[M680x0] Add CycleTimer implementation for M68K architecture

[myhsu]

Summary

This is a cherrypick of the upstream fix commit ffe1342 onto
llvm/utils/benchmakr and libcxx/utils/google-benchmark.
This adds CycleTimer implementation for M680x0, and simply
uses gettimeofday to implement, same as MIPS.

[lebedev.ri]

Thanks, LG, please feel free to commit this.
There's also one in test-suite, you might want to patch that too.

[myhsu]

In D88868#2313527, @lebedev.ri wrote:

Thanks, LG, please feel free to commit this.
There's also one in test-suite, you might want to patch that too.

Thanks for the swift review. I've created the test-suite part review https://reviews.llvm.org/D89011

[RKSimon]

@ldionne @mclow.lists @EricWF - can someone from the libc++ group accept this to unblock it please?

[lebedev.ri]

I don't believe this needs libc++'s signoff, it only mechanically replicates
the exact same change to the same external dependency into all the duplicate copies of said dependency.
@myhsu i still maintain that you can just commit this now.

[ldionne]

Indeed, this is fine without libc++'s signoff. Libc++ is added automatically because it touches files in libcxx/. Thanks for asking, though.

[myhsu]

In D88868#2442287, @lebedev.ri wrote:

I don't believe this needs libc++'s signoff, it only mechanically replicates
the exact same change to the same external dependency into all the duplicate copies of said dependency.
@myhsu i still maintain that you can just commit this now.

As @ldionne mentioned libcxx was added automatically and I couldn't remove it. Otherwise I've already committed this patch a while ago: https://github.com/llvm/llvm-test-suite/commit/25883edc2f615a84c4fa741e997904293474bbc0

[RKSimon]

In D88868#2443692, @myhsu wrote:

In D88868#2442287, @lebedev.ri wrote:

I don't believe this needs libc++'s signoff, it only mechanically replicates
the exact same change to the same external dependency into all the duplicate copies of said dependency.
@myhsu i still maintain that you can just commit this now.

As @ldionne mentioned libcxx was added automatically and I couldn't remove it. Otherwise I've already committed this patch a while ago: https://github.com/llvm/llvm-test-suite/commit/25883edc2f615a84c4fa741e997904293474bbc0

Isn't that just testsuite's copy of the code from D89011? It looks like at least the libcxx parts of this patch still need committing.

[myhsu]

In D88868#2445067, @RKSimon wrote:

In D88868#2443692, @myhsu wrote:

In D88868#2442287, @lebedev.ri wrote:

I don't believe this needs libc++'s signoff, it only mechanically replicates
the exact same change to the same external dependency into all the duplicate copies of said dependency.
@myhsu i still maintain that you can just commit this now.

As @ldionne mentioned libcxx was added automatically and I couldn't remove it. Otherwise I've already committed this patch a while ago: https://github.com/llvm/llvm-test-suite/commit/25883edc2f615a84c4fa741e997904293474bbc0

Isn't that just testsuite's copy of the code from D89011? It looks like at least the libcxx parts of this patch still need committing.

You’re right, this a a different one. I’ll commit this shortly

[myhsu]

It turns out I've already committed this at the same time with changes in llvm-test (after @lebedev.ri approved this) and I totally forgot about it :-P
https://github.com/llvm/llvm-project/commit/bd5fe7b010ea0d6c4b1e5e3740085f5eabf7def6

libcxx/utils/google-benchmark/AUTHORS

 Radoslav Yovchev <[e-mail]>
 Roman Lebedev <[e-mail]>
 Shuo Chen <[e-mail]>
 Steinar H. Gunderson <[e-mail]>
 Stripe, Inc.
 Yixuan Qiu <[e-mail]>
 Yusuke Suzuki <[e-mail]>
 Zbigniew Skowron <[e-mail]>
+Min-Yih Hsu <[e-mail]>

libcxx/utils/google-benchmark/CONTRIBUTORS

 Robert Guo <[e-mail]>
 Roman Lebedev <[e-mail]>
 Shuo Chen <[e-mail]>
 Tobias Ulvgård <[e-mail]>
 Tom Madams <[e-mail]> <[e-mail]>
 Yixuan Qiu <[e-mail]>
 Yusuke Suzuki <[e-mail]>
 Zbigniew Skowron <[e-mail]>
+Min-Yih Hsu <[e-mail]>

libcxx/utils/google-benchmark/README.LLVM

 * https://github.com/google/benchmark/commit/a77d5f70efaebe2b7e8c10134526a23a7ce7ef35
   and
   https://github.com/google/benchmark/commit/ecc1685340f58f7fe6b707036bc0bb1fccabb0c1
   are applied on top of
   https://github.com/google/benchmark/commit/8e48105d465c586068dd8e248fe75a8971c6ba3a
   to fix timestamp-related inline asm issues and 32-bit RISC-V build failures.
   The second cherrypicked commit fixes formatting issues introduced by the
   preceding change.
+* https://github.com/google/benchmark/commit/ffe1342eb2faa7d2e7c35b4db2ccf99fab81ec20
+  is applied to add the CycleTimer implementation for M680x0

libcxx/utils/google-benchmark/src/cycleclock.h

 // data might be inaccurate due to out of order instruction execution.
 // ----------------------------------------------------------------------
 
