[v2] xlnx-zynqmp: add support to boot on RPUs

[PATCH v2 0/4] xlnx-zynqmp: add support to boot on RPUs

Posted by Clément Chigot 8 months, 2 weeks ago

This series enhances Xilinx ZynqMP support to allow booting on RPUs and
ease gdb connections to them.

It was validated with home-made binaries. FreeRTOS was tested but without
success: outputs/IRQ seems broken. AFAICT, FreeRTOS is expecting Xilinx's
QEMU thus I didn't investigate further, but I'd still like advice
on the 3rd patch ("wire a second GIC") since it could be related to it.

Changes since v1:
 - add doc for "first-cpu-index" new property in arm_gic.h.

Clément Chigot (2):
  hw/arm: make cpu targeted by arm_load_kernel the primary CPU.
  hw/arm/xlnx-zynqmp: adapt cluster-id based on the boot cpu

Frederic Konrad (2):
  hw/intc/arm_gic: introduce a first-cpu-index property
  hw/arm/xlnx-zynqmp: wire a second GIC for the Cortex-R5

 hw/arm/boot.c                    |  15 +++--
 hw/arm/xlnx-zynqmp.c             | 104 ++++++++++++++++++++++++++++---
 hw/intc/arm_gic.c                |   2 +-
 hw/intc/arm_gic_common.c         |   1 +
 include/hw/arm/boot.h            |   3 +
 include/hw/arm/xlnx-zynqmp.h     |   6 ++
 include/hw/intc/arm_gic.h        |   2 +
 include/hw/intc/arm_gic_common.h |   2 +
 8 files changed, 117 insertions(+), 18 deletions(-)

-- 
2.34.1

Re: [PATCH v2 0/4] xlnx-zynqmp: add support to boot on RPUs

Posted by Peter Maydell 7 months, 4 weeks ago

On Mon, 26 May 2025 at 09:55, Clément Chigot <chigot@adacore.com> wrote:
>
> This series enhances Xilinx ZynqMP support to allow booting on RPUs and
> ease gdb connections to them.
>
> It was validated with home-made binaries. FreeRTOS was tested but without
> success: outputs/IRQ seems broken. AFAICT, FreeRTOS is expecting Xilinx's
> QEMU thus I didn't investigate further, but I'd still like advice
> on the 3rd patch ("wire a second GIC") since it could be related to it.
>
> Changes since v1:
>  - add doc for "first-cpu-index" new property in arm_gic.h.
>
> Clément Chigot (2):
>   hw/arm: make cpu targeted by arm_load_kernel the primary CPU.
>   hw/arm/xlnx-zynqmp: adapt cluster-id based on the boot cpu
>
> Frederic Konrad (2):
>   hw/intc/arm_gic: introduce a first-cpu-index property
>   hw/arm/xlnx-zynqmp: wire a second GIC for the Cortex-R5

Hi; I've taken the first 2 patches in this series into
target-arm.next. I left some review comments on patch 3.
For patch 4, I still feel this is something that should
be dealt with at the gdb end (e.g. gdb should support
"connect to this target and connect to remote-inferior 2,
not inferior 1, and don't fall over just because you
don't know the architecture of inferior 1").

thanks
-- PMM

Re: [PATCH v2 0/4] xlnx-zynqmp: add support to boot on RPUs

Posted by Clément Chigot 7 months, 4 weeks ago

On Fri, Jun 13, 2025 at 1:58 PM Peter Maydell <peter.maydell@linaro.org> wrote:
>
> On Mon, 26 May 2025 at 09:55, Clément Chigot <chigot@adacore.com> wrote:
> >
> > This series enhances Xilinx ZynqMP support to allow booting on RPUs and
> > ease gdb connections to them.
> >
> > It was validated with home-made binaries. FreeRTOS was tested but without
> > success: outputs/IRQ seems broken. AFAICT, FreeRTOS is expecting Xilinx's
> > QEMU thus I didn't investigate further, but I'd still like advice
> > on the 3rd patch ("wire a second GIC") since it could be related to it.
> >
> > Changes since v1:
> >  - add doc for "first-cpu-index" new property in arm_gic.h.
> >
> > Clément Chigot (2):
> >   hw/arm: make cpu targeted by arm_load_kernel the primary CPU.
> >   hw/arm/xlnx-zynqmp: adapt cluster-id based on the boot cpu
> >
> > Frederic Konrad (2):
> >   hw/intc/arm_gic: introduce a first-cpu-index property
> >   hw/arm/xlnx-zynqmp: wire a second GIC for the Cortex-R5
>
> Hi; I've taken the first 2 patches in this series into
> target-arm.next. I left some review comments on patch 3.

Thanks a lot. I've just sent a v3 addressing them.

> For patch 4, I still feel this is something that should
> be dealt with at the gdb end (e.g. gdb should support
> "connect to this target and connect to remote-inferior 2,
> not inferior 1, and don't fall over just because you
> don't know the architecture of inferior 1").

Yeah, I'm totally fine with that, I've removed it from v3.

PS: I'll be off for a few weeks but should be able to make a v4 if
needed before 10.1 soft freeze.

Thanks, Clément