Re: mostly successful swsusp on I4150, Radeon video
- Just a quick followup to myself: using apm instead of acpi
(and still using swsusp) works even better. Then I use
echo "1 0 0" > /proc/sys/kernel/swsusp
to suspend to disk, and my battery monitoring, etc, all work fine.
The suspend process is faster too. The only remaining wish is for me
to be able to detect when the lid closes, and invoke swsusp when this
Dan Christensen <jdc@...> writes:
> This is just to let others know that the software suspend patch works
> on an Inspiron 4150 with the Radeon Mobility 7500 32M video card.
> With it, I can do a fast suspend to disk. [I never tried the BIOS
> level suspend to disk. I did try hard to get the BIOS level suspend
> to RAM working, but had the problem that my screen would be white
> while the machine was suspended.]
> What I did:
> I applied acpi-20021212-2.4.20.diff patch and then the
> patch-acpi-acpi20021205-swsusp16.gz patch to a stock 2.4.20
> kernel tree, turned on all acpi options (except the Toshiba
> and debugging options) and turned on swsusp.
> Initiating suspend with
> echo 4 > /proc/acpi/sleep
> hiberates the machine (quite quickly, too, even with 512M ram)
> and it resumes fine, as long as DRI is turned off in XF86Config.
> I'm using Debian's XFree86 4.2.1-3. Pcmcia doesn't need any
> special treatment.
> The other way to suspend is with
> echo "1 0 0" > /proc/sys/kernel/swsusp
> but this seemed to not work as well.
> One odd thing: the echo 4 command usually never returns; I
> have to ctrl-c it.
> - Without the swsusp patch, /proc/acpi had lots of info about my
> system. But with the swsusp patch, most of the info is missing.
> So no battery info, for example. Also, acpi doesn't register
> any buttons (power, lid, Fn-Suspend, etc) and neither Fn-Setup nor
> Fn-Battery work anymore (but I think they were fine with just the
> acpi patch).
> - The i8k kernel module now thinks the machine is always on battery
> power. This could be related to the above, or maybe i8k needs to
> use apm?
> - suspend to ram and sleep don't work under acpi
> Other facts:
> - need "resume=/dev/hdax" on boot line to suspend or resume,
> where hdax is your swap partition;
> can add this to lilo.conf, for example
> - add "noresume" to boot line to avoid resuming. Then need
> to "mkswap /dev/hdax; swapon -a" before trying to suspend
> again (else locks up, gives errors about being out of memory
> and trying to kill processes, etc)
> Useful URLS:
> /proc/acpi/* docs:
> swsusp patches, scripts, etc:
> Feel free to write with any questions, or to let me know if you
> have suggestions on fixing the problems. Should I try swsusp with