The check for the number of slots shold be done before allocating

They passed away...

This reverts commit 5febaac4.

it's not the culprit of the allocation bug

The i2c core in the golden SVEC firmware is the OpenCores i2c
master, and differs from the one used in the fine delay and
the SPEC. So we grab this code from Tom to read the EEPROM
(originally at /user/twlostow/svec/software/sveclib/golden_i2c.c,
and probably adapted from the opencores sources).

polling removed, now it sleeps between retries.

Because vmebase1 does not exist anymore...

Rubini's 'check-fmc-bus' script added.
vmebridge local includes removed.