在 2024/12/17 下午12:30, Xi Ruoyao 写道:
On Tue, 2024-12-17 at 11:27 +0800, Lulu Cheng wrote:
在 2024/12/16 下午9:20, Xi Ruoyao 写道:
/* snip */
+;; For HImode it's a little complicated...
+(define_expand "rbithi"
I didn't find rtithi's template description. Are there any test cases
?
No, it's not a standard name. I just used the same name as AArch64.
And due to PR50481 is unimplemented yet, there's no way to directly test
it :(.
Bitreverse is a standard operation. if "define_expand "rbithi""
implementated using
bitreverse, will it trigger optimization in some cases?
If so, can we add an operation here?
The functionality can be tested with the existing test case,
gcc/testsuite/gcc.dg/torture/crc-coremark32-data16.c, with the next
patch (adding crc_revsi<mode>4) applied. The generated code uses the
bitrev instruction to reverse bits:
bitrev.d $r13,$r4
bitrev.w $r5,$r5
srli.d $r13,$r13,48
I.e. r5 will contain the 32-bit reversal of the original value, and r13
will contain the 16-bit reversal of r4.
This is a "dg-do run" test so the correctness is already verified at
runtime (I've ran the regression test with the entire series applied).
