Author Topic: So who here is running the "testing" 2.6.32 kernel? How's it going?  (Read 8580 times)

Offline marian

  • Hero Member
  • *****
  • Posts: 608
  • Hi
Re: So who here is running the "testing" 2.6.32 kernel? How's it going?
« Reply #15 on: December 27, 2009, 12:44:13 AM »
Beware if you have an intel graphics driver or using the fglrx driver. Also the madwifi driver needs update to build against new kernel or use the ath5k driver.

The intel graphics seems to work ok, except for the boot screen. (I have compiz running, kde3.5, no glitches).

Galen

The 2.6.31.9 has better intel graphic support.


Then maybe I'll try that one because with 2.6.32 I've had problems with proprietary driver for my Nvidia.
Freedom is the only way!

Offline daniel

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 3735
  • God knows, i'm not an Angel!
    • Tipps und Tricks
Re: So who here is running the "testing" 2.6.32 kernel? How's it going?
« Reply #16 on: December 27, 2009, 04:45:20 AM »
So after install, and remove kernel version 2.6.31.9, because no grafic (intel810 and later), is my wireless connection broken.
DKMS-broadcom-wl is installed. What is needed to repair my wireless connection?


EDIT:
Drama, drama, drama...
After i read with
http://www.pclinuxos.com/forum/index.php/topic,65070.0.html
I installed 2.6.26.8 and removed 2.6.27.31, but no connection (dkms-broadcom-wl can not be installed, but it is installed). So installed kernel 2.6.27.31 and now my connection work's again...
I don't more testing a kernel  >:(
« Last Edit: December 27, 2009, 06:12:28 AM by Leiche »

Offline coolbreeze

  • Hero Member
  • *****
  • Posts: 2668
  • Error #152 - Windows not found: (C)heer (P)arty (D
Re: So who here is running the "testing" 2.6.32 kernel? How's it going?
« Reply #17 on: December 27, 2009, 06:09:42 AM »
What is needed to repair my wireless connection?

Basically the same here, keeps dropping connection and reconnecting so fast that the pop-up can't keep pace, it's continuous thing.  ???
Linux user #440309
PCLOS IS THE KING  Please Donate to the cause. PCLinuxOS
My mind is so full that there's no room to think.
M5A78L-M mobo, AMD Phenom IIx 6 1055T, 4Gig ram,nvidia GeForce GT 240 1 Gig, Netgear DGN2200

Offline Was_Just19

  • Hero Member
  • *****
  • Posts: 6852
  • MLU
Re: So who here is running the "testing" 2.6.32 kernel? How's it going?
« Reply #18 on: December 27, 2009, 07:19:34 AM »
Had a look at installing this and got the following

Quote
To be removed
dkms-cdfs
dkms-kqemu
dkms-ndiswrapper
dkms-squashfs-lzma
qemu
qemu-launcher
qemulater
qtemu

To be Upgraded
kernel-firmware

Decided against it for the present   :o

KDE3 ..... fully updated.

Is this normal?

Offline pinoc

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 2842
    • other projects...
Re: So who here is running the "testing" 2.6.32 kernel? How's it going?
« Reply #19 on: December 27, 2009, 09:50:09 AM »
all full updates from testing, including new udev and kernel 2.6.32:
- updated a VMware Fusion virtual machine, all perfect and FAST...
- updated a laptop to 2.6.32 with Nvidia 7300GO, reboot, the only dkms with a failed build was madwifi (got "error 10"), everything else works fine, newest Nvidia build and works perfectly, wlan with ipw3945, all running nicely!

just don't forget to also add the kernel-devel package, to have the dkms build.
great job!

Online Old-Polack

  • Administrator
  • Super Villain
  • *****
  • Posts: 11580
  • ----IOFLU----
Re: So who here is running the "testing" 2.6.32 kernel? How's it going?
« Reply #20 on: December 27, 2009, 12:58:47 PM »
Neither of the testing kernels will build the dkms-nvidia173 driver. Trying with the same driver from nVidia d/l the installer complains /lib/modules/2.6.31.9-pclos3/source/include/linux/version.h missing, or /lib/modules/2.6.32.2-pclos5/source/include/linux/version.h missing.

This is with the full kernel sources installed, rather than the kernel-devel packages. The newer nVidia drivers build, but won't work properly with a GeForce 6800 GT card; gets a decent desktop, but locks up the various video players. This is common to all kernels when this driver is used.

Needs a way to get the nvidia 173.xx driver to build on either of these kernels before they can be used with this graphics card.

Back to the 2.6.27.31.tex5 for now.
Old-Polack

Of what use be there for joy, if not for the sharing thereof?



Lest we forget...

Offline MBantz

  • PCLinuxOS Tester
  • Hero Member
  • *******
  • Posts: 1317
Re: So who here is running the "testing" 2.6.32 kernel? How's it going?
« Reply #21 on: December 27, 2009, 02:24:24 PM »
hm.. just about to get a nvidia card to test vdpau and try the new kernel out with this, looks like I can look forward to some fine hours getting it to work :-)

O-P (and not OP :-) have you tried http://www.nvidia.co.uk/object/linux_display_ia32_190.53_uk.html - that is the 190 version? I'll check out if anyone is working on newer nvidia drivers for packaging and eventually give it a go :-)

cheers,
MBantz

Online Old-Polack

  • Administrator
  • Super Villain
  • *****
  • Posts: 11580
  • ----IOFLU----
Re: So who here is running the "testing" 2.6.32 kernel? How's it going?
« Reply #22 on: December 27, 2009, 04:38:31 PM »
hm.. just about to get a nvidia card to test vdpau and try the new kernel out with this, looks like I can look forward to some fine hours getting it to work :-)

O-P (and not OP :-) have you tried http://www.nvidia.co.uk/object/linux_display_ia32_190.53_uk.html - that is the 190 version? I'll check out if anyone is working on newer nvidia drivers for packaging and eventually give it a go :-)

cheers,
MBantz


That's one that breaks the video players. The 173.xx is now listed as the newest legacy driver. They actually use that term on the site, denoting the change of card families. The GeForce 6xxx, 7xxx, 8xxx, 9xxx all being closely related, but fundamentally different from those cards with the newest designations, will use the 173xx driver. I don't have any cards in the 9xxx group, but do have GF 6800, 7600, and 8600 cards that all work perfectly with the 173xx driver, and all show the same problem with video players trying to use the 180xx, 185xx and 190xx drivers.

I would guess, without digging deeper, that all production of cards, of that family of cards, has ceased now, and that is the reason the 173xx is now considered legacy
Old-Polack

Of what use be there for joy, if not for the sharing thereof?



Lest we forget...

Offline tschommer

  • PCLinuxOS Tester
  • Hero Member
  • *******
  • Posts: 1890
  • MLU and BLU (Bacon lovin' user)
Re: So who here is running the "testing" 2.6.32 kernel? How's it going?
« Reply #23 on: December 28, 2009, 05:25:44 AM »
hm.. just about to get a nvidia card to test vdpau and try the new kernel out with this, looks like I can look forward to some fine hours getting it to work :-)

O-P (and not OP :-) have you tried http://www.nvidia.co.uk/object/linux_display_ia32_190.53_uk.html - that is the 190 version? I'll check out if anyone is working on newer nvidia drivers for packaging and eventually give it a go :-)

cheers,
MBantz


That's one that breaks the video players. The 173.xx is now listed as the newest legacy driver. They actually use that term on the site, denoting the change of card families. The GeForce 6xxx, 7xxx, 8xxx, 9xxx all being closely related, but fundamentally different from those cards with the newest designations, will use the 173xx driver. I don't have any cards in the 9xxx group, but do have GF 6800, 7600, and 8600 cards that all work perfectly with the 173xx driver, and all show the same problem with video players trying to use the 180xx, 185xx and 190xx drivers.

I would guess, without digging deeper, that all production of cards, of that family of cards, has ceased now, and that is the reason the 173xx is now considered legacy

That's strange, because I am currently using a 9800GT without any issues with the 190xx driver. All tested with 2.6.27.31 as well as the new kernels (2.6.31.9 and 2.6.32.1).  Even had it for a week with a 2.6.32 kernel I built myself.

I also used an 8600GT with the 180xx driver for several weeks before switching cards. And the 190xx is the best driver I've used for my 9800GT... ???
Our defense is in the preservation of the spirit which prizes liberty as the heritage of all men, in all lands, everywhere."
Abraham Lincoln --September 11, 1858 Speech at Edwardsville, Illinois

Offline AndrzejL

  • PCLinuxOS Tester
  • Super Villain
  • *******
  • Posts: 12797
  • RLU #490933
    • Wordpress On The Wardrobe...
Re: So who here is running the "testing" 2.6.32 kernel? How's it going?
« Reply #24 on: December 28, 2009, 02:05:53 PM »
*** THIS IS NOT A HOWTO *** I AM USING KDE4 AS A TEST INSTALL ONLY *** IF YOU WILL FOLLOW IT AND DO NOT UNDERSTAND THE COMMANDS USED YOU MAY AND PROLLY WILL LOOSE VITAL DATA ***

NEWBIES HAVE BEEN WARNED!


Beware if you have an intel graphics driver or using the fglrx driver. Also the madwifi driver needs update to build against new kernel or use the ath5k driver.

The intel graphics seems to work ok, except for the boot screen. (I have compiz running, kde3.5, no glitches).

Galen

The 2.6.31.9 has better intel graphic support.

Info about graphics from the KDE3 installation with 27.31 kernel.

Code: [Select]
00:02.0 VGA compatible controller: Intel Corporation Mobile 915GM/GMS/910GML Express Graphics Controller (rev 03) (prog-if 00 [VGA controller])
        Subsystem: Acer Incorporated [ALI] Device 006a
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap+ 66MHz- UDF- FastB2B+ ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0
        Interrupt: pin A routed to IRQ 16
        Region 0: Memory at b0080000 (32-bit, non-prefetchable) [size=512K]
        Region 1: I/O ports at 1800 [size=8]
        Region 2: Memory at c0000000 (32-bit, prefetchable) [size=256M]
        Region 3: Memory at b0000000 (32-bit, non-prefetchable) [size=256K]
        Expansion ROM at <unassigned> [disabled]
        Capabilities: [d0] Power Management version 2
                Flags: PMEClk- DSI+ D1- D2- AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot-,D3cold-)
                Status: D0 NoSoftRst- PME-Enable- DSel=0 DScale=0 PME-
        Kernel modules: intelfb

KDE4 from the repos fully updated + bazillion applications added.

Kernel 32.2 installed wouldn't boot (with autologin enabled via pcc) to my user account.

I have solved it by

a) booting into the single user mode by pressing [F2] and typing in 1 at the end of the boot options for this kernel
b) userdel andrzejl <<< removed my old account
c) rm -r /home/andrzejl <<< removed my old account folder
d) rm /etc/X11/xorg.conf <<< removed xorg.conf
e) useradd -m -d /home/andrzejl -p password <<< added new user + created home and password
f) video <<< setup new xserver
g) reboot <<< selfexplanatory...

Older xorg caused problems when booting to the old account. Some settings would not allow me to boot in. Autologin was working fine. I was able to see splash screen and then puter stopped loading anything after the tools icon. The world icon was blurred. Nothing was happening. Just the magic keys combo allowed me to reboot till the moment that I did the above. Removing and recreating from the scratch xorg.conf and user account made it usable again. Also when rebooting I dont see the shutdown screen BUT vertical white stripes. This happens as well if xorg crashes or [ALT]+[CTRL]+[BACKSPACE] combo is used.

Hope this means something to someone.

