Old posts from the morph-sources thread

Message

nightfrost · Post by **nightfrost** » Mon May 23, 2005 7:13 pm

I'm still missing /proc/acpi/button, otherwise everything else seems to work fine. Any ideas what the reason can be? Here's the relevant part of .config

Code: Select all

#
# ACPI (Advanced Configuration and Power Interface) Support
#
CONFIG_ACPI=y
CONFIG_ACPI_BOOT=y
CONFIG_ACPI_INTERPRETER=y
CONFIG_ACPI_SLEEP=y
CONFIG_ACPI_SLEEP_PROC_FS=y
# CONFIG_ACPI_SLEEP_PROC_SLEEP is not set
CONFIG_ACPI_AC=y
CONFIG_ACPI_BATTERY=y
CONFIG_ACPI_BUTTON=y
CONFIG_ACPI_VIDEO=y
CONFIG_ACPI_FAN=y
CONFIG_ACPI_PROCESSOR=y
CONFIG_ACPI_THERMAL=y
CONFIG_ACPI_ASUS=m
CONFIG_ACPI_IBM=m
CONFIG_ACPI_TOSHIBA=m
# CONFIG_ACPI_CUSTOM_DSDT is not set
CONFIG_ACPI_BLACKLIST_YEAR=0
# CONFIG_ACPI_DEBUG is not set
CONFIG_ACPI_BUS=y
CONFIG_ACPI_EC=y
CONFIG_ACPI_POWER=y
CONFIG_ACPI_PCI=y
CONFIG_ACPI_SYSTEM=y
# CONFIG_X86_PM_TIMER is not set
# CONFIG_ACPI_CONTAINER is not set

yardbird · Post by **yardbird** » Mon May 23, 2005 9:54 pm

omidxo wrote:morph-sources 2.6.11-morph9 is working well,but one messages here:

Code: Select all

Enabling fast FPU save and restore... done.
Enabling unmasked SIMD FPU exception support... done.
Checking 'hlt' instruction... OK.
    ACPI-0498: *** Warning: Encountered executable code at module level, [AE_NOT_CONFIGURED]
    ACPI-0498: *** Warning: Encountered executable code at module level, [AE_NOT_CONFIGURED]
    ACPI-0498: *** Warning: Encountered executable code at module level, [AE_NOT_CONFIGURED]
CPU0: Intel(R) Celeron(R) CPU 2.53GHz stepping 03
per-CPU timeslice cutoff: 731.40 usecs.

How to fix it?

You may want to try to fix your DSDT:

http://forums.gentoo.org/viewtopic-t-12 ... ommon.html

Nightfrost wrote:I'm still missing /proc/acpi/button, otherwise everything else seems to work fine. Any ideas what the reason can be?

This (as well as the message above) could be related to the updated ACPI patches. A new ACPI patch has just come out, I'll put it into next morph to see if it solves. Otherwise we can try to reverse it (unfortunately the "old" patch has been removed and we cannot reverse it right now).

EDIT: Whoops, too fast, the old patch is still there and has not been updated. Can you guys try to reverse it?

http://www.kernel.org/pub/linux/kernel/ ... 1.diff.bz2

nightfrost · Post by **nightfrost** » Tue May 24, 2005 6:04 am

Yep, that was it

Works just fine now. Thanks.

k_max · Post by **k_max** » Tue May 24, 2005 9:50 am

2 yardbird
could you please include swsusp 2.1.8.9 (or 2.1.9 when it comes out) into the next release? the swapwriter is no longer deprecated since 2.1.8.8 and it's faster than the filewriter...

omidxo · Post by **omidxo** » Tue May 24, 2005 1:00 pm

yardbird wrote:

omidxo wrote:morph-sources 2.6.11-morph9 is working well,but one messages here:

Code: Select all

Enabling fast FPU save and restore... done.
Enabling unmasked SIMD FPU exception support... done.
Checking 'hlt' instruction... OK.
    ACPI-0498: *** Warning: Encountered executable code at module level, [AE_NOT_CONFIGURED]
    ACPI-0498: *** Warning: Encountered executable code at module level, [AE_NOT_CONFIGURED]
    ACPI-0498: *** Warning: Encountered executable code at module level, [AE_NOT_CONFIGURED]
CPU0: Intel(R) Celeron(R) CPU 2.53GHz stepping 03
per-CPU timeslice cutoff: 731.40 usecs.

How to fix it?

You may want to try to fix your DSDT:

http://forums.gentoo.org/viewtopic-t-12 ... ommon.html

There is a bug-report about acpi-20050408 as http://sourceforge.net/mailarchive/foru ... m_id=35685

yardbird · Post by **yardbird** » Tue May 24, 2005 2:53 pm

k_max wrote:2 yardbird
could you please include swsusp 2.1.8.9 (or 2.1.9 when it comes out) into the next release? the swapwriter is no longer deprecated since 2.1.8.8 and it's faster than the filewriter...

