There are claims that glibc cannot be built on an EABI armv4 system:
https://sourceware.org/bugzilla/show_bug.cgi?id=12097
But there was an interesting comment posted at that link above:
Quote:
Ignore my previous note.
-march=armv5 -mtune=xscale
produces bad binaries for xscale.
./hello
Illegal instruction
building with -mcpu=xscale (no -march or -mtune) does not need this patch.
|
Perhaps uClibc is the only option for working toolchains for the K1? And for some reason Rob Landly kept aboriginal armv4l using uClibc even after converting all else to musl. I have not stumbled upon his rationale for that build choice yet though... But that comment in the link above makes me wonder.
I am also curious if one of the OLD versions of aboriginal armv4l might have the right version of GCC for building K1-compatible kernel modules. If not, I will need to succeed at building the toolchain from the source included with the K1 GPL source code from amazon (if in fact that is even the right GCC version for building K1 FW 1.2 kernel modules).
I do like this comment in one of Rob Landley's README files:
Quote:
|
If you don't do cross compiling already, this project isn't going to teach you how, and in fact recommends you just don't go there. We cross compile so you don't have to.
|