Module Name: src
Committed By: matt
Date: Sun Jun 12 05:32:38 UTC 2011
Modified Files:
src/sys/arch/powerpc/booke: booke_pmap.c trap.c
src/sys/common/pmap/tlb: pmap.c pmap.h
Log Message:
For debugging, use UVMHIST to create kernel histories pmaphist and
pmapexechist. This replaces the old pmapdebug/PDB stuff.
In ddb, you can just use kernhist_dump(pmaphist) to view the history
or vmstat -H from userland.
To generate a diff of this commit:
cvs rdiff -u -r1.4 -r1.5 src/sys/arch/powerpc/booke/booke_pmap.c
cvs rdiff -u -r1.7 -r1.8 src/sys/arch/powerpc/booke/trap.c
cvs rdiff -u -r1.4 -r1.5 src/sys/common/pmap/tlb/pmap.c
cvs rdiff -u -r1.3 -r1.4 src/sys/common/pmap/tlb/pmap.h
Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.
Modified files:
Index: src/sys/arch/powerpc/booke/booke_pmap.c
diff -u src/sys/arch/powerpc/booke/booke_pmap.c:1.4 src/sys/arch/powerpc/booke/booke_pmap.c:1.5
--- src/sys/arch/powerpc/booke/booke_pmap.c:1.4 Sun Jun 5 16:52:24 2011
+++ src/sys/arch/powerpc/booke/booke_pmap.c Sun Jun 12 05:32:38 2011
@@ -37,7 +37,7 @@
#include <sys/cdefs.h>
-__KERNEL_RCSID(0, "$NetBSD: booke_pmap.c,v 1.4 2011/06/05 16:52:24 matt Exp $");
+__KERNEL_RCSID(0, "$NetBSD: booke_pmap.c,v 1.5 2011/06/12 05:32:38 matt Exp $");
#include <sys/param.h>
#include <sys/kcore.h>
@@ -263,8 +263,9 @@
void
pmap_zero_page(paddr_t pa)
{
-// printf("%s(%#lx): calling dcache_zero_page(%#lx)\n", __func__, pa, pa);
dcache_zero_page(pa);
+
+ KASSERT(!VM_PAGE_MD_EXECPAGE_P(PHYS_TO_VM_PAGE(pa)));
}
void
@@ -289,6 +290,8 @@
"r28", "r29", "r30", "r31");
}
}
+
+ KASSERT(!VM_PAGE_MD_EXECPAGE_P(PHYS_TO_VM_PAGE(dst - PAGE_SIZE)));
}
void
Index: src/sys/arch/powerpc/booke/trap.c
diff -u src/sys/arch/powerpc/booke/trap.c:1.7 src/sys/arch/powerpc/booke/trap.c:1.8
--- src/sys/arch/powerpc/booke/trap.c:1.7 Sun Jun 5 16:52:24 2011
+++ src/sys/arch/powerpc/booke/trap.c Sun Jun 12 05:32:38 2011
@@ -1,4 +1,4 @@
-/* $NetBSD: trap.c,v 1.7 2011/06/05 16:52:24 matt Exp $ */
+/* $NetBSD: trap.c,v 1.8 2011/06/12 05:32:38 matt Exp $ */
/*-
* Copyright (c) 2010, 2011 The NetBSD Foundation, Inc.
* All rights reserved.
@@ -39,7 +39,7 @@
#include <sys/cdefs.h>
-__KERNEL_RCSID(1, "$NetBSD: trap.c,v 1.7 2011/06/05 16:52:24 matt Exp $");
+__KERNEL_RCSID(1, "$NetBSD: trap.c,v 1.8 2011/06/12 05:32:38 matt Exp $");
#include <sys/param.h>
#include <sys/systm.h>
@@ -287,11 +287,20 @@
KASSERT(ptep != NULL);
pt_entry_t pte = *ptep;
+ UVMHIST_FUNC(__func__); UVMHIST_CALLED(pmapexechist);
+
if ((pte & PTE_UNSYNCED) == PTE_UNSYNCED) {
const paddr_t pa = pte_to_paddr(pte);
struct vm_page * const pg = PHYS_TO_VM_PAGE(pa);
KASSERT(pg);
+ UVMHIST_LOG(pmapexechist,
+ "srr0=%#x pg=%p (pa %#"PRIxPADDR"): %s",
+ tf->tf_srr0, pg, pa,
+ (VM_PAGE_MD_EXECPAGE_P(pg)
+ ? "no syncicache (already execpage)"
+ : "performed syncicache (now execpage)"));
+
if (!VM_PAGE_MD_EXECPAGE_P(pg)) {
ci->ci_softc->cpu_ev_exec_trap_sync.ev_count++;
dcache_wb_page(pa);
@@ -301,10 +310,12 @@
pte &= ~PTE_UNSYNCED;
pte |= PTE_xX;
*ptep = pte;
+
pmap_tlb_update_addr(faultmap->pmap, trunc_page(faultva),
pte, 0);
kpreempt_enable();
- return false;
+ UVMHIST_LOG(pmapexechist, "<- 0", 0,0,0,0);
+ return 0;
}
kpreempt_enable();
@@ -323,6 +334,7 @@
ksi->ksi_code = SEGV_ACCERR;
ksi->ksi_addr = (void *)tf->tf_srr0; /* not truncated */
}
+ UVMHIST_LOG(pmapexechist, "<- %d", rv, 0,0,0);
return rv;
}
@@ -442,6 +454,12 @@
if (!usertrap_p(tf))
return rv;
+ UVMHIST_FUNC(__func__); UVMHIST_CALLED(pmapexechist);
+
+ UVMHIST_LOG(pmapexechist, " srr0/1=%#x/%#x esr=%#x pte=%#x",
+ tf->tf_srr0, tf->tf_srr1, tf->tf_esr,
+ *trap_pte_lookup(tf, trunc_page(tf->tf_srr0), PSL_IS));
+
ci->ci_ev_pgm.ev_count++;
if (tf->tf_esr & ESR_PTR) {
Index: src/sys/common/pmap/tlb/pmap.c
diff -u src/sys/common/pmap/tlb/pmap.c:1.4 src/sys/common/pmap/tlb/pmap.c:1.5
--- src/sys/common/pmap/tlb/pmap.c:1.4 Sun Jun 5 16:52:27 2011
+++ src/sys/common/pmap/tlb/pmap.c Sun Jun 12 05:32:38 2011
@@ -1,4 +1,4 @@
-/* $NetBSD: pmap.c,v 1.4 2011/06/05 16:52:27 matt Exp $ */
+/* $NetBSD: pmap.c,v 1.5 2011/06/12 05:32:38 matt Exp $ */
/*-
* Copyright (c) 1998, 2001 The NetBSD Foundation, Inc.
