Re: [linux-linus test] 61295: regressions - FAIL

[Ian Campbell <ian.campbell@citrix.com>]

On Wed, 2015-09-09 at 11:31 +0100, Ian Campbell wrote:
> On Thu, 2015-09-03 at 23:29 +0000, osstest service owner wrote:
> > flight 61295 linux-linus real [real]
> > http://logs.test-lab.xenproject.org/osstest/logs/61295/
> > 
> > Regressions :-(
> > 
> > Tests which did not succeed and are blocking,
> > including tests which could not be run:
> > [...]
> >  test-armhf-armhf-xl-cubietruck  6 xen-boot                fail REGR.
> > vs. 59254
> > [...]
> 
> The bisector has fingered[0] a merge commit, f36fc04e4cdd, from the clk
> tree.
> 
[...]
> Between 26f8b7edc9ea and f36fc04e4cdd there is:
> 
>  * no diff to arch/arm/mach-sunxi/ 
>  * addition or changes of "clock-indices" property in many
>    arch/arm/boot/dts/*sun?i*
>  * a bunch of changes in drivers/clk/*sun?i*

This turned out to be an incompatible change between the kernel and the
DTB. osstest currently keeps using the Debian supplied DTB for all tests,
rather than picking up the one from the kernel it has built.

I've reported this upstream:
    http://lists.infradead.org/pipermail/linux-arm-kernel/2015-October/373753.html

In theory the DTB are supposed to be a stable API and the kernel is
supposed to work with shipped DTBS. The reality doesn't really live up to
this ideal though and I don't think it is really our job to fight this, so
I plan to adjust osstest to use the DTBs built by ts-kernel-build. In fact
this is probably preferable for us since it means we will be testing Xen
with a variety of DTBs over the different kernel versions, instead of just
the ones which Debian supplies.

I'll cook up some osstest patches today.

Ian.