[pptp-server] Maximum number of users

[David Luyer]

> True, but you forgot the "please also fix the algorithm" bit.

I did a smarter algorithnm for it, and subitted it to the kernel guys
years ago[1] (search linux-kernel for dev_alloc_hint and you might find
it - it was only called by devices which expected to do numerous
allocations and kept a count)... who informed me that they saw no need
for >100 devices.  As with many IMHO useful patches I've submitted, I
gave up, used it myself, and was happy.

David.

[1] actually I did two solutions to it, over time... neither was
accepted

       1.  dev_alloc_hint, where the device provides a hint to the
kernel
           as to starting number

       2.  since that involved an extra exported symbol, an approach
           which didn't and was more automatic:

           if (no free device name in first 10 guesses)
               alloc_based_on_global_hint

           if two devices went beyond 10 instances, this would cause
gaps
           to occur in device names, with no important implication most
           likely...