TBS6984 Kernel error on close

[TankSlappa]

We have a problem with the tbs6984 quad S2 card. The card tunes and
gives data fine for long periods. However sometimes when closing
/dev/dvb/adapter?/dvr0 we get the following kernel error:

Code: Select All Code

[ 4395.288391] irq 16: nobody cared (try booting with the "irqpoll" option)
[ 4395.288416] Pid: 524, comm: irq/16-SAA716x  Tainted: P         C O
3.2.0-23-lowlatency #31-Ubuntu
[ 4395.288417] Call Trace:
[ 4395.288423]  [<ffffffff810c858d>] __report_bad_irq+0x3d/0xe0
[ 4395.288425]  [<ffffffff810c8815>] note_interrupt+0x135/0x190
[ 4395.288427]  [<ffffffff810c6d9d>] irq_thread+0x1ed/0x210
[ 4395.288429]  [<ffffffff810c6bb0>] ? irq_finalize_oneshot+0x20/0x20
[ 4395.288431]  [<ffffffff81084a9c>] kthread+0x8c/0xa0
[ 4395.288435]  [<ffffffff81656ef4>] kernel_thread_helper+0x4/0x10
[ 4395.288436]  [<ffffffff81084a10>] ? kthread_worker_fn+0x190/0x190
[ 4395.288438]  [<ffffffff81656ef0>] ? gs_change+0x13/0x13
[ 4395.288439] handlers:
[ 4395.288481] [<ffffffff810c6250>] irq_default_primary_handler
threaded [<ffffffffa0245cd0>] saa716x_tbs6984_pci_irq
[ 4395.288551] Disabling IRQ #16

After that, the card can be tuned but will no longer give any data
until the system is rebooted.

Tested with on ubuntu with kernels 3.2.0-23-lowlatency and 3.2.0-30-generic.
Tested with MSI H61M-E23 (MS-7680) and Gigabyte Z77X-UD5H motherboards.

dmesg and lspci logs attached (hopefully)

[TankSlappa]

Logs attached, eventually found the board would accept .zip extensions...

[steven]

Hi TankSlappa

Why do you need to close /dev/dvb/adapter?/dvr0 ,because this is needed to
received data.So you should not close this device.

Kind Regards

steven

[TankSlappa]

We open the card in our application. Sometimes our application needs to restart when its settings are changed, so we need to stop using the card when our application closes so it can reopen it again when it restarts.

Surely a player application for example would open and close the card when it is started and run? Is this not correct?

[steven]

Hi TankSlappa

Can you make sure that you have released the /dev/dvb/adapter?/dvr0 at all?
Maybe you have not released it at all so next time you can not use it normally.

Kind Regards

steven

[TankSlappa]

Yes we "release" the device completely. This being linux we just close() the file descriptors for the demux, frontend and dvr devices. If that is insufficient then there is a driver bug. The kernel error occurs around the time we close the devices (when our application exits), not when we re-open them. The re-open could happen at any time after the close, but never sooner than 30s. 10 restarts of our application are enough to reproduce this problem 99% of the time.

[SergioDanielG]

Hi TankSlappa
Did you check if some adapter is busy before close it and specially before re-open it?

Best regards.

[gladroger]

My kernel.log

Sep 27 03:38:39 tvheadend kernel: [213830.131699] irq 17: nobody cared (try booting with the "irqpoll" option)
Sep 27 03:38:39 tvheadend kernel: [213830.131736] Pid: 0, comm: swapper/1 Tainted: P O 3.2.0-31-generic #50-Ubuntu
Sep 27 03:38:39 tvheadend kernel: [213830.131739] Call Trace:
Sep 27 03:38:39 tvheadend kernel: [213830.131741] <IRQ> [<ffffffff810dafdd>] __report_bad_irq+0x3d/0xe0
Sep 27 03:38:39 tvheadend kernel: [213830.131752] [<ffffffff810db265>] note_interrupt+0x135/0x190
Sep 27 03:38:39 tvheadend kernel: [213830.131755] [<ffffffff810d8ada>] handle_irq_event_percpu+0xaa/0x210
Sep 27 03:38:39 tvheadend kernel: [213830.131759] [<ffffffff810d8c91>] handle_irq_event+0x51/0x80
Sep 27 03:38:39 tvheadend kernel: [213830.131762] [<ffffffff810dbcba>] handle_fasteoi_irq+0x6a/0x110
Sep 27 03:38:39 tvheadend kernel: [213830.131766] [<ffffffff81015282>] handle_irq+0x22/0x40
Sep 27 03:38:39 tvheadend kernel: [213830.131771] [<ffffffff8166561a>] do_IRQ+0x5a/0xe0
Sep 27 03:38:39 tvheadend kernel: [213830.131774] [<ffffffff8165a9ae>] common_interrupt+0x6e/0x6e
Sep 27 03:38:39 tvheadend kernel: [213830.131776] <EOI> [<ffffffff8101be45>] ? mwait_idle+0x95/0x210
Sep 27 03:38:39 tvheadend kernel: [213830.131783] [<ffffffff81012236>] cpu_idle+0xd6/0x120
Sep 27 03:38:39 tvheadend kernel: [213830.131787] [<ffffffff81637f9d>] start_secondary+0xd9/0xdb
Sep 27 03:38:39 tvheadend kernel: [213830.131789] handlers:
Sep 27 03:38:39 tvheadend kernel: [213830.131803] [<ffffffff81466b70>] ata_bmdma_interrupt
Sep 27 03:38:39 tvheadend kernel: [213830.131832] [<ffffffffa0229cf0>] saa716x_tbs6984_pci_irq
Sep 27 03:38:39 tvheadend kernel: [213830.131857] Disabling IRQ #17

I'm also getting this IRQ error frequently after a couple of days. Only a reboot will solve it.

I have a Gigabyte P35-DS4 with Intel Q6600, 4gb ram and I got also the latest driver for the tbs6984.

[gladroger]

I think it's solved by using the MSI solution by creating and editing tbs.conf. 3 days now without error.

viewtopic.php?f=52&t=7631#p24903