[Bug build/30740] New: [m68k] undefined reference to `_wordcopy_fwd_dest_aligned'

[Bug build/30740] New: [m68k] undefined reference to `_wordcopy_fwd_dest_aligned'

[wbx at openadk dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30740

            Bug ID: 30740
           Summary: [m68k] undefined reference to
                    `_wordcopy_fwd_dest_aligned'
           Product: glibc
           Version: 2.38
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: build
          Assignee: unassigned at sourceware dot org
          Reporter: wbx at openadk dot org
                CC: carlos at redhat dot com
  Target Milestone: ---

Hi,

I am trying to create a cross-compile toolchain targeting qemu-system-m68k with
emulates a q800 machine with a m68040 cpu. The build of glibc fails with:

/home/wbx/openadk/toolchain_qemu-m68k-q800_glibc_68040/usr/bin/m68k-openadk-linux-gnu-gcc
  -nostdlib -nostartfiles -r -o
/home/wbx/openadk/toolchain_build_qemu-m68k-q800_glibc_68040/w-glibc-2.38-1/glibc-2.38-final/elf/librtld.os
'-Wl,-('
/home/wbx/openadk/toolchain_build_qemu-m68k-q800_glibc_68040/w-glibc-2.38-1/glibc-2.38-final/elf/dl-allobjs.os
/home/wbx/openadk/toolchain_build_qemu-m68k-q800_glibc_68040/w-glibc-2.38-1/glibc-2.38-final/elf/rtld-libc.a
-lgcc '-Wl,-)' \
         
-Wl,-Map,/home/wbx/openadk/toolchain_build_qemu-m68k-q800_glibc_68040/w-glibc-2.38-1/glibc-2.38-final/elf/librtld.os.map
/home/wbx/openadk/toolchain_qemu-m68k-q800_glibc_68040/usr/bin/m68k-openadk-linux-gnu-gcc
  -nostdlib -nostartfiles -shared -o
/home/wbx/openadk/toolchain_build_qemu-m68k-q800_glibc_68040/w-glibc-2.38-1/glibc-2.38-final/elf/ld.so.new
          \
          -Wl,-z,relro -Wl,-z,defs      \
           \
         
/home/wbx/openadk/toolchain_build_qemu-m68k-q800_glibc_68040/w-glibc-2.38-1/glibc-2.38-final/elf/librtld.os
-Wl,--version-script=/home/wbx/openadk/toolchain_build_qemu-m68k-q800_glibc_68040/w-glibc-2.38-1/glibc-2.38-final/ld.map
             \
          -Wl,-soname=ld.so.1
/home/wbx/openadk/toolchain_qemu-m68k-q800_glibc_68040/usr/lib/gcc/m68k-openadk-linux-gnu/12.3.0/../../../../m68k-openadk-linux-gnu/bin/ld:
/home/wbx/openadk/toolchain_build_qemu-m68k-q800_glibc_68040/w-glibc-2.38-1/glibc-2.38-final/elf/librtld.os:
in function `memcpy':
(.text+0x17bb4): undefined reference to `_wordcopy_fwd_aligned'
/home/wbx/openadk/toolchain_qemu-m68k-q800_glibc_68040/usr/lib/gcc/m68k-openadk-linux-gnu/12.3.0/../../../../m68k-openadk-linux-gnu/bin/ld:
(.text+0x17bf2): undefined reference to `_wordcopy_fwd_dest_aligned'
/home/wbx/openadk/toolchain_qemu-m68k-q800_glibc_68040/usr/lib/gcc/m68k-openadk-linux-gnu/12.3.0/../../../../m68k-openadk-linux-gnu/bin/ld:
/home/wbx/openadk/toolchain_build_qemu-m68k-q800_glibc_68040/w-glibc-2.38-1/glibc-2.38-final/elf/librtld.os:
in function `memmove':
(.text+0x17c68): undefined reference to `_wordcopy_fwd_aligned'
/home/wbx/openadk/toolchain_qemu-m68k-q800_glibc_68040/usr/lib/gcc/m68k-openadk-linux-gnu/12.3.0/../../../../m68k-openadk-linux-gnu/bin/ld:
(.text+0x17cda): undefined reference to `_wordcopy_bwd_aligned'
/home/wbx/openadk/toolchain_qemu-m68k-q800_glibc_68040/usr/lib/gcc/m68k-openadk-linux-gnu/12.3.0/../../../../m68k-openadk-linux-gnu/bin/ld:
(.text+0x17d24): undefined reference to `_wordcopy_bwd_dest_aligned'
/home/wbx/openadk/toolchain_qemu-m68k-q800_glibc_68040/usr/lib/gcc/m68k-openadk-linux-gnu/12.3.0/../../../../m68k-openadk-linux-gnu/bin/ld:
(.text+0x17d44): undefined reference to `_wordcopy_fwd_dest_aligned'
/home/wbx/openadk/toolchain_qemu-m68k-q800_glibc_68040/usr/lib/gcc/m68k-openadk-linux-gnu/12.3.0/../../../../m68k-openadk-linux-gnu/bin/ld:
/home/wbx/openadk/toolchain_build_qemu-m68k-q800_glibc_68040/w-glibc-2.38-1/glibc-2.38-final/elf/ld.so.new:
hidden symbol `_wordcopy_fwd_aligned' isn't defined
/home/wbx/openadk/toolchain_qemu-m68k-q800_glibc_68040/usr/lib/gcc/m68k-openadk-linux-gnu/12.3.0/../../../../m68k-openadk-linux-gnu/bin/ld:
final link failed: bad value
collect2: error: ld returned 1 exit status
gmake[8]: *** [Makefile:1328:
/home/wbx/openadk/toolchain_build_qemu-m68k-q800_glibc_68040/w-glibc-2.38-1/glibc-2.38-final/elf/ld.so]
Error 1
gmake[7]: *** [Makefile:484: elf/subdir_lib] Error 2
gmake[6]: *** [Makefile:9: all] Error 2
gmake[5]: *** [Makefile:70:
/home/wbx/openadk/toolchain_build_qemu-m68k-q800_glibc_68040/w-glibc-2.38-1/glibc-2.38/.compiled]
Error 2
gmake[4]: *** [Makefile:106: glibc-compile] Error 2
gmake[3]: *** [mk/build.mk:227: toolchain/final] Error 2
gmake[2]: *** [/home/wbx/openadk/mk/build.mk:176: world] Error 2

If it matters, I tried with -O2 and -Os, but same error occurs.

gcc -v:
./toolchain_qemu-m68k-q800_glibc_68040/usr/bin/m68k-openadk-linux-gnu-gcc -v
Using built-in specs.
COLLECT_GCC=./toolchain_qemu-m68k-q800_glibc_68040/usr/bin/m68k-openadk-linux-gnu-gcc
COLLECT_LTO_WRAPPER=/home/wbx/openadk/toolchain_qemu-m68k-q800_glibc_68040/usr/libexec/gcc/m68k-openadk-linux-gnu/12.3.0/lto-wrapper
Target: m68k-openadk-linux-gnu
Configured with:
/home/wbx/openadk/toolchain_build_qemu-m68k-q800_glibc_68040/w-gcc-12.3.0-1/gcc-12.3.0/configure
--prefix=/home/wbx/openadk/toolchain_qemu-m68k-q800_glibc_68040/usr
--with-bugurl=https://openadk.org --build=x86_64-linux-gnu
--host=x86_64-linux-gnu --target=m68k-openadk-linux-gnu
--with-gmp=/home/wbx/openadk/host_x86_64-linux-gnu/usr
--with-mpfr=/home/wbx/openadk/host_x86_64-linux-gnu/usr --enable-__cxa_atexit
--with-system-zlib --with-gnu-ld --with-gnu-as --disable-libsanitizer
--disable-install-libiberty --disable-libitm --disable-libmudflap
--disable-libgomp --disable-libcc1 --disable-libmpx --disable-libcilkrts
--disable-libquadmath --disable-libquadmath-support --disable-decimal-float
--disable-gcov --disable-libstdcxx-pch --disable-ppl-version-check
--disable-cloog-version-check --without-ppl --without-cloog --without-isl
--disable-werror --disable-nls --enable-obsolete --disable-lto --with-cpu=68040
--enable-tls --enable-languages=c --with-newlib --disable-shared
--disable-threads --disable-multilib --without-headers
Thread model: single
Supported LTO compression algorithms: zlib
gcc version 12.3.0 (GCC) 

./toolchain_qemu-m68k-q800_glibc_68040/usr/bin/m68k-openadk-linux-gnu-ld -v
GNU ld (GNU Binutils) 2.40

Linux 6.1.41 is used.

Anyone could help here?
 best regards
   Waldemar

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug build/30740] [m68k] undefined reference to `_wordcopy_fwd_dest_aligned'

[adhemerval.zanella at linaro dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30740

Adhemerval Zanella <adhemerval.zanella at linaro dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |adhemerval.zanella at linaro dot o
                   |                            |rg

--- Comment #1 from Adhemerval Zanella <adhemerval.zanella at linaro dot org> ---
The issue is that we only test for m68020, which should use the optimized
macros from sysdeps/m68k/memcopy.h and not use the wordcopy.c routines. To fix
it should be as simple as:

diff --git a/sysdeps/m68k/m680x0/m68020/wordcopy.S
b/sysdeps/m68k/m680x0/m68020/wordcopy.S
deleted file mode 100644
index 4fb1a4518f..0000000000
--- a/sysdeps/m68k/m680x0/m68020/wordcopy.S
+++ /dev/null
@@ -1 +0,0 @@
-/* Empty, not needed.  */
diff --git a/sysdeps/m68k/wordcopy.c b/sysdeps/m68k/wordcopy.c
new file mode 100644
index 0000000000..e459e0e403
--- /dev/null
+++ b/sysdeps/m68k/wordcopy.c
@@ -0,0 +1,21 @@
+/* Definitions for memory copy functions.  Motorola 68020 version.
+   Copyright (C) 2023 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library.  If not, see
+   <https://www.gnu.org/licenses/>.  */
+
+#if !defined(__mc68020__) && !defined(mc68020)
+#include <string/wordcopy.c>
+#endif

However, since we do not actually build everything for anything different than
m68020; the math code assumes m68020 FLT_EVAL_METHOD (which is 2 since it uses
the external 68881 math processor).  And this is not correct fro m68040 which
now has a builtin FP processor and thus GCC uses a different  FLT_EVAL_METHOD
(0).

The FP issues should be simple to fix by removing
sysdeps/m68k/m680x0/bits/flt-eval-method.h and letting using the generic one,
but I don't have a way to actually test it.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug build/30740] [m68k] undefined reference to `_wordcopy_fwd_dest_aligned'

[wbx at openadk dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30740

--- Comment #2 from wbx at openadk dot org ---
Hi,

that worked fine for the toolchain creation and a simple bootup in Qemu works
as well.
Thanks for the fast answer.

There is one minor issue, this oneliner does not work, I use it to show the ip
address of the system:
ipaddr=$(ip addr show $(ip route show|awk '/default/ { print $5 }')|awk '/inet
/ { print $2 }'|tail -1)

With strace I get:
Fatal glibc error: printf_fp.c:501 (__printf_fp_buffer_1): assertion failed: cy
== 1 || (p.frac[p.fracsize - 2] == 0 && p.frac[0] == 0)
Aborted

Does this have something todo with the FP implementation?

best regards
 Waldemar

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug build/30740] [m68k] undefined reference to `_wordcopy_fwd_dest_aligned'

[adhemerval.zanella at linaro dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30740

--- Comment #3 from Adhemerval Zanella <adhemerval.zanella at linaro dot org> ---
(In reply to wbx from comment #2)
> Hi,
> 
> that worked fine for the toolchain creation and a simple bootup in Qemu
> works as well.
> Thanks for the fast answer.
> 
> There is one minor issue, this oneliner does not work, I use it to show the
> ip address of the system:
> ipaddr=$(ip addr show $(ip route show|awk '/default/ { print $5 }')|awk
> '/inet / { print $2 }'|tail -1)
> 
> With strace I get:
> Fatal glibc error: printf_fp.c:501 (__printf_fp_buffer_1): assertion failed:
> cy == 1 || (p.frac[p.fracsize - 2] == 0 && p.frac[0] == 0)
> Aborted
> 
> Does this have something todo with the FP implementation?
> 
> best regards
>  Waldemar

I am not sure, it indeed could be related since the assert being triggered is
related to floating point printing. Do you know exactly what awk issues with
printf?

