[systemd-devel] sysfs regression: wrong link counts

[Jiri Slaby]

On 01/31/2012 02:27 AM, Eric W. Biederman wrote:
> Greg KH <greg at kroah.com> writes:
> 
>> On Mon, Jan 30, 2012 at 10:10:59PM +0000, Alan Cox wrote:
>>>> Isn't there some other "proper" way of doing this in userspace, or is
>>>> this really the correct way?
>>>
>>> You can look at the S_IFMT bits and stuff however link count indicating
>>> number of subdirectories is a standard Unix thing and used by many quite
>>> mundane tools as an optimisation.
>>
>> Ah, yeah, that is easier.
>>
>> Eric, care to fix this or want me to revert it?
> 
> With respect to sensors the conversation has already been had, and I fix
> is already queued and should come out in the sensors release due out in
> a week or so.  So sensors should be fixed before this code is merged
> into Linus's kernel.

Oh, we are not going to break userspace with 3.3, are we? I understand
that what sensors do is nothing but sh*t. But this change should wait
until everybody has a chance to have fixed sensors package in their
distribution.

To be clear, what I'm suggesting is to postpone the change and schedule
it for something like 3.7.

> Now why something minor like sensors that nothing seems to depend
> strongly on should stop a system from booting is a huge question.

sensors do not prevent boot at all, of course. The other
cannot-start-network bug does.

> So I am going to have the conversation to triage what is up with
> systemd.  That seems totally unexpected.

Looks like netdev renaming problem, see my other mail.

thanks,
-- 
js
suse labs
.