[lld] [llvm] [lld][MachO] Follow-up to use madvise() for threaded file page-in. (PR #157917)
Ellis Hoag via llvm-commits
llvm-commits at lists.llvm.org
Thu Sep 11 08:41:58 PDT 2025
================
@@ -334,29 +339,35 @@ class SerialBackgroundQueue {
// This code forces the page-ins on multiple threads so
// the process is not stalled waiting on disk buffer i/o.
void multiThreadedPageInBackground(DeferredFiles &deferred) {
+ using namespace std::chrono;
static const size_t pageSize = Process::getPageSizeEstimate();
static const size_t largeArchive = 10 * 1024 * 1024;
-#ifndef NDEBUG
- using namespace std::chrono;
- std::atomic_int numDeferedFilesTouched = 0;
static std::atomic_uint64_t totalBytes = 0;
+ std::atomic_int numDeferedFilesAdvised = 0;
auto t0 = high_resolution_clock::now();
-#endif
auto preloadDeferredFile = [&](const DeferredFile &deferredFile) {
const StringRef &buff = deferredFile.buffer.getBuffer();
if (buff.size() > largeArchive)
return;
-#ifndef NDEBUG
+
totalBytes += buff.size();
- numDeferedFilesTouched += 1;
-#endif
+ numDeferedFilesAdvised += 1;
+#if _WIN32
// Reference all file's mmap'd pages to load them into memory.
for (const char *page = buff.data(), *end = page + buff.size(); page < end;
page += pageSize)
LLVM_ATTRIBUTE_UNUSED volatile char t = *page;
+#else
+#define DEBUG_TYPE "lld-madvise"
+ auto aligned = ((uintptr_t)buff.data() + (pageSize - 1)) & ~(pageSize - 1);
----------------
ellishg wrote:
I think you can use llvm:alignAddr()
https://github.com/llvm/llvm-project/pull/157917
More information about the llvm-commits
mailing list