xen/mm: do direct hypercall in xen_set_pte() if batching is unavailable

In xen_set_pte() if batching is unavailable (because the caller is in
an interrupt context such as handling a page fault) it would fall back
to using native_set_pte() and trapping and emulating the PTE write.

On 32-bit guests this requires two traps for each PTE write (one for
each dword of the PTE).  Instead, do one mmu_update hypercall
directly.

During construction of the initial page tables, continue to use
native_set_pte() because most of the PTEs being set are in writable
and unpinned pages (see phys_pmd_init() in arch/x86/mm/init_64.c) and
using a hypercall for this is very expensive.

This significantly improves page fault performance in 32-bit PV
guests.

lmbench3 test  Before    After     Improvement
----------------------------------------------
lat_pagefault  3.18 us   2.32 us   27%
lat_proc fork  356 us    313.3 us  11%

Signed-off-by: David Vrabel <david.vrabel@citrix.com>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>

diff --git a/arch/x86/xen/mmu.c b/arch/x86/xen/mmu.c
index 3a73785631ce5f7ae8dc0dad1119ebf922e233e0..3f1783a79a3caad992448baec69b095869934725 100644 (file)
--- a/arch/x86/xen/mmu.c
+++ b/arch/x86/xen/mmu.c
@@ -308,8 +308,20 @@ static bool xen_batched_set_pte(pte_t *ptep, pte_t pteval)
 
 static inline void __xen_set_pte(pte_t *ptep, pte_t pteval)
 {
-       if (!xen_batched_set_pte(ptep, pteval))
-               native_set_pte(ptep, pteval);
+       if (!xen_batched_set_pte(ptep, pteval)) {
+               /*
+                * Could call native_set_pte() here and trap and
+                * emulate the PTE write but with 32-bit guests this
+                * needs two traps (one for each of the two 32-bit
+                * words in the PTE) so do one hypercall directly
+                * instead.
+                */
+               struct mmu_update u;
+
+               u.ptr = virt_to_machine(ptep).maddr | MMU_NORMAL_PT_UPDATE;
+               u.val = pte_val_ma(pteval);
+               HYPERVISOR_mmu_update(&u, 1, NULL, DOMID_SELF);
+       }
 }
 
 static void xen_set_pte(pte_t *ptep, pte_t pteval)
@@ -1416,13 +1428,21 @@ static pte_t __init mask_rw_pte(pte_t *ptep, pte_t pte)
 }
 #endif /* CONFIG_X86_64 */
 
-/* Init-time set_pte while constructing initial pagetables, which
-   doesn't allow RO pagetable pages to be remapped RW */
+/*
+ * Init-time set_pte while constructing initial pagetables, which
+ * doesn't allow RO page table pages to be remapped RW.
+ *
+ * Many of these PTE updates are done on unpinned and writable pages
+ * and doing a hypercall for these is unnecessary and expensive.  At
+ * this point it is not possible to tell if a page is pinned or not,
+ * so always write the PTE directly and rely on Xen trapping and
+ * emulating any updates as necessary.
+ */
 static void __init xen_set_pte_init(pte_t *ptep, pte_t pte)
 {
        pte = mask_rw_pte(ptep, pte);
 
-       xen_set_pte(ptep, pte);
+       native_set_pte(ptep, pte);
 }
 
 static void pin_pagetable_pfn(unsigned cmd, unsigned long pfn)