From paczesiowa at gmail.com Mon Dec 1 03:53:34 2008 From: paczesiowa at gmail.com (Bartek =?utf-8?q?=C4=86wik=C5=82owski?=) Date: Mon, 1 Dec 2008 04:53:34 +0100 Subject: [TuxOnIce-users] nvidia-legacy and >=2.6.25 In-Reply-To: <1227817979.6160.7.camel@nigel-laptop> References: <1227817979.6160.7.camel@nigel-laptop> Message-ID: <200812010453.35041.paczesiowa@dw.pl> > Hi Bartek. > > Just in case you haven't seen the developments, would you please try > (in /etc/hibernate/suspend2.conf): > > powerdownmethod 4 > procsetting extra_pages_allowance 2000 > > (In place of any existing values). > > Regards, > > Nigel Hello Nigel, that powerdownmethod 4 solved my problem. hibernate -r now works great with 2.6.27 and tuxonice-current, however regular hibernate (that should shutdown my machine) instead of shutting it down, reboots just like "-r". The only difference is, that my hdd spins down and immediately starts again (suspend and resume works great though). can you tell me how to fix it, or at least how do you start that S4 thingie, so I can google/experiment on my own, without risking my data on disk (I usually test things with booting from network, with hdd disconnected). I would also like to know why it worked with 2.6.24 and with swsusp. thanks for your help Bartek Cwiklowski PS. sorry for first mail sent from another email account - I'm still not paying enough attention. From ncunningham at crca.org.au Mon Dec 1 06:44:25 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Mon, 01 Dec 2008 17:44:25 +1100 Subject: [TuxOnIce-users] nvidia-legacy and >=2.6.25 In-Reply-To: <200812010453.35041.paczesiowa@dw.pl> References: <1227817979.6160.7.camel@nigel-laptop> <200812010453.35041.paczesiowa@dw.pl> Message-ID: <1228113865.27441.43.camel@nigel-laptop> Hi. On Mon, 2008-12-01 at 04:53 +0100, Bartek ?wik?owski wrote: > > Hi Bartek. > > > > Just in case you haven't seen the developments, would you please try > > (in /etc/hibernate/suspend2.conf): > > > > powerdownmethod 4 > > procsetting extra_pages_allowance 2000 > > > > (In place of any existing values). > > > > Regards, > > > > Nigel > > Hello Nigel, > > that powerdownmethod 4 solved my problem. Good to hear. > hibernate -r now works great with 2.6.27 and tuxonice-current, however regular > hibernate (that should shutdown my machine) instead of shutting it down, > reboots just like "-r". The only difference is, that my hdd spins down and I'm not sure what you mean by "regular hibernate" - would you give more information, please? > immediately starts again (suspend and resume works great though). can you > tell me how to fix it, or at least how do you start that S4 thingie, so I can > google/experiment on my own, without risking my data on disk (I usually test > things with booting from network, with hdd disconnected). > > I would also like to know why it worked with 2.6.24 and with swsusp. swsusp uses platform mode (= powerdownmethod 4) by default. We are currently defaulting to powerdownmethod 0; I'll change that shortly. > thanks for your help You're welcome. Regards, Nigel From paczesiowa at gmail.com Mon Dec 1 17:18:49 2008 From: paczesiowa at gmail.com (Bartek =?utf-8?q?=C4=86wik=C5=82owski?=) Date: Mon, 1 Dec 2008 18:18:49 +0100 Subject: [TuxOnIce-users] nvidia-legacy and >=2.6.25 In-Reply-To: <1228113865.27441.43.camel@nigel-laptop> References: <200812010453.35041.paczesiowa@dw.pl> <1228113865.27441.43.camel@nigel-laptop> Message-ID: <200812011818.49495.paczesiowa@dw.pl> Hi Nigel, > I'm not sure what you mean by "regular hibernate" - would you give more > information, please? I meant hibernate without "-r" argument. > swsusp uses platform mode (= powerdownmethod 4) by default. We are > currently defaulting to powerdownmethod 0; I'll change that shortly. swsusp shutdowns my machine successfully, but it cannot reboot it after suspending (it doesn't have that setting), but tuxonice reboots ok, but instead of shutting down, it reboots as well. Regards, Bartek Cwiklowski From ncunningham at crca.org.au Mon Dec 1 23:31:54 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Tue, 02 Dec 2008 10:31:54 +1100 Subject: [TuxOnIce-users] nvidia-legacy and >=2.6.25 In-Reply-To: <200812011818.49495.paczesiowa@dw.pl> References: <200812010453.35041.paczesiowa@dw.pl> <1228113865.27441.43.camel@nigel-laptop> <200812011818.49495.paczesiowa@dw.pl> Message-ID: <1228174314.27441.48.camel@nigel-laptop> Hi. On Mon, 2008-12-01 at 18:18 +0100, Bartek ?wik?owski wrote: > Hi Nigel, > > > I'm not sure what you mean by "regular hibernate" - would you give more > > information, please? > > I meant hibernate without "-r" argument. Okay. Thanks. > > swsusp uses platform mode (= powerdownmethod 4) by default. We are > > currently defaulting to powerdownmethod 0; I'll change that shortly. > > swsusp shutdowns my machine successfully, but it cannot reboot it after > suspending (it doesn't have that setting), but tuxonice reboots ok, but > instead of shutting down, it reboots as well. I wonder if the reboot option is getting reset. After trying a cycle in which you don't want the system to reboot, would you please cat /sys/power/tuxonice/reboot? If it returns 1, try echoing (as root) 0 into that file before giving it another try. Regards, Nigel From paczesiowa at gmail.com Tue Dec 2 00:35:01 2008 From: paczesiowa at gmail.com (Bartek =?utf-8?q?=C4=86wik=C5=82owski?=) Date: Tue, 2 Dec 2008 01:35:01 +0100 Subject: [TuxOnIce-users] nvidia-legacy and >=2.6.25 In-Reply-To: <1228174314.27441.48.camel@nigel-laptop> References: <200812011818.49495.paczesiowa@dw.pl> <1228174314.27441.48.camel@nigel-laptop> Message-ID: <200812020135.01571.paczesiowa@dw.pl> Hello, > I wonder if the reboot option is getting reset. After trying a cycle in > which you don't want the system to reboot, would you please > cat /sys/power/tuxonice/reboot? > > If it returns 1, try echoing (as root) 0 into that file before giving it > another try. unfortunately, before and after hibernate command (2 attempts), /sys/power/tuxonice/reboot contains 0. Regards, Bartek Cwiklowski From ncunningham at crca.org.au Tue Dec 2 00:50:37 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Tue, 02 Dec 2008 11:50:37 +1100 Subject: [TuxOnIce-users] nvidia-legacy and >=2.6.25 In-Reply-To: <200812020135.01571.paczesiowa@dw.pl> References: <200812011818.49495.paczesiowa@dw.pl> <1228174314.27441.48.camel@nigel-laptop> <200812020135.01571.paczesiowa@dw.pl> Message-ID: <1228179037.27441.59.camel@nigel-laptop> Hi. On Tue, 2008-12-02 at 01:35 +0100, Bartek ?wik?owski wrote: > Hello, > > > I wonder if the reboot option is getting reset. After trying a cycle in > > which you don't want the system to reboot, would you please > > cat /sys/power/tuxonice/reboot? > > > > If it returns 1, try echoing (as root) 0 into that file before giving it > > another try. > > unfortunately, before and after hibernate command (2 attempts), > /sys/power/tuxonice/reboot contains 0. Okay. That's not really unfortunate - it just shows that the issue is something else. Are you using a userui? If so, would you try enabling pausing between steps (press 'P' while hibernating). Then, when you get to the end of writing an image, please check that it says "Powering down." and not "Ready to reboot". Then we'll be completely sure it's not deliberately rebooting. After that, I'd check the powerdown method so we can know how it's trying to power down. Regards, Nigel From paczesiowa at gmail.com Tue Dec 2 01:44:29 2008 From: paczesiowa at gmail.com (Bartek =?utf-8?q?=C4=86wik=C5=82owski?=) Date: Tue, 2 Dec 2008 02:44:29 +0100 Subject: [TuxOnIce-users] nvidia-legacy and >=2.6.25 In-Reply-To: <1228179037.27441.59.camel@nigel-laptop> References: <200812020135.01571.paczesiowa@dw.pl> <1228179037.27441.59.camel@nigel-laptop> Message-ID: <200812020244.29338.paczesiowa@dw.pl> Hi, > Are you using a userui? If so, would you try enabling > pausing between steps (press 'P' while hibernating). Then, when you get > to the end of writing an image, please check that it says "Powering > down." and not "Ready to reboot". Then we'll be completely sure it's not > deliberately rebooting. It says "Powering Down". > After that, I'd check the powerdown method so we can know how it's > trying to power down. I don't know if this is of any help, but both "/sbin/shutdown -h now" and "/sbin/poweroff" work fine. I also tried swsusp again to confirm, and it too reboots instead of shutting down. maybe it is something 2.6.27 related? Regards, Bartek Cwiklowski From paczesiowa at gmail.com Tue Dec 2 02:52:55 2008 From: paczesiowa at gmail.com (Bartek =?utf-8?q?=C4=86wik=C5=82owski?=) Date: Tue, 2 Dec 2008 03:52:55 +0100 Subject: [TuxOnIce-users] nvidia-legacy and >=2.6.25 In-Reply-To: <1228179037.27441.59.camel@nigel-laptop> References: <200812020135.01571.paczesiowa@dw.pl> <1228179037.27441.59.camel@nigel-laptop> Message-ID: <200812020352.55446.paczesiowa@dw.pl> Hi, I had some fun with tuxonice sources, and in kernel/power/tuxonice_power_off.c I commented out hibernation_platform_enter(); from __toi_power_down(), so I use that "alternative power off method", and now I can succesfully reboot AND shutdown after suspending. It is amazing what you and that guy from nvidia forums, who figured agp hack, can make with evil, blackbox, proprietary drivers. Thanks for all your help with nvidia and tuxonice. Regards, Bartek Cwiklowski From ncunningham at crca.org.au Tue Dec 2 02:57:28 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Tue, 02 Dec 2008 13:57:28 +1100 Subject: [TuxOnIce-users] nvidia-legacy and >=2.6.25 In-Reply-To: <200812020352.55446.paczesiowa@dw.pl> References: <200812020135.01571.paczesiowa@dw.pl> <1228179037.27441.59.camel@nigel-laptop> <200812020352.55446.paczesiowa@dw.pl> Message-ID: <1228186648.14432.7.camel@nigel-laptop> Hi again. On Tue, 2008-12-02 at 03:52 +0100, Bartek ?wik?owski wrote: > Hi, > > I had some fun with tuxonice sources, and in > kernel/power/tuxonice_power_off.c I commented out > hibernation_platform_enter(); from __toi_power_down(), so I use that > "alternative power off method", and now I can succesfully reboot AND shutdown > after suspending. It is amazing what you and that guy from nvidia forums, who > figured agp hack, can make with evil, blackbox, proprietary drivers. Thanks > for all your help with nvidia and tuxonice. Okay. So, then, the problem is that you need to use S4 for doing the atomic copy and restore, but need to not use it for powering down. That's certainly doable. I have a patch in the works that I'll send to you in a while, if you're willing to test it. You should then be able to select whatever your old powerdown method was and everything should be fine. Regards, Nigel From paczesiowa at gmail.com Tue Dec 2 03:04:14 2008 From: paczesiowa at gmail.com (Bartek =?utf-8?q?=C4=86wik=C5=82owski?=) Date: Tue, 2 Dec 2008 04:04:14 +0100 Subject: [TuxOnIce-users] nvidia-legacy and >=2.6.25 In-Reply-To: <1228186648.14432.7.camel@nigel-laptop> References: <200812020352.55446.paczesiowa@dw.pl> <1228186648.14432.7.camel@nigel-laptop> Message-ID: <200812020404.14208.paczesiowa@dw.pl> Hi, > I have a patch in the works that I'll send to > you in a while, if you're willing to test it. You should then be able to > select whatever your old powerdown method was and everything should be > fine. no problem, I can test whatever patch you need. Regards, Bartek Cwiklowski From dion at inhex.net Tue Dec 2 09:10:25 2008 From: dion at inhex.net (Dmitry Nezhevenko) Date: Tue, 2 Dec 2008 11:10:25 +0200 Subject: [TuxOnIce-users] nvidia-legacy and >=2.6.25 In-Reply-To: <1228186648.14432.7.camel@nigel-laptop> References: <200812020135.01571.paczesiowa@dw.pl> <1228179037.27441.59.camel@nigel-laptop> <200812020352.55446.paczesiowa@dw.pl> <1228186648.14432.7.camel@nigel-laptop> Message-ID: <20081202091025.GA2935@laptop.local> On Tue, Dec 02, 2008 at 01:57:28PM +1100, Nigel Cunningham wrote: > > Okay. So, then, the problem is that you need to use S4 for doing the > atomic copy and restore, but need to not use it for powering down. > That's certainly doable. I have a patch in the works that I'll send to > you in a while, if you're willing to test it. Am I right, that with this patch it will be possible to use PowerdownMethod 5 as final shutdown for recent NVIDIA drivers (that lockups xorg just after resume without S4)? Currently I'm using PowerdownMethod 5 as suggested by you. I've tried to unplug laptop battery and power supply after such S4. Resuming is done correctly after this. -- WBR, Dmitry -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 197 bytes Desc: Digital signature Url : http://lists.tuxonice.net/pipermail/tuxonice-users/attachments/20081202/88d174c6/attachment.pgp From ncunningham at crca.org.au Tue Dec 2 11:25:32 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Tue, 02 Dec 2008 22:25:32 +1100 Subject: [TuxOnIce-users] nvidia-legacy and >=2.6.25 In-Reply-To: <20081202091025.GA2935@laptop.local> References: <200812020135.01571.paczesiowa@dw.pl> <1228179037.27441.59.camel@nigel-laptop> <200812020352.55446.paczesiowa@dw.pl> <1228186648.14432.7.camel@nigel-laptop> <20081202091025.GA2935@laptop.local> Message-ID: <1228217132.25380.329.camel@nigel-laptop> Hi. On Tue, 2008-12-02 at 11:10 +0200, Dmitry Nezhevenko wrote: > On Tue, Dec 02, 2008 at 01:57:28PM +1100, Nigel Cunningham wrote: > > > > Okay. So, then, the problem is that you need to use S4 for doing the > > atomic copy and restore, but need to not use it for powering down. > > That's certainly doable. I have a patch in the works that I'll send to > > you in a while, if you're willing to test it. > > Am I right, that with this patch it will be possible to use > PowerdownMethod 5 as final shutdown for recent NVIDIA drivers (that > lockups xorg just after resume without S4)? > > Currently I'm using PowerdownMethod 5 as suggested by you. I've tried to > unplug laptop battery and power supply after such S4. Resuming is done > correctly after this. Yes, that's the aim. The patch will make TuxOnIce always run the S4 code when doing the atomic copy and restore, but use the code for your chosen powerdown method when doing the actual power off. Given what you say, it sounds like it shouldn't cause any problems. Regards, Nigel From ncunningham at crca.org.au Tue Dec 2 11:29:31 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Tue, 02 Dec 2008 22:29:31 +1100 Subject: [TuxOnIce-users] nvidia-legacy and >=2.6.25 In-Reply-To: <200812020244.29338.paczesiowa@dw.pl> References: <200812020135.01571.paczesiowa@dw.pl> <1228179037.27441.59.camel@nigel-laptop> <200812020244.29338.paczesiowa@dw.pl> Message-ID: <1228217371.25380.334.camel@nigel-laptop> Hi. On Tue, 2008-12-02 at 02:44 +0100, Bartek ?wik?owski wrote: > Hi, > > > Are you using a userui? If so, would you try enabling > > pausing between steps (press 'P' while hibernating). Then, when you get > > to the end of writing an image, please check that it says "Powering > > down." and not "Ready to reboot". Then we'll be completely sure it's not > > deliberately rebooting. > > It says "Powering Down". Great. > > After that, I'd check the powerdown method so we can know how it's > > trying to power down. > > I don't know if this is of any help, but both "/sbin/shutdown -h now" and > "/sbin/poweroff" work fine. > > I also tried swsusp again to confirm, and it too reboots instead of shutting > down. maybe it is something 2.6.27 related? Sounds like you might have some sort of regression. It worked fine with 2.6.26 or earlier? Regards, Nigel From acelists at atlas.sk Tue Dec 2 15:51:20 2008 From: acelists at atlas.sk (ace) Date: Tue, 02 Dec 2008 16:51:20 +0100 Subject: [TuxOnIce-users] Hibernation support among graphics cards Message-ID: <49355978.4030704@atlas.sk> Hi guys. I am planning to get a new PC and a new graphics card. I am wondering which brands and drivers are working with tuxonice. Maybe it will help more people to have an overview list like this. My experiences so far: - Nvidia old era (TNT2) + free nv driver (no 3D accel) = no problems. - Nvidia old era (TNT2) + binary closed driver (3D accel) = it is blacklisted by tuxonice, no hibernation. The driver itself doesn't even work lately on my machine. - Radeon middle era (9550) + free radeon driver (3D accel) = no problems. My observations from this list: - Nvidia new era (8xxx - 9xxxx) + binary closed driver (3D accel) = varying success. Depends on driver version, card and other factors. My questions: - Radeon new era (radeon HD 3xxxx - 4xxx) + free radeonHD driver (3D accel?)