diff mbox series

[5.10+5.4,2/3] sched/topology: Fix sched_domain_topology_level alloc in sched_init_numa()

Message ID 20220305164430.245125-3-dann.frazier@canonical.com
State Superseded
Headers show
Series None | expand

Commit Message

dann frazier March 5, 2022, 4:44 p.m. UTC 
From: Dietmar Eggemann <dietmar.eggemann@arm.com>

commit 71e5f6644fb2f3304fcb310145ded234a37e7cc1 upstream.

Commit "sched/topology: Make sched_init_numa() use a set for the
deduplicating sort" allocates 'i + nr_levels (level)' instead of
'i + nr_levels + 1' sched_domain_topology_level.

This led to an Oops (on Arm64 juno with CONFIG_SCHED_DEBUG):

sched_init_domains
  build_sched_domains()
    __free_domain_allocs()
      __sdt_free() {
	...
        for_each_sd_topology(tl)
	  ...
          sd = *per_cpu_ptr(sdd->sd, j); <--
	  ...
      }

Signed-off-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Vincent Guittot <vincent.guittot@linaro.org>
Tested-by: Barry Song <song.bao.hua@hisilicon.com>
Link: https://lkml.kernel.org/r/6000e39e-7d28-c360-9cd6-8798fd22a9bf@arm.com
Signed-off-by: dann frazier <dann.frazier@canonical.com>
---
 kernel/sched/topology.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Comments

Joe Perches March 5, 2022, 5:13 p.m. UTC | #1 
On Sat, 2022-03-05 at 09:44 -0700, dann frazier wrote:
> From: Dietmar Eggemann <dietmar.eggemann@arm.com>
> 
> commit 71e5f6644fb2f3304fcb310145ded234a37e7cc1 upstream.
> 
> Commit "sched/topology: Make sched_init_numa() use a set for the
> deduplicating sort" allocates 'i + nr_levels (level)' instead of
> 'i + nr_levels + 1' sched_domain_topology_level.
[]
> diff --git a/kernel/sched/topology.c b/kernel/sched/topology.c
[]
> @@ -1655,7 +1655,7 @@ void sched_init_numa(void)
>  	/* Compute default topology size */
>  	for (i = 0; sched_domain_topology[i].mask; i++);

Thanks.

Couple trivial notes:

A trailing semicolon in a for loop, "for (...);" can be error prone
and is also the only usage of that style in kernel/ path.

A more common usage might be:

	i = 0;
	while (sched_domain_topology[i].mask)
		i++;

> -	tl = kzalloc((i + nr_levels) *
> +	tl = kzalloc((i + nr_levels + 1) *
>  			sizeof(struct sched_domain_topology_level), GFP_KERNEL);

kcalloc would be better, although the array is completely set
by the loop below so the zeroing isn't necessary.
Maybe use kmalloc_array.

Doubtful there's an overall impact though.
diff mbox series

Patch

diff --git a/kernel/sched/topology.c b/kernel/sched/topology.c
index f2951eba9592..ff2c6d3ba6c7 100644
--- a/kernel/sched/topology.c
+++ b/kernel/sched/topology.c
@@ -1655,7 +1655,7 @@  void sched_init_numa(void)
 	/* Compute default topology size */
 	for (i = 0; sched_domain_topology[i].mask; i++);
 
-	tl = kzalloc((i + nr_levels) *
+	tl = kzalloc((i + nr_levels + 1) *
 			sizeof(struct sched_domain_topology_level), GFP_KERNEL);
 	if (!tl)
 		return;