|
|
heads up: cache alignment panic on boot on sparc
Date: Mon, 29 Aug 2005 18:00:38 -0700
From: John Beck <jbeck at eng dot sun dot com>
To: ONNV gate staff <gatekeeper at onnv dot eng dot sun dot com>
Subject: heads up: cache alignment panic on boot on sparc
Note: 1) I am sending this on behalf of the gate staff, 2) I am sending this
to the gate staff and blind copying on-all and onnv-gate so people who reply-
all when they shouldn't don't spam the entire ON community...
PIT reported a panic "during bfu" on SPARC recently:
panic[cpu0]/thread=180e000: assertion failed: L2CACHE_ALIGN <= L2CACHE_ALIGN_MAX, file: ../../../common/vm/vm_pagelist.c, line: 333
Then this morning the ONNV sparc build machine hit the same on boot, with
the following stack trace.
000000000180b730 genunix:assfail+74 (10bb4f8, 10bb520, 14d, 18c6400, 1295800, 0)
%l0-3: 00000000000000bb 0000000000000010 000000000190c800 000000000190c800
%l4-7: 0000000001295800 0000000000000000 0000000001917000 0000000000000000
000000000180b7e0 unix:cpu_vm_data_init+44 (30080412000, 6e, 6000014b928, 200, 0, 200)
%l0-3: 00000000010bb400 00000000010bb400 000000000000006e 000000000180c000
%l4-7: 0000000000000000 0000000001899c00 000000000000014a 000000000000006e
000000000180b890 unix:setup_cpu_common+288 (30080412000, 1068800, 1, ffffffffffffffff, 4, 1000)
%l0-3: 000000000189bf18 00000000018cc800 000000000180e380 00000000018a6000
%l4-7: 0000000000000000 0000000000000000 0000000000000001 00000000018d9380
000000000180b950 unix:start_other_cpus+1b8 (0, 1, 0, 18a1fd0, 18af950, 189abe8)
%l0-3: 0000000000000001 00000000018af800 0000000001099000 00000000f142a400
%l4-7: 0000000000000000 0000000000000001 ffffffffffffffff 0000000000000000
000000000180ba00 genunix:main+24c (4312fc6e, 1902000, 189a000, 1957000, 0, 1950800)
%l0-3: 0000000001957000 0000000000000002 0000000001957b58 0000000001957800
%l4-7: 0000000001954a68 0000000001954800 0000000001954a58 0000000001950800
panic: entering debugger (no dump device, continue to reboot)
Type 'go' to resume
debugger entered.
{0} ok
This has been brought to the attention of the ONNV triage team, who is looking
into it, but given the serious nature of the problem, we wanted to let people
know about it. We hope to have an update tomorrow: presumably a new high-
priority bug being filed, possibly accompanied by a back-out.
-- John
|