INFO: trying to register non-static key in nilfs_bmap_lookup_at_level
syzbot
syzbotHello,
syzbot found the following issue on:
HEAD commit: 3800a713b607 Merge tag 'mm-hotfixes-stable-2022-09-26' of .. git tree: upstream console+strace: https://syzkaller.appspot.com/x/log.txt?x=12704b40880000 kernel config: https://syzkaller.appspot.com/x/.config?x=ba0d23aa7e1ffaf5 dashboard link: https://syzkaller.appspot.com/bug?extid=2b32eb36c1a825b7a74c compiler: Debian clang version 13.0.1-++20220126092033+75e33f71c2da-1~exp1~20220126212112.63, GNU ld (GNU Binutils for Debian) 2.35.2 syz repro: https://syzkaller.appspot.com/x/repro.syz?x=14e930c0880000 C reproducer: https://syzkaller.appspot.com/x/repro.c?x=157ce604880000
IMPORTANT: if you fix the issue, please add the following tag to the commit:
loop0: detected capacity change from 0 to 264192
NILFS (loop0): broken superblock, retrying with spare superblock (blocksize = 1024)
INFO: trying to register non-static key.
The code is fine but needs lockdep annotation, or maybe
you didn't initialize this object before use?
turning off the locking correctness validator.
CPU: 0 PID: 3614 Comm: syz-executor231 Not tainted 6.0.0-rc7-syzkaller-00029-g3800a713b607 #0 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 08/26/2022 Call Trace: <TASK> __dump_stack lib/dump_stack.c:88 [inline] dump_stack_lvl+0x1b1/0x28e lib/dump_stack.c:106 assign_lock_key+0x212/0x220 kernel/locking/lockdep.c:979 register_lock_class+0x18d/0x930 kernel/locking/lockdep.c:1292 __lock_acquire+0xe4/0x1f60 kernel/locking/lockdep.c:4932 lock_acquire+0x182/0x3c0 kernel/locking/lockdep.c:5666 down_read+0x39/0x50 kernel/locking/rwsem.c:1499 nilfs_bmap_lookup_at_level+0x88/0x350 fs/nilfs2/bmap.c:68 nilfs_bmap_lookup fs/nilfs2/bmap.h:170 [inline] nilfs_mdt_submit_block+0x23d/0x520 fs/nilfs2/mdt.c:142 nilfs_mdt_read_block+0xa3/0x460 fs/nilfs2/mdt.c:176 nilfs_mdt_get_block+0x104/0x820 fs/nilfs2/mdt.c:251 nilfs_palloc_get_block+0x142/0x240 fs/nilfs2/alloc.c:216 nilfs_palloc_get_entry_block+0x1e8/0x2e0 fs/nilfs2/alloc.c:318 nilfs_ifile_get_inode_block+0xf0/0x1a0 fs/nilfs2/ifile.c:143 __nilfs_read_inode fs/nilfs2/inode.c:477 [inline] nilfs_iget+0x253/0x8b0 fs/nilfs2/inode.c:603 nilfs_get_root_dentry+0x28/0x230 fs/nilfs2/super.c:904 nilfs_fill_super+0x3dd/0x5d0 fs/nilfs2/super.c:1078 nilfs_mount+0x613/0x9b0 fs/nilfs2/super.c:1317 legacy_get_tree+0xea/0x180 fs/fs_context.c:610 vfs_get_tree+0x88/0x270 fs/super.c:1530 do_new_mount+0x289/0xad0 fs/namespace.c:3040 do_mount fs/namespace.c:3383 [inline] __do_sys_mount fs/namespace.c:3591 [inline] __se_sys_mount+0x2d3/0x3c0 fs/namespace.c:3568 do_syscall_x64 arch/x86/entry/common.c:50 [inline] do_syscall_64+0x3d/0xb0 arch/x86/entry/common.c:80 entry_SYSCALL_64_after_hwframe+0x63/0xcd RIP: 0033:0x7efd97e9462a Code: 83 c4 08 5b 5d c3 66 2e 0f 1f 84 00 00 00 00 00 c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 49 89 ca b8 a5 