Optimize redundant vsetvl instructions

[DingliZhang]

According to the discussion[1], we may need to simplify the redundant opcode vsetvli. Almost every instruct in riscv_v.ad uses this opcode (vsetvli) at the beginning, and it does look like there is a need for optimization.

[1] openjdk#11344 (comment)

[DingliZhang]

In V8[1], each block will be used as a unit, and the same vset in the same block will not emit.

[1] https://chromium-review.googlesource.com/c/v8/v8/+/3917324

[DingliZhang]

@zhongjuzhe submitted VSETVL PASS support, using LCM for perfect adaptation:
https://gcc.gnu.org/git/?p=gcc.git;a=commit;h=3b16afeb3f6aacf64b9f9c50b7cb9805a9dfff63
https://gcc.gnu.org/git/?p=gcc.git;a=commit;h=9243c3d1b63b9092a82178392145f9e9d62423d9

A series of subsequent patches submitted by zhongjuzhe to fix RVV vsetvl optimization issues:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108185
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108270
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108271