* [PATCH] sched/topology: fix potential memoryleak in sched_init_numa()
@ 2023-06-21 6:38 Miaohe Lin
2023-06-21 8:19 ` Abel Wu
0 siblings, 1 reply; 2+ messages in thread
From: Miaohe Lin @ 2023-06-21 6:38 UTC (permalink / raw
To: mingo, peterz, juri.lelli, vincent.guittot
Cc: dietmar.eggemann, rostedt, bsegall, mgorman, bristot, vschneid,
linux-kernel, linmiaohe
When sched_init_numa() fails to allocate enough memory for sched domains
numa masks, it forgot to free the allocated memory leading to memoryleak.
Add a helper to help release the resource.
Fixes: cb83b629bae0 ("sched/numa: Rewrite the CONFIG_NUMA sched domain support")
Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
---
kernel/sched/topology.c | 22 ++++++++++++++++++++--
1 file changed, 20 insertions(+), 2 deletions(-)
diff --git a/kernel/sched/topology.c b/kernel/sched/topology.c
index 290509383419..dcec4d653ae3 100644
--- a/kernel/sched/topology.c
+++ b/kernel/sched/topology.c
@@ -1807,6 +1807,20 @@ static void init_numa_topology_type(int offline_node)
#define NR_DISTANCE_VALUES (1 << DISTANCE_BITS)
+static void sched_free_numa_mask(struct cpumask ***masks, int nr_levels)
+{
+ int i, j;
+
+ for (i = 0; i < nr_levels; i++) {
+ if (!masks[i])
+ continue;
+ for_each_node(j)
+ kfree(masks[i][j]);
+ kfree(masks[i]);
+ }
+ kfree(masks);
+}
+
void sched_init_numa(int offline_node)
{
struct sched_domain_topology_level *tl;
@@ -1886,15 +1900,19 @@ void sched_init_numa(int offline_node)
*/
for (i = 0; i < nr_levels; i++) {
masks[i] = kzalloc(nr_node_ids * sizeof(void *), GFP_KERNEL);
- if (!masks[i])
+ if (!masks[i]) {
+ sched_free_numa_mask(masks, nr_levels);
return;
+ }
for_each_cpu_node_but(j, offline_node) {
struct cpumask *mask = kzalloc(cpumask_size(), GFP_KERNEL);
int k;
- if (!mask)
+ if (!mask) {
+ sched_free_numa_mask(masks, nr_levels);
return;
+ }
masks[i][j] = mask;
--
2.27.0
^ permalink raw reply related [flat|nested] 2+ messages in thread
* Re: [PATCH] sched/topology: fix potential memoryleak in sched_init_numa()
2023-06-21 6:38 [PATCH] sched/topology: fix potential memoryleak in sched_init_numa() Miaohe Lin
@ 2023-06-21 8:19 ` Abel Wu
0 siblings, 0 replies; 2+ messages in thread
From: Abel Wu @ 2023-06-21 8:19 UTC (permalink / raw
To: Miaohe Lin, mingo, peterz, juri.lelli, vincent.guittot
Cc: dietmar.eggemann, rostedt, bsegall, mgorman, bristot, vschneid,
linux-kernel
On 6/21/23 2:38 PM, Miaohe Lin wrote:
> When sched_init_numa() fails to allocate enough memory for sched domains
> numa masks, it forgot to free the allocated memory leading to memoryleak.
> Add a helper to help release the resource.
>
> Fixes: cb83b629bae0 ("sched/numa: Rewrite the CONFIG_NUMA sched domain support")
> Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
> ---
> kernel/sched/topology.c | 22 ++++++++++++++++++++--
> 1 file changed, 20 insertions(+), 2 deletions(-)
>
> diff --git a/kernel/sched/topology.c b/kernel/sched/topology.c
> index 290509383419..dcec4d653ae3 100644
> --- a/kernel/sched/topology.c
> +++ b/kernel/sched/topology.c
> @@ -1807,6 +1807,20 @@ static void init_numa_topology_type(int offline_node)
>
> #define NR_DISTANCE_VALUES (1 << DISTANCE_BITS)
>
> +static void sched_free_numa_mask(struct cpumask ***masks, int nr_levels)
> +{
> + int i, j;
> +
> + for (i = 0; i < nr_levels; i++) {
> + if (!masks[i])
> + continue;
> + for_each_node(j)
> + kfree(masks[i][j]);
> + kfree(masks[i]);
> + }
> + kfree(masks);
> +}
> +
> void sched_init_numa(int offline_node)
> {
> struct sched_domain_topology_level *tl;
> @@ -1886,15 +1900,19 @@ void sched_init_numa(int offline_node)
> */
> for (i = 0; i < nr_levels; i++) {
> masks[i] = kzalloc(nr_node_ids * sizeof(void *), GFP_KERNEL);
> - if (!masks[i])
> + if (!masks[i]) {
> + sched_free_numa_mask(masks, nr_levels);
> return;
> + }
>
> for_each_cpu_node_but(j, offline_node) {
> struct cpumask *mask = kzalloc(cpumask_size(), GFP_KERNEL);
> int k;
>
> - if (!mask)
> + if (!mask) {
> + sched_free_numa_mask(masks, nr_levels);
> return;
> + }
>
> masks[i][j] = mask;
>
Allocation can also fail in @tl (topology level), and if that is the
case, masks[][] IMHO also needs be freed. So I think it might be better
if call sched_reset_numa() at proper place.
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2023-06-21 8:20 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2023-06-21 6:38 [PATCH] sched/topology: fix potential memoryleak in sched_init_numa() Miaohe Lin
2023-06-21 8:19 ` Abel Wu
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).