Bug 4169

Summary: Texas Instruments PCI1450 / Airo 340 Kernel freeze
Product: Drivers Reporter: Hadmut Danisch (hadmut)
Component: PCMCIAAssignee: linux-pcmcia
Status: REJECTED INSUFFICIENT_DATA    
Severity: normal CC: bunk
Priority: P2    
Hardware: i386   
OS: Linux   
Kernel Version: 2.6.10 Subsystem:
Regression: --- Bisected commit-id:
Attachments: config.opts
dmesg 2.6.11 rc5
Patch that might fix this problem

Description Hadmut Danisch 2005-02-05 02:30:48 UTC
Distribution: Debian
Hardware Environment:

Compaq Armada M700 (quite old notebook), 686 processor
0000:00:00.0 Host bridge: Intel Corp. 440BX/ZX/DX - 82443BX/ZX/DX Host bridge
(rev 03)
0000:00:01.0 PCI bridge: Intel Corp. 440BX/ZX/DX - 82443BX/ZX/DX AGP bridge (rev 03)
0000:00:04.0 CardBus bridge: Texas Instruments PCI1450 (rev 03)
0000:00:04.1 CardBus bridge: Texas Instruments PCI1450 (rev 03)
0000:00:07.0 Bridge: Intel Corp. 82371AB/EB/MB PIIX4 ISA (rev 02)
0000:00:07.1 IDE interface: Intel Corp. 82371AB/EB/MB PIIX4 IDE (rev 01)
0000:00:07.2 USB Controller: Intel Corp. 82371AB/EB/MB PIIX4 USB (rev 01)
0000:00:07.3 Bridge: Intel Corp. 82371AB/EB/MB PIIX4 ACPI (rev 02)
0000:00:08.0 Multimedia audio controller: ESS Technology ES1978 Maestro 2E (rev 10)
0000:00:09.0 Communication controller: Lucent Microelectronics WinModem 56k (rev 01)
0000:01:00.0 VGA compatible controller: ATI Technologies Inc Rage Mobility P/M
AGP 2x (rev 64)


Cisco Aironet 340 WLAN PCMCIA 



Software Environment:
Debian unstable, kernel 2.6.10, pcmcia-cs 3.2.5
Problem Description:

When system boots with airo card or if card is inserted later
the kernel immediately freezes. No hint when plugging in the card
into the running machine, kernel is dead immediately.

When booting the famous last words are:

cardmgr[1835]: starting, version is 3.2.5
cs: memory probe 0x0c0000-0x0fffff: excluding 0xc0000-0xcffff 0xf0000-0xfffff
cardmgr[1835]: socket 0: Aironet PC4800
done.
cardmgr[1835]: executing: 'modprobe airo_cs'
airo: Probing for PCI adapters
airo: Finished probing for PCI adapters

And that's it.

When trying this several times I once saw that the kernel 
was complaining about an unexpected NMI and telling me, that 
I probably have a hardware problem with my memory (memory works
well under 2.4)



The system works under 2.4.26 (while the airo driver had a little 
bug which sometimes lowers the bandwidht, that's why I wanted to 
upgrade to 2.6). When booting the 2.4.26, it says

cs: IO port probe 0x0100-0x04ff: excluding 0x100-0x107 0x170-0x177 0x370-0x37f
0x4d0-0x4d7
cs: IO port probe 0x0800-0x08ff: clean.
cs: IO port probe 0x0c00-0x0cff: clean.
cs: IO port probe 0x0a00-0x0aff: clean.
cs: memory probe 0xa0000000-0xa0ffffff: clean.

which is obviously a different memory range. 


Is the memory range the problem?

regards
Hadmut









Steps to reproduce:
Comment 1 Dominik Brodowski 2005-02-05 03:52:35 UTC
Did /etc/pcmcia/config.opts change in between? Can you post it current
content here, please? Also, try out removing the line

include memory 0xc0000-0xfffff

and add a line

include memory 0xa0000000-0xa0ffffff

in case the latter isn't already there.
Comment 2 Hadmut Danisch 2005-02-05 09:57:09 UTC
Created attachment 4517 [details]
config.opts

This is the config.opts used for both 2.4 and 2.6
Comment 3 Hadmut Danisch 2005-02-05 09:59:50 UTC
Hi,

the attached config.opts was used in both cases.

