Bug#450601: pm-utils suck hard on non-ACPI systems
Johannes Berg wrote: Are you saying that for S4 we need something similar (eg. some binary that pokes some ioclts?) No. For suspend (to RAM) all it needs is s2ram installed. For suspend to disk you need to fix the assumption that /sys/power/disk has 'platform' as a valid option. What's the correct assumption for hibernate? Just accept what is set as default in /sys/power/disk by the kernel and do a simple echo disk /sys/power/state? Michael -- Why is it that all of the instruments seeking intelligent life in the universe are pointed away from Earth? signature.asc Description: OpenPGP digital signature
Bug#450601: pm-utils suck hard on non-ACPI systems
No. For suspend (to RAM) all it needs is s2ram installed. For suspend to disk you need to fix the assumption that /sys/power/disk has 'platform' as a valid option. What's the correct assumption for hibernate? Just accept what is set as default in /sys/power/disk by the kernel and do a simple echo disk /sys/power/state? You can take whichever one you like, but you should first *check* that it *exists* by reading /sys/power/disk and seeing if it's listed in there, and only then try to use it. 'platform' may be necessary on some brain-dead ACPI machines, but it doesn't exist on powerpc. johannes signature.asc Description: This is a digitally signed message part
Bug#450601: pm-utils suck hard on non-ACPI systems
Just wondering.. Why isn't your patch so suspend via /sys/power/state instead of the PMU ioctl in the upstream kernel yet :) IMHO it would be good to have that interface work on ppc too (It would save us a lot of special casing) Because Paul Mackerras dislikes the freezer which that currently implies. It may go into 2.6.25. Yesterday i did a rebuild/upload of uswsusp on powerpc, so uswsusp is now available and working on ppc (works nicely on my powerbook).. pm-utils still needs a little tweak though, but nothing big. :) johannes signature.asc Description: This is a digitally signed message part
Bug#450601: pm-utils suck hard on non-ACPI systems
On Thu, Nov 15, 2007 at 08:31:05PM +0100, Johannes Berg wrote: I was just reading http://lwn.net/Articles/197603/. :) Could you send us the output of cat /sys/power/disk and cat /sys/power/state Mind you, I'm talking about two systems here. On my powermac, /sys/power/disk contains something like reboot shutdown test testproc or so. Not sure, but it's identical to your system except lacking platform. /sys/power/state has only disk. On a regular powerbook without my patches /sys/power/state is empty afaik. I have mem there and it works, but that's because I patched the kernel. Just wondering.. Why isn't your patch so suspend via /sys/power/state instead of the PMU ioctl in the upstream kernel yet :) IMHO it would be good to have that interface work on ppc too (It would save us a lot of special casing) Yesterday i did a rebuild/upload of uswsusp on powerpc, so uswsusp is now available and working on ppc (works nicely on my powerbook).. pm-utils still needs a little tweak though, but nothing big. Sjoerd -- Everything should be made as simple as possible, but not simpler. -- Albert Einstein -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#450601: pm-utils suck hard on non-ACPI systems
pm-hibernate has no such option and thus breaks. Not sure why s2disk is involved at all? s2disk isn't involved per se, but it can bring your machine in hibernation. It can work as a `back end' in pm-hibernate. Right. That much I understand. I'm not sure we're understanding each other. In my understanding, on a powermac, the interface /sys/power/state was not working for suspend-to-ram. Correct, on a G5 powermac there is no suspend to RAM implemented at all. On a G4 powerbook/powermac suspend to RAM is implemented but not via /sys/power/state. If I read your e-mails to this report it isn't working for hibernation either. Right? /sys/power/state works for hibernation on a G5 powermac, on a G4 powermac/powerbook it also works but is broken when DRI is enabled. Now, for S3 there were some ioctl that had to be used. This is now in the s2ram binary. Yes, for suspend to RAM (there is no S3 on powermacs/powerbooks! stop confusing everybody with ACPI terminology :) ) there are ioctls on /dev/pmu to be used. And now I just reviewed the s2ram sources and indeed, it does implement this. So I guess the whole bug is s2ram isn't installed by default. Are you saying that for S4 we need something similar (eg. some binary that pokes some ioclts?) No. For suspend (to RAM) all it needs is s2ram installed. For suspend to disk you need to fix the assumption that /sys/power/disk has 'platform' as a valid option. johannes signature.asc Description: This is a digitally signed message part
Bug#450601: pm-utils suck hard on non-ACPI systems
On Thu, 15 Nov 2007 17:49:30 +0100 Johannes Berg [EMAIL PROTECTED] wrote: I thought s2disk was also tested and working, but apparently not. If you could point me at some code that does the job or some docs, I could add it to s2disk. s2disk needs to not try to set the platform method if that method is unavailable. I think it does by default. It's also an config option. pm-hibernate has no such option and thus breaks. Not sure why s2disk is involved at all? s2disk isn't involved per se, but it can bring your machine in hibernation. It can work as a `back end' in pm-hibernate. ... I'm not sure we're understanding each other. In my understanding, on a powermac, the interface /sys/power/state was not working for suspend-to-ram. If I read your e-mails to this report it isn't working for hibernation either. Right? Now, for S3 there were some ioctl that had to be used. This is now in the s2ram binary. Are you saying that for S4 we need something similar (eg. some binary that pokes some ioclts?) If yes, I think we'll have to add that to uswsusp. pm-utils is just a bunch of (arch: all) scripts, that uses the /sys/power/state interface. grts Tim signature.asc Description: PGP signature
Bug#450601: pm-utils suck hard on non-ACPI systems
