/*H:480
* (vi) Mapping the Switcher when the Guest is about to run.
*
* The Switcher and the two pages for this CPU need to be visible in the Guest
* (and not the pages for other CPUs).
*
* The pages for the pagetables have all been allocated before: we just need
* to make sure the actual PTEs are up-to-date for the CPU we're about to run
* on.
*/
void map_switcher_in_guest(struct lg_cpu *cpu, struct lguest_pages *pages)
{
unsigned long base;
struct page *percpu_switcher_page, *regs_page;
pte_t *pte;
struct pgdir *pgdir = &cpu->lg->pgdirs[cpu->cpu_pgd];
/* Switcher page should always be mapped by now! */
BUG_ON(!pgdir->switcher_mapped);
/*
* Remember that we have two pages for each Host CPU, so we can run a
* Guest on each CPU without them interfering. We need to make sure
* those pages are mapped correctly in the Guest, but since we usually
* run on the same CPU, we cache that, and only update the mappings
* when we move.
*/
if (pgdir->last_host_cpu == raw_smp_processor_id())
return;
/* -1 means unknown so we remove everything. */
if (pgdir->last_host_cpu == -1) {
unsigned int i;
for_each_possible_cpu(i)
remove_switcher_percpu_map(cpu, i);
} else {
/* We know exactly what CPU mapping to remove. */
remove_switcher_percpu_map(cpu, pgdir->last_host_cpu);
}
/*
* When we're running the Guest, we want the Guest's "regs" page to
* appear where the first Switcher page for this CPU is. This is an
* optimization: when the Switcher saves the Guest registers, it saves
* them into the first page of this CPU's "struct lguest_pages": if we
* make sure the Guest's register page is already mapped there, we
* don't have to copy them out again.
*/
/* Find the shadow PTE for this regs page. */
base = switcher_addr + PAGE_SIZE
+ raw_smp_processor_id() * sizeof(struct lguest_pages);
pte = find_spte(cpu, base, false, 0, 0);
regs_page = pfn_to_page(__pa(cpu->regs_page) >> PAGE_SHIFT);
get_page(regs_page);
set_pte(pte, mk_pte(regs_page, __pgprot(__PAGE_KERNEL & ~_PAGE_GLOBAL)));
/*
* We map the second page of the struct lguest_pages read-only in
* the Guest: the IDT, GDT and other things it's not supposed to
* change.
*/
pte = find_spte(cpu, base + PAGE_SIZE, false, 0, 0);
percpu_switcher_page
= lg_switcher_pages[1 + raw_smp_processor_id()*2 + 1];
get_page(percpu_switcher_page);
set_pte(pte, mk_pte(percpu_switcher_page,
__pgprot(__PAGE_KERNEL_RO & ~_PAGE_GLOBAL)));
pgdir->last_host_cpu = raw_smp_processor_id();
}
static int
__change_page_attr(struct page *page, pgprot_t prot)
{
pte_t *kpte;
unsigned long address;
struct page *kpte_page;
BUG_ON(PageHighMem(page));
address = (unsigned long)page_address(page);
kpte = lookup_address(address);
if (!kpte)
return -EINVAL;
kpte_page = virt_to_page(kpte);
if (pgprot_val(prot) != pgprot_val(PAGE_KERNEL)) {
if ((pte_val(*kpte) & _PAGE_PSE) == 0) {
set_pte_atomic(kpte, mk_pte(page, prot));
} else {
pgprot_t ref_prot;
struct page *split;
ref_prot =
((address & LARGE_PAGE_MASK) < (unsigned long)&_etext)
? PAGE_KERNEL_EXEC : PAGE_KERNEL;
split = split_large_page(address, prot, ref_prot);
if (!split)
return -ENOMEM;
set_pmd_pte(kpte,address,mk_pte(split, ref_prot));
kpte_page = split;
}
page_private(kpte_page)++;
} else if ((pte_val(*kpte) & _PAGE_PSE) == 0) {
set_pte_atomic(kpte, mk_pte(page, PAGE_KERNEL));
BUG_ON(page_private(kpte_page) == 0);
page_private(kpte_page)--;
} else
BUG();
/*
* If the pte was reserved, it means it was created at boot
* time (not via split_large_page) and in turn we must not
* replace it with a largepage.