Sure.

yardbird · Post by **yardbird** » Tue May 24, 2005 2:54 pm

nightfrost wrote:Yep, that was it Works just fine now. Thanks.

omidxo wrote:There is a bug-report about acpi-20050408 as http://sourceforge.net/mailarchive/foru ... m_id=35685

If anyone hasn't a better idea I'll try to put ACPI as optional USE flag in the next release, so you can choose whether you want the updated patch or not.

yardbird · Post by **yardbird** » Tue May 24, 2005 2:58 pm

ahorn wrote:I tried certain patchsets in the last days, but morph-sources are the best mix! i never went in trouble with your stuff, that's the reason why i call your kernel stable. so here's another point, why don't include the use-flags in the kernel with a kernel menu, possibly with some more unstable/dangerous things like timeslices? also, it would be nice, if you include some mm-stuff or whatever which makes the desktop runs faster, not as standard, but as an option (maybe with a second menu).

Urgh! Sorry for the late reply, I missed your post

Certainly it would be nice to do that, but this involves some dirty hacking of the kernel source. You should first place an entry in the menus (Kconfig files) and then place _big_ #ifdef's in the files affected by the patches, so that the support for certain features is compiled in only if the feature is enabled in kernel config. It'd take a _lot_ of time and would be a very very error-prone process. IMHO it is not worth to do such thing.

chouimat · Post by **chouimat** » Wed May 25, 2005 2:41 pm

any plan to include Xen into the patch?

yardbird · Post by **yardbird** » Wed May 25, 2005 3:24 pm

chouimat wrote:any plan to include Xen into the patch?

As I wrote on the website some time ago, it does not make much sense to include Xen into morph-sources. Mainly because a Xen-enabled kernel is a minimal kernel which allows you to run other kernels on top of it (including morph-sources

). So the features of morph-sources would not make much sense in this situation (maybe only the scheduler patches would be worth including). Also, many features of morph-sources would simply not be available in a Xen kernel, which includes just the stuff to get the machine booted + the virtual machine layer.

mrfree · Post by **mrfree** » Wed May 25, 2005 5:09 pm

I'm trying to use schedtool but...

Code: Select all

# schedtool -I 11626
ERROR: could not set PID 11626 to I: SCHED_ISO - value out of range / policy not implemented

Does morph-sources have SCHED_BATCH and SCHED_ISO patch (This is distinctive feature of ck-source)?

Edit: I think you should apply, with staircase useflag, schedbatch and schediso too

chouimat · Post by **chouimat** » Thu May 26, 2005 1:36 am

I asked about Xen because I need a xen kernel with reiser4 ... ok thanks

baeksu · Post by **baeksu** » Fri May 27, 2005 9:32 am

mrfree wrote:I'm trying to use schedtool but...
Code: Select all
# schedtool -I 11626
ERROR: could not set PID 11626 to I: SCHED_ISO - value out of range / policy not implemented
Does morph-sources have SCHED_BATCH and SCHED_ISO patch (This is distinctive feature of ck-source)?

Edit: I think you should apply, with staircase useflag, schedbatch and schediso too

metoo!

Now I had to apply the three sched-batches myself, and I don't think I did a clean job with that. Is there any particular reason those patches weren't included? They're not too ricerish, are they? ~gasp~.

Combined with schedtool they give a nice way of really putting some processes (like file archiving and merging) in the background where they don't interfere with desktop use.

By the way, this is the first time I'm trying out the morph-sources. Seems quite nice. I'm happy being able to use the newest staircase with swsusp2 without having to go to the 2.6.12-rc* kernels.

Those kernels seem to break a *lot* of stuff (like all pre-1.0.7* nvidia-kernels and swsusp2). Frankly I'm a little curious why nitro and love sources insist on releasing patches for .12. I'd tried patching vanilla kernels myself to get the options I wanted, but it's hella difficult to add swsusp2 when you don't know what you're doing.

Morph gives all the new patches I need while letting me stick to the .11-series. Thank you for that, and keep doing the great job!

El_Goretto · Post by **El_Goretto** » Fri May 27, 2005 9:52 am

Hi, glad to see another morph release

I just want to know: is the kernel base still 2.6.11.8?
Is it planned to upgrade to 2.6.11.10 for the next realease?

It's just that I'm worrying about the severe security issues of the < 2.6.11.10...

yardbird · Post by **yardbird** » Sat May 28, 2005 12:41 am

Sched-tools and the other ck-goodies will be in the next morph-sources as USE flags.

2.6.11-morph9 is based on 2.6.11.10, so it has the security fixes.

I've been quite busy transferring the site from sourceforge to my home server, but tomorrow or at most sunday I will release 2.6.11-morph10. Bye!

mrfree · Post by **mrfree** » Sat May 28, 2005 11:48 am

yardbird wrote:Sched-tools and the other ck-goodies will be in the next morph-sources as USE flags.

Ohohohohohoh yes!!!

nightfrost · Post by **nightfrost** » Sat May 28, 2005 3:30 pm

I'm not able to compile ipw2200-drivers against the latest morph-sources. Anyone else having this problem? (Note: I haven't had time to try this against vanilla, so can't say for sure whether it's morph or vanilla).

