Re: kernel BUG at net/core/net-sysfs.c:LINE!

[wanghai (M)]

thanks , Can it be fixed like this?

diff --git a/net/core/net-sysfs.c b/net/core/net-sysfs.c
index 4ff661f..e609c8d 100644
--- a/net/core/net-sysfs.c
+++ b/net/core/net-sysfs.c
@@ -1745,16 +1745,21 @@ int netdev_register_kobject(struct net_device *ndev)

ÂÂÂÂÂÂÂ error = device_add(dev);
ÂÂÂÂÂÂÂ if (error)
-ÂÂÂÂÂÂÂÂÂÂÂÂÂÂ return error;
+ÂÂÂÂÂÂÂÂÂÂÂÂÂÂ goto error_put_device;

ÂÂÂÂÂÂÂ error = register_queue_kobjects(ndev);
-ÂÂÂÂÂÂ if (error) {
-ÂÂÂÂÂÂÂÂÂÂÂÂÂÂ device_del(dev);
-ÂÂÂÂÂÂÂÂÂÂÂÂÂÂ return error;
-ÂÂÂÂÂÂ }
+ÂÂÂÂÂÂ if (error)
+ÂÂÂÂÂÂÂÂÂÂÂÂÂÂ goto error_device_del;

ÂÂÂÂÂÂÂ pm_runtime_set_memalloc_noio(dev, true);

+ÂÂÂÂÂÂ return 0;
+
+error_device_del:
+ÂÂÂÂÂÂ device_del(dev);
+error_put_device:
+ÂÂÂÂÂÂ ndev->reg_state = NETREG_RELEASED;
+ÂÂÂÂÂÂ put_device(dev);
ÂÂÂÂÂÂÂ return error;
Â}

å 2019/3/24 1:16, Andy Shevchenko åé:
> Nice.
>
> I looked briefly in the flow of this report and it looks like the patch above
> should be reverted.
>
> The problem is not so easy to fix. One approach is to initialize device
> (and thus kobject) somewhere in alloc_netdev() and put device in free_netdev()
> respectively, but this might produce more interesting regressions.