public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
* [PATCH] POSIX timers RCU optimization fix for clock_gettime
@ 2007-02-04  4:15 Paul E. McKenney
  0 siblings, 0 replies; only message in thread
From: Paul E. McKenney @ 2007-02-04  4:15 UTC (permalink / raw)
  To: linux-kernel; +Cc: oleg, akpm, johnstul, mingo, tglx, zippel

Fixes an exit()/posix_cpu_clock_get() race spotted by Oleg Nesterov.
Solution is to move the p->signal check under the tasklist_lock.

Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
---

 posix-cpu-timers.c |    9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

diff -urpNa -X dontdiff linux-2.6.20-rc6-clockrcu/kernel/posix-cpu-timers.c linux-2.6.20-rc6-clockrcu-fixes/kernel/posix-cpu-timers.c
--- linux-2.6.20-rc6-clockrcu/kernel/posix-cpu-timers.c	2007-02-01 13:03:57.000000000 -0800
+++ linux-2.6.20-rc6-clockrcu-fixes/kernel/posix-cpu-timers.c	2007-02-03 16:53:41.000000000 -0800
@@ -312,10 +312,13 @@ int posix_cpu_clock_get(const clockid_t 
 					error = cpu_clock_sample(which_clock,
 								 p, &rtn);
 				}
-			} else if (p->tgid == pid && p->signal) {
+			} else {
 				read_lock(&tasklist_lock);
-				error = cpu_clock_sample_group(which_clock,
-							       p, &rtn);
+				if (p->tgid == pid && p->signal) {
+					error =
+					    cpu_clock_sample_group(which_clock,
+							           p, &rtn);
+				}
 				read_unlock(&tasklist_lock);
 			}
 		}

^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2007-02-04  4:15 UTC | newest]

Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2007-02-04  4:15 [PATCH] POSIX timers RCU optimization fix for clock_gettime Paul E. McKenney

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox