Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 29 Oct 2023 20:05:14 +0100 (CET)
From:      Wojciech Puchar <wojtek@puchar.net>
To:        Mark Millard <marklmi@yahoo.com>
Cc:        Dimitry Andric <dim@FreeBSD.org>, FreeBSD Toolchain <freebsd-toolchain@freebsd.org>, FreeBSD Hackers <freebsd-hackers@freebsd.org>
Subject:   Re: problem linking kernel 
Message-ID:  <8e5b7eea-cfb-1cbb-8eb-18ea18b230d4@puchar.net>
In-Reply-To: <FA0E9EEA-B282-4DBD-8FF8-183DAD91CFE4@yahoo.com>
References:  <FA0E9EEA-B282-4DBD-8FF8-183DAD91CFE4.ref@yahoo.com> <FA0E9EEA-B282-4DBD-8FF8-183DAD91CFE4@yahoo.com>

next in thread | previous in thread | raw e-mail | index | archive | help
  This message is in MIME format.  The first part should be readable text,
  while the remaining parts are likely unreadable without MIME-aware tools.

--3974694515-1688657752-1698606316=:1940
Content-Type: text/plain; charset=US-ASCII; format=flowed
Content-Transfer-Encoding: 8BIT

right but anyway it was my fault i for some reason some time ago changed 
$PATH so /usr/local had precedence.

On Sun, 29 Oct 2023, Mark Millard wrote:

> From: Wojciech Puchar <wojtek_at_puchar.net> wrote on
> Date: Sun, 29 Oct 2023 12:16:31 UTC :
>
>> reason found.
>>
>>
>> /usr/local/bin/ld which is part of gcc from ports takes precedence
>
> I've run into the FreeBSD system toolchain using a linker
> that was below /usr/local/ multiple times before. I try to
> avoid building ports that results in the conflict.
>
> But, in my view, the FreeBSD toolchain should require an
> explicit command line option to be willing to use part
> of a toolchain that is someplace under /usr/local/ instead
> of using its own toolchain materials for everything.
>
> See, for example, my report from 2022-Sept:
>
> https://lists.freebsd.org/archives/freebsd-toolchain/2022-September/000898.html
>
> where, in an aarch64 context, it used:
>
> /usr/local/bin/aarch64-unknown-freebsd14.0-ld
>
> instead of:
>
> /sur/bin/ld
> or:
> /usr/bin/dl.lld
>
> It did not do this if -fuse-ld=lld was explicitly indicated on
> the command line: one way to avoid the problem.
>
> What is used can be seen in the extra output from use of:
>
> clang++ -v
>
> Back then the example was:
>
> clang++ -v -std=c++20 -Wpedantic -Wall -Wextra   -I../other_src_used  -pedantic  -g3 -O3 -mcpu=cortex-a72 -flto=thin  -pthread -flto=thin ../objs/cpp_thousandslocale-clang++_14_O3lto-libc++.o  ../objs/cpp_clockinfo-clang++_14_O3lto-libc++.o -o ../cpp_clockinfo_main-HoneyComb-65536MiB-threads_16-LP64-FreeBSD_main_n256584_5bc926af9fd1_64bit-clang++_14_O3lto-libc++  -DCPPCLOCKINFO_VERS='"acpphint_0.1.24"'  ../other_src_used/cpp_clockinfo_main.cpp
> FreeBSD clang version 14.0.5 (https://github.com/llvm/llvm-project.git llvmorg-14.0.5-0-gc12386ae247c)
> Target: aarch64-unknown-freebsd14.0
> Thread model: posix
> InstalledDir: /usr/bin
> "/usr/bin/clang++" -cc1 -triple aarch64-unknown-freebsd14.0 -emit-llvm-bc -flto=thin -flto-unit -disable-free -clear-ast-before-backend -disable-llvm-verifier -discard-value-names -main-file-name cpp_clockinfo_main.cpp -mrelocation-model static -mframe-pointer=non-leaf -ffp-contract=on -fno-rounding-math -mconstructor-aliases -funwind-tables=2 -target-cpu cortex-a72 -target-feature +v8a -target-feature +fp-armv8 -target-feature +neon -target-feature +crc -target-feature +crypto -target-feature +sha2 -target-feature +aes -target-abi aapcs -fallow-half-arguments-and-returns -mllvm -treat-scalable-fixed-error-as-warning -debug-info-kind=standalone -dwarf-version=4 -debugger-tuning=gdb -v -fcoverage-compilation-dir=/root/acpphint/acpphint_src -resource-dir /usr/lib/clang/14.0.5 -I ../other_src_used -D "CPPCLOCKINFO_VERS=\"acpphint_0.1.24\"" -internal-isystem /usr/include/c++/v1 -O3 -Wpedantic -Wall -Wextra -pedantic -std=c++20 -fdeprecated-macro -fdebug-compilation-dir=/root/a!
 cpphint/acpphint_src -ferror-limit 19 -pthread -fno-signed-char -fgnuc-version=4.2.1 -fno-implicit-modules -fcxx-exceptions -fexceptions -fcolor-diagnostics -vectorize-loops -vectorize-slp -faddrsig -D__GCC_HAVE_DWARF2_CFI_ASM=1 -o /tmp/cpp_clockinfo_main-56273f.o -x c++ ../other_src_used/cpp_clockinfo_main.cpp
> clang -cc1 version 14.0.5 based upon LLVM 14.0.5 default target aarch64-unknown-freebsd14.0
> #include "..." search starts here:
> #include <...> search starts here:
> ../other_src_used
> /usr/include/c++/v1
> /usr/lib/clang/14.0.5/include
> /usr/include
> End of search list.
> "/usr/local/bin/aarch64-unknown-freebsd14.0-ld" --eh-frame-hdr -dynamic-linker /libexec/ld-elf.so.1 --enable-new-dtags -o ../cpp_clockinfo_main-HoneyComb-65536MiB-threads_16-LP64-FreeBSD_main_n256584_5bc926af9fd1_64bit-clang++_14_O3lto-libc++ /usr/lib/crt1.o /usr/lib/crti.o /usr/lib/crtbegin.o -L/usr/lib -plugin /usr/bin/../lib/LLVMgold.so -plugin-opt=mcpu=cortex-a72 -plugin-opt=O3 -plugin-opt=thinlto ../objs/cpp_thousandslocale-clang++_14_O3lto-libc++.o ../objs/cpp_clockinfo-clang++_14_O3lto-libc++.o /tmp/cpp_clockinfo_main-56273f.o -lc++ -lm -lgcc --as-needed -lgcc_s --no-as-needed -lpthread -lc -lgcc --as-needed -lgcc_s --no-as-needed /usr/lib/crtend.o /usr/lib/crtn.o
> /usr/local/bin/aarch64-unknown-freebsd14.0-ld: /usr/bin/../lib/LLVMgold.so: error loading plugin: Cannot open "/usr/bin/../lib/LLVMgold.so"
> clang++: error: linker command failed with exit code 1 (use -v to see invocation)
> *** Error code 1
>
> Note the: "/usr/local/bin/aarch64-unknown-freebsd14.0-ld" after the search list.
>
> It looks like I did not report my PATH back then. But finding and using
> aarch64-unknown-freebsd14.0-ld is not going to be found under any
> system path for the system toolchain. So my  view is that the name
> should not be looked for by the system toolchain --or should be last
> if the normal system names are not found.
>
> FYI:
>
> # echo $PATH
> /sbin:/bin:/usr/sbin:/usr/bin:/usr/local/sbin:/usr/local/bin:/usr/home/root/bin
>
> and it is unlikely to have had any changes since back then.
>
>> PATH=/sbin:/bin:/usr/sbin:/usr/bin:/usr/games make
>>
>> and all builds
>>
>> On Sun, 29 Oct 2023, Dimitry Andric wrote:
>>
>>> On 29 Oct 2023, at 11:34, Wojciech Puchar <wojtek@puchar.net> wrote:
>>>>
>>>> i get this linking my custom kernel. No errors until link
>>>>
>>>> linking kernel
>>>> ld: dynamic STT_GNU_IFUNC symbol `rdtsc_ordered' with pointer equality in `cpu_machdep.o' can not be used when making an executable; recompile with -fPIE and relink with -pie
>>>>
>>>>
>>>> kernel config attached, latest FreeBSD-13 downloaded today.
>>>>
>>>> can you help?<conf.txt>
>>>
>>> I just tried it here, and it works fine:
>>>
>>> ...
>>> linking kernel
>>> objcopy --strip-debug kernel
>>> text data bss dec hex filename
>>> 7925073 779012 3411328 12115413 0xb8ddd5 kernel
>>> --------------------------------------------------------------
>>>>>> Kernel build for PUCHAR completed on Sun Oct 29 12:06:08 UTC 2023
>>> --------------------------------------------------------------
>>>>>> Kernel(s) PUCHAR built in 28 seconds, ncpu: 32, make -j32
>>> --------------------------------------------------------------
>>>
>
> ===
> Mark Millard
> marklmi at yahoo.com
>
>
>
--3974694515-1688657752-1698606316=:1940--



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?8e5b7eea-cfb-1cbb-8eb-18ea18b230d4>