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 c0 ff ff ff f7 d8 64 89 01 48 RSP: 002b:00007ffc213b3aa8 EFLAGS: 00000286 ORIG_RAX: 00000000000000a5 RAX: ffffffffffffffda RBX: 00007ffc213b3b00 RCX: 00007efd97e9462a RDX: 0000000020000000 RSI: 0000000020000100 RDI: 00007ffc213b3ac0 RBP: 00007ffc213b3ac0 R08: 00007ffc213b3b00 R09: 0000000000000000 R10: 0000000000000000 R11: 0000000000000286 R12: 0000000020000788 R13: 0000000000000003 R14: 0000000000000004 R15: 000000000000003b </TASK> general protection fault, probably for non-canonical address 0xdffffc0000000000: 0000 [#1] PREEMPT SMP KASAN KASAN: null-ptr-deref in range [0x0000000000000000-0x0000000000000007] CPU: 1 PID: 3614 Comm: syz-executor231 Not tainted 6.0.0-rc7-syzkaller-00029-g3800a713b607 #0 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 08/26/2022 RIP: 0010:nilfs_bmap_lookup_at_level+0xb9/0x350 fs/nilfs2/bmap.c:69 Code: 48 89 d8 48 c1 e8 03 42 80 3c 30 00 74 08 48 89 df e8 4b 9b 94 fe 4c 8b 33 4c 89 f0 48 c1 e8 03 48 b9 00 00 00 00 00 fc ff df <80> 3c 08 00 74 08 4c 89 f7 e8 29 9b 94 fe 4c 89 ff 4c 89 ee 44 89 RSP: 0018:ffffc9000385f440 EFLAGS: 00010246 RAX: 0000000000000000 RBX: ffff888074ff2760 RCX: dffffc0000000000 RDX: 0000000000000001 RSI: 0000000000000008 RDI: ffff888074ff26c8 RBP: ffffc9000385f4f0 R08: dffffc0000000000 R09: ffffed100e9fe4da R10: ffffed100e9fe4da R11: 1ffff1100e9fe4d9 R12: 0000000000000001 R13: 0000000000000002 R14: 0000000000000000 R15: ffff888074ff2688 FS: 0000555555be2300(0000) GS:ffff8880b9b00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00007ffc213b4000 CR3: 00000000725e8000 CR4: 00000000003506e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Call Trace: <TASK> nilfs_bmap_lookup fs/nilfs2/bmap.h:170 [inline] nilfs_mdt_submit_block+0x23d/0x520 fs/nilfs2/mdt.c:142 nilfs_mdt_read_block+0xa3/0x460 fs/nilfs2/mdt.c:176 nilfs_mdt_get_block+0x104/0x820 fs/nilfs2/mdt.c:251 nilfs_palloc_get_block+0x142/0x240 fs/nilfs2/alloc.c:216 nilfs_palloc_get_entry_block+0x1e8/0x2e0 fs/nilfs2/alloc.c:318 nilfs_ifile_get_inode_block+0xf0/0x1a0 fs/nilfs2/ifile.c:143 __nilfs_read_inode fs/nilfs2/inode.c:477 [inline] nilfs_iget+0x253/0x8b0 fs/nilfs2/inode.c:603 nilfs_get_root_dentry+0x28/0x230 fs/nilfs2/super.c:904 nilfs_fill_super+0x3dd/0x5d0 fs/nilfs2/super.c:1078 nilfs_mount+0x613/0x9b0 fs/nilfs2/super.c:1317 legacy_get_tree+0xea/0x180 fs/fs_context.c:610 vfs_get_tree+0x88/0x270 fs/super.c:1530 do_new_mount+0x289/0xad0 fs/namespace.c:3040 do_mount fs/namespace.c:3383 [inline] __do_sys_mount fs/namespace.c:3591 [inline] __se_sys_mount+0x2d3/0x3c0 fs/namespace.c:3568 do_syscall_x64 arch/x86/entry/common.c:50 [inline] do_syscall_64+0x3d/0xb0 arch/x86/entry/common.c:80 entry_SYSCALL_64_after_hwframe+0x63/0xcd RIP: 0033:0x7efd97e9462a Code: 83 c4 08 5b 5d c3 66 2e 0f 1f 84 00 00 00 00 00 c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 49 89 ca b8 a5 