On Thu, Feb 27, 2025 at 06:15:00PM +0800, Herbert Xu wrote: > This adds request chaining and virtual address support to the > acomp interface. > > It is identical to the ahash interface, except that a new flag > CRYPTO_ACOMP_REQ_NONDMA has been added to indicate that the > virtual addresses are not suitable for DMA. This is because As with ahash, the request chaining support is a huge mistake. The problem is much better solved by adding methods that explicitly take an array of data buffers (i.e., batch_compress and batch_decompress in the case of compression). "Request chaining" is unnecessarily complex and creates a lot of edge cases and ambiguities. E.g. it would make it ambiguous whether every function operates on a single request or on the whole list of requests. It also doesn't even work in the expected way, as submitting a list of requests won't be equivalent to submitting them individually. And as usual for your submissions it's also undocumented and has no tests. - Eric
