Breaking news: Need to start over on AIX

02-10-2011, 02:32 PM

Hi AIX users!

Right now I've checked in the changes to use the new shared library
model called "aix-soname" I've proposed recently along bug#213277 [1].

Due to filenames staying the same but having different kind of content now,
unfortunately @preserved-rebuild will not work.

So there is no upgrade path from the previous shared library model,
and things will break badly if you update system packages, especially
those ones python depends on.

You will have to start over not earlier than tomorrow (after the mirrors got updated).

Sorry for the inconvenience, but as far as I can tell, the number of
affected users isn't so high anyway (self, Perry, Benda(still?), anyone?).

Thank you!

/haubi/
--
Michael Haubenwallner
Gentoo on a different level

02-10-2011, 08:11 PM

On Feb 10, 2011, at 9:32 AM, Michael Haubenwallner wrote:

> Hi AIX users!
>
> Right now I've checked in the changes to use the new shared library
> model called "aix-soname" I've proposed recently along bug#213277 [1].

Can you tell us which version you picked? At one point you had about
five ideas you were weighing.

02-11-2011, 08:23 AM

On 02/10/2011 10:11 PM, Perry Smith wrote:
> On Feb 10, 2011, at 9:32 AM, Michael Haubenwallner wrote:
>>
>> Right now I've checked in the changes to use the new shared library
>> model called "aix-soname" I've proposed recently along bug#213277 [1].
>
> Can you tell us which version you picked? At one point you had about
> five ideas you were weighing.

Ohw sorry, sure: The one I've posted at libtool mailing list[1][2] as well as
in the IBM AIX Developers forum[3], and recently to the binutils mailing list[4].
While the big agreement on that variant is still not there - maybe because of
less interest (there's not much response anywhere), I've implemented that in
Prefix now for both prove-of-concept as well as to identify possible problems.

How I prefer doing "shared libraries" with "soname" on AIX now is:

*) Create the shared object "shr.o" (using '-G' linker flag).
*) Set the LOADONLY flag for "shr.o" (using 'strip -e').
*) Create the Import File "shr.imp", containing
- this header line, being the "soname" emulation:
#! libNAME.so.1(shr.o)
- another header line telling the bits used:
# 32
or
# 64
- the list of symbols exported.
*) Create the archive library "libNAME.so.1.2.3" from both
"shr.imp" and "shr.o".
*) Create the symlinks as usual:
libNAME.so.1 -> libNAME.so.1.2.3
libNAME.so -> libNAME.so.1.2.3
*) Create the static library archive "libNAME.a" from static objects.

[1] http://lists.gnu.org/archive/html/libtool/2011-01/msg00023.html
[2] http://lists.gnu.org/archive/html/libtool/2011-02/msg00000.html
[3] http://www.ibm.com/developerworks/forums/thread.jspa?threadID=358893
[4] http://sourceware.org/ml/binutils/2011-02/msg00099.html

/haubi/
--
Michael Haubenwallner
Gentoo on a different level

Fri Feb 11 10:30:02 2011
Return-path: <redhat-list-bounces@redhat.com>
Envelope-to: tom@linux-archive.org
Delivery-date: Fri, 11 Feb 2011 10:17:33 +0200
Received: from mx3-phx2.redhat.com ([209.132.183.24]:59259)
by s2.java-tips.org with esmtp (Exim 4.69)
(envelope-from <redhat-list-bounces@redhat.com>)
id 1PnoC4-0002tD-VC
for tom@linux-archive.org; Fri, 11 Feb 2011 10:17:33 +0200
Received: from lists01.pubmisc.prod.ext.phx2.redhat.com (lists01.pubmisc.prod.ext.phx2.redhat.com [10.5.19.33])
by mx3-phx2.redhat.com (8.13.8/8.13.8) with ESMTP id p1B9KEh7004676;
Fri, 11 Feb 2011 04:20:38 -0500
Received: from int-mx12.intmail.prod.int.phx2.redhat.com
(int-mx12.intmail.prod.int.phx2.redhat.com [10.5.11.25])
by lists01.pubmisc.prod.ext.phx2.redhat.com (8.13.8/8.13.8) with ESMTP
id p1B9KCP0005429 for <redhat-list@listman.util.phx.redhat.com>;
Fri, 11 Feb 2011 04:20:12 -0500
Received: from mx1.redhat.com (ext-mx14.extmail.prod.ext.phx2.redhat.com
[10.5.110.19])
by int-mx12.intmail.prod.int.phx2.redhat.com (8.14.4/8.14.4) with ESMTP
id p1B9K751010513
for <redhat-list@redhat.com>; Fri, 11 Feb 2011 04:20:07 -0500
Received: from mail-fx0-f46.google.com (mail-fx0-f46.google.com
[209.85.161.46])
by mx1.redhat.com (8.14.4/8.14.4) with ESMTP id p1B9Juhg031868
for <redhat-list@redhat.com>; Fri, 11 Feb 2011 04:19:57 -0500
Received: by fxm20 with SMTP id 20so2607178fxm.33
for <redhat-list@redhat.com>; Fri, 11 Feb 2011 01:19:56 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma;
h=domainkey-signature:mime-version:date:message-id:subject:from:to
:content-type:content-transfer-encoding;
bh=BLzKZr3neM9WO87VpCSG5/mLB4/5rSW7/CEr74qGTDw=;
b=wrmrlEntijz6FQuoGIhC4E/meH2bKHz8SLG/KMS2i6DaRWAgv6iNsdnmmQ6XhlkLk4
25bhl/A2Dsm7kkAc995t3jBeixFb+BerxLQMUA3KtjYhNgqwTnFNKITT sm7pHn6tCXBV
12Zqr8nSrCNu40rRLJ+2aqo/umTC8n17trLuo=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma;
h=mime-version:date:message-id:subject:from:to:content-type
:content-transfer-encoding;
b=CcDhotgcbNKMulJbOPWysP7RnzuREHfDtRX6KerPFmgyBIxM mYhr70YaJ1bd7LxyOz
pdJHd8NxbdEt7hvqyJOi6dZgzyd3fnT5VEFEEwqicSSpRjr8gX aAHFi0ulFTVwS2KDqH
25KAS36lqkS0qFxzt/EWwLMimjcGigQTOTwoM=
MIME-Version: 1.0
Received: by 10.223.83.6 with SMTP id d6mr279938fal.95.1297415995443; Fri, 11
Feb 2011 01:19:55 -0800 (PST)
Received: by 10.223.112.133 with HTTP; Fri, 11 Feb 2011 01:19:55 -0800 (PST)
Date: Fri, 11 Feb 2011 10:19:55 +0100
Message-ID: <AANLkTinkhWhpcO695U2gQhQF7wp9NWVnnejc69HeCdkW@mai l.gmail.com>
Subject: Re: Is the default red hat 5.4 NFS server multithreaded? (was:
redhat-list Digest, Vol 84, Issue 10)
From: =?UTF-8?B?KEltZWQgQ2hpaGkpINi52YXYp9ivINin2YTYtNmK2K3Zig ==?=
<imed.chihi@gmail.com>
To: redhat-list@redhat.com
X-RedHat-Spam-Score: -0.789 (DKIM_SIGNED, DKIM_VALID, DKIM_VALID_AU,
FREEMAIL_FROM, RCVD_IN_DNSWL_LOW, T_TO_NO_BRKTS_FREEMAIL)
X-Scanned-By: MIMEDefang 2.68 on 10.5.11.25
X-Scanned-By: MIMEDefang 2.68 on 10.5.110.19
X-MIME-Autoconverted: from quoted-printable to 8bit by
lists01.pubmisc.prod.ext.phx2.redhat.com id p1B9KCP0005429
X-loop: redhat-list@redhat.com
X-BeenThere: redhat-list@redhat.com
X-Mailman-Version: 2.1.12
Precedence: junk
Reply-To: General Red Hat Linux discussion list <redhat-list@redhat.com>
List-Id: General Red Hat Linux discussion list <redhat-list.redhat.com>
List-Unsubscribe: <https://www.redhat.com/mailman/options/redhat-list>,
<mailto:redhat-list-request@redhat.com?subject=unsubscribe>
List-Archive: <https://www.redhat.com/archives/redhat-list>
List-Post: <mailto:redhat-list@redhat.com>
List-Help: <mailto:redhat-list-request@redhat.com?subject=help>
List-Subscribe: <https://www.redhat.com/mailman/listinfo/redhat-list>,
<mailto:redhat-list-request@redhat.com?subject=subscribe>
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: base64
Sender: redhat-list-bounces@redhat.com
Errors-To: redhat-list-bounces@redhat.com

