View Single Post
Old 05-19-2016, 08:02 PM   #261
geekmaster
Carpe diem, c'est la vie.
geekmaster ought to be getting tired of karma fortunes by now.geekmaster ought to be getting tired of karma fortunes by now.geekmaster ought to be getting tired of karma fortunes by now.geekmaster ought to be getting tired of karma fortunes by now.geekmaster ought to be getting tired of karma fortunes by now.geekmaster ought to be getting tired of karma fortunes by now.geekmaster ought to be getting tired of karma fortunes by now.geekmaster ought to be getting tired of karma fortunes by now.geekmaster ought to be getting tired of karma fortunes by now.geekmaster ought to be getting tired of karma fortunes by now.geekmaster ought to be getting tired of karma fortunes by now.
 
geekmaster's Avatar
 
Posts: 6,433
Karma: 10773670
Join Date: Nov 2011
Location: Multiverse 6627A
Device: K1 to PW3
Rob Landley was talking about how ancient and unmaintained a lot of that code was (though mostly the parts of linux that wrap around the uClibc) in his toybox roadmap. The main reason he was using it was for nommu (arms without memory management) support. Most linux distros abandoned nommu support aeons ago. The only reason K1 used it is that lab126 (probably) began software development on E-Ink's E-Paper Display (EPD) dev kit, which was based on the Apollo eink controller and gumstix board with uClibc. Apparently they learned enough from the three years spent bringing that to market to switch to a freescale SoC and epson Broadsheet eink controller and glibc for the K2.

Regarding musl builds for the K1, the toybox roadmap says this:
Quote:
Note: toybox can autodetect nommu support when building with a uClibc toolchain such as the one from Aboriginal Linux, but with musl-libc you'll have to enable CONFIG_TOYBOX_MUSL_NOMMU_IS_BROKEN to work around the fact they provide a non-functional fork() implementation that always returns -ENOSYS, to prevent you from compile-time probing for nommu support when cross-compiling. Unfortunately "preventing you from probing" seems to be an explicit policy with musl, they also don't provide an "#ifdef __MUSL__" because their library is perfect and you're only ever allowed to work around other people's bugs, not theirs. So we have to use menuconfig to manually enable musl-specific bug workarounds.
And regarding K2s, I just received a K2 that was bulged on the front and the back, with the side seam split open. On opening, its battery is swelled to literally THREE times its original thickness. Thankfully, the battery was mounted behind the keyboard and not the fragile eink display. I will take photos, and investigate replacing the internal LiPo inside the plastic battery frame with a cheap-ass replacement from a cheap-ass remote control toy (keeping the little circuit board that identifies it as a genuine amazon battery).

P.S. I have KUAL and usbnet running fine on another K2, along with my Video Player KUAL extension. Though I did not yet add a "press a key to abort early" to the extension, unlike my K1 where I do not have to sit and wait for "Big Buck Bunny" to play to completion (or abort it from an SSH shell). I also have this video player extension on my PW3 (where the 800x600 video only occupies about a quarter of the display).

Last edited by geekmaster; 05-19-2016 at 10:19 PM.
geekmaster is offline   Reply With Quote