[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] [PATCH 07/10] xen/mmu: Work with 1-1 mappings when allocating new top/middle entries.



On Tue, 2010-12-21 at 21:37 +0000, Konrad Rzeszutek Wilk wrote: 
> We were not properly taking under advisement the 1-1 mappings
> when a large area of memory was ballooned out.

Are we lazily allocating the p2m tree nodes for regions initially
covered by the balloon? (perhaps we have always done this and it isn't
new with this series)

Would it be simpley to always populate enough tree nodes to cover the
ballooned area as well as nr_pages at start of day and therefore avoid
worrying about it later on (except for memory hotplug which is special
in this way already)?

Ian.

> 
> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@xxxxxxxxxx>
> ---
>  arch/x86/xen/mmu.c |   13 ++++++++-----
>  1 files changed, 8 insertions(+), 5 deletions(-)
> 
> diff --git a/arch/x86/xen/mmu.c b/arch/x86/xen/mmu.c
> index bd02e7d..92f4fec 100644
> --- a/arch/x86/xen/mmu.c
> +++ b/arch/x86/xen/mmu.c
> @@ -464,7 +464,7 @@ static void free_p2m_page(void *p)
>  static bool alloc_p2m(unsigned long pfn)
>  {
>       unsigned topidx, mididx;
> -     unsigned long ***top_p, **mid;
> +     unsigned long ***top_p, **mid, **mid_orig;
>       unsigned long *top_mfn_p, *mid_mfn;
>  
>       topidx = p2m_top_index(pfn);
> @@ -473,15 +473,16 @@ static bool alloc_p2m(unsigned long pfn)
>       top_p = &p2m_top[topidx];
>       mid = *top_p;
>  
> -     if (mid == p2m_mid_identity) {
> +     if (mid == p2m_mid_identity || mid == p2m_mid_missing) {
>               /* Mid level is missing, allocate a new one */
> +             mid_orig = mid;
>               mid = alloc_p2m_page();
>               if (!mid)
>                       return false;
>  
>               p2m_mid_init(mid, p2m_identity);
>  
> -             if (cmpxchg(top_p, p2m_mid_identity, mid) != p2m_mid_identity)
> +             if (cmpxchg(top_p, mid_orig, mid) != mid_orig)
>                       free_p2m_page(mid);
>       }
>  
> @@ -510,9 +511,11 @@ static bool alloc_p2m(unsigned long pfn)
>                       p2m_top_mfn_p[topidx] = mid_mfn;
>       }
>  
> -     if (p2m_top[topidx][mididx] == p2m_identity) {
> +     if (p2m_top[topidx][mididx] == p2m_identity ||
> +         p2m_top[topidx][mididx] == p2m_missing) {
>               /* p2m leaf page is missing */
>               unsigned long *p2m;
> +             unsigned long *p2m_orig = p2m_top[topidx][mididx];
>  
>               p2m = alloc_p2m_page();
>               if (!p2m)
> @@ -520,7 +523,7 @@ static bool alloc_p2m(unsigned long pfn)
>  
>               p2m_init(p2m);
>  
> -             if (cmpxchg(&mid[mididx], p2m_identity, p2m) != p2m_identity)
> +             if (cmpxchg(&mid[mididx], p2m_orig, p2m) != p2m_orig)
>                       free_p2m_page(p2m);
>               else
>                       mid_mfn[mididx] = virt_to_mfn(p2m);



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.