PROBLEM: ISDN CAPI (avm b1pci) doesn't work, occasionally freezes Kernel (2.6.1)

Hi all,

I've tried to get my AVM B1 V4 ISDN card to run with Kernel 2.6.1; it produced lots of noise on the logs, didn't work and occasionally froze the kernel.

Regardless of whether I compile everything (or parts of it ) into the kernel or as a module, the symptoms/effects remain the same. The card appears to be detected just fine

______________________________________________________________________

Jan 18 16:59:11 mybox kernel: capi20_register: 
Jan 18 16:59:11 mybox kernel: kcapi: appl 1 up
Jan 18 16:59:11 mybox kernel: capidrv: Rev 1.39.6.7: loaded
Jan 18 16:59:11 mybox kernel: capi20: Rev 1.1.4.1.2.2: started up with major 68 (middleware+capifs)
Jan 18 16:59:11 mybox kernel: b1dma: revision 1.1.4.1.2.1
Jan 18 16:59:11 mybox kernel: PCI: Found IRQ 11 for device 0000:00:0c.0
Jan 18 16:59:11 mybox kernel: b1pci: PCI BIOS reports AVM-B1 V4 at i/o 0x8000, irq 11, mem 0xd9000000
Jan 18 16:59:11 mybox kernel: kcapi: Controller 1: b1pciv4-8000 attached
Jan 18 16:59:11 mybox kernel: b1pci: AVM B1 PCI V4 at i/o 0x8000, irq 11, mem 0 xd9000000, revision 4 (dma)

______________________________________________________________________
and I can load the firmware into the card [*], but here it is starting to get noisy

______________________________________________________________________

Jan 18 16:59:38 mybox kernel: b1pciv4-8000: card 1 "B1" ready.
Jan 18 16:59:38 mybox kernel: b1pciv4-8000: card 1 Protocol: DSS1
Jan 18 16:59:38 mybox kernel: b1pciv4-8000: card 1 Linetype: point to multipoint
Jan 18 16:59:38 mybox kernel: b1pciv4-8000: B1-card (3.11-03) now active
Jan 18 16:59:38 mybox kernel: kcapi: card 1 "b1pciv4-8000" ready.
Jan 18 16:59:38 mybox kernel: kcapi: notify up contr 1
Jan 18 16:59:38 mybox kernel: capidrv: controller 1 up
Jan 18 16:59:38 mybox kernel: get_drv 0: 0 -> 1
Jan 18 16:59:38 mybox kernel: capidrv-1: State ST_DRV_NULL Event EV_DRV_REGISTER
Jan 18 16:59:38 mybox kernel: capidrv-1: ChangeState ST_DRV_LOADED
Jan 18 16:59:38 mybox kernel: get_drv 0: 1 -> 2
Jan 18 16:59:38 mybox kernel: b1pciv4-8000: b1dma_interrupt: 0x1 ???
Jan 18 16:59:38 mybox kernel: capidrv-1: State ST_DRV_LOADED Event EV_STAT_RUN
Jan 18 16:59:38 mybox kernel: capidrv-1: ChangeState ST_DRV_RUNNING
Jan 18 16:59:38 mybox kernel: put_drv 0: 2 -> 1
Jan 18 16:59:38 mybox kernel: capi20_put_message: applid 0x1
Jan 18 16:59:38 mybox kernel: capidrv-1: now up (2 B channels)
Jan 18 16:59:38 mybox kernel: capidrv-1: D2 trace enabled
Jan 18 16:59:38 mybox kernel: capi20_put_message: applid 0x1

______________________________________________________________________
When starting up "old style" isdn0/ipppd interfaces, I see lots of lines like

______________________________________________________________________

Jan 18 13:08:51 mybox kernel: get_drv 0: 1 -> 2
Jan 18 13:08:51 mybox kernel: put_drv 0: 2 -> 1

______________________________________________________________________
with the occasional reports that things have been set up correctly, like

