Rework some code in the powerpc pmap module.

Previously it was possible to remove multiple entries on a single pass thru
the pv deletion code in pmap_page_protect. Also when it did this, it
was not properly tracking of decrementing pm->pm_stats.resident_count.

By storing an additional piece of information in the pv structure,
the pmap pointer it is possible to call pmap_remove on the
entries rather than duplicating the pte removal code (again).

This fixes a problem seen where the system panics or hangs in
pmap_remove_pv due to the item not being on the list. Now
it is not possible for mix entries.

Because the pv entry only contained the pteidx, it was possible
to remove an incorrect entry due to ambiguity. multiple pmaps
having mappings at the same va of the same pa. Multipe pmap containing similar
entries will occur frequenty with shared libaries. Because of the hash
entries that have the same result of (sr & ptab_mask) will alias in the pv
list..

Since the pv_idx is now recomputable, should it be removed?

1 files changed, 32 insertions, 37 deletions

diff --git a/sys/arch/powerpc/powerpc/pmap.c b/sys/arch/powerpc/powerpc/pmap.c
index d99e5b3a9a2..44a2f5a999d 100644
--- a/sys/arch/powerpc/powerpc/pmap.c
+++ b/sys/arch/powerpc/powerpc/pmap.c
@@ -1,4 +1,4 @@
-/*	$OpenBSD: pmap.c,v 1.17 2000/06/15 03:20:53 rahnds Exp $	*/
+/*	$OpenBSD: pmap.c,v 1.18 2000/07/12 13:49:54 rahnds Exp $	*/
 /*	$NetBSD: pmap.c,v 1.1 1996/09/30 16:34:52 ws Exp $	*/
 
 /*
@@ -102,6 +102,7 @@ struct pv_entry {
 	struct pv_entry *pv_next;	/* Linked list of mappings */
 	int pv_idx;			/* Index into ptable */
 	vm_offset_t pv_va;		/* virtual address of mapping */
+	struct pmap *pv_pmap;		/* pmap associated with this map */
 };
 
 struct pv_entry *pv_table;
@@ -671,10 +672,12 @@ pmap_pinit(pm)
 	struct pmap *pm;
 {
 	int i, j;
+	int s;
 	
 	/*
 	 * Allocate some segment registers for this pmap.
 	 */
+	s = splimp();
 	pm->pm_refs = 1;
 	for (i = 0; i < sizeof usedsr / sizeof usedsr[0]; i++)
 		if (usedsr[i] != 0xffffffff) {
@@ -683,8 +686,10 @@ pmap_pinit(pm)
 			pm->pm_sr[0] = (i * sizeof usedsr[0] * 8 + j) * 16;
 			for (i = 1; i < 16; i++)
 				pm->pm_sr[i] = pm->pm_sr[i - 1] + 1;
+			splx(s);
 			return;
 		}
+	splx(s);
 	panic("out of segments");
 }
 
@@ -942,7 +947,8 @@ pofree(po, freepage)
  * This returns whether this is the first mapping of a page.
  */
 static inline int
-pmap_enter_pv(pteidx, va, pind)
+pmap_enter_pv(pm, pteidx, va, pind)
+	struct pmap *pm;
 	int pteidx;
 	vm_offset_t va;
 	u_int pind;
@@ -962,6 +968,7 @@ pmap_enter_pv(pteidx, va, pind)
 		 */
 		pv->pv_va = va;
 		pv->pv_idx = pteidx;
+		pv->pv_pmap = pm;
 		pv->pv_next = NULL;
 	} else {
 		/*
@@ -971,6 +978,7 @@ pmap_enter_pv(pteidx, va, pind)
 		npv = pmap_alloc_pv();
 		npv->pv_va = va;
 		npv->pv_idx = pteidx;
+		npv->pv_pmap = pm;
 		npv->pv_next = pv->pv_next;
 		pv->pv_next = npv;
 	}
@@ -1007,7 +1015,7 @@ pmap_remove_pv(pm, pteidx, va, pind, pte)
 	 * to the header.  Otherwise we must search the list for
 	 * the entry.  In either case we free the now unused entry.
 	 */
-	if (pteidx == pv->pv_idx && va == pv->pv_va) {
+	if (pteidx == pv->pv_idx && va == pv->pv_va && pm == pv->pv_pmap) {
 		npv = pv->pv_next;
 		if (npv) {
 			*pv = *npv;
@@ -1016,17 +1024,25 @@ pmap_remove_pv(pm, pteidx, va, pind, pte)
 			pv->pv_idx = -1;
 		}
 	} else {
-		for (; npv = pv->pv_next; pv = npv)
-			if (pteidx == npv->pv_idx && va == npv->pv_va)
+		for (; npv = pv->pv_next; pv = npv) {
+			if (pteidx == npv->pv_idx && va == npv->pv_va &&
+				pm == npv->pv_pmap)
+			{
 				break;
+			}
+		}
 		if (npv) {
 			pv->pv_next = npv->pv_next;
 			pmap_free_pv(npv);
 		}
-#if 0
+#if 1
 #ifdef	DIAGNOSTIC
-		else
+		else {
+			printf("pmap_remove_pv: not on list");
+			/*
 			panic("pmap_remove_pv: not on list");
+			*/
+		}
 #endif
 #endif
 	}
@@ -1083,7 +1099,7 @@ pmap_enter(pm, va, pa, prot, wired, access_type)
 	 * Now record mapping for later back-translation.
 	 */
 	if (pmap_initialized && (i = pmap_page_index(pa)) != -1)
-		if (pmap_enter_pv(idx, va, i)) {
+		if (pmap_enter_pv(pm, idx, va, i)) {
 			/* 
 			 * Flush the real memory from the cache.
 			 */
@@ -1376,6 +1392,9 @@ pmap_page_protect(pa, prot)
 	pte_t *ptp;
 	struct pte_ovfl *po, *npo;
 	int i, s, pind, idx;
+	int found;
+	struct pmap *pm;
+	struct pv_entry *pv;
 	
 	pa &= ~ADDR_POFF;
 	if (prot & VM_PROT_READ) {
@@ -1387,36 +1406,12 @@ pmap_page_protect(pa, prot)
 	if (pind < 0)
 		return;
 
+	pv = &pv_table[pind];
 	s = splimp();
-	while (pv_table[pind].pv_idx >= 0) {
-		idx = pv_table[pind].pv_idx;
-		va = pv_table[pind].pv_va;
-		for (ptp = ptable + idx * 8, i = 8; --i >= 0; ptp++)
-			if ((ptp->pte_hi & PTE_VALID)
-			    && (ptp->pte_lo & PTE_RPGN) == pa) {
-				pmap_remove_pv(NULL, idx, va, pind, ptp);
-				ptp->pte_hi &= ~PTE_VALID;
-				asm volatile ("sync");
-				tlbie(va);
-				tlbsync();
-			}
-		for (ptp = ptable + (idx ^ ptab_mask) * 8, i = 8; --i >= 0; ptp++)
-			if ((ptp->pte_hi & PTE_VALID)
-			    && (ptp->pte_lo & PTE_RPGN) == pa) {
-				pmap_remove_pv(NULL, idx, va, pind, ptp);
-				ptp->pte_hi &= ~PTE_VALID;
-				asm volatile ("sync");
-				tlbie(va);
-				tlbsync();
-			}
-		for (po = potable[idx].lh_first; po; po = npo) {
-			npo = po->po_list.le_next;
-			if ((po->po_pte.pte_lo & PTE_RPGN) == pa) {
-				pmap_remove_pv(NULL,idx, va, pind, &po->po_pte);
-				LIST_REMOVE(po, po_list);
-				pofree(po, 1);
-			}
-		}
+	while (pv->pv_idx != -1) {
+		va = pv->pv_va;
+		pm = pv->pv_pmap;
+		pmap_remove(pm, va, va + NBPG);
 	}
 	splx(s);
 }