netdev - Re: [PATCH net-next 06/12] iavf: Add trace while removing device

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: <20211125075701.7b67ae7f@kicinski-fedora-pc1c0hjn.dhcp.thefacebook.com>
Date:   Thu, 25 Nov 2021 07:57:01 -0800
From:   Jakub Kicinski <kuba@...nel.org>
To:     Stefan Assmann <sassmann@...hat.com>
Cc:     Tony Nguyen <anthony.l.nguyen@...el.com>, davem@...emloft.net,
        Jedrzej Jagielski <jedrzej.jagielski@...el.com>,
        netdev@...r.kernel.org,
        Arkadiusz Kubalewski <arkadiusz.kubalewski@...el.com>,
        Konrad Jankowski <konrad0.jankowski@...el.com>
Subject: Re: [PATCH net-next 06/12] iavf: Add trace while removing device

On Thu, 25 Nov 2021 16:43:49 +0100 Stefan Assmann wrote:
> On 2021-11-25 07:13, Jakub Kicinski wrote:
> > On Thu, 25 Nov 2021 07:50:49 +0100 Stefan Assmann wrote:  
> > > From personal experience I'd say this piece of information has value,
> > > especially when debugging it can be interesting to know exactly when
> > > the driver was removed.  
> > 
> > But there isn't anything specific to iavf here, right? If it really 
> > is important then core should be doing the printing for all drivers.
> > 
> > Actually, I can't come up with any uses for this print on the spot.
> > What debugging scenarios do you have in mind?  
> 
> There was a lot of trouble with iavf in terms of device reset, device
> unbinding (DPDK), stress testing of driver load/unload issues. When
> looking through the crash logs it was not always easy to determine if
> the driver was still loaded.
> Especially on problems that weren't easy to reproduce.

That's a slippery slope, historically we were always pushing for
avoiding informational prints. Otherwise every driver reconfig would
result in a line in the logs.

> So for iavf having that information would have been valuable. Not sure
> if that justifies a PCI core message or if others might find that too
> verbose.

So what you're saying is from your experience iavf is of significantly
poorer quality than other vendors' drivers?