qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-devel] [PATCH v3 01/15] tests: add fp-test, a floating point t


From: Emilio G. Cota
Subject: Re: [Qemu-devel] [PATCH v3 01/15] tests: add fp-test, a floating point test suite
Date: Wed, 11 Apr 2018 17:36:48 -0400
User-agent: Mutt/1.5.24 (2015-08-30)

On Wed, Apr 11, 2018 at 02:20:49 +0100, Alex Bennée wrote:
> Emilio G. Cota <address@hidden> writes:
> So with the attached patch and my proposed cross build we can now get:
> 
> 02:15:54 address@hidden:~/l/q/qemu.git] softfloat-fixes-for-2.12-v1 ± find . 
> -iname "fp-test" | xargs file
> ./ppc64-linux-user/tests/fp-test:      ELF 64-bit LSB executable, 64-bit 
> PowerPC or cisco 7500, version 1 (GNU/Linux), statically linked, for 
> GNU/Linux 3.2.0, not stripped
(snip)
> But it did mean having to hack about a little, mainly to get rid of
> glib.

That will let us build fp-test using a cross-compiler. My initial
thinking was that since we'd end up testing on a real host
(with "-t host" mode), cross-compiling wouldn't be necessary since we
could just compile natively on said host.

But since we seem to be moving towards supporting cross-compilers,
it takes little effort to cross-compile fp-test as well. The main
hurdle is to remove the glib dependence as you pointed out. I just
wrote a few patches to do this:

$ git log --oneline -5 --reverse
48e802b osdep: disable glib-compat.h include with QEMU_NO_GLIB
d3c78c7 softfloat: do not include glib headers
744a9c4 tests/tcg/Makefile: define _GNU_SOURCE
661c0e2 tests/fp: fixup
e057d45 tests/tcg/Makefile: fp-test build fixup

The main difference with your attached patch is that we remove ifdef's
from fp-test.c while keeping the osdep.h include.

You can fetch the patches from
  https://github.com/cota/qemu/tree/softfloat-fixes-for-2.12-v1

[BTW the name of the branch is just to keep your original branch name;
I'm in now way intending for this to be part of 2.12 :>]

Thanks,

                Emilio



reply via email to

[Prev in Thread] Current Thread [Next in Thread]