C++ pte_unmap函数代码示例

static int unuse_pte_range(struct vm_area_struct *vma, pmd_t *pmd,
				unsigned long addr, unsigned long end,
				swp_entry_t entry, struct page *page)
{
	pte_t swp_pte = swp_entry_to_pte(entry);
	pte_t *pte;
	int ret = 0;

	/*
	 * We don't actually need pte lock while scanning for swp_pte: since
	 * we hold page lock and mmap_sem, swp_pte cannot be inserted into the
	 * page table while we're scanning; though it could get zapped, and on
	 * some architectures (e.g. x86_32 with PAE) we might catch a glimpse
	 * of unmatched parts which look like swp_pte, so unuse_pte must
	 * recheck under pte lock.  Scanning without pte lock lets it be
	 * preemptible whenever CONFIG_PREEMPT but not CONFIG_HIGHPTE.
	 */
	pte = pte_offset_map(pmd, addr);
	do {
		/*
		 * swapoff spends a _lot_ of time in this loop!
		 * Test inline before going to call unuse_pte.
		 */
		if (unlikely(pte_same(*pte, swp_pte))) {
			pte_unmap(pte);
			ret = unuse_pte(vma, pmd, addr, entry, page);
			if (ret)
				goto out;
			pte = pte_offset_map(pmd, addr);
		}
	} while (pte++, addr += PAGE_SIZE, addr != end);
	pte_unmap(pte - 1);
out:
	return ret;
}

/*
 * need to get a 16k page for level 1
 */
pgd_t *pgd_alloc(struct mm_struct *mm)
{
	pgd_t *new_pgd, *init_pgd;
	pud_t *new_pud, *init_pud;
	pmd_t *new_pmd, *init_pmd;
	pte_t *new_pte, *init_pte;

	new_pgd = (pgd_t *)__get_free_pages(GFP_KERNEL, 2);
	if (!new_pgd)
		goto no_pgd;

	memset(new_pgd, 0, USER_PTRS_PER_PGD * sizeof(pgd_t));

	/*
	 * Copy over the kernel and IO PGD entries
	 */
	init_pgd = pgd_offset_k(0);
	memcpy(new_pgd + USER_PTRS_PER_PGD, init_pgd + USER_PTRS_PER_PGD,
		       (PTRS_PER_PGD - USER_PTRS_PER_PGD) * sizeof(pgd_t));

	clean_dcache_area(new_pgd, PTRS_PER_PGD * sizeof(pgd_t));

	if (!vectors_high()) {
		/*
		 * On ARM, first page must always be allocated since it
		 * contains the machine vectors.
		 */
		new_pud = pud_alloc(mm, new_pgd, 0);
		if (!new_pud)
			goto no_pud;

		new_pmd = pmd_alloc(mm, new_pud, 0);
		if (!new_pmd)
			goto no_pmd;

		new_pte = pte_alloc_map(mm, NULL, new_pmd, 0);
		if (!new_pte)
			goto no_pte;

		init_pud = pud_offset(init_pgd, 0);
		init_pmd = pmd_offset(init_pud, 0);
		init_pte = pte_offset_map(init_pmd, 0);
		set_pte_ext(new_pte, *init_pte, 0);
		pte_unmap(init_pte);
		pte_unmap(new_pte);
	}

	return new_pgd;

no_pte:
	pmd_free(mm, new_pmd);
no_pmd:
	pud_free(mm, new_pud);
no_pud:
	free_pages((unsigned long)new_pgd, 2);
no_pgd:
	return NULL;
}

