页表映射的疑惑

tc1989tc

/**
* remap_pfn_range - remap kernel memory to userspace
* @vma: user vma to map to
* @addr: target user address to start at
* @pfn: physical address of kernel memory
* @size: size of map area
* @prot: page protection flags for this mapping
*
*  Note: this is only safe if the mm semaphore is held when called.
*/
int remap_pfn_range(struct vm_area_struct *vma, unsigned long addr,
                    unsigned long pfn, unsigned long size, pgprot_t prot)
{
        pgd_t *pgd;
        unsigned long next;
        unsigned long end = addr + PAGE_ALIGN(size);
        struct mm_struct *mm = vma->vm_mm;
        int err;

        /*
         * Physically remapped pages are special. Tell the
         * rest of the world about it:
         *   VM_IO tells people not to look at these pages
         *        (accesses can have side effects).
         *   VM_RESERVED is specified all over the place, because
         *        in 2.4 it kept swapout's vma scan off this vma; but
         *        in 2.6 the LRU scan won't even find its pages, so this
         *        flag means no more than count its pages in reserved_vm,
         *         and omit it from core dump, even when VM_IO turned off.
         *   VM_PFNMAP tells the core MM that the base pages are just
         *        raw PFN mappings, and do not have a "struct page" associated
         *        with them.
         *
         * There's a horrible special case to handle copy-on-write
         * behaviour that some programs depend on. We mark the "original"
         * un-COW'ed pages by matching them up with "vma->vm_pgoff".
         */
        if (is_cow_mapping(vma->vm_flags)) {
                if (addr != vma->vm_start || end != vma->vm_end)
                        return -EINVAL;
                vma->vm_pgoff = pfn;
        }

        vma->vm_flags |= VM_IO | VM_RESERVED | VM_PFNMAP;

        BUG_ON(addr >= end);
        pfn -= addr >> PAGE_SHIFT;
        pgd = pgd_offset(mm, addr);
        flush_cache_range(vma, addr, end);
        do {
                next = pgd_addr_end(addr, end);
                err = remap_pud_range(mm, pgd, addr, next,
                                pfn + (addr >> PAGE_SHIFT), prot);
                if (err)
                        break;
        } while (pgd++, addr = next, addr != end);
        return err;
}
上面加黑处的：pfn -= addr >> PAGE_SHIFT；表示什么意思？
我的看法是一个页帧号减去一个用户空间虚拟地址右移12？？ 不懂

kiongf

addr >>　PAGE_SHIFT  : 获得物理地址所在的page frame的 PFN。

tc1989tc

回复 2# kiongf


    但是addr是用户空间的虚拟地址哦、、

firkraag

See that?
  next = pgd_addr_end(addr, end);
             err = remap_pud_range(mm, pgd, addr, next,
                                pfn + (addr >> PAGE_SHIFT), prot);
   while (pgd++, addr = next, addr != end);

That's a way of making use of the arg 'addr'(the counter of do while loop) as an increment of the PFN to map.

tc1989tc

回复 4# firkraag


    thank you！！