HP Pavilion dv6103eu laptop - all Ubuntu versions gets frozen during X startup until I specify "noapic" boot option

$ uname-a:
Linux ubuntu 2.6.20-9-generic #2 SMP Mon Feb 26 03:01:44 UTC 2007 i686 GNU/Linux

Kernel package on Ubuntu Feisty Herd 5 is:

Package: linux-image-2.6.20-9-generic
Status: install ok installed
Installed-Size: 67408
Maintainer: Ben Collins <email address hidden>
Architecture: i386
Source: linux-source-2.6.20
Version: 2.6.20-9.16

Confirmed bugs should have their Importance set.

Baltix currently has an alternative boot option (with added noapic nolapic after --) in LiveCD startup menu and this is temporary workaround for such errors.

Ubuntu still freezes during boot on HP Pavilion DV 6003. I've tested with newest Ubuntu Feisty daily, dated 2007-04-10.
Btw, when I choose "Safe graphic mode" from LiveCD startup menu, then Ubuntu starts on my system.
Are you planing to solve this important bug before Ubuntu 7.04 'Feisty' final will be released ?

This also occurs on the Compaq Presario V6000Z laptop.

I'll try to get the equivalent information without noapic tomorrow, but I don't know if it'll stay up long enough.

Should've mentioned above - this is from my Presario V6000

I managed to boot without 'noapic' for long enough to get a kernel log output. Notably, I did not start X, nor load any tainting modules. The freeze actually occured in the midst of a 'cat /proc/interrupts', but this may have been a coincidence. There was no keyboard LED flashing nor console output. Attached is the kern.log output from that boot.

Note that the dmidecode output is unchanged without noapic.

I have the same problem with my HP pavilion dv9000.

I have to specify noapic or apci=off otherwise the computer freezes randomly.

When I specify noapic the usb devices don't work correctly and when I specify acpi=off acpi and nvidia card don't work.

I have seen a lot of users having similar problem.

@marcus, Booting with noirqdebug noapic fixes the USB problems for me - but if I have ehci_hcd.ko loaded, one of my CPUs is pegged at 100% hi in top, causing my CPU temperature to rise, and my battery life to drop significantly.

Marcus, Bryan:
Are there BIOS updates available for your exact laptops?

Bryan,

I havn't tried with noirqdebug, but I have exactly the same problem with the ehci_hcd module. If I manage to start the computer without the noapic option I see that the ehci and ohci modules receives the same IRQ and no cpu load. But with noapic ehci receives a different IRQ and I can see in /proc/interupts that the ehci-hcd revieced a lot of interrupts and when I run then 'top' command the cpuload is 100% + cpu fan goes high all the time.

Sitsofe,

I havn't checked for bios upgrade, but my compter is from may 2007.

I upgraded to latest bios from july 2007. No change.

Hey guys,

First off, I use gentoo, so I realize I'm a little out of place here.

However, I'm experiencing the exact same problems with my dv9317cl.

Booted normally on 2.6.22, my system functions okay. However, running programs that produce a lot of text output (cat /proc/interrupts, and especially dmesg) freeze the system in text mode. Booting with noapic causes one of the cores to get tied up processing spurious interrupts from the USB controller (unloading ehci-hcd causes this to stop at the expense of USB 2.0 support) and I experience intermittent freezes in Xorg while using OpenGL apps (according to nvidia this is caused by the video card not receiving interrupts).

I'd be willing to provide any logs/info requested. I'd even be willing to install ubuntu to troubleshoot this.

Now, I had this problem in Feisty Fawn final, but after testing Gutsy Tribe 2, I noticed that it still has the same problem.

When booting up, it halts at a blank screen. The common workaround is to add "noapic irqpoll noirqdebug" to the kernel parameters.

I have the same problem with hp pavilion a1547c

I have an HP Pavilion dv6125om which is having this issue in both Feisty and Gutsy Tribe 4, both running in amd64 mode with the ubuntu generic kernel. If I boot without special boot params, it will hang with the splash screen progress bar at around 13%. If I boot with "noapic noirqpoll" things boot up ok, but about 75% of my total CPU (as shown in top, both cores combined) is used in hard interrupt state, and it appears ehci_hcd is the culprit based on /proc/interrupts. I'm running the latest bios (F.38 from July 2007).

I upgraded to Gutsy because based on the resolution of the following bug report I thought this issue may have been fixed:
http://bugzilla.kernel.org/show_bug.cgi?id=7562

I too am willing to provide what I can to help get this resolved. I'd really like to see this box performing the way it should. Thanks.

Have you tried to disable RTC support in the kernel? This is a dirty fix that makes the computer boot more often.

I have reported rtc bug in http://bugzilla.kernel.org/show_bug.cgi?id=8870.

There are still other things that makes my computer crash, but the RTC was one of the major things.

Running programs that produces a lot of output to the console (dmesg etc.) also makes my computer crash.

I have done a more comprehensive test of boot parameters too see the behavior of my system. These were all done running Gutsy Tribe 4 with linux-image-2.6.22-9-generic version 2.6.22-9.25 on amd64 architecture. My HP dv6000 series laptop has a dual-core AMD Turion64 x2. I'm running the proprietary nvidia driver.

No special boot params: hangs during "Loading hardware drivers." Does not respond to ctl-alt-del.
noapic: boots up properly. top shows 70% of combined CPU usage in hard interrupt state. After a minute or two, kernel says "disabling IRQ #7." According to /proc/interrupts, IRQ 7 is for ehci_hcd:usb1.
noapic noirqdebug: boots up properly. top shows 70% of combined CPU usage in hard interrupt state. In /proc/interrupts, interrupts appear unbalanced to CPUs. Also, appears that the 70% interrupts are being caused by ehci_hcd.
acpi=noirq: problems loading X. Screen turns black and does not progress further. Still responds to ctl-alt-del and reboots the box.
acpi=off: same symptoms as acpi=noirq.
pci=noacpi: same symptoms as acpi=noirq.
nmi_watchdog=0: same symptoms as using no special boot params.
nmi_watchdog=2: same symptoms as using no special boot params.
maxcpus=1: boots and appears to work well (only with 1 core). However, shutdown appears to hang. Once it hung with "System is restarting, please wait..." on the screen. Another time it hung after printing "Running local boot scripts (/etc/rc.local) [OK]".

Currently I am run with "noapic noirqdebug" for normal usage and I don't see any hangs or other issues besides the fact that performance is poor.

Robert, Have you tried to compile without RTC support? And then run without any extra kernel options.

I tried recompiling the Gutsy 2.6.22-9 kernel with all RTC options completely disabled. Unfortunately, it didn't seem to change anything for me. On bootup with "noapic noirqdebug" I still had over 65% CPU usage in hard interrupts (according to top). When I booted without those flags, the machine hung during boot again and was non-responsive to the keyboard.

Any suggestions on how to debug this problem?

I have tried using powertop to see the interrupt usage. With no USB devices plugged in, my machine shows about 450 ehci_hcd interrupts per second. When I unload the ehci_hcd module these interrupts stop. This is all with boot parameters "noapic noirqdebug". My machine still won't boot without those parameters.

Robert:
If you are capable of building your own kernels, I think your best bet of having this problem resolved (given that you have said it persists into Gutsy) is to test a vanilla kernel and report the issue in the upstream kernel bugzilla ( http://bugzilla.kernel.org/ ). Your interrupt routing sounds quirky and might need a workaround (or a BIOS update).

I have been following this bug for a while since I am encountering a very similar problems with a HP dv6408nr laptop (Turion 64 X2 TL-56 CPU).

I have found that the excessive hard interrupts, in my case, are coming from the sdhci driver and the SD/SDIO/MMC/MS/MSPro Host Adapter, especially if it is auto-loaded. By blacklisting the device and loading it later by hand, it is only occasionally that this device will then generate excessive interrupts.

