NetBSD Problem Report #51174

From www@NetBSD.org  Sat May 28 02:58:28 2016
Return-Path: <www@NetBSD.org>
Received: from mail.netbsd.org (mail.netbsd.org [199.233.217.200])
	(using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits))
	(Client CN "mail.netbsd.org", Issuer "Postmaster NetBSD.org" (verified OK))
	by mollari.NetBSD.org (Postfix) with ESMTPS id 4BFAC7A495
	for <gnats-bugs@gnats.NetBSD.org>; Sat, 28 May 2016 02:58:28 +0000 (UTC)
Message-Id: <20160528025826.EFAB87AABA@mollari.NetBSD.org>
Date: Sat, 28 May 2016 02:58:26 +0000 (UTC)
From: jdbaker@mylinuxisp.com
Reply-To: jdbaker@mylinuxisp.com
To: gnats-bugs@NetBSD.org
Subject: panic: init died i386-7.99.29
X-Send-Pr-Version: www-1.0

>Number:         51174
>Category:       kern
>Synopsis:       panic: init died i386-7.99.29
>Confidential:   no
>Severity:       critical
>Priority:       high
>Responsible:    kern-bug-people
>State:          closed
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sat May 28 03:00:00 +0000 2016
>Closed-Date:    Mon May 30 03:25:21 +0000 2016
>Last-Modified:  Mon May 30 03:25:21 +0000 2016
>Originator:     John D. Baker
>Release:        NetBSD/i386-7.99.29 (sources from 20160526)
>Organization:
>Environment:
NetBSD 7.99.29 (GENERIC) #24: Thu May 26 11:42:11 CDT 2016

>Description:
After building a release from sources updated 27 May 2016, i386 kernels
panic as follows:

