5.4 Hanging, have some DDB output, unable to get panic dump

[David Kirchner]

On 10/3/05, Philip Kizer <pckizer at nostrum.com> wrote:
> The system was updated (from uname):
>     FreeBSD 5.4-STABLE #2: Wed Sep 21 15:55:00 CDT 2005
>
> And the console log is (2 events):
>
> [note: machine hung, could not do anything from current logins or running daemons]
> KDB: enter: Line break on console
> [thread pid 11 tid 100003 ]
> Stopped at      0xc062672f = kdb_enter+0x2b:    nop
> db> set $lines=0
> db> trace
> Tracing pid 11 tid 100003 td 0xc4b1e480
> kdb_enter(c08286fc) at 0xc062672f = kdb_enter+0x2b
> siointr1(c4c9c000,c0915780,0,c08284c0,6ad) at 0xc0781696 = siointr1+0xce
> siointr(c4c9c000) at 0xc078148d = siointr+0x21
> intr_execute_handlers(c4b16490,e9379ca0,4,e9379ce4,c0794f63) at 0xc079874d = intr_execute_handlers+0x7d
> lapic_handle_intr(34) at 0xc079a95e = lapic_handle_intr+0x2e
> Xapic_isr1() at 0xc0794f63 = Xapic_isr1+0x33
> --- interrupt, eip = 0xc079f402, esp = 0xe9379ce4, ebp = 0xe9379ce4 ---
> mp_grab_cpu_hlt(e9379cfc,c05ffc15,0,c4b1dc5c,e9379d24) at 0xc079f402 = mp_grab_cpu_hlt+0x16
> cpu_idle(0,c4b1dc5c,e9379d24,c05ffa16,0) at 0xc079c018 = cpu_idle+0x8
> idle_proc(0,e9379d38,c08cf000,0,c080c943) at 0xc05ffc15 = idle_proc+0x4d
> fork_exit(c05ffbc8,0,e9379d38) at 0xc05ffa16 = fork_exit+0x66
> fork_trampoline() at 0xc0794c0c = fork_trampoline+0x8
> --- trap 0x1, eip = 0, esp = 0xe9379d6c, ebp = 0 ---

I was seeing this error with 5.4-RELEASE, on a machine with 4GB, and
when using PAE. Are you using PAE? If so, you'll need to either
upgrade to 5-STABLE or apply a 2 line patch to your 5.4-STABLE system
(the patch hasn't been committed to that branch yet, it should
probably be committed to 5-4.RELEASE-p*). You can find the patch by
searching the archives for "PAE" and "pmap.c".