The linux-pf thread; BFS/CK, TuxOnIce, BFQ, AUFS3

jakob · 2011-07-20 13:52:00

see kernel26-pf AUR page, where Phate666 and I have reported hibernating or rather resuming problems with pf4…

however, pf3 does the job

kalpik · 2011-07-20 13:54:43

jakob wrote:

see kernel26-pf AUR page, where Phate666 and I have reported hibernating or rather resuming problems with pf4…
however, pf3 does the job

Ah! Thank you so much! I was breaking my head with this since 4-5 hours

xclaude · 2011-08-03 23:30:41

as deescribed in the gentoo wiki, the lzo comprossion should not be loaded as a module, but in your standard-config (and maybe your precompiled kernels) you use it as a module.

http://en.gentoo-wiki.com/wiki/TuxOnIce … figuration

3. LZO compression support may speed up the hibernation process if you have a fast CPU. Do not compile LZO compression algorithm as a module.
You will likely not be able to resume from the image on boot up.

I think this is the reason, why the hibernation is not working with lzo and your kernel26-pf as described in the arch wiki.

this user had the same problem.
https://bbs.archlinux.org/viewtopic.php?id=122889

Edit:
i recompiled the kernel with static lzo enabled, but the error -22 when resuming still exists, later this day i'll try lzf compression and/or other kernel.

Last edited by xclaude (2011-08-04 09:11:28)

post-factum · 2011-08-04 21:27:09

We've moved to new server. I hope, you haven't noticed that . In case of some troubles mail me, please.

nous · 2011-08-04 21:33:35

xclaude wrote:

as deescribed in the gentoo wiki, the lzo comprossion should not be loaded as a module, but in your standard-config (and maybe your precompiled kernels) you use it as a module.
http://en.gentoo-wiki.com/wiki/TuxOnIce … figuration
3. LZO compression support may speed up the hibernation process if you have a fast CPU. Do not compile LZO compression algorithm as a module.
You will likely not be able to resume from the image on boot up. 
I think this is the reason, why the hibernation is not working with lzo and your kernel26-pf as described in the arch wiki.
this user had the same problem.
https://bbs.archlinux.org/viewtopic.php?id=122889
Edit:
i recompiled the kernel with static lzo enabled, but the error -22 when resuming still exists, later this day i'll try lzf compression and/or other kernel.

LZO is built-in kernel by default in kernel26-pf; I'm sure it would also work if compiled as module and added in the MODULES array of /etc/mkinitcpio.conf, though I haven't tested it as that because I don't use an initramfs in my laptop. OTOH, turning off LZO beats (well, kinda) the purpose of tuxonice... I'm still with 2.6.39-pf3.

Last edited by nous (2011-08-04 21:38:14)

xclaude · 2011-08-06 12:28:25

after some compiling-orgies :-) i have found a solution for me, using tux-on-ice with compression. using suspend whithout compression was no option for me.

i recompiled the kernel using the pkgbuild from aur. modified the kernel-cofiguration using your config.x86_64 set as base and changed the processor-typ to Opteron/Athlon64 (i'm using Turion64 x2) and under crypto-api i enabled lzf static (*).
now it works perfect for me.

don't know why its not working with lzo, even whith static or loaded as modul in initramfs.

nous · 2011-08-06 13:23:57

Well, I compiled pf4 with a tailored config for my laptop (Dell Vostro 1500), no initramfs, lzo compression. I had an uptime of quite a few hibernation cycles with pf3, so I hibernated with pf3 and resumed with pf4. It went just fine. To rule out the possibility of pf4 corrupting somehow the image, I re-hibernated with pf4 and resumed again without problem. So, it appears that the -22 error hits specific hardware. As a side note, my laptop sports an Intel Core 2 Duo.

[EDIT]
Here's some more info:

Extended CMOS year: 2000
Enabling non-boot CPUs ...
Booting Node 0 Processor 1 APIC 0x1
smpboot cpu 1: start_ip = 9a000
Switched to NOHz mode on CPU #1
CPU1 is up
ACPI: Waking up from system sleep state S4
HDA Intel 0000:00:1b.0: restoring config space at offset 0x1 (was 0x100106, writing 0x100102)
ata_piix 0000:00:1f.1: restoring config space at offset 0x1 (was 0x2880001, writing 0x2880005)
ahci 0000:00:1f.2: restoring config space at offset 0x1 (was 0x2b00403, writing 0x2b00407)
nvidia 0000:01:00.0: restoring config space at offset 0xc (was 0xfea00000, writing 0x0)
nvidia 0000:01:00.0: restoring config space at offset 0x3 (was 0x10, writing 0x0)
PM: early restore of devices complete after 1.763 msecs
HDA Intel 0000:00:1b.0: PCI INT A -> GSI 21 (level, low) -> IRQ 21
HDA Intel 0000:00:1b.0: setting latency timer to 64
HDA Intel 0000:00:1b.0: irq 45 for MSI/MSI-X
ata_piix 0000:00:1f.1: PCI INT A -> GSI 16 (level, low) -> IRQ 16
ata_piix 0000:00:1f.1: setting latency timer to 64
PM: restore of devices complete after 1036.900 msecs
...20%...40%...60%...80%
Waited for i/o due to readahead not ready 516 times.
Waited for i/o due to throughput_throttle 451 times.
Restarting all filesystems ...
Restarting tasks ... done.
video LNXVIDEO:00: Restoring backlight state
TuxOnIce debugging info:
- TuxOnIce core  : 3.2
- Kernel Version : 2.6.39-pf
- Compiler vers. : 4.6
- Attempt number : 7
- Parameters     : 0 667657 0 1 -2 0
- Overall expected compression percentage: 0.
- Compressor is 'lzo'.
  Compressed 875954176 bytes into 340785543 (61 percent compression).
- Block I/O active.
  Used 84047 pages from swap on /dev/sda8.
- Max outstanding reads 1378. Max writes 1528.
  Memory_needed: 1024 x (4096 + 336 + 104) = 4644864 bytes.
  Free mem throttle point reached 0.
- Swap Allocator enabled.
  Swap available for image: 634558 pages.
- I/O speed: Write 108 MB/s, Read 136 MB/s.
- Extra pages    : 862 used/7500.
- Result         : Succeeded.

[SECOND EDIT]
A subsequent resume failed with the usual -22 error. Back to pf3.

Last edited by nous (2011-08-07 20:56:17)

kalpik · 2011-08-08 02:05:54

Ok, seeing that linux-3.0 is out, what's the recommended upgrade path? Its all too confusing right now with kernel26-pf and linux-pf

nous · 2011-08-08 22:43:10

kalpik wrote:

Ok, seeing that linux-3.0 is out, what's the recommended upgrade path? Its all too confusing right now with kernel26-pf and linux-pf

Just wait for linux-pf 3.0.1, as Con Kolivas hasn't yet updated his BFS patchset for linux 3. The 3.0.0 version of linux-pf in the AUR is just a placeholder. To update when ready, simply install linux-pf and it will (hopefully) replace kernel26-pf.

kalpik · 2011-08-09 01:57:05

nous wrote:

kalpik wrote:
Ok, seeing that linux-3.0 is out, what's the recommended upgrade path? Its all too confusing right now with kernel26-pf and linux-pf
Just wait for linux-pf 3.0.1, as Con Kolivas hasn't yet updated his BFS patchset for linux 3. The 3.0.0 version of linux-pf in the AUR is just a placeholder. To update when ready, simply install linux-pf and it will (hopefully) replace kernel26-pf.

Thanks! Will wait for linux-pf to hit your repo

nous · 2011-08-11 21:57:37

The first pf patchset for linux-pf is online! Grab yours from the AUR and post your impressions!

lucak3 · 2011-08-12 10:42:56

I got nvidia-pf up and running but we have a problem: linux-pf returns 'Linux arch 3.0.1-pf' with uname and that makes the nvidia builder require that i set the _kernver to 3.0.1. Here's the problem, i would need to update the package for every 3.0.x kernel and REQUIRE that you have that version. Please scour the official linux PKGBUILD to see how they make it so to return 3.0-ARCH with uname.

nous · 2011-08-12 14:24:50

lucak3 wrote:

I got nvidia-pf up and running but we have a problem: linux-pf returns 'Linux arch 3.0.1-pf' with uname and that makes the nvidia builder require that i set the _kernver to 3.0.1. Here's the problem, i would need to update the package for every 3.0.x kernel and REQUIRE that you have that version. Please scour the official linux PKGBUILD to see how they make it so to return 3.0-ARCH with uname.

Hmmm, I see what the problem is. I missed it because I use nvidia-beta-all (which builds fine for both -ARCH and -pf). Fixed.

nous · 2011-08-13 16:54:33

I've added indexes to the dropbox repository of pfkernel. It can now be browsed as a regular web page.

kalpik · 2011-08-13 17:12:12

Thanks!

Tuxonice still not functional?

nous · 2011-08-13 17:40:46

kalpik wrote:

Thanks!
Tuxonice still not functional?

