On Thu, May 14, 2015 at 09:22:01AM +0000, Gujulan Elango, Hari Prasath (H.) wrote:
> The return value of copy_to_user() isn't checked for failure.Hence
> return -EFAULT if it fails.
>
> Signed-off-by: Hari Prasath Gujulan Elango <hgujulan@xxxxxxxxxxx>
> ---
> drivers/staging/lustre/lustre/lov/lov_pack.c | 7 ++++---
> 1 file changed, 4 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/staging/lustre/lustre/lov/lov_pack.c b/drivers/staging/lustre/lustre/lov/lov_pack.c
> index 5356d53..aa3d6de 100644
> --- a/drivers/staging/lustre/lustre/lov/lov_pack.c
> +++ b/drivers/staging/lustre/lustre/lov/lov_pack.c
> @@ -448,9 +448,10 @@ int lov_getstripe(struct obd_export *exp, struct lov_stripe_md *lsm,
> (lum.lmm_stripe_count < lsm->lsm_stripe_count)) {
> /* Return right size of stripe to user */
> lum.lmm_stripe_count = lsm->lsm_stripe_count;
> - rc = copy_to_user(lump, &lum, lum_size);
> - rc = -EOVERFLOW;
> - goto out_set;
> + if (copy_to_user(lump, &lum, lum_size)) {
> + rc = -EFAULT;
> + goto out_set;
> + }
I'm not sure this is right, and I don't think we should take it without
some lustre people signing off.
The original code looks deliberate, like the error happened earlier and
if the copy_to_user() succeed that's fine because it gives them a hint
what went wrong, but we want to return an error -EOVERFLOW regardless of
if the copy_to_user() works or not.
regards,
dan carpenter
_______________________________________________
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxx
http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
