ax25: fix NPD and UAF bugs when detaching ax25 device
Subsystems (2)
cover ax25: fix NPD and UAF bugs when detaching ax25 device
There are NPD and UAF bugs when detaching ax25 device, we use lock and refcount to mitigate these bugs.
Duoming Zhou (2):
ax25: improve the incomplete fix to avoid UAF and NPD bugs
ax25: add refcount in ax25_dev to avoid UAF bugs
include/net/ax25.h | 10 ++++++++++ net/ax25/af_ax25.c | 11 ++++++++--- net/ax25/ax25_dev.c | 12 ++++++++++-- net/ax25/ax25_route.c | 3 +++ 4 files changed, 31 insertions(+), 5 deletions(-)
Signature · 1 line
2.17.1
PATCH 1/2 ax25: improve the incomplete fix to avoid UAF and NPD bugs
The previous commit 1ade48d0c27d ("ax25: NPD bug when detaching AX25 device") introduce lock_sock() into ax25_kill_by_device to prevent NPD bug. But the concurrency NPD or UAF bug will occur, when lock_sock() or release_sock() dereferences the ax25_cb->sock.
The NULL pointer dereference bug can be shown as below:
ax25_kill_by_device() | ax25_release()
| ax25_destroy_socket()
| ax25_cb_del()
... | ...
| ax25->sk=NULL;
lock_sock(s->sk); //(1) |
s->ax25_dev = NULL; | ...
release_sock(s->sk); //(2) |
... |The root cause is that the sock is set to null before dereference site (1) or (2). Therefore, this patch extracts the ax25_cb->sock in advance, and uses ax25_list_lock to protect it, which can synchronize with ax25_cb_del() and ensure the value of sock is not null before dereference sites.
The concurrency UAF bug can be shown as below:
ax25_kill_by_device() | ax25_release()
| ax25_destroy_socket()
... | ...
| sock_put(sk); //FREE
lock_sock(s->sk); //(1) |
s->ax25_dev = NULL; | ...
release_sock(s->sk); //(2) |
... |The root cause is that the sock is released before dereference site (1) or (2). Therefore, this patch uses sock_hold() to increase the refcount of sock and uses ax25_list_lock to protect it, which can synchronize with ax25_cb_del() in ax25_destroy_socket() and ensure the sock wil not be released before dereference sites.
--- net/ax25/af_ax25.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-)
net/ax25/af_ax25.c +6 -3
@@ -77,6 +77,7 @@ static void ax25_kill_by_device(struct net_device *dev) {ax25_dev *ax25_dev;ax25_cb *s; +struct sock *sk;if ((ax25_dev =ax25_dev_ax25dev (dev)) ==NULL )return ;
@@ -85,13 +86,15 @@ static void ax25_kill_by_device(struct net_device *dev) again:ax25_for_each (s, &ax25_list) {if (s->ax25_dev == ax25_dev) { + sk = s->sk; +sock_hold (sk);spin_unlock_bh (&ax25_list_lock); -lock_sock (s->sk); +lock_sock (sk); s->ax25_dev =NULL ; -release_sock (s->sk); +release_sock (sk);ax25_disconnect (s,ENETUNREACH );spin_lock_bh (&ax25_list_lock); - +sock_put (sk);/* The entry could have been deleted from the * list meanwhile and thus the next pointer is * no longer valid. Play it safe and restart
Signature · 1 line
2.17.1
PATCH 2/2 ax25: add refcount in ax25_dev to avoid UAF bugs
In reply to Duoming ZhouIf we dereference ax25_dev after we call kfree(ax25_dev) in ax25_dev_device_down(), it will lead to concurrency UAF bugs. There are eight syscall functions suffer from UAF bugs, include ax25_bind(), ax25_release(), ax25_connect(), ax25_ioctl(), ax25_getname(), ax25_sendmsg(), ax25_getsockopt() and ax25_info_show().
One of the concurrency UAF can be shown as below:
(USE) | (FREE)
| ax25_device_event
| ax25_dev_device_down
ax25_bind | ...
