netdev - Re: [patch net-next v3 05/13] net: devlink: track netdev with devlink

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

Date:   Wed, 2 Nov 2022 12:13:18 +0100
From:   Jiri Pirko <jiri@...nulli.us>
To:     Jakub Kicinski <kuba@...nel.org>
Cc:     netdev@...r.kernel.org, davem@...emloft.net, pabeni@...hat.com,
        edumazet@...gle.com, tariqt@...dia.com, moshe@...dia.com,
        saeedm@...dia.com, linux-rdma@...r.kernel.org
Subject: Re: [patch net-next v3 05/13] net: devlink: track netdev with
 devlink_port assigned

Tue, Nov 01, 2022 at 05:25:42PM CET, kuba@...nel.org wrote:
>On Mon, 31 Oct 2022 13:42:40 +0100 Jiri Pirko wrote:
>> +/*
>> + * Driver should use this to assign devlink port instance to a netdevice
>> + * before it registers the netdevice. Therefore devlink_port is static
>> + * during the netdev lifetime after it is registered.
>> + */
>> +#define SET_NETDEV_DEVLINK_PORT(dev, _devlink_port)		\
>> +({								\
>> +	WARN_ON(dev->reg_state != NETREG_UNINITIALIZED);	\
>> +	((dev)->devlink_port = (_devlink_port));		\
>> +})
>
>The argument wrapping is inconsistent here - dev is in brackets
>on the second line but not on the first. _devlink_port is prefixed 
>with underscore and dev is not. Let's make this properly secure
>and define a local var for dev.

"_" is there because the struct field is named the same. I will change
this to "port" if you don't like "_", no problem.

I will put the first dev to (). Not sure what you mean by "local var",
but I believe that "(dev)" is secure.

>
>> @@ -10107,6 +10107,7 @@ int register_netdevice(struct net_device *dev)
>>  	return ret;
>>  
>>  err_uninit:
>> +	call_netdevice_notifiers(NETDEV_PRE_UNINIT, dev);
>
>I think we should make this symmetric with POST_INIT.
>IOW PRE_UNINIT should only be called if POST_INIT was called.