[Xen-devel] [PATCH v2 1/2] xen: enter/exit lazy_mmu_mode around m2p_override calls

[Stefano Stabellini <stefano.stabellini@xxxxxxxxxxxxx>]

This patch is a significant performance improvement for the
m2p_override: about 6% using the gntdev device.

Each m2p_add/remove_override call issues a MULTI_grant_table_op and a
__flush_tlb_single if kmap_op != NULL.  Batching all the calls together
is a great performance benefit because it means issuing one hypercall
total rather than two hypercall per page.
If paravirt_lazy_mode is set PARAVIRT_LAZY_MMU, all these calls are
going to be batched together, otherwise they are issued one at a time.

Adding arch_enter_lazy_mmu_mode/arch_leave_lazy_mmu_mode around the
m2p_add/remove_override calls forces paravirt_lazy_mode to
PARAVIRT_LAZY_MMU, therefore makes sure that they are always batched.

Signed-off-by: Stefano Stabellini <stefano.stabellini@xxxxxxxxxxxxx>
---
 drivers/block/xen-blkback/blkback.c |    4 ++++
 drivers/xen/grant-table.c           |    8 ++++++++
 2 files changed, 12 insertions(+), 0 deletions(-)

diff --git a/drivers/block/xen-blkback/blkback.c 
b/drivers/block/xen-blkback/blkback.c
index 0088bf6..60abff6 100644
--- a/drivers/block/xen-blkback/blkback.c
+++ b/drivers/block/xen-blkback/blkback.c
@@ -342,6 +342,7 @@ static void xen_blkbk_unmap(struct pending_req *req)
         * Note, we use invcount, so nr->pages, so we can't index
         * using vaddr(req, i).
         */
+       arch_enter_lazy_mmu_mode();
        for (i = 0; i < invcount; i++) {
                ret = m2p_remove_override(
                        virt_to_page(unmap[i].host_addr), false);
@@ -351,6 +352,7 @@ static void xen_blkbk_unmap(struct pending_req *req)
                        continue;
                }
        }
+       arch_leave_lazy_mmu_mode();
 }
 
 static int xen_blkbk_map(struct blkif_request *req,
@@ -386,6 +388,7 @@ static int xen_blkbk_map(struct blkif_request *req,
         * so that when we access vaddr(pending_req,i) it has the contents of
         * the page from the other domain.
         */
+       arch_enter_lazy_mmu_mode();
        for (i = 0; i < nseg; i++) {
                if (unlikely(map[i].status != 0)) {
                        pr_debug(DRV_PFX "invalid buffer -- could not remap 
it\n");
@@ -410,6 +413,7 @@ static int xen_blkbk_map(struct blkif_request *req,
                seg[i].buf  = map[i].dev_bus_addr |
                        (req->u.rw.seg[i].first_sect << 9);
        }
+       arch_leave_lazy_mmu_mode();
        return ret;
 }
 
diff --git a/drivers/xen/grant-table.c b/drivers/xen/grant-table.c
index b4d4eac..c7dc2d6 100644
--- a/drivers/xen/grant-table.c
+++ b/drivers/xen/grant-table.c
@@ -751,6 +751,8 @@ int gnttab_map_refs(struct gnttab_map_grant_ref *map_ops,
        if (xen_feature(XENFEAT_auto_translated_physmap))
                return ret;
 
+       arch_enter_lazy_mmu_mode();
+
        for (i = 0; i < count; i++) {
                /* Do not add to override if the map failed. */
                if (map_ops[i].status)
@@ -769,6 +771,8 @@ int gnttab_map_refs(struct gnttab_map_grant_ref *map_ops,
                        return ret;
        }
 
+       arch_leave_lazy_mmu_mode();
+
        return ret;
 }
 EXPORT_SYMBOL_GPL(gnttab_map_refs);
@@ -785,12 +789,16 @@ int gnttab_unmap_refs(struct gnttab_unmap_grant_ref 
*unmap_ops,
        if (xen_feature(XENFEAT_auto_translated_physmap))
                return ret;
 
+       arch_enter_lazy_mmu_mode();
+
        for (i = 0; i < count; i++) {
                ret = m2p_remove_override(pages[i], clear_pte);
                if (ret)
                        return ret;
        }
 
+       arch_leave_lazy_mmu_mode();
+
        return ret;
 }
 EXPORT_SYMBOL_GPL(gnttab_unmap_refs);
-- 
1.7.2.5


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