Summary: | sys-power/pm-utils-1.4.1-r6 don't suspend and hibernate ThinkPad Z61m with 1.4.1-inhibit-on-right-status.patch | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Ognian Tenchev <drJeckyll> |
Component: | [OLD] Core system | Assignee: | Freedesktop bugs <freedesktop-bugs> |
Status: | RESOLVED INVALID | ||
Severity: | normal | CC: | nikoli |
Priority: | Normal | ||
Version: | unspecified | ||
Hardware: | x86 | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: |
pm-suspend not working
pm-suspend working pm-suspend.log not working from /var/log/pm-suspend.log pm-suspend.log working from /var/log/pm-suspend.log |
Description
Ognian Tenchev
2014-06-22 19:19:17 UTC
Reduce the patches one by one in the ebuild to pinpoint the exact problematic patch. It might be correctly refusing to suspend/hibernate, some of the patches applied, would do that. Added patches between -r2 and -r5: + "${FILESDIR}"/${PV}-fix-alpm-typo.patch \ + "${FILESDIR}"/${PV}-inhibit-on-right-status.patch \ + "${FILESDIR}"/${PV}-ignore-led-failure.patch \ + "${FILESDIR}"/${PV}-run-hook-logging.patch \ + "${FILESDIR}"/${PV}-suspend-hybrid.patch \ + "${FILESDIR}"/${PV}-uswsusp-hibernate-mode.patch \ + "${FILESDIR}"/${PV}-xfs_buffer_arguments.patch Also new files are installed you can try to drop one by one /usr/lib/pm-utils/power.d/pci_devices /usr/lib/pm-utils/power.d/usb_bluetooth Newly removed (outdated) file between -r2 and -r5 you can try to restore by removing the removal line from the ebuild: /usr/lib/pm-utils/sleep.d/49bluetooth Because saying "It doesn't work." really doesn't help anything removing /usr/lib/pm-utils/power.d/pci_devices and/or /usr/lib/pm-utils/power.d/usb_bluetooth don't make suspend to work with -r5 adding back /usr/lib/pm-utils/sleep.d/49bluetooth also don't make suspend to work with -r5 removing "${FILESDIR}"/${PV}-fix-alpm-typo.patch don't fix issue if I remove "${FILESDIR}"/${PV}-fix-alpm-typo.patch ebuild become broken with error that run-hook-logging.patch can't be applied so I also remove epatch "${FILESDIR}"/${PV}-run-hook-logging.patch this make suspend to work ... I don't make tests with other patches. (In reply to Ognian Tenchev from comment #2) > if I remove > "${FILESDIR}"/${PV}-fix-alpm-typo.patch > ebuild become broken with error that run-hook-logging.patch can't be applied > so I also remove > epatch "${FILESDIR}"/${PV}-run-hook-logging.patch > this make suspend to work ... That doesn't make any sense, both of these are only related to printing out messages or logging, they don't affect real functionality at all There is also PM_DEBUG="true" environment variable you can set, same as USE="debug" in pm-utils, that might help you track the real reason down, since I suspect it's something else that just happened to happen at the same time you removed the 2 patches Also, notice that the patches are from Debian, they've been using them for years without issues, so you can understand why I'm suspicious :) It must have been accident that it started working for you at the same time you removed the 2 patches that only effect printed out messages and logs. There is nothing to go on with in the bug. Please continue digging up and reopen the bug (that means setting the bug back to UNCONFIRMED from RESOLVED, NEEDINFO) once you have something tangible to work with. Thanks! I'm sorry. There is mistake by me in comment 2. Actually I remove these two patches: "${FILESDIR}"/${PV}-inhibit-on-right-status.patch (not "${FILESDIR}"/${PV}-fix-alpm-typo.patch) and "${FILESDIR}"/${PV}-run-hook-logging.patch I'm attaching logs with PM_DEBUG="true" (working and not working), but there is no difference in them. Created attachment 379444 [details, diff]
pm-suspend not working
Created attachment 379446 [details, diff]
pm-suspend working
Next are logs from /var/log/pm-suspend.log which are different between working and not working Created attachment 379448 [details, diff]
pm-suspend.log not working from /var/log/pm-suspend.log
Created attachment 379450 [details, diff]
pm-suspend.log working from /var/log/pm-suspend.log
1.4.1-inhibit-on-right-status.patch that was applied as pm-utils-1.4.1-r3 is propably the culprit. https://bugs.gentoo.org/show_bug.cgi?id=360673#c6 Can you verify -r3 fails for you too? -r3 fails too Wow. I found bunch of bugs in -r5, and removed it from tree, please try 1.4.1-r6 too. As in, propably will remove the problematic patch or something if -r6 is still broken for you, and create -r7 with a dropped patch. However I really would like to figure better solution than removing the patch, since like said, Debian has used it succesfully for quite a bit. However there are of course similar bugs in Debian's bugs, like: https://bugs.debian.org/cgi-bin/pkgreport.cgi?dist=unstable;package=pm-utils Does any of those resemble the problem you are having? I'm sorry but -r6 still don't suspend my poor laptop :) I will have to take my night to read all bugs on Debian ... I'm not sure and really don't read logs until now ... may be it is better do dig in logs first and play around before remove this patch ... may be it is just my old hardware which cause problems. Can you just leave this bug as it is for day and give me some time to look around. If it is just me I will fork ebuild in my local overlay and ... let it be :) Sure, I will investigate myself too, lets put the bug in hold for a few days After looking at logs I found that /var/run/pm-utils/pm-suspend/storage/inhibit is always created and which prevent suspend Founding cause was just game of trying but I remove all hooks from /usr/lib/pm-utils/sleep.d which don't resolve issue. Then I just start to look at scripts of which pm-utils is composed and this lead me to one forgotten file from years ago which I crated in /etc/pm/sleep.d called resume.sh which have wrong case close tag ... so ... I feed ashamed and guilty but now -r6 works ... and I learn something about pm-utils :) Sorry for your time and thanks for support. So this bug should be closed as INVALID? yes ah, good. now I can move forward with stabilizing the new version, this bug was the only reason I haven't done it yet. |