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug build/30740] [m68k] undefined reference to `_wordcopy_fwd_dest_aligned'

[adhemerval.zanella at linaro dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30740

--- Comment #4 from Adhemerval Zanella <adhemerval.zanella at linaro dot org> ---
Could you try to remove sysdeps/m68k/m680x0/lshift.S and see if program does
not trigger the issue anymore?

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug build/30740] [m68k] undefined reference to `_wordcopy_fwd_dest_aligned'

[wbx at openadk dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30740

--- Comment #5 from wbx at openadk dot org ---
Hi,

wonderful, this fixed the awk issue for me.

Any chance to get these changes upstream for the next release?

best regards
 Waldemar

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug build/30740] [m68k] undefined reference to `_wordcopy_fwd_dest_aligned'

[adhemerval.zanella at linaro dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30740

--- Comment #6 from Adhemerval Zanella <adhemerval.zanella at linaro dot org> ---
(In reply to wbx from comment #5)
> Hi,
> 
> wonderful, this fixed the awk issue for me.
> 
> Any chance to get these changes upstream for the next release?
> 
> best regards
>  Waldemar

Good to know and I hope to get this sort out. Do you have a easy way to
bootstrap a m68k using qemu? I am trying to use debian, but it keep breaking
which is really time consuming... I would like to check if both 68020 and 68040
works with these changes.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug build/30740] [m68k] undefined reference to `_wordcopy_fwd_dest_aligned'

[wbx at openadk dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30740

--- Comment #7 from wbx at openadk dot org ---
If you like to cross-compile, you can use www.openadk.org, it supports 68020
and 68040
and both can be booted in Qemu. Buildroot works, too, but it only supports
68040, so you would need to hack 68020 toolchain support into it. And you would
need this:
https://lists.buildroot.org/pipermail/buildroot/2023-August/672637.html

I don't know any recent native Linux distro, which can be used in Qemu M68k.

best regards
 Waldemar

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug build/30740] [m68k] undefined reference to `_wordcopy_fwd_dest_aligned'

[wbx at openadk dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30740

--- Comment #8 from wbx at openadk dot org ---
Hi,

You can download following two files:
https://debug.openadk.org/m68k/qemu-m68k.img
https://debug.openadk.org/m68k/qemu-m68k-virt-archive-kernel

It contains everything you need to compile glibc.
gcc 13.2.0, binutils 2.41, ...

And start Qemu like this:
qemu-system-m68k -nographic -M virt -m 512M -netdev user,id=eth0 -device
virtio-net-device,netdev=eth0 -append "rootwait root=/dev/vda1 ro" -drive
file=qemu-m68k.img,format=raw,id=hd0 -device virtio-blk-device,drive=hd0
-kernel qemu-m68k-virt-archive-kernel

User: root
Password: linux123

Under /opt is a build directory, which contains some precompiled code.

Unfortunately the native build fails with following error:
make[2]: Entering directory '/opt/glibc-2.38/malloc'
gcc   -shared -static-libgcc -Wl,-O1  -Wl,-z,defs
-Wl,-dynamic-linker=/opt/lib/ld.so.1   -B/opt/build/csu/ 
-Wl,--version-script=/opt/build/libc_malloc_debug.map
-Wl,-soname=libc_malloc_debug.so.0 -Wl,-z,relro   
-L/opt/build -L/opt/build/math -L/opt/build/elf -L/opt/build/dlfcn
-L/opt/build/nss -L/opt/build/nis -L/opt/build/rt -L/opt/build/resolv
-L/opt/build/mathvec -L/opt/build/support -L/opt/build/nptl -Wl,-rpath-link
=/opt/build:/opt/build/math:/opt/build/elf:/opt/build/dlfcn:/opt/build/nss:/opt/build/nis:/opt/build/rt:/opt/build/resolv:/opt/build/mathvec:/opt/build/support:/opt/build/nptl
-o /opt/build/malloc/libc_malloc_deb
ug.so /opt/build/csu/abi-note.o -Wl,--whole-archive
/opt/build/malloc/libc_malloc_debug_pic.a -Wl,--no-whole-archive  
-Wl,--start-group /opt/build/libc.so /opt/build/libc_nonshared.a
-Wl,--as-needed /opt/build/e
lf/ld.so -Wl,--no-as-needed -Wl,--end-group
/usr/bin/ld: /opt/build/malloc/libc_malloc_debug_pic.a(malloc-debug.os): in
function `do_mtrace':
/opt/glibc-2.38/malloc/mtrace-impl.c:193:(.text+0x163c): undefined reference to
`__dso_handle'
/usr/bin/ld: /opt/build/malloc/libc_malloc_debug.so: hidden symbol
`__dso_handle' isn't defined
/usr/bin/ld: final link failed: bad value
collect2: error: ld returned 1 exit status
make[2]: *** [../Makerules:533: /opt/build/malloc/libc_malloc_debug.so] Error 1
make[2]: Leaving directory '/opt/glibc-2.38/malloc'
make[1]: *** [Makefile:484: malloc/others] Error 2
make[1]: Leaving directory '/opt/glibc-2.38'
make: *** [Makefile:9: all] Error 2

Any idea why it fails here?
best regards
 Waldemar

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug build/30740] [m68k] undefined reference to `_wordcopy_fwd_dest_aligned'

[adhemerval.zanella at linaro dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30740

--- Comment #9 from Adhemerval Zanella <adhemerval.zanella at linaro dot org> ---
(In reply to wbx from comment #8)
> Hi,
> 
> You can download following two files:
> https://debug.openadk.org/m68k/qemu-m68k.img
> https://debug.openadk.org/m68k/qemu-m68k-virt-archive-kernel
> 
> It contains everything you need to compile glibc.
> gcc 13.2.0, binutils 2.41, ...
> 
> And start Qemu like this:
> qemu-system-m68k -nographic -M virt -m 512M -netdev user,id=eth0 -device
> virtio-net-device,netdev=eth0 -append "rootwait root=/dev/vda1 ro" -drive
> file=qemu-m68k.img,format=raw,id=hd0 -device virtio-blk-device,drive=hd0
> -kernel qemu-m68k-virt-archive-kernel
> 
> User: root
> Password: linux123
> 
> Under /opt is a build directory, which contains some precompiled code.
> 
> Unfortunately the native build fails with following error:
> make[2]: Entering directory '/opt/glibc-2.38/malloc'
> gcc   -shared -static-libgcc -Wl,-O1  -Wl,-z,defs
> -Wl,-dynamic-linker=/opt/lib/ld.so.1   -B/opt/build/csu/ 
> -Wl,--version-script=/opt/build/libc_malloc_debug.map
> -Wl,-soname=libc_malloc_debug.so.0 -Wl,-z,relro   
> -L/opt/build -L/opt/build/math -L/opt/build/elf -L/opt/build/dlfcn
> -L/opt/build/nss -L/opt/build/nis -L/opt/build/rt -L/opt/build/resolv
> -L/opt/build/mathvec -L/opt/build/support -L/opt/build/nptl -Wl,-rpath-link
> =/opt/build:/opt/build/math:/opt/build/elf:/opt/build/dlfcn:/opt/build/nss:/
> opt/build/nis:/opt/build/rt:/opt/build/resolv:/opt/build/mathvec:/opt/build/
> support:/opt/build/nptl -o /opt/build/malloc/libc_malloc_deb
> ug.so /opt/build/csu/abi-note.o -Wl,--whole-archive
> /opt/build/malloc/libc_malloc_debug_pic.a -Wl,--no-whole-archive  
> -Wl,--start-group /opt/build/libc.so /opt/build/libc_nonshared.a
> -Wl,--as-needed /opt/build/e
> lf/ld.so -Wl,--no-as-needed -Wl,--end-group
> /usr/bin/ld: /opt/build/malloc/libc_malloc_debug_pic.a(malloc-debug.os): in
> function `do_mtrace':
> /opt/glibc-2.38/malloc/mtrace-impl.c:193:(.text+0x163c): undefined reference
> to `__dso_handle'
> /usr/bin/ld: /opt/build/malloc/libc_malloc_debug.so: hidden symbol
> `__dso_handle' isn't defined
> /usr/bin/ld: final link failed: bad value
> collect2: error: ld returned 1 exit status
> make[2]: *** [../Makerules:533: /opt/build/malloc/libc_malloc_debug.so]
> Error 1
> make[2]: Leaving directory '/opt/glibc-2.38/malloc'
> make[1]: *** [Makefile:484: malloc/others] Error 2
> make[1]: Leaving directory '/opt/glibc-2.38'
> make: *** [Makefile:9: all] Error 2
> 
> Any idea why it fails here?
> best regards
>  Waldemar

