View previous topic :: View next topic |
Author |
Message |
iphitus Apprentice
Joined: 03 Aug 2005 Posts: 226
|
|
Back to top |
|
|
_sys/sid Guru
Joined: 27 Aug 2004 Posts: 346 Location: Asola (Mantova)
|
Posted: Tue Oct 11, 2005 2:02 pm Post subject: |
|
|
Code: |
Linux nebula 2.6.13-archck8 #1 SMP Tue Oct 11 15:10:44 CEST 2005
i686 Intel(R) Pentium(R) 4 CPU 2.60GHz GenuineIntel GNU/Linux
Linux version 2.6.13-archck8 (root@nebula)
(gcc version 3.3.6 (Gentoo 3.3.6, ssp-3.3.6-1.0, pie-8.7.8))
#1 SMP Tue Oct 11 15:10:44 CEST 2005
|
Thanks... _________________ http://th30z.netsons.org/ |
|
Back to top |
|
|
iphitus Apprentice
Joined: 03 Aug 2005 Posts: 226
|
Posted: Wed Oct 12, 2005 12:16 pm Post subject: |
|
|
_sys/sid wrote: | Code: |
Linux nebula 2.6.13-archck8 #1 SMP Tue Oct 11 15:10:44 CEST 2005
i686 Intel(R) Pentium(R) 4 CPU 2.60GHz GenuineIntel GNU/Linux
Linux version 2.6.13-archck8 (root@nebula)
(gcc version 3.3.6 (Gentoo 3.3.6, ssp-3.3.6-1.0, pie-8.7.8))
#1 SMP Tue Oct 11 15:10:44 CEST 2005
|
Thanks... |
Awesome, good to see people using it.
Anyone else using it? Any problems to report?
iphitus |
|
Back to top |
|
|
human_shape n00b
Joined: 06 Jan 2003 Posts: 43
|
Posted: Wed Oct 12, 2005 4:00 pm Post subject: |
|
|
...coming from nitro-sources I'm using it since the 2.6.13 kernel release and have no problems....I'm very,very satisfied.. you did a good job... |
|
Back to top |
|
|
DocterD Tux's lil' helper
Joined: 15 May 2004 Posts: 129
|
Posted: Wed Oct 12, 2005 4:46 pm Post subject: |
|
|
iphitus wrote: | _sys/sid wrote: | Code: |
Linux nebula 2.6.13-archck8 #1 SMP Tue Oct 11 15:10:44 CEST 2005
i686 Intel(R) Pentium(R) 4 CPU 2.60GHz GenuineIntel GNU/Linux
Linux version 2.6.13-archck8 (root@nebula)
(gcc version 3.3.6 (Gentoo 3.3.6, ssp-3.3.6-1.0, pie-8.7.8))
#1 SMP Tue Oct 11 15:10:44 CEST 2005
|
Thanks... |
Awesome, good to see people using it.
Anyone else using it? Any problems to report?
iphitus |
works great as always |
|
Back to top |
|
|
rcb1974 n00b
Joined: 12 Mar 2003 Posts: 56 Location: Ithaca, NY, USA
|
Posted: Thu Oct 13, 2005 2:41 pm Post subject: Do you have an ebuild for this? |
|
|
Is there an ebuild for this? How hard would it be to make one? I've been using morph-sources but there hasn't been a new release for awhile so I'm looking for a substitute. This one looks promising! |
|
Back to top |
|
|
ericsson n00b
Joined: 08 Oct 2004 Posts: 24
|
Posted: Thu Oct 13, 2005 5:54 pm Post subject: 2.6.13-archck8 |
|
|
Hello,
Works very well here. I have been using it for quite some time without any problems whatsoever. I have tested adding the genetic library and i/o scheduler patches from Jake Moilanens site without any problems. The genetic antipacitory scheduler seems to perform well with reiser4.
Hope you continue
Thanks |
|
Back to top |
|
|
Varuna Tux's lil' helper
Joined: 10 Jun 2004 Posts: 97 Location: Connecticut, United States
|
Posted: Thu Oct 13, 2005 7:27 pm Post subject: Re: Do you have an ebuild for this? |
|
|
rcb1974 wrote: | Is there an ebuild for this? How hard would it be to make one? I've been using morph-sources but there hasn't been a new release for awhile so I'm looking for a substitute. This one looks promising! |
This bug has ebuilds for ArchCK sources. |
|
Back to top |
|
|
baeksu l33t
Joined: 26 Sep 2004 Posts: 609 Location: Seoul, Korea
|
Posted: Fri Oct 14, 2005 8:51 am Post subject: |
|
|
Great patchset. I Absolutely love it. Too early to say there's no bugs for me yet, but at least interactivity is really really good.
I'm abusing my system now (load >6), music doesn't skip, windows get drawn properly, everything is lovely responsive.
Thank you very much! _________________ Gnome:
1. A legendary being.
2. A never ending quest to make unix friendly to people who don't want unix and excruciating for those that do. |
|
Back to top |
|
|
Ignatich n00b
Joined: 14 Sep 2004 Posts: 18
|
Posted: Fri Oct 14, 2005 5:09 pm Post subject: |
|
|
Using this since cko died, very nice and stable patchset, thanks iphitus. But now I'm trying *cough* to migrate to amd64 and I get random crashes seemingly related to reiser4 (they do not happen if I do not mount reiser4 partitions). Any ideas what may be causing this? archck8 + mppe-mppc + routes patch. |
|
Back to top |
|
|
iphitus Apprentice
Joined: 03 Aug 2005 Posts: 226
|
Posted: Fri Oct 14, 2005 10:48 pm Post subject: |
|
|
Ignatich wrote: | Using this since cko died, very nice and stable patchset, thanks iphitus. But now I'm trying *cough* to migrate to amd64 and I get random crashes seemingly related to reiser4 (they do not happen if I do not mount reiser4 partitions). Any ideas what may be causing this? archck8 + mppe-mppc + routes patch. |
How frequently do these occur? This is the first time anyone has reported issues on AMD64 with this reiser4.
Are these issues new with this particular release?
Did you have the problem with any patches before 7.1?
iphitus. |
|
Back to top |
|
|
butters Guru
Joined: 13 May 2002 Posts: 427 Location: Poughkeepsie, NY
|
Posted: Sat Oct 15, 2005 1:49 pm Post subject: |
|
|
What are the main differences between archck and nitro? I'm on an aging 2.6.11-cko5-ck7 kernel and need a migration path.
It's good to see that the arch community is picking up the patchset slack. I distinctly remember when the kernel and hardware section of the gentoo forums would typically have 2 or 3 patchset threads on the first page. _________________ If tugboats were bigger, they'd be the ones getting tugged. |
|
Back to top |
|
|
butters Guru
Joined: 13 May 2002 Posts: 427 Location: Poughkeepsie, NY
|
Posted: Sat Oct 15, 2005 2:48 pm Post subject: |
|
|
butters wrote: | What are the main differences between archck and nitro? I'm on an aging 2.6.11-cko5-ck7 kernel and need a migration path.
It's good to see that the arch community is picking up the patchset slack. I distinctly remember when the kernel and hardware section of the gentoo forums would typically have 2 or 3 patchset threads on the first page. |
From my own comparison, it seems like the nitro patchset is based on the -mm tree and currently uses a git snapshot of 2.6.14-rc3. Otherwise the features are very similar. I'm gunna go with archck because nitro seems to be a little balls-to-the-wall right now. The make oldconfig from my previous cko .config wasn't too bad, either.
I went with the complete preempt, including the BKL. I heard that Con himself was debating the benefits of preemption vs. the overhead of the preemption code. Anyone have any further insite or subjective performance anecdotes? _________________ If tugboats were bigger, they'd be the ones getting tugged. |
|
Back to top |
|
|
a13x Apprentice
Joined: 05 Jan 2005 Posts: 215
|
Posted: Sat Oct 15, 2005 2:53 pm Post subject: |
|
|
wassup iphitus !
I'm using archck8 right now. It seems stable so far. I compiled non-stop for about 4 hours (KDE 3.4.3 + KOffice 1.4.2) without problems. Thx for the release. |
|
Back to top |
|
|
butters Guru
Joined: 13 May 2002 Posts: 427 Location: Poughkeepsie, NY
|
Posted: Sat Oct 15, 2005 3:19 pm Post subject: |
|
|
how does ipw2100 work with this kernel? There doesn't seem to be any support in the kernel, yet remerging ipw2100 externally fails due to an incompatible ieee80211 subsystem:
Code: | * Found sources for kernel version:
* 2.6.13-archck8
* Checking for suitable kernel configuration options:
*
* You kernel source contains an incompatible version of the
* ieee80211 subsystem, which needs to be removed before
* ieee80211-1.0.3 can be installed. This can be accomplished by running:
*
* # rm -i /usr/src/linux/include/net/ieee80211.h
*
* Please note that this will make it impossible to use some of the
* in-kernel IEEE 802.11 wireless LAN drivers (eg. orinoco).
