From mboxrd@z Thu Jan 1 00:00:00 1970 From: martin.petersen@oracle.com (Martin K. Petersen) Date: Wed, 20 Sep 2017 16:01:58 -0400 Subject: [PATCH] nvme: set physical block size to value discovered in Identify Namespace In-Reply-To: <20170920190752.GC1379@localhost.localdomain> (Keith Busch's message of "Wed, 20 Sep 2017 15:07:52 -0400") References: <20170920170605.42161-1-andrzej.jakowski@intel.com> <20170920174842.GB1379@localhost.localdomain> <20170920174814.GA2556@infradead.org> <20170920190752.GC1379@localhost.localdomain> Message-ID: Keith, > I don't think it's about "reasonable" performance; it's about getting > extra relative performance. What else can the best performing LBAF > indicate other than the device's preferred access alignment/granularity? > The spec provides this hint, so it's not really a guess, but maybe > there's a better way to make use of it instead of considering it to be > the physical block size? io_opt? It should really report io_min and not physical_block_size. pbs is a special case for spinning media. I agree it's a bit weird to walk the format list as a heuristic. Would be nicer to have a smallest preferred I/O size in the protocol. > The first and last are the exact same NVMe format, but they're > reported differently. Yeah, we should set pbs/io_min as well. -- Martin K. Petersen Oracle Linux Engineering