Gentoo Archives: gentoo-dev

From: Leno Hou <lenohou@×××××.com>
To: Leno Hou <lenohou@×××××.com>, gentoo-dev@l.g.o, "Anthony G. Basile" <blueness@g.o>, toolchain@g.o
Subject: Re: [gentoo-dev][PATCH v2 1/2] eclass/libtool: fix unable read SONAME for ppc64le #580792
Date: Sun, 24 Apr 2016 11:32:07
Message-Id: CAGQVrL_USJT=R6FeMQTxit-FL4mEUUK0TU6mt2=EN91gmWc_Ow@mail.gmail.com
In Reply to: Re: [gentoo-dev][PATCH v2 1/2] eclass/libtool: fix unable read SONAME for ppc64le #580792 by Mike Frysinger
1 On Sat, Apr 23, 2016 at 2:16 PM, Mike Frysinger <vapier@g.o> wrote:
2
3 > On 22 Apr 2016 03:44, Leno Hou wrote:
4 > > 2) The problem is due to the wrong $LD="ld -m elf64ppc" when link.
5 > > This patch sets $LD="ld -m elf64lppc" in m4/libtool.m4 on ppc64le.
6 >
7 > i'm looking at the upstream libtool project and i'm not seeing
8 > these changes in there. have you sent patches to them to update
9 > libtool ? otherwise, all packages will continue to be created
10 > with bad versions in them ...
11 >
12 >
13 1) I have reported the libtool in upstream. See here
14 http://debbugs.gnu.org/cgi/bugreport.cgi?bug=23348
15
16 2) The new PATCH V3 which called
17 [gentoo-dev][PATCH V3] libtool.eclass:elibtoolize: add ppc64le handling
18 #580792
19 was sent to mailing list and cc'ed you
20 has update libtool-2.4.6
21
22 what upstream has right now is:
23 > if AC_TRY_EVAL(ac_compile); then
24 > case `/usr/bin/file conftest.o` in
25 > *32-bit*)
26 > case $host in
27 > ...
28 > powerpc64le-*linux*)
29 > LD="${LD-ld} -m elf32lppclinux"
30 > ;;
31 > powerpc64-*linux*)
32 > LD="${LD-ld} -m elf32ppclinux"
33 > ;;
34 > ...
35 > *64-bit*)
36 > case $host in
37 > ...
38 > powerpcle-*linux*)
39 > LD="${LD-ld} -m elf64lppc"
40 > ;;
41 > powerpc-*linux*)
42 > LD="${LD-ld} -m elf64ppc"
43 > ;;
44 > ...
45 >
46 > your 2.4.2 patch makes sense in that it had no LE support, so you
47 > need to add the tuple matching. however, the 2.4.4 & 2.4.6 patches
48 > are applying to libtool versions that know how to handle LE.
49 >
50
51 3) The 2.4.4 and 2.4.6 are applied correct lib tool version that know how
52 to handle LE
53
54
55 > keep in mind that the logic here is just for handling mismatch in
56 > the expected bitness and the tuple. i.e. when you're using a tuple
57 > like powerpc-linux-gnu (which is normally 32bit) but you're actually
58 > producing 64-bit code.
59 >
60
61 4) Is tuple
62 ....
63 powerpcle-*linux*)
64 LD="${LD-ld} -m elf64lppc"
65
66 should be changed to
67 ....
68 powerpcle-*linux*)
69 LD="${LD-ld} -m elf32lppc"
70
71 to producing 32-bit code?
72
73 If yes, I'll send new PATCH to cover this.
74
75 -Leno Hou