Re: 5.0 BETA WAPBL related crash

["Gao Ya'nan" <abutter.gao%gmail.com@localhost>]

I got a similar crash on the last Sat. when installing the current
version from 200812110002Z, and I am sure that the instructions caused
fault and the RAX register value are as same as mentioned above.

2008/12/16 Juraj Hercek <nbsd%hck.sk@localhost>:
> Hello,
>
> I've experienced the crash when cvs updating pkgsrc. Intel C2D runs
> NetBSD/amd64 5.0_BETA (XEN3_DOM0 running on xen-3.3.0) built yesterday Mon
> Dec 15 12:13:02 CET 2008:
>
> --- hand rewritten ---
> uvm_fault (0xffffffff80c6ea20, 0x0, 1) -> e
> fatal page fault in supervisor mode
> trap type 6 code 0 rip ffffffff803b38fb c e030 rflags 10246 cf2  19 cpl 0
> rsp ffffa0002868baf0
> kernel: page fault trap, code=0
> Stopped in pid 0.52 (system) at netbsd:ffs_itimes+0x10b:     cmpq
> $0x1,0x18(%rax)
> ffs_itimes() at netbsd:ffs_itimes+0x10b
> ffs_update() at netbsd:ffs_update+0x50
> ffs_full_fsync() at netbsd:ffs_full_fsync+0x3d1
> spec_fsync() at netbsd:spec_fsync+0x59
> VOP_FSYNC() at netbsd:VOP_FSYNC+0x34
> sched_sync() at netbsd:sched_sync+0x14d
> ds    0x5684
> es    0xbaf0
> fs     0x4
> gs    0xd5c1
> rdi   0
> rsi    0
> rbp  0xffffa0002868bb30
> rbx  0xffffa0002875ad38
> rdx  0xf09a00d816af3517
> rcx  0xd5555684
> rax  0x1
> r8   0
> r9   0
> r10 0
> r11 0
> r12 0
> r13 0
> r14 0xffffa0002868baf0
> r15 0xffffa0002868baf0
> rip  0xffffffff803b38fb  ffs_itimes+0x10b
> cs   0xe030
> rflags 0x10246
> rsp  0xffffa0002868baf0
> ss    0xe02b
> netbsd:ffs_itimes+0x10b:    cmpq   $0x1,0x18(%rax)
> db>
>
> This one occured when I rebooted system:
> umounting /mnt/data (/dev/wd0p)...uvm_fault(0xffffa00026e8a5c0, 0x0, 1) -> e
> fatal page fault in supervisor mode
> trap type 6 code 0 rip ffffffff803b3854 cs e030 rflags 10246 cr2  19 cpl 0
> rsp ffffa00029007910
> kernel: page fault trap, code=0
> Stopped in pid 390.1 (reboot) at   netbsd:ffs_itimes+0x64: cmpq
> $0x1,0x18(%rax)
> ffs_itimes() at netbsd:ffs_itimes+0x64
> ffs_update() at netbsd:ffs_update+0x50
> ffs_fullfsync() at netbsd:ffs_full_fsync+0x3d1
> spec_fsync() at netbsd:spec_fsync+0x59
> VOP_FSYNC() at netbsd:VOP_FSYNC+0x34
> ffs_flushfiles() at netbsd:ffs_flushfiles+0xf7
> ffs_unmount() at netbsd:ffs_unmount+0x55
> dounmount() at netbsd:dounmount+0xd5
> vfs_unmountall() at netbsd:vfs_unmountall+0x7c
> cpu_reboot() at netbsd:cpu_reboot+0xdc
> sys_reboot() at netbsd:sys_reboot+0x52
> syscall() at netbsd:syscall+0xb4
> ds 0x56e7
> es 0x7910
> fs 0x7900
> gs 0xd5c1
> rdi 0
> rsi 0
> rbp 0xffffa00029007950
> rbx 0xffffa0002875ad38
> rdx 0x5bfa51b4c9cffd7f
> rcx 0xd55556e7
> rax 0x1
> r8 0
> r9 0x100000
> r10 0
> r11 0xffffa000289f34d8
> r12 0xffffa00029007910
> r13 0
> r14 0
> r15 0xffffa00029007910
> rip 0xffffffff803b3854 ffs_itimes+0x64
> cs 0xe030
> rflags 0x10246
> rsp 0xffffa00029007910
> ss 0xe02b
> netbsd:ffs_itimes+0x64: cmpq  $0x1,0x18(%rax)
> db>
>
> The partition is mounted this way (if it matters):
> $ mount | grep mnt
> /local/mnt on /mnt type null (hidden, local)
> /dev/wd0p on /mnt/data type ffs (log, local)
>
> Is the crash known already/being worked on?
>
> -- Juraj
>