... | kfree(ax25_dev)
ax25_fillin_cb() | ...
ax25_fillin_cb_from_dev() |
... |The root cause of UAF bugs is that kfree(ax25_dev) in ax25_dev_device_down() is not protected by any locks. When ax25_dev, which there are still pointers point to, is released, the concurrency UAF bug will happen.
This patch introduces refcount into ax25_dev in order to guarantee that there are no pointers point to it when ax25_dev is released.
--- include/net/ax25.h | 10 ++++++++++ net/ax25/af_ax25.c | 2 ++ net/ax25/ax25_dev.c | 12 ++++++++++-- net/ax25/ax25_route.c | 3 +++ 4 files changed, 25 insertions(+), 2 deletions(-)
include/net/ax25.h +10 -0
@@ -239,6 +239,7 @@ typedef struct ax25_dev {#if defined(CONFIG_AX25_DAMA_SLAVE ) || defined(CONFIG_AX25_DAMA_MASTER )ax25_dama_info dama;#endif +refcount_t refcount; } ax25_dev; typedef struct ax25_cb {
@@ -293,6 +294,15 @@ static __inline__ void ax25_cb_put(ax25_cb *ax25) } } +#define ax25_dev_hold (__ax25_dev) \ + refcount_inc(&((__ax25_dev)->refcount)) + +static __inline__void ax25_dev_put (ax25_dev *ax25_dev) +{ +if (refcount_dec_and_test (&ax25_dev->refcount)) { +kfree (ax25_dev); + } +}static inline __be16 ax25_type_trans (struct sk_buff *skb,struct net_device *dev) { skb->dev = dev;
net/ax25/af_ax25.c +2 -0
@@ -91,6 +91,7 @@ static void ax25_kill_by_device(struct net_device *dev)spin_unlock_bh (&ax25_list_lock);lock_sock (sk); s->ax25_dev =NULL ; +ax25_dev_put (ax25_dev);release_sock (sk);ax25_disconnect (s,ENETUNREACH );spin_lock_bh (&ax25_list_lock);
@@ -439,6 +440,7 @@ static int ax25_ctl_ioctl(const unsigned int cmd, void __user *arg) } out_put: +ax25_dev_put (ax25_dev);ax25_cb_put (ax25);return ret;
net/ax25/ax25_dev.c +10 -2
@@ -37,6 +37,7 @@ ax25_dev *ax25_addr_ax25dev(ax25_address *addr)for (ax25_dev = ax25_dev_list; ax25_dev !=NULL ; ax25_dev = ax25_dev->next)if (ax25cmp (addr, (const ax25_address *)ax25_dev->dev->dev_addr) ==0 ) { res = ax25_dev; +ax25_dev_hold (ax25_dev); }spin_unlock_bh (&ax25_dev_lock);
@@ -56,6 +57,7 @@ void ax25_dev_device_up(struct net_device *dev)return ; } +refcount_set (&ax25_dev->refcount,1 ); dev->ax25_ptr = ax25_dev; ax25_dev->dev = dev;dev_hold_track (dev, &ax25_dev->dev_tracker,GFP_ATOMIC );
@@ -83,6 +85,7 @@ void ax25_dev_device_up(struct net_device *dev)spin_lock_bh (&ax25_dev_lock); ax25_dev->next = ax25_dev_list; ax25_dev_list = ax25_dev; +ax25_dev_hold (ax25_dev);spin_unlock_bh (&ax25_dev_lock);ax25_register_dev_sysctl (ax25_dev);
