You are not logged in.

#26 2009-05-14 13:01:33

Tera
Member
From: Finland
Registered: 2007-01-25
Posts: 81

Re: kernel26-one-dev development thread

RayNbow wrote:
psch wrote:

I do have trouble with cpufreq, too.
Edit: something weird i just noticed: The announcement says, WiFi LED support has completely been dropped, yet they are functional for me.

I haven't upgraded to this one-dev version, but I think Tera means he has removed the LED patches since 2.6.30 now supports the WiFi LED out of the box:

Correct. I update the announcement.

Offline

#27 2009-05-14 14:16:15

Tera
Member
From: Finland
Registered: 2007-01-25
Posts: 81

Re: kernel26-one-dev development thread

psch wrote:

I do have trouble with cpufreq, too.

According to:

$ zcat /proc/config.gz | grep -i cpufreq
# CPUFreq processor drivers
# CONFIG_X86_ACPI_CPUFREQ is not set
# CONFIG_X86_CPUFREQ_NFORCE2 is not set

It seems to simply be not enabled. Unless of course I'm looking in the wrong spot.
Anyway, any pointers would be appreciated.

Add: I did compile the kernel myself, via the PKGBUILD on my desktop system, in case that could be relevant.

Add2: recompiling the kernel with

CONFIG_X86_ACPI_CPUFREQ=y

set in the config resolved the cpufreq issue. Seems like the uploaded config in the tarball wasn't correct.

Thanks for noticing! I am already preparing a fixed release.

Offline

#28 2009-05-15 15:03:51

Zom
Member
From: Sweden
Registered: 2007-10-27
Posts: 430

Re: kernel26-one-dev development thread

I went back to my older solution for suspending the system when having files on a SDHC-card. Don't mount the card as /home, instead store all .-files there and symlink other folders to it.

$ ls -al ~
.gnome/
.gconf/
.whatever/
Pictures -> /media/sdhc/Pictures
Downloads -> /media/sdhc/Downloads
Notes -> /media/sdhc/Notes

Then add this script to /etc/pm/sleep.d/75sdhc

#!/bin/bash
case "$1" in
        suspend)
                umount /dev/mmcblk0p1
                ;;
        resume)
                ;;
        *)
                ;;
esac

and chmod +x it. You can add whatever actions or scripts you want to run there too, or in a separate file. It should be quite easy to understand.

This is more of a workaround than a solution, but if you're like me, then this is really the only way I got it to work somewhat. Obviously, this way won't reduce all wear-and-tear on the SSD, and apps like firefox will be a bit slower. Also, if you use kdm, you might want to change the default path from .xsession-errors to /var/log/xsession-errors since it could get quite large otherwise. Check in /usr/share/config/kdm/kdmrc.

Last edited by Zom (2009-05-15 15:04:47)

Offline

#29 2009-05-16 06:32:34

Tera
Member
From: Finland
Registered: 2007-01-25
Posts: 81

Re: kernel26-one-dev development thread

Zom wrote:

I went back to my older solution for suspending the system when having files on a SDHC-card.

Does that mean suspend works for you now with one-dev?

Offline

#30 2009-05-16 06:48:58

Tera
Member
From: Finland
Registered: 2007-01-25
Posts: 81

Re: kernel26-one-dev development thread

tom_g wrote:

One request I have for future versions is to include CONFIG_BT_RFCOMM_TTY=y for rfcomm serial connections (eg. for accessing a dial-up service (GPRS/HSDPA) through a bluetooth enabled phone).

Hey, can you try modprobing a module called "rfcomm" and see does it do the trick for you? Because the driver you are asking me to compile statically, is in fact already available as a module. I have used GPRS/HSDPA through a Bluetooth enabled phone via pand just fine with the drivers already available on the kernel.

Offline

#31 2009-05-16 07:30:15

Tera
Member
From: Finland
Registered: 2007-01-25
Posts: 81

Re: kernel26-one-dev development thread

I don't have access to my laptop right now so I compiled the new kernel with CPU frequency fix and rc6 on my server remotely. I uploaded it to http://dolly.rulex.net/~giga/arch/kerne … pkg.tar.gz. Can someone install it and test it works before I make an official announcement? Thanks.

Last edited by Tera (2009-05-16 07:30:51)

Offline

#32 2009-05-16 08:26:29

psch
Member
Registered: 2009-05-14
Posts: 4

Re: kernel26-one-dev development thread

Tera wrote:

Can someone install it and test it works before I make an official announcement? Thanks.

As far as I can tell, everything seems to be working. Suspend (via s2ram -f) works fine for me, just like it used to, cpufreq works, acerhdf works, KMS work fine, too.
I can't comment on suspend and corrupted filesystems on storage expension as I don't have an expansion card, though.

Offline

#33 2009-05-16 20:35:22

bulat
Member
From: Kazan, Russia
Registered: 2008-05-26
Posts: 18

Re: kernel26-one-dev development thread

It would be great if you add support for external USB devices as modules.
Also, what do you think about adding a patch to ext2 compression support (http://e2compr.sourceforge.net/)? It needs some code fixing to support new kernels. Or maybe you can suggest ohter FS with compression support?

Offline

#34 2009-05-18 01:48:27

tom_g
Member
Registered: 2009-05-12
Posts: 3

Re: kernel26-one-dev development thread

Tera wrote:
tom_g wrote:

One request I have for future versions is to include CONFIG_BT_RFCOMM_TTY=y for rfcomm serial connections (eg. for accessing a dial-up service (GPRS/HSDPA) through a bluetooth enabled phone).

Hey, can you try modprobing a module called "rfcomm" and see does it do the trick for you? Because the driver you are asking me to compile statically, is in fact already available as a module. I have used GPRS/HSDPA through a Bluetooth enabled phone via pand just fine with the drivers already available on the kernel.

You have rfcomm included, but not TTY support. I'm not using pand, just rfcomm & pppd, so maybe TTY isn't required when pand is used.

rfcomm> cat Kconfig
config BT_RFCOMM
    tristate "RFCOMM protocol support"
    depends on BT && BT_L2CAP
    help
      RFCOMM provides connection oriented stream transport.  RFCOMM
      support is required for Dialup Networking, OBEX and other Bluetooth
      applications.

      Say Y here to compile RFCOMM support into the kernel or say M to
      compile it as module (rfcomm).

config BT_RFCOMM_TTY
    bool "RFCOMM TTY support"
    depends on BT_RFCOMM
    help
      This option enables TTY emulation support for RFCOMM channels.

Your config

kernel26-one-dev> cat config | grep RFCOMM
CONFIG_BT_RFCOMM=m
# CONFIG_BT_RFCOMM_TTY is not set

My config

kernel26-one-dev> cat config | grep RFCOMM
CONFIG_BT_RFCOMM=m
CONFIG_BT_RFCOMM_TTY=y

Offline

#35 2009-05-18 19:11:22

Zom
Member
From: Sweden
Registered: 2007-10-27
Posts: 430

Re: kernel26-one-dev development thread

Tera wrote:

Does that mean suspend works for you now with one-dev?

Yep, and since I no longer use a splash and initrd, it boots super fast. I suspect it's just the card that's flaky, since I (still) randomly get Error -110 blah blah in dmesg when inserting the card. Apparently, that particular brand is recommended against using on a wii too.

Well, thanks for the help and the kernel. Will test the latest release right away.

Offline

#36 2009-05-19 07:35:45

el_cammello
Member
From: Italy
Registered: 2009-03-20
Posts: 70
Website

Re: kernel26-one-dev development thread

nfs4 works with this release?
Because:

[~] sudo mount -t nfs4 192.168.0.2:/ /media/nfs/
mount.nfs4: No such device

I've tried to load the modules for nfs4, but for these:

[~] sudo modprobe aes_generic
FATAL: Module aes_generic not found.
[~] sudo modprobe auth_rpcgss
FATAL: Module auth_rpcgss not found.
[~] sudo modprobe nfs_acl    
FATAL: Module nfs_acl not found.

P.S: Sorry for my bad english


Sorry for my bad english :-)

Offline

#37 2009-05-20 19:08:28

RayNbow
Member
From: The Netherlands
Registered: 2009-04-04
Posts: 7

Re: kernel26-one-dev development thread

I've tried installing kernel26-one-dev-2.6.30rc6-1 today, but now X no longer works. When starting X, the screen just turns off and the system hangs.

Since I noticed the new kernel uses uvesafb, I followed the uvesafb steps on the AOne wiki page, but I'm puzzled at the mkinicpio / grub2 step. Which step is needed?