Partly. Resuming a LZO compressed image doesn't work (at least in my setup). For the time being I use 'Compressor none' in /etc/hibernate/tuxonice.conf which is 60% slower than lzo but does the trick.

Last edited by nous (2011-08-26 22:16:06)

kalpik · 2011-08-13 17:50:34

nous wrote:

kalpik wrote:
Thanks!
Tuxonice still not functional?
Partly. Resuming a LZO compressed image doesn't work (at least in my setup). For the time being I use 'Compressor none' in /etc/hibernate/tuxonice.conf which is 60% slower than lzo but does the trick. If anyone has the time to find the guilty commit to the LZO code from 2.6.39.2 to .3...

Is it possible to use lzf in the meantime?

kalpik · 2011-08-13 17:57:39

Hey one more thing I've noticed is that with this kernel, I get dmesg output in between my deamon startup sequence. Similar to this: https://bbs.archlinux.org/viewtopic.php?id=34582

It does not happen with the arch kernel.

nous · 2011-08-13 17:58:09

kalpik wrote:

nous wrote:
kalpik wrote:
Tuxonice still not functional?
Partly. Resuming a LZO compressed image doesn't work (at least in my setup). For the time being I use 'Compressor none' in /etc/hibernate/tuxonice.conf which is 60% slower than lzo but does the trick. If anyone has the time to find the guilty commit to the LZO code from 2.6.39.2 to .3...
Is it possible to use lzf in the meantime?

I've tried and failed with lzf built-in.

jakob · 2011-08-13 21:55:24

I’ve had no success with encrypted hibernation, although I’ve CONFIG_CRYPTO_LZO=y, so for the time being, I’m using tuxonice unencrypted and it does work…

However, I’m having the same dmesg output in tty1 as kalpik reported in #168… and it doesn’t appear when using current linux-ARCH kernel…

xclaude · 2011-08-16 08:41:19

for me, lzf compression for hibernating/resume is working.
i have built the kernel from aur using the config.x86_64 as base and changed only optimize to Opteron/K8 and enabled lzf as module.

i'm using encrypted partitions.

/dev/sda1                 /boot
/dev/sda2                 luks encrypted lvm-pv
/dev/mapper/vg0-root      /
/dev/mapper/vg0-home      /home
/dev/mapper/vg0-swap      swap

in mkinitcpio.conf i loaded lzf as module and userui as hook, changed compression to lzf and the right path to tuxoniceui in the hibernate-scipts, in my kernel cmd-line i added resume=/dev/mapper/vg0-swap
after then i rebuilt the inintramfs image.

@nous
P.S.: maybe there is something wrong with your repository when trying to install linux-pf with pacman it downloads only 5.5K and brings errormessage

 
[root@myhost vm-test]# pacman -S linux-pf-k8
Löse Abhängigkeiten auf...
Suche nach Zwischenkonflikten...

Pakete (1): linux-pf-k8-3.0.1-4

Gesamtgröße des Downloads: 39,17 MB
Gesamtgröße der zu installierenden Pakete: 93,76 MB

Installation fortsetzen?  [J/n] 
:: Empfange Pakete von pfkernel...
 linux-pf-k8-3.0.1-4-x86_64                          5,5K   33,1K/s 00:00:00 [############################################] 100%
:: Die Datei linux-pf-k8-3.0.1-4-x86_64.pkg.tar.xz ist beschädigt. Möchten Sie sie löschen? [J/n]

when downloading the file with webbrowser the file seems to be o.k. (i only checked the size)

Last edited by xclaude (2011-08-16 09:50:42)

snack · 2011-08-16 09:01:20

Hello, I'd like to try the BFS, but on my i686 system pae is a must for me. Is there any PKGBUILD (or better, repo) sporting both BFS and pae? Thanks.

nous · 2011-08-16 09:18:02

xclaude wrote:

for me, lzf compression for hibernating/resume is working.
i have built the kernel from aur using the config.x86_64 as base and changed only optimize to Opteron/K8 and enabled lzf as module.
in mkinitcpio.conf i loaded lzf as module and userui as hook,
changed compression to lzo

You meant lzf here, correct?

