[PATCH v3 0/2] bfs: fix null-ptr-deref and possible warning in bfs_move_block() func

[PATCH v3 0/2] bfs: fix null-ptr-deref and possible warning in bfs_move_block() func

[kovalev]

https://syzkaller.appspot.com/bug?extid=d98fd19acd08b36ff422

[PATCH v3 1/2] bfs: prevent null pointer dereference in bfs_move_block()
v3: Changed the error handling

[PATCH v3 2/2] bfs: ensure buffer is marked uptodate before marking it dirty
v3: Replaced the buffer up-to-date check with an error exit by forcefully
setting the buffer as up-to-date before call mark_buffer_dirty()

[PATCH v3 1/2] bfs: prevent null pointer dereference in bfs_move_block()

[kovalev]

From: Vasiliy Kovalev <kovalev@altlinux.org>

Detect a failed sb_getblk() call (before copying data)
so that null pointer dereferences should not happen any more.
We also decrement (brelse) the bh counter in this case.

Found when launching the reproducer generated by Syzkaller:

KASAN: null-ptr-deref in range [0x0000000000000028-0x000000000000002f]
Comm: mark_buffer_dir Tainted: G  W  6.1.105-un-def-alt1.kasan #1
RIP: 0010:bfs_get_block+0x35b/0xdf0 (fs/bfs/file.c:42) [bfs]
Call Trace:
<TASK>
 __die_body.cold (arch/x86/kernel/dumpstack.c:478)
 die_addr (arch/x86/kernel/dumpstack.c:462)
 exc_general_protection (arch/x86/kernel/traps.c:787)
 asm_exc_general_protection (./arch/x86/include/asm/idtentry.h:564)
 __getblk_gfp (fs/buffer.c:1335)
 bfs_get_block (fs/bfs/file.c:42) bfs
 bfs_get_block (fs/bfs/file.c:56) bfs
 bfs_get_block (fs/bfs/file.c:125) bfs
 bfs_write_begin (fs/bfs/file.c:66) bfs
 __block_write_begin_int (fs/buffer.c:1991)
 bfs_write_begin (fs/bfs/file.c:66) bfs
 invalidate_bh_lrus_cpu (fs/buffer.c:1955)
 fault_in_readable (mm/gup.c:1898)
 PageHeadHuge (mm/hugetlb.c:2123)
 bfs_write_begin (fs/bfs/file.c:66) bfs
 block_write_begin (fs/buffer.c:2103)
 bfs_write_begin (fs/bfs/file.c:178) bfs
 generic_perform_write (mm/filemap.c:3817)
 generic_file_readonly_mmap (mm/filemap.c:3781)
 new_inode (fs/inode.c:2126)
 generic_write_checks (fs/read_write.c:1687)
 __generic_file_write_iter (mm/filemap.c:3946)
 generic_file_write_iter (./include/linux/fs.h:763 mm/filemap.c:3978)
 vfs_write (./include/linux/fs.h:2265 fs/read_write.c:491)
 kernel_write (fs/read_write.c:565)
 __fget_files (fs/file.c:918)
 ksys_write (fs/read_write.c:638)

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Reported-by: syzbot+d98fd19acd08b36ff422@syzkaller.appspotmail.com
Link: https://syzkaller.appspot.com/bug?extid=d98fd19acd08b36ff422
Link: https://syzkaller.appspot.com/text?tag=ReproC&x=16515ba3e80000
Cc: stable@vger.kernel.org
Signed-off-by: Vasiliy Kovalev <kovalev@altlinux.org>
---
 fs/bfs/file.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/fs/bfs/file.c b/fs/bfs/file.c
index 57ae5ee6deec12..23773e62994024 100644
--- a/fs/bfs/file.c
+++ b/fs/bfs/file.c
@@ -39,6 +39,10 @@ static int bfs_move_block(unsigned long from, unsigned long to,
 	if (!bh)
 		return -EIO;
 	new = sb_getblk(sb, to);
+	if (!new) {
+		brelse(bh);
+		return -EIO;
+	}
 	memcpy(new->b_data, bh->b_data, bh->b_size);
 	mark_buffer_dirty(new);
 	bforget(bh);
-- 
2.33.8

[PATCH v3 2/2] bfs: ensure buffer is marked uptodate before marking it dirty

[kovalev]

From: Vasiliy Kovalev <kovalev@altlinux.org>

Add a call to `set_buffer_uptodate(new)` in `bfs_move_block` before
marking the buffer as dirty. This change ensures the buffer is correctly
flagged as containing valid data after copying, preventing potential data
inconsistencies or corruption during writeback.

Found by Syzkaller:

WARNING: CPU: 1 PID: 5055 at fs/buffer.c:1176 mark_buffer_dirty+0x37b/0x3f0 fs/buffer.c:1176
Modules linked in:
CPU: 1 PID: 5055 Comm: syz-executor162 Not tainted 6.8.0-rc1-syzkaller-00049-g6098d87eaf31 #0
Hardware name: Google Compute Engine/Google Compute Engine, BIOS Google 11/17/2023
RIP: 0010:mark_buffer_dirty+0x37b/0x3f0 fs/buffer.c:1176
Call Trace:

 bfs_move_block fs/bfs/file.c:44 [inline]
 bfs_move_blocks fs/bfs/file.c:57 [inline]
 bfs_get_block+0x3e5/0xeb0 fs/bfs/file.c:126
 __block_write_begin_int+0x4fb/0x16e0 fs/buffer.c:2103
 __block_write_begin fs/buffer.c:2152 [inline]
 block_write_begin+0xb1/0x490 fs/buffer.c:2211
 bfs_write_begin+0x31/0xd0 fs/bfs/file.c:179
 generic_perform_write+0x278/0x600 mm/filemap.c:3930
 __generic_file_write_iter+0x1f9/0x240 mm/filemap.c:4025
 generic_file_write_iter+0xe3/0x350 mm/filemap.c:4051
 call_write_iter include/linux/fs.h:2085 [inline]
 new_sync_write fs/read_write.c:497 [inline]
 vfs_write+0x6d5/0x1100 fs/read_write.c:590
 ksys_write+0x12f/0x250 fs/read_write.c:643
 do_syscall_x64 arch/x86/entry/common.c:52 [inline]
 do_syscall_64+0xd3/0x250 arch/x86/entry/common.c:83
 entry_SYSCALL_64_after_hwframe+0x63/0x6b

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Reported-by: syzbot+d98fd19acd08b36ff422@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=d98fd19acd08b36ff422
Cc: stable@vger.kernel.org
Signed-off-by: Vasiliy Kovalev <kovalev@altlinux.org>
---
 fs/bfs/file.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/fs/bfs/file.c b/fs/bfs/file.c
index 23773e62994024..3f0c506584560e 100644
--- a/fs/bfs/file.c
+++ b/fs/bfs/file.c
@@ -44,6 +44,7 @@ static int bfs_move_block(unsigned long from, unsigned long to,
 		return -EIO;
 	}
 	memcpy(new->b_data, bh->b_data, bh->b_size);
+	set_buffer_uptodate(new);
 	mark_buffer_dirty(new);
 	bforget(bh);
 	brelse(new);
-- 
2.33.8