[PATCH 1/7] binder: fix alloc->vma_vm_mm null-ptr dereference

[PATCH 1/7] binder: fix alloc->vma_vm_mm null-ptr dereference

[Carlos Llamas]

Syzbot reported a couple issues introduced by commit 44e602b4e52f
("binder_alloc: add missing mmap_lock calls when using the VMA"), in
which we attempt to acquire the mmap_lock when alloc->vma_vm_mm has not
been initialized yet.

This can happen if a binder_proc receives a transaction without having
previously called mmap() to setup the binder_proc->alloc space in [1].
Also, a similar issue occurs via binder_alloc_print_pages() when we try
to dump the debugfs binder stats file in [2].

Sample of syzbot's crash report:
  ==================================================================
  KASAN: null-ptr-deref in range [0x0000000000000128-0x000000000000012f]
  CPU: 0 PID: 3755 Comm: syz-executor229 Not tainted 6.0.0-rc1-next-20220819-syzkaller #0
  syz-executor229[3755] cmdline: ./syz-executor2294415195
  Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 07/22/2022
  RIP: 0010:__lock_acquire+0xd83/0x56d0 kernel/locking/lockdep.c:4923
  [...]
  Call Trace:
   <TASK>
   lock_acquire kernel/locking/lockdep.c:5666 [inline]
   lock_acquire+0x1ab/0x570 kernel/locking/lockdep.c:5631
   down_read+0x98/0x450 kernel/locking/rwsem.c:1499
   mmap_read_lock include/linux/mmap_lock.h:117 [inline]
   binder_alloc_new_buf_locked drivers/android/binder_alloc.c:405 [inline]
   binder_alloc_new_buf+0xa5/0x19e0 drivers/android/binder_alloc.c:593
   binder_transaction+0x242e/0x9a80 drivers/android/binder.c:3199
   binder_thread_write+0x664/0x3220 drivers/android/binder.c:3986
   binder_ioctl_write_read drivers/android/binder.c:5036 [inline]
   binder_ioctl+0x3470/0x6d00 drivers/android/binder.c:5323
   vfs_ioctl fs/ioctl.c:51 [inline]
   __do_sys_ioctl fs/ioctl.c:870 [inline]
   __se_sys_ioctl fs/ioctl.c:856 [inline]
   __x64_sys_ioctl+0x193/0x200 fs/ioctl.c:856
   do_syscall_x64 arch/x86/entry/common.c:50 [inline]
   do_syscall_64+0x35/0xb0 arch/x86/entry/common.c:80
   entry_SYSCALL_64_after_hwframe+0x63/0xcd
   [...]
  ==================================================================

Fix these issues by setting up alloc->vma_vm_mm pointer during open()
and caching directly from current->mm. This guarantees we have a valid
reference to take the mmap_lock during scenarios described above.

[1] https://syzkaller.appspot.com/bug?extid=f7dc54e5be28950ac459
[2] https://syzkaller.appspot.com/bug?extid=a75ebe0452711c9e56d9

Fixes: 44e602b4e52f ("binder_alloc: add missing mmap_lock calls when using the VMA")
Reported-by: syzbot+f7dc54e5be28950ac459@syzkaller.appspotmail.com
Reported-by: syzbot+a75ebe0452711c9e56d9@syzkaller.appspotmail.com
Cc: <stable@vger.kernel.org> # v5.15+
Cc: Liam R. Howlett <Liam.Howlett@oracle.com>
Signed-off-by: Carlos Llamas <cmllamas@google.com>
---
 drivers/android/binder_alloc.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/android/binder_alloc.c b/drivers/android/binder_alloc.c
index 51f4e1c5cd01..9b1778c00610 100644
--- a/drivers/android/binder_alloc.c
+++ b/drivers/android/binder_alloc.c
@@ -322,7 +322,6 @@ static inline void binder_alloc_set_vma(struct binder_alloc *alloc,
 	 */
 	if (vma) {
 		vm_start = vma->vm_start;
-		alloc->vma_vm_mm = vma->vm_mm;
 		mmap_assert_write_locked(alloc->vma_vm_mm);
 	} else {
 		mmap_assert_locked(alloc->vma_vm_mm);
@@ -795,7 +794,6 @@ int binder_alloc_mmap_handler(struct binder_alloc *alloc,
 	binder_insert_free_buffer(alloc, buffer);
 	alloc->free_async_space = alloc->buffer_size / 2;
 	binder_alloc_set_vma(alloc, vma);
-	mmgrab(alloc->vma_vm_mm);
 
 	return 0;
 
@@ -1091,6 +1089,8 @@ static struct shrinker binder_shrinker = {
 void binder_alloc_init(struct binder_alloc *alloc)
 {
 	alloc->pid = current->group_leader->pid;
+	alloc->vma_vm_mm = current->mm;
+	mmgrab(alloc->vma_vm_mm);
 	mutex_init(&alloc->mutex);
 	INIT_LIST_HEAD(&alloc->buffers);
 }
-- 
2.37.2.672.g94769d06f0-goog

Re: [PATCH 1/7] binder: fix alloc->vma_vm_mm null-ptr dereference

[Andrew Morton]

On Mon, 29 Aug 2022 20:12:48 +0000 Carlos Llamas <cmllamas@google.com> wrote:

> Syzbot reported a couple issues introduced by commit 44e602b4e52f
> ("binder_alloc: add missing mmap_lock calls when using the VMA"), in
> which we attempt to acquire the mmap_lock when alloc->vma_vm_mm has not
> been initialized yet.
> 
> This can happen if a binder_proc receives a transaction without having
> previously called mmap() to setup the binder_proc->alloc space in [1].
> Also, a similar issue occurs via binder_alloc_print_pages() when we try
> to dump the debugfs binder stats file in [2].

Thanks.  I assume you'll be merging all these into mainline?

> 
> Fixes: 44e602b4e52f ("binder_alloc: add missing mmap_lock calls when using the VMA")
> Reported-by: syzbot+f7dc54e5be28950ac459@syzkaller.appspotmail.com
> Reported-by: syzbot+a75ebe0452711c9e56d9@syzkaller.appspotmail.com
> Cc: <stable@vger.kernel.org> # v5.15+

44e602b4e52f is only present in 6.0-rcX?

Re: [PATCH 1/7] binder: fix alloc->vma_vm_mm null-ptr dereference

[Carlos Llamas]

On Mon, Aug 29, 2022 at 01:34:52PM -0700, Andrew Morton wrote:
> On Mon, 29 Aug 2022 20:12:48 +0000 Carlos Llamas <cmllamas@google.com> wrote:
> 
> > Syzbot reported a couple issues introduced by commit 44e602b4e52f
> > ("binder_alloc: add missing mmap_lock calls when using the VMA"), in
> > which we attempt to acquire the mmap_lock when alloc->vma_vm_mm has not
> > been initialized yet.
> > 
> > This can happen if a binder_proc receives a transaction without having
> > previously called mmap() to setup the binder_proc->alloc space in [1].
> > Also, a similar issue occurs via binder_alloc_print_pages() when we try
> > to dump the debugfs binder stats file in [2].
> 
> Thanks.  I assume you'll be merging all these into mainline?

Yes, I believe Greg will pick up these patches into his char-misc tree.

> 
> > 
> > Fixes: 44e602b4e52f ("binder_alloc: add missing mmap_lock calls when using the VMA")
> > Reported-by: syzbot+f7dc54e5be28950ac459@syzkaller.appspotmail.com
> > Reported-by: syzbot+a75ebe0452711c9e56d9@syzkaller.appspotmail.com
> > Cc: <stable@vger.kernel.org> # v5.15+
> 
> 44e602b4e52f is only present in 6.0-rcX?

Right, it was just added to the stable queue earlier today:
https://lore.kernel.org/all/20220829105814.857786586@linuxfoundation.org/
https://lore.kernel.org/all/20220829105809.855177179@linuxfoundation.org/

Re: [PATCH 1/7] binder: fix alloc->vma_vm_mm null-ptr dereference

[Liam Howlett]

* Carlos Llamas <cmllamas@google.com> [220829 16:13]:
> Syzbot reported a couple issues introduced by commit 44e602b4e52f
> ("binder_alloc: add missing mmap_lock calls when using the VMA"), in
> which we attempt to acquire the mmap_lock when alloc->vma_vm_mm has not
> been initialized yet.
> 
> This can happen if a binder_proc receives a transaction without having
> previously called mmap() to setup the binder_proc->alloc space in [1].
> Also, a similar issue occurs via binder_alloc_print_pages() when we try
> to dump the debugfs binder stats file in [2].
> 
> Sample of syzbot's crash report:
>   ==================================================================
>   KASAN: null-ptr-deref in range [0x0000000000000128-0x000000000000012f]
>   CPU: 0 PID: 3755 Comm: syz-executor229 Not tainted 6.0.0-rc1-next-20220819-syzkaller #0
>   syz-executor229[3755] cmdline: ./syz-executor2294415195
>   Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 07/22/2022
>   RIP: 0010:__lock_acquire+0xd83/0x56d0 kernel/locking/lockdep.c:4923
>   [...]
>   Call Trace:
>    <TASK>
>    lock_acquire kernel/locking/lockdep.c:5666 [inline]
>    lock_acquire+0x1ab/0x570 kernel/locking/lockdep.c:5631
>    down_read+0x98/0x450 kernel/locking/rwsem.c:1499
>    mmap_read_lock include/linux/mmap_lock.h:117 [inline]
>    binder_alloc_new_buf_locked drivers/android/binder_alloc.c:405 [inline]
>    binder_alloc_new_buf+0xa5/0x19e0 drivers/android/binder_alloc.c:593
>    binder_transaction+0x242e/0x9a80 drivers/android/binder.c:3199
>    binder_thread_write+0x664/0x3220 drivers/android/binder.c:3986
>    binder_ioctl_write_read drivers/android/binder.c:5036 [inline]
>    binder_ioctl+0x3470/0x6d00 drivers/android/binder.c:5323
>    vfs_ioctl fs/ioctl.c:51 [inline]
>    __do_sys_ioctl fs/ioctl.c:870 [inline]
>    __se_sys_ioctl fs/ioctl.c:856 [inline]
>    __x64_sys_ioctl+0x193/0x200 fs/ioctl.c:856
>    do_syscall_x64 arch/x86/entry/common.c:50 [inline]
>    do_syscall_64+0x35/0xb0 arch/x86/entry/common.c:80
>    entry_SYSCALL_64_after_hwframe+0x63/0xcd
>    [...]
>   ==================================================================
> 
> Fix these issues by setting up alloc->vma_vm_mm pointer during open()
> and caching directly from current->mm. This guarantees we have a valid
> reference to take the mmap_lock during scenarios described above.
> 
> [1] https://syzkaller.appspot.com/bug?extid=f7dc54e5be28950ac459
> [2] https://syzkaller.appspot.com/bug?extid=a75ebe0452711c9e56d9
> 
> Fixes: 44e602b4e52f ("binder_alloc: add missing mmap_lock calls when using the VMA")
> Reported-by: syzbot+f7dc54e5be28950ac459@syzkaller.appspotmail.com
> Reported-by: syzbot+a75ebe0452711c9e56d9@syzkaller.appspotmail.com
> Cc: <stable@vger.kernel.org> # v5.15+
> Cc: Liam R. Howlett <Liam.Howlett@oracle.com>
> Signed-off-by: Carlos Llamas <cmllamas@google.com>
> ---
>  drivers/android/binder_alloc.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/android/binder_alloc.c b/drivers/android/binder_alloc.c
> index 51f4e1c5cd01..9b1778c00610 100644
> --- a/drivers/android/binder_alloc.c
> +++ b/drivers/android/binder_alloc.c
> @@ -322,7 +322,6 @@ static inline void binder_alloc_set_vma(struct binder_alloc *alloc,
>  	 */
>  	if (vma) {
>  		vm_start = vma->vm_start;
> -		alloc->vma_vm_mm = vma->vm_mm;

Is this really the null pointer dereference?  We check for vma above..?

>  		mmap_assert_write_locked(alloc->vma_vm_mm);
>  	} else {
>  		mmap_assert_locked(alloc->vma_vm_mm);
> @@ -795,7 +794,6 @@ int binder_alloc_mmap_handler(struct binder_alloc *alloc,
>  	binder_insert_free_buffer(alloc, buffer);
>  	alloc->free_async_space = alloc->buffer_size / 2;
>  	binder_alloc_set_vma(alloc, vma);
> -	mmgrab(alloc->vma_vm_mm);
>  
>  	return 0;
>  
> @@ -1091,6 +1089,8 @@ static struct shrinker binder_shrinker = {
>  void binder_alloc_init(struct binder_alloc *alloc)
>  {
>  	alloc->pid = current->group_leader->pid;
> +	alloc->vma_vm_mm = current->mm;
> +	mmgrab(alloc->vma_vm_mm);
>  	mutex_init(&alloc->mutex);
>  	INIT_LIST_HEAD(&alloc->buffers);
>  }
> -- 
> 2.37.2.672.g94769d06f0-goog
> 

Re: [PATCH 1/7] binder: fix alloc->vma_vm_mm null-ptr dereference

[Carlos Llamas]

On Tue, Aug 30, 2022 at 07:06:37PM +0000, Liam Howlett wrote:
> > diff --git a/drivers/android/binder_alloc.c b/drivers/android/binder_alloc.c
> > index 51f4e1c5cd01..9b1778c00610 100644
> > --- a/drivers/android/binder_alloc.c
> > +++ b/drivers/android/binder_alloc.c
> > @@ -322,7 +322,6 @@ static inline void binder_alloc_set_vma(struct binder_alloc *alloc,
> >  	 */
> >  	if (vma) {
> >  		vm_start = vma->vm_start;
> > -		alloc->vma_vm_mm = vma->vm_mm;
> 
> Is this really the null pointer dereference?  We check for vma above..?
> 

Not here. The sequence leading to the null-ptr-deref happens when we try
to take alloc->vma_vm_mm->mmap_lock in binder_alloc_new_buf_locked() and
in binder_alloc_print_pages() without initializing alloc->vma_vm_mm
first (e.g. mmap() was never called). These sequences are described in
the commit message but basically they translate to mmap_read_lock(NULL)
calls.

Re: [PATCH 1/7] binder: fix alloc->vma_vm_mm null-ptr dereference

[Liam Howlett]

* Carlos Llamas <cmllamas@google.com> [220830 15:41]:
> On Tue, Aug 30, 2022 at 07:06:37PM +0000, Liam Howlett wrote:
> > > diff --git a/drivers/android/binder_alloc.c b/drivers/android/binder_alloc.c
> > > index 51f4e1c5cd01..9b1778c00610 100644
> > > --- a/drivers/android/binder_alloc.c
> > > +++ b/drivers/android/binder_alloc.c
> > > @@ -322,7 +322,6 @@ static inline void binder_alloc_set_vma(struct binder_alloc *alloc,
> > >  	 */
> > >  	if (vma) {
> > >  		vm_start = vma->vm_start;
> > > -		alloc->vma_vm_mm = vma->vm_mm;
> > 
> > Is this really the null pointer dereference?  We check for vma above..?
> > 
> 
> Not here. The sequence leading to the null-ptr-deref happens when we try
> to take alloc->vma_vm_mm->mmap_lock in binder_alloc_new_buf_locked() and
> in binder_alloc_print_pages() without initializing alloc->vma_vm_mm
> first (e.g. mmap() was never called). These sequences are described in
> the commit message but basically they translate to mmap_read_lock(NULL)
> calls.

Ah, this is unnecessary with the rest of the change.

Feel free to add my reviewed-by if you want.


Reviewed-by: Liam R. Howlett <Liam.Howlett@oracle.com>

Re: [PATCH 1/7] binder: fix alloc->vma_vm_mm null-ptr dereference

[Todd Kjos]

