[PATCH net] ptr_ring: fail early if queue occupies more than KMALLOC_MAX_SIZE

To avoid slab to warn about exceeded size, fail early if queue
occupies more than KMALLOC_MAX_SIZE.

Reported-by: syzbot+e4d4f9ddd4295539735d@syzkaller.appspotmail.com
Signed-off-by: Jason Wang <redacted>
---
 include/linux/ptr_ring.h | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/include/linux/ptr_ring.h b/include/linux/ptr_ring.h
index 1883d61..4b862da 100644
--- a/include/linux/ptr_ring.h
+++ b/include/linux/ptr_ring.h

@@ -466,6 +466,8 @@ static inline int ptr_ring_consume_batched_bh(struct ptr_ring *r,
 
 static inline void **__ptr_ring_init_queue_alloc(unsigned int size, gfp_t gfp)
 {
+	if (size > KMALLOC_MAX_SIZE)
+		return NULL;
 	return kcalloc(size, sizeof(void *), gfp);
 }
 

-- 
2.7.4

On 2018年02月07日 16:08, Jason Wang wrote:

Fixes: 2e0ab8ca83c12 ("ptr_ring: array based FIFO for pointers")

On Wed, Feb 07, 2018 at 05:17:59PM +0800, Jason Wang wrote:

On 2018年02月07日 16:08, Jason Wang wrote:

quoted

To avoid slab to warn about exceeded size, fail early if queue
occupies more than KMALLOC_MAX_SIZE.

Reported-by: syzbot+e4d4f9ddd4295539735d@syzkaller.appspotmail.com
Signed-off-by: Jason Wang <redacted>
---
  include/linux/ptr_ring.h | 2 ++
  1 file changed, 2 insertions(+)

diff --git a/include/linux/ptr_ring.h b/include/linux/ptr_ring.h
index 1883d61..4b862da 100644
--- a/include/linux/ptr_ring.h
+++ b/include/linux/ptr_ring.h

@@ -466,6 +466,8 @@ static inline int ptr_ring_consume_batched_bh(struct ptr_ring *r,
  static inline void **__ptr_ring_init_queue_alloc(unsigned int size, gfp_t gfp)
  {
+	if (size > KMALLOC_MAX_SIZE)
+		return NULL;
  	return kcalloc(size, sizeof(void *), gfp);
  }

Fixes: 2e0ab8ca83c12 ("ptr_ring: array based FIFO for pointers")

That's probably not enough. How about switching to kvmalloc and limiting
this drastically to e.g. 64K. I vaguely remember that some wise man said
that should be enough for everybody :)

-- 
MST

On 2018年02月07日 23:15, Michael S. Tsirkin wrote:

On Wed, Feb 07, 2018 at 05:17:59PM +0800, Jason Wang wrote:

quoted

On 2018年02月07日 16:08, Jason Wang wrote:

quoted

To avoid slab to warn about exceeded size, fail early if queue
occupies more than KMALLOC_MAX_SIZE.

Reported-by: syzbot+e4d4f9ddd4295539735d@syzkaller.appspotmail.com
Signed-off-by: Jason Wang <redacted>
---
   include/linux/ptr_ring.h | 2 ++
   1 file changed, 2 insertions(+)

diff --git a/include/linux/ptr_ring.h b/include/linux/ptr_ring.h
index 1883d61..4b862da 100644
--- a/include/linux/ptr_ring.h
+++ b/include/linux/ptr_ring.h

@@ -466,6 +466,8 @@ static inline int ptr_ring_consume_batched_bh(struct ptr_ring *r,
   static inline void **__ptr_ring_init_queue_alloc(unsigned int size, gfp_t gfp)
   {
+	if (size > KMALLOC_MAX_SIZE)
+		return NULL;
   	return kcalloc(size, sizeof(void *), gfp);
   }

Fixes: 2e0ab8ca83c12 ("ptr_ring: array based FIFO for pointers")

That's probably not enough. How about switching to kvmalloc and limiting
this drastically to e.g. 64K. I vaguely remember that some wise man said
that should be enough for everybody :)

Okay, let me post a V2 to see if there's any objection.

Thanks

On Wed, Feb 07, 2018 at 04:08:25PM +0800, Jason Wang wrote:

I guess this approach does begin to make more sense
at least as a temporary stop-gap.

But does this actually prevent the crash in all cases?

size is in void* entry units, KMALLOC_MAX_SIZE is in bytes.

 
-- 
2.7.4