Bugzilla – Bug 131713
Suspend-to-ram stopped working
Last modified: 2006-01-22 20:31:29 UTC
As of SuSE 10 Final, suspend to ram has stopped working on my laptop. This feature worked fine up to beta 4. This means that I cannot use this on my laptop anymore. It appears to suspend without a problem, but freezes when attempting to resume. This appears to happen when returning to X, but I am unable to change VTs or get any response from the system, prompting a forced shutdown. HP Pavilion ze2000z Mainboard: ATI RADEON XPRESS 200M Processor: AMD Turion 64 ML-30 RAM: 1.5 gb
Some questions to narrow down the problem :-) 1) did you have to configure any special hacks to enable suspend to RAM to work at all? 2) Did you update from beta4 to final or did you reinstall? 3) are the hacks from 1) (if necessary) still configured? An update might remove some kernel bootparameters, it shouldn't but better check it :-) 4) is this with the open source radeon driver or some binary only ATI stuff? 5) does it work if you booted to runlevel 3 instead of 5? (means: without X, you can suspend from the command line with "powersave -u") 6) does suspend to disk still work / work at all?
suspend-to-ram is unsupported/experimental in suse 10.0.
This is a fresh install with no suspend-to-ram hacks/workarounds. The only extra kernel parameters I'm using are ec_burst=0 and reboot=force (workarounds for other bugs). I tested with the october 31 kotd and it appears to be working again. Unfortunately, the x86_64 source package was damaged on the ftp server (I downloaded it more than once) so I could not compile ndiswrapper module (needed for my wifi adapter). November 2 kotd is appearing in the ftp, but the binary package for x86_64 has not been uploaded, it appears. I may just download the source and compile it with the configuration provided with it or make oldconfig. I'm trying to get things working that I need before I start tweaking it so this is with the provided radeon drivers. Suspend-to-disk never worked with the kernel provided, I'm already subscribed to a bug report on that. That isn't a major problem to me as I don't even use it in windows (it's too buggy there too). In my testing, it appears that it worked in older kernels, however, so it has worked previously, at least for the kernel side of things, not for suse 10. I'll have to try powersave -u from runelevel 3 later tonight.
Just downloaded kotd for nov 4th (testing various patches). Suspend to ram locks up before returning the display regardless if I am using runlevel 3 or 5. This appears to have been broken again from kotd for oct 31.
Runlevel 3 is not enough, system with minimal modules loaded (init=/bin/bash) would be more useful. Or better try mainline kernel, suspend-to-ram is still very experimental.
I'll test with init=/bin/bash when I get home tonight. I should point out that I have removed noapictimer and reboot=force from my kernel parameters due to other bugs being fixed in the kotd I am running, so I'm no longer using those when testing suspend-to-ram. I would also like to emphasize that suspend-to-ram functionality seemed to return in the oct 31 kotd, then started showing the current symptoms with nov 4 kotd. Suspend-to-ram seems to work fine in the mainline kernel as this has been tested previously. If I do test again, would you prefer that I stick to the 2.6.13 version since that is what SuSE 10 is based on?
Testing with init=/bin/bash: Error(9): dbus error And it doesn't suspend.
powersaved is not running with init=/bin/bash. I have written a guide on how to debug acpi suspend "lowlevel", find it at http://www.susewiki.org/index.php?title=ACPI_suspend
It suspended, but when I attempted resume, it appeared to have gone into a hardware initialization loop (hardware kept re-initilizing according to keyboard lights, usb key light, etc).
There's could be small enough diff between oct31 kotd and nov4 kotd to perhaps locate the offending patch?
Suspend to ram is unsupported. If you want it to get fixed, 3days differences should not be that much to try...
Suspend to ram is unsupported and it seems original reporter is not interested any more.