cmd/link: don't disable memory profiling when pprof.WriteHeapProfile is used

We have an optimization that if the memory profile is not consumed
anywhere, we set the memory profiling rate to 0 to disable the
"background" low-rate profiling. We detect whether the memory
profile is used by checking whether the runtime.MemProfile function
is reachable at link time. Previously, all APIs that access the
memory profile go through runtime.MemProfile. But the code was
refactored in CL 572396, and now the legacy entry point
WriteHeapProfile uses pprof_memProfileInternal without going
through runtime.MemProfile. In fact, even with the recommended
runtime/pprof.Profile API (pprof.Lookup or pprof.Profiles),
runtime.MemProfile is only (happen to be) reachable through
countHeap.

Change the linker to check runtime.memProfileInternal instead,
which is on all code paths that retrieve the memory profile. Add
a test case for WriteHeapProfile, so we cover all entry points.

Fixes #68136.

Change-Id: I075c8d45c95c81825a1822f032e23107aea4303c
Reviewed-on: https://go-review.googlesource.com/c/go/+/596538
Reviewed-by: Than McIntosh <[email protected]>
LUCI-TryBot-Result: Go LUCI <[email protected]>

Author: cherrymui

src/cmd/link/internal/ld/ld_test.go

@@ -304,7 +304,34 @@ package main
 import "runtime"
 import "runtime/pprof"
 func main() {
-        _ = pprof.Profiles()
+	_ = pprof.Profiles()
+	println(runtime.MemProfileRate)
+}
+`,
+			"524288",
+		},
+		{
+			"with_memprofile_runtime_pprof_writeheap",
+			`
+package main
+import "io"
+import "runtime"
+import "runtime/pprof"
+func main() {
+	_ = pprof.WriteHeapProfile(io.Discard)
+	println(runtime.MemProfileRate)
+}
+`,
+			"524288",
+		},
+		{
+			"with_memprofile_runtime_pprof_lookupheap",
+			`
+package main
+import "runtime"
+import "runtime/pprof"
+func main() {
+	_ = pprof.Lookup("heap")
 	println(runtime.MemProfileRate)
 }
 `,

src/cmd/link/internal/ld/lib.go

@@ -892,9 +892,9 @@ func (ctxt *Link) linksetup() {
 		}
 
 		// Set runtime.disableMemoryProfiling bool if
-		// runtime.MemProfile is not retained in the binary after
+		// runtime.memProfileInternal is not retained in the binary after
 		// deadcode (and we're not dynamically linking).
-		memProfile := ctxt.loader.Lookup("runtime.MemProfile", abiInternalVer)
+		memProfile := ctxt.loader.Lookup("runtime.memProfileInternal", abiInternalVer)
 		if memProfile != 0 && !ctxt.loader.AttrReachable(memProfile) && !ctxt.DynlinkingGo() {
 			memProfSym := ctxt.loader.LookupOrCreateSym("runtime.disableMemoryProfiling", 0)
 			sb := ctxt.loader.MakeSymbolUpdater(memProfSym)

src/runtime/mprof.go

@@ -892,9 +892,10 @@ func (r *StackRecord) Stack() []uintptr {
 // at the beginning of main).
 var MemProfileRate int = 512 * 1024
 
-// disableMemoryProfiling is set by the linker if runtime.MemProfile
+// disableMemoryProfiling is set by the linker if memory profiling
 // is not used and the link type guarantees nobody else could use it
 // elsewhere.
+// We check if the runtime.memProfileInternal symbol is present.
 var disableMemoryProfiling bool
 
 // A MemProfileRecord describes the live objects allocated
@@ -955,6 +956,13 @@ func MemProfile(p []MemProfileRecord, inuseZero bool) (n int, ok bool) {
 // memProfileInternal returns the number of records n in the profile. If there
 // are less than size records, copyFn is invoked for each record, and ok returns
 // true.
+//
+// The linker set disableMemoryProfiling to true to disable memory profiling
+// if this function is not reachable. Mark it noinline to ensure the symbol exists.
+// (This function is big and normally not inlined anyway.)
+// See also disableMemoryProfiling above and cmd/link/internal/ld/lib.go:linksetup.
+//
+//go:noinline
 func memProfileInternal(size int, inuseZero bool, copyFn func(profilerecord.MemProfileRecord)) (n int, ok bool) {
 	cycle := mProfCycle.read()
 	// If we're between mProf_NextCycle and mProf_Flush, take care

src/runtime/pprof/pprof.go

@@ -586,7 +586,8 @@ func writeHeapInternal(w io.Writer, debug int, defaultSampleType string) error {
 		runtime.ReadMemStats(memStats)
 	}
 
-	// Find out how many records there are (MemProfile(nil, true)),
+	// Find out how many records there are (the call
+	// pprof_memProfileInternal(nil, true) below),
 	// allocate that many records, and get the data.
 	// There's a race—more records might be added between
 	// the two calls—so allocate a few extra records for safety