Andy

BTW: My xorg.conf for kernel 32.2 and kde4 now looks like this

Code: [Select]
# File generated by XFdrake (rev 247269)

# **********************************************************************
# Refer to the xorg.conf man page for details about the format of
# this file.
# **********************************************************************

Section "ServerFlags"
    #DontZap # disable <Ctrl><Alt><BS> (server abort)
    AllowMouseOpenFail # allows the server to start up even if the mouse does not work
    #DontZoom # disable <Ctrl><Alt><KP_+>/<KP_-> (resolution switching)
EndSection

Section "Module"
    Load "dbe" # Double-Buffering Extension
    Load "v4l" # Video for Linux
    Load "extmod"
    Load "type1"
    Load "freetype"
    Load "glx" # 3D layer
    Load "dri" # direct rendering
EndSection

Section "InputDevice"
    Identifier "Keyboard1"
    Driver "kbd"
    Option "XkbModel" "pc105"
    Option "XkbLayout" "ie"
    Option "XkbOptions" "compose:rwin"
EndSection

Section "InputDevice"
    Identifier "Mouse1"
    Driver "mouse"
    Option "Protocol" "ExplorerPS/2"
    Option "Device" "/dev/mouse"
EndSection

Section "InputDevice"
    Identifier "Mouse2"
    Driver "evdev"
    Option "device" "/dev/input/by-id/usb-MLK_Silvercrest_OM1008-event-mouse"
    Option "HWheelRelativeAxisButtons" "7 6"
EndSection

Section "InputDevice"
    Identifier "SynapticsMouse1"
    Driver "synaptics"
    Option "SHMConfig" "on"
EndSection

Section "Monitor"
    Identifier "monitor1"
    VendorName "Generic"
    ModelName "Flat Panel 1280x800"
    HorizSync 31.5-50.0
    VertRefresh 56.0-65.0

    # Monitor preferred modeline (60.0 Hz vsync, 48.9 kHz hsync, ratio 16/10)
    ModeLine "1280x800" 68.9 1280 1301 1333 1408 800 804 808 816 -hsync -vsync

    # TV fullscreen mode or DVD fullscreen output.
    # 768x576 @ 79 Hz, 50 kHz hsync
    ModeLine "768x576"     50.00  768  832  846 1000   576  590  595  630

    # 768x576 @ 100 Hz, 61.6 kHz hsync
    ModeLine "768x576"     63.07  768  800  960 1024   576  578  590  616

    # modeline generated by gtf(1) [handled by XFdrake]
    ModeLine "1280x800_120"  181.21  1280 1376 1520 1760  800 801 804 858  -HSync +Vsync

    # modeline generated by gtf(1) [handled by XFdrake]
    ModeLine "1280x800_100"  147.89  1280 1376 1512 1744  800 801 804 848  -HSync +Vsync

    # modeline generated by gtf(1) [handled by XFdrake]
    ModeLine "1280x800_85"  123.38  1280 1368 1504 1728  800 801 804 840  -HSync +Vsync

    # modeline generated by gtf(1) [handled by XFdrake]
    ModeLine "1280x800_75"  107.21  1280 1360 1496 1712  800 801 804 835  -HSync +Vsync

    # modeline generated by gtf(1) [handled by XFdrake]
    ModeLine "1280x800_60"  83.46  1280 1344 1480 1680  800 801 804 828  -HSync +Vsync

    # modeline generated by gtf(1) [handled by XFdrake]
    ModeLine "1280x800_50"  68.56  1280 1336 1472 1664  800 801 804 824  -HSync +Vsync
EndSection

Section "Device"
    Identifier "device1"
    VendorName "Intel Corporation"
    BoardName "Intel 810 and later"
    Driver "intel"
    Option "DPMS"
    Option "AccelMethod" "EXA"
EndSection

Section "Screen"
    Identifier "screen1"
    Device "device1"
    Monitor "monitor1"
    DefaultColorDepth 16

    Subsection "Display"
        Depth 8
        Modes "1280x800"
    EndSubsection

    Subsection "Display"
        Depth 15
        Modes "1280x800"
    EndSubsection

    Subsection "Display"
        Depth 16
        Modes "1280x800"
    EndSubsection

    Subsection "Display"
        Depth 24
        Modes "1280x800"
    EndSubsection
EndSection

Section "ServerLayout"
    Identifier "layout1"
    InputDevice "Keyboard1" "CoreKeyboard"
    InputDevice "Mouse1" "CorePointer"
    InputDevice "Mouse2" "SendCoreEvents"
    InputDevice "SynapticsMouse1" "SendCoreEvents"
    Screen "screen1"
EndSection

Performance is weak BUT I am able to boot to the system...