/**
 * Replace the PFN of a PTE with the address of the actual page.
 *
 * The caller maps a reserved dummy page at the address with the desired access
 * and flags.
 *
 * This hack is required for older Linux kernels which don't provide
 * remap_pfn_range().
 *
 * @returns 0 on success, -ENOMEM on failure.
 * @param   mm          The memory context.
 * @param   ulAddr      The mapping address.
 * @param   Phys        The physical address of the page to map.
 */
static int rtR0MemObjLinuxFixPte(struct mm_struct *mm, unsigned long ulAddr, RTHCPHYS Phys)
{
    int rc = -ENOMEM;
    pgd_t *pgd;

    spin_lock(&mm->page_table_lock);

    pgd = pgd_offset(mm, ulAddr);
    if (!pgd_none(*pgd) && !pgd_bad(*pgd))
    {
        pmd_t *pmd = pmd_offset(pgd, ulAddr);
        if (!pmd_none(*pmd))
        {
            pte_t *ptep = pte_offset_map(pmd, ulAddr);
            if (ptep)
            {
                pte_t pte = *ptep;
                pte.pte_high &= 0xfff00000;
                pte.pte_high |= ((Phys >> 32) & 0x000fffff);
                pte.pte_low  &= 0x00000fff;
                pte.pte_low  |= (Phys & 0xfffff000);
                set_pte(ptep, pte);
                pte_unmap(ptep);
                rc = 0;
            }
        }

/* Ensure all existing pages follow the policy. */
static int
verify_pages(unsigned long addr, unsigned long end, unsigned long *nodes)
{
	while (addr < end) {
		struct page *p;
		pte_t *pte;
		pmd_t *pmd;
		pgd_t *pgd = pgd_offset_k(addr);
		if (pgd_none(*pgd)) {
			addr = (addr + PGDIR_SIZE) & PGDIR_MASK;
			continue;
		}
		pmd = pmd_offset(pgd, addr);
		if (pmd_none(*pmd)) {
			addr = (addr + PMD_SIZE) & PMD_MASK;
			continue;
		}
		p = NULL;
		pte = pte_offset_map(pmd, addr);
		if (pte_present(*pte))
			p = pte_page(*pte);
		pte_unmap(pte);
		if (p) {
			unsigned nid = page_to_nid(p);
			if (!test_bit(nid, nodes))
				return -EIO;
		}
		addr += PAGE_SIZE;
	}
	return 0;
}

int kthread_wss(void *data)
{
	unsigned long va;
	int ret;
	int wss;

	pgd_t *pgd;
	pmd_t *pmd;
	pud_t *pud;
	pte_t *ptep;
	
	struct task_struct *task;
	while(!kthread_should_stop())
	{
		printk(KERN_INFO "Checking process' WSS.\n");
		for_each_process(task)
		{
			wss = 0;
			if(task->mm != NULL)
			{
				struct vm_area_struct *temp = task->mm->mmap;
				while(temp)
				{
					if(temp->vm_flags & VM_IO){}
					else
					{
						for(va = temp->vm_start; va < temp->vm_end; va+=PAGE_SIZE)
						{
				  			pgd = pgd_offset(task->mm,va);
			 		  		if(pgd_none(*pgd))
								break;
							pud = pud_offset(pgd,va);
							if(pud_none(*pud))
								break;
							pmd = pmd_offset(pud,va);
							if(pmd_none(*pmd))
								break;
							ptep = pte_offset_map(pmd,va);
							ret = 0;
							if(pte_young(*ptep))
							{
								ret = test_and_clear_bit(_PAGE_BIT_ACCESSED,												(unsigned long *) &ptep->pte);
								wss++;
							}
							if(ret)
							{
								pte_update(task->mm, va, ptep);
							}
							pte_unmap(ptep);
						}
					}
					temp = temp->vm_next;
				}
				printk(KERN_INFO "%i: %i\n", task->pid, wss);
			}
		}
	msleep(1000);
	}
	return 0;
}

static int filemap_sync_pte_range(pmd_t * pmd,
	unsigned long address, unsigned long end, 
	struct vm_area_struct *vma, unsigned int flags)
{
	pte_t *pte;
	int error;

	if (pmd_none(*pmd))
		return 0;
	if (pmd_bad(*pmd)) {
		pmd_ERROR(*pmd);
		pmd_clear(pmd);
		return 0;
	}
	pte = pte_offset_map(pmd, address);
	if ((address & PMD_MASK) != (end & PMD_MASK))
		end = (address & PMD_MASK) + PMD_SIZE;
	error = 0;
	do {
		error |= filemap_sync_pte(pte, vma, address, flags);
		address += PAGE_SIZE;
		pte++;
	} while (address && (address < end));

