runtime: ensure free and unscavenged spans may be backed by huge pages

This change adds a new sysHugePage function to provide the equivalent of
Linux's madvise(MADV_HUGEPAGE) support to the runtime. It then uses
sysHugePage to mark a newly-coalesced free span as backable by huge
pages to make the freeHugePages approximation a bit more accurate.

The problem being solved here is that if a large free span is composed
of many small spans which were coalesced together, then there's a chance
that they have had madvise(MADV_NOHUGEPAGE) called on them at some point,
which makes freeHugePages less accurate.

For #30333.

Change-Id: Idd4b02567619fc8d45647d9abd18da42f96f0522
Reviewed-on: https://go-review.googlesource.com/c/go/+/173338
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Austin Clements <austin@google.com>

src/runtime/mem_aix.go

@@ -35,6 +35,9 @@ func sysUnused(v unsafe.Pointer, n uintptr) {
 func sysUsed(v unsafe.Pointer, n uintptr) {
 }
 
+func sysHugePage(v unsafe.Pointer, n uintptr) {
+}
+
 // Don't split the stack as this function may be invoked without a valid G,
 // which prevents us from allocating more stack.
 //go:nosplit

src/runtime/mem_bsd.go

@@ -29,6 +29,9 @@ func sysUnused(v unsafe.Pointer, n uintptr) {
 func sysUsed(v unsafe.Pointer, n uintptr) {
 }
 
+func sysHugePage(v unsafe.Pointer, n uintptr) {
+}
+
 // Don't split the stack as this function may be invoked without a valid G,
 // which prevents us from allocating more stack.
 //go:nosplit

src/runtime/mem_darwin.go

@@ -33,6 +33,9 @@ func sysUsed(v unsafe.Pointer, n uintptr) {
 	madvise(v, n, _MADV_FREE_REUSE)
 }
 
+func sysHugePage(v unsafe.Pointer, n uintptr) {
+}
+
 // Don't split the stack as this function may be invoked without a valid G,
 // which prevents us from allocating more stack.
 //go:nosplit

src/runtime/mem_js.go

@@ -26,6 +26,9 @@ func sysUnused(v unsafe.Pointer, n uintptr) {
 func sysUsed(v unsafe.Pointer, n uintptr) {
 }
 
+func sysHugePage(v unsafe.Pointer, n uintptr) {
+}
+
 // Don't split the stack as this function may be invoked without a valid G,
 // which prevents us from allocating more stack.
 //go:nosplit

src/runtime/mem_linux.go

@@ -117,16 +117,19 @@ func sysUnused(v unsafe.Pointer, n uintptr) {
 }
 
 func sysUsed(v unsafe.Pointer, n uintptr) {
-	if physHugePageSize != 0 {
-		// Partially undo the NOHUGEPAGE marks from sysUnused
-		// for whole huge pages between v and v+n. This may
-		// leave huge pages off at the end points v and v+n
-		// even though allocations may cover these entire huge
-		// pages. We could detect this and undo NOHUGEPAGE on
-		// the end points as well, but it's probably not worth
-		// the cost because when neighboring allocations are
-		// freed sysUnused will just set NOHUGEPAGE again.
+	// Partially undo the NOHUGEPAGE marks from sysUnused
+	// for whole huge pages between v and v+n. This may
+	// leave huge pages off at the end points v and v+n
+	// even though allocations may cover these entire huge
+	// pages. We could detect this and undo NOHUGEPAGE on
+	// the end points as well, but it's probably not worth
+	// the cost because when neighboring allocations are
+	// freed sysUnused will just set NOHUGEPAGE again.
+	sysHugePage(v, n)
+}
 
+func sysHugePage(v unsafe.Pointer, n uintptr) {
+	if physHugePageSize != 0 {
 		// Round v up to a huge page boundary.
 		beg := (uintptr(v) + (physHugePageSize - 1)) &^ (physHugePageSize - 1)
 		// Round v+n down to a huge page boundary.

src/runtime/mem_plan9.go

@@ -173,6 +173,9 @@ func sysUnused(v unsafe.Pointer, n uintptr) {
 func sysUsed(v unsafe.Pointer, n uintptr) {
 }
 
+func sysHugePage(v unsafe.Pointer, n uintptr) {
+}
+
 func sysMap(v unsafe.Pointer, n uintptr, sysStat *uint64) {
 	// sysReserve has already allocated all heap memory,
 	// but has not adjusted stats.

src/runtime/mem_windows.go

@@ -81,6 +81,9 @@ func sysUsed(v unsafe.Pointer, n uintptr) {
 	}
 }
 
+func sysHugePage(v unsafe.Pointer, n uintptr) {
+}
+
 // Don't split the stack as this function may be invoked without a valid G,
 // which prevents us from allocating more stack.
 //go:nosplit

src/runtime/mheap.go

@@ -502,6 +502,8 @@ func (h *mheap) coalesce(s *mspan) {
 		h.free.insert(other)
 	}
 
+	hpBefore := s.hugePages()
+
 	// Coalesce with earlier, later spans.
 	if before := spanOf(s.base() - 1); before != nil && before.state == mSpanFree {
 		if s.scavenged == before.scavenged {
@@ -519,6 +521,18 @@ func (h *mheap) coalesce(s *mspan) {
 			realign(s, after, after)
 		}
 	}
+
+	if !s.scavenged && s.hugePages() > hpBefore {
+		// If s has grown such that it now may contain more huge pages than it
+		// did before, then mark the whole region as huge-page-backable.
+		//
+		// Otherwise, on systems where we break up huge pages (like Linux)
+		// s may not be backed by huge pages because it could be made up of
+		// pieces which are broken up in the underlying VMA. The primary issue
+		// with this is that it can lead to a poor estimate of the amount of
+		// free memory backed by huge pages for determining the scavenging rate.
+		sysHugePage(unsafe.Pointer(s.base()), s.npages*pageSize)
+	}
 }
 
 // hugePages returns the number of aligned physical huge pages in the memory