This is the mail archive of the
newlib@sourceware.org
mailing list for the newlib project.
Re: Building newlib without -mhard-float
On Sep 4, 2007, at 5:25 PM, Jeff Johnston wrote:
Yes, with CFLAGS spelled correctly :)
Heh, yes, I noticed that after the follow-up email I sent, and didn't
want to send YET ANOTHER minor correction. I meant to show that
TARGET_CFLAGS assignment in front of "make", which is what I intended
to ask about.
Which brings up another question I've had before: should I just
use -mcpu=xscale, or just -march=armv5te, or both? At one point, I
got errors in something when I used both, saying that they were in
conflict (IIRC--strikes me as incorrect).
That's a question for the ARM experts out there. I don't know
exactly what options you want to specify. I would assume the
original --print-multi-lib output would show you compatible
permutations.
Sadly, only a couple of the permutations I have get at all cpu-
specific, and none specify an arch.
How does newlib choose the "machine" subdirectory implementation (I
already asked this in another post, so if it gets answered there,
feel free to ignore the question).
Thank you very much for your help, and patience. I'm under the gun to
get this stuff working and am struggling to find all the right answers.
--
Rick