This was reported once before. The installation of Beta 4 undid my kludgy
patch for the problem, so I did more research.
My laptop has the hardware clock set to Localtime, which is GMT -5. After a
suspend/resume cycle, the system clock "loses" exactly 5 hours. The clock is
being reset as if UTC=true. I have found that:
/etc/rc.d/rc.sysinit handles the clock correctly.
/etc/sysconfig/apm-scripts/resume.d/7clock Only tests for UTC=true. If
true, then force -u, else default (which for hwclock is 'last used')
It appears the apm module in the kernel is compiled with GMT=yes.
The net result is on a resume, apm sets the clock to UTC mode, then 7clock
resets it to the same thing. So I'm back to my kludgy patch to 7clock:
...
CLOCK=""
[ "$UTC" = "yes" -o "$UTC" = "true" -o "$UTC" = 1 ] && CLOCK="-u"
Becomes
CLOCK="--localtime"
[ "$UTC" = "yes" -o "$UTC" = "true" -o "$UTC" = 1 ] && CLOCK="--utc"
This patch still allows a small window as the machine is restarting where the
clock is set incorrectly. This is visible in the various log timestamps. Does
anyone have a better generic fix?
On a related subject... while diddling around, I noticed that it appears
10usb is the first script run in /etc/sysconfig/apm-scripts/resume.d. Is this
intentional?
--
Randy K. Wilson