*/
if (!PageReserved(kpte_page)) {
if (cpu_has_pse && (page_private(kpte_page) == 0)) {
ClearPagePrivate(kpte_page);
list_add(&kpte_page->lru, &df_list);
revert_page(kpte_page, address);
}
}
return 0;
}
/*
* This routine is used to map in a page into an address space: needed by
* execve() for the initial stack and environment pages.
*/
unsigned long put_dirty_page(struct task_struct * tsk, unsigned long page, unsigned long address)
{
pgd_t * pgd;
pmd_t * pmd;
pte_t * pte;
if (page >= high_memory)
printk("put_dirty_page: trying to put page %08lx at %08lx\n",page,address);
if (mem_map[MAP_NR(page)].count != 1)
printk("mem_map disagrees with %08lx at %08lx\n",page,address);
pgd = pgd_offset(tsk->mm,address);
pmd = pmd_alloc(pgd, address);
if (!pmd) {
free_page(page);
oom(tsk);
return 0;
}
pte = pte_alloc(pmd, address);
if (!pte) {
free_page(page);
oom(tsk);
return 0;
}
if (!pte_none(*pte)) {
printk("put_dirty_page: page already exists\n");
free_page(page);
return 0;
}
flush_page_to_ram(page);
set_pte(pte, pte_mkwrite(pte_mkdirty(mk_pte(page, PAGE_COPY))));
/* no need for invalidate */
return page;
}
static int vmap_pte_range(pmd_t *pmd, unsigned long addr,
unsigned long end, pgprot_t prot, struct page **pages, int *nr)
{
pte_t *pte;
/*
* nr is a running index into the array which helps higher level
* callers keep track of where we're up to.
*/
pte = pte_alloc_kernel(pmd, addr);
if (!pte)
return -ENOMEM;
do {
struct page *page = pages[*nr];
if (WARN_ON(!pte_none(*pte)))
return -EBUSY;
if (WARN_ON(!page))
return -ENOMEM;
set_pte_at(&init_mm, addr, pte, mk_pte(page, prot));
(*nr)++;
} while (pte++, addr += PAGE_SIZE, addr != end);
return 0;
}
/*
* page not present ... go through shm_pages
*/
static unsigned long shm_nopage(struct vm_area_struct * shmd, unsigned long address, int no_share)
{
pte_t pte;
struct shmid_kernel *shp;
unsigned int id, idx;
id = SWP_OFFSET(shmd->vm_pte) & SHM_ID_MASK;
idx = (address - shmd->vm_start + shmd->vm_offset) >> PAGE_SHIFT;
#ifdef DEBUG_SHM
if (id > max_shmid) {
printk ("shm_nopage: id=%d too big. proc mem corrupted\n", id);
return 0;
}
#endif
shp = shm_segs[id];
#ifdef DEBUG_SHM
if (shp == IPC_UNUSED || shp == IPC_NOID) {
printk ("shm_nopage: id=%d invalid. Race.\n", id);
return 0;
}
#endif
/* This can occur on a remap */
if (idx >= shp->shm_npages) {
return 0;
}
pte = __pte(shp->shm_pages[idx]);
if (!pte_present(pte)) {
unsigned long page = get_free_page(GFP_USER);
if (!page)
return -1;
pte = __pte(shp->shm_pages[idx]);
if (pte_present(pte)) {
free_page (page); /* doesn't sleep */
goto done;
}
if (!pte_none(pte)) {
rw_swap_page_nocache(READ, pte_val(pte), (char *)page);
pte = __pte(shp->shm_pages[idx]);
if (pte_present(pte)) {
free_page (page); /* doesn't sleep */
goto done;
}
swap_free(pte_val(pte));
shm_swp--;
}
shm_rss++;
pte = pte_mkdirty(mk_pte(page, PAGE_SHARED));
shp->shm_pages[idx] = pte_val(pte);
} else
--current->maj_flt; /* was incremented in do_no_page */
done: /* pte_val(pte) == shp->shm_pages[idx] */
current->min_flt++;
atomic_inc(&mem_map[MAP_NR(pte_page(pte))].count);
return pte_page(pte);
}
/*
* We hold the mm semaphore for reading and vma->vm_mm->page_table_lock
*/
static inline void break_cow(struct vm_area_struct * vma, struct page * new_page, unsigned long address,
pte_t *page_table)
{
flush_page_to_ram(new_page);
flush_cache_page(vma, address);
establish_pte(vma, address, page_table, pte_mkwrite(pte_mkdirty(mk_pte(new_page, vma->vm_page_prot))));
}
/*
* Trying to stop swapping from a file is fraught with races, so
* we repeat quite a bit here when we have to pause. swapoff()
* isn't exactly timing-critical, so who cares (but this is /really/
* inefficient, ugh).
*
* We return 1 after having slept, which makes the process start over
* from the beginning for this process..
*/
static inline int unuse_pte(struct vm_area_struct * vma, unsigned long address,
pte_t *dir, unsigned int type, unsigned long page)
{
pte_t pte = *dir;
if (pte_none(pte))
return 0;
if (pte_present(pte)) {
unsigned long page_nr = MAP_NR(pte_page(pte));
if (page_nr >= MAP_NR(high_memory))
return 0;
if (!in_swap_cache(page_nr))
return 0;
if (SWP_TYPE(in_swap_cache(page_nr)) != type)
return 0;
delete_from_swap_cache(page_nr);
set_pte(dir, pte_mkdirty(pte));
return 0;
}
if (SWP_TYPE(pte_val(pte)) != type)
return 0;
read_swap_page(pte_val(pte), (char *) page);
#if 0 /* Is this really needed here, hasn't it been solved elsewhere? */
flush_page_to_ram(page);
#endif
if (pte_val(*dir) != pte_val(pte)) {
free_page(page);
return 1;
}
set_pte(dir, pte_mkwrite(pte_mkdirty(mk_pte(page, vma->vm_page_prot))));
flush_tlb_page(vma, address);
++vma->vm_mm->rss;
swap_free(pte_val(pte));
return 1;
}
/*
* Map the 32bit vsyscall page on demand.
*
* RED-PEN: This knows too much about high level VM.
*
* Alternative would be to generate a vma with appropriate backing options
* and let it be handled by generic VM.
*/
int __map_syscall32(struct mm_struct *mm, unsigned long address)
{
pgd_t *pgd;
pud_t *pud;
pte_t *pte;
pmd_t *pmd;
int err = -ENOMEM;
spin_lock(&mm->page_table_lock);
pgd = pgd_offset(mm, address);
pud = pud_alloc(mm, pgd, address);
if (pud) {
pmd = pmd_alloc(mm, pud, address);
if (pmd && (pte = pte_alloc_map(mm, pmd, address)) != NULL) {
if (pte_none(*pte)) {
set_pte(pte,
mk_pte(virt_to_page(syscall32_page),
PAGE_KERNEL_VSYSCALL32));
}
/* Flush only the local CPU. Other CPUs taking a fault
will just end up here again
This probably not needed and just paranoia. */
__flush_tlb_one(address);
err = 0;
}
}
spin_unlock(&mm->page_table_lock);
return err;
}
/**
* 建立临时内核映射
* type和CPU共同确定用哪个固定映射的线性地址映射请求页。
*/
void *kmap_atomic(struct page *page, enum km_type type)
{
enum fixed_addresses idx;
unsigned long vaddr;
/* even !CONFIG_PREEMPT needs this, for in_atomic in do_page_fault */
inc_preempt_count();
/**
* 如果被映射的页不属于高端内存，当然用不着映射。直接返回线性地址就行了。
*/
if (!PageHighMem(page))
return page_address(page);
/**
* 通过type和CPU确定线性地址。
*/
idx = type + KM_TYPE_NR*smp_processor_id();
vaddr = __fix_to_virt(FIX_KMAP_BEGIN + idx);
#ifdef CONFIG_DEBUG_HIGHMEM
if (!pte_none(*(kmap_pte-idx)))
BUG();
#endif
/**
* 将线性地址与页表项建立映射。
*/
set_pte(kmap_pte-idx, mk_pte(page, kmap_prot));
/**
* 当然，最后必须刷新一下TLB。然后才能返回线性地址。
*/
__flush_tlb_one(vaddr);
return (void*) vaddr;
}
int zeromap_page_range(unsigned long address, unsigned long size, pgprot_t prot)
{
int error = 0;
pgd_t * dir;
unsigned long beg = address;
unsigned long end = address + size;
pte_t zero_pte;
zero_pte = pte_wrprotect(mk_pte(ZERO_PAGE, prot));
dir = pgd_offset(current->mm, address);
flush_cache_range(current->mm, beg, end);
while (address < end) {
pmd_t *pmd = pmd_alloc(dir, address);
error = -ENOMEM;
if (!pmd)
break;
error = zeromap_pmd_range(pmd, address, end - address, zero_pte);
if (error)
break;
