Re: [PATCH] iscsi-target: Make buf param ofiscsit_do_crypto_hash_buf() const void *

[Nicholas A. Bellinger]

On Fri, 2013-05-03 at 23:22 +0200, Geert Uytterhoeven wrote:
> On Fri, May 3, 2013 at 11:15 PM, Geert Uytterhoeven
> <geert@xxxxxxxxxxxxxx> wrote:
> > --- a/drivers/target/iscsi/iscsi_target.c
> > +++ b/drivers/target/iscsi/iscsi_target.c
> 
> > @@ -3585,9 +3575,8 @@ static int iscsit_send_reject(
> >         }
> >
> >         if (conn->conn_ops->DataDigest) {
> > -               iscsit_do_crypto_hash_buf(&conn->conn_tx_hash,
> > -                               (unsigned char *)cmd->buf_ptr, ISCSI_HDR_LEN,
> > -                               0, NULL, (u8 *)&cmd->data_crc);
> > +               iscsit_do_crypto_hash_buf(&conn->conn_tx_hash, cmd->buf_ptr,
> > +                               ISCSI_HDR_LEN, 0, NULL, (u8 *)&cmd->data_crc);
> 
> BTW, there are serious issues with casting pointers to u32s to pointers to u8s
> for the last two output parameters of iscsit_do_crypto_hash_buf() (pad_bytes
> and data_crc):
>   1. It's not endian-safe,
>   2. In many cases, the original u32 was not initialized at all, so the other 24
>       bits contain garbage. This will cause 32-bit comparisons like
> 
>         if (checksum != data_crc)
> 
>      to fail in unpredictable ways.
> 

Ugh, I can see how that would be problematic for BE..

Changing these to use u8 data_crc[ISCSI_DIGEST_SIZE] and
pad_bytes[ISCSI_PAD_LEN] now, and will post a patch soon.

Thanks again,

--nab


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/