Thanks for your reply.
I am using Mandrake 7.0 which is heavily updated from the cooker site.
Currently I am on :
kernel-2.2.15-0.27mdk
XFree86-3.3.6-14mdk, with the Mach64 server
apmd-3.0final-3mdk
kapm-0.3.1-3mdk
kcmlaptop-0.82-8mdk
Since I have all software suspends turned off I think this is as a result of
the BIOS suspend setting. The CMOS setting is to suspend after 30 minutes, or
there abouts. Is it possible that having a suspend come in from the BIOS is
enough to 'phase' Linux / X and cause this crash ?
the /etc/sysconfig/apm-scripts has three files:
apmscript, resume and suspend.
The contents of suspend :
#!/bin/bash
# DO NOT EDIT THIS SCRIPT, CREATE AND EDIT APMCONTINUE IN THIS DIRECTORY; you
# can put your stuff into apmcontinue for every link you create to apmscript;
# for a start and definitely enough for most laptops we have two links and
# according subroutines defined here: suspend and resume; all other links
# will be redirected directly to apmcontinue which you can create; also
# suspend and resume call apmscript, so you can also do other things (like
# reinitialising some PCMCIA-Card) there; apmcontinue will get the name
# as which it was called as $1; for debugging see the logfiles
#
# This script can be controlled by editing /etc/sysconfig/apmd.
PROG=`basename "$0"`
LOCKFILE=/var/lock/subsys/resume
[ -e /etc/sysconfig/keyboard ] && . /etc/sysconfig/keyboard
[ -e /etc/sysconfig/apmd ] && . /etc/sysconfig/apmd
[ -e /etc/sysconfig/clock ] && . /etc/sysconfig/clock
[ "$UTC" = "" ] && UTC="no"
[ "$CHANGEVT" = "" ] && CHANGEVT="0"
[ "$CLOCK_SYNC" = "" ] && CLOCK_SYNC="no"
case "$PROG" in
suspend)
# Some broken disks won't return from a suspend to disk
# unless they're set to 100% failsafe settings with hdparm.
if test "x$HDPARM_AT_SUSPEND" != "x"; then
for i in /proc/ide/hd*; do
DRIVE=/dev/`echo $i |sed -e "s,.*/,,g"`
if test "x`cat $i/media`" = "xdisk"; then
hdparm $HDPARM_AT_SUSPEND $DRIVE
fi
done
fi
sync
[ "$PCMCIARESTART" = "yes" ] && /etc/rc.d/init.d/pcmcia stop
if [ "$PCMCIAWAIT" = "yes" ]; then
until [ `grep "Socket .: empty" /var/run/stab|wc -l` = `grep
"Socket" /var/run/stab|wc -l` ]; do
logger "Waiting for pcmcia-device to be removed..."
sleep 10
done
fi
while [ -f "$LOCKFILE" ]; do
RESUMEPID=`cat "$LOCKFILE"`
logger "Waiting for resume to be finished ($LOCKFILE gives PID
$RESUMEPID)..."
sleep 10
done
[ "$CHANGEVT" != "0" ] && chvt 1
sync
[ -f /etc/sysconfig/apm-scripts/apmcontinue ] && .
/etc/sysconfig/apm-scripts/apmcontinue "$PROG"
sync
;;
resume)
# If HDPARM_AT_RESUME is set, the user has a broken disk.
# We'd better wake it up manually. :/
if test "x$HDPARM_AT_RESUME" != "x"; then
for i in /proc/ide/hd*; do
DRIVE=/dev/`echo $i |sed -e "s,.*/,,g"`
if test "x`cat $i/media`" = "xdisk"; then
hdparm -q -S0 $DRIVE
fi
done
fi
# as some displays don't like being waked up from power-off
# state directly into graphics mode, we can switch to console
# 1 during suspend and back to console n (where we assume X11
# is running) after a resume; also the setting of the console
# beep and the whole sound system can be confused; see
# /etc/sysconfig/apm for a detailed description of the
# variables used here.
echo $$ >> "$LOCKFILE"
if [ "$TERMINALBEEP" != "" ]; then
# first we set the beep-length
for t in 1 2 3 4 5 6 7 8 9; do
setterm -blength $BEEPLENGTH >/dev/tty$t
done
fi
if [ "$RESTORESOUND" = "yes" ]; then
# then we have to find all programs that use sound, because
# we have to stop them in order to restart the
# sound-system-driver
SOUNDPROGS=""
for n in `lsof |grep /dev|grep ' 14,'|sed -e 's/ \+/|/g'|cut '-d|'
-f1-3`; do
SOUNDPROG=`echo $n|cut '-d|' -f1`
SOUNDPID=`echo $n|cut '-d|' -f2`
SOUNDUSER=`echo $n|cut '-d|' -f3`
SOUNDDISPLAY=`perl -e 's/\000/\n/g;' -p
/proc/$SOUNDPID/environ|grep DISPLAY=|cut -d= -f2`
SOUNDPROGS="$n|$SOUNDDISPLAY $SOUNDPROGS"
kill "$SOUNDPID"
ps -ax "$SOUNDPID" &>/dev/null
[ "$?" = "0" ] && kill -9 "$SOUNDPID"
logger "apmscript: Program $SOUNDPROG ($SOUNDPID) of user
$SOUNDUSER on display $SOUNDDISPLAY terminated."
done
# let's restart the sound-system-driver now
for m in $SOUNDMODULES; do
rmmod $m
done
for m in $SOUNDMODULES; do
insmod $m
[ "$?" = "0" ] || modprobe $m
done
fi
if [ "$RESTORESOUNDPROGS" = "yes" ]; then
# at least we can start the soundprograms again using
# the right display
for n in $SOUNDPROGS; do
SOUNDPROG=`echo $n|cut '-d|' -f1`
SOUNDPID=`echo $n|cut '-d|' -f2`
SOUNDUSER=`echo $n|cut '-d|' -f3`
SOUNDDISPLAY=`echo $n|cut '-d|' -f4`
su - "$SOUNDUSER" -c "source /etc/profile;[ -f ~/.bashrc ] &&
source ~/.bashrc;[ -f ~/.bash_profile ] && source ~/.bash_profile;export
DISPLAY="$SOUNDDISPLAY";$SOUNDPROG &"
logger "apmscript: Program $SOUNDPROG started as user $SOUNDUSER
on display $SOUNDDISPLAY."
done
fi
[ "$CHANGEVT" != "0" ] && chvt "$CHANGEVT"
if [ "$PCMCIARESTART" = "yes" ] ; then
/etc/rc.d/init.d/pcmcia start
/etc/rc.d/init.d/network restart
fi
sync
[ -f /etc/sysconfig/apm-scripts/apmcontinue ] && .
/etc/sysconfig/apm-scripts/apmcontinue "$PROG"
# Read the hardware clock
if test $CLOCK_SYNC != "no"; then
if test $UTC = yes -o "$UTC" = true; then
hwclock --utc --hctosys
else
hwclock --hctosys
fi
fi
# Finally, run anacron to catch up cron stuff we missed...
# If anacron is installed.
if test -x /usr/bin/anacron; then
/usr/bin/anacron
fi
rm -f "$LOCKFILE"
;;
*)
if [ -f /etc/sysconfig/apm-scripts/apmcontinue ]; then
. /etc/sysconfig/apm-scripts/apmcontinue "$PROG"
else
logger "FAIL: Wrong parameter \"$PROG\" in apmscript!"
exit 1
fi
;;
esac
exit 0
the contents of apmscript:
