[OE-core] [PATCH 1/1] tune-cortexa9.inc: add vfpv3 tunes
Martin Jansa
martin.jansa at gmail.com
Fri Feb 12 15:16:42 UTC 2016
On Fri, Feb 12, 2016 at 05:05:12PM +0200, Ioan-Adrian Ratiu wrote:
> On Fri, 12 Feb 2016 15:25:01 +0100
> Martin Jansa <martin.jansa at gmail.com> wrote:
>
> > On Fri, Feb 12, 2016 at 03:58:57PM +0200, Ioan-Adrian Ratiu wrote:
> > > Hello
> > >
> > > On Thu, 11 Feb 2016 17:41:54 +0100
> > > Martin Jansa <martin.jansa at gmail.com> wrote:
> > >
> > > > On Thu, Feb 11, 2016 at 05:18:56PM +0200, Ioan-Adrian Ratiu wrote:
> > > > > From: Richard Tollerton <rich.tollerton at ni.com>
> > > > >
> > > > > Define tunnings to enable 32 register VFPv3 for cortexa9 processor cores
> > > >
> > > > Can you execute the test from:
> > > > http://git.openembedded.org/openembedded-core-contrib/log/?h=jansa/tune2-test
> > > > just to verify that all these new combinations work ok and to update the
> > > > test results for next person modifying the tune files?
> > >
> > > I've tested with that script and everything looks ok (it took a while for all those openssl builds to run the first time).
> >
> > Thanks, it doesn't build them, only parses.
> >
> > > I also had to modify that hardcoded path to the openembedded-core location in the script. I asume this test script was created for the poky layout?
> >
> > No I don't use poky, but the layout isn't set in stone, I'm not
> > surprised you're using different.
> >
> > > (also I've done a grep "\(-vfpv3\)" on the AVAILTUNES var, but I don't think it matters that much though, run the test without it also)
> >
> > You mean in this line:
> > AVAILTUNES=`grep '^AVAILTUNES=' ${TUNE_TEST_DIR}/env.${M} | sed 's/^[^"]*"\([^"]*\)"/\1/g' | tr ' ' '\n'`;
> >
> > to restrict the testing only to new -vfpv3 tune?
>
> Yes, at the commit pointed by the URL in your previous mail the line had a grep neon-vfpv4 appended:
>
> AVAILTUNES=`grep '^AVAILTUNES=' ${TUNE_TEST_DIR}/env.${M} | sed 's/^[^"]*"\([^"]*\)"/\1/g' | tr ' ' '\n' | grep "\(neon-vfpv4\)"`;
>
> >
> > > There is also a discrepancy between the value of TUNE_PKGARCH == armv7athf and the file name from the env.fake-cortexa9.armv7athf-vfpv3. I believe there may be a typo in one of the tunes, ommited the "t" for thumb.
> >
> > I'm not sure where the typo is, can you push your results somewhere? So
> > it's easy to see the diff with latest jansa/tune2-test ?
>
> I did some more digging and that divergence seems to be caused by the following line in arch-armv7a.inc:
>
> PACKAGE_EXTRA_ARCHS_tune-armv7athf = "${PACKAGE_EXTRA_ARCHS_tune-armv6thf} armv7ahf-vfp armv7at2hf-vfp"
>
> Shouldn't it be the following: (add a t for thumb)?
>
> PACKAGE_EXTRA_ARCHS_tune-armv7athf = "${PACKAGE_EXTRA_ARCHS_tune-armv6thf} armv7athf-vfp armv7at2hf-vfp"
>
> The same applies to other armv7athf tunes like armv7athf-vfpv3d16, they all seem to be missing a t.
No they aren't missing "t", because thumb for armv7 is appended as "t2"
which is already there and tune-armv7athf needs to include arm tune,
otherwise recipes which set ARM_INSTRUCTION_SET to "arm" won't be in
PACKAGE_EXTRA_ARCHS and image build will fail.
> > > > > More details: http://www.arm.com/products/processors/technologies/vector-floating-point.php
> > > > >
> > > > > Signed-off-by: Richard Tollerton <rich.tollerton at ni.com>
> > > > > Signed-off-by: Ioan-Adrian Ratiu <adrian.ratiu at ni.com>
> > > > > ---
> > > > > meta/conf/machine/include/tune-cortexa9.inc | 15 +++++++++++++++
> > > > > 1 file changed, 15 insertions(+)
> > > > >
> > > > > diff --git a/meta/conf/machine/include/tune-cortexa9.inc b/meta/conf/machine/include/tune-cortexa9.inc
> > > > > index 965f2d9..3d8dc1d 100644
> > > > > --- a/meta/conf/machine/include/tune-cortexa9.inc
> > > > > +++ b/meta/conf/machine/include/tune-cortexa9.inc
> > > > > @@ -34,3 +34,18 @@ PACKAGE_EXTRA_ARCHS_tune-cortexa9hf = "${PACKAGE_EXTRA_ARCHS_tune-ar
> > > > > PACKAGE_EXTRA_ARCHS_tune-cortexa9thf = "${PACKAGE_EXTRA_ARCHS_tune-armv7athf} cortexa9hf-vfp cortexa9t2hf-vfp"
> > > > > PACKAGE_EXTRA_ARCHS_tune-cortexa9hf-neon = "${PACKAGE_EXTRA_ARCHS_tune-armv7ahf-neon} cortexa9hf-vfp cortexa9hf-neon"
> > > > > PACKAGE_EXTRA_ARCHS_tune-cortexa9thf-neon = "${PACKAGE_EXTRA_ARCHS_tune-armv7athf-neon} cortexa9hf-vfp cortexa9hf-neon cortexa9t2hf-vfp cortexa9t2hf-neon"
> > > > > +
> > > > > +# VFPv3 Tunes
> > > > > +AVAILTUNES += "cortexa9-vfpv3 cortexa9t-vfpv3 cortexa9hf-vfpv3 cortexa9thf-vfpv3"
> > > > > +ARMPKGARCH_tune-cortexa9-vfpv3 = "cortexa9"
> > > > > +ARMPKGARCH_tune-cortexa9t-vfpv3 = "cortexa9"
> > > > > +ARMPKGARCH_tune-cortexa9hf-vfpv3 = "cortexa9"
> > > > > +ARMPKGARCH_tune-cortexa9thf-vfpv3 = "cortexa9"
> > > > > +TUNE_FEATURES_tune-cortexa9-vfpv3 = "${TUNE_FEATURES_tune-armv7a-vfpv3} cortexa9"
> > > > > +TUNE_FEATURES_tune-cortexa9t-vfpv3 = "${TUNE_FEATURES_tune-armv7at-vfpv3} cortexa9"
> > > > > +TUNE_FEATURES_tune-cortexa9hf-vfpv3 = "${TUNE_FEATURES_tune-armv7ahf-vfpv3} cortexa9"
> > > > > +TUNE_FEATURES_tune-cortexa9thf-vfpv3 = "${TUNE_FEATURES_tune-armv7athf-vfpv3} cortexa9"
> > > > > +PACKAGE_EXTRA_ARCHS_tune-cortexa9-vfpv3 = "${PACKAGE_EXTRA_ARCHS_tune-armv7a-vfpv3} cortexa9-vfp cortexa9-vfpv3"
> > > > > +PACKAGE_EXTRA_ARCHS_tune-cortexa9t-vfpv3 = "${PACKAGE_EXTRA_ARCHS_tune-armv7at-vfpv3} cortexa9-vfp cortexa9-vfpv3 cortexa9t2-vfp cortexa9t2-vfpv3"
> > > > > +PACKAGE_EXTRA_ARCHS_tune-cortexa9hf-vfpv3 = "${PACKAGE_EXTRA_ARCHS_tune-armv7ahf-vfpv3} cortexa9hf-vfp cortexa9hf-vfpv3"
> > > > > +PACKAGE_EXTRA_ARCHS_tune-cortexa9thf-vfpv3 = "${PACKAGE_EXTRA_ARCHS_tune-armv7athf-vfpv3} cortexa9hf-vfp cortexa9hf-vfpv3 cortexa9t2hf-vfp cortexa9t2hf-vfpv3"
> > > > > --
> > > > > 2.7.0
> > > > >
> > > > > --
> > > > > _______________________________________________
> > > > > Openembedded-core mailing list
> > > > > Openembedded-core at lists.openembedded.org
> > > > > http://lists.openembedded.org/mailman/listinfo/openembedded-core
> > > >
> > >
> >
>
--
Martin 'JaMa' Jansa jabber: Martin.Jansa at gmail.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 181 bytes
Desc: Digital signature
URL: <http://lists.openembedded.org/pipermail/openembedded-core/attachments/20160212/79a3fcea/attachment-0002.sig>
More information about the Openembedded-core
mailing list