Bug 201368

Summary: Kernel panic vm_fault: fault on nofault entry
Product: Base System Reporter: Jake Guffey <jake.guffey>
Component: kernAssignee: freebsd-bugs (Nobody) <bugs>
Status: New ---    
Severity: Affects Only Me CC: op
Priority: ---    
Version: 10.1-RELEASE   
Hardware: i386   
OS: Any   

Description Jake Guffey 2015-07-05 21:53:57 UTC
I just experienced a kernel panic/reboot due to a reported vm_fault. /var/log/messages provides:

Jul  5 02:19:37 shemuel syslogd: kernel boot file is /boot/kernel/kernel
Jul  5 02:19:37 shemuel kernel: panic: vm_fault: fault on nofault entry, addr: c8f45000
Jul  5 02:19:37 shemuel kernel: KDB: stack backtrace:
Jul  5 02:19:37 shemuel kernel: #0 0xc06f07e2 at kdb_backtrace+0x52
Jul  5 02:19:37 shemuel kernel: #1 0xc06bafe5 at panic+0x85
Jul  5 02:19:37 shemuel kernel: #2 0xc096439e at vm_fault_hold+0x211e
Jul  5 02:19:37 shemuel kernel: #3 0xc0962241 at vm_fault+0x81
Jul  5 02:19:37 shemuel kernel: #4 0xc0a02931 at trap_pfault+0x291
Jul  5 02:19:37 shemuel kernel: #5 0xc0a020e1 at trap+0x5f1
Jul  5 02:19:37 shemuel kernel: #6 0xc09ee5ec at calltrap+0x6
Jul  5 02:19:37 shemuel kernel: #7 0xc085eb8a at sysctl_ip6_temppltime+0x3a
Jul  5 02:19:37 shemuel kernel: #8 0xc06c5a4f at sysctl_root+0x2af
Jul  5 02:19:37 shemuel kernel: #9 0xc06c6049 at userland_sysctl+0x1f9
Jul  5 02:19:37 shemuel kernel: #10 0xc06c5e08 at sys___sysctl+0x98
Jul  5 02:19:37 shemuel kernel: #11 0xc0a02f4e at syscall+0x38e
Jul  5 02:19:37 shemuel kernel: #12 0xc09ee651 at Xint0x80_syscall+0x21
Jul  5 02:19:37 shemuel kernel: Uptime: 35d19h44m32s
Jul  5 02:19:37 shemuel kernel: Physical memory: 787 MB
Jul  5 02:19:37 shemuel kernel: Dumping 133 MB: 118 102 86 70 54 38 22 6 
Jul  5 02:19:37 shemuel kernel: Dump complete
Jul  5 02:19:37 shemuel kernel: Automatic reboot in 15 seconds - press a key on the console to abort
Jul  5 02:19:37 shemuel kernel: Rebooting...

I do have a dump that I can provide if necessary.
I have a custom kernel and can provide the configuration as well.

This particular box runs two jails and that’s about it. At the time of the panic, I issued "sysctl -a” from within one of my jails. I was able to reproduce the bug by doing the same thing after the reboot. This resulted in a corrupted filesystem which I'm fixing right now.
Comment 1 Oliver Pinter freebsd_committer freebsd_triage 2015-07-05 21:56:47 UTC
Are you sure about the architecture? From the addresses seems like it's an i386 (32 bit), rather then a amd64 (64 bit).
Comment 2 Jake Guffey 2015-07-05 22:00:17 UTC
Good catch. I'm on a VPS and forgot that it's 32-bit. Sorry for the confusion.
Comment 3 Jake Guffey 2015-07-06 03:05:35 UTC
After rescuing the filesystem, I executed "sysctl -a" from *outside* a jail and got another kernel panic. Same issue vm_fault on nofault.