x86: Speed up PV-guest superpage mapping

The current version of superpage mapping takes a PGT_writable
reference to every page in a superpage each time it is mapped.  This
is extremely slow, so slow that applications become unusable.

My solution for this is to introduce a superpage table in the
hypervisor, similar to the frametable structure for pages.  Currently
this table only has a type_info element.  There are three types a
superpage can have, SGT_mark, SGT_dynamic, or SGT_none.

In normal operation, the first time a superpage is mapped, a
PGT_writable reference is taken to each page in the superpage, and the
superpage is set to type SGT_dynamic and the superpage typecount is
incremented.  On subsequent mappings and unmappings, only the
superpage typecount changes.  On the last unmap, the PGT_writable
reference on each page is removed.

The SGT_mark type is set and cleared through two new MMUEXT
hypercalls, mark_super and unmark_super.  When the hypercall is made,
the superpage's type is set to SGT_mark and a PGT_writable reference
is taken to its pages.  On unmark, the type is cleared and the
reference removed.

If a page is already set to SGT_dynamic when mark_super is called, the
type is changed to SGT_mark and no additional PGT_writable reference
is taken.  If there are still outstanding mappings of this superpage
when unmark_super is called, the type is set to SGT_dynamic and the
PGT_writable reference is not removed.

Fast superpage mapping is only supported on 64 bit hypervisors.  For
32 bit hyperviors, superpage mapping is supported but will be
extremely slow.

Signed-off-by: Dave McCracken <dave.mccracken@oracle.com>

1 files changed, 23 insertions, 1 deletions

diff --git a/xen/include/asm-x86/mm.h b/xen/include/asm-x86/mm.h
index 26a1a680f2..f1d6c8dc1c 100644
--- a/xen/include/asm-x86/mm.h
+++ b/xen/include/asm-x86/mm.h
@@ -214,6 +214,23 @@ struct page_info
 #define PGC_count_width   PG_shift(9)
 #define PGC_count_mask    ((1UL<<PGC_count_width)-1)
 
+#ifdef __x86_64__
+struct spage_info
+{
+       unsigned long type_info;
+};
+
+ /* The following page types are MUTUALLY EXCLUSIVE. */
+#define SGT_none          PG_mask(0, 2)  /* superpage not in use */
+#define SGT_mark          PG_mask(1, 2)  /* Marked as a superpage */
+#define SGT_dynamic       PG_mask(2, 2)  /* has been dynamically mapped as a superpage */
+#define SGT_type_mask     PG_mask(3, 2)  /* Bits 30-31 or 62-63. */
+
+ /* Count of uses of this superpage as its current type. */
+#define SGT_count_width   PG_shift(3)
+#define SGT_count_mask    ((1UL<<SGT_count_width)-1)
+#endif
+
 #if defined(__i386__)
 #define is_xen_heap_page(page) is_xen_heap_mfn(page_to_mfn(page))
 #define is_xen_heap_mfn(mfn) ({                         \
@@ -262,6 +279,9 @@ extern void share_xen_page_with_privileged_guests(
     struct page_info *page, int readonly);
 
 #define frame_table ((struct page_info *)FRAMETABLE_VIRT_START)
+#ifdef __x86_64__
+#define spage_table ((struct spage_info *)SPAGETABLE_VIRT_START)
+#endif
 extern unsigned long max_page;
 extern unsigned long total_pages;
 void init_frametable(void);
@@ -305,6 +325,8 @@ void cleanup_page_cacheattr(struct page_info *page);
 
 int is_iomem_page(unsigned long mfn);
 
+void clear_superpage_mark(struct page_info *page);
+
 struct domain *page_get_owner_and_reference(struct page_info *page);
 void put_page(struct page_info *page);
 int  get_page(struct page_info *page, struct domain *domain);
@@ -370,7 +392,7 @@ pae_copy_root(struct vcpu *v, l3_pgentry_t *l3tab);
 
 int check_descriptor(const struct domain *, struct desc_struct *d);
 
-extern int opt_allow_hugepage;
+extern int opt_allow_superpage;
 extern int mem_hotplug;
 
 /******************************************************************************