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 c0 ff ff ff f7 d8 64 89 01 48 RSP: 002b:00007ffc213b3aa8 EFLAGS: 00000286 ORIG_RAX: 00000000000000a5 RAX: ffffffffffffffda RBX: 00007ffc213b3b00 RCX: 00007efd97e9462a RDX: 0000000020000000 RSI: 0000000020000100 RDI: 00007ffc213b3ac0 RBP: 00007ffc213b3ac0 R08: 00007ffc213b3b00 R09: 0000000000000000 R10: 0000000000000000 R11: 0000000000000286 R12: 0000000020000788 R13: 0000000000000003 R14: 0000000000000004 R15: 000000000000003b </TASK> Modules linked in: ---[ end trace 0000000000000000 ]--- RIP: 0010:nilfs_bmap_lookup_at_level+0xb9/0x350 fs/nilfs2/bmap.c:69 Code: 48 89 d8 48 c1 e8 03 42 80 3c 30 00 74 08 48 89 df e8 4b 9b 94 fe 4c 8b 33 4c 89 f0 48 c1 e8 03 48 b9 00 00 00 00 00 fc ff df <80> 3c 08 00 74 08 4c 89 f7 e8 29 9b 94 fe 4c 89 ff 4c 89 ee 44 89 RSP: 0018:ffffc9000385f440 EFLAGS: 00010246 RAX: 0000000000000000 RBX: ffff888074ff2760 RCX: dffffc0000000000 RDX: 0000000000000001 RSI: 0000000000000008 RDI: ffff888074ff26c8 RBP: ffffc9000385f4f0 R08: dffffc0000000000 R09: ffffed100e9fe4da R10: ffffed100e9fe4da R11: 1ffff1100e9fe4d9 R12: 0000000000000001 R13: 0000000000000002 R14: 0000000000000000 R15: ffff888074ff2688 FS: 0000555555be2300(0000) GS:ffff8880b9a00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000560cba9be408 CR3: 00000000725e8000 CR4: 00000000003506f0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 ---------------- Code disassembly (best guess): 0: 48 89 d8 mov %rbx,%rax 3: 48 c1 e8 03 shr $0x3,%rax 7: 42 80 3c 30 00 cmpb $0x0,(%rax,%r14,1) c: 74 08 je 0x16 e: 48 89 df mov %rbx,%rdi 11: e8 4b 9b 94 fe callq 0xfe949b61 16: 4c 8b 33 mov (%rbx),%r14 19: 4c 89 f0 mov %r14,%rax 1c: 48 c1 e8 03 shr $0x3,%rax 20: 48 b9 00 00 00 00 00 movabs $0xdffffc0000000000,%rcx 27: fc ff df * 2a: 80 3c 08 00 cmpb $0x0,(%rax,%rcx,1) <-- trapping instruction 2e: 74 08 je 0x38 30: 4c 89 f7 mov %r14,%rdi 33: e8 29 9b 94 fe callq 0xfe949b61 38: 4c 89 ff mov %r15,%rdi 3b: 4c 89 ee mov %r13,%rsi 3e: 44 rex.R 3f: 89 .byte 0x89
---
This report is generated by a bot. It may contain errors.
See https://goo.gl/tpsmEJ for more information about syzbot.
syzbot engineers can be reached at syzkaller-/JYPxA39Uh5TLH3MbocFF+G/[email protected]
syzbot will keep track of this issue. See:
https://goo.gl/tpsmEJ#status for how to communicate with syzbot. syzbot can test patches for this issue, for details see:
https://goo.gl/tpsmEJ#testing-patches
PATCH 1/1 nilfs: check filetype/uid/gid at nilfs_read_inode_common()
In reply to a message not in the archivesyzbot is reporting lockdep warning followed by NULL pointer dereference at nilfs_bmap_lookup_at_level() [1], for a crafted filesystem which contains raw_inode->i_mode == 0 is poisoning checkpoint inode at nilfs_read_inode_common() from nilfs_ifile_read() from nilfs_attach_checkpoint() from nilfs_fill_super() from nilfs_mount(). Check that filetype/uid/gid are valid as well as i_nlink is valid.
