I've got a laptop for which $NET_RESTART is not good enough. I have
to unload the driver also:
[ "$NET_RESTART" = "yes" ] && /etc/rc.d/init.d/network stop &&
/sbin/rmmod 3c59x && logger "apmscript unloaded 3c59x network
[ "$NET_RESTART" = "yes" ] && modprobe 3c59x &&
logger "apmscript started 3c59x" && /etc/rc.d/init.d/network start
But frankly, this seems like a good idea anyway. Why not do a /sbin/rmmod
at the end of the process, and force ALL the unused drivers out of memory?
I fee that the "apmcontinue" script should be capped at the END of shutdown
at the START of restore. Right now it's always started at the end, which
makes it hard to get the sequencing of things correct.
That's "called" not "capped"
Possibility to reload network modules added in 3.0final-32.
I disagree about starting apmcontinue first after a resume because it's too
easy to mess up (remember that almost all hardware is still unusable at the
beginning of resume).
Then start apmcontine FIRST on shutdown, and LAST on restore.
Look at your own script -- shutdown and restore are mirrors of each other.
The same need is present in apmcontinue.
Same problem. If the user already shuts down a piece of hardware and the apmd
script tries to do it again, the system is messed up.
What do you think about creating an additional "apmstart" script that is run
at the beginning of suspends and resumes?
(This part has already been suggested and done in 3.0final-33)
*** This bug has been marked as a duplicate of 44603 ***