On Sun, Dec 21, 2014 at 07:40:48AM -0600, Mahmoud Al-Qudsi
wrote:> Hello all,
>
> I have two very-similarily configured virtual machines, one of which will
> instantly go into a kernel panic and reboots when I attempt to launch
Chrome
> (as a normal user, without root priviliges).
>
> This is running on FreeBSD 10.1-RELEASE i386, under XOrg with the VESA
driver.
>
> Even though my kernel is built with BREAK_TO_DEBUGGER, Chromimum always
causes
> a hard restart of the entire machine - though KDB does print the kernel
> stacktrace. Any idea how I can catch this?
>
> Without ddb, I'm not sure where to go next to debug the issue.
>
> This is the stack trace as reported by KDB on the serial console, and
I've
> attached the output of dtruss for the process in question.
>
> Many thanks,
>
> Mahmoud Al-Qudsi
>
>
>
> processor eflags = interrupt enabled, resume, IOPL = 0
> current process = 820 (chrome)
> trld-elf.so.1`0x2dap n746cfb
> umber = 12
> panic: page fault
> cpuid = 5
> KDB: stack backtrace:
> db_trace_self_wrapper(c11ca238,c141f650,c65fc620,c5b26478,c095dc76,...) at
db_trace_self_wrapper+0x2d/frame 0xc5b26448
> kdb_backtrace(c138ccff,5,c1095b1c,c5b26504,c5db0cc8,...) at
kdb_backtrace+0x30/frame 0xc5b264b0
> vpanic(c5b26504,c5b26548,c104cd24,c1095b1c,c138e74d,...) at
vpanic+0x11d/frame 0xc5b264ec
> panic(c1095b1c,c138e74d,c65fc7c8,1,1,...) at panic+0x12/frame 0xc5b264f8
> trap_fatal(5,0,c138e746,246,c107be3f,...) at trap_fatal+0x324/frame
0xc5b26548
> trap_pfault(c,c,c65fc620,c5b265e0,246,...) at trap_pfault+0x72/frame
0xc5b265c0
> trap(c5b26714) at trap+0x75e/frame 0xc5b26708
> calltrap() at calltrap+0x6/frame 0xc5b26708
> --- trap 0xc, eip = 0xc0c0363d, esp = 0xc5b26754, ebp = 0xc5b2675c ---
> strlen(f,c5b26848,c11c3301,246,c11c32a8,...) at strlen+0xd/frame 0xc5b2675c
> kvprintf(c11c328f,c0b7b3b0,c5b26848,a,c5b26890,...) at kvprintf+0x81e/frame
0xc5b26828
> vsnprintf(c153c0b2,100,c11c328f,c5b26890,c5b26890,...) at
vsnprintf+0x45/frame 0xc5b26860
> kassert_panic(c11c328f,f,c655d0ef,b5,0,...) at kassert_panic+0x2d/frame
0xc5b26884
> __mtx_lock_flags(c66ccba4,0,c655d0ef,b5,c65fc620,...) at
__mtx_lock_flags+0x17e/frame 0xc5b268b8
> fdesc_allocvp(1,0,3,c65ca000,c5b26918,...) at fdesc_allocvp+0xad/frame
0xc5b268e8
> fdesc_lookup(c5b269a8,c1395423,c11fbb35,c11d6ef8,c14798f0,...) at
fdesc_lookup+0x1f0/frame 0xc5b26940
> VOP_LOOKUP_APV(c655e2f8,c5b269a8,c5b26b64,31b,0,...) at
VOP_LOOKUP_APV+0x12f/frame 0xc5b26970
> lookup(c5b26b08,c5b26a40,110,d5,1101,...) at lookup+0x514/frame 0xc5b269d0
> namei(c5b26b08,0,4000144,0,bf9fd6d1,...) at namei+0x4fd/frame 0xc5b26a68
> kern_statat_vnhook(c65fc620,0,ffffff9c,bf9fd6d1,0,...) at
kern_statat_vnhook+0xbd/frame 0xc5b26ba8
> sys_stat(c65fc620,c5b26cc8,c14385b4,c5b26cc8,0,...) at sys_stat+0x5a/frame
0xc5b26c40
> syscall(c5b26d08) at syscall+0x31b/frame 0xc5b26cfc
> Xint0x80_syscall() at Xint0x80_syscall+0x21/frame 0xc5b26cfc
> --- syscall (188, FreeBSD ELF32, sys_stat), eip = 0x2f0536e7, esp =
0xbf9fd68c, ebp = 0xbf9fd898 ???
>
What is the source line for address fdesc_allocvp+0xad ?
Easiest is to run 'kgdb /boot/kernel/kernel /dev/mem' on the system
which panics and do "list *fdesc_allocvp+0xad".