netdev - Re: [PATCH net 1/2] s390/qeth: update cached link

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:   Thu, 4 Aug 2022 13:27:42 -0700
From:   Jakub Kicinski <kuba@...nel.org>
To:     Andrew Lunn <andrew@...n.ch>,
        Alexandra Winter <wintera@...ux.ibm.com>
Cc:     David Miller <davem@...emloft.net>, netdev@...r.kernel.org,
        linux-s390@...r.kernel.org, Heiko Carstens <hca@...ux.ibm.com>,
        Thorsten Winkler <twinkler@...ux.ibm.com>
Subject: Re: [PATCH net 1/2] s390/qeth: update cached link_info for ethtool

On Thu, 4 Aug 2022 15:08:11 +0200 Andrew Lunn wrote:
> On Thu, Aug 04, 2022 at 10:53:33AM +0200, Alexandra Winter wrote:
> > Thank you Andrew for the review. I fully understand your point.
> > I would like to propose that I put that on my ToDo list and fix
> > that in a follow-on patch to net-next.
> > 
> > The fields in the link_info control blocks are used today to generate
> > other values (e.g. supported speed) which will not work with *_UNKNOWN,
> > so the follow-on patch will be more than just 2 lines.  
> 
> So it sounds like your code is all backwards around. If you know what
> the hardware is, you know the supported link modes are, assuming its
> not an SFP and the SFP module is not plugged in. Those link modes
> should be independent of if the link is up or not. speed/duplex is
> only valid when the link is up and negotiation has finished.

To make sure I understand - the code depends on the speed and duplex
being set to something specific when the device is _down_? Can this be
spelled out more clearly in the commit message?

> Since this is for net, than yes, maybe it would be best to go with a
> minimal patch to make your backwards around code work. But for
> net-next, you really should fix this properly. 

Then again this patch doesn't look like a regression fix (and does not
have a fixes tag). Channeling my inner Greg I'd say - fix this right and
then worry about backports later.