mm/mincore.c

   1 // SPDX-License-Identifier: GPL-2.0
   2 /*
   3  *      linux/mm/mincore.c
   4  *
   5  * Copyright (C) 1994-2006  Linus Torvalds
   6  */
   7
   8 /*
   9  * The mincore() system call.
  10  */
  11 #include <linux/pagemap.h>
  12 #include <linux/gfp.h>
  13 #include <linux/mm.h>
  14 #include <linux/mman.h>
  15 #include <linux/syscalls.h>
  16 #include <linux/swap.h>
  17 #include <linux/swapops.h>
  18 #include <linux/shmem_fs.h>
  19 #include <linux/hugetlb.h>
  20
  21 #include <linux/uaccess.h>
  22 #include <asm/pgtable.h>
  23
  24 static int mincore_hugetlb(pte_t *pte, unsigned long hmask, unsigned long addr,
  25                         unsigned long end, struct mm_walk *walk)
  26 {
  27 #ifdef CONFIG_HUGETLB_PAGE
  28         unsigned char present;
  29         unsigned char *vec = walk->private;
  30
  31         /*
  32          * Hugepages under user process are always in RAM and never
  33          * swapped out, but theoretically it needs to be checked.
  34          */
  35         present = pte && !huge_pte_none(huge_ptep_get(pte));
  36         for (; addr != end; vec++, addr += PAGE_SIZE)
  37                 *vec = present;
  38         walk->private = vec;
  39 #else
  40         BUG();
  41 #endif
  42         return 0;
  43 }
  44
  45 /*
  46  * Later we can get more picky about what "in core" means precisely.
  47  * For now, simply check to see if the page is in the page cache,
  48  * and is up to date; i.e. that no page-in operation would be required
  49  * at this time if an application were to map and access this page.
  50  */
  51 static unsigned char mincore_page(struct address_space *mapping, pgoff_t pgoff)
  52 {
  53         unsigned char present = 0;
  54         struct page *page;
  55
  56         /*
  57          * When tmpfs swaps out a page from a file, any process mapping that
  58          * file will not get a swp_entry_t in its pte, but rather it is like
  59          * any other file mapping (ie. marked !present and faulted in with
  60          * tmpfs's .fault). So swapped out tmpfs mappings are tested here.
  61          */
  62 #ifdef CONFIG_SWAP
  63         if (shmem_mapping(mapping)) {
  64                 page = find_get_entry(mapping, pgoff);
  65                 /*
  66                  * shmem/tmpfs may return swap: account for swapcache
  67                  * page too.
  68                  */
  69                 if (xa_is_value(page)) {
  70                         swp_entry_t swp = radix_to_swp_entry(page);
  71                         page = find_get_page(swap_address_space(swp),
  72                                              swp_offset(swp));
  73                 }
  74         } else
  75                 page = find_get_page(mapping, pgoff);
  76 #else
  77         page = find_get_page(mapping, pgoff);
  78 #endif
  79         if (page) {
  80                 present = PageUptodate(page);
  81                 put_page(page);
  82         }
  83
  84         return present;
  85 }
  86
  87 static int __mincore_unmapped_range(unsigned long addr, unsigned long end,
  88                                 struct vm_area_struct *vma, unsigned char *vec)
  89 {
  90         unsigned long nr = (end - addr) >> PAGE_SHIFT;
  91         int i;
  92
  93         if (vma->vm_file) {
  94                 pgoff_t pgoff;
  95
  96                 pgoff = linear_page_index(vma, addr);
  97                 for (i = 0; i < nr; i++, pgoff++)
  98                         vec[i] = mincore_page(vma->vm_file->f_mapping, pgoff);
  99         } else {
 100                 for (i = 0; i < nr; i++)
 101                         vec[i] = 0;
 102         }
 103         return nr;
 104 }
 105
 106 static int mincore_unmapped_range(unsigned long addr, unsigned long end,
 107                                    struct mm_walk *walk)
 108 {
 109         walk->private += __mincore_unmapped_range(addr, end,
 110                                                   walk->vma, walk->private);
 111         return 0;
 112 }
 113
 114 static int mincore_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end,
 115                         struct mm_walk *walk)
 116 {
 117         spinlock_t *ptl;
 118         struct vm_area_struct *vma = walk->vma;
 119         pte_t *ptep;
 120         unsigned char *vec = walk->private;
 121         int nr = (end - addr) >> PAGE_SHIFT;
 122
 123         ptl = pmd_trans_huge_lock(pmd, vma);
 124         if (ptl) {
 125                 memset(vec, 1, nr);
 126                 spin_unlock(ptl);
 127                 goto out;
 128         }
 129
 130         if (pmd_trans_unstable(pmd)) {
 131                 __mincore_unmapped_range(addr, end, vma, vec);
 132                 goto out;
 133         }
 134
 135         ptep = pte_offset_map_lock(walk->mm, pmd, addr, &ptl);
 136         for (; addr != end; ptep++, addr += PAGE_SIZE) {
 137                 pte_t pte = *ptep;
 138
 139                 if (pte_none(pte))
 140                         __mincore_unmapped_range(addr, addr + PAGE_SIZE,
 141                                                  vma, vec);
 142                 else if (pte_present(pte))
 143                         *vec = 1;
 144                 else { /* pte is a swap entry */
 145                         swp_entry_t entry = pte_to_swp_entry(pte);
 146
 147                         if (non_swap_entry(entry)) {
 148                                 /*
 149                                  * migration or hwpoison entries are always
 150                                  * uptodate
 151                                  */
 152                                 *vec = 1;
 153                         } else {
 154 #ifdef CONFIG_SWAP
 155                                 *vec = mincore_page(swap_address_space(entry),
 156                                                     swp_offset(entry));
 157 #else
 158                                 WARN_ON(1);
 159                                 *vec = 1;
 160 #endif
 161                         }
 162                 }
 163                 vec++;
 164         }
 165         pte_unmap_unlock(ptep - 1, ptl);
 166 out:
 167         walk->private += nr;
 168         cond_resched();
 169         return 0;
 170 }
 171
 172 static inline bool can_do_mincore(struct vm_area_struct *vma)
 173 {
 174         if (vma_is_anonymous(vma))
 175                 return true;
 176         if (!vma->vm_file)
 177                 return false;
 178         /*
 179          * Reveal pagecache information only for non-anonymous mappings that
 180          * correspond to the files the calling process could (if tried) open
 181          * for writing; otherwise we'd be including shared non-exclusive
 182          * mappings, which opens a side channel.
 183          */
 184         return inode_owner_or_capable(file_inode(vma->vm_file)) ||
 185                 inode_permission(file_inode(vma->vm_file), MAY_WRITE) == 0;
 186 }
 187
 188 /*
 189  * Do a chunk of "sys_mincore()". We've already checked
 190  * all the arguments, we hold the mmap semaphore: we should
 191  * just return the amount of info we're asked for.
 192  */
 193 static long do_mincore(unsigned long addr, unsigned long pages, unsigned char *vec)
 194 {
 195         struct vm_area_struct *vma;
 196         unsigned long end;
 197         int err;
 198         struct mm_walk mincore_walk = {
 199                 .pmd_entry = mincore_pte_range,
 200                 .pte_hole = mincore_unmapped_range,
 201                 .hugetlb_entry = mincore_hugetlb,
 202                 .private = vec,
 203         };
 204
 205         vma = find_vma(current->mm, addr);
 206         if (!vma || addr < vma->vm_start)
 207                 return -ENOMEM;
 208         end = min(vma->vm_end, addr + (pages << PAGE_SHIFT));
 209         if (!can_do_mincore(vma)) {
 210                 unsigned long pages = DIV_ROUND_UP(end - addr, PAGE_SIZE);
 211                 memset(vec, 1, pages);
 212                 return pages;
 213         }
 214         mincore_walk.mm = vma->vm_mm;
 215         err = walk_page_range(addr, end, &mincore_walk);
 216         if (err < 0)
 217                 return err;
 218         return (end - addr) >> PAGE_SHIFT;
 219 }
 220
 221 /*
 222  * The mincore(2) system call.
 223  *
 224  * mincore() returns the memory residency status of the pages in the
 225  * current process's address space specified by [addr, addr + len).
 226  * The status is returned in a vector of bytes.  The least significant
 227  * bit of each byte is 1 if the referenced page is in memory, otherwise
 228  * it is zero.
 229  *
 230  * Because the status of a page can change after mincore() checks it
 231  * but before it returns to the application, the returned vector may
 232  * contain stale information.  Only locked pages are guaranteed to
 233  * remain in memory.
 234  *
 235  * return values:
 236  *  zero    - success
 237  *  -EFAULT - vec points to an illegal address
 238  *  -EINVAL - addr is not a multiple of PAGE_SIZE
 239  *  -ENOMEM - Addresses in the range [addr, addr + len] are
 240  *              invalid for the address space of this process, or
 241  *              specify one or more pages which are not currently
 242  *              mapped
 243  *  -EAGAIN - A kernel resource was temporarily unavailable.
 244  */
 245 SYSCALL_DEFINE3(mincore, unsigned long, start, size_t, len,
 246                 unsigned char __user *, vec)
 247 {
 248         long retval;
 249         unsigned long pages;
 250         unsigned char *tmp;
 251
 252         /* Check the start address: needs to be page-aligned.. */
 253         if (start & ~PAGE_MASK)
 254                 return -EINVAL;
 255
 256         /* ..and we need to be passed a valid user-space range */
 257         if (!access_ok((void __user *) start, len))
 258                 return -ENOMEM;
 259
 260         /* This also avoids any overflows on PAGE_ALIGN */
 261         pages = len >> PAGE_SHIFT;
 262         pages += (offset_in_page(len)) != 0;
 263
 264         if (!access_ok(vec, pages))
 265                 return -EFAULT;
 266
 267         tmp = (void *) __get_free_page(GFP_USER);
 268         if (!tmp)
 269                 return -EAGAIN;
 270
 271         retval = 0;
 272         while (pages) {
 273                 /*
 274                  * Do at most PAGE_SIZE entries per iteration, due to
 275                  * the temporary buffer size.
 276                  */
 277                 down_read(&current->mm->mmap_sem);
 278                 retval = do_mincore(start, min(pages, PAGE_SIZE), tmp);
 279                 up_read(&current->mm->mmap_sem);
 280
 281                 if (retval <= 0)
 282                         break;
 283                 if (copy_to_user(vec, tmp, retval)) {
 284                         retval = -EFAULT;
 285                         break;
 286                 }
 287                 pages -= retval;
 288                 vec += retval;
 289                 start += retval << PAGE_SHIFT;
 290                 retval = 0;
 291         }
 292         free_page((unsigned long) tmp);
 293         return retval;
 294 }