runtime: acquire stack lock in traceEvent

traceEvent records system call events after a G has already entered _Gsyscall, which means the garbage collector could be installing stack barriers in the G's stack during the traceEvent. If traceEvent attempts to capture the user stack during this, it may observe a inconsistent stack barriers and panic. Fix this by acquiring the stack lock around the stack walk in traceEvent. Fixes golang#14101. Change-Id: I15f0ab0c70c04c6e182221f65a6f761c5a896459 Reviewed-on: https://go-review.googlesource.com/18973 Run-TryBot: Austin Clements <[email protected]> TryBot-Result: Gobot Gobot <[email protected]> Reviewed-by: Russ Cox <[email protected]>
ConradIrwin · Jan 27, 2016 · 08594ac · 08594ac
1 parent eb3b183
commit 08594ac
Showing 1 changed file with 6 additions and 1 deletion.
diff --git a/src/runtime/trace.go b/src/runtime/trace.go
@@ -529,7 +529,12 @@ func traceEvent(ev byte, skip int, args ...uint64) {
 			nstk = callers(skip, buf.stk[:])
 		} else if gp != nil {
 			gp = mp.curg
-			nstk = gcallers(gp, skip, buf.stk[:])
+			// This may happen when tracing a system call,
+			// so we must lock the stack.
+			if gcTryLockStackBarriers(gp) {
+				nstk = gcallers(gp, skip, buf.stk[:])
+				gcUnlockStackBarriers(gp)
+			}
 		}
 		if nstk > 0 {
 			nstk-- // skip runtime.goexit