nvidia 100.14.19 + ipw3945: system freezes

[mlfarrell]

Holy crap. I didn't notitce that until now!

ACPI: MCFG 7FE92F02, 003C (r1 HP NISSAN 6040000 LOHR 5A)
ACPI: BOOT 7FE92FD8, 0028 (r1 PTLTD $SBFTBL$ 6040000 LTP 1)
ACPI: APIC 7FE92F70, 0068 (r1 PTLTD APIC 6040000 LTP 0)
ACPI: SSDT 7FE8C85F, 020A (r1 SataRe SataAhci 1000 INTL 20060217)
ACPI: SSDT 7FE8BED1, 04F6 (r1 PmRef CpuPm 3000 INTL 20060217)
ACPI: BIOS bug: multiple APIC/MADT found, using 0
ACPI: If "acpi_apic_instance=2" works better, notify linux-acpi@vger.kernel.org
ACPI: PM-Timer IO Port: 0x1008
ACPI: Local APIC address 0xfee00000


Have you tried the acpi_apic_instance=2 yet? I might try that later today

Also note the NISSAN compiled DSDT. What the heck? Did a car company compile my dsdt. I use the iasl decompile and recompile test that I found online on these forums and I get like 30 errors about undefined things. Dunno if that has to do with the xid freezes though


~mike

[david_e]

Originally Posted by mlfarrell

Have you tried the acpi_apic_instance=2 yet? I might try that later today

I am keeping my computer uptime as much as I can without the ipw3945 modules to see what happens... I will try it later...

[mlfarrell]

Ok, I'm on the acpi_apic_instance=2 now. will let you know how it works

dmesg|grep -i bios
...
ACPI: BIOS bug: multiple APIC/MADT found, using 2
...


Speaking of bios bug messages. I've always gotten this one below. NVIDIA admins - any idea if this could be a problem?

0000:00:1d.7 EHCI: BIOS handoff failed (BIOS bug ?) 01010001

[netllama]

See: http://lkml.org/lkml/2006/9/11/129

[mlfarrell]

ok, my usb works fine mostly apart from an intermittent freezing ipod annoyance that I can deal with.

Sofar on acpi_apic_instance=2 I'm noticing short pauses when I switch desktops via the "expo" plugin but this is normal as I have many windows open right now firefox, vmware, gkrellm, evolution, and scene editor containing lots of textures and 15k polys, so I think its just pausing to swap vram in and out from the main ram. Perhaps this is what the nv driver was freezing on, and I'm just never coming back from those pauses.

Keeping my fingers crossed.

EDIT: ooh new stuff. still no freeze. When running my game i'm developing 1400x900 under compiz

rtc: lost 1 interrupts
rtc: lost 1 interrupts
rtc: lost 1 interrupts
rtc: lost 1 interrupts
rtc: lost 1 interrupts
rtc: lost 1 interrupts
rtc: lost 1 interrupts
rtc: lost 1 interrupts
rtc: lost 1 interrupts
rtc: lost 1 interrupts
printk: 26 messages suppressed.
rtc: lost 1 interrupts

no noticeable bad behavior during these msgs, (except vmware skipping and popping audio output on itunes which happens sometimes anyway)

[david_e]

You will always see the BIOS bug message, as Linux will always find two APIC/MADT:

http://bugzilla.kernel.org/show_bug.cgi?id=7465

As I didn't had any problems when working without ipw3945 I am testing the acpi option now. I have the same EHCI message...

davide@gentoo ~ $ dmesg | grep -i ehci
0000:00:1d.7 EHCI: BIOS handoff failed (BIOS bug ?) 01010001
ehci_hcd 0000:00:1d.7: EHCI Host Controller
ehci_hcd 0000:00:1d.7: new USB bus registered, assigned bus number 1
ehci_hcd 0000:00:1d.7: debug port 1
ehci_hcd 0000:00:1d.7: irq 20, io mem 0xd2404000
ehci_hcd 0000:00:1d.7: USB 2.0 started, EHCI 1.00, driver 10 Dec 2004

*** EDIT ***
Sorry I didn't see netlama post... thanks for the infos with the EHCI.

[mlfarrell]

Save yourself the trouble. Complete lockup just now and screen shutoff

Sep 27 17:04:08 slickbox kernel: NVRM: Xid (0001:00): 6, PE0000 0400 ff000000 0000f848 00ffffff 00ffffff
Sep 27 17:04:08 slickbox kernel: NVRM: Xid (0001:00): 30, L2 -> L1

Not the fix

Edit: I just realized that the log snippit above doest not have todays date, so I suppose it froze before it got a chance to write to the log. Either way, the acpi_apic_instance thing does not solve the problem

[rmayer]

Greetings,

Don't know if this helps or hinders, but the latest version of iwlwifi works fine on my laptop, not an ibm, but very similar problems and errors according to your posts and googling. I had to fix my dsdt and load it at boot to get past acpi problems (but iwlwifi always worked great), everything works great now, very stable using sidux (with compiz). If anything with iwlwifi, wpasupplicant may have to be configured. Good luck!

