Woof-CE version I used had this kind of section in /sbin/pup_event_backend_modprobe:
Code: Select all
#####BAD HACKS SECTION#####
if [ "$MODULE" = "mwave" ];then
#only install firmware tarball, do not load module (firmware script does it).
firmware_tarball_func
exit 1Code: Select all
firmware_tarball_func() {
fPATTERN='[:,]'"${MODULE}"'\.ko|[:,]'"${MODULEx}"'\.ko'
FIRMPKG="`cat /etc/modules/firmware.dep.${KERNVER} | grep -v '^#' | grep ':' | grep -E "$fPATTERN" | cut -f 1 -d ':' | head -n 1`"
if [ "$FIRMPKG" != "" ];then
iPATTERN='^'"${FIRMPKG}"'$'
if [ "`grep "$iPATTERN" /etc/modules/firmware.dep.inst.${KERNVER}`" = "" ];then
FLAGFIRM='no'
if [ -d /lib/modules/all-firmware/${FIRMPKG} ];then #111106 support firmware directories.
cp -a -f --remove-destination /lib/modules/all-firmware/${FIRMPKG}/* /
FLAGFIRM='yes'
else
if [ -f /lib/modules/all-firmware/${FIRMPKG}.tar.gz ];then
tar -z -x --strip=1 --directory=/ -f /lib/modules/all-firmware/${FIRMPKG}.tar.gz > /dev/null 2>&1
FLAGFIRM='yes'
fi
fi
if [ "$FLAGFIRM" = "yes" ];then
#execute any post-install script...
if [ -f /pinstall.${FIRMPKG}.sh ];then
BRKCNT=0
while [ 1 ];do #serialise execution of pinstall scripts...
PINSTALLCNT=`find / -maxdepth 1 -type f -name 'pinstall.*.sh' | wc -l`
[ $PINSTALLCNT -eq 1 ] && break
usleep ${SLEEPU}0 #110509
BRKCNT=$(($BRKCNT + 1))
[ $BRKCNT -gt 5 ] && break
done
cd /
/pinstall.${FIRMPKG}.sh >/dev/null 2>&1
rm -f /pinstall.${FIRMPKG}.sh >/dev/null 2>&1
fi
echo "$FIRMPKG" >> /etc/modules/firmware.dep.inst.${KERNVER} #101202 120823 moved.
fi
fi
fi
}So...I suggest that people who know that they need firmware for their wireless...copy manually the firmwares to the /lib/firmware. I suppose that is not intended mechanism...but also it looks like that is how it works in Dpup Wheezy 3.5.2.10.
It also explains the failure in firmware dependent kernel module drivers for wireless devices.
I think I have to download latest master brach and check how things is in it now....
I used whole my weekend with latest build and fixing pinstall.sh scripts with coders. I take a rest now. Firmware problem seems solved.
