Re: [Xen-devel] [PATCH v2 2/4] libxc: support of linear p2m list for migration of pv-domains

[Andrew Cooper <andrew.cooper3@xxxxxxxxxx>]

On 15/12/15 06:33, Juergen Gross wrote:
> In order to be able to migrate pv-domains with more than 512 GB of RAM
> the p2m information can be specified by the guest kernel via a virtual
> mapped linear p2m list instead of a 3 level tree.
>
> Add support for this new p2m format in libxc.
>
> As the sanity checking of the virtual p2m address needs defines for the
> xen regions use those defines when doing page table checks as well.
>
> Signed-off-by: Juergen Gross <jgross@xxxxxxxx>

Reviewed-by: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>, with a few
comments/suggestions inline.

> ---
>  tools/libxc/xc_sr_save_x86_pv.c | 193 
> ++++++++++++++++++++++++++++++++++++++--
>  1 file changed, 184 insertions(+), 9 deletions(-)
>
> diff --git a/tools/libxc/xc_sr_save_x86_pv.c b/tools/libxc/xc_sr_save_x86_pv.c
> index d7acd37..98e9011 100644
> --- a/tools/libxc/xc_sr_save_x86_pv.c
> +++ b/tools/libxc/xc_sr_save_x86_pv.c
> @@ -3,6 +3,18 @@
>  
>  #include "xc_sr_common_x86_pv.h"
>  
> +/* Check a 64 bit virtual address for being canonical. */
> +static inline bool is_canonical_address(xen_vaddr_t vaddr)
> +{
> +    return ((int64_t)vaddr >> 47) == ((int64_t)vaddr >> 63);
> +}
> +
> +#define HYPERVISOR_VIRT_START_X86_64 0xFFFF800000000000ULL
> +#define HYPERVISOR_VIRT_END_X86_64   0xFFFF87FFFFFFFFFFULL
> +
> +#define HYPERVISOR_VIRT_START_X86_32 0x00000000F5800000ULL
> +#define HYPERVISOR_VIRT_END_X86_32   0x00000000FFFFFFFFULL

These might be better in the x86_pv header file, as they could plausibly
be useful on the restore side as well.

Ideally the virtual ranges would come from the Xen public header files,
but that would involve a non-trivial modifications to be able to get the
all the information.

> +
>  /*
>   * Maps the guests shared info page.
>   */
> @@ -116,7 +128,7 @@ static int map_p2m_leaves(struct xc_sr_context *ctx, 
> xen_pfn_t *mfns,
>   * frames making up the guests p2m table.  Construct a list of pfns making up
>   * the table.
>   */
> -static int map_p2m(struct xc_sr_context *ctx)
> +static int map_p2m_tree(struct xc_sr_context *ctx)
>  {
>      /* Terminology:
>       *
> @@ -681,8 +855,8 @@ static int normalise_pagetable(struct xc_sr_context *ctx, 
> const uint64_t *src,
>          /* 64bit guests only have Xen mappings in their L4 tables. */
>          if ( type == XEN_DOMCTL_PFINFO_L4TAB )
>          {
> -            xen_first = 256;
> -            xen_last = 271;
> +            xen_first = (HYPERVISOR_VIRT_START_X86_64 >> 39) & 511;
> +            xen_last = (HYPERVISOR_VIRT_END_X86_64 >> 39) & 511;

Would you mind using appropriate PAGETABLE_SHIFT defines from xg_private.h ?

>          }
>      }
>      else
> @@ -698,7 +872,7 @@ static int normalise_pagetable(struct xc_sr_context *ctx, 
> const uint64_t *src,
>              /* 32bit guests can only use the first 4 entries of their L3 
> tables.
>               * All other are potentially used by Xen. */
>              xen_first = 4;
> -            xen_last = 512;
> +            xen_last = 511;
>              break;
>  
>          case XEN_DOMCTL_PFINFO_L2TAB:
> @@ -709,10 +883,11 @@ static int normalise_pagetable(struct xc_sr_context 
> *ctx, const uint64_t *src,
>               *
>               * ...which is conveniently unavailable to us in a 64bit build.
>               */

The comment partially in context is now stale.

~Andrew

> -            if ( pte_to_frame(src[428]) == ctx->x86_pv.compat_m2p_mfn0 )
> +            i = (HYPERVISOR_VIRT_START_X86_32 >> 21) & 511;
> +            if ( pte_to_frame(src[i]) == ctx->x86_pv.compat_m2p_mfn0 )
>              {
> -                xen_first = 428;
> -                xen_last = 512;
> +                xen_first = i;
> +                xen_last = (HYPERVISOR_VIRT_END_X86_32 >> 21) & 511;
>              }
>              break;
>          }


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