address = (address + PGDIR_SIZE) & PGDIR_MASK;
#ifdef CONFIG_BESTA
if (!address) break; /* unsigned overflow */
#endif
dir++;
}
flush_tlb_range(current->mm, beg, end);
return error;
}
void kunmap_high_l1_vipt(struct page *page, pte_t saved_pte)
{
unsigned int idx, cpu = smp_processor_id();
int *depth = &per_cpu(kmap_high_l1_vipt_depth, cpu);
unsigned long vaddr, flags;
pte_t pte, *ptep;
idx = KM_L1_CACHE + KM_TYPE_NR * cpu;
vaddr = __fix_to_virt(FIX_KMAP_BEGIN + idx);
ptep = TOP_PTE(vaddr);
pte = mk_pte(page, kmap_prot);
BUG_ON(pte_val(*ptep) != pte_val(pte));
BUG_ON(*depth <= 0);
raw_local_irq_save(flags);
(*depth)--;
if (*depth != 0 && pte_val(pte) != pte_val(saved_pte)) {
set_pte_ext(ptep, saved_pte, 0);
local_flush_tlb_kernel_page(vaddr);
}
raw_local_irq_restore(flags);
if (!in_interrupt())
preempt_enable();
}
void *kmap_atomic(struct page *page, enum km_type type)
{
unsigned long idx;
unsigned long vaddr;
/* even !CONFIG_PREEMPT needs this, for in_atomic in do_page_fault */
pagefault_disable();
if (!PageHighMem(page))
return page_address(page);
debug_kmap_atomic(type);
idx = type + KM_TYPE_NR*smp_processor_id();
vaddr = __fix_to_virt(FIX_KMAP_BEGIN + idx);
/* XXX Fix - Anton */
#if 0
__flush_cache_one(vaddr);
#else
flush_cache_all();
#endif
#ifdef CONFIG_DEBUG_HIGHMEM
BUG_ON(!pte_none(*(kmap_pte-idx)));
#endif
set_pte(kmap_pte-idx, mk_pte(page, kmap_prot));
/* XXX Fix - Anton */
#if 0
__flush_tlb_one(vaddr);
#else
flush_tlb_all();
#endif
return (void*) vaddr;
}
static int init_stub_pte(struct mm_struct *mm, unsigned long proc,
unsigned long kernel)
{
pgd_t *pgd;
pud_t *pud;
pmd_t *pmd;
pte_t *pte;
pgd = pgd_offset(mm, proc);
pud = pud_alloc(mm, pgd, proc);
if (!pud)
goto out;
pmd = pmd_alloc(mm, pud, proc);
if (!pmd)
goto out_pmd;
pte = pte_alloc_map(mm, NULL, pmd, proc);
if (!pte)
goto out_pte;
*pte = mk_pte(virt_to_page(kernel), __pgprot(_PAGE_PRESENT));
*pte = pte_mkread(*pte);
return 0;
out_pte:
pmd_free(mm, pmd);
out_pmd:
pud_free(mm, pud);
out:
return -ENOMEM;
}
void *kmap_atomic(struct page *page, enum km_type type)
{
unsigned int idx;
unsigned long vaddr;
void *kmap;
pagefault_disable();
if (!PageHighMem(page))
return page_address(page);
debug_kmap_atomic(type);
kmap = kmap_high_get(page);
if (kmap)
return kmap;
idx = type + KM_TYPE_NR * smp_processor_id();
vaddr = __fix_to_virt(FIX_KMAP_BEGIN + idx);
#ifdef CONFIG_DEBUG_HIGHMEM
/*
* With debugging enabled, kunmap_atomic forces that entry to 0.
* Make sure it was indeed properly unmapped.
*/
BUG_ON(!pte_none(*(TOP_PTE(vaddr))));
#endif
set_pte_ext(TOP_PTE(vaddr), mk_pte(page, kmap_prot), 0);
/*
* When debugging is off, kunmap_atomic leaves the previous mapping
* in place, so this TLB flush ensures the TLB is updated with the
* new mapping.
*/
local_flush_tlb_kernel_page(vaddr);
return (void *)vaddr;
}
/*H:501
* We do need the Switcher code mapped at all times, so we allocate that
* part of the Guest page table here. We map the Switcher code immediately,
* but defer mapping of the guest register page and IDT/LDT etc page until
* just before we run the guest in map_switcher_in_guest().