	pte_unmap(pte - 1);

	return error;
}

/*
 * Dump out the page tables associated with 'addr' in mm 'mm'.
 */
void show_pte(struct mm_struct *mm, unsigned long addr)
{
    pgd_t *pgd;

    if (!mm)
        mm = &init_mm;

    pr_alert("pgd = %p\n", mm->pgd);
    pgd = pgd_offset(mm, addr);
    pr_alert("[%08lx] *pgd=%016llx", addr, pgd_val(*pgd));

    do {
        pud_t *pud;
        pmd_t *pmd;
        pte_t *pte;

        if (pgd_none(*pgd) || pgd_bad(*pgd))
            break;

        pud = pud_offset(pgd, addr);
        if (pud_none(*pud) || pud_bad(*pud))
            break;

        pmd = pmd_offset(pud, addr);
        printk(", *pmd=%016llx", pmd_val(*pmd));
        if (pmd_none(*pmd) || pmd_bad(*pmd))
            break;

        pte = pte_offset_map(pmd, addr);
        printk(", *pte=%016llx", pte_val(*pte));
        pte_unmap(pte);
    } while(0);

    printk("\n");
}

/* Ensure all existing pages follow the policy. */
static int check_pte_range(struct mm_struct *mm, pmd_t *pmd,
		unsigned long addr, unsigned long end, unsigned long *nodes)
{
	pte_t *orig_pte;
	pte_t *pte;

	spin_lock(&mm->page_table_lock);
	orig_pte = pte = pte_offset_map(pmd, addr);
	do {
		unsigned long pfn;
		unsigned int nid;

		if (!pte_present(*pte))
			continue;
		pfn = pte_pfn(*pte);
		if (!pfn_valid(pfn))
			continue;
		nid = pfn_to_nid(pfn);
		if (!test_bit(nid, nodes))
			break;
	} while (pte++, addr += PAGE_SIZE, addr != end);
	pte_unmap(orig_pte);
	spin_unlock(&mm->page_table_lock);
	return addr != end;
}

/*
 * need to get a 16k page for level 1
 */
pgd_t *get_pgd_slow(struct mm_struct *mm)
{
	pgd_t *new_pgd, *init_pgd;
	pmd_t *new_pmd, *init_pmd;
	pte_t *new_pte, *init_pte;
	unsigned long flags;

	new_pgd = (pgd_t *)__get_free_pages(GFP_KERNEL, 2);
	if (!new_pgd)
		goto no_pgd;

	memset(new_pgd, 0, FIRST_KERNEL_PGD_NR * sizeof(pgd_t));

	/*
	 * Copy over the kernel and IO PGD entries
	 */
	init_pgd = pgd_offset_k(0);
	pgd_list_lock(flags);
	memcpy(new_pgd + FIRST_KERNEL_PGD_NR, init_pgd + FIRST_KERNEL_PGD_NR,
		       (PTRS_PER_PGD - FIRST_KERNEL_PGD_NR) * sizeof(pgd_t));
	pgd_list_add(new_pgd);
	pgd_list_unlock(flags);

	clean_dcache_area(new_pgd, PTRS_PER_PGD * sizeof(pgd_t));

	if (!vectors_high()) {
#ifdef CONFIG_ARM_FCSE
		/* FCSE does not work without high vectors. */
		BUG();
#endif /* CONFIG_ARM_FCSE */

		/*
		 * On ARM, first page must always be allocated since it
		 * contains the machine vectors.
		 */
		new_pmd = pmd_alloc(mm, new_pgd, 0);
		if (!new_pmd)
			goto no_pmd;

		new_pte = pte_alloc_map(mm, new_pmd, 0);
		if (!new_pte)
			goto no_pte;

		init_pmd = pmd_offset(init_pgd, 0);
		init_pte = pte_offset_map_nested(init_pmd, 0);
		set_pte_ext(new_pte, *init_pte, 0);
		pte_unmap_nested(init_pte);
		pte_unmap(new_pte);
	}

	return new_pgd;

no_pte:
	pmd_free(mm, new_pmd);
no_pmd:
	free_pages((unsigned long)new_pgd, 2);
no_pgd:
	return NULL;
}