= ??? (any experieces?) - Intel new era (9xx) + (free?) driver (3D) = ??? (any experieces?) This is on a desktop PC. Laptops probably have their own share of problems with ACPI and gfx cards. Has anybody experiences with these last 2 families? That would help me. Thanks in advance, Peter From paczesiowa at gmail.com Tue Dec 2 22:35:09 2008 From: paczesiowa at gmail.com (Bartek =?utf-8?q?=C4=86wik=C5=82owski?=) Date: Tue, 2 Dec 2008 23:35:09 +0100 Subject: [TuxOnIce-users] nvidia-legacy and >=2.6.25 In-Reply-To: <1228217371.25380.334.camel@nigel-laptop> References: <200812020244.29338.paczesiowa@dw.pl> <1228217371.25380.334.camel@nigel-laptop> Message-ID: <200812022335.09259.paczesiowa@dw.pl> Hi, > > I also tried swsusp again to confirm, and it too reboots instead of > > shutting down. maybe it is something 2.6.27 related? > > Sounds like you might have some sort of regression. It worked fine with > 2.6.26 or earlier? Today I did some testing of suspending followed by shutdown, with various kernel versions and s2disk. I also modified "shutdown method" from /etc/suspend.conf. here are the results: A -> shutdown method = shutdown B -> shutdown method = platform now 2.6.26A hangs on suspending just like tuxonice with default powerdownmethod 2.6.26B works great 2.6.27A suspends, shutdowns, but hangs with black screen after resume 2.6.27B reboots instead of shutting down - otherwise works. 2.6.28A hangs on suspending just like tuxonice with default powerdownmethod 2.6.28B reboots instead of shutting down - otherwise works. so looks like platform thingie is broken since 2.6.27 and shutdown method since 2.6.25. Anyway it doesn't look look like it is tuxonice fault (check https://kerneltrap.org/mailarchive/linux-kernel/2008/7/16/2530114/thread ), so you don't have to write any patches just for me, I'm perfectly fine with commenting out that one line, I wrote about earlier, in my own 2.6.27 with tuxonice. Regards, Bartek Cwiklowski From paczesiowa at gmail.com Wed Dec 3 00:02:15 2008 From: paczesiowa at gmail.com (Bartek =?utf-8?q?=C4=86wik=C5=82owski?=) Date: Wed, 3 Dec 2008 01:02:15 +0100 Subject: [TuxOnIce-users] Can I leave network media mounted ro? Message-ID: <200812030102.15938.paczesiowa@dw.pl> Hi, http://www.tuxonice.net/HOWTO-4.html mentions that I have to unmount nfs/smbfs mounted media during hibernation because I "risk losing data". Is it true if I have some shares mounted read-only (they are also exported read-only)? I have some movies and music exported from my file-server and I don't like killing mediaplayers just because I have to unmount their data. thanks in advance, Bartek Cwiklowski From ncunningham at crca.org.au Wed Dec 3 00:48:16 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Wed, 03 Dec 2008 11:48:16 +1100 Subject: [TuxOnIce-users] Can I leave network media mounted ro? In-Reply-To: <200812030102.15938.paczesiowa@dw.pl> References: <200812030102.15938.paczesiowa@dw.pl> Message-ID: <1228265296.25380.525.camel@nigel-laptop> Gidday. On Wed, 2008-12-03 at 01:02 +0100, Bartek ?wik?owski wrote: > Hi, > > http://www.tuxonice.net/HOWTO-4.html mentions that I have to unmount nfs/smbfs > mounted media during hibernation because I "risk losing data". Is it true if I > have some shares mounted read-only (they are also exported read-only)? I have > some movies and music exported from my file-server and I don't like killing > mediaplayers just because I have to unmount their data. If data on the server changes while you're hibernated and the client on your side doesn't refresh its data when resuming, you might run into trouble. If you can tell your computer to invalidate local caches, it might work. Regards, Nigel From ncunningham at crca.org.au Wed Dec 3 00:50:04 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Wed, 03 Dec 2008 11:50:04 +1100 Subject: [TuxOnIce-users] Hibernation support among graphics cards In-Reply-To: <49355978.4030704@atlas.sk> References: <49355978.4030704@atlas.sk> Message-ID: <1228265404.25380.527.camel@nigel-laptop> Hi. On Tue, 2008-12-02 at 16:51 +0100, ace wrote: > Hi guys. > > I am planning to get a new PC and a new graphics card. I am wondering > which brands and drivers are working with tuxonice. Maybe it > will help more people to have an overview list like this. > > My experiences so far: > - Nvidia old era (TNT2) + free nv driver (no 3D accel) = no problems. > - Nvidia old era (TNT2) + binary closed driver (3D accel) = it is blacklisted by tuxonice, no hibernation. > The driver itself doesn't even work lately on my machine. It shouldn't be blacklisted anymore. I've just removed the blacklisting in SVN (with kernels post 2.6.25). > - Radeon middle era (9550) + free radeon driver (3D accel) = no problems. > > My observations from this list: > - Nvidia new era (8xxx - 9xxxx) + binary closed driver (3D accel) = varying success. > Depends on driver version, card and other factors. They should all be reliable, hence the removal above. > My questions: > - Radeon new era (radeon HD 3xxxx - 4xxx) + free radeonHD driver (3D accel?)= ??? (any experieces?) > - Intel new era (9xx) + (free?) driver (3D) = ??? (any experieces?) Not here. Hope this helps. Nigel From paczesiowa at gmail.com Wed Dec 3 02:52:57 2008 From: paczesiowa at gmail.com (Bartek =?utf-8?q?=C4=86wik=C5=82owski?=) Date: Wed, 3 Dec 2008 03:52:57 +0100 Subject: [TuxOnIce-users] Can I leave network media mounted ro? In-Reply-To: <1228265296.25380.525.camel@nigel-laptop> References: <200812030102.15938.paczesiowa@dw.pl> <1228265296.25380.525.camel@nigel-laptop> Message-ID: <200812030352.57557.paczesiowa@dw.pl> Hi, > If data on the server changes while you're hibernated and the client on > your side doesn't refresh its data when resuming, you might run into > trouble. But what trouble exactly? If I hibernate while playing some mp3 file and it is deleted/changed during poweroff time (highly unlikely) will it crash my player (perfectly acceptable for those rare situations) or something worse? > If you can tell your computer to invalidate local caches, it might work. I googled "sync && echo N > /proc/sys/vm/drop_caches", but N depends on what exactly I want to free (pagecache, dentries, inodes). Also, when it should be run? OnResume 0 is started after processes are unfreezed, so it would be too late, right? Regards, Bartek Cwiklowski From ncunningham at crca.org.au Wed Dec 3 03:10:00 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Wed, 03 Dec 2008 14:10:00 +1100 Subject: [TuxOnIce-users] Can I leave network media mounted ro? In-Reply-To: <200812030352.57557.paczesiowa@dw.pl> References: <200812030102.15938.paczesiowa@dw.pl> <1228265296.25380.525.camel@nigel-laptop> <200812030352.57557.paczesiowa@dw.pl> Message-ID: <1228273800.5357.7.camel@nigel-laptop> Hi again. On Wed, 2008-12-03 at 03:52 +0100, Bartek ?wik?owski wrote: > Hi, > > > If data on the server changes while you're hibernated and the client on > > your side doesn't refresh its data when resuming, you might run into > > trouble. > > But what trouble exactly? If I hibernate while playing some mp3 file and it is > deleted/changed during poweroff time (highly unlikely) will it crash my player > (perfectly acceptable for those rare situations) or something worse? I don't know the details of how NFS etc work. If they're paranoid about validating requests from clients (as I'd expect them to be), the main potential for problems would be in your client being told that data it has is suddenly invalid. If those error paths are properly debugged, you should be right. > > If you can tell your computer to invalidate local caches, it might work. > > I googled "sync && echo N > /proc/sys/vm/drop_caches", but N depends on what > exactly I want to free (pagecache, dentries, inodes). Also, when it should be > run? OnResume 0 is started after processes are unfreezed, so it would be too > late, right? TuxOnIce already has support for dropping page cache - set /sys/power/tuxonice/image_size_limit to -2. It's done when preparing the image rather than when resuming, but I don't think that should matter. Regards, Nigel From karlis.repsons at gmail.com Wed Dec 3 14:11:49 2008 From: karlis.repsons at gmail.com (=?utf-8?q?K=C4=81rlis_Repsons?=) Date: Wed, 3 Dec 2008 14:11:49 +0000 Subject: [TuxOnIce-users] forced shutdown of UPS after RAM contents are saved? In-Reply-To: <1228081994.27441.24.camel@nigel-laptop> References: <200811291029.16495.Karlis.Repsons@gmail.com> <200811301252.24443.Karlis.Repsons@gmail.com> <1228081994.27441.24.camel@nigel-laptop> Message-ID: <200812031411.49571.Karlis.Repsons@gmail.com> On Sunday 30 November 2008 21:53:14 Nigel Cunningham wrote: > Hi again. .. > > 1. Keep Image Mode. > > The first thing to do is prepare a kernel that has TuxOnIce's keep image > mode enabled. ready. > We'll use this below to allow you to hibernate our special > 'power down the UPS' initrd/ramfs once how will that happen? So I have to include everything hibernate related into initramfs or that is kernel supported? > , then resume it over and over to > do the actual powering down. clear... > 2. Storage of the images. > > You need to two places to store images. One will need the normal amount > of storage that you'd use for an image, and the other can be quite small > (say 20 meg). It will probably be simplest to make the small one a > partition. Ah, now I learn, is not a bad idea to have extended partition and free space.. Is it hard / risky to use files? I have sufficient amount of free space in boot partition and no real will to delete my existing partitions! .. > > Hope this is clear and helpful! Thanks. > Nigel How stable is tuxonice with nvidia cards? Last time I used tuxonice with them, I got hangs at resume instead of enhancements... k. From acelists at atlas.sk Wed Dec 3 20:24:56 2008 From: acelists at atlas.sk (ace) Date: Wed, 03 Dec 2008 21:24:56 +0100 Subject: [TuxOnIce-users] Hibernation support among graphics cards In-Reply-To: <1228265404.25380.527.camel@nigel-laptop> References: <49355978.4030704@atlas.sk> <1228265404.25380.527.camel@nigel-laptop> Message-ID: <4936EB18.10809@atlas.sk> Nigel Cunningham wrote / nap?sal(a): > Hi. > > On Tue, 2008-12-02 at 16:51 +0100, ace wrote: >> Hi guys. >> >> I am planning to get a new PC and a new graphics card. I am wondering >> which brands and drivers are working with tuxonice. Maybe it >> will help more people to have an overview list like this. >> >> My experiences so far: >> - Nvidia old era (TNT2) + free nv driver (no 3D accel) = no problems. >> - Nvidia old era (TNT2) + binary closed driver (3D accel) = it is blacklisted by tuxonice, no hibernation. >> The driver itself doesn't even work lately on my machine. > > It shouldn't be blacklisted anymore. I've just removed the blacklisting > in SVN (with kernels post 2.6.25). But you should probably still distinguish between driver versions. I am talking about the 71.86.xx series (oldest legacy class). That one oficially supports only APM suspend and ACPI S3... If you remove blacklist, it is probably safe only for the newest 1xx.xx range. > >> - Radeon middle era (9550) + free radeon driver (3D accel) = no problems. >> >> My observations from this list: >> - Nvidia new era (8xxx - 9xxxx) + binary closed driver (3D accel) = varying success. >> Depends on driver version, card and other factors. > > They should all be reliable, hence the removal above. > Hope this helps. Yes, thanks for your hints. Peter From pierre at mouraf.org Wed Dec 3 21:08:12 2008 From: pierre at mouraf.org (Pierre-Alexandre Meyer) Date: Wed, 3 Dec 2008 22:08:12 +0100 Subject: [TuxOnIce-users] Dual images Message-ID: <20081203210812.GI17410@panda> Hi all, I am currently trying to use the kiosk feature (keep_image) in order to speed up my boot time: echo /dev/sda6 > /sys/power/tuxonice/resume echo 1 > /sys/power/tuxonice/keep_image This works great and allows me to have a pristine image of my system on /dev/sda6 (swap partition) from which I can resume. But I would also like to be able to suspend-to-disk on another partition (/dev/sda7), using it as a regular hibernate partition, that is not permanent across reboots. "Boot" is performed via /dev/sda6 and if I want to save temporarily the state of my machine, I hibernate on /dev/sda7. This can be done in the initrd easily: ... # First, try to resume from the swap partition (suspend triggered by he calling hibernate) echo /dev/sda7 > /sys/power/tuxonice/resume echo 1 > /sys/power/tuxonice/do_resume # If the initrd comes here, I didn't hibernate, it was probably a shutdown # Let's try to use the kiosk image echo /dev/sda6 > /sys/power/tuxonice/resume echo 1 > /sys/power/tuxonice/do_resume # If the initrd comes here, the file wasn't valid. Fresh install maybe? ... Both partitions are swapped on in the kiosk image (/dev/sda7 with highest priority, /dev/sda6 with lowest). The issue I run into is that TuxOnIce will remove the kiosk image on /dev/sda6 if I try to: echo /dev/sda7 > /sys/power/tuxonice/resume echo 0 > /sys/power/tuxonice/keep_image On resume both swap partitions are empty. Ideally I would disable /dev/sda6 when trying to hibernate on /dev/sda7. I can't use swapoff (swap/swapfilename cannot do the trick) - since the kiosk pages on /dev/sda6 are "used". Any idea how to "disable" this swap partition? Or is there a way to tell TuxOnIce to ignore one partition? If there is a special TuxOnIce signature on each page in an image stored on the swap area, I could write a patch to ignore some of these. Thanks, -- Pierre-Alexandre Meyer From ncunningham at crca.org.au Wed Dec 3 21:56:59 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Thu, 04 Dec 2008 08:56:59 +1100 Subject: [TuxOnIce-users] Dual images In-Reply-To: <20081203210812.GI17410@panda> References: <20081203210812.GI17410@panda> Message-ID: <1228341419.8648.29.camel@nigel-laptop> Hi Pierre-Alexandre. On Wed, 2008-12-03 at 22:08 +0100, Pierre-Alexandre Meyer wrote: > Hi all, > > I am currently trying to use the kiosk feature (keep_image) in order > to speed up my boot time: > > echo /dev/sda6 > /sys/power/tuxonice/resume > echo 1 > /sys/power/tuxonice/keep_image > > This works great and allows me to have a pristine image of my system on > /dev/sda6 (swap partition) from which I can resume. Is the data stored in /dev/sda6 from unchanging filesystems? If there's any chance of the data stored there being changed, using kiosk mode is not safe. > But I would also like to be able to suspend-to-disk on another partition > (/dev/sda7), using it as a regular hibernate partition, that is not > permanent across reboots. "Boot" is performed via /dev/sda6 and if I > want to save temporarily the state of my machine, I hibernate on > /dev/sda7. This can be done in the initrd easily: > > ... > > # First, try to resume from the swap partition (suspend triggered by > he calling hibernate) > echo /dev/sda7 > /sys/power/tuxonice/resume > echo 1 > /sys/power/tuxonice/do_resume > > # If the initrd comes here, I didn't hibernate, it was probably > a shutdown > # Let's try to use the kiosk image > echo /dev/sda6 > /sys/power/tuxonice/resume > echo 1 > /sys/power/tuxonice/do_resume > > # If the initrd comes here, the file wasn't valid. Fresh install maybe? > > ... > > Both partitions are swapped on in the kiosk image (/dev/sda7 with highest priority, > /dev/sda6 with lowest). > > The issue I run into is that TuxOnIce will remove the kiosk image on > /dev/sda6 if I try to: > > echo /dev/sda7 > /sys/power/tuxonice/resume > echo 0 > /sys/power/tuxonice/keep_image > > On resume both swap partitions are empty. Hmmm.... the difficulty comes from the fact that I never imagined anyone doing something like this, and so never wrote code to support having multiple valid images at the same time from one context. It's doable, just not something that's supported at the moment. > Ideally I would disable /dev/sda6 when trying to hibernate on /dev/sda7. > I can't use swapoff (swap/swapfilename cannot do the trick) - since the > kiosk pages on /dev/sda6 are "used". > > Any idea how to "disable" this swap partition? Or is there a way to tell > TuxOnIce to ignore one partition? > If there is a special TuxOnIce signature on each page in an image stored > on the swap area, I could write a patch to ignore some of these. The simplest way to get around this would be to use the file allocator instead of swap. Would you consider that? Regards, Nigel From ncunningham at crca.org.au Wed Dec 3 22:03:45 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Thu, 04 Dec 2008 09:03:45 +1100 Subject: [TuxOnIce-users] forced shutdown of UPS after RAM contents are saved? In-Reply-To: <200812031411.49571.Karlis.Repsons@gmail.com> References: <200811291029.16495.Karlis.Repsons@gmail.com> <200811301252.24443.Karlis.Repsons@gmail.com> <1228081994.27441.24.camel@nigel-laptop> <200812031411.49571.Karlis.Repsons@gmail.com> Message-ID: <1228341825.8648.35.camel@nigel-laptop> Hi. On Wed, 2008-12-03 at 14:11 +0000, K?rlis Repsons wrote: > On Sunday 30 November 2008 21:53:14 Nigel Cunningham wrote: > > Hi again. > .. > > > > > 1. Keep Image Mode. > > > > The first thing to do is prepare a kernel that has TuxOnIce's keep image > > mode enabled. > ready. > > > We'll use this below to allow you to hibernate our special > > 'power down the UPS' initrd/ramfs once > how will that happen? So I have to include everything hibernate related into > initramfs or that is kernel supported? You need to put everything related to telling the UPS to power down into that initramfs, and anything necessary to set up for hibernating. The simplest case would be that you have a kernel with all of the drivers needed for accessing your hibernation storage compiled in, and your binary for powering off your UPS is statically compiled. You'd then do something along the lines of: echo > /sys/power/tuxonice/do_hibernate /bin/my-ups-command --powerdown This initramfs could even be your normal one if it was make to understand an extra commandline option and use that to trigger the above commands, provided that they were run prior to mount root (and other) filesystems. > > , then resume it over and over to > > do the actual powering down. > clear... > > > 2. Storage of the images. > > > > You need to two places to store images. One will need the normal amount > > of storage that you'd use for an image, and the other can be quite small > > (say 20 meg). It will probably be simplest to make the small one a > > partition. > Ah, now I learn, is not a bad idea to have extended partition and free space.. > Is it hard / risky to use files? I have sufficient amount of free space in > boot partition and no real will to delete my existing partitions! Both swap files and ordinary files (using the file allocator) should be fine. > How stable is tuxonice with nvidia cards? Last time I used tuxonice with them, > I got hangs at resume instead of enhancements... Try powerdown method 4. It has fixed a number of issues people have seen recently. Regards, Nigel From ncunningham at crca.org.au Wed Dec 3 23:30:31 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Thu, 04 Dec 2008 10:30:31 +1100 Subject: [TuxOnIce-users] Dual images In-Reply-To: <20081203232215.GJ17410@panda> References: <20081203210812.GI17410@panda> <1228341419.8648.29.camel@nigel-laptop> <20081203232215.GJ17410@panda> Message-ID: <1228347031.8648.39.camel@nigel-laptop> Hi again. On Thu, 2008-12-04 at 00:22 +0100, Pierre-Alexandre Meyer wrote: > On Thursday 04 December 2008 ? 08:56:59AM, Nigel Cunningham wrote: > > Hi Pierre-Alexandre. > > Hi Nigel, > > > Is the data stored in /dev/sda6 from unchanging filesystems? If there's > > any chance of the data stored there being changed, using kiosk mode is > > not safe. > > I think (correct me if I am wrong) my installation is safe. > > The kiosk image has been created right after a pristine installation, with no > writable filesystem mounted. I mount my local data storage at each resume from > this image. > The Linux distribution itself is simply a squash file union mounted with a tmpfs > (basic Ubuntu). That sounds like it should be safe. > > The simplest way to get around this would be to use the file allocator > > instead of swap. Would you consider that? > > Sure. I actually ran across one of your posts on the archives: > > http://lists.tuxonice.net/lurker/message/20080725.061913.5ff50296.en.html > > You are stating: > > > By writing the image to a file, you don't have to worry about whether > > swap is enabled or not, and can still use keep-image mode. > > I can definitively try it but I don't understand how that would be > different. My understanding was that swap partition or swap file is > actually the same for the kernel - a swap area. You need to swapon (wathever > that actually means) this area either case. Am I wrong? I don't mean a swapfile, but an ordinary file, using the file allocator rather than the swap allocator. In this case, swap is not in the picture at all - the file is only used for hibernation. Regards, Nigel From pierre at mouraf.org Wed Dec 3 23:22:15 2008 From: pierre at mouraf.org (Pierre-Alexandre Meyer) Date: Thu, 4 Dec 2008 00:22:15 +0100 Subject: [TuxOnIce-users] Dual images In-Reply-To: <1228341419.8648.29.camel@nigel-laptop> References: <20081203210812.GI17410@panda> <1228341419.8648.29.camel@nigel-laptop> Message-ID: <20081203232215.GJ17410@panda> On Thursday 04 December 2008 ? 08:56:59AM, Nigel Cunningham wrote: > Hi Pierre-Alexandre. Hi Nigel, > Is the data stored in /dev/sda6 from unchanging filesystems? If there's > any chance of the data stored there being changed, using kiosk mode is > not safe. I think (correct me if I am wrong) my installation is safe. The kiosk image has been created right after a pristine installation, with no writable filesystem mounted. I mount my local data storage at each resume from this image. The Linux distribution itself is simply a squash file union mounted with a tmpfs (basic Ubuntu). > The simplest way to get around this would be to use the file allocator > instead of swap. Would you consider that? Sure. I actually ran across one of your posts on the archives: http://lists.tuxonice.net/lurker/message/20080725.061913.5ff50296.en.html You are stating: > By writing the image to a file, you don't have to worry about whether > swap is enabled or not, and can still use keep-image mode. I can definitively try it but I don't understand how that would be different. My understanding was that swap partition or swap file is actually the same for the kernel - a swap area. You need to swapon (wathever that actually means) this area either case. Am I wrong? Thanks for your answer, -- Pierre-Alexandre Meyer From ncunningham at crca.org.au Wed Dec 3 23:46:24 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Thu, 04 Dec 2008 10:46:24 +1100 Subject: [TuxOnIce-users] End of line for 2.6.22, 2.6.23, Ubuntu Feisty support. Message-ID: <1228347984.8648.41.camel@nigel-laptop> Hi all. My installed headers are causing me grief with compiling older kernels, and they are quite out of date now anyway, so I'm going to discontinue support for the above kernels, unless someone gives me a good reason not to. Regards, Nigel From pierre at mouraf.org Wed Dec 3 23:42:08 2008 From: pierre at mouraf.org (Pierre-Alexandre Meyer) Date: Thu, 4 Dec 2008 00:42:08 +0100 Subject: [TuxOnIce-users] Dual images In-Reply-To: <1228347031.8648.39.camel@nigel-laptop> References: <20081203210812.GI17410@panda> <1228341419.8648.29.camel@nigel-laptop> <20081203232215.GJ17410@panda> <1228347031.8648.39.camel@nigel-laptop> Message-ID: <20081203234208.GN17410@panda> On Thursday 04 December 2008 ? 10:30:31AM, Nigel Cunningham wrote: > I don't mean a swapfile, but an ordinary file, using the file allocator > rather than the swap allocator. In this case, swap is not in the picture > at all - the file is only used for hibernation. Oh, my bad - I mixed the two of them. I'll try that! Best, -- Pierre-Alexandre Meyer From karlis.repsons at gmail.com Thu Dec 4 09:37:27 2008 From: karlis.repsons at gmail.com (=?utf-8?q?K=C4=81rlis_Repsons?=) Date: Thu, 4 Dec 2008 09:37:27 +0000 Subject: [TuxOnIce-users] forced shutdown of UPS after RAM contents are saved? In-Reply-To: <1228341825.8648.35.camel@nigel-laptop> References: <200811291029.16495.Karlis.Repsons@gmail.com> <200812031411.49571.Karlis.Repsons@gmail.com> <1228341825.8648.35.camel@nigel-laptop> Message-ID: <200812040937.27480.Karlis.Repsons@gmail.com> On Wednesday 03 December 2008 22:03:45 Nigel Cunningham wrote: > Hi. > > On Wed, 2008-12-03 at 14:11 +0000, K?rlis Repsons wrote: > > On Sunday 30 November 2008 21:53:14 Nigel Cunningham wrote: > > > Hi again. > > > > .. > > > > > 1. Keep Image Mode. > > > > > > The first thing to do is prepare a kernel that has TuxOnIce's keep > > > image mode enabled. > > > > ready. > > > > > We'll use this below to allow you to hibernate our special > > > 'power down the UPS' initrd/ramfs once > > > > how will that happen? So I have to include everything hibernate related > > into initramfs or that is kernel supported? > > You need to put everything related to telling the UPS to power down into > that initramfs, and anything necessary to set up for hibernating. The > simplest case would be that you have a kernel with all of the drivers > needed for accessing your hibernation storage compiled in, and your > binary for powering off your UPS is statically compiled. You'd then do > something along the lines of: > > echo > /sys/power/tuxonice/do_hibernate > /bin/my-ups-command --powerdown But how to put hibernation things in there? If I have all drivers in kernel, is it sufficient to do $(echo > /sys/power/tuxonice/do_hibernate)? Why is hibernate script using all those monstrous code lines?? k. From ncunningham at crca.org.au Thu Dec 4 21:23:54 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Fri, 05 Dec 2008 08:23:54 +1100 Subject: [TuxOnIce-users] forced shutdown of UPS after RAM contents are saved? In-Reply-To: <200812040937.27480.Karlis.Repsons@gmail.com> References: <200811291029.16495.Karlis.Repsons@gmail.com> <200812031411.49571.Karlis.Repsons@gmail.com> <1228341825.8648.35.camel@nigel-laptop> <200812040937.27480.Karlis.Repsons@gmail.com> Message-ID: <1228425834.8648.100.camel@nigel-laptop> Hi. On Thu, 2008-12-04 at 09:37 +0000, K?rlis Repsons wrote: > On Wednesday 03 December 2008 22:03:45 Nigel Cunningham wrote: > > On Wed, 2008-12-03 at 14:11 +0000, K?rlis Repsons wrote: > > > On Sunday 30 November 2008 21:53:14 Nigel Cunningham wrote: > > > > Hi again. > > > > > > .. > > > > > > > 1. Keep Image Mode. > > > > > > > > The first thing to do is prepare a kernel that has TuxOnIce's keep > > > > image mode enabled. > > > > > > ready. > > > > > > > We'll use this below to allow you to hibernate our special > > > > 'power down the UPS' initrd/ramfs once > > > > > > how will that happen? So I have to include everything hibernate related > > > into initramfs or that is kernel supported? > > > > You need to put everything related to telling the UPS to power down into > > that initramfs, and anything necessary to set up for hibernating. The > > simplest case would be that you have a kernel with all of the drivers > > needed for accessing your hibernation storage compiled in, and your > > binary for powering off your UPS is statically compiled. You'd then do > > something along the lines of: > > > > echo > /sys/power/tuxonice/do_hibernate > > /bin/my-ups-command --powerdown > > But how to put hibernation things in there? If I have all drivers in kernel, > is it sufficient to do $(echo > /sys/power/tuxonice/do_hibernate)? Why is > hibernate script using all those monstrous code lines?? The hibernate script seeks to deal with services that need unloading, modules that might need unloading and so on, in a more generic way. If you're starting a hibernation at boot time, those services won't have been started yet, and modules won't have been loaded. Hence, if you have all the drivers you need in the kernel, a simple echo (or echo something) should do the trick. (I say "echo something" because some shells that are used in initramfses do nothing if told "echo > do_hibernate" - they need "echo something > do_hibernate" instead. Hope that helps. Nigel From ncunningham at crca.org.au Thu Dec 4 21:25:34 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Fri, 05 Dec 2008 08:25:34 +1100 Subject: [TuxOnIce-users] Hibernation support among graphics cards In-Reply-To: <4936EB18.10809@atlas.sk> References: <49355978.4030704@atlas.sk> <1228265404.25380.527.camel@nigel-laptop> <4936EB18.10809@atlas.sk> Message-ID: <1228425934.8648.102.camel@nigel-laptop> Hi. On Wed, 2008-12-03 at 21:24 +0100, ace wrote: > Nigel Cunningham wrote / nap?sal(a): > > Hi. > > > > On Tue, 2008-12-02 at 16:51 +0100, ace wrote: > >> Hi guys. > >> > >> I am planning to get a new PC and a new graphics card. I am wondering > >> which brands and drivers are working with tuxonice. Maybe it > >> will help more people to have an overview list like this. > >> > >> My experiences so far: > >> - Nvidia old era (TNT2) + free nv driver (no 3D accel) = no problems. > >> - Nvidia old era (TNT2) + binary closed driver (3D accel) = it is blacklisted by tuxonice, no hibernation. > >> The driver itself doesn't even work lately on my machine. > > > > It shouldn't be blacklisted anymore. I've just removed the blacklisting > > in SVN (with kernels post 2.6.25). > But you should probably still distinguish between driver versions. I am > talking about the 71.86.xx series (oldest legacy class). That one > oficially supports only APM suspend and ACPI S3... If you remove > blacklist, it is probably safe only for the newest 1xx.xx range. Okay. How about this?... @nvidia 0.0 100.00 Regards, Nigel From acelists at atlas.sk Thu Dec 4 21:44:11 2008 From: acelists at atlas.sk (ace) Date: Thu, 04 Dec 2008 22:44:11 +0100 Subject: [TuxOnIce-users] Hibernation support among graphics cards In-Reply-To: <1228425934.8648.102.camel@nigel-laptop> References: <49355978.4030704@atlas.sk> <1228265404.25380.527.camel@nigel-laptop> <4936EB18.10809@atlas.sk> <1228425934.8648.102.camel@nigel-laptop> Message-ID: <49384F2B.4070906@atlas.sk> Nigel Cunningham wrote / nap?sal(a): > Hi. > > On Wed, 2008-12-03 at 21:24 +0100, ace wrote: >> Nigel Cunningham wrote / nap?sal(a): >>> Hi. >>> >>> On Tue, 2008-12-02 at 16:51 +0100, ace wrote: >>>> Hi guys. >>>> >>>> I am planning to get a new PC and a new graphics card. I am wondering >>>> which brands and drivers are working with tuxonice. Maybe it >>>> will help more people to have an overview list like this. >>>> >>>> My experiences so far: >>>> - Nvidia old era (TNT2) + free nv driver (no 3D accel) = no problems. >>>> - Nvidia old era (TNT2) + binary closed driver (3D accel) = it is blacklisted by tuxonice, no hibernation. >>>> The driver itself doesn't even work lately on my machine. >>> It shouldn't be blacklisted anymore. I've just removed the blacklisting >>> in SVN (with kernels post 2.6.25). >> But you should probably still distinguish between driver versions. I am >> talking about the 71.86.xx series (oldest legacy class). That one >> oficially supports only APM suspend and ACPI S3... If you remove >> blacklist, it is probably safe only for the newest 1xx.xx range. > > Okay. How about this?... > > @nvidia 0.0 100.00 Yes, let's try that. Of course, the 100.00 number is just arbitrary. Users will have to experiment since which driver release is hibernation working fine. And please release a new hibernate script package someday, so that we can use it, without going to SVN :) Bye, Peter From pierre at mouraf.org Fri Dec 5 01:02:09 2008 From: pierre at mouraf.org (Pierre-Alexandre Meyer) Date: Fri, 5 Dec 2008 02:02:09 +0100 Subject: [TuxOnIce-users] Dual images In-Reply-To: <1228347031.8648.39.camel@nigel-laptop> References: <20081203210812.GI17410@panda> <1228341419.8648.29.camel@nigel-laptop> <20081203232215.GJ17410@panda> <1228347031.8648.39.camel@nigel-laptop> Message-ID: <20081205010209.GQ17410@panda> On Thursday 04 December 2008 ? 10:30:31AM, Nigel Cunningham wrote: > I don't mean a swapfile, but an ordinary file, using the file allocator > rather than the swap allocator. In this case, swap is not in the picture > at all - the file is only used for hibernation. I managed to hibernate using a regular file on a ext2 filesystem - the file allocator seems to work on my system - but I have issues when I try to use a raw device (specifying a block device): This is what I tried: echo "TuxOnIce" > /dev/sda6 dd if=/dev/zero bs=1M count=950 >> /dev/sda6 I can double check: 16:58 pierre at ubuntu ~% od -N20 -c /dev/sda6 0000000 T u x O n I c e \n \0 \0 \0 \0 \0 \0 \0 0000020 \0 \0 \0 \0 0000024 But if I try to hibernate by: echo /dev/sda6 > /sys/power/tuxonice/file/target and specifying resume=file:/dev/sda6 (returned by /sys/power/tuxonice/resume), I trigger a BUG() on resume (no error when hibernating): [ 25.267903] TuxOnIce: Found an hibernate image. [ 25.327129] Failed to launch userspace program '/usr/local/sbin/tuxonice_fbsplash': Error -2 [ 25.428342] Launch userspace program failed. [ 25.479541] Freeze processes. [ 25.515118] Stopping fuse filesystems. [ 25.560043] Freezing user space processes ... (elapsed 0.00 seconds) <6>done. [ 25.645919] Stopping normal filesystems. [ 25.692912] Freezing remaining freezable tasks ... (elapsed 0.00 seconds) done. [ 26.229761] Reading kernel & process data... [ 26.598142] 20%...40%...60%...80%...100%...done. [ 27.123970] Waited for i/o due to readahead not ready 39 times. [ 27.194761] Waited for i/o due to synchronous I/O 7 times. [ 27.305047] Atomic restore. [ 27.338402] Doing atomic copy/restore. [ 27.383190] Suspending console(s) (use no_console_suspend to debug) [ 229.068387] Post atomic. [ 229.098613] Reading caches... [ 229.170236] Decompression yielded 25 bytes instead of 4096. [ 229.170238] [ 229.391226] ------------[ cut here ]------------ [ 229.395133] kernel BUG at /sandbox/kernel/power/tuxonice_io.c:689! [ 229.395133] invalid opcode: 0000 [#1] SMP [ 229.395133] Modules linked in: rfkill_input vmnet vmblock vmci vmmon ipv6 speedstep_centrino cpufreq_userspace cpufreq_stats cpufreq_powersave cpufreq_ondemand freq_table cpufreq_conservative w] [ 229.395133] [ 229.395133] Pid: 8167, comm: hibernate Not tainted (2.6.27.2 #12) [ 229.395133] EIP: 0060:[] EFLAGS: 00210282 CPU: 0 [ 229.395133] EIP is at do_rw_loop+0x272/0x2e0 [ 229.395133] EAX: 00000073 EBX: 0001fe72 ECX: ffffffff EDX: 00200046 [ 229.395133] ESI: c0165240 EDI: 0001b070 EBP: c1869e84 ESP: c1869e60 [ 229.395133] DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068 [ 229.395133] Process hibernate (pid: 8167, ti=c1868000 task=f2082580 task.ti=c1868000) [ 229.395133] Stack: c046ad1c 0001b070 0001b06f 000001fe 000001fe c04d1430 c04d0f20 0001fe72 [ 229.395133] ffffbacb c1869eb0 c015b0e4 00004e02 0001fe72 00000002 00004e02 0001b070 [ 229.395133] c04d1430 00000000 f20de824 c04f6824 c1869ebc c015b1bc 00000000 c1869ed0 [ 229.395133] Call Trace: [ 229.395133] [] ? read_pageset+0x134/0x1e0 [ 229.395133] [] ? read_pageset2+0x2c/0x50 [ 229.395133] [] ? copyback_post+0x71/0xa0 [ 229.395133] [] ? do_toi_step+0x5ac/0x790 [ 229.395133] [] ? _toi_try_hibernate+0xce/0x120 [ 229.395133] [] ? toi_main_wrapper+0xf/0x20 [ 229.395133] [] ? toi_attr_store+0x59/0x1f0 [ 229.395133] [] ? sysfs_write_file+0xbb/0x100 [ 229.395133] [] ? vfs_write+0x9b/0x170 [ 229.395133] [] ? fput+0x8/0x30 [ 229.395133] [] ? sysfs_write_file+0x0/0x100 [ 229.395133] [] ? sys_write+0x42/0x70 [ 229.395133] [] ? syscall_call+0x7/0xb [ 229.395133] [] ? quirk_vt8235_acpi+0x70/0xa0 [ 229.395133] ======================= [ 229.395133] Code: 00 8b 15 88 1b 5e c0 83 c2 01 39 d0 0f 84 70 ff ff ff a1 b4 74 5b c0 89 7c 24 04 c7 04 24 1c ad 46 c0 89 44 24 08 e8 9e 4b fd ff <0f> 0b eb fe c7 04 24 f8 07 46 c0 e8 8e 4b fd [ 229.395133] EIP: [] do_rw_loop+0x272/0x2e0 SS:ESP 0068:c1869e60 [ 232.558573] tg3: eth0: Link is up at 1000 Mbps, full duplex. [ 232.626245] tg3: eth0: Flow control is on for TX and on for RX. [ 232.697040] ---[ end trace 014b421dcdd8104c ]--- For info, the Tuxonice patch was created from the tuxonice git tree at kernel.ubuntu.com, reset to the changeset that performed the merge with the intrepid tag (2.6.27-7.15): git reset --hard ac8ccf1c012b9ceb953cd942fdca0ebd0b346622 Also, interestingly, the TuxOnIce signature is tainted: 16:58 pierre at ubuntu ~% od -N20 -c /dev/sda6 0000000 355 303 002 351 n I c e \n \0 \0 \0 \0 \0 \0 \0 0000020 \b \0 \0 \0 0000024 Shouldn't the string "TuxOnIce" be replaced by "HaveImage"? Am I missing something? Thanks, -- Pierre-Alexandre Meyer From ncunningham at crca.org.au Fri Dec 5 01:08:18 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Fri, 05 Dec 2008 12:08:18 +1100 Subject: [TuxOnIce-users] Dual images In-Reply-To: <20081205010209.GQ17410@panda> References: <20081203210812.GI17410@panda> <1228341419.8648.29.camel@nigel-laptop> <20081203232215.GJ17410@panda> <1228347031.8648.39.camel@nigel-laptop> <20081205010209.GQ17410@panda> Message-ID: <1228439298.13883.1.camel@nigel-laptop> Hi Pierre. On Fri, 2008-12-05 at 02:02 +0100, Pierre-Alexandre Meyer wrote: > On Thursday 04 December 2008 ? 10:30:31AM, Nigel Cunningham wrote: > > I don't mean a swapfile, but an ordinary file, using the file allocator > > rather than the swap allocator. In this case, swap is not in the picture > > at all - the file is only used for hibernation. > > I managed to hibernate using a regular file on a ext2 filesystem - the file > allocator seems to work on my system - but I have issues when I try to use > a raw device (specifying a block device): > > This is what I tried: > > echo "TuxOnIce" > /dev/sda6 > dd if=/dev/zero bs=1M count=950 >> /dev/sda6 > > I can double check: > > 16:58 pierre at ubuntu ~% od -N20 -c /dev/sda6 > 0000000 T u x O n I c e \n \0 \0 \0 \0 \0 \0 \0 > 0000020 \0 \0 \0 \0 > 0000024 > > But if I try to hibernate by: > > echo /dev/sda6 > /sys/power/tuxonice/file/target > > and specifying resume=file:/dev/sda6 (returned by > /sys/power/tuxonice/resume), I trigger a BUG() on resume (no error when hibernating): > > [ 25.267903] TuxOnIce: Found an hibernate image. > [ 25.327129] Failed to launch userspace program '/usr/local/sbin/tuxonice_fbsplash': Error -2 > [ 25.428342] Launch userspace program failed. > [ 25.479541] Freeze processes. > [ 25.515118] Stopping fuse filesystems. > [ 25.560043] Freezing user space processes ... (elapsed 0.00 seconds) <6>done. > [ 25.645919] Stopping normal filesystems. > [ 25.692912] Freezing remaining freezable tasks ... (elapsed 0.00 seconds) done. > [ 26.229761] Reading kernel & process data... > [ 26.598142] 20%...40%...60%...80%...100%...done. > [ 27.123970] Waited for i/o due to readahead not ready 39 times. > [ 27.194761] Waited for i/o due to synchronous I/O 7 times. > [ 27.305047] Atomic restore. > [ 27.338402] Doing atomic copy/restore. > [ 27.383190] Suspending console(s) (use no_console_suspend to debug) > [ 229.068387] Post atomic. > [ 229.098613] Reading caches... > [ 229.170236] Decompression yielded 25 bytes instead of 4096. > [ 229.170238] > [ 229.391226] ------------[ cut here ]------------ > [ 229.395133] kernel BUG at /sandbox/kernel/power/tuxonice_io.c:689! > [ 229.395133] invalid opcode: 0000 [#1] SMP > [ 229.395133] Modules linked in: rfkill_input vmnet vmblock vmci vmmon ipv6 speedstep_centrino cpufreq_userspace cpufreq_stats cpufreq_powersave cpufreq_ondemand freq_table cpufreq_conservative w] > [ 229.395133] > [ 229.395133] Pid: 8167, comm: hibernate Not tainted (2.6.27.2 #12) > [ 229.395133] EIP: 0060:[] EFLAGS: 00210282 CPU: 0 > [ 229.395133] EIP is at do_rw_loop+0x272/0x2e0 > [ 229.395133] EAX: 00000073 EBX: 0001fe72 ECX: ffffffff EDX: 00200046 > [ 229.395133] ESI: c0165240 EDI: 0001b070 EBP: c1869e84 ESP: c1869e60 > [ 229.395133] DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068 > [ 229.395133] Process hibernate (pid: 8167, ti=c1868000 task=f2082580 task.ti=c1868000) > [ 229.395133] Stack: c046ad1c 0001b070 0001b06f 000001fe 000001fe c04d1430 c04d0f20 0001fe72 > [ 229.395133] ffffbacb c1869eb0 c015b0e4 00004e02 0001fe72 00000002 00004e02 0001b070 > [ 229.395133] c04d1430 00000000 f20de824 c04f6824 c1869ebc c015b1bc 00000000 c1869ed0 > [ 229.395133] Call Trace: > [ 229.395133] [] ? read_pageset+0x134/0x1e0 > [ 229.395133] [] ? read_pageset2+0x2c/0x50 > [ 229.395133] [] ? copyback_post+0x71/0xa0 > [ 229.395133] [] ? do_toi_step+0x5ac/0x790 > [ 229.395133] [] ? _toi_try_hibernate+0xce/0x120 > [ 229.395133] [] ? toi_main_wrapper+0xf/0x20 > [ 229.395133] [] ? toi_attr_store+0x59/0x1f0 > [ 229.395133] [] ? sysfs_write_file+0xbb/0x100 > [ 229.395133] [] ? vfs_write+0x9b/0x170 > [ 229.395133] [] ? fput+0x8/0x30 > [ 229.395133] [] ? sysfs_write_file+0x0/0x100 > [ 229.395133] [] ? sys_write+0x42/0x70 > [ 229.395133] [] ? syscall_call+0x7/0xb > [ 229.395133] [] ? quirk_vt8235_acpi+0x70/0xa0 > [ 229.395133] ======================= > [ 229.395133] Code: 00 8b 15 88 1b 5e c0 83 c2 01 39 d0 0f 84 70 ff ff ff a1 b4 74 5b c0 89 7c 24 04 c7 04 24 1c ad 46 c0 89 44 24 08 e8 9e 4b fd ff <0f> 0b eb fe c7 04 24 f8 07 46 c0 e8 8e 4b fd > [ 229.395133] EIP: [] do_rw_loop+0x272/0x2e0 SS:ESP 0068:c1869e60 > [ 232.558573] tg3: eth0: Link is up at 1000 Mbps, full duplex. > [ 232.626245] tg3: eth0: Flow control is on for TX and on for RX. > [ 232.697040] ---[ end trace 014b421dcdd8104c ]--- > > For info, the Tuxonice patch was created from the tuxonice git tree at kernel.ubuntu.com, > reset to the changeset that performed the merge with the intrepid tag (2.6.27-7.15): > > git reset --hard ac8ccf1c012b9ceb953cd942fdca0ebd0b346622 > > Also, interestingly, the TuxOnIce signature is tainted: > > 16:58 pierre at ubuntu ~% od -N20 -c /dev/sda6 > 0000000 355 303 002 351 n I c e \n \0 \0 \0 \0 \0 \0 \0 > 0000020 \b \0 \0 \0 > 0000024 > > Shouldn't the string "TuxOnIce" be replaced by "HaveImage"? > > Am I missing something? No, it looks like I have a bug I need to look into. I'll try to get on to it as soon as I can, but that might be a few days. Nigel From pierre at mouraf.org Fri Dec 5 01:11:53 2008 From: pierre at mouraf.org (Pierre-Alexandre Meyer) Date: Fri, 5 Dec 2008 02:11:53 +0100 Subject: [TuxOnIce-users] Dual images In-Reply-To: <1228439298.13883.1.camel@nigel-laptop> References: <20081203210812.GI17410@panda> <1228341419.8648.29.camel@nigel-laptop> <20081203232215.GJ17410@panda> <1228347031.8648.39.camel@nigel-laptop> <20081205010209.GQ17410@panda> <1228439298.13883.1.camel@nigel-laptop> Message-ID: <20081205011153.GS17410@panda> On Friday 05 December 2008 ? 12:08:18PM, Nigel Cunningham wrote: > No, it looks like I have a bug I need to look into. > > I'll try to get on to it as soon as I can, but that might be a few days. No worries. Let me know how I can help. I will try to see if the bug occurs using the latest git tree on kernel.org Thanks, -- Pierre-Alexandre Meyer From cheoppy at gmail.com Fri Dec 5 22:51:35 2008 From: cheoppy at gmail.com (=?UTF-8?B?R2VyZ2VseSBDc8OpcMOhbnk=?=) Date: Fri, 05 Dec 2008 23:51:35 +0100 Subject: [TuxOnIce-users] Resume speed from S3 Message-ID: <4939B077.2040500@gmail.com> Hi! I was curious and did some tests targeting my laptop's resume speed from the S3 state (suspend-to-ram). The tested systems were: - Ubuntu Hardy with stock kernel - Ubuntu Hardy with Tuxonice patched kernel - Ubuntu Intrepid with stock kernel - Windows Vista SP1. With the Tuxoniced kernel I also did a hybrid-suspend test (which means echoing 3 to powerdown_method prior to hibernating). I found that it takes the Tuxoniced kernel ~1-2 seconds more time to come alive from the simple suspend then the stock kernel does. If I do a hybrid-suspend, then the difference is up to ~4 seconds. I also noticed that when resuming from a hybrid-suspend, Tuxonice "reads caches" (according to the user-ui), but I cannot understand why, since it's only suspended to ram, so there shouldn't be any caches removed/freed from ram. The Vista still wins by far, it resumes twice as fast as the fastest Ubuntu kernel. I measured the time from pressing the power button to the time when I was able to enter text to an opened text-editor (this way I can be sure that my usb keyboard is waked up and works). The measured data: Hardy: hybrid-suspend (with Tuxoniced kernel obviously :) ): 13,7 13,2 13,5 avg=13,47 s simple suspend with Tuxoniced kernel: 9,4 10,4 10,4 avg=10,07 s simple suspend with stock kernel: 8,6 9,2 8,6 avg=8,8 s Intrepid (only simple suspend with stock kernel): 7,7 8,1 8,3 avg=8,03 s Vista: 4 4,6 3,9 avg=4,17 s My questions are: - What's that "Reading caches" state when resuming from a hybrid-suspend? - Why a simple suspend's resume takes 1-2s longer with a Tuxoniced kernel? - Is there any way to shorten this time? I know it doesn't take too long to resume, but it would be nice to just open my laptop's lid, press the power button and start working immediately after I sit down :) (I know it depends on the hardware also, but if Vista can do it in 4s, then Linux could do it as well.) Thanks in advance, Gergely Cs?p?ny From ncunningham at crca.org.au Sat Dec 6 05:09:43 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Sat, 06 Dec 2008 16:09:43 +1100 Subject: [TuxOnIce-users] Resume speed from S3 In-Reply-To: <4939B077.2040500@gmail.com> References: <4939B077.2040500@gmail.com> Message-ID: <1228540183.10689.8.camel@nigel-laptop> Hi! On Fri, 2008-12-05 at 23:51 +0100, Gergely Cs?p?ny wrote: > Hi! > > I was curious and did some tests targeting my laptop's resume speed from > the S3 state (suspend-to-ram). The tested systems were: > - Ubuntu Hardy with stock kernel > - Ubuntu Hardy with Tuxonice patched kernel > - Ubuntu Intrepid with stock kernel > - Windows Vista SP1. > > With the Tuxoniced kernel I also did a hybrid-suspend test (which means > echoing 3 to powerdown_method prior to hibernating). > > I found that it takes the Tuxoniced kernel ~1-2 seconds more time to > come alive from the simple suspend then the stock kernel does. If I do a > hybrid-suspend, then the difference is up to ~4 seconds. I also noticed > that when resuming from a hybrid-suspend, Tuxonice "reads caches" > (according to the user-ui), but I cannot understand why, since it's only > suspended to ram, so there shouldn't be any caches removed/freed from ram. > The Vista still wins by far, it resumes twice as fast as the fastest > Ubuntu kernel. > > I measured the time from pressing the power button to the time when I > was able to enter text to an opened text-editor (this way I can be sure > that my usb keyboard is waked up and works). > The measured data: > > Hardy: > hybrid-suspend (with Tuxoniced kernel obviously :) ): > 13,7 > 13,2 > 13,5 > avg=13,47 s > simple suspend with Tuxoniced kernel: > 9,4 > 10,4 > 10,4 > avg=10,07 s > simple suspend with stock kernel: > 8,6 > 9,2 > 8,6 > avg=8,8 s > > Intrepid (only simple suspend with stock kernel): > 7,7 > 8,1 > 8,3 > avg=8,03 s > > Vista: > 4 > 4,6 > 3,9 > avg=4,17 s > > > My questions are: > - What's that "Reading caches" state when resuming from a hybrid-suspend? > - Why a simple suspend's resume takes 1-2s longer with a Tuxoniced kernel? > - Is there any way to shorten this time? > > I know it doesn't take too long to resume, but it would be nice to just > open my laptop's lid, press the power button and start working > immediately after I sit down :) > (I know it depends on the hardware also, but if Vista can do it in 4s, > then Linux could do it as well.) I'll take a look, but off the top of my head, I would think the difference is probably due to the fuse filesystem freezing support we add in the TuxOnIce patch. Regarding "Reading caches", that's because when we're saving the hibernation image, we overwrite part of the caches memory with the atomically copied part of the image. When resuming, we need to read the original contents of that memory back in before we can restart processes. This could be done prior to doing the suspend to ram, but it isn't done that way at the moment. Regards, Nigel From kenny at panix.com Mon Dec 8 21:02:26 2008 From: kenny at panix.com (Kenneth Crudup) Date: Mon, 8 Dec 2008 13:02:26 -0800 (PST) Subject: [TuxOnIce-users] [TuxOnIce-devel] NVidia + TuxOnIce hang: solved? In-Reply-To: <1227824356.6482.1.camel@nigel-laptop> References: <1227783453.6160.3.camel@nigel-laptop> <1227823620.6482.0.camel@nigel-laptop> <1227824356.6482.1.camel@nigel-laptop> Message-ID: On Fri, 28 Nov 2008, Nigel Cunningham wrote: > Ah, okay. I also find the driver is a bit hit and miss when I have a > projector connected while hibernating. FWIW, as long as I remember to not connect my external monitor when back home (via the HDMI out to an HDMI-DVI cable to my monitor) until I've sucessfully suspended, the latest TuxOnIce GIT for 2.6.27 (.8) and the latest NVidia driver (180.11) have been suspending and resuming for me with no issues. -Kenny -- Kenneth R. Crudup Sr. SW Engineer, Scott County Consulting, Los Angeles O: 3630 S. Sepulveda Blvd. #138, L.A., CA 90034-6809 (888) 454-8181 From kenny at panix.com Mon Dec 8 21:13:42 2008 From: kenny at panix.com (Kenneth Crudup) Date: Mon, 8 Dec 2008 13:13:42 -0800 (PST) Subject: [TuxOnIce-users] Can I leave network media mounted ro? In-Reply-To: <1228273800.5357.7.camel@nigel-laptop> References: <200812030102.15938.paczesiowa@dw.pl> <1228265296.25380.525.camel@nigel-laptop> <200812030352.57557.paczesiowa@dw.pl> <1228273800.5357.7.camel@nigel-laptop> Message-ID: On Wed, 3 Dec 2008, Nigel Cunningham wrote: > I don't know the details of how NFS etc work. NFS 2 and 3 are meant to be stateless; it was designed with the notion that your server might go away every now and then. Depending on your mount options, you'll get something like "NFS server foobar not responding, still trying" to "NFS server foobar: stale file handle". -Kenny -- Kenneth R. Crudup Sr. SW Engineer, Scott County Consulting, Los Angeles O: 3630 S. Sepulveda Blvd. #138, L.A., CA 90034-6809 (888) 454-8181 From moptop99 at gmail.com Tue Dec 9 04:36:03 2008 From: moptop99 at gmail.com (Matt Price) Date: Mon, 8 Dec 2008 23:36:03 -0500 Subject: [TuxOnIce-users] tuxonice discussion at UDS jaunty Message-ID: hi, thought list members might want to know that ubuntu developers are discussing including tuxonice as a kernel flavour for the next ubuntu release, jaunty jackalope. there's a wiki page here: https://wiki.ubuntu.com/KernelTeam/Specs/FasterBootFromImage but the actual discussion seems to be using a different document as its basis, which is available through gobby (host:gobby.ubuntu.com, port 6522). the main session on this issue takes place on thursday, and i believe you can listen in and edit the minutes as they scroll by. see the schedule at : http://summit.ubuntu.com/uds-jaunty/adal/ . i think you need a launchpad account to look at that last page. nigel, i wonder if you in particular want to join in. i think this would be a big step forward and i'd be very happy to try to help out with it. matt From ncunningham at crca.org.au Tue Dec 9 07:55:46 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Tue, 09 Dec 2008 18:55:46 +1100 Subject: [TuxOnIce-users] tuxonice discussion at UDS jaunty In-Reply-To: References: Message-ID: <1228809346.6681.13.camel@nigel-laptop> Hi Matt. On Mon, 2008-12-08 at 23:36 -0500, Matt Price wrote: > hi, > > thought list members might want to know that ubuntu developers are > discussing including tuxonice as a kernel flavour for the next ubuntu > release, jaunty jackalope. there's a wiki page here: > https://wiki.ubuntu.com/KernelTeam/Specs/FasterBootFromImage > but the actual discussion seems to be using a different document as > its basis, which is available through gobby (host:gobby.ubuntu.com, > port 6522). the main session on this issue takes place on thursday, > and i believe you can listen in and edit the minutes as they scroll > by. see the schedule at : http://summit.ubuntu.com/uds-jaunty/adal/ . > i think you need a launchpad account to look at that last page. > nigel, i wonder if you in particular want to join in. > > i think this would be a big step forward and i'd be very happy to try > to help out with it. > > matt What timezone as the meeting times in? Regards, Nigel From matt.price at utoronto.ca Tue Dec 9 16:09:25 2008 From: matt.price at utoronto.ca (Matt Price) Date: Tue, 09 Dec 2008 11:09:25 -0500 Subject: [TuxOnIce-users] tuxonice discussion at UDS jaunty In-Reply-To: <1228809346.6681.13.camel@nigel-laptop> References: <1228809346.6681.13.camel@nigel-laptop> Message-ID: <1228838965.6194.415.camel@localhost> On Tue, 2008-12-09 at 18:55 +1100, Nigel Cunningham wrote: > Hi Matt. > > On Mon, 2008-12-08 at 23:36 -0500, Matt Price wrote: > > hi, > > > > thought list members might want to know that ubuntu developers are > > discussing including tuxonice as a kernel flavour for the next ubuntu > > release, jaunty jackalope. there's a wiki page here: > > https://wiki.ubuntu.com/KernelTeam/Specs/FasterBootFromImage > > but the actual discussion seems to be using a different document as > > its basis, which is available through gobby (host:gobby.ubuntu.com, > > port 6522). the main session on this issue takes place on thursday, > > and i believe you can listen in and edit the minutes as they scroll > > by. see the schedule at : http://summit.ubuntu.com/uds-jaunty/adal/ . > > i think you need a launchpad account to look at that last page. > > nigel, i wonder if you in particular want to join in. > > > > i think this would be a big step forward and i'd be very happy to try > > to help out with it. > > > > matt > > What timezone as the meeting times in? > hi nigel, the meeting is in california, so that should be GMT +8, or GMT -8, i forget which way the notation works. as i send this message it is 8:08 AM in california... looks cool, doesn't it? did you check out the gobby session? matt > Regards, > > Nigel > -- Matt Price matt.price at utoronto.ca should be GMT -8 (or +8? i forget From ncunningham at crca.org.au Tue Dec 9 21:26:15 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Wed, 10 Dec 2008 08:26:15 +1100 Subject: [TuxOnIce-users] tuxonice discussion at UDS jaunty In-Reply-To: <1228838965.6194.415.camel@localhost> References: <1228809346.6681.13.camel@nigel-laptop> <1228838965.6194.415.camel@localhost> Message-ID: <1228857975.6681.33.camel@nigel-laptop> Hi Matt. On Tue, 2008-12-09 at 11:09 -0500, Matt Price wrote: > On Tue, 2008-12-09 at 18:55 +1100, Nigel Cunningham wrote: > > Hi Matt. > > > > On Mon, 2008-12-08 at 23:36 -0500, Matt Price wrote: > > > hi, > > > > > > thought list members might want to know that ubuntu developers are > > > discussing including tuxonice as a kernel flavour for the next ubuntu > > > release, jaunty jackalope. there's a wiki page here: > > > https://wiki.ubuntu.com/KernelTeam/Specs/FasterBootFromImage > > > but the actual discussion seems to be using a different document as > > > its basis, which is available through gobby (host:gobby.ubuntu.com, > > > port 6522). the main session on this issue takes place on thursday, > > > and i believe you can listen in and edit the minutes as they scroll > > > by. see the schedule at : http://summit.ubuntu.com/uds-jaunty/adal/ . > > > i think you need a launchpad account to look at that last page. > > > nigel, i wonder if you in particular want to join in. > > > > > > i think this would be a big step forward and i'd be very happy to try > > > to help out with it. > > > > > > matt > > > > What timezone as the meeting times in? > > > > hi nigel, > > the meeting is in california, so that should be GMT +8, or GMT -8, i > forget which way the notation works. as i send this message it is 8:08 > AM in california... Okay. You sent the message at 3am my time, and the meeting is 9am? So I'd have to get up at 4am Friday (I'm a day ahead). I've got a pretty full day on Friday already, so I don't think I'll be able to promise to do that. I can however give you some key points, and I'm doing some work to make merging TuxOnIce more of a possibility. > looks cool, doesn't it? did you check out the gobby session? Looks promising, but I've seen things look promising before. Regards, Nigel From oopla at users.sf.net Tue Dec 9 21:36:38 2008 From: oopla at users.sf.net (Paolo) Date: Tue, 9 Dec 2008 22:36:38 +0100 Subject: [TuxOnIce-users] tuxonice discussion at UDS jaunty In-Reply-To: <1228838965.6194.415.camel@localhost> References: <1228809346.6681.13.camel@nigel-laptop> <1228838965.6194.415.camel@localhost> Message-ID: <20081209213637.GA26128@localhost> On Tue, Dec 09, 2008 at 11:09:25AM -0500, Matt Price wrote: > > the meeting is in california, so that should be GMT +8, or GMT -8, i GMT +x -> x hours eastward from Greenwich (UK) - because they see the sun x hours earlier than in Greenwich (UK) GMT -x -> x hours westward from Greenwich (UK) - because they see the sun x hours later > forget which way the notation works. as i send this message it is 8:08 > AM in california... so that's be 8:08 -0800, which is consistent with 11:09:25AM -0500, time of the node in Toronto area your msg came from, which is +3 hours eastward than your location. I'm sure I'll screw on that myself next time, no matter what :P -- paolo From ncunningham at crca.org.au Tue Dec 9 22:15:09 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Wed, 10 Dec 2008 09:15:09 +1100 Subject: [TuxOnIce-users] Plans for rc8. Message-ID: <1228860909.6681.63.camel@nigel-laptop> Hi all. I hope to release rc8 within a week or so. Prior to that, I'm going to start some new branches in the git trees, and separate out additional/standalone changes to these branches. This means that rc8 and later will again have multiple patches, coming in a bzipped tarball with the apply script we used to use. With all the patches applied, there will be little difference to what is currently in git. The branches are: - core: Core TuxOnIce files only (I expect this to be 97% new files, and hopefully something can Ubuntu can use for Jaunty, together with lzf if lzo can't be made reliable). - modules: Additional changes for building TuxOnIce as loadable kernel modules. - freezer enhancements: Filesystem freezing support etc. - lzf: LZF cryptoapi module. I'd still like to use LZO (which is already merged), but haven't gotten it to work reliably under SMP yet (haven't tried that much either). - Load average save/restore In addition to these changes, I'll look into replacing the dynamically allocated pageflags code with the swsusp bitmap code Rafael has implemented. Regards, Nigel From xavier.gnata at gmail.com Tue Dec 9 23:39:12 2008 From: xavier.gnata at gmail.com (Xavier Gnata) Date: Wed, 10 Dec 2008 00:39:12 +0100 Subject: [TuxOnIce-users] [TuxOnIce-devel] Plans for rc8. In-Reply-To: <1228860909.6681.63.camel@nigel-laptop> References: <1228860909.6681.63.camel@nigel-laptop> Message-ID: <493F01A0.4050506@gmail.com> Hi Nigel, tuxonice by default in Januty? It would be great! Xavier > Hi all. > > I hope to release rc8 within a week or so. Prior to that, I'm going to > start some new branches in the git trees, and separate out > additional/standalone changes to these branches. This means that rc8 and > later will again have multiple patches, coming in a bzipped tarball with > the apply script we used to use. With all the patches applied, there > will be little difference to what is currently in git. > > The branches are: > - core: Core TuxOnIce files only (I expect this to be 97% new files, and > hopefully something can Ubuntu can use for Jaunty, together with lzf if > lzo can't be made reliable). > - modules: Additional changes for building TuxOnIce as loadable kernel > modules. > - freezer enhancements: Filesystem freezing support etc. > - lzf: LZF cryptoapi module. I'd still like to use LZO (which is already > merged), but haven't gotten it to work reliably under SMP yet (haven't > tried that much either). > - Load average save/restore > > In addition to these changes, I'll look into replacing the dynamically > allocated pageflags code with the swsusp bitmap code Rafael has > implemented. > > Regards, > > Nigel > > _______________________________________________ > TuxOnIce-devel mailing list > TuxOnIce-devel at lists.tuxonice.net > http://lists.tuxonice.net/mailman/listinfo/tuxonice-devel From pierre at vmware.com Wed Dec 10 00:33:15 2008 From: pierre at vmware.com (Pierre-Alexandre Meyer) Date: Tue, 9 Dec 2008 16:33:15 -0800 Subject: [TuxOnIce-users] [TuxOnIce-devel] Plans for rc8. In-Reply-To: <493F01A0.4050506@gmail.com> References: <1228860909.6681.63.camel@nigel-laptop> <493F01A0.4050506@gmail.com> Message-ID: <20081210003315.GC26449@vmware.com> On Tuesday 09 December 2008 ? 