netdev - Re: [RFC PATCH v3 3/5] Add ndo_hwtstamp_get/set support to vlan code path

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]

Message-ID: <20230406165055.egz32amam6o2bmqu@skbuf>
Date:   Thu, 6 Apr 2023 19:50:55 +0300
From:   Vladimir Oltean <vladimir.oltean@....com>
To:     Max Georgiev <glipus@...il.com>
Cc:     Jakub Kicinski <kuba@...nel.org>, kory.maincent@...tlin.com,
        netdev@...r.kernel.org, maxime.chevallier@...tlin.com,
        vadim.fedorenko@...ux.dev, richardcochran@...il.com,
        gerhard@...leder-embedded.com
Subject: Re: [RFC PATCH v3 3/5] Add ndo_hwtstamp_get/set support to vlan code
 path

On Thu, Apr 06, 2023 at 10:18:36AM -0600, Max Georgiev wrote:
> If I may, there are other ways to work around this inefficiency.
> Since kernel_hwtstamp_config was meant to be easily extendable,
> we can abuse it and add a flag field there. One of the flag values
> can indicate that the operation result structure was already copied
> to kernel_config->ifr by the function that received this kernel_config
> instance as a parameter, and that the content of the
> hwtstamp_config-related fields in kernel_config structure must
> be ignored when the function returns. It would complicate the
> implementation logic, but we'd avoid some unnecessary copy
> operations while converting *vlan components to the newer interface.
> Would it be a completely unreasonable approach?

No, I think that's fair game and a good idea. It would make the best
case better (SET request on a converted real_dev drops from 3 copies to 2),
while keeping the worst case the same (SET request on an unconverted
real_dev remains at 3 copies).