[mlfarrell]

I'll switch to iwlwifi and see if the xid's still occur in the nvidia driver. My dsdt is attached though the most recent version of iasl seems to compile it fine (earlier iasl got errors).

slickbox:~ # acpica-unix-20061109/compiler/iasl -c dsdt.dsl

Intel ACPI Component Architecture
ASL Optimizing Compiler version 20061109 [Oct 5 2007]
Copyright (C) 2000 - 2006 Intel Corporation
Supports ACPI Specification Revision 3.0a

ASL Input: dsdt.dsl - 5876 lines, 222927 bytes, 2200 keywords
AML Output: DSDT.aml - 25473 bytes 507 named objects 1693 executable opcodes

Compilation complete. 0 Errors, 0 Warnings, 0 Remarks, 2 Optimizations

[rmayer]

Greetings,

I did find something that will not show up as an error or warning that may help you. I have a Toshiba, Centrino duo, 3945 and everything else pretty much intel, and have seen your yours while looking to fix mine. The relevant sections of our dsdt's are almost identical. Check it out, details here:

[david_e]

Originally Posted by rmayer

Greetings,

I did find something that will not show up as an error or warning that may help you. I have a Toshiba, Centrino duo, 3945 and everything else pretty much intel, and have seen your yours while looking to fix mine. The relevant sections of our dsdt's are almost identical. Check it out, details here:

I think that you could use the acpi_os_name="Windows 2006" option instead of rebuilding your dsdt...

By the way here is mine. I think that it's ok...

davide@gentoo temp $ iasl -tc dsdt.dsl

Intel ACPI Component Architecture
ASL Optimizing Compiler version 20060912 [Oct 5 2007]
Copyright (C) 2000 - 2006 Intel Corporation
Supports ACPI Specification Revision 3.0a

ASL Input: dsdt.dsl - 6436 lines, 251565 bytes, 2433 keywords
AML Output: dsdt.aml - 27188 bytes 552 named objects 1881 executable opcodes

Compilation complete. 0 Errors, 0 Warnings, 0 Remarks, 32 Optimizations

I am tring to see if I get better results with "acpi=noirq" and "irqpoll" to make the system ignore the ACPI IRQ table (maybe there is an issue there), and (from the kernel Documentation):

irqfixup [HW]
When an interrupt is not handled search all handlers
for it. Intended to get systems with badly broken
firmware running.

I am attaching the new dmesg: you can see that there are some error message related with the PCI:

PCI BIOS passed nonexistent PCI bus 4!

....

PCI BIOS passed nonexistent PCI bus 5!
PCI BIOS passed nonexistent PCI bus 5!
PCI BIOS passed nonexistent PCI bus 5!
PCI BIOS passed nonexistent PCI bus 5!
PCI BIOS passed nonexistent PCI bus 5!

....

PCI: Cannot allocate resource region 7 of bridge 0000:00:1c.0
PCI: Cannot allocate resource region 8 of bridge 0000:00:1c.0
PCI: Cannot allocate resource region 7 of bridge 0000:00:1c.1
PCI: Cannot allocate resource region 8 of bridge 0000:00:1c.1
PCI: Cannot allocate resource region 7 of bridge 0000:00:1c.2
PCI: Cannot allocate resource region 8 of bridge 0000:00:1c.2
PCI: Cannot allocate resource region 0 of device 0000:04:00.0
PCI: Failed to allocate mem resource #6:20000@d0000000 for 0000:01:00.0

[mlfarrell]

hmm, I'm still testing with the new with the new wireless driver. Does my kernel 2.6.22.8 allow me to insert custom dsdt's without patching or rebuilding? If it involves messing with the initrd, I'm a little weary as I panicked my kernel on boot before by messing with the initrd image.

I'm curious about this acpi_os_name=Windows2006 option. What does that do? Isn't it bad for the bios to think that we're running windows?