Re: [kvm-unit-tests RFC PATCH 2/6] configure: Make arch_libdir a first-class entity

[Thomas Huth <thuth@redhat.com>]

On 03/07/2026 08.46, Chinmay Rath wrote:
> 
> On 6/10/26 17:00, Thomas Huth wrote:
>> On 02/06/2026 08.48, Chinmay Rath wrote:
>>> From: Nicholas Piggin <npiggin@gmail.com>
>>>
>>> arch_libdir was brought in to improve the heuristic determination of
>>> the lib/ directory based on arch and testdir names, but it did not
>>> entirely clean that mess up.
>>>
>>> Remove the arch_libdir->arch->testdir heuristic and just require
>>> everybody sets arch_libdir correctly. Fail if the lib/arch or
>>> lib/arch/asm directories can not be found.
>>>
>>> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
>>> Signed-off-by: Chinmay Rath <rathc@linux.ibm.com>
>>> ---
>>>   Makefile  |  2 +-
>>>   configure | 20 ++++++++++++--------
>>>   2 files changed, 13 insertions(+), 9 deletions(-)
>>
>> FYI, this patch seems to trigger yet another issue with Clang (I used v21):
>>
>> clang -no-integrated-as -std=gnu99 -ffreestanding -O2 -msoft-float -mno- 
>> altivec  -I /root/kvm-unit-tests/lib -I /root/kvm-unit-tests/lib/libfdt -I 
>> lib -Wa,-mregnames -fverbose-asm -S -o lib/powerpc/asm-offsets.s lib/ 
>> powerpc/asm-offsets.c
>> lib/powerpc/asm-offsets.c:10:10: error: 'asm/smp.h' file not found with 
>> <angled> include; use
>>       "quotes" instead
>>    10 | #include <asm/smp.h>
>>       |          ^~~~~~~~~~~
>>       |          "asm/smp.h"
>> In file included from lib/powerpc/asm-offsets.c:10:
>> lib/powerpc/asm/smp.h:5:10: fatal error: 'asm/processor.h' file not found
>>     5 | #include <asm/processor.h>
>>       |          ^~~~~~~~~~~~~~~~~
>> 2 errors generated.
>>
>> Could you please have a look?
> 
> Hi Thomas, back from traveling and looked into this.
> I am using Clang v21 as well :
> # clang --version
> clang version 21.1.8 (Fedora 21.1.8-4.fc43)
> Target: ppc64le-redhat-linux-gnu
> Thread model: posix
> InstalledDir: /usr/bin
> Configuration file: /etc/clang/ppc64le-redhat-linux-gnu-clang.cfg
> 
> I see the above file, asm-offsets.s being generated properly without any 
> error during the make process (used the extra flags for make that I have 
> mentioned in my other reply).
> Snippet from make :
> cat powerpc/boot_rom.bin.tmp >> powerpc/boot_rom.bin
> rm -f powerpc/boot_rom.bin.tmp
> clang -no-integrated-as -Wa,-mpower10 -std=gnu99 -ffreestanding -O2 -msoft- 
> float -mno-altivec  -I /root/chins/git/kvm-unit-tests/lib -I /root/chins/ 
> git/kvm-unit-tests/lib/libfdt -I lib -Wa,-mregnames -fverbose-asm -S -o lib/ 
> powerpc/asm-offsets.s /root/chins/git/kvm-unit-tests/lib/powerpc/asm-offsets.c
> (set -e; echo "#ifndef __ASM_OFFSETS_H__"; echo "#define __ASM_OFFSETS_H__"; 
> echo "/*"; echo " * Generated file. DO NOT MODIFY."; echo " *"; echo " */"; 
> echo ""; sed -ne  's:^[[:space:]]*\.ascii[[:space:]]*"\(.*\)".*:\1:; /^->/ 
> {s:->#\(.*\):/* \1 */:; s:^->\([^ ]*\) [\$#]*\([-0-9]*\) \(.*\):#define \1 
> \2 /* \3 */:; s:^->\([^ ]*\) [\$#]*\([^ ]*\) \(.*\):#define \1 \2 /* \3 */:; 
> s:->::; p;}' lib/powerpc/asm-offsets.s; echo ""; echo "#endif" ) > lib/ 
> powerpc/asm-offsets.h
> cp -f lib/powerpc/asm-offsets.h lib/generated/
> clang -no-integrated-as -Wa,-mpower10 -std=gnu99 -ffreestanding -O2 -msoft- 
> float -mno-altivec  -I /root/chins/git/kvm-unit-tests/lib -I /root/chins/ 
> git/kvm-unit-tests/lib/libfdt -I lib -Wa,-mregnames -g -MMD -MP -MF 
> powerpc/.selftest.d -fno-strict-aliasing -fno-common -Wall -Wwrite-strings - 
> Wempty-body -Wuninitialized -Wignored-qualifiers -Wno-missing-braces -Werror 
> -fomit-frame-pointer  -fno-stack-protector   -Wno-frame-address  -fno-pic    
> -Wunused-but-set-parameter -Wno-override-init  -Wmissing-prototypes - 
> Wstrict-prototypes -mlittle-endian   -c -o powerpc/selftest.o /root/chins/ 
> git/kvm-unit-tests/powerpc/selftest.c
> 
> Infact, if I just execute that single command independently to build asm- 
> offsets.s, it works as well.

Ok, I finally got access to a POWER system again and tried to reproduce the 
issue, but I also cannot reproduce it anymore. No clue how I managed to get 
into that situation, but looks like it was rather due to a messed up build 
tree than due to your patch. So sorry for the trouble, and please go ahead 
and simply ignore my objection here.

  Thanks,
   Thomas