netdev - Re: [RFC 1/3] net: dsa: add basic support for VLAN ndo

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:	Sun, 31 May 2015 09:14:01 -0700
From:	Scott Feldman <sfeldma@...il.com>
To:	Vivien Didelot <vivien.didelot@...oirfairelinux.com>
Cc:	Guenter Roeck <linux@...ck-us.net>,
	Or Gerlitz <gerlitz.or@...il.com>,
	netdev <netdev@...r.kernel.org>,
	Florian Fainelli <f.fainelli@...il.com>,
	Andrew Lunn <andrew@...n.ch>,
	Jérome Oufella 
	<jerome.oufella@...oirfairelinux.com>,
	Chris Healy <cphealy@...il.com>, Jiri Pirko <jiri@...nulli.us>,
	Roopa Prabhu <roopa@...ulusnetworks.com>
Subject: Re: [RFC 1/3] net: dsa: add basic support for VLAN ndo

On Fri, May 29, 2015 at 3:57 PM, Vivien Didelot
<vivien.didelot@...oirfairelinux.com> wrote:

> Scott,
>
> Can I ask details about this NETIF_F_HW_SWITCH_OFFLOAD flag? Why is it
> unneeded?

Sorry, the commit message for that changes wasn't very good because it
didn't capture the email discussion leading up.

The issue I had with the flag was it could be set/cleared at any time
and it wasn't clear what the driver should do in response.

Let's say flag is on for swp1, on startup.  swp1 driver offloads L2/L3
to the device.  Next user clears flag with ethtool.  What does that
mean?  That existing offloads should be removed from device?  Or just
stop adding/removing offloads while flag is off?  Stopping offloads
can cause problems for L3 offload by breaking LPM rules.  (You could
end up with a shorter prefix route in HW and a longer prefix route in
SW, and the HW route will win, which would be bad).

So Roopa wants to revisit the flag.  I might be that tightening up
when the flag can be set/cleared will fix the sequencing issues.

-scott
--
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