On Fri, Jan 12, 2018 at 6:10 PM, Colin King <colin.king@xxxxxxxxxxxxx> wrote:
> From: Colin Ian King <colin.king@xxxxxxxxxxxxx>
>
> The current error exit path in function cc_ivgen_init via label
> 'out' free's resources from the drvdata->ivgen_handle context.
> However, drvdata->ivgen_handle has not been assigned to the
> context ivgen_ctx at this point, so the resources are not freed.
> Fix this by setting drvdata->ivgen_handle to ivgen_ctx as early
> as possible so that the clean up error exit return path can free
> the resources.
>
> Detected by CoveritScan, CID#1463795 ("Resource leak")
>
> Signed-off-by: Colin Ian King <colin.king@xxxxxxxxxxxxx>
> ---
> drivers/staging/ccree/cc_ivgen.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/staging/ccree/cc_ivgen.c b/drivers/staging/ccree/cc_ivgen.c
> index 25a3131a93ce..c47f419b277b 100644
> --- a/drivers/staging/ccree/cc_ivgen.c
> +++ b/drivers/staging/ccree/cc_ivgen.c
> @@ -178,6 +178,8 @@ int cc_ivgen_init(struct cc_drvdata *drvdata)
> if (!ivgen_ctx)
> return -ENOMEM;
>
> + drvdata->ivgen_handle = ivgen_ctx;
> +
> /* Allocate pool's header for initial enc. key/IV */
> ivgen_ctx->pool_meta = dma_alloc_coherent(device, CC_IVPOOL_META_SIZE,
> &ivgen_ctx->pool_meta_dma,
> @@ -196,8 +198,6 @@ int cc_ivgen_init(struct cc_drvdata *drvdata)
> goto out;
> }
>
> - drvdata->ivgen_handle = ivgen_ctx;
> -
> return cc_init_iv_sram(drvdata);
>
> out:
> --
> 2.15.1
>
Good catch!
Acked-by: Gilad Ben-Yossef <gilad@xxxxxxxxxxxxx>
Thanks,
Gilad
--
Gilad Ben-Yossef
Chief Coffee Drinker
"If you take a class in large-scale robotics, can you end up in a
situation where the homework eats your dog?"
-- Jean-Baptiste Queru
