Bug 105591

I now started kdetv and it has started scanning for available channels, so at
least the driver is working *somehow*...

root@kallisto:~/saa7143-bug # cat /proc/interrupts
           CPU0
  0:    2613613          XT-PIC  timer
  2:          0          XT-PIC  cascade
  5:     472196          XT-PIC  mga@pci:0000:01:05.0
  7:          0          XT-PIC  parport0
  8:          3          XT-PIC  rtc
  9:          1          XT-PIC  acpi
 10:      93543          XT-PIC  Ensoniq AudioPCI, eth0
 11:     205824          XT-PIC  ide2, uhci_hcd:usb1, uhci_hcd:usb2, saa7134[0]
 14:      89930          XT-PIC  ide0
 15:     159801          XT-PIC  ide1
NMI:          0
LOC:          0
ERR:          0
MIS:          0

Let me know, if you need any additional info. Thanks!

Now I even received an Ooops during bootup:

[SNIP]
saa7134[0]: found at 0000:00:0d.0, rev: 1, irq: 255, latency: 255, mmio: 0xda020000
saa7134[0]: subsystem: 153b:1142, board: Terratec Cinergy 400 TV
[card=8,autodetected]
saa7134[0]: board init: gpio is 50000
input.c: calling hotplug without a hotplug agent defined
saa7134[0]: registered input device for IR
saa7134[0]: can't get IRQ 255
input.c: calling hotplug without a hotplug agent defined
Unable to handle kernel NULL pointer dereference at virtual address 00000000
 printing eip:
c01195a8
*pde = 00000000
Oops: 0000 [#1]
Modules linked in: saa7134 video_buf v4l2_common v4l1_compat ir_common
i2c_viapro via686a i2c_sensor i2c_core videodev prism2_plx snd_rawmidi
snd_seq_device snd_ac97_codec snd_pcm snd_timer snd soundcore snd_page_alloc
snd_ac97_bus p80211 matrox_w1 wire uhci_hcd usbcore pci_hotplug amd_k7_agp
agpgart generic parport_pc lp parport reiserfs ide_cd cdrom ide_disk dm_snapshot
dm_mod raid1 fan thermal processor pdc202xx_old via82cxxx ide_core
CPU:    0
EIP:    0060:[<c01195a8>]    Tainted: GF    U VLI
EFLAGS: 00010017   (2.6.13-rc6-git7-3-default)
EIP is at __wake_up_common+0x18/0x60
eax: 00000000   ebx: 00000246   ecx: 00000001   edx: eff095ec
esi: dff5dc00   edi: ffffffea   ebp: eb1bfef0   esp: eb1bfed8
ds: 007b   es: 007b   ss: 0068
Process modprobe (pid: 2148, threadinfo=eb1be000 task=eb128020)
Stack: 00000001 00000001 eff095ec 00000246 dff5dc00 ffffffea eb1bff04 c0119601
       00000000 00000000 eff09000 eff090d0 f15ce22e 00000000 eff09000 f15c91dd
       eff09000 f15c99a4 f15dfdb0 dff5dc00 f15e08a0 dff5dc44 c0257630 c01eae40
Call Trace:
 [<c0119601>] __wake_up+0x11/0x20
 [<f15ce22e>] saa7134_tvaudio_fini+0x2e/0x60 [saa7134]
 [<f15c91dd>] saa7134_hwfini+0x6d/0x90 [saa7134]
 [<f15c99a4>] saa7134_initdev+0x5a4/0x710 [saa7134]
 [<c0257630>] __driver_attach+0x0/0x40
 [<c01eae40>] __pci_device_probe+0x30/0x40
 [<c01eae6e>] pci_device_probe+0x1e/0x40
 [<c025756f>] driver_probe_device+0x2f/0x90
 [<c0257669>] __driver_attach+0x39/0x40
 [<c0256dd8>] bus_for_each_dev+0x38/0x60
 [<c0257681>] driver_attach+0x11/0x20
 [<c0257630>] __driver_attach+0x0/0x40
 [<c02571da>] bus_add_driver+0x5a/0xa0
 [<c01eb028>] pci_register_driver+0x58/0x80
 [<c0134161>] sys_init_module+0xa1/0x180
 [<c0102d1b>] sysenter_past_esp+0x54/0x79
Code: 00 00 00 55 8b 40 04 89 e5 5d e9 b4 f6 ff ff 8d 74 26 00 55 89 e5 57 56 53
83 ec 0c 89 55 ec 89 45 f0 89 4d e8 89 c2 8b 00 39 d0 <8b> 30 75 06 eb 31 89 f6
89 d6 8d 58 f4 8b 78 f4 8b 45 0c 8b 4d
 <6>usbcore: registered new driver hiddev
[SNIP]

This is the /etc/modprobe.d/tv file that YaST created:

[SNIP]
alias char-major-81 videodev
options i2c-algo-bit bit_test=1

# YaST configured TV card
# qnJ_.AIy4B8S35H6:Terratec Cinergy 400 TV
alias char-major-81-0 saa7134
alias char-major-81-1 off
alias char-major-81-2 off
alias char-major-81-3 off
[SNIP]

Andreas, not sure if Kraxel is around at the moment - can you please assign this
one? Thanks!

Chris, something for your team.

Still happens with Beta3 :(

looks like any read access to the hardware returns 0xffffffff, guess it is some
serious pci config problem (or the card is broken).  Works the machine fine
otherwise?

Please try the usual acpi/apic/pci kernel command line options and see if that
helps ...

The oops looks like a bug in the error path, the driver failed to get IRQ 255
(0xff again ...), whereas it got IRQ 11 assigned on the first boot.

Could this be another problem with the pci_unassigned_resources issue  
that olh talked about on kernel@suse? May be worth trying the patch he 
posted there. 

Quite possible.  Is the card in question actually a cardbus one or normal PCI?
[ still busy catching up with email after vacation ... ]

Answer to Comment#8: Yes the system works fine otherwise. Removing the saa7134
kernel module makes the machine usable again (syslog bogs it down a bit). I last
used that card fine on SUSE 9.3. Could you elaborate on the "usual acpi/apic/pci
kernel command line options"? Which ones would you recommend to try?

Answer to Comment#10: This is a "Terratec Cinergy 400 TV" PCI card.

The stuff used for the "failsafe" boot, especially acpi=off,
don't know all of them offhand.

I've also started a bunch of mbuilds with the patch referenced
in comment #9 (dkernel-default: kalman-kraxel-163,
kernel-smp: kalman-kraxel-169), you can try whenever the issue
is fixed there.

Gerd, could you please post these somewhere Lenz can download them?

was ill three days, back now.

Seems the patch has been mergen into cvs meanwhile, so using a recent
kernel-of-the-day should do ;)

This still happens in RC1 and actually results in the system to not be able to
finish the installation - it simply freezes randomly after the first reboot.
Again, syslog is flooded with the same error message. Will try a failsafe boot now.

Same messages appear in failsafe mode, but I was able to rmmod the saa7134
module before the box froze again. Will try to finish the installation and test
a KOTD afterwards.

Any news with kotd?

Ping

Sorry for the long delay :(

I gave the 10.1b1 installation a try, but I still see the same errors. In the meanwhile I suspect that there is either a resource problem (IRQ routing), or the card is definitly broken. Feel free to close this bug for now, there is no point in keeping it open, as long as I can't provide you with a reproducable test case.

[x] done ;)