| Message ID | ZGc7hCaDrnEFG8Lr@gondor.apana.org.au (mailing list archive) |
|---|---|
| State | Not Applicable |
| Headers | show |
| Series | crypto: shash - Allow cloning on algorithms with no init_tfm | expand |
| Context | Check | Description |
|---|---|---|
| netdev/tree_selection | success | Not a local patch |
On Fri, 19 May 2023 at 11:04, Herbert Xu <herbert@gondor.apana.org.au> wrote: > > On Fri, May 19, 2023 at 10:54:11AM +0200, Ard Biesheuvel wrote: > > > > Does this imply that the cmac-aes-ce and cmac-aes-neon implementations > > for arm64 need a similar treatment? > > Good catch. Since these don't have init functions we can deal > with them at a higher level: > > ---8<--- > Some shash algorithms are so simple that they don't have an init_tfm > function. These can be cloned trivially. Check this before failing > in crypto_clone_shash. > OK. So IIUC, cloning a keyless hash just shares the TFM and bumps the refcount, but here we must actually allocate a new TFM referring to the same algo, and this new TFM needs its key to be set before use, as it doesn't inherit it from the clonee, right? And this works in the same way as cloning an instance of the generic HMAC template, as this will just clone the inner shash too, and will also leave the key unset. If so, Acked-by: Ard Biesheuvel <ardb@kernel.org> If not, could you explain it to me again? :-) > Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> > > diff --git a/crypto/shash.c b/crypto/shash.c > index 717b42df3495..1fadb6b59bdc 100644 > --- a/crypto/shash.c > +++ b/crypto/shash.c > @@ -597,7 +597,7 @@ struct crypto_shash *crypto_clone_shash(struct crypto_shash *hash) > return hash; > } > > - if (!alg->clone_tfm) > + if (!alg->clone_tfm && (alg->init_tfm || alg->base.cra_init)) > return ERR_PTR(-ENOSYS); > > nhash = crypto_clone_tfm(&crypto_shash_type, tfm); > @@ -606,10 +606,12 @@ struct crypto_shash *crypto_clone_shash(struct crypto_shash *hash) > > nhash->descsize = hash->descsize; > > - err = alg->clone_tfm(nhash, hash); > - if (err) { > - crypto_free_shash(nhash); > - return ERR_PTR(err); > + if (alg->clone_tfm) { > + err = alg->clone_tfm(nhash, hash); > + if (err) { > + crypto_free_shash(nhash); > + return ERR_PTR(err); > + } > } > > return nhash; > -- > Email: Herbert Xu <herbert@gondor.apana.org.au> > Home Page: http://gondor.apana.org.au/~herbert/ > PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
On Fri, May 19, 2023 at 11:31:30AM +0200, Ard Biesheuvel wrote: > > OK. So IIUC, cloning a keyless hash just shares the TFM and bumps the > refcount, but here we must actually allocate a new TFM referring to > the same algo, and this new TFM needs its key to be set before use, as > it doesn't inherit it from the clonee, right? And this works in the > same way as cloning an instance of the generic HMAC template, as this > will just clone the inner shash too, and will also leave the key > unset. Yes that's pretty much it. Cloning a tfm is basically exactly the same as allocating a tfm, except that instead of going through the init_tfm code-path it executes clone_tfm instead (thus allowing any internal data structures to either be shared or allocated with GFP_ATOMIC). The key will be unset just like a freshly allocated tfm. > If so, > > Acked-by: Ard Biesheuvel <ardb@kernel.org> Thanks,
diff --git a/crypto/shash.c b/crypto/shash.c index 717b42df3495..1fadb6b59bdc 100644 --- a/crypto/shash.c +++ b/crypto/shash.c @@ -597,7 +597,7 @@ struct crypto_shash *crypto_clone_shash(struct crypto_shash *hash) return hash; } - if (!alg->clone_tfm) + if (!alg->clone_tfm && (alg->init_tfm || alg->base.cra_init)) return ERR_PTR(-ENOSYS); nhash = crypto_clone_tfm(&crypto_shash_type, tfm); @@ -606,10 +606,12 @@ struct crypto_shash *crypto_clone_shash(struct crypto_shash *hash) nhash->descsize = hash->descsize; - err = alg->clone_tfm(nhash, hash); - if (err) { - crypto_free_shash(nhash); - return ERR_PTR(err); + if (alg->clone_tfm) { + err = alg->clone_tfm(nhash, hash); + if (err) { + crypto_free_shash(nhash); + return ERR_PTR(err); + } } return nhash;