From mboxrd@z Thu Jan 1 00:00:00 1970 From: ggm@algebras.org (George Michaelson) Date: Tue, 6 Feb 2018 09:52:42 +1000 Subject: [TUHS] Windows roots and Unix influence (was Re: Happy birthday, Ken Thompson!) In-Reply-To: References: <20180204091435.GA22841@indra.papnet.eu> <00d001d39ddc$a069a380$e13cea80$@ronnatalie.com> <0B2B7C84-12D8-49A6-BAA3-BD434823D41D@cheswick.com> <013001d39ecc$614dc500$23e94f00$@ronnatalie.com> <5a838d87-4acb-e2f4-f09e-458d6c5eae49@kilonet.net> Message-ID: My understanding is that FAA rated 'ethernet' switches for aircraft are modified ASICs which do TDM. You program in slot reservations, and this is used to give bounded-time delivery guarantees for flight critical stuff. Cheaper to make by modifying existing stuff, the physical and link layer code is almost identical, its just put into a timeslot regime. Its fully isochronous for the flight control and its best effort delivery for the entertainment. So, when people talk about the entertainment being 'isolated' its not opto-isolated and its not airgap discrete switch isolated: the TDM bitfield excludes the customers from talking in the timeslots for the flight control logic. I think the imputed hack, was to "see" the flight control sequences because you can probably fake out read-side, and get passive view of them if you can make the ports broadcast-receive. I am really unconvinced anyone succeeded in write-mode into this model. Not that it couldn't happen and not that I believe I'm smarted than bad people: I just thing the descriptions have the quality of urban myth right now. On Tue, Feb 6, 2018 at 9:36 AM, Arthur Krewat wrote: > Sorry, Thunderbird strikes again. I highlighted the text to include, and it > didn't put the correct address on the quote. > > My apologies. > > > On 2/5/2018 6:28 PM, Dave Horsfall wrote: >> >> On Mon, 5 Feb 2018, Arthur Krewat wrote: >> >>> On 2/5/2018 4:57 PM, Ron Natalie wrote: >>>> >>>> There's certainly been demonstrations of vehicles being taken over via >>>> the entertainment system; why the stereo needs to talk to the engine >>>> computer I'll never know... I know, wind up the volume the faster you go >>>> etc, but surely it ought to be one-way? >> >> >> Umm, I wrote that, not Ron... >> >