netdev - Re: [patch net 3/3] devlink: fix a deadlock with nested instances during namespace remove

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date: Wed, 10 May 2023 11:27:29 -0700
From: Jacob Keller <jacob.e.keller@...el.com>
To: Jakub Kicinski <kuba@...nel.org>, Jiri Pirko <jiri@...nulli.us>
CC: <netdev@...r.kernel.org>, <pabeni@...hat.com>, <davem@...emloft.net>,
	<edumazet@...gle.com>, <saeedm@...dia.com>, <moshe@...dia.com>
Subject: Re: [patch net 3/3] devlink: fix a deadlock with nested instances
 during namespace remove



On 5/9/2023 8:24 PM, Jakub Kicinski wrote:
> On Tue,  9 May 2023 12:09:36 +0200 Jiri Pirko wrote:
>> The commit 565b4824c39f ("devlink: change port event netdev notifier
>> from per-net to global") changed original per-net notifier to be global
>> which fixed the issue of non-receiving events of netdev uninit if that
>> moved to a different namespace. That worked fine in -net tree.
>>
>> However, later on when commit ee75f1fc44dd ("net/mlx5e: Create
>> separate devlink instance for ethernet auxiliary device") and
>> commit 72ed5d5624af ("net/mlx5: Suspend auxiliary devices only in
>> case of PCI device suspend") were merged, a deadlock was introduced
>> when removing a namespace with devlink instance with another nested
>> instance.
>>
>> Here there is the bad flow example resulting in deadlock with mlx5:
>> net_cleanup_work -> cleanup_net (takes down_read(&pernet_ops_rwsem) ->
>> devlink_pernet_pre_exit() -> devlink_reload() ->
>> mlx5_devlink_reload_down() -> mlx5_unload_one_devl_locked() ->
>> mlx5_detach_device() -> del_adev() -> mlx5e_remove() ->
>> mlx5e_destroy_devlink() -> devlink_free() ->
>> unregister_netdevice_notifier() (takes down_write(&pernet_ops_rwsem)
> 
> Why don't we have a single, static notifier for all of devlink?
> Why the per device/per port notifiers?
> 
> We have the devlink port pointer in struct net_device, resolving from
> a global event to the correct devlink instance is trivial.

Ok, so if I think through all the possibilities:

1. Originally we had a namespace specific notifier for each struct devlink.

2. Then we added a global notifier for all namespaces, but still had one
for each devlink.

3. Then Jiri's proposal here was a per-namespace notifier per port, but
we then we follow the namespace when the netdev changes namespaces.



But its simpler to just have a single global notifier that is setup once
for all devlinks. Then, when we get a notification, instead of looking
up the devlink instance from the notifier using container_of we look it
up through the netdev->devlink_port connection.

Ya that seems a lot simpler and requires only one notifier instead, and
wouldn't require namespace following code.

I think that makes a lot of sense.