From: Tetsuo Handa <penguin-kernel@xxxxxxxxxxxxxxxxxxx> [ Upstream commit c7c6363ca186747ebc2df10c8a1a51e66e0e32d9 ] When the volume header contains erroneous values that do not reflect the actual state of the filesystem, hfsplus_fill_super() assumes that the attributes file is not yet created, which later results in hitting BUG_ON() when hfsplus_create_attributes_file() is called. Replace this BUG_ON() with -EIO error with a message to suggest running fsck tool. Reported-by: syzbot <syzbot+1107451c16b9eb9d29e6@xxxxxxxxxxxxxxxxxxxxxxxxx> Closes: https://syzkaller.appspot.com/bug?extid=1107451c16b9eb9d29e6 Signed-off-by: Tetsuo Handa <penguin-kernel@xxxxxxxxxxxxxxxxxxx> Reviewed-by: Viacheslav Dubeyko <slava@xxxxxxxxxxx> Link: https://lore.kernel.org/r/7b587d24-c8a1-4413-9b9a-00a33fbd849f@xxxxxxxxxxxxxxxxxxx Signed-off-by: Viacheslav Dubeyko <slava@xxxxxxxxxxx> Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx> --- LLM Generated explanations, may be completely bogus: **Backport Status: YES** This commit should be backported to stable kernel trees for the following reasons: 1. **Fixes a Critical Bug**: The commit replaces a `BUG_ON()` that could crash the entire system with proper error handling. The original code at line 175 had `BUG_ON(i_size_read(attr_file) != 0);` which would trigger a kernel panic when the HFS+ filesystem metadata is corrupted or inconsistent. 2. **Real-World Issue**: This was reported by syzbot (a kernel fuzzer), indicating it's a reachable code path that can be triggered with malformed filesystem data. The bug report link confirms this is a reproducible issue. 3. **Simple and Contained Fix**: The change is minimal - it replaces the BUG_ON with: ```c if (i_size_read(attr_file) != 0) { err = -EIO; pr_err("detected inconsistent attributes file, running fsck.hfsplus is recommended.\n"); goto end_attr_file_creation; } ``` This is a straightforward conversion that maintains the same logic but handles the error gracefully. 4. **Prevents System Crashes**: BUG_ON() causes a kernel panic, which is particularly severe. Converting it to return -EIO allows the system to continue running and provides users with actionable guidance (run fsck.hfsplus). 5. **Low Risk of Regression**: The change only affects error handling when filesystem corruption is detected. It doesn't alter normal operation paths and follows established error handling patterns in the kernel. 6. **Follows Stable Tree Rules**: This is a classic example of a stable- worthy fix: - Fixes a real bug (system crash on corrupted filesystem) - Small, easily reviewable change - No new features or architectural changes - High benefit (prevents crashes) with minimal risk The commit message clearly indicates this addresses filesystem corruption scenarios where "the volume header contains erroneous values that do not reflect the actual state of the filesystem," making it an important robustness improvement for stable kernels. fs/hfsplus/xattr.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/fs/hfsplus/xattr.c b/fs/hfsplus/xattr.c index d91f76ef18d9..2438cd759620 100644 --- a/fs/hfsplus/xattr.c +++ b/fs/hfsplus/xattr.c @@ -172,7 +172,11 @@ static int hfsplus_create_attributes_file(struct super_block *sb) return PTR_ERR(attr_file); } - BUG_ON(i_size_read(attr_file) != 0); + if (i_size_read(attr_file) != 0) { + err = -EIO; + pr_err("detected inconsistent attributes file, running fsck.hfsplus is recommended.\n"); + goto end_attr_file_creation; + } hip = HFSPLUS_I(attr_file); -- 2.39.5