Before loading the sdhci driver I see that it is states IRQ 11 in the lspci output:
# lspci -s 07:05.01 -v
07:05.1 Generic system peripheral [Class 0805]: Ricoh Co Ltd R5C822 SD/SDIO/MMC/MS/MSPro Host Adapter (rev 19)
        Subsystem: Hewlett-Packard Company Presario V6133CL
        Flags: bus master, medium devsel, latency 64, IRQ 11
        Memory at b8000800 (32-bit, non-prefetchable) [size=256]
        Capabilities: [80] Power Management version 2

But after loading sdhci, lspci states IRQ7 is in use:
# lspci -s 07:05.01 -v
07:05.1 Generic system peripheral [Class 0805]: Ricoh Co Ltd R5C822 SD/SDIO/MMC/MS/MSPro Host Adapter (rev 19)
        Subsystem: Hewlett-Packard Company Presario V6133CL
        Flags: bus master, medium devsel, latency 64, IRQ 7
        Memory at b8000800 (32-bit, non-prefetchable) [size=256]
        Capabilities: [80] Power Management version 2

This is confirmed by /proc/interrupts:
# cat /proc/interrupts | grep sdhci
  7: 2 821 IO-APIC-fasteoi sdhci:slot0

I also encounter the intermittent RTC issues. I have blacklisted all the RTC modules: rtc_cmos, rtc_core and rtc_lib, but even so, the hwclock command (or any bootscript that calls hwclock) will frequently cause a hard hang.

The 3rd issue is an unknown hang when running openSUSE 10.3 -default kernels... The boot option, pollirq, can delay the hang, sometimes for hours. This is most frustrating since it does not occur while running 2.6.22.5-*-vanilla kernels.

Hope this helps someone get to the bottom of this... even though my experiences are on a different distribution.

Same issue with HP Pavilion dv9430us laptop.

 noapic was the only change needed to get the live cd to boot. I will report all the bugs I find under HP Pavilion dv9430us laptop.

I have the same problem with a dv6000us laptop. I had feisty running 'ok,' it would boot about half the times, and decided to upgrade to gutsy to see if the problems would stop. With gutsy my broadcom card 'dissapeared' and did not show up on the lspci but everything else seemed to work, so I opted to install the 64bit tribe 5. This has been nothing but a headache and the laptop will at the most boot to a command line. I've read that this might be a problem with the hwclock (most of the times the system locks up after checking the clock), and have just found a temporary fix which I will try later since I had to get a hold of a windows pc first because my laptop is decomissioned...

The funny thing is that knoppix will boot without a hitch, and feisty seemed to boot ok. Oh, and when the gutsy install has worked (I've installed it about 5 times with different degrees of success) and let me into the gui, the updates completely crashed the system, so hopefully in the final version this will stop happening,
Jack

The computer boots with noapic, but the bcm43xx chipset does not show up and irq7 (usb) becomes disabled... On a side note, the other options such as "acpi=noirq," etc which have been previously mentioned didn't work either. I hope someone can solve this, and if anyone would like me to test something I would be more than willing to. The final release is just 1 month away and it would be nice if this problem could be fixed... I would hate to have to give up and go back to xp,
Jack

I have run some more bootup tests. I am now booting with the options "noapic irqfixup" instead of my previous "noapic noirqdebug". My system runs very stable with either of those, but there are fewer spurious ehci interrupts with irqfixup. powertop shows the ehci interrupts drop from 450 per second to 250 per second with irqfixup.

I have also found that the LiveCD of SimplyMEPIS 6.5.02 (64-bit) boots fine on my system. I believe it is running kernel 2.6.15. Perhaps an attempt at bisecting the differences between that kernel and the current version will yield what changes are causing the problem.

Sitsofe: thanks for the advice, I'm getting closer to doing that.

I tried the irqfixup and my laptop failed to boot 2/3 times, telling me that there was a problem with the hda not being ready
[ 182.906747] ide: failed opcode was: unknown
[ 182.906750] hda: drive not ready for command
it has booted now, but I still have no wireless chip in the lspci, which I would assume is due to the noapic option. In case it helps I'm running the following kernel: 2.6.22-11-generic #1 SMP Fri Sep 7 04:31:16 GMT 2007 x86_64 GNU/Linux.
Jack

i haven't tried it yet but it seems to be the response to our problem..
if it really works, i'll comment again immediately..

i have had some search results and when i've had a look at them; i've seen anybody, who has the problem, uses a laptop with amd turion x2 cpu - branded as "hp"..

http://wiki.cchtml.com/index.php/Ubuntu_Dapper_Installation_Guide#Black_Screen_when_supposed_to_see_login_screen

and you (i mean "we") should initially boot up with "noapic" kernel parameter by live cd..