Johannes Berg schrieb: I thought s2disk was also tested and working, but apparently not. If you could point me at some code that does the job or some docs, I could add it to s2disk. s2disk needs to not try to set the platform method if that method is unavailable. I think it does by default. It's also an config option. pm-hibernate has no such option and thus breaks. Not sure why s2disk is involved at all? Hi Johannes, I was just reading http://lwn.net/Articles/197603/. Could you send us the output of cat /sys/power/disk and cat /sys/power/state If your platform doesn't support this interface, what is the correct way to detect a PMU and how is it used correctly. Michael -- Why is it that all of the instruments seeking intelligent life in the universe are pointed away from Earth? signature.asc Description: OpenPGP digital signature
Bug#450601: pm-utils suck hard on non-ACPI systems
I thought s2disk was also tested and working, but apparently not. If you could point me at some code that does the job or some docs, I could add it to s2disk. s2disk needs to not try to set the platform method if that method is unavailable. I think it does by default. It's also an config option. pm-hibernate has no such option and thus breaks. Not sure why s2disk is involved at all? johannes signature.asc Description: This is a digitally signed message part
Bug#450601: pm-utils suck hard on non-ACPI systems
I was just reading http://lwn.net/Articles/197603/. :) Could you send us the output of cat /sys/power/disk and cat /sys/power/state Mind you, I'm talking about two systems here. On my powermac, /sys/power/disk contains something like reboot shutdown test testproc or so. Not sure, but it's identical to your system except lacking platform. /sys/power/state has only disk. On a regular powerbook without my patches /sys/power/state is empty afaik. I have mem there and it works, but that's because I patched the kernel. If your platform doesn't support this interface, what is the correct way to detect a PMU and how is it used correctly. the PMU is powerbook specific. Look at the original hal helpers or pbbuttonsd for how to use it. joahnnes signature.asc Description: This is a digitally signed message part
Bug#450601: pm-utils suck hard on non-ACPI systems
On Tue, 13 Nov 2007 14:09:37 +0100 Johannes Berg [EMAIL PROTECTED] wrote: I consider the non-functioning /sys/power/state interface a kernel bug. That's fine for the part about suspend to RAM. However, with suspend to disk, the interface is functioning *perfectly*, it just doesn't offer the platform method. ah, ok. I'll look into that. I thought s2disk was also tested and working, but apparently not. If you could point me at some code that does the job or some docs, I could add it to s2disk. s2disk needs to not try to set the platform method if that method is unavailable. I think it does by default. It's also an config option. grts Tim signature.asc Description: PGP signature
Bug#450601: pm-utils suck hard on non-ACPI systems
I consider the non-functioning /sys/power/state interface a kernel bug. That's fine for the part about suspend to RAM. However, with suspend to disk, the interface is functioning *perfectly*, it just doesn't offer the platform method. I thought s2disk was also tested and working, but apparently not. If you could point me at some code that does the job or some docs, I could add it to s2disk. s2disk needs to not try to set the platform method if that method is unavailable. johannes signature.asc Description: This is a digitally signed message part
Bug#450601: pm-utils suck hard on non-ACPI systems
On Fri, 09 Nov 2007 13:10:40 +0100 Johannes Berg [EMAIL PROTECTED] wrote: uswsusp should work nowadays. If not, please file a bug there. Actually, it doesn't, but that's unrelated, I don't want to use it. I consider the non-functioning /sys/power/state interface a kernel bug. s2ram and s2disk provide extra features and workarounds for machines where the bare interface doesn't work. It is mostly video-adapter related hacks but in the case of powerpc s2ram it uses some ioctls to get it to suspend to ram. I thought s2disk was also tested and working, but apparently not. If you could point me at some code that does the job or some docs, I could add it to s2disk. grts Tim signature.asc Description: PGP signature
Bug#450601: pm-utils suck hard on non-ACPI systems
uswsusp should work nowadays. If not, please file a bug there. Actually, it doesn't, but that's unrelated, I don't want to use it. johannes signature.asc Description: This is a digitally signed message part
Bug#450601: pm-utils suck hard on non-ACPI systems
Package: pm-utils Version: 0.99.2-3 Severity: normal pm-utils are unable to hibernate my powermac. I do not have a 'platform' /sys/power/disk mode and this makes pm-utils give up on hibernating. pm-utils are unable to suspend my powerbook. powerbooks don't support suspending via /sys/power/state but need to use the PMU ioctls instead. -- System Information: Debian Release: lenny/sid APT prefers unstable APT policy: (500, 'unstable'), (1, 'experimental') Architecture: powerpc (ppc) Kernel: Linux 2.6.24-rc1-g7f3640ca-dirty (PREEMPT) Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/bash Versions of packages pm-utils depends on: ii powermgmt-base1.29 Common utils and configs for power Versions of packages pm-utils recommends: ii hal0.5.10-2 Hardware Abstraction Layer pn radeontool none(no description available) ii uswsusp0.6~cvs20070202-2 tools to use userspace software su pn vbetoolnone(no description available) -- no debconf information -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#450601: pm-utils suck hard on non-ACPI systems
On Thu, 08 Nov 2007 14:55:21 +0100 Johannes Berg [EMAIL PROTECTED] wrote: Package: pm-utils Version: 0.99.2-3 Severity: normal pm-utils are unable to hibernate my powermac. I do not have a 'platform' /sys/power/disk mode and this makes pm-utils give up on hibernating. pm-utils are unable to suspend my powerbook. powerbooks don't support suspending via /sys/power/state but need to use the PMU ioctls instead. uswsusp should work nowadays. If not, please file a bug there. grts Tim signature.asc Description: PGP signature