linux-kernel - [PATCH v3] vc_screen: reload load of struct vc_data pointer in vcs

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>] [thread-next>] [day] [month] [year] [list]

Message-Id: <1683889728-10411-1-git-send-email-george.kennedy@oracle.com>
Date:   Fri, 12 May 2023 06:08:48 -0500
From:   George Kennedy <george.kennedy@...cle.com>
To:     linux@...ssschuh.net, gregkh@...uxfoundation.org,
        jirislaby@...nel.org
Cc:     george.kennedy@...cle.com, sfr@...b.auug.org.au,
        akpm@...ux-foundation.org, linux-kernel@...r.kernel.org,
        linux-serial@...r.kernel.org, regressions@...ts.linux.dev
Subject: [PATCH v3] vc_screen: reload load of struct vc_data pointer in vcs_write() to avoid UAF

After a call to console_unlock() in vcs_write() the vc_data struct can be
freed by vc_port_destruct(). Because of that, the struct vc_data pointer
must be reloaded in the while loop in vcs_write() after console_lock() to
avoid a UAF when vcs_size() is called.

Syzkaller reported a UAF in vcs_size().

BUG: KASAN: slab-use-after-free in vcs_size (drivers/tty/vt/vc_screen.c:215)
Read of size 4 at addr ffff8880beab89a8 by task repro_vcs_size/4119

Call Trace:
 <TASK>
__asan_report_load4_noabort (mm/kasan/report_generic.c:380)
vcs_size (drivers/tty/vt/vc_screen.c:215)
vcs_write (drivers/tty/vt/vc_screen.c:664)
vfs_write (fs/read_write.c:582 fs/read_write.c:564)
...
 <TASK>

Allocated by task 1213:
kmalloc_trace (mm/slab_common.c:1064)
vc_allocate (./include/linux/slab.h:559 ./include/linux/slab.h:680
    drivers/tty/vt/vt.c:1078 drivers/tty/vt/vt.c:1058)
con_install (drivers/tty/vt/vt.c:3334)
tty_init_dev (drivers/tty/tty_io.c:1303 drivers/tty/tty_io.c:1415
    drivers/tty/tty_io.c:1392)
tty_open (drivers/tty/tty_io.c:2082 drivers/tty/tty_io.c:2128)
chrdev_open (fs/char_dev.c:415)
do_dentry_open (fs/open.c:921)
vfs_open (fs/open.c:1052)
...

Freed by task 4116:
kfree (mm/slab_common.c:1016)
vc_port_destruct (drivers/tty/vt/vt.c:1044)
tty_port_destructor (drivers/tty/tty_port.c:296)
tty_port_put (drivers/tty/tty_port.c:312)
vt_disallocate_all (drivers/tty/vt/vt_ioctl.c:662 (discriminator 2))
vt_ioctl (drivers/tty/vt/vt_ioctl.c:903)
tty_ioctl (drivers/tty/tty_io.c:2778)
...

The buggy address belongs to the object at ffff8880beab8800
 which belongs to the cache kmalloc-1k of size 1024
The buggy address is located 424 bytes inside of
 freed 1024-byte region [ffff8880beab8800, ffff8880beab8c00)

The buggy address belongs to the physical page:
page:00000000afc77580 refcount:1 mapcount:0 mapping:0000000000000000
    index:0x0 pfn:0xbeab8
head:00000000afc77580 order:3 entire_mapcount:0 nr_pages_mapped:0
    pincount:0
flags: 0xfffffc0010200(slab|head|node=0|zone=1|lastcpupid=0x1fffff)
page_type: 0xffffffff()
raw: 000fffffc0010200 ffff888100042dc0 ffffea000426de00 dead000000000002
raw: 0000000000000000 0000000000100010 00000001ffffffff 0000000000000000
page dumped because: kasan: bad access detected

Memory state around the buggy address:
 ffff8880beab8880: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
 ffff8880beab8900: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
>ffff8880beab8980: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
                                  ^
 ffff8880beab8a00: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
 ffff8880beab8a80: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
==================================================================
Disabling lock debugging due to kernel taint

Fixes: ac751efa6a0d ("console: rename acquire/release_console_sem() to console_lock/unlock()")
Reported-by: syzkaller <syzkaller@...glegroups.com>
Signed-off-by: George Kennedy <george.kennedy@...cle.com>
---
v3: changed to "ret = -ENXIO;"

v2: added "ret = -EINVAL;"

v1: This patch is similar to the recent UAF vcs_read() patches (226fae124b2d & 46d733d0efc7).
vcs_write() should have been checked for the same UAF issue at the same time the fixes went
into vcs_read().

 drivers/tty/vt/vc_screen.c | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)

diff --git a/drivers/tty/vt/vc_screen.c b/drivers/tty/vt/vc_screen.c
index 498ba9c0ee93..829c4be66f3b 100644
--- a/drivers/tty/vt/vc_screen.c
+++ b/drivers/tty/vt/vc_screen.c
@@ -656,10 +656,17 @@ vcs_write(struct file *file, const char __user *buf, size_t count, loff_t *ppos)
 			}
 		}
 
-		/* The vcs_size might have changed while we slept to grab
-		 * the user buffer, so recheck.
+		/* The vc might have been freed or vcs_size might have changed
+		 * while we slept to grab the user buffer, so recheck.
 		 * Return data written up to now on failure.
 		 */
+		vc = vcs_vc(inode, &viewed);
+		if (!vc) {
+			if (written)
+				break;
+			ret = -ENXIO;
+			goto unlock_out;
+		}
 		size = vcs_size(vc, attr, false);
 		if (size < 0) {
 			if (written)
-- 
2.31.1