PiBEYXRlOiBXZWQsIDkgRmViIDIwMTEgMTc6MjA6MDQgLTA1MD AKPiBGcm9tOiBNYXR0eSBTYXJy
byA8bXNhcnJvQGdtYWlsLmNvbT4KPiBUbzogR2VuZXJhbCBSZW QgSGF0IExpbnV4IGRpc2N1c3Np
b24gbGlzdCA8cmVkaGF0LWxpc3RAcmVkaGF0LmNvbT4KPiBTdW JqZWN0OiBJcyB0aGUgZGVmYXVs
dCByZWQgaGF0IDUuNCBORlMgc2VydmVyIG11bHRpdGhyZWFkZW Q/Cj4KPiBIZXkgZXZlcnlvbmUs
Cj4gVGhpcyBpcyBhIHByZXR0eSBzaW1wbGUgcXVlc3Rpb24gYn V0IEkgY2FuJ3Qgc2VlbSB0byBm
aW5kIHRoZSBhbnN3ZXIuCj4gSSdtIGxvb2tpbmcgdG8gcHVyY2 hhc2UgYSBkZWRpY2F0ZWQgTkZT
IHNlcnZlci4gTW9zdCBvZiB0aGUgdmVuZG9ycwo+IGFyZSBwdX NoaW5nIHF1YWQvaGV4IGNvcmUg
cHJvY2Vzc29ycy4gVGhlcmUncyBvbmUgbW9kZWwgbGVmdCB0aG F0Cj4gc3RpbGwgaGFzIGEgZHVh
bCBjb3JlLiBTbyBmYXIgYXMgSSBjYW4gdGVsbCwgdGhlIGRlZm F1bHQgTkZTIHNlcnZlcgo+IGlz
bid0IG11bHRpdGhyZWFkZWQsIHNvIGV2ZW4gZHVhbCBjb3JlIG lzIGdvaW5nIHRvIGJlIG92ZXJr
aWxsLiBUaGUKPiBzZXJ2ZXIgaXMgZ29pbmcgdG8gYmUgZ2V0dG luZyBMb2cgZGF0YSBjb3BpZWQg
dG8gaXQgTiB0aW1lcyBhIGRheSwKPiB3aGVyZSBOIGlzIHRoZS BudW1iZXIgb2YgdGltZXMgdGhl
IGZpbGVzIHJlYWNoIGEgY2VydGFpbiBzaXplLgo+Cj4gVGhlIG 9ubHkgb3RoZXIgdGhpbmcgdGhl
IHNlcnZlciB3aWxsIGJlIGRvaW5nIGlzIHJ1bm5pbmcgYSBjcm 9uIGpvYgo+IHdoaWNoIGRlbGV0
ZXMgZmlsZXMgdGhhdCBhcmUgbW9yZSB0aGFuIDYwIGRheXMgb2 xkLiBQcmV0dHkgc2ltcGxlLAo+
IHJpZ2h0Pwo+IFRoZSBvbmx5IHJlYXNvbiBJIGNhbiBzZWUgan VzdGlmeWluZyBhIHF1YWQgY29y
ZSBwcm9jZXNzb3IgaXMgaWYgTkZTCj4gaXMgbXVsdGl0aHJlYW RlZC4gVGhvdWdodHM/CgpUaGUg
TkZTIHNlcnZlciBiaW5hcnkgSSBoYXZlIGNoZWNrZWQgaXMgbm 90IGxpbmtlZCB0byBsaWJwdGhy