#!/bin/bash
# DO NOT EDIT THIS SCRIPT, CREATE AND EDIT APMCONTINUE IN THIS DIRECTORY; you
# can put your stuff into apmcontinue for every link you create to apmscript;
# for a start and definitely enough for most laptops we have two links and
# according subroutines defined here: suspend and resume; all other links
# will be redirected directly to apmcontinue which you can create; also
# suspend and resume call apmscript, so you can also do other things (like
# reinitialising some PCMCIA-Card) there; apmcontinue will get the name
# as which it was called as $1; for debugging see the logfiles
#
# This script can be controlled by editing /etc/sysconfig/apmd.
PROG=`basename "$0"`
LOCKFILE=/var/lock/subsys/resume
[ -e /etc/sysconfig/keyboard ] && . /etc/sysconfig/keyboard
[ -e /etc/sysconfig/apmd ] && . /etc/sysconfig/apmd
[ -e /etc/sysconfig/clock ] && . /etc/sysconfig/clock
[ "$UTC" = "" ] && UTC="no"
[ "$CHANGEVT" = "" ] && CHANGEVT="0"
[ "$CLOCK_SYNC" = "" ] && CLOCK_SYNC="no"
case "$PROG" in
suspend)
# Some broken disks won't return from a suspend to disk
# unless they're set to 100% failsafe settings with hdparm.
if test "x$HDPARM_AT_SUSPEND" != "x"; then
for i in /proc/ide/hd*; do
DRIVE=/dev/`echo $i |sed -e "s,.*/,,g"`
if test "x`cat $i/media`" = "xdisk"; then
hdparm $HDPARM_AT_SUSPEND $DRIVE
fi
done
fi
sync
[ "$PCMCIARESTART" = "yes" ] && /etc/rc.d/init.d/pcmcia stop
if [ "$PCMCIAWAIT" = "yes" ]; then
until [ `grep "Socket .: empty" /var/run/stab|wc -l` = `grep
"Socket" /var/run/stab|wc -l` ]; do
logger "Waiting for pcmcia-device to be removed..."
sleep 10
done
fi
while [ -f "$LOCKFILE" ]; do
RESUMEPID=`cat "$LOCKFILE"`
logger "Waiting for resume to be finished ($LOCKFILE gives PID
$RESUMEPID)..."
sleep 10
done
[ "$CHANGEVT" != "0" ] && chvt 1
sync
[ -f /etc/sysconfig/apm-scripts/apmcontinue ] && .
/etc/sysconfig/apm-scripts/apmcontinue "$PROG"
sync
;;
resume)
# If HDPARM_AT_RESUME is set, the user has a broken disk.
# We'd better wake it up manually. :/
if test "x$HDPARM_AT_RESUME" != "x"; then
for i in /proc/ide/hd*; do
DRIVE=/dev/`echo $i |sed -e "s,.*/,,g"`
if test "x`cat $i/media`" = "xdisk"; then
hdparm -q -S0 $DRIVE
fi
done
fi
# as some displays don't like being waked up from power-off
# state directly into graphics mode, we can switch to console
# 1 during suspend and back to console n (where we assume X11
# is running) after a resume; also the setting of the console
# beep and the whole sound system can be confused; see
# /etc/sysconfig/apm for a detailed description of the
# variables used here.
echo $$ >> "$LOCKFILE"
if [ "$TERMINALBEEP" != "" ]; then
# first we set the beep-length
for t in 1 2 3 4 5 6 7 8 9; do
setterm -blength $BEEPLENGTH >/dev/tty$t
done
fi
if [ "$RESTORESOUND" = "yes" ]; then
# then we have to find all programs that use sound, because
# we have to stop them in order to restart the
# sound-system-driver
SOUNDPROGS=""
for n in `lsof |grep /dev|grep ' 14,'|sed -e 's/ \+/|/g'|cut '-d|'
-f1-3`; do
SOUNDPROG=`echo $n|cut '-d|' -f1`
SOUNDPID=`echo $n|cut '-d|' -f2`
SOUNDUSER=`echo $n|cut '-d|' -f3`
SOUNDDISPLAY=`perl -e 's/\000/\n/g;' -p
/proc/$SOUNDPID/environ|grep DISPLAY=|cut -d= -f2`
SOUNDPROGS="$n|$SOUNDDISPLAY $SOUNDPROGS"
kill "$SOUNDPID"
ps -ax "$SOUNDPID" &>/dev/null
[ "$?" = "0" ] && kill -9 "$SOUNDPID"
logger "apmscript: Program $SOUNDPROG ($SOUNDPID) of user
$SOUNDUSER on display $SOUNDDISPLAY terminated."
done
# let's restart the sound-system-driver now
for m in $SOUNDMODULES; do
rmmod $m
done
for m in $SOUNDMODULES; do
insmod $m
[ "$?" = "0" ] || modprobe $m
done
fi
if [ "$RESTORESOUNDPROGS" = "yes" ]; then
# at least we can start the soundprograms again using
# the right display
for n in $SOUNDPROGS; do
SOUNDPROG=`echo $n|cut '-d|' -f1`
SOUNDPID=`echo $n|cut '-d|' -f2`
SOUNDUSER=`echo $n|cut '-d|' -f3`
SOUNDDISPLAY=`echo $n|cut '-d|' -f4`
su - "$SOUNDUSER" -c "source /etc/profile;[ -f ~/.bashrc ] &&
source ~/.bashrc;[ -f ~/.bash_profile ] && source ~/.bash_profile;export
DISPLAY="$SOUNDDISPLAY";$SOUNDPROG &"
logger "apmscript: Program $SOUNDPROG started as user $SOUNDUSER
on display $SOUNDDISPLAY."
done
fi
[ "$CHANGEVT" != "0" ] && chvt "$CHANGEVT"
if [ "$PCMCIARESTART" = "yes" ] ; then
/etc/rc.d/init.d/pcmcia start
/etc/rc.d/init.d/network restart
fi
sync
[ -f /etc/sysconfig/apm-scripts/apmcontinue ] && .
/etc/sysconfig/apm-scripts/apmcontinue "$PROG"
# Read the hardware clock
if test $CLOCK_SYNC != "no"; then
if test $UTC = yes -o "$UTC" = true; then
hwclock --utc --hctosys
else
hwclock --hctosys
fi
fi
# Finally, run anacron to catch up cron stuff we missed...
# If anacron is installed.
if test -x /usr/bin/anacron; then
/usr/bin/anacron
fi
rm -f "$LOCKFILE"
;;
*)
if [ -f /etc/sysconfig/apm-scripts/apmcontinue ]; then
. /etc/sysconfig/apm-scripts/apmcontinue "$PROG"
else
logger "FAIL: Wrong parameter \"$PROG\" in apmscript!"
exit 1
fi
;;
esac
exit 0
Regards,
Owen
On Tue, 02 May 2000, you wrote:
> I had that behavior with my Dell Inspiron 7500 (ATI Rage Pro Mobility
> chipset) when I ran RedHat 6.1 upgraded to XFree86 3.3.6 with the Mach64
> drivers, but installed Mandrake 7.0 fixed it. I don't actually run
> cooker; one would hope that it didn't retrograde.
>
> Can you fill us in on what *you* are running (O/S, hardward, X version)?
>
> The essential bit here is that /etc/sysconfig/apm-scripts should contain
> scripts that switch out of X before suspending . . .
>
> I still can only suspend *once* between reboots, though; it doesn't lock
> if I try, but it refuses to suspend again. Any ideas about that one
> apprecited.
>
> OS wrote:
> >
> > Hello,
> >
> > There appears to be a problem with suspending to disk, well at least that what
> > I think it is.
> >
> > As far as I knew this was setup using kcmlaptop (0.82-8mdk). My current setup is
> > Not Powered : Suspend after 5 mins, Powered : Off
> >
> > However I think it may also be done from somewhere else as well ?!? because it
> > was powered at the time it went wrong.
> >
> > After leaving my laptop, for supper I believe :), I returned to find the system
> > dead. The majority of the screen was a dark grey, and there was a thick blue
> > stripe running down the left hand edge. The system was completely unresponsive,
> > I had to power off and on again.
> >
> > I have noticed this now on several occasions, but it 'always used to work',
> > unfortunately I cannot say at what versions of whatever package is responsible
> > for suspending laptops it started to go wrong at.
> >
> > Hope someone can shed some light on this fairly fatal problem,
> >
> > Owen
>
> --
> "Brian, the man from babble-on" [EMAIL PROTECTED]
> Brian T. Schellenberger http://www.babbleon.org
> Support http://www.eff.org. Support decss defendents.
> Support http://www.programming-freedom.org. Boycott amazon.com.
