Hi,

my ss10 running stock 1.6 with a

	NetBSD 1.6_STABLE (PIZZA) #3: Sat Oct 19 00:45:25 CEST 2002

kernel locks up fairly reproducably under high load, like doing a full
build and the daily cron job or an Amanda backup run kicking in. The
file-systems are ffs mounted with softdep.

Breaking into the debugger (serial console) gives me

Stopped in pid 7 (aiodoned) at  cpu_Debugger+0x4:       jmpl
[%o7 + 0x8], %g0
db{0}> t
zstty_stint(0xf08bbc68, 0x0, 0xf01bfaec, 0xf024d0e8, 0xf0286000, 0xce7fcb)
at zstty_stint+0x88
zsc_intr_hard(0x8, 0xf08b4eb0, 0xf0246000, 0x7b9, 0xffff, 0x8) at
zsc_intr_hard+0x68
zshard(0x0, 0xf01b579c, 0x0, 0xf0335000, 0xf029b360, 0xf029b364) at zshard+0x40
sparc_interrupt44c(0xff000000, 0xf029dd40, 0x3691, 0xf0335010, 0x3df0,
0xf0241aa8) at sparc_interrupt44c+0x128
uvm_km_pgremove_intrsafe(0xf0c2b000, 0x8000, 0xf0c33000, 0xf0281208,
0xf0256c14, 0xf0256bb8) at uvm_km_pgremove_intrsafe+0xf8
uvm_unmap_remove(0xf0256b90, 0x8000, 0xf0c33000, 0xf0241b8c, 0xf024ce84,
0x0) at uvm_unmap_remove+0x164
uvm_unmap(0xf0256b90, 0xf0c2b000, 0xf0c33000, 0xf029d138, 0xf0, 0x3c) at
uvm_unmap+0x100
uvm_km_free(0xf0256b90, 0xf0c2b000, 0x8000, 0xf0249000, 0xf08b1de0, 0x8000)
at uvm_km_free+0x14
free(0xf0c2b000, 0x68, 0xf, 0xf0286400, 0xf092f840, 0x7064e000) at free+0x88
softdep_disk_write_complete(0xf210db60, 0xf0073ab4, 0xf210db60, 0xf0932fd8,
0xfc070fff, 0x8) at softdep_disk_write_complete+0x2c0
biodone(0xf210db60, 0x500, 0x0, 0xf01cc6f8, 0xfff7, 0xfc071000) at biodone+0x74
scsipi_complete(0x0, 0x22009, 0x200000, 0x0, 0x0, 0x0) at scsipi_complete+0x46c
scsipi_done(0xf09f4020, 0x1a1187, 0xf01bb224, 0xf024ce84, 0x1, 0x1) at
scsipi_done+0x154
ncr53c9x_done(0xf0930000, 0xf0932540, 0x14, 0xf01bb238, 0xf024ce84, 0x0) at
ncr53c9x_done+0x248
ncr53c9x_intr(0xf0930000, 0xf0010e84, 0xf029b0cc, 0xf029d138, 0xf0, 0x3c)
at ncr53c9x_intr+0x19c4
sparc_interrupt44c(0x0, 0xf02ed058, 0xa7e69a, 0x0, 0x0, 0x0) at
sparc_interrupt44c+0x128
pmap_kremove4m(0xf0f16000, 0xffff0f16, 0xf0335010, 0x0, 0x100a9,
0xf66a3e48) at pmap_kremove4m+0x17c
uvm_pagermapout(0xf0f11000, 0x8000, 0xf6d95b78, 0xf0284000, 0xf, 0x8) at
uvm_pagermapout+0xc
uvm_aio_aiodone(0xf0bb8160, 0xf66a3eb8, 0xffef, 0x0, 0xfff7, 0xffbf) at
uvm_aio_aiodone+0xd0
uvm_aiodone_daemon(0xf5e18b10, 0x200, 0xf0007530, 0xf029dd38, 0xf0286000,
0xf0249c00) at uvm_aiodone_daemon+0x110
proc_trampoline(0x0, 0x0, 0x0, 0x0, 0x0, 0x0) at proc_trampoline+0x18
db{0}>

A 'reboot' results in another panic during sync, and the machine locks up
terminally trying to dump core, so I have to toggle the power switch. Ouch.

Is this a 'known problem'? And/or should I send-pr?

	hauke


--
/~\  The ASCII Ribbon Campaign
\ /    No HTML/RTF in email
 X     No Word docs in email
/ \  Respect for open standards