03:39:12PM, Xavier Gnata wrote: > tuxonice by default in Januty? It would be great! Did you talk to the Ubuntu kernel team Nigel? During the meeting yesterday, they were saying that they may only be interested in multithreaded_io/compression support. They don't want to maintain the entire core. -- Pierre-Alexandre Meyer PIT - OSC - CVP Host OS pierre at vmware.com From ncunningham at crca.org.au Wed Dec 10 02:52:59 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Wed, 10 Dec 2008 13:52:59 +1100 Subject: [TuxOnIce-users] [TuxOnIce-devel] Plans for rc8. In-Reply-To: <493F01A0.4050506@gmail.com> References: <1228860909.6681.63.camel@nigel-laptop> <493F01A0.4050506@gmail.com> Message-ID: <1228877579.7268.6.camel@nigel-laptop> Hi. On Wed, 2008-12-10 at 00:39 +0100, Xavier Gnata wrote: > Hi Nigel, > > tuxonice by default in Januty? It would be great! It would be, but I don't think that's what's going to happen. Apparently they're having a meeting in a day or two, and will discuss it at that. I'm just seeking to make it as easy as possible for TuxOnIce to be merged into vanilla or any distro. Regards, Nigel From ncunningham at crca.org.au Wed Dec 10 02:57:13 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Wed, 10 Dec 2008 13:57:13 +1100 Subject: [TuxOnIce-users] [TuxOnIce-devel] Plans for rc8. In-Reply-To: <20081210003315.GC26449@vmware.com> References: <1228860909.6681.63.camel@nigel-laptop> <493F01A0.4050506@gmail.com> <20081210003315.GC26449@vmware.com> Message-ID: <1228877833.7268.11.camel@nigel-laptop> Hi. On Tue, 2008-12-09 at 16:33 -0800, Pierre-Alexandre Meyer wrote: > On Tuesday 09 December 2008 ? 03:39:12PM, Xavier Gnata wrote: > > tuxonice by default in Januty? It would be great! > > Did you talk to the Ubuntu kernel team Nigel? During the meeting > yesterday, they were saying that they may only be interested in > multithreaded_io/compression support. They don't want to maintain the > entire core. They don't know what they're talking about then. If they want multithreaded I/O and compression, they may as well take the whole lot because it's all in completely new files that will be easiest to maintain if they don't go reworking it all. Once I separate out the fuse support and build-as-modules support, I hope to have a diff stat that looks something along the lines of 46 new files + ~10 files (at a guess) with minimal changes (mostly in kernel/power). The other thing to mention is that in a year or so of adding TuxOnIce to their kernels, I've only rarely had rejects - it's not going to be nearly as much of a pain to add as they seem to be thinking. Regards, Nigel From pierre at vmware.com Wed Dec 10 03:29:29 2008 From: pierre at vmware.com (Pierre-Alexandre Meyer) Date: Tue, 9 Dec 2008 19:29:29 -0800 Subject: [TuxOnIce-users] [TuxOnIce-devel] Plans for rc8. In-Reply-To: <1228877833.7268.11.camel@nigel-laptop> References: <1228860909.6681.63.camel@nigel-laptop> <493F01A0.4050506@gmail.com> <20081210003315.GC26449@vmware.com> <1228877833.7268.11.camel@nigel-laptop> Message-ID: <20081210032929.GA19476@vmware.com> On Tuesday 09 December 2008 ? 06:57:13PM, Nigel Cunningham wrote: > They don't know what they're talking about then. If they want > multithreaded I/O and compression, they may as well take the whole lot > because it's all in completely new files that will be easiest to > maintain if they don't go reworking it all. Once I separate out the fuse > support and build-as-modules support, I hope to have a diff stat that > looks something along the lines of 46 new files + ~10 files (at a guess) > with minimal changes (mostly in kernel/power). This is what I told them. The discussion was about boot time improvement. The only minimal feature I can see they could integrate is keep_image mode (not the patch itself but to make the idea work with swsusp) and make Ubuntu resume from a pristine image. I reduced Intrepid boot time to 20sec with this method (using TuxOnIce as a late init call). The real work then is not in the kernel but rather in the user space level (mounting the rootfs ro to take the image, retaking snapshots after each apt-get install, ...). Also, regarding multithreaded i/o, my understanding is that it is not proved that it is actually faster in all situations. Am I wrong? > The other thing to mention is that in a year or so of adding TuxOnIce to > their kernels, I've only rarely had rejects - it's not going to be > nearly as much of a pain to add as they seem to be thinking. They don't want to maintain some patches (even if it is just adding files into kernel/power) which are not upstreamed: TuxOnIce * Resumes from disk a lot faster due to LZF compression * Nigel maintains a TuxOnIce kernel for all Ubuntu flavours (550Kb patch) o Git tree on kernel.ubuntu.com * Offer a .deb package to users? Drawbacks * Work is not going upstream. EVER. o Upstream working at a kexec-based solution as successor to in-kernel suspend From ncunningham at crca.org.au Wed Dec 10 03:36:11 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Wed, 10 Dec 2008 14:36:11 +1100 Subject: [TuxOnIce-users] [TuxOnIce-devel] Plans for rc8. In-Reply-To: <20081210032929.GA19476@vmware.com> References: <1228860909.6681.63.camel@nigel-laptop> <493F01A0.4050506@gmail.com> <20081210003315.GC26449@vmware.com> <1228877833.7268.11.camel@nigel-laptop> <20081210032929.GA19476@vmware.com> Message-ID: <1228880171.7268.21.camel@nigel-laptop> Hi. On Tue, 2008-12-09 at 19:29 -0800, Pierre-Alexandre Meyer wrote: > On Tuesday 09 December 2008 ? 06:57:13PM, Nigel Cunningham wrote: > > They don't know what they're talking about then. If they want > > multithreaded I/O and compression, they may as well take the whole lot > > because it's all in completely new files that will be easiest to > > maintain if they don't go reworking it all. Once I separate out the fuse > > support and build-as-modules support, I hope to have a diff stat that > > looks something along the lines of 46 new files + ~10 files (at a guess) > > with minimal changes (mostly in kernel/power). > > This is what I told them. > > The discussion was about boot time improvement. The only minimal feature I > can see they could integrate is keep_image mode (not the patch itself > but to make the idea work with swsusp) and make Ubuntu resume > from a pristine image. I reduced Intrepid boot time to 20sec with this > method (using TuxOnIce as a late init call). > > The real work then is not in the kernel but rather in the user space level > (mounting the rootfs ro to take the image, retaking snapshots after each > apt-get install, ...). > > Also, regarding multithreaded i/o, my understanding is that it is not > proved that it is actually faster in all situations. Am I wrong? That's right. If the bottleneck is compressing, then multithreaded will be faster. If it's the disk (as it often is), then it won't help. We go as fast as the slowest component. > > The other thing to mention is that in a year or so of adding TuxOnIce to > > their kernels, I've only rarely had rejects - it's not going to be > > nearly as much of a pain to add as they seem to be thinking. > > They don't want to maintain some patches (even if it is just adding > files into kernel/power) which are not upstreamed: > > TuxOnIce > > * Resumes from disk a lot faster due to LZF compression It's not just LZF though. I think the fact that TuxOnIce doesn't batch things like swsusp also helps. > * Nigel maintains a TuxOnIce kernel for all Ubuntu flavours > (550Kb patch) > o Git tree on kernel.ubuntu.com > * Offer a .deb package to users? > > Drawbacks > > * Work is not going upstream. EVER. That's not proven. The reason it hasn't gone upstream so far is that I haven't really pushed for it to go upstream. Pavel has resisted, but given that they've merged other stuff, there's no reason it shouldn't. > o Upstream working at a kexec-based solution as successor > to in-kernel suspend Intel (IIRC) is working on kexec, not 'upstream'. Pavel just acks patches, Rafael is focused on lowerlevel stuff more than swsusp and certainly not on kexec. The other thing to mention here is that I'm starting to think about using kexec. > >From https://wiki.ubuntu.com/KernelTeam/Specs/FasterBootFromImage > > About this last point, what will be the impact of the kexec work on > TuxOnIce? I'm starting to think about using kexec. At this stage, I image that the biggest issue would be passing data between the two kernels (I still want to use storage that's accessible from the kernel being hibernated, and so would need to pass major/minor numbers and lists of blocks to read/write. Assuming that the list of pages used in the kernel being hibernated is already handled by kexec, I shouldn't have a huge job in front of me. Nigel From ml at lalamuhkuh.de Wed Dec 10 13:55:14 2008 From: ml at lalamuhkuh.de (Roland Tapken) Date: Wed, 10 Dec 2008 14:55:14 +0100 Subject: [TuxOnIce-users] fs corruption through filewriter? Message-ID: <200812101455.14723.ml@lalamuhkuh.de> Hi, I'm using tuxonice on conjunction with filewriter. But from time to time the (regular) boot process stops because of a filesystem corruption. An ext3 system can be repaired using a simple fsck, but another system using reiferfs requires reiserfsck with '--rebuild-tree'-option. I'm experiences this only when I hibernated before so I assume this problem is caused by filewriter. I'm using my own hibernation scripts which integrates into pm-utils: http://www.blogs.uni-osnabrueck.de/rotapken/2008/11/14/integrate-tuxonice- into-ubuntus-hibernation-process/ The main parts of this scripts (/etc/pm/sleep.d/01tuxonice) are the generation of the hibernation file (if neccessary): dd if=/dev/zero bs="1k" count="${mem}" of="${hibernation}" The initializion of this file: echo "TuxOnIce" | dd conv=notrunc of="${hibernation}" sync And the storing of the target information into a file on the root partition: echo "Writing resume info..." echo 1 > /sys/power/tuxonice/replace_swsusp echo > /sys/power/tuxonice/file/target echo "${hibernation}" > /sys/power/tuxonice/file/target cat /sys/power/tuxonice/resume > "${RESUME_INFO}" The resuming is done in initramfs: mount_root # Mounts the root fs READ-ONLY using "mount -r" resume="$(cat ${rootmnt}/${RESUME_INFO} 2> /dev/null)" umount_root echo "${resume}" > /sys/power/tuxonice/resume echo > /sys/power/tuxonice/do_resume Might this problem be related to the use of "mount -r"? Or could this be a serious bug in TuxOnIce' filewriter subsystem? Kind regards, Roland Tapken -- Privacy is not a crime! From kenny at panix.com Wed Dec 10 20:35:02 2008 From: kenny at panix.com (Kenneth Crudup) Date: Wed, 10 Dec 2008 12:35:02 -0800 (PST) Subject: [TuxOnIce-users] Did someone here say they were using ReiserFS? Message-ID: I just remember they were saying how fast it is. I was thinking about experimenting with it, since I have a box with a removable drive bay, so that having Reiser around doesn't kill my box (*cough*too*cough*). Seriously, though- is anyone using it with ToI? Which version, any issues? -Kenny -- Kenneth R. Crudup Sr. SW Engineer, Scott County Consulting, Los Angeles O: 3630 S. Sepulveda Blvd. #138, L.A., CA 90034-6809 (888) 454-8181 From kenny at panix.com Wed Dec 10 20:57:19 2008 From: kenny at panix.com (Kenneth Crudup) Date: Wed, 10 Dec 2008 12:57:19 -0800 (PST) Subject: [TuxOnIce-users] [TuxOnIce-devel] Did someone here say they were using ReiserFS? In-Reply-To: <49cc99580812101251v54af275qac2cbd3912be4910@mail.gmail.com> References: <49cc99580812101251v54af275qac2cbd3912be4910@mail.gmail.com> Message-ID: On Wed, 10 Dec 2008, Demeter Tam?s wrote: > I am using reiserfs with tuxonice on two laptops (Fujitsu-Siemens Amilo Pro > v2055 (from 2.6.18 till 2.6.25) and Lenovo Thinkpad R500 (2.6.27)) for every > partition except for the /boot. I never had problems with it. Thanks. The Reiser3 that comes with the kernel, or 4 from external patches? -Kenny -- Kenneth R. Crudup Sr. SW Engineer, Scott County Consulting, Los Angeles O: 3630 S. Sepulveda Blvd. #138, L.A., CA 90034-6809 (888) 454-8181 From Demeter.Tamas at stud.u-szeged.hu Wed Dec 10 20:51:22 2008 From: Demeter.Tamas at stud.u-szeged.hu (=?UTF-8?Q?Demeter_Tam=C3=A1s?=) Date: Wed, 10 Dec 2008 21:51:22 +0100 Subject: [TuxOnIce-users] [TuxOnIce-devel] Did someone here say they were using ReiserFS? In-Reply-To: References: Message-ID: <49cc99580812101251v54af275qac2cbd3912be4910@mail.gmail.com> Hi! I am using reiserfs with tuxonice on two laptops (Fujitsu-Siemens Amilo Pro v2055 (from 2.6.18 till 2.6.25) and Lenovo Thinkpad R500 (2.6.27)) for every partition except for the /boot. I never had problems with it. Best regards, Tamas On Wed, Dec 10, 2008 at 9:35 PM, Kenneth Crudup wrote: > > I just remember they were saying how fast it is. I was thinking about > experimenting with it, since I have a box with a removable drive bay, > so that having Reiser around doesn't kill my box (*cough*too*cough*). > > Seriously, though- is anyone using it with ToI? Which version, any issues? > > -Kenny > > -- > Kenneth R. Crudup Sr. SW Engineer, Scott County Consulting, Los Angeles > O: 3630 S. Sepulveda Blvd. #138, L.A., CA 90034-6809 (888) 454-8181 > _______________________________________________ > TuxOnIce-devel mailing list > TuxOnIce-devel at lists.tuxonice.net > http://lists.tuxonice.net/mailman/listinfo/tuxonice-devel > -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.tuxonice.net/pipermail/tuxonice-users/attachments/20081210/69c11f37/attachment.htm From Demeter.Tamas at stud.u-szeged.hu Wed Dec 10 21:08:33 2008 From: Demeter.Tamas at stud.u-szeged.hu (=?UTF-8?Q?Demeter_Tam=C3=A1s?=) Date: Wed, 10 Dec 2008 22:08:33 +0100 Subject: [TuxOnIce-users] [TuxOnIce-devel] Did someone here say they were using ReiserFS? In-Reply-To: References: <49cc99580812101251v54af275qac2cbd3912be4910@mail.gmail.com> Message-ID: <49cc99580812101308p3202cd5bv698631abafaa7f25@mail.gmail.com> reiser3 On Wed, Dec 10, 2008 at 9:57 PM, Kenneth Crudup wrote: > > On Wed, 10 Dec 2008, Demeter Tam?s wrote: > > > I am using reiserfs with tuxonice on two laptops (Fujitsu-Siemens Amilo > Pro > > v2055 (from 2.6.18 till 2.6.25) and Lenovo Thinkpad R500 (2.6.27)) for > every > > partition except for the /boot. I never had problems with it. > > Thanks. The Reiser3 that comes with the kernel, or 4 from external patches? > > -Kenny > > -- > Kenneth R. Crudup Sr. SW Engineer, Scott County Consulting, Los Angeles > O: 3630 S. Sepulveda Blvd. #138, L.A., CA 90034-6809 (888) 454-8181 > _______________________________________________ > TuxOnIce-devel mailing list > TuxOnIce-devel at lists.tuxonice.net > http://lists.tuxonice.net/mailman/listinfo/tuxonice-devel > -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.tuxonice.net/pipermail/tuxonice-users/attachments/20081210/4c63b7e1/attachment.htm From ncunningham at crca.org.au Wed Dec 10 21:41:59 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Thu, 11 Dec 2008 08:41:59 +1100 Subject: [TuxOnIce-users] fs corruption through filewriter? In-Reply-To: <200812101455.14723.ml@lalamuhkuh.de> References: <200812101455.14723.ml@lalamuhkuh.de> Message-ID: <1228945319.7268.54.camel@nigel-laptop> Hi Roland. On Wed, 2008-12-10 at 14:55 +0100, Roland Tapken wrote: > Hi, > > I'm using tuxonice on conjunction with filewriter. But from time to time the > (regular) boot process stops because of a filesystem corruption. An ext3 system > can be repaired using a simple fsck, but another system using reiferfs > requires reiserfsck with '--rebuild-tree'-option. > > I'm experiences this only when I hibernated before so I assume this problem is > caused by filewriter. > > I'm using my own hibernation scripts which integrates into pm-utils: > http://www.blogs.uni-osnabrueck.de/rotapken/2008/11/14/integrate-tuxonice- > into-ubuntus-hibernation-process/ > > The main parts of this scripts (/etc/pm/sleep.d/01tuxonice) are the generation > of the hibernation file (if neccessary): > > dd if=/dev/zero bs="1k" count="${mem}" of="${hibernation}" > > The initializion of this file: > > echo "TuxOnIce" | dd conv=notrunc of="${hibernation}" > sync > > And the storing of the target information into a file on the root partition: > > echo "Writing resume info..." > echo 1 > /sys/power/tuxonice/replace_swsusp > echo > /sys/power/tuxonice/file/target > echo "${hibernation}" > /sys/power/tuxonice/file/target > cat /sys/power/tuxonice/resume > "${RESUME_INFO}" > > The resuming is done in initramfs: > > mount_root # Mounts the root fs READ-ONLY using "mount -r" > resume="$(cat ${rootmnt}/${RESUME_INFO} 2> /dev/null)" > umount_root > echo "${resume}" > /sys/power/tuxonice/resume > echo > /sys/power/tuxonice/do_resume > > Might this problem be related to the use of "mount -r"? Or could this be a > serious bug in TuxOnIce' filewriter subsystem? Even mount -r replays the journal, iirc. This would result in corruption because the filesystem wouldn't then match the state in the image. It would be slower, but you'd probably want to put ${RESUME_INFO} directly into the initramfs. Regards, Nigel From ml at lalamuhkuh.de Wed Dec 10 21:57:14 2008 From: ml at lalamuhkuh.de (Roland Tapken) Date: Wed, 10 Dec 2008 22:57:14 +0100 Subject: [TuxOnIce-users] fs corruption through filewriter? Message-ID: <200812102257.14619.ml@lalamuhkuh.de> Hi Nigel, > Even mount -r replays the journal, iirc. This would result in corruption > because the filesystem wouldn't then match the state in the image. That would explain it. > It > would be slower, but you'd probably want to put ${RESUME_INFO} directly > into the initramfs. I'm using this info file because I tried to avoid updating the image everytime the hibernation file changed (e.g. because it has been deleted). But if there is really no way to mount a filesystem without changing it's state I have to bite the bullet. Viele Gr??e, Roland Tapken -- Privacy is not a crime! From roland at dau-sicher.de Wed Dec 10 21:54:41 2008 From: roland at dau-sicher.de (Roland Tapken) Date: Wed, 10 Dec 2008 22:54:41 +0100 Subject: [TuxOnIce-users] fs corruption through filewriter? In-Reply-To: <1228945319.7268.54.camel@nigel-laptop> References: <200812101455.14723.ml@lalamuhkuh.de> <1228945319.7268.54.camel@nigel-laptop> Message-ID: <200812102254.41567.roland@dau-sicher.de> Hi Nigel, > Even mount -r replays the journal, iirc. This would result in corruption > because the filesystem wouldn't then match the state in the image. That would explain it. > It > would be slower, but you'd probably want to put ${RESUME_INFO} directly > into the initramfs. I'm using this info file because I tried to avoid updating the image everytime the hibernation file changed (e.g. because it has been deleted). But if there is really no way to mount a filesystem without changing it's state I have to bite the bullet. Viele Gr??e, Roland Tapken -- Privacy is not a crime! From ncunningham at crca.org.au Wed Dec 10 22:03:21 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Thu, 11 Dec 2008 09:03:21 +1100 Subject: [TuxOnIce-users] fs corruption through filewriter? In-Reply-To: <200812102257.14619.ml@lalamuhkuh.de> References: <200812102257.14619.ml@lalamuhkuh.de> Message-ID: <1228946601.7268.69.camel@nigel-laptop> Hi. On Wed, 2008-12-10 at 22:57 +0100, Roland Tapken wrote: > Hi Nigel, > > > Even mount -r replays the journal, iirc. This would result in corruption > > because the filesystem wouldn't then match the state in the image. > > That would explain it. > > > It > > would be slower, but you'd probably want to put ${RESUME_INFO} directly > > into the initramfs. > > I'm using this info file because I tried to avoid updating the image everytime > the hibernation file changed (e.g. because it has been deleted). But if there > is really no way to mount a filesystem without changing it's state I have to > bite the bullet. Why delete the hibernation file (unless you're low on disk space)? If you can leave it the same all the time, then you'll never have to change that file or rebuild the initramfs for that reason. Regards, Nigel From ncunningham at crca.org.au Thu Dec 11 03:41:09 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Thu, 11 Dec 2008 14:41:09 +1100 Subject: [TuxOnIce-users] tuxonice discussion at UDS jaunty In-Reply-To: <1228962827.18010.1041.camel@localhost> References: <1228809346.6681.13.camel@nigel-laptop> <1228838965.6194.415.camel@localhost> <1228857975.6681.33.camel@nigel-laptop> <1228962827.18010.1041.camel@localhost> Message-ID: <1228966869.7268.103.camel@nigel-laptop> Hi. On Wed, 2008-12-10 at 21:33 -0500, Matt Price wrote: > On Wed, 2008-12-10 at 08:26 +1100, Nigel Cunningham wrote: > > Hi Matt. > > > > On Tue, 2008-12-09 at 11:09 -0500, Matt Price wrote: > > > On Tue, 2008-12-09 at 18:55 +1100, Nigel Cunningham wrote: > > > > Hi Matt. > > > > > > > > On Mon, 2008-12-08 at 23:36 -0500, Matt Price wrote: > > > > > hi, > > > > > > > > > > thought list members might want to know that ubuntu developers are > > > > > discussing including tuxonice as a kernel flavour for the next ubuntu > > > > > release, jaunty jackalope. there's a wiki page here: > > > > > https://wiki.ubuntu.com/KernelTeam/Specs/FasterBootFromImage > > > > > but the actual discussion seems to be using a different document as > > > > > its basis, which is available through gobby (host:gobby.ubuntu.com, > > > > > port 6522). the main session on this issue takes place on thursday, > > > > > and i believe you can listen in and edit the minutes as they scroll > > > > > by. see the schedule at : http://summit.ubuntu.com/uds-jaunty/adal/ . > > > > > i think you need a launchpad account to look at that last page. > > > > > nigel, i wonder if you in particular want to join in. > > > > > > > > > > i think this would be a big step forward and i'd be very happy to try > > > > > to help out with it. > > > > > > > > > > matt > > > > > > > > What timezone as the meeting times in? > > > > > > > > > > hi nigel, > > > > > > the meeting is in california, so that should be GMT +8, or GMT -8, i > > > forget which way the notation works. as i send this message it is 8:08 > > > AM in california... > > > > Okay. You sent the message at 3am my time, and the meeting is 9am? So > > I'd have to get up at 4am Friday (I'm a day ahead). I've got a pretty > > full day on Friday already, so I don't think I'll be able to promise to > > do that. I can however give you some key points, and I'm doing some work > > to make merging TuxOnIce more of a possibility. > > > > hey that soundsgood. i will try to attend -- i also am a little > overscheduled that day but i should be able to spend about half the > session online. what are the main points that might need to be raised? > and what can you say about the merge status, etc? Most of the TuxOnIce patch is adding new files. I'm currently working on splitting into separate branches the LZF support, load average save/restore, building as modules, paranoia code (checksumming etc), cluster development work and filesystem freezing. With that done, a diffstat of changes to existing files looks incredibly boring. I've started with 2.6.24, so it has stuff that's not there in head. Removing the stuff I know isn't in head, here's a diffstat: diffstat current-2.6.24-tuxonice.patch MAINTAINERS | 7 + include/linux/freezer.h | 3 include/linux/kernel.h | 2 include/linux/netlink.h | 2 include/linux/suspend.h | 58 +++++++++++++++ include/linux/swap.h | 1 init/do_mounts.c | 2 init/do_mounts_initrd.c | 6 + kernel/power/Kconfig | 176 ++++++++++++++++++++++++++++++++++++++++++++++++ kernel/power/Makefile | 20 +++++ kernel/power/disk.c | 34 +++++++-- kernel/power/power.h | 48 ++++++++++++- kernel/power/process.c | 5 + kernel/power/snapshot.c | 46 ++++++------ lib/vsprintf.c | 23 ++++++ mm/Makefile | 2 mm/memory_hotplug.c | 3 mm/page_alloc.c | 28 +++++++ 18 files changed, 435 insertions(+), 31 deletions(-) (LZF is just 1 new files + additional lines in crypto/Makefile and crypto/Kconfig). Regarding merge status: I'm hoping to do 3.0-rc8 next week and 3.0 shortly afterwards. Once we hit 3.0, I'm planning on focusing exclusively on submitting the code for review and doing suggested cleanups etc. > > > looks cool, doesn't it? did you check out the gobby session? > > > > Looks promising, but I've seen things look promising before. > > > fair enough. anyway i'll tune in tomorrow. thanks, > matt Thanks Matt. I'll see if I can wake up in time, but don't count on me. Nigel From matt.price at utoronto.ca Thu Dec 11 02:33:47 2008 From: matt.price at utoronto.ca (Matt Price) Date: Wed, 10 Dec 2008 21:33:47 -0500 Subject: [TuxOnIce-users] tuxonice discussion at UDS jaunty In-Reply-To: <1228857975.6681.33.camel@nigel-laptop> References: <1228809346.6681.13.camel@nigel-laptop> <1228838965.6194.415.camel@localhost> <1228857975.6681.33.camel@nigel-laptop> Message-ID: <1228962827.18010.1041.camel@localhost> On Wed, 2008-12-10 at 08:26 +1100, Nigel Cunningham wrote: > Hi Matt. > > On Tue, 2008-12-09 at 11:09 -0500, Matt Price wrote: > > On Tue, 2008-12-09 at 18:55 +1100, Nigel Cunningham wrote: > > > Hi Matt. > > > > > > On Mon, 2008-12-08 at 23:36 -0500, Matt Price wrote: > > > > hi, > > > > > > > > thought list members might want to know that ubuntu developers are > > > > discussing including tuxonice as a kernel flavour for the next ubuntu > > > > release, jaunty jackalope. there's a wiki page here: > > > > https://wiki.ubuntu.com/KernelTeam/Specs/FasterBootFromImage > > > > but the actual discussion seems to be using a different document as > > > > its basis, which is available through gobby (host:gobby.ubuntu.com, > > > > port 6522). the main session on this issue takes place on thursday, > > > > and i believe you can listen in and edit the minutes as they scroll > > > > by. see the schedule at : http://summit.ubuntu.com/uds-jaunty/adal/ . > > > > i think you need a launchpad account to look at that last page. > > > > nigel, i wonder if you in particular want to join in. > > > > > > > > i think this would be a big step forward and i'd be very happy to try > > > > to help out with it. > > > > > > > > matt > > > > > > What timezone as the meeting times in? > > > > > > > hi nigel, > > > > the meeting is in california, so that should be GMT +8, or GMT -8, i > > forget which way the notation works. as i send this message it is 8:08 > > AM in california... > > Okay. You sent the message at 3am my time, and the meeting is 9am? So > I'd have to get up at 4am Friday (I'm a day ahead). I've got a pretty > full day on Friday already, so I don't think I'll be able to promise to > do that. I can however give you some key points, and I'm doing some work > to make merging TuxOnIce more of a possibility. > hey that soundsgood. i will try to attend -- i also am a little overscheduled that day but i should be able to spend about half the session online. what are the main points that might need to be raised? and what can you say about the merge status, etc? > > looks cool, doesn't it? did you check out the gobby session? > > Looks promising, but I've seen things look promising before. > fair enough. anyway i'll tune in tomorrow. thanks, matt > Regards, > > Nigel > -- Matt Price matt.price at utoronto.ca From paczesiowa at gmail.com Thu Dec 11 08:01:10 2008 From: paczesiowa at gmail.com (Bartek =?utf-8?q?=C4=86wik=C5=82owski?=) Date: Thu, 11 Dec 2008 09:01:10 +0100 Subject: [TuxOnIce-users] Did someone here say they were using ReiserFS? In-Reply-To: References: Message-ID: <200812110901.10487.paczesiowa@dw.pl> > I just remember they were saying how fast it is. I was thinking about > experimenting with it, since I have a box with a removable drive bay, > so that having Reiser around doesn't kill my box (*cough*too*cough*). > > Seriously, though- is anyone using it with ToI? Which version, any issues? > > -Kenny I'm using reiserfs on my desktop and laptop machines for both root and home partitions and I don't have any problems with tuxonice (at least fs related problems). I've seen my share of kernel panics related to reiserfs code though, so one of these days I'll go back to extN. And ext4 is going stable, since it's faster than reiserfs it's probably a better choice. Regards, Bartek Cwiklowski From ncunningham at crca.org.au Thu Dec 11 11:07:05 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Thu, 11 Dec 2008 22:07:05 +1100 Subject: [TuxOnIce-users] tuxonice discussion at UDS jaunty In-Reply-To: <1228966869.7268.103.camel@nigel-laptop> References: <1228809346.6681.13.camel@nigel-laptop> <1228838965.6194.415.camel@localhost> <1228857975.6681.33.camel@nigel-laptop> <1228962827.18010.1041.camel@localhost> <1228966869.7268.103.camel@nigel-laptop> Message-ID: <1228993626.7268.106.camel@nigel-laptop> Hi again. I'm going to leave my laptop on, logged in to Gobby. Hopefully I'll get a log, even if I don't wake up. Nigel From huffcslists at gmail.com Thu Dec 11 14:04:15 2008 From: huffcslists at gmail.com (Craig Huff) Date: Thu, 11 Dec 2008 08:04:15 -0600 Subject: [TuxOnIce-users] [mythtv-users] kernel-suspend2 for EL5 on atrpms? In-Reply-To: <6f9afebd0812110009v2867ab39i9fa6c2be8fd1ab1f@mail.gmail.com> References: <6f9afebd0812110009v2867ab39i9fa6c2be8fd1ab1f@mail.gmail.com> Message-ID: <8f2ba1e90812110604h30a02021ldf42f54020617ea6@mail.gmail.com> On Thu, Dec 11, 2008 at 2:09 AM, Brad Benson wrote: > I'm trying to get suspend-to-ram working on my newly-installed CentOS 5.2 > backend and I keep running across references to this repo: > > http://atrpms.net/dist/el5/kernel-suspend2/ > > However, I can't find it. I get 404 errors from that URL. I've found a > similar one: > > http://www.atrpms.net/dist/el5/kernel-tuxonice/ > > But that one currently seems to be empty. Anyone have any idea where I can > find a kernel RPM that will support suspend-to-ram for CentOS 5? Preferably > a PAE kernel as this box has 4GB of ram, but at this point I'll take any > kernel so I can at least test and make sure I can get STR to work at all. > > Thanks, > Brad Benson Brad, The name of the Suspend2 project was changed to TuxOnIce a while back to avoid confusion with other efforts. ATRPMs doesn't always have all the RPM pieces for CentOS support of TuxOnIce, but the missing ones, specifically the kernel-tuxonice RPM can usually be found at Matthias Hensler's website (which I understand is the source for Axel's TuxOnIce RPMs, anyway). Check out http://mhensler.de/swsusp/. In addition, I am sending your inquiry to the TuxOnIce mailing list where you can find additional support for resolving issues. HTH. Craig. From kenny at panix.com Thu Dec 11 18:41:50 2008 From: kenny at panix.com (Kenneth Crudup) Date: Thu, 11 Dec 2008 10:41:50 -0800 (PST) Subject: [TuxOnIce-users] Did someone here say they were using ReiserFS? In-Reply-To: <200812110901.10487.paczesiowa@dw.pl> References: <200812110901.10487.paczesiowa@dw.pl> Message-ID: On Thu, 11 Dec 2008, Bartek ?~Fwik?~Bowski wrote: > I've seen my share of kernel panics related to reiserfs code though, > so one of these days I'll go back to extN. And ext4 is going stable, > since it's faster than reiserfs it's probably a better choice. Huh. About EXT4 being faster than ReiserFS- have they published a paper, or is this just the general wisdom? -Kenny -- Kenneth R. Crudup Sr. SW Engineer, Scott County Consulting, Los Angeles O: 3630 S. Sepulveda Blvd. #138, L.A., CA 90034-6809 (888) 454-8181 From paczesiowa at gmail.com Thu Dec 11 19:13:41 2008 From: paczesiowa at gmail.com (Bartek =?utf-8?q?