/*
 * The performance critical leaf functions are made noinline otherwise gcc
 * inlines everything into a single function which results in too much
 * register pressure.
 */
static noinline int gup_pte_range(pmd_t pmd, unsigned long addr,
		unsigned long end, int write, struct page **pages, int *nr)
{
	u64 mask, result;
	pte_t *ptep;

#ifdef CONFIG_X2TLB
	result = _PAGE_PRESENT | _PAGE_EXT(_PAGE_EXT_KERN_READ | _PAGE_EXT_USER_READ);
	if (write)
		result |= _PAGE_EXT(_PAGE_EXT_KERN_WRITE | _PAGE_EXT_USER_WRITE);
#elif defined(CONFIG_SUPERH64)
	result = _PAGE_PRESENT | _PAGE_USER | _PAGE_READ;
	if (write)
		result |= _PAGE_WRITE;
#else
	result = _PAGE_PRESENT | _PAGE_USER;
	if (write)
		result |= _PAGE_RW;
#endif

	mask = result | _PAGE_SPECIAL;

	ptep = pte_offset_map(&pmd, addr);
	do {
		pte_t pte = gup_get_pte(ptep);
		struct page *page;

		if ((pte_val(pte) & mask) != result) {
			pte_unmap(ptep);
			return 0;
		}
		VM_BUG_ON(!pfn_valid(pte_pfn(pte)));
		page = pte_page(pte);
		get_page(page);
		__flush_anon_page(page, addr);
		flush_dcache_page(page);
		pages[*nr] = page;
		(*nr)++;

	} while (ptep++, addr += PAGE_SIZE, addr != end);
	pte_unmap(ptep - 1);

	return 1;
}

// Return true (!= 0) if any referenced bits are set.
static int ref_bits_set (int exclude_irqhandler) {
    void *cur_addr;
    pte_t *pte;
    int i;
    int ret_val = 0;

    for (i = 0; i < cr_num_drivers; i++) {
        if (exclude_irqhandler) uprintk ("i %d: ", i);
        for (cur_addr = cr_base_address[i];
             cur_addr < cr_base_address[i] + cr_module_size[i];
             cur_addr += PAGE_SIZE) {
            
            pte = virt_to_pte (cur_addr);
            if (pte != NULL) {
                // See if we're excluding the interrupt handler
                // from this check.
                if (exclude_irqhandler &&
                    addr_contains_irq_handler (cur_addr)) {
                    pte_unmap(pte);
                    if (exclude_irqhandler) uprintk ("X");
                    continue;
                }

                // See if the page was referenced lately.
                if (pte_young(*pte) != 0) {
                    // kunmap_atomic (page, KM_IRQ1);
                    pte_unmap(pte);
                    if (exclude_irqhandler) uprintk ("1");
                    ret_val = 1;
                    continue;
                }

                if (exclude_irqhandler) uprintk ("0");
                
                // kunmap_atomic (page, KM_IRQ1);
                pte_unmap(pte);
            }
        }

        if (exclude_irqhandler) uprintk ("\n");
    }

    return ret_val;
}

asmlinkage long sys_my_syscall( int pid, unsigned long address)
{
   
   struct task_struct* task;
   struct mm_struct* mm;
   pgd_t* pgd;
   pud_t* pud;
   pmd_t* pmd;
   pte_t* pte;
   unsigned long pte_val ;
   printk(KERN_INFO "PID: %d, VIRTUAL_ADDR: 0x%lx\n", pid, address);
   for_each_process(task)
   {
     if(task->pid == pid)
     {           
      printk(KERN_INFO "Task %d found\n", task->pid);
      mm = task->mm;
      