______________________________________________________________________

Jan 18 13:08:52 mybox ipppd[2578]: Found 2 devices: , 
Jan 18 13:08:52 mybox ipppd[2589]: ipppd i2.2.12 (isdn4linux version of pppd by
 MH) started
Jan 18 13:08:52 mybox ipppd[2589]: init_unit: 0
Jan 18 13:08:52 mybox kernel: ipppd_get: 1
Jan 18 13:08:52 mybox ipppd[2589]: Connect[0]: /dev/ippp1, fd: 9
Jan 18 13:08:52 mybox ipppd[2589]: init_unit: 1
Jan 18 13:08:52 mybox kernel: ipppd_get: 1
Jan 18 13:08:52 mybox ipppd[2589]: Connect[1]: /dev/ippp0, fd: 10
[...]
Jan 18 13:08:53 mybox isdnlog: Holiday Version 1.10-Germany [12-Apr-1999] loaded [11 entries from /usr/share/isdn/holiday-de.dat]
Jan 18 13:08:53 mybox isdnlog: Dest V1.01: File '/usr/share/isdn/dest.cdb' opened fine - Dest 1.0 int (+h) AT DE NL CH BE
Jan 18 13:08:53 mybox isdnlog: Zone V1.25: Provider 0 File '/usr/share/isdn/zone-de-dtag.cdb' opened fine - V1.25 K2 C2 N256 T157147 O1 L5
Jan 18 13:08:53 mybox isdnlog: Rates   Version 3.06 [07-Aug-2003 22:35:40] loaded [70 Providers, 738 Zones, 2667 Areas, 42 Services, 509 Comments, 13 eXceptions, 35 Redirects, 2091 Rates from /usr/share/isdn/rate-de.dat]
[...]
Jan 18 13:08:53 mybox isdnlog: (ISDN subsystem with ISDN_MAX_CHANNELS > 16 detected, ioctl(IIOCNETGPN) is available)
Jan 18 13:08:53 mybox isdnlog: isdn.conf:2 active channels, 6 MSN/SI entries
Jan 18 13:08:53 mybox isdnlog: (Data versions: iprofd=0x06  net_cfg=0x06  /dev/isdninfo=0x01)
Jan 18 13:08:53 mybox isdnlog: Everything is fine, isdnlog-4.67 is running in full featured mode.
Jan 18 13:08:53 mybox isdnlog: 
Jan 18 13:08:53 mybox + Jan 18 13:08:53 -----------------------------------------
Jan 18 13:08:53 mybox isdnlog: | capidrv-1#0 : free
Jan 18 13:08:53 mybox isdnlog: | capidrv-1#1 : free

[...]

______________________________________________________________________
These have several get_drv/put_drv lines thrown in (as above).

But then very quickly, the following block shows up, which gets repeated occasionally

______________________________________________________________________

Jan 18 13:09:30 mybox kernel: capidrv-1: controller dead ??
Jan 18 13:09:30 mybox kernel: capi20_put_message: applid 0x1
Jan 18 13:09:30 mybox kernel: capidrv-1: listen_change_state state=1 event=1 ????

______________________________________________________________________
Trying to access the ISDN, or trying to call the cards (to check the answering machine) produce many more of the get_drv/put_drv lines with different numbers at the end (e.g. "3 -> 4", "4 -> 3") but nothing appears to work.

I've tried for hours, checking the kernel config and recompiling with different options (including acpi/no acpi, pnp/no pnp and so on), but the behaviour was always the same.

The card ought to be fine -- it works flawlessly with kernel 2.4.x.

So may this be a problem with kernel 2.6.x?

I hope this helps, if you need more info, please let me know.

Help to get it to work would be appreciated.

Regards,
Georg

[*] Yes, the firmware should be correct: the cards works perfectly (with that firmware) under kernel 2.4.24.

P.S. Distribution: Debian GNU/Linux Testing
CPU: AMD Athlon(tm) Processor (700MHz)

P.P.S. Possibly related (not sure):

When trying the compiled-in CAPI with dynamically loaded "old style" ISDN interfaces, I saw some kernel oopses. When everything was compiled-in, I didn't see Oopses, although the above is still exactly the same.

Even though it may be unrelated, in case it helps here are the syslog entries for reference:

______________________________________________________________________

Jan 18 14:25:53 mybox kernel: Unable to handle kernel paging request at virtual address 04800600
Jan 18 14:25:53 mybox kernel:  printing eip:
Jan 18 14:25:53 mybox kernel: c0139a40
Jan 18 14:25:53 mybox kernel: *pde = 00000000
Jan 18 14:25:53 mybox kernel: Oops: 0000 [#1]
Jan 18 14:25:53 mybox kernel: CPU:    0
Jan 18 14:25:53 mybox kernel: EIP:    0060:[shrink_list+400/1136]    Not tainted
Jan 18 14:25:53 mybox kernel: EFLAGS: 00010282
Jan 18 14:25:53 mybox kernel: EIP is at shrink_list+0x190/0x470
Jan 18 14:25:53 mybox kernel: eax: 04800600   ebx: 00000000   ecx: 00000000   edx: 00000002
Jan 18 14:25:53 mybox kernel: esi: c13b7568   edi: d7d67e64   ebp: d7c8db4c   esp: d7d67da0
Jan 18 14:25:53 mybox kernel: ds: 007b   es: 007b   ss: 0068
Jan 18 14:25:53 mybox kernel: Process kswapd0 (pid: 8, threadinfo=d7d66000 task=d7d6ccc0)
Jan 18 14:25:53 mybox kernel: Stack: d7d4c080 00000082 c0429680 00000001 00000000 0000004e c13b7a08 c13be0b0 
Jan 18 14:25:53 mybox kernel:        20000001 00000000 d7d67e10 c010ae0a 0000000e d7d4c080 d7d67e10 0000000e 
Jan 18 14:25:53 mybox kernel:        00000000 00000001 00000087 c010b13c 0000000e c03e0d80 00000001 c036ed34 
Jan 18 14:25:53 mybox kernel: Call Trace:
Jan 18 14:25:53 mybox kernel:  [handle_IRQ_event+58/112] handle_IRQ_event+0x3a/0x70
Jan 18 14:25:53 mybox kernel:  [do_IRQ+140/240] do_IRQ+0x8c/0xf0
Jan 18 14:25:53 mybox kernel:  [common_interrupt+24/32] common_interrupt+0x18/0x20
Jan 18 14:25:53 mybox kernel:  [shrink_cache+352/640] shrink_cache+0x160/0x280
Jan 18 14:25:53 mybox kernel:  [balance_pgdat+372/496] balance_pgdat+0x174/0x1f0
Jan 18 14:25:53 mybox kernel:  [kswapd+252/256] kswapd+0xfc/0x100
Jan 18 14:25:53 mybox kernel:  [autoremove_wake_function+0/80] autoremove_wake_function+0x0/0x50
Jan 18 14:25:53 mybox kernel:  [autoremove_wake_function+0/80] autoremove_wake_function+0x0/0x50
Jan 18 14:25:53 mybox kernel:  [kswapd+0/256] kswapd+0x0/0x100
Jan 18 14:25:53 mybox kernel:  [kernel_thread_helper+5/12] kernel_thread_helper+0x5/0xc
Jan 18 14:25:53 mybox kernel: 
Jan 18 14:25:53 mybox kernel: Code: 8b 08 85 c9 0f 84 9b 01 00 00 8b 54 24 0c 85 d2 74 3e 8b 45

