Re: [Xen-devel] [PATCH v4 4/7] xen/arm: move setup_mm right after setup_pagetables

[Ian Campbell <Ian.Campbell@xxxxxxxxxx>]

On Tue, 2013-01-08 at 20:03 +0000, Stefano Stabellini wrote:
> At the moment we destroy the DTB mappings we have in setup_pagetables
> and we restore them only in setup_mm.
> 
> Move setup_mm right after setup_pagetables so that we don't leave over
> portions of the initialization code without a working DTB mapping.

"leave over"? I think you just mean to say "This ensures we have a valid
DTB mapping while running the subsequent initialization code" or
something like that.

Do you still get the :
        printk("Xen heap: %lu pages  Dom heap: %lu pages\n
message with this change? Or do all the printks in this function need to
become early_printk/early_panic?

Actually, this function has 1x early_printk, 1x printk and 1x panic,
it's a bit confused I think ;-)

> 
> Signed-off-by: Stefano Stabellini <stefano.stabellini@xxxxxxxxxxxxx>
> ---
>  xen/arch/arm/setup.c |    3 +--
>  1 files changed, 1 insertions(+), 2 deletions(-)
> 
> diff --git a/xen/arch/arm/setup.c b/xen/arch/arm/setup.c
> index 2076724..8882514 100644
> --- a/xen/arch/arm/setup.c
> +++ b/xen/arch/arm/setup.c
> @@ -212,6 +212,7 @@ void __init start_xen(unsigned long boot_phys_offset,
>      cmdline_parse(device_tree_bootargs(fdt));
>  
>      setup_pagetables(boot_phys_offset, get_xen_paddr());
> +    setup_mm(atag_paddr, fdt_size);
>  
>  #ifdef EARLY_UART_ADDRESS
>      /* TODO Need to get device tree or command line for UART address */
> @@ -229,8 +230,6 @@ void __init start_xen(unsigned long boot_phys_offset,
>  
>      init_xen_time();
>  
> -    setup_mm(atag_paddr, fdt_size);
> -
>      /* Setup Hyp vector base */
>      WRITE_CP32((uint32_t) hyp_traps_vector, HVBAR);
>      printk("Set hyp vector base to %"PRIx32" (expected %p)\n",



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel