netdev - Re: [net 1/8] igb: Update queue reinit function to call dev

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: <1386004696.1516.15.camel@bwh-desktop.uk.level5networks.com>
Date:	Mon, 2 Dec 2013 17:18:16 +0000
From:	Ben Hutchings <bhutchings@...arflare.com>
To:	"Wyborny, Carolyn" <carolyn.wyborny@...el.com>
CC:	David Miller <davem@...emloft.net>,
	"Kirsher, Jeffrey T" <jeffrey.t.kirsher@...el.com>,
	"netdev@...r.kernel.org" <netdev@...r.kernel.org>,
	"gospo@...hat.com" <gospo@...hat.com>,
	"sassmann@...hat.com" <sassmann@...hat.com>
Subject: Re: [net 1/8] igb: Update queue reinit function to call dev_close
 when init of queues fails

On Mon, 2013-12-02 at 16:55 +0000, Wyborny, Carolyn wrote:
> > -----Original Message-----
> > From: David Miller [mailto:davem@...emloft.net]
> > Sent: Friday, November 29, 2013 12:48 PM
> > To: Kirsher, Jeffrey T
> > Cc: Wyborny, Carolyn; netdev@...r.kernel.org; gospo@...hat.com;
> > sassmann@...hat.com
> > Subject: Re: [net 1/8] igb: Update queue reinit function to call dev_close when
> > init of queues fails
> > 
> > From: Jeff Kirsher <jeffrey.t.kirsher@...el.com>
> > Date: Wed, 27 Nov 2013 23:35:55 -0800
> > 
> > > From: Carolyn Wyborny <carolyn.wyborny@...el.com>
> > >
> > > This patch adds a call to dev_close if the queue reinit fails in order
> > > to make clearer to the user that the device is down.
> > >
> > > Signed-off-by: Carolyn Wyborny <carolyn.wyborny@...el.com>
> > > Tested-by: Aaron Brown <aaron.f.brown@...el.com>
> > > Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@...el.com>
> > 
> > This is a very bad approach to this problem.
> > 
> > Users absolutely do not expect their entire interface to go down simply because
> > an ethtool request cannot be satisfied.  This is extremely poor quality of
> > implementation.
> > 
> > And in this specific case it absolutely is not necessary.
> > 
> > The only thing that can fail is the queue allocation, so make a function which can
> > preserve the previous configuration if the queue allocation fails.  How about
> > "igb_reinit_interrupt_scheme".
> > 
> > Don't free the q vectors until the very last moment, when you know that the
> > allocation of the new q vectors has succeeded.
> > 
> > I'm not applying this patch, it needs to be reimplemented more sanely, using the
> > above suggestions or similar.
> > 
> 
> I did this per Ben's suggestion on Laura's original patch.  You didn't
> argue with it at the time.  I don't think I misunderstood his
> suggestion, but if so, I'll rework it.

My concerns were that after a failure where the interface is effectively
down,
- the interface must be in a consistent state where is it safe to
reconfigure the interface again or to unbind the driver, and
- it should be obviously down so the user can then try to bring it up
again.

David is saying that you should implement the reconfiguration in such a
way that you can always roll back to the previous working state in case
of a failure (which would make my concerns moot).  This is definitely a
good goal but I'm not convinced that it's always possible.

Ben.

-- 
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html