[wasm] Deserialization: adjust batch size

Instead of processing batches with a fixed number of functions, process
batches with approximately the same number of bytes. This prevents
disproportionately large batches to block the pipeline.

R=ahaas@chromium.org

Bug: v8:11164
Change-Id: I7fe57abac13c5fb749a002e339c5a9b2dab607be
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2567699Reviewed-by: Andreas Haas <ahaas@chromium.org>
Commit-Queue: Thibaud Michaud <thibaudm@chromium.org>
Cr-Commit-Position: refs/heads/master@{#71530}

src/wasm/wasm-serialization.cc

@@ -608,16 +608,19 @@ bool NativeModuleDeserializer::Read(Reader* reader) {
 
   auto batch = std::make_unique<std::vector<DeserializationUnit>>();
   int num_batches = 0;
+  const byte* batch_start = reader->current_location();
   for (uint32_t i = first_wasm_fn; i < total_fns; ++i) {
     DeserializationUnit unit = ReadCodeAndAlloc(i, reader);
     if (unit.code) {
       batch->push_back(std::move(unit));
     }
-    constexpr int kBatchSize = 100;
-    if (batch->size() == kBatchSize) {
+    uint64_t batch_size_in_bytes = reader->current_location() - batch_start;
+    constexpr int kMinBatchSizeInBytes = 100000;
+    if (batch_size_in_bytes >= kMinBatchSizeInBytes) {
       reloc_queue.Add(std::move(batch));
       num_batches++;
       batch = std::make_unique<std::vector<DeserializationUnit>>();
+      batch_start = reader->current_location();
     }
   }