_WW_MUTEX_SLOWPATH in conjunction with this test harness. - Say M if you want these self tests to build as a module. + Say M if you want these tests to build as a module. Say N if you are unsure. config SCF_TORTURE_TEST diff --git a/tools/testing/selftests/locking/ww_mutex.sh b/tools/testing/selftests/locking/ww_mutex.sh deleted file mode 100755 index 91e4ac7566af..000000000000 --- a/tools/testing/selftests/locking/ww_mutex.sh +++ /dev/null @@ -1,19 +0,0 @@ -#!/bin/sh -# SPDX-License-Identifier: GPL-2.0 - -# Kselftest framework requirement - SKIP code is 4. -ksft_skip=4 - -# Runs API tests for struct ww_mutex (Wait/Wound mutexes) -if ! /sbin/modprobe -q -n test-ww_mutex; then - echo "ww_mutex: module test-ww_mutex is not found [SKIP]" - exit $ksft_skip -fi - -if /sbin/modprobe -q test-ww_mutex; then - /sbin/modprobe -q -r test-ww_mutex - echo "locking/ww_mutex: ok" -else - echo "locking/ww_mutex: [FAIL]" - exit 1 -fi --- base-commit: 7b7a883c7f4de1ee5040bd1c32aabaafde54d209 change-id: 20250208-ww_mutex-kunit-convert-71842a7a6be2 Best regards, -- Tamir Duberstein <tamird(a)gmail.com>

4 months, 1 week

2
5
0 0

[PATCH bpf-next v2] arm64, bpf: Add 12-argument support for bpf trampoline

by Puranjay Mohan

The arm64 bpf JIT currently supports attaching the trampoline to functions with <= 8 arguments. This is because up to 8 arguments can be passed in registers r0-r7. If there are more than 8 arguments then the 9th and later arguments are passed on the stack, with SP pointing to the first stacked argument. See aapcs64[1] for more details. If the 8th argument is a structure of size > 8B, then it is passed fully on stack and r7 is not used for passing any argument. If there is a 9th argument, it will be passed on the stack, even though r7 is available. Add the support of storing and restoring arguments passed on the stack to the arm64 bpf trampoline. This will allow attaching the trampoline to functions that take up to 12 arguments. [1] https://github.com/ARM-software/abi-aa/blob/main/aapcs64/aapcs64.rst#parame… Signed-off-by: Puranjay Mohan <puranjay(a)kernel.org> --- Changes in V1 -> V2: V1: https://lore.kernel.org/all/20240704173227.130491-1-puranjay@kernel.org/ - Fixed the argument handling for composite types (structs) --- arch/arm64/net/bpf_jit_comp.c | 139 ++++++++++++++----- tools/testing/selftests/bpf/DENYLIST.aarch64 | 3 - 2 files changed, 107 insertions(+), 35 deletions(-) diff --git a/arch/arm64/net/bpf_jit_comp.c b/arch/arm64/net/bpf_jit_comp.c index 751331f5ba90..063bf5e11fc6 100644 --- a/arch/arm64/net/bpf_jit_comp.c +++ b/arch/arm64/net/bpf_jit_comp.c @@ -30,6 +30,8 @@ #define TMP_REG_3 (MAX_BPF_JIT_REG + 3) #define FP_BOTTOM (MAX_BPF_JIT_REG + 4) #define ARENA_VM_START (MAX_BPF_JIT_REG + 5) +/* Up to eight function arguments are passed in registers r0-r7 */ +#define ARM64_MAX_REG_ARGS 8 #define check_imm(bits, imm) do { \ if ((((imm) > 0) && ((imm) >> (bits))) || \ @@ -2001,26 +2003,51 @@ static void invoke_bpf_mod_ret(struct jit_ctx *ctx, struct bpf_tramp_links *tl, } } -static void save_args(struct jit_ctx *ctx, int args_off, int nregs) +static void save_args(struct jit_ctx *ctx, int args_off, int orig_sp_off, + int nargs, int nreg_args) { + const u8 tmp = bpf2a64[TMP_REG_1]; + int arg_pos; int i; - for (i = 0; i < nregs; i++) { - emit(A64_STR64I(i, A64_SP, args_off), ctx); + for (i = 0; i < nargs; i++) { + if (i < nreg_args) { + emit(A64_STR64I(i, A64_SP, args_off), ctx); + } else { + arg_pos = orig_sp_off + (i - nreg_args) * 8; + emit(A64_LDR64I(tmp, A64_SP, arg_pos), ctx); + emit(A64_STR64I(tmp, A64_SP, args_off), ctx); + } args_off += 8; } } -static void restore_args(struct jit_ctx *ctx, int args_off, int nregs) +static void restore_args(struct jit_ctx *ctx, int args_off, int nreg_args) { int i; - for (i = 0; i < nregs; i++) { + for (i = 0; i < nreg_args; i++) { emit(A64_LDR64I(i, A64_SP, args_off), ctx); args_off += 8; } } +static void restore_stack_args(struct jit_ctx *ctx, int args_off, int stk_arg_off, + int nargs, int nreg_args) +{ + const u8 tmp = bpf2a64[TMP_REG_1]; + int arg_pos; + int i; + + for (i = nreg_args; i < nargs; i++) { + arg_pos = args_off + i * 8; + emit(A64_LDR64I(tmp, A64_SP, arg_pos), ctx); + emit(A64_STR64I(tmp, A64_SP, stk_arg_off), ctx); + + stk_arg_off += 8; + } +} + /* Based on the x86's implementation of arch_prepare_bpf_trampoline(). * * bpf prog and function entry before bpf trampoline hooked: @@ -2034,15 +2061,17 @@ static void restore_args(struct jit_ctx *ctx, int args_off, int nregs) */ static int prepare_trampoline(struct jit_ctx *ctx, struct bpf_tramp_image *im, struct bpf_tramp_links *tlinks, void *func_addr, - int nregs, u32 flags) + int nargs, int nreg_args, u32 flags) { int i; int stack_size; + int stk_arg_off; + int orig_sp_off; int retaddr_off; int regs_off; int retval_off; int args_off; - int nregs_off; + int nargs_off; int ip_off; int