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  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]
Date:	Fri, 20 Feb 2015 22:07:55 -0500
From:	Andy Gospodarek <gospo@...ulusnetworks.com>
To:	David Miller <davem@...emloft.net>
Cc:	stephen@...workplumber.org, Yanjun.Zhu@...driver.com,
	romieu@...zoreil.com, netdev@...r.kernel.org, mst@...hat.com,
	jasowang@...hat.com, viro@...iv.linux.org.uk,
	sergei.shtylyov@...entembedded.com, jonathon.reinhart@...il.com
Subject: Re: [PATCH 1/1] tun: change speed from 10M to dynamically configured

On Fri, Feb 20, 2015 at 03:07:49PM -0500, David Miller wrote:
> From: Andy Gospodarek <gospo@...ulusnetworks.com>
> Date: Thu, 19 Feb 2015 21:35:00 -0500
> 
> > On Thu, Feb 19, 2015 at 03:40:41PM -0500, David Miller wrote:
> >> I see three realistic options:
> >> 
> >> 1) Create a link state indication which means "I am a software device,
> >>    so I don't really have a link state in the traditional sense"
> >> 
> >> 2) Don't implement the link set/get operations at all on software
> >>    devices.
> >> 
> >>    People can use ETHTOOL_GLINK to see if the thing is "up"
> >> 
> >> 3) Propagate the ultimate physical transport parameters into what
> >>    the software device advertises.
> >> 
> >> It's important to carefully pick one of these, and consistently apply
> >> it to all of our software devices.
> >> 
> >> I don't want TUN doing one thing, ipv4 tunnels doing another, etc.
> > 
> > I would prefer option #3 as it relates to the ability to transport
> > offload statistics and parameters to software devices.  This could be
> > useful for hardware with some form of offload vxlan/gre/etc that may be
> > backed by hardare statistics (a callback to the upper device's ndo op
> > could be made by the native hardware driver) or any other device without
> > a proper in-kernel driver like a userspace user of tuntap that might be
> > backed by something like OpenVPN.
> > 
> > That might help provide some more detailed, ethtool-like statistics in
> > the format that is more easily readable by common monitoring tools
> > without the need to have those applications look at anything except an
> > ethtool/kernel interface.
> 
> netdev_feature_t like propagation is tricky because offloading
> depends upon whether the device can do the offload "even in the
> presence of encapsulation X".  So I don't see it as useful for
> that.
> 
> But for things that are largely advisory types of information like
> "physical link state", it works in a much more straightforward manner.
> 
> Link duplex and speed being wrong on a software device doesn't stop
> communication, but incorrect hw offload feature flags might.

Ah yes, that is an excellent point.  Trying to get too fancy is probably
a bad idea when thinking about the exception cases.  In my example the
driver for the vxlan offload device should be the one tracking this
rather than propagating information up to a upper-layer software device.

The software devices that may not have any true hardware backing in the
kernel (OpenVPN, etc) are the best consumers for this type of
infrastructure.

--
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

Powered by blists - more mailing lists