ZWFkLCBidXQKeW91IG1heSBub3QgbmVlZCB0byBhbnN3ZXIgdG hlIHN1YmplY3QgcXVlc3Rpb24g
aW4geW91ciBjb250ZXh0OiAgbW9zdApvZiB0aGUgTkZTIHdvcm sgSSBoYXZlIHNlZW4gaW4gcGFz
dCB5ZWFycyBzZWVtcyB0byBiZSBvbiB0aGUgTkZTCmNsaWVudC BpbiBhcmVhcyBsaWtlIHBlcmZv
cm1hbmNlLCBjYWNoZSBjb25zaXN0ZW5jeSBhbmQgcHJvdG9jb2 wKZWZmaWNpZW5jeS4gIE5GUyBz
ZXJ2ZXIgd29yayBzZWVtcyB0byBnbyBpbiB0aGUgaW1wbGVtZW 50YXRpb24gb2YKTkZTdjQgYW5k
IEtlcmJlcm9zIHN1cHBvcnQuICBJbiBzaG9ydCwgSSBiZWxpZX ZlIHRoYXQgd2hldGhlciB0aGUg
TkZTCnNlcnZlciBpcyBtdWx0aS10aHJlYWRlZCBvciBub3QgbW F5IG5vdCBtYXR0ZXIuCgpVbmRl
ciBSZWQgSGF0IEVudGVycHJpc2UgTGludXggYW5kIGRlcml2YX RpdmVzLCB5b3UgY2FuIGNvbmZp
Z3VyZSB0aGUKbnVtYmVyIG9mIHBhcmFsbGVsIE5GUyBzZXJ2ZX IgcHJvY2Vzc2VzIGJ5IHNldHRp
bmcgUlBDTkZTRENPVU5UIGluCi9ldGMvc3lzY29uZmlnL25mcy 4gIFRoaXMgc2hvdWxkIHRha2Ug
YWR2YW50YWdlIG9mIHdoYXRldmVyIG51bWJlciBvZgpjb3Jlcy B5b3UgaGF2ZS4KCkkgc3VnZ2Vz
dCB5b3Ugd2F0Y2ggdGhlICJyZXRyYW5zIiBjb3VudCBpbiB0aG Ugb3V0cHV0IG9mICJuZnNzdGF0
Ci1yYyIuICBJZiB0aGlzIGtlZXBzIGluY3JlYXNpbmcgdW5kZX IgbG9hZCwgdGhlbiB5b3UgY291
bGQgaW5jcmVhc2UKdGhlIFJQQ05GU0RDT1VOVCB1bnRpbCB5b3 UgZ2V0IGEgY29tZm9ydGFibGUg
dmFsdWUuCgpHb29kIGx1Y2ssCgogLUltZWQKCi0tIApJbWVkIE NoaWhpIC0g2LnZhdin2K8g2KfZ
hNi02YrYrdmKCmh0dHA6Ly9wZXJzby5oZXhhYnl0ZS50bi9pY2 hpaGkvCgotLSAKcmVkaGF0LWxp
c3QgbWFpbGluZyBsaXN0CnVuc3Vic2NyaWJlIG1haWx0bzpyZW RoYXQtbGlzdC1yZXF1ZXN0QHJl
ZGhhdC5jb20/c3ViamVjdD11bnN1YnNjcmliZQpodHRwczovL3d3dy5yZWRoYX QuY29tL21haWxt
YW4vbGlzdGluZm8vcmVkaGF0LWxpc3Q=

