GSoC 2018 (x86_64, pc386, SMP improvements, rump kernels)

Amaan Cheval amaan.cheval at
Sat Mar 10 17:02:35 UTC 2018


I've introduced myself in the past, but just to reiterate; I'm a final year
Bachelor's in Engineering student at Thakur College of Engineering, in
Mumbai. I work part-time on an Intel x86 emulator, written primarily in C
and JavaScript (that's right :P) - we also have a JIT compiler for
x86->WebAssembly. I've had an interest in reverse engineering, compiler
theory, and operating systems, and Linux specifically, for quite a while

Anyway, given that the time to work on our GSoC proposals is nigh, I
figured I'd solicit more information for the projects that I'm interested

- x86_64 BSP[1]

This seems like a fairly large undertaking, with potential overlaps with
the next project (pc386 improvement). Does anyone know what subset of the
listed items may be considered the "core", for completion within the GSoC
period? Is there anything I can look into right now to (1) better gauge the
kind of work involved in this project and (2) to better estimate how the
work will need to be split up in terms of time?

I'm looking into the BSP porting guide, and just the source code for BSPs
and no_cpu in specific to get a better feel for how RTEMS structures these
things at the moment.

- Improve PC386 BSP[2]

Seems to contain a fair bit of information already - I'm hoping to look
into #2183 as time permits to figure out better what some of the work on
this project will be like. Are there any other potential next steps I
should consider?

- Improve RTEMS SMP[3]

What kinds of improvements to SMP are we considering? As noted earlier, SMP
support on i386 is lagging. Is there any interest in bringing that up to
par with the other architectures?

- Rump kernel hypercall interface[4]

Given that this project hasn't been converted to a ticket, I imagine that
this is lower in priority. Let me know if that's not the case, though?


I'm hoping to get a proposal or two prepared within the coming week to give
the community a fair amount of time to critique it and propose changes /
additions to it before the deadline. I'd like to figure out what's
realistic to propose, and what's required most by RTEMS, so any information
would be useful.



More information about the devel mailing list