Re: armel qualification for Wheezy
On Wed, 2012-05-23 at 17:45 -0400, Lennart Sorensen wrote:
> On Wed, May 23, 2012 at 09:02:33PM +0100, Tixy wrote:
> > I may be being naive, but could an X86 PC be used with an ARM chroot and
> > qemu-arm-static to emulate ARM instructions? Or is qemu not stable
> > enough, or the emulated environment different enough that package
> > building would fail (e.g. through use of uname)?
>
> It is _horribly_ slow.
Not that horrible. I just did a kernel build on my laptop in an ARM
chroot and it took 19m43s, doing it as a cross-build took 1m14s. I
haven't got my Pandaboard setup to do a comparison, but I
suspect it wouldn't be much faster than my emulated ARM run.
> > PCs have the advantage of RAM (assuming QEMU can handle 2GB+), fast
> > hardware and multiple cores.
>
> Yes, but qemu doesn't really use more than one cpu, can't (last I checked)
> emulate more than one cpu core, and since it is emulating is rather slow
> (although it is fast as emulators go).
I'm not talking about using QEMU as a system emulator, just an
instruction set emulator. So ARM processes are running and scheduled as
native X86 PC processes, just using QEMU to interpret the instructions
in ARM ELF files. (There may be other magic going on, all I really know
about QEMU is how to make use of it following cut'n'paste instruction
from the web).
In the kernel building example I mentioned I was using "make -j8" and
that went a _lot_ faster than -j1; I didn't wait to get final timings
for a single threaded build.
--
Tixy
Reply to: