{
unsigned long address = (unsigned long)vaddr;
unsigned int level;
- pte_t *pte = lookup_address(address, &level);
- unsigned offset = address & ~PAGE_MASK;
+ pte_t *pte;
+ unsigned offset;
- BUG_ON(pte == NULL);
+ /*
+ * if the PFN is in the linear mapped vaddr range, we can just use
+ * the (quick) virt_to_machine() p2m lookup
+ */
+ if (virt_addr_valid(vaddr))
+ return virt_to_machine(vaddr);
+
+ /* otherwise we have to do a (slower) full page-table walk */
+ pte = lookup_address(address, &level);
+ BUG_ON(pte == NULL);
+ offset = address & ~PAGE_MASK;
return XMADDR(((phys_addr_t)pte_mfn(*pte) << PAGE_SHIFT) + offset);
}
xen_mc_batch();
- u.ptr = virt_to_machine(ptep).maddr | MMU_PT_UPDATE_PRESERVE_AD;
+ u.ptr = arbitrary_virt_to_machine(ptep).maddr | MMU_PT_UPDATE_PRESERVE_AD;
u.val = pte_val_ma(pte);
xen_extend_mmu_update(&u);
* For 64-bit, we must skip the Xen hole in the middle of the address
* space, just after the big x86-64 virtual hole.
*/
-static int xen_pgd_walk(pgd_t *pgd, int (*func)(struct page *, enum pt_level),
+static int xen_pgd_walk(struct mm_struct *mm,
+ int (*func)(struct mm_struct *mm, struct page *,
+ enum pt_level),
unsigned long limit)
{
+ pgd_t *pgd = mm->pgd;
int flush = 0;
unsigned hole_low, hole_high;
unsigned pgdidx_limit, pudidx_limit, pmdidx_limit;
pud = pud_offset(&pgd[pgdidx], 0);
if (PTRS_PER_PUD > 1) /* not folded */
- flush |= (*func)(virt_to_page(pud), PT_PUD);
+ flush |= (*func)(mm, virt_to_page(pud), PT_PUD);
for (pudidx = 0; pudidx < PTRS_PER_PUD; pudidx++) {
pmd_t *pmd;
pmd = pmd_offset(&pud[pudidx], 0);
if (PTRS_PER_PMD > 1) /* not folded */
- flush |= (*func)(virt_to_page(pmd), PT_PMD);
+ flush |= (*func)(mm, virt_to_page(pmd), PT_PMD);
for (pmdidx = 0; pmdidx < PTRS_PER_PMD; pmdidx++) {
struct page *pte;
continue;
pte = pmd_page(pmd[pmdidx]);
- flush |= (*func)(pte, PT_PTE);
+ flush |= (*func)(mm, pte, PT_PTE);
}
}
}
out:
/* Do the top level last, so that the callbacks can use it as
a cue to do final things like tlb flushes. */
- flush |= (*func)(virt_to_page(pgd), PT_PGD);
+ flush |= (*func)(mm, virt_to_page(pgd), PT_PGD);
return flush;
}
/* If we're using split pte locks, then take the page's lock and
return a pointer to it. Otherwise return NULL. */
-static spinlock_t *xen_pte_lock(struct page *page)
+static spinlock_t *xen_pte_lock(struct page *page, struct mm_struct *mm)
{
spinlock_t *ptl = NULL;
#if USE_SPLIT_PTLOCKS
ptl = __pte_lockptr(page);
- spin_lock(ptl);
+ spin_lock_nest_lock(ptl, &mm->page_table_lock);
#endif
return ptl;
MULTI_mmuext_op(mcs.mc, op, 1, NULL, DOMID_SELF);
}
-static int xen_pin_page(struct page *page, enum pt_level level)
+static int xen_pin_page(struct mm_struct *mm, struct page *page,
+ enum pt_level level)
{
unsigned pgfl = TestSetPagePinned(page);
int flush;
*/
ptl = NULL;
if (level == PT_PTE)
- ptl = xen_pte_lock(page);
+ ptl = xen_pte_lock(page, mm);
MULTI_update_va_mapping(mcs.mc, (unsigned long)pt,
pfn_pte(pfn, PAGE_KERNEL_RO),
/* This is called just after a mm has been created, but it has not
been used yet. We need to make sure that its pagetable is all
read-only, and can be pinned. */
-void xen_pgd_pin(pgd_t *pgd)
+static void __xen_pgd_pin(struct mm_struct *mm, pgd_t *pgd)
{
+ vm_unmap_aliases();
+
xen_mc_batch();
- if (xen_pgd_walk(pgd, xen_pin_page, USER_LIMIT)) {
- /* re-enable interrupts for kmap_flush_unused */
+ if (xen_pgd_walk(mm, xen_pin_page, USER_LIMIT)) {
+ /* re-enable interrupts for flushing */
xen_mc_issue(0);
+
kmap_flush_unused();
+
xen_mc_batch();
}
xen_do_pin(MMUEXT_PIN_L4_TABLE, PFN_DOWN(__pa(pgd)));
if (user_pgd) {
- xen_pin_page(virt_to_page(user_pgd), PT_PGD);
+ xen_pin_page(mm, virt_to_page(user_pgd), PT_PGD);
xen_do_pin(MMUEXT_PIN_L4_TABLE, PFN_DOWN(__pa(user_pgd)));
}
}
#else /* CONFIG_X86_32 */
#ifdef CONFIG_X86_PAE
/* Need to make sure unshared kernel PMD is pinnable */
- xen_pin_page(virt_to_page(pgd_page(pgd[pgd_index(TASK_SIZE)])), PT_PMD);
+ xen_pin_page(mm, virt_to_page(pgd_page(pgd[pgd_index(TASK_SIZE)])),
+ PT_PMD);
#endif
xen_do_pin(MMUEXT_PIN_L3_TABLE, PFN_DOWN(__pa(pgd)));
#endif /* CONFIG_X86_64 */
xen_mc_issue(0);
}
+static void xen_pgd_pin(struct mm_struct *mm)
+{
+ __xen_pgd_pin(mm, mm->pgd);
+}
+
/*
* On save, we need to pin all pagetables to make sure they get their
* mfns turned into pfns. Search the list for any unpinned pgds and pin
* them (unpinned pgds are not currently in use, probably because the
* process is under construction or destruction).
+ *
+ * Expected to be called in stop_machine() ("equivalent to taking
+ * every spinlock in the system"), so the locking doesn't really
+ * matter all that much.
*/
void xen_mm_pin_all(void)
{
list_for_each_entry(page, &pgd_list, lru) {
if (!PagePinned(page)) {
- xen_pgd_pin((pgd_t *)page_address(page));
+ __xen_pgd_pin(&init_mm, (pgd_t *)page_address(page));
SetPageSavePinned(page);
}
}
* that's before we have page structures to store the bits. So do all
* the book-keeping now.
*/
-static __init int xen_mark_pinned(struct page *page, enum pt_level level)
+static __init int xen_mark_pinned(struct mm_struct *mm, struct page *page,
+ enum pt_level level)
{
SetPagePinned(page);
return 0;
void __init xen_mark_init_mm_pinned(void)
{
- xen_pgd_walk(init_mm.pgd, xen_mark_pinned, FIXADDR_TOP);
+ xen_pgd_walk(&init_mm, xen_mark_pinned, FIXADDR_TOP);
}
-static int xen_unpin_page(struct page *page, enum pt_level level)
+static int xen_unpin_page(struct mm_struct *mm, struct page *page,
+ enum pt_level level)
{
unsigned pgfl = TestClearPagePinned(page);
* partially-pinned state.
*/
if (level == PT_PTE) {
- ptl = xen_pte_lock(page);
+ ptl = xen_pte_lock(page, mm);
if (ptl)
xen_do_pin(MMUEXT_UNPIN_TABLE, pfn);
}
/* Release a pagetables pages back as normal RW */
-static void xen_pgd_unpin(pgd_t *pgd)
+static void __xen_pgd_unpin(struct mm_struct *mm, pgd_t *pgd)
{
xen_mc_batch();
if (user_pgd) {
xen_do_pin(MMUEXT_UNPIN_TABLE, PFN_DOWN(__pa(user_pgd)));
- xen_unpin_page(virt_to_page(user_pgd), PT_PGD);
+ xen_unpin_page(mm, virt_to_page(user_pgd), PT_PGD);
}
}
#endif
#ifdef CONFIG_X86_PAE
/* Need to make sure unshared kernel PMD is unpinned */
- xen_unpin_page(virt_to_page(pgd_page(pgd[pgd_index(TASK_SIZE)])), PT_PMD);
+ xen_unpin_page(mm, virt_to_page(pgd_page(pgd[pgd_index(TASK_SIZE)])),
+ PT_PMD);
#endif
- xen_pgd_walk(pgd, xen_unpin_page, USER_LIMIT);
+ xen_pgd_walk(mm, xen_unpin_page, USER_LIMIT);
xen_mc_issue(0);
}
+static void xen_pgd_unpin(struct mm_struct *mm)
+{
+ __xen_pgd_unpin(mm, mm->pgd);
+}
+
/*
* On resume, undo any pinning done at save, so that the rest of the
* kernel doesn't see any unexpected pinned pagetables.
list_for_each_entry(page, &pgd_list, lru) {
if (PageSavePinned(page)) {
BUG_ON(!PagePinned(page));
- xen_pgd_unpin((pgd_t *)page_address(page));
+ __xen_pgd_unpin(&init_mm, (pgd_t *)page_address(page));
ClearPageSavePinned(page);
}
}
void xen_activate_mm(struct mm_struct *prev, struct mm_struct *next)
{
spin_lock(&next->page_table_lock);
- xen_pgd_pin(next->pgd);
+ xen_pgd_pin(next);
spin_unlock(&next->page_table_lock);
}
void xen_dup_mmap(struct mm_struct *oldmm, struct mm_struct *mm)
{
spin_lock(&mm->page_table_lock);
- xen_pgd_pin(mm->pgd);
+ xen_pgd_pin(mm);
spin_unlock(&mm->page_table_lock);
}
/* pgd may not be pinned in the error exit path of execve */
if (xen_page_pinned(mm->pgd))
- xen_pgd_unpin(mm->pgd);
+ xen_pgd_unpin(mm);
spin_unlock(&mm->page_table_lock);
}
projects / linux-2.6-omap-h63xx.git / blobdiff