Re: Panic in "irq15: ata1" (perhaps just ATAPICAM...?)

On Thu, Jul 01, 2004 at 01:49:46PM -0400, Damian Gerow wrote:
> Okay, I can't exactly reproduce the orignal panic (I think -- these might
> all be related), but here's another cdparanoia panic.  I am using ATAPICAM,
> which I vaguely remember has caused some issues with other people.  Note
> that this ocurred after I let bgfsck run its course.
> 
> First off, a 'dmesg | grep ata1':
> 
>     ata1: at 0x170 irq 15 on atapci0
>     acd0: CDRW <SONY CD-RW CRX230E> at ata1-master PIO4
>     acd1: DVDROM <Pioneer DVD-ROM ATAPIModel DVD-120S> at ata1-slave PIO4
>     cd0 at ata1 bus 0 target 0 lun 0
>     cd1 at ata1 bus 0 target 1 lun 0
> 
> Now the actual panic:
> 
>     Fatal trap 12: page fault while in kernel mode
>     cpuod = 0; apic id = 00
> 
>     fault virtual address      = 0xc1cad220
>     fault code                 = supervisor read, page not present
>     instruction pointer        = 0x8:0xc049a4a9
>     stack pointer              = 0x10:0xd42bac74
>     frame pointer              = 0x10:0xd42baca8
>     code segment               = base 0x0, limit 0xfffff, type 0x1b
>                                = DPL 0, pres 1, def32 1, gran 1
>     processor eflags           = interrupt enabled, resume, IOPL = 0
>     current process            = 26 (irq15: ata1)
> 
>     kernel: type 12 trap, code=0
>     Stopped at      ata_pio_read+0x64:         repe insw       %dx,%es:(%edi)
>     db> examine
>     ata_pio_read+0x64:      eb6d66f3
>     db> trace
>     ata_pio_read(c1ca5168,12,52e19674,d42bacdc,12) at ata_pio_read+0x64
>     ata_generic_interrupt(c1a7d600,0,0,0,0) at ata_generic_interrupt+0x473
>     ithread_loop(c19c2400,d42bad48,0,0,c19c2400) at ithread_loop+0x19c
>     fork_exit(c05217c3,c19c2400,d42bad48) at fork_exit+0x7a
>     fork_trampoline() at fork_trampoline+0x8
>     --- trap 0x1, eip = 0, esp = 0xd42bad7c, ebp = 0 ---
>     db> reset
> 
> Me being a n00b, I have no idea what's going on.  But I /do/ know that I can
> reliably produce a panic while ripping this CD.

Can you get a crashdump?  They seem to be broken for me (and lock up the
system instead), but if you can do "call doadump" from the ddb> prompt
you should be able to get it so we can further debug what's happening.
It might be telling that you're able to get a panic without using DMA
whereas I get a lockup with DMA, and quite possibly the same problem.
Is there anything in dmesg, like messages about resetting devices?

I bet this is a buggy reset issue.

-- 
Brian Fundakowski Feldman                           \'[ FreeBSD ]''''''''''\
  <> green_at_FreeBSD.org                               \  The Power to Serve! \
 Opinions expressed are my own.                       \,,,,,,,,,,,,,,,,,,,,,,\