*
* We *could* do this setup in map_switcher_in_guest(), but at that point
* we've interrupts disabled, and allocating pages like that is fraught: we
* can't sleep if we need to free up some memory.
*/
static bool allocate_switcher_mapping(struct lg_cpu *cpu)
{
int i;
for (i = 0; i < TOTAL_SWITCHER_PAGES; i++) {
pte_t *pte = find_spte(cpu, switcher_addr + i * PAGE_SIZE, true,
CHECK_GPGD_MASK, _PAGE_TABLE);
if (!pte)
return false;
/*
* Map the switcher page if not already there. It might
* already be there because we call allocate_switcher_mapping()
* in guest_set_pgd() just in case it did discard our Switcher
* mapping, but it probably didn't.
*/
if (i == 0 && !(pte_flags(*pte) & _PAGE_PRESENT)) {
/* Get a reference to the Switcher page. */
get_page(lg_switcher_pages[0]);
/* Create a read-only, exectuable, kernel-style PTE */
set_pte(pte,
mk_pte(lg_switcher_pages[0], PAGE_KERNEL_RX));
}
}
cpu->lg->pgdirs[cpu->cpu_pgd].switcher_mapped = true;
return true;
}
void *kmap_high_l1_vipt(struct page *page, pte_t *saved_pte)
{
unsigned int idx, cpu = smp_processor_id();
int *depth = &per_cpu(kmap_high_l1_vipt_depth, cpu);
unsigned long vaddr, flags;
pte_t pte, *ptep;
idx = KM_L1_CACHE + KM_TYPE_NR * cpu;
vaddr = __fix_to_virt(FIX_KMAP_BEGIN + idx);
ptep = TOP_PTE(vaddr);
pte = mk_pte(page, kmap_prot);
if (!in_interrupt())
preempt_disable();
raw_local_irq_save(flags);
(*depth)++;
if (pte_val(*ptep) == pte_val(pte)) {
*saved_pte = pte;
} else {
*saved_pte = *ptep;
set_pte_ext(ptep, pte, 0);
local_flush_tlb_kernel_page(vaddr);
}
raw_local_irq_restore(flags);
return (void *)vaddr;
}
/*
* maps a range of vmalloc()ed memory into the requested pages. the old
* mappings are removed.
*/
static inline void
vmap_pte_range (pte_t *pte, unsigned long address, unsigned long size, unsigned long vaddr)
{
unsigned long end;
pgd_t *vdir;
pmd_t *vpmd;
pte_t *vpte;
address &= ~PMD_MASK;
end = address + size;
if (end > PMD_SIZE)
end = PMD_SIZE;
do {
pte_t oldpage = *pte;
struct page * page;
pte_clear(pte);
vdir = pgd_offset_k (vaddr);
vpmd = pmd_offset (vdir, vaddr);
vpte = pte_offset (vpmd, vaddr);
page = pte_page (*vpte);
set_pte(pte, mk_pte(page, PAGE_USERIO));
forget_pte(oldpage);
address += PAGE_SIZE;
vaddr += PAGE_SIZE;
pte++;
} while (address < end);
}
/*H:520
* Setting up the Switcher PTE page for given CPU is fairly easy, given
* the CPU number and the "struct page"s for the Switcher code itself.
*
* Currently the Switcher is less than a page long, so "pages" is always 1.
*/
static __init void populate_switcher_pte_page(unsigned int cpu,
struct page *switcher_page[],
unsigned int pages)
{
unsigned int i;
pte_t *pte = switcher_pte_page(cpu);
/* The first entries are easy: they map the Switcher code. */
for (i = 0; i < pages; i++) {
set_pte(&pte[i], mk_pte(switcher_page[i],
__pgprot(_PAGE_PRESENT|_PAGE_ACCESSED)));
}
/* The only other thing we map is this CPU's pair of pages. */
i = pages + cpu*2;
/* First page (Guest registers) is writable from the Guest */
set_pte(&pte[i], pfn_pte(page_to_pfn(switcher_page[i]),
__pgprot(_PAGE_PRESENT|_PAGE_ACCESSED|_PAGE_RW)));
/*
* The second page contains the "struct lguest_ro_state", and is
* read-only.
