Hi Alan, Porting to debug devices on eval boards is not something I’m considering. They’re usually either locked down, or have insufficient on-chip resources to run Maven. Then there’s the sheer number of eval boards, all slightly different...
Steve > On 20 Jul 2021, at 11:32, Alan Carvalho de Assis <acas...@gmail.com> wrote: > > Hi Steve, > > Nice work! I'm looking forward to test it. > > Is it possible to use it on MCU-Link existing on some NXP board? Like some > board already supported by NuttX? > > BR, > > Alan > > On Tuesday, July 20, 2021, Steve Woodford <st...@mctavish.co.uk.invalid> > wrote: > >> Thanks for the intro, Dave. >> >> NuttX has moved on a fair bit from the last time I was subscribed to the >> mailing list, between 2010 and 2017. Yahoo groups, anyone? >> >> Anyhow, I’ve been adding FreeRTOS thread-awareness to Maven over the last >> few weeks and decided the next target would be NuttX. Dave seems to think >> this would be a big deal, and tried to encourage me to do something for the >> workshop in August, but I don’t see that happening due to time constraints. >> In any case, a quick intro to Maven... >> >> Maven on its native hardware has been around for a few years now, mostly >> as a hobby project in my retirement, but available for sale via Tindie and >> my website. I’ve since discontinued the original hardware as Maven firmware >> has outgrown the Flash/SRAM on its SoC. I do have an updated hardware >> design in prototype form, but this is not likely to see the light of day - >> I’m not a hardware guy! >> >> When I got wind of NXP’s el-cheapo MCU-Link I decided a Maven port was in >> order, and that it would be a free download. A few people have tried out >> the early versions over the last few months. At least, I think they have - >> I’ve had no negative feedback! >> >> There’s a NuttX-specific ReadMe file in the download describing the small >> additional data structure required for thread awareness. This is not set in >> stone; I’m happy to change it to something more akin to the OpenOCD method >> if that’s the consensus. This would not require you to rebuild Maven >> firmware (unlike with OpenOCD), as there is already support for >> RTOS-specific configuration parameters. In fact that’s the method used for >> FreeRTOS. NuttX has so many CONFIG_* options which change the sizes and >> offsets of the interesting data structures that it looked to be too much of >> a chore. >> >> Anyhow, let me know what you think. >> >> Steve >> >>> On 20 Jul 2021, at 10:02, Dave Marples <d...@marples.net> wrote: >>> >>> Folks, >>> >>> Steve Woodford has just integrated NuttX task awareness into his Maven >> debugger. This initial version runs on the NXP MCU-Link, replacing the >> 'stock' firmware. Fortunately the MCULink is only about $10 and I'm sure >> it'll get ported to anything with a clock tick pretty quickly. There's a >> couple of READMEs in the archive which provide information about what it >> does and how it's been done (including a small NuttX patch to make it >> easier for debuggers to find the TCB....for discussion.) >>> >>> Steve has just joined the mailing list but hasn't introduced himself to >> you hairy lot yet...and clicking on links from newly signed up users is not >> something to do lightly, hence my posting this initial intro; >>> >>> https://www.maverick-embedded.co.uk/downloads/maven-mculink-0.5.tgz >>> >>> I'm sure he'll welcome comments and (limited) brickbats. Hopefully he'll >> be able to show it off at the NuttX Workshop. >>> >>> ...and for those of you _still_ waiting for Debug and Parallel Trace, a >> morsel to whet your appetites; >>> >>> https://cdn.discordapp.com/attachments/614885210395508738/ >> 863565020158754816/1625960319475.jpg >>> >>> Regards >>> >>> DAVE >>> >>> >> >>
