Update configure.ac and sync configure.

[pz9115]

Use autoupdate updates the configure.ac and sync the configure file with configure.ac with autoconf.

[cmuellner]

This is caused by a forgotten change in configure.ac.
Can you update configure.ac?

[pz9115]

Yes, the new configure file is just generated by $ autoconf configure.ac :)

[cmuellner]

I got that, but with this PR we end up with a wrong message in configure.
So, please fix configure.ac (remove the "only RV64 for musl libc" there).

[pz9115]

I see the in the configure.ac, we still set rv64imafdc as default input with --with-arch option, Do you think it would be more appropriate to replace it with gc?

And since now gcc supports vector extension, do we need to update the --enable-multilib option to support this commonly used extension?

[cmuellner]

I see the in the configure.ac, we still set rv64imafdc as default input with --with-arch option, Do you think it would be more appropriate to replace it with gc?

I would leave it as it is now.
For sure it is out of scope of this PR.

And since now gcc supports vector extension, do we need to update the --enable-multilib option to support this commonly used extension?

Yes, but that's also out of this PR's scope (it should be a different PR).

[TommyMurphyTM1234]

I see the in the configure.ac, we still set rv64imafdc as default input with --with-arch option, Do you think it would be more appropriate to replace it with gc?

Or would rv64gc_zicsr_zifencei be even more appropriate?
Or does rv64gc automatically/silently get expanded to rv64gc_zicsr_zifencei or something similar?

Edit: my comment crossed with @cmuellner's previous one which renders mine moot...

[pz9115]

I see the in the configure.ac, we still set rv64imafdc as default input with --with-arch option, Do you think it would be more appropriate to replace it with gc?

I would leave it as it is now. For sure it is out of scope of this PR.

And since now gcc supports vector extension, do we need to update the --enable-multilib option to support this commonly used extension?

Yes, but that's also out of this PR's scope (it should be a different PR).

Okay, I will update them in the new PR.

[cmuellner]

I see the in the configure.ac, we still set rv64imafdc as default input with --with-arch option, Do you think it would be more appropriate to replace it with gc?

Or would rv64gc_zicsr_zifencei be even more appropriate? Or does rv64gc automatically/silently get expanded to rv64gc_zicsr_zifencei or something similar?

I don't mind changing the default.
However, the change should have an observable benefit and not trigger complaints.
And of course such a change needs to come in its own PR.

[pz9115]

I see the in the configure.ac, we still set rv64imafdc as default input with --with-arch option, Do you think it would be more appropriate to replace it with gc?

Or would rv64gc_zicsr_zifencei be even more appropriate? Or does rv64gc automatically/silently get expanded to rv64gc_zicsr_zifencei or something similar?

Edit: my comment crossed with @cmuellner's previous one which renders mine moot...

@TommyMurphyTM1234 It's up to the ISA spec version, when set 2.2 or 20190608， gc equals to imafdc
when use 20191213 gc equals to imafdc_zicsr_zifencei. Currently gcc set 20191213 as the default ISA spec version.

[cmuellner]

This still needs the change in configure.ac (removal of the "only RV64 for musl libc").

[pz9115]

This still needs the change in configure.ac (removal of the "only RV64 for musl libc").

Understood! Update it with the new commit.