Arch Novice
Aspire One A150-Aw

Offline

#38 2009-05-20 19:41:00

nimoworm
Member
From: Regensburg, Germany
Registered: 2009-05-03
Posts: 3

Re: kernel26-one-dev development thread

RayNbow wrote:

I've tried installing kernel26-one-dev-2.6.30rc6-1 today, but now X no longer works. When starting X, the screen just turns off and the system hangs.

Since I noticed the new kernel uses uvesafb, I followed the uvesafb steps on the AOne wiki page, but I'm puzzled at the mkinicpio / grub2 step. Which step is needed?

I had to switch to from EXA to UXA in my xorg.conf in the videocard section. But this was for kernel26-one-dev 2.6.30rc5-1.

Offline

#39 2009-05-20 20:16:34

psch
Member
Registered: 2009-05-14
Posts: 4

Re: kernel26-one-dev development thread

RayNbow wrote:

I've tried installing kernel26-one-dev-2.6.30rc6-1 today, but now X no longer works. When starting X, the screen just turns off and the system hangs.

This might be related to KMS enabled in the kernel.
Do you have xf86-video-intel-newest and libdrm-newest from AUR installed?
KMS will remove any need for a dedicated framebuffer driver, too, as it automatically sets the vts to the displays native resolution. I'd advice to use this solution rather than uvesafb.
As far as I know using uvesafb as X driver you lose any direct rendering, don't you? Though maybe that - and the inability to suspend - doesn't matter to you.

If you want to go with uvesafb, you don't need to mkinitcpio but go with the grub update and altered menu.lst, as we don't use a initramfs with kernel26-one-dev (I think, I don't want to sound over-confident... wink).

FYI: the Device section in xorg.conf doesn't really need anything else but the Identifier and the Driver, in case you choose to go with xf86-video-intel-newest;

Section "Device"
        Identifier "VideoCard0"
        Driver "intel"
EndSection

works flawlessly for me; UXA is automatically used by X, it seems;

# grep XA /var/log/Xorg.0.log
(II) UXA(0): Driver registered support for the following operations:
(II)         solid
(II)         copy
(II)         composite (RENDER acceleration)

HTH

Edit:

el_cammello wrote:

nfs4 works with this release?

No, NFS4 is not enabled in the kernel config:

$ zcat /proc/config.gz | grep -E "NFS.*4"
# CONFIG_NFS_V4 is not set

You'll have to change the config and recompile; or wait for Tera to add support (though I have no clue if he will add it).

Last edited by psch (2009-05-20 20:23:02)

Offline

#40 2009-05-20 20:52:02

RayNbow
Member
From: The Netherlands
Registered: 2009-04-04
Posts: 7

Re: kernel26-one-dev development thread

nimoworm, pcsh:
I found out that X hanged because of my virtual desktop resolution:

Section "Screen"
  Identifier "Screen0"
  Device "Videocard0"
  Monitor "Monitor0"
  DefaultDepth 24
  SubSection "Display"
    Viewport 0 0
    Depth 24
    Modes "1024x600" "800x600" "640x480"
    Virtual 1024 600
    # below no longer works with newer kernel
    #Modes "1280x1024" "1024x768" "1024x600" "800x600" "640x480"
    #Virtual 2304 1024
  EndSubSection
EndSection

This virtual desktop resolution is used in the Setting the right resolution on an external monitor trick.


Arch Novice
Aspire One A150-Aw

Offline

#41 2009-05-21 05:33:05

el_cammello
Member
From: Italy
Registered: 2009-03-20
Posts: 70
Website

Re: kernel26-one-dev development thread

psch wrote:

You'll have to change the config and recompile; or wait for Tera to add support (though I have no clue if he will add it).

I've recompile, now works smile


Sorry for my bad english :-)

Offline

#42 2009-05-24 19:37:40

Tera
Member
From: Finland
Registered: 2007-01-25
Posts: 81

Re: kernel26-one-dev development thread

New release based on 2.6.30rc7 is out.

Available from AUR and my dump dir as usual.

Changes includes the cpufreq fix from the last release and CONFIG_BT_RFCOMM_TTY=y as suggested by tom_g

MD5sum a77f8196f895a4f1ebb5c49a0642e036

Offline

#43 2009-05-24 19:44:32

psch
Member
Registered: 2009-05-14
Posts: 4

Re: kernel26-one-dev development thread

There does seem to be some kind of problem with DRI, KMS and suspend, apparently it did happen before (there were multiple traces in dmesg) but I only noticed it now because it did have an effect:

resize_wonk.png
(resized to not make me/others browsing on their aa1 scroll horizontally)

Here's the mentioned trace from dmesg:

[23769.796349] ------------[ cut here ]------------
[23769.796357] WARNING: at drivers/gpu/drm/i915/i915_gem_tiling.c:320 i915_gem_set_t
[23769.796364] Hardware name: AOA110
[23769.796368] failed to unbind object for tiling switchModules linked in: ipv6 uhci
[23769.796385] Pid: 3689, comm: X Tainted: G        W  2.6.30-one-dev #1
[23769.796391] Call Trace:
[23769.796405]  [<c0129437>] ? warn_slowpath_fmt+0x87/0xc0
[23769.796419]  [<c04cc029>] ? printk+0x17/0x1e
[23769.796428]  [<c02ff33a>] ? i915_gem_object_unbind+0x22a/0x230
[23769.796438]  [<c0303963>] ? i915_gem_set_tiling+0x6d3/0x6f0
[23769.796449]  [<c0106461>] ? print_trace_warning+0x11/0x20
[23769.796459]  [<c02e8ae3>] ? drm_ioctl+0x123/0x360
[23769.796469]  [<c0303290>] ? i915_gem_set_tiling+0x0/0x6f0
[23769.796479]  [<c01f2eee>] ? ext4_file_write+0x4e/0x190
[23769.796490]  [<c0188e32>] ? do_sync_write+0xd2/0x110
[23769.796499]  [<c0196249>] ? vfs_ioctl+0x89/0xa0
[23769.796508]  [<c0106461>] ? print_trace_warning+0x11/0x20
[23769.796517]  [<c01962d2>] ? do_vfs_ioctl+0x72/0x5d0
[23769.796527]  [<c023f65c>] ? security_file_permission+0xc/0x10
[23769.796536]  [<c0188fe3>] ? rw_verify_area+0x63/0xd0
[23769.796546]  [<c01735c0>] ? handle_mm_fault+0x3b0/0x660
[23769.796555]  [<c0189a90>] ? vfs_write+0x100/0x170
[23769.796564]  [<c0188d60>] ? do_sync_write+0x0/0x110
[23769.796574]  [<c0106461>] ? print_trace_warning+0x11/0x20
[23769.796583]  [<c01968a6>] ? sys_ioctl+0x76/0x90
[23769.796592]  [<c0102eb4>] ? sysenter_do_call+0x12/0x26
[23769.796601]  [<c0106461>] ? print_trace_warning+0x11/0x20
[23769.796609] ---[ end trace 76f7bbbe2dc9d7e2 ]---

I'm not actually sure, where I should put this info (except here), as I feel this is not directly kernel-related, or possibly out of the scope or Tera to fix, so any advice regarding that is welcome, too.

If there is any fix, I'd like that as well, obviously... smile

Edit: Whoa, that's some timing...

This problem was with the previous release, 2.6.30-r6, NOT the latest one, I'll try the latest one...

Though common sense suddenly came up and tells me this is most likely not at all kernel related and I should start googling around about libdrm/xf86-video-intel related stuff.
Yeah, anyway, building the latest kernel.

Another edit:

It does not (yet?) seem reproducible with the 2.6.30-rc7 kernel, but but as it took me some time to notice at first too, who knows.

I guess I'll see about this after prolonged use...

Last edited by psch (2009-05-24 21:26:49)

Offline

#44 2009-05-25 07:55:59

el_cammello
Member
From: Italy
Registered: 2009-03-20
Posts: 70
Website

Re: kernel26-one-dev development thread

@Tera: you can also change:

# CONFIG_MEMSTICK is not set

with:

CONFIG_MEMSTICK=y
# CONFIG_MEMSTICK_DEBUG is not set

#
# MemoryStick drivers
#
# CONFIG_MEMSTICK_UNSAFE_RESUME is not set
CONFIG_MSPRO_BLOCK=m

#
# MemoryStick Host Controller Drivers
#
CONFIG_MEMSTICK_TIFM_MS=m
CONFIG_MEMSTICK_JMICRON_38X=m
CONFIG_NEW_LEDS=y
CONFIG_LEDS_CLASS=y

thx smile


Sorry for my bad english :-)

Offline

#45 2009-06-05 17:03:16

EasyTarget
Member
From: Amsterdam
Registered: 2009-02-16
Posts: 17
Website

Re: kernel26-one-dev development thread

el_cammello;

I've recompiled the latest (24 may) one-dev release from terra with these options,  it does not enable the memorystick slot.. (other cards types, SD, CF, etc. work fine, only Sony's stupid proprietory memorystick fails.. no great loss to the world really).

Behaviour is exactly as I described in http://bbs.archlinux.org/viewtopic.php? … 86#p555886

I have now recompiled with CONFIG_MEMSTICK_DEBUG enabled, and plan to take it to the kernel devs if I can be bothered.. (sorry, but v.busy so not able to followup immediately)
- I am not aware of anyone making this work with non-linpus linux based aspire Ones.

EZ

Last edited by EasyTarget (2009-06-08 10:53:31)


My anger management class pissed me off. - anon

Offline

#46 2009-06-18 19:16:53

patrickthebold
Member
Registered: 2008-12-15
Posts: 55

Re: kernel26-one-dev development thread

I just installed rc7 and xrandr doesn't work.  I couldn't find any errors in the logs and by "doesn't" work I mean it does nothing when I try and switch to the external vga port or vice versa.  Any Idea what is going on?

Offline

#47 2009-08-11 12:50:18

rasteenb
Member
Registered: 2009-04-20
Posts: 19

Re: kernel26-one-dev development thread

[Solved] - The problem reappeared so I decided to add a "rootdelay=10" to the GRUB kernel line since I was working from a usb stick.  That seemed to help.

(earlier today):  a little time without power (AC/Batt) seemed to cure the problem.  Still not sure why it showed up.  I'm running Arch from a usb stick and I wonder if the delay created a problem.

(even earlier): Hello Tera:

I installed the latest -one-dev but acerhdf is not working.

I check dmesg | grep acerhdf and find "failed reading fan state, disabling kernelmode"

Attempts to manually enter kernelmode or usermode fail.

Have you heard of this problem?  Do you have any ideas for a fix?

Thanks,
Rob

Last edited by rasteenb (2009-08-12 04:04:26)

Offline

#48 2009-08-12 05:34:27

Tera
Member
From: Finland
Registered: 2007-01-25
Posts: 81

Re: kernel26-one-dev development thread

I'm glad to hear you solved it smile

I'm in fact in progress of rolling out a new release soon. I have already a working PKGBUILD ported for kernel version 2.6.31-rc5-git7. I'm in the progress of Q/A'ing it but don't expect me to release it anytime soon as there's one major issue with it which I'm working with.

The biggest change from the last release is that acerhdf is now part of the official vanilla kernel so I'm dropping the acerhdf patch from it, as there is no need for it anymore.

Last edited by Tera (2009-08-12 05:41:28)

Offline

#49 2009-08-12 06:11:32

AndyRTR
Developer
From: Magdeburg/Germany
Registered: 2005-10-07
Posts: 1,641

Re: kernel26-one-dev development thread

My girlfriend has noticed when moving from kernel 2.6.29 to a .30.x one that the internal fan doesn run and it's getting much more hot. Anyone else having this, maybe some acpi changes or have I accidently changed something in the config :S

Offline

#50 2009-08-12 09:19:46

Tera
Member
From: Finland
Registered: 2007-01-25
Posts: 81

Re: kernel26-one-dev development thread

AndyRTR wrote:

maybe some acpi changes or have I accidently changed something in the config :S

If the fan is not running you are most likely using acerhdf, otherwise the fan runs non-stop. I believe the default limits for acerhdf when it kicks the fan in were in about 60-70'C. The chipset of Intel Atom is (in my believe) rated to somewhere around 90'C operational temperature (feel free to correct me).

Talking from personal view, the machine indeed gets a bit hot with acerhdf enabled but the fan does work. I am using my Acer Aspire as a 3G router at the moment of writing and it's in a thick slot with very little air space. The fan is running happily and blowing hot air from the inside.

Offline

Board footer

Powered by FluxBB