It seems to be a misconfigured toolchain, where gcc crtbegin files do not
contain the __dso_handle:

$ objdump -t /usr/lib/gcc/m68k-openadk-linux-gnu/13.2.0/crtbeginS.o
[...]
00000000         *UND*  00000000 _GLOBAL_OFFSET_TABLE_
00000000         *UND*  00000000 .hidden __TMC_END__
00000000  w      *UND*  00000000 _ITM_deregisterTMCloneTable
00000000  w      *UND*  00000000 _ITM_registerTMCloneTable
[...]

With a compiler built with build-many-glibcs.py:

$ objdump -t
/home/azanella/toolchain/install/compilers/13/m68k-linux-gnu/bin/../lib/gcc/m68k-glibc-linux-gnu/13.1.1/crtbeginS.o
00000000         *UND*  00000000 _GLOBAL_OFFSET_TABLE_
00000000         *UND*  00000000 .hidden __TMC_END__
00000000  w      *UND*  00000000 _ITM_deregisterTMCloneTable
00000000  w      *UND*  00000000 _ITM_registerTMCloneTable
00000000 g     O .data.rel.local        00000000 .hidden __dso_handle
00000000  w      *UND*  00000000 __cxa_finalize

And checking the GCC configuration, it seems to be from
'--disable-__cxa_atexit'. From gcc/config.in:

/* Define if you want to use __cxa_atexit, rather than atexit, to register C++
   destructors for local statics and global objects. This is essential for
   fully standards-compliant handling of destructors, but requires
   __cxa_atexit in libc. */
#ifndef USED_FOR_TARGET
#undef DEFAULT_USE_CXA_ATEXIT
#endif

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug build/30740] [m68k] undefined reference to `_wordcopy_fwd_dest_aligned'

[wbx at openadk dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30740

--- Comment #10 from wbx at openadk dot org ---
Hi,
You are right. I figured it out myself, too and rebuild the gcc in the
meanwhile.
You find an updated image on the webserver. This time the build of glibc
succeeded!

best regards
 Waldemar

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug build/30740] [m68k] undefined reference to `_wordcopy_fwd_dest_aligned'

[adhemerval.zanella at linaro dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30740

--- Comment #11 from Adhemerval Zanella <adhemerval.zanella at linaro dot org> ---
Thanks for confirming it.  In any case, I will install the straightforward fix
that breaks the m68k build.  Could you open another one for the lshift.S issue
on 68040?

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug build/30740] [m68k] undefined reference to `_wordcopy_fwd_dest_aligned'

[wbx at openadk dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30740

--- Comment #12 from wbx at openadk dot org ---
https://sourceware.org/bugzilla/show_bug.cgi?id=30773

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug build/30740] [m68k] undefined reference to `_wordcopy_fwd_dest_aligned'

[wbx at openadk dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30740

--- Comment #13 from wbx at openadk dot org ---
Hi,
running make check fails with:

