[PATCHv12 0/5] gnttab: Improve scaleability

[PATCHv12 0/5] gnttab: Improve scaleability

[David Vrabel]

The series builds on the original series by Matt Wilson and Christoph
Egger from Amazon.

Performance results for aggregate intrahost network throughput
(between 20 VM pairs, with 16 dom0 VCPUs) show substantial
improvements.

                             Throughput/Gbit/s

Base                          9.7
Split locks                  25.8
Split locks                  42.9
 + per-VCPU maptrack lists

v12:
  * Check for same refs in __gnttab_swap_grant_ref() to avoid deadlock.
  * Assert local grant table lock is held in mapcount().
  * Document that maptrack_lock is only for the free list.
  * Document locking needed for maptrack entries (the active entry
    lock).
  * Try and steal maptrack entries if a VCPU runs out (RFC).

v11:
  * Fix exists test in gnttab_map_exists().
  * Call gnttab_need_iommu_mapping() once for each lock/unlock.
  * Take active entry lock in gnttab_transfer().

  * Use read_atomic() when checking maptrack flags for validity (note:
    all other reads of map->flags are either after it is validated as
    stable or protected by the active entry lock and thus don't need
    read_atomic()).
  * Add comment to double_gt_lock().
  * Allocate maptrack array with vzalloc() since it may be >1 page.

v10:
  * Reduce scope of act in grant_map_exists().
  * Make unlock sequence in error paths consistent in
    __acquire_grant_for_copy().
  * gnt_unlock_out -> to gt_unlock_out in __acquire_grant_for_copy().
  * In gnttab_map_grant_ref(), keep double lock around maptrack update
    if gnttab_need_iommu_mapping().  Use a wmb(), otherwise.
  * Divide max_maptrack_frames evenly amongst the VCPUs.
  * Increase default max_maptrack_frames to compensate.

v9:
  * Refactor the locking patches into three commits:
      1. Add per-active entry locking.
      2. Add maptrack_lock.
      3. Make the grant table lock a read-write lock.
  * Keep the double lock around IOMMU updates, but adjust their scope
    to only the IOMMU updates.

v8:
  * Misc locking fixes.
  * Use read/write_atomic() in per-VCPU maptrack free list.
  * Improve locking docs.

v7:
  * Re-add grant table rwlock (locking was broken without it and the
    gains are minimal anyway).
  * Remove unneeded locks from grant_table_destroy().
  * Fix get_maptrack_handle() locking.
v6:
  * Remove most uses of the grant table lock.
  * Make the grant table lock a spin lock again (there were only
    writers left after the above)
  * Add per-VCPU maptrack free lists.
v5:
  * Addressed locking issue pointed out by Jan Beulich
  * Fixed git rebase merge issue introduced in v4
    (acquiring locking twice)
  * Change for ()-loop in grant_map_exists
  * Coding style fixes
v4:
  * Coding style nits from Jan Beulich
  * Fixup read locks pointed out by Jan Beulich
  * renamed double_gt_(un)lock to double_maptrack_(un)lock
    per request from Jan Beulich
  * Addressed ASSERT()'s from Jan Beulich
  * Addressed locking issues in unmap_common pointed out
    by Jan Beulich
v3:
  * Addressed gnttab_swap_grant_ref() comment from Andrew Cooper
v2:
  * Add arm part per request from Julien Grall

David

[PATCHv12 1/5] gnttab: per-active entry locking

[David Vrabel]

Introduce a per-active entry spin lock to protect active entry state
The grant table lock must be locked before acquiring (locking) an
active entry.

This is a step in reducing contention on the grant table lock, but
will only do so once the grant table lock is turned into a read-write
lock.

Based on a patch originally by Matt Wilson <msw@amazon.com>.

Signed-off-by: David Vrabel <david.vrabel@citrix.com>
---
v12:
- Check for same refs in __gnttab_swap_grant_ref() to avoid deadlock.
- Assert local grant table lock is held in mapcount().

v11:
- Fix exists test in gnttab_map_exists().

v10:
- Reduce scope of act in grant_map_exists().
- Make unlock sequence in error paths consistent in
  __acquire_grant_for_copy().
- gnt_unlock_out -> to gt_unlock_out in __acquire_grant_for_copy().
---
 docs/misc/grant-tables.txt |   42 +++++++++++-
 xen/common/grant_table.c   |  155 ++++++++++++++++++++++++++++++++------------
 2 files changed, 156 insertions(+), 41 deletions(-)

diff --git a/docs/misc/grant-tables.txt b/docs/misc/grant-tables.txt
index 19db4ec..2a3f591 100644
--- a/docs/misc/grant-tables.txt
+++ b/docs/misc/grant-tables.txt
@@ -63,6 +63,7 @@ is complete.
   act->domid : remote domain being granted rights
   act->frame : machine frame being granted
   act->pin   : used to hold reference counts
+  act->lock  : spinlock used to serialize access to active entry state
 
  Map tracking
  ~~~~~~~~~~~~
@@ -74,7 +75,46 @@ is complete.
  matching map track entry is then removed, as if unmap had been invoked.
  These are not used by the transfer mechanism.
   map->domid         : owner of the mapped frame
-  map->ref_and_flags : grant reference, ro/rw, mapped for host or device access
+  map->ref           : grant reference
+  map->flags         : ro/rw, mapped for host or device access
+
+********************************************************************************
+ Locking
+ ~~~~~~~
+ Xen uses several locks to serialize access to the internal grant table state.
+
+  grant_table->lock          : lock used to prevent readers from accessing
+                               inconsistent grant table state such as current
+                               version, partially initialized active table pages,
+                               etc.
+  active_grant_entry->lock   : spinlock used to serialize modifications to
+                               active entries
+
+ The primary lock for the grant table is a spinlock. All functions
+ that access members of struct grant_table must acquire the lock
+ around critical sections.
+
+ Active entries are obtained by calling active_entry_acquire(gt, ref).
+ This function returns a pointer to the active entry after locking its
+ spinlock. The caller must hold the grant table lock for the gt in
+ question before calling active_entry_acquire(). This is because the
+ grant table can be dynamically extended via gnttab_grow_table() while
+ a domain is running and must be fully initialized. Once all access to
+ the active entry is complete, release the lock by calling
+ active_entry_release(act).
+
+ Summary of rules for locking:
+  active_entry_acquire() and active_entry_release() can only be
+  called when holding the relevant grant table's lock. I.e.:
+    spin_lock(&gt->lock);
+    act = active_entry_acquire(gt, ref);
+    ...
+    active_entry_release(act);
+    spin_unlock(&gt->lock);
+
+ Active entries cannot be acquired while holding the maptrack lock.
+ Multiple active entries can be acquired while holding the grant table
+ lock.
 
 ********************************************************************************
 
diff --git a/xen/common/grant_table.c b/xen/common/grant_table.c
index ca4c973..917b46b 100644
--- a/xen/common/grant_table.c
+++ b/xen/common/grant_table.c
@@ -157,10 +157,13 @@ struct active_grant_entry {
                                in the page.                           */
     unsigned      length:16; /* For sub-page grants, the length of the
                                 grant.                                */
+    spinlock_t    lock;      /* lock to protect access of this entry.
+                                see docs/misc/grant-tables.txt for
+                                locking protocol                      */
 };
 
 #define ACGNT_PER_PAGE (PAGE_SIZE / sizeof(struct active_grant_entry))
-#define active_entry(t, e) \
+#define _active_entry(t, e) \
     ((t)->active[(e)/ACGNT_PER_PAGE][(e)%ACGNT_PER_PAGE])
 
 static inline void gnttab_flush_tlb(const struct domain *d)
@@ -188,6 +191,24 @@ nr_active_grant_frames(struct grant_table *gt)
     return num_act_frames_from_sha_frames(nr_grant_frames(gt));
 }
 
+static inline struct active_grant_entry *
+active_entry_acquire(struct grant_table *t, grant_ref_t e)
+{
+    struct active_grant_entry *act;
+
+    ASSERT(spin_is_locked(&t->lock));
+
+    act = &_active_entry(t, e);
+    spin_lock(&act->lock);
+
+    return act;
+}
+
+static inline void active_entry_release(struct active_grant_entry *act)
+{
+    spin_unlock(&act->lock);
+}
+
 /* Check if the page has been paged out, or needs unsharing. 
    If rc == GNTST_okay, *page contains the page struct with a ref taken.
    Caller must do put_page(*page).
@@ -505,7 +526,6 @@ static int grant_map_exists(const struct domain *ld,
                             unsigned long mfn,
                             unsigned int *ref_count)
 {
-    const struct active_grant_entry *act;
     unsigned int ref, max_iter;
     
     ASSERT(spin_is_locked(&rgt->lock));
@@ -514,18 +534,19 @@ static int grant_map_exists(const struct domain *ld,
                    nr_grant_entries(rgt));
     for ( ref = *ref_count; ref < max_iter; ref++ )
     {
-        act = &active_entry(rgt, ref);
+        struct active_grant_entry *act;
+        bool_t exists;
 
-        if ( !act->pin )
-            continue;
+        act = active_entry_acquire(rgt, ref);
 
-        if ( act->domid != ld->domain_id )
-            continue;
+        exists = act->pin
+            && act->domid == ld->domain_id
+            && act->frame == mfn;
 
-        if ( act->frame != mfn )
-            continue;
+        active_entry_release(act);
 
-        return 0;
+        if ( exists )
+            return 0;
     }
 
     if ( ref < nr_grant_entries(rgt) )
@@ -546,13 +567,24 @@ static void mapcount(
 
     *wrc = *rdc = 0;
 
+    /*
+     * Must have the local domain's grant table lock when iterating
+     * over its maptrack entries.
+     */
+    ASSERT(spin_is_locked(&lgt->lock));
+    /*
+     * Must have the remote domain's grant table lock while counting
+     * its active entries.
+     */
+    ASSERT(spin_is_locked(&rd->grant_table->lock));
+
     for ( handle = 0; handle < lgt->maptrack_limit; handle++ )
     {
         map = &maptrack_entry(lgt, handle);
         if ( !(map->flags & (GNTMAP_device_map|GNTMAP_host_map)) ||
              map->domid != rd->domain_id )
             continue;
-        if ( active_entry(rd->grant_table, map->ref).frame == mfn )
+        if ( _active_entry(rd->grant_table, map->ref).frame == mfn )
             (map->flags & GNTMAP_readonly) ? (*rdc)++ : (*wrc)++;
     }
 }
@@ -639,7 +671,7 @@ __gnttab_map_grant_ref(
     if ( unlikely(op->ref >= nr_grant_entries(rgt)))
         PIN_FAIL(unlock_out, GNTST_bad_gntref, "Bad ref (%d).\n", op->ref);
 
-    act = &active_entry(rgt, op->ref);
+    act = active_entry_acquire(rgt, op->ref);
     shah = shared_entry_header(rgt, op->ref);
     if (rgt->gt_version == 1) {
         sha1 = &shared_entry_v1(rgt, op->ref);
@@ -656,7 +688,7 @@ __gnttab_map_grant_ref(
          ((act->domid != ld->domain_id) ||
           (act->pin & 0x80808080U) != 0 ||
           (act->is_sub_page)) )
-        PIN_FAIL(unlock_out, GNTST_general_error,
+        PIN_FAIL(act_release_out, GNTST_general_error,
                  "Bad domain (%d != %d), or risk of counter overflow %08x, or subpage %d\n",
                  act->domid, ld->domain_id, act->pin, act->is_sub_page);
 
@@ -667,7 +699,7 @@ __gnttab_map_grant_ref(
         if ( (rc = _set_status(rgt->gt_version, ld->domain_id,
                                op->flags & GNTMAP_readonly,
                                1, shah, act, status) ) != GNTST_okay )
-             goto unlock_out;
+            goto act_release_out;
 
         if ( !act->pin )
         {
@@ -702,6 +734,7 @@ __gnttab_map_grant_ref(
 
     cache_flags = (shah->flags & (GTF_PAT | GTF_PWT | GTF_PCD) );
 
+    active_entry_release(act);
     spin_unlock(&rgt->lock);
 
     /* pg may be set, with a refcount included, from __get_paged_frame */
@@ -839,7 +872,7 @@ __gnttab_map_grant_ref(
 
     spin_lock(&rgt->lock);
 
-    act = &active_entry(rgt, op->ref);
+    act = active_entry_acquire(rgt, op->ref);
 
     if ( op->flags & GNTMAP_device_map )
         act->pin -= (op->flags & GNTMAP_readonly) ?
@@ -856,6 +889,9 @@ __gnttab_map_grant_ref(
     if ( !act->pin )
         gnttab_clear_flag(_GTF_reading, status);
 
+ act_release_out:
+    active_entry_release(act);
+
  unlock_out:
     spin_unlock(&rgt->lock);
     op->status = rc;
@@ -950,7 +986,7 @@ __gnttab_unmap_common(
     }
 
     op->rd = rd;
-    act = &active_entry(rgt, op->map->ref);
+    act = active_entry_acquire(rgt, op->map->ref);
 
     if ( op->frame == 0 )
     {
@@ -959,7 +995,7 @@ __gnttab_unmap_common(
     else
     {
         if ( unlikely(op->frame != act->frame) )
-            PIN_FAIL(unmap_out, GNTST_general_error,
+            PIN_FAIL(act_release_out, GNTST_general_error,
                      "Bad frame number doesn't match gntref. (%lx != %lx)\n",
                      op->frame, act->frame);
         if ( op->flags & GNTMAP_device_map )
@@ -978,7 +1014,7 @@ __gnttab_unmap_common(
         if ( (rc = replace_grant_host_mapping(op->host_addr,
                                               op->frame, op->new_addr, 
                                               op->flags)) < 0 )
-            goto unmap_out;
+            goto act_release_out;
 
         ASSERT(act->pin & (GNTPIN_hstw_mask | GNTPIN_hstr_mask));
         op->map->flags &= ~GNTMAP_host_map;
@@ -1000,7 +1036,7 @@ __gnttab_unmap_common(
         if ( err )
         {
             rc = GNTST_general_error;
-            goto unmap_out;
+            goto act_release_out;
         }
     }
 
@@ -1008,8 +1044,11 @@ __gnttab_unmap_common(
     if ( !(op->flags & GNTMAP_readonly) )
          gnttab_mark_dirty(rd, op->frame);
 
+ act_release_out:
+    active_entry_release(act);
  unmap_out:
     double_gt_unlock(lgt, rgt);
+
     op->status = rc;
     rcu_unlock_domain(rd);
 }
@@ -1042,9 +1081,9 @@ __gnttab_unmap_common_complete(struct gnttab_unmap_common *op)
     spin_lock(&rgt->lock);
 
     if ( rgt->gt_version == 0 )
-        goto unmap_out;
+        goto unlock_out;
 
-    act = &active_entry(rgt, op->map->ref);
+    act = active_entry_acquire(rgt, op->map->ref);
     sha = shared_entry_header(rgt, op->map->ref);
 
     if ( rgt->gt_version == 1 )
@@ -1058,7 +1097,7 @@ __gnttab_unmap_common_complete(struct gnttab_unmap_common *op)
          * Suggests that __gntab_unmap_common failed early and so
          * nothing further to do
          */
-        goto unmap_out;
+        goto act_release_out;
     }
 
     pg = mfn_to_page(op->frame);
@@ -1082,7 +1121,7 @@ __gnttab_unmap_common_complete(struct gnttab_unmap_common *op)
              * Suggests that __gntab_unmap_common failed in
              * replace_grant_host_mapping() so nothing further to do
              */
-            goto unmap_out;
+            goto act_release_out;
         }
 
         if ( !is_iomem_page(op->frame) ) 
@@ -1103,8 +1142,11 @@ __gnttab_unmap_common_complete(struct gnttab_unmap_common *op)
     if ( act->pin == 0 )
         gnttab_clear_flag(_GTF_reading, status);
 
- unmap_out:
+ act_release_out:
+    active_entry_release(act);
+ unlock_out:
     spin_unlock(&rgt->lock);
+
     if ( put_handle )
     {
         op->map->flags = 0;
@@ -1296,7 +1338,7 @@ gnttab_grow_table(struct domain *d, unsigned int req_nr_frames)
     /* d's grant table lock must be held by the caller */
 
     struct grant_table *gt = d->grant_table;
-    unsigned int i;
+    unsigned int i, j;
 
     ASSERT(req_nr_frames <= max_grant_frames);
 
@@ -1311,6 +1353,8 @@ gnttab_grow_table(struct domain *d, unsigned int req_nr_frames)
         if ( (gt->active[i] = alloc_xenheap_page()) == NULL )
             goto active_alloc_failed;
         clear_page(gt->active[i]);
+        for ( j = 0; j < ACGNT_PER_PAGE; j++ )
+            spin_lock_init(&gt->active[i][j].lock);
     }
 
     /* Shared */
@@ -1806,7 +1850,7 @@ __release_grant_for_copy(
 
     spin_lock(&rgt->lock);
 
-    act = &active_entry(rgt, gref);
+    act = active_entry_acquire(rgt, gref);
     sha = shared_entry_header(rgt, gref);
     r_frame = act->frame;
 
@@ -1845,6 +1889,7 @@ __release_grant_for_copy(
         released_read = 1;
     }
 
+    active_entry_release(act);
     spin_unlock(&rgt->lock);
 
     if ( td != rd )
@@ -1906,14 +1951,14 @@ __acquire_grant_for_copy(
     spin_lock(&rgt->lock);
 
     if ( rgt->gt_version == 0 )
-        PIN_FAIL(unlock_out, GNTST_general_error,
+        PIN_FAIL(gt_unlock_out, GNTST_general_error,
                  "remote grant table not ready\n");
 
     if ( unlikely(gref >= nr_grant_entries(rgt)) )
-        PIN_FAIL(unlock_out, GNTST_bad_gntref,
+        PIN_FAIL(gt_unlock_out, GNTST_bad_gntref,
                  "Bad grant reference %ld\n", gref);
 
-    act = &active_entry(rgt, gref);
+    act = active_entry_acquire(rgt, gref);
     shah = shared_entry_header(rgt, gref);
     if ( rgt->gt_version == 1 )
     {
@@ -1972,6 +2017,13 @@ __acquire_grant_for_copy(
                 PIN_FAIL(unlock_out_clear, GNTST_general_error,
                          "transitive grant referenced bad domain %d\n",
                          trans_domid);
+
+            /*
+             * __acquire_grant_for_copy() could take the lock on the
+             * remote table (if rd == td), so we have to drop the lock
+             * here and reacquire
+             */
+            active_entry_release(act);
             spin_unlock(&rgt->lock);
 
             rc = __acquire_grant_for_copy(td, trans_gref, rd->domain_id,
@@ -1979,9 +2031,12 @@ __acquire_grant_for_copy(
                                           &trans_page_off, &trans_length, 0);
 
             spin_lock(&rgt->lock);
+            act = active_entry_acquire(rgt, gref);
+
             if ( rc != GNTST_okay ) {
                 __fixup_status_for_copy_pin(act, status);
                 rcu_unlock_domain(td);
+                active_entry_release(act);
                 spin_unlock(&rgt->lock);
                 return rc;
             }
@@ -1994,6 +2049,7 @@ __acquire_grant_for_copy(
             {
                 __fixup_status_for_copy_pin(act, status);
                 rcu_unlock_domain(td);
+                active_entry_release(act);
                 spin_unlock(&rgt->lock);
                 put_page(*page);
                 return __acquire_grant_for_copy(rd, gref, ldom, readonly,
@@ -2062,6 +2118,7 @@ __acquire_grant_for_copy(
     *length = act->length;
     *frame = act->frame;
 