EDIT: Sorry, I was too fast there. I don't seem to be able to compile it against any kernel now, including those I've been comliling it against before. Problem is elsewhere..

nightfrost · Post by **nightfrost** » Sun May 29, 2005 3:09 pm

Lately, I've been trying a bunch of different distros - mostly in order feel at home with the most common distros since I constantly encourage people to try out linux. Anyway, long story short; I just tried out Ubuntu and was surprised to find out that the old problem of losing drm after a resume from suspend-to-ram wasn't there. I assume the ubuntu guys have some sort of a patch for their kernel - I haven't been able to find out anything about it but I'm looking around. Does anyone know anything about this? And I'm wondering if yardbird could possibly imagine adding any such patch to morph? Cheers.

yardbird · Post by **yardbird** » Mon May 30, 2005 10:46 pm

Hello,

2.6.11-morph10 is out. Changelog is at the website. I've added schedbatch and schediso from -ck (let me know how it goes!), and made ACPI updated patch optional. Plus the usual updates. Soon a 2.6.12-rcX morph patch will appear (I've been testing it for some days).

Cheers

yardbird · Post by **yardbird** » Mon May 30, 2005 10:47 pm

nightfrost wrote:Lately, I've been trying a bunch of different distros - mostly in order feel at home with the most common distros since I constantly encourage people to try out linux. Anyway, long story short; I just tried out Ubuntu and was surprised to find out that the old problem of losing drm after a resume from suspend-to-ram wasn't there. I assume the ubuntu guys have some sort of a patch for their kernel - I haven't been able to find out anything about it but I'm looking around. Does anyone know anything about this? And I'm wondering if yardbird could possibly imagine adding any such patch to morph? Cheers.

Are you using swsusp2 or "vanilla" swsusp on Ubuntu? I'm using swsusp2 and DRI is just fine after a resume from disk (suspen to RAM does not work, even in Windows

). If this is a swsusp2 bug I can report it to the list.

El_Goretto · Post by **El_Goretto** » Tue May 31, 2005 7:59 am

Ok, i'll try morph10 in a couple of minutes. But I noticed that Win4Lin was gone in morph9 :'(
Was it on purpose?

--
edit: same on morph10, no win4lin, erf.

nightfrost · Post by **nightfrost** » Tue May 31, 2005 9:23 am

yardbird wrote:
nightfrost wrote:Lately, I've been trying a bunch of different distros - mostly in order feel at home with the most common distros since I constantly encourage people to try out linux. Anyway, long story short; I just tried out Ubuntu and was surprised to find out that the old problem of losing drm after a resume from suspend-to-ram wasn't there. I assume the ubuntu guys have some sort of a patch for their kernel - I haven't been able to find out anything about it but I'm looking around. Does anyone know anything about this? And I'm wondering if yardbird could possibly imagine adding any such patch to morph? Cheers.
Are you using swsusp2 or "vanilla" swsusp on Ubuntu? I'm using swsusp2 and DRI is just fine after a resume from disk (suspen to RAM does not work, even in Windows ). If this is a swsusp2 bug I can report it to the list.

Sorry, I should've given more info. I'm using the ubuntu stock kernel, and it's not swsusp/swsusp2. It is actually suspend-to-ram via "echo mem > /sys/power/state". I have a Fujitsu-Siemens lifebook p7010 (it has a forum here: http://www.leog.net/fujp_forum/forum.asp?FORUM_ID=8, if you'd like to see that), and everyone seems to have problem with DRI after resume. And, like said, I was surprised to find out that I didn't have this problem under ubuntu.

Actually, the loss of DRI is there after a swsusp2 resume as well, but I think this is due to the i915-module. Which module do you use?

Anyway, I'm gonna try out the new patch set now

Thanks again for your good work by the way!

baeksu · Post by **baeksu** » Tue May 31, 2005 9:29 am

I tried morph10. Swsusp2 didn't work very well. On suspension, it gave me some error messages (can't find them in logs, though). Resuming from suspension failed once. Successful resumes were followed by hardlocks after a few minutes.

I used the staircase and acpi patches. Removing the acpi patch didn't remove the problems (though it did remove my powerbutton). Maybe the swsusp 2.1.8.9 caused the problems.

Right now I'm using morph9 with the staircase patch from -10. Tried to suspend a couple of times, everything works fine.

[edit]Forgot to thank for the patchset. Thank you[/edit]

mrfree · Post by **mrfree** » Tue May 31, 2005 9:29 am

Wonderful yardbird

Using morph10 now swsusp2 works well (but I need to use alsa 1.0.9 in portage instead of 1.0.8 integrated into the kernel) and I enjoy with the performace introduced with the ck goodies too... good work!

Ciao ciao

dan2003 · Post by **dan2003** » Tue May 31, 2005 10:32 am

I tried morph9 and found sws2 didnt work, complined about lack of space, i then tried the file writer support but tt didnt work either, cant remember why but i'm still using morph 6 which works very well, had an uptime of nearly 2 weeks b4 i tried morph 9 the other day

Edit : thats using suspend at least twice a day every day