| Message ID | 20170814131725.9402-3-clabbe.montjoie@gmail.com (mailing list archive) |
|---|---|
| State | Superseded |
| Delegated to: | Herbert Xu |
| Headers | show |
Hi Corentin, Since I have just sent a patch to add the support of "aead_request" to crypto engine, I am wondering if your proposed change (checking cra_type instead of crypto_tfm_alg_type) and mine are compatible. It looks like they are (assuming we export crypto_aead_type): can you confirm? BR Fabien. >-----Original Message----- >From: linux-crypto-owner@vger.kernel.org [mailto:linux-crypto- >owner@vger.kernel.org] On Behalf Of Corentin Labbe >Sent: lundi 14 août 2017 15:17 >To: herbert@gondor.apana.org.au; davem@davemloft.net >Cc: linux-crypto@vger.kernel.org; linux-kernel@vger.kernel.org; Corentin Labbe ><clabbe.montjoie@gmail.com> >Subject: [PATCH 2/3] crypto: engine - find request type with cra_type > >The current method for finding request type is based on crypto_tfm_alg_type. > >But in case of skcipher, it is the same than ablkcipher. >Using cra_type for this work permits to make the distinction between the two. > >Signed-off-by: Corentin Labbe <clabbe.montjoie@gmail.com> >--- > crypto/crypto_engine.c | 19 ++++++++----------- > 1 file changed, 8 insertions(+), 11 deletions(-) > >diff --git a/crypto/crypto_engine.c b/crypto/crypto_engine.c index >61e7c4e02fd2..74b840749074 100644 >--- a/crypto/crypto_engine.c >+++ b/crypto/crypto_engine.c >@@ -38,7 +38,8 @@ static void crypto_pump_requests(struct crypto_engine >*engine, > struct ablkcipher_request *breq; > unsigned long flags; > bool was_busy = false; >- int ret, rtype; >+ int ret; >+ const struct crypto_type *cratype; > > spin_lock_irqsave(&engine->queue_lock, flags); > >@@ -94,7 +95,7 @@ static void crypto_pump_requests(struct crypto_engine >*engine, > > spin_unlock_irqrestore(&engine->queue_lock, flags); > >- rtype = crypto_tfm_alg_type(engine->cur_req->tfm); >+ cratype = engine->cur_req->tfm->__crt_alg->cra_type; > /* Until here we get the request need to be encrypted successfully */ > if (!was_busy && engine->prepare_crypt_hardware) { > ret = engine->prepare_crypt_hardware(engine); >@@ -104,8 +105,7 @@ static void crypto_pump_requests(struct crypto_engine >*engine, > } > } > >- switch (rtype) { >- case CRYPTO_ALG_TYPE_AHASH: >+ if (cratype == &crypto_ahash_type) { > hreq = ahash_request_cast(engine->cur_req); > if (engine->prepare_hash_request) { > ret = engine->prepare_hash_request(engine, hreq); @@ >-122,7 +122,7 @@ static void crypto_pump_requests(struct crypto_engine >*engine, > goto req_err; > } > return; >- case CRYPTO_ALG_TYPE_ABLKCIPHER: >+ } else if (cratype == &crypto_ablkcipher_type) { > breq = ablkcipher_request_cast(engine->cur_req); > if (engine->prepare_cipher_request) { > ret = engine->prepare_cipher_request(engine, breq); >@@ -139,21 +139,18 @@ static void crypto_pump_requests(struct >crypto_engine *engine, > goto req_err; > } > return; >- default: >+ } else { > dev_err(engine->dev, "failed to prepare request of unknown >type\n"); > return; > } > > req_err: >- switch (rtype) { >- case CRYPTO_ALG_TYPE_AHASH: >+ if (cratype == &crypto_ahash_type) { > hreq = ahash_request_cast(engine->cur_req); > crypto_finalize_hash_request(engine, hreq, ret); >- break; >- case CRYPTO_ALG_TYPE_ABLKCIPHER: >+ } else if (cratype == &crypto_ablkcipher_type) { > breq = ablkcipher_request_cast(engine->cur_req); > crypto_finalize_cipher_request(engine, breq, ret); >- break; > } > return; > >-- >2.13.0
On Tue, Aug 15, 2017 at 07:51:14AM +0000, Fabien DESSENNE wrote: > Hi Corentin, > > Since I have just sent a patch to add the support of "aead_request" to crypto engine, I am wondering if your proposed change (checking cra_type instead of crypto_tfm_alg_type) and mine are compatible. > It looks like they are (assuming we export crypto_aead_type): can you confirm? > BR > > Fabien. > Hello My change is incompatible with yours since I remove a switch(crypto_tfm_alg_type) that you use. Anyway you will need my change:) because you use ablkcipher which is obsolete and you need to convert to skcipher. regards Corentin Labbe
On Mon, Aug 14, 2017 at 03:17:24PM +0200, Corentin Labbe wrote: > The current method for finding request type is based on crypto_tfm_alg_type. > > But in case of skcipher, it is the same than ablkcipher. > Using cra_type for this work permits to make the distinction between the two. > > Signed-off-by: Corentin Labbe <clabbe.montjoie@gmail.com> I think you misunderstood my suggestion. I'm not saying that you should use cra_type to distinguish all crypto types, it should only be used to distinguish ablkcipher from skcipher. Thanks,
On Fri, Aug 18, 2017 at 04:28:37PM +0800, Herbert Xu wrote: > On Mon, Aug 14, 2017 at 03:17:24PM +0200, Corentin Labbe wrote: > > The current method for finding request type is based on crypto_tfm_alg_type. > > > > But in case of skcipher, it is the same than ablkcipher. > > Using cra_type for this work permits to make the distinction between the two. > > > > Signed-off-by: Corentin Labbe <clabbe.montjoie@gmail.com> > > I think you misunderstood my suggestion. I'm not saying that > you should use cra_type to distinguish all crypto types, it should > only be used to distinguish ablkcipher from skcipher. > Will change that in v4 Thanks
diff --git a/crypto/crypto_engine.c b/crypto/crypto_engine.c index 61e7c4e02fd2..74b840749074 100644 --- a/crypto/crypto_engine.c +++ b/crypto/crypto_engine.c @@ -38,7 +38,8 @@ static void crypto_pump_requests(struct crypto_engine *engine, struct ablkcipher_request *breq; unsigned long flags; bool was_busy = false; - int ret, rtype; + int ret; + const struct crypto_type *cratype; spin_lock_irqsave(&engine->queue_lock, flags); @@ -94,7 +95,7 @@ static void crypto_pump_requests(struct crypto_engine *engine, spin_unlock_irqrestore(&engine->queue_lock, flags); - rtype = crypto_tfm_alg_type(engine->cur_req->tfm); + cratype = engine->cur_req->tfm->__crt_alg->cra_type; /* Until here we get the request need to be encrypted successfully */ if (!was_busy && engine->prepare_crypt_hardware) { ret = engine->prepare_crypt_hardware(engine); @@ -104,8 +105,7 @@ static void crypto_pump_requests(struct crypto_engine *engine, } } - switch (rtype) { - case CRYPTO_ALG_TYPE_AHASH: + if (cratype == &crypto_ahash_type) { hreq = ahash_request_cast(engine->cur_req); if (engine->prepare_hash_request) { ret = engine->prepare_hash_request(engine, hreq); @@ -122,7 +122,7 @@ static void crypto_pump_requests(struct crypto_engine *engine, goto req_err; } return; - case CRYPTO_ALG_TYPE_ABLKCIPHER: + } else if (cratype == &crypto_ablkcipher_type) { breq = ablkcipher_request_cast(engine->cur_req); if (engine->prepare_cipher_request) { ret = engine->prepare_cipher_request(engine, breq); @@ -139,21 +139,18 @@ static void crypto_pump_requests(struct crypto_engine *engine, goto req_err; } return; - default: + } else { dev_err(engine->dev, "failed to prepare request of unknown type\n"); return; } req_err: - switch (rtype) { - case CRYPTO_ALG_TYPE_AHASH: + if (cratype == &crypto_ahash_type) { hreq = ahash_request_cast(engine->cur_req); crypto_finalize_hash_request(engine, hreq, ret); - break; - case CRYPTO_ALG_TYPE_ABLKCIPHER: + } else if (cratype == &crypto_ablkcipher_type) { breq = ablkcipher_request_cast(engine->cur_req); crypto_finalize_cipher_request(engine, breq, ret); - break; } return;
The current method for finding request type is based on crypto_tfm_alg_type. But in case of skcipher, it is the same than ablkcipher. Using cra_type for this work permits to make the distinction between the two. Signed-off-by: Corentin Labbe <clabbe.montjoie@gmail.com> --- crypto/crypto_engine.c | 19 ++++++++----------- 1 file changed, 8 insertions(+), 11 deletions(-)