Re: [Qemu-devel] [PATCH 06/24] QEMUMachine: pass address space to machine init function

[Avi Kivity <avi@redhat.com>]

On 08/24/2011 01:53 PM, Peter Maydell wrote:
> >  The purpose here is to allow removal of get_system_memory() from
> >  the general code base.
>
> The right way to remove get_system_memory() from the general code base
> is to actually model things correctly, for instance by having machine
> models create a container memory region into which they insert the
> memory regions for all the devices which currently use sysbus_mmio_map
> to map themselves, and then pass the container memory region to the
> "master" end of the bus, ie the CPU.
>

I think you're right.  This also allows eventual removal of system_io on 
anything non-x86.

So a replacement would look like:

(before)

-static void pc_init_isa(ram_addr_t ram_size,
+static void pc_init_isa(MemoryRegion *address_space_mem,
+                        MemoryRegion *address_space_io,
+                        ram_addr_t ram_size,
                          const char *boot_device,
                          const char *kernel_filename,
                          const char *kernel_cmdline,
@@ -259,15 +265,17 @@ static void pc_init_isa(ram_addr_t ram_size,
  {
      if (cpu_model == NULL)
          cpu_model = "486";
-    pc_init1(get_system_memory(),
-             get_system_io(),
+    pc_init1(address_space_mem,
+             address_space_io,
               ram_size, boot_device,
               kernel_filename, kernel_cmdline,
               initrd_filename, cpu_model, 0, 1);
  }

(after)
@@ -259,15 +265,17 @@ static void pc_init_isa(ram_addr_t ram_size,
  {
+    MemoryRegion *address_space_mem, *address_space_io;
+
+    setup_system_memory(&address_space_mem,&address_space_io);
      if (cpu_model == NULL)
          cpu_model = "486";
-    pc_init1(get_system_memory(),
-             get_system_io(),
+    pc_init1(address_space_mem,
+             address_space_io,
               ram_size, boot_device,
               kernel_filename, kernel_cmdline,
               initrd_filename, cpu_model, 0, 1);
  }


Later on, we'd refine the setup_system_memory() calls, for example not 
to create the io space on non-x86.

A possible complication is whether anything currently uses system_memory 
before ->init is called.  Anyone know?

-- 
I have a truly marvellous patch that fixes the bug which this
signature is too narrow to contain.