From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Disseldorp Date: Wed, 28 Nov 2018 16:28:55 +0000 Subject: Re: [PATCH v3 3/4] target: add device vendor_id configfs attribute Message-Id: <20181128172855.2dfdd7d1@suse.de> List-Id: References: <20181119210636.22979-4-ddiss@suse.de> In-Reply-To: <20181119210636.22979-4-ddiss@suse.de> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: target-devel@vger.kernel.org On Wed, 28 Nov 2018 08:08:30 -0800, Bart Van Assche wrote: > > Just a follow up here. I think it's safer to retain strncpy() in this > > function for the purpose of zero filling. scsi_dump_inquiry() and > > target_stat_lu_vend_show() walk the entire length of the t10_wwn.vendor > > buffer. > > How about adding a comment about above struct t10_wwn that vendor[], model[] > and revision[] in that structure may but do not have to be terminated with a > trailing '\0' and also that unit_serial[] is '\0'-terminated? > > It would make me happy if the size of the character arrays in that structure > would be increased by one and if the target code would be modified such that > these strings are always '\0'-terminated. I'm working on the +1 length increase for those t10_wwn members ATM, but just thought I'd mention that the zeroing of unused bytes is still required due to the scsi_dump_inquiry() + target_stat_lu_vend_show() behaviour. Cheers, David