jbd2: fix off-by-one while erasing journal

In __jbd2_journal_erase(), the block_stop parameter includes the last
block of a contiguous region; however, the calculation of byte_stop is
incorrect, as it does not account for the bytes in that last block.
Consequently, the page cache is not cleared properly, which occasionally
causes the ext4/050 test to fail.

Since block_stop operates on inclusion semantics, it involves repeated
increments and decrements by 1, significantly increasing the complexity
of the calculations. Optimize the calculation and fix the incorrect
byte_stop by make both block_stop and byte_stop to use exclusion
semantics.

This fixes a failure in fstests ext4/050.

Fixes: 01d5d96542fd ("ext4: add discard/zeroout flags to journal flush")
Signed-off-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Baokun Li <libaokun1@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Link: https://patch.msgid.link/20250217065955.3829229-1-yi.zhang@huaweicloud.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>

diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
index ff41a31e2bc9f03d9448a4e9542b4d0f2bb8a808..445f2bddd17cbc1d9514637771062447a0a1ce1b 100644 (file)
--- a/fs/jbd2/journal.c
+++ b/fs/jbd2/journal.c
@@ -1964,17 +1964,15 @@ static int __jbd2_journal_erase(journal_t *journal, unsigned int flags)
                        return err;
                }
 
-               if (block_start == ~0ULL) {
-                       block_start = phys_block;
-                       block_stop = block_start - 1;
-               }
+               if (block_start == ~0ULL)
+                       block_stop = block_start = phys_block;
 
                /*
                 * last block not contiguous with current block,
                 * process last contiguous region and return to this block on
                 * next loop
                 */
-               if (phys_block != block_stop + 1) {
+               if (phys_block != block_stop) {
                        block--;
                } else {
                        block_stop++;
@@ -1993,11 +1991,10 @@ static int __jbd2_journal_erase(journal_t *journal, unsigned int flags)
                 */
                byte_start = block_start * journal->j_blocksize;
                byte_stop = block_stop * journal->j_blocksize;
-               byte_count = (block_stop - block_start + 1) *
-                               journal->j_blocksize;
+               byte_count = (block_stop - block_start) * journal->j_blocksize;
 
                truncate_inode_pages_range(journal->j_dev->bd_mapping,
-                               byte_start, byte_stop);
+                               byte_start, byte_stop - 1);
 
                if (flags & JBD2_JOURNAL_FLUSH_DISCARD) {
                        err = blkdev_issue_discard(journal->j_dev,
@@ -2012,7 +2009,7 @@ static int __jbd2_journal_erase(journal_t *journal, unsigned int flags)
                }
 
                if (unlikely(err != 0)) {
-                       pr_err("JBD2: (error %d) unable to wipe journal at physical blocks %llu - %llu",
+                       pr_err("JBD2: (error %d) unable to wipe journal at physical blocks [%llu, %llu)",
                                        err, block_start, block_stop);
                        return err;
                }