Bug#481766: pm-utils: Please fix this before lenny becomes stable!
reassign 481766 laptop-mode-tools retitle 481776 please provide integration into pm-utils thanks Bart Samwel wrote: BTW, in an earlier mail you suggested that this should actually be fixed in laptop-mode-tools. While I understand the reasoning behind this policy, and while I support this design concept in general, there are simply too many suspend methods in Linux to support, each of which have their own set of scripts which have to be tweaked. The ones I know of are hibernate, pm-utils, pmud, pbbuttonsd, and acpi-support. I'm afraid this situation needs to be resolved before such scripts can be put in the resume-action-requiring packages. :-/ Really imho, the only important suspend framework is pm-utils (and I personally don't care about the others). I'd argue that the vast majority of laptop users is using pm-utils via HAL. I'm personally against trying to support a dozen different power management script suites and I don't think it makes sense if we attempt to do so. We had that before in hal, and it was a bloddy mess. (FWIW, acpi-support afaik is dead) Now we only support pm-utils within hal and we finally begin to have a sane (and manageable) power management stack with HAL/pm-utils and frontends like kpowersave/gnome-power-manager talking to HAL via D-Bus. FWIW, the Ubuntu laptop-mode-tools package has just added the pm-utils hooks. Maybe you can consider to just copy their work. I decided to reassign this bug to laptop-mode-tools. Cheers, 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#481766: pm-utils: Please fix this before lenny becomes stable!
Hi Michael, Michael Biebl wrote: Michael Biebl wrote: The problem is, that the bug is about adding support for laptop-tools, but I am missing proper justification, why this is necessary, what the underlying problem is and why laptop-mode-tools is the correct solution. Bart, can you elaborate a bit on this. The linked launchpad bug is a bit hard to digest. Sorry about the delay, I've just returned from some scheduled downtime (aka a holiday). I understand that you can't read the launchpad thing, but it's not all that important really. Let me try to explain the situation. The basic problem is the following combination of things: 1. Laptop mode tools tweaks hardware settings. Most of these are lost over suspend. So, on resume, it must re-tweak them. The settings that are lost over suspend include the ones that are required for the basic functionality of laptop mode, i.e., spinning down the hard drive. So it is *very* important that laptop mode tools is restarted on resume, or else it is entirely broken on resume. This is especially important since suspending is used most frequently on laptops which are running on battery power, exactly the target audience of laptop mode tools. 2. Until recently, the package acpi-support handled suspend itself. It has a directory of on-resume scripts in /etc/acpi/resume.d. This directory contains a script that restarts laptop mode tools. 3. The package acpi-support has since been changed to delegate suspend to pm-utils instead (by default -- the old functionality is still present). 4. The pm-utils suspend functionality does not contain a script that restarts laptop mode tools on resume. So technically: a) This is not a regression in pm-utils when viewed in isolation. The pm-utils package has never supported this. b) This is not a regression in laptop-mode-tools. Laptop mode tools has never cared about suspend. c) This IS a regression in acpi-support, because it used to do this in sarge, and now it doesn't in lenny. So it's all a bit strange -- an important regression in the functionality of acpi-support, which can only be fixed by changing pm-utils. BTW, in an earlier mail you suggested that this should actually be fixed in laptop-mode-tools. While I understand the reasoning behind this policy, and while I support this design concept in general, there are simply too many suspend methods in Linux to support, each of which have their own set of scripts which have to be tweaked. The ones I know of are hibernate, pm-utils, pmud, pbbuttonsd, and acpi-support. I'm afraid this situation needs to be resolved before such scripts can be put in the resume-action-requiring packages. :-/ On a side note, I've just received another bug report (#501804) for laptop-mode-tools mentioning exactly this -- indicated priority by the submitter as important. I've CC'ed the submitter, as he may want do some advocacy as well. :-) Cheers, Bart -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#481766: pm-utils: Please fix this before lenny becomes stable!
Package: pm-utils Version: 1.1.2.3-1 Followup-For: Bug #481766 I would _really_ like to see this bug fixed before Lenny becomes stable, because of the following reasons: 1) This bug is really enjoying. 2) The suspend/hibernate capability of Linux has a bad reputation because of bugs like that. 3) It is very hard for a novice to find out what is going wrong. 4) It could be fixed in just a second, as the only thing that is to be done is to include a simple text file in the package. 5) There is a patch available since 18 May 2008 (!). Regards, Christian Gogolin -- System Information: Debian Release: lenny/sid APT prefers testing APT policy: (800, 'testing'), (70, 'unstable') Architecture: i386 (i686) Kernel: Linux 2.6.26-1-686 (SMP w/1 CPU core) 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 kbd 1.14.1-4 Linux console font and keytable ut ii powermgmt-base1.30 Common utils and configs for power Versions of packages pm-utils recommends: ii hal 0.5.11-3 Hardware Abstraction Layer ii radeontool1.5-5 utility to control ATI Radeon back ii uswsusp 0.7-1.2tools to use userspace software su ii vbetool 1.0-3 run real-mode video BIOS code to a Versions of packages pm-utils suggests: ii cpufrequtils 004-2 utilities to deal with the cpufreq -- no debconf information -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#481766: pm-utils: Please fix this before lenny becomes stable!
Christian Gogolin wrote: Package: pm-utils Version: 1.1.2.3-1 Followup-For: Bug #481766 I would _really_ like to see this bug fixed before Lenny becomes stable, because of the following reasons: Lenny is frozen. It is really unlikely that our RMs will grant a freeze exception. I will take the blame for not responding earlier, but see below. 1) This bug is really enjoying. Do you mean annoying? Anyways, statements like this do generally not help to get an issue addressed, as they add no valuable information. The problem is, that the bug is about adding support for laptop-tools, but I am missing proper justification, why this is necessary, what the underlying problem is and why laptop-mode-tools is the correct solution. Bart was referencing a launchpad bug number, but reading through it, it seems laptop-mode-tools can even worsen the situation regarding the harddisk load/unload cycles. Would simply installing laptop-mode-tools do the right thing out-of-the-box or does it require configuration? If the latter, it would be basically useless, imho. 2) The suspend/hibernate capability of Linux has a bad reputation because of bugs like that. 3) It is very hard for a novice to find out what is going wrong. 4) It could be fixed in just a second, as the only thing that is to be done is to include a simple text file in the package. 5) There is a patch available since 18 May 2008 (!). It's obviously important to know if the patch is the right fix. Would it actually help, if we added this hook to pm-utils, given that laptop-mode-tools is not installed by default. I guess we won't add a depedency on pm-utils, given how big this package is and looking through the file list it seems to duplicate functionality from pm-utils (wireless control, lcd brightness, ...) would those two packages actually stomp on each others feet? As a side-note, it is no problem, if this hook is added to the laptop-mode-tools package itself (actually, upstream prefers if such hooks are shipped by the corresponding packages) Cheers, 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#481766: pm-utils: Please fix this before lenny becomes stable!
Michael Biebl wrote: The problem is, that the bug is about adding support for laptop-tools, but I am missing proper justification, why this is necessary, what the underlying problem is and why laptop-mode-tools is the correct solution. Bart, can you elaborate a bit on this. The linked launchpad bug is a bit hard to digest. Thanks, 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#481766: pm-utils: Please fix this before lenny becomes stable!
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Hi, Do you mean annoying? Anyways, statements like this do generally not help to get an issue addressed, as they add no valuable information. Yes, I did mean annoying (auto correction...). I must apologize for this statement. My intention was to provoke a reaction (and this seems to have worked just fine) but perhaps I was a bit too unkind... Bart was referencing a launchpad bug number, but reading through it, it seems laptop-mode-tools can even worsen the situation regarding the harddisk load/unload cycles. Would simply installing laptop-mode-tools do the right thing out-of-the-box or does it require configuration? If the latter, it would be basically useless, imho. I don't know which bug you mean , but I think the harddisk load/unload cycle thing is a totally different issue. The package laptop-mode-tools offers a lot more functionality than just setting the (potentially dangerous) hdparm -S option. The problem is, that the bug is about adding support for laptop-tools, but I am missing proper justification, why this is necessary, what the underlying problem is and why laptop-mode-tools is the correct solution. I think that all packages (and especially those that are likely to be used together) should support each other as well as possible. The present setup just does not work properly without edition conf files, which i regard as acceptable for the testing branch but unaccepatble for a stable release. The problem is, that laptop-mode-tools sets some options (especially hdparm options for acoustic management, spin down, ...) that are not persistent. Therefor it needs to be noticed when the computer resumes from suspend/hibernate. As mentioned by Bart Samwel acpi-support does just that and it works fine. To exactly mimic the behavior of acpi-support one would have to add something like #!/bin/sh # # 99laptop-mode: Reload laptop mode case $1 in hibernate|suspend) if [ -e /etc/init.d/laptop-mode ] ; then invoke-rc.d laptop-mode stop fi ;; thaw|resume) if [ -e /etc/init.d/laptop-mode ] ; then invoke-rc.d laptop-mode restart fi ;; *) exit $NA ;; esac to /usr/lib/pm-utils/sleep.d . I personally have my machine set up like this and everything works flawless. Would it actually help, if we added this hook to pm-utils, given that laptop-mode-tools is not installed by default. If laptop-mode-tools is not installed this hook would do nothing at all. I guess we won't add a depedency on pm-utils, given how big this package is. I dont think this is necessary. ...looking through the file list it seems to duplicate functionality from pm-utils (wireless control, lcd brightness, ...) would those two packages actually stomp on each others feet? I don't think that they will come in each others way (at least they do not on my computer), but perhaps Bart can comment on this. As a side-note, it is no problem, if this hook is added to the laptop-mode-tools package itself (actually, upstream prefers if such hooks are shipped by the corresponding packages) For a user this doesn't matter, but either of you should add it, as the laptop-mode-tools package is rather useless if it fails after each suspend. Regards, Christian -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iEYEARECAAYFAkjaasIACgkQOYJY/BTv8jGIDQCff7X4LgobSVFV1jrR0e7BtNCM 6dgAnins/AEXaFkTmwRk7qIZfTwEJSI3 =iBjj -END PGP SIGNATURE- begin:vcard fn:Christian Gogolin n:Gogolin;Christian adr;quoted-printable;quoted-printable:Schillerstra=C3=9Fe 2;;Christian Gogolin;Veitsh=C3=B6chheim;;97209;Germany email;internet:[EMAIL PROTECTED] x-mozilla-html:FALSE url:www.cgogolin.de version:2.1 end:vcard