From: Eric Biggers ebiggers@google.com
Due to several different off-by-one errors, or perhaps due to a late change in design that wasn't fully reflected in the code that was actually merged, there are several very strange constraints on how fast-commit blocks are filled with tlv entries:
- tlvs must start at least 10 bytes before the end of the block, even though the minimum tlv length is 8. Otherwise, the replay code will ignore them. (BUG: ext4_fc_reserve_space() could violate this requirement if called with a len of blocksize - 9 or blocksize - 8. Fortunately, this doesn't seem to happen currently.)
- tlvs must end at least 1 byte before the end of the block. Otherwise the replay code will consider them to be invalid. This quirk contributed to a bug (fixed by an earlier commit) where uninitialized memory was being leaked to disk in the last byte of blocks.
Also, strangely these constraints don't apply to the replay code in e2fsprogs, which will accept any tlvs in the blocks (with no bounds checks at all, but that is a separate issue...).
Given that this all seems to be a bug, let's fix it by just filling blocks with tlv entries in the natural way.
Note that old kernels will be unable to replay fast-commit journals created by kernels that have this commit.
Fixes: aa75f4d3daae ("ext4: main fast-commit commit path") Cc: stable@vger.kernel.org # v5.10+ Signed-off-by: Eric Biggers ebiggers@google.com --- fs/ext4/fast_commit.c | 66 +++++++++++++++++++++---------------------- 1 file changed, 33 insertions(+), 33 deletions(-)
diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c index 892fa7c7a768b..7ed71c652f67f 100644 --- a/fs/ext4/fast_commit.c +++ b/fs/ext4/fast_commit.c @@ -714,43 +714,43 @@ static u8 *ext4_fc_reserve_space(struct super_block *sb, int len, u32 *crc) struct buffer_head *bh; int bsize = sbi->s_journal->j_blocksize; int ret, off = sbi->s_fc_bytes % bsize; - int pad_len; + int remaining; u8 *dst;
/* - * After allocating len, we should have space at least for a 0 byte - * padding. + * If 'len' is too long to fit in any block alongside a PAD tlv, then we + * cannot fulfill the request. */ - if (len + EXT4_FC_TAG_BASE_LEN > bsize) + if (len > bsize - EXT4_FC_TAG_BASE_LEN) return NULL;
- if (bsize - off - 1 > len + EXT4_FC_TAG_BASE_LEN) { - /* - * Only allocate from current buffer if we have enough space for - * this request AND we have space to add a zero byte padding. - */ - if (!sbi->s_fc_bh) { - ret = jbd2_fc_get_buf(EXT4_SB(sb)->s_journal, &bh); - if (ret) - return NULL; - sbi->s_fc_bh = bh; - } - sbi->s_fc_bytes += len; - return sbi->s_fc_bh->b_data + off; + if (!sbi->s_fc_bh) { + ret = jbd2_fc_get_buf(EXT4_SB(sb)->s_journal, &bh); + if (ret) + return NULL; + sbi->s_fc_bh = bh; } - /* Need to add PAD tag */ dst = sbi->s_fc_bh->b_data + off; + + /* + * Allocate the bytes in the current block if we can do so while still + * leaving enough space for a PAD tlv. + */ + remaining = bsize - EXT4_FC_TAG_BASE_LEN - off; + if (len <= remaining) { + sbi->s_fc_bytes += len; + return dst; + } + + /* + * Else, terminate the current block with a PAD tlv, then allocate a new + * block and allocate the bytes at the start of that new block. + */ + tl.fc_tag = cpu_to_le16(EXT4_FC_TAG_PAD); - pad_len = bsize - off - 1 - EXT4_FC_TAG_BASE_LEN; - tl.fc_len = cpu_to_le16(pad_len); + tl.fc_len = cpu_to_le16(remaining); ext4_fc_memcpy(sb, dst, &tl, EXT4_FC_TAG_BASE_LEN, crc); - dst += EXT4_FC_TAG_BASE_LEN; - if (pad_len > 0) { - ext4_fc_memzero(sb, dst, pad_len, crc); - dst += pad_len; - } - /* Don't leak uninitialized memory in the unused last byte. */ - *dst = 0; + ext4_fc_memzero(sb, dst + EXT4_FC_TAG_BASE_LEN, remaining, crc);
ext4_fc_submit_bh(sb, false);
@@ -758,7 +758,7 @@ static u8 *ext4_fc_reserve_space(struct super_block *sb, int len, u32 *crc) if (ret) return NULL; sbi->s_fc_bh = bh; - sbi->s_fc_bytes = (sbi->s_fc_bytes / bsize + 1) * bsize + len; + sbi->s_fc_bytes += bsize - off + len; return sbi->s_fc_bh->b_data; }
@@ -789,7 +789,7 @@ static int ext4_fc_write_tail(struct super_block *sb, u32 crc) off = sbi->s_fc_bytes % bsize;
tl.fc_tag = cpu_to_le16(EXT4_FC_TAG_TAIL); - tl.fc_len = cpu_to_le16(bsize - off - 1 + sizeof(struct ext4_fc_tail)); + tl.fc_len = cpu_to_le16(bsize - off + sizeof(struct ext4_fc_tail)); sbi->s_fc_bytes = round_up(sbi->s_fc_bytes, bsize);
ext4_fc_memcpy(sb, dst, &tl, EXT4_FC_TAG_BASE_LEN, &crc); @@ -2056,7 +2056,7 @@ static int ext4_fc_replay_scan(journal_t *journal, state = &sbi->s_fc_replay_state;
start = (u8 *)bh->b_data; - end = (__u8 *)bh->b_data + journal->j_blocksize - 1; + end = start + journal->j_blocksize;
if (state->fc_replay_expected_off == 0) { state->fc_cur_tag = 0; @@ -2077,7 +2077,7 @@ static int ext4_fc_replay_scan(journal_t *journal, }
state->fc_replay_expected_off++; - for (cur = start; cur < end - EXT4_FC_TAG_BASE_LEN; + for (cur = start; cur <= end - EXT4_FC_TAG_BASE_LEN; cur = cur + EXT4_FC_TAG_BASE_LEN + tl.fc_len) { ext4_fc_get_tl(&tl, cur); val = cur + EXT4_FC_TAG_BASE_LEN; @@ -2195,9 +2195,9 @@ static int ext4_fc_replay(journal_t *journal, struct buffer_head *bh, #endif
start = (u8 *)bh->b_data; - end = (__u8 *)bh->b_data + journal->j_blocksize - 1; + end = start + journal->j_blocksize;
- for (cur = start; cur < end - EXT4_FC_TAG_BASE_LEN; + for (cur = start; cur <= end - EXT4_FC_TAG_BASE_LEN; cur = cur + EXT4_FC_TAG_BASE_LEN + tl.fc_len) { ext4_fc_get_tl(&tl, cur); val = cur + EXT4_FC_TAG_BASE_LEN;