On Mon, Aug 29, 2022 at 1:13 PM 'Carlos Llamas' via kernel-team
<kernel-team@android.com> wrote:
>
> Syzbot reported a couple issues introduced by commit 44e602b4e52f
> ("binder_alloc: add missing mmap_lock calls when using the VMA"), in
> which we attempt to acquire the mmap_lock when alloc->vma_vm_mm has not
> been initialized yet.
>
> This can happen if a binder_proc receives a transaction without having
> previously called mmap() to setup the binder_proc->alloc space in [1].
> Also, a similar issue occurs via binder_alloc_print_pages() when we try
> to dump the debugfs binder stats file in [2].
>
> Sample of syzbot's crash report:
>   ==================================================================
>   KASAN: null-ptr-deref in range [0x0000000000000128-0x000000000000012f]
>   CPU: 0 PID: 3755 Comm: syz-executor229 Not tainted 6.0.0-rc1-next-20220819-syzkaller #0
>   syz-executor229[3755] cmdline: ./syz-executor2294415195
>   Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 07/22/2022
>   RIP: 0010:__lock_acquire+0xd83/0x56d0 kernel/locking/lockdep.c:4923
>   [...]
>   Call Trace:
>    <TASK>
>    lock_acquire kernel/locking/lockdep.c:5666 [inline]
>    lock_acquire+0x1ab/0x570 kernel/locking/lockdep.c:5631
>    down_read+0x98/0x450 kernel/locking/rwsem.c:1499
>    mmap_read_lock include/linux/mmap_lock.h:117 [inline]
>    binder_alloc_new_buf_locked drivers/android/binder_alloc.c:405 [inline]
>    binder_alloc_new_buf+0xa5/0x19e0 drivers/android/binder_alloc.c:593
>    binder_transaction+0x242e/0x9a80 drivers/android/binder.c:3199
>    binder_thread_write+0x664/0x3220 drivers/android/binder.c:3986
>    binder_ioctl_write_read drivers/android/binder.c:5036 [inline]
>    binder_ioctl+0x3470/0x6d00 drivers/android/binder.c:5323
>    vfs_ioctl fs/ioctl.c:51 [inline]
>    __do_sys_ioctl fs/ioctl.c:870 [inline]
>    __se_sys_ioctl fs/ioctl.c:856 [inline]
>    __x64_sys_ioctl+0x193/0x200 fs/ioctl.c:856
>    do_syscall_x64 arch/x86/entry/common.c:50 [inline]
>    do_syscall_64+0x35/0xb0 arch/x86/entry/common.c:80
>    entry_SYSCALL_64_after_hwframe+0x63/0xcd
>    [...]
>   ==================================================================
>
> Fix these issues by setting up alloc->vma_vm_mm pointer during open()
> and caching directly from current->mm. This guarantees we have a valid
> reference to take the mmap_lock during scenarios described above.
>
> [1] https://syzkaller.appspot.com/bug?extid=f7dc54e5be28950ac459
> [2] https://syzkaller.appspot.com/bug?extid=a75ebe0452711c9e56d9
>
> Fixes: 44e602b4e52f ("binder_alloc: add missing mmap_lock calls when using the VMA")
> Reported-by: syzbot+f7dc54e5be28950ac459@syzkaller.appspotmail.com
> Reported-by: syzbot+a75ebe0452711c9e56d9@syzkaller.appspotmail.com
> Cc: <stable@vger.kernel.org> # v5.15+
> Cc: Liam R. Howlett <Liam.Howlett@oracle.com>
> Signed-off-by: Carlos Llamas <cmllamas@google.com>

Acked-by: Todd Kjos <tkjos@google.com>

> ---
>  drivers/android/binder_alloc.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/android/binder_alloc.c b/drivers/android/binder_alloc.c
> index 51f4e1c5cd01..9b1778c00610 100644
> --- a/drivers/android/binder_alloc.c
> +++ b/drivers/android/binder_alloc.c
> @@ -322,7 +322,6 @@ static inline void binder_alloc_set_vma(struct binder_alloc *alloc,
>          */
>         if (vma) {
>                 vm_start = vma->vm_start;
> -               alloc->vma_vm_mm = vma->vm_mm;
>                 mmap_assert_write_locked(alloc->vma_vm_mm);
>         } else {
>                 mmap_assert_locked(alloc->vma_vm_mm);
> @@ -795,7 +794,6 @@ int binder_alloc_mmap_handler(struct binder_alloc *alloc,
>         binder_insert_free_buffer(alloc, buffer);
>         alloc->free_async_space = alloc->buffer_size / 2;
>         binder_alloc_set_vma(alloc, vma);
> -       mmgrab(alloc->vma_vm_mm);
>
>         return 0;
>
> @@ -1091,6 +1089,8 @@ static struct shrinker binder_shrinker = {
>  void binder_alloc_init(struct binder_alloc *alloc)
>  {
>         alloc->pid = current->group_leader->pid;
> +       alloc->vma_vm_mm = current->mm;
> +       mmgrab(alloc->vma_vm_mm);
>         mutex_init(&alloc->mutex);
>         INIT_LIST_HEAD(&alloc->buffers);
>  }
> --
> 2.37.2.672.g94769d06f0-goog
>
> --
> To unsubscribe from this group and stop receiving emails from it, send an email to kernel-team+unsubscribe@android.com.
>