NetBSD 7.99.29 (GENERIC) #24: Thu May 26 11:42:11 CDT 2016
[...]
boot device: ex0
root on ex0
[...]
root file system type: nfs
kern.module.path=/stand/i386/7.99.29/modules
panic: init died (signal 0, exit 1)
fatal breakpoint trap in supervisor mode
trap type 1 code 0 eip c010e764 cs 8 eflags 246 cr2 b4bdc9cd ilevel 0 esp d9246e9c
curlwp 0xc21b7000 pid 1 lid 1 lowest kstack 0xd92442c0
Stopped in pid 1.1 (init) at    netbsd:breakpoint+0x4:  popl    %ebp
db{0}> bt
breakpoint(c0f6e886,c1300560,c0fca750,d9246eb8,c21d7da0,c21b7000,d9246fa8,d9246e
ac,c090c4fa,c0fca750) at netbsd:breakpoint+0x4
vpanic(c0fca750,d9246eb8,d9246f1c,c08be839,c0fca750,0,1,c21dcef8,efe000,c1e80e70
) at netbsd:vpanic+0x121
snprintf(c0fca750,0,1,c21dcef8,efe000,c1e80e70,ef8000,d9246f14,2,c21d8b80) at ne
tbsd:snprintf
exit1(c21b7000,1,0,c21b7000,c12024f4,d9246f9c,c014e73c,c21b7000,d9246f68,d9246f6
0) at netbsd:exit1+0x6de
sys_exit(c21b7000,d9246f68,d9246f60,c21dcee8,b4bdc000,d9246f60,d9246f68,1,0,0) a
t netbsd:sys_exit+0x3b
syscall() at netbsd:syscall+0x13c
--- syscall (number 1) ---
baea6637:
db{0}> 
 ps
PID    LID S CPU     FLAGS       STRUCT LWP *               NAME WAIT
1    >   1 7   0         0           c21b7000               init
0       51 3   0       200           c232a2a0           aiodoned aiodoned
0       50 3   0       200           c232a540            ioflush syncer
0       49 3   0       200           c232a7e0           pgdaemon pgdaemon
0       48 5   0       200           c232aa80           (zombie)
0       46 3   0       200           c21b8020              nfsio nfsiod
0       45 3   0       200           c21b82c0              nfsio nfsiod
0       44 3   0       200           c21b8560              nfsio nfsiod
0       43 3   0       200           c21b7540              nfsio nfsiod
0       42 3   0       200           c21b8800          atapibus0 sccomp
0       39 3   0       200           c21b77e0               usb3 usbevt
0       38 3   0       200           c21b7a80               usb2 usbevt
0       37 3   0       200           c21b7d20               usb1 usbevt
0       36 3   0       200           c21b72a0               usb0 usbevt
0       35 3   0       200           c21b8d40              unpgc unpgc
0       34 3   0       200           c21b8aa0        vmem_rehash vmem_rehash
0       25 3   0       200           c1f96020            atabus1 atath
0       24 3   0       200           c1f962c0            atabus0 atath
0       23 3   0       280           c1f96560           fw0probe ieee1394
0       22 3   0       200           c1f96800         usbtask-dr usbtsk
0       21 3   0       200           c1f96aa0         usbtask-hc usbtsk
0       20 3   0       200           c1f96d40               pms0 pmsreset
0       19 3   0       200           c1f6a000               apm0 apmev
0       18 3   0       200           c1f6a2a0           lnxsyswq lnxsyswq
0       17 3   0       200           c1f6a540          cryptoret crypto_w
0       16 3   0       200           c1f6a7e0             sysmon smtaskq
0       15 3   0       200           c1f6aa80         pmfsuspend pmfsuspend
0       14 3   0       200           c1f6ad20           pmfevent pmfevent
0       13 3   0       200           c1f60020         sopendfree sopendfr
0       12 3   0       200           c1f602c0           nfssilly nfssilly
0       11 3   0       200           c1f60560            cachegc cachegc
0       10 3   0       200           c1f60800              vrele vrele
0        9 3   0       200           c1f60aa0             vdrain vdrain
0        8 3   0       200           c1f60d40          modunload mod_unld
0        7 3   0       200           c1eb7000            xcall/0 xcall
0        6 1   0       200           c1eb72a0          softser/0
0        5 1   0       200           c1eb7540          softclk/0
0        4 1   0       200           c1eb77e0          softbio/0
0        3 1   0       200           c1eb7a80          softnet/0
0        2 1   0       201           c1eb7d20             idle/0
0        1 3   0       200           c1205640            swapper uvm
db{0}> 

>How-To-Repeat:
See above.

Thinking it might be related to kern/51173, I replied to that PR.

Seeing that it's not, christos@ suggested:

> Very different; you are dying later. Compile a kernel without ASLR and
> see if that works. It could be the change to randomize the location of
> rtld.

Building a kernel which includes GENERIC but then does:

  no options PAX_ASLR_DEBUG
  no options PAX_ASLR

Did not work.
>Fix:

>Release-Note:

>Audit-Trail:
From: "John D. Baker" <jdbaker@mylinuxisp.com>
To: gnats-bugs@NetBSD.org
Cc: 
Subject: Re: kern/51174: panic: init died i386-7.99.29
Date: Fri, 27 May 2016 22:03:28 -0500 (CDT)

 I should note that the last known working kernel was built from
 sources around 19 May 2016:

 NetBSD 7.99.29 (GENERIC) #23: Thu May 19 22:23:05 CDT 2016

 So that sets the bounds on the window for bisection.

 -- 
 |/"\ John D. Baker, KN5UKS               NetBSD     Darwin/MacOS X
 |\ / jdbaker[snail]mylinuxisp[flyspeck]com    OpenBSD            FreeBSD
 | X  No HTML/proprietary data in email.   BSD just sits there and works!
 |/ \ GPGkeyID:  D703 4A7E 479F 63F8 D3F4  BD99 9572 8F23 E4AD 1645

From: "John D. Baker" <jdbaker@mylinuxisp.com>
To: gnats-bugs@NetBSD.org
Cc: 
Subject: Re: kern/51174: panic: init died i386-7.99.29
Date: Sat, 28 May 2016 17:46:12 -0500 (CDT)

 Duh.  I think the problem is that I neglected to 'make cleandir' or nuke
 $OBJDIR and just did an update build.  I downloaded the latest snapshot
 (NetBSD 7.99.29 (GENERIC.201605281150Z)) and it has no trouble with
 'init'.

 Perhaps a note should be added to UPDATING for i386 when the switch was
 flipped as was done for amd64.  (Possibly for arm too, but its switch to
 GCC 5.3 at the same time makes that something of a given.)

 This seems also to be at the root of '/usr/libexec/cc1{,plus}' failures
 due to inability to write-enable the text segment I was seeing.  Possibly
 also related to a couple of package build failures.

 I'll rebuild properly and try again, but the above is likely the case.

 -- 
 |/"\ John D. Baker, KN5UKS               NetBSD     Darwin/MacOS X
 |\ / jdbaker[snail]mylinuxisp[flyspeck]com    OpenBSD            FreeBSD
 | X  No HTML/proprietary data in email.   BSD just sits there and works!
 |/ \ GPGkeyID:  D703 4A7E 479F 63F8 D3F4  BD99 9572 8F23 E4AD 1645

From: "John D. Baker" <jdbaker@mylinuxisp.com>
To: gnats-bugs@NetBSD.org
Cc: 
Subject: Re: kern/51174: panic: init died i386-7.99.29
Date: Sun, 29 May 2016 14:21:36 -0500 (CDT)

 On Sat, 28 May 2016, John D. Baker wrote:

 > Duh.  I think the problem is that I neglected to 'make cleandir' or nuke
 > $OBJDIR and just did an update build.  I downloaded the latest snapshot
 > (NetBSD 7.99.29 (GENERIC.201605281150Z)) and it has no trouble with
 > 'init'.

 This seems to be the case.  After nuking everything i[34]86-related and
 rebuilding, everything works again.

 This also clears up the problems with '/usr/libexec/cc1{,plus}' and
 build failures with "sysutils/cdrtools" (and probably "graphics/cairo").

 So, operator error.  I think a note should be added to UPDATING.

 -- 
 |/"\ John D. Baker, KN5UKS               NetBSD     Darwin/MacOS X
 |\ / jdbaker[snail]mylinuxisp[flyspeck]com    OpenBSD            FreeBSD
 | X  No HTML/proprietary data in email.   BSD just sits there and works!
 |/ \ GPGkeyID:  D703 4A7E 479F 63F8 D3F4  BD99 9572 8F23 E4AD 1645

From: "David A. Holland" <dholland@netbsd.org>
To: gnats-bugs@gnats.NetBSD.org
Cc: 
Subject: PR/51174 CVS commit: src
Date: Mon, 30 May 2016 03:11:48 +0000

 Module Name:	src
 Committed By:	dholland
 Date:		Mon May 30 03:11:48 UTC 2016

 Modified Files:
 	src: UPDATING

 Log Message:
 Note that i386 needs a PIE-related cleandir. Should avoid having PR 51174
 happen to anyone else. :-/


 To generate a diff of this commit:
 cvs rdiff -u -r1.271 -r1.272 src/UPDATING

 Please note that diffs are not public domain; they are subject to the
 copyright notices on the relevant files.

From: David Holland <dholland-bugs@netbsd.org>
To: gnats-bugs@NetBSD.org
Cc: 
Subject: Re: kern/51174: panic: init died i386-7.99.29
Date: Mon, 30 May 2016 03:12:02 +0000

 On Sun, May 29, 2016 at 07:25:01PM +0000, John D. Baker wrote:
  >  So, operator error.  I think a note should be added to UPDATING.

 Yes, it should.

 -- 
 David A. Holland
 dholland@netbsd.org

State-Changed-From-To: open->closed
State-Changed-By: dholland@NetBSD.org
State-Changed-When: Mon, 30 May 2016 03:25:21 +0000
State-Changed-Why:
UPDATING entry added - let me know if I got the date wrong or anything.


>Unformatted:

Home	PR Database Search