Bug 2846

*** Bug 2799 has been marked as a duplicate of this bug. ***

Just a confirmation. Hardware: nforce2 motherboard (abit nf7s), Athlon XP 2600+, 512Mb RAM, vanilla 
kernel 2.6.7
Initially observed symptoms were that network (forcedeth driver) does not work on kernel 2.6.7 whereas 
it does work on kernel 2.6.6. It seems nothing has changed in the code for forcedeth between 2.6.6 
and 2.6.7 though, looking at /var/log/messages seems to indicate that it could be a problem with the 
ehci_hcd:

Jun 18 09:34:03 geel kernel: ehci_hcd 0000:00:02.2: nVidia Corporation nForce2 USB Controller
Jun 18 09:34:03 geel kernel: ehci_hcd 0000:00:02.2: irq 10, pci mem e08bc000
Jun 18 09:34:03 geel kernel: ehci_hcd 0000:00:02.2: new USB bus registered, assigned bus number 3
Jun 18 09:34:03 geel kernel: irq 10: nobody cared!
Jun 18 09:34:03 geel kernel:  [<c010783a>] __report_bad_irq+0x2a/0x90
Jun 18 09:34:03 geel kernel:  [<c010792c>] note_interrupt+0x6c/0xa0
Jun 18 09:34:03 geel kernel:  [<c0107b2e>] do_IRQ+0xee/0x110
Jun 18 09:34:03 geel kernel:  [<c01060a8>] common_interrupt+0x18/0x20
Jun 18 09:34:03 geel kernel:  [<c011ffb0>] __do_softirq+0x30/0x80
Jun 18 09:34:03 geel kernel:  [<c0120026>] do_softirq+0x26/0x30
Jun 18 09:34:03 geel kernel:  [<c0107b15>] do_IRQ+0xd5/0x110
Jun 18 09:34:03 geel kernel:  [<c01060a8>] common_interrupt+0x18/0x20
Jun 18 09:34:03 geel kernel:  [<c01bda4b>] pci_bus_read_config_byte+0x4b/0x60
Jun 18 09:34:03 geel kernel:  [<e08b88cc>] ehci_start+0x2cc/0x360 [ehci_hcd]
Jun 18 09:34:03 geel kernel:  [<c011cea0>] printk+0x100/0x140
Jun 18 09:34:03 geel kernel:  [<c02631eb>] usb_register_bus+0x14b/0x170
Jun 18 09:34:03 geel kernel:  [<c0267f0b>] usb_hcd_pci_probe+0x2ab/0x4e0
Jun 18 09:34:03 geel kernel:  [<c01c11c2>] pci_device_probe_static+0x52/0x70
Jun 18 09:34:03 geel kernel:  [<c01c121b>] __pci_device_probe+0x3b/0x50
Jun 18 09:34:03 geel kernel:  [<c01c125c>] pci_device_probe+0x2c/0x50
Jun 18 09:34:03 geel kernel:  [<c02113af>] bus_match+0x3f/0x70
Jun 18 09:34:03 geel kernel:  [<c02114d9>] driver_attach+0x59/0x90
Jun 18 09:34:03 geel kernel:  [<c02117d5>] bus_add_driver+0xa5/0xc0
Jun 18 09:34:03 geel kernel:  [<c0211c9f>] driver_register+0x2f/0x40
Jun 18 09:34:03 geel kernel:  [<c01c14dc>] pci_register_driver+0x5c/0x90
Jun 18 09:34:03 geel kernel:  [<e087b023>] init+0x23/0x30 [ehci_hcd]
Jun 18 09:34:03 geel kernel:  [<c0130f2a>] sys_init_module+0x10a/0x200
Jun 18 09:34:03 geel kernel:  [<c0105ee9>] sysenter_past_esp+0x52/0x71
Jun 18 09:34:03 geel kernel:
Jun 18 09:34:03 geel kernel: handlers:
Jun 18 09:34:03 geel kernel: [<c0263ea0>] (usb_hcd_irq+0x0/0x70)
Jun 18 09:34:03 geel kernel: Disabling IRQ #10
Jun 18 09:34:03 geel kernel: ehci_hcd 0000:00:02.2: USB 2.0 enabled, EHCI 1.00, driver 2004-May-10
.
.
.
Jun 18 09:34:03 geel kernel: forcedeth.c: Reverse Engineered nForce ethernet driver. Version 0.25.
Jun 18 09:34:03 geel kernel: eth1: forcedeth.c: subsystem: 0147b:1c00 bound to 0000:00:04.0
.
.
.
(after trying to ping on the network card)
Jun 18 09:35:01 geel kernel: NETDEV WATCHDOG: eth1: transmit timed out
Jun 18 09:37:36 geel kernel: NETDEV WATCHDOG: eth1: transmit timed out

I tested this with 2.6.7-bk4, same problem (noticed in the changelog that there was some bug w.r.t. ehci 
fixed, but apparently not the same bug)

Apparently the same happens here. 

