Re: git: 53a2e2635ab2 - main - zfs: merge openzfs/zfs@a9410ccbd

Hi

I have committed a fix one minute before Dag's email.
I don't understand why my make universe on a amd64 with a clean obj 
could pass without that fix.

On 20. 8. 2025 15:50, John Baldwin wrote:
> On 8/20/25 09:44, Dag-Erling Smørgrav wrote:
>> Martin Matuska <mm@FreeBSD.org> writes:
>>>      zfs: merge openzfs/zfs@a9410ccbd
>>
>> This doesn't build, please either fix asap or revert.
>>
>>      cc -target x86_64-unknown-freebsd15.0 
>> --sysroot=/usr/obj/usr/src/amd64.amd64/tmp 
>> -B/usr/obj/usr/src/amd64.amd64/tmp/usr/bin -O2 -pipe -fno-common 
>> -DIN_BASE -I/usr/src/sys/contrib/openzfs/include 
>> -I/usr/src/sys/contrib/openzfs/lib/libspl/include 
>> -I/usr/src/sys/contrib/openzfs/lib/libspl/include/os/freebsd 
>> -I/usr/src/sys/contrib/openzfs/lib/libzpool/include 
>> -I/usr/src/cddl/compat/opensolaris/include 
>> -I/usr/src/sys/contrib/openzfs/module/icp/include -include 
>> /usr/src/sys/contrib/openzfs/include/os/freebsd/spl/sys/ccompile.h 
>> -DHAVE_ISSETUGID -include /usr/src/sys/modules/zfs/zfs_config.h -g 
>> -DDEBUG=1 -Wno-format -DZFS_DEBUG=1 -DNEED_SOLARIS_BOOLEAN 
>> -DHAVE_STRLCAT -DHAVE_STRLCPY -fPIE -g -gz=zlib -std=iso9899:1999 
>> -Wno-format-zero-length -fstack-protector-strong -Wsystem-headers 
>> -Werror -Wall -Wno-format-y2k -Wno-uninitialized -Wno-pointer-sign 
>> -Wdate-time -Wno-unknown-pragmas -Wno-empty-body -Wno-string-plus-int 
>> -Wno-unused-const-variable -Wno-error=unused-but-set-parameter 
>> -Wno-error=cast-function-type-mismatch -Wno-tautological-compare 
>> -Wno-unused-value -Wno-parentheses-equality -Wno-unused-function 
>> -Wno-enum-conversion -Wno-unused-local-typedef 
>> -Wno-address-of-packed-member -Wno-switch -Wno-switch-enum 
>> -Wno-knr-promoted-parameter -Qunused-arguments  -Wl,-zrelro -pie    
>> -o ztest.full ztest.pieo   -lgeom  -lm  -lnvpair -lumem  -lzpool   
>> -lavl  -lzfs_core  -lspl  -lzutil  -lzfs -luutil  -licp -lpthread
>>      ld: error: undefined reference: gcm_ghash_vpclmulqdq_avx2
>>      >>> referenced by 
>> /usr/obj/usr/src/amd64.amd64/tmp/usr/lib/libicp.so (disallowed by 
>> --no-allow-shlib-undefined)
>>           ld: error: undefined reference: gcm_init_vpclmulqdq_avx2
>>      >>> referenced by 
>> /usr/obj/usr/src/amd64.amd64/tmp/usr/lib/libicp.so (disallowed by 
>> --no-allow-shlib-undefined)
>>           ld: error: undefined reference: aes_gcm_enc_update_vaes_avx2
>>      >>> referenced by 
>> /usr/obj/usr/src/amd64.amd64/tmp/usr/lib/libicp.so (disallowed by 
>> --no-allow-shlib-undefined)
>>           ld: error: undefined reference: aes_gcm_dec_update_vaes_avx2
>>      >>> referenced by 
>> /usr/obj/usr/src/amd64.amd64/tmp/usr/lib/libicp.so (disallowed by 
>> --no-allow-shlib-undefined)
>>      cc: error: linker command failed with exit code 1 (use -v to see 
>> invocation)
>>      *** Error code 1
>>           Stop.
>>      make[5]: stopped making "all" in /usr/src/cddl/usr.bin/ztest
>
> Probably just needs the new file added to SRC in our bespoke build glue?
>
> However, it does raise the question of how this was tested (during 
> code slush
> no less)?
>