[PATCH] blkcg: check pol->cpd_free_fn before free cpd

[PATCH] blkcg: check pol->cpd_free_fn before free cpd

[weiping zhang]

check pol->cpd_free_fn() instead of pol->cpd_alloc_fn() when free cpd.

Signed-off-by: weiping zhang <zhangweiping@didichuxing.com>
---
 block/blk-cgroup.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
index 0480892..adcbc3e 100644
--- a/block/blk-cgroup.c
+++ b/block/blk-cgroup.c
@@ -1044,7 +1044,7 @@ static void blkcg_css_free(struct cgroup_subsys_state *css)
 	list_del(&blkcg->all_blkcgs_node);
 
 	for (i = 0; i < BLKCG_MAX_POLS; i++)
-		if (blkcg->cpd[i])
+		if (blkcg->cpd[i] && blkcg_policy[i]->cpd_free_fn)
 			blkcg_policy[i]->cpd_free_fn(blkcg->cpd[i]);
 
 	mutex_unlock(&blkcg_pol_mutex);
@@ -1109,7 +1109,7 @@ blkcg_css_alloc(struct cgroup_subsys_state *parent_css)
 
 free_pd_blkcg:
 	for (i--; i >= 0; i--)
-		if (blkcg->cpd[i])
+		if (blkcg->cpd[i] && blkcg_policy[i]->cpd_free_fn)
 			blkcg_policy[i]->cpd_free_fn(blkcg->cpd[i]);
 free_blkcg:
 	kfree(blkcg);
@@ -1450,7 +1450,7 @@ int blkcg_policy_register(struct blkcg_policy *pol)
 	return 0;
 
 err_free_cpds:
-	if (pol->cpd_alloc_fn) {
+	if (pol->cpd_free_fn) {
 		list_for_each_entry(blkcg, &all_blkcgs, all_blkcgs_node) {
 			if (blkcg->cpd[pol->plid]) {
 				pol->cpd_free_fn(blkcg->cpd[pol->plid]);
@@ -1490,7 +1490,7 @@ void blkcg_policy_unregister(struct blkcg_policy *pol)
 	/* remove cpds and unregister */
 	mutex_lock(&blkcg_pol_mutex);
 
-	if (pol->cpd_alloc_fn) {
+	if (pol->cpd_free_fn) {
 		list_for_each_entry(blkcg, &all_blkcgs, all_blkcgs_node) {
 			if (blkcg->cpd[pol->plid]) {
 				pol->cpd_free_fn(blkcg->cpd[pol->plid]);
-- 
2.9.4

[PATCH] blkcg: check pol->cpd_free_fn before free cpd

[weiping zhang]

check pol->cpd_free_fn() instead of pol->cpd_alloc_fn() when free cpd.

Signed-off-by: weiping zhang <zhangweiping@didichuxing.com>
---
 block/blk-cgroup.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
index d3f56ba..e7ec676 100644
--- a/block/blk-cgroup.c
+++ b/block/blk-cgroup.c
@@ -1452,7 +1452,7 @@ int blkcg_policy_register(struct blkcg_policy *pol)
 	return 0;
 
 err_free_cpds:
-	if (pol->cpd_alloc_fn) {
+	if (pol->cpd_free_fn) {
 		list_for_each_entry(blkcg, &all_blkcgs, all_blkcgs_node) {
 			if (blkcg->cpd[pol->plid]) {
 				pol->cpd_free_fn(blkcg->cpd[pol->plid]);
@@ -1492,7 +1492,7 @@ void blkcg_policy_unregister(struct blkcg_policy *pol)
 	/* remove cpds and unregister */
 	mutex_lock(&blkcg_pol_mutex);
 
-	if (pol->cpd_alloc_fn) {
+	if (pol->cpd_free_fn) {
 		list_for_each_entry(blkcg, &all_blkcgs, all_blkcgs_node) {
 			if (blkcg->cpd[pol->plid]) {
 				pol->cpd_free_fn(blkcg->cpd[pol->plid]);
-- 
2.9.4

Re: [PATCH] blkcg: check pol->cpd_free_fn before free cpd

[Johannes Thumshirn]

Good catch,
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
-- 
Johannes Thumshirn                                          Storage
jthumshirn@suse.de                                +49 911 74053 689
SUSE LINUX GmbH, Maxfeldstr. 5, 90409 N�rnberg
GF: Felix Imend�rffer, Jane Smithard, Graham Norton
HRB 21284 (AG N�rnberg)
Key fingerprint = EC38 9CAB C2C4 F25D 8600 D0D0 0393 969D 2D76 0850

Re: [PATCH] blkcg: check pol->cpd_free_fn before free cpd

[Jens Axboe]

On 10/10/2017 08:53 AM, weiping zhang wrote:
> check pol->cpd_free_fn() instead of pol->cpd_alloc_fn() when free cpd.

In practice this shouldn't make a difference, since if you have an
alloc_fn, you better also have a free_fn. I'd argue a better
patch would be ensuring that's the case, since the current situation
would at least oops and show you there's an issue, but with the patch
we'll just leak the memory.

I'll apply it, but would be nice if this was handled a bit more
proactively.

-- 
Jens Axboe

Re: [PATCH] blkcg: check pol->cpd_free_fn before free cpd

[weiping zhang]

On Tue, Oct 10, 2017 at 09:04:39AM -0600, Jens Axboe wrote:
> On 10/10/2017 08:53 AM, weiping zhang wrote:
> > check pol->cpd_free_fn() instead of pol->cpd_alloc_fn() when free cpd.
> 
> In practice this shouldn't make a difference, since if you have an
> alloc_fn, you better also have a free_fn. I'd argue a better
> patch would be ensuring that's the case, since the current situation
> would at least oops and show you there's an issue, but with the patch
> we'll just leak the memory.
> 
> I'll apply it, but would be nice if this was handled a bit more
> proactively.
> 
Thanks your comments, I'll add a sanity check in blkcg_policy_register
function.