You are not logged in.
archman-cro wrote:You guys talking about pm-hibernate? Is it storing all cache on swap now? Or drops all like before?
Till 2.6.36.2 I used pm-hibernate (which I personally dislike), now I switched to "echo [disk|mem] > /sys/power/state" which works great.
I thought that's basically what PM does, only with the possibility of running scripts on suspend/hibernate or resume/thaw.
Offline
I also noticed that swap is now almost empty after resuming.
Yeah, that's not really a "hibernation" behaviour. That's why I use TuxOnIce for a fair amount of time now. Cheers for replying to my question!
Offline
What is so special about tuxonice?
Last edited by DonVla (2011-01-09 18:12:15)
Offline
What is so special about tuxonice?
It provides a true hibernation? Means puts all cache on swap and copies it back to ram on resume. So you have all your cache back, unlike with pm-hibernate.
Offline
sweet, the new brcm80211 driver works with the macbook air3,2 (bcm43224)!
// sluggo
Offline
DonVla wrote:What is so special about tuxonice?
It provides a true hibernation? Means puts all cache on swap and copies it back to ram on resume. So you have all your cache back, unlike with pm-hibernate.
The pm-hibernate is not doing anything to your cache. It calls actual hibernation methods, one of which can be tuxonice, to perform the suspend.
You need to install an RTFM interface.
Offline
sweet, the new brcm80211 driver works with the macbook air3,2 (bcm43224)!
Yes, audio and backlight also work out of the box. Keyboard and touchpad still seem to need patches..
ᶘ ᵒᴥᵒᶅ
Offline
@tomk carl9170 works great see https://wiki.archlinux.org/index.php/Wireless#carl9170
thanks amirs - I'm glad it works for you. Its a location issue for me - the D-Link dongle that I'm using doesn't have sufficient range to reliably pick up my AP in the attic.
Sticking with my rock-solid b43 for now, but considering relocating the router.
Offline
DonVla wrote:What is so special about tuxonice?
It provides a true hibernation? Means puts all cache on swap and copies it back to ram on resume. So you have all your cache back, unlike with pm-hibernate.
I've come to correct myself, it seems like the kernel hibernation _does_ restore caches, on resuming this morning I had some 600M of cache. I can't compare with my caches before hibernation, but I'll do so the next time
Offline
archman-cro wrote:DonVla wrote:What is so special about tuxonice?
It provides a true hibernation? Means puts all cache on swap and copies it back to ram on resume. So you have all your cache back, unlike with pm-hibernate.
I've come to correct myself, it seems like the kernel hibernation _does_ restore caches, on resuming this morning I had some 600M of cache. I can't compare with my caches before hibernation, but I'll do so the next time
Thanks!
Btw, can you describe how you use pm-hibernate? You just start it or do you have some custom config? If the caches are reloaded upon resuming, I would really be interested in pm, since then I don't need TOI.
Last edited by archman-cro (2011-01-10 11:31:08)
Offline
Ramses de Norre wrote:archman-cro wrote:It provides a true hibernation? Means puts all cache on swap and copies it back to ram on resume. So you have all your cache back, unlike with pm-hibernate.
I've come to correct myself, it seems like the kernel hibernation _does_ restore caches, on resuming this morning I had some 600M of cache. I can't compare with my caches before hibernation, but I'll do so the next time
Thanks!
Btw, can you describe how you use pm-hibernate? You just start it or do you have some custom config? If the caches are reloaded upon resuming, I would really be interested in pm, since then I don't need TOI.
I use it through xfce4-power-manager to be able to lock the screen before suspension/hibernation. But you can also easily run it via acpi scrips or by just using a keybind to run pm-suspend or pm-hibernate.
BTW, you can also use pm to suspend with TOI. Take a look at the pm-utils wiki article, I think it explains all this fairly well.
Offline
Interesting. I subscribed to this topic to see how long it takes to have the new kernel in the non-testing repos. Research is fun :-D
We just had a kernel update yesterday (?). I was amazed... it was quite a while since I saw grub the last time on my desktop. Maybe again, soon.
Offline
Ramses de Norre wrote:archman-cro wrote:It provides a true hibernation? Means puts all cache on swap and copies it back to ram on resume. So you have all your cache back, unlike with pm-hibernate.
I've come to correct myself, it seems like the kernel hibernation _does_ restore caches, on resuming this morning I had some 600M of cache. I can't compare with my caches before hibernation, but I'll do so the next time
Thanks!
Btw, can you describe how you use pm-hibernate? You just start it or do you have some custom config? If the caches are reloaded upon resuming, I would really be interested in pm, since then I don't need TOI.
I use s2disk from uswsusp and in /etc/suspend.conf i just made my "image size" parameter large enough.
That way, caches and buffers are not dropped, it is a feature.
Last edited by kokoko3k (2011-01-10 14:10:42)
Help me to improve ssh-rdp !
Retroarch User? Try my koko-aio shader !
Offline
Ramses de Norre wrote:archman-cro wrote:It provides a true hibernation? Means puts all cache on swap and copies it back to ram on resume. So you have all your cache back, unlike with pm-hibernate.
I've come to correct myself, it seems like the kernel hibernation _does_ restore caches, on resuming this morning I had some 600M of cache. I can't compare with my caches before hibernation, but I'll do so the next time
Thanks!
Btw, can you describe how you use pm-hibernate? You just start it or do you have some custom config? If the caches are reloaded upon resuming, I would really be interested in pm, since then I don't need TOI.
pm-utils doesn't do the hibernating, it handles the pre-hibernate business (alternative to hibernate script).