      pgd = pgd_offset(mm, address);
      printk(KERN_INFO "PGD INFO: PRESENT: %d, BAD: %d, NONE: %d\n", pgd_present(*pgd), pgd_bad(*pgd), pgd_none(*pgd));
      if(!(pgd_none(*pgd) || pgd_bad(*pgd)) && pgd_present(*pgd))
      {
        printk(KERN_INFO "PGD INFO: PRESENT: %d, BAD: %d, NONE: %d\n", pgd_present(*pgd), pgd_bad(*pgd), pgd_none(*pgd));
        pud = pud_offset(pgd, address);
        printk(KERN_INFO "PUD INFO: PRESENT: %d, BAD: %d, NONE: %d\n", pud_present(*pud), pud_bad(*pud), pud_none(*pud));
           
        if(!(pud_none(*pud) || pud_bad(*pud)) && pud_present(*pud))
        {
          printk(KERN_INFO "PUD INFO: PRESENT: %d, BAD: %d, NONE: %d\n", pud_present(*pud), pud_bad(*pud), pud_none(*pud));
          pmd = pmd_offset(pud, address);
          printk(KERN_INFO "PMD INFO: PRESENT: %d, BAD: %d, NONE: %d\n", pmd_present(*pmd), pmd_bad(*pmd), pmd_none(*pmd));
           
          if(!(pmd_none(*pmd) || pmd_bad(*pmd)) && pmd_present(*pmd))
          {
            printk(KERN_INFO "PMD INFO: PRESENT: %d, BAD: %d, NONE: %d\n", pmd_present(*pmd), pmd_bad(*pmd), pmd_none(*pmd));
            pte = pte_offset_map(pmd, address);
            printk(KERN_INFO "PTE INFO: PRESENT: %d PTE: 0x%lx \n ", pte_present(*pte), pte->pte);


            pte_val = pte->pte;
            
            if(pte_val == 0)
                pte_val = __pte_to_swp_entry(*pte).val;
            
            pte_unmap(pte);
            printk(KERN_INFO "pte_val: %lx\n" , pte_val);
            return pte_val;
            
            }
        }
      }
    }
  }
  printk(KERN_INFO "Data not found!\n");
  return 0;
}

/*
 * get_pgd_slow:申请一个pgd项
 * notice:一个pgd占用4个页框,每一个pgt项大小为8字节
 */
pgd_t *get_pgd_slow(struct mm_struct *mm)
{
	pgd_t *new_pgd, *init_pgd;
	pmd_t *new_pmd, *init_pmd;
	pte_t *new_pte, *init_pte;

	/*pgd占用四个页框*/
	new_pgd = (pgd_t *)__get_free_pages(GFP_KERNEL, 2);
	if (!new_pgd)
		goto no_pgd;

	memset(new_pgd, 0, FIRST_KERNEL_PGD_NR * sizeof(pgd_t));

	/*
	 * 复制内核与I/O PGD entries
	 */
	init_pgd = pgd_offset_k(0);
	memcpy(new_pgd + FIRST_KERNEL_PGD_NR, init_pgd + FIRST_KERNEL_PGD_NR,
		       (PTRS_PER_PGD - FIRST_KERNEL_PGD_NR) * sizeof(pgd_t));

	clean_dcache_area(new_pgd, PTRS_PER_PGD * sizeof(pgd_t));

	if (!vectors_high()) {
		/*
		 * On ARM, first page must always be allocated since it
		 * contains the machine vectors.
		 */
		new_pmd = pmd_alloc(mm, new_pgd, 0);
		if (!new_pmd)
			goto no_pmd;

		/*返回pmd的第0项页表项,因为第0项用于映射中断向量*/
		new_pte = pte_alloc_map(mm, new_pmd, 0);
		if (!new_pte)
			goto no_pte;

