On Mon, 10 Jan 2022 18:38:18 +0100
Raimo Niskanen <[email protected]> wrote:

> I took pictures of ddb with, I hope, enough information to start with.
> They are at http://erlang.org/~raimo/obsd-panic_2022-01-10/image(0).jpg
> with (0) ranging from (0) through (12).

image(0).jpg:
| ddb{0}> show panic
| *cpu0: uvm_fault(0xfffffd83f348c338, 0x0, 0, 2) -> e

This message is from amd64/trap.c:kpageflttrap() line 275, below a
comment that says "bad memory access in the kernel".  Your "ps" shows
some multi-threaded beam.smp processes.  I don't know, but I suspect
a bug with multi-threaded processes and page faults on amd64.  (People
who run single-threaded processes, like clang, might not see the bug.)

I have seen uvm_fault failures on macppc, but this amd64 bug looks
different.  My macppc was running clang.

I have transcribed part of your screenshots.
--George

image(0).jpg:
| ddb{0}> trace
| _rb_remove(...) at _rb_remove+0x1eb
| nfs_reclaim(...) at nfs_reclaim+0x7d
| VOP_RECLAIM(...) at VOP_RECLAIM+0x59
| vclean(...) at vclean+0x156
| vgonel(...) at vgonel+0x5f
| getnewvnode(...) at getnewvnode+0x1eb
| ffs_vget(...) at ffs_vget+0x8a
| ufs_lookup() at ufs_lookup+0xc5e
| VOP_LOOKUP(...) at VOP_LOOKUP+0x46
| vfs_lookup(...) at vfs_lookup+0x359
| namei(...) at namei+0x245
| dofstatat(...) at dofstat+0x86
| syscall(...) at syscall+0x3a9
| Xsyscall() at Xsyscall+0x128
| end of kernel
| end trace frame: 0x7f7ffffeddc0, count: -14

image(1).jpg:
| ddb{0}> mach ddbcpu 1
| Stoppped at     x86_ipi_db+0x12:       leave
| x86_ipi_db(...) at xf86_ipi_db+0x12
| x86_ipi_handler() at x86_ipi_handler+0x80
| Xresume_lapic_ipi() at Xresume_lapic_ipi+0x23
| _kernel_lock() at _kernel_lock+0xb2
| uvm_fault(...) at uvm_fault+0x1a3
| upageflttrap(...) at upageflttrap+0x62
| usertrap(...) at usertrap+0x18b
| recall_trap() at recall_trap+0x8
| end of kernel
| end trace frame: 0x2, count: 7

image({4,5,7,8,9}).jpg, most lines omitted, SRUN = 2, SONPROC = 7:
| ddb{3}> ps
|    PID     TID   PPID    UID  S       FLAGS  WAIT        COMMAND
|  87043  105308  21865      0  7         0x2              perl
|  47217  107500  62779    421  7   0x4000002              beam.smp
|  47217  210707  62779    421  2   0x4000482              beam.smp
|  91976  308297      1    421  2       0x480              epmd
|  62779  510479  13890    421  3        0x82  select      erl_child_setup
|  13890  165605  63135    421  2   0x4000482              beam.smp
|  63135  510289  96576    421  3        0x82  wait        perl
|  71939  387441      1      0  2    0x100498              cron
|  74005  146341      1      0  2       0x480              amd
|  90642  106325      1      0  2       0x480              ypbind
|  42044  471963      1     83  2    0x100492              ntpd
|  46342  112358  78016     74  2    0x100492              pflogd
|  97590  489052  91921     73  2    0x100490              syslogd
| *39569  297939      0      0  7  0x40014200              idle3
|  99230  137026      0      0  7  0x40014200              idle2

Reply via email to