=C4=86wik=C5=82owski?=) Date: Thu, 11 Dec 2008 20:13:41 +0100 Subject: [TuxOnIce-users] Did someone here say they were using ReiserFS? In-Reply-To: References: <200812110901.10487.paczesiowa@dw.pl> Message-ID: <200812112013.41087.paczesiowa@dw.pl> Hi, > Huh. About EXT4 being faster than ReiserFS- have they published a paper, > or is this just the general wisdom? take a look at this: http://www.phoronix.com/scan.php?page=article&item=ext4_benchmarks&num=1 Regards, Bartek Cwiklowski From rasasi78 at gmail.com Thu Dec 11 23:35:00 2008 From: rasasi78 at gmail.com (=?iso-8859-1?q?Ra=FAl_S=E1nchez_Siles?=) Date: Fri, 12 Dec 2008 00:35:00 +0100 Subject: [TuxOnIce-users] Did someone here say they were using ReiserFS? In-Reply-To: References: Message-ID: <200812120035.03842.rasasi78@gmail.com> El Mi?rcoles 10 Diciembre 2008, Kenneth Crudup escribi?: > I just remember they were saying how fast it is. I was thinking about > experimenting with it, since I have a box with a removable drive bay, > so that having Reiser around doesn't kill my box (*cough*too*cough*). > > Seriously, though- is anyone using it with ToI? Which version, any issues? > > -Kenny Hello: Reiser user on all my machines, former laptop, current, several dasktops. No problem with either reiser and/or reiser+ToI(or suspend2) combination. Well, TBH, if I happen to had any crash that could be remotely charged to reiser if has been very rare (<2 cases on 4 years). HTH, -- Ra?l S?nchez Siles ----->Proud Debian user<----- Linux registered user #416098 -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 197 bytes Desc: This is a digitally signed message part. Url : http://lists.tuxonice.net/pipermail/tuxonice-users/attachments/20081212/281bf89a/attachment.pgp From rasasi78 at gmail.com Thu Dec 11 23:55:20 2008 From: rasasi78 at gmail.com (=?utf-8?q?Ra=C3=BAl_S=C3=A1nchez_Siles?=) Date: Fri, 12 Dec 2008 00:55:20 +0100 Subject: [TuxOnIce-users] Can I leave network media mounted ro? In-Reply-To: <200812030102.15938.paczesiowa@dw.pl> References: <200812030102.15938.paczesiowa@dw.pl> Message-ID: <200812120055.21298.rasasi78@gmail.com> El Mi?rcoles 03 Diciembre 2008, Bartek ?wik?owski escribi?: > Hi, > > http://www.tuxonice.net/HOWTO-4.html mentions that I have to unmount > nfs/smbfs mounted media during hibernation because I "risk losing data". Is > it true if I have some shares mounted read-only (they are also exported > read-only)? I have some movies and music exported from my file-server and I > don't like killing mediaplayers just because I have to unmount their data. > > thanks in advance, > Bartek Cwiklowski Hello: In case it is useful I'll give you my impressions. I use at work CIFS (samba) shared mounts. I use them as regular dirs and I have some of them mounted permanently. I don't use TOI in that situation, but instead vanilla suspend. In this case I hibernate when I have even some pdf opened and being viewed from the shared mounts. On resume I get some CIFS warnings on dmesg but everything is worked correctly. If vanilla suspend does it I can't think of a reason TOI couldn't HTH, regards, -- Ra?l S?nchez Siles ----->Proud Debian user<----- Linux registered user #416098 -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 197 bytes Desc: This is a digitally signed message part. Url : http://lists.tuxonice.net/pipermail/tuxonice-users/attachments/20081212/358f675e/attachment.pgp From lggagnon at uniserve.com Fri Dec 12 21:38:43 2008 From: lggagnon at uniserve.com (Larry Gagnon) Date: Fri, 12 Dec 2008 13:38:43 -0800 Subject: [TuxOnIce-users] S3 resume problem: Acer Aspire 1355LCi with VT8378 S3 Unichrome Video Message-ID: I have an Acer Aspire 1355LCi laptop, runing kernel 2.6.22.19, the Via VT8378 S3 UniChrome Video chip and the latest and greatest hibernate script. Suspend to disk works great (thanks for all the hard work). Suspend to RAM (S3) hangs on resume with no video. Need to hard reset the laptop. I have tried the following to no avail: - attempt to run from single user mode with no modules loaded: hangs - attempted following kernel boot parameters: acpi_sleep=s3_bios then acpi_sleep=s3_mode then acpi_sleep=s3_bios,s3_mode : hangs - attempted using video_post utility: stii hangs From a few hours of googling around it does appear to be a bit of a problem with this video chip and this series of Aspire laptop - I have not seen a single post of anyone having it work. This is a longshot but thought I would use this list to see if there is anything else anyone could suggest I try? thanks.... Larry Gagnon From rasasi78 at gmail.com Sat Dec 13 00:02:26 2008 From: rasasi78 at gmail.com (=?utf-8?q?Ra=C3=BAl_S=C3=A1nchez_Siles?=) Date: Sat, 13 Dec 2008 01:02:26 +0100 Subject: [TuxOnIce-users] Crash on resume. rc7a Message-ID: <200812130102.32878.rasasi78@gmail.com> Hello: I've got this backtrace of an oops preventing my machine to resume on a 2.6.27.8 with ToI 3rc7a: read_pageset+0x110/0x196 read_pageset2+0x23/0x46 copyback_post+0x71/0x8d do_toi_step+0x478/0x6e0 _toi_try_hibernate+0xb5/0xfa toi_main_wrapper+0xb/0xd toi_attr_store+0x1d0/0x216 ?alloc_pages_current+0xbe/0xc6 sysfs_write_file+0xdf/0x114 vfs_write+0xae/0x157 sys_write+0x47/0x70 sys_call_fastpath+0x16/0x1d The beggining of the oops pointed to kernel/power/tuxonice_userui.c:655, but don't take this too serious since I can't remember exactly the line. I also got that the oops was for process pm-hibernate The backtrace should be correct with the execption of any typo I have made. Feel free to ask for more info, but I have to say that I'm not confident I can reproduce this. I can't remember of a ToI problem/oops since months. HTH, -- Ra?l S?nchez Siles ----->Proud Debian user<----- Linux registered user #416098 -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 197 bytes Desc: This is a digitally signed message part. Url : http://lists.tuxonice.net/pipermail/tuxonice-users/attachments/20081213/5042b2a2/attachment.pgp From kenny at panix.com Sat Dec 13 00:16:49 2008 From: kenny at panix.com (Kenneth Crudup) Date: Fri, 12 Dec 2008 16:16:49 -0800 (PST) Subject: [TuxOnIce-users] Crash on resume. rc7a In-Reply-To: <200812130102.32878.rasasi78@gmail.com> References: <200812130102.32878.rasasi78@gmail.com> Message-ID: On Sat, 13 Dec 2008, Ra?l S?nchez Siles wrote: > Feel free to ask for more info, but I have to say that I'm not confident > I can reproduce this. I can't remember of a ToI problem/oops since months. Every now and then a ToI setup that's suspended and resumed many times and had no problems will just "go left" every blue moon on suspend or resume (and it doesn't have to be at the 100th cycle, sometimes it's the first). When you consider the complexity and all that goes on during a cycle, I just figure that as long as it's not consistent, it's just "one of those things". My GF has a Vista laptop (I'm not gonna take the time to bring her to Linux just yet :)) and she's averaging a higher failure rate on Hibernate cycles, though (and her box is nearly as-shipped). -Kenny -- Kenneth R. Crudup Sr. SW Engineer, Scott County Consulting, Los Angeles O: 3630 S. Sepulveda Blvd. #138, L.A., CA 90034-6809 (888) 454-8181 From ncunningham at crca.org.au Sat Dec 13 00:41:18 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Sat, 13 Dec 2008 11:41:18 +1100 Subject: [TuxOnIce-users] Crash on resume. rc7a In-Reply-To: <200812130102.32878.rasasi78@gmail.com> References: <200812130102.32878.rasasi78@gmail.com> Message-ID: <1229128878.7268.135.camel@nigel-laptop> Hi Raul. On Sat, 2008-12-13 at 01:02 +0100, Ra?l S?nchez Siles wrote: > Hello: > > I've got this backtrace of an oops preventing my machine to resume on a > 2.6.27.8 with ToI 3rc7a: > > read_pageset+0x110/0x196 > read_pageset2+0x23/0x46 > copyback_post+0x71/0x8d > do_toi_step+0x478/0x6e0 > _toi_try_hibernate+0xb5/0xfa > toi_main_wrapper+0xb/0xd > toi_attr_store+0x1d0/0x216 > ?alloc_pages_current+0xbe/0xc6 > sysfs_write_file+0xdf/0x114 > vfs_write+0xae/0x157 > sys_write+0x47/0x70 > sys_call_fastpath+0x16/0x1d > > The beggining of the oops pointed to kernel/power/tuxonice_userui.c:655, but > don't take this too serious since I can't remember exactly the line. I also > got that the oops was for process pm-hibernate > > The backtrace should be correct with the execption of any typo I have made. > > Feel free to ask for more info, but I have to say that I'm not confident I > can reproduce this. I can't remember of a ToI problem/oops since months. > > HTH, The reference to the line number is the most useful part, so I can't do much with your report, I'm afraid. If it were to happen again and you got a screenshot with a digital camera, that would be good. Regards, Nigel From ncunningham at crca.org.au Sat Dec 13 00:42:16 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Sat, 13 Dec 2008 11:42:16 +1100 Subject: [TuxOnIce-users] Crash on resume. rc7a In-Reply-To: References: <200812130102.32878.rasasi78@gmail.com> Message-ID: <1229128936.7268.136.camel@nigel-laptop> Hi Kenny. On Fri, 2008-12-12 at 16:16 -0800, Kenneth Crudup wrote: > On Sat, 13 Dec 2008, Ra?l S?nchez Siles wrote: > > > Feel free to ask for more info, but I have to say that I'm not confident > > I can reproduce this. I can't remember of a ToI problem/oops since months. > > Every now and then a ToI setup that's suspended and resumed many times > and had no problems will just "go left" every blue moon on suspend or > resume (and it doesn't have to be at the 100th cycle, sometimes it's > the first). When you consider the complexity and all that goes on during > a cycle, I just figure that as long as it's not consistent, it's just > "one of those things". Please give me reports if/when this happens. I don't want TuxOnIce code to ever oops. I have no control over drivers, but I want to make what I do have control over as good code as possible. Regards, Nigel From kenny at panix.com Sat Dec 13 00:46:44 2008 From: kenny at panix.com (Kenneth Crudup) Date: Fri, 12 Dec 2008 16:46:44 -0800 (PST) Subject: [TuxOnIce-users] Crash on resume. rc7a In-Reply-To: <1229128936.7268.136.camel@nigel-laptop> References: <200812130102.32878.rasasi78@gmail.com> <1229128936.7268.136.camel@nigel-laptop> Message-ID: On Sat, 13 Dec 2008, Nigel Cunningham wrote: > I don't want TuxOnIce code to ever oops. I have no control over drivers ... but that's why I don't sweat it- I'm almost sure that the problem lies in some driver (I've had stuff like that happen to my own code before I've fixed it) because of its intermitent and unpredictable nature. -Kenny -- Kenneth R. Crudup Sr. SW Engineer, Scott County Consulting, Los Angeles O: 3630 S. Sepulveda Blvd. #138, L.A., CA 90034-6809 (888) 454-8181 From ncunningham at crca.org.au Sat Dec 13 06:13:14 2008 From: ncunningham at crca.org.au (Nigel Cunningham) Date: Sat, 13 Dec 2008 17:13:14 +1100 Subject: [TuxOnIce-users] Crash on resume. rc7a In-Reply-To: References: <200812130102.32878.rasasi78@gmail.com> <1229128936.7268.136.camel@nigel-laptop> Message-ID: <1229148794.7268.137.camel@nigel-laptop> Hi. On Fri, 2008-12-12 at 16:46 -0800, Kenneth Crudup wrote: > On Sat, 13 Dec 2008, Nigel Cunningham wrote: > > > I don't want TuxOnIce code to ever oops. I have no control over drivers > > ... but that's why I don't sweat it- I'm almost sure that the problem > lies in some driver (I've had stuff like that happen to my own code before > I've fixed it) because of its intermitent and unpredictable nature. Ah, okay then. :) Nigel From rasasi78 at gmail.com Sat Dec 13 16:18:05 2008 From: rasasi78 at gmail.com (=?utf-8?q?Ra=C3=BAl_S=C3=A1nchez_Siles?=) Date: Sat, 13 Dec 2008 17:18:05 +0100 Subject: [TuxOnIce-users] Crash on resume. rc7a In-Reply-To: <1229128878.7268.135.camel@nigel-laptop> References: <200812130102.32878.rasasi78@gmail.com> <1229128878.7268.135.camel@nigel-laptop> Message-ID: <200812131718.10118.rasasi78@gmail.com> Hello All, Nigel: El S?bado 13 Diciembre 2008, Nigel Cunningham escribi?: > Hi Raul. > > On Sat, 2008-12-13 at 01:02 +0100, Ra?l S?nchez Siles wrote: > > Hello: > > > > The beggining of the oops pointed to > > kernel/power/tuxonice_userui.c:655, but don't take this too serious since > > I can't remember exactly the line. I also got that the oops was for > > process pm-hibernate > > > > The reference to the line number is the most useful part, so I can't do > much with your report, I'm afraid. If it were to happen again and you > got a screenshot with a digital camera, that would be good. > > Regards, > > Nigel > I inspected a liitle that dir files and looks that there's an oops mention in that line, so I thought it may be that. I'm sorry the report is not accurate enough. You know, oops happen in the worst moment. Next time, I'll try to be more reliable in reporting. Thanks for your time and effort. Regards, -- Ra?l S?nchez Siles ----->Proud Debian user<----- Linux registered user #416098 -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 197 bytes Desc: This is a digitally signed message part. Url : http://lists.tuxonice.net/pipermail/tuxonice-users/attachments/20081213/2c54b825/attachment.pgp From jvromans at squirrel.nl Sat Dec 13 19:40:32 2008 From: jvromans at squirrel.nl (Johan Vromans) Date: 13 Dec 2008 20:40:32 +0100 Subject: [TuxOnIce-users] [Not completely on-topic] FbSplash Message-ID: On my notebooks I always boot with kernel parameter vga=791. This gives me more text, and enables ToI to use the fbsplash_ui. However, on my desktops this doesn't seem to work. It seems they do not understand the vga values. Is this common for desktops? The desktops have ATI and nVidia video card (if that matters). -- Johan From molecularbiophysics at gmail.com Tue Dec 16 11:33:12 2008 From: molecularbiophysics at gmail.com (Karthik Paithankar) Date: Tue, 16 Dec 2008 11:33:12 +0000 Subject: [TuxOnIce-users] /lib/firmware problems with intrepid git toi Message-ID: Dear all, After updating to intrepid, I can see some conflicts in installing with the new git-toi kernel_image. The problem being - the firmware is installed directly in /lib/firmware and not in sub-directory /lib/firmware/. https://bugs.launchpad.net/ubuntu/+source/linux/+bug/262115 Has this fix been applied to the git repository ? Best regards, Karthik (I do not have the laptop at the moment, so I do not have the exact message) From rasasi78 at gmail.com Fri Dec 19 23:42:09 2008 From: rasasi78 at gmail.com (=?utf-8?q?Ra=C3=BAl_S=C3=A1nchez_Siles?=) Date: Sat, 20 Dec 2008 00:42:09 +0100 Subject: [TuxOnIce-users] Crash on resume. rc7a In-Reply-To: <1229128878.7268.135.camel@nigel-laptop> References: <200812130102.32878.rasasi78@gmail.com> <1229128878.7268.135.camel@nigel-laptop> Message-ID: <200812200042.12341.rasasi78@gmail.com> Hello: I have new bits about this one. El S?bado 13 Diciembre 2008, Nigel Cunningham escribi?: > Hi Raul. > > On Sat, 2008-12-13 at 01:02 +0100, Ra?l S?nchez Siles wrote: > > Hello: > > > > I've got this backtrace of an oops preventing my machine to resume on a > > 2.6.27.8 with ToI 3rc7a: > > > > read_pageset+0x110/0x196 > > read_pageset2+0x23/0x46 > > copyback_post+0x71/0x8d > > do_toi_step+0x478/0x6e0 > > _toi_try_hibernate+0xb5/0xfa > > toi_main_wrapper+0xb/0xd > > toi_attr_store+0x1d0/0x216