Removing the 
include memory 0xc0000-0xfffff

didn't make any difference: I still saw the NMI error message
about memory hardware error and the kernel was dead.

regards
Hadmut
Comment 4 Dominik Brodowski 2005-02-09 12:28:08 UTC
does disabling the airo driver (switching it off in the kernel config, make,
make modules_install, reboot) let the lockup go away?
Comment 5 Hadmut Danisch 2005-02-10 13:50:02 UTC
2.6.10 kernel:

Machine with driver and without card works well.

Machine with card inserted and module airo_cs.ko rename to _airo_cs.ko works 
too (obviously the card does not work without driver).

Machine with card inserted and module in place this time showed
a strange behavior. This time it repeatedly printed error messages from ide bus
that the hda disk did not work anymore for about 2-3 minutes (dma failed).

Then the machine booted normally but without detecting the card.
Since the machine was running this time, I was able to save
dmesg output:

cs: memory probe 0x60000000-0x60ffffff: clean.
airo:  Probing for PCI adapters
airo:  Finished probing for PCI adapters
Uhhuh. NMI received. Dazed and confused, but trying to continue
You probably have a hardware problem with your RAM chips
airo: Max tries exceeded when issueing command
airo: MAC could not be enabled
airo_cs: RequestConfiguration: Operation succeeded
hda: dma_timer_expiry: dma status == 0x61
hda: DMA timeout error
hda: dma timeout error: status=0x58 { DriveReady SeekComplete DataRequest }

ide: failed opcode was: unknown
hda: dma_timer_expiry: dma status == 0x61
hda: DMA timeout error
hda: dma timeout error: status=0x58 { DriveReady SeekComplete DataRequest }

ide: failed opcode was: unknown
hda: status error: status=0x58 { DriveReady SeekComplete DataRequest }

ide: failed opcode was: unknown
hda: drive not ready for command
hda: status error: status=0x58 { DriveReady SeekComplete DataRequest }

... repeated several times ...

ide: failed opcode was: unknown
hda: drive not ready for command
ide0: reset: success


Then the machine was booting. Maybe it's a problem with interrupts.

regards
Hadmut
Comment 6 Hadmut Danisch 2005-02-10 14:02:44 UTC
Sorry, my last comment was a little bit misleading. 
The machine was not booting again, it was continuing with the 
boot sequence and starting X11 normally (just without WLAN).

regards
Hadmut
Comment 7 Dominik Brodowski 2005-02-27 03:40:02 UTC
could you re-test with 2.6.11-rc5, please, and if it fails, post a dmesg here?
Comment 8 Hadmut Danisch 2005-02-27 09:22:46 UTC
Created attachment 4608 [details]
dmesg  2.6.11 rc5

dmesg as requested for 2.6.11 rc5:
System boots and hangs for a while (1-2 minutes or so) in the 
ide failures. Then boots normally and works, except for the 
airo card support. Airo card flashes one LED while booting (just before the
problem occurs), but then gets dark again and does not work.

Same problem as with 2.6.10
Comment 9 Adrian Bunk 2005-11-05 08:02:22 UTC
Created attachment 6479 [details]
Patch that might fix this problem

Can you check whether your problem is a side effect of the bug fixed by this
patch?
Comment 10 Dominik Brodowski 2006-01-10 09:18:15 UTC
I'd be interested to know whether this problem is still present in recent
kernels, like 2.6.15, with or without the patch posted by Adrian in comment #9.
Comment 11 Dominik Brodowski 2006-03-27 13:18:29 UTC
Could you verify whether this problem still exists in 2.6.16, please?
Comment 12 Adrian Bunk 2006-08-01 06:39:21 UTC
Please reopen this bug if it's still present in kernel 2.6.17.