Skip to contentSkip to conversation

cfg80211: free the object allocated in wiphy_apply_custom_regulatory

PATCH In review Dongliang Mu 3 participants 12 messages Jul 23, 2021

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 cfg80211: free the object allocated in wiphy_apply_custom_regulatory

The commit beee24695157 ("cfg80211: Save the regulatory domain when setting custom regulatory") forgets to free the newly allocated regd object.

Fix this by freeing the regd object in the error handling code and deletion function - mac80211_hwsim_del_radio.

Reported-by [email protected]
Fixes beee24695157 ("cfg80211: Save the regulatory domain when setting custom regulatory")
---
 drivers/net/wireless/mac80211_hwsim.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/net/wireless/mac80211_hwsim.c +4 -0
index ffa894f7312a..20b870af6356 100644
@@ -3404,6 +3404,8 @@ static int mac80211_hwsim_new_radio(struct genl_info *info,
 	debugfs_remove_recursive(data->debugfs);
 	ieee80211_unregister_hw(data->hw);
 failed_hw:
+	if (param->regd)
+		kfree_rcu(get_wiphy_regdom(data->hw->wiphy));
 	device_release_driver(data->dev);
 failed_bind:
 	device_unregister(data->dev);
@@ -3454,6 +3456,8 @@ static void mac80211_hwsim_del_radio(struct mac80211_hwsim_data *data,
 {
 	hwsim_mcast_del_radio(data->idx, hwname, info);
 	debugfs_remove_recursive(data->debugfs);
+	if (data->regd)
+		kfree_rcu(get_wiphy_regdom(data->hw->wiphy));
 	ieee80211_unregister_hw(data->hw);
 	device_release_driver(data->dev);
 	device_unregister(data->dev);
Signature · 1 line
2.25.1
#1 lore raw
Quoted 20 lines · Dongliang Mu

The commit beee24695157 ("cfg80211: Save the regulatory domain when setting custom regulatory") forgets to free the newly allocated regd object.

Fix this by freeing the regd object in the error handling code and deletion function - mac80211_hwsim_del_radio.

Reported-by [email protected]
Fixes beee24695157 ("cfg80211: Save the regulatory domain when setting custom regulatory")
---
 drivers/net/wireless/mac80211_hwsim.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/net/wireless/mac80211_hwsim.c +0 -0
index ffa894f7312a..20b870af6356 100644

drivers/net/wireless/mac80211_hwsim.c static int mac80211_hwsim_new_radio(struct genl_info *info, · line 3408

Quoted 5 lines
drivers/net/wireless/mac80211_hwsim.c +2 -0
@@ -3404,6 +3404,8 @@ static int mac80211_hwsim_new_radio(struct genl_info *info,
        debugfs_remove_recursive(data->debugfs);
        ieee80211_unregister_hw(data->hw);
 failed_hw:
+       if (param->regd)
+               kfree_rcu(get_wiphy_regdom(data->hw->wiphy));

Here kfree_rcu supports only 1 parameter. But some example code uses the 2nd parameter: rcu_head or rcu. For example,

static void rcu_free_regdom(const struct ieee80211_regdomain *r)

{
        ......
        kfree_rcu((struct ieee80211_regdomain *)r, rcu_head);
}

If this patch has any issues, please let me know.

Quoted 15 lines
drivers/net/wireless/mac80211_hwsim.c +2 -1
        device_release_driver(data->dev);
 failed_bind:
        device_unregister(data->dev);
@@ -3454,6 +3456,8 @@ static void mac80211_hwsim_del_radio(struct mac80211_hwsim_data *data,
 {
        hwsim_mcast_del_radio(data->idx, hwname, info);
        debugfs_remove_recursive(data->debugfs);
+       if (data->regd)
+               kfree_rcu(get_wiphy_regdom(data->hw->wiphy));
        ieee80211_unregister_hw(data->hw);
        device_release_driver(data->dev);
        device_unregister(data->dev);
--

2.25.1

maintainer
#2 lore raw
In reply to Dongliang Mu
Quoted 3 lines · Dongliang Mu

The commit beee24695157 ("cfg80211: Save the regulatory domain when setting custom regulatory") forgets to free the newly allocated regd object.

Not really? It's not forgetting it, it just saves it?

+ new_regd = reg_copy_regd(regd);
+ if (IS_ERR(new_regd))
+ return;
+

+       tmp = get_wiphy_regdom(wiphy);
+       rcu_assign_pointer(wiphy->regd, new_regd);
+       rcu_free_regdom(tmp);
Quoted 2 lines

Fix this by freeing the regd object in the error handling code and deletion function - mac80211_hwsim_del_radio.

This can't be right - the same would affect all other users of that function, no?

Perhaps somewhere we have a case where wiphy->regd is leaked, but than that should be fixed more generally in cfg80211?

johannes

#3 lore raw
Quoted 7 lines · Johannes Berg
Quoted 3 lines · Dongliang Mu

The commit beee24695157 ("cfg80211: Save the regulatory domain when setting custom regulatory") forgets to free the newly allocated regd object.

Not really? It's not forgetting it, it just saves it?

Yes, it saves the regd object in the function wiphy_apply_custom_regulatory.

But its parent function - mac80211_hwsim_new_radio forgets to free this object when the ieee80211_register_hw fails.

Quoted 9 lines
(quoted diff) +7 -0
+       new_regd = reg_copy_regd(regd);
+       if (IS_ERR(new_regd))
+               return;
+
+       tmp = get_wiphy_regdom(wiphy);
+       rcu_assign_pointer(wiphy->regd, new_regd);
+       rcu_free_regdom(tmp);
Quoted 5 lines
Quoted 2 lines

Fix this by freeing the regd object in the error handling code and deletion function - mac80211_hwsim_del_radio.

This can't be right - the same would affect all other users of that function, no?

The problem occurs in the error handling code of mac80211_hwsim_new_radio, not wiphy_apply_custom_regulatory. My commit message may be not very clear.

So I think the code in the mac80211_hwsim_del_radio paired with mac80211_hwsim_new_radio should be changed correspondingly. If I miss any problems, please let me know.

I have successfully tested my patch in the syzbot dashboard [1].

[1] https://syzkaller.appspot.com/bug?extid=1638e7c770eef6b6c0d0

Quoted 6 lines

Perhaps somewhere we have a case where wiphy->regd is leaked, but than that should be fixed more generally in cfg80211?

johannes

maintainer
#4 lore raw
Quoted 10 lines · Dongliang Mu
Quoted 7 lines · Johannes Berg
Quoted 3 lines · Dongliang Mu

The commit beee24695157 ("cfg80211: Save the regulatory domain when setting custom regulatory") forgets to free the newly allocated regd object.

Not really? It's not forgetting it, it just saves it?

Yes, it saves the regd object in the function wiphy_apply_custom_regulatory.

Right.

Quoted 2 lines

But its parent function - mac80211_hwsim_new_radio forgets to free this object when the ieee80211_register_hw fails.

But why is this specific to mac80211-hwsim?

Any other code calling wiphy_apply_custom_regulatory() and then failing the subsequent wiphy_register() or otherwise calling wiphy_free() will run into the same situation.

So why wouldn't we free this in wiphy_free(), if it exists?

johannes

#5 lore raw
Quoted 18 lines · Johannes Berg
Quoted 10 lines · Dongliang Mu
Quoted 7 lines · Johannes Berg
Quoted 3 lines · Dongliang Mu

The commit beee24695157 ("cfg80211: Save the regulatory domain when setting custom regulatory") forgets to free the newly allocated regd object.

Not really? It's not forgetting it, it just saves it?

Yes, it saves the regd object in the function wiphy_apply_custom_regulatory.

Right.

Quoted 2 lines

But its parent function - mac80211_hwsim_new_radio forgets to free this object when the ieee80211_register_hw fails.

Quoted 8 lines

But why is this specific to mac80211-hwsim?

Any other code calling wiphy_apply_custom_regulatory() and then failing the subsequent wiphy_register() or otherwise calling wiphy_free() will run into the same situation.

So why wouldn't we free this in wiphy_free(), if it exists?

Hi Johannes,

if zhao in the thread is right, we don't need to add this free operation to wiphy_free().

What we should do is to only handle regd in the error handling code of mac80211_hwsim_new_radio. This will not affect other users of mac80211-hwsim. Any idea?

Quoted 2 lines

johannes

maintainer
#6 lore raw

Hi,

Quoted 2 lines · Dongliang Mu

if zhao in the thread is right, we don't need to add this free operation to wiphy_free().

Actually, no, that statement is not true.

All that zhao claimed was that the free happens correctly during unregister (or later), and that is indeed true, since it happens from

ieee80211_unregister_hw()
-> wiphy_unregister()
-> wiphy_regulatory_deregister()

However, syzbot of course is also correct. Abstracting a bit and ignoring mac80211, the problem is that here we assign it before wiphy_register(), then wiphy_register() doesn't get called or fails, and therefore we don't call wiphy_unregister(), only wiphy_free().

Hence the leak.

But you can also easily see from that description that it's not related to hwsim - we should add a secondary round of cleanups in wiphy_free() or even move the call to wiphy_regulatory_deregister() into wiphy_free(), we need to look what else this does to see if we can move it or not.

johannes

#7 lore raw
Quoted 7 lines · Johannes Berg

Hi,

Quoted 2 lines · Dongliang Mu

if zhao in the thread is right, we don't need to add this free operation to wiphy_free().

Quoted 9 lines

Actually, no, that statement is not true.

All that zhao claimed was that the free happens correctly during unregister (or later), and that is indeed true, since it happens from

ieee80211_unregister_hw()
-> wiphy_unregister()
-> wiphy_regulatory_deregister()

Thanks for your explanation. Now the situation is more clear.

Quoted 5 lines

However, syzbot of course is also correct. Abstracting a bit and ignoring mac80211, the problem is that here we assign it before wiphy_register(), then wiphy_register() doesn't get called or fails, and therefore we don't call wiphy_unregister(), only wiphy_free().

Yes, you're right. In this case, wiphy_register is not called. We should not call wiphy_unregister() to clean up anything.

Quoted 8 lines

Hence the leak.

But you can also easily see from that description that it's not related to hwsim - we should add a secondary round of cleanups in wiphy_free() or even move the call to wiphy_regulatory_deregister() into wiphy_free(), we need to look what else this does to see if we can move it or not.

I agree to move the cleanup operation of regd to wiphy_free API. That's the partial functionability of this function.

Quoted 3 lines

johannes

#8 lore raw
In reply to Dongliang Mu
Quoted 19 lines · Dongliang Mu

The commit beee24695157 ("cfg80211: Save the regulatory domain when setting custom regulatory") forgets to free the newly allocated regd object.

Fix this by freeing the regd object in the error handling code and deletion function - mac80211_hwsim_del_radio.

Reported-by [email protected]
Fixes beee24695157 ("cfg80211: Save the regulatory domain when setting custom regulatory")
---
 drivers/net/wireless/mac80211_hwsim.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/net/wireless/mac80211_hwsim.c +0 -0
index ffa894f7312a..20b870af6356 100644

drivers/net/wireless/mac80211_hwsim.c static int mac80211_hwsim_new_radio(struct genl_info *info, · line 3409

Quoted 6 lines
drivers/net/wireless/mac80211_hwsim.c +2 -0
@@ -3404,6 +3404,8 @@ static int mac80211_hwsim_new_radio(struct genl_info *info,
 	debugfs_remove_recursive(data->debugfs);
 	ieee80211_unregister_hw(data->hw);
 failed_hw:
+	if (param->regd)
+		kfree_rcu(get_wiphy_regdom(data->hw->wiphy));
 	device_release_driver(data->dev);

hw->wiphy->regd may be NULL if previous reg_copy_regd failed, so how about:
if (hw->wiphy->regd)

	rcu_free_regdom(get_wiphy_regdom(hw->wiphy))	

drivers/net/wireless/mac80211_hwsim.c static void mac80211_hwsim_del_radio(struct mac80211_hwsim_data *data, · line 3460

Quoted 8 lines
drivers/net/wireless/mac80211_hwsim.c +2 -0
 failed_bind:
 	device_unregister(data->dev);
@@ -3454,6 +3456,8 @@ static void mac80211_hwsim_del_radio(struct mac80211_hwsim_data *data,
 {
 	hwsim_mcast_del_radio(data->idx, hwname, info);
 	debugfs_remove_recursive(data->debugfs);
+	if (data->regd)
+		kfree_rcu(get_wiphy_regdom(data->hw->wiphy));

this is not correct, because ieee80211_unregister_hw below will free data->hw_wiphy->regd

Quoted 4 lines
drivers/net/wireless/mac80211_hwsim.c +0 -0
 	ieee80211_unregister_hw(data->hw);
 	device_release_driver(data->dev);
 	device_unregister(data->dev);
#9 lore raw
Quoted 32 lines · xiaoqiang zhao
Quoted 19 lines · Dongliang Mu

The commit beee24695157 ("cfg80211: Save the regulatory domain when setting custom regulatory") forgets to free the newly allocated regd object.

Fix this by freeing the regd object in the error handling code and deletion function - mac80211_hwsim_del_radio.

Reported-by [email protected]
Fixes beee24695157 ("cfg80211: Save the regulatory domain when setting custom regulatory")
---
 drivers/net/wireless/mac80211_hwsim.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/net/wireless/mac80211_hwsim.c +0 -0
index ffa894f7312a..20b870af6356 100644

drivers/net/wireless/mac80211_hwsim.c static int mac80211_hwsim_new_radio(struct genl_info *info, · line 3409

Quoted 6 lines
drivers/net/wireless/mac80211_hwsim.c +2 -0
@@ -3404,6 +3404,8 @@ static int mac80211_hwsim_new_radio(struct genl_info *info,
      debugfs_remove_recursive(data->debugfs);
      ieee80211_unregister_hw(data->hw);
 failed_hw:
+     if (param->regd)
+             kfree_rcu(get_wiphy_regdom(data->hw->wiphy));
      device_release_driver(data->dev);

hw->wiphy->regd may be NULL if previous reg_copy_regd failed, so how about:
if (hw->wiphy->regd)

Quoted 1 line
        rcu_free_regdom(get_wiphy_regdom(hw->wiphy))

Previously I would like to use this API(rcu_free_regdom), but it is static and located in non-global header file - reg.h.

Quoted 11 lines

drivers/net/wireless/mac80211_hwsim.c static void mac80211_hwsim_del_radio(struct mac80211_hwsim_data *data, · line 3460

Quoted 9 lines
drivers/net/wireless/mac80211_hwsim.c +2 -0
 failed_bind:
      device_unregister(data->dev);
@@ -3454,6 +3456,8 @@ static void mac80211_hwsim_del_radio(struct mac80211_hwsim_data *data,
 {
      hwsim_mcast_del_radio(data->idx, hwname, info);
      debugfs_remove_recursive(data->debugfs);
+     if (data->regd)
+             kfree_rcu(get_wiphy_regdom(data->hw->wiphy));

this is not correct, because ieee80211_unregister_hw below will free data->hw_wiphy->regd

Can you point out the concrete code releasing regd? Maybe the link to elixir.

Quoted 4 lines
Quoted 4 lines
drivers/net/wireless/mac80211_hwsim.c +0 -0
      ieee80211_unregister_hw(data->hw);
      device_release_driver(data->dev);
      device_unregister(data->dev);
#10 lore raw
Quoted 6 lines · Dongliang Mu

Can you point out the concrete code releasing regd? Maybe the link to elixir.

Quoted 4 lines
Quoted 4 lines
      ieee80211_unregister_hw(data->hw);
      device_release_driver(data->dev);
      device_unregister(data->dev);

call graph seems like this:

ieee80211_unregister_hw
(https://elixir.bootlin.com/linux/v5.14-rc2/source/net/mac80211/main.c#L1368)
wiphy_unregister
(https://elixir.bootlin.com/linux/v5.14-rc2/source/net/wireless/core.c#L1011)
wiphy_regulatory_deregister
(https://elixir.bootlin.com/linux/v5.14-rc2/source/net/wireless/reg.c#L4057)
rcu_free_regdom

#11 lore raw
Quoted 13 lines · xiaoqiang zhao
Quoted 6 lines · Dongliang Mu

Can you point out the concrete code releasing regd? Maybe the link to elixir.

Quoted 4 lines
Quoted 4 lines
      ieee80211_unregister_hw(data->hw);
      device_release_driver(data->dev);
      device_unregister(data->dev);

call graph seems like this:

Quoted 7 lines

ieee80211_unregister_hw
(https://elixir.bootlin.com/linux/v5.14-rc2/source/net/mac80211/main.c#L1368)
wiphy_unregister
(https://elixir.bootlin.com/linux/v5.14-rc2/source/net/wireless/core.c#L1011)
wiphy_regulatory_deregister
(https://elixir.bootlin.com/linux/v5.14-rc2/source/net/wireless/reg.c#L4057)
rcu_free_regdom

Thanks very much. This is really helpful.

Quoted 1 line