NetBSD Problem Report #52924

From  Sat Jan 13 18:58:37 2018
Return-Path: <>
Received: from ( [])
	(using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits))
	(Client CN "", Issuer " CA" (not verified))
	by (Postfix) with ESMTPS id CD4BE7A1B2
	for <>; Sat, 13 Jan 2018 18:58:36 +0000 (UTC)
Subject: sparc kernel dies with strange fault in pool_cache_get_slow
X-Send-Pr-Version: 3.95

>Number:         52924
>Category:       kern
>Synopsis:       sparc kernel dies with strange fault in pool_cache_get_slow
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    kern-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sat Jan 13 19:00:00 +0000 2018
>Last-Modified:  Sun Jan 14 11:55:00 +0000 2018
>Originator:     Martin Husemann
>Release:        NetBSD 8.99.12
The NetBSD Foundation, Inc.
System: NetBSD 8.99.12 (MP3) #55: Sat Jan 13 18:29:09 CET 2018 sparc
Architecture: sparc
Machine: sparc

My krups dies when trying to boot -current:

audiocs0 at ebus0 bar 14 offset 0x200000 line 3: CS4231A
audio0 at audiocs0: full duplex, playback, capture, mmap
audiocs0: Virtual format configured - Format SLINEAR, precision 16, channels 2, frequency 48000
audiocs0: Latency: 256 milliseconds
trap type 0x7: pc=0xe8306b14 npc=0xe8306b18 psr=0x4401fc6<EF,S,PS>
kernel: alignment fault trap
Stopped in pid 0.1 (system) at  netbsd:pool_cache_get_slow+0x1a4:       ld              [
%g1 + 0x100], %g1

Backtrace is:

pool_cache_get_slow(0xf0919840, 0xf00, 0xe8505d40, 0x0, 0x0, 0x44) at netbsd:cpu
cpu_intr_p(0xf09196c0, 0x1, 0x0, 0xa, 0x0, 0xe85063e7) at netbsd:vmem_alloc+0x20
vmem_alloc(0xe852ee50, 0x1000, 0x1001, 0xe8505e54, 0xfffffff8, 0xf09196c0) at ne
uvm_km_kmem_alloc(0xe852ee50, 0x1000, 0x1001, 0xe8505ecc, 0x4401fe3, 0x4401fe3) a
t netbsd:pool_page_alloc+0x6c
pool_page_alloc(0xf091c018, 0x1, 0xe8506500, 0x60, 0xe8506380, 0x4401fe4) at net
pool_allocator_alloc(0xf091c018, 0x1, 0xe85064a0, 0xe850649c, 0xe8506500, 0xe850
5f8f) at netbsd:pool_grow+0x148
pool_grow(0xf091c018, 0x1, 0x0, 0xe850697c, 0x1, 0xe8506007) at netbsd:pool_get+
pool_get(0xf091c018, 0x1, 0xe8506b74, 0x0, 0x0, 0xe85060a7) at netbsd:pool_cache
pool_cache_get_slow(0xf091c198, 0xf00, 0xe8506118, 0x0, 0x0, 0x4) at netbsd:pool
pool_cache_get_paddr(0xf091c018, 0x1, 0x0, 0xa, 0x0, 0xe85063e7) at netbsd:kmem_
kmem_intr_alloc(0xc08, 0x1, 0xfffffff8, 0x4, 0xfffffff8, 0xe8505cc0) at netbsd:k
kern_malloc(0xc00, 0x4, 0xe8506380, 0x60, 0x4401fe3, 0x4401fe3) at netbsd:sysctl
sysctl_realloc(0xf092b1f0, 0xf092e5b8, 0xe8506500, 0x60, 0xe8506380, 0x4401fe4) a
t netbsd:sysctl_create+0x1038
sysctl_create(0xe8506568, 0x1, 0xe85064a0, 0xe850649c, 0xe8506500, 0x60) at netb
sysctl_createv(0xf09a767c, 0x0, 0x0, 0xe850697c, 0x1, 0xe8506500) at netbsd:audi
audioattach(0xf09a8ac8, 0xf09a8908, 0xe8506b74, 0x0, 0x0, 0xf09a767c) at netbsd:
config_attach_loc(0xf09a8ac8, 0xe85180a0, 0x0, 0xe8506b74, 0xe814e47c, 0xf09a892
4) at netbsd:config_found_sm_loc+0x60
config_found_sm_loc(0xf09a8ac8, 0x0, 0x0, 0xe8506b74, 0xe814e47c, 0x0) at netbsd
config_found(0xf09a8ac8, 0xe8506b74, 0xe814e47c, 0x0, 0xe8506bc8, 0x0) at netbsd
audio_attach_mi(0xe8450bf0, 0xf0996a88, 0xf09a8ac8, 0xe8012900, 0xf0996af0, 0xf0
996bb0) at netbsd:cs4231_ebus_attach+0x2b8
cs4231_ebus_attach(0xf09621c8, 0xf09a8ac8, 0xe8506e98, 0x6, 0x14, 0x0) at netbsd

which is:

(gdb) list *pool_cache_get_slow+0x1a4
0xe8306b14 is in pool_cache_get_slow (../../../../kern/subr_pool.c:2194).
2189                            mutex_exit(&pc->pc_lock);
2190                            return true;
2191                    }
2192            }
2194            if (__predict_true((pcg = pc->pc_fullgroups) != NULL)) {
2195                    /*
2196                     * If there's a full group, release our empty
2197                     * group back to the cache.  Install the full
2198                     * group as cc_current and return.

and the value of pc is 3.

Try to boot -current on sparc/Krups


From: Martin Husemann <>
Subject: Re: kern/52924: sparc kernel dies with strange fault in
Date: Sun, 14 Jan 2018 11:54:05 +0000

 This happens when (accidentaly) compiling subr_vmem.c without -O2.

 Not sure we want to track this further (but something definitively is wrong).


NetBSD Home
NetBSD PR Database Search

(Contact us) $NetBSD: query-full-pr,v 1.39 2013/11/01 18:47:49 spz Exp $
$NetBSD:,v 1.8 2006/05/07 09:23:38 tsutsui Exp $
Copyright © 1994-2014 The NetBSD Foundation, Inc. ALL RIGHTS RESERVED.