RFR: 8366777: Build fails unknown pseudo-op with old AS on linux-aarch64 [v6]

Wed Sep 10 10:03:02 UTC 2025

On Wed, 10 Sep 2025 02:20:03 GMT, SendaoYan <syan at openjdk.org> wrote:

>> make/autoconf/flags-cflags.m4 line 963:
>> 
>>> 961:       AC_MSG_RESULT([$sve_supported])
>>> 962:       $2SVE_CFLAGS=""
>>> 963:       if test "x$enable_aarch64_sve" = "xyes"; then
>> 
>> `UTIL_ARG_ENABLE` will return `true`, not `yes`. Have you even tested this code?
>
> I did the test before. I was also confused that the `enable_aarch64_sve` sometimes return `yes`/`no`, sometimes return `true`/`false`, it seems depends the code implementations.
> 
> The log snippet after run bash configure.
> 
> 
>> grep "enable_aarch64_sve=" build/linux-aarch64-server-release/configure.log -n
> 1117:+++ eval 'enable_aarch64_sve=$ac_optarg'
> 1118:++++ enable_aarch64_sve=yes
> 20015:+++ enable_aarch64_sve=true
> 
> 
> [configure.log](https://github.com/user-attachments/files/22246010/configure.log)

So if you used the AC_ version, it will set enable_aarch64_sve with yes/no, but then you switched to the UTIL_ version, but you kept the same name as argument. That just causes a mess. I guess the UTIL_ version, which calls the AC_function at its core, will override the value. But just don't use that name.

-------------

PR Review Comment: https://git.openjdk.org/jdk/pull/27073#discussion_r2336237203