------------- dmesg output 2.6.7 -----------------
ohci1394: $Rev: 1223 $ Ben Collins <bcollins@debian.org>
PCI: Setting latency timer of device 0000:00:0d.0 to 64
ohci1394: fw-host0: OHCI-1394 1.1 (PCI): IRQ=[11]  MMIO=[d9083000-d90837ff]  Max
Packet=[2048]
ehci_hcd 0000:00:02.2: nVidia Corporation nForce2 USB Controller
PCI: Setting latency timer of device 0000:00:02.2 to 64
ehci_hcd 0000:00:02.2: irq 12, pci mem f8d3e000
ehci_hcd 0000:00:02.2: new USB bus registered, assigned bus number 1
irq 12: nobody cared!
 [<c0107a1a>] __report_bad_irq+0x2a/0x90
 [<c0107b10>] note_interrupt+0x70/0xa0
 [<c0107db1>] do_IRQ+0x121/0x130
 [<c01061f4>] common_interrupt+0x18/0x20
 [<c011d7f0>] __do_softirq+0x30/0x80
 [<c011d866>] do_softirq+0x26/0x30
 [<c0107d8d>] do_IRQ+0xfd/0x130
 [<c01061f4>] common_interrupt+0x18/0x20
 [<c0249d9f>] pci_bus_read_config_byte+0x5f/0x90
 [<c033381e>] ehci_start+0x2ce/0x360
 [<c011a3fd>] printk+0x10d/0x170
 [<c03250b7>] usb_register_bus+0x137/0x160
 [<c032a10b>] usb_hcd_pci_probe+0x2ab/0x4e0
 [<c024d6c2>] pci_device_probe_static+0x52/0x70
 [<c024d71b>] __pci_device_probe+0x3b/0x50
 [<c024d75c>] pci_device_probe+0x2c/0x50
 [<c02a124f>] bus_match+0x3f/0x70
 [<c02a1379>] driver_attach+0x59/0x90
 [<c02a1621>] bus_add_driver+0x91/0xb0
 [<c02a1adf>] driver_register+0x2f/0x40
 [<c024d4e7>] pci_create_newid_file+0x27/0x30
 [<c024d9dc>] pci_register_driver+0x5c/0x90
 [<c05379e3>] init+0x23/0x30
 [<c052067c>] do_initcalls+0x2c/0xc0
 [<c0128877>] init_workqueues+0x17/0x60
 [<c0100410>] init+0x0/0x170
 [<c0100445>] init+0x35/0x170
 [<c010428c>] kernel_thread_helper+0x0/0x14
 [<c0104291>] kernel_thread_helper+0x5/0x14

handlers:
[<c0325ea0>] (usb_hcd_irq+0x0/0x70)
Disabling IRQ #12
-----------------------------------------------------------------------------------

Problem shows with kernel 2.6.7-rc1, rc2, rc3 and 2.6.7. Kernel 2.6.6 works
fine. I'll attach a dmesg diff between 2.6.6 and 2.6.7 (unfortunately 2.6.7
output is missing a few lines because the buffer was full).

This in an Athlon XP system, with an ASUS A7N8X Deluxe nforce2 motherboard.

Having a look at the dmesg usb diff's between 2.6.7 and 2.6.6 I see that pci mem
changed for ohci and uhci...

Created attachment 3245 [details]
dmesg diff between

I have same problems with my computer ASUS A7N8X Deluxe with Athlon XP 2600+ 512
Mo RAM. The following message appear (about ehci-hcd) during my boot process :
irq11 - nobody cared Disabled IRQ11.

I have tested 2.6.7-bk10, the problem seems to be present again, no fix found.

I just upgraded to kernel 2.6.7-bk14 and the problem seems to be solved here, my
usb-stick is working fine again.

Would be nice if someone could confirm this ( especially for the
nforce2-onboard-ethernet ).

Anyway, thanks to the developers for fixing this bug!

I can confirm that onboard ethernet now works on nforce2 chipset! Took me a while to get the kernel 
booting though, since with this kernel my hard disk is now seen as /dev/sda whereas it was /dev/hdc on 
2.6.6 and 2.6.7-bk4, after tweaking some config files everything is good again. I guess the SATA is now 
seen to be SCSI or something...

I guess it looks like this bug is fixed?

Could someone who's having this problem in current 2.6.7-bk please 
send /var/log/dmesg (for PCI/ACPI IRQ setup information) along 
with /proc/interrupts after this "shutdown". 

I'm not seeing the problem anymore, seems to have been fixed somewhere between bk10 and bk14.

Yes, it seems the bug has been solved.

I think we can close this bug and mark it as "PATCH_ALREADY_AVAILABLE"?

I just wanted to wait a while before doing this to avoid that there is still
someone having this problem and unable to post it.

solved here too (with 2.6.8-rc1)

I'm marking this resolved, as the systems where it "suddenly" 
appeared don't see it any more. 
 
There may be other systems which have such problems, but 
not "suddenly" with that release -- different root cause. 

Reply-To: postmaster@liu.se

You have tried to reach daner964@student.liu.se.
This student is no longer active at Linkoping University.
For more information, see www.student.liu.se/en/itstod/gallring