*/
set_pte(&pte[i+1], pfn_pte(page_to_pfn(switcher_page[i+1]),
__pgprot(_PAGE_PRESENT|_PAGE_ACCESSED)));
}
/*
* No need to decide whether this PTE shares the swap entry with others,
* just let do_wp_page work it out if a write is requested later - to
* force COW, vm_page_prot omits write permission from any private vma.
*/
static int unuse_pte(struct vm_area_struct *vma, pmd_t *pmd,
unsigned long addr, swp_entry_t entry, struct page *page)
{
spinlock_t *ptl;
pte_t *pte;
int ret = 1;
if (mem_cgroup_charge(page, vma->vm_mm, GFP_KERNEL))
ret = -ENOMEM;
pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl);
if (unlikely(!pte_same(*pte, swp_entry_to_pte(entry)))) {
if (ret > 0)
mem_cgroup_uncharge_page(page);
ret = 0;
goto out;
}
inc_mm_counter(vma->vm_mm, anon_rss);
get_page(page);
set_pte_at(vma->vm_mm, addr, pte,
pte_mkold(mk_pte(page, vma->vm_page_prot)));
page_add_anon_rmap(page, vma, addr);
swap_free(entry);
/*
* Move the page to the active list so it is not
* immediately swapped out again after swapon.
*/
activate_page(page);
out:
pte_unmap_unlock(pte, ptl);
return ret;
}
void *kmap_atomic(struct page *page)
{
unsigned long vaddr;
long idx, type;
preempt_disable();
pagefault_disable();
if (!PageHighMem(page))
return page_address(page);
type = kmap_atomic_idx_push();
idx = type + KM_TYPE_NR*smp_processor_id();
vaddr = __fix_to_virt(FIX_KMAP_BEGIN + idx);
/* XXX Fix - Anton */
#if 0
__flush_cache_one(vaddr);
#else
flush_cache_all();
#endif
#ifdef CONFIG_DEBUG_HIGHMEM
BUG_ON(!pte_none(*(kmap_pte-idx)));
#endif
set_pte(kmap_pte-idx, mk_pte(page, kmap_prot));
/* XXX Fix - Anton */
#if 0
__flush_tlb_one(vaddr);
#else
flush_tlb_all();
#endif
return (void*) vaddr;
}
static int
__change_page_attr(struct page *page, pgprot_t prot)
{
pte_t *kpte;
unsigned long address;
struct page *kpte_page;
BUG_ON(PageHighMem(page));
address = (unsigned long)page_address(page);
kpte = lookup_address(address);
if (!kpte)
return -EINVAL;
kpte_page = virt_to_page(kpte);
if (pgprot_val(prot) != pgprot_val(PAGE_KERNEL)) {
if ((pte_val(*kpte) & _PAGE_PSE) == 0) {
set_pte_atomic(kpte, mk_pte(page, prot));
} else {
struct page *split = split_large_page(address, prot);
if (!split)
return -ENOMEM;
set_pmd_pte(kpte,address,mk_pte(split, PAGE_KERNEL));
kpte_page = split;
}
get_page(kpte_page);
} else if ((pte_val(*kpte) & _PAGE_PSE) == 0) {
set_pte_atomic(kpte, mk_pte(page, PAGE_KERNEL));
__put_page(kpte_page);
} else
BUG();
/*
* If the pte was reserved, it means it was created at boot
* time (not via split_large_page) and in turn we must not
* replace it with a largepage.
*/
if (!PageReserved(kpte_page)) {
/* memleak and potential failed 2M page regeneration */
BUG_ON(!page_count(kpte_page));
if (cpu_has_pse && (page_count(kpte_page) == 1)) {
list_add(&kpte_page->lru, &df_list);
revert_page(kpte_page, address);
}
}
return 0;
}
/* this routine handles present pages, when users try to write
to a shared page.
*/
void do_wp_page(struct vm_area_struct *vma, unsigned long address, int write_access)
{
pgd_t *pgd;
pmd_t *pmd;
pte_t *page_table,pte;
unsigned long old_page, new_page;
new_page = get_free_page(GFP_KERNEL);
pgd = pgd_offset(vma->vm_task, address);
if(pgd_none(*pgd))
goto end_wp_page;
if(pgd_bad(*pgd))
goto bad_wp_page;
pmd = pmd_offset(pgd,address);
if(pmd_none(*pmd))
goto end_wp_page;
if(pmd_bad(*pmd))
goto bad_wp_page;
page_table = pte_offset(pmd,address);
pte = *page_table;
if(!pte_present(pte))
goto end_wp_page;
if(pte_write(pte))
goto end_wp_page;
old_page = pte_page(pte);
if(old_page >= main_memory_end)
goto bad_wp_page;
(vma->vm_task->mm->min_flt)++;
if(mem_map[MAP_NR(old_page)].flags & PAGE_PRESENT)
{
if(new_page)
{
if(mem_map[MAP_NR(old_page)].flags & MAP_PAGE_RESERVED)
++(vma->vm_task->mm->rss);
copy_page(old_page, new_page);
*page_table = pte_mkwrite(pte_mkdirty(mk_pte((unsigned long)&new_page, vma->vm_page_prot)));
free_page(old_page);
return;
}
pte_val(*page_table) &= PAGE_BAD;
free_page(old_page);
oom();
return;
}
*page_table = pte_mkdirty(pte_mkwrite(pte));
if(new_page)
free_page(new_page);
return;
bad_wp_page:
printk("do_wp_page: bogus page at address %08lx (%08lx)\n",address,old_page);
goto end_wp_page;
end_wp_page:
if(new_page)
free_page(new_page);
return;
}
pte_t __bad_page(void)
{
extern char empty_bad_page[PAGE_SIZE];
unsigned long page = (unsigned long)empty_bad_page;
clear_page(page);
return pte_mkdirty(mk_pte(page, PAGE_SHARED));
}
/*
* This only needs the MM semaphore
*/
static int do_anonymous_page(struct mm_struct * mm, struct vm_area_struct * vma, pte_t *page_table, int write_access, unsigned long addr)
{
struct page *page = NULL;
pte_t entry = pte_wrprotect(mk_pte(ZERO_PAGE(addr), vma->vm_page_prot));
if (write_access) {
page = alloc_page(GFP_HIGHUSER);
if (!page)
return -1;
clear_user_highpage(page, addr);
entry = pte_mkwrite(pte_mkdirty(mk_pte(page, vma->vm_page_prot)));
mm->rss++;
flush_page_to_ram(page);
}
set_pte(page_table, entry);
/* No need to invalidate - it was non-present before */
update_mmu_cache(vma, addr, entry);
return 1; /* Minor fault */
}
static int __dma_update_pte(pte_t *pte, pgtable_t token, unsigned long addr,
void *data)
{
struct page *page = virt_to_page(addr);
pgprot_t prot = *(pgprot_t *)data;
set_pte(pte, mk_pte(page, prot));
return 0;
}
static void *
__dma_alloc_remap(struct page *page, size_t size, gfp_t gfp, pgprot_t prot,
const void *caller)
{
struct arm_vmregion *c;
size_t align;
int bit;
if (!consistent_pte) {
printk(KERN_ERR "%s: not initialised\n", __func__);
dump_stack();
return NULL;
}
/*
* Align the virtual region allocation - maximum alignment is
* a section size, minimum is a page size. This helps reduce
* fragmentation of the DMA space, and also prevents allocations
* smaller than a section from crossing a section boundary.
*/
bit = fls(size - 1);
if (bit > SECTION_SHIFT)
bit = SECTION_SHIFT;
align = 1 << bit;
/*
* Allocate a virtual address in the consistent mapping region.
*/
c = arm_vmregion_alloc(&consistent_head, align, size,
gfp & ~(__GFP_DMA | __GFP_HIGHMEM), caller);
if (c) {
pte_t *pte;
int idx = CONSISTENT_PTE_INDEX(c->vm_start);
u32 off = CONSISTENT_OFFSET(c->vm_start) & (PTRS_PER_PTE-1);
pte = consistent_pte[idx] + off;
c->vm_pages = page;
do {
BUG_ON(!pte_none(*pte));
set_pte_ext(pte, mk_pte(page, prot), 0);
page++;
pte++;
off++;
if (off >= PTRS_PER_PTE) {
off = 0;
pte = consistent_pte[++idx];
}
} while (size -= PAGE_SIZE);
dsb();
return (void *)c->vm_start;
}
return NULL;
}
/**
* __replace_page - replace page in vma by new page.
* based on replace_page in mm/ksm.c
*
* @vma: vma that holds the pte pointing to page
* @addr: address the old @page is mapped at
* @page: the cowed page we are replacing by kpage
* @kpage: the modified page we replace page by
*
* Returns 0 on success, -EFAULT on failure.