		/*返回中断向量的页表项,中断向量位于低地址空间时(0地址开始处)*/
		init_pmd = pmd_offset(init_pgd, 0);
		init_pte = pte_offset_map_nested(init_pmd, 0);
		/*给新的页表项映射中断向量的页表项*/
		set_pte_ext(new_pte, *init_pte, 0);
		pte_unmap_nested(init_pte);
		/*取消new_pte的高端内存的页表映射*/
		pte_unmap(new_pte);
	}

	return new_pgd;

no_pte:
	pmd_free(mm, new_pmd);
no_pmd:
	free_pages((unsigned long)new_pgd, 2);
no_pgd:
	return NULL;
}

unsigned int pmem_user_v2p_video(unsigned int va)
{
    unsigned int pageOffset = (va & (PAGE_SIZE - 1)); 
    pgd_t *pgd;
    pud_t *pud;
    pmd_t *pmd;
    pte_t *pte;
    unsigned int pa;

    if(NULL==current)
    {
    	  MFV_LOGE("[ERROR] pmem_user_v2p_video, current is NULL! \n");
    	  return 0;
    }
    
    if(NULL==current->mm)
    {
    	  MFV_LOGE("[ERROR] pmem_user_v2p_video, current->mm is NULL! tgid=0x%x, name=%s \n", current->tgid, current->comm);
    	  return 0;
    }
    
    pgd = pgd_offset(current->mm, va); /* what is tsk->mm */
    if(pgd_none(*pgd)||pgd_bad(*pgd))
    {    
        MFV_LOGE("[ERROR] pmem_user_v2p(), va=0x%x, pgd invalid! \n", va); 
        return 0;
    }    

    pud = pud_offset(pgd, va); 
    if(pud_none(*pud)||pud_bad(*pud))
    {    
        MFV_LOGE("[ERROR] pmem_user_v2p(), va=0x%x, pud invalid! \n", va); 
        return 0;
    }    
    
    pmd = pmd_offset(pud, va); 
    if(pmd_none(*pmd)||pmd_bad(*pmd))
    {    
        MFV_LOGE("[ERROR] pmem_user_v2p(), va=0x%x, pmd invalid! \n", va); 
        return 0;
    }    
      
    pte = pte_offset_map(pmd, va); 
    if(pte_present(*pte)) 
    {    
        pa=(pte_val(*pte) & (PAGE_MASK)) | pageOffset; 
        pte_unmap(pte);
        return pa;  
    }     

    MFV_LOGE("[ERROR] pmem_user_v2p(), va=0x%x, pte invalid! \n", va); 
    return 0;
}

static long kgsl_cache_range_op(unsigned long addr, int size,
					unsigned int flags)
{
#ifdef CONFIG_OUTER_CACHE
	unsigned long end;
#endif
	BUG_ON(addr & (KGSL_PAGESIZE - 1));
	BUG_ON(size & (KGSL_PAGESIZE - 1));

	if (flags & KGSL_CACHE_FLUSH)
		dmac_flush_range((const void *)addr,
				(const void *)(addr + size));
	else
		if (flags & KGSL_CACHE_CLEAN)
			dmac_clean_range((const void *)addr,
					(const void *)(addr + size));
		else
			dmac_inv_range((const void *)addr,
					(const void *)(addr + size));

#ifdef CONFIG_OUTER_CACHE
	for (end = addr; end < (addr + size); end += KGSL_PAGESIZE) {
		pte_t *pte_ptr, pte;
		unsigned long physaddr;
		if (flags & KGSL_CACHE_VMALLOC_ADDR)
			physaddr = vmalloc_to_pfn((void *)end);
		else
			if (flags & KGSL_CACHE_USER_ADDR) {
				pte_ptr = kgsl_get_pte_from_vaddr(end);
				if (!pte_ptr)
					return -EINVAL;
				pte = *pte_ptr;
				physaddr = pte_pfn(pte);
				pte_unmap(pte_ptr);
			} else
				return -EINVAL;

		physaddr <<= PAGE_SHIFT;
		if (flags & KGSL_CACHE_FLUSH)
			outer_flush_range(physaddr, physaddr + KGSL_PAGESIZE);
		else
			if (flags & KGSL_CACHE_CLEAN)
				outer_clean_range(physaddr,
					physaddr + KGSL_PAGESIZE);
			else
				outer_inv_range(physaddr,
					physaddr + KGSL_PAGESIZE);
	}
#endif
	return 0;
}