netdev - Re: Please revert removal of /sys/class/net/*/features

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: <20120410134633.4456a1b0@nehalam.linuxnetplumber.net>
Date:	Tue, 10 Apr 2012 13:46:33 -0700
From:	Stephen Hemminger <shemminger@...tta.com>
To:	Ben Hutchings <bhutchings@...arflare.com>
Cc:	Michał Mirosław <mirq-linux@...e.qmqm.pl>,
	David Miller <davem@...emloft.net>, <netdev@...r.kernel.org>
Subject: Re: Please revert removal of /sys/class/net/*/features

On Tue, 10 Apr 2012 21:25:47 +0100
Ben Hutchings <bhutchings@...arflare.com> wrote:

> On Tue, 2012-04-10 at 21:59 +0200, Michał Mirosław wrote:
> > On Tue, Apr 10, 2012 at 11:08:51AM -0700, Stephen Hemminger wrote:
> > > This commit needs to be reverted. It removed an available sysfs file, and sysfs
> > > files are part of the ABI. It caused a bug in our current release of Vyatta because
> > > the shell script was using the sysfs file to see if VLAN was supported on a device.
> > > 
> > > An API maybe redundant, but you can't just remove it.
> > 
> > There was a discussion about it before the patch was accepted. It was
> > concluded that due to features flags changes, the /sys/class/net/*/features
> > was never a stable API.
> 
> For reference:
> http://thread.gmane.org/gmane.linux.network/199115/focus=199136
> 
> But note that the VLAN-related feature flags never actually changed
> value.
> 
> Ben.
> 

It is actually not a big deal for our product since 3.3 is
really early in development cycle and the script can be fixed in
that release. Also, we release an image with utilities, scripts and
kernel in one release, and don't have to support running new kernels with
older userspace.

I am more worried that other people might be hit 
the same thing. What about a patch to keep the old flags values
available for backwards compatiablity. Also, when removing a sysfs
file, shouldn't the documented feature removal process be used?
--
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