Re: [PATCH 12/13] ppc440_pcix: Don't use iomem for regs

[BALATON Zoltan]

On Tue, 4 Jul 2023, Philippe Mathieu-Daudé wrote:

On 4/7/23 00:02, BALATON Zoltan wrote:

The iomem memory region is better used for the PCI IO space but
currently used for registers. Stop using it for that to allow this to
be cleaned up in the next patch.

Signed-off-by: BALATON Zoltan <bala...@eik.bme.hu>
---
  hw/ppc/ppc440_pcix.c | 7 ++++---
  1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/hw/ppc/ppc440_pcix.c b/hw/ppc/ppc440_pcix.c
index adfecf1e76..ee2dc44f67 100644
--- a/hw/ppc/ppc440_pcix.c
+++ b/hw/ppc/ppc440_pcix.c

@@ -484,6 +484,7 @@ static void ppc440_pcix_realize(DeviceState *dev, Error **errp)

      SysBusDevice *sbd = SYS_BUS_DEVICE(dev);
      PPC440PCIXState *s;
      PCIHostState *h;
+    MemoryRegion *regs = g_new(MemoryRegion, 1);

Why not hold it within PPC440PCIXState?

Because it's never needed after this function.

Regards,
BALATON Zoltan

      h = PCI_HOST_BRIDGE(dev);
      s = PPC440_PCIX_HOST(dev);

@@ -507,11 +508,11 @@ static void ppc440_pcix_realize(DeviceState *dev, Error **errp)

                            h, "pci-conf-idx", 4);
      memory_region_init_io(&h->data_mem, OBJECT(s), &pci_host_data_le_ops,
                            h, "pci-conf-data", 4);
-    memory_region_init_io(&s->iomem, OBJECT(s), &pci_reg_ops, s,
-                          "pci.reg", PPC440_REG_SIZE);
+    memory_region_init_io(regs, OBJECT(s), &pci_reg_ops, s, "pci-reg",
+                          PPC440_REG_SIZE);

memory_region_add_subregion(&s->container, PCIC0_CFGADDR, &h->conf_mem); memory_region_add_subregion(&s->container, PCIC0_CFGDATA, &h->data_mem); - memory_region_add_subregion(&s->container, PPC440_REG_BASE, &s->iomem);

+    memory_region_add_subregion(&s->container, PPC440_REG_BASE, regs);
      sysbus_init_mmio(sbd, &s->container);
  }