On 7/18/24 20:13, Kamal Heib wrote:
Add support for reporting fw status via the devlink health report.
Example:
# devlink health show pci/0000:02:00.0 reporter fw
pci/0000:02:00.0:
reporter fw
state healthy error 0 recover 0
# devlink health diagnose pci/0000:02:00.0 reporter fw
Mode: normal
Signed-off-by: Kamal Heib <kh...@redhat.com>
---
v2:
- Address comments from Jiri.
- Move the creation of the health report.
---
drivers/net/ethernet/intel/i40e/i40e.h | 1 +
.../net/ethernet/intel/i40e/i40e_devlink.c | 57 +++++++++++++++++++
.../net/ethernet/intel/i40e/i40e_devlink.h | 2 +
drivers/net/ethernet/intel/i40e/i40e_main.c | 14 +++++
4 files changed, 74 insertions(+)
[...]
+int i40e_devlink_create_health_reporter(struct i40e_pf *pf)
+{
+ struct devlink *devlink = priv_to_devlink(pf);
+ struct device *dev = &pf->pdev->dev;
+ int rc = 0;
+
+ devl_lock(devlink);
if you are going to have just one health reporter, you could just use
devlink_health_reporter_create() and avoid lock+unlock calls here
+ pf->fw_health_report =
+ devl_health_reporter_create(devlink, &i40e_fw_reporter_ops, 0,
pf);
+ if (IS_ERR(pf->fw_health_report)) {
+ rc = PTR_ERR(pf->fw_health_report);
+ dev_err(dev, "Failed to create fw reporter, err = %d\n", rc);
you are not zeroing pf->fw_health_report here, so ...
+ }
+ devl_unlock(devlink);
+
+ return rc;
+}
[snip]
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -15370,6 +15370,9 @@ static bool i40e_check_recovery_mode(struct i40e_pf *pf)
dev_crit(&pf->pdev->dev, "Firmware recovery mode detected. Limiting
functionality.\n");
dev_crit(&pf->pdev->dev, "Refer to the Intel(R) Ethernet Adapters
and Devices User Guide for details on firmware recovery mode.\n");
set_bit(__I40E_RECOVERY_MODE, pf->state);
+ if (pf->fw_health_report)
+ devlink_health_report(pf->fw_health_report,
... you could possibly call devlink_health_report() with ERR_PTR as
the first argument
+ "recovery mode detected", pf);
return true;
}
