On Thu, 26 Sep 2019, Douglas Gilbert wrote: > On 2019-09-26 7:57 p.m., Alan Stern wrote: > > PS: The correct term is "residue", not "residual". I know that the > > code sometimes uses the wrong word. > > Digging into my T10 document archive I found this cam3r03.pdf : > > − cam_resid; > The data residual length member contains the difference in twos > complement form of the number of data bytes transferred by the > HA for the SCSI command compared with the number of bytes > requested by the CCB cam_dxfer_len member. This is calculated > by the total number of bytes requested to be transferred by the > CCB minus the actual number of bytes transferred by the HA. > > CAM is a now withdrawn T10 standard from the 1990s that was influential > at the time. FreeBSD's SCSI subsystem is (still) based on CAM. I was going by my old copy of X3T9.2 Project 375D Rev 10L, Working Draft, "Information technology - Small Computer System Interface - 2", from back in 1996. Yes, it's thoroughly out of date, but you would think that the nomenclature would still be accurate. At any rate, it includes 24 instances of the word "residue" and no instances of "residual". On the other hand, my copy of X3T10/792D Rev 12b, draft proposed, "Information technology - SCSI-2 Common access method transport and SCSI interface module", dated 1995, contains 24 instances of "residual" and only 2 instances of "residue". So I guess this was never defined precisely. > For a more recent standard/draft there is fcp5r00.pdf that uses the > term "residual value" when defining its fcp_resid. > > The only reference to the term "residue" that I found is in CAM: an > optional message: IGNORE WIDE RESIDUE . > > So I would leave the naming up to the patch author. It is pretty > clear what is being referred to in either case. True enough. Alan Stern > Doug Gilbert > > > P.S. I prefer "residual" because it is more flexible being both > an adjective and a noun. > [Ref: https://www.lexico.com/en/definition/residual]
