| View previous topic :: View next topic |
| Author |
Message |
magnesium
Apprentice
Joined: 28 Oct 2003
Posts: 280
Location: Toronto, Canada
|
 Posted: Tue Oct 28, 2003 4:58 pm Post subject: Kernel Panic installing Beta AMD 64 Gentoo full install |
 |
|
I've experienced a kernel panic while begining installation of the beta AMD 64 full install. My system specs are at the end.
The Goal of the install was to get Gentoo Linux for AMD64 to install and work with existing Windows 2000 partitions in RAID 0 Array.
The steps taken:
Downloaded and burned gentoo-amd64-2.6-full-20031024.iso to CD-R
Booted from CD and typed in gentoo doataraid for support of promise ata raid controller. Note that the exact same error occurred (including same registry and pointer values) with a host of different boot options.
after flashing through a whole bunch of normal kernel initialization info it failed just after using anticipary io scheduler and ide0 at 0xblahblahblah on irq14, displaying the following. Keep in mind that this is all that could be seen on a non-scrollable screen.
| Code: | Oops: 0010 [1]
CPU 0
Pid: 0, comm: swapper Not tainted
RIP: 0010:[<000000008010f784>] [<000000008010f784>]
RSP: 0000:ffffffff80485fd8 EFLAGS: 00010246
RAX: 0000000000000000 RBX: ffffffff8010f784 RCX: 0000000000000000
RDX: 000001003fb240c0 RSI: ffffffff803a3ba0 RDI: ffffffff803a37e0
RBP: 0000000000000000 R08: ffffffff80484000 R09: 000001003fa51e9c
R10: 0000000000000078 R11: 0000001003fb240c0 R12: 0000000000000000
R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000000
FS: 0000000000000000(0000) GS:ffffffff8047dd40(0000) knlGS: 0000000000000000
CS: 0010 DS: 0018 ES: 0018 CR0: 000000008005003b
CR2: 000000008010f784 CR3: 0000000000101000 CR4: 000000000000006a0
Process swapper (pid: 0, stackpage=ffffffff803a4ba0)
Stack: ffffffff8010f80a 000000000000010a ffffffff8048972a ffffffff80448500
0000000000000000
Call TraceL<ffffffff8010f80a>{cpu_idle+26} <ffffffff8048972a>{start_kernel+378}
Code: Bad RIP value.
RIP [<000000008010f784>] RSP <ffffffff80485fd8>
CR2: 000000008010f784
<0>Kernel panic: Attempted to kill the idle task!
In idle task not syncing
|
Any sort of explaination at where this is failing could be helpfull. My personal guess is that it has to do with my promise ATA RAID drives failing to be detected and resulting in a kernel panic.
System Specs:
AMD Athlon 64 3200+
Asus K8V Deluxe Motherboard w/ Onboard
- 3Com Gigabit Ethernet NIC (3c940)
- Promise SATA RAID controller
- Promise ATA RAID controller (FastTrack 378)
- 6.1 Channel Sound Card
- IEEE 1392 (FireWire) controller
- USB 2.0 controller
- Standard IDE and Floppy stuff
ATI Radeon 9800 Pro video card
LG 32x16x10x CD-RW drive
NEC (Masushita) 32x CD-ROM Drive
2 40GB ATA 133 7200 RPM Maxtor Hard Drives in RAID 0 Array
1GB PC400 CL3 RAM
Logitech Mx700 Mouse connected via USB port
Logitech wireless keyboard connected via PS2 port |
|
| Back to top |
|
 |
Yarrick
Bodhisattva
Joined: 05 Jun 2002
Posts: 304
Location: Malmö, Sweden
|
| Posted: Tue Oct 28, 2003 6:15 pm Post subject: |
|
|
| i am trying to help my friend to run 64bit gentoo on his new athlon64. we got the same error. |
|
| Back to top |
|
|
magnesium
Apprentice
Joined: 28 Oct 2003
Posts: 280
Location: Toronto, Canada
|
| Posted: Tue Oct 28, 2003 7:47 pm Post subject: |
|
|
Is your friend running an ATA RAID like I am as well, or is s/he running a standard IDE drive in non-raid configuration?
