Hi,

There is no point in exposing an internal MMIO region via
SysBus and directly mapping it in the very same device.

This series replaces a sequence of:
- sysbus_init_mmio()
- sysbus_mmio_map()
by a single call to memory_region_add_subregion().

Philippe Mathieu-Daudé (6):
  hw/ppc/pnv_xscom: Rename pnv_xscom_realize(Error **) ->
    pnv_xscom_init()
  hw/ppc/pnv_xscom: Move sysbus_mmio_map() call within pnv_xscom_init()
  hw/ppc/pnv_xscom: Do not use SysBus API to map local MMIO region
  hw/ppc/pnv: Do not use SysBus API to map local MMIO region
  hw/intc/spapr_xive: Move sysbus_init_mmio() calls around
  hw/intc/spapr_xive: Do not use SysBus API to map local MMIO region

 include/hw/ppc/pnv_xscom.h |  2 +-
 hw/intc/spapr_xive.c       | 12 ++++++------
 hw/ppc/pnv.c               | 26 +++++---------------------
 hw/ppc/pnv_xscom.c         |  5 ++---
 4 files changed, 14 insertions(+), 31 deletions(-)

-- 
2.41.0

On 2023/10/18 21:30, Philippe Mathieu-Daudé wrote:
> Hi,
>
> There is no point in exposing an internal MMIO region via
> SysBus and directly mapping it in the very same device.
>
> This series replaces a sequence of:
> - sysbus_init_mmio()
> - sysbus_mmio_map()
> by a single call to memory_region_add_subregion().

Reviewed-by: LIU Zhiwei <zhiwei_liu@linux.alibaba.com>

Zhiwei

>
> Philippe Mathieu-Daudé (6):
>    hw/ppc/pnv_xscom: Rename pnv_xscom_realize(Error **) ->
>      pnv_xscom_init()
>    hw/ppc/pnv_xscom: Move sysbus_mmio_map() call within pnv_xscom_init()
>    hw/ppc/pnv_xscom: Do not use SysBus API to map local MMIO region
>    hw/ppc/pnv: Do not use SysBus API to map local MMIO region
>    hw/intc/spapr_xive: Move sysbus_init_mmio() calls around
>    hw/intc/spapr_xive: Do not use SysBus API to map local MMIO region
>
>   include/hw/ppc/pnv_xscom.h |  2 +-
>   hw/intc/spapr_xive.c       | 12 ++++++------
>   hw/ppc/pnv.c               | 26 +++++---------------------
>   hw/ppc/pnv_xscom.c         |  5 ++---
>   4 files changed, 14 insertions(+), 31 deletions(-)
>

On 10/18/23 06:30, Philippe Mathieu-Daudé wrote:
> Hi,
> 
> There is no point in exposing an internal MMIO region via
> SysBus and directly mapping it in the very same device.
> 
> This series replaces a sequence of:
> - sysbus_init_mmio()
> - sysbus_mmio_map()
> by a single call to memory_region_add_subregion().
> 
> Philippe Mathieu-Daudé (6):
>    hw/ppc/pnv_xscom: Rename pnv_xscom_realize(Error **) ->
>      pnv_xscom_init()
>    hw/ppc/pnv_xscom: Move sysbus_mmio_map() call within pnv_xscom_init()
>    hw/ppc/pnv_xscom: Do not use SysBus API to map local MMIO region
>    hw/ppc/pnv: Do not use SysBus API to map local MMIO region
>    hw/intc/spapr_xive: Move sysbus_init_mmio() calls around
>    hw/intc/spapr_xive: Do not use SysBus API to map local MMIO region
> 
>   include/hw/ppc/pnv_xscom.h |  2 +-
>   hw/intc/spapr_xive.c       | 12 ++++++------
>   hw/ppc/pnv.c               | 26 +++++---------------------
>   hw/ppc/pnv_xscom.c         |  5 ++---
>   4 files changed, 14 insertions(+), 31 deletions(-)
> 

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>

r~

On 18/10/23 15:30, Philippe Mathieu-Daudé wrote:
> Hi,
> 
> There is no point in exposing an internal MMIO region via
> SysBus and directly mapping it in the very same device.
> 
> This series replaces a sequence of:
> - sysbus_init_mmio()
> - sysbus_mmio_map()
> by a single call to memory_region_add_subregion().
> 
> Philippe Mathieu-Daudé (6):
>    hw/ppc/pnv_xscom: Rename pnv_xscom_realize(Error **) ->
>      pnv_xscom_init()
>    hw/ppc/pnv_xscom: Move sysbus_mmio_map() call within pnv_xscom_init()
>    hw/ppc/pnv_xscom: Do not use SysBus API to map local MMIO region
>    hw/ppc/pnv: Do not use SysBus API to map local MMIO region
>    hw/intc/spapr_xive: Move sysbus_init_mmio() calls around
>    hw/intc/spapr_xive: Do not use SysBus API to map local MMIO region

Cc'ing Thomas :)