Zen-Kernel Discussion/Support Thread - Part 4

[kernelOfTruth]

it's my own day-to-day experience and reading the comments & reports from the devs and users on lkml

And yet...
https://bbs.archlinux.org/viewtopic.php?id=36019
http://www.webhostingtalk.com/showthread.php?t=913888

+1 for SLQB, if only someone ported the patch to 2.6.37...

currently I'm trying a heavy-patched 2.6.37 with SLUB and it seems to work fine - I need to try it out against SLAB (as a reference)

but it seems to have improved nicely since the last time I tested it

yeah, SLQB really was a kind of light-weight juggernaut - working perfectly under heavy load and improving responsiveness

hopefully Nick will find some time to update it - in the meantime he has come up with lots of amazing improvements:

*) Quick merge window note..

some of the performance numbers are pretty stunning. For example, a
hot-cache "find . -size" on my home directory (which basically just
does name lookups to get the stat information for every file
recursively) became 35% faster. And that's the _unthreaded_ case. Not
some odd high-end scalability thing, and not some recompiled binary
taking advantage of new facilities. Pathname lookup is just simply
faster.

*) [patch 0/8] Inode data integrity patches

only to name some ...

[ponciarello]

FYI, I got this since the latest bfs merge

Code: Select all

In file included from kernel/sched.c:2:0:
kernel/sched_bfs.c:2011:6: error: conflicting types for 'calc_global_load'
include/linux/sched.h:152:13: note: previous declaration of 'calc_global_load' was here

thanks for the work on 2.6.37-zen too

[kernelOfTruth]

FYI, I got this since the latest bfs merge

Code: Select all

In file included from kernel/sched.c:2:0:
kernel/sched_bfs.c:2011:6: error: conflicting types for 'calc_global_load'
include/linux/sched.h:152:13: note: previous declaration of 'calc_global_load' was here

thanks for the work on 2.6.37-zen too

take a look at Ralph Ulrich's comment on Con's blog

[ponciarello]

thanks kOT, tried that and it builds, but got a (dunno if related) bad hang soon while playing Enemy Territory (nvidia 256.53)

as a side note, 2.6.37-ck1 works fine (but I haven't tried it on any intel gpu yet).

[jbouzan]

I'm trying to get the git version of Zen, but checkouts always seem to fail. I tried these each about 5 minutes apart.

Code: Select all

amelia src # git clone --depth 1 git://zen-kernel.org/kernel/zen.git linux-2.6-zen-git
Cloning into linux-2.6-zen-git...
fatal: read error: Connection reset by peer
amelia src # git clone --depth 1 git://zen-kernel.org/kernel/zen.git linux-2.6-zen-git
Cloning into linux-2.6-zen-git...
fatal: read error: Connection reset by peer
amelia src # layman -a THE
* Running... # /usr/bin/git clone git://zen-sources.org/zen/THE.git /var/lib/layman/THE
Cloning into /var/lib/layman/THE...
zen-sources.org[0: 8.5.1.1]: errno=Connection timed out
fatal: unable to connect a socket (Connection timed out)
* Failed to add overlay "THE".
* Error was: Adding overlay "THE" failed!
amelia src # git clone --depth 1 git://zen-kernel.org/kernel/zen.git linux-2.6-zen-git
Cloning into linux-2.6-zen-git...
fatal: read error: Connection reset by peer

Does anyone have any clue why it always does this for me? The same error has happened when I've tried to add the THE overlay before, several times in weeks.

[kernelOfTruth]

I'm trying to get the git version of Zen, but checkouts always seem to fail. I tried these each about 5 minutes apart.

Code: Select all

amelia src # git clone --depth 1 git://zen-kernel.org/kernel/zen.git linux-2.6-zen-git
Cloning into linux-2.6-zen-git...
fatal: read error: Connection reset by peer
amelia src # git clone --depth 1 git://zen-kernel.org/kernel/zen.git linux-2.6-zen-git
Cloning into linux-2.6-zen-git...
fatal: read error: Connection reset by peer
amelia src # layman -a THE
* Running... # /usr/bin/git clone git://zen-sources.org/zen/THE.git /var/lib/layman/THE
Cloning into /var/lib/layman/THE...
zen-sources.org[0: 8.5.1.1]: errno=Connection timed out
fatal: unable to connect a socket (Connection timed out)
* Failed to add overlay "THE".
* Error was: Adding overlay "THE" failed!
amelia src # git clone --depth 1 git://zen-kernel.org/kernel/zen.git linux-2.6-zen-git
Cloning into linux-2.6-zen-git...
fatal: read error: Connection reset by peer

Does anyone have any clue why it always does this for me? The same error has happened when I've tried to add the THE overlay before, several times in weeks.

zen.git hasn't been updated in 5 months

you'd rather use: git://zen-kernel.org/kernel/zen-stable.git

[Ormaaj]

git seems to have been down for several days. I've switched to liquorix 2.6.37 now which seems alright so far. IDK why zen insists on running it's own git server though... gitorious/github are so much more reliable.

[jbouzan]

I'm trying to get the git version of Zen, but checkouts always seem to fail. I tried these each about 5 minutes apart.

Code: Select all

amelia src # git clone --depth 1 git://zen-kernel.org/kernel/zen.git linux-2.6-zen-git
Cloning into linux-2.6-zen-git...
fatal: read error: Connection reset by peer
amelia src # git clone --depth 1 git://zen-kernel.org/kernel/zen.git linux-2.6-zen-git
Cloning into linux-2.6-zen-git...
fatal: read error: Connection reset by peer
amelia src # layman -a THE
* Running... # /usr/bin/git clone git://zen-sources.org/zen/THE.git /var/lib/layman/THE
Cloning into /var/lib/layman/THE...
zen-sources.org[0: 8.5.1.1]: errno=Connection timed out
fatal: unable to connect a socket (Connection timed out)
* Failed to add overlay "THE".
* Error was: Adding overlay "THE" failed!
amelia src # git clone --depth 1 git://zen-kernel.org/kernel/zen.git linux-2.6-zen-git
Cloning into linux-2.6-zen-git...
fatal: read error: Connection reset by peer

Does anyone have any clue why it always does this for me? The same error has happened when I've tried to add the THE overlay before, several times in weeks.

zen.git hasn't been updated in 5 months

you'd rather use: git://zen-kernel.org/kernel/zen-stable.git

Thank you for that. Still not working, but as Ormaaj says it's down for him too I'll try tomorrow.

Code: Select all

src # git clone --depth 1 git://zen-kernel.org/kernel/zen-stable.git linux-2.6-zen
Cloning into linux-2.6-zen...
fatal: read error: Connection reset by peer

[ponciarello]

alternatively you can get a snapshot from cgit

Code: Select all

wget http://git.zen-kernel.org/zen-stable/snapshot/zen-stable-master.tar.bz2

[wrc1944]

Just saw this on lwn.net kernel patch page. http://lwn.net/Articles/422333

Yet another scheduler. Anyone taken a look at this? I looked at the links provided on the page, but still am still basically unsure of what this one is all about. The name sounds cool "sched: QFQ - quick fair queue scheduler."

After all, "quick" must be better, right?

[Ant P.]

From reading the link they provided it sounds like it'd be useful on a media server streaming live content due to the low packet latency.

[kernelOfTruth]

when you guys are updating / bumping to 2.6.37

kztmem: http://marc.info/?l=linux-kernel&m=129537135109667&w=2

[ponciarello]

tnx for the git updates

tried them on x86_64 and all ok, but on a 32bit host I got

Code: Select all

  CC      arch/x86/kernel/asm-offsets.s
In file included from include/linux/suspend.h:8:0,
                 from arch/x86/kernel/asm-offsets_32.c:11,
                 from arch/x86/kernel/asm-offsets.c:2:
include/linux/mm.h:515:2: error: #error SECTIONS_WIDTH+NODES_WIDTH+ZONES_WIDTH > BITS_PER_LONG - NR_PAGEFLAGS
make[1]: *** [arch/x86/kernel/asm-offsets.s] Error 1
make: *** [prepare0] Error 2

[tranquilcool]

tnx for the git updates

tried them on x86_64 and all ok, but on a 32bit host I got

Code: Select all

  CC      arch/x86/kernel/asm-offsets.s
In file included from include/linux/suspend.h:8:0,
                 from arch/x86/kernel/asm-offsets_32.c:11,
                 from arch/x86/kernel/asm-offsets.c:2:
include/linux/mm.h:515:2: error: #error SECTIONS_WIDTH+NODES_WIDTH+ZONES_WIDTH > BITS_PER_LONG - NR_PAGEFLAGS
make[1]: *** [arch/x86/kernel/asm-offsets.s] Error 1
make: *** [prepare0] Error 2

same errors here with 32bit

[desultory]

Split off "[topic=860915]Fortress Linux evangelism[/topic]", as it seemed wildly off topic.

[da_monumental_1]

Was AUFS removed?

[kernelOfTruth]

Was AUFS removed?

probably it wasn't added yet

tuxonice also is missing

and keep in mind - it's still 2.6.37-zen0

[Ant P.]

Anyone else using zram on 2.6.37? Does it run for more than 5 minutes without an Oops now?

[da_monumental_1]

Was AUFS removed?

probably it wasn't added yet

tuxonice also is missing

and keep in mind - it's still 2.6.37-zen0

I patched 2.6.37 with aufs (git standalone) and tuxonice (two line reject I had to fix). Everything seems to be working OK.

Kztmem sounds like fun, but I couldn't get a completed kernel compile. I'll play around with it later.

[Kollin]

When 2.6.37-zen* is planned to hit portage ?

[kernelOfTruth]

When 2.6.37-zen* is planned to hit portage ?

probably when some other (outstanding) features have been merged

zen now includes inode integrity patches from Nick Piggin !

Thanks guys !

[kernelOfTruth]

Anyone else using zram on 2.6.37? Does it run for more than 5 minutes without an Oops now?

yeah, seems to work fine (and stable) here

with kztmem


I tried out both the zram provided by zen and the vanilla one from 2.6.37 - both worked for me

[kernelOfTruth]

@zen devs:

please check the last merge of the inode integrity patches - especially patch 8 (fs: add i_op->sync_inode)

it either seems to be missing or didn't merge correctly (in my humble opinion)

thanks !

[benneque]

Hi!

Today I visited zen-kernel.org and recognized the statement @ frontpage.

Now... can somebody explain what exactly that means? (My English doesn't seem good enough to get it... )

I don't know what Zen tag is ?!

Is it now "better" to use zen-sources-9999 , because there won't be much "stable" updates in future ?

[albright]

Does that (see previous post) mean that zen won't be going
into portage anymore? (Well, it's almost like that already )