+    active_entry_release(act);
     spin_unlock(&rgt->lock);
     return rc;
  
@@ -2074,7 +2131,11 @@ __acquire_grant_for_copy(
         gnttab_clear_flag(_GTF_reading, status);
 
  unlock_out:
+    active_entry_release(act);
+
+ gt_unlock_out:
     spin_unlock(&rgt->lock);
+
     return rc;
 }
 
@@ -2374,7 +2435,6 @@ gnttab_set_version(XEN_GUEST_HANDLE_PARAM(gnttab_set_version_t) uop)
     gnttab_set_version_t op;
     struct domain *d = current->domain;
     struct grant_table *gt = d->grant_table;
-    struct active_grant_entry *act;
     grant_entry_v1_t reserved_entries[GNTTAB_NR_RESERVED_ENTRIES];
     long res;
     int i;
@@ -2399,8 +2459,7 @@ gnttab_set_version(XEN_GUEST_HANDLE_PARAM(gnttab_set_version_t) uop)
     {
         for ( i = GNTTAB_NR_RESERVED_ENTRIES; i < nr_grant_entries(gt); i++ )
         {
-            act = &active_entry(gt, i);
-            if ( act->pin != 0 )
+            if ( read_atomic(&_active_entry(gt, i).pin) != 0 )
             {
                 gdprintk(XENLOG_WARNING,
                          "tried to change grant table version from %d to %d, but some grant entries still in use\n",
@@ -2587,7 +2646,8 @@ __gnttab_swap_grant_ref(grant_ref_t ref_a, grant_ref_t ref_b)
 {
     struct domain *d = rcu_lock_current_domain();
     struct grant_table *gt = d->grant_table;
-    struct active_grant_entry *act;
+    struct active_grant_entry *act_a = NULL;
+    struct active_grant_entry *act_b = NULL;
     s16 rc = GNTST_okay;
 
     spin_lock(&gt->lock);
@@ -2601,12 +2661,16 @@ __gnttab_swap_grant_ref(grant_ref_t ref_a, grant_ref_t ref_b)
     if ( unlikely(ref_b >= nr_grant_entries(d->grant_table)))
         PIN_FAIL(out, GNTST_bad_gntref, "Bad ref-b (%d).\n", ref_b);
 
-    act = &active_entry(gt, ref_a);
-    if ( act->pin )
+    /* Swapping the same ref is a no-op. */
+    if ( ref_a == ref_b )
+        goto out;
+
+    act_a = active_entry_acquire(gt, ref_a);
+    if ( act_a->pin )
         PIN_FAIL(out, GNTST_eagain, "ref a %ld busy\n", (long)ref_a);
 
-    act = &active_entry(gt, ref_b);
-    if ( act->pin )
+    act_b = active_entry_acquire(gt, ref_b);
+    if ( act_b->pin )
         PIN_FAIL(out, GNTST_eagain, "ref b %ld busy\n", (long)ref_b);
 
     if ( gt->gt_version == 1 )
@@ -2633,6 +2697,10 @@ __gnttab_swap_grant_ref(grant_ref_t ref_a, grant_ref_t ref_b)
     }
 
 out:
+    if ( act_b != NULL )
+        active_entry_release(act_b);
+    if ( act_a != NULL )
+        active_entry_release(act_a);
     spin_unlock(&gt->lock);
 
     rcu_unlock_domain(d);
@@ -2942,7 +3010,7 @@ grant_table_create(
     struct domain *d)
 {
     struct grant_table *t;
-    int                 i;
+    unsigned int i, j;
 
     if ( (t = xzalloc(struct grant_table)) == NULL )
         goto no_mem_0;
@@ -2961,6 +3029,8 @@ grant_table_create(
         if ( (t->active[i] = alloc_xenheap_page()) == NULL )
             goto no_mem_2;
         clear_page(t->active[i]);
+        for ( j = 0; j < ACGNT_PER_PAGE; j++ )
+            spin_lock_init(&t->active[i][j].lock);
     }
 
     /* Tracking of mapped foreign frames table */
@@ -3057,7 +3127,7 @@ gnttab_release_mappings(
         rgt = rd->grant_table;
         spin_lock(&rgt->lock);
 
-        act = &active_entry(rgt, ref);
+        act = active_entry_acquire(rgt, ref);
         sha = shared_entry_header(rgt, ref);
         if (rgt->gt_version == 1)
             status = &sha->flags;
@@ -3115,6 +3185,7 @@ gnttab_release_mappings(
         if ( act->pin == 0 )
             gnttab_clear_flag(_GTF_reading, status);
 
+        active_entry_release(act);
         spin_unlock(&rgt->lock);
 
         rcu_unlock_domain(rd);
@@ -3177,9 +3248,12 @@ static void gnttab_usage_print(struct domain *rd)
         uint16_t status;
         uint64_t frame;
 
-        act = &active_entry(gt, ref);
+        act = active_entry_acquire(gt, ref);
         if ( !act->pin )
+        {
+            active_entry_release(act);
             continue;
+        }
 
         sha = shared_entry_header(gt, ref);
 
@@ -3209,6 +3283,7 @@ static void gnttab_usage_print(struct domain *rd)
         printk("[%3d]    %5d 0x%06lx 0x%08x      %5d 0x%06"PRIx64" 0x%02x\n",
                ref, act->domid, act->frame, act->pin,
                sha->domid, frame, status);
+        active_entry_release(act);
     }
 
  out:
-- 
1.7.10.4

[PATCHv12 2/5] gnttab: introduce maptrack lock

[David Vrabel]

Split grant table lock into two separate locks. One to protect
maptrack free list (maptrack_lock) and one for everything else (lock).

Based on a patch originally by Matt Wilson <msw@amazon.com>.

Signed-off-by: David Vrabel <david.vrabel@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
---
v12:
- Document that maptrack_lock is only for the free list.
---
 docs/misc/grant-tables.txt    |    4 ++++
 xen/common/grant_table.c      |    9 +++++----
 xen/include/xen/grant_table.h |    2 ++
 3 files changed, 11 insertions(+), 4 deletions(-)

diff --git a/docs/misc/grant-tables.txt b/docs/misc/grant-tables.txt
index 2a3f591..4a62bf9 100644
--- a/docs/misc/grant-tables.txt
+++ b/docs/misc/grant-tables.txt
@@ -87,6 +87,7 @@ is complete.
                                inconsistent grant table state such as current
                                version, partially initialized active table pages,
                                etc.
+  grant_table->maptrack_lock : spinlock used to protect the maptrack free list
   active_grant_entry->lock   : spinlock used to serialize modifications to
                                active entries
 
@@ -94,6 +95,9 @@ is complete.
  that access members of struct grant_table must acquire the lock
  around critical sections.
 
+ The maptrack free list is protected by its own spinlock. The maptrack
+ lock may be locked while holding the grant table lock.
+
  Active entries are obtained by calling active_entry_acquire(gt, ref).
  This function returns a pointer to the active entry after locking its
  spinlock. The caller must hold the grant table lock for the gt in
diff --git a/xen/common/grant_table.c b/xen/common/grant_table.c
index 917b46b..25da735 100644
--- a/xen/common/grant_table.c
+++ b/xen/common/grant_table.c
@@ -288,10 +288,10 @@ static inline void
 put_maptrack_handle(
     struct grant_table *t, int handle)
 {
-    spin_lock(&t->lock);
+    spin_lock(&t->maptrack_lock);
     maptrack_entry(t, handle).ref = t->maptrack_head;
     t->maptrack_head = handle;
-    spin_unlock(&t->lock);
+    spin_unlock(&t->maptrack_lock);
 }
 
 static inline int
@@ -303,7 +303,7 @@ get_maptrack_handle(
     struct grant_mapping *new_mt;
     unsigned int          new_mt_limit, nr_frames;
 
-    spin_lock(&lgt->lock);
+    spin_lock(&lgt->maptrack_lock);
 
     while ( unlikely((handle = __get_maptrack_handle(lgt)) == -1) )
     {
@@ -332,7 +332,7 @@ get_maptrack_handle(
                  nr_frames + 1);
     }
 
-    spin_unlock(&lgt->lock);
+    spin_unlock(&lgt->maptrack_lock);
 
     return handle;
 }
@@ -3017,6 +3017,7 @@ grant_table_create(
 
     /* Simple stuff. */
     spin_lock_init(&t->lock);
+    spin_lock_init(&t->maptrack_lock);
     t->nr_grant_frames = INITIAL_NR_GRANT_FRAMES;
 
     /* Active grant table. */
diff --git a/xen/include/xen/grant_table.h b/xen/include/xen/grant_table.h
index 32f5786..0b35a5e 100644
--- a/xen/include/xen/grant_table.h
+++ b/xen/include/xen/grant_table.h
@@ -82,6 +82,8 @@ struct grant_table {
     struct grant_mapping **maptrack;
     unsigned int          maptrack_head;
     unsigned int          maptrack_limit;
+    /* Lock protecting the maptrack page list, head, and limit */
+    spinlock_t            maptrack_lock;
     /* Lock protecting updates to active and shared grant tables. */
     spinlock_t            lock;
     /* The defined versions are 1 and 2.  Set to 0 if we don't know
-- 
1.7.10.4

[PATCHv12 3/5] gnttab: make the grant table lock a read-write lock

[David Vrabel]

In combination with the per-active entry locks, the grant table lock
can be made a read-write lock since the majority of cases only the
read lock is required. The grant table read lock protects against
changes to the table version or size (which are done with the write
lock held).

The write lock is also required when two active entries must be
acquired.

The double lock is still required when updating IOMMU page tables.

