Can networking trees carry temporary patches?

Can networking trees carry temporary patches?

[Tetsuo Handa]

I was carrying debug printk() patches [1][2] in linux-next tree via my tree
in order to debug "unregister_netdevice: waiting for DEV to become free"
problem [3].

While these patches had been a lot helpful, now that easily reproducible cases
has been fixed, this problem is hardly reproduced in linux-next tree recently;
giving me no hints for debugging remaining cases. However, this problem is still
reproduced in networking trees.

Therefore, I appreciate if networking trees can carry temporary patches
(patches which are not intended to be merged into the linux.git tree).

[1] https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit?id=58b4f8f612187b0d9583b5549044734ecd7071eb
[2] https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit?id=02fb7fbf5db47bf5c1736b692888232f5f6bf80e
[3] https://syzkaller.appspot.com/bug?extid=881d65229ca4f9ae8c84

Re: Can networking trees carry temporary patches?

[Jakub Kicinski]

On Tue, 24 Mar 2026 19:55:18 +0900 Tetsuo Handa wrote:
> I was carrying debug printk() patches [1][2] in linux-next tree via my tree
> in order to debug "unregister_netdevice: waiting for DEV to become free"
> problem [3].
> 
> While these patches had been a lot helpful, now that easily reproducible cases
> has been fixed, this problem is hardly reproduced in linux-next tree recently;
> giving me no hints for debugging remaining cases. However, this problem is still
> reproduced in networking trees.
> 
> Therefore, I appreciate if networking trees can carry temporary patches
> (patches which are not intended to be merged into the linux.git tree).
> 
> [1] https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit?id=58b4f8f612187b0d9583b5549044734ecd7071eb
> [2] https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit?id=02fb7fbf5db47bf5c1736b692888232f5f6bf80e
> [3] https://syzkaller.appspot.com/bug?extid=881d65229ca4f9ae8c84

Have you see netdev_hold() / netdev_put() ?

Re: Can networking trees carry temporary patches?

[Tetsuo Handa]

On 2026/03/26 11:22, Jakub Kicinski wrote:
> On Tue, 24 Mar 2026 19:55:18 +0900 Tetsuo Handa wrote:
>> I was carrying debug printk() patches [1][2] in linux-next tree via my tree
>> in order to debug "unregister_netdevice: waiting for DEV to become free"
>> problem [3].
>>
>> While these patches had been a lot helpful, now that easily reproducible cases
>> has been fixed, this problem is hardly reproduced in linux-next tree recently;
>> giving me no hints for debugging remaining cases. However, this problem is still
>> reproduced in networking trees.
>>
>> Therefore, I appreciate if networking trees can carry temporary patches
>> (patches which are not intended to be merged into the linux.git tree).
>>
>> [1] https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit?id=58b4f8f612187b0d9583b5549044734ecd7071eb
>> [2] https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit?id=02fb7fbf5db47bf5c1736b692888232f5f6bf80e
>> [3] https://syzkaller.appspot.com/bug?extid=881d65229ca4f9ae8c84
> 
> Have you see netdev_hold() / netdev_put() ?

Of course. I made these patches because netdev_hold() / netdev_put() are not
sufficient for debugging. There are many reports which contains only
"unregister_netdevice: waiting for DEV to become free" line.