Re: sysfs root link count broken in 2.6.22-git5

[Jean Delvare <khali@linux-fr.org>]

Hi Kay,

On Thu, 19 Jul 2007 02:44:54 +0200, Kay Sievers wrote:
> On 7/18/07, Jean Delvare <khali@linux-fr.org> wrote:
> > On Tue, 17 Jul 2007 20:38:28 -0700, Greg KH wrote:
> > > On Tue, Jul 17, 2007 at 11:05:30PM +0200, Jean Delvare wrote:
> > > > The code looks like:
> > > >
> > > >        if (sysfs_get_mnt_path(sensors_sysfs_mount, NAME_MAX)
> > > >          || stat(sensors_sysfs_mount, &statbuf) < 0
> > > >          || statbuf.st_nlink <= 2)      /* Empty directory */
> > > >                 return 0;           /* Failure */
> > > >
> > > > This works OK with 2.6.22.1, but the last test fails with the current
> > > > git kernel even when sysfs is mounted.
> > >
> > > Yeah, but is checking the number of hard links in the directory a safe
> > > way to always verify that it isn't empty?
> >
> > I think so, yes. To the best of my knowledge, it has worked on all
> > Unix-like systems for decades. There are other ways, but this is by far
> > the less expensive.
> 
> Well, just check if /sys/devices/ exists, that should be cheap enough. :)

Yes, this is a possibility, and one I had considered at first. But I
wasn't sure which subdirectory to check. sysfs isn't well known for its
stability, and I didn't know which directories exist since the
early days of sysfs, and which do not. For example, fs, kernel and
module were not present in 2.6.5. I am also not sure if directories
which exist today are guaranteed to exist forever. This is the reason
why I decided to check the link count instead, basically checking that
at least one subdirectory exists, without having to name it.

-- 
Jean Delvare