Caliphra, do you mean that we should try the: Option "BusType" "PCI" for the fglrx? I have an nvidia card, and fglrx is for ati I believe, so unfortunately this won't help.
But yes, I do have an X2 turion, which seems to be the main problem with the apic which dual cores need.
I got my usb ports working by using the irqpoll option today, which has been letting my notebook boot most of the times, but still have no wireless card.
If anyone has any advice of how I can get my broadcom card working I would much appreciate it. For some reason it even disappeared from the lspci output.
I'm currently booting with noapic nolapic irqpoll
Thanks

I have been seeing some intermittent issues with my DVD drive when booting with "noapic irqfixup". I have switched back to "noapic noirqdebug" as my best boot options.

Robert, I had some intermittent problems with my dvd drive as well; building my own kernel with the old IDE drivers disabled, and the experimental libata drivers enabled, fixes it. I still need to boot with "noapic noirqdebug" however.

I checked out the ubuntu-gutsy kernel tree using git. I compiled a kernel using the make-kpkg method, installed the resulting .deb, and verified that it required noapic to boot correctly. This was done by roughly following the instructions here for a git build:
  https://help.ubuntu.com/community/Kernel/Compile
After getting the tools installed, this was roughly:
  add my user to the "src" group
  logout/login again for this change to take effect
  cd /usr/src
  git clone git://kernel.ubuntu.com/ubuntu/ubuntu-gutsy.git ubuntu-gutsy
  ln -s linux ubuntu-gutsy
  cd ubuntu-gutsy
  cp /boot/config-2.6.22-11-generic .config
  make menuconfig
  remove kernel debugging from options (in kernel hacking)
  save new .config (by exiting menuconfig)
  # the build command procedure using AUTOBUILD=1 didn't seem to work for me.
  export CONCURRENCY_LEVEL=4
  make-kpkg --rootcmd fakeroot --initrd --append-to-version=-test kernel-image kernel-headers
  cd ..
  sudo dpkg -i linux-image-blah-blah-blah.deb
attempt to boot new kernel without noapic
  Since I am able to verify whether the kernel boots correctly without using all the restricted modules, I didn't bother trying to compile them. My plan is that when I get a kernel to boot without noapic, I will also run powertop to verify that the ehci interrupts are under control. After rebooting back into my stock gutsy kernel, I did this:
  cd /usr/src/ubuntu-gutsy
  git bisect start
  git bisect bad
Then I tried to build 2.6.15, since that kernel seemed to boot ok without noapic from the SimplyMEPIS liveCD:
  git checkout v2.6.15
  # git complained about some file being modified or something, so I did this...
  git checkout -f v2.6.15
  added -fno-stack-protector to CFLAGS in top-level build Makefile
  cp /boot/config-2.6.22-11-generic .config
  make menuconfig
  removed kernel debug symbols and saved .config again
  # Unfortunately, trying to build with make-kpkg immediately failed. The only way I could build was like this:
  make clean
  make
  sudo make modules_install
  cd /boot
  sudo cp /usr/src/ubuntu-gutsy/arch/x86_64/boot/bzImage vmlinuz-2.6.15
  sudo update-initramfs -c -k 2.6.15
  # Unfortunately, update-initramfs complained that udev hooks needs kernel 2.6.17.
  # So I was unable to try booting 2.6.15. I decided to see if 2.6.17 would work.
  cd /usr/src/ubuntu-gutsy
  git checkout v2.6.17
  added -fno-stack-protector to CFLAGS in top-level Makefile
  cp /boot/config-2.6.22-11-generic .config
  make menuconfig
  removed kernel debug symbols from build
  # make-kpkg still failed immediately, so reverted to old-school style:
  make clean
  make
  sudo make modules_install
  cd /boot
  sudo cp /usr/src/ubuntu-gutsy/arch/x86_64/boot/bzImage vmlinuz-2.6.17
  sudo update-initramfs -c -k 2.6.17
  sudo vi grub/menu.lst
  added lines for 2.6.17 above "BEGIN AUTOMAGIC KERNELS LIST"
  rebooted into 2.6.17 kernel without "quiet splash noapic noirqdebug" boot flags

Unfortunately, the 2.6.17 boot consistently just hung shortly after initializing USB, although the keyboard was still responsive. Sometimes the last message was about "Uniform CD-ROM driver" and other times it was from the USB subsystem. I tried...

I have a dv9000 with turion 64 X2. Running 64 bit Ubuntu, kernel 2.6.20-16 with the same results.

I have tried all suggested alternatives, but I am currently booting with " noapic nosmp", which sacrifices one core. However doing this resolves the interrupt problem, which lowers CPU usage to 3% when idle, and allows frequency scaling to work (and hence makes the system usable as a laptop).

Everything I have tried works, (including wireless through ndiswrapper); so until this issue is resolved (kernel is fixed upstream), I will continue to boot with nosmp. Without this option the CPUs get wasted servicing interrupts, so disabling one CPU actually makes the system faster (and usable!).

I confirm this with hp pavillion dv6000. IT is happenining since edgy. I also heard this happens on Suse too.

Same here, most of idle time is wasted on irqs.

HP dv6328eu /Gutsy-2.6.22-14/amd64 with noapic irqpoll boot options.

i have tried to 'start kubuntu with safe graphics' and it didn't encounter a problem like 'just a black screen'..
then i installed kubuntu on hard drive while running in safe graphics mode.. after reboot, kubuntu loaded with no problem..
but whenever i alter the screen resolution to 1280*800, the black screen comes again and after any boot, it says 'hi' to me; i cannot escape it..

i think it could be a problem related to vertical refresh rate of the screen.. 'cause i have had a glance at the refresh rate of screen on windows vista and it is 59hz.. on kubuntu, there is no option as 59hz.. briefly, it could be a problem 'caused by monitor driver or gpu driver..

thanks for all comments..

This also happens in my HP Pavilion dv6230br, which has a 2.0 ghz AMD Turion 64 MK-36 (single core!).

It happens on both Feisty and Gutsy Ubuntus, Opensuse 10.3 and Fedora 7. So far the only distro that booted flawlessly out-of-the-box was Mandriva One (Spring 2007), although the broadcom wireless didn't work with it.

I have an HP Pavilion dv9398eu, amd turion64 x2 and I got the exactly same problem. I have also the latest BIOS installed. (jully 2007). What I found interesting is that I have a friend that have "exactly" the same computer but with an intel processor instead and doesn't got the problem. I found also on the net (http://www.networkworld.com/community/node/6340) that there have been a lot of issues on the HP bios/Amd configuration. So is that a bug in the HP bios or is that a bug in the linux kernel or in the AMD processor? Another question I ask myself is why apic woks under vista, did they got a patch? I asked HP and they said, of course, that "HP does not recommend you to change the Operating system on your notebook". Asking about the bug, the response was "Hence we do not have any information or tested information on this"... Puting some pressure on HP could be good if the bug is actualy in the bios!

I have tried to disable rtc and using vesa console driver. I don't have any freezing problems anymore. Have anyone else tried this?

I have the same problem on my Compaq 6715s, but I tested Kernel 2.6.24-rc3 and it looks like this bug is fixed there !
Maybe this affects the Pavilion series too !

I disagree, I compiled 2.6.24-rc3 twice, and it still hangs up where it used to (restricted modules). I have a dv611?us turion X2. Also, take into account that if you compile your own kernel you will lose a lot of support from your distro, I have no sound, no wifi, and had to install the nvidia driver again. Just my 2 cents, I hope this bug gets fixed some time. Jack

I have HP Pavillon dv6120eu and i have the same problem for install ubuntu (7.04 or 7.10) BLACK SCREEN
But install and option "noapic nolapic" it's OK for me

I have the same problem reported here with my HP pavillion dv6000z with AMD Turion64 X2 TL-64. and bcmwl5 with ndiswrapper for wireless.
The issue is that have to use noapic to boot otherwise it does not boot. Disabling apic start lots of irq conflicts and if i don't use irqfixup ir irqpoll, after 5 minutes the the irq's associated with usb's are disbled (that means the webcam stop working, and any hotplugable usb device is no longer detected), all usb devices connected from the begining continue to work, like my usb mouse for example. Now I use noapic irqpoll as boot options. Using irqpoll has a down side that is I can't use the nvidia restricted drivers, because the machine hangs when the display sends many irq's together (i.e. when fast displaying data).
The root of the problem is the use of noapic. I've been searching for a while and found no solution. Hope this is solved and willing to try other options.

Had the same problem running 32-bit Ubuntu on a dv6000 - actually a dv6105ca, AMD Turion64 X2 TL-64. Switched to 64-bit Ubuntu and everything works perfectly, including the native wireless driver, ACPI, sleep, hibernate, NVidia drivers, etc. Try using amd64 Ubuntu and see if it helps you.

I'm running Ubuntu 7.10 amd64 on a dv6125om and must boot with noapic
noirqdebug. Unfortunately it seems there are different things that work for
different variants of these laptops. For myself and some others, amd64 is
not the magic bullet. When I do that, all devices works ok, but I have a
high rate of interrupts on one processor.

On Dec 23, 2007 6:14 PM, Zin Slan <email address hidden> wrote:

> Had the same problem running 32-bit Ubuntu on a dv6000 - actually a
> dv6105ca, AMD Turion64 X2 TL-64. Switched to 64-bit Ubuntu and
> everything works perfectly, including the native wireless driver, ACPI,
> sleep, hibernate, NVidia drivers, etc. Try using amd64 Ubuntu and see if
> it helps you.
>
> --
> HP Pavilion dv6103eu laptop - all Ubuntu versions gets frozen during X
> startup until I specify "noapic" boot option
> https://bugs.launchpad.net/bugs/89746
> You received this bug notification because you are a direct subscriber
> of the bug.
>

--
Robert W. Brewer

I suffer the same problems of other pavilion Turion64 X2 users: system can be started only with noapic option (sometimes it can boot without it, but only if I use the NVIDIA proprietary drivers: in that case is very unstable, expecially during the boot and the shutdown).
Using the noapic boot option USB does not work well: only peripherals already connected before Ubuntu is booted do their works, all the others, if plugged when Ubuntu is running, are simply and completely ignored. For example: if i plug an USB memory key when the computer is running, that is completely ignored and is impossible to mount it. If i plug it when the computer is switched off, and then I boot, Ubuntu is able to mount it. But even in that case, if i umount the key and unplug it, then there will be no way to remount it when re-plugged. The same occurs with USB mouses: if i plug it when Ubuntu is already booted, the mouse is ignored. Only a peripheral already connected before the boot is recognized and configurated.

here is my "cat /proc/interrupts" :
           CPU0 CPU1
  0: 380780 12695 XT-PIC-XT timer
  1: 5418 186 XT-PIC-XT i8042
  2: 0 0 XT-PIC-XT cascade
  5: 75173 3535 XT-PIC-XT sata_nv
  7: 34838 955961 XT-PIC-XT ehci_hcd:usb1
  8: 4 3 XT-PIC-XT rtc
  9: 1888 238 XT-PIC-XT acpi, HDA Intel
 10: 180282 7640 XT-PIC-XT bcm43xx, eth2
 11: 299800 10441 XT-PIC-XT ohci_hcd:usb2, ohci1394, sdhci:slot0
 12: 70 53 XT-PIC-XT i8042
 14: 13316 585 XT-PIC-XT ide0
NMI: 0 0
LOC: 12678 380637
ERR: 891921 <---------------------STRANGE....
MIS: 0

