From mboxrd@z Thu Jan 1 00:00:00 1970 From: dave.martin@linaro.org (Dave Martin) Date: Mon, 3 Sep 2012 13:21:34 +0100 Subject: [PATCH v2] ARM: Build dtb files in all target In-Reply-To: <20120901144922.GA25277@lunn.ch> References: <1346303687-7795-1-git-send-email-andrew@lunn.ch> <20120901144922.GA25277@lunn.ch> Message-ID: <20120903122134.GA3290@linaro.org> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Sat, Sep 01, 2012 at 04:49:22PM +0200, Andrew Lunn wrote: > On Thu, Aug 30, 2012 at 07:14:47AM +0200, Andrew Lunn wrote: > > Tools like kisskb are good at finding build regressions in the kernel > > sources. However, regressions in the DT desscriptions are not found, > > because generally these build systems don't build the DT binary blobs. > > > > Extend the ARM all target to build all enabled DTB files. > > > > Signed-off-by: Andrew Lunn > > --- > > v2: * in the help test > > Only invoke dtbs if CONFIGUSE_OF=y > > > > arch/arm/Makefile | 9 +++++++-- > > 1 file changed, 7 insertions(+), 2 deletions(-) > > > > diff --git a/arch/arm/Makefile b/arch/arm/Makefile > > index 30eae87..0457ef4 100644 > > --- a/arch/arm/Makefile > > +++ b/arch/arm/Makefile > > @@ -268,7 +268,12 @@ else > > KBUILD_IMAGE := zImage > > endif > > > > -all: $(KBUILD_IMAGE) > > +# Build the DT binary blobs if we have OF configured > > +ifeq ($(CONFIG_USE_OF),y) > > +KBUILD_DTBS := dtbs > > +endif > > + > > +all: $(KBUILD_IMAGE) $(KBUILD_DTBS) > > > > boot := arch/arm/boot > > > > @@ -306,7 +311,7 @@ define archhelp > > echo ' uImage - U-Boot wrapped zImage' > > echo ' bootpImage - Combined zImage and initial RAM disk' > > echo ' (supply initrd image via make variable INITRD=)' > > - echo ' dtbs - Build device tree blobs for enabled boards' > > + echo '* dtbs - Build device tree blobs for enabled boards' > > echo ' install - Install uncompressed kernel' > > echo ' zinstall - Install compressed kernel' > > echo ' uinstall - Install U-Boot wrapped compressed kernel' > > -- > > 1.7.10.4 > > > > Hi Folks > > Would anybody like to comment about the patch itself? > > Assuming the patch is acceptable, through which maintainers tree > should it be submitted? > > Thanks > Andrew My only concern here is that because of the missing make dependency between dtbs and dtc, this could cause unexpected build failures during parallel builds. However, so far as I can see this is not a problem unless explicit targets are put on the command line (which is already problematic with or without this patch). There is a separate patch under discussion to fix that, so your patch need not address it. This looks like a sensible fix for something which has been niggling at me, so Reviewed-by: Dave Martin This is ARM core build system stuff, so I would expect that it should be submitted via Russell. Because this is related to DT, I suggest that you repost first, CC'ing devicetree-discuss at lists.ozlabs.org so those guys have a chance to comment (though the relevant people will generally be reading alkml anyway). They might have a view on whether there are better ways to write the patch ... but probably not. I'll leave it for the stable guys to judge whether this patch is appropriate for them. The patch itself is a feature addition, which would normally exclude it. But its inclusion might still be considered justified on grounds of improving build regression coverage of dtbs and dtc on stable. Cheers ---Dave