On Mon, Jan 29, 2024 at 08:31:51AM +0100, Christoph Hellwig wrote:
> Check the 32-bit limits using sizeof instead of cpp ifdefs so that we
> can get rid of BITS_PER_LONG.
>
> Signed-off-by: Christoph Hellwig <hch@xxxxxx>
> ---
> repair/bmap.c | 23 +++++++++++++++--------
> repair/bmap.h | 13 -------------
> 2 files changed, 15 insertions(+), 21 deletions(-)
>
> diff --git a/repair/bmap.c b/repair/bmap.c
> index cd1a8b07b..d1b2faaec 100644
> --- a/repair/bmap.c
> +++ b/repair/bmap.c
> @@ -22,6 +22,15 @@
> pthread_key_t dblkmap_key;
> pthread_key_t ablkmap_key;
>
> +/*
> + * For 32 bit platforms, we are limited to extent arrays of 2^31 bytes, which
> + * limits the number of extents in an inode we can check. If we don't limit the
> + * valid range, we can overflow the BLKMAP_SIZE() calculation and allocate less
> + * memory than we think we needed, and hence walk off the end of the array and
> + * corrupt memory.
> + */
> +#define BLKMAP_NEXTS32_MAX ((INT_MAX / sizeof(bmap_ext_t)) - 1)
> +
> blkmap_t *
> blkmap_alloc(
> xfs_extnum_t nex,
> @@ -35,8 +44,7 @@ blkmap_alloc(
> if (nex < 1)
> nex = 1;
>
> -#if (BITS_PER_LONG == 32) /* on 64-bit platforms this is never true */
> - if (nex > BLKMAP_NEXTS_MAX) {
> + if (sizeof(long) == 32 && nex > BLKMAP_NEXTS32_MAX) {
That's a really, really big long. sizeof(long) = 4, perhaps?
-Dave.
--
Dave Chinner
david@xxxxxxxxxxxxx
