Re: [PATCH] sched: Do not re-read h_load_next during hierarchical load calculation

[Peter Zijlstra <peterz@infradead.org>]

On Tue, Mar 19, 2019 at 09:35:18AM +0000, Mel Gorman wrote:
> A NULL pointer dereference bug was reported on a distribution kernel but
> the same issue should be present on mainline kernel. It occured on s390
> but should not be arch-specific.  A partial oops looks like
> 
> [775277.408564] Unable to handle kernel pointer dereference in virtual kernel address space
> ...
> [775277.408759] Call Trace:
> [775277.408763] ([<0002c11c56899c61>] 0x2c11c56899c61)
> [775277.408766]  [<0000000000177bb4>] try_to_wake_up+0xfc/0x450
> [775277.408773]  [<000003ff81ede872>] vhost_poll_wakeup+0x3a/0x50 [vhost]
> [775277.408777]  [<0000000000194ae4>] __wake_up_common+0xbc/0x178
> [775277.408779]  [<0000000000194f86>] __wake_up_common_lock+0x9e/0x160
> [775277.408780]  [<00000000001950de>] __wake_up_sync_key+0x4e/0x60
> [775277.408785]  [<00000000005d911e>] sock_def_readable+0x5e/0x98
> 
> The bug hits any time between 1 hour to 3 days. The dereference occurs
> in update_cfs_rq_h_load when accumulating h_load. The problem is that
> cfq_rq->h_load_next is not protected by any locking and can be updated
> by parallel calls to task_h_load.

Hurpmh, right.

> Depending on the compiler, code may be
> generated that re-reads cfq_rq->h_load_next after the check for NULL and
> then oops when reading se->avg.load_avg. The dissassembly showed that it
> was possible to reread h_load_next after the check for NULL.
> 
> While this does not appear to be an issue for later compilers, it's still
> an accident if the correct code is generated. Full locking in this path
> would have high overhead so this patch uses READ_ONCE to read h_load_next
> only once and check for NULL before dereferencing. It was confirmed that
> there were no further oops after 10 days of testing.
> 
> Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
> ---
>  kernel/sched/fair.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> index 310d0637fe4b..34aeb40e69d2 100644
> --- a/kernel/sched/fair.c
> +++ b/kernel/sched/fair.c
> @@ -7726,7 +7726,7 @@ static void update_cfs_rq_h_load(struct cfs_rq *cfs_rq)
>  		cfs_rq->last_h_load_update = now;
>  	}
>  
> -	while ((se = cfs_rq->h_load_next) != NULL) {
> +	while ((se = READ_ONCE(cfs_rq->h_load_next)) != NULL) {
>  		load = cfs_rq->h_load;
>  		load = div64_ul(load * se->avg.load_avg,
>  			cfs_rq_load_avg(cfs_rq) + 1);

Where there is a READ_ONCE there should also be a corresponding
WRITE_ONCE(). Otherwise the compiler can still screw us over by doing
store-tearing.

So something like the below. But looking at this, we probably also want
ONCE treatment on cfs_rq->h_load itself, but that's another patch.

And I think we can do something with cfs_rq->last_h_load_update.

---
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index fdab7eb6f351..40bd1e27b1b7 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -7784,10 +7784,10 @@ static void update_cfs_rq_h_load(struct cfs_rq *cfs_rq)
 	if (cfs_rq->last_h_load_update == now)
 		return;
 
-	cfs_rq->h_load_next = NULL;
+	WRITE_ONCE(cfs_rq->h_load_next, NULL);
 	for_each_sched_entity(se) {
 		cfs_rq = cfs_rq_of(se);
-		cfs_rq->h_load_next = se;
+		WRITE_ONCE(cfs_rq->h_load_next, se);
 		if (cfs_rq->last_h_load_update == now)
 			break;
 	}
@@ -7797,7 +7797,7 @@ static void update_cfs_rq_h_load(struct cfs_rq *cfs_rq)
 		cfs_rq->last_h_load_update = now;
 	}
 
-	while ((se = cfs_rq->h_load_next) != NULL) {
+	while ((se = READ_ONCE(cfs_rq->h_load_next)) != NULL) {
 		load = cfs_rq->h_load;
 		load = div64_ul(load * se->avg.load_avg,
 			cfs_rq_load_avg(cfs_rq) + 1);