sched/numa: Document usages of mm->numa_scan_seq
authorJason Low <jason.low2@hp.com>
Fri, 1 May 2015 00:28:14 +0000 (17:28 -0700)
committerIngo Molnar <mingo@kernel.org>
Fri, 8 May 2015 10:13:13 +0000 (12:13 +0200)
The p->mm->numa_scan_seq is accessed using READ_ONCE/WRITE_ONCE
and modified without exclusive access. It is not clear why it is
accessed this way. This patch provides some documentation on that.

Suggested-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Jason Low <jason.low2@hp.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Rik van Riel <riel@redhat.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Aswin Chandramouleeswaran <aswin@hp.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Mike Galbraith <umgwanakikbuti@gmail.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Preeti U Murthy <preeti@linux.vnet.ibm.com>
Cc: Scott J Norton <scott.norton@hp.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Waiman Long <waiman.long@hp.com>
Link: http://lkml.kernel.org/r/1430440094.2475.61.camel@j-VirtualBox
Signed-off-by: Ingo Molnar <mingo@kernel.org>
kernel/sched/fair.c

index d6915a038d8a08780e1331e4b6386e39e19e52b4..f18ddb72fe8863f5c24020f6d8c1d83fdfdf3a6b 100644 (file)
@@ -1794,6 +1794,11 @@ static void task_numa_placement(struct task_struct *p)
        u64 runtime, period;
        spinlock_t *group_lock = NULL;
 
+       /*
+        * The p->mm->numa_scan_seq field gets updated without
+        * exclusive access. Use READ_ONCE() here to ensure
+        * that the field is read in a single access:
+        */
        seq = READ_ONCE(p->mm->numa_scan_seq);
        if (p->numa_scan_seq == seq)
                return;
@@ -2107,6 +2112,14 @@ void task_numa_fault(int last_cpupid, int mem_node, int pages, int flags)
 
 static void reset_ptenuma_scan(struct task_struct *p)
 {
+       /*
+        * We only did a read acquisition of the mmap sem, so
+        * p->mm->numa_scan_seq is written to without exclusive access
+        * and the update is not guaranteed to be atomic. That's not
+        * much of an issue though, since this is just used for
+        * statistical sampling. Use READ_ONCE/WRITE_ONCE, which are not
+        * expensive, to avoid any form of compiler optimizations:
+        */
        WRITE_ONCE(p->mm->numa_scan_seq, READ_ONCE(p->mm->numa_scan_seq) + 1);
        p->mm->numa_scan_offset = 0;
 }