From 53020fe81eecd0b7be295868ce5850ef8f41074e Mon Sep 17 00:00:00 2001
From: Peter Zijlstra <a.p.zijlstra@chello.nl>
Date: Wed, 13 May 2009 21:26:19 +0200
Subject: perf_counter: Fix perf_output_copy() WARN to account for overflow

The simple reservation test in perf_output_copy() failed to take
unsigned int overflow into account, fix this.

[ Impact: fix false positive warning with more than 4GB of profiling data ]

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 kernel/perf_counter.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/kernel/perf_counter.c b/kernel/perf_counter.c
index ff166c11b69..985be0b662a 100644
--- a/kernel/perf_counter.c
+++ b/kernel/perf_counter.c
@@ -1927,7 +1927,11 @@ static void perf_output_copy(struct perf_output_handle *handle,
 
 	handle->offset = offset;
 
-	WARN_ON_ONCE(handle->offset > handle->head);
+	/*
+	 * Check we didn't copy past our reservation window, taking the
+	 * possible unsigned int wrap into account.
+	 */
+	WARN_ON_ONCE(((int)(handle->head - handle->offset)) < 0);
 }
 
 #define perf_output_put(handle, x) \
-- 
cgit v1.2.3