[MPlayer-dev-eng] IRQ Sharing and System Freeze

[keithcsl at ihug.co.nz]

Hi

We are experiencing random freezes (varies from hours to
days), and managed to capture the stack trace from the
serial console. We are using mencoder to record video from
the bttv, and we also have a USB WIFI driver sharing the
same IRQ. This is the latest crash we have:

mencoder      S 00018C00  5680 31138  31134              
31135 (NOTLB)
f4bd5ba0 00000000 0001ef00 00018c00 f899d4b3 00000160
00000001 c0f88250
       c0f881e0 c0f88180 00000000 00000000 42b6e000 003d7cc6
f718d560 f718d688
       c784a1c0 f4bd4000 f4bd4000 00000001 f8964978 00000000
00000160 00000000
Call Trace:
 [<f899d4b3>] bttv_risc_planar+0x53/0x350 [bttv]
 [<f8964978>] videobuf_waiton+0xb8/0x150 [video_buf]
 [<c0118ba0>] default_wake_function+0x0/0x10
 [<f8964a64>] videobuf_iolock+0x54/0x110 [video_buf]
 [<c0118ba0>] default_wake_function+0x0/0x10
 [<f89946ce>] bttv_prepare_buffer+0x10e/0x1b0 [bttv]
 [<f8996c0b>] bttv_do_ioctl+0x111b/0x1630 [bttv]
 [<c0333ff9>] _spin_lock_irqsave+0x9/0x10
 [<f8a1ad71>] SendPkt+0x141/0x1c0 [zd1211]
 [<f8a165aa>] zd_SendPkt+0x10a/0x210 [zd1211]
 [<f8a0cd48>] zd1205_xmit_frame+0x98/0x490 [zd1211]
 [<c0252068>] dma_pool_alloc+0x98/0x170
 [<c02cea51>] dev_queue_xmit+0x291/0x310
 [<c02ece90>] ip_output+0x120/0x2e0
 [<c0106a4a>] do_gettimeofday+0x1a/0xd0
 [<c02ce4bd>] __net_timestamp+0xd/0x20
 [<f8c6202b>] tcp_in_window+0x30b/0x510 [ip_conntrack]
 [<c01181bf>] activate_task+0x5f/0x80
 [<c01182a8>] try_to_wake_up+0x88/0xd0
 [<c0118be7>] __wake_up_common+0x37/0x70
 [<c0118c58>] __wake_up+0x38/0x50
 [<c0333ff9>] _spin_lock_irqsave+0x9/0x10
 [<c0124b20>] lock_timer_base+0x20/0x50
 [<c0124bec>] __mod_timer+0x9c/0xd0
 [<f8a19bab>] TxCompleted+0xeb/0x120 [zd1211]
 [<f89743ec>] video_usercopy+0xec/0x160 [videodev]
 [<c02975d5>] urb_unlink+0x35/0x60
 [<c0118c58>] __wake_up+0x38/0x50
 [<c0118c58>] __wake_up+0x38/0x50
 [<f8997120>] bttv_ioctl+0x0/0x70 [bttv]
 [<f8997152>] bttv_ioctl+0x32/0x70 [bttv]
 [<f8995af0>] bttv_do_ioctl+0x0/0x1630 [bttv]
 [<c0171450>] do_ioctl+0x50/0x80
 [<c01715be>] vfs_ioctl+0x5e/0x1c0
 [<c013d62f>] __do_IRQ+0xaf/0xf0
 [<c0171797>] sys_ioctl+0x77/0x90
 [<c0102fff>] sysenter_past_esp+0x54/0x75


Pid: 0, comm:              swapper
EIP: 0060:[<c0111d14>] CPU: 0
EIP is at delay_pmtmr+0x14/0x20
 EFLAGS: 00000287    Not tainted  (2.6.14.2)
EAX: a8b46c9a EBX: 001b4440 ECX: a8a27df1 EDX: 0000f3fe
ESI: ffffffff EDI: 0000000b EBP: 00000802 DS: 007b ES: 007b
CR0: 8005003b CR2: 00000802 CR3: 37206000 CR4: 000006d0
 [<c01d58a9>] __delay+0x9/0x10
 [<c011c225>] panic+0xd5/0xe0
 [<c010396b>] die+0x13b/0x150
 [<c0334ef0>] do_page_fault+0x0/0x5c0
 [<c0334ef0>] do_page_fault+0x0/0x5c0
 [<c0335281>] do_page_fault+0x391/0x5c0
 [<c0334ef0>] do_page_fault+0x0/0x5c0
 [<c0103283>] error_code+0x4f/0x54
 [<c011007b>] centrino_target+0x7b/0x180
 [<f8998a27>] bttv_irq+0x157/0x3a0 [bttv]
 [<c0118c58>] __wake_up+0x38/0x50
 [<c013d542>] handle_IRQ_event+0x32/0x70
 [<c013d601>] __do_IRQ+0x81/0xf0
 [<c01046a6>] do_IRQ+0x36/0x70
 [<c0103226>] common_interrupt+0x1a/0x20
 [<c0103915>] die+0xe5/0x150
 [<c0334ef0>] do_page_fault+0x0/0x5c0
 [<c0334ef0>] do_page_fault+0x0/0x5c0
 [<c0335281>] do_page_fault+0x391/0x5c0
 [<c0334ef0>] do_page_fault+0x0/0x5c0
 [<c0103283>] error_code+0x4f/0x54
 [<f8998a27>] bttv_irq+0x157/0x3a0 [bttv]
 [<c0118c58>] __wake_up+0x38/0x50
 [<c013d542>] handle_IRQ_event+0x32/0x70
 [<c013d601>] __do_IRQ+0x81/0xf0
 [<c01046a6>] do_IRQ+0x36/0x70
 [<c0103226>] common_interrupt+0x1a/0x20
 [<c020f803>] acpi_processor_idle+0x0/0x27f
 [<c020f904>] acpi_processor_idle+0x101/0x27f
 [<c01010e2>] cpu_idle+0x42/0x60
 [<c03ea7e5>] start_kernel+0x175/0x1c0
 [<c03ea330>] unknown_bootoption+0x0/0x1e0

What we don't understand is why is the [zd1211] (wifi
driver) methods being serviced from the mencoder's stack
trace?

Regards
Keith