Thanks for the feedback, I'll test this setup on a desktop pc and see how it goes.

xclaude wrote:

@nous
P.S.: maybe there is something wrong with your repository when trying to install linux-pf with pacman it downloads only 5.5K and brings errormessage

I just downloaded linux-pf-k8 with pacman -Sw and it is OK. Maybe you encountered a glitch in dropbox or some proxy leftover cache?

xclaude · 2011-08-16 09:49:50

yes your right, i'm using LZF.
LZO is still with error -22 when resuming.

and the error with the repository still exists in my vmware even after -Syy, but works correct on my notebook - i've no idea why it is so ?!?
both, the vm (bridged network) and my notebook behind the same router without a proxy.

nous · 2011-08-16 10:19:10

snack wrote:

Hello, I'd like to try the BFS, but on my i686 system pae is a must for me. Is there any PKGBUILD (or better, repo) sporting both BFS and pae? Thanks.

I can't alter the stock -ARCH config any further, because it'll become difficult to maintain (and I'll keep forgetting the changes). It's easy to build your own kernel though. Download the linux-pf package, untar and run 'makepkg -i'.

At this prompt press <M> + <ENTER>:

==> Kernel configuration options before build:
    <M> make menuconfig (console menu)
    <N> make nconfig (newer alternative to menuconfig)
    <G> make gconfig (needs gtk)
    <X> make xconfig (needs qt)
    <O> make oldconfig
    <ENTER> to skip configuration and start compiling

At the menu select "Processor type and features --->", then "High Memory Support", then "64 GB", exit and save.

Arch Linux

#151 2011-07-20 13:52:00

Re: The linux-pf thread; BFS/CK, TuxOnIce, BFQ, AUFS3

#152 2011-07-20 13:54:43

Re: The linux-pf thread; BFS/CK, TuxOnIce, BFQ, AUFS3

#153 2011-08-03 23:30:41

Re: The linux-pf thread; BFS/CK, TuxOnIce, BFQ, AUFS3

#154 2011-08-04 21:27:09

Re: The linux-pf thread; BFS/CK, TuxOnIce, BFQ, AUFS3

#155 2011-08-04 21:33:35

Re: The linux-pf thread; BFS/CK, TuxOnIce, BFQ, AUFS3

#156 2011-08-06 12:28:25

Re: The linux-pf thread; BFS/CK, TuxOnIce, BFQ, AUFS3

#157 2011-08-06 13:23:57

Re: The linux-pf thread; BFS/CK, TuxOnIce, BFQ, AUFS3

#158 2011-08-08 02:05:54

Re: The linux-pf thread; BFS/CK, TuxOnIce, BFQ, AUFS3

#159 2011-08-08 22:43:10

Re: The linux-pf thread; BFS/CK, TuxOnIce, BFQ, AUFS3

#160 2011-08-09 01:57:05

Re: The linux-pf thread; BFS/CK, TuxOnIce, BFQ, AUFS3

#161 2011-08-11 21:57:37

Re: The linux-pf thread; BFS/CK, TuxOnIce, BFQ, AUFS3

#162 2011-08-12 10:42:56

Re: The linux-pf thread; BFS/CK, TuxOnIce, BFQ, AUFS3

#163 2011-08-12 14:24:50

Re: The linux-pf thread; BFS/CK, TuxOnIce, BFQ, AUFS3

#164 2011-08-13 16:54:33

Re: The linux-pf thread; BFS/CK, TuxOnIce, BFQ, AUFS3

#165 2011-08-13 17:12:12

Re: The linux-pf thread; BFS/CK, TuxOnIce, BFQ, AUFS3

#166 2011-08-13 17:40:46

Re: The linux-pf thread; BFS/CK, TuxOnIce, BFQ, AUFS3

#167 2011-08-13 17:50:34

Re: The linux-pf thread; BFS/CK, TuxOnIce, BFQ, AUFS3

#168 2011-08-13 17:57:39

Re: The linux-pf thread; BFS/CK, TuxOnIce, BFQ, AUFS3

#169 2011-08-13 17:58:09

Re: The linux-pf thread; BFS/CK, TuxOnIce, BFQ, AUFS3

#170 2011-08-13 21:55:24

Re: The linux-pf thread; BFS/CK, TuxOnIce, BFQ, AUFS3

#171 2011-08-16 08:41:19

Re: The linux-pf thread; BFS/CK, TuxOnIce, BFQ, AUFS3

#172 2011-08-16 09:01:20

Re: The linux-pf thread; BFS/CK, TuxOnIce, BFQ, AUFS3

#173 2011-08-16 09:18:02

Re: The linux-pf thread; BFS/CK, TuxOnIce, BFQ, AUFS3

#174 2011-08-16 09:49:50

Re: The linux-pf thread; BFS/CK, TuxOnIce, BFQ, AUFS3

#175 2011-08-16 10:19:10

Re: The linux-pf thread; BFS/CK, TuxOnIce, BFQ, AUFS3

Board footer