kernel/linux.git/drivers/net/ethernet/amazon, branch v6.18.21

net/ena: fix missing lock when update devlink params

2026-01-17T15:35:24+00:00

[ Upstream commit 8da901ffe497a53fa4ecc3ceed0e6d771586f88e ] Fix assert lock warning while calling devl_param_driverinit_value_set() in ena. WARNING: net/devlink/core.c:261 at devl_assert_locked+0x62/0x90, CPU#0: kworker/0:0/9 CPU: 0 UID: 0 PID: 9 Comm: kworker/0:0 Not tainted 6.19.0-rc2+ #1 PREEMPT(lazy) Hardware name: Amazon EC2 m8i-flex.4xlarge/, BIOS 1.0 10/16/2017 Workqueue: events work_for_cpu_fn RIP: 0010:devl_assert_locked+0x62/0x90 Call Trace: devl_param_driverinit_value_set+0x15/0x1c0 ena_devlink_alloc+0x18c/0x220 [ena] ? __pfx_ena_devlink_alloc+0x10/0x10 [ena] ? trace_hardirqs_on+0x18/0x140 ? lockdep_hardirqs_on+0x8c/0x130 ? __raw_spin_unlock_irqrestore+0x5d/0x80 ? __raw_spin_unlock_irqrestore+0x46/0x80 ? devm_ioremap_wc+0x9a/0xd0 ena_probe+0x4d2/0x1b20 [ena] ? __lock_acquire+0x56a/0xbd0 ? __pfx_ena_probe+0x10/0x10 [ena] ? local_clock+0x15/0x30 ? __lock_release.isra.0+0x1c9/0x340 ? mark_held_locks+0x40/0x70 ? lockdep_hardirqs_on_prepare.part.0+0x92/0x170 ? trace_hardirqs_on+0x18/0x140 ? lockdep_hardirqs_on+0x8c/0x130 ? __raw_spin_unlock_irqrestore+0x5d/0x80 ? __raw_spin_unlock_irqrestore+0x46/0x80 ? __pfx_ena_probe+0x10/0x10 [ena] ...... Fixes: 816b52624cf6 ("net: ena: Control PHC enable through devlink") Signed-off-by: Frank Liang Reviewed-by: David Arinzon Reviewed-by: Jiri Pirko Link: https://patch.msgid.link/20251231145808.6103-1-xiliang@redhat.com Signed-off-by: Jakub Kicinski Signed-off-by: Sasha Levin

net: ena: return 0 in ena_get_rxfh_key_size() when RSS hash key is not configurable

2025-09-30T01:33:53+00:00

In EC2 instances where the RSS hash key is not configurable, ethtool shows bogus RSS hash key since ena_get_rxfh_key_size() unconditionally returns ENA_HASH_KEY_SIZE. Commit 6a4f7dc82d1e ("net: ena: rss: do not allocate key when not supported") added proper handling for devices that don't support RSS hash key configuration, but ena_get_rxfh_key_size() has been unchanged. When the RSS hash key is not configurable, return 0 instead of ENA_HASH_KEY_SIZE to clarify getting the value is not supported. Tested on m5 instance families. Without patch: # ethtool -x ens5 | grep -A 1 "RSS hash key" RSS hash key: 00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00 With patch: # ethtool -x ens5 | grep -A 1 "RSS hash key" RSS hash key: Operation not supported Fixes: 6a4f7dc82d1e ("net: ena: rss: do not allocate key when not supported") Signed-off-by: Kohei Enju Link: https://patch.msgid.link/20250929050247.51680-1-enjuk@amazon.com Signed-off-by: Jakub Kicinski

net: Fix typos

2025-07-25T17:29:07+00:00

Fix typos in comments and error messages. Signed-off-by: Bjorn Helgaas Reviewed-by: David Arinzon Reviewed-by: Simon Horman Link: https://patch.msgid.link/20250723201528.2908218-1-helgaas@kernel.org Signed-off-by: Jakub Kicinski

net: ena: View PHC stats using debugfs

2025-06-19T01:57:29+00:00

Add an entry named `phc_stats` to view the PHC statistics. If PHC is enabled, the stats are printed, as below: phc_cnt: 0 phc_exp: 0 phc_skp: 0 phc_err_dv: 0 phc_err_ts: 0 If PHC is disabled, no statistics will be displayed. Signed-off-by: David Arinzon Link: https://patch.msgid.link/20250617110545.5659-9-darinzon@amazon.com Signed-off-by: Jakub Kicinski