@@ -67,7 +67,7 @@
#include <sys/cdefs.h>
-__KERNEL_RCSID(0, "$NetBSD: pmap.c,v 1.4 2011/06/05 16:52:27 matt Exp $");
+__KERNEL_RCSID(0, "$NetBSD: pmap.c,v 1.5 2011/06/12 05:32:38 matt Exp $");
/*
* Manages physical address maps.
@@ -189,21 +189,6 @@
PMAP_COUNTER(protect, "protects");
PMAP_COUNTER(page_protect, "page_protects");
-#define PDB_FOLLOW 0x0001
-#define PDB_INIT 0x0002
-#define PDB_ENTER 0x0004
-#define PDB_REMOVE 0x0008
-#define PDB_CREATE 0x0010
-#define PDB_PTPAGE 0x0020
-#define PDB_PVENTRY 0x0040
-#define PDB_BITS 0x0080
-#define PDB_PROTECT 0x0200
-#define PDB_TLBPID 0x0400
-#define PDB_PARANOIA 0x2000
-#define PDB_WIRING 0x4000
-#define PDB_PVDUMP 0x8000
-int pmapdebug = 0;
-
#define PMAP_ASID_RESERVED 0
CTASSERT(PMAP_ASID_RESERVED == 0);
@@ -243,6 +228,11 @@
} poolpage;
#endif
+#ifdef UVMHIST
+static struct kern_history_ent pmapexechistbuf[10000];
+static struct kern_history_ent pmaphistbuf[10000];
+#endif
+
/*
* The pools from which pmap structures and sub-structures are allocated.
*/
@@ -441,10 +431,10 @@
void
pmap_init(void)
{
-#ifdef DEBUG
- if (pmapdebug & (PDB_FOLLOW|PDB_INIT))
- printf("pmap_init()\n");
-#endif
+ UVMHIST_INIT_STATIC(pmapexechist, pmapexechistbuf);
+ UVMHIST_INIT_STATIC(pmaphist, pmaphistbuf);
+
+ UVMHIST_FUNC(__func__); UVMHIST_CALLED(pmaphist);
/*
* Set a low water mark on the pv_entry pool, so that we are
@@ -478,11 +468,8 @@
{
pmap_t pmap;
+ UVMHIST_FUNC(__func__); UVMHIST_CALLED(pmaphist);
PMAP_COUNT(create);
-#ifdef DEBUG
- if (pmapdebug & (PDB_FOLLOW|PDB_CREATE))
- printf("pmap_create()\n");
-#endif
pmap = pool_get(&pmap_pmap_pool, PR_WAITOK);
memset(pmap, 0, PMAP_SIZE);
@@ -495,6 +482,7 @@
pmap_segtab_alloc(pmap);
+ UVMHIST_LOG(pmaphist, "<- pmap %p", pmap,0,0,0);
return pmap;
}
@@ -506,10 +494,9 @@
void
pmap_destroy(pmap_t pmap)
{
-#ifdef DEBUG
- if (pmapdebug & (PDB_FOLLOW|PDB_CREATE))
- printf("pmap_destroy(%p)\n", pmap);
-#endif
+ UVMHIST_FUNC(__func__); UVMHIST_CALLED(pmaphist);
+ UVMHIST_LOG(pmaphist, "(pmap=%p)", pmap, 0,0,0);
+
if (atomic_dec_uint_nv(&pmap->pm_count) > 0) {
PMAP_COUNT(dereference);
return;
@@ -523,6 +510,8 @@
pool_put(&pmap_pmap_pool, pmap);
kpreempt_enable();
+
+ UVMHIST_LOG(pmaphist, "<- done", 0,0,0,0);
}
/*
@@ -532,14 +521,15 @@
pmap_reference(pmap_t pmap)
{
-#ifdef DEBUG
- if (pmapdebug & PDB_FOLLOW)
- printf("pmap_reference(%p)\n", pmap);
-#endif
+ UVMHIST_FUNC(__func__); UVMHIST_CALLED(pmaphist);
+ UVMHIST_LOG(pmaphist, "(pmap=%p)", pmap, 0,0,0);
+ PMAP_COUNT(reference);
+
if (pmap != NULL) {
atomic_inc_uint(&pmap->pm_count);
}
- PMAP_COUNT(reference);
+
+ UVMHIST_LOG(pmaphist, "<- done", 0,0,0,0);
}
/*
@@ -550,6 +540,8 @@
{
pmap_t pmap = l->l_proc->p_vmspace->vm_map.pmap;
+ UVMHIST_FUNC(__func__); UVMHIST_CALLED(pmaphist);
+ UVMHIST_LOG(pmaphist, "(l=%p (pmap=%p))", l, pmap, 0,0);