From my dmesg:
....
[ 0.632000] PCI: BIOS BUG #81[49435000] found
.....
[ 129.776000] irq 7: nobody cared (try booting with the "irqpoll" option)
[ 129.776000] [<c015b594>] __report_bad_irq+0x24/0x80
[ 129.776000] [<c013972e>] __rcu_process_callbacks+0x11e/0x1d0
[ 129.776000] [<c015b852>] note_interrupt+0x262/0x2a0
[ 129.776000] [<f88f96c2>] usb_hcd_irq+0x22/0x60 [usbcore]
[ 129.776000] [<c015aab0>] handle_IRQ_event+0x30/0x60
[ 129.776000] [<c015c46b>] handle_level_irq+0xdb/0x110
[ 129.776000] [<c0106b1b>] do_IRQ+0x3b/0x70
[ 129.776000] [<c0118735>] smp_apic_timer_interrupt+0x55/0x80
[ 129.776000] [<f884f74b>] acpi_processor_idle+0x0/0x425 [processor]
[ 129.776000] [<c0105223>] common_interrupt+0x23/0x30
[ 129.776000] [<f884f74b>] acpi_processor_idle+0x0/0x425 [processor]
[ 129.776000] [<c011dba2>] native_safe_halt+0x2/0x10
[ 129.776000] [<f884f6d1>] acpi_safe_halt+0x1c/0x29 [processor]
[ 129.776000] [<f884f8dd>] acpi_processor_idle+0x192/0x425 [processor]
[ 129.776000] [<f884f74b>] acpi_processor_idle+0x0/0x425 [processor]
[ 129.776000] [<c0102413>] cpu_idle+0x53/0xe0
[ 129.776000] =======================
[ 129.776000] handlers:
[ 129.776000] [<f88f96a0>] (usb_hcd_irq+0x0/0x60 [usbcore])
[ 129.776000] Disabling IRQ #7
[ 131.000000] Clocksource tsc unstable (delta = -172025479 ns)

....

The patch in this message may help: http://lkml.org/lkml/2007/12/15/130
Be sure to insert your own DMI IDs; then try booting without noapic.

I'd try it on my own laptop, but it's away for warranty repair at the moment.

I've been following this thread for a long time now and although I am not an expert I wanted to share a relative success I've had in order to boot without the noapic option.

This is part of the message to Leann Ogasawara in the thread https://bugs.launchpad.net/ubuntu/+source/linux/+bug/160718

I wanted to share it with you guys in hope it might work for you too. I'm very thankful to you all for this thread, and felt I owed you something. Best of luck

----------------------------------------------------------------------------------------------------------------------------------------------------------------------

I've been partially successful in booting without the noapic option using the newer kernel 2.6.24

My system is a rare HP dv9343eu (Basically an AMD Turion X2 Broadcomm 4312 wireless card (as detected by linux))

While on Feisty or Gutsy (using 2.6.22) I had to specify noapic noirqdebug in order to get a somewhat stable system. That is the combination that best worked for me, because it booted and gave me USB support [although I don't really know whether it was USB 2.0]

Anyhow, I followed your tip and adventured myself into upgrading the kernel manually to linux-image-2.6.24-3-generic (instead of linux-image-2.6.24-1-generic you suggested). At fisrt that made no difference for me, so I decided to make sure and proceeded to install Hardy through a " update-manager -d". I must point out that before proceeding I had uninstalled bcm43xx-fwcutter and the bcm43xx firmware from the restricted firmware manager.

When I rebooted I was glad to see that even though it took considerably a longer time (precisely there was a very long idle time when loading Hardware drivers) it managed to boot up fine!.

I got a message telling me that my wlan card (detected this time as a Bcm4311) was not properly configured and that I needed to install the v.4 firmware for the b43 cards. So I went ahead and did that. First using the Restricted Manager (which installed a non-working bcm43xx-fwcutter) and then manually installing the b43-fwcutter which correctly enabled my wlan and even allowed me join a wireless network without needing ndiswrapper, wcid, etc...

Problem is that when I rebooted the comp, it once again refused to boot properly.

In Conclusion:

I am finally able to boot withouth noapic using 2.6.24 if I don't have any firmware installed for my wireless card. Other things such as suspend/resume/etc still don't work. I still have to try USB, but CPU Freqs seem fine....

Please let me know how I could provide you guys with more ideal information. I so want this to work that I am keen on trying whatever it takes to fix it..

Regards,

alej

It looks like there is a patch in 2.6.24-git21 (or earlier) for this problem.

Have a look at http://bugzilla.kernel.org/show_bug.cgi?id=8870 or the changelog for the latest kernel at www.kernel.org and search for io_delay

I tested Hardy Heron Alpha 4 and it booted without any options, although it did stall for quite a while after renaming the WLAN, if anyone else would like to try it perhaps we could finally close this bug,
kudos, Jack

I have tried Hardy Heron Alpha 4 too. It booted without problems (no extra options), but running hwclock makes the computer freeze.

I really think the patch in 2.6.24-git21 (now 2.6.25-rc1) is necessary. Is there any chance that 2.6.25 will be in Hardy Heron?

I tried the iodelay patch some time ago, and while the laptop booted, once X started it still froze. I recall reading that X uses the same delay hack, so there might need to be some fixes there? Not sure. I'll try again with -rc1 shortly (I don't remember if I was on the open source nv driver or the closed source nvidia at the time)

Marcus, I tried installing RC1 (it compiled ok) but when I did the dpkg install of the new kernel I got a file not found (something about drivers), then a process is busy, then a failure... I noticed later that there was a dpkg update, but when I try a dpkg reinstall it tries to reinstall the kernel whith the same results... Any ideas? I think that I need to reinstall dpkg... anyways, I know this isnt the right place to post but someone here might have an idea, thank you, Jack

I didn't use dpkg. I did "make" "make modules" "make install" and "make modules_install" (with -j4 option) and then I had to modify /boot/grub/menu.lst. The kernel will install it self in /boot and /lib/modules so its not an difficult task to remove it later.

I managed to compile linux-image-2.6.24-8-generic with the io_delay patch for hardy (64bit) and it worked fine for me. Write me an email and I will send you the .deb file.

Hi Guys,

Examining the upstream bugzilla there were the upstream commits referenced. One of which is in the latest Hardy Alpha kernel:

commit c399da0d97e06803e51085ec076b63a3168aad1b
Author: David P. Reed <email address hidden>
Date: Wed Nov 14 17:47:35 2007 -0500

    x86: fix freeze in x86_64 RTC update code in time_64.c

The other two are not yet in Hardy:

commit b02aae9cf52956dfe1bec73f77f81a3d05d3902b
Author: Rene Herman <email address hidden>
Date: Wed Jan 30 13:30:05 2008 +0100

    x86: provide a DMI based port 0x80 I/O delay override.

commit f9fc58910ebc448b0b7d37af1bf57a896a78e9c4
Author: Ingo Molnar <email address hidden>
Date: Wed Jan 30 13:30:05 2008 +0100

    x86: add DMI quirk for io-delay hangs on Compaq Presario V6000 laptops

If you can, please test the latest 2.6.24-11 Hardy kernel and verify if you are still experiencing issues. If you still notice issues, care to test applying the two commits still not yet pulled into Hardy and verify if one or both are needed to resolve your issue. I can then have the kernel team take and look and consider pulling either one or both them in for the final Hardy release. Otherwise they'll automatically be released with Hardy+1 - the Intrepid Ibex. Also please note we'll keep this report open against the actively developed kernel but this will be closed against 2.6.20. Thanks!

https://bugs.launchpad.net/ubuntu/+source/linux-meta/+bug/200057

This user says problem is fixed on his computer.
I am going to try this patch myself, this problem has been around way too long and needs to end.

Fix committed! See LP#200057.

Even after boot with the described option removed the system freezes during install somewhere after installing all the packages. There are few steps after the packages are installed and I have it frozen on each of them in separate attempts. It could be related to the IRQ remapping problem reported previously or in another bug.

Fixes for bug 200057 went into the 2.6.24-14.24 . Can someone please confrim that a 2.6.24-14-.24 kernel or later version resolves this issue. 'cat /proc/version_signature' to verify the version of the kernel you are running. Thanks.

Ubuntu 8.04 with kernel 2.6.24-16.30 now works with no modifications for me! (never tried 2.6.24-14.24) Great job! I have a pavilion dv9000.

Ubuntu 8.04 works well on new Amd Turion 64x2 HP Pavilion DV9000 without noapic since kernel version 2.6.24-14.24 !
Thank you very much! Great job.

Ubuntu 8.04 also corrects this problem for me too. I am able to boot
without special boot flags, and I have both cores functioning, USB works
well, CPU speed stepping, etc. I'm running an HP Pavilion dv6125om.

Marking this "Fix Released" per the last few comments. Thanks.

Marking this Fix Released for the Baltix kernel. Mantas, feel free to change this if this isn't correct.

Per a decision made by the Ubuntu Kernel Team, bugs will longer be assigned to the ubuntu-kernel-team in Launchpad as part of the bug triage process. The ubuntu-kernel-team is being unassigned from this bug report. Refer to https://wiki.ubuntu.com/KernelTeamBugPolicies for more information. Thanks.