x86/PoD: Eager sweep for zeroed pages
Based on the contents of a guests physical address space,
p2m_pod_emergency_sweep() could degrade into a linear memcmp() from 0 to
max_gfn, which runs non-preemptibly.
As p2m_pod_emergency_sweep() runs behind the scenes in a number of contexts,
making it preemptible is not feasible.
Instead, a different approach is taken. Recently-populated pages are eagerly
checked for reclaimation, which amortises the p2m_pod_emergency_sweep()
operation across each p2m_pod_demand_populate() operation.
Note that in the case that a 2M superpage can't be reclaimed as a superpage,
it is shattered if 4K pages of zeros can be reclaimed. This is unfortunate
but matches the previous behaviour, and is required to avoid regressions
(domain crash from PoD exhaustion) with VMs configured close to the limit.
This is CVE-2015-7970 / XSA-150.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: George Dunlap <george.dunlap@citrix.com>
--- a/xen/arch/x86/mm/p2m-pod.c
+++ b/xen/arch/x86/mm/p2m-pod.c
@@ -920,28 +920,6 @@ p2m_pod_zero_check(struct p2m_domain *p2
}
#define POD_SWEEP_LIMIT 1024
-
-/* When populating a new superpage, look at recently populated superpages
- * hoping that they've been zeroed. This will snap up zeroed pages as soon as
- * the guest OS is done with them. */
-static void
-p2m_pod_check_last_super(struct p2m_domain *p2m, unsigned long gfn_aligned)
-{
- unsigned long check_gfn;
-
- ASSERT(p2m->pod.last_populated_index < POD_HISTORY_MAX);
-
- check_gfn = p2m->pod.last_populated[p2m->pod.last_populated_index];
-
- p2m->pod.last_populated[p2m->pod.last_populated_index] = gfn_aligned;
-
- p2m->pod.last_populated_index =
- ( p2m->pod.last_populated_index + 1 ) % POD_HISTORY_MAX;
-
- p2m_pod_zero_check_superpage(p2m, check_gfn);
-}
-
-
#define POD_SWEEP_STRIDE 16
static void
p2m_pod_emergency_sweep(struct p2m_domain *p2m)
@@ -982,7 +960,7 @@ p2m_pod_emergency_sweep(struct p2m_domai
* NB that this is a zero-sum game; we're increasing our cache size
* by re-increasing our 'debt'. Since we hold the pod lock,
* (entry_count - count) must remain the same. */
- if ( p2m->pod.count > 0 && i < limit )
+ if ( i < limit && (p2m->pod.count > 0 || hypercall_preempt_check()) )
break;
}
@@ -994,6 +972,58 @@ p2m_pod_emergency_sweep(struct p2m_domai
}
+static void pod_eager_reclaim(struct p2m_domain *p2m)
+{
+ struct pod_mrp_list *mrp = &p2m->pod.mrp;
+ unsigned int i = 0;
+
+ /*
+ * Always check one page for reclaimation.
+ *
+ * If the PoD pool is empty, keep checking some space is found, or all
+ * entries have been exhaused.
+ */
+ do
+ {
+ unsigned int idx = (mrp->idx + i++) % ARRAY_SIZE(mrp->list);
+ unsigned long gfn = mrp->list[idx];
+
+ if ( gfn != INVALID_GFN )
+ {
+ if ( gfn & POD_LAST_SUPERPAGE )
+ {
+ gfn &= ~POD_LAST_SUPERPAGE;
+
+ if ( p2m_pod_zero_check_superpage(p2m, gfn) == 0 )
+ {
+ unsigned int x;
+
+ for ( x = 0; x < SUPERPAGE_PAGES; ++x, ++gfn )
+ p2m_pod_zero_check(p2m, &gfn, 1);
+ }
+ }
+ else
+ p2m_pod_zero_check(p2m, &gfn, 1);
+
+ mrp->list[idx] = INVALID_GFN;
+ }
+
+ } while ( (p2m->pod.count == 0) && (i < ARRAY_SIZE(mrp->list)) );
+}
+
+static void pod_eager_record(struct p2m_domain *p2m,
+ unsigned long gfn, unsigned int order)
+{
+ struct pod_mrp_list *mrp = &p2m->pod.mrp;
+
+ ASSERT(mrp->list[mrp->idx] == INVALID_GFN);
+ ASSERT(gfn != INVALID_GFN);
+
+ mrp->list[mrp->idx++] =
+ gfn | (order == PAGE_ORDER_2M ? POD_LAST_SUPERPAGE : 0);
+ mrp->idx %= ARRAY_SIZE(mrp->list);
+}
+
int
p2m_pod_demand_populate(struct p2m_domain *p2m, unsigned long gfn,
unsigned int order,
@@ -1034,6 +1064,8 @@ p2m_pod_demand_populate(struct p2m_domai
return 0;
}
+ pod_eager_reclaim(p2m);
+
/* Only sweep if we're actually out of memory. Doing anything else
* causes unnecessary time and fragmentation of superpages in the p2m. */
if ( p2m->pod.count == 0 )
@@ -1070,6 +1102,8 @@ p2m_pod_demand_populate(struct p2m_domai
p2m->pod.entry_count -= (1 << order);
BUG_ON(p2m->pod.entry_count < 0);
+ pod_eager_record(p2m, gfn_aligned, order);
+
if ( tb_init_done )
{
struct {
@@ -1085,12 +1119,6 @@ p2m_pod_demand_populate(struct p2m_domai
__trace_var(TRC_MEM_POD_POPULATE, 0, sizeof(t), &t);
}
- /* Check the last guest demand-populate */
- if ( p2m->pod.entry_count > p2m->pod.count
- && (order == PAGE_ORDER_2M)
- && (q & P2M_ALLOC) )
- p2m_pod_check_last_super(p2m, gfn_aligned);
-
pod_unlock(p2m);
return 0;
out_of_memory:
--- a/xen/arch/x86/mm/p2m.c
+++ b/xen/arch/x86/mm/p2m.c
@@ -58,6 +58,7 @@ boolean_param("hap_2mb", opt_hap_2mb);
/* Init the datastructures for later use by the p2m code */
static int p2m_initialise(struct domain *d, struct p2m_domain *p2m)
{
+ unsigned int i;
int ret = 0;
mm_rwlock_init(&p2m->lock);
@@ -73,6 +74,9 @@ static int p2m_initialise(struct domain
p2m->np2m_base = P2M_BASE_EADDR;
+ for ( i = 0; i < ARRAY_SIZE(p2m->pod.mrp.list); ++i )
+ p2m->pod.mrp.list[i] = INVALID_GFN;
+
if ( hap_enabled(d) && cpu_has_vmx )
ret = ept_p2m_init(p2m);
else
--- a/xen/include/asm-x86/p2m.h
+++ b/xen/include/asm-x86/p2m.h
@@ -292,10 +292,20 @@ struct p2m_domain {
entry_count; /* # of pages in p2m marked pod */
unsigned long reclaim_single; /* Last gpfn of a scan */
unsigned long max_guest; /* gpfn of max guest demand-populate */
-#define POD_HISTORY_MAX 128
- /* gpfn of last guest superpage demand-populated */
- unsigned long last_populated[POD_HISTORY_MAX];
- unsigned int last_populated_index;
+
+ /*
+ * Tracking of the most recently populated PoD pages, for eager
+ * reclamation.
+ */
+ struct pod_mrp_list {
+#define NR_POD_MRP_ENTRIES 32
+
+/* Encode ORDER_2M superpage in top bit of GFN */
+#define POD_LAST_SUPERPAGE (INVALID_GFN & ~(INVALID_GFN >> 1))
+
+ unsigned long list[NR_POD_MRP_ENTRIES];
+ unsigned int idx;
+ } mrp;
mm_lock_t lock; /* Locking of private pod structs, *
* not relying on the p2m lock. */
} pod;