NetBSD Problem Report #48623

From www@NetBSD.org  Thu Feb 27 10:06:54 2014
Return-Path: <www@NetBSD.org>
Received: from mail.netbsd.org (mail.netbsd.org [149.20.53.66])
	(using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits))
	(Client CN "mail.netbsd.org", Issuer "Postmaster NetBSD.org" (verified OK))
	by mollari.NetBSD.org (Postfix) with ESMTPS id 49DA8A5830
	for <gnats-bugs@gnats.NetBSD.org>; Thu, 27 Feb 2014 10:06:54 +0000 (UTC)
Message-Id: <20140227100652.B3754A64B2@mollari.NetBSD.org>
Date: Thu, 27 Feb 2014 10:06:52 +0000 (UTC)
From: isaki@NetBSD.org
Reply-To: isaki@NetBSD.org
To: gnats-bugs@NetBSD.org
Subject: x68k-current build failure on FreeBSD/amd64
X-Send-Pr-Version: www-1.0

>Number:         48623
>Category:       toolchain
>Synopsis:       x68k-current build failure on FreeBSD/amd64
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    toolchain-manager
>State:          closed
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Thu Feb 27 10:10:00 +0000 2014
>Closed-Date:    Sun Mar 02 06:38:21 +0000 2014
>Last-Modified:  Sun Mar 02 06:38:21 +0000 2014
>Originator:     Tetsuya Isaki
>Release:        NetBSD-current
>Organization:
>Environment:
FreeBSD 8.2-RELEASE-p9 amd64
>Description:
x68k-current (20140227) build fails at sys/arch/x68k/stand/boot
with a strange error message on FreeBSD-8.2/amd64 host.
This failure continues from at least 20140101.
(I have no build log during from 201308 to 201312 :(
Last successful build is 20130801 on this host.

Of course, I tried a clean build, but got the same result.

On the other hand, the build on NetBSD-6.1/i386 succeeded
with the same source tree (20140227) shared by NFS.

Therefore, I think that some problem happened in a crossbuild
tool during from 201308 to 201312.

The uname which build fails is "FreeBSD 8.2-RELEASE-p9 amd64".

build log is
:
dependall ===> sys/arch/x68k/stand/boot
    compile  boot/boot.o
    compile  boot/conf.o
       link  boot/boot
/var/obj/current/x68k/tools/bin/m68k--netbsdelf-ld: boot.sym: can not represent
section `.SUNW_ctf' in a.out object file format
/var/obj/current/x68k/tools/bin/m68k--netbsdelf-ld: final link failed: Nonrepres
entable section on output

*** Failed target:  boot
*** Failed command: /var/obj/current/x68k/tools/bin/m68k--netbsdelf-ld -N -stati
c -T /autobuild/current/src/sys/arch/x68k/stand/boot/boot.ldscript -o boot.sym s
rt0.o boot.o conf.o exec_image.o vers.o /var/obj/current/x68k/obj/sys/arch/x68k/
stand/libsa/lib/sa/libsa.a /var/obj/current/x68k/obj/sys/arch/x68k/stand/libsa/l
ib/kern/libkern.a /var/obj/current/x68k/obj/sys/arch/x68k/stand/libsa/lib/z/libz
.a /var/obj/current/x68k/obj/sys/arch/x68k/stand/libiocs/libiocs.a
*** Error code 1

Stop.
nbmake[10]: stopped in /autobuild/current/src/sys/arch/x68k/stand/boot

*** Failed target:  dependall
*** Failed command: cd "/autobuild/current/src/sys/arch/x68k/stand/boot"; /var/o
bj/current/x68k/tools/bin/nbmake realall
*** Error code 1
:

>How-To-Repeat:
build.sh -m x68k release on FreeBSD/amd64
>Fix:
I don't know.

>Release-Note:

>Audit-Trail:
From: Martin Husemann <martin@duskware.de>
To: gnats-bugs@NetBSD.org
Cc: 
Subject: Re: toolchain/48623: x68k-current build failure on FreeBSD/amd64
Date: Thu, 27 Feb 2014 11:28:26 +0100

 On Thu, Feb 27, 2014 at 10:10:00AM +0000, isaki@NetBSD.org wrote:
 > dependall ===> sys/arch/x68k/stand/boot
 >     compile  boot/boot.o
 >     compile  boot/conf.o
 >        link  boot/boot
 > /var/obj/current/x68k/tools/bin/m68k--netbsdelf-ld: boot.sym: can not represent
 > section `.SUNW_ctf' in a.out object file format
 > /var/obj/current/x68k/tools/bin/m68k--netbsdelf-ld: final link failed: Nonrepres
 > entable section on output

 .SUNW_ctf sounds like some dtrace related thing - I am not sure why it
 ends up in the object files here (and we should probably avoid that),
 but ignoring that for a moment, this should allow it to build:

 Index: boot.ldscript
 ===================================================================
 RCS file: /cvsroot/src/sys/arch/x68k/stand/boot/boot.ldscript,v
 retrieving revision 1.5
 diff -u -r1.5 boot.ldscript
 --- boot.ldscript	12 Aug 2013 16:32:39 -0000	1.5
 +++ boot.ldscript	27 Feb 2014 10:27:24 -0000
 @@ -63,5 +63,6 @@
        *(.debug_str)
        *(.debug_frame)
        *(.eh_frame)
 +      *(.SUNW_ctf)
    }
  }


 Martin

From: Martin Husemann <martin@duskware.de>
To: Tetsuya Isaki <isaki@NetBSD.org>
Cc: gnats-bugs@NetBSD.org
Subject: Re: toolchain/48623: x68k-current build failure on FreeBSD/amd64
Date: Fri, 28 Feb 2014 14:53:22 +0100

 On Fri, Feb 28, 2014 at 10:35:59PM +0900, Tetsuya Isaki wrote:
 > This patch solved my problem, thank you!
 > btw, is this an ad hoc solution?

 We should understand wethere it is a bug that those sections are generated,
 anf if it is, it should be fixed.

 If it is not a bug, the patch is a good solution.

 Martin

From: Tetsuya Isaki <isaki@NetBSD.org>
To: gnats-bugs@NetBSD.org, martin@duskware.de
Cc: toolchain-manager@netbsd.org,
	gnats-admin@netbsd.org,
	netbsd-bugs@netbsd.org, isaki@NetBSD.org
Subject: Re: toolchain/48623: x68k-current build failure on FreeBSD/amd64
Date: Fri, 28 Feb 2014 22:35:59 +0900

 At Thu, 27 Feb 2014 10:30:00 +0000 (UTC),
 Martin Husemann wrote:
 >  On Thu, Feb 27, 2014 at 10:10:00AM +0000, isaki@NetBSD.org wrote:
 >  > dependall ===> sys/arch/x68k/stand/boot
 >  >     compile  boot/boot.o
 >  >     compile  boot/conf.o
 >  >        link  boot/boot
 >  > /var/obj/current/x68k/tools/bin/m68k--netbsdelf-ld: boot.sym: can not represent
 >  > section `.SUNW_ctf' in a.out object file format
 >  > /var/obj/current/x68k/tools/bin/m68k--netbsdelf-ld: final link failed: Nonrepres
 >  > entable section on output
 >  
 >  .SUNW_ctf sounds like some dtrace related thing - I am not sure why it
 >  ends up in the object files here (and we should probably avoid that),
 >  but ignoring that for a moment, this should allow it to build:

 This patch solved my problem, thank you!
 btw, is this an ad hoc solution?
 ---
 Tetsuya Isaki <isaki@pastel-flower.jp / isaki@NetBSD.org>

 >  Index: boot.ldscript
 >  ===================================================================
 >  RCS file: /cvsroot/src/sys/arch/x68k/stand/boot/boot.ldscript,v
 >  retrieving revision 1.5
 >  diff -u -r1.5 boot.ldscript
 >  --- boot.ldscript	12 Aug 2013 16:32:39 -0000	1.5
 >  +++ boot.ldscript	27 Feb 2014 10:27:24 -0000
 >  @@ -63,5 +63,6 @@
 >         *(.debug_str)
 >         *(.debug_frame)
 >         *(.eh_frame)
 >  +      *(.SUNW_ctf)
 >     }
 >   }
 >  
 >  
 >  Martin
 >  
 > 

From: Tetsuya Isaki <isaki@NetBSD.org>
To: gnats-bugs@NetBSD.org, martin@duskware.de
Cc: toolchain-manager@netbsd.org,
	gnats-admin@netbsd.org,
	netbsd-bugs@netbsd.org,
	isaki@NetBSD.org
Subject: Re: toolchain/48623: x68k-current build failure on FreeBSD/amd64
Date: Sat, 01 Mar 2014 16:29:12 +0900

 At Fri, 28 Feb 2014 13:55:01 +0000 (UTC),
 Martin Husemann wrote:
 >  On Fri, Feb 28, 2014 at 10:35:59PM +0900, Tetsuya Isaki wrote:
 >  > This patch solved my problem, thank you!
 >  > btw, is this an ad hoc solution?
 >  
 >  We should understand wethere it is a bug that those sections are generated,
 >  anf if it is, it should be fixed.
 >  
 >  If it is not a bug, the patch is a good solution.