@Awebb, the kernel update in [core] isn't 2.6.37 but a security update to 2.6.36. 2.6.37 will probably sit AT LEAST a month in [testing], like all previous kernels...
Allan-Volunteer on the (topic being discussed) mailn lists. You never get the people who matters attention on the forums.
jasonwryan-Installing Arch is a measure of your literacy. Maintaining Arch is a measure of your diligence. Contributing to Arch is a measure of your competence.
Griemak-Bleeding edge, not bleeding flat. Edge denotes falls will occur from time to time. Bring your own parachute.
Offline
[offtopic] Is it possible to run a command upon resuming (using just uswsusp) without having pm or hibernate-script? I want to run "sdparm -B" after resuming back into my session. [/offtopic]
Last edited by archman-cro (2011-01-10 17:33:47)
Offline
Offline
Useless, cause I said "without pm" . Anyway, a dirty fix is here: just alias a set of commands: "s2disk && hdparm -B". Tested&Works. ^^
Offline
I have a lot of trouble with this kernel, leading to strong instability. System crashes (freezes) completely, especially when using video (mplayer switching from fullscreen to windowed) or watching Youtube videos in Firefox. The freezes occur at random, quite often but are not 100% reproducible. I have to do a hard reboot, with no message in the logs. The problem is described in this other thread: https://bbs.archlinux.org/viewtopic.php?id=111465 where I tried several package combinations with the nvidia proprietary driver. It is not related to the udev-165 issue also described in the Kernel & Hardware sub-section.
I have a x86_64 system (i7 870, chipset P55), with a GeForce GTX 260 card.
Last edited by zebulon (2011-01-10 20:59:24)
Offline
@zebulon: Thats bad news. I am having similar symptoms with any combination newer than kernel 2.6.34, xorg 1.8 and nvidia 256.x. (Yes it doesn't feel like I am using Arch anymore, but anyway.) I was hoping some kernel magic will fix it in this release. You can follow more about it on nvnews, like here and here, to see if you have a similar problem.
Offline
@zebulon: Thats bad news. I am having similar symptoms with any combination newer than kernel 2.6.34, xorg 1.8 and nvidia 256.x. (Yes it doesn't feel like I am using Arch anymore, but anyway.) I was hoping some kernel magic will fix it in this release. You can follow more about it on nvnews, like here and here, to see if you have a similar problem.
Well, that's very strange I do not experience the issue before kernel version 2.6.37. Is it dpendent on some combination of hardware, kernel and nvidia driver? Do you solve your problem if you use nouveau?
Offline
My machine seems to run fine with the new kernel. Hell, I'd say it runs better with this kernel than the previous one (or maybe it's just the numerous performance-related optimism I had read that is altering my perception).
Okay, I lied. ID mapping doesn't work with NVSv4 it seems. But it is purported that this issue is not intrinsic to the kernel, but to nfs-utils (whose update I'm consequently eagerly awaiting).
(As an aside, I strongly encourage people who are reading this thread that are not subscribed to the testing repositories to do so. It's far less broken that you think )
Last edited by orlandu63 (2011-01-11 02:38:30)
Offline
(As an aside, I strongly encourage people who are reading this thread that are not subscribed to the testing repositories to do so. It's far less broken that you think )
Never post this without also adding that they should subscribe to [arch-dev-public]. Arch users SHOULD know the prerequisites (or search the wiki to find out), but the number of random complaints about [testing] where it turns out user A isn't subscribed to [arch-dev-public] is disheartening
Allan-Volunteer on the (topic being discussed) mailn lists. You never get the people who matters attention on the forums.
jasonwryan-Installing Arch is a measure of your literacy. Maintaining Arch is a measure of your diligence. Contributing to Arch is a measure of your competence.
Griemak-Bleeding edge, not bleeding flat. Edge denotes falls will occur from time to time. Bring your own parachute.
Offline
Well, that's very strange I do not experience the issue before kernel version 2.6.37.
Thats why I am not sure you have the same problem as I am having. But the problem is similar, random lockups requiring hard reset.
Is it dpendent on some combination of hardware, kernel and nvidia driver?
Also include xorg in this list. Apparently the mechanism of switching virtual terminals was changed in xorg 1.9 which introduced this regression, since Nvidia is still using the old mechanism. Not all, only some chips are affected, which shows that it is also related to the hardware and perhaps to the kernel as well.
Do you solve your problem if you use nouveau?
I don't have any problems in Fedora with nouveau module. So I guess it should also work in Arch.
Offline
orlandu63 wrote:(As an aside, I strongly encourage people who are reading this thread that are not subscribed to the testing repositories to do so. It's far less broken that you think )
Never post this without also adding that they should subscribe to [arch-dev-public]. Arch users SHOULD know the prerequisites (or search the wiki to find out), but the number of random complaints about [testing] where it turns out user A isn't subscribed to [arch-dev-public] is disheartening
Well, I am subscribed to the ML, but wanted to share the issue with others on the testing section of the forum, to see if others noticed the issue, which seems to be the case. Is not that the point of a [testing] sub-section in the forum?
Offline
zebulon wrote:Well, that's very strange I do not experience the issue before kernel version 2.6.37.
Thats why I am not sure you have the same problem as I am having. But the problem is similar, random lockups requiring hard reset.
Do you have anything in /var/log/everything.log ?
Offline