netdev - Re: [RFC PATCH net-next 5/6] net: ethtool: add a mutex protecting RSS contexts

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:   Mon, 3 Apr 2023 15:03:12 -0700
From:   Jakub Kicinski <kuba@...nel.org>
To:     <edward.cree@....com>
Cc:     <linux-net-drivers@....com>, <davem@...emloft.net>,
        <pabeni@...hat.com>, <edumazet@...gle.com>,
        Edward Cree <ecree.xilinx@...il.com>, <netdev@...r.kernel.org>,
        <habetsm.xilinx@...il.com>, <sudheer.mogilappagari@...el.com>
Subject: Re: [RFC PATCH net-next 5/6] net: ethtool: add a mutex protecting
 RSS contexts

On Mon, 3 Apr 2023 17:33:02 +0100 edward.cree@....com wrote:
>  	u32			rss_ctx_max_id;
>  	struct idr		rss_ctx;
> +	struct mutex		rss_lock;

Argh, the mutex doubles the size of the state, and most drivers don't
implement this feature.  My thinking was to add a "ethtool state"
pointer to net_device which will be allocated by ethtool on demand
and can hold all ethtool related state.

For psychological reasons primarily (IOW I feel like people shy away
from storing state in the core because it feels expensive to add stuff
to net_device while it would not seem expensive to add it to struct
ethtool_netdev_state).

Maybe we can do the on-demand allocation later - but could we at least
wrap the ethtool-related fields in a separate struct to hold them
together?