---
This patch solves crash but I don't know whether this patch is correct.
Commit 05fe58fdc10df9eb ("nilfs2: inode operations") did
+ if (inode->i_nlink == 0 && (inode->i_mode == 0 || ii->i_dtime))
and commit 612392307cb09e49 ("nilfs2: support nanosecond timestamp") did
- if (inode->i_nlink == 0 && (inode->i_mode == 0 || ii->i_dtime))
+ if (inode->i_nlink == 0 && inode->i_mode == 0)
and commit 705304a863cc4158 ("nilfs2: fix the nilfs_iget() vs. nilfs_new_inode() races") did
- if (inode->i_nlink == 0 && inode->i_mode == 0)
+ if (inode->i_nlink == 0)
but I don't know which commit was wrong. Is inode->i_mode == 0 valid in nilfs?
fs/nilfs2/inode.c | 14 ++++++++++++++ 1 file changed, 14 insertions(+)
fs/nilfs2/inode.c +14 -0
@@ -442,6 +442,20 @@ int nilfs_read_inode_common(struct inode *inode, inode->i_mtime.tv_nsec =le32_to_cpu (raw_inode->i_mtime_nsec);if (inode->i_nlink ==0 )return -ESTALE ;/* this inode is deleted */ +switch (inode->i_mode &S_IFMT ) { +case S_IFLNK : +case S_IFREG : +case S_IFDIR : +case S_IFCHR : +case S_IFBLK : +case S_IFIFO : +case S_IFSOCK : +break ; +default : +return -EINVAL ;/* Filetype is invalid */ + } +if (!uid_valid (inode->i_uid) || !gid_valid (inode->i_gid)) +return -EINVAL ;/* UID/GID is invalid */ inode->i_blocks =le64_to_cpu (raw_inode->i_blocks); ii->i_flags =le32_to_cpu (raw_inode->i_flags);
Signature · 1 line
2.34.1
Re: [PATCH] nilfs: check filetype/uid/gid at nilfs_read_inode_common()
In reply to a message not in the archiveQuoted 7 lines · Tetsuo Handa
syzbot is reporting lockdep warning followed by NULL pointer dereference at
nilfs_bmap_lookup_at_level()[1], for a crafted filesystem which containsraw_inode->i_mode== 0 is poisoning checkpoint inode atnilfs_read_inode_common()fromnilfs_ifile_read()fromnilfs_attach_checkpoint()fromnilfs_fill_super()fromnilfs_mount(). Check that filetype/uid/gid are valid as well asi_nlinkis valid.
Quoted 5 lines
Reported-by syzot <syzbot+2b32eb36c1a825b7a74c-Pl5Pbv+GP7P466ipTTIvnc23WoclnBCfAL8bYrjMMd8@public.gmane.org>Signed-off-by Tetsuo Handa <penguin-kernel-JPay3/[email protected]>---
This patch solves crash but I don't know whether this patch is correct.
Hmm, since bmap->b_sem and bmap->b_ops are initialized by nilfs_bmap_read(), and nilfs_bmap_read() is called from nilfs_read_inode_common() only if S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) || S_ISLNK(inode->i_mode) is true, I guess that the caller needs to verify that the returned inode is a regular file? But where to add that check?
Re: [PATCH] nilfs: check filetype/uid/gid at nilfs_read_inode_common()
In reply to a message not in the archiveQuoted 15 lines · Tetsuo Handa
Quoted 7 lines · Tetsuo Handa
syzbot is reporting lockdep warning followed by NULL pointer dereference at
nilfs_bmap_lookup_at_level()[1], for a crafted filesystem which containsraw_inode->i_mode== 0 is poisoning checkpoint inode atnilfs_read_inode_common()fromnilfs_ifile_read()fromnilfs_attach_checkpoint()fromnilfs_fill_super()fromnilfs_mount(). Check that filetype/uid/gid are valid as well asi_nlinkis valid.Quoted 5 lines
Reported-by syzot <syzbot+2b32eb36c1a825b7a74c-Pl5Pbv+GP7P466ipTTIvnc23WoclnBCfAL8bYrjMMd8@public.gmane.org>Signed-off-by Tetsuo Handa <penguin-kernel-JPay3/[email protected]>---
This patch solves crash but I don't know whether this patch is correct.
