Hello Thomas My experience isĀ not only way We pick the processor then the tools ( RTOS,JTAG. Compiler/ IDE if the stuff doesn't work or if vendor's don't support properly we don't use their product( I determine this usually). Like you I expect answers š¤
So pick chip, OS then architecture then FINALLY from Hardware System's engineer document IO is assigned and documentedĀ in table and layed out in boardĀ Low level Coder starts pin muxing codingĀ and Interrupt and drivers. Application is divided in to tasks or threads by software architect at same timeĀ I do board support so JTAG for me is something I insist on. ( New hardware). I like IDE one that allows RTOS awareness Greenhills has one of the Best I have used CCS maybe a dozen time's projects used their DSP I am familiar with CCS but have worked where you build by command line and make files. These are huge fortune 200 companies so money is no problems for tools. One big Medical company my title was tool's software. I played around. I'm a mercenary hired gun. If I don't like their approach andĀ they lie or waste my time I tactfully explain bad ideas. A few times I know there designĀ approach will fail I mention to them at beginning thenĀ I do what I'm told. then it fails they get angry š¤£Ā What's a$$ backwards to you is ( your last email) is normal. For what you do linux makes sense buy hardware don't need JTAG or CCS. Ā here in lies the flaw when you use free software the approach to IPC may be less than ideal but getting it fixed no one cares with an RTOS it costs $50K you get support and attention. or Else In open source you get ignored and runaround in circles and your support is this group which isn't support. This approach is for hobbyists with big dreams and no budget. I will say many jobs are using Linux I see them daily doesn't mean I have to. There are free alternative but why bother if Linux works for you and this IPC is flawed you have my attention. This is another part of my job. Improve RTOS supplied BSP driver's add new ones and find the crap that's not hard real time flawed.Ā An example Green Hill's Integrity BSP for omapl138 had examples code of IPC between ARM and DSP given to customer as I mentioned the PhD design fast queues didn't use it improved it made it fast like what your doing. Now TI had driver's and code and .h files and Green Hills had their own .c .h files doing same thing. I think TI did custom first stage loader not sure.It did what about and mlo does on Linux but much better. There's no one way to design thing's perhaps you can share what's flawed in Rpmsg in details if that's true so you don't get sued call it RP Don't Worky nice.š¤£š¤ and bring attention to it's limitations. You sound like smart Guy maybe it's true. In closing these dev board's or hobby boards serve a purpose. The EVM ones the HW design was very good but I have issues with long term support I'm not going to mention in this group ( poor practice) the early .org boards by Gerald very nice he's gone. The open source community won't support or don't care about JTAG at least in this group but TI does.You really have no choices in this group beyond Linux.Ā Barebones and RTOS on these boards isn't a concern on this group so I'm wasting my time but it's sometimes good reading. BTW good luck even if you demonstrate this IPCĀ isn't ideal because that's not the Agenda here.Ā It's marketing getting student to Love Linux and this board. Unfortunately that's not the Real world not ever be used in Missles,Satellites or air. Boeing and Lockheed have the best engineering in United States. Both of these company uses TI DSP chips they don't use Linux my friend. Linux is good but not a fix all for everything and getting ignored is something big company won't tolerate just the hobbyistsĀ they have no choices the dire hard open source will save the world fanatic are attracted to all the glamour like sheep. Then one day they show up in this group posting it don't Worky help me LoL š Great group it's focus is easy building of mainstream linux and keep buying more boards. What's the next board coming? Was nice when Gerald was onboard nice guy very talented and always had time for questions you see the bestĀ engineer don't feel threatened by questions and are willing to help there's some exception some of the best are over worked micro managed by manager's who have no technical skill and want to advance in company. Good engineer's leave those environments. Mark Sent from Yahoo Mail on Android On Mon, Apr 19, 2021 at 12:46 AM, TJF<jeli.freih...@gmail.com> wrote: Hi Mark, let me first point out that this discussion isn't in my mother tongue. Sometimes I read your statements several times still not understanding what exactly you mean. Mark Lazarewicz schrieb am Sonntag, 18. April 2021 um 16:49:22 UTC+2: who you want to believe Texas Instrumentr or TJF? I believe none of them. The hardware is reality. I make experiments (ask the hardware) to find out what's really correct. There isn't one great TI company. Hardware engineers made a great AM335x CPU design. It's not bug free, but well usable. And software engineers do everything to block the hardware benefits. So the question should be: blinking a led, do you want to follow an CCS example (changing with each new kernel version) or an libpruio example published years before (2014, still working)? Do you want to learn from coders making small examples only, or are you interested in hands-on experience form people who did already fill up the IRam of both PRUSS? You're lucky, I'm waiting for feedback before starting my next project. So I've time to waste of good mojo. -- For more options, visit http://beagleboard.org/discuss --- You received this message because you are subscribed to the Google Groups "BeagleBoard" group. To unsubscribe from this group and stop receiving emails from it, send an email to beagleboard+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/beagleboard/0646b780-3ac1-41fe-8c7d-3ee2b2156b66n%40googlegroups.com. -- For more options, visit http://beagleboard.org/discuss --- You received this message because you are subscribed to the Google Groups "BeagleBoard" group. To unsubscribe from this group and stop receiving emails from it, send an email to beagleboard+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/beagleboard/476722279.2100658.1618849009331%40mail.yahoo.com.
