Re: [PATCH] sysfs: Don't allow the creation of symlinks we can't remove

[ebiederm@xmission.com (Eric W. Biederman)]

Johannes Berg <johannes@sipsolutions.net> writes:

> On Thu, 2010-07-08 at 09:31 -0700, Eric W. Biederman wrote:
>> Recently my tagged sysfs support revealed a flaw in the device core
>> that a few rare drivers are running into such that we don't always put
>> network devices in a class subdirectory named net/.
>> 
>> Since we are not creating the class directory the network devices wind
>> up in a non-tagged directory, but the symlinks to the network devices
>> from /sys/class/net are in a tagged directory.  All of which works
>> until we go to remove or rename the symlink.  When we remove or rename
>> a symlink we look in the namespace of the target of the symlink.
>> Since the target of the symlink is in a non-tagged sysfs directory we
>> don't have a namespace to look in, and we fail to remove the symlink.
>> 
>> Detect this problem up front and simply don't create symlinks we won't
>> be able to remove later.  This prevents symlink leakage and fails in
>> a much clearer and more understandable way.
>
> Eric, I was looking into sysfs netns support for wireless, and with this
> patch applied I just get the warning and no network interfaces.

The warning patch just makes things fail faster.  Although I get some of the
wireless interfaces for hwsim when I use this one.

> Was there any patch that was supposed to fix hwsim?

- If you have my patches that fix CONFIG_SYSFS_DEPRECATED,
  you should find everything works there.

As for a proper fix I have just resent my one liner to
drives/base/core.c I can't think of a better option right now.

For hwsim it is arguable, but the behaviour of sysfs for the
bluetooth bnep driver is very clearly a 3 year old regression,
and the cause is exactly the same.

Eric