>stevei@earthworks.com.au (Steven Ireland) writes:
>
>> With a bare 2.1.98, this system boots sometimes. When it fails to boot, it stops
>> just after the FDC identification message. When it completes the boot it seems
>
>I also have problems with the floppy driver. When inserting it as a
>module, I never get beyond the floppy drive identification message --
>NOT to be confused with the FDC identification message. The floppy
>driver just says "fd0 is 1.44MB" or something which it does after
>querying the BIOS setup, without really probing the hardware.
>
>Afterwards it hangs somewhere, after calling user_reset_fdc() (or
>however it is called). I think this is a bug in the floppy driver.
>
>Otherwise my machine seems to run, at least no problems with Linus' new
>ioapic patches.
>
>Claus
>
In my case it may be it stops at the FDC identification message because the next
stage is the SCSI identification and SCSI seems badly mangled by this mod.
I've been trying to reproduce a meaningful traceback, but every boot is
different. If it gets past the SCSI identification, it always dies when the
network is brought up. I think the death occurs when the first packet is
received, hence it is asynchronous and causes the kernel to die in many splendid
ways, like
>eth0: Reentering the interrupt handler! isr=0x0 imr=0x0.
>eth0: Reentering the interrupt handler! isr=0x2 imr=0x0.
>eth0: Reentering the interrupt handler! isr=0x2 imr=0x0.
>page fault from irq handler: 0000
>Unable to handle kernel paging requestCPU: 0
>EIP: 0010:[<c0111972>]
>EFLAGS: 00010086
> at virtual address 0030d0d0
>eax: c4a764d8 ebx: c4a76000 ecx: c0111958 edx: 00000018
>current->tss.cr3 = 04a9b000, %cr3 = 04a9b000
>esi: d2c27254 edi: 0000df40 ebp: c4a76430 esp: c4a76038
>*pde = 00000000
>ds: 0018 es: 0018 ss: 0018
>Process (pid: 0, process nr: 9, stackpage=c4a75000)
>Stack: c4a76000 d2c27254 0000df40 c4a76430 00000000 00000000 00000000 c0109fdc
> c4a76060 00000000 c4a76000 c0111958 00000018 d2c27254 0000df40 c4a76430
> c4a764d8 00000018 00000018 ffffffff c0111972 00000010 00010086 c4a76000
>Call Trace: [<c0109fdc>] [<c0111958>] [<c0111972>] [<c0109fdc>] [<c0111958>] [<c0111972>] [<c0109fdc>]
> [<c0111958>] [<c0111972>] [<c0109fdc>] [<c0111958>] [<c0111972>] [<c0109fdc>] [<c0111958>] [<c0111972>]
> [<c0109fdc>] [<c0111958>] [<c0111972>] [<c0109fdc>] [<c0111958>] [<c0111972>] [<c0109fdc>] [<c0111958>]
> [<c0111972>] [<c0109fdc>] [<c0111958>] [<c0111972>] [<c0109fdc>] [<c0111958>] [<c0111972>] [<c0109fdc>]
> [<c010bd77>] [<c0109f64>] [<c01c0961>] [<c010ba29>] [<c010bd27>] [<c010bd8f>] [<c0109f64>] [<c01c0961>]
Each death is different. I tried to find a common thread, but couldn't and
decided I'd have to leave it in sager (if that IS a word) hands.
Steve.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu