NetBSD Problem Report #51925

From www@NetBSD.org  Sun Jan 29 21:13:30 2017
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 C2E9F7A16B
	for <gnats-bugs@gnats.NetBSD.org>; Sun, 29 Jan 2017 21:13:29 +0000 (UTC)
Message-Id: <20170129211328.C2DEE7A2AB@mollari.NetBSD.org>
Date: Sun, 29 Jan 2017 21:13:28 +0000 (UTC)
From: guillem@debian.org
Reply-To: guillem@debian.org
To: gnats-bugs@NetBSD.org
Subject: The sparc/sparc64 ports consider EM_SPARC32PLUS equivalent to EM_SPARCV9
X-Send-Pr-Version: www-1.0

>Number:         51925
>Category:       port-sparc64
>Synopsis:       The sparc/sparc64 ports consider EM_SPARC32PLUS equivalent to EM_SPARCV9
>Confidential:   no
>Severity:       non-critical
>Priority:       low
>Responsible:    martin
>State:          closed
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sun Jan 29 21:15:00 +0000 2017
>Closed-Date:    Tue Jan 31 16:14:50 +0000 2017
>Last-Modified:  Sun Feb 05 14:55:00 +0000 2017
>Originator:     Guillem Jover
>Release:        HEAD
>Organization:
Debian
>Environment:
>Description:
While implementing ABI matching based on the ELF header data in dpkg-shlibdeps, and checking how various implementations handled mapping the alternative or legacy ELF machine IDs. I noticed that the NetBSD sparc and sparc64 ports consider the EM_SPARC32PLUS to be equivalent to EM_SPARCV9 when the object is ELFCLASS64.

This seems wrong and inconsistent with any of the other implementations I've checked, including Linux, FreeBSD, OpenBSD, GNU binutils and LLVM.

It also does not seem like anything will emit that kind of ELF object? So it would seem safe to drop the equivalence?

The dpkg code, for which I've had to add the special casing is:

  <https://anonscm.debian.org/cgit/dpkg/dpkg.git/tree/scripts/Dpkg/Shlibs/Objdump.pm#n216>

The NetBSD code involved is:

 <http://cvsweb.netbsd.org/bsdweb.cgi/src/sys/arch/sparc/include/elf_machdep.h?rev=1.6.122.1&content-type=text/x-cvsweb-markup>
 <http://cvsweb.netbsd.org/bsdweb.cgi/src/sys/arch/sparc64/include/elf_machdep.h?rev=1.11&content-type=text/x-cvsweb-markup>

in the MACHDEP_ID_CASES macros. In any case I thought I'd drop a bug in case you agree, and I can eventually remove that special case. :)
>How-To-Repeat:

>Fix:

>Release-Note:

>Audit-Trail:

Responsible-Changed-From-To: port-sparc64-maintainer->martin
Responsible-Changed-By: martin@NetBSD.org
Responsible-Changed-When: Tue, 31 Jan 2017 16:14:50 +0000
Responsible-Changed-Why:
Fixed, thanks for the report!


State-Changed-From-To: open->closed
State-Changed-By: martin@NetBSD.org
State-Changed-When: Tue, 31 Jan 2017 16:14:50 +0000
State-Changed-Why:
I fixed it


From: "Martin Husemann" <martin@netbsd.org>
To: gnats-bugs@gnats.NetBSD.org
Cc: 
Subject: PR/51925 CVS commit: src/sys/arch/sparc64/include
Date: Tue, 31 Jan 2017 16:13:56 +0000

 Module Name:	src
 Committed By:	martin
 Date:		Tue Jan 31 16:13:56 UTC 2017

 Modified Files:
 	src/sys/arch/sparc64/include: elf_machdep.h

 Log Message:
 PR port-sparc64/51925: remove EM_SPARC32PLUS from the 64 bit ELF machine
 ids we accept for our executables - that combination makes no sense
 at all (probably copy&pasto).


 To generate a diff of this commit:
 cvs rdiff -u -r1.11 -r1.12 src/sys/arch/sparc64/include/elf_machdep.h

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

From: Guillem Jover <guillem@debian.org>
To: gnats-bugs@NetBSD.org
Cc: martin@NetBSD.org, port-sparc64-maintainer@netbsd.org,
	netbsd-bugs@netbsd.org, gnats-admin@netbsd.org
Subject: Re: port-sparc64/51925 (The sparc/sparc64 ports consider
 EM_SPARC32PLUS equivalent to EM_SPARCV9)
Date: Thu, 2 Feb 2017 04:47:55 +0100

 Hi!

 On Tue, 2017-01-31 at 16:14:50 +0000, martin@NetBSD.org wrote:
 > Synopsis: The sparc/sparc64 ports consider EM_SPARC32PLUS equivalent to EM_SPARCV9
 > 
 > Responsible-Changed-From-To: port-sparc64-maintainer->martin
 > Responsible-Changed-By: martin@NetBSD.org
 > Responsible-Changed-When: Tue, 31 Jan 2017 16:14:50 +0000
 > Responsible-Changed-Why:
 > Fixed, thanks for the report!
 > 
 > 
 > State-Changed-From-To: open->closed
 > State-Changed-By: martin@NetBSD.org
 > State-Changed-When: Tue, 31 Jan 2017 16:14:50 +0000
 > State-Changed-Why:
 > I fixed it

 Thanks! But I see you only modified the sparc64 tree. A similar entry
 can be found in the sparc port, and although that one will just
 fall through to an error case (I assume), it seems confusing given the
 comment underneath. Also I think this is what caused the issue in the
 sparc64, as the sparc tree was imported as a clone initially? So
 fixing this would probably also avoid further confusion. :)

 For reference, I'm talking about the following code:

   ,--- sys/arch/sparc/include/elf_machdep.h ---
   #define	ELF64_MACHDEP_ENDIANNESS	ELFDATA2MSB
   #define	ELF64_MACHDEP_ID_CASES					\
   		case EM_SPARC32PLUS:					\
   		case EM_SPARCV9:					\
   		/* no 64-bit ELF machine types supported */
   `---

 Thanks,
 Guillem

From: "Martin Husemann" <martin@netbsd.org>
To: gnats-bugs@gnats.NetBSD.org
Cc: 
Subject: PR/51925 CVS commit: src/sys/arch/sparc/include
Date: Thu, 2 Feb 2017 08:12:08 +0000

 Module Name:	src
 Committed By:	martin
 Date:		Thu Feb  2 08:12:08 UTC 2017

 Modified Files:
 	src/sys/arch/sparc/include: elf_machdep.h

 Log Message:
 PR port-sparc64/51925: cosmetic cleanup, no functional change


 To generate a diff of this commit:
 cvs rdiff -u -r1.7 -r1.8 src/sys/arch/sparc/include/elf_machdep.h

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

From: Martin Husemann <martin@duskware.de>
To: Guillem Jover <guillem@debian.org>
Cc: gnats-bugs@NetBSD.org
Subject: Re: port-sparc64/51925 (The sparc/sparc64 ports consider
 EM_SPARC32PLUS equivalent to EM_SPARCV9)
Date: Thu, 2 Feb 2017 09:14:03 +0100

 On Thu, Feb 02, 2017 at 04:47:55AM +0100, Guillem Jover wrote:
 > Thanks! But I see you only modified the sparc64 tree. A similar entry
 > can be found in the sparc port, and although that one will just
 > fall through to an error case (I assume), it seems confusing given the
 > comment underneath.

 That was dead code, I just removed it.

 It is very unlikely  we would ever see a COMPAT_NETBSD64 on 32bit sparc
 machines ;-}

 Martin

From: matthew green <mrg@eterna.com.au>
To: Guillem Jover <guillem@debian.org>
Cc: martin@NetBSD.org, port-sparc64-maintainer@netbsd.org,
    netbsd-bugs@netbsd.org, gnats-admin@netbsd.org,
    gnats-bugs@NetBSD.org
Subject: re: port-sparc64/51925 (The sparc/sparc64 ports consider EM_SPARC32PLUS equivalent to EM_SPARCV9)
Date: Mon, 06 Feb 2017 01:53:32 +1100

 ideally, the sparc* elf_machdep.h could be shared to avoid
 having multiple versions of the mostly same thing.


 .mrg.

>Unformatted:

Home	PR Database Search