netdev - Re: [PATCH net-next 2/3] nfp: add support for link auto negotiation

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]

Message-ID: <20220927015226.GA22642@nj-rack01-04.nji.corigine.com>
Date:   Tue, 27 Sep 2022 09:52:26 +0800
From:   Yinjun Zhang <yinjun.zhang@...igine.com>
To:     Jakub Kicinski <kuba@...nel.org>
Cc:     Simon Horman <simon.horman@...igine.com>,
        David Miller <davem@...emloft.net>,
        Paolo Abeni <pabeni@...hat.com>,
        "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
        oss-drivers <oss-drivers@...igine.com>,
        Fei Qin <fei.qin@...igine.com>
Subject: Re: [PATCH net-next 2/3] nfp: add support for link auto negotiation

On Mon, Sep 26, 2022 at 06:38:40PM -0700, Jakub Kicinski wrote:
> On Tue, 27 Sep 2022 09:13:53 +0800 Yinjun Zhang wrote:
> > On Mon, Sep 26, 2022 at 09:25:47AM -0700, Jakub Kicinski wrote:
> > > On Sat, 24 Sep 2022 10:45:30 +0800 Yinjun Zhang wrote:  
> > > > Not only check if it's flower, but also check if it's sp_indiff when
> > > > it's not flower by parsing the tlv caps.  
> > > 
> > > Seems bogus. The speed independence is a property of the whole FW image,
> > > you record it in the pf structure.  
> > 
> > It's indeed a per-fw property, but we don't have existing way to expose
> > per-fw capabilities to driver currently, so use per-vnic tlv caps here.
> > Maybe define a new fw symbol is a choice, but my concern is it's not
> > visible to netvf driver.
> > Any suggestion is welcomed.
> 
> Why not put an rtsym with the value in the FW? That'd be my first
> go-to way of communicating information about the FW as a whole.

Like I said, the VF driver cannot read the rtsym, so it's not a perfect
way as exposing a per-FW property. But for this case, it's OK since VF
doesn't need this info. I'll try this way. Thanks.