Re: [Qemu-devel] [PATCH v3 03/10] aspeed-soc: provide a framework to add new SoCs

[Peter Maydell]

On 2 August 2016 at 18:15, Cédric Le Goater <address@hidden> wrote:
> Let's define an object class for each Aspeed SoC we support. A
> AspeedSoCInfo struct gathers the SoC specifications which can later be
> used by an instance of the class or by a board using the SoC.
>
> Signed-off-by: Cédric Le Goater <address@hidden>

> diff --git a/hw/arm/palmetto-bmc.c b/hw/arm/palmetto-bmc.c
> index 4d11905cfb18..531c266d9449 100644
> --- a/hw/arm/palmetto-bmc.c
> +++ b/hw/arm/palmetto-bmc.c
> @@ -22,8 +22,6 @@
>  #include "sysemu/blockdev.h"
>
>  static struct arm_boot_info palmetto_bmc_binfo = {
> -    .loader_start = AST2400_SDRAM_BASE,
> -    .board_id = 0,
>      .nb_cpus = 1,
>  };
>
> @@ -61,14 +59,17 @@ static void palmetto_bmc_init_flashes(AspeedSMCState *s, 
> const char *flashtype,
>  static void palmetto_bmc_init(MachineState *machine)
>  {
>      PalmettoBMCState *bmc;
> +    AspeedSoCClass *sc;
>
>      bmc = g_new0(PalmettoBMCState, 1);
> -    object_initialize(&bmc->soc, (sizeof(bmc->soc)), TYPE_ASPEED_SOC);
> +    object_initialize(&bmc->soc, (sizeof(bmc->soc)), "ast2400-a0");
>      object_property_add_child(OBJECT(machine), "soc", OBJECT(&bmc->soc),
>                                &error_abort);
>
> +    sc = ASPEED_SOC_GET_CLASS(&bmc->soc);
> +
>      memory_region_allocate_system_memory(&bmc->ram, NULL, "ram", ram_size);
> -    memory_region_add_subregion(get_system_memory(), AST2400_SDRAM_BASE,
> +    memory_region_add_subregion(get_system_memory(), sc->info->sdram_base,
>                                  &bmc->ram);
>      object_property_add_const_link(OBJECT(&bmc->soc), "ram", 
> OBJECT(&bmc->ram),
>                                     &error_abort);
> @@ -84,6 +85,9 @@ static void palmetto_bmc_init(MachineState *machine)
>      palmetto_bmc_binfo.initrd_filename = machine->initrd_filename;
>      palmetto_bmc_binfo.kernel_cmdline = machine->kernel_cmdline;
>      palmetto_bmc_binfo.ram_size = ram_size;
> +    palmetto_bmc_binfo.board_id = sc->info->silicon_rev;
> +    palmetto_bmc_binfo.loader_start = sc->info->sdram_base;
> +

This changes the behaviour from passing in the board_id
as 0 to passing in the silicon rev. Neither of those
things is actually a valid board ID value, which must
be one of:
(a) for legacy pre-device-tree boards, a value listed in
the official database at:
http://www.arm.linux.org.uk/developer/machines/download.php
(b) for device-tree-only boards, -1

board_id 0 means "I am an EBSA110", which this isn't,
and your silicon rev values are completely out of range.

thanks
-- PMM