Impact
What kind of vulnerability is it? Who is impacted?
This vulnerability impacts all the initialization functions on the Heap
and LockedHeap
types, including Heap::new
, Heap::init
, Heap::init_from_slice
, and LockedHeap::new
. It also affects multiple uses of the Heap::extend
method.
Initialization Functions
The heap initialization methods were missing a minimum size check for the given heap size argument. This could lead to out-of-bound writes when a heap was initialized with a size smaller than 3 * size_of::<usize>
because of metadata write operations.
Heap::extend
This vulnerability impacts three specific uses of the Heap::extend
method:
- When calling
Heap::extend
with a size smaller than two usize
s (e.g., 16 on x86_64
), the size was erroneously rounded up to the minimum size, which could result in an out-of-bounds write.
- Calling
Heap::extend
on an empty heap tried to construct a heap starting at address 0, which is also an out-of-bounds write.
- One specific way to trigger this accidentally is to call
Heap::new
(or a similar constructor) with a heap size that is smaller than two usize
s. This was treated as an empty heap as well.
- Calling
Heap::extend
on a heap whose size is not a multiple of the size of two usize
s resulted in unaligned writes. It also left the heap in an unexpected state, which might lead to subsequent issues. We did not find a way to exploit this undefined behavior yet (apart from DoS on platforms that fault on unaligned writes).
Patches
Has the problem been patched? What versions should users upgrade to?
We published a patch in version 0.10.2
and recommend all users to upgrade to it. This patch release includes the following changes:
- The initialization functions now panic if the given size is not large enough to store the necessary metadata. Depending on the alignment of the heap bottom pointer, the minimum size is between
2 * size_of::<usize>
and 3 * size_of::<usize>
.
- The
extend
method now panics when trying to extend an unitialized heap.
- Extend calls with a size smaller than
size_of::<usize>() * 2
are now buffered internally and not added to the list directly. The buffered region will be merged with future extend
calls.
- The
size()
method now returns the usable size of the heap, which might be slightly smaller than the top() - bottom()
difference because of alignment constraints.
Workarounds
Is there a way for users to fix or remediate the vulnerability without upgrading?
To avoid this issue, ensure that the heap is only initialized with a size larger than 3 * size_of::<usize>
and that the Heap::extend
method is only called with sizes larger than 2 * size_of::<usize>()
. Also, ensure that the total heap size is (and stays) a multiple of 2 * size_of::<usize>()
.
For more information
If you have any questions or comments about this advisory:
Acknowledgements
This issue was responsibly reported by Evan Richter at ForAllSecure and found with Mayhem and cargo fuzz.
Impact
What kind of vulnerability is it? Who is impacted?
This vulnerability impacts all the initialization functions on the
Heap
andLockedHeap
types, includingHeap::new
,Heap::init
,Heap::init_from_slice
, andLockedHeap::new
. It also affects multiple uses of theHeap::extend
method.Initialization Functions
The heap initialization methods were missing a minimum size check for the given heap size argument. This could lead to out-of-bound writes when a heap was initialized with a size smaller than
3 * size_of::<usize>
because of metadata write operations.Heap::extend
This vulnerability impacts three specific uses of the
Heap::extend
method:Heap::extend
with a size smaller than twousize
s (e.g., 16 onx86_64
), the size was erroneously rounded up to the minimum size, which could result in an out-of-bounds write.Heap::extend
on an empty heap tried to construct a heap starting at address 0, which is also an out-of-bounds write.Heap::new
(or a similar constructor) with a heap size that is smaller than twousize
s. This was treated as an empty heap as well.Heap::extend
on a heap whose size is not a multiple of the size of twousize
s resulted in unaligned writes. It also left the heap in an unexpected state, which might lead to subsequent issues. We did not find a way to exploit this undefined behavior yet (apart from DoS on platforms that fault on unaligned writes).Patches
Has the problem been patched? What versions should users upgrade to?
We published a patch in version
0.10.2
and recommend all users to upgrade to it. This patch release includes the following changes:2 * size_of::<usize>
and3 * size_of::<usize>
.extend
method now panics when trying to extend an unitialized heap.size_of::<usize>() * 2
are now buffered internally and not added to the list directly. The buffered region will be merged with futureextend
calls.size()
method now returns the usable size of the heap, which might be slightly smaller than thetop() - bottom()
difference because of alignment constraints.Workarounds
Is there a way for users to fix or remediate the vulnerability without upgrading?
To avoid this issue, ensure that the heap is only initialized with a size larger than
3 * size_of::<usize>
and that theHeap::extend
method is only called with sizes larger than2 * size_of::<usize>()
. Also, ensure that the total heap size is (and stays) a multiple of2 * size_of::<usize>()
.For more information
If you have any questions or comments about this advisory:
Acknowledgements
This issue was responsibly reported by Evan Richter at ForAllSecure and found with Mayhem and cargo fuzz.