Re: [PATCH 3/3] misc: use a proper range for minor number dynamic allocation

[cascardo@holoscopio.com]

On Tue, Dec 15, 2009 at 02:42:30PM -0800, H. Peter Anvin wrote:
> On 12/15/2009 02:34 PM, Andrew Morton wrote:
> >>>>
> >>>> The proposed solution uses the not yet reserved range from 64 to 127. If
> >>>> more devices are needed, we may push 64 to 16.
> >>>>
> >>>
> >>> Again, why not push these up above 256?
> >>>
> >>
> >> I merged this patch, but made a note-to-self that there are remaining
> >> open issues..
> > 
> > And nothing else happened.  Can we revisit this please?
> > 
> 
> There seem to be people still worried about breaking userspace with
> majors/minors >= 256.  I'm starting to think it is time to actually
> break userspace, and dynamic majors/minors seem as good as any place to
> start, especially since they by definition has to be managed by
> something like udev.  We have had large dev_t for something like six
> years now, and most pieces of software isn't affected at all -- only the
> stuff that manages /dev.
> 
> 	-hpa

I see no problem in this. Can we make this configurable like the random
minor for block devices was done? Or, perhaps, before making a decision,
we can check what kind of devices are using misc dynamic minor, which
was what I did before writing this patch to count the number of current
in-tree users.

I am currently downloading source code from git since I've lost my disk
recently. That's why I'm late replying. I may send a list of current
users if that's interesting for a better analysis.

Regards,
Cascardo.