Any info you can provide might help us both troubleshoot and get back to the developer where the bug / uncompatible area is. |
|
| Back to top |
|
|
Yarrick
Bodhisattva
Joined: 05 Jun 2002
Posts: 304
Location: Malmö, Sweden
|
| Posted: Wed Oct 29, 2003 12:26 am Post subject: |
|
|
we were using regular ide disks. we settled so far with 32 bit, running 2.6.0-test9 with onboard sound and net working. getting accelerated 3d from a radeon 9600 is our next task.
we use the same motherboard, asus k8v, with an athlon64 3200+. 512mb ddr400 ram, 40 gb ide disk. |
|
| Back to top |
|
|
linkan
n00b
Joined: 31 Aug 2003
Posts: 8
Location: Stockholm, Sweden
|
| Posted: Thu Oct 30, 2003 1:04 pm Post subject: |
|
|
I had the same problem, turned out that disabling legacy USB support in the BIOS took care of the problem. At least for me.
MB: Asus K8V Deluxe
CPU: Athlon 64 XP 3200+
RAM: 2x512 Crucial matched
guide:
Enter BIOS
go to Advanced -> USB Configuration
Legacy USB Support -> Disable
press F10, answer Yes.
Please reply if it works for you to.
Any ideas as to why this happens?
/Linkan |
|
| Back to top |
|
|
riprjak
Tux's lil' helper
Joined: 26 Apr 2003
Posts: 116
Location: Adelaide, Au
|
| Posted: Wed Nov 19, 2003 1:32 am Post subject: |
|
|
| linkan wrote: | I had the same problem, turned out that disabling legacy USB support in the BIOS took care of the problem. At least for me.
MB: Asus K8V Deluxe
CPU: Athlon 64 XP 3200+
RAM: 2x512 Crucial matched
guide:
Enter BIOS
go to Advanced -> USB Configuration
Legacy USB Support -> Disable
press F10, answer Yes.
Please reply if it works for you to.
Any ideas as to why this happens?
/Linkan |
Mate!!
you are a bloody genius. Problem solved; and it appears that the kernel panic was at mouse detection.
I will go back and try legacy support "no mice" setting (so my USB keyboard will work in grub). Ill keep in touch on this!
EDIT
I have tried enable, disable and "no mice" setting for legacy usb support in bios; the Live CD boots only with "disable"... once again, cheers to linkan for spotting this, I never woulda thunk it  Course, now I have to get a usb/ps2 adaptor, plug the usb keyboard in to go through grub, then whack it back in the USB port before linux gets to loading, but Ill see if I cant build a kernel that doesn't bork with legacy support enabled.
/EDIT
EDIT the second
I reported this to kernel bugzilla and was told it relates to a CPU bug which requires a bios workaround and does indeed relate to the "USB legacy mode". I haven't seen a bios upgrade in the wild for my MSI KT8 neo yet.
/EDIT
Cheers,
jak
_________________
---
$> cd /pub
$> more beer > /home/riprjak/.mouth
Last edited by riprjak on Thu Nov 27, 2003 10:51 pm; edited 1 time in total |
|
| Back to top |
|
|
riprjak
Tux's lil' helper
Joined: 26 Apr 2003
Posts: 116
Location: Adelaide, Au
|
| Posted: Thu Nov 20, 2003 12:14 am Post subject: |
|
|
| Yarrick wrote: | we were using regular ide disks. we settled so far with 32 bit, running 2.6.0-test9 with onboard sound and net working. getting accelerated 3d from a radeon 9600 is our next task.
we use the same motherboard, asus k8v, with an athlon64 3200+. 512mb ddr400 ram, 40 gb ide disk. |
 Let me know how you go with the radeon 9600... Im struggling right now.