* |
Any hints? _________________ If tugboats were bigger, they'd be the ones getting tugged. |
|
Back to top |
|
|
iphitus Apprentice
Joined: 03 Aug 2005 Posts: 226
|
Posted: Sat Oct 15, 2005 4:34 pm Post subject: |
|
|
From my observation, nitro has become more experimental since when I used to use it back around 2.6.9 with seppe maintaining it. Nowadays releases often have little or no testing, resulting in kernels that occasionally dont compile, and have a range of bugs.
With ArchCK, I have really tried to focus on providing extra features like ricer4, suspend2 and fbsplash, while retaining stability, keeping good working patches going in it. I realise some stuff is more cutting edge than others, and isnt quite stable, but those things can easily be disabled for those who do not wish to have them.
ArchCK also includes unionfs and squashfs, as the ArchCK patchset is used in the Archie LiveCD. Again, easily disabled in your config. Not sure why -nitro, -morph and friends are including them though, as they are of little relevance outside a LiveCD.
As for pre-empt, Con says turn it off. That's right, select 'Server (No Pre-emption)' in menuconfig.
I've run with both for extended periods of time, and really, there's no noticable difference, and any difference I do notice, may well just be placebo. So, there's no point adding more code into the kernel to make it do more, that does nothing.
And ipw2100, For the ieee80211 stack, it will ask you if you want to replace a file, let it, and it will compile and install fine. ipw2100 and ipw2200 both need patches, which apply to the driver, not the kernel, that are available on my website, iphitus.loudas.com/archck.php
iphitus |
|
Back to top |
|
|
Ignatich n00b
Joined: 14 Sep 2004 Posts: 18
|
Posted: Sat Oct 15, 2005 5:32 pm Post subject: |
|
|
iphitus wrote: |
How frequently do these occur? This is the first time anyone has reported issues on AMD64 with this reiser4.
Are these issues new with this particular release?
Did you have the problem with any patches before 7.1?
| Very frequently. Not sure about particular release because I just switched to amd64, so this is the first archck release I tried on it. I began my experiments: manually compiled vanilla 2.6.13 kernel with ck8 and reiser4 patch 2.6.13-2 from namesys, added acpi and libaata fixes, vesafb-tng, squashfs, shfs from your patch. It locked sometimes but much less frequently. Tried 2.6.14-rc4 with namesys reiser4 2.6.13-2 (i saw ppl complaining about version from rc2-mm2), ck, squashfs, shfs = no luck, opps at boot in reiser4 code. Friend suggested me 2.6.12-ck2, tried it with manually patched forcedeth driver + shfs + squashfs = opps at boot in reiser4 code. Tried to fsck my partition. All clean, still opps at boot. Removing shfs and squashfs from that kernel helped. Running 2.6.12-ck2 + routes + mppe-mppc + patched forcedeth w/o any problems now.
Conclusion: something really strange happens on amd64 with shfs and squashfs compiled in.
Btw preempt and swap prefetch was always set to off. |
|
Back to top |
|
|
iphitus Apprentice
Joined: 03 Aug 2005 Posts: 226
|
Posted: Sun Oct 16, 2005 1:09 am Post subject: |
|
|
You shouldnt be compiling squashfs, its irrelevant.
But neither of them make any changes that will cause instability in the kernel. /me thinks it's just placebo.
Are you using custom cflags on the kernel? This will break it.
Are you sure it's the kernel causing it? Nobody on the R4, CK, Suspend2, lists, nor some of my amd64 users have had any problems.
iphitus |
|
Back to top |
|
|
butters Guru
Joined: 13 May 2002 Posts: 427 Location: Poughkeepsie, NY
|
Posted: Sun Oct 16, 2005 8:32 am Post subject: |
|
|
iphitus wrote: | As for pre-empt, Con says turn it off. That's right, select 'Server (No Pre-emption)' in menuconfig.
I've run with both for extended periods of time, and really, there's no noticable difference, and any difference I do notice, may well just be placebo. So, there's no point adding more code into the kernel to make it do more, that does nothing.
And ipw2100, For the ieee80211 stack, it will ask you if you want to replace a file, let it, and it will compile and install fine. ipw2100 and ipw2200 both need patches, which apply to the driver, not the kernel, that are available on my website, iphitus.loudas.com/archck.php
iphitus |
Thanks for the info on preemption. I wonder if kernel preemption is undesirable for desktops in theory, or only in this particular style of implementation. Maybe the staircase proc scheduler is good enough at preempting userspace that it is no longer necessary to preempt the kernel. I also assume that kernel preemption becomes less advantageous when there are multiple hardware threads available, and this is the apparent trend.
After removing the ieee80211 kernel header file, ipw2100 (1.1.2) compiled fine without the patch. This bothers me on principle (because now the builtin 802.11 modules won't work), but I guess we'll have things like this until ipw2100 and the updated ieee80211 subsystem get merged into mainline. The ebuild (1.1.2-r3) applies the following patches for kernel versions 2.6.12 or newer, which seem to fix what your patch does, plus some more stuff:
ipw2100-1.1.2-header_data-fix.patch
ipw2100-1.1.2-pci_link-fix.patch
ipw2100-1.1.2-suspend2.patch
ipw2100-1.1.2-wext-crash.patch
Thanks for your help, rebooting now... _________________ If tugboats were bigger, they'd be the ones getting tugged. |
|
Back to top |
|
|
butters Guru
Joined: 13 May 2002 Posts: 427 Location: Poughkeepsie, NY
|
Posted: Sun Oct 16, 2005 9:01 am Post subject: |
|
|
Working great now, very snappy. I can say for sure it's at least as responsive as my previous 2.6.11-cko5-ck7. It seems to handle memory better, too. Good stuff.
Looking forward to 2.6.14-archck1... _________________ If tugboats were bigger, they'd be the ones getting tugged. |
|
Back to top |
|
|
iphitus Apprentice
Joined: 03 Aug 2005 Posts: 226
|
Posted: Sun Oct 16, 2005 10:42 pm Post subject: |
|
|
ipw2x00 is in the vanilla kernel as of 2.6.14-rc1
expect a 2.6.14-rcX-archie closer to the 2.6.14 release |
|
Back to top |
|
|
lucke n00b
Joined: 11 May 2005 Posts: 35
|
Posted: Mon Oct 17, 2005 6:59 am Post subject: |
|
|
AFAIK, rc4 is meant to be the last rc before final, so you'd better hurry, iphitus |
|
Back to top |
|
|
iphitus Apprentice
Joined: 03 Aug 2005 Posts: 226
|
Posted: Mon Oct 17, 2005 7:16 am Post subject: |
|
|
lucke wrote: | AFAIK, rc4 is meant to be the last rc before final, so you'd better hurry, iphitus |
I might just wait for final then |
|
Back to top |
|
|
Ignatich n00b
Joined: 14 Sep 2004 Posts: 18
|
Posted: Mon Oct 17, 2005 7:37 am Post subject: |
|
|
iphitus wrote: | You shouldnt be compiling squashfs, its irrelevant.
But neither of them make any changes that will cause instability in the kernel. /me thinks it's just placebo.
Are you using custom cflags on the kernel? This will break it.
Are you sure it's the kernel causing it? Nobody on the R4, CK, Suspend2, lists, nor some of my amd64 users have had any problems.
iphitus | Well, I wanted to experiment with portage-in-squashfs. No custom cflags, and yes, very slim possibility of hardware problem because I booted from windows partition and stress tested my system with S&M, Super PI and Prime 95. And it worked just fine until I mount reiser4 drive (true for archck8 at least).
I have no idea how squash and shfs can relate to reiser4 problems, but it looks like they do in some obscure way. Anyway it's good that nobody else experiencing such problems, it took me some time to make it work. |
|
Back to top |
|
|
butters Guru
Joined: 13 May 2002 Posts: 427 Location: Poughkeepsie, NY
|
Posted: Mon Oct 17, 2005 12:05 pm Post subject: |
|
|
iphitus wrote: | lucke wrote: | AFAIK, rc4 is meant to be the last rc before final, so you'd better hurry, iphitus |
I might just wait for final then |
yeah, that's what I was implying. The new kernel development cycle changes make things a little more predictable. The rc's for 2.6.14 have been coming at 1 week intervals, with the last one coming 6 days ago. Linus merged davem's SPARC branch (supposed to be the only large change allowed after the two-week submission window) and Andrew fixed an obscure build conflict involving ipv6 and netfilter a couple days ago, so I wouldn't be surprised to see 2.6.14 final before wednesday.
Only 4 .y kernels for 2.6.13, that's certainly an improvement!! A combination of increasing stability and shortening development cycles.
It's a small world... nice to see you here, lucke, and thanks again for hosting archbootstrap (and removing the trailing whitespace). I did my Arch install off the Damn Small Linux LiveCD, worked perfectly. _________________ If tugboats were bigger, they'd be the ones getting tugged. |
|
Back to top |
|
|
|