PMAP_COUNT(activate);
kpreempt_disable();
@@ -558,6 +550,8 @@
pmap_segtab_activate(pmap, l);
}
kpreempt_enable();
+
+ UVMHIST_LOG(pmaphist, "<- done", 0,0,0,0);
}
/*
@@ -566,39 +560,50 @@
void
pmap_deactivate(struct lwp *l)
{
+ pmap_t pmap = l->l_proc->p_vmspace->vm_map.pmap;
+
+ UVMHIST_FUNC(__func__); UVMHIST_CALLED(pmaphist);
+ UVMHIST_LOG(pmaphist, "(l=%p (pmap=%p))", l, pmap, 0,0);
PMAP_COUNT(deactivate);
kpreempt_disable();
curcpu()->ci_pmap_user_segtab = PMAP_INVALID_SEGTAB_ADDRESS;
- pmap_tlb_asid_deactivate(l->l_proc->p_vmspace->vm_map.pmap);
+ pmap_tlb_asid_deactivate(pmap);
kpreempt_enable();
+
+ UVMHIST_LOG(pmaphist, "<- done", 0,0,0,0);
}
void
-pmap_update(struct pmap *pm)
+pmap_update(struct pmap *pmap)
{
+
+ UVMHIST_FUNC(__func__); UVMHIST_CALLED(pmaphist);
+ UVMHIST_LOG(pmaphist, "(pmap=%p)", pmap, 0,0,0);
PMAP_COUNT(update);
kpreempt_disable();
#ifdef MULTIPROCESSOR
- u_int pending = atomic_swap_uint(&pm->pm_shootdown_pending, 0);
- if (pending && pmap_tlb_shootdown_bystanders(pm))
+ u_int pending = atomic_swap_uint(&pmap->pm_shootdown_pending, 0);
+ if (pending && pmap_tlb_shootdown_bystanders(pmap))
PMAP_COUNT(shootdown_ipis);
#endif
#ifdef DEBUG
- pmap_tlb_check(pm);
+ pmap_tlb_check(pmap);
#endif /* DEBUG */
/*
* If pmap_remove_all was called, we deactivated ourselves and nuked
* our ASID. Now we have to reactivate ourselves.
*/
- if (__predict_false(pm->pm_flags & PMAP_DEFERRED_ACTIVATE)) {
- pm->pm_flags ^= PMAP_DEFERRED_ACTIVATE;
- pmap_tlb_asid_acquire(pm, curlwp);
- pmap_segtab_activate(pm, curlwp);
+ if (__predict_false(pmap->pm_flags & PMAP_DEFERRED_ACTIVATE)) {
+ pmap->pm_flags ^= PMAP_DEFERRED_ACTIVATE;
+ pmap_tlb_asid_acquire(pmap, curlwp);
+ pmap_segtab_activate(pmap, curlwp);
}
kpreempt_enable();
+
+ UVMHIST_LOG(pmaphist, "<- done", 0,0,0,0);
}
/*
@@ -615,12 +620,12 @@
const pt_entry_t npte = flags;
const bool is_kernel_pmap_p = (pmap == pmap_kernel());
-#ifdef DEBUG
- if (pmapdebug & (PDB_FOLLOW|PDB_REMOVE|PDB_PROTECT)) {
- printf("%s: %p, %"PRIxVADDR", %"PRIxVADDR", %p, %"PRIxPTR"\n",
- __func__, pmap, sva, eva, ptep, flags);
- }
-#endif
+ UVMHIST_FUNC(__func__); UVMHIST_CALLED(pmaphist);
+ UVMHIST_LOG(pmaphist, "(pmap=%p %sva=%"PRIxVADDR"..%"PRIxVADDR,
+ pmap, (is_kernel_pmap_p ? "(kernel) " : ""), sva, eva);
+ UVMHIST_LOG(pmaphist, "ptep=%p, flags(npte)=%#"PRIxPTR")",
+ ptep, flags, 0, 0);
+
KASSERT(kpreempt_disabled());
for (; sva < eva; sva += NBPG, ptep++) {
@@ -654,10 +659,9 @@
const bool is_kernel_pmap_p = (pmap == pmap_kernel());
const pt_entry_t npte = pte_nv_entry(is_kernel_pmap_p);
-#ifdef DEBUG
- if (pmapdebug & (PDB_FOLLOW|PDB_REMOVE|PDB_PROTECT))
- printf("pmap_remove(%p, %#"PRIxVADDR", %#"PRIxVADDR")\n", pmap, sva, eva);
-#endif
+ UVMHIST_FUNC(__func__); UVMHIST_CALLED(pmaphist);
+ UVMHIST_LOG(pmaphist, "(pmap=%p, va=%#"PRIxVADDR"..%#"PRIxVADDR")",
+ pmap, sva, eva, 0);
if (is_kernel_pmap_p)
PMAP_COUNT(remove_kernel_calls);
@@ -679,6 +683,8 @@
kpreempt_disable();
pmap_pte_process(pmap, sva, eva, pmap_pte_remove, npte);
kpreempt_enable();
+
+ UVMHIST_LOG(pmaphist, "<- done", 0,0,0,0);
}
/*
@@ -692,13 +698,11 @@
pv_entry_t pv;
vaddr_t va;
+ UVMHIST_FUNC(__func__); UVMHIST_CALLED(pmaphist);
+ UVMHIST_LOG(pmaphist, "(pg=%p (pa %#"PRIxPADDR") prot=%#x)",
+ pg, VM_PAGE_TO_PHYS(pg), prot, 0);
PMAP_COUNT(page_protect);
-#ifdef DEBUG
- if ((pmapdebug & (PDB_FOLLOW|PDB_PROTECT)) ||
- (prot == VM_PROT_NONE && (pmapdebug & PDB_REMOVE)))
- printf("pmap_page_protect(%#"PRIxPADDR", %x)\n",
- VM_PAGE_TO_PHYS(pg), prot);
-#endif
+
switch (prot) {
case VM_PROT_READ|VM_PROT_WRITE:
case VM_PROT_ALL:
@@ -738,6 +742,8 @@
* won't try to sync the icache.
*/
if (pmap_page_clear_attributes(pg, VM_PAGE_MD_EXECPAGE)) {
+ UVMHIST_LOG(pmapexechist, "pg %p (pa %#"PRIxPADDR
+ "): execpage cleared", pg, VM_PAGE_TO_PHYS(pg),0,0);
PMAP_COUNT(exec_uncached_page_protect);
}
(void)VM_PAGE_PVLIST_LOCK(pg, false);
@@ -752,6 +758,8 @@
}
VM_PAGE_PVLIST_UNLOCK(pg);
}
+
+ UVMHIST_LOG(pmaphist, "<- done", 0,0,0,0);
}
static bool
@@ -760,6 +768,12 @@
{
const vm_prot_t prot = (flags & VM_PROT_ALL);
+ UVMHIST_FUNC(__func__); UVMHIST_CALLED(pmaphist);
+ UVMHIST_LOG(pmaphist, "(pmap=%p %sva=%"PRIxVADDR"..%"PRIxVADDR,
+ pmap, (pmap == pmap_kernel() ? "(kernel) " : ""), sva, eva);
+ UVMHIST_LOG(pmaphist, "ptep=%p, flags(npte)=%#"PRIxPTR")",
+ ptep, flags, 0, 0);
+
KASSERT(kpreempt_disabled());
/*
* Change protection on every valid mapping within this segment.
@@ -774,6 +788,10 @@
pmap_md_vca_clean(pg, sva, PMAP_WBINV);
if (VM_PAGE_MD_EXECPAGE_P(pg)) {
KASSERT(VM_PAGE_TO_MD(pg)->mdpg_first.pv_pmap != NULL);
+ UVMHIST_LOG(pmapexechist,
+ "pg %p (pa %#"PRIxPADDR"): %s",
+ pg, VM_PAGE_TO_PHYS(pg),
+ "syncicached performed", 0);
if (pte_cached_p(pt_entry)) {
pmap_page_syncicache(pg);
PMAP_COUNT(exec_synced_protect);
@@ -801,14 +819,15 @@
pmap_protect(pmap_t pmap, vaddr_t sva, vaddr_t eva, vm_prot_t prot)
{
+ UVMHIST_FUNC(__func__); UVMHIST_CALLED(pmaphist);
+ UVMHIST_LOG(pmaphist,
+ " pmap=%p, va=%#"PRIxVADDR"..%#"PRIxVADDR" port=%#x)",
+ pmap, sva, eva, prot);
PMAP_COUNT(protect);
-#ifdef DEBUG
- if (pmapdebug & (PDB_FOLLOW|PDB_PROTECT))
- printf("pmap_protect(%p, %#"PRIxVADDR", %#"PRIxVADDR", %x)\n",
- pmap, sva, eva, prot);
-#endif
+
if ((prot & VM_PROT_READ) == VM_PROT_NONE) {
pmap_remove(pmap, sva, eva);
+ UVMHIST_LOG(pmaphist, "<- done", 0,0,0,0);
return;
}
@@ -832,6 +851,8 @@
kpreempt_disable();
pmap_pte_process(pmap, sva, eva, pmap_pte_protect, prot);
kpreempt_enable();
+
+ UVMHIST_LOG(pmaphist, "<- done", 0,0,0,0);
}
#if defined(VM_PAGE_MD_CACHED)
