On Sat 02 Jul 2016 02:36:38 PM CEST, Max Reitz wrote: >> This was the only exceptional module init function that does >> something else than a simple list of bdrv_register() calls, in all >> the block drivers. > > This sounds like this patch specifically wants to drop the check from > bdrv_quorum_init(). > > I think keeping the check would be better; while we can probably > assume that SHA256 is supported if CONFIG_GNUTLS_HASH is defined, it > would still technically be better to actually check for that support.
Right, but SHA256 was added to GNUTLS almost 10 years ago, so I don't think that adding such check has any practical effect. Berto