Squashfs: fix variable overflow in squashfs_readpage_block

Syzbot reports a slab out of bounds access in squashfs_readpage_block().

This is caused by an attempt to read page index 0x2000000000.  This value
(start_index) is stored in an integer loop variable which overflows
producing a value of 0.  This causes a loop which iterates over pages
start_index -> end_index to iterate over 0 -> end_index, which ultimately
causes an out of bounds page array access.

Fix by changing variable to a loff_t, and rename to index to make it
clearer it is a page index, and not a loop count.

Link: https://lkml.kernel.org/r/20241020232200.837231-1-phillip@squashfs.org.uk
Signed-off-by: Phillip Lougher <phillip@squashfs.org.uk>
Reported-by: "Lai, Yi" <yi1.lai@linux.intel.com>
Closes: https://lore.kernel.org/all/ZwzcnCAosIPqQ9Ie@ly-workstation/
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

diff --git a/fs/squashfs/file_direct.c b/fs/squashfs/file_direct.c
index 22251743fadfcd5ebfb8aa2fe40f91e2a3fb0735..d19d4db74af83b4e4ccff0cf0242d8c738c73d3d 100644 (file)
--- a/fs/squashfs/file_direct.c
+++ b/fs/squashfs/file_direct.c
@@ -30,7 +30,8 @@ int squashfs_readpage_block(struct page *target_page, u64 block, int bsize,
        int mask = (1 << (msblk->block_log - PAGE_SHIFT)) - 1;
        loff_t start_index = folio->index & ~mask;
        loff_t end_index = start_index | mask;
-       int i, n, pages, bytes, res = -ENOMEM;
+       loff_t index;
+       int i, pages, bytes, res = -ENOMEM;
        struct page **page, *last_page;
        struct squashfs_page_actor *actor;
        void *pageaddr;
@@ -45,9 +46,9 @@ int squashfs_readpage_block(struct page *target_page, u64 block, int bsize,
                return res;
 
        /* Try to grab all the pages covered by the Squashfs block */
-       for (i = 0, n = start_index; n <= end_index; n++) {
-               page[i] = (n == folio->index) ? target_page :
-                       grab_cache_page_nowait(target_page->mapping, n);
+       for (i = 0, index = start_index; index <= end_index; index++) {
+               page[i] = (index == folio->index) ? target_page :
+                       grab_cache_page_nowait(target_page->mapping, index);
 
                if (page[i] == NULL)
                        continue;