PCMCIA trouble with 2.6 kernel

tomk

I have a 3Com Bluetooth PCMCIA card (not Cardbus) which works fine with kernel 2.4.26. I've compiled a 2.6.6 kernel with the same Bluetooth config, but when I boot the card doesn't come up and I get the following in my dmesg (this is an extract- I can post other sections if needed):

cs: IO port probe 0x0100-0x04ff: excluding 0x4d0-0x4d7
cs: IO port probe 0x0800-0x08ff: clean.
cs: IO port probe 0x0c00-0x0cff: clean.
cs: IO port probe 0x0a00-0x0aff: clean.
Bluetooth: Core ver 2.4
NET: Registered protocol family 31
Bluetooth: HCI device and connection manager initialized
Bluetooth: HCI socket layer initialized
Unable to handle kernel NULL pointer dereference at virtual address 00000000
printing eip:
c0181166
*pde = 00000000
Oops: 0000 [#1]
PREEMPT
CPU: 0
EIP: 0060:[<c0181166>] Not tainted
EFLAGS: 00010246 (2.6.6.i5k030704)
EIP is at object_path_length+0x26/0x40
eax: 00000000 ebx: 00000000 ecx: ffffffff edx: c90ab8e4
esi: 00000001 edi: 00000000 ebp: ffffffff esp: c6cc16a8
ds: 007b es: 007b ss: 0068
Process cardmgr (pid: 1488, threadinfo=c6cc0000 task=c7446d30)
Stack: 00000003 00000000 c90ab8e4 c9025a80 c018120a c90ab8e4 c9025a94 c9025a80
c01ff777 c03a2185 c67bca54 c9025ae0 c06a4d90 c9025ae0 c9025a80 c022a6c1
c06a4d98 c90ab8e4 c03b3f7d c022a9dd c06a4d90 c06a4dd8 c9025acc 00000000
Call Trace:
[<c018120a>] sysfs_create_link+0x2a/0x150
[<c01ff777>] kobject_hotplug+0x67/0x70
[<c022a6c1>] class_device_dev_link+0x31/0x40
[<c022a9dd>] class_device_add+0xed/0x130
[<c90245c3>] fw_register_class_device+0x113/0x180 [firmware_class]
[<c01065d8>] common_interrupt+0x18/0x20
[<c9024665>] fw_setup_class_device+0x35/0xd0 [firmware_class]
[<c9024768>] request_firmware+0x68/0x170 [firmware_class]
[<c90a9790>] bt3c_hci_send_frame+0x0/0x110 [bt3c_cs]
[<c90a9d32>] bt3c_open+0xc2/0x190 [bt3c_cs]
[<c90aa456>] bt3c_config+0x2d6/0x420 [bt3c_cs]
[<c0133ba0>] __rmqueue+0xd0/0x120
[<c0130911>] file_read_actor+0xd1/0xf0
[<c0134035>] buffered_rmqueue+0xf5/0x1d0
[<c012fffd>] find_get_page+0x2d/0x60
[<c01341b4>] __alloc_pages+0xa4/0x320
[<c013e536>] do_no_page+0x1c6/0x330
[<c0133ba0>] __rmqueue+0xd0/0x120
[<c01bbf83>] do_get_write_access+0x253/0x5e0
[<c0116690>] autoremove_wake_function+0x0/0x50
[<c02ec4cf>] pcmcia_align+0x9f/0xb0
[<c011ca69>] find_resource+0x99/0xf0
[<c011cb27>] allocate_resource+0x67/0x90
[<c02e9fdc>] pcmcia_get_next_tuple+0x24c/0x2b0
[<c02e9abc>] pcmcia_get_first_tuple+0xbc/0x170
[<c02f334e>] yenta_set_mem_map+0x19e/0x1f0
[<c90aa668>] bt3c_event+0x68/0xe0 [bt3c_cs]
[<c02efde4>] pcmcia_register_client+0x244/0x290
[<c02f334e>] yenta_set_mem_map+0x19e/0x1f0
[<c0114f0d>] scheduler_tick+0x6d/0x520
[<c0120054>] update_process_times+0x44/0x50
[<c90a9f55>] bt3c_attach+0xf5/0x140 [bt3c_cs]
[<c90aa600>] bt3c_event+0x0/0xe0 [bt3c_cs]
[<c02f2b94>] get_pcmcia_driver+0x34/0x50
[<c02f1d02>] bind_request+0xd2/0x170
[<c02f2801>] ds_ioctl+0x4d1/0x5f0
[<c0319b71>] sock_def_readable+0x81/0x90
[<c03742ee>] unix_dgram_sendmsg+0x44e/0x540
[<c0178aad>] proc_alloc_inode+0x4d/0x80
[<c0316918>] sock_sendmsg+0x98/0xd0
[<c01653d5>] iget_locked+0x95/0xc0
[<c0178a37>] proc_read_inode+0x17/0x40
[<c0165def>] wake_up_inode+0xf/0x30
[<c0116904>] __mmdrop+0x34/0x50
[<c0389857>] schedule+0x3b7/0x5c0
[<c013cca3>] zap_pmd_range+0x63/0x80
[<c013cd13>] unmap_page_range+0x53/0x80
[<c013ce16>] unmap_vmas+0xd6/0x190
[<c014064a>] unmap_region+0x9a/0xe0
[<c0140538>] unmap_vma+0x48/0x90
[<c014059f>] unmap_vma_list+0x1f/0x30
[<c01409da>] do_munmap+0x14a/0x190
[<c015dc70>] sys_ioctl+0x100/0x270
[<c0140a64>] sys_munmap+0x44/0x70
[<c0105c6b>] syscall_call+0x7/0xb

Code: f2 ae f7 d1 49 8b 52 24 8d 74 31 01 85 d2 75 ea 5b 89 f0 5e
<6>Bluetooth: L2CAP ver 2.1
Bluetooth: L2CAP socket layer initialized
Bluetooth: RFCOMM ver 1.2
Bluetooth: RFCOMM socket layer initialized
Bluetooth: RFCOMM TTY layer initialized
mtrr: 0xfd000000,0x800000 overlaps existing 0xfd000000,0x400000
mtrr: 0xfd000000,0x800000 overlaps existing 0xfd000000,0x400000

Anyone got an idea what's going on here?

TIA.

Typedef

Hrrm.

Looks like this guy

Bluetooth: HCI socket layer initialized

has oops'd your kernel.

Try looking at the list of modules loaded by the kernel during a normal 2.4.x boot, then turning off cardmgr at boot time.

Next modprobe for the same module order you observed cardmgr had inserted in the working 2.4.x image.

Hopefully you can get around the oops by intelligently loading the modules in the correct order.

Else, ... maybe send the driver author an email.

tomk

Found the answer - thanks for pointing me in the right direction, typedef.

Did what you suggested, and got the same error when I modprobed the module for the card (bt3c_cs). Checked around the bluez.org site for anything relevant, and found this:

Second Bluetooth kernel patch for 2.6.6 available
With this patch a bug in the 3Com PCMCIA card driver was fixed along with another module reference counting problem in the RFCOMM protocol layer. The HID support was updated and should now also work built-in.

The patch has since been rolled into 2.6.7, so guess what I'm downloading right now?

tomk

I may be talking to myself here, but for the sake of completeness, and in case anyone else comes along with a similar problem, I needed kernel 2.6.7 and the latest patch from Bluez.org.

So my laptop is totally wireless now - WiFi at home, and wherever I can find an access point, and Bluetooth/GPRS with my mobile otherwise.:D