Merge bitcoin/bitcoin#27189: util: Use steady clock in SeedStrengthen, FindBestImplementation, FlushStateToDisk

fa1b4e5c3294fc9aec033892a4a4d7b5cfc015aa Use steady clock in FlushStateToDisk (MarcoFalke)
1111e2f8b43cd9ed62dcf6b571a224b84fc421fd Use steady clock in SeedStrengthen and FindBestImplementation (MarcoFalke)

Pull request description:

  There may be a theoretical deadlock for the duration of the offset when the system clock is adjusted into a past time while executing `SeedStrengthen`.

  Fix this by using steady clock.

  Do the same in `FindBestImplementation`, which shouldn't be affected, because it discards outlier measurements. However, doing the same there for consistency seems fine.

  Do the same in `FlushStateToDisk`, which should make the flushes more steady, if the system clock is adjusted by a large offset.

ACKs for top commit:
  john-moffett:
    ACK fa1b4e5c3294fc9aec033892a4a4d7b5cfc015aa
  willcl-ark:
    ACK fa1b4e5c3

Tree-SHA512: cc625e796b186accd53222bd64eb57d0512bc7e588312d254349b542bbc5e5daac348ff2b3b3f7dc5ae0bbbae2ec11fdbf3022cf2164211633765a4b0108e83e

1 files changed, 4 insertions, 4 deletions

diff --git a/src/node/minisketchwrapper.cpp b/src/node/minisketchwrapper.cpp
index 67e823cb68..96707f7a0a 100644
--- a/src/node/minisketchwrapper.cpp
+++ b/src/node/minisketchwrapper.cpp
@@ -23,17 +23,17 @@ static constexpr uint32_t BITS = 32;
 
 uint32_t FindBestImplementation()
 {
-    std::optional<std::pair<int64_t, uint32_t>> best;
+    std::optional<std::pair<SteadyClock::duration, uint32_t>> best;
 
     uint32_t max_impl = Minisketch::MaxImplementation();
     for (uint32_t impl = 0; impl <= max_impl; ++impl) {
-        std::vector<int64_t> benches;
+        std::vector<SteadyClock::duration> benches;
         uint64_t offset = 0;
         /* Run a little benchmark with capacity 32, adding 184 entries, and decoding 11 of them once. */
         for (int b = 0; b < 11; ++b) {
             if (!Minisketch::ImplementationSupported(BITS, impl)) break;
             Minisketch sketch(BITS, impl, 32);
-            auto start = GetTimeMicros();
+            auto start = SteadyClock::now();
             for (uint64_t e = 0; e < 100; ++e) {
                 sketch.Add(e*1337 + b*13337 + offset);
             }
@@ -41,7 +41,7 @@ uint32_t FindBestImplementation()
                 sketch.Add(e*1337 + b*13337 + offset);
             }
             offset += (*sketch.Decode(32))[0];
-            auto stop = GetTimeMicros();
+            auto stop = SteadyClock::now();
             benches.push_back(stop - start);
         }
         /* Remember which implementation has the best median benchmark time. */