02-13-2011, 02:04 PM

On Feb 11, 2011, at 3:23 AM, Michael Haubenwallner wrote:

>
> On 02/10/2011 10:11 PM, Perry Smith wrote:
>> On Feb 10, 2011, at 9:32 AM, Michael Haubenwallner wrote:
>>>
>>> Right now I've checked in the changes to use the new shared library
>>> model called "aix-soname" I've proposed recently along bug#213277 [1].
>>
>> Can you tell us which version you picked? At one point you had about
>> five ideas you were weighing.
>
> Ohw sorry, sure: The one I've posted at libtool mailing list[1][2] as well as
> in the IBM AIX Developers forum[3], and recently to the binutils mailing list[4].
> While the big agreement on that variant is still not there - maybe because of
> less interest (there's not much response anywhere), I've implemented that in
> Prefix now for both prove-of-concept as well as to identify possible problems.
>
> How I prefer doing "shared libraries" with "soname" on AIX now is:
>
> *) Create the shared object "shr.o" (using '-G' linker flag).
> *) Set the LOADONLY flag for "shr.o" (using 'strip -e').
> *) Create the Import File "shr.imp", containing
> - this header line, being the "soname" emulation:
> #! libNAME.so.1(shr.o)
> - another header line telling the bits used:
> # 32
> or
> # 64
> - the list of symbols exported.
> *) Create the archive library "libNAME.so.1.2.3" from both
> "shr.imp" and "shr.o".
> *) Create the symlinks as usual:
> libNAME.so.1 -> libNAME.so.1.2.3
> libNAME.so -> libNAME.so.1.2.3
> *) Create the static library archive "libNAME.a" from static objects.

Do you recall the advantage of having a import file plus the shared
object in the archive? What problem does that solve that having
just the shared object creates?

Perry

02-21-2011, 03:29 PM

On 02/13/2011 04:04 PM, Perry Smith wrote:
> On Feb 11, 2011, at 3:23 AM, Michael Haubenwallner wrote:
>>
>> How I prefer doing "shared libraries" with "soname" on AIX now is:
>>
>> *) Create the shared object "shr.o" (using '-G' linker flag).
>> *) Set the LOADONLY flag for "shr.o" (using 'strip -e').
>> *) Create the Import File "shr.imp", containing
>> - this header line, being the "soname" emulation:
>> #! libNAME.so.1(shr.o)
>> - another header line telling the bits used:
>> # 32
>> or
>> # 64
>> - the list of symbols exported.
>> *) Create the archive library "libNAME.so.1.2.3" from both
>> "shr.imp" and "shr.o".
>> *) Create the symlinks as usual:
>> libNAME.so.1 -> libNAME.so.1.2.3
>> libNAME.so -> libNAME.so.1.2.3
>> *) Create the static library archive "libNAME.a" from static objects.
>
> Do you recall the advantage of having a import file plus the shared
> object in the archive? What problem does that solve that having
> just the shared object creates?

Actually I've used to do standalone both Import File libNAME.so and
shared object libNAME.so.1 - which indeed did work for Prefix so far.
However, to get a better chance for upstream integration - especially
into libtool, I've had to look at the Big Picture:

*) Multilib:
Using an archive, it is possible to provide both 32/64 bit shared
objects within one single file - much like AIX libc.a and others do.

We don't do multilib in Prefix.

*) Dynamically load both with and without version number:
dlopen("libNAME.so(shr.o)", RTLD_MEMBER), besides the preferred
dlopen("libNAME.so.1(shr.o)", RTLD_MEMBER), and even
dlopen("libNAME.so.1.2.3(shr.o)", RTLD_MEMBER) does work.

Must admit that I've not seen a package I've built on AIX so far
where this actually is necessary.

*) All-in-one file to call "shared library (with soname)":
This is necessary to allow wrapping the linker to support the '-soname' linker
flag, so existing packages without explicit AIX build support can work right
out of the box. This even includes those symlinks known from ELF-world.

For Prefix this does help a lot as sys-devel/native-cctools provides this
ld-wrapper now: I've been able to drop *all* AIX-specific patches and
handling from within app-arch/bzip2-1.0.6-r1 now.

These (and more) reasons are also listed in libtool-list archives:
http://lists.gnu.org/archive/html/libtool/2011-01/msg00023.html
http://lists.gnu.org/archive/html/libtool/2011-02/msg00000.html

/haubi/
--
Michael Haubenwallner
Gentoo on a different level