Quoted 5 lines
Hmm, since
bmap->b_semandbmap->b_opsare initialized bynilfs_bmap_read(), andnilfs_bmap_read()is called fromnilfs_read_inode_common()only ifS_ISREG(inode->i_mode)||S_ISDIR(inode->i_mode)||S_ISLNK(inode->i_mode)is true, I guess that the caller needs to verify that the returned inode is a regular file? But where to add that check?
Exactly, this looks like the root cause of the issue.
For metadata files, which are internal persistent files of nilfs2 and used to store metadata
on disk, must have a file type and must call nilfs_bmap_read(), but that sanity check
is missing.
As you pointed out, we need to add a sanity check for i_mode of metadata files.
The simple way is to insert a conditional sanity check in nilfs_read_inode_common()
using nilfs_is_metadata_file_inode() and S_ISREG(), for instance:
if (nilfs_is_metadata_file_inode(inode) && !S_ISREG(inode->i_mode))
return -EIO; /* On-disk inode of the metadata file is
corrupted. */
if (inode->i_nlink == 0)
return -ESTALE; /* ... */Or, we can add the i_mode sanity checks separately in nilfs_dat_read(), nilfs_ifile_read(), nilfs_sufile_read(), and nilfs_cpfile_read() before they call nilfs_read_inode_common().
I think the former is fine because we don't have the idea to add metadata file inodes other
than file type. If you make a revised patch, I will approve or
handle it. Or if you are busy,
I will take over the fix.
Thank you,
Ryusuke Konishi
Re: [PATCH] nilfs: check filetype/uid/gid at nilfs_read_inode_common()
In reply to a message not in the archiveQuoted 5 lines · Ryusuke Konishi
I think the former is fine because we don't have the idea to add metadata file inodes other
than file type. If you make a revised patch, I will approve or
handle it. Or if you are busy,
I will take over the fix.
Please take over the fix. Thank you.
Re: [PATCH] nilfs: check filetype/uid/gid at nilfs_read_inode_common()
In reply to a message not in the archiveQuoted 9 lines · Tetsuo Handa
Quoted 5 lines · Ryusuke Konishi
I think the former is fine because we don't have the idea to add metadata file inodes other
than file type. If you make a revised patch, I will approve or
handle it. Or if you are busy,
I will take over the fix.Please take over the fix. Thank you.
All right, thank you for your help so far!
Thanks,
Ryusuke Konishi
Ryusuke Konishi
maintainerPATCH 1/1 nilfs2: fix NULL pointer dereference at nilfs_bmap_lookup_at_level()
In reply to syzbotIf the i_mode field in inode of metadata files is corrupted on disk, it can cause the initialization of bmap structure, which should have been called from nilfs_read_inode_common(), not to be called. This causes a lockdep warning followed by a NULL pointer dereference at nilfs_bmap_lookup_at_level().
This patch fixes these issues by adding a missing sanitiy check for the i_mode field of metadata file's inode.
--- fs/nilfs2/inode.c | 2 ++ 1 file changed, 2 insertions(+)
fs/nilfs2/inode.c +2 -0
@@ -440,6 +440,8 @@ int nilfs_read_inode_common(struct inode *inode, inode->i_atime.tv_nsec =le32_to_cpu (raw_inode->i_mtime_nsec); inode->i_ctime.tv_nsec =le32_to_cpu (raw_inode->i_ctime_nsec); inode->i_mtime.tv_nsec =le32_to_cpu (raw_inode->i_mtime_nsec); +if (nilfs_is_metadata_file_inode (inode) && !S_ISREG (inode->i_mode)) +return -EIO ;/* this inode is for metadata and corrupted */ if (inode->i_nlink ==0 )return -ESTALE ;/* this inode is deleted */
Signature · 1 line
2.34.1