@@ -843,12 +864,11 @@
static void
pmap_page_cache(struct vm_page *pg, bool cached)
{
+ UVMHIST_FUNC(__func__); UVMHIST_CALLED(pmaphist);
+ UVMHIST_LOG(pmaphist, "(pg=%p (pa %#"PRIxPADDR") cached=%s)",
+ pg, VM_PAGE_TO_PHYS(pg), cached ? "true" : "false", 0);
KASSERT(kpreempt_disabled());
-#ifdef DEBUG
- if (pmapdebug & (PDB_FOLLOW|PDB_ENTER))
- printf("pmap_page_uncache(%#"PRIxPADDR")\n", VM_PAGE_TO_PHYS(pg));
-#endif
-
+
if (cached) {
pmap_page_clear_attributes(pg, VM_PAGE_MD_UNCACHED);
PMAP_COUNT(page_cache_restorations);
@@ -878,6 +898,7 @@
PMAP_TLB_NEED_IPI);
}
}
+ UVMHIST_LOG(pmaphist, "<- done", 0,0,0,0);
}
#endif /* VM_PAGE_MD_CACHED */
@@ -899,12 +920,20 @@
pt_entry_t npte;
const bool wired = (flags & PMAP_WIRED) != 0;
const bool is_kernel_pmap_p = (pmap == pmap_kernel());
-
-#ifdef DEBUG
- if (pmapdebug & (PDB_FOLLOW|PDB_ENTER))
- printf("pmap_enter(%p, %#"PRIxVADDR", %#"PRIxPADDR", %x, %x)\n",
- pmap, va, pa, prot, wired);
+#ifdef UVMHIST
+ struct kern_history * const histp =
+ ((prot & VM_PROT_EXECUTE) ? &pmapexechist : &pmaphist);
#endif
+
+ UVMHIST_FUNC(__func__);
+#define VM_PROT_STRING(prot) \
+ &"\0 (R)\0 (W)\0 (RW)\0 (X)\0 (RX)\0 (WX)\0 (RWX)\0"[UVM_PROTECTION(prot)*6]
+ UVMHIST_CALLED(*histp);
+ UVMHIST_LOG(*histp, "(pmap=%p, va=%#"PRIxVADDR", pa=%#"PRIxPADDR,
+ pmap, va, pa, 0);
+ UVMHIST_LOG(*histp, "prot=%#x%s flags=%#x%s)",
+ prot, VM_PROT_STRING(prot), flags, VM_PROT_STRING(flags));
+
const bool good_color = PMAP_PAGE_COLOROK_P(pa, va);
if (is_kernel_pmap_p) {
PMAP_COUNT(kernel_mappings);
@@ -921,8 +950,8 @@
__func__, is_kernel_pmap_p ? "kva" : "uva", va);
#endif
- if (!(prot & VM_PROT_READ))
- panic("%s: no READ in prot %u ", __func__, prot);
+ KASSERTMSG(prot & VM_PROT_READ,
+ ("%s: no READ (%#x) in prot %#x", __func__, VM_PROT_READ, prot));
struct vm_page * const pg = PHYS_TO_VM_PAGE(pa);
@@ -954,6 +983,7 @@
pt_entry_t * const ptep = pmap_pte_reserve(pmap, va, flags);
if (__predict_false(ptep == NULL)) {
kpreempt_enable();
+ UVMHIST_LOG(*histp, "<- ENOMEM", 0,0,0,0);
return ENOMEM;
}
pt_entry_t opte = *ptep;
@@ -971,12 +1001,8 @@
pmap->pm_stats.wired_count++;
npte |= pte_wired_entry();
}
-#if defined(DEBUG)
- if (pmapdebug & PDB_ENTER) {
- printf("pmap_enter: %p: %#"PRIxVADDR": new pte %#x (pa %#"PRIxPADDR")", pmap, va, npte, pa);
- printf("\n");
- }
-#endif
+
+ UVMHIST_LOG(*histp, "new pte %#x (pa %#"PRIxPADDR")", npte, pa, 0,0);
if (pte_valid_p(opte) && pte_to_paddr(opte) != pa) {
pmap_remove(pmap, va, va + NBPG);
@@ -995,22 +1021,44 @@
kpreempt_enable();
if (pg != NULL && (prot == (VM_PROT_READ | VM_PROT_EXECUTE))) {
-#ifdef DEBUG
- if (pmapdebug & PDB_ENTER)
- printf("pmap_enter: flush I cache va %#"PRIxVADDR" (%#"PRIxPADDR")\n",
- va, pa);
-#endif
PMAP_COUNT(exec_mappings);
- if ((flags & VM_PROT_EXECUTE)
- && !pte_deferred_exec_p(npte)
- && !VM_PAGE_MD_EXECPAGE_P(pg)
- && pte_cached_p(npte)) {
- pmap_page_syncicache(pg);
- pmap_page_set_attributes(pg, VM_PAGE_MD_EXECPAGE);
- PMAP_COUNT(exec_synced_mappings);
+ if (!VM_PAGE_MD_EXECPAGE_P(pg) && pte_cached_p(npte)) {
+ if (!pte_deferred_exec_p(npte)) {
+ UVMHIST_LOG(*histp,
+ "va=%#"PRIxVADDR" pg %p: %s syncicache%s",
+ va, pg, "immediate", "");
+ pmap_page_syncicache(pg);
+ pmap_page_set_attributes(pg,
+ VM_PAGE_MD_EXECPAGE);
+ PMAP_COUNT(exec_synced_mappings);
+ } else {
+ UVMHIST_LOG(*histp, "va=%#"PRIxVADDR
+ " pg %p: %s syncicache: pte %#x",
+ va, pg, "defer", npte);
+ }
+ } else {
+ UVMHIST_LOG(*histp,
+ "va=%#"PRIxVADDR" pg %p: %s syncicache%s",
+ va, pg, "no",
+ (pte_cached_p(npte)
+ ? " (already exec)"
+ : " (uncached)"));
}
+ } else if (pg != NULL && (prot & VM_PROT_EXECUTE)) {
+ KASSERT(prot & VM_PROT_WRITE);
+ PMAP_COUNT(exec_mappings);
+ pmap_page_syncicache(pg);
+ pmap_page_clear_attributes(pg, VM_PAGE_MD_EXECPAGE);
+ UVMHIST_LOG(pmapexechist,
+ "va=%#"PRIxVADDR" pg %p: %s syncicache%s",
+ va, pg, "immediate", " (writeable)");
}
+ if (prot & VM_PROT_EXECUTE) {
+ UVMHIST_LOG(pmapexechist, "<- 0 (OK)", 0,0,0,0);
+ } else {
+ UVMHIST_LOG(pmaphist, "<- 0 (OK)", 0,0,0,0);
+ }
return 0;
}
@@ -1019,11 +1067,11 @@
{
struct vm_page * const pg = PHYS_TO_VM_PAGE(pa);
-#ifdef DEBUG
- if (pmapdebug & (PDB_FOLLOW|PDB_ENTER))
- printf("pmap_kenter_pa(%#"PRIxVADDR", %#"PRIxPADDR", %x, %x)\n", va, pa, prot, flags);
-#endif
+ UVMHIST_FUNC(__func__); UVMHIST_CALLED(pmaphist);
+ UVMHIST_LOG(pmaphist, "(va=%#"PRIxVADDR" pa=%#"PRIxPADDR
+ ", prot=%#x, flags=%#x)", va, pa, prot, flags);
PMAP_COUNT(kenter_pa);
+
if (!PMAP_PAGE_COLOROK_P(pa, va) && pg != NULL)
PMAP_COUNT(kenter_pa_bad);
@@ -1053,6 +1101,7 @@
((long *)va)[i], va, ((long *)pa)[i], pa);
}
#endif
+ UVMHIST_LOG(pmaphist, "<- done", 0,0,0,0);
}
static bool
@@ -1091,14 +1140,15 @@
const vaddr_t sva = trunc_page(va);
const vaddr_t eva = round_page(va + len);
-#ifdef DEBUG
- if (pmapdebug & (PDB_FOLLOW|PDB_REMOVE))
- printf("pmap_kremove(%#"PRIxVADDR", %#"PRIxVSIZE")\n", va, len);
-#endif
+ UVMHIST_FUNC(__func__); UVMHIST_CALLED(pmaphist);
+ UVMHIST_LOG(pmaphist, "(va=%#"PRIxVADDR" len=%#"PRIxVSIZE")",
+ va, len, 0,0);
kpreempt_disable();
pmap_pte_process(pmap_kernel(), sva, eva, pmap_pte_kremove, 0);
kpreempt_enable();
+
+ UVMHIST_LOG(pmaphist, "<- done", 0,0,0,0);
}
void
@@ -1129,17 +1179,16 @@
pmap_unwire(pmap_t pmap, vaddr_t va)
{
+ UVMHIST_FUNC(__func__); UVMHIST_CALLED(pmaphist);
+ UVMHIST_LOG(pmaphist, "(pmap=%p va=%#"PRIxVADDR")", pmap, va, 0,0);
PMAP_COUNT(unwire);
-#ifdef DEBUG
- if (pmapdebug & (PDB_FOLLOW|PDB_WIRING))
- printf("pmap_unwire(%p, %#"PRIxVADDR")\n", pmap, va);
-#endif
+
/*
* Don't need to flush the TLB since PG_WIRED is only in software.
*/
#ifdef PARANOIADIAG
- if (va < pmap->pm_minaddr || pmap->pm_maxaddr <= va)
- panic("pmap_unwire");
+ if (va < pmap->pm_minaddr || pmap->pm_maxaddr <= va)
+ panic("pmap_unwire");
#endif
kpreempt_disable();
pt_entry_t * const ptep = pmap_pte_lookup(pmap, va);
@@ -1180,10 +1229,8 @@
{
paddr_t pa;
-#ifdef DEBUG
- if (pmapdebug & PDB_FOLLOW)
- printf("pmap_extract(%p, %#"PRIxVADDR") -> ", pmap, va);
-#endif
+ //UVMHIST_FUNC(__func__); UVMHIST_CALLED(pmaphist);
+ //UVMHIST_LOG(pmaphist, "(pmap=%p va=%#"PRIxVADDR")", pmap, va, 0,0);
if (pmap == pmap_kernel()) {
if (pmap_md_direct_mapped_vaddr_p(va)) {
pa = pmap_md_direct_mapped_vaddr_to_paddr(va);
@@ -1195,18 +1242,12 @@
kpreempt_disable();
pt_entry_t * const ptep = pmap_pte_lookup(pmap, va);
if (ptep == NULL) {
-#ifdef DEBUG
- if (pmapdebug & PDB_FOLLOW)
- printf("not in segmap\n");
-#endif
+ //UVMHIST_LOG(pmaphist, "<- false (not in segmap)", 0,0,0,0);
kpreempt_enable();
return false;
}
if (!pte_valid_p(*ptep)) {
-#ifdef DEBUG
- if (pmapdebug & PDB_FOLLOW)
- printf("PTE not valid\n");
-#endif
+ //UVMHIST_LOG(pmaphist, "<- false (PTE not valid)", 0,0,0,0);
kpreempt_enable();
return false;
}
@@ -1216,10 +1257,7 @@
if (pap != NULL) {
*pap = pa;
}
-#ifdef DEBUG
- if (pmapdebug & PDB_FOLLOW)
- printf("pa %#"PRIxPADDR"\n", pa);
-#endif
+ //UVMHIST_LOG(pmaphist, "<- true (pa %#"PRIxPADDR")", pa, 0,0,0);
return true;
}
@@ -1235,12 +1273,8 @@
vaddr_t src_addr)
{
+ UVMHIST_FUNC(__func__); UVMHIST_CALLED(pmaphist);
PMAP_COUNT(copy);
-#ifdef DEBUG
- if (pmapdebug & PDB_FOLLOW)
- printf("pmap_copy(%p, %p, %#"PRIxVADDR", %#"PRIxVSIZE", %#"PRIxVADDR")\n",
- dst_pmap, src_pmap, dst_addr, len, src_addr);
-#endif
}
/*
@@ -1251,12 +1285,15 @@
bool
pmap_clear_reference(struct vm_page *pg)
{
-#ifdef DEBUG
- if (pmapdebug & PDB_FOLLOW)
- printf("pmap_clear_reference(%#"PRIxPADDR")\n",
- VM_PAGE_TO_PHYS(pg));
-#endif
- return pmap_page_clear_attributes(pg, VM_PAGE_MD_REFERENCED);
+ UVMHIST_FUNC(__func__); UVMHIST_CALLED(pmaphist);
+ UVMHIST_LOG(pmaphist, "(pg=%p (pa %#"PRIxPADDR"))",
+ pg, VM_PAGE_TO_PHYS(pg), 0,0);
+
+ bool rv = pmap_page_clear_attributes(pg, VM_PAGE_MD_REFERENCED);
+
+ UVMHIST_LOG(pmaphist, "<- %s", rv ? "true" : "false", 0,0,0);
+
+ return rv;
}
/*
@@ -1282,23 +1319,32 @@
pv_entry_t pv_next;
uint16_t gen;
+ UVMHIST_FUNC(__func__); UVMHIST_CALLED(pmaphist);
+ UVMHIST_LOG(pmaphist, "(pg=%p (%#"PRIxPADDR"))",
+ pg, VM_PAGE_TO_PHYS(pg), 0,0);
PMAP_COUNT(clear_modify);
-#ifdef DEBUG
- if (pmapdebug & PDB_FOLLOW)
- printf("pmap_clear_modify(%#"PRIxPADDR")\n", VM_PAGE_TO_PHYS(pg));
-#endif
+
if (VM_PAGE_MD_EXECPAGE_P(pg)) {
if (pv->pv_pmap == NULL) {
+ UVMHIST_LOG(pmapexechist,
+ "pg %p (pa %#"PRIxPADDR"): %s",
+ pg, VM_PAGE_TO_PHYS(pg), "execpage cleared", 0);
pmap_page_clear_attributes(pg, VM_PAGE_MD_EXECPAGE);
PMAP_COUNT(exec_uncached_clear_modify);
} else {
+ UVMHIST_LOG(pmapexechist,
+ "pg %p (pa %#"PRIxPADDR"): %s",
+ pg, VM_PAGE_TO_PHYS(pg), "syncicache performed", 0);
pmap_page_syncicache(pg);
PMAP_COUNT(exec_synced_clear_modify);
}
}
- if (!pmap_page_clear_attributes(pg, VM_PAGE_MD_MODIFIED))
+ if (!pmap_page_clear_attributes(pg, VM_PAGE_MD_MODIFIED)) {
+ UVMHIST_LOG(pmaphist, "<- false", 0,0,0,0);
return false;
+ }
if (pv->pv_pmap == NULL) {
+ UVMHIST_LOG(pmaphist, "<- true (no mappings)", 0,0,0,0);
return true;
}
@@ -1333,6 +1379,8 @@
}
VM_PAGE_PVLIST_UNLOCK(pg);
kpreempt_enable();
+
+ UVMHIST_LOG(pmaphist, "<- true (mappings changed)", 0,0,0,0);
