Re: [Xen-devel] [PATCH v2 4/4] x86/shim: only mark special pages as RAM in pvshim mode

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 28.12.18 at 13:04, <roger.pau@xxxxxxxxxx> wrote:
> @@ -113,6 +115,47 @@ uint64_t pv_shim_mem(uint64_t avail)
>      return shim_nrpages;
>  }
>  
> +static void __init mark_pfn_as_ram(struct e820map *e820, uint64_t pfn)
> +{
> +    if ( !e820_add_range(e820, pfn << PAGE_SHIFT,
> +                         (pfn << PAGE_SHIFT) + PAGE_SIZE, E820_RAM) )
> +        if ( !e820_change_range_type(e820, pfn << PAGE_SHIFT,
> +                                     (pfn << PAGE_SHIFT) + PAGE_SIZE,
> +                                     E820_RESERVED, E820_RAM) )
> +            panic("Unable to add/change memory type of pfn %#lx to RAM\n", 
> pfn);

While doing the move it would have been nice if you folded the two
if()-s. (Arguably pfn_to_paddr() could also be used here.)

> +void pv_shim_fixup_e820(struct e820map *e820)

You've lost the __init here.

> @@ -91,6 +93,16 @@ static inline uint64_t pv_shim_mem(uint64_t avail)
>      ASSERT_UNREACHABLE();
>      return 0;
>  }
> +static inline void pv_shim_fixup_e820(struct e820map *e820)
> +{
> +    ASSERT_UNREACHABLE();
> +}
> +static inline const struct platform_bad_page
> +*pv_shim_reserved_pages(unsigned int *s)

I can see why you split the previously too long line, but surely it
shouldn't be split in the middle of the return type.

I've taken the liberty and made all three modifications in preparation
for committing.

Jan



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel