Re: cgdstrategy: divide fault in supervisor mode

[Alexander Nasonov <alnsn%yandex.ru@localhost>]

Michael van Elst wrote:
> That would require dg_secsize to be 0 which is difficult to do
> because the drivers initialize the value and the common disk_set_info()
> function fixes up a zero value.

I can reproduce division by zero but not when rebooting. If I take
an unconfigured cgd device, i.e. cgd2 and run

	mount /dev/cgd2d /tmp

the kernel will panic instead of returning ENXIO.

> But maybe the dg pointer is bad? Please have a look at the %rdi
> register.

I don't know what was rdi's value when it crashed during reboot but
crashes when mounting /dev/cgd2d all have good kernel-space values.
I can examine data at 0x34 offset and it's indeed 0.

$ crash -M netbsd.12.core
Crash version 7.99.36, image version 7.99.36.
System panicked: dump forced via kernel debugger
Backtrace from time of crash is available.
crash> dmesg|tail
iwm0: 11g rates: 1Mbps 2Mbps 5.5Mbps 11Mbps 6Mbps 9Mbps 12Mbps 18Mbps 24Mbps 36M
bps 48Mbps 54Mbps
acpibat0: normal capacity on 'charge state'
fatal integer divide fault in supervisor mode
trap type 8 code 0 rip ffffffff808db36f cs 8 rflags 10246 cr2 4d8000 ilevel 0 rs
p fffffe8116cfba50
curlwp 0xfffffe836fcbab00 pid 13.1 lowest kstack 0xfffffe8116cf82c0

dumping to dev 20,17 (offset=212951, size=3119109):
dump
crash> bt
_KERNEL_OPT_NARCNET() at 0
_KERNEL_OPT_NARCNET() at 0
db_reboot_cmd() at db_reboot_cmd
db_command() at db_command+0xeb
db_command_loop() at db_command_loop+0x90
db_trap() at db_trap+0xe3
kdb_trap() at kdb_trap+0xe1
trap() at trap+0x574
--- trap (number 8) ---
cgdstrategy() at cgdstrategy+0x26
bdev_strategy() at bdev_strategy+0x68
spec_strategy() at spec_strategy+0x81
VOP_STRATEGY() at VOP_STRATEGY+0x33
bio_doread() at bio_doread+0x98
bread() at bread+0x1a
ffs_mountfs() at ffs_mountfs+0x170
ffs_mount() at ffs_mount+0x227
VFS_MOUNT() at VFS_MOUNT+0x34
mount_domount() at mount_domount+0x122
do_sys_mount() at do_sys_mount+0x20f
sys___mount50() at sys___mount50+0x33
syscall() at syscall+0x15b
--- syscall (number 410) ---
75c7da:
crash> ps
PID    LID S CPU     FLAGS       STRUCT LWP *               NAME WAIT
13   >   1 7   1         0   fffffe836fcbab00          mount_ffs
12       1 2   1   8020000   fffffe811681d2a0              mount
8        1 2   1   8020000   fffffe811681d6c0                ksh
2        1 2   1   8020000   fffffe811681dae0                ksh
1        1 2   1   8020000   fffffe81163f5680               init
...

Alex