RFR: 8291649: multiple tests failing with -Xcomp after JDK-8290034 [v3]

[John R Rose]

On Fri, 26 Aug 2022 13:42:16 GMT, Jatin Bhateja <jbhateja at openjdk.org> wrote:

>> Hi,
>> 
>> This bug fix patch fixes the incorrect computation during constant folding of Reverse[IL] IR node.
>> A new utility class has been added to optimize computation of reverse_bit operation using GCC built-ins.
>> 
>> Kindly review and share your feedback.
>> 
>> Best Regards,
>> Jatin
>
> Jatin Bhateja has updated the pull request incrementally with one additional commit since the last revision:
> 
>   8291649: Incorporating code changes from Kim and John.

test/hotspot/gtest/opto/test_moveBits.cpp line 77:

> 75: }
> 76: 
> 77: // here is some object code to look at:

Here's a more useful comment:


// Here is some object code to look at if we want to do a manual
// study.  One could find the build file named test_moveBits.o.cmdline
// and hand-edit the command line to produce assembly code in
// test_moveBits.s.
//
// Or, given the two empty "fence functions", one could do a
// quick scan like this:
//
// $ objdump -D $(find build/*release -name test_moveBits.o) \
//   | sed -n /start_code_quality/,/end_code_quality/p \
//   | egrep -B10 bswap  # or grep -B20 cfi_endproc

void start_code_quality_moveBits() { }

int16_t code_quality_reverse_bytes_16(int16_t x) {
…

int64_t code_quality_reverse_bits_64(int64_t x) {
  return reverse_bits(x);  // should compile to andq/shrq with 64-bit masks
}

void end_code_quality_moveBits() { }

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

PR: https://git.openjdk.org/jdk/pull/9987