Message-ID: <CAGyP=7eR9nQEdKhfLEzQ7kV9WbHeZV5gUQPL2Kf5OYXTu8tqjQ@mail.gmail.com>
Date:   Sat, 3 Jun 2023 04:39:25 -0700
From:   Palash Oswal <oswalpalash@...il.com>
To:     Anton Altaparmakov <anton@...era.com>,
        linux-ntfs-dev@...ts.sourceforge.net,
        LKML <linux-kernel@...r.kernel.org>,
        syzkaller-bugs <syzkaller-bugs@...glegroups.com>
Subject: KASAN: slab-use-after-free Read in ntfs_test_inode

Hello,
I found the following issue using syzkaller with enriched corpus[1] on:
HEAD commit : 0bcc4025550403ae28d2984bddacafbca0a2f112
git tree: linux
C Reproducer : I do not have a C reproducer yet. I will update this
thread when I get one.
Kernel .config :
https://gist.github.com/oswalpalash/d9580b0bfce202b37445fa5fd426e41f

Link:
1. https://github.com/cmu-pasta/linux-kernel-enriched-corpus

Console log :


loop2: detected capacity change from 0 to 4096
ntfs: volume version 3.1.
==================================================================
BUG: KASAN: slab-use-after-free in ntfs_test_inode+0x9a/0x2f0
Read of size 8 at addr ffff888042df75e0 by task syz-executor.2/5020
CPU: 1 PID: 5020 Comm: syz-executor.2 Not tainted
6.3.0-rc6-pasta-00035-g0bcc40255504 #1     Hardware name: QEMU
Standard PC (i440FX + PIIX, 1996), BIOS 1.13.0-1ubuntu1.1 04/01/2014

Call Trace:
 <TASK>
 dump_stack_lvl+0xd9/0x150
 print_address_description.constprop.0+0x2c/0x3c0
 kasan_report+0x11c/0x130
 kasan_check_range+0x141/0x190
 ntfs_test_inode+0x9a/0x2f0
 find_inode+0xe4/0x220
 iget5_locked+0xb5/0x2c0
 ntfs_iget+0xa1/0x180
 ntfs_fill_super+0x44bd/0x9400
 mount_bdev+0x351/0x410
 legacy_get_tree+0x109/0x220
 vfs_get_tree+0x8d/0x350
 path_mount+0x1342/0x1e40
 __x64_sys_mount+0x283/0x300
 do_syscall_64+0x39/0xb0
 entry_SYSCALL_64_after_hwframe+0x63/0xcd
RIP: 0033:0x7f138809146e
Code: 48 c7 c1 b8 ff ff ff f7 d8 64 89 01 48 83 c8 ff c3 66 2e 0f 1f
84 00 00 00 00 00 90 f3 0f 1e fa 49 89 ca b8 a5 00 00 00 0f 05 <48> 3d
01 f0 ff ff 73 01 c3 48 c7 c1 b8 ff ff ff f7 d8 64 89 01 48
RSP: 002b:00007f1388e1da08 EFLAGS: 00000206 ORIG_RAX: 00000000000000a5
RAX: ffffffffffffffda RBX: 0000000020000200 RCX: 00007f138809146e
RDX: 0000000020000000 RSI: 0000000020000100 RDI: 00007f1388e1da60
RBP: 00007f1388e1daa0 R08: 00007f1388e1daa0 R09: 0000000020000000
R10: 0000000000000000 R11: 0000000000000206 R12: 0000000020000000
R13: 0000000020000100 R14: 00007f1388e1da60 R15: 0000000020079ca0
 </TASK>

Allocated by task 15621:
 kasan_save_stack+0x22/0x40
 kasan_set_track+0x25/0x30
 __kasan_slab_alloc+0x7f/0x90
 kmem_cache_alloc_lru+0x1e3/0x640
 reiserfs_alloc_inode+0x28/0xc0
 alloc_inode+0x61/0x230
 new_inode+0x2b/0x280
 reiserfs_mkdir+0x1f1/0x990
 reiserfs_xattr_init+0x57e/0xc30
 reiserfs_fill_super+0x2127/0x2ea0
 mount_bdev+0x351/0x410
 legacy_get_tree+0x109/0x220
 vfs_get_tree+0x8d/0x350
 path_mount+0x1342/0x1e40
 __x64_sys_mount+0x283/0x300
 do_syscall_64+0x39/0xb0
 entry_SYSCALL_64_after_hwframe+0x63/0xcd

Last potentially related work creation:
 kasan_save_stack+0x22/0x40
 __kasan_record_aux_stack+0x7b/0x90
 __call_rcu_common.constprop.0+0x99/0x7e0
 destroy_inode+0x129/0x1b0
 iput.part.0+0x50a/0x740
 iput+0x5c/0x80
 reiserfs_new_inode+0x402/0x2190
 reiserfs_mkdir+0x4b8/0x990
 reiserfs_xattr_init+0x57e/0xc30
 reiserfs_fill_super+0x2127/0x2ea0
 mount_bdev+0x351/0x410
 legacy_get_tree+0x109/0x220
 vfs_get_tree+0x8d/0x350
 path_mount+0x1342/0x1e40
 __x64_sys_mount+0x283/0x300
 do_syscall_64+0x39/0xb0
 entry_SYSCALL_64_after_hwframe+0x63/0xcd
The buggy address belongs to the object at ffff888042df7080
 which belongs to the cache reiser_inode_cache of size 1568
The buggy address is located 1376 bytes inside of
 freed 1568-byte region [ffff888042df7080, ffff888042df76a0)
The buggy address belongs to the physical page:
page:ffffea00010b7dc0 refcount:1 mapcount:0 mapping:0000000000000000
index:0xffff888042df7ffe pfn:0x42df7
flags: 0xfff00000000200(slab|node=0|zone=1|lastcpupid=0x7ff)
raw: 00fff00000000200 ffff8881031dba00 ffffea00010b2110 ffff888015cfb940
raw: ffff888042df7ffe ffff888042df7080 0000000100000001 0000000000000000
page dumped because: kasan: bad access detected
page_owner tracks the page as allocated
page last allocated via order 0, migratetype Reclaimable, gfp_mask
0x2420d0(__GFP_IO|__GFP_FS|__GFP_NOWARN|__GFP_COMP|__GFP_THISNODE|__GFP_RECLAIMABLE),
pid 15621, tgid 15617 (syz-executor.0), ts 208467737645, f
ree_ts 0
 get_page_from_freelist+0x1190/0x2e20
 __alloc_pages+0x1cb/0x4a0
 cache_grow_begin+0x9b/0x3b0
 cache_alloc_refill+0x27f/0x380
 kmem_cache_alloc_lru+0x540/0x640
 reiserfs_alloc_inode+0x28/0xc0
 alloc_inode+0x61/0x230
 iget5_locked+0x1cf/0x2c0
 reiserfs_fill_super+0xf84/0x2ea0
 mount_bdev+0x351/0x410
 legacy_get_tree+0x109/0x220
 vfs_get_tree+0x8d/0x350
 path_mount+0x1342/0x1e40
 __x64_sys_mount+0x283/0x300
 do_syscall_64+0x39/0xb0
 entry_SYSCALL_64_after_hwframe+0x63/0xcd
page_owner free stack trace missing

Memory state around the buggy address:
 ffff888042df7480: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
 ffff888042df7500: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
>ffff888042df7580: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
                                                       ^
 ffff888042df7600: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
 ffff888042df7680: fb fb fb fb fc fc fc fc fc fc fc fc fc fc fc fc
==================================================================

Powered by blists - more mailing lists