On Thu, Sep 16, 2021 at 08:39:54AM +0100, Christoph Hellwig wrote: > On Wed, Sep 15, 2021 at 10:50:13AM -0700, Eric Biggers wrote: > > Using "blk_crypto_fallback_*" for all these variables results in some pretty > > long names, e.g. "blk_crypto_fallback_crypt_ctx_cache" and > > "blk_crypto_fallback_num_prealloc_crypt_ctxs". This proposal gives the best of > > both worlds; the names are properly "namespaced" but there is also a shortcut to > > refer to them (struct blk_crypto_fallback *fallback = &blk_crypto_fallback). > > I'd just drop the second crypt in those. > > > If this is going to be controversial I can just drop this patch, but I was > > hoping there would be a way to make things more consistent. > > I personally detest that pattern. Not sure if that counts as > controversial or even matters :) The names are still pretty long even with the second "crypt" dropped from those two. How about just "fallback_*"? It might be clear enough from the context. Anyway, I've dropped this patch from the series for now. This can be done later we can agree on which approach to take. - Eric