@@ -112,20 +115,22 @@ void ax25_dev_device_down(struct net_device *dev)if ((s = ax25_dev_list) == ax25_dev) { ax25_dev_list = s->next; +ax25_dev_put (ax25_dev);spin_unlock_bh (&ax25_dev_lock); dev->ax25_ptr =NULL ;dev_put_track (dev, &ax25_dev->dev_tracker); -kfree (ax25_dev); +ax25_dev_put (ax25_dev);return ; }while (s !=NULL && s->next !=NULL ) {if (s->next == ax25_dev) { s->next = ax25_dev->next; +ax25_dev_put (ax25_dev);spin_unlock_bh (&ax25_dev_lock); dev->ax25_ptr =NULL ;dev_put_track (dev, &ax25_dev->dev_tracker); -kfree (ax25_dev); +ax25_dev_put (ax25_dev);return ; }
@@ -133,6 +138,7 @@ void ax25_dev_device_down(struct net_device *dev) }spin_unlock_bh (&ax25_dev_lock); dev->ax25_ptr =NULL ; +ax25_dev_put (ax25_dev); }int ax25_fwd_ioctl (unsigned int cmd,struct ax25_fwd_struct *fwd)
@@ -149,6 +155,7 @@ int ax25_fwd_ioctl(unsigned int cmd, struct ax25_fwd_struct *fwd)if (ax25_dev->forward !=NULL )return -EINVAL ; ax25_dev->forward = fwd_dev->dev; +ax25_dev_put (fwd_dev);break ;case SIOCAX25DELFWD :
@@ -161,6 +168,7 @@ int ax25_fwd_ioctl(unsigned int cmd, struct ax25_fwd_struct *fwd)return -EINVAL ; } +ax25_dev_put (ax25_dev);return 0 ; }
net/ax25/ax25_route.c +3 -0
@@ -116,6 +116,7 @@ static int __must_check ax25_rt_add(struct ax25_routes_struct *route) ax25_rt->dev = ax25_dev->dev; ax25_rt->digipeat =NULL ; ax25_rt->ip_mode =' ' ; +ax25_dev_put (ax25_dev);if (route->digi_count !=0 ) {if ((ax25_rt->digipeat =kmalloc (sizeof (ax25_digi),GFP_ATOMIC )) ==NULL ) {write_unlock_bh (&ax25_route_lock);
@@ -172,6 +173,7 @@ static int ax25_rt_del(struct ax25_routes_struct *route) } } } +ax25_dev_put (ax25_dev);write_unlock_bh (&ax25_route_lock);return 0 ;
@@ -214,6 +216,7 @@ static int ax25_rt_opt(struct ax25_route_opt_struct *rt_option) } out: +ax25_dev_put (ax25_dev);write_unlock_bh (&ax25_route_lock);return err; }
Signature · 1 line
2.17.1
Re: [PATCH 2/2] ax25: add refcount in ax25_dev to avoid UAF bugs
Quoted 13 lines · Duoming Zhou
include/net/ax25.h +1
-0@@ -239,6 +239,7 @@ typedef struct ax25_dev {#if defined(CONFIG_AX25_DAMA_SLAVE ) || defined(CONFIG_AX25_DAMA_MASTER )ax25_dama_info dama;#endif +refcount_t refcount; } ax25_dev; typedef struct ax25_cb {
include/net/ax25.h static __inline__ void ax25_cb_put(ax25_cb *ax25) · line 298
Quoted 5 lines
include/net/ax25.h +2
-0@@ -293,6 +294,15 @@ static __inline__ void ax25_cb_put(ax25_cb *ax25) } } +#define ax25_dev_hold (__ax25_dev) \ + refcount_inc(&((__ax25_dev)->refcount))
Make this an inline function.
include/net/ax25.h static __inline__ void ax25_cb_put(ax25_cb *ax25) · line 300
Quoted 2 lines
include/net/ax25.h +2
-0+ +static __inline__void ax25_dev_put (ax25_dev *ax25_dev)
Please run checkpatch.pl --strict on your patches. s/__inline__/inline/
include/net/ax25.h static __inline__ void ax25_cb_put(ax25_cb *ax25) · line 304
Quoted 4 lines
include/net/ax25.h +4
-0+{ +if (refcount_dec_and_test (&ax25_dev->refcount)) { +kfree (ax25_dev); + }
Delete the extra curly braces.
