[ltp] vlock + suspend to RAM fails
Felix E. Klee
linux-thinkpad@linux-thinkpad.org
Sat, 21 Jan 2006 20:28:40 +0100
On Sat, 21 Jan 2006 15:47:36 +0100, "Daniel Maier" <nusse@teamidiot.de>
said:
> > Can anyone reproduce the problem? What may be the cause?
> According to some smart guy, it is known and he also made a fix.
> http://www.formorer.de/~formorer/lock
Great. Thanks for the link!
And, by looking at the skript, I also learned something new concerning
shell programming: "set -e" - very valuable!
> 15:43 <@formorer> und 63 ist das zauberwort "and 63 is the magic key"
Yeah, this seems to be the trick. May I ask what IRC channel on which
server you were chatting on? I'm interested in channels where one can
discuss Thinkpad / laptop related issues.
Anyway, my current suspend to RAM script:
#!/bin/sh
# Puts the system into ACPI mode S3 (suspend to RAM), unless the first
# parameter is "on_closed_lid" and the lid is open.
# Sample ACPI deamon event file for executing this script (the "&" is
# necessary so that the deamon will call the script if the event is
# triggered several times in a row, e.g. when the lid has been opened,
# and suddenly needs to be closed again, before the password can be
# entered - happens sometimes at the railway station):
# event=button/lid.*
# action=/usr/local/lib/t41-tools/sleep on_closed_lid &
# Note that sometimes after waking up, the password prompt is not
# properly displayed. In this case: Just type a couple of characters
# and hit return.
# There are no checks preventing vlock from getting called twice: I'm
# worried about race conditions weakening security.
# Felix E. Klee <felix.klee@inka.de>
[ $1 = "on_closed_lid" ] && \
grep -q open /proc/acpi/button/lid/LID/state && exit 0
sync # For the case that the system does not wake up again
echo mem > /sys/power/state
openvt -f -c 63 -w -s -- su - felix -c "/usr/bin/vlock -a 2>&1 |
logger"
Now, I only wonder what will happen if the system is under high load
when
waking up: In this case the consoles may stay unlocked for a
considerable
amount of time, giving an intruder the possibility to read what's on
screen, or perhaps even to get control of the system.
Any idea how to avoid that?
--
Felix E. Klee