Re: ahash alg that does not implement import/export failed to register

2015-12-18 Thread Herbert Xu
On Fri, Dec 18, 2015 at 05:31:34PM -0800, Tim Chen wrote: > Herbert, > > There are some ahash algorithms like x86's sha1-mb and > ghash that failed to register because of the newly added > check of non-zero statesize from commit 8996eafd. But > since there are algorithms that do not implement an

ahash alg that does not implement import/export failed to register

2015-12-18 Thread Tim Chen
Herbert, There are some ahash algorithms like x86's sha1-mb and ghash that failed to register because of the newly added check of non-zero statesize from commit 8996eafd. But since there are algorithms that do not implement an import or export, there is no state required for them. Wonder if the

Re: WARNING in crypto_wait_for_test

2015-12-18 Thread Herbert Xu
On Fri, Dec 18, 2015 at 04:49:17PM +0800, Herbert Xu wrote: > > OK turns out I was looking at the wrong WARN_ON line. > > This particular bug is a legacy of the geniv construct we have > for ablkcipher. > > It should disappear once we remove geniv which I will do as part > of the skcipher conver

Re: WARNING in crypto_wait_for_test

2015-12-18 Thread Herbert Xu
On Mon, Dec 14, 2015 at 11:45:02PM +0100, Stephan Mueller wrote: > > But with the given code, when you remove the bogus mask setting (which in > turn > leaves it as 0), the code works flawless. OK turns out I was looking at the wrong WARN_ON line. This particular bug is a legacy of the geniv co