Quoted 17 lines
include/net/ax25.h +1
-0+}static inline __be16 ax25_type_trans (struct sk_buff *skb,struct net_device *dev) { skb->dev = dev;net/ax25/af_ax25.c +1
-0@@ -91,6 +91,7 @@ static void ax25_kill_by_device(struct net_device *dev)spin_unlock_bh (&ax25_list_lock);lock_sock (sk); s->ax25_dev =NULL ; +ax25_dev_put (ax25_dev);release_sock (sk);ax25_disconnect (s,ENETUNREACH );spin_lock_bh (&ax25_list_lock);
net/ax25/af_ax25.c static int ax25_ctl_ioctl(const unsigned int cmd, void __user *arg) · line 443
Quoted 4 lines
net/ax25/af_ax25.c +1
-0@@ -439,6 +440,7 @@ static int ax25_ctl_ioctl(const unsigned int cmd, void __user *arg) } out_put: +ax25_dev_put (ax25_dev);
The ax25_ctl_ioctl() has a ton of reference leak paths now. Almost every return -ESOMETHING needs to be fixed.
Quoted 32 lines
net/ax25/af_ax25.c +0
-0ax25_cb_put (ax25);return ret;net/ax25/ax25_dev.c +3
-0@@ -37,6 +37,7 @@ ax25_dev *ax25_addr_ax25dev(ax25_address *addr)for (ax25_dev = ax25_dev_list; ax25_dev !=NULL ; ax25_dev = ax25_dev->next)if (ax25cmp (addr, (const ax25_address *)ax25_dev->dev->dev_addr) ==0 ) { res = ax25_dev; +ax25_dev_hold (ax25_dev); }spin_unlock_bh (&ax25_dev_lock);@@ -56,6 +57,7 @@ void ax25_dev_device_up(struct net_device *dev)return ; } +refcount_set (&ax25_dev->refcount,1 ); dev->ax25_ptr = ax25_dev; ax25_dev->dev = dev;dev_hold_track (dev, &ax25_dev->dev_tracker,GFP_ATOMIC );@@ -83,6 +85,7 @@ void ax25_dev_device_up(struct net_device *dev)spin_lock_bh (&ax25_dev_lock); ax25_dev->next = ax25_dev_list; ax25_dev_list = ax25_dev; +ax25_dev_hold (ax25_dev);spin_unlock_bh (&ax25_dev_lock);ax25_register_dev_sysctl (ax25_dev);
net/ax25/ax25_dev.c void ax25_dev_device_down(struct net_device *dev) · line 118
Quoted 4 lines
net/ax25/ax25_dev.c +1
-0@@ -112,20 +115,22 @@ void ax25_dev_device_down(struct net_device *dev)if ((s = ax25_dev_list) == ax25_dev) { ax25_dev_list = s->next; +ax25_dev_put (ax25_dev);
Do we not have to call ax25_dev_hold(s->next)?
net/ax25/ax25_dev.c void ax25_dev_device_down(struct net_device *dev) · line 129
Quoted 12 lines
net/ax25/ax25_dev.c +2
-1spin_unlock_bh (&ax25_dev_lock); dev->ax25_ptr =NULL ;dev_put_track (dev, &ax25_dev->dev_tracker); -kfree (ax25_dev); +ax25_dev_put (ax25_dev);return ; }while (s !=NULL && s->next !=NULL ) {if (s->next == ax25_dev) { s->next = ax25_dev->next; +ax25_dev_put (ax25_dev);
ax25_dev_hold(ax25_dev->next)?
Quoted 17 lines
net/ax25/ax25_dev.c +2
-1spin_unlock_bh (&ax25_dev_lock); dev->ax25_ptr =NULL ;dev_put_track (dev, &ax25_dev->dev_tracker); -kfree (ax25_dev); +ax25_dev_put (ax25_dev);return ; }@@ -133,6 +138,7 @@ void ax25_dev_device_down(struct net_device *dev) }spin_unlock_bh (&ax25_dev_lock); dev->ax25_ptr =NULL ; +ax25_dev_put (ax25_dev); }int ax25_fwd_ioctl (unsigned int cmd,struct ax25_fwd_struct *fwd)
net/ax25/ax25_dev.c int ax25_fwd_ioctl(unsigned int cmd, struct ax25_fwd_struct *fwd) · line 156
Quoted 2 lines
net/ax25/ax25_dev.c +0
-0@@ -149,6 +155,7 @@ int ax25_fwd_ioctl(unsigned int cmd, struct ax25_fwd_struct *fwd)if (ax25_dev->forward !=NULL )return -EINVAL ;
Every return -ERROR; in this function leaks reference counts. This one should drop the reference for both fwd_dev and ax25_dev.
net/ax25/ax25_dev.c int ax25_fwd_ioctl(unsigned int cmd, struct ax25_fwd_struct *fwd) · line 161
Quoted 5 lines
net/ax25/ax25_dev.c +1
-0ax25_dev->forward = fwd_dev->dev; +ax25_dev_put (fwd_dev);break ;case SIOCAX25DELFWD :
regards,
dan carpenter
Re: Re: [PATCH 2/2] ax25: add refcount in ax25_dev to avoid UAF bugs
Thank you very much for your time and pointing out problems in my patch. Another two questions you asked is shown below:
net/ax25/ax25_dev.c void ax25_dev_device_down(struct net_device *dev) · line 118
Quoted 5 lines
net/ax25/ax25_dev.c +1
-0@@ -112,20 +115,22 @@ void ax25_dev_device_down(struct net_device *dev)if ((s = ax25_dev_list) == ax25_dev) { ax25_dev_list = s->next; +ax25_dev_put (ax25_dev);
Do we not have to call ax25_dev_hold(s->next)?
net/ax25/ax25_dev.c void ax25_dev_device_down(struct net_device *dev) · line 129
Quoted 12 lines
net/ax25/ax25_dev.c +2
-1spin_unlock_bh (&ax25_dev_lock); dev->ax25_ptr =NULL ;dev_put_track (dev, &ax25_dev->dev_tracker); -kfree (ax25_dev); +ax25_dev_put (ax25_dev);return ; }while (s !=NULL && s->next !=NULL ) {if (s->next == ax25_dev) { s->next = ax25_dev->next; +ax25_dev_put (ax25_dev);
ax25_dev_hold(ax25_dev->next)?
Answer:
We don't have to call ax25_dev_hold(s->next) or ax25_dev_hold(ax25_dev->next) in ax25_dev_device_down() because we have already increased the refcount when we insert ax25_dev into the linked list in ax25_dev_device_up().
net/ax25/ax25_dev.c void ax25_dev_device_up(struct net_device *dev) · line 89
Quoted 6 lines
net/ax25/ax25_dev.c +1
-0@@ -83,6 +85,7 @@ void ax25_dev_device_up(struct net_device *dev)spin_lock_bh (&ax25_dev_lock); ax25_dev->next = ax25_dev_list; ax25_dev_list = ax25_dev; +ax25_dev_hold (ax25_dev);spin_unlock_bh (&ax25_dev_lock);
Best wishes,
Duoming Zhou
Re: [PATCH 2/2] ax25: add refcount in ax25_dev to avoid UAF bugs
In reply to Duoming ZhouQuoted 23 lines · Duoming Zhou
If we dereference
ax25_devafter we callkfree(ax25_dev)inax25_dev_device_down(), it will lead to concurrency UAF bugs. There are eight syscall functions suffer from UAF bugs, includeax25_bind(),ax25_release(),ax25_connect(),ax25_ioctl(),ax25_getname(),ax25_sendmsg(),ax25_getsockopt()andax25_info_show().One of the concurrency UAF can be shown as below:
(USE) | (FREE) | ax25_device_event | ax25_dev_device_down ax25_bind | ... ... | kfree(ax25_dev) ax25_fillin_cb() | ... ax25_fillin_cb_from_dev() | ... |The root cause of UAF bugs is that
kfree(ax25_dev)inax25_dev_device_down()is not protected by any locks. Whenax25_dev, which there are still pointers point to, is released, the concurrency UAF bug will happen.
Quoted 5 lines
This patch introduces refcount into
ax25_devin order to guarantee that there are no pointers point to it whenax25_devis released.Signed-off-by Duoming Zhou <[email protected]>
I pointed out a few bugs in my previous email. I've had more time to look at it now.