gcc -O   -D_GNU_SOURCE -DIS_IN_build -include /opt/build/config.h isomac.c -o
/opt/build/stdlib/isomac
In file included from isomac.c:71:
/usr/include/stdlib.h:153:8: error: '_Float128' is not supported on this target
  153 | extern _Float128 strtof128 (const char *__restrict __nptr,
      |        ^~~~~~~~~
/usr/include/stdlib.h:165:8: error: '_Float64x' is not supported on this target
  165 | extern _Float64x strtof64x (const char *__restrict __nptr,
      |        ^~~~~~~~~
/usr/include/stdlib.h:311:25: error: '_Float128' is not supported on this
target
  311 |                         _Float128 __f)
      |                         ^~~~~~~~~
/usr/include/stdlib.h:323:25: error: '_Float64x' is not supported on this
target
  323 |                         _Float64x __f)
      |                         ^~~~~~~~~
/usr/include/stdlib.h:450:8: error: '_Float128' is not supported on this target
  450 | extern _Float128 strtof128_l (const char *__restrict __nptr,
      |        ^~~~~~~~~
/usr/include/stdlib.h:464:8: error: '_Float64x' is not supported on this target
  464 | extern _Float64x strtof64x_l (const char *__restrict __nptr,
      |        ^~~~~~~~~
make[2]: *** [Makefile:477: /opt/build/stdlib/isomac] Error 1
rm /opt/build/stdlib/test-as-const-ucontext_i.c
make[2]: Leaving directory '/opt/glibc-2.38/stdlib'
make[1]: *** [Makefile:484: stdlib/tests] Error 2
make[1]: Leaving directory '/opt/glibc-2.38'
make: *** [Makefile:9: check] Error 2
2|root@openadk:/opt/build #

What is wrong here?

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug build/30740] [m68k] undefined reference to `_wordcopy_fwd_dest_aligned'

[adhemerval.zanella at linaro dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30740

--- Comment #14 from Adhemerval Zanella <adhemerval.zanella at linaro dot org> ---
(In reply to wbx from comment #13)
> Hi,
> running make check fails with:
> 
> gcc -O   -D_GNU_SOURCE -DIS_IN_build -include /opt/build/config.h isomac.c
> -o /opt/build/stdlib/isomac
> In file included from isomac.c:71:
> /usr/include/stdlib.h:153:8: error: '_Float128' is not supported on this
> target
>   153 | extern _Float128 strtof128 (const char *__restrict __nptr,
>       |        ^~~~~~~~~
> /usr/include/stdlib.h:165:8: error: '_Float64x' is not supported on this
> target
>   165 | extern _Float64x strtof64x (const char *__restrict __nptr,
>       |        ^~~~~~~~~
> /usr/include/stdlib.h:311:25: error: '_Float128' is not supported on this
> target
>   311 |                         _Float128 __f)
>       |                         ^~~~~~~~~
> /usr/include/stdlib.h:323:25: error: '_Float64x' is not supported on this
> target
>   323 |                         _Float64x __f)
>       |                         ^~~~~~~~~
> /usr/include/stdlib.h:450:8: error: '_Float128' is not supported on this
> target
>   450 | extern _Float128 strtof128_l (const char *__restrict __nptr,
>       |        ^~~~~~~~~
> /usr/include/stdlib.h:464:8: error: '_Float64x' is not supported on this
> target
>   464 | extern _Float64x strtof64x_l (const char *__restrict __nptr,
>       |        ^~~~~~~~~
> make[2]: *** [Makefile:477: /opt/build/stdlib/isomac] Error 1
> rm /opt/build/stdlib/test-as-const-ucontext_i.c
> make[2]: Leaving directory '/opt/glibc-2.38/stdlib'
> make[1]: *** [Makefile:484: stdlib/tests] Error 2
> make[1]: Leaving directory '/opt/glibc-2.38'
> make: *** [Makefile:9: check] Error 2
> 2|root@openadk:/opt/build #
> 
> What is wrong here?

This seems another toolchain misconfiguration, the VM
/usr/include/bits/floatn.h seems to be the x86 one instead of the generic one
which the expected one for m68k.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug build/30740] [m68k] undefined reference to `_wordcopy_fwd_dest_aligned'

[chewi at gentoo dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30740

James Le Cuirot <chewi at gentoo dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |chewi at gentoo dot org

--- Comment #15 from James Le Cuirot <chewi at gentoo dot org> ---
Thank you for looking into this, Adhemerval. Please could you confirm that
you've considered the 68030 case. This does not have a built-in FPU, so I think
it should be treated like the 68020? I'm no expert here though.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug build/30740] [m68k] undefined reference to `_wordcopy_fwd_dest_aligned'

[wbx at openadk dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30740

--- Comment #16 from wbx at openadk dot org ---
Hi,
I fixed the misconfiguration and added the required bash, now make check
completes
with following output:
Summary of test results:
    949 FAIL
   3344 PASS
     99 UNSUPPORTED
     16 XFAIL
      2 XPASS

A lot of math tests are failing, is it because of a missing ULPs file?

