Re: 8.0-BETA1 Fatal trap 12 during boot

Kostik Belousov wrote:
> On Thu, Jul 16, 2009 at 10:39:43AM +0200, Henri Hennebert wrote:
>> Hello,
>>
>> On one of my computer I get at the end of the boot:
>>
>> cd0 at ata1 bus 0 target 0 lun 0 
>>
>> cd0: <LITE-ON DVDRW SHM-165P6S MS0F> Removable CD-ROM SCSI-0 device 
>>
>> cd0: 33.000MB/s transfers 
>>
>> cd0: Attempt to query device size failed: NOT READY, Medium not present 
>>
>> flowtable clean(cd1:ata1:0:1:0): error 6
>> (cd1:ata1:0:1:0): Unretryable Error
>> cd1 at ata1 bus 0 target 1 lun 0
>> cd1: <HP CD-Writer+ 8200 1.0f> Removable CD-ROM SCSI-0 device
>> cd1: 16.000MB/s transfers
>> cd1: Attempt to query device size failed: NOT READY, Medium not present
>> er started
>> GEOM: new disk cd0
>> GEOM: new disk cd1
>> (cd0:ata1:0:0:0): error 6
>> (cd0:ata1:0:0:0): Unretryable Error
>> (cd0:ata1:0:0:0): error 6
>> (cd0:ata1:0:0:0): Unretryable Error
>> (cd1:ata1:0:1:0): error 6
>> (cd1:ata1:0:1:0): Unretryable Error
>> (cd1:ata1:0:1:0): error 6
>> (cd1:ata1:0:1:0): Unretryable Error
>>
>>
>> Fatal trap 12: page fault while in kernel mode
>> cpuid = 0; apic id = 00
>> fault virtual address   = 0x0
>> fault code              = supervisor write, page not present
>> instruction pointer     = 0x20:0xc057ecb1
>> stack pointer           = 0x28:0xc2e918fc
>> frame pointer           = 0x28:0xc2e91930
>> 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         = 1 (kernel)
>> [thread pid 1 tid 100002 ]
>> Stopped at      devfs_populate_loop+0x201:      movl    %edi,0(%eax)
>> db> bt
>> Tracing pid 1 tid 100002 td 0xc3150d80
>> devfs_populate_loop(0,c2e91a10,c0582d57,c347da40,c2e919a8,...) at 
>> devfs_populate_loop+0x201
>> devfs_populate(c347da40,c2e919a8,61,c148a08c,c2e91978,...) at 
>> devfs_populate+0x1b
>> devfs_lookup(c2e91a84,c3351d9c,80400,c08e779c,1ec,...) at devfs_lookup+0x2c7
>> VOP_LOOKUP_APV(c0923980,c2e91a84,c2e91be0,1ec,c3350060,...) at 
>> VOP_LOOKUP_APV+0x53
>> lookup(c2e91bb4,c3486000,400,c2e91bd4,0,...) at lookup+0x3fe
>> namei(c2e91bb4,c3351e10,0,0,0,...) at namei+0x642
>> kern_symlinkat(c3150d80,c08d9d99,ffffff9c,c08e7ab4,1,...) at 
>> kern_symlinkat+0xe6
>> kern_symlink(c3150d80,c08d9d99,c08e7ab4,1,c068ef10,...) at kern_symlink+0x2e
>> vfs_mountroot(c098c510,c3150d80,0,0,0,...) at vfs_mountroot+0x5dd
>> start_init(0,c2e91d38,0,0,0,...) at start_init+0x3c
>> fork_exit(c05b6340,0,c2e91d38) at fork_exit+0x88
>> fork_trampoline() at fork_trampoline+0x8
>> --- trap 0, eip = 0, esp = 0xc2e91d70, ebp = 0 ---
>> db>
>>
>> Is it a hardware problem ?
>>
>> I boot it with a 6.1 CD without problem.
> 
> No, this seems to be a driver problem. The panic in devfs_populate_loop()
> is usually caused by driver mistakes with cdev ref counting.
> 
> Can you look up source line for devfs_populate_loop+0x201 ?
> (Load kernel.debug into gdb and do
> 	list *devfs_populate_loop+0x201
> ; you do not need vmcore for this, only kernel.debug).

I try to debug it... but ...

First of all, the root partition is only 128MB (yes I know, but it
was a old config that I want to upgrade...) so my kernel was without
symbols. I try to boot with a fresh burned 7.2 LiveFS which come up with
just the same error! Strangly enough, I was sure that a 7.2 kernel was
booting this thing without problem...

So I remember that I was doing some glabel under 8.0-BETA1 on ad0s2[a-h]
when I encounter a crash... The bsdlabel being in bad shape under
8.0-BETA1, I have make my ad0s2[a-h] `bsdlabel' with gpart. Anyway,
I was suspicious of this glabel...

I boot with my 6.1 disk. Scratch the bsdlabel on ad0s2, and the glabel
with it.

I reboot, and bingo, all is back to normal.

Just for the record.

Henri