Reduce the number of global TLB shootdowns generated by pmap_qenter().

Specifically, teach pmap_qenter() to recognize the case when it is being
asked to replace a mapping with the very same mapping and not generate
a shootdown.  Unfortunately, the buffer cache commonly passes an entire
buffer to pmap_qenter() when only a subset of the mappings are changing.
For the extension of buffers in allocbuf() this was resulting in
unnecessary shootdowns.  The addition of new pages to the end of the
buffer need not and did not trigger a shootdown, but overwriting the
initial mappings with the very same mappings was seen as a change that
necessitated a shootdown.  With this change, that is no longer so.

For a "buildworld" on amd64, this change eliminates 14-15% of the
pmap_invalidate_range() shootdowns, and about 4% of the overall
shootdowns.

MFC after:	3 weeks

1 files changed, 9 insertions, 6 deletions

diff --git a/sys/amd64/amd64/pmap.c b/sys/amd64/amd64/pmap.c
index 3b03d08..74731ab 100644
--- a/sys/amd64/amd64/pmap.c
+++ b/sys/amd64/amd64/pmap.c
@@ -1331,19 +1331,22 @@ pmap_map(vm_offset_t *virt, vm_paddr_t start, vm_paddr_t end, int prot)
 void
 pmap_qenter(vm_offset_t sva, vm_page_t *ma, int count)
 {
-	pt_entry_t *endpte, oldpte, *pte;
+	pt_entry_t *endpte, oldpte, pa, *pte;
+	vm_page_t m;
 
 	oldpte = 0;
 	pte = vtopte(sva);
 	endpte = pte + count;
 	while (pte < endpte) {
-		oldpte |= *pte;
-		pte_store(pte, VM_PAGE_TO_PHYS(*ma) | PG_G |
-		    pmap_cache_bits((*ma)->md.pat_mode, 0) | PG_RW | PG_V);
+		m = *ma++;
+		pa = VM_PAGE_TO_PHYS(m) | pmap_cache_bits(m->md.pat_mode, 0);
+		if ((*pte & (PG_FRAME | PG_PTE_CACHE)) != pa) {
+			oldpte |= *pte;
+			pte_store(pte, pa | PG_G | PG_RW | PG_V);
+		}
 		pte++;
-		ma++;
 	}
-	if ((oldpte & PG_V) != 0)
+	if (__predict_false((oldpte & PG_V) != 0))
 		pmap_invalidate_range(kernel_pmap, sva, sva + count *
 		    PAGE_SIZE);
 }