On Wed, 2 Nov 2016, Ondrej Kozina wrote:

> on crypt_message() key set code path while loading new key
> replacing the old one in case tfm->setkey() fails
> we would end having invalid key in crypto layer but
> DM_CRYPT_KEY_VALID flag set.
> 
> Signed-off-by: Ondrej Kozina <[email protected]>

Reviewed-by: Mikulas Patocka <[email protected]>

> ---
>  drivers/md/dm-crypt.c | 7 +++++--
>  1 file changed, 5 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/md/dm-crypt.c b/drivers/md/dm-crypt.c
> index a276883..0aedd0e 100644
> --- a/drivers/md/dm-crypt.c
> +++ b/drivers/md/dm-crypt.c
> @@ -1503,12 +1503,15 @@ static int crypt_set_key(struct crypt_config *cc, 
> char *key)
>       if (!cc->key_size && strcmp(key, "-"))
>               goto out;
>  
> +     /* clear the flag since following operations may invalidate previously 
> valid key */
> +     clear_bit(DM_CRYPT_KEY_VALID, &cc->flags);
> +
>       if (cc->key_size && crypt_decode_key(cc->key, key, cc->key_size) < 0)
>               goto out;
>  
> -     set_bit(DM_CRYPT_KEY_VALID, &cc->flags);
> -
>       r = crypt_setkey_allcpus(cc);
> +     if (!r)
> +             set_bit(DM_CRYPT_KEY_VALID, &cc->flags);
>  
>  out:
>       /* Hex key string not needed after here, so wipe it. */
> -- 
> 2.7.4
> 

--
dm-devel mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/dm-devel

Reply via email to