qemu-arm
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [PATCH v6 02/11] hw/intc/xilinx_intc: Make device endianness configu


From: Philippe Mathieu-Daudé
Subject: Re: [PATCH v6 02/11] hw/intc/xilinx_intc: Make device endianness configurable
Date: Wed, 12 Feb 2025 12:44:23 +0100
User-agent: Mozilla Thunderbird

On 12/2/25 12:42, Thomas Huth wrote:
On 12/02/2025 12.24, Philippe Mathieu-Daudé wrote:
Replace the DEVICE_NATIVE_ENDIAN MemoryRegionOps by a pair
of DEVICE_LITTLE_ENDIAN / DEVICE_BIG_ENDIAN.
Add the "little-endian" property to select the device
endianness, defaulting to little endian.
Set the proper endianness for each machine using the device.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
---
  hw/intc/xilinx_intc.c                    | 60 ++++++++++++++++++------
  hw/microblaze/petalogix_ml605_mmu.c      |  1 +
  hw/microblaze/petalogix_s3adsp1800_mmu.c |  3 ++
  hw/ppc/virtex_ml507.c                    |  1 +
  hw/riscv/microblaze-v-generic.c          |  1 +
  5 files changed, 52 insertions(+), 14 deletions(-)

diff --git a/hw/intc/xilinx_intc.c b/hw/intc/xilinx_intc.c
index 6930f83907a..523402b688c 100644
--- a/hw/intc/xilinx_intc.c
+++ b/hw/intc/xilinx_intc.c
@@ -3,6 +3,9 @@
   *
   * Copyright (c) 2009 Edgar E. Iglesias.
   *
+ * https://docs.amd.com/v/u/en-US/xps_intc
+ * DS572: LogiCORE IP XPS Interrupt Controller (v2.01a)
+ *
   * Permission is hereby granted, free of charge, to any person obtaining a copy    * of this software and associated documentation files (the "Software"), to deal    * in the Software without restriction, including without limitation the rights
@@ -23,10 +26,12 @@
   */
  #include "qemu/osdep.h"
+#include "qapi/error.h"
  #include "hw/sysbus.h"
  #include "qemu/module.h"
  #include "hw/irq.h"
  #include "hw/qdev-properties.h"
+#include "hw/qdev-properties-system.h"
  #include "qom/object.h"
  #define D(x)
@@ -49,6 +54,7 @@ struct XpsIntc
  {
      SysBusDevice parent_obj;
+    EndianMode model_endianness;
      MemoryRegion mmio;
      qemu_irq parent_irq;
@@ -140,18 +146,29 @@ static void pic_write(void *opaque, hwaddr addr,
      update_irq(p);
  }
-static const MemoryRegionOps pic_ops = {
-    .read = pic_read,
-    .write = pic_write,
-    .endianness = DEVICE_NATIVE_ENDIAN,
-    .impl = {
-        .min_access_size = 4,
-        .max_access_size = 4,
+static const MemoryRegionOps pic_ops[2] = {
+    [0 ... 1] = {

Hmm, ok, so here we have now an assumption that ENDIAN_MODE_BIG and ENDIAN_MODE_LITTLE match 0 and 1, which would not work anymore when using 0 as unspecified... a little bit ugly ... so maybe instead of changing pic_ops into an array ....

+        .read = pic_read,
+        .write = pic_write,
+        .endianness = DEVICE_BIG_ENDIAN,
+        .impl = {
+            .min_access_size = 4,
+            .max_access_size = 4,
+        },
+        .valid = {
+            /*
+             * All XPS INTC registers are accessed through the PLB interface.
+             * The base address for these registers is provided by the
+             * configuration parameter, C_BASEADDR. Each register is 32 bits +             * although some bits may be unused and is accessed on a 4-byte
+             * boundary offset from the base address.
+             */
+            .min_access_size = 4,
+            .max_access_size = 4,
+        },
      },
-    .valid = {
-        .min_access_size = 4,
-        .max_access_size = 4
-    }
+    [ENDIAN_MODE_BIG].endianness = DEVICE_BIG_ENDIAN,
+    [ENDIAN_MODE_LITTLE].endianness = DEVICE_LITTLE_ENDIAN,
  };
  static void irq_handler(void *opaque, int irq, int level)
@@ -174,13 +191,27 @@ static void xilinx_intc_init(Object *obj)
      qdev_init_gpio_in(DEVICE(obj), irq_handler, 32);
      sysbus_init_irq(SYS_BUS_DEVICE(obj), &p->parent_irq);
-
-    memory_region_init_io(&p->mmio, obj, &pic_ops, p, "xlnx.xps-intc",
-                          R_MAX * 4);
      sysbus_init_mmio(SYS_BUS_DEVICE(obj), &p->mmio);
  }
+static void xilinx_intc_realize(DeviceState *dev, Error **errp)
+{
+    XpsIntc *p = XILINX_INTC(dev);
+
+    if (p->model_endianness == ENDIAN_MODE_UNSPECIFIED) {
+        error_setg(errp, TYPE_XILINX_INTC " property 'endianness'"
+                         " must be set to 'big' or 'little'");
+        return;
+    }

... would it be possible to patch in the right value for pic_ops.endianness
here instead?

Ah, clever than my reply on the previous patch. I'll give it a try, thanks!


  Thomas




reply via email to

[Prev in Thread] Current Thread [Next in Thread]