I have Athlon64 3200+, MSI K8T Neo FISR2, 120GB EIDE HDD, 40 GB EIDE HDD (no raid), Liteon 48,24,48,16 CDR/RW/read/DVD, Powercolour Radeon9600pro, 1 GB (2 x 512MB PC3200) SDRAM, SB Audigy, Netgear 802.11b PCI card. Ive also got a nobeless 6in1 usb card reader in place of an FDD. I intend to get all of these bits working (and get the USB legacy support NOT crashing the os... as I have a MS wireless office usb kb/mouse combo)... Ill post configs and such when I achieve this lofty goal
I *was* most of the way through a stage 2 build when I left for work this morning... man it builds fast Prior to the rebuild, I was running in in an Athlon-XP optimised 32 bit environment and had much trouble getting the AGPGART to work (I got an error "too many northbridges for GART" error in kernel log, pesky on chip garp not playing nice with the via northbridge!!); so I am not optimistic about geting DRI running quickly.
Oh yeah, I played a bit with compiling gcc last night and it looks like the Athlon64 (as a single processor) gets benefit from -j3 and -j4 settings for make... -j5 caused a significant slowdown (I really shoulda written down the "time" output, but I didn't). I settled on -j3 and it is rocketing along.
EDIT
"rocketing along" = emerge -e world (stage 3 system + metalog + vcron + wireless-tools + gnome 2.4, 222 packages in all) in 4.5 hours!!! so tasty.
/EDIT
err!
jak.
_________________
---
$> cd /pub
$> more beer > /home/riprjak/.mouth |
|
| Back to top |
|
|
longint
Tux's lil' helper
Joined: 27 Aug 2002
Posts: 100
Location: good old germany
|
| Posted: Sun Nov 30, 2003 4:46 pm Post subject: onboard sound working? |
|
|
Acknowledged. Also using an k8v an disabling USB Legacy does the trick. BTW: There is just enable, disable or auto, nothing like just mouse or keyboard (bios 1003).
An other point: Someone get the onboard sound working? If so, please let me know your kernel config (for sound) und modules configuration (if needed). I've tried oss and alsa, but none of them worked for me.... |
|
| Back to top |
|
|
linkan
n00b
Joined: 31 Aug 2003
Posts: 8
Location: Stockholm, Sweden
|
| Posted: Sun Nov 30, 2003 6:32 pm Post subject: Re: onboard sound working? |
|
|
| longint wrote: |
An other point: Someone get the onboard sound working? If so, please let me know your kernel config (for sound) und modules configuration (if needed). I've tried oss and alsa, but none of them worked for me.... |
I haven't gotten that far yet, actually i can't even remember if the liveCD had sound. It's not my machine so I can't really play around as much as I would like.
But, on to other issues, does anyone have OpenOffice or Abiword up and running? My OO just bugs out complaining about Java... So it's actually a java issue, I'm running blackdown. I'll update with the errors. |
|
| Back to top |
|
|
longint
Tux's lil' helper
Joined: 27 Aug 2002
Posts: 100
Location: good old germany
|
| Posted: Sun Nov 30, 2003 8:08 pm Post subject: |
|
|
no probs with OO so far, but with sound...
Which java/jdk do you use? I guess blackdown is the right for you... |
|
| Back to top |
|
|
longint
Tux's lil' helper
Joined: 27 Aug 2002
Posts: 100
Location: good old germany
|
| Posted: Fri Dec 05, 2003 10:51 pm Post subject: |
|
|
Might it be of help for anyone else:
sound is working now using alsa and unmuted pcm and surround |
|
| Back to top |
|
|
riprjak
Tux's lil' helper
Joined: 26 Apr 2003
Posts: 116
Location: Adelaide, Au
|
| Posted: Fri Dec 12, 2003 1:58 am Post subject: |
|
|
I have discovered another work around for the legacy USB bug; It is a nasty hack; but set "idle=poll" as a kernel flag.
This bybasses the use of HLT, which is the cause of the top 32bits being zeroed in SMM mode. Of course, your CPU will use more power and run hotter... JFYI really
err!
jak.
