fix wrong hds-thresh value setting
Subsystems (4)
v2 cover fix wrong hds-thresh value setting
A hds-thresh value is not set correctly if input value is 0. The cause is that ethtool_ringparam_get_cfg(), which is a internal function that returns ringparameters from both ->get_ringparam() and dev->cfg can't return a correct hds-thresh value.
The first patch fixes ethtool_ringparam_get_cfg() to set hds-thresh value correcltly.
The second patch adds random test for hds-thresh value. So that we can test 0 value for a hds-thresh properly.
v2:
- Skips set_hds_thresh_random test when hds-thresh-max value is too small. (2/2)
- Change random range from 1-MAX to 1-(MAX-1). (2/2)
Taehee Yoo (2):
net: ethtool: fix ethtool_ringparam_get_cfg() returns a hds_thresh
value always as 0.
selftests: drv-net: test random value for hds-thresh
net/ethtool/common.c | 1 + tools/testing/selftests/drivers/net/hds.py | 33 +++++++++++++++++++++- 2 files changed, 33 insertions(+), 1 deletion(-)
Signature · 1 line
2.34.1
PATCH v2 1/2 net: ethtool: fix ethtool_ringparam_get_cfg() returns a hds_thresh value always as 0.
When hds-thresh is configured, ethnl_set_rings() is called, and it calls ethtool_ringparam_get_cfg() to get ringparameters from .get_ringparam() callback and dev->cfg.
Both hds_config and hds_thresh values should be set from dev->cfg, not from .get_ringparam().
But ethtool_ringparam_get_cfg() sets only hds_config from dev->cfg. So, ethtool_ringparam_get_cfg() returns always a hds_thresh as 0.
If an input value of hds-thresh is 0, a hds_thresh value from ethtool_ringparam_get_cfg() are same. So ethnl_set_rings() does nothing and returns immediately.
It causes a bug that setting a hds-thresh value to 0 is not working.
Reproducer:
modprobe netdevsim
echo 1 > /sys/bus/netdevsim/new_device ethtool -G eth0 hds-thresh 100
ethtool -G eth0 hds-thresh 0
ethtool -g eth0
#hds-thresh value should be 0, but it shows 100.
The tools/testing/selftests/drivers/net/hds.py can test it too with applying a following patch for hds.py.
---
The ethtool command doesn't support hds-thresh option yet. So, RFC[1] patch is needed to be applied.
v2:
- No changes.
net/ethtool/common.c | 1 + 1 file changed, 1 insertion(+)
net/ethtool/common.c +1 -0
@@ -830,6 +830,7 @@ void ethtool_ringparam_get_cfg(struct net_device *dev,/* Driver gives us current state, we want to return current config */ kparam->tcp_data_split = dev->cfg->hds_config; + kparam->hds_thresh = dev->cfg->hds_thresh; }static void ethtool_init_tsinfo (struct kernel_ethtool_ts_info *info)
Signature · 1 line
2.34.1
PATCH v2 2/2 selftests: drv-net: test random value for hds-thresh
In reply to Taehee Yoohds.py has been testing 0(set_hds_thresh_zero()), MAX(set_hds_thresh_max()), GT(set_hds_thresh_gt()) values for hds-thresh. However if a hds-thresh value was already 0, set_hds_thresh_zero() can't test properly.
So, it tests random value first and then tests 0, MAX, GT values.
Testing bnxt:
TAP version 13
1..13
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_random
ok 6 hds.set_hds_thresh_zero
ok 7 hds.set_hds_thresh_max
ok 8 hds.set_hds_thresh_gt
ok 9 hds.set_xdp
ok 10 hds.enabled_set_xdp
ok 11 hds.ioctl
ok 12 hds.ioctl_set_xdp
ok 13 hds.ioctl_enabled_set_xdp
# Totals: pass:12 fail:0 xfail:0 xpass:0 skip:1 error:0
Testing lo:
TAP version 13
1..13
ok 1 hds.get_hds # SKIP tcp-data-split not supported by device ok 2 hds.get_hds_thresh # SKIP hds-thresh not supported by device ok 3 hds.set_hds_disable # SKIP ring-set not supported by the device ok 4 hds.set_hds_enable # SKIP ring-set not supported by the device ok 5 hds.set_hds_thresh_random # SKIP hds-thresh not supported by device
ok 6 hds.set_hds_thresh_zero # SKIP ring-set not supported by the device
ok 7 hds.set_hds_thresh_max # SKIP hds-thresh not supported by device
ok 8 hds.set_hds_thresh_gt # SKIP hds-thresh not supported by device ok 9 hds.set_xdp # SKIP tcp-data-split not supported by device ok 10 hds.enabled_set_xdp # SKIP tcp-data-split not supported by device
ok 11 hds.ioctl # SKIP tcp-data-split not supported by device ok 12 hds.ioctl_set_xdp # SKIP tcp-data-split not supported by device
ok 13 hds.ioctl_enabled_set_xdp # SKIP tcp-data-split not supported by device
# Totals: pass:0 fail:0 xfail:0 xpass:0 skip:13 error:0
---
v2:
- Skips set_hds_thresh_random test when hds-thresh-max value is too small.