Basically you just want to audit all the calls sites which call ax25_dev_ax25dev() and make sure all the error paths decrement. Most of them are buggy. I'm testing a new Smatch check which is supposed to detect these sorts of bugs.
I think the refcount in ax25_bind() needs a matching decrement. Where is that? I don't know networking well enough to know the answer to this...
net/ax25/ax25_dev.c void ax25_dev_device_down(struct net_device *dev) · line 118
Quoted 5 lines
net/ax25/ax25_dev.c +1
-0@@ -112,20 +115,22 @@ void ax25_dev_device_down(struct net_device *dev)if ((s = ax25_dev_list) == ax25_dev) { ax25_dev_list = s->next; +ax25_dev_put (ax25_dev);
It would be more readable to do ax25_dev_put(ax25_dev_list). It's weird to put ax25_dev here and then a couple lines later
net/ax25/ax25_dev.c void ax25_dev_device_down(struct net_device *dev) · line 122
Quoted 5 lines
net/ax25/ax25_dev.c +1
-1spin_unlock_bh (&ax25_dev_lock); dev->ax25_ptr =NULL ;dev_put_track (dev, &ax25_dev->dev_tracker); -kfree (ax25_dev); +ax25_dev_put (ax25_dev);
Here
net/ax25/ax25_dev.c void ax25_dev_device_down(struct net_device *dev) · line 129
Quoted 7 lines
net/ax25/ax25_dev.c +1
-0return ; }while (s !=NULL && s->next !=NULL ) {if (s->next == ax25_dev) { s->next = ax25_dev->next; +ax25_dev_put (ax25_dev);
Same.
net/ax25/ax25_dev.c void ax25_dev_device_down(struct net_device *dev) · line 125
Quoted 8 lines
net/ax25/ax25_dev.c +1
-1spin_unlock_bh (&ax25_dev_lock); dev->ax25_ptr =NULL ;dev_put_track (dev, &ax25_dev->dev_tracker); -kfree (ax25_dev); +ax25_dev_put (ax25_dev);return ; }
regards,
dan carpenter
Re: Re: [PATCH 2/2] ax25: add refcount in ax25_dev to avoid UAF bugs
Thank you very much for your time and pointing out problems in my patch.
The decrement of ax25_bind() is in ax25_kill_by_device(). If we don't call ax25_bind() before ax25_kill_by_device(), the ax25_list will be empty and ax25_dev_put() in ax25_kill_by_device() will not execute.
net/ax25/af_ax25.c static void ax25_kill_by_device(struct net_device *dev) · line 97
Quoted 8 lines
net/ax25/af_ax25.c +1
-0@@ -91,6 +91,7 @@ static void ax25_kill_by_device(struct net_device *dev)spin_unlock_bh (&ax25_list_lock);lock_sock (sk); s->ax25_dev =NULL ; +ax25_dev_put (ax25_dev);release_sock (sk);ax25_disconnect (s,ENETUNREACH );spin_lock_bh (&ax25_list_lock);
I will send the improved patch as soon as possible.
Best wishes,
Duoming Zhou
This series was applied to netdev/net.git (master) by David S. Miller <[email protected]>:
Hello:
This series was applied to netdev/net.git (master) by David S. Miller <[email protected]>:
Quoted 8 lines · On Fri, 28 Jan 2022 12:47:14 +0800 you wrote:
There are NPD and UAF bugs when detaching ax25 device, we use lock and refcount to mitigate these bugs.
Duoming Zhou (2):
ax25: improve the incomplete fix to avoid UAF and NPD bugs
ax25: add refcount inax25_devto avoid UAF bugs[...]
Here is the summary with links:
- [1/2] ax25: improve the incomplete fix to avoid UAF and NPD bugs
https://git.kernel.org/netdev/net/c/4e0f718daf97 - [2/2] ax25: add refcount in ax25_dev to avoid UAF bugs https://git.kernel.org/netdev/net/c/d01ffb9eee4a
You are awesome, thank you!
Signature · 2 lines
Deet-doot-dot, I am a bot. https://korg.docs.kernel.org/patchwork/pwbot.html