 #ifndef BENCHMARK_CYCLECLOCK_H_
 #define BENCHMARK_CYCLECLOCK_H_
 
 #include <cstdint>
 
 #include "benchmark/benchmark.h"
 #include "internal_macros.h"
 
 #if defined(BENCHMARK_OS_MACOSX)
 #include <mach/mach_time.h>
 #endif
 // For MSVC, we want to use '_asm rdtsc' when possible (since it works
 // with even ancient MSVC compilers), and when not possible the
 // __rdtsc intrinsic, declared in <intrin.h>.  Unfortunately, in some
if (pmcntenset & 0x80000000ul) {  // Is it counting?
       // The counter is set up to count every 64th cycle
       return static_cast<int64_t>(pmccntr) * 64;  // Should optimize to << 6
     }
   }
 #endif
   struct timeval tv;
   gettimeofday(&tv, nullptr);
   return static_cast<int64_t>(tv.tv_sec) * 1000000 + tv.tv_usec;
-#elif defined(__mips__)
+#elif defined(__mips__) || defined(__m68k__)
   // mips apparently only allows rdtsc for superusers, so we fall
   // back to gettimeofday.  It's possible clock_gettime would be better.
   struct timeval tv;
   gettimeofday(&tv, nullptr);
   return static_cast<int64_t>(tv.tv_sec) * 1000000 + tv.tv_usec;
 #elif defined(__s390__)  // Covers both s390 and s390x.
   // Return the CPU clock.
   uint64_t tsc;

llvm/utils/benchmark/AUTHORS

 Radoslav Yovchev <[e-mail]>
 Roman Lebedev <[e-mail]>
 Shuo Chen <[e-mail]>
 Steinar H. Gunderson <[e-mail]>
 Stripe, Inc.
 Yixuan Qiu <[e-mail]>
 Yusuke Suzuki <[e-mail]>
 Zbigniew Skowron <[e-mail]>
+Min-Yih Hsu <[e-mail]>

llvm/utils/benchmark/CONTRIBUTORS

 Robert Guo <[e-mail]>
 Roman Lebedev <[e-mail]>
 Shuo Chen <[e-mail]>
 Tobias Ulvgård <[e-mail]>
 Tom Madams <[e-mail]> <[e-mail]>
 Yixuan Qiu <[e-mail]>
 Yusuke Suzuki <[e-mail]>
 Zbigniew Skowron <[e-mail]>
+Min-Yih Hsu <[e-mail]>

llvm/utils/benchmark/README.LLVM

 * https://github.com/google/benchmark/commit/8e48105d465c586068dd8e248fe75a8971c6ba3a
   is applied on top of v1.4.1 to fix cross-build from linux to windows via MinGW.
 * https://github.com/google/benchmark/commit/a77d5f70efaebe2b7e8c10134526a23a7ce7ef35
   and
   https://github.com/google/benchmark/commit/ecc1685340f58f7fe6b707036bc0bb1fccabb0c1
   are applied on top of the previous cherrypick to fix timestamp-related inline
   asm issues and 32-bit RISC-V build failures. The second cherrypicked commit
   fixes formatting issues introduced by the preceding change.
+* https://github.com/google/benchmark/commit/ffe1342eb2faa7d2e7c35b4db2ccf99fab81ec20
+  is applied to add the CycleTimer implementation for M680x0

llvm/utils/benchmark/src/cycleclock.h

if (pmcntenset & 0x80000000ul) {  // Is it counting?
       // The counter is set up to count every 64th cycle
       return static_cast<int64_t>(pmccntr) * 64;  // Should optimize to << 6
     }
   }
 #endif
   struct timeval tv;
   gettimeofday(&tv, nullptr);
   return static_cast<int64_t>(tv.tv_sec) * 1000000 + tv.tv_usec;
-#elif defined(__mips__)
+#elif defined(__mips__) || defined(__m68k__)
   // mips apparently only allows rdtsc for superusers, so we fall
   // back to gettimeofday.  It's possible clock_gettime would be better.
   struct timeval tv;
   gettimeofday(&tv, nullptr);
   return static_cast<int64_t>(tv.tv_sec) * 1000000 + tv.tv_usec;
 #elif defined(__s390__) // Covers both s390 and s390x.
   // Return the CPU clock.
   uint64_t tsc;