 I see, thank you for explanation.
 ---
 Tetsuya Isaki <isaki@pastel-flower.jp / isaki@NetBSD.org>

From: Martin Husemann <martin@duskware.de>
To: Tetsuya Isaki <isaki@NetBSD.org>
Cc: gnats-bugs@NetBSD.org
Subject: Re: toolchain/48623: x68k-current build failure on FreeBSD/amd64
Date: Sat, 1 Mar 2014 14:07:07 +0100

 The .SUNW_ctf section is added by running ctfconvert, which gets added
 to each compile step basically starting here in sys.mk:

 # We have to define these here, because if we don't the rules below will
 # not work
 .if exists(/usr/bin/ctfconvert) && exists(/usr/bin/ctfmerge)
 CTFCONVERT      ?=      ctfconvert
 CTFMERGE        ?=      ctfmerge
 .endif


 Now on a FreeBSD host, you will have /usr/bin/ctf* installed, and voila:
 you end up with dtrace info in the object files.

 The condition there should be changed to not refer /usr/bin on the host.
 Just MKDTRACE or soemthing?

 Martin

From: Martin Husemann <martin@duskware.de>
To: Tetsuya Isaki <isaki@NetBSD.org>
Cc: gnats-bugs@NetBSD.org
Subject: Re: toolchain/48623: x68k-current build failure on FreeBSD/amd64
Date: Sat, 1 Mar 2014 15:37:07 +0100

 I filed a separate PR about the bogus CTFCONVERT setting, please
 commit the patch (if only just as an additional safety layer).

 Martin

From: christos@zoulas.com (Christos Zoulas)
To: gnats-bugs@NetBSD.org, toolchain-manager@netbsd.org, 
	gnats-admin@netbsd.org, netbsd-bugs@netbsd.org, isaki@NetBSD.org
Cc: 
Subject: Re: toolchain/48623: x68k-current build failure on FreeBSD/amd64
Date: Sat, 1 Mar 2014 10:17:31 -0500

 On Mar 1,  1:10pm, martin@duskware.de (Martin Husemann) wrote:
 -- Subject: Re: toolchain/48623: x68k-current build failure on FreeBSD/amd64

 | The following reply was made to PR toolchain/48623; it has been noted by GNATS.
 | 
 | From: Martin Husemann <martin@duskware.de>
 | To: Tetsuya Isaki <isaki@NetBSD.org>
 | Cc: gnats-bugs@NetBSD.org
 | Subject: Re: toolchain/48623: x68k-current build failure on FreeBSD/amd64
 | Date: Sat, 1 Mar 2014 14:07:07 +0100
 | 
 |  The .SUNW_ctf section is added by running ctfconvert, which gets added
 |  to each compile step basically starting here in sys.mk:
 |  
 |  # We have to define these here, because if we don't the rules below will
 |  # not work
 |  .if exists(/usr/bin/ctfconvert) && exists(/usr/bin/ctfmerge)
 |  CTFCONVERT      ?=      ctfconvert
 |  CTFMERGE        ?=      ctfmerge
 |  .endif
 |  
 |  
 |  Now on a FreeBSD host, you will have /usr/bin/ctf* installed, and voila:
 |  you end up with dtrace info in the object files.
 |  
 |  The condition there should be changed to not refer /usr/bin on the host.
 |  Just MKDTRACE or soemthing?

 Perhaps do it only for NetBSD?

 christos

From: "Tetsuya Isaki" <isaki@netbsd.org>
To: gnats-bugs@gnats.NetBSD.org
Cc: 
Subject: PR/48623 CVS commit: src/sys/arch/x68k/stand/boot
Date: Sun, 2 Mar 2014 06:16:27 +0000

 Module Name:	src
 Committed By:	isaki
 Date:		Sun Mar  2 06:16:27 UTC 2014

 Modified Files:
 	src/sys/arch/x68k/stand/boot: boot.ldscript

 Log Message:
 Discard .SUNW_ctf section just in case.
 patch from martin@, see PR/48623.


 To generate a diff of this commit:
 cvs rdiff -u -r1.5 -r1.6 src/sys/arch/x68k/stand/boot/boot.ldscript

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

State-Changed-From-To: open->closed
State-Changed-By: isaki@NetBSD.org
State-Changed-When: Sun, 02 Mar 2014 06:38:21 +0000
State-Changed-Why:
I committed it.


>Unformatted:

Home	PR Database Search