I have a Fujitsu Siemens Amilo 1650 with 64 bit Hardy and Intrepid installed on it.
Under Gutsy, standby worked flawlessy (when using free graphic ati drivers instead of fglrx).
Under Hardy (I still use free graphic ati drivers), standby usually works; sometimes, instead, computer correctly goes in suspend mode but then doesn't wake up (power LED is on, fan is on but the screen doesn't come up and the system doesn't react to "Ctrl+Alt+Del" or "Ctrl+Alt+Sys+B").
Under Intrepid (still free graphic drivers), after suspend computer NEVER wakes up (same symptoms as when it doesn't under Hardy).
Though I cant' tell if the problem in Hardy and Intrepid is the same, it may be an interesting information to know that in Hardy, when the computer doesn't come up from suspend (I think it happens around one time out of 5 or 10), I can still wake it if I press many times (let's say something between 4 and 8) the power button. The system seems to load normally, and I can login, but it seems like the CPU gets stuck very easily (blocking everything, from the GUI to the clock, that becomes late) until I give some interrupt (i.e: moving the mouse); when I do, I can see with dmesg a message such
soft lockup - CPU #1 stuck for 11s
, so system is practically unusable, and I have to shutdown it (still moving touchpad, or it will pause!).
That said, under Intrepid computer just doesn't get usable, no matter how many times I press the power button.
(Still under Intrepid) I did some testing: following the procedure given in https://wiki.ubuntu.com/DebuggingKernelSuspend, I gave the command:
sync; echo 1 > /sys/power/pm_trace; /etc/acpi/sleep.sh force
after reboot, I looked at the dmesg, but no hash was found (I'm attaching it anyway as useless_dmesg.txt, but I think it's useless).
So I tried to just manually remove any module I could remove: i gave the command:
for i in `lsmod | sed 's# .*##'`; do rmmod $i; done
(two times, to also remove "dependencies blocking" modules). The modules left on are reported in the "modules_left_on.txt" file attached.
Finally, I tried to manually remove all removable modules, as above, and THEN give the
sync; echo 1 > /sys/power/pm_trace; /etc/acpi/sleep.sh force
command; this time, on reboot, I COULD find a pair of hash matches in dmesg:
[ 4.188003] Magic number: 4:281:579
[ 4.188003] mem kmem: hash matches
[ 4.188003] pci_bus 0000:00: hash matches
so I'm reporting it as dmesg_lucky.txt.
That said, I have no idea what to do, but I can do any testing requested.
Notice that under Intrepid I'm using 2.6.27-1-generic kernel, but resume from suspend also used to fail with 2.6.26-5 (every time).
Notice that hibernation works perfectly