Re: [PATCH v3 00/16] hw/arm/raspi: Add thermal/timer, improve address space, run U-boot

[Peter Maydell]

On Thu, 24 Oct 2019 at 14:42, Peter Maydell <address@hidden> wrote:
>
> On Sun, 20 Oct 2019 at 00:47, Philippe Mathieu-Daudé <address@hidden> wrote:
> >
> > Since v2:
> > - fixed issue in videocore address space
> > - allow to start with some cores OFF (to boot firmwares)
> > - add proof-of-concept test for '-smp cores=1' and U-boot
> > - fixed my email setup
> >
> > Previous cover:
> >
> > Hi,
> >
> > Some patches from v1 are already merged. This v2 addresses the
> > review comment from v1, and add patches to clean the memory
> > space when using multiple cores.
> >
> > Laurent, if you test U-Boot with this patchset again, do you mind
> > replying with a "Tested-by:" tag?
> >
> > The next patchset is probably about the interrupt controller blocks,
> > then will come another one about the MBox/Properties.
> >
> > The last patch is unrelated to the series, but since I cleaned this
> > for the raspi and the highbank is the only board with the same issue,
> > I included the patch in this series.
>
> I'm going to apply 1-10 and 14 to target-arm.next.
> (I've reviewed 10, and the rest have been reviewed.)

...but that causes tests/boot-serial-test to throw
a clang sanitizer error and then hang:

e104462:bionic:clang$ QTEST_QEMU_BINARY=arm-softmmu/qemu-system-arm
./tests/boot-serial-test
/arm/boot-serial/raspi2:
/home/petmay01/linaro/qemu-from-laptop/qemu/memory.c:2517:27: runtime
error: null pointer passed as argument 2, which is declared to never
be null
/usr/include/stdlib.h:819:6: note: nonnull attribute specified here

The offending patch is "hw/arm/bcm2836: Use per CPU address spaces"
(patch 7). So I'm dropping 7/8/9.

thanks
-- PMM