*/
static int __replace_page(struct vm_area_struct *vma, unsigned long addr,
struct page *old_page, struct page *new_page)
{
struct mm_struct *mm = vma->vm_mm;
spinlock_t *ptl;
pte_t *ptep;
int err;
/* For mmu_notifiers */
const unsigned long mmun_start = addr;
const unsigned long mmun_end = addr + PAGE_SIZE;
struct mem_cgroup *memcg;
err = mem_cgroup_try_charge(new_page, vma->vm_mm, GFP_KERNEL, &memcg,
false);
if (err)
return err;
/* For try_to_free_swap() and munlock_vma_page() below */
lock_page(old_page);
mmu_notifier_invalidate_range_start(mm, mmun_start, mmun_end);
err = -EAGAIN;
ptep = page_check_address(old_page, mm, addr, &ptl, 0);
if (!ptep) {
mem_cgroup_cancel_charge(new_page, memcg, false);
goto unlock;
}
get_page(new_page);
page_add_new_anon_rmap(new_page, vma, addr, false);
mem_cgroup_commit_charge(new_page, memcg, false, false);
lru_cache_add_active_or_unevictable(new_page, vma);
if (!PageAnon(old_page)) {
dec_mm_counter(mm, mm_counter_file(old_page));
inc_mm_counter(mm, MM_ANONPAGES);
}
flush_cache_page(vma, addr, pte_pfn(*ptep));
ptep_clear_flush_notify(vma, addr, ptep);
set_pte_at_notify(mm, addr, ptep, mk_pte(new_page, vma->vm_page_prot));
page_remove_rmap(old_page, false);
if (!page_mapped(old_page))
try_to_free_swap(old_page);
pte_unmap_unlock(ptep, ptl);
if (vma->vm_flags & VM_LOCKED)
munlock_vma_page(old_page);
put_page(old_page);
err = 0;
unlock:
mmu_notifier_invalidate_range_end(mm, mmun_start, mmun_end);
unlock_page(old_page);
return err;
}
/*
* We hold the mm semaphore for reading and vma->vm_mm->page_table_lock
*/
static inline void break_cow(struct vm_area_struct * vma, struct page * new_page, unsigned long address,
pte_t *page_table)
{
flush_page_to_ram(new_page);
#ifndef CONFIG_SUPERH
/* Not needed for VIPT cache (need better API for caches) */
flush_cache_page(vma, address);
#endif
establish_pte(vma, address, page_table, pte_mkwrite(pte_mkdirty(mk_pte(new_page, vma->vm_page_prot))));
}
/*
* We are called with the MM semaphore and page_table_lock
* spinlock held to protect against concurrent faults in
* multithreaded programs.
*/
static int do_anonymous_page(struct mm_struct * mm, struct vm_area_struct * vma, pte_t *page_table, int write_access, unsigned long addr)
{
pte_t entry;
/* Read-only mapping of ZERO_PAGE. */
entry = pte_wrprotect(mk_pte(ZERO_PAGE(addr), vma->vm_page_prot));
/* ..except if it's a write access */
if (write_access) {
struct page *page;
/* Allocate our own private page. */
spin_unlock(&mm->page_table_lock);
page = alloc_page(GFP_HIGHUSER);
if (!page)
goto no_mem;
clear_user_highpage(page, addr);
spin_lock(&mm->page_table_lock);
if (!pte_none(*page_table)) {
page_cache_release(page);
spin_unlock(&mm->page_table_lock);
return 1;
}
mm->rss++;
flush_page_to_ram(page);
entry = pte_mkwrite(pte_mkdirty(mk_pte(page, vma->vm_page_prot)));
lru_cache_add(page);
mark_page_accessed(page);
}
set_pte(page_table, entry);
/* No need to invalidate - it was non-present before */
update_mmu_cache(vma, addr, entry);
spin_unlock(&mm->page_table_lock);
return 1; /* Minor fault */
no_mem:
return -1;
}
pte_t __bad_page(void)
{
extern char empty_bad_page[PAGE_SIZE];
__asm__ __volatile__("cld ; rep ; stosl":
:"a" (0),
"D" ((long) empty_bad_page),
"c" (PAGE_SIZE/4)
:"di","cx");
return pte_mkdirty(mk_pte((unsigned long) empty_bad_page, PAGE_SHARED));
}
