fs/jfs: Use full 40 bits offset and address for a data extent

An extent's logical offset and address are represented as a 40-bit value
split into two parts: the most significant 8 bits and the least
significant 32 bits. Currently the JFS code uses only the least
significant 32 bits value for offsets and addresses assuming the data
size will never exceed the 32-bit range. This approach ignores the most
significant 8 bits potentially leading to incorrect offsets and
addresses for larger values. The patch fixes it by incorporating the
most significant 8 bits into the calculation to get the full 40-bits
value for offsets and addresses.

https://jfs.sourceforge.net/project/pub/jfslayout.pdf

  "off1,off2 is a 40-bit field, containing the logical offset of the first
   block in the extent.
   ...
   addr1,addr2 is a 40-bit field, containing the address of the extent."

Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Ross Philipson <ross.philipson@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>

diff --git a/grub-core/fs/jfs.c b/grub-core/fs/jfs.c
index 7a68fcbe3bf3bac3cd439759719b0ceb98546d51..3026d5a0bbba61f302a72359acb737e6ae1275be 100644 (file)
--- a/grub-core/fs/jfs.c
+++ b/grub-core/fs/jfs.c
@@ -265,6 +265,20 @@ static grub_dl_t my_mod;
 \f
 static grub_err_t grub_jfs_lookup_symlink (struct grub_jfs_data *data, grub_uint32_t ino);
 
+/*
+ * An extent's offset, physical and logical, is represented as a 40-bit value.
+ * This 40-bit value is split into two parts:
+ *   - offset1: the most signficant 8 bits of the offset,
+ *   - offset2: the least significant 32 bits of the offset.
+ *
+ * This function calculates and returns the 64-bit offset of an extent.
+ */
+static grub_uint64_t
+get_ext_offset (grub_uint8_t offset1, grub_uint32_t offset2)
+{
+  return (((grub_uint64_t) offset1 << 32) | grub_le_to_cpu32 (offset2));
+}
+
 static grub_int64_t
 getblk (struct grub_jfs_treehead *treehead,
        struct grub_jfs_tree_extent *extents,
@@ -274,22 +288,25 @@ getblk (struct grub_jfs_treehead *treehead,
 {
   int found = -1;
   int i;
+  grub_uint64_t ext_offset, ext_blk;
 
   for (i = 0; i < grub_le_to_cpu16 (treehead->count) - 2 &&
              i < max_extents; i++)
     {
+      ext_offset = get_ext_offset (extents[i].offset1, extents[i].offset2);
+      ext_blk = get_ext_offset (extents[i].extent.blk1, extents[i].extent.blk2);
+
       if (treehead->flags & GRUB_JFS_TREE_LEAF)
        {
          /* Read the leafnode.  */
-         if (grub_le_to_cpu32 (extents[i].offset2) <= blk
+         if (ext_offset <= blk
              && ((grub_le_to_cpu16 (extents[i].extent.length))
                  + (extents[i].extent.length2 << 16)
-                 + grub_le_to_cpu32 (extents[i].offset2)) > blk)
-           return (blk - grub_le_to_cpu32 (extents[i].offset2)
-                   + grub_le_to_cpu32 (extents[i].extent.blk2));
+                 + ext_offset) > blk)
+           return (blk - ext_offset + ext_blk);
        }
       else
-       if (blk >= grub_le_to_cpu32 (extents[i].offset2))
+       if (blk >= ext_offset)
          found = i;
     }
 
@@ -307,10 +324,9 @@ getblk (struct grub_jfs_treehead *treehead,
        return -1;
 
       if (!grub_disk_read (data->disk,
-                          ((grub_disk_addr_t) grub_le_to_cpu32 (extents[found].extent.blk2))
-                          << (grub_le_to_cpu16 (data->sblock.log2_blksz)
-                              - GRUB_DISK_SECTOR_BITS), 0,
-                          sizeof (*tree), (char *) tree))
+                          (grub_disk_addr_t) ext_blk
+                          << (grub_le_to_cpu16 (data->sblock.log2_blksz) - GRUB_DISK_SECTOR_BITS),
+                          0, sizeof (*tree), (char *) tree))
        {
          if (grub_memcmp (&tree->treehead, treehead, sizeof (struct grub_jfs_treehead)) ||
              grub_memcmp (&tree->extents, extents, 254 * sizeof (struct grub_jfs_tree_extent)))
@@ -361,7 +377,7 @@ grub_jfs_read_inode (struct grub_jfs_data *data, grub_uint32_t ino,
                      sizeof (iag_inodes), &iag_inodes))
     return grub_errno;
 
-  inoblk = grub_le_to_cpu32 (iag_inodes[inoext].blk2);
+  inoblk = get_ext_offset (iag_inodes[inoext].blk1, iag_inodes[inoext].blk2);
   inoblk <<= (grub_le_to_cpu16 (data->sblock.log2_blksz)
              - GRUB_DISK_SECTOR_BITS);
   inoblk += inonum;
@@ -490,7 +506,8 @@ grub_jfs_opendir (struct grub_jfs_data *data, struct grub_jfs_inode *inode)
       return 0;
     }
 
-  blk = grub_le_to_cpu32 (de[inode->dir.header.sorted[0]].ex.blk2);
+  blk = get_ext_offset (de[inode->dir.header.sorted[0]].ex.blk1,
+                     de[inode->dir.header.sorted[0]].ex.blk2);
   blk <<= (grub_le_to_cpu16 (data->sblock.log2_blksz) - GRUB_DISK_SECTOR_BITS);
 
   /* Read in the nodes until we are on the leaf node level.  */
@@ -508,7 +525,7 @@ grub_jfs_opendir (struct grub_jfs_data *data, struct grub_jfs_inode *inode)
 
       de = (struct grub_jfs_internal_dirent *) diro->dirpage->dirent;
       index = diro->dirpage->sorted[diro->dirpage->header.sindex * 32];
-      blk = (grub_le_to_cpu32 (de[index].ex.blk2)
+      blk = (get_ext_offset (de[index].ex.blk1, de[index].ex.blk2)
             << (grub_le_to_cpu16 (data->sblock.log2_blksz)
                 - GRUB_DISK_SECTOR_BITS));
     } while (!(diro->dirpage->header.flags & GRUB_JFS_TREE_LEAF));