Hello everyone, I am new to this mailing list, and to Genode generally; I will probably be doing most of my experiments in the near future with a Raspberry Pi, and I'm likely to be fairly quiet about it, but I haven't actually fired anything up yet.
My background is with Linux, and I anticipate a bit of discomfort as I try to adapt to the Genode Way...
My interest is largely in Amateur Radio; The Linux community has a long history of supporting Amateur Radio with protocol stacks (Such as AX.25) and both hardware and virtual Terminal Node Controllers in the kernel. There are also userland implimentations. Some of those run on Windows... some are open source... In other words, lots of code is out there... In various levels of loyalty to the Unix philosophy...
Putting it in the kernel may have never made sense... And the "Genode Way" has value in organizing the complexity of the AX.25 networking stack, similar to how it separates the organization of the Ethernet stack... How much, if any, actually belongs in Genode and/or Sculpt OS is beyond me... It all probably deserves to be an installable package, rather than a core package... But it probably also benefits from Capabilities, as it will interface with hardware, so probably more than just an untrusted Linux instance...
Fldigi also seems like it could benefit from being broken apart into smaller components, with a stronger capability based trust model... Or not, I really don't know...
Anyway... I'm new... And I'm still trying to wrap my head around the "Right" way to think about these things... Are there any other hams out there interested in bouncing around ideas?
73, Tyrell Jentink, KD7TKJ
Hello Tyrell,
On Tuesday, January 29, 2019 5:27:18 PM CET, Tyrell Jentink wrote:
My interest is largely in Amateur Radio; The Linux community has a long history of supporting Amateur Radio with protocol stacks (Such as AX.25) and both hardware and virtual Terminal Node Controllers in the kernel. There are also userland implimentations. Some of those run on Windows... some are open source... In other words, lots of code is out there... In various levels of loyalty to the Unix philosophy...
I have some experience with amateur radios, but it has been many years so my knowledge may no longer be relevant. I think using Genode for something simple like PSK31 or unnumbered AX.25 could be a fun project, but sending arbitrary IP over AX.25 would be quite complicated. We would not want to increase the complexity of the current networking stacks with any alternatives to ethernet, so you would need to provide everything from AX.25 and up yourself.
Putting it in the kernel may have never made sense... And the "Genode Way" has value in organizing the complexity of the AX.25 networking stack, similar to how it separates the organization of the Ethernet stack... How much, if any, actually belongs in Genode and/or Sculpt OS is beyond me... It all probably deserves to be an installable package, rather than a core package...
Yes AX.25 can certainly fit within Genode and Sculpt, but I don't how portable the existing AX.25 implementations are.
Fldigi also seems like it could benefit from being broken apart into smaller components, with a stronger capability based trust model... Or not, I really don't know...
I have used Fldigi before, but there is no FLTK GUI support for Genode so Fldigi is not immediately portable. However breaking it into smaller components does sounds interesting, I imagine the transmit and receive signals would carried by audio sessions and routed to different signal processors or out to hardware. If the hardware can be controlled by RS-232 then it shouldn't be too hard to interface with through a terminal session, but again, I'm not sure how things are done with new radios.
Best of luck, Emery
-
Emery Hemingway -
Tyrell Jentink