[ltp] swsusp / t23

Michael Perry linux-thinkpad@linux-thinkpad.org
Sun, 17 Oct 2004 08:16:05 -0700


On Sun, 17 Oct 2004 07:57:52 -0400, Christopher Hubbell
<cgh@rochester.rr.com> wrote:
> James,
> 
> Thanks for your response, but unfortunately I'm not using any PC cards, nor am I running a PPP connection.  I'm just using the internal network port.  I've tried stopping PCMCIA services, I always unload UHCI, I've compiled kernels both with and without a preemptible kernel, I've tried 2.6.7, 2.6.8, and all the rcX versions of 2.6.9.  I have tried the kernel's default swsusp, and I've tried using swsup2 via patches.  Both swsusp & swsusp2 go through the motions, writes to disk, and then returns without stopping.  Grrr....  I'm hoping that when the next major kernel comes out I can try again, but I'm beginning to think that it's not possible to suspend to disk on a T23.
> 
> If anyone has a T20 or T23 on a 2.6 kernel that can suspend to both RAM and disk, I'd love to see your .config and your acpi action for suspend to disk.
> 
>         /cgh
> 
> > Christopher:
> >
> > Are you using a PC-Card?  If so, you must either eject it or suspend
> > it,  Also, if you are using a network connection, even a dial-up ppp
> > connection, you must do the same to that connection or the first data
> > packet will stop the suspend action.  Just something to look at.
> >
> > James McKenzie
> >
> > Christopher Hubbell wrote:
> 
> --
> The linux-thinkpad mailing list home page is at:
> http://mailman.linux-thinkpad.org/mailman/listinfo/linux-thinkpad
> 

I have a T23 which now does regular old acpi sleeps and it can do
suspend to ram on lid events and other stuff.  My steps:

1. downloaded and installed the 2.6.8.1 kernel
2. found out that the laptop would fail and kernel panic upon resume
from suspend.
3. downloaded and installed latest ibm bios and power management
upgrade from ibm's website
4. tried again... .still kernel panic
5. applied the latest acpi patches from acpi.sf.net
6. wrote small script to make the laptop suspend with a lod close event

Now where things are is that the suspend to ram2 thing will not patch
cleanly because I think I applied the kernel acpi patch first.  I had
gone through this before with a fresh kernel source and did it the
opposite way before and I got it all patched.  I can successfully do
an echo 3 to suspend the system, the little moon thing on the laptop
shows its suspending.  I can do a suspend with a lid close and the
system comes back every time.
1
Some things which I tried and were problems for me.  The wireless card
driver I use is the linuxant driver which allows me to use a windows
only card under linux.  For some reason, the laptop will not suspend
if this driver is in the kernel.  It starts but I get a few different
warnings and then it simply fails to suspend.  To get around this, I
wrote a smallish script that includes the suspend event but stops the
driverloader using a small command line tool that linuxant wrote and
then starts the driverloader when it comes back up.

I have not found any problems with the wired ethernet card when it
suspends and usb or pcmcia also do not stop the laptop from suspending
whatsoever.

I think the major thing for my success was the patching of the 2.6.8.1
kernel with the latest acpi.sf.net patch and then recompiling.  I also
found this small acpi tool called acpitool :) which manages things a
bit for me.

BTW, running Debian unstable here with manically applied dist-upgrades
to keep things current.


-- 
Michael Perry
meperry@gmail.com