linux-kernel - [PATCH] Fix slab-out-of-bounds in r

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [day] [month] [year] [list]

Message-Id: <20211217084045.363818-1-yinxiujiang@kylinos.cn>
Date:   Fri, 17 Dec 2021 16:40:45 +0800
From:   Yin Xiujiang <yinxiujiang@...inos.cn>
To:     thomas.kuehnel@....de, almaz.alexandrovich@...agon-software.com
Cc:     kbuild-all@...ts.01.org, ntfs3@...ts.linux.dev,
        linux-kernel@...r.kernel.org, n.schier@....de
Subject: [PATCH] Fix slab-out-of-bounds in r_page

When PAGE_SIZE is 64K, if read_log_page is called by log_read_rst for
the first time, the size of *buffer would be equal to
DefaultLogPageSize(4K).But for *buffer operations like memcpy,
if the memory area size(n) which being assigned to buffer is larger
than 4K (log->page_size(64K) or bytes(64K-page_off)), it will cause
an out of boundary error.
 Call trace:
  [...]
  kasan_report+0x44/0x130
  check_memory_region+0xf8/0x1a0
  memcpy+0xc8/0x100
  ntfs_read_run_nb+0x20c/0x460
  read_log_page+0xd0/0x1f4
  log_read_rst+0x110/0x75c
  log_replay+0x1e8/0x4aa0
  ntfs_loadlog_and_replay+0x290/0x2d0
  ntfs_fill_super+0x508/0xec0
  get_tree_bdev+0x1fc/0x34c
  [...]

Fix this by setting variable r_page to NULL in log_read_rst.

Signed-off-by: Yin Xiujiang <yinxiujiang@...inos.cn>
---
 fs/ntfs3/fslog.c | 24 +-----------------------
 1 file changed, 1 insertion(+), 23 deletions(-)

diff --git a/fs/ntfs3/fslog.c b/fs/ntfs3/fslog.c
index 06492f088d60..4fdb5bdfc2ee 100644
--- a/fs/ntfs3/fslog.c
+++ b/fs/ntfs3/fslog.c
@@ -1180,10 +1180,7 @@ static int log_read_rst(struct ntfs_log *log, u32 l_size, bool first,
 			struct restart_info *info)
 {
 	u32 skip, vbo;
-	struct RESTART_HDR *r_page = kmalloc(DefaultLogPageSize, GFP_NOFS);
-
-	if (!r_page)
-		return -ENOMEM;
+	struct RESTART_HDR *r_page = NULL;
 
 	memset(info, 0, sizeof(struct restart_info));
 
@@ -1199,7 +1196,6 @@ static int log_read_rst(struct ntfs_log *log, u32 l_size, bool first,
 	/* Loop continuously until we succeed. */
 	for (; vbo < l_size; vbo = 2 * vbo + skip, skip = 0) {
 		bool usa_error;
-		u32 sys_page_size;
 		bool brst, bchk;
 		struct RESTART_AREA *ra;
 
@@ -1253,24 +1249,6 @@ static int log_read_rst(struct ntfs_log *log, u32 l_size, bool first,
 			goto check_result;
 		}
 
-		/* Read the entire restart area. */
-		sys_page_size = le32_to_cpu(r_page->sys_page_size);
-		if (DefaultLogPageSize != sys_page_size) {
-			kfree(r_page);
-			r_page = kzalloc(sys_page_size, GFP_NOFS);
-			if (!r_page)
-				return -ENOMEM;
-
-			if (read_log_page(log, vbo,
-					  (struct RECORD_PAGE_HDR **)&r_page,
-					  &usa_error)) {
-				/* Ignore any errors. */
-				kfree(r_page);
-				r_page = NULL;
-				continue;
-			}
-		}
-
 		if (is_client_area_valid(r_page, usa_error)) {
 			info->valid_page = true;
 			ra = Add2Ptr(r_page, le16_to_cpu(r_page->ra_off));
-- 
2.30.0