Le 27/12/2016 à 14:58, Thomas Huth a écrit :
> Am Tue, 27 Dec 2016 12:52:30 +0100
> schrieb Laurent Vivier <laur...@vivier.eu>:
>
>> Le 20/12/2016 à 15:32, Thomas Huth a écrit :
>>> You can get an empty machine with "-M none" nowadays, so the
>>> m68k dummy board (introduced in 2007) seems to be pretty
>>> redundant since the "none" machine has been added in 2012.
>>>
>>> Signed-off-by: Thomas Huth <h...@tuxfamily.org>
>>> ---
>>> MAINTAINERS | 4 ---
>>> hw/m68k/Makefile.objs | 2 --
>>> hw/m68k/dummy_m68k.c | 84
>>> --------------------------------------------------- 3 files
>>> changed, 90 deletions(-) delete mode 100644 hw/m68k/dummy_m68k.c
>>>
>>> diff --git a/MAINTAINERS b/MAINTAINERS
>>> index 4a60579..88ee8cd 100644
>>> --- a/MAINTAINERS
>>> +++ b/MAINTAINERS
>>> @@ -552,10 +552,6 @@ S: Orphan
>>> F: hw/m68k/an5206.c
>>> F: hw/m68k/mcf5206.c
>>>
>>> -dummy_m68k
>>> -S: Orphan
>>> -F: hw/m68k/dummy_m68k.c
>>> -
>>> mcf5208
>>> S: Orphan
>>> F: hw/m68k/mcf5208.c
>>> diff --git a/hw/m68k/Makefile.objs b/hw/m68k/Makefile.objs
>>> index c4352e7..d1f089c 100644
>>> --- a/hw/m68k/Makefile.objs
>>> +++ b/hw/m68k/Makefile.objs
>>> @@ -1,4 +1,2 @@
>>> obj-y += an5206.o mcf5208.o
>>> -obj-y += dummy_m68k.o
>>> -
>>> obj-y += mcf5206.o mcf_intc.o
>>> diff --git a/hw/m68k/dummy_m68k.c b/hw/m68k/dummy_m68k.c
>>> deleted file mode 100644
>>> index 0b11d20..0000000
>>> --- a/hw/m68k/dummy_m68k.c
>>> +++ /dev/null
>>> @@ -1,84 +0,0 @@
>>> -/*
>>> - * Dummy board with just RAM and CPU for use as an ISS.
>>> - *
>>> - * Copyright (c) 2007 CodeSourcery.
>>> - *
>>> - * This code is licensed under the GPL
>>> - */
>>> -
>>> -#include "qemu/osdep.h"
>>> -#include "qemu-common.h"
>>> -#include "cpu.h"
>>> -#include "hw/hw.h"
>>> -#include "hw/boards.h"
>>> -#include "hw/loader.h"
>>> -#include "elf.h"
>>> -#include "exec/address-spaces.h"
>>> -
>>> -#define KERNEL_LOAD_ADDR 0x10000
>>> -
>>> -/* Board init. */
>>> -
>>> -static void dummy_m68k_init(MachineState *machine)
>>> -{
>>> - ram_addr_t ram_size = machine->ram_size;
>>> - const char *cpu_model = machine->cpu_model;
>>> - const char *kernel_filename = machine->kernel_filename;
>>> - M68kCPU *cpu;
>>> - CPUM68KState *env;
>>> - MemoryRegion *address_space_mem = get_system_memory();
>>> - MemoryRegion *ram = g_new(MemoryRegion, 1);
>>> - int kernel_size;
>>> - uint64_t elf_entry;
>>> - hwaddr entry;
>>> -
>>> - if (!cpu_model)
>>> - cpu_model = "cfv4e";
>>> - cpu = cpu_m68k_init(cpu_model);
>>> - if (!cpu) {
>>> - fprintf(stderr, "Unable to find m68k CPU definition\n");
>>> - exit(1);
>>> - }
>>> - env = &cpu->env;
>>> -
>>> - /* Initialize CPU registers. */
>>> - env->vbr = 0;
>>> -
>>> - /* RAM at address zero */
>>> - memory_region_allocate_system_memory(ram, NULL,
>>> "dummy_m68k.ram",
>>> - ram_size);
>>> - memory_region_add_subregion(address_space_mem, 0, ram);
>>> -
>>> - /* Load kernel. */
>>> - if (kernel_filename) {
>>> - kernel_size = load_elf(kernel_filename, NULL, NULL,
>>> &elf_entry,
>>> - NULL, NULL, 1, EM_68K, 0, 0);
>>> - entry = elf_entry;
>>> - if (kernel_size < 0) {
>>> - kernel_size = load_uimage(kernel_filename, &entry,
>>> NULL, NULL,
>>> - NULL, NULL);
>>> - }
>>> - if (kernel_size < 0) {
>>> - kernel_size = load_image_targphys(kernel_filename,
>>> - KERNEL_LOAD_ADDR,
>>> - ram_size -
>>> KERNEL_LOAD_ADDR);
>>> - entry = KERNEL_LOAD_ADDR;
>>> - }
>>> - if (kernel_size < 0) {
>>> - fprintf(stderr, "qemu: could not load kernel '%s'\n",
>>> - kernel_filename);
>>> - exit(1);
>>> - }
>>> - } else {
>>> - entry = 0;
>>> - }
>>> - env->pc = entry;
>>> -}
>>> -
>>> -static void dummy_m68k_machine_init(MachineClass *mc)
>>> -{
>>> - mc->desc = "Dummy board";
>>> - mc->init = dummy_m68k_init;
>>> -}
>>> -
>>> -DEFINE_MACHINE("dummy", dummy_m68k_machine_init)
>>>
>>
>> I'm not sure the "none" machine can replace the "dummy" machine as the
>> "dummy" machine can load a kernel whereas it seems the "none" one
>> can't.
>
> Oh, that's true ... I did not notice it 'cause there was no error
> message when I tried to start the "none" machine with "-kernel". But
> looking at the code in hw/core/null-machine.c, it really seems that
> there is nothing in here :-( So never mind, please forget about this
> patch. (but now I wonder whether the "none" machine could be improved to
> provide the features of the "dummy" machine, too?)
The "none" machine is defined to be be filled with QMP commands, perhaps
the generic-loader object can replace the "-kernel" parameters?
Laurent