With the lock contention being only on the maptrack lock (unless IOMMU
updates are required), performance and scalability is improved.

Based on a patch originally by Matt Wilson <msw@amazon.com>.

Signed-off-by: David Vrabel <david.vrabel@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
---
v12:
- Document locking needed for maptrack entries (the active entry
  lock).

v11:
- Call gnttab_need_iommu_mapping() once for each lock/unlock.
- Take active entry lock in gnttab_transfer().
- Use read_atomic() when checking maptrack flags for validity (note:
  all other reads of map->flags are either after it is validated as
  stable or protected by the active entry lock and thus don't need
  read_atomic()).
- Add comment to double_gt_lock().

v10:
- In gnttab_map_grant_ref(), keep double lock around maptrack update
  if gnttab_need_iommu_mapping().  Use a wmb(), otherwise.
---
 docs/misc/grant-tables.txt    |   36 ++++++---
 xen/arch/arm/mm.c             |    4 +-
 xen/arch/x86/mm.c             |    4 +-
 xen/common/grant_table.c      |  177 ++++++++++++++++++++++++-----------------
 xen/include/xen/grant_table.h |    9 ++-
 5 files changed, 135 insertions(+), 95 deletions(-)

diff --git a/docs/misc/grant-tables.txt b/docs/misc/grant-tables.txt
index 4a62bf9..fad74de 100644
--- a/docs/misc/grant-tables.txt
+++ b/docs/misc/grant-tables.txt
@@ -83,7 +83,7 @@ is complete.
  ~~~~~~~
  Xen uses several locks to serialize access to the internal grant table state.
 
-  grant_table->lock          : lock used to prevent readers from accessing
+  grant_table->lock          : rwlock used to prevent readers from accessing
                                inconsistent grant table state such as current
                                version, partially initialized active table pages,
                                etc.
@@ -91,34 +91,44 @@ is complete.
   active_grant_entry->lock   : spinlock used to serialize modifications to
                                active entries
 
- The primary lock for the grant table is a spinlock. All functions
- that access members of struct grant_table must acquire the lock
- around critical sections.
+ The primary lock for the grant table is a read/write spinlock. All
+ functions that access members of struct grant_table must acquire a
+ read lock around critical sections. Any modification to the members
+ of struct grant_table (e.g., nr_status_frames, nr_grant_frames,
+ active frames, etc.) must only be made if the write lock is
+ held. These elements are read-mostly, and read critical sections can
+ be large, which makes a rwlock a good choice.
 
  The maptrack free list is protected by its own spinlock. The maptrack
  lock may be locked while holding the grant table lock.
 
  Active entries are obtained by calling active_entry_acquire(gt, ref).
  This function returns a pointer to the active entry after locking its
- spinlock. The caller must hold the grant table lock for the gt in
- question before calling active_entry_acquire(). This is because the
- grant table can be dynamically extended via gnttab_grow_table() while
- a domain is running and must be fully initialized. Once all access to
- the active entry is complete, release the lock by calling
+ spinlock. The caller must hold the grant table read lock before
+ calling active_entry_acquire(). This is because the grant table can
+ be dynamically extended via gnttab_grow_table() while a domain is
+ running and must be fully initialized. Once all access to the active
+ entry is complete, release the lock by calling
  active_entry_release(act).
 
  Summary of rules for locking:
   active_entry_acquire() and active_entry_release() can only be
-  called when holding the relevant grant table's lock. I.e.:
-    spin_lock(&gt->lock);
+  called when holding the relevant grant table's read lock. I.e.:
+    read_lock(&gt->lock);
     act = active_entry_acquire(gt, ref);
     ...
     active_entry_release(act);
-    spin_unlock(&gt->lock);
+    read_unlock(&gt->lock);
 
  Active entries cannot be acquired while holding the maptrack lock.
  Multiple active entries can be acquired while holding the grant table
- lock.
+ _write_ lock.
+
+ Maptrack entries are protected by the corresponding active entry
+ lock.  As an exception, new maptrack entries may be populated without
+ holding the lock, provided the flags field is written last.  This
+ requires any maptrack entry user validates the flags field as
+ non-zero first.
 
 ********************************************************************************
 
diff --git a/xen/arch/arm/mm.c b/xen/arch/arm/mm.c
index feabe34..ff1b330 100644
--- a/xen/arch/arm/mm.c
+++ b/xen/arch/arm/mm.c
@@ -1050,7 +1050,7 @@ int xenmem_add_to_physmap_one(
     switch ( space )
     {
     case XENMAPSPACE_grant_table:
-        spin_lock(&d->grant_table->lock);
+        write_lock(&d->grant_table->lock);
 
         if ( d->grant_table->gt_version == 0 )
             d->grant_table->gt_version = 1;
@@ -1080,7 +1080,7 @@ int xenmem_add_to_physmap_one(
 
         t = p2m_ram_rw;
 
-        spin_unlock(&d->grant_table->lock);
+        write_unlock(&d->grant_table->lock);
         break;
     case XENMAPSPACE_shared_info:
         if ( idx != 0 )
diff --git a/xen/arch/x86/mm.c b/xen/arch/x86/mm.c
index 472c494..9e08c9b 100644
--- a/xen/arch/x86/mm.c
+++ b/xen/arch/x86/mm.c
@@ -4595,7 +4595,7 @@ int xenmem_add_to_physmap_one(
                 mfn = virt_to_mfn(d->shared_info);
             break;
         case XENMAPSPACE_grant_table:
-            spin_lock(&d->grant_table->lock);
+            write_lock(&d->grant_table->lock);
 
             if ( d->grant_table->gt_version == 0 )
                 d->grant_table->gt_version = 1;
@@ -4617,7 +4617,7 @@ int xenmem_add_to_physmap_one(
                     mfn = virt_to_mfn(d->grant_table->shared_raw[idx]);
             }
 
-            spin_unlock(&d->grant_table->lock);
+            write_unlock(&d->grant_table->lock);
             break;
         case XENMAPSPACE_gmfn_range:
         case XENMAPSPACE_gmfn:
diff --git a/xen/common/grant_table.c b/xen/common/grant_table.c
index 25da735..e134f13 100644
--- a/xen/common/grant_table.c
+++ b/xen/common/grant_table.c
@@ -196,7 +196,7 @@ active_entry_acquire(struct grant_table *t, grant_ref_t e)
 {
     struct active_grant_entry *act;
 
-    ASSERT(spin_is_locked(&t->lock));
+    ASSERT(rw_is_locked(&t->lock));
 
     act = &_active_entry(t, e);
     spin_lock(&act->lock);
@@ -252,25 +252,29 @@ static int __get_paged_frame(unsigned long gfn, unsigned long *frame, struct pag
 static inline void
 double_gt_lock(struct grant_table *lgt, struct grant_table *rgt)
 {
+    /*
+     * See mapcount() for why the write lock is also required for the
+     * remote domain.
+     */
     if ( lgt < rgt )
     {
-        spin_lock(&lgt->lock);
-        spin_lock(&rgt->lock);
+        write_lock(&lgt->lock);
+        write_lock(&rgt->lock);
     }
     else
     {
         if ( lgt != rgt )
-            spin_lock(&rgt->lock);
-        spin_lock(&lgt->lock);
+            write_lock(&rgt->lock);
+        write_lock(&lgt->lock);
     }
 }
 
 static inline void
 double_gt_unlock(struct grant_table *lgt, struct grant_table *rgt)
 {
-    spin_unlock(&lgt->lock);
+    write_unlock(&lgt->lock);
     if ( lgt != rgt )
-        spin_unlock(&rgt->lock);
+        write_unlock(&rgt->lock);
 }
 
 static inline int
@@ -528,7 +532,7 @@ static int grant_map_exists(const struct domain *ld,
 {
     unsigned int ref, max_iter;
     
-    ASSERT(spin_is_locked(&rgt->lock));
+    ASSERT(rw_is_locked(&rgt->lock));
 
     max_iter = min(*ref_count + (1 << GNTTABOP_CONTINUATION_ARG_SHIFT),
                    nr_grant_entries(rgt));
@@ -568,15 +572,15 @@ static void mapcount(
     *wrc = *rdc = 0;
 
     /*
-     * Must have the local domain's grant table lock when iterating
-     * over its maptrack entries.
+     * Must have the local domain's grant table write lock when
+     * iterating over its maptrack entries.
      */
-    ASSERT(spin_is_locked(&lgt->lock));
+    ASSERT(rw_is_write_locked(&lgt->lock));
     /*
-     * Must have the remote domain's grant table lock while counting
-     * its active entries.
+     * Must have the remote domain's grant table write lock while
+     * counting its active entries.
      */
-    ASSERT(spin_is_locked(&rd->grant_table->lock));
+    ASSERT(rw_is_write_locked(&rd->grant_table->lock));
 
     for ( handle = 0; handle < lgt->maptrack_limit; handle++ )
     {
@@ -616,6 +620,7 @@ __gnttab_map_grant_ref(
     grant_entry_v2_t *sha2;
     grant_entry_header_t *shah;
     uint16_t *status;
+    bool_t need_iommu;
 
     led = current;
     ld = led->domain;
@@ -661,7 +666,7 @@ __gnttab_map_grant_ref(
     }
 
     rgt = rd->grant_table;
-    spin_lock(&rgt->lock);
+    read_lock(&rgt->lock);
 
     if ( rgt->gt_version == 0 )
         PIN_FAIL(unlock_out, GNTST_general_error,
@@ -735,7 +740,7 @@ __gnttab_map_grant_ref(
     cache_flags = (shah->flags & (GTF_PAT | GTF_PWT | GTF_PCD) );
 
     active_entry_release(act);
-    spin_unlock(&rgt->lock);
+    read_unlock(&rgt->lock);
 
     /* pg may be set, with a refcount included, from __get_paged_frame */
     if ( !pg )
@@ -811,12 +816,14 @@ __gnttab_map_grant_ref(
         goto undo_out;
     }
 
-    double_gt_lock(lgt, rgt);
-
-    if ( gnttab_need_iommu_mapping(ld) )
+    need_iommu = gnttab_need_iommu_mapping(ld);
+    if ( need_iommu )
     {
         unsigned int wrc, rdc;
         int err = 0;
+
+        double_gt_lock(lgt, rgt);
+
         /* We're not translated, so we know that gmfns and mfns are
            the same things, so the IOMMU entry is always 1-to-1. */
         mapcount(lgt, rd, frame, &wrc, &rdc);
@@ -842,12 +849,22 @@ __gnttab_map_grant_ref(
 
     TRACE_1D(TRC_MEM_PAGE_GRANT_MAP, op->dom);
 
+    /*
+     * All maptrack entry users check mt->flags first before using the
+     * other fields so just ensure the flags field is stored last.
+     *
+     * However, if gnttab_need_iommu_mapping() then this would race
+     * with a concurrent mapcount() call (on an unmap, for example)
+     * and a lock is required.
+     */
     mt = &maptrack_entry(lgt, handle);
     mt->domid = op->dom;
     mt->ref   = op->ref;
-    mt->flags = op->flags;
+    wmb();
+    write_atomic(&mt->flags, op->flags);
 
-    double_gt_unlock(lgt, rgt);
+    if ( need_iommu )
+        double_gt_unlock(lgt, rgt);
 
     op->dev_bus_addr = (u64)frame << PAGE_SHIFT;
     op->handle       = handle;
@@ -870,7 +887,7 @@ __gnttab_map_grant_ref(
         put_page(pg);
     }
 
-    spin_lock(&rgt->lock);
+    read_lock(&rgt->lock);
 
     act = active_entry_acquire(rgt, op->ref);
 
@@ -893,7 +910,7 @@ __gnttab_map_grant_ref(
     active_entry_release(act);
 
  unlock_out:
-    spin_unlock(&rgt->lock);
+    read_unlock(&rgt->lock);
     op->status = rc;
     put_maptrack_handle(lgt, handle);
     rcu_unlock_domain(rd);
@@ -943,18 +960,19 @@ __gnttab_unmap_common(
     }
 
     op->map = &maptrack_entry(lgt, op->handle);
-    spin_lock(&lgt->lock);
 
-    if ( unlikely(!op->map->flags) )
+    read_lock(&lgt->lock);
+
+    if ( unlikely(!read_atomic(&op->map->flags)) )
     {
-        spin_unlock(&lgt->lock);
+        read_unlock(&lgt->lock);
         gdprintk(XENLOG_INFO, "Zero flags for handle (%d).\n", op->handle);
         op->status = GNTST_bad_handle;
         return;
     }
 
     dom = op->map->domid;
-    spin_unlock(&lgt->lock);
+    read_unlock(&lgt->lock);
 
     if ( unlikely((rd = rcu_lock_domain_by_id(dom)) == NULL) )
     {
@@ -975,9 +993,10 @@ __gnttab_unmap_common(
     TRACE_1D(TRC_MEM_PAGE_GRANT_UNMAP, dom);
 
     rgt = rd->grant_table;
-    double_gt_lock(lgt, rgt);
 
-    op->flags = op->map->flags;
+    read_lock(&rgt->lock);
+
+    op->flags = read_atomic(&op->map->flags);
     if ( unlikely(!op->flags) || unlikely(op->map->domid != dom) )
     {
         gdprintk(XENLOG_WARNING, "Unstable handle %u\n", op->handle);
@@ -1024,31 +1043,34 @@ __gnttab_unmap_common(
             act->pin -= GNTPIN_hstw_inc;
     }
 
-    if ( gnttab_need_iommu_mapping(ld) )
+ act_release_out:
+    active_entry_release(act);
+ unmap_out:
+    read_unlock(&rgt->lock);
+
+    if ( rc == GNTST_okay && gnttab_need_iommu_mapping(ld) )
     {
         unsigned int wrc, rdc;
         int err = 0;
+
+        double_gt_lock(lgt, rgt);
+
         mapcount(lgt, rd, op->frame, &wrc, &rdc);
         if ( (wrc + rdc) == 0 )
             err = iommu_unmap_page(ld, op->frame);
         else if ( wrc == 0 )
             err = iommu_map_page(ld, op->frame, op->frame, IOMMUF_readable);
+
+        double_gt_unlock(lgt, rgt);
+
         if ( err )
-        {
             rc = GNTST_general_error;
-            goto act_release_out;
-        }
     }
 
     /* If just unmapped a writable mapping, mark as dirtied */
-    if ( !(op->flags & GNTMAP_readonly) )
+    if ( rc == GNTST_okay && !(op->flags & GNTMAP_readonly) )
          gnttab_mark_dirty(rd, op->frame);
 
- act_release_out:
-    active_entry_release(act);
- unmap_out:
-    double_gt_unlock(lgt, rgt);
-
     op->status = rc;
     rcu_unlock_domain(rd);
 }
@@ -1078,8 +1100,8 @@ __gnttab_unmap_common_complete(struct gnttab_unmap_common *op)
 
     rcu_lock_domain(rd);
     rgt = rd->grant_table;
-    spin_lock(&rgt->lock);
 
+    read_lock(&rgt->lock);
     if ( rgt->gt_version == 0 )
         goto unlock_out;
 
@@ -1145,7 +1167,7 @@ __gnttab_unmap_common_complete(struct gnttab_unmap_common *op)
  act_release_out:
     active_entry_release(act);
  unlock_out:
-    spin_unlock(&rgt->lock);
+    read_unlock(&rgt->lock);
 
     if ( put_handle )
     {
@@ -1332,11 +1354,13 @@ gnttab_unpopulate_status_frames(struct domain *d, struct grant_table *gt)
     gt->nr_status_frames = 0;
 }
 
+/*
+ * Grow the grant table. The caller must hold the grant table's
+ * write lock before calling this function.
+ */
 int
 gnttab_grow_table(struct domain *d, unsigned int req_nr_frames)
 {
-    /* d's grant table lock must be held by the caller */
-
     struct grant_table *gt = d->grant_table;
     unsigned int i, j;
 
@@ -1442,7 +1466,7 @@ gnttab_setup_table(
     }
 
     gt = d->grant_table;
-    spin_lock(&gt->lock);
+    write_lock(&gt->lock);
 
     if ( gt->gt_version == 0 )
         gt->gt_version = 1;
@@ -1470,7 +1494,7 @@ gnttab_setup_table(
     }
 
  out3:
-    spin_unlock(&gt->lock);
+    write_unlock(&gt->lock);
  out2:
     rcu_unlock_domain(d);
  out1:
@@ -1512,13 +1536,13 @@ gnttab_query_size(
         goto query_out_unlock;
     }
 
-    spin_lock(&d->grant_table->lock);
+    read_lock(&d->grant_table->lock);
 
     op.nr_frames     = nr_grant_frames(d->grant_table);
     op.max_nr_frames = max_grant_frames;
     op.status        = GNTST_okay;
 
-    spin_unlock(&d->grant_table->lock);
+    read_unlock(&d->grant_table->lock);
 
  
  query_out_unlock:
@@ -1544,7 +1568,7 @@ gnttab_prepare_for_transfer(
     union grant_combo   scombo, prev_scombo, new_scombo;
     int                 retries = 0;
 
-    spin_lock(&rgt->lock);
+    read_lock(&rgt->lock);
 
     if ( rgt->gt_version == 0 )
     {
@@ -1595,11 +1619,11 @@ gnttab_prepare_for_transfer(
         scombo = prev_scombo;
     }
 
-    spin_unlock(&rgt->lock);
+    read_unlock(&rgt->lock);
     return 1;
 
  fail:
-    spin_unlock(&rgt->lock);
+    read_unlock(&rgt->lock);
     return 0;
 }
 
@@ -1614,6 +1638,7 @@ gnttab_transfer(
     struct gnttab_transfer gop;
     unsigned long mfn;
     unsigned int max_bitsize;
+    struct active_grant_entry *act;
 
     for ( i = 0; i < count; i++ )
     {
@@ -1792,7 +1817,8 @@ gnttab_transfer(
         TRACE_1D(TRC_MEM_PAGE_GRANT_TRANSFER, e->domain_id);
 
         /* Tell the guest about its new page frame. */
-        spin_lock(&e->grant_table->lock);
+        read_lock(&e->grant_table->lock);
+        act = active_entry_acquire(e->grant_table, gop.ref);
 
         if ( e->grant_table->gt_version == 1 )
         {
@@ -1810,7 +1836,8 @@ gnttab_transfer(
         shared_entry_header(e->grant_table, gop.ref)->flags |=
             GTF_transfer_completed;
 
-        spin_unlock(&e->grant_table->lock);
+        active_entry_release(act);
+        read_unlock(&e->grant_table->lock);
 
         rcu_unlock_domain(e);
 
@@ -1848,7 +1875,7 @@ __release_grant_for_copy(
     released_read = 0;
     released_write = 0;
 
-    spin_lock(&rgt->lock);
+    read_lock(&rgt->lock);
 
     act = active_entry_acquire(rgt, gref);
     sha = shared_entry_header(rgt, gref);
@@ -1890,7 +1917,7 @@ __release_grant_for_copy(
     }
 
     active_entry_release(act);
-    spin_unlock(&rgt->lock);
+    read_unlock(&rgt->lock);
 
     if ( td != rd )
     {
@@ -1948,7 +1975,7 @@ __acquire_grant_for_copy(
 
     *page = NULL;
 
-    spin_lock(&rgt->lock);
+    read_lock(&rgt->lock);
 
     if ( rgt->gt_version == 0 )
         PIN_FAIL(gt_unlock_out, GNTST_general_error,
@@ -2024,20 +2051,20 @@ __acquire_grant_for_copy(
              * here and reacquire
              */
             active_entry_release(act);
-            spin_unlock(&rgt->lock);
+            read_unlock(&rgt->lock);
 
             rc = __acquire_grant_for_copy(td, trans_gref, rd->domain_id,
                                           readonly, &grant_frame, page,
                                           &trans_page_off, &trans_length, 0);
 
-            spin_lock(&rgt->lock);
+            read_lock(&rgt->lock);
             act = active_entry_acquire(rgt, gref);
 
             if ( rc != GNTST_okay ) {
                 __fixup_status_for_copy_pin(act, status);
                 rcu_unlock_domain(td);
                 active_entry_release(act);
-                spin_unlock(&rgt->lock);
+                read_unlock(&rgt->lock);
                 return rc;
             }
 
@@ -2050,7 +2077,7 @@ __acquire_grant_for_copy(
                 __fixup_status_for_copy_pin(act, status);
                 rcu_unlock_domain(td);
                 active_entry_release(act);
-                spin_unlock(&rgt->lock);
+                read_unlock(&rgt->lock);
                 put_page(*page);
                 return __acquire_grant_for_copy(rd, gref, ldom, readonly,
                                                 frame, page, page_off, length,
@@ -2119,7 +2146,7 @@ __acquire_grant_for_copy(
     *frame = act->frame;
 
     active_entry_release(act);
-    spin_unlock(&rgt->lock);
+    read_unlock(&rgt->lock);
     return rc;
  
  unlock_out_clear:
@@ -2134,7 +2161,7 @@ __acquire_grant_for_copy(
     active_entry_release(act);
 
  gt_unlock_out:
-    spin_unlock(&rgt->lock);
+    read_unlock(&rgt->lock);
 
     return rc;
 }
@@ -2450,7 +2477,7 @@ gnttab_set_version(XEN_GUEST_HANDLE_PARAM(gnttab_set_version_t) uop)
     if ( gt->gt_version == op.version )
         goto out;
 
-    spin_lock(&gt->lock);
+    write_lock(&gt->lock);
     /* Make sure that the grant table isn't currently in use when we
        change the version number, except for the first 8 entries which
        are allowed to be in use (xenstore/xenconsole keeps them mapped).
@@ -2535,7 +2562,7 @@ gnttab_set_version(XEN_GUEST_HANDLE_PARAM(gnttab_set_version_t) uop)
     gt->gt_version = op.version;
 
 out_unlock:
-    spin_unlock(&gt->lock);
+    write_unlock(&gt->lock);
 
 out:
     op.version = gt->gt_version;
@@ -2591,7 +2618,7 @@ gnttab_get_status_frames(XEN_GUEST_HANDLE_PARAM(gnttab_get_status_frames_t) uop,
 
     op.status = GNTST_okay;
 
-    spin_lock(&gt->lock);
+    read_lock(&gt->lock);
 
     for ( i = 0; i < op.nr_frames; i++ )
     {
@@ -2600,7 +2627,7 @@ gnttab_get_status_frames(XEN_GUEST_HANDLE_PARAM(gnttab_get_status_frames_t) uop,
             op.status = GNTST_bad_virt_addr;
     }
 
-    spin_unlock(&gt->lock);
+    read_unlock(&gt->lock);
 out2:
     rcu_unlock_domain(d);
 out1:
@@ -2650,7 +2677,7 @@ __gnttab_swap_grant_ref(grant_ref_t ref_a, grant_ref_t ref_b)
     struct active_grant_entry *act_b = NULL;
     s16 rc = GNTST_okay;
 
-    spin_lock(&gt->lock);
+    write_lock(&gt->lock);
 
     if ( gt->gt_version == 0 )
         PIN_FAIL(out, GNTST_general_error, "grant table not yet set up\n");
@@ -2701,7 +2728,7 @@ out:
         active_entry_release(act_b);
     if ( act_a != NULL )
         active_entry_release(act_a);
-    spin_unlock(&gt->lock);
+    write_unlock(&gt->lock);
 
     rcu_unlock_domain(d);
 
@@ -2772,12 +2799,12 @@ static int __gnttab_cache_flush(gnttab_cache_flush_t *cflush,
 
     if ( d != owner )
     {
-        spin_lock(&owner->grant_table->lock);
+        read_lock(&owner->grant_table->lock);
 
         ret = grant_map_exists(d, owner->grant_table, mfn, ref_count);
         if ( ret != 0 )
         {
-            spin_unlock(&owner->grant_table->lock);
+            read_unlock(&owner->grant_table->lock);
             rcu_unlock_domain(d);
             put_page(page);
             return ret;
@@ -2797,7 +2824,7 @@ static int __gnttab_cache_flush(gnttab_cache_flush_t *cflush,
         ret = 0;
 
     if ( d != owner )
-        spin_unlock(&owner->grant_table->lock);
+        read_unlock(&owner->grant_table->lock);
     unmap_domain_page(v);
     put_page(page);
 
@@ -3016,7 +3043,7 @@ grant_table_create(
         goto no_mem_0;
 
     /* Simple stuff. */
-    spin_lock_init(&t->lock);
+    rwlock_init(&t->lock);
     spin_lock_init(&t->maptrack_lock);
     t->nr_grant_frames = INITIAL_NR_GRANT_FRAMES;
 
@@ -3126,7 +3153,7 @@ gnttab_release_mappings(
         }
 
         rgt = rd->grant_table;
-        spin_lock(&rgt->lock);
+        read_lock(&rgt->lock);
 
         act = active_entry_acquire(rgt, ref);
         sha = shared_entry_header(rgt, ref);
@@ -3187,7 +3214,7 @@ gnttab_release_mappings(
             gnttab_clear_flag(_GTF_reading, status);
 
         active_entry_release(act);
-        spin_unlock(&rgt->lock);
+        read_unlock(&rgt->lock);
 
         rcu_unlock_domain(rd);
 
@@ -3235,7 +3262,7 @@ static void gnttab_usage_print(struct domain *rd)
     printk("      -------- active --------       -------- shared --------\n");
     printk("[ref] localdom mfn      pin          localdom gmfn     flags\n");
 
-    spin_lock(&gt->lock);
+    read_lock(&gt->lock);
 
     if ( gt->gt_version == 0 )
         goto out;
@@ -3288,7 +3315,7 @@ static void gnttab_usage_print(struct domain *rd)
     }
 
  out:
-    spin_unlock(&gt->lock);
+    read_unlock(&gt->lock);
 
     if ( first )
         printk("grant-table for remote domain:%5d ... "
diff --git a/xen/include/xen/grant_table.h b/xen/include/xen/grant_table.h
index 0b35a5e..f22ebd0 100644
--- a/xen/include/xen/grant_table.h
+++ b/xen/include/xen/grant_table.h
@@ -64,6 +64,11 @@ struct grant_mapping {
 
 /* Per-domain grant information. */
 struct grant_table {
+    /*
+     * Lock protecting updates to grant table state (version, active
+     * entry list, etc.)
+     */
+    rwlock_t              lock;
     /* Table size. Number of frames shared with guest */
     unsigned int          nr_grant_frames;
     /* Shared grant table (see include/public/grant_table.h). */
@@ -84,8 +89,6 @@ struct grant_table {
     unsigned int          maptrack_limit;
     /* Lock protecting the maptrack page list, head, and limit */
     spinlock_t            maptrack_lock;
-    /* Lock protecting updates to active and shared grant tables. */
-    spinlock_t            lock;
     /* The defined versions are 1 and 2.  Set to 0 if we don't know
        what version to use yet. */
     unsigned              gt_version;
@@ -103,7 +106,7 @@ gnttab_release_mappings(
     struct domain *d);
 
 /* Increase the size of a domain's grant table.
- * Caller must hold d's grant table lock.
+ * Caller must hold d's grant table write lock.
  */
 int
 gnttab_grow_table(struct domain *d, unsigned int req_nr_frames);
-- 
1.7.10.4

[PATCHv12 4/5] gnttab: use per-VCPU maptrack free lists

[David Vrabel]

From: Malcolm Crossley <malcolm.crossley@citrix.com>

Performance analysis of aggregate network throughput with many VMs
shows that performance is signficantly limited by contention on the
maptrack lock when obtaining/releasing maptrack handles from the free
list.

Instead of a single free list use a per-VCPU list. This avoids any
contention when obtaining a handle.  Handles must be released back to
their original list and since this may occur on a different VCPU there
is some contention on the destination VCPU's free list tail pointer
(but this is much better than a per-domain lock).

Increase the default maximum number of maptrack frames by 4 times
because: a) struct grant_mapping is now 16 bytes (instead of 8); and
b) a guest may not evenly distribute all the grant map operations
across the VCPUs (meaning some VCPUs need more maptrack entries than
others).

Signed-off-by: Malcolm Crossley <malcolm.crossley@citrix.com>
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
---
v11:
- Allocate maptrack array with vzalloc() since it may be >1 page.

v10:
- Divide max_maptrack_frames evenly amongst the VCPUs.
- Increase default max_maptrack_frames to compensate.
---
 xen/common/grant_table.c      |  149 +++++++++++++++++++++++++----------------
 xen/include/xen/grant_table.h |    8 ++-
 xen/include/xen/sched.h       |    5 ++
 3 files changed, 101 insertions(+), 61 deletions(-)

diff --git a/xen/common/grant_table.c b/xen/common/grant_table.c
index e134f13..41d4e0f 100644
--- a/xen/common/grant_table.c
+++ b/xen/common/grant_table.c
@@ -37,6 +37,7 @@
 #include <xen/iommu.h>
 #include <xen/paging.h>
 #include <xen/keyhandler.h>
+#include <xen/vmap.h>
 #include <xsm/xsm.h>
 #include <asm/flushtlb.h>
 
@@ -57,7 +58,7 @@ integer_param("gnttab_max_frames", max_grant_frames);
  * New options allow to set max_maptrack_frames and
  * map_grant_table_frames independently.
  */
-#define DEFAULT_MAX_MAPTRACK_FRAMES 256
+#define DEFAULT_MAX_MAPTRACK_FRAMES 1024
 
 static unsigned int __read_mostly max_maptrack_frames;
 integer_param("gnttab_max_maptrack_frames", max_maptrack_frames);
@@ -117,12 +118,6 @@ struct gnttab_unmap_common {
 #define maptrack_entry(t, e) \
     ((t)->maptrack[(e)/MAPTRACK_PER_PAGE][(e)%MAPTRACK_PER_PAGE])
 
-static inline unsigned int
-nr_maptrack_frames(struct grant_table *t)
-{
-    return t->maptrack_limit / MAPTRACK_PER_PAGE;
-}
-
 #define MAPTRACK_TAIL (~0u)
 
 #define SHGNT_PER_PAGE_V1 (PAGE_SIZE / sizeof(grant_entry_v1_t))
@@ -279,66 +274,104 @@ double_gt_unlock(struct grant_table *lgt, struct grant_table *rgt)
 
 static inline int
 __get_maptrack_handle(
-    struct grant_table *t)
+    struct grant_table *t,
+    struct vcpu *v)
 {
-    unsigned int h;
-    if ( unlikely((h = t->maptrack_head) == MAPTRACK_TAIL) )
+    unsigned int head, next;
+
+    /* No maptrack pages allocated for this VCPU yet? */
+    head = v->maptrack_head;
+    if ( unlikely(head == MAPTRACK_TAIL) )
         return -1;
-    t->maptrack_head = maptrack_entry(t, h).ref;
-    return h;
+
+    /*
+     * Always keep one entry in the free list to make it easier to add
+     * free entries to the tail.
+     */
+    next = read_atomic(&maptrack_entry(t, head).ref);
+    if ( unlikely(next == MAPTRACK_TAIL) )
+        return -1;
+
+    v->maptrack_head = next;
+
+    return head;
 }
 
 static inline void
 put_maptrack_handle(
     struct grant_table *t, int handle)
 {
-    spin_lock(&t->maptrack_lock);
-    maptrack_entry(t, handle).ref = t->maptrack_head;
-    t->maptrack_head = handle;
-    spin_unlock(&t->maptrack_lock);
+    struct domain *d = current->domain;
+    struct vcpu *v;
+    unsigned int prev_tail, cur_tail;
+
+    /* 1. Set entry to be a tail. */
+    maptrack_entry(t, handle).ref = MAPTRACK_TAIL;
+
+    /* 2. Add entry to the tail of the list on the original VCPU. */
+    v = d->vcpu[maptrack_entry(t,handle).vcpu];
+
+    cur_tail = read_atomic(&v->maptrack_tail);
+    do {
+        prev_tail = cur_tail;
+        cur_tail = cmpxchg(&v->maptrack_tail, prev_tail, handle);
+    } while ( cur_tail != prev_tail );
+
+    /* 3. Update the old tail entry to point to the new entry. */
+    write_atomic(&maptrack_entry(t, prev_tail).ref, handle);
 }
 
 static inline int
 get_maptrack_handle(
     struct grant_table *lgt)
 {
+    struct vcpu          *v = current;
     int                   i;
     grant_handle_t        handle;
     struct grant_mapping *new_mt;
-    unsigned int          new_mt_limit, nr_frames;
 
-    spin_lock(&lgt->maptrack_lock);
+    handle = __get_maptrack_handle(lgt, v);
+    if ( likely(handle != -1) )
+        return handle;
 
-    while ( unlikely((handle = __get_maptrack_handle(lgt)) == -1) )
-    {
-        nr_frames = nr_maptrack_frames(lgt);
-        if ( nr_frames >= max_maptrack_frames )
-            break;
+    /*
+     * max_maptrack_frames is per domain so each VCPU gets a share of
+     * the maximum, but allow at least one frame per VCPU.
+     */
+    if ( v->maptrack_frames
+         && v->maptrack_frames >= max_maptrack_frames / v->domain->max_vcpus )
+        return -1;
 
-        new_mt = alloc_xenheap_page();
-        if ( !new_mt )
-            break;
+    new_mt = alloc_xenheap_page();
+    if ( !new_mt )
+        return -1;
+    clear_page(new_mt);
 
-        clear_page(new_mt);
+    spin_lock(&lgt->maptrack_lock);
 
-        new_mt_limit = lgt->maptrack_limit + MAPTRACK_PER_PAGE;
+    for ( i = 1; i < MAPTRACK_PER_PAGE; i++ )
+    {
+        new_mt[i - 1].ref = (lgt->maptrack_pages * MAPTRACK_PER_PAGE) + i;
+        new_mt[i - 1].vcpu = v->vcpu_id;
+    }
+    /* Set last entry vcpu and ref */
+    new_mt[i - 1].ref = v->maptrack_head;
+    new_mt[i - 1].vcpu = v->vcpu_id;
 
-        for ( i = 1; i < MAPTRACK_PER_PAGE; i++ )
-            new_mt[i - 1].ref = lgt->maptrack_limit + i;
-        new_mt[i - 1].ref = lgt->maptrack_head;
-        lgt->maptrack_head = lgt->maptrack_limit;
+    v->maptrack_head = lgt->maptrack_pages * MAPTRACK_PER_PAGE;
 
-        lgt->maptrack[nr_frames] = new_mt;
-        smp_wmb();
-        lgt->maptrack_limit      = new_mt_limit;
+    /* Set tail directly if this is the first page for this VCPU. */
+    if ( v->maptrack_tail == MAPTRACK_TAIL )
+        v->maptrack_tail = (lgt->maptrack_pages * MAPTRACK_PER_PAGE)
+            + MAPTRACK_PER_PAGE - 1;
 
-        gdprintk(XENLOG_INFO, "Increased maptrack size to %u frames\n",
-                 nr_frames + 1);
-    }
+    lgt->maptrack[lgt->maptrack_pages++] = new_mt;
 
     spin_unlock(&lgt->maptrack_lock);
 
-    return handle;
+    v->maptrack_frames++;
+
+    return __get_maptrack_handle(lgt, v);
 }
 
 /* Number of grant table entries. Caller must hold d's grant table lock. */
@@ -582,7 +615,8 @@ static void mapcount(
      */
     ASSERT(rw_is_write_locked(&rd->grant_table->lock));
 
-    for ( handle = 0; handle < lgt->maptrack_limit; handle++ )
+    for ( handle = 0; handle < lgt->maptrack_pages * MAPTRACK_PER_PAGE; 
+          handle++ )
     {
         map = &maptrack_entry(lgt, handle);
         if ( !(map->flags & (GNTMAP_device_map|GNTMAP_host_map)) ||
@@ -952,7 +986,7 @@ __gnttab_unmap_common(
 
     op->frame = (unsigned long)(op->dev_bus_addr >> PAGE_SHIFT);
 
-    if ( unlikely(op->handle >= lgt->maptrack_limit) )
+    if ( unlikely(op->handle >= lgt->maptrack_pages * MAPTRACK_PER_PAGE) )
     {
         gdprintk(XENLOG_INFO, "Bad handle (%d).\n", op->handle);
         op->status = GNTST_bad_handle;
@@ -1427,6 +1461,7 @@ gnttab_setup_table(
     struct gnttab_setup_table op;
     struct domain *d;
     struct grant_table *gt;
+    struct vcpu *v;
     int            i;
     xen_pfn_t  gmfn;
 
@@ -1468,6 +1503,17 @@ gnttab_setup_table(
     gt = d->grant_table;
     write_lock(&gt->lock);
 
+    /* Tracking of mapped foreign frames table */
+    gt->maptrack = vzalloc(max_maptrack_frames * sizeof(*gt->maptrack));
+    if ( gt->maptrack == NULL )
+        goto out3;
+    for_each_vcpu( d, v )
+    {
+        v->maptrack_head = MAPTRACK_TAIL;
+        v->maptrack_tail = MAPTRACK_TAIL;
+    }
+    gt->maptrack_pages = 0;
+
     if ( gt->gt_version == 0 )
         gt->gt_version = 1;
 
@@ -3061,18 +3107,6 @@ grant_table_create(
             spin_lock_init(&t->active[i][j].lock);
     }
 
-    /* Tracking of mapped foreign frames table */
-    if ( (t->maptrack = xzalloc_array(struct grant_mapping *,
-                                      max_maptrack_frames)) == NULL )
-        goto no_mem_2;
-    if ( (t->maptrack[0] = alloc_xenheap_page()) == NULL )
-        goto no_mem_3;
-    clear_page(t->maptrack[0]);
-    t->maptrack_limit = MAPTRACK_PER_PAGE;
-    for ( i = 1; i < MAPTRACK_PER_PAGE; i++ )
-        t->maptrack[0][i - 1].ref = i;
-    t->maptrack[0][i - 1].ref = MAPTRACK_TAIL;
-
     /* Shared grant table. */
     if ( (t->shared_raw = xzalloc_array(void *, max_grant_frames)) == NULL )
         goto no_mem_3;
@@ -3103,8 +3137,6 @@ grant_table_create(
         free_xenheap_page(t->shared_raw[i]);
     xfree(t->shared_raw);
  no_mem_3:
-    free_xenheap_page(t->maptrack[0]);
-    xfree(t->maptrack);
  no_mem_2:
     for ( i = 0;
           i < num_act_frames_from_sha_frames(INITIAL_NR_GRANT_FRAMES); i++ )
@@ -3132,7 +3164,8 @@ gnttab_release_mappings(
 
     BUG_ON(!d->is_dying);
 
-    for ( handle = 0; handle < gt->maptrack_limit; handle++ )
+    for ( handle = 0; handle < gt->maptrack_pages * MAPTRACK_PER_PAGE;
+          handle++ )
     {
         map = &maptrack_entry(gt, handle);
         if ( !(map->flags & (GNTMAP_device_map|GNTMAP_host_map)) )
@@ -3237,9 +3270,9 @@ grant_table_destroy(
         free_xenheap_page(t->shared_raw[i]);
     xfree(t->shared_raw);
 
-    for ( i = 0; i < nr_maptrack_frames(t); i++ )
+    for ( i = 0; i < t->maptrack_pages; i++ )
         free_xenheap_page(t->maptrack[i]);
-    xfree(t->maptrack);
+    vfree(t->maptrack);
 
     for ( i = 0; i < nr_active_grant_frames(t); i++ )
         free_xenheap_page(t->active[i]);
diff --git a/xen/include/xen/grant_table.h b/xen/include/xen/grant_table.h
index f22ebd0..c6e4ebf 100644
--- a/xen/include/xen/grant_table.h
+++ b/xen/include/xen/grant_table.h
@@ -60,6 +60,8 @@ struct grant_mapping {
     u32      ref;           /* grant ref */
     u16      flags;         /* 0-4: GNTMAP_* ; 5-15: unused */
     domid_t  domid;         /* granting domain */
+    u32      vcpu;          /* vcpu which created the grant mapping */
+    u32      pad;           /* round size to a power of 2 */
 };
 
 /* Per-domain grant information. */
@@ -83,10 +85,10 @@ struct grant_table {
     grant_status_t       **status;
     /* Active grant table. */
     struct active_grant_entry **active;
-    /* Mapping tracking table. */
+    /* Mapping tracking table per vcpu. */
     struct grant_mapping **maptrack;
-    unsigned int          maptrack_head;
-    unsigned int          maptrack_limit;
+    /* Total pages used for mapping tracking table */
+    unsigned int          maptrack_pages;
     /* Lock protecting the maptrack page list, head, and limit */
     spinlock_t            maptrack_lock;
     /* The defined versions are 1 and 2.  Set to 0 if we don't know
diff --git a/xen/include/xen/sched.h b/xen/include/xen/sched.h
index 80c6f62..d46a561 100644
--- a/xen/include/xen/sched.h
+++ b/xen/include/xen/sched.h
@@ -219,6 +219,11 @@ struct vcpu
     /* VCPU paused by system controller. */
     int              controller_pause_count;
 
+    /* Maptrack */
+    unsigned int     maptrack_head;
+    unsigned int     maptrack_tail;
+    unsigned int     maptrack_frames;
+
     /* IRQ-safe virq_lock protects against delivering VIRQ to stale evtchn. */
     evtchn_port_t    virq_to_evtchn[NR_VIRQS];
     spinlock_t       virq_lock;
-- 
1.7.10.4

[PATCHv12 5/5] [RFC] gnttab: steal maptrack entries from other VCPUs

[David Vrabel]

If a guest is not evenly grant mapping across its VCPUs one of the
VCPUs may run out of free maptrack entries even though other VCPUs
have many free.

If this happens, "steal" free entries from other VCPUs.  We want to
steal entries such that:

a) We avoid ping-ponging stolen entries between VCPUs.

b) The number of free entries owned by each VCPUs tends (over time) to
   the number it uses.

So when stealing, we select a VCPU at random (reducing (a)) and we
transfer the stolen entries to the thief VCPU (aiming for (b)).

Signed-off-by: David Vrabel <david.vrabel@citrix.com>
---
v12:
- New.
---
 xen/common/grant_table.c |   73 ++++++++++++++++++++++++++++++++++++----------
 1 file changed, 58 insertions(+), 15 deletions(-)

diff --git a/xen/common/grant_table.c b/xen/common/grant_table.c
index 41d4e0f..5d91d38 100644
--- a/xen/common/grant_table.c
+++ b/xen/common/grant_table.c
@@ -277,26 +277,65 @@ __get_maptrack_handle(
     struct grant_table *t,
     struct vcpu *v)
 {
-    unsigned int head, next;
+    unsigned int head, next, prev_head;
 
-    /* No maptrack pages allocated for this VCPU yet? */
-    head = v->maptrack_head;
-    if ( unlikely(head == MAPTRACK_TAIL) )
-        return -1;
+    do {
+        /* No maptrack pages allocated for this VCPU yet? */
+        head = read_atomic(&v->maptrack_head);
+        if ( unlikely(head == MAPTRACK_TAIL) )
+            return -1;
 
-    /*
-     * Always keep one entry in the free list to make it easier to add
-     * free entries to the tail.
-     */
-    next = read_atomic(&maptrack_entry(t, head).ref);
-    if ( unlikely(next == MAPTRACK_TAIL) )
-        return -1;
+        /*
+         * Always keep one entry in the free list to make it easier to
+         * add free entries to the tail.
+         */
+        next = read_atomic(&maptrack_entry(t, head).ref);
+        if ( unlikely(next == MAPTRACK_TAIL) )
+            return -1;
 
-    v->maptrack_head = next;
+        prev_head = head;
+        head = cmpxchg(&v->maptrack_head, prev_head, next);
+    } while ( head != prev_head );
 
     return head;
 }
 
+/*
+ * Try to "steal" a free maptrack entry from another VCPU.
+ *
+ * A stolen entry is transferred to the thief, so the number of
+ * entries for each VCPU should tend to the usage pattern.
+ *
+ * To avoid having to atomically count the number of free entries on
+ * each VCPU and to avoid two VCPU repeatedly stealing entries from
+ * each other, the initial victim VCPU is selected randomly.
+ */
+static int steal_maptrack_handle(struct grant_table *t, struct vcpu *v)
+{
+    struct domain *d = v->domain;
+    unsigned int first, i;
+
+    /* Find an initial victim. */
+    first = i = NOW() % d->max_vcpus;
+
+    do {
+        unsigned int handle;
+
+        handle = __get_maptrack_handle(t, d->vcpu[i]);
+        if ( handle >= 0 ) {
+            maptrack_entry(t, handle).vcpu = v->vcpu_id;
+            return handle;
+        }
+
+        i++;
+        if ( i == d->max_vcpus )
+            i = 0;
+    } while ( i != first );
+
+    /* No free handles on any VCPU. */
+    return -1;
+}
+
 static inline void
 put_maptrack_handle(
     struct grant_table *t, int handle)
@@ -337,10 +376,14 @@ get_maptrack_handle(
     /*
      * max_maptrack_frames is per domain so each VCPU gets a share of
      * the maximum, but allow at least one frame per VCPU.
+     *
+     * If this VCPU has run out of frames, try stealing an entry from
+     * another VCPU (in case the guest isn't mapping across its VCPUs
+     * evenly).
      */
     if ( v->maptrack_frames
          && v->maptrack_frames >= max_maptrack_frames / v->domain->max_vcpus )
-        return -1;
+        return steal_maptrack_handle(lgt, v);
 
     new_mt = alloc_xenheap_page();
     if ( !new_mt )
@@ -358,7 +401,7 @@ get_maptrack_handle(
     new_mt[i - 1].ref = v->maptrack_head;
     new_mt[i - 1].vcpu = v->vcpu_id;
 
-    v->maptrack_head = lgt->maptrack_pages * MAPTRACK_PER_PAGE;
+    write_atomic(&v->maptrack_head, lgt->maptrack_pages * MAPTRACK_PER_PAGE);
 
     /* Set tail directly if this is the first page for this VCPU. */
     if ( v->maptrack_tail == MAPTRACK_TAIL )
-- 
1.7.10.4

Re: [PATCHv12 1/5] gnttab: per-active entry locking

[Jan Beulich]

>>> On 12.06.15 at 15:33, <david.vrabel@citrix.com> wrote:
> Introduce a per-active entry spin lock to protect active entry state
> The grant table lock must be locked before acquiring (locking) an
> active entry.
> 
> This is a step in reducing contention on the grant table lock, but
> will only do so once the grant table lock is turned into a read-write
> lock.
> 
> Based on a patch originally by Matt Wilson <msw@amazon.com>.
> 
> Signed-off-by: David Vrabel <david.vrabel@citrix.com>
> ---
> v12:
> - Check for same refs in __gnttab_swap_grant_ref() to avoid deadlock.

I'm glad you spotted this, because the check having been there
until v8 I didn't notice it got lost later on.

Jan

Re: [PATCHv12 5/5] [RFC] gnttab: steal maptrack entries from other VCPUs

[Jan Beulich]

>>> On 12.06.15 at 15:33, <david.vrabel@citrix.com> wrote:
> +static int steal_maptrack_handle(struct grant_table *t, struct vcpu *v)
> +{
> +    struct domain *d = v->domain;
> +    unsigned int first, i;
> +
> +    /* Find an initial victim. */
> +    first = i = NOW() % d->max_vcpus;
> +
> +    do {
> +        unsigned int handle;
> +
> +        handle = __get_maptrack_handle(t, d->vcpu[i]);
> +        if ( handle >= 0 ) {

Coding style.

> @@ -337,10 +376,14 @@ get_maptrack_handle(
>      /*
>       * max_maptrack_frames is per domain so each VCPU gets a share of
>       * the maximum, but allow at least one frame per VCPU.
> +     *
> +     * If this VCPU has run out of frames, try stealing an entry from
> +     * another VCPU (in case the guest isn't mapping across its VCPUs
> +     * evenly).
>       */
>      if ( v->maptrack_frames
>           && v->maptrack_frames >= max_maptrack_frames / v->domain->max_vcpus )
> -        return -1;
> +        return steal_maptrack_handle(lgt, v);

This "allow at least one frame per vCPU" imo needs to go away
again now.

Looks okay to me apart from that; quite nice an idea to utilize
the fact that ownership of each entry is already being tracked.

Jan