Re: [PATCH v2 16/18] mini-os: map page allocator's bitmap to virtual kernel area for ballooning

[Samuel Thibault <samuel.thibault@ens-lyon.org>]

It looks a bit odd to be remapping something that was just allocated,
but I guess it makes portability easier.

Juergen Gross, on Fri 05 Aug 2016 19:36:00 +0200, wrote:
> diff --git a/balloon.c b/balloon.c
> index 4c18c5c..75b87c8 100644
> --- a/balloon.c
> +++ b/balloon.c
> @@ -44,3 +44,20 @@ void get_max_pages(void)
>      nr_max_pages = ret;
>      printk("Maximum memory size: %ld pages\n", nr_max_pages);
>  }
> +
> +void alloc_bitmap_remap(void)
> +{
> +    unsigned long i;
> +
> +    if ( alloc_bitmap_size >= ((nr_max_pages + 1) >> (PAGE_SHIFT + 3)) )
> +        return;
> +
> +    for ( i = 0; i < alloc_bitmap_size; i += PAGE_SIZE )
> +    {
> +        map_frame_rw(virt_kernel_area_end + i,
> +                     virt_to_mfn((unsigned long)(alloc_bitmap) + i));
> +    }
> +
> +    alloc_bitmap = (unsigned long *)virt_kernel_area_end;
> +    virt_kernel_area_end += round_pgup((nr_max_pages + 1) >> (PAGE_SHIFT + 3));

Ditto here, better check against hitting VIRT_DEMAND_AREA.

> diff --git a/include/balloon.h b/include/balloon.h
> index b8d9335..0e2340b 100644
> --- a/include/balloon.h
> +++ b/include/balloon.h
> @@ -31,11 +31,13 @@ extern unsigned long virt_kernel_area_end;
>  
>  void get_max_pages(void);
>  void arch_remap_p2m(unsigned long max_pfn);
> +void alloc_bitmap_remap(void);
>  
>  #else /* CONFIG_BALLOON */
>  
>  static inline void get_max_pages(void) { }
>  static inline void arch_remap_p2m(unsigned long max_pfn) { }
> +static inline void alloc_bitmap_remap(void) { }

I'd say call it rather mm_alloc_bitmap_remap(). We have C namespace
issues with the stubdom applications, and the alloc_bitmap_ prefix seems
quite generic (even if less that bitmap_), mm_ adds some kernelish
notion.

>  extern unsigned long nr_free_pages;
>  
> +extern unsigned long *alloc_bitmap;
> +extern unsigned long alloc_bitmap_size;

Ditto, mm_bitmap and mm_bitmap_size.

Otherwise it looks good.

Samuel

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