[PATCH] malloc: Add size check when moving fastbin->tcache

[Adhemerval Zanella Netto]

On 06/02/25 18:37, Ben Kallus wrote:
> By overwriting a forward link in a fastbin chunk that is subsequently
> moved into the tcache, it's possible to get malloc to return an
> arbitrary address [0].
> 
> When a chunk is fetched from a fastbin, its size is checked against the
> expected chunk size for that fastbin (see malloc.c:3991). This patch
> adds a similar check for chunks being moved from a fastbin to tcache,
> which renders obsolete the exploitation technique described above.
> 
> [0]: https://github.com/shellphish/how2heap/blob/master/glibc_2.39/fastbin_reverse_into_tcache.c
> 
> Signed-off-by: Ben Kallus <benjamin.p.kallus.gr@dartmouth.edu>
> ---
>  malloc/malloc.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/malloc/malloc.c b/malloc/malloc.c
> index 27dfd1eb90..3e8a20ae87 100644
> --- a/malloc/malloc.c
> +++ b/malloc/malloc.c
> @@ -4005,6 +4005,9 @@ _int_malloc (mstate av, size_t bytes)
>  		    {
>  		      if (__glibc_unlikely (misaligned_chunk (tc_victim)))
>  			malloc_printerr ("malloc(): unaligned fastbin chunk detected 3");
> +		      size_t victim_tc_idx = csize2tidx (chunksize (tc_victim));
> +		      if (__builtin_expect (tc_idx != victim_tc_idx, 0))
> +			malloc_printerr ("malloc(): chunk size mismatch in fastbin");

This test looks ok and should be inexpensive.  Use __glibc_unlikely,
as the test above.

>  		      if (SINGLE_THREAD_P)
>  			*fb = REVEAL_PTR (tc_victim->fd);
>  		      else