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug build/30740] [m68k] undefined reference to `_wordcopy_fwd_dest_aligned'

[adhemerval.zanella at linaro dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30740

--- Comment #17 from Adhemerval Zanella <adhemerval.zanella at linaro dot org> ---
(In reply to James Le Cuirot from comment #15)
> Thank you for looking into this, Adhemerval. Please could you confirm that
> you've considered the 68030 case. This does not have a built-in FPU, so I
> think it should be treated like the 68020? I'm no expert here though.

glibc assumes the existence of FPU for m68k, and GCC defaults to 68881:

$ m68k-glibc-linux-gnu-gcc -mcpu=68030 -E -dD - < /dev/null  | grep 68881
#define __HAVE_68881__ 1

The math tests do all build because different than mc68020, GCC defines
__DEC_EVAL_METHOD__ to 2 for mc68030.

The soft-float is only support for coldfire (__mcoldfire__) cpus.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug build/30740] [m68k] undefined reference to `_wordcopy_fwd_dest_aligned'

[adhemerval.zanella at linaro dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30740

--- Comment #18 from Adhemerval Zanella <adhemerval.zanella at linaro dot org> ---
(In reply to wbx from comment #16)
> Hi,
> I fixed the misconfiguration and added the required bash, now make check
> completes
> with following output:
> Summary of test results:
>     949 FAIL
>    3344 PASS
>      99 UNSUPPORTED
>      16 XFAIL
>       2 XPASS
> 
> A lot of math tests are failing, is it because of a missing ULPs file?

I am not sure, the last m68k testsuite report where from 2.26 release [1]
running under ARAnyM.  A lot of tests have been added since then, so we will
need to check each failure.

For ULPs update, you can try "make regen-ulps" to generate a diff file with
newer UPs values (the regeneration ignores bogus ULPs values).  But I guess the
math failures are from the extra rounding and exception testing, which requires
a fully compliant IEEE 754 fp unit (which I guess m68k does not provide).

[1] https://sourceware.org/glibc/wiki/Release/2.26#M68K

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug build/30740] [m68k] undefined reference to `_wordcopy_fwd_dest_aligned'

[adhemerval.zanella at linaro dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30740

--- Comment #19 from Adhemerval Zanella <adhemerval.zanella at linaro dot org> ---
(In reply to Adhemerval Zanella from comment #17)
> The math tests do all build because different than mc68020, GCC defines
> __DEC_EVAL_METHOD__ to 2 for mc68030.

*Some* math tests do *not* build ...

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug build/30740] [m68k] undefined reference to `_wordcopy_fwd_dest_aligned'

[adhemerval.zanella at linaro dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30740

--- Comment #20 from Adhemerval Zanella <adhemerval.zanella at linaro dot org> ---
(In reply to Adhemerval Zanella from comment #19)
> (In reply to Adhemerval Zanella from comment #17)
> > The math tests do all build because different than mc68020, GCC defines
> > __DEC_EVAL_METHOD__ to 2 for mc68030.
> 
> *Some* math tests do *not* build ...

Another correction, the __DEC_EVAL_METHOD__ issues is for mc68040 in fact (not
mc68030).

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug build/30740] [m68k] undefined reference to `_wordcopy_fwd_dest_aligned'

[adhemerval.zanella at linaro dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30740

Adhemerval Zanella <adhemerval.zanella at linaro dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |FIXED
   Target Milestone|---                         |2.39
             Status|UNCONFIRMED                 |RESOLVED

--- Comment #21 from Adhemerval Zanella <adhemerval.zanella at linaro dot org> ---
The build issues was fixed on 2.39 (b85880633f69b737050c22eede858f8c13bb2248).

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug build/30740] [m68k] undefined reference to `_wordcopy_fwd_dest_aligned'

[sam at gentoo dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30740

Sam James <sam at gentoo dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sam at gentoo dot org

-- 
You are receiving this mail because:
You are on the CC list for the bug.