CRYPTO_TFM_REQ_MASK and CRYPTO_TFM_RES_MASK

CRYPTO_TFM_REQ_MASK and CRYPTO_TFM_RES_MASK

[Dimitrios Siganos]

Hi,

Could someone explain what the masks CRYPTO_TFM_REQ_MASK and 
CRYPTO_TFM_RES_MASK do and why they must be manipulated before and after 
crypto_cipher_setkey(...)?

Here is an example use (from crypto_pcbc_setkey):
    crypto_cipher_clear_flags(child, CRYPTO_TFM_REQ_MASK);
    crypto_cipher_set_flags(child, crypto_tfm_get_flags(parent) &
                CRYPTO_TFM_REQ_MASK);
    err = crypto_cipher_setkey(child, key, keylen);
    crypto_tfm_set_flags(parent, crypto_cipher_get_flags(child) &
                 CRYPTO_TFM_RES_MASK);

It would be very useful if someone added some comments to these 
definitions (found in linux/crypto.h):

/*
 * Transform masks and values (for crt_flags).
 */
#define CRYPTO_TFM_REQ_MASK        0x000fff00
#define CRYPTO_TFM_RES_MASK        0xfff00000

#define CRYPTO_TFM_REQ_WEAK_KEY        0x00000100
#define CRYPTO_TFM_REQ_MAY_SLEEP    0x00000200
#define CRYPTO_TFM_REQ_MAY_BACKLOG    0x00000400
#define CRYPTO_TFM_RES_WEAK_KEY        0x00100000
#define CRYPTO_TFM_RES_BAD_KEY_LEN       0x00200000
#define CRYPTO_TFM_RES_BAD_KEY_SCHED     0x00400000
#define CRYPTO_TFM_RES_BAD_BLOCK_LEN     0x00800000
#define CRYPTO_TFM_RES_BAD_FLAGS     0x01000000

Thanks,
Dimitris

Re: CRYPTO_TFM_REQ_MASK and CRYPTO_TFM_RES_MASK

[Dimitrios Siganos]

After sending the email, I realised what's happening... 
crypto_cipher_setkey seems to be used that way only when a child cipher 
is used by another (parent) cipher as a way to delegate work.

It seems that crypto_cipher_setkey accepts a number of request flags 
CRYPTO_TFM_REQ_* which must be passed to the child object and after 
completion of crypto_cipher_setkey, it returns some results flags 
CRYPTO_TFM_RES_* back which must be passed to the parent object.

So the flag manipulation before and after is because we are using an 
internal cipher object to delegate the work.

Please correct me, if I am wrong.

Dimitris

Dimitrios Siganos wrote:
> Hi,
>
> Could someone explain what the masks CRYPTO_TFM_REQ_MASK and 
> CRYPTO_TFM_RES_MASK do and why they must be manipulated before and 
> after crypto_cipher_setkey(...)?
>
> Here is an example use (from crypto_pcbc_setkey):
>    crypto_cipher_clear_flags(child, CRYPTO_TFM_REQ_MASK);
>    crypto_cipher_set_flags(child, crypto_tfm_get_flags(parent) &
>                CRYPTO_TFM_REQ_MASK);
>    err = crypto_cipher_setkey(child, key, keylen);
>    crypto_tfm_set_flags(parent, crypto_cipher_get_flags(child) &
>                 CRYPTO_TFM_RES_MASK);
>
> It would be very useful if someone added some comments to these 
> definitions (found in linux/crypto.h):
>
> /*
> * Transform masks and values (for crt_flags).
> */
> #define CRYPTO_TFM_REQ_MASK        0x000fff00
> #define CRYPTO_TFM_RES_MASK        0xfff00000
>
> #define CRYPTO_TFM_REQ_WEAK_KEY        0x00000100
> #define CRYPTO_TFM_REQ_MAY_SLEEP    0x00000200
> #define CRYPTO_TFM_REQ_MAY_BACKLOG    0x00000400
> #define CRYPTO_TFM_RES_WEAK_KEY        0x00100000
> #define CRYPTO_TFM_RES_BAD_KEY_LEN       0x00200000
> #define CRYPTO_TFM_RES_BAD_KEY_SCHED     0x00400000
> #define CRYPTO_TFM_RES_BAD_BLOCK_LEN     0x00800000
> #define CRYPTO_TFM_RES_BAD_FLAGS     0x01000000
>
> Thanks,
> Dimitris
> -- 
> To unsubscribe from this list: send the line "unsubscribe 
> linux-crypto" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: CRYPTO_TFM_REQ_MASK and CRYPTO_TFM_RES_MASK

[Herbert Xu]

Dimitrios Siganos <dimitris@siganos.org> wrote:
> After sending the email, I realised what's happening... 
> crypto_cipher_setkey seems to be used that way only when a child cipher 
> is used by another (parent) cipher as a way to delegate work.
> 
> It seems that crypto_cipher_setkey accepts a number of request flags 
> CRYPTO_TFM_REQ_* which must be passed to the child object and after 
> completion of crypto_cipher_setkey, it returns some results flags 
> CRYPTO_TFM_RES_* back which must be passed to the parent object.
> 
> So the flag manipulation before and after is because we are using an 
> internal cipher object to delegate the work.

You're absolutely correct.  Feel free to send patches to add
comments too.

Cheers,
-- 
Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmV>HI~} <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt