Le Tue, Mar 02, 2021 at 04:34:46PM +0000, Colin King a écrit :
> From: Colin Ian King <colin.k...@canonical.com>
>
> In the case where the dma_iv mapping fails, the return error path leaks
> the memory allocated to object d. Fix this by adding a new error return
> label and jumping to this to ensure d is free'd before the return.
>
> Addresses-Coverity: ("Resource leak")
> Fixes: ac2614d721de ("crypto: sun8i-ss - Add support for the PRNG")
> Signed-off-by: Colin Ian King <colin.k...@canonical.com>
> ---
> drivers/crypto/allwinner/sun8i-ss/sun8i-ss-prng.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/crypto/allwinner/sun8i-ss/sun8i-ss-prng.c
> b/drivers/crypto/allwinner/sun8i-ss/sun8i-ss-prng.c
> index 08a1473b2145..3191527928e4 100644
> --- a/drivers/crypto/allwinner/sun8i-ss/sun8i-ss-prng.c
> +++ b/drivers/crypto/allwinner/sun8i-ss/sun8i-ss-prng.c
> @@ -103,7 +103,8 @@ int sun8i_ss_prng_generate(struct crypto_rng *tfm, const
> u8 *src,
> dma_iv = dma_map_single(ss->dev, ctx->seed, ctx->slen, DMA_TO_DEVICE);
> if (dma_mapping_error(ss->dev, dma_iv)) {
> dev_err(ss->dev, "Cannot DMA MAP IV\n");
> - return -EFAULT;
> + err = -EFAULT;
> + goto err_free;
> }
>
> dma_dst = dma_map_single(ss->dev, d, todo, DMA_FROM_DEVICE);
> @@ -167,6 +168,7 @@ int sun8i_ss_prng_generate(struct crypto_rng *tfm, const
> u8 *src,
> memcpy(ctx->seed, d + dlen, ctx->slen);
> }
> memzero_explicit(d, todo);
> +err_free:
> kfree(d);
>
> return err;
> --
> 2.30.0
>
Hello
Tested-by: Corentin Labbe <clabbe.montj...@gmail.com>
Acked-by: Corentin Labbe <clabbe.montj...@gmail.com>
Thanks