- Change random range from 1-MAX to 1-(MAX-1).
tools/testing/selftests/drivers/net/hds.py | 33 +++++++++++++++++++++- 1 file changed, 32 insertions(+), 1 deletion(-)
tools/testing/selftests/drivers/net/hds.py +32 -1
@@ -6,7 +6,7 @@ import os from lib.py import ksft_run, ksft_exit, ksft_eq, ksft_raises, KsftSkipEx from lib.py import CmdExitFailure, EthtoolFamily, NlError from lib.py import NetDrvEnv -from lib.py import defer, ethtool, ip +from lib.py import defer, ethtool, ip, random def _get_hds_mode(cfg, netnl) -> str:
@@ -109,6 +109,36 @@ def set_hds_thresh_zero(cfg, netnl) -> None: ksft_eq(0, rings['hds-thresh']) +def set_hds_thresh_random(cfg, netnl) -> None: + try: + rings = netnl.rings_get({'header': {'dev-index': cfg.ifindex}}) + except NlError as e: + raise KsftSkipEx('ring-get not supported by device') + if 'hds-thresh' not in rings: + raise KsftSkipEx('hds-thresh not supported by device') + if 'hds-thresh-max' not in rings: + raise KsftSkipEx('hds-thresh-max not defined by device') + + if rings['hds-thresh-max'] < 2: + raise KsftSkipEx('hds-thresh-max is too small') + elif rings['hds-thresh-max'] == 2: + hds_thresh = 1 + else: + while True: + hds_thresh = random.randint(1, rings['hds-thresh-max'] - 1) + if hds_thresh != rings['hds-thresh']: + break + + try: + netnl.rings_set({'header': {'dev-index': cfg.ifindex}, 'hds-thresh': hds_thresh}) + except NlError as e: + if e.error == errno.EINVAL: + raise KsftSkipEx("hds-thresh-set not supported by the device") + elif e.error == errno.EOPNOTSUPP: + raise KsftSkipEx("ring-set not supported by the device") + rings = netnl.rings_get({'header': {'dev-index': cfg.ifindex}}) + ksft_eq(hds_thresh, rings['hds-thresh']) + def set_hds_thresh_max(cfg, netnl) -> None: try: rings = netnl.rings_get({'header': {'dev-index': cfg.ifindex}})
@@ -243,6 +273,7 @@ def main() -> None: get_hds_thresh, set_hds_disable, set_hds_enable, + set_hds_thresh_random, set_hds_thresh_zero, set_hds_thresh_max, set_hds_thresh_gt,
Signature · 1 line
2.34.1
This series was applied to netdev/net.git (main) by Jakub Kicinski <[email protected]>:
Hello:
This series was applied to netdev/net.git (main)
by Jakub Kicinski <[email protected]>:
Quoted 9 lines · On Fri, 4 Apr 2025 12:21:24 +0000 you wrote:
A hds-thresh value is not set correctly if input value is 0. The cause is that
ethtool_ringparam_get_cfg(), which is a internal function that returns ringparameters from both ->get_ringparam() anddev->cfgcan't return a correct hds-thresh value.The first patch fixes
ethtool_ringparam_get_cfg()to set hds-thresh value correcltly.[...]
Here is the summary with links:
- [v2,net,1/2] net: ethtool: fix ethtool_ringparam_get_cfg() returns a hds_thresh value always as 0.
https://git.kernel.org/netdev/net/c/216a61d33c07 - [v2,net,2/2] selftests: drv-net: test random value for hds-thresh https://git.kernel.org/netdev/net/c/22d3a63d5321
You are awesome, thank you!
Signature · 2 lines
Deet-doot-dot, I am a bot. https://korg.docs.kernel.org/patchwork/pwbot.html