[geom] page fault in g_mbr_config()

[pluknet]

2009/7/24 pluknet <pluknet at gmail.com>:
> 2009/7/24 pluknet <pluknet at gmail.com>:
>> Hi.
>>
>> I got a panic while performing a repetitive  'fdisk -BI aacd0',
>> where aacd0 is a disk on aac0: <IBM ServeRAID-8k>.
>> This means that the command was issued after filesystems
>> were already created on aacd (after the first fdisk -BI aacd0
>> iteration), and are in umount'ed state.
>>
>> This is on 7.2-R, amd64. Config is a GENERIC plus ddb, carp, ipfw, quota.
>>
>> Fatal trap 12: page fault while in kernel mode
>> cpuid = 5; apic id = 05
>> fault virtual address   = 0x20
>> fault code              = supervisor read data, page not present
>> instruction pointer     = 0x8:0xffffffff804cc554
>> stack pointer           = 0x10:0xfffffffe80079b80
>> frame pointer           = 0x10:0xfffffffe80079bd0
>> code segment            = base 0x0, limit 0xfffff, type 0x1b
>>                        = DPL 0, pres 1, long 1, def32 0, gran 1
>> processor eflags        = interrupt enabled, resume, IOPL = 0
>> current process         = 2 (g_event)
>> [thread pid 2 tid 100013 ]
>> Stopped at      g_mbr_config+0x64:      movq    0x20(%rax),%r15
>> db> bt
>> Tracing pid 2 tid 100013 td 0xffffff000144da50
>> g_mbr_config() at g_mbr_config+0x64
>> g_run_events() at g_run_events+0x1b8
>> g_event_procbody() at g_event_procbody+0x57
>> fork_exit() at fork_exit+0x11f
>> fork_trampoline() at fork_trampoline+0xe
>> --- trap 0, rip = 0, rsp = 0xfffffffe80079d30, rbp = 0 ---
>
> And, of course...
>
> db> show proc 818
> Process 818 (fdisk) at 0xffffff0004ed1000:
>  state: NORMAL
>  uid: 0  gids: 0, 0, 5
>  parent: pid 814 at 0xffffff00045c0478
>  ABI: FreeBSD ELF64
>  arguments: fdisk
>  threads: 1
> 100169                   D       g_waitfo 0xffffff0004ec9100 fdisk
> db> bt 818
> Tracing pid 818 tid 100169 td 0xffffff0004fbf6e0
> sched_switch() at sched_switch+0x1fe
> mi_switch() at mi_switch+0x18e
> sleepq_timedwait() at sleepq_timedwait+0x31
> _sleep() at _sleep+0x354
> g_waitfor_event() at g_waitfor_event+0x9a
> g_ctl_ioctl() at g_ctl_ioctl+0x2df
> giant_ioctl() at giant_ioctl+0x7d
> devfs_ioctl_f() at devfs_ioctl_f+0x77
> kern_ioctl() at kern_ioctl+0xa2
> ioctl() at ioctl+0xf9
> syscall() at syscall+0x256
> Xfast_syscall() at Xfast_syscall+0xab
> --- syscall (54, FreeBSD ELF64, ioctl), rip = 0x8008200ec, rsp =
> 0x7fffffffe1d8, rbp = 0x4 ---
>

This makes me tied to GEOM_* -> GEOM_PART_* scheme (which is in 8+ in
DEFAULTS now).
After this replacement in DEFAULTS I see no more panics in 'fdisk -BI aacd0'.

-- 
wbr,
pluknet