EDIT: Also Acer Travelmate 2420 that I own comes with atheros - madwifi card. I replaced it with other card - also atheros based BUT I dont use it. I know that lan driver and huawei E160G works so my Internet Connection Sharing was not interrupted...

Info about wifi and lan cards from the kde3 install with 27.31 kernel:

Code: [Select]
06:05.0 Network controller: Atheros Communications Inc. AR5008 Wireless Network Adapter (rev 01)
        Subsystem: Apple Computer Inc. Device 0087
        Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV+ VGASnoop- ParErr- Stepping- SERR+ FastB2B- DisINTx-
        Status: Cap+ 66MHz+ UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 96, Cache Line Size: 32 bytes
        Interrupt: pin A routed to IRQ 21
        Region 0: Memory at b0100000 (32-bit, non-prefetchable) [size=64K]
        Capabilities: [40] #80 [0000]
        Kernel driver in use: ath_pci
        Kernel modules: ath9k, ath_pci

06:07.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL-8139/8139C/8139C+ (rev 10)
        Subsystem: Acer Incorporated [ALI] Device 006a
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR+ FastB2B- DisINTx-
        Status: Cap+ 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 32 (8000ns min, 16000ns max)
        Interrupt: pin A routed to IRQ 20
        Region 0: I/O ports at 3000 [size=256]
        Region 1: Memory at b0110000 (32-bit, non-prefetchable) [size=256]
        Capabilities: [50] Power Management version 2
                Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=375mA PME(D0-,D1+,D2+,D3hot+,D3cold+)
                Status: D0 NoSoftRst- PME-Enable- DSel=0 DScale=0 PME-
        Kernel driver in use: 8139too
        Kernel modules: 8139cp, 8139too

And another information... Kernel 27.31tex5 + the kde3 are really working as a charm on this machine... And the same goes for Dell Latitude D600... Remastered it and am ready for xorg testing...

Andy
« Last Edit: December 28, 2009, 03:42:14 PM by AndrzejL »

Offline Joble

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 6804
  • USA - Mountain Time
Re: So who here is running the "testing" 2.6.32 kernel? How's it going?
« Reply #25 on: December 28, 2009, 06:06:51 PM »
I got no technical stuff to share.  Old ATI graphics 250M or some such chipset, Broadcom PCMCIA 4318 wireless.  New Kernel, dbus, ndiswrapper and wpa_supplicant, I like it!  BUT:  live-usb fails, course I knew that would happen, or suspected it anyway, hoping to help those in the know to solve that issue.

Your Mileage May Vary.
Search First.
Forum Rules
Hero means I talk a lot, nothing more, nothing less!
Have an Awesome Day!
Healthy System

Offline daniel

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 3735
  • God knows, i'm not an Angel!
    • Tipps und Tricks
Re: So who here is running the "testing" 2.6.32 kernel? How's it going?
« Reply #26 on: December 29, 2009, 03:06:54 AM »
Install 2.6.31.9 again, with all dependencies on my testsystem, and grafic server starts, but broadcom-wl is missing...

Identifikation
Hersteller: ‎Broadcom Corporation

Beschreibung: ‎BCM4312 802.11b/g

Medienklasse: ‎Network controller

DKMS-BROADCOM-WL is installed, and dkms is enabled at start up.