run_ctx_off; struct bpf_tramp_links *fentry = &tlinks[BPF_TRAMP_FENTRY]; @@ -2052,6 +2081,7 @@ static int prepare_trampoline(struct jit_ctx *ctx, struct bpf_tramp_image *im, __le32 **branches = NULL; /* trampoline stack layout: + * SP + orig_sp_off [ first stack arg ] if nargs > 8 * [ parent ip ] * [ FP ] * SP + retaddr_off [ self ip ] @@ -2069,14 +2099,24 @@ static int prepare_trampoline(struct jit_ctx *ctx, struct bpf_tramp_image *im, * [ ... ] * SP + args_off [ arg reg 1 ] * - * SP + nregs_off [ arg regs count ] + * SP + nargs_off [ arg count ] * * SP + ip_off [ traced function ] BPF_TRAMP_F_IP_ARG flag * * SP + run_ctx_off [ bpf_tramp_run_ctx ] + * + * [ stack_argN ] + * [ ... ] + * SP + stk_arg_off [ stack_arg1 ] BPF_TRAMP_F_CALL_ORIG */ stack_size = 0; + stk_arg_off = stack_size; + if ((flags & BPF_TRAMP_F_CALL_ORIG) && (nargs - nreg_args > 0)) { + /* room for saving arguments passed on stack */ + stack_size += (nargs - nreg_args) * 8; + } + run_ctx_off = stack_size; /* room for bpf_tramp_run_ctx */ stack_size += round_up(sizeof(struct bpf_tramp_run_ctx), 8); @@ -2086,13 +2126,13 @@ static int prepare_trampoline(struct jit_ctx *ctx, struct bpf_tramp_image *im, if (flags & BPF_TRAMP_F_IP_ARG) stack_size += 8; - nregs_off = stack_size; + nargs_off = stack_size; /* room for args count */ stack_size += 8; args_off = stack_size; /* room for args */ - stack_size += nregs * 8; + stack_size += nargs * 8; /* room for return value */ retval_off = stack_size; @@ -2110,6 +2150,11 @@ static int prepare_trampoline(struct jit_ctx *ctx, struct bpf_tramp_image *im, /* return address locates above FP */ retaddr_off = stack_size + 8; + /* original SP position + * stack_size + parent function frame + patched function frame + */ + orig_sp_off = stack_size + 32; + /* bpf trampoline may be invoked by 3 instruction types: * 1. bl, attached to bpf prog or kernel function via short jump * 2. br, attached to bpf prog or kernel function via long jump @@ -2135,12 +2180,12 @@ static int prepare_trampoline(struct jit_ctx *ctx, struct bpf_tramp_image *im, emit(A64_STR64I(A64_R(10), A64_SP, ip_off), ctx); } - /* save arg regs count*/ - emit(A64_MOVZ(1, A64_R(10), nregs, 0), ctx); - emit(A64_STR64I(A64_R(10), A64_SP, nregs_off), ctx); + /* save argument count */ + emit(A64_MOVZ(1, A64_R(10), nargs, 0), ctx); + emit(A64_STR64I(A64_R(10), A64_SP, nargs_off), ctx); - /* save arg regs */ - save_args(ctx, args_off, nregs); + /* save arguments passed in regs and on the stack */ + save_args(ctx, args_off, orig_sp_off, nargs, nreg_args); /* save callee saved registers */ emit(A64_STR64I(A64_R(19), A64_SP, regs_off), ctx); @@ -2167,7 +2212,10 @@ static int prepare_trampoline(struct jit_ctx *ctx, struct bpf_tramp_image *im, } if (flags & BPF_TRAMP_F_CALL_ORIG) { - restore_args(ctx, args_off, nregs); + /* restore arguments that were passed in registers */ + restore_args(ctx, args_off, nreg_args); + /* restore arguments that were passed on the stack */ + restore_stack_args(ctx, args_off, stk_arg_off, nargs, nreg_args); /* call original func */ emit(A64_LDR64I(A64_R(10), A64_SP, retaddr_off), ctx); emit(A64_ADR(A64_LR, AARCH64_INSN_SIZE * 2), ctx); @@ -2196,7 +2244,7 @@ static int prepare_trampoline(struct jit_ctx *ctx, struct bpf_tramp_image *im, } if (flags & BPF_TRAMP_F_RESTORE_REGS) - restore_args(ctx, args_off, nregs); + restore_args(ctx, args_off, nreg_args); /* restore callee saved register x19 and x20 */ emit(A64_LDR64I(A64_R(19), A64_SP, regs_off), ctx); @@ -2228,19 +2276,42 @@ static int prepare_trampoline(struct jit_ctx *ctx, struct bpf_tramp_image *im, return ctx->idx; } -static int btf_func_model_nregs(const struct btf_func_model *m) +static int btf_func_model_nargs(const struct btf_func_model *m) { - int nregs = m->nr_args; + int nargs = m->nr_args; int i; - /* extra registers needed for struct argument */ + /* extra registers or stack slots needed for struct argument */ for (i = 0; i < MAX_BPF_FUNC_ARGS; i++) { /* The arg_size is at most 16 bytes, enforced by the verifier. */ if (m->arg_flags[i] & BTF_FMODEL_STRUCT_ARG) - nregs += (m->arg_size[i] + 7) / 8 - 1; + nargs += (m->arg_size[i] + 7) / 8 - 1; } - return nregs; + return nargs; +} + +/* get the count of the regs that are used to pass arguments */ +static int btf_func_model_nreg_args(const struct btf_func_model *m) +{ + int nargs = m->nr_args; + int nreg_args = 0; + int i; + + for (i = 0; i < nargs; i++) { + /* The arg_size is at most 16 bytes, enforced by the verifier. */ + if (m->arg_flags[i] & BTF_FMODEL_STRUCT_ARG) { + /* struct members are all in the registers or all + * on the stack. + */ + if (nreg_args + ((m->arg_size[i] + 7) / 8 - 1) > 7) + break; + nreg_args += (m->arg_size[i] + 7) / 8 - 1; + } + nreg_args++; + } + + return (nreg_args > ARM64_MAX_REG_ARGS ? ARM64_MAX_REG_ARGS : nreg_args); } int arch_bpf_trampoline_size(const struct btf_func_model *m, u32 flags, @@ -2251,14 +2322,16 @@ int arch_bpf_trampoline_size(const struct btf_func_model *m, u32 flags, .idx = 0, }; struct bpf_tramp_image im; - int nregs, ret; + int nargs, nreg_args, ret; - nregs = btf_func_model_nregs(m); - /* the first 8 registers are used for arguments */ - if (nregs > 8) + nargs = btf_func_model_nargs(m); + if (nargs > MAX_BPF_FUNC_ARGS) return -ENOTSUPP; - ret = prepare_trampoline(&ctx, &im, tlinks, func_addr, nregs, flags); + nreg_args = btf_func_model_nreg_args(m); + + ret = prepare_trampoline(&ctx, &im, tlinks, func_addr, nargs, nreg_args, + flags); if (ret < 0) return ret; @@ -2285,7 +2358,7 @@ int arch_prepare_bpf_trampoline(struct bpf_tramp_image *im, void *ro_image, u32 flags, struct bpf_tramp_links *tlinks, void *func_addr) { - int ret, nregs; + int ret, nargs, nreg_args; void *image, *tmp; u32 size = ro_image_end - ro_image; @@ -2302,13 +2375,15 @@ int arch_prepare_bpf_trampoline(struct bpf_tramp_image *im, void *ro_image, .idx = 0, }; - nregs = btf_func_model_nregs(m); - /* the first 8 registers are used for arguments */ - if (nregs > 8) + nargs = btf_func_model_nargs(m); + if (nargs > MAX_BPF_FUNC_ARGS) return -ENOTSUPP; + nreg_args = btf_func_model_nreg_args(m); + jit_fill_hole(image, (unsigned int)(ro_image_end - ro_image)); - ret = prepare_trampoline(&ctx, im, tlinks, func_addr, nregs, flags); + ret = prepare_trampoline(&ctx, im, tlinks, func_addr, nargs, nreg_args, + flags); if (ret > 0 && validate_code(&ctx) < 0) { ret = -EINVAL; diff --git a/tools/testing/selftests/bpf/DENYLIST.aarch64 b/tools/testing/selftests/bpf/DENYLIST.aarch64 index 3c7c3e79aa93..e865451e90d2 100644 --- a/tools/testing/selftests/bpf/DENYLIST.aarch64 +++ b/tools/testing/selftests/bpf/DENYLIST.aarch64 @@ -4,9 +4,6 @@ fexit_sleep # The test never returns. The r kprobe_multi_bench_attach # needs CONFIG_FPROBE kprobe_multi_test # needs CONFIG_FPROBE module_attach # prog 'kprobe_multi': failed to auto-attach: -95 -fentry_test/fentry_many_args # fentry_many_args:FAIL:fentry_many_args_attach unexpected error: -524 -fexit_test/fexit_many_args # fexit_many_args:FAIL:fexit_many_args_attach unexpected error: -524 -tracing_struct/struct_many_args # struct_many_args:FAIL:tracing_struct_many_args__attach unexpected error: -524 fill_link_info/kprobe_multi_link_info # bpf_program__attach_kprobe_multi_opts unexpected error: -95 fill_link_info/kretprobe_multi_link_info # bpf_program__attach_kprobe_multi_opts unexpected error: -95 fill_link_info/kprobe_multi_invalid_ubuff # bpf_program__attach_kprobe_multi_opts unexpected error: -95 -- 2.40.1

4 months, 1 week

3
4
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-kselftest-mirror