net: ena: Add debugfs support to the ENA driver

2025-06-19T01:57:29+00:00

Adding the base directory of debugfs to the driver. In order for the folder to be unique per driver instantiation, the chosen name is the device name. This commit contains the initialization and the base folder. The creation of the base folder may fail, but is considered non-fatal. Signed-off-by: David Arinzon Link: https://patch.msgid.link/20250617110545.5659-8-darinzon@amazon.com Signed-off-by: Jakub Kicinski

net: ena: Control PHC enable through devlink

2025-06-19T01:57:29+00:00

Add the capability to set parameters through the devlink framework. The parameter used for controlling PHC (enable/disable) details are as follows: - Name: enable_phc - Type: Boolean (true - enable/false - disable) - Mode: DEVLINK_PARAM_CMODE_DRIVERINIT - Effect: Changes take place during driver initialization, any changes require a devlink reload to take effect. Signed-off-by: David Arinzon Link: https://patch.msgid.link/20250617110545.5659-7-darinzon@amazon.com Signed-off-by: Jakub Kicinski

net: ena: Add devlink port support

2025-06-19T01:57:28+00:00

Add the basic functionality to support devlink port for devlink model completeness purposes. Current support is for registration/un-registration. Signed-off-by: David Arinzon Link: https://patch.msgid.link/20250617110545.5659-5-darinzon@amazon.com Signed-off-by: Jakub Kicinski

net: ena: Add device reload capability through devlink

2025-06-19T01:57:28+00:00

Adding basic devlink capability support of reloading the driver. This capability is required to support driver init type devlink params (DEVLINK_PARAM_CMODE_DRIVERINIT). Such params require reloading of the driver (destroy/restore sequence). The reloading is done by the devlink framework using the hooks provided by the driver. Signed-off-by: David Arinzon Link: https://patch.msgid.link/20250617110545.5659-4-darinzon@amazon.com Signed-off-by: Jakub Kicinski

net: ena: PHC silent reset

2025-06-19T01:57:28+00:00

Each PHC device kernel registration receives a unique kernel index, which is associated with a new PHC device file located at "/dev/ptp". This device file serves as an interface for obtaining PHC timestamps. Examples of tools that use "/dev/ptp" include testptp [1] and chrony [2]. A reset flow may occur in the ENA driver while PHC is active. During a reset, the driver will unregister and then re-register the PHC device with the kernel. Under race conditions, particularly during heavy PHC loads, the driver’s reset flow might complete faster than the kernel’s PHC unregister/register process. This can result in the PHC index being different from what it was prior to the reset, as the PHC index is selected using kernel ID allocation [3]. While driver rmmod/insmod are done by the user, a reset may occur at anytime, without the user awareness, consequently, the driver might receive a new PHC index after the reset, potentially compromising the user experience. To prevent this issue, the PHC flow will detect the reset during PHC destruction and will skip the PHC unregister/register calls to preserve the kernel PHC index. During the reset flow, any attempt to get a PHC timestamp will fail as expected, but the kernel PHC index will remain unchanged. [1]: https://github.com/torvalds/linux/blob/v6.1/tools/testing/selftests/ptp/testptp.c [2]: https://github.com/mlichvar/chrony [3]: https://www.kernel.org/doc/html/latest/core-api/idr.html Signed-off-by: Amit Bernstein Signed-off-by: David Arinzon Link: https://patch.msgid.link/20250617110545.5659-3-darinzon@amazon.com Signed-off-by: Jakub Kicinski

net: ena: Add PHC support in the ENA driver

2025-06-19T01:57:28+00:00

The ENA driver will be extended to support the new PHC feature using ptp_clock interface [1]. this will provide timestamp reference for user space to allow measuring time offset between the PHC and the system clock in order to achieve nanosecond accuracy. [1] - https://www.kernel.org/doc/html/latest/driver-api/ptp.html Signed-off-by: Amit Bernstein Signed-off-by: David Arinzon Link: https://patch.msgid.link/20250617110545.5659-2-darinzon@amazon.com Signed-off-by: Jakub Kicinski