When i open netapplet is not show me any wireless connection, nothing, only lan  ::)

Offline pinoc

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 2842
    • other projects...
Re: So who here is running the "testing" 2.6.32 kernel? How's it going?
« Reply #27 on: December 30, 2009, 03:56:04 AM »
Neither of the testing kernels will build the dkms-nvidia173 driver. Trying with the same driver from nVidia d/l the installer complains /lib/modules/2.6.31.9-pclos3/source/include/linux/version.h missing, or /lib/modules/2.6.32.2-pclos5/source/include/linux/version.h missing.

This is with the full kernel sources installed, rather than the kernel-devel packages. The newer nVidia drivers build, but won't work properly with a GeForce 6800 GT card; gets a decent desktop, but locks up the various video players. This is common to all kernels when this driver is used.

Needs a way to get the nvidia 173.xx driver to build on either of these kernels before they can be used with this graphics card.

Back to the 2.6.27.31.tex5 for now.

Don't know anything about dkms so I'm just guessing here but maybe the following assumption makes sense: the Nvidia 190 dkms compiles and works fine because it has been set up to use the new gcc which is also used for the new kernels. The 173 Nvidia dkms was set up with the old gcc and hence only works with the older kernels up to 2.6.27 which were compiled with the old gcc. So to have the 173 Nvidia dkms working with the newer kernels maybe the dkms needs to be reworked to use the newer gcc or it may work if "export IGNORE_CC_MISMATCH=1" is set before compiling the dkms?? Just a thought...

Offline DeBaas

  • Hero Member
  • *****
  • Posts: 1518
    • PCLinuxOS.nl
Re: So who here is running the "testing" 2.6.32 kernel? How's it going?
« Reply #28 on: December 30, 2009, 04:46:31 AM »
Got it all working, only one....................
Seen this before, all IDE drives are now sda/sdb drives instead of hda/hdb
No problem because UUID= or LABEL= is the solution ;)

kernel-devel-2.6.32.2-pclos5 and dkms-nvidia-current nvidia_190.xx
fully updated from:

http://ftp.nl.freebsd.org/os/Linux/distr/texstar/pclinuxos/apt/
pclinuxos/2007
main extra nonfree kde4 gnome testing

Offline Ɗα√ϵς§

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 5524
  • PCLinuxOS: A revolution in personal computing
Re: So who here is running the "testing" 2.6.32 kernel? How's it going?
« Reply #29 on: December 30, 2009, 06:55:16 AM »
I started by installing Mini-Me 2009.1. Then did full update.  Went OK.

Tried to update at that point to KDE4 using Tex's instructions but had to reboot first. Even though the network showed as still connected, Synaptic couldn't see it, even after an X-system restart. Once I rebooted, update to KDE4 went ahead just fine.

After a reboot, got into KDE4, added testing and upgraded to full testing. This brought in the 2.6.32 kernel but not the headers, which I had to do separately. At one point, I could not get into the desktop anymore, kept throwing me back to the login screen.

I got into the failsafe desktop, and tried to run kwin. The error I got was a missing library libdrm.so.2 - I ran synaptic, and libdrm2 was installed! I re-installed it, tried to run kwin again, this time it was successful. Something obviously went wrong somewhere along the line.

Had to go into PCC to make dkms run on startup. Installed dkms-nvidia-current which uses version 190, which works well with my card.

Some of the utility scripts do not work, because they demand that the kernel name ends with "texn" whereas the testing kernels end with "pclosn". Had to do localisation through control centre, had to add the openoffice repo to install OOo.

Once all this was done, it appears to be working nicely. Had to add the apps I wanted, so far no further bugs. Runs OK, is it seriously faster? Not sure about that.

A few "ease of use" bugs around then, but once sorted out, seems very good.

I would like to try it on my Samsung NC10, but not sure having read the stuff about Intel graphics drivers.
« Last Edit: December 30, 2009, 06:57:01 AM by davecs »
PCLinuxOS has no wealthy sponsors and can only survive with your donations! Don't wait it might be too late.

Running PCLOS on Athlon II/250 with 2Gb Ram using ASUS M2N68-AM Plus Mobo with Nvidia GF7600GT graphics, and Samsung NC10 Netbook