_________________
---
$> cd /pub
$> more beer > /home/riprjak/.mouth |
|
| Back to top |
|
|
fo0bar
n00b
Joined: 15 Aug 2003
Posts: 4
|
| Posted: Sat Dec 13, 2003 5:46 am Post subject: |
|
|
| FYI, the bios workaround is also needed on the new eMachines T6000. Thanks for the info! |
|
| Back to top |
|
|
DarkBlackFox
n00b
Joined: 13 Jun 2003
Posts: 12
Location: Connecticut
|
| Posted: Wed Dec 17, 2003 4:43 am Post subject: |
|
|
I've had this same problem on an MSI K8T Neo FISR2. Legacy USB support was already disabled in the BIOS, and it does the same thing even passing the nousb and nohotswap params to the kernel.
Any other solutions out there? |
|
| Back to top |
|
|
riprjak
Tux's lil' helper
Joined: 26 Apr 2003
Posts: 116
Location: Adelaide, Au
|
| Posted: Wed Dec 17, 2003 10:50 pm Post subject: |
|
|
| DarkBlackFox wrote: | I've had this same problem on an MSI K8T Neo FISR2. Legacy USB support was already disabled in the BIOS, and it does the same thing even passing the nousb and nohotswap params to the kernel.
Any other solutions out there? |
I use exactly the same board; and it was the legacy usb setting alone that caused the crash... you may have a bigger problem. Do you have any USB devices connected?? ram type?? (does it pass a memtest??) power supply brand/wattage??
BTW I have used both bios 1.0 and 1.1 on the same board.
Try the "idle=poll" kernel parameter and double check the USB Legacy is set to "Disabled" and not "No Mice" or "Enabled". I say doublecheck as the usb legacy setting should be the only cause of this smm mode bug (if you arent running in smm mode, it *shouldn't* occur...).
Otherwise try a newer livedisk; may help...
err!
jak.
_________________
---
$> cd /pub
$> more beer > /home/riprjak/.mouth |
|
| Back to top |
|
|
DarkBlackFox
n00b
Joined: 13 Jun 2003
Posts: 12
Location: Connecticut
|
| Posted: Thu Dec 18, 2003 6:14 pm Post subject: |
|
|
I didn't have a boatload of time to play with it, but I did triple check that Legacy USB support was disabled. If I have more time tonight I'll try it with USB completely disabled in the bios. Now that you mention it, I do have a few things plugged into the USB ports- mouse, iPaq cradle, and a joystick. I'll have to try with those unplugged.
I know the RAM isn't a problem, it's genuine Samsung, and has passed memtest overnight multiple times. I have one gig of memory, and I know that amount has had some issue with kernels in the past (2.4.x, had to configure the large memory settings in the kernel when compiling.. never got it to recognize the full one gig, always only saw 640mb or so).
Thanks for the suggestions! I'll post my results when i get home. |
|
| Back to top |
|
|
floffe
Guru
Joined: 24 Nov 2003
Posts: 414
Location: Linköping, Sweden
|
| Posted: Thu Dec 18, 2003 6:31 pm Post subject: |
|
|
You should enable High Memory Support under Processor Type and Features. Setting it to 4GB should do for 1GB RAM + swap. In 2.4 there was a help item on this, but it seems to be gone on my 2.6 menu  |
|
| Back to top |
|
|
DarkBlackFox
n00b
Joined: 13 Jun 2003
Posts: 12
Location: Connecticut
|
| Posted: Fri Dec 19, 2003 3:22 am Post subject: |
|
|
| Yeah, whenever I enabled High Memory Support for 4 gigs, I would get a kernel panic booting. (This was a while back, so I don't remember specifics). But that's besides the point. Any idea when a new livecd will be out with the 2.6.0 final kernel? |
|
| Back to top |
|
|
riprjak
Tux's lil' helper
Joined: 26 Apr 2003
Posts: 116
Location: Adelaide, Au
|
| Posted: Sun Dec 21, 2003 11:54 pm Post subject: |
|
|
| DarkBlackFox wrote: | I didn't have a boatload of time to play with it, but I did triple check that Legacy USB support was disabled. If I have more time tonight I'll try it with USB completely disabled in the bios. Now that you mention it, I do have a few things plugged into the USB ports- mouse, iPaq cradle, and a joystick. I'll have to try with those unplugged.
I know the RAM isn't a problem, it's genuine Samsung, and has passed memtest overnight multiple times. I have one gig of memory, and I know that amount has had some issue with kernels in the past (2.4.x, had to configure the large memory settings in the kernel when compiling.. never got it to recognize the full one gig, always only saw 640mb or so).
Thanks for the suggestions! I'll post my results when i get home. |
I use 1 gig (2 x 512 pc3200), samsung too... So (barring the dreaded static or good old fashioned manufacturing variation) there shouldn't be an issue, unless you have 1 x 512 and 2 x 256 or something; board doesnt play well with all three dimm slots populated.
Also, what kernel version?? I used the 2.6 live cd with the "bootsplash" most recently and wouldn't even *think* of trying 2.4.x after the trouble I had getting it to work with an old via kt400 chipset.
If you need more help, let me know and I will send you my .configs and a breakdown of bios settings...
err!
jak.
_________________
---
$> cd /pub
$> more beer > /home/riprjak/.mouth |
|
| Back to top |
|
|
DarkBlackFox
n00b
Joined: 13 Jun 2003
Posts: 12
Location: Connecticut
|
| Posted: Mon Dec 22, 2003 3:54 am Post subject: |
|
|
I did finally get the livecd to boot! Passing the idle=poll option seems to have done the trick. Didn't even have to disable USB at all. The livecd I'm using is the latest (released in november I believe), and thus far hasn't had any major issue (once I got past the panic).
What exactly does that parameter (idle=poll) do to the kernel? I assume it has to do with idle CPU power, ie, what the system does when it's not doing anything else.
As a side note, the real gentoo installation I work with was in use with an nforce2 based board (MSI K7N2G), before I upgraded to the Athlon 64. Since then, I moved the hard drive from the nforce board to an MSI KT4 Ultra, (sold the nforce board) using the original release of the KT400 chipset, and that has been running like a champ. Booted itself without a recompile, but I needed modules for the IDE bus, sound, and network, so I recompiled anyway. That system only has 512 megs of ram, but it is running happily with kernel 2.4.20. |
|
| Back to top |
|
|
riprjak
Tux's lil' helper
Joined: 26 Apr 2003
Posts: 116
Location: Adelaide, Au
|
| Posted: Tue Dec 23, 2003 1:17 am Post subject: |
|
|
| DarkBlackFox wrote: | What exactly does that parameter (idle=poll) do to the kernel? I assume it has to do with idle CPU power, ie, what the system does when it's not doing anything else.
|
In very basic terms, it tells the kernel not to execute the HTL instruction when the CPU is idle. It has the side effect of making tools like oprofile more effective, but it does increase overall power usage of the CPU. Since the CPU bug that is the root cause of the problem relates to executing HLT in SMM mode, it is a complete work around and no other action is needed.
Essentially it tells the idle loop in the kernel to poll the "need reschedule" flag instead of executing HLT and waiting for an interrupt (AFAIK); so the CPU taps its virtual foot and waits impatiently for something to do rather than quietly meditating until someone disturbs it
However, as long as the SMM mode error causes the top 32bits of a 64bit address to get clobbered by HLT, Im willing to pay the extra power bill.
err!
jak.
_________________
---
$> cd /pub
$> more beer > /home/riprjak/.mouth |
|
| Back to top |
|
|
soth
Apprentice
Joined: 12 Sep 2003
Posts: 207
|
| Posted: Tue Jan 31, 2006 9:28 am Post subject: |
|
|
This looks kinda promising. Will try this. Who wants mouse in grub anyways?
_________________
- Never argue with an idiot. They just drag you down to your level and beat you with experience.
Join the adopt an unanswered post initiative today |
|
| Back to top |
|
|
|
Installing Gentoo
