This message is from the T13 list server.
Found! And indeed N was 8. Thanks again to the so many of you who volunteered offline to help me hunt, and especially to the one of you who hunted successfully. Courtesy your work, now on the web we have: --- http://ide-byte-counting.blog-city.com/index.cfm?d=7&m=1&y=2004 ATA/PI PIO can count bytes, ATA/PI DMA (i.e. SWDMA/ MWDMA/ UDMA) can't, says Pat LaVarre. A guru among the folk of t13.org has tried to refute Pat in a cascade of eight points. Here now in this blog, we have dedicated a blog entry to each of these points, so that we can work separately on each to reach a coherent resolution. ... --- In review, I think I dispute only the point 6 of 8, the claim: "There can be only 0 or 1 pad bytes in a command's data transfer." I agree 7 of 8 and 8 of 8 logically follow if we accept 6 of 8, and I essentially agree with 1 thru 5 of 8 as written, though there still I quibble over jargon. Interesting to see I could have been heard to be disputing 1 thru 5 and/or the logical necessity of 7/8 of 8 when given 6 of 8. English is slippery. Please let's NOT here discuss the merits of any of these Eight contentions yet: a significant fraction of our subscribers here have no reasonable way to register their lack of interest in that long-running debate. Pat LaVarre P.S. Google indexing/ archive of t13 may be interestingly incomplete. Specifically as yet we get no hits for: http://www.google.com/search?q=%22parallel+ATA%2FATAPI%22+%2216-bit+data%22+PIO+DMA+%22even+number+of+bytes%22
