From: Chao Yu chao@kernel.org
[ Upstream commit 1773f63d108b1b9b9d053d8c95f8300c556f93b8 ]
F2FS-fs (dm-55): access invalid blkaddr:972878540 Call trace: dump_backtrace+0xec/0x128 show_stack+0x18/0x28 dump_stack_lvl+0x40/0x88 dump_stack+0x18/0x24 __f2fs_is_valid_blkaddr+0x360/0x3b4 f2fs_is_valid_blkaddr+0x10/0x20 f2fs_get_node_info+0x21c/0x60c __write_node_page+0x15c/0x734 f2fs_sync_node_pages+0x4f8/0x700 f2fs_write_checkpoint+0x4a8/0x99c __checkpoint_and_complete_reqs+0x7c/0x20c issue_checkpoint_thread+0x4c/0xd8 kthread+0x11c/0x1b0 ret_from_fork+0x10/0x20
If nat.blkaddr is corrupted, during checkpoint, f2fs_sync_node_pages() will loop to flush node page w/ corrupted nat.blkaddr.
Although, it tags SBI_NEED_FSCK, checkpoint can not persist it due to deadloop.
Let's call f2fs_handle_error(, ERROR_INCONSISTENT_NAT) to record such error into superblock, it expects fsck can detect the error and repair inconsistent nat.blkaddr after device reboot.
Note that, let's add sanity check in f2fs_get_node_info() to detect in-memory nat.blkaddr inconsistency, but only if CONFIG_F2FS_CHECK_FS is enabled.
Signed-off-by: Chao Yu chao@kernel.org Signed-off-by: Jaegeuk Kim jaegeuk@kernel.org Signed-off-by: Sasha Levin sashal@kernel.org ---
LLM Generated explanations, may be completely bogus:
Based on my analysis of the commit, here is my assessment:
**Backport Status: YES**
This commit should be backported to stable kernel trees for the following reasons:
1. **Critical Bug Fix**: This fixes a **deadloop/infinite loop condition** that can occur during checkpoint operations when NAT (Node Address Table) entries have corrupted block addresses. The commit message clearly shows a real-world stack trace where `f2fs_sync_node_pages()` enters an infinite loop trying to flush nodes with corrupted nat.blkaddr during checkpoint.
2. **Data Integrity and System Availability**: The deadloop prevents the checkpoint from completing, which means: - The filesystem cannot persist the `SBI_NEED_FSCK` flag that would trigger repair on next mount - The system may hang or become unresponsive during checkpoint operations - Filesystem metadata corruption cannot be properly handled
3. **Small and Contained Fix**: The patch is relatively small (21 insertions, 8 deletions) and focused on a single function `f2fs_get_node_info()`. The changes are: - Move the sanity check earlier in the code path - Add proper error handling with `f2fs_handle_error()` and `ERROR_INCONSISTENT_NAT` - Return `-EFSCORRUPTED` instead of `-EFAULT` for better error semantics - Add conditional caching logic to avoid caching corrupted entries
4. **Low Risk of Regression**: - The main sanity check logic already existed but was incomplete - The new error handling path only triggers when corruption is detected - The CONFIG_F2FS_CHECK_FS conditional check adds extra validation only when debugging is enabled - The change doesn't alter normal operation flow for valid NAT entries
5. **Follows Stable Rules**: This fix meets stable kernel criteria: - Fixes a real bug that users have hit (stack trace provided) - The fix is minimal and targeted - No new features are introduced - The risk of regression is low
6. **Corruption Handling**: The commit properly handles filesystem corruption by: - Setting the `SBI_NEED_FSCK` flag - Recording the error in the superblock via `f2fs_handle_error()` - Providing detailed error logging for debugging - Returning appropriate error codes to prevent further damage
The deadloop condition this fixes is particularly severe as it can make the system unresponsive and prevent proper error recovery, making this an important candidate for stable backporting.
fs/f2fs/node.c | 29 +++++++++++++++++++++-------- 1 file changed, 21 insertions(+), 8 deletions(-)
diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c index bfe104db284e..2fd287f2bca4 100644 --- a/fs/f2fs/node.c +++ b/fs/f2fs/node.c @@ -555,8 +555,8 @@ int f2fs_get_node_info(struct f2fs_sb_info *sbi, nid_t nid, struct f2fs_nat_entry ne; struct nat_entry *e; pgoff_t index; - block_t blkaddr; int i; + bool need_cache = true;
ni->flag = 0; ni->nid = nid; @@ -569,6 +569,10 @@ int f2fs_get_node_info(struct f2fs_sb_info *sbi, nid_t nid, ni->blk_addr = nat_get_blkaddr(e); ni->version = nat_get_version(e); f2fs_up_read(&nm_i->nat_tree_lock); + if (IS_ENABLED(CONFIG_F2FS_CHECK_FS)) { + need_cache = false; + goto sanity_check; + } return 0; }
@@ -594,7 +598,7 @@ int f2fs_get_node_info(struct f2fs_sb_info *sbi, nid_t nid, up_read(&curseg->journal_rwsem); if (i >= 0) { f2fs_up_read(&nm_i->nat_tree_lock); - goto cache; + goto sanity_check; }
/* Fill node_info from nat page */ @@ -609,14 +613,23 @@ int f2fs_get_node_info(struct f2fs_sb_info *sbi, nid_t nid, ne = nat_blk->entries[nid - start_nid]; node_info_from_raw_nat(ni, &ne); f2fs_folio_put(folio, true); -cache: - blkaddr = le32_to_cpu(ne.block_addr); - if (__is_valid_data_blkaddr(blkaddr) && - !f2fs_is_valid_blkaddr(sbi, blkaddr, DATA_GENERIC_ENHANCE)) - return -EFAULT; +sanity_check: + if (__is_valid_data_blkaddr(ni->blk_addr) && + !f2fs_is_valid_blkaddr(sbi, ni->blk_addr, + DATA_GENERIC_ENHANCE)) { + set_sbi_flag(sbi, SBI_NEED_FSCK); + f2fs_err_ratelimited(sbi, + "f2fs_get_node_info of %pS: inconsistent nat entry, " + "ino:%u, nid:%u, blkaddr:%u, ver:%u, flag:%u", + __builtin_return_address(0), + ni->ino, ni->nid, ni->blk_addr, ni->version, ni->flag); + f2fs_handle_error(sbi, ERROR_INCONSISTENT_NAT); + return -EFSCORRUPTED; + }
/* cache nat entry */ - cache_nat_entry(sbi, nid, &ne); + if (need_cache) + cache_nat_entry(sbi, nid, &ne); return 0; }