______________________________________________________________________

Jan 18 15:11:37 mybox kernel: Unable to handle kernel paging request at virtual address 01001f02
Jan 18 15:11:37 mybox kernel:  printing eip:
Jan 18 15:11:37 mybox kernel: c01732b1
Jan 18 15:11:37 mybox kernel: *pde = 00000000
Jan 18 15:11:37 mybox kernel: Oops: 0000 [#2]
Jan 18 15:11:37 mybox kernel: CPU:    0
Jan 18 15:11:37 mybox kernel: EIP:    0060:[proc_lookup+49/192]    Not tainted
Jan 18 15:11:37 mybox kernel: EFLAGS: 00010206
Jan 18 15:11:37 mybox kernel: EIP is at proc_lookup+0x31/0xc0
Jan 18 15:11:37 mybox kernel: eax: d5632900   ebx: 01001f00   ecx: 00000082   edx: d5632900
Jan 18 15:11:37 mybox kernel: esi: d563297d   edi: d7c82e89   ebp: 00000003   esp: d54d9e84
Jan 18 15:11:37 mybox kernel: ds: 007b   es: 007b   ss: 0068
Jan 18 15:11:37 mybox kernel: Process cat (pid: 17165, threadinfo=d54d8000 task=d601ed40)
Jan 18 15:11:37 mybox kernel: Stack: d563297c d4a62ec0 c015dfe3 fffffffe 00000000 fffffff4 d69e82b8 d69e8250 
Jan 18 15:11:37 mybox kernel:        d5632900 c0154cdc d69e8250 d5632900 d54d9f70 00000000 d54d9f70 d7ff0840 
Jan 18 15:11:37 mybox kernel:        d54d9f00 c0154f46 d4a62ec0 d54d9f08 d54d9f70 c489b00a d54d9f00 d69e8250 
Jan 18 15:11:37 mybox kernel: Call Trace:
Jan 18 15:11:37 mybox kernel:  [d_lookup+35/80] d_lookup+0x23/0x50
Jan 18 15:11:37 mybox kernel:  [real_lookup+204/240] real_lookup+0xcc/0xf0
Jan 18 15:11:37 mybox kernel:  [do_lookup+150/176] do_lookup+0x96/0xb0
Jan 18 15:11:37 mybox kernel:  [link_path_walk+1036/2016] link_path_walk+0x40c/0x7e0
Jan 18 15:11:37 mybox kernel:  [open_namei+131/1008] open_namei+0x83/0x3f0
Jan 18 15:11:37 mybox kernel:  [filp_open+62/112] filp_open+0x3e/0x70
Jan 18 15:11:37 mybox kernel:  [sys_open+91/144] sys_open+0x5b/0x90
Jan 18 15:11:37 mybox kernel:  [syscall_call+7/11] syscall_call+0x7/0xb
Jan 18 15:11:37 mybox kernel: 
Jan 18 15:11:37 mybox kernel: Code: 0f b7 4b 02 39 e9 74 47 8b 5b 28 85 db 75 f1 8b 44 24 10 85 

______________________________________________________________________

 || On Sun, 18 Jan 2004 21:35:27 -0200
 || fcp <[email protected]> wrote: 

f> I had similar problems. Running RH9, 2.6.1, W6692 pci card. Spent f> quite some time chasing this and in the end installed mISDN beta f> and it worked the first time. No nonsense. Hope this helps

Thanks.

I wonder. If it is a known problem that ISDN doesn't work for multiple cards in 2.6.1: Are there plans to incorporate mISDN officially?

Regards,
Georg

 || On Mon, 19 Jan 2004 14:59:48 -0200
 || fcp <[email protected]> wrote: 

f> He advised me this morning that the latest i4l cvs branch kernel26 f> version (old isdn system) fixes the problems we are having with f> the 2.6.1 kernel. Haven't had a chance to test this yet.

Ah, so 2.6.2 should fix the problem? Good.