return true;
}
@@ -1386,17 +1434,19 @@
{
pv_entry_t pv, npv, apv;
int16_t gen;
+ bool first = false;
+
+ UVMHIST_FUNC(__func__); UVMHIST_CALLED(pmaphist);
+ UVMHIST_LOG(pmaphist,
+ "(pmap=%p va=%#"PRIxVADDR" pg=%p (%#"PRIxPADDR")",
+ pmap, va, pg, VM_PAGE_TO_PHYS(pg));
+ UVMHIST_LOG(pmaphist, "nptep=%p (%#x))", npte, *npte, 0, 0);
KASSERT(kpreempt_disabled());
KASSERT(pmap != pmap_kernel() || !pmap_md_direct_mapped_vaddr_p(va));
apv = NULL;
pv = &VM_PAGE_TO_MD(pg)->mdpg_first;
-#ifdef DEBUG
- if (pmapdebug & PDB_ENTER)
- printf("pmap_enter: pv %p: was %#"PRIxVADDR"/%p/%p\n",
- pv, pv->pv_va, pv->pv_pmap, pv->pv_next);
-#endif
gen = VM_PAGE_PVLIST_LOCK(pg, true);
pmap_check_pvlist(pg);
again:
@@ -1405,13 +1455,9 @@
/*
* No entries yet, use header as the first entry
*/
-#ifdef DEBUG
- if (pmapdebug & PDB_PVENTRY)
- printf("pmap_enter_pv: first pv: pmap %p va %#"PRIxVADDR"\n",
- pmap, va);
-#endif
PMAP_COUNT(primary_mappings);
PMAP_COUNT(mappings);
+ first = true;
#ifdef VM_PAGE_MD_UNCACHED
pmap_page_clear_attributes(pg, VM_PAGE_MD_UNCACHED);
#endif
@@ -1450,11 +1496,6 @@
return;
}
}
-#ifdef DEBUG
- if (pmapdebug & PDB_PVENTRY)
- printf("pmap_enter_pv: new pv: pmap %p va %#"PRIxVADDR"\n",
- pmap, va);
-#endif
if (__predict_true(apv == NULL)) {
#ifdef MULTIPROCESSOR
/*
@@ -1491,6 +1532,9 @@
VM_PAGE_PVLIST_UNLOCK(pg);
if (__predict_false(apv != NULL))
pmap_pv_free(apv);
+
+ UVMHIST_LOG(pmaphist, "<- done pv=%p%s",
+ pv, first ? " (first pv)" : "",0,0);
}
/*
@@ -1506,11 +1550,12 @@
pv_entry_t pv, npv;
bool last;
-#ifdef DEBUG
- if (pmapdebug & (PDB_FOLLOW|PDB_PVENTRY))
- printf("pmap_remove_pv(%p, %#"PRIxVADDR", %#"PRIxPADDR")\n", pmap, va,
- VM_PAGE_TO_PHYS(pg));
-#endif
+ UVMHIST_FUNC(__func__); UVMHIST_CALLED(pmaphist);
+ UVMHIST_LOG(pmaphist,
+ "(pmap=%p va=%#"PRIxVADDR" pg=%p (pa %#"PRIxPADDR")\n",
+ pmap, va, pg, VM_PAGE_TO_PHYS(pg));
+ UVMHIST_LOG(pmaphist, "dirty=%s)", dirty ? "true" : "false", 0,0,0);
+
KASSERT(kpreempt_disabled());
pv = &VM_PAGE_TO_MD(pg)->mdpg_first;
@@ -1564,6 +1609,11 @@
* If this was the page's last mapping, we no longer
* care about its execness.
*/
+ UVMHIST_LOG(pmapexechist,
+ "pg %p (pa %#"PRIxPADDR")%s: %s",
+ pg, VM_PAGE_TO_PHYS(pg),
+ last ? " [last mapping]" : "",
+ "execpage cleared");
pmap_page_clear_attributes(pg, VM_PAGE_MD_EXECPAGE);
PMAP_COUNT(exec_uncached_remove);
} else {
@@ -1571,10 +1621,16 @@
* Someone still has it mapped as an executable page
* so we must sync it.
*/
+ UVMHIST_LOG(pmapexechist,
+ "pg %p (pa %#"PRIxPADDR")%s: %s",
+ pg, VM_PAGE_TO_PHYS(pg),
+ last ? " [last mapping]" : "",
+ "performed syncicache");
pmap_page_syncicache(pg);
PMAP_COUNT(exec_synced_remove);
}
}
+ UVMHIST_LOG(pmaphist, "<- done", 0,0,0,0);
}
#ifdef MULTIPROCESSOR
Index: src/sys/common/pmap/tlb/pmap.h
diff -u src/sys/common/pmap/tlb/pmap.h:1.3 src/sys/common/pmap/tlb/pmap.h:1.4
--- src/sys/common/pmap/tlb/pmap.h:1.3 Sun Jun 5 16:52:27 2011
+++ src/sys/common/pmap/tlb/pmap.h Sun Jun 12 05:32:38 2011
@@ -1,4 +1,4 @@
-/* $NetBSD: pmap.h,v 1.3 2011/06/05 16:52:27 matt Exp $ */
+/* $NetBSD: pmap.h,v 1.4 2011/06/12 05:32:38 matt Exp $ */
/*
* Copyright (c) 1992, 1993
@@ -74,6 +74,12 @@
#ifndef _COMMON_PMAP_H_
#define _COMMON_PMAP_H_
+#include <uvm/uvm_stat.h>
+#ifdef UVMHIST
+UVMHIST_DECL(pmapexechist);
+UVMHIST_DECL(pmaphist);
+#endif
+
/*
* The user address space is mapped using a two level structure where
* virtual address bits 31..22 are used to index into a segment table which
