opticsmon: Fix runaway loop in on_link_change()

When on_link_change() gets called with a netdev that would be monitored
but hasn't entered zpci_list yet, reloads is 1 after the loops and
a reload occurs. Then the netdev is found in the list and reloads
becomes -1 which incorrectly triggers more reloads until underflow.
Fix this by returning once the device is found. Also just check for
reloads being larger than zero.

Fixes: c34adb9cab ("opticsmon: Introduce opticsmon tool")
Reviewed-by: Halil Pasic <pasic@linux.ibm.com>
Signed-off-by: Niklas Schnelle <schnelle@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
This commit is contained in:
Niklas Schnelle
2024-12-06 15:28:08 +01:00
committed by Jan Höppner
parent 79de37bbc8
commit dff965465c

View File

@@ -280,16 +280,15 @@ void on_link_change(struct zpci_netdev *netdev, void *arg)
if (!ctx->zpci_list || util_list_is_empty(ctx->zpci_list))
zpci_list_reload(&ctx->zpci_list);
reload:
find:
util_list_iterate(ctx->zpci_list, zdev) {
for (i = 0; i < zdev->num_netdevs; i++) {
if (!strcmp(zdev->netdevs[i].name, netdev->name)) {
reloads--;
/* Skip data collection if operational state is
* unchanged
*/
if (zdev->netdevs[i].operstate == netdev->operstate)
continue;
return;
/* Update operation state for VFs even though
* they are skipped just for a consistent view
*/
@@ -297,14 +296,15 @@ reload:
/* Only collect optics data for PFs */
if (!zpci_is_vf(zdev))
dump_adapter_data(ctx, zdev);
return;
}
}
}
/* Might be a new device, reload list of devices and retry */
if (reloads) {
if (reloads > 0) {
zpci_list_reload(&ctx->zpci_list);
reloads--;
goto reload;
goto find;
}
}