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

gee · 2011-11-01 04:35:23

nous, how do you install aufs3-utils ? do you use a different pkgbuild or something?

I am curious

nous · 2011-11-01 05:49:27

gee wrote:

nous, how do you install aufs3-utils ? do you use a different pkgbuild or something?
I am curious

I don't :-) I incorporated aufs3 as a request...

gee · 2011-11-01 06:15:36

Ooooooooh

Then same question to whoever uses aufs3

olansley · 2011-11-03 07:57:36

I managed to get aufs-util to compile by patching the makefile with CPPFLAGS to include the pf kernel include path.
Instructions to do this are in the readme in the src of aufs3-util.

The patch added the following line to the makefile:

CPPFLAGS += -I/lib/modules/3.0-pf/build/include

Hope this helps

nous · 2011-11-03 12:19:31

olansley wrote:

I managed to get aufs-util to compile by patching the makefile with CPPFLAGS to include the pf kernel include path.
Instructions to do this are in the readme in the src of aufs3-util.
The patch added the following line to the makefile:

CPPFLAGS += -I/lib/modules/3.0-pf/build/include
Hope this helps

Thaks, I'm posting this to the aufs3-util AUR page, to alert the maintainer .

lkraav · 2011-11-03 14:19:31

BFS for 3.1.0 was released last night. yay.

gee · 2011-11-04 04:07:55

awesome guys thank you!
It'd be interesting to understand why we need that though.

kalpik · 2011-11-04 17:17:05

Pardon my ignorance, but is compression working fine now for TuxOnIce?

kokoko3k · 2011-11-05 14:11:13

Just a little warning for anyone using bfs + vmware (player,workstation) 8.
The default io scheduler, bfq, doesn't play nice with it.
Guest operating system (xp in my case) can freeze the host indefinitely when doing hd i/o, this is evident in boot stage (guest also take ages to boot).

Switching from bfq to noop or cfq makes the machine resurrect, instantly.

Tried with 3.0.7-pf

-EDIT-
I've to add that this is the first time i noticed that bug, i mean that it only happens with vmware.

Last edited by kokoko3k (2011-11-05 14:29:44)

post-factum · 2011-11-05 23:16:40

BFQ for 3.1 is known for bugs. Please, wait for updated release from its developers.

kalpik · 2011-11-06 12:07:47

Why is the nvidia-pf in the repo so old?

el mariachi · 2011-11-06 12:21:53

kokoko3k wrote:

Just a little warning for anyone using bfs + vmware (player,workstation) 8.
The default io scheduler, bfq, doesn't play nice with it.
Guest operating system (xp in my case) can freeze the host indefinitely when doing hd i/o, this is evident in boot stage (guest also take ages to boot).
Switching from bfq to noop or cfq makes the machine resurrect, instantly.
Tried with 3.0.7-pf
-EDIT-
I've to add that this is the first time i noticed that bug, i mean that it only happens with vmware.

I confirm this. After switching to noop the issue was gone.

nous · 2011-11-06 12:58:20

kalpik wrote:

Why is the nvidia-pf in the repo so old?

Because I have the Alzheimer's

Last edited by nous (2011-11-06 12:58:33)

kalpik · 2011-11-06 13:03:20

nous wrote:

kalpik wrote:
Why is the nvidia-pf in the repo so old?
Because I have the Alzheimer's

No worries, I'll bug you whenever it's out of date

lkraav · 2011-11-10 23:04:31

as a side note, is it... asking for trouble to patch linux-vserver on top of pf-sources? seems to apply quite cleanly.

$ patch -s -p1 --dry-run < /mnt/datapool/gentoo/distfiles/patch-3.0.7-vs2.3.1.diff
1 out of 1 hunk FAILED -- saving rejects to file Makefile.rej
2 out of 16 hunks FAILED -- saving rejects to file include/linux/fs.h.rej
$

edit: Makefile is obviously for "-pf" and fs.h is about only some #defines getting rearranged with pf-patch going first. added the missing lines from vserver-diff .rej and all looks safe at first sight, i'll start building in the morning :>

edit2: built and booted this monstrosity :> got a strange hang rebuilding kernel on a zfs-fuse mount.

felt quite similar to above:

The default io scheduler, bfq, doesn't play nice with it.

otoh 3.0.2-pf hasn't had a single hickup with it.. went elevator=cfq for the time being just in case.

Last edited by lkraav (2011-11-11 21:25:53)

el mariachi · 2011-11-12 14:45:05

This patch seems useful, is anyone thinking about adding it to the mix while it doesn't get adopted upstream?
https://lkml.org/lkml/2011/11/11/178

nous · 2011-11-12 21:06:19

el mariachi wrote:

This patch seems useful, is anyone thinking about adding it to the mix while it doesn't get adopted upstream?
https://lkml.org/lkml/2011/11/11/178

I think post-factum has already added it; see http://pf.natalenko.name/.

nous · 2011-11-12 22:33:34

Hi people.
I created this package (starting with kernel26-pf) because I used both BFS and tuxonice, patching tediously my kernel, before I found out that post-factum had actually set up a git repository doing the same thing the right and elegant way. That was when I decided to provide a package here, to let other Archers benefit from his effort.
That being said and as "real life" obligations keep squeezing my free time to extinction, I've decided to seek a devoted and capable maintainer for this package. I still use linux-pf exclusively on all my boxes and will do so, therefore I need someone to take good care of it. At the very least, one should have:
1. A firm grasp of bash shell scripting (for handling the PKGBUILD)
2. Adequate experience in linux generally and in Arch linux specifically
3. Some knowledge of kernel compilation tasks (e.g. patching, resolving gcc errors, understanding kernel configuration options etc)

I'll keep providing precompiled binaries in my repo, as building them is scripted and trivial once the PKGBUILD is updated and the 2 kernel configs from the stock -ARCH kernel are adapted to -pf every time the minor version number changes. I'll also help the new maintainer with the whys and wtfs of the PKGBUILD.

lkraav · 2011-11-12 23:01:57

lkraav wrote:

looking forward to 3.1.0-pf. in the meanwhile, i'm running 3.0.2-pf, does anyone happen to know why my laptop system time stops while in suspend?
i thought only PM_DEBUG and friends did that since they save their magic into RTC, but after disabling these, system time still stops in suspend. this used to work fine in 2.6.39 for sure. i suspect it has something to do with the new RTC subsystem redesign in 3.0.
could this be related:
$ sudo hwclock -r
hwclock: select() to /dev/rtc to wait for clock tick timed out: Success
adding --directisa makes hwclock work, but is this how it's supposed to be?
full kernel config is at http://paste.pocoo.org/show/500367/, perhaps someone knows what i'm doing wrong?
these seem interesting:
CONFIG_RTC_HCTOSYS=y
CONFIG_RTC_HCTOSYS_DEVICE="rtc0"

installed 3.1.1-pf. it did not help with my stopping time issue. resetting clock in BIOS ended up fixing it.

i eventually also tried going back to CONFIG_RTC [2]. did not help and hwclock -r still failed.

after finding [1], i thought i'd give resetting BIOS
a try. arriving in BIOS i noticed the time shown there was off for even UTC, so
i set it right, and saved the result - i.e. did not even do a complete "load
defaults".

after next boot, hwclock -r seemed to work without issues and
testing suspend also resulted in correct time after resume. hopefully problem
solved.

[1]: http://linux.derkeiler.com/Mailing-List … 01326.html
[2]: https://bugs.gentoo.org/show_bug.cgi?id=212588

mrbrich · 2011-11-15 01:05:32

I hope someone does continue to maintain is as having this binary package is really useful. I also wanted to reiterate the request of kokoko3k regarding CONFIG_TIMER_STATS and powertop. Tuxonice is particularly useful on laptops and powertop is also useful on laptops for monitoring power consumption, so I would like to see this option enabled. Many thanks.

kokoko3k · 2011-11-15 08:24:23

It appeasrs config_timer_stats and TOI are included in linux-3.1-pf:

Gozer tmp # grep 'CONFIG_TIMER_STATS\|TOI' /usr/src/linux-3.1-pf/.config
CONFIG_TOI_CORE=m
CONFIG_TOI_FILE=m
CONFIG_TOI_SWAP=m
CONFIG_TOI_CRYPTO=m
CONFIG_TOI_USERUI=m
CONFIG_TOI_USERUI_DEFAULT_PATH="/usr/sbin/tuxoniceui"
CONFIG_TOI_DEFAULT_IMAGE_SIZE_LIMIT=-2
# CONFIG_TOI_KEEP_IMAGE is not set
CONFIG_TOI_REPLACE_SWSUSP=y
# CONFIG_TOI_IGNORE_LATE_INITCALL is not set
CONFIG_TOI_DEFAULT_WAIT=25
CONFIG_TOI_DEFAULT_EXTRA_PAGES_ALLOWANCE=7500
# CONFIG_TOI_CHECKSUM is not set
CONFIG_TOI=y
CONFIG_TOI_EXPORTS=y
CONFIG_TOI_ZRAM_SUPPORT=y
CONFIG_TOIM3232_DONGLE=m
CONFIG_TIMER_STATS=y

Thanks!

Last edited by kokoko3k (2011-11-15 08:24:40)

nous · 2011-11-15 12:55:19

kokoko3k wrote:

It appeasrs config_timer_stats and TOI are included in linux-3.1-pf

Indeed!

mrbrich wrote:

I hope someone does continue to maintain is as having this binary package is really useful. I also wanted to reiterate the request of kokoko3k regarding CONFIG_TIMER_STATS and powertop. Tuxonice is particularly useful on laptops and powertop is also useful on laptops for monitoring power consumption, so I would like to see this option enabled. Many thanks.

I'm finding it hard to orphan this package. It seems that I'd rather spend my few precious hours of rest working on it than letting it go...

sxe · 2011-11-15 18:36:02

Hey guys,
building linux-pf failes with the new PKGBUILD.
It generated two packages. "-3.1.1-5-x86_64.pkg.tar.xz" and "linux-pf-headers-3.1.1-5-x86_64.pkg.tar.xz".

I'm using the "L" option when asked.

Does anyone know whats going wrong in my system?

nous · 2011-11-15 20:53:29

sxe wrote:

Hey guys,
building linux-pf failes with the new PKGBUILD.
It generated two packages. "-3.1.1-5-x86_64.pkg.tar.xz" and "linux-pf-headers-3.1.1-5-x86_64.pkg.tar.xz".
I'm using the "L" option when asked.
Does anyone know whats going wrong in my system?

Nothing is wrong with your system, I botched the PKGBUILD a little moving to the split package scheme. 3.1.1-6 which I just uploaded should be OK.

sxe · 2011-11-16 09:27:52

Hi nous,
tried it again but still the same behavior.

Arch Linux

#251 2011-11-01 04:35:23

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

#252 2011-11-01 05:49:27

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

#253 2011-11-01 06:15:36

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

#254 2011-11-03 07:57:36

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

#255 2011-11-03 12:19:31

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

#256 2011-11-03 14:19:31

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

#257 2011-11-04 04:07:55

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

#258 2011-11-04 17:17:05

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

#259 2011-11-05 14:11:13

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

#260 2011-11-05 23:16:40

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

#261 2011-11-06 12:07:47

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

#262 2011-11-06 12:21:53

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

#263 2011-11-06 12:58:20

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

#264 2011-11-06 13:03:20

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

#265 2011-11-10 23:04:31

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

#266 2011-11-12 14:45:05

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

#267 2011-11-12 21:06:19

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

#268 2011-11-12 22:33:34

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

#269 2011-11-12 23:01:57

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

#270 2011-11-15 01:05:32

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

#271 2011-11-15 08:24:23

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

#272 2011-11-15 12:55:19

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

#273 2011-11-15 18:36:02

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

#274 2011-11-15 20:53:29

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

#275 2011-11-16 09:27:52

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

Board footer