On Wed, Jun 29, 2016 at 08:26:43AM -0700, Andy Lutomirski wrote: > > Two reasons: > > 1. Code like tcp md5 would be simpler if it could pass a scatterlist > to hash the skb but use a virtual address for the header.
True. But I bet we can make it simpler in other ways without creating special interfaces for it. Look at how we do IPsec encryption/hashing, this is what TCP md5 should look like. But nobody in their right mind would bother doing this because this is dead code. > 2. The actual calling sequence and the amount of indirection is much > less for shash, so hashing short buffer is probably *much* faster. Really? Have you measured the speed difference between the ahash and shash interfaces? Are you sure that this would matter when compared against the speed of hashing a single MD5 block? Cheers, -- Email: Herbert Xu <herb...@gondor.apana.org.au> Home Page: http://gondor.apana.org.au/~herbert/ PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt