Skip to contentSkip to conversation

jbd2: avoid transaction reuse after reformatting

PATCH New 常凤楠 4 participants 10 messages Sep 10, 2020 – Oct 3, 2020 1 message not in the archive Journalling layer for block devices (JBD2)

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

常凤楠

author
#1 lore raw

PATCH 1/1 jbd2: avoid transaction reuse after reformatting

When format ext4 with lazy_journal_init=1, the previous transaction is still on disk, it is possible that the previous transaction will be used again during jbd2 recovery. Because the seed is changed, the CRC check will fail.

---
 fs/jbd2/recovery.c | 60 +++++++++++++++++++++++++++++++++++++++++-----
 1 file changed, 54 insertions(+), 6 deletions(-)
fs/jbd2/recovery.c +2 -0
@@ -33,6 +33,8 @@ struct recovery_info
 intnr_replays;
 intnr_revokes;
 intnr_revoke_hits;
+unsigned long  ri_commit_block;
+__be64  last_trans_commit_time;
 };
 
 enum passtype {PASS_SCAN, PASS_REVOKE, PASS_REPLAY}; @@ -412,7 +414,27 @@ static int jbd2_block_tag_csum_verify(journal_t *j, journal_block_tag_t *tag,

else
return tag->t_checksum == cpu_to_be16(csum32); }
+/*
+ * We check the commit time and compare it with the commit time of
+ * the previous transaction, if it's smaller than previous,
+ * We think it's not belong to same journal.
+ */
+static bool is_same_journal(journal_t *journal, struct buffer_head *bh, +unsigned long blocknr, __u64 last_commit_sec) {
+unsigned long commit_block = blocknr + count_tags(journal, bh) + 1; +struct buffer_head *nbh;
+struct commit_header *cbh;
+__u64commit_sec;
+int err = jread(&nbh, journal, commit_block);

+if (err)
+return true;
+
+cbh = (struct commit_header *)nbh->b_data;
+commit_sec = be64_to_cpu(cbh->h_commit_sec);
+
+return commit_sec >= last_commit_sec;
+}
static int do_one_pass(journal_t *journal,
struct recovery_info *info, enum passtype pass) { @@ -514,18 +536,29 @@ static int do_one_pass(journal_t *journal, switch(blocktype) {
case JBD2_DESCRIPTOR_BLOCK:
/* Verify checksum first */
+if (pass == PASS_SCAN)
+info->ri_commit_block = 0;
+
if (jbd2_journal_has_csum_v2or3(journal))
descr_csum_size =
sizeof(struct jbd2_journal_block_tail);
if (descr_csum_size > 0 &&

     !jbd2_descriptor_block_csum_verify(journal,
        bh->b_data)) {

-printk(KERN_ERR "JBD2: Invalid checksum "
- "recovering block %lu in log\n",
+if (is_same_journal(journal, bh, next_log_block-1, info->last_trans_commit_time)) { +printk(KERN_ERR "JBD2: Invalid checksum recovering block %lu in +log\n",

        next_log_block);

-err = -EFSBADCRC;
-brelse(bh);
-goto failed;
+err = -EFSBADCRC;
+brelse(bh);
+goto failed;
+} else {
+/*it's not belong to same journal, just end this recovery with success*/ +jbd_debug(1, "JBD2: Invalid checksum found in block %lu in log, but not same journal %d\n",
+ next_log_block, next_commit_ID);
+err = 0;
+brelse(bh);
+goto done;
+}
}

/* If it is a valid descriptor block, replay it @@ -688,6 +721,17 @@ static int do_one_pass(journal_t *journal,
* are present verify them in PASS_SCAN; else not
* much to do other than move on to the next sequence
* number. */
+if (pass == PASS_SCAN) {
+struct commit_header *cbh =
+(struct commit_header *)bh->b_data;
+if (info->ri_commit_block) {
+jbd_debug(1, "invalid commit block found in %lu, stop here.\n", next_log_block);
+brelse(bh);
+goto done;
+}
+info->ri_commit_block = next_log_block;
+info->last_trans_commit_time = be64_to_cpu(cbh->h_commit_sec);
+}
if (pass == PASS_SCAN &&

     jbd2_has_feature_checksum(journal)) {

int chksum_err, chksum_seen;
@@ -761,7 +805,11 @@ static int do_one_pass(journal_t *journal,
brelse(bh);
continue;
}
-
+if (pass != PASS_SCAN && info->ri_commit_block) {
+jbd_debug(1, "invalid revoke block found in %lu, stop here.\n", next_log_block); +brelse(bh);
+goto done;
+}
err = scan_revoke_records(journal, bh,
next_commit_ID, info);
brelse(bh);
--
2.27.0.windows.1

________________________________

CONFIDENTIALITY NOTICE: This electronic message is intended to be viewed only by the individual or entity to whom it is addressed. It may contain information that is privileged, confidential and exempt from disclosure under applicable law. Any dissemination, distribution or copying of this communication is strictly prohibited without our prior permission. If the reader of this message is not the intended recipient, or the employee or agent responsible for delivering the message to the intended recipient, or if you have received this communication in error, please notify us immediately by return e-mail and delete the original message and any copies of it from your computer system. For further information about Hikvision company. please see our website at www.hikvision.com<http://www.hikvision.com>

Reviewed-by
#2 lore raw
In reply to a message not in the archive
Quoted 9 lines · changfengnan

From: Fengnan Chang <[email protected]>

When format ext4 with lazy_journal_init=1, the previous transaction is still on disk, it is possible that the previous transaction will be used again during jbd2 recovery.Because the seed is changed, the CRC check will fail.

This version of the patch looks OK, with one trivial indentation problem after the "jbd_debug(1, "JBD2: Invalid checksum" line that could be fixed.

Quoted 63 lines

---
fs/jbd2/recovery.c | 60 +++++++++++++++++++++++++++++++++++++++++-----

1 file changed, 54 insertions(+), 6 deletions(-)
fs/jbd2/recovery.c +27 -2
index a4967b27ffb6..8a6bc322a06d 100644
@@ -33,6 +33,8 @@ struct recovery_info
 	int		nr_replays;
 	int		nr_revokes;
 	int		nr_revoke_hits;
+	unsigned long  ri_commit_block;
+	__be64  last_trans_commit_time;
 };
 
 enum passtype {PASS_SCAN, PASS_REVOKE, PASS_REPLAY};
@@ -412,7 +414,27 @@ static int jbd2_block_tag_csum_verify(journal_t *j, journal_block_tag_t *tag,
 	else
 		return tag->t_checksum == cpu_to_be16(csum32);
 }
+/*
+ * We check the commit time and compare it with the commit time of
+ * the previous transaction, if it's smaller than previous,
+ * We think it's not belong to same journal.
+ */
+static bool is_same_journal(journal_t *journal, struct buffer_head *bh, unsigned long blocknr, __u64 last_commit_sec)
+{
+	unsigned long commit_block = blocknr + count_tags(journal, bh) + 1;
+	struct buffer_head *nbh;
+	struct commit_header *cbh;
+	__u64	commit_sec;
+	int err = jread(&nbh, journal, commit_block);
 
+	if (err)
+		return true;
+
+	cbh = (struct commit_header *)nbh->b_data;
+	commit_sec = be64_to_cpu(cbh->h_commit_sec);
+
+	return commit_sec >= last_commit_sec;
+}
 static int do_one_pass(journal_t *journal,
 			struct recovery_info *info, enum passtype pass)
 {
@@ -514,18 +536,29 @@ static int do_one_pass(journal_t *journal,
 		switch(blocktype) {
 		case JBD2_DESCRIPTOR_BLOCK:
 			/* Verify checksum first */
+			if (pass == PASS_SCAN)
+				info->ri_commit_block = 0;
+
 			if (jbd2_journal_has_csum_v2or3(journal))
 				descr_csum_size =
 					sizeof(struct jbd2_journal_block_tail);
 			if (descr_csum_size > 0 &&
 			    !jbd2_descriptor_block_csum_verify(journal,
 							       bh->b_data)) {
-				printk(KERN_ERR "JBD2: Invalid checksum "
-				       "recovering block %lu in log\n",
+				if (is_same_journal(journal, bh, next_log_block-1, info->last_trans_commit_time)) {
+					printk(KERN_ERR "JBD2: Invalid checksum recovering block %lu in log\n",
 				       next_log_block);

fs/jbd2/recovery.c static int do_one_pass(journal_t *journal, · line 557

Quoted 10 lines
fs/jbd2/recovery.c +7 -3
-				err = -EFSBADCRC;
-				brelse(bh);
-				goto failed;
+					err = -EFSBADCRC;
+					brelse(bh);
+					goto failed;
+				} else {
+					/*it's not belong to same journal, just end this recovery with success*/
+					jbd_debug(1, "JBD2: Invalid checksum found in block %lu in log, but not same journal %d\n",
+				       next_log_block, next_commit_ID);

The continued line should be indented one more tab.

Quoted 39 lines
fs/jbd2/recovery.c +20 -2
+					err = 0;
+					brelse(bh);
+					goto done;
+				}
 			}
 
 			/* If it is a valid descriptor block, replay it
@@ -688,6 +721,17 @@ static int do_one_pass(journal_t *journal,
 			 * are present verify them in PASS_SCAN; else not
 			 * much to do other than move on to the next sequence
 			 * number. */
+			if (pass == PASS_SCAN) {
+				struct commit_header *cbh =
+					(struct commit_header *)bh->b_data;
+				if (info->ri_commit_block) {
+					jbd_debug(1, "invalid commit block found in %lu, stop here.\n", next_log_block);
+					brelse(bh);
+					goto done;
+				}
+				info->ri_commit_block = next_log_block;
+				info->last_trans_commit_time = be64_to_cpu(cbh->h_commit_sec);
+			}
 			if (pass == PASS_SCAN &&
 			    jbd2_has_feature_checksum(journal)) {
 				int chksum_err, chksum_seen;
@@ -761,7 +805,11 @@ static int do_one_pass(journal_t *journal,
 				brelse(bh);
 				continue;
 			}
-
+			if (pass != PASS_SCAN && info->ri_commit_block) {
+				jbd_debug(1, "invalid revoke block found in %lu, stop here.\n", next_log_block);
+				brelse(bh);
+				goto done;
+			}
 			err = scan_revoke_records(journal, bh,
 						  next_commit_ID, info);
 			brelse(bh);
--
Quoted 3 lines

2.27.0.windows.1

Cheers, Andreas

#3 lore raw
In reply to a message not in the archive
Quoted 8 lines · changfengnan

From: Fengnan Chang <[email protected]>

When format ext4 with lazy_journal_init=1, the previous transaction is still on disk, it is possible that the previous transaction will be used again during jbd2 recovery.Because the seed is changed, the CRC check will fail.

Thanks for the patch! I have a few functional comments below, also on a coding style side your patch contains lines a lot longer than 80 characters. Although that is not a strict requirement anymore it is still prefered that lines are wrapped to fit 80 columns to make code more easily readable...

Quoted 9 lines
---
 fs/jbd2/recovery.c | 60 +++++++++++++++++++++++++++++++++++++++++-----
 1 file changed, 54 insertions(+), 6 deletions(-)
fs/jbd2/recovery.c +0 -0
index a4967b27ffb6..8a6bc322a06d 100644

fs/jbd2/recovery.c struct recovery_info · line 37

Quoted 5 lines
fs/jbd2/recovery.c +2 -0
@@ -33,6 +33,8 @@ struct recovery_info
 	int		nr_replays;
 	int		nr_revokes;
 	int		nr_revoke_hits;
+	unsigned long  ri_commit_block;
+	__be64  last_trans_commit_time;

Is there a need for these in recovery_info? They are needed only in PASS_SCAN AFAICT so they can be as well just local variables in do_one_pass()...

Quoted 9 lines
fs/jbd2/recovery.c +2 -0
 };
 
 enum passtype {PASS_SCAN, PASS_REVOKE, PASS_REPLAY};
@@ -412,7 +414,27 @@ static int jbd2_block_tag_csum_verify(journal_t *j, journal_block_tag_t *tag,
 	else
 		return tag->t_checksum == cpu_to_be16(csum32);
 }
+/*
+ * We check the commit time and compare it with the commit time of

fs/jbd2/recovery.c static int jbd2_block_tag_csum_verify(journal_t *j, journal_block_tag_t *tag, · line 428

Quoted 10 lines
fs/jbd2/recovery.c +10 -0
+ * the previous transaction, if it's smaller than previous,
+ * We think it's not belong to same journal.
+ */
+static bool is_same_journal(journal_t *journal, struct buffer_head *bh, unsigned long blocknr, __u64 last_commit_sec)
+{
+	unsigned long commit_block = blocknr + count_tags(journal, bh) + 1;
+	struct buffer_head *nbh;
+	struct commit_header *cbh;
+	__u64	commit_sec;
+	int err = jread(&nbh, journal, commit_block);

So there are two flaws in this logic. First, the computed commit_block may be beyond end of the journal. You need to wrap the computed value using wrap() macro. Second, you seem to assume that each transaction has only a single descriptor block. That is definitely not true. There can be arbitrary number (well, up to a maximum transaction size which depends on the total size of the journal) of descriptor blocks in a transaction before a commit block. So to locate commit block, you need to read the block, check its type (it can be another descriptor block, revoke block, or commit block) and then decide accordingly what to do. And you have to be rather careful when going through the journal like this because you cannot check checksums and blocks can be containing arbitrary garbage. Probably the cleanest way how to implement this would be to use the scanning loop in do_one_pass(), just indicate in some state variable that "we are now searching for next commit block to determine whether a checksum failure is due to lazy init or storage error". In this state we will continue PASS_SCAN, just not checking checksums anymore until we find a commit block or reach end of journal and then based on what we found decide whether to report checksum error or just silently end PASS_SCAN pass...

Quoted 52 lines
fs/jbd2/recovery.c +25 -5
+	if (err)
+		return true;
+
+	cbh = (struct commit_header *)nbh->b_data;
+	commit_sec = be64_to_cpu(cbh->h_commit_sec);
+
+	return commit_sec >= last_commit_sec;
+}
 static int do_one_pass(journal_t *journal,
 			struct recovery_info *info, enum passtype pass)
 {
@@ -514,18 +536,29 @@ static int do_one_pass(journal_t *journal,
 		switch(blocktype) {
 		case JBD2_DESCRIPTOR_BLOCK:
 			/* Verify checksum first */
+			if (pass == PASS_SCAN)
+				info->ri_commit_block = 0;
+
 			if (jbd2_journal_has_csum_v2or3(journal))
 				descr_csum_size =
 					sizeof(struct jbd2_journal_block_tail);
 			if (descr_csum_size > 0 &&
 			    !jbd2_descriptor_block_csum_verify(journal,
 							       bh->b_data)) {
-				printk(KERN_ERR "JBD2: Invalid checksum "
-				       "recovering block %lu in log\n",
+				if (is_same_journal(journal, bh, next_log_block-1, info->last_trans_commit_time)) {
+					printk(KERN_ERR "JBD2: Invalid checksum recovering block %lu in log\n",
 				       next_log_block);
-				err = -EFSBADCRC;
-				brelse(bh);
-				goto failed;
+					err = -EFSBADCRC;
+					brelse(bh);
+					goto failed;
+				} else {
+					/*it's not belong to same journal, just end this recovery with success*/
+					jbd_debug(1, "JBD2: Invalid checksum found in block %lu in log, but not same journal %d\n",
+				       next_log_block, next_commit_ID);
+					err = 0;
+					brelse(bh);
+					goto done;
+				}
 			}
 
 			/* If it is a valid descriptor block, replay it
@@ -688,6 +721,17 @@ static int do_one_pass(journal_t *journal,
 			 * are present verify them in PASS_SCAN; else not
 			 * much to do other than move on to the next sequence
 			 * number. */
+			if (pass == PASS_SCAN) {

fs/jbd2/recovery.c static int do_one_pass(journal_t *journal, · line 734

Quoted 10 lines
fs/jbd2/recovery.c +10 -0
+				struct commit_header *cbh =
+					(struct commit_header *)bh->b_data;
+				if (info->ri_commit_block) {
+					jbd_debug(1, "invalid commit block found in %lu, stop here.\n", next_log_block);
+					brelse(bh);
+					goto done;
+				}
+				info->ri_commit_block = next_log_block;
+				info->last_trans_commit_time = be64_to_cpu(cbh->h_commit_sec);
+			}

I don't think you need to update ri_commit_block and last_trans_commit_time if we are not checking checksums (lazy journal init cannot happen without journal checksumming being used). So I'd just move this into the

                        if (pass == PASS_SCAN &&
                            jbd2_has_feature_checksum(journal)) {

branch below. Also we could validate the commit block the same way (using timestamp) as the descriptor block. Then you will not need the 'ri_commit_block' logic and things will be nicely uniform...

fs/jbd2/recovery.c static int do_one_pass(journal_t *journal, · line 808

Quoted 9 lines
fs/jbd2/recovery.c +1 -1
 			if (pass == PASS_SCAN &&
 			    jbd2_has_feature_checksum(journal)) {
 				int chksum_err, chksum_seen;
@@ -761,7 +805,11 @@ static int do_one_pass(journal_t *journal,
 				brelse(bh);
 				continue;
 			}
-
+			if (pass != PASS_SCAN && info->ri_commit_block) {

I don't understand this. We get here only if "pass == PASS_REVOKE" so "pass != PASS_SCAN" is guaranteed to be true. But also info->ri_commit_block will likely be != 0 from the previous PASS_SCAN pass. So this will always skip revoke block handling?

fs/jbd2/recovery.c static int do_one_pass(journal_t *journal, · line 815

Quoted 7 lines
fs/jbd2/recovery.c +4 -0
+				jbd_debug(1, "invalid revoke block found in %lu, stop here.\n", next_log_block);
+				brelse(bh);
+				goto done;
+			}
 			err = scan_revoke_records(journal, bh,
 						  next_commit_ID, info);
 			brelse(bh);

Honza

Signature · 2 lines
Jan Kara <[email protected]>
SUSE Labs, CR
author
#4 lore raw

Thank you for your review comments, there are indeed some problems.
Let's discuss the function problems first, and I will modify the code style later.
1. About add two members in struct recovery_info, it can be just local variables in do_one_pass().
2. About is_same_journal(), I indded I did not consider the situation of serverl
descriptor blocks in a transaction,I have moved the work of is_same_journal() outside to do it as you suggest.
3. You suggest add jbd2_has_feature_checksum() judge in PASS_SCAN, But jbd2_has_feature_checksum() only check checksum VER1, not include VER2 and VER3. And I don't understand why lazy journal init cannot happen without journal checksumming being used, so i didn't add jbd2_has_feature_checksum() for now.
4. The pass == PASS_REVOKE case, this is a low-level error, my previous test did not cover it. I have changed it now.

fs/jbd2/recovery.c +47 -8
index a4967b27ffb6..f7702e14077f 100644
@@ -417,7 +417,7 @@ static int do_one_pass(journal_t *journal,
 struct recovery_info *info, enum passtype pass)
 {
 unsigned intfirst_commit_ID, next_commit_ID;
-unsigned longnext_log_block;
+unsigned longnext_log_block, ri_commit_block = 0;
 interr, success = 0;
 journal_superblock_t *sb;
 journal_header_t *tmp;
@@ -428,7 +428,9 @@ static int do_one_pass(journal_t *journal,
 __u32crc32_sum = ~0; /* Transactional Checksums */
 intdescr_csum_size = 0;
 intblock_error = 0;
-
+boolneed_check_commit_time = false;
+__be64last_trans_commit_time;
+
 /*
  * First thing is to establish what we expect to find in the log
  * (in terms of transaction IDs), and where (in terms of log
@@ -514,18 +516,18 @@ static int do_one_pass(journal_t *journal,
 switch(blocktype) {
 case JBD2_DESCRIPTOR_BLOCK:
 /* Verify checksum first */
+if(pass == PASS_SCAN)
+ri_commit_block = 0;
+
 if (jbd2_journal_has_csum_v2or3(journal))
 descr_csum_size =
 sizeof(struct jbd2_journal_block_tail);
 if (descr_csum_size > 0 &&
     !jbd2_descriptor_block_csum_verify(journal,
        bh->b_data)) {
-printk(KERN_ERR "JBD2: Invalid checksum "
-       "recovering block %lu in log\n",
-       next_log_block);
-err = -EFSBADCRC;
-brelse(bh);
-goto failed;
+need_check_commit_time = true;
+jbd_debug(1, "invalid descriptor block found in %lu, continue recovery first.\n",next_log_block);
+
 }
 
 /* If it is a valid descriptor block, replay it
@@ -535,6 +537,7 @@ static int do_one_pass(journal_t *journal,
 if (pass != PASS_REPLAY) {
 if (pass == PASS_SCAN &&
     jbd2_has_feature_checksum(journal) &&
+    !need_check_commit_time &&
     !info->end_transaction) {
 if (calc_chksums(journal, bh,
 &next_log_block,
@@ -688,6 +691,36 @@ static int do_one_pass(journal_t *journal,
  * are present verify them in PASS_SCAN; else not
  * much to do other than move on to the next sequence
  * number. */
+if(pass == PASS_SCAN) {
+struct commit_header *cbh =
+(struct commit_header *)bh->b_data;
+if(need_check_commit_time) {
+__be64 commit_time = be64_to_cpu(cbh->h_commit_sec);
+if(commit_time >= last_trans_commit_time) {
+printk(KERN_ERR "JBD2: Invalid checksum found in log, %d\n",
+next_commit_ID);
+err = -EFSBADCRC;
+brelse(bh);
+goto failed;
+}
+else
+{
+/*it's not belong to same journal, just end this recovery with success*/
+jbd_debug(1, "JBD2: Invalid checksum found in block in log, but not same journal %d\n",
+next_commit_ID);
+err = 0;
+brelse(bh);
+goto done;
+}
+}
+if(ri_commit_block) {
+jbd_debug(1, "invalid commit block found in %lu, stop here.\n",next_log_block);
+brelse(bh);
+goto done;
+}
+ri_commit_block = next_log_block;
+last_trans_commit_time = be64_to_cpu(cbh->h_commit_sec);
+}
 if (pass == PASS_SCAN &&
     jbd2_has_feature_checksum(journal)) {
 int chksum_err, chksum_seen;
@@ -755,6 +788,12 @@ static int do_one_pass(journal_t *journal,
 continue;
 
 case JBD2_REVOKE_BLOCK:
+if (pass == PASS_SCAN &&
+ri_commit_block) {
+jbd_debug(1, "invalid revoke block found in %lu, stop here.\n",next_log_block);
+brelse(bh);
+goto done;
+}
 /* If we aren't in the REVOKE pass, then we can
  * just skip over this block. */
 if (pass != PASS_REVOKE) {

-----邮件原件-----
发件人: Jan Kara <[email protected]>
发送时间: 2020年9月11日 18:06
收件人: changfengnan <[email protected]>
抄送: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; 常凤楠 <[email protected]> 主题: Re: [PATCH] jbd2: avoid transaction reuse after reformatting

Quoted 8 lines · changfengnan

From: Fengnan Chang <[email protected]>

When format ext4 with lazy_journal_init=1, the previous transaction is still on disk, it is possible that the previous transaction will be used again during jbd2 recovery.Because the seed is changed, the CRC check will fail.

Thanks for the patch! I have a few functional comments below, also on a coding style side your patch contains lines a lot longer than 80 characters. Although that is not a strict requirement anymore it is still prefered that lines are wrapped to fit 80 columns to make code more easily readable...

Quoted 10 lines
---
 fs/jbd2/recovery.c | 60
+++++++++++++++++++++++++++++++++++++++++-----
 1 file changed, 54 insertions(+), 6 deletions(-)
fs/jbd2/recovery.c index renamed from fs/jbd2/recovery.c +0 -0

a4967b27ffb6..8a6bc322a06d 100644

fs/jbd2/recovery.c struct recovery_info · line 37

Quoted 5 lines
fs/jbd2/recovery.c +2 -0
@@ -33,6 +33,8 @@ struct recovery_info
 intnr_replays;
 intnr_revokes;
 intnr_revoke_hits;
+unsigned long  ri_commit_block;
+__be64  last_trans_commit_time;

Is there a need for these in recovery_info? They are needed only in PASS_SCAN AFAICT so they can be as well just local variables in do_one_pass()...

Quoted 8 lines
fs/jbd2/recovery.c +3 -0
 };
 
 enum passtype {PASS_SCAN, PASS_REVOKE, PASS_REPLAY}; @@ -412,7
+414,27 @@ static int jbd2_block_tag_csum_verify(journal_t *j, journal_block_tag_t *tag,
 else
 return tag->t_checksum == cpu_to_be16(csum32);  }
+/*
+ * We check the commit time and compare it with the commit time of

fs/jbd2/recovery.c struct recovery_info

Quoted 10 lines
fs/jbd2/recovery.c +10 -0
+ * the previous transaction, if it's smaller than previous,
+ * We think it's not belong to same journal.
+ */
+static bool is_same_journal(journal_t *journal, struct buffer_head
+*bh, unsigned long blocknr, __u64 last_commit_sec) {
+unsigned long commit_block = blocknr + count_tags(journal, bh) + 1;
+struct buffer_head *nbh;
+struct commit_header *cbh;
+__u64commit_sec;
+int err = jread(&nbh, journal, commit_block);

So there are two flaws in this logic. First, the computed commit_block may be beyond end of the journal. You need to wrap the computed value using
wrap() macro. Second, you seem to assume that each transaction has only a single descriptor block. That is definitely not true. There can be arbitrary number (well, up to a maximum transaction size which depends on the total size of the journal) of descriptor blocks in a transaction before a commit block. So to locate commit block, you need to read the block, check its type (it can be another descriptor block, revoke block, or commit
block) and then decide accordingly what to do. And you have to be rather careful when going through the journal like this because you cannot check checksums and blocks can be containing arbitrary garbage. Probably the cleanest way how to implement this would be to use the scanning loop in do_one_pass(), just indicate in some state variable that "we are now searching for next commit block to determine whether a checksum failure is due to lazy init or storage error". In this state we will continue PASS_SCAN, just not checking checksums anymore until we find a commit block or reach end of journal and then based on what we found decide whether to report checksum error or just silently end PASS_SCAN pass...

Quoted 51 lines
fs/jbd2/recovery.c +25 -5
+if (err)
+return true;
+
+cbh = (struct commit_header *)nbh->b_data;
+commit_sec = be64_to_cpu(cbh->h_commit_sec);
+
+return commit_sec >= last_commit_sec; }
 static int do_one_pass(journal_t *journal,
 struct recovery_info *info, enum passtype pass)  { @@ -514,18
+536,29 @@ static int do_one_pass(journal_t *journal,
 switch(blocktype) {
 case JBD2_DESCRIPTOR_BLOCK:
 /* Verify checksum first */
+if (pass == PASS_SCAN)
+info->ri_commit_block = 0;
+
 if (jbd2_journal_has_csum_v2or3(journal))
 descr_csum_size =
 sizeof(struct jbd2_journal_block_tail);
 if (descr_csum_size > 0 &&
     !jbd2_descriptor_block_csum_verify(journal,
        bh->b_data)) {
-printk(KERN_ERR "JBD2: Invalid checksum "
-       "recovering block %lu in log\n",
+if (is_same_journal(journal, bh, next_log_block-1, info->last_trans_commit_time)) {
+printk(KERN_ERR "JBD2: Invalid checksum recovering block %lu in
+log\n",
        next_log_block);
-err = -EFSBADCRC;
-brelse(bh);
-goto failed;
+err = -EFSBADCRC;
+brelse(bh);
+goto failed;
+} else {
+/*it's not belong to same journal, just end this recovery with success*/
+jbd_debug(1, "JBD2: Invalid checksum found in block %lu in log, but not same journal %d\n",
+       next_log_block, next_commit_ID);
+err = 0;
+brelse(bh);
+goto done;
+}
 }
 
 /* If it is a valid descriptor block, replay it @@ -688,6 +721,17

@@ static int do_one_pass(journal_t *journal,
* are present verify them in PASS_SCAN; else not
* much to do other than move on to the next sequence
* number. */

fs/jbd2/recovery.c +1 -0
+if (pass == PASS_SCAN) {

fs/jbd2/recovery.c

Quoted 10 lines
fs/jbd2/recovery.c +10 -0
+struct commit_header *cbh =
+(struct commit_header *)bh->b_data;
+if (info->ri_commit_block) {
+jbd_debug(1, "invalid commit block found in %lu, stop here.\n", next_log_block);
+brelse(bh);
+goto done;
+}
+info->ri_commit_block = next_log_block;
+info->last_trans_commit_time = be64_to_cpu(cbh->h_commit_sec);
+}

I don't think you need to update ri_commit_block and last_trans_commit_time if we are not checking checksums (lazy journal init cannot happen without journal checksumming being used). So I'd just move this into the

                        if (pass == PASS_SCAN &&
                            jbd2_has_feature_checksum(journal)) {

branch below. Also we could validate the commit block the same way (using
timestamp) as the descriptor block. Then you will not need the 'ri_commit_block' logic and things will be nicely uniform...

fs/jbd2/recovery.c static int do_one_pass(journal_t *journal, · line 808

Quoted 9 lines

if (pass == PASS_SCAN &&

     jbd2_has_feature_checksum(journal)) {

int chksum_err, chksum_seen;

fs/jbd2/recovery.c +1 -1
@@ -761,7 +805,11 @@ static int do_one_pass(journal_t *journal,
 brelse(bh);
 continue;
 }
-
+if (pass != PASS_SCAN && info->ri_commit_block) {

I don't understand this. We get here only if "pass == PASS_REVOKE" so "pass != PASS_SCAN" is guaranteed to be true. But also info->ri_commit_block will likely be != 0 from the previous PASS_SCAN pass. So this will always skip revoke block handling?

fs/jbd2/recovery.c static int do_one_pass(journal_t *journal, · line 815

Quoted 7 lines
fs/jbd2/recovery.c +4 -0
+jbd_debug(1, "invalid revoke block found in %lu, stop here.\n", next_log_block);
+brelse(bh);
+goto done;
+}
 err = scan_revoke_records(journal, bh,
   next_commit_ID, info);
 brelse(bh);

Honza
--
Jan Kara <[email protected]>
SUSE Labs, CR

________________________________

CONFIDENTIALITY NOTICE: This electronic message is intended to be viewed only by the individual or entity to whom it is addressed. It may contain information that is privileged, confidential and exempt from disclosure under applicable law. Any dissemination, distribution or copying of this communication is strictly prohibited without our prior permission. If the reader of this message is not the intended recipient, or the employee or agent responsible for delivering the message to the intended recipient, or if you have received this communication in error, please notify us immediately by return e-mail and delete the original message and any copies of it from your computer system. For further information about Hikvision company. please see our website at www.hikvision.com<http://www.hikvision.com>

#5 lore raw
Quoted 9 lines · 常凤楠

Thank you for your review comments, there are indeed some problems.
Let's discuss the function problems first, and I will modify the code style later.
1. About add two members in struct recovery_info, it can be just local variables in do_one_pass().
2. About is_same_journal(), I indded I did not consider the situation of serverl
descriptor blocks in a transaction,I have moved the work of is_same_journal() outside to do it as you suggest.
3. You suggest add jbd2_has_feature_checksum() judge in PASS_SCAN, But jbd2_has_feature_checksum() only check checksum VER1, not include VER2 and VER3. And I don't understand why lazy journal init cannot happen without journal checksumming being used, so i didn't add jbd2_has_feature_checksum() for now.
4. The pass == PASS_REVOKE case, this is a low-level error, my previous test did not cover it. I have changed it now.

Please, send the patch as an attachment if your mailer is mangling whitespace like this. The patch is very hard to read in this form. Thanks!

Honza

Quoted 266 lines
fs/jbd2/recovery.c +47 -9
index a4967b27ffb6..f7702e14077f 100644
@@ -417,7 +417,7 @@ static int do_one_pass(journal_t *journal,
 struct recovery_info *info, enum passtype pass)
 {
 unsigned intfirst_commit_ID, next_commit_ID;
-unsigned longnext_log_block;
+unsigned longnext_log_block, ri_commit_block = 0;
 interr, success = 0;
 journal_superblock_t *sb;
 journal_header_t *tmp;
@@ -428,7 +428,9 @@ static int do_one_pass(journal_t *journal,
 __u32crc32_sum = ~0; /* Transactional Checksums */
 intdescr_csum_size = 0;
 intblock_error = 0;
-
+boolneed_check_commit_time = false;
+__be64last_trans_commit_time;
+
 /*
  * First thing is to establish what we expect to find in the log
  * (in terms of transaction IDs), and where (in terms of log
@@ -514,18 +516,18 @@ static int do_one_pass(journal_t *journal,
 switch(blocktype) {
 case JBD2_DESCRIPTOR_BLOCK:
 /* Verify checksum first */
+if(pass == PASS_SCAN)
+ri_commit_block = 0;
+
 if (jbd2_journal_has_csum_v2or3(journal))
 descr_csum_size =
 sizeof(struct jbd2_journal_block_tail);
 if (descr_csum_size > 0 &&
     !jbd2_descriptor_block_csum_verify(journal,
        bh->b_data)) {
-printk(KERN_ERR "JBD2: Invalid checksum "
-       "recovering block %lu in log\n",
-       next_log_block);
-err = -EFSBADCRC;
-brelse(bh);
-goto failed;
+need_check_commit_time = true;
+jbd_debug(1, "invalid descriptor block found in %lu, continue recovery first.\n",next_log_block);
+
 }
 
 /* If it is a valid descriptor block, replay it
@@ -535,6 +537,7 @@ static int do_one_pass(journal_t *journal,
 if (pass != PASS_REPLAY) {
 if (pass == PASS_SCAN &&
     jbd2_has_feature_checksum(journal) &&
+    !need_check_commit_time &&
     !info->end_transaction) {
 if (calc_chksums(journal, bh,
 &next_log_block,
@@ -688,6 +691,36 @@ static int do_one_pass(journal_t *journal,
  * are present verify them in PASS_SCAN; else not
  * much to do other than move on to the next sequence
  * number. */
+if(pass == PASS_SCAN) {
+struct commit_header *cbh =
+(struct commit_header *)bh->b_data;
+if(need_check_commit_time) {
+__be64 commit_time = be64_to_cpu(cbh->h_commit_sec);
+if(commit_time >= last_trans_commit_time) {
+printk(KERN_ERR "JBD2: Invalid checksum found in log, %d\n",
+next_commit_ID);
+err = -EFSBADCRC;
+brelse(bh);
+goto failed;
+}
+else
+{
+/*it's not belong to same journal, just end this recovery with success*/
+jbd_debug(1, "JBD2: Invalid checksum found in block in log, but not same journal %d\n",
+next_commit_ID);
+err = 0;
+brelse(bh);
+goto done;
+}
+}
+if(ri_commit_block) {
+jbd_debug(1, "invalid commit block found in %lu, stop here.\n",next_log_block);
+brelse(bh);
+goto done;
+}
+ri_commit_block = next_log_block;
+last_trans_commit_time = be64_to_cpu(cbh->h_commit_sec);
+}
 if (pass == PASS_SCAN &&
     jbd2_has_feature_checksum(journal)) {
 int chksum_err, chksum_seen;
@@ -755,6 +788,12 @@ static int do_one_pass(journal_t *journal,
 continue;
 
 case JBD2_REVOKE_BLOCK:
+if (pass == PASS_SCAN &&
+ri_commit_block) {
+jbd_debug(1, "invalid revoke block found in %lu, stop here.\n",next_log_block);
+brelse(bh);
+goto done;
+}
 /* If we aren't in the REVOKE pass, then we can
  * just skip over this block. */
 if (pass != PASS_REVOKE) {
 
-----邮件原件-----

发件人: Jan Kara <[email protected]>
发送时间: 2020年9月11日 18:06
收件人: changfengnan <[email protected]>
抄送: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; 常凤楠 <[email protected]> 主题: Re: [PATCH] jbd2: avoid transaction reuse after reformatting

Quoted 8 lines · changfengnan

From: Fengnan Chang <[email protected]>

When format ext4 with lazy_journal_init=1, the previous transaction is still on disk, it is possible that the previous transaction will be used again during jbd2 recovery.Because the seed is changed, the CRC check will fail.

Thanks for the patch! I have a few functional comments below, also on a coding style side your patch contains lines a lot longer than 80 characters. Although that is not a strict requirement anymore it is still prefered that lines are wrapped to fit 80 columns to make code more easily readable...

Quoted 10 lines
---
 fs/jbd2/recovery.c | 60
+++++++++++++++++++++++++++++++++++++++++-----
 1 file changed, 54 insertions(+), 6 deletions(-)
fs/jbd2/recovery.c index renamed from fs/jbd2/recovery.c +0 -0

a4967b27ffb6..8a6bc322a06d 100644

fs/jbd2/recovery.c struct recovery_info · line 37

Quoted 5 lines
fs/jbd2/recovery.c +2 -0
@@ -33,6 +33,8 @@ struct recovery_info
 intnr_replays;
 intnr_revokes;
 intnr_revoke_hits;
+unsigned long  ri_commit_block;
+__be64  last_trans_commit_time;

Is there a need for these in recovery_info? They are needed only in PASS_SCAN AFAICT so they can be as well just local variables in do_one_pass()...

Quoted 8 lines
fs/jbd2/recovery.c +3 -0
 };
 
 enum passtype {PASS_SCAN, PASS_REVOKE, PASS_REPLAY}; @@ -412,7
+414,27 @@ static int jbd2_block_tag_csum_verify(journal_t *j, journal_block_tag_t *tag,
 else
 return tag->t_checksum == cpu_to_be16(csum32);  }
+/*
+ * We check the commit time and compare it with the commit time of

fs/jbd2/recovery.c struct recovery_info

Quoted 10 lines
fs/jbd2/recovery.c +10 -0
+ * the previous transaction, if it's smaller than previous,
+ * We think it's not belong to same journal.
+ */
+static bool is_same_journal(journal_t *journal, struct buffer_head
+*bh, unsigned long blocknr, __u64 last_commit_sec) {
+unsigned long commit_block = blocknr + count_tags(journal, bh) + 1;
+struct buffer_head *nbh;
+struct commit_header *cbh;
+__u64commit_sec;
+int err = jread(&nbh, journal, commit_block);

So there are two flaws in this logic. First, the computed commit_block may be beyond end of the journal. You need to wrap the computed value using
wrap() macro. Second, you seem to assume that each transaction has only a single descriptor block. That is definitely not true. There can be arbitrary number (well, up to a maximum transaction size which depends on the total size of the journal) of descriptor blocks in a transaction before a commit block. So to locate commit block, you need to read the block, check its type (it can be another descriptor block, revoke block, or commit
block) and then decide accordingly what to do. And you have to be rather careful when going through the journal like this because you cannot check checksums and blocks can be containing arbitrary garbage. Probably the cleanest way how to implement this would be to use the scanning loop in do_one_pass(), just indicate in some state variable that "we are now searching for next commit block to determine whether a checksum failure is due to lazy init or storage error". In this state we will continue PASS_SCAN, just not checking checksums anymore until we find a commit block or reach end of journal and then based on what we found decide whether to report checksum error or just silently end PASS_SCAN pass...

Quoted 51 lines
fs/jbd2/recovery.c +25 -5
+if (err)
+return true;
+
+cbh = (struct commit_header *)nbh->b_data;
+commit_sec = be64_to_cpu(cbh->h_commit_sec);
+
+return commit_sec >= last_commit_sec; }
 static int do_one_pass(journal_t *journal,
 struct recovery_info *info, enum passtype pass)  { @@ -514,18
+536,29 @@ static int do_one_pass(journal_t *journal,
 switch(blocktype) {
 case JBD2_DESCRIPTOR_BLOCK:
 /* Verify checksum first */
+if (pass == PASS_SCAN)
+info->ri_commit_block = 0;
+
 if (jbd2_journal_has_csum_v2or3(journal))
 descr_csum_size =
 sizeof(struct jbd2_journal_block_tail);
 if (descr_csum_size > 0 &&
     !jbd2_descriptor_block_csum_verify(journal,
        bh->b_data)) {
-printk(KERN_ERR "JBD2: Invalid checksum "
-       "recovering block %lu in log\n",
+if (is_same_journal(journal, bh, next_log_block-1, info->last_trans_commit_time)) {
+printk(KERN_ERR "JBD2: Invalid checksum recovering block %lu in
+log\n",
        next_log_block);
-err = -EFSBADCRC;
-brelse(bh);
-goto failed;
+err = -EFSBADCRC;
+brelse(bh);
+goto failed;
+} else {
+/*it's not belong to same journal, just end this recovery with success*/
+jbd_debug(1, "JBD2: Invalid checksum found in block %lu in log, but not same journal %d\n",
+       next_log_block, next_commit_ID);
+err = 0;
+brelse(bh);
+goto done;
+}
 }
 
 /* If it is a valid descriptor block, replay it @@ -688,6 +721,17

@@ static int do_one_pass(journal_t *journal,
* are present verify them in PASS_SCAN; else not
* much to do other than move on to the next sequence
* number. */

fs/jbd2/recovery.c +1 -0
+if (pass == PASS_SCAN) {

fs/jbd2/recovery.c static int do_one_pass(journal_t *journal, · line 715

Quoted 10 lines
fs/jbd2/recovery.c +10 -0
+struct commit_header *cbh =
+(struct commit_header *)bh->b_data;
+if (info->ri_commit_block) {
+jbd_debug(1, "invalid commit block found in %lu, stop here.\n", next_log_block);
+brelse(bh);
+goto done;
+}
+info->ri_commit_block = next_log_block;
+info->last_trans_commit_time = be64_to_cpu(cbh->h_commit_sec);
+}

I don't think you need to update ri_commit_block and last_trans_commit_time if we are not checking checksums (lazy journal init cannot happen without journal checksumming being used). So I'd just move this into the

                        if (pass == PASS_SCAN &&
                            jbd2_has_feature_checksum(journal)) {

branch below. Also we could validate the commit block the same way (using
timestamp) as the descriptor block. Then you will not need the 'ri_commit_block' logic and things will be nicely uniform...

fs/jbd2/recovery.c static int do_one_pass(journal_t *journal, · line 808

Quoted 9 lines

if (pass == PASS_SCAN &&

     jbd2_has_feature_checksum(journal)) {

int chksum_err, chksum_seen;

fs/jbd2/recovery.c +1 -1
@@ -761,7 +805,11 @@ static int do_one_pass(journal_t *journal,
 brelse(bh);
 continue;
 }
-
+if (pass != PASS_SCAN && info->ri_commit_block) {

I don't understand this. We get here only if "pass == PASS_REVOKE" so "pass != PASS_SCAN" is guaranteed to be true. But also info->ri_commit_block will likely be != 0 from the previous PASS_SCAN pass. So this will always skip revoke block handling?

fs/jbd2/recovery.c static int do_one_pass(journal_t *journal, · line 815

Quoted 7 lines
fs/jbd2/recovery.c +4 -0
+jbd_debug(1, "invalid revoke block found in %lu, stop here.\n", next_log_block);
+brelse(bh);
+goto done;
+}
 err = scan_revoke_records(journal, bh,
   next_commit_ID, info);
 brelse(bh);

Honza
--
Jan Kara <[email protected]>
SUSE Labs, CR

________________________________

CONFIDENTIALITY NOTICE: This electronic message is intended to be viewed only by the individual or entity to whom it is addressed. It may contain information that is privileged, confidential and exempt from disclosure under applicable law. Any dissemination, distribution or copying of this communication is strictly prohibited without our prior permission. If the reader of this message is not the intended recipient, or the employee or agent responsible for delivering the message to the intended recipient, or if you have received this communication in error, please notify us immediately by return e-mail and delete the original message and any copies of it from your computer system. For further information about Hikvision company. please see our website at www.hikvision.com<http://www.hikvision.com>

Signature · 2 lines
Jan Kara <[email protected]>
SUSE Labs, CR
author
#6 lore raw

Sorry about my mailer, the patch is in the attachment.

-----邮件原件-----
发件人: Jan Kara <[email protected]>
发送时间: 2020年9月17日 18:45
收件人: 常凤楠 <[email protected]>
抄送: Jan Kara <[email protected]>; changfengnan <[email protected]>; [email protected]; [email protected]; [email protected]; [email protected]; [email protected] 主题: Re: 答复: [PATCH] jbd2: avoid transaction reuse after reformatting

Quoted 9 lines · 常凤楠

Thank you for your review comments, there are indeed some problems.
Let's discuss the function problems first, and I will modify the code style later.
1. About add two members in struct recovery_info, it can be just local variables in do_one_pass().
2. About is_same_journal(), I indded I did not consider the situation
of serverl descriptor blocks in a transaction,I have moved the work of is_same_journal() outside to do it as you suggest.
3. You suggest add jbd2_has_feature_checksum() judge in PASS_SCAN, But
jbd2_has_feature_checksum() only check checksum VER1, not include VER2
and VER3. And I don't understand why lazy journal init cannot happen without journal checksumming being used, so i didn't add jbd2_has_feature_checksum() for now.
4. The pass == PASS_REVOKE case, this is a low-level error, my previous test did not cover it. I have changed it now.

Please, send the patch as an attachment if your mailer is mangling whitespace like this. The patch is very hard to read in this form. Thanks!

Honza

Quoted 243 lines
fs/jbd2/recovery.c index renamed from fs/jbd2/recovery.c +0 -0

a4967b27ffb6..f7702e14077f 100644

fs/jbd2/recovery.c +0 -0
@@ -417,7 +417,7 @@ static int do_one_pass(journal_t *journal,  struct

recovery_info *info, enum passtype pass) { unsigned intfirst_commit_ID, next_commit_ID; -unsigned longnext_log_block;

fs/jbd2/recovery.c +1 -0
+unsigned longnext_log_block, ri_commit_block = 0;
 interr, success = 0;
 journal_superblock_t *sb;
 journal_header_t *tmp;
@@ -428,7 +428,9 @@ static int do_one_pass(journal_t *journal,

__u32crc32_sum = ~0; /* Transactional Checksums */ intdescr_csum_size = 0; intblock_error = 0;

fs/jbd2/recovery.c +10 -8
-
+boolneed_check_commit_time = false;
+__be64last_trans_commit_time;
+
 /*
  * First thing is to establish what we expect to find in the log
  * (in terms of transaction IDs), and where (in terms of log @@
-514,18 +516,18 @@ static int do_one_pass(journal_t *journal,
 switch(blocktype) {
 case JBD2_DESCRIPTOR_BLOCK:
 /* Verify checksum first */
+if(pass == PASS_SCAN)
+ri_commit_block = 0;
+
 if (jbd2_journal_has_csum_v2or3(journal))
 descr_csum_size =
 sizeof(struct jbd2_journal_block_tail);  if (descr_csum_size > 0 &&
     !jbd2_descriptor_block_csum_verify(journal,
        bh->b_data)) {
-printk(KERN_ERR "JBD2: Invalid checksum "
-       "recovering block %lu in log\n",
-       next_log_block);
-err = -EFSBADCRC;
-brelse(bh);
-goto failed;
+need_check_commit_time = true;
+jbd_debug(1, "invalid descriptor block found in %lu, continue
+recovery first.\n",next_log_block);
+
 }
 
 /* If it is a valid descriptor block, replay it @@ -535,6 +537,7 @@

static int do_one_pass(journal_t *journal, if (pass != PASS_REPLAY) { if (pass == PASS_SCAN &&

     jbd2_has_feature_checksum(journal) &&
fs/jbd2/recovery.c +17 -0
+    !need_check_commit_time &&
     !info->end_transaction) {
 if (calc_chksums(journal, bh,
 &next_log_block,
@@ -688,6 +691,36 @@ static int do_one_pass(journal_t *journal,
  * are present verify them in PASS_SCAN; else not
  * much to do other than move on to the next sequence
  * number. */
+if(pass == PASS_SCAN) {
+struct commit_header *cbh =
+(struct commit_header *)bh->b_data;
+if(need_check_commit_time) {
+__be64 commit_time = be64_to_cpu(cbh->h_commit_sec); if(commit_time
+>= last_trans_commit_time) { printk(KERN_ERR "JBD2: Invalid checksum
+found in log, %d\n", next_commit_ID); err = -EFSBADCRC; brelse(bh);
+goto failed; } else { /*it's not belong to same journal, just end
+this recovery with success*/ jbd_debug(1, "JBD2: Invalid checksum
+found in block in log, but not same journal %d\n", next_commit_ID);
+err = 0; brelse(bh); goto done; } }
+if(ri_commit_block) {
+jbd_debug(1, "invalid commit block found in %lu, stop
+here.\n",next_log_block); brelse(bh); goto done; } ri_commit_block =
+next_log_block; last_trans_commit_time =
+be64_to_cpu(cbh->h_commit_sec); }
 if (pass == PASS_SCAN &&
     jbd2_has_feature_checksum(journal)) {  int chksum_err,

chksum_seen; @@ -755,6 +788,12 @@ static int do_one_pass(journal_t *journal, continue;

case JBD2_REVOKE_BLOCK:

fs/jbd2/recovery.c +4 -1
+if (pass == PASS_SCAN &&
+ri_commit_block) {
+jbd_debug(1, "invalid revoke block found in %lu, stop
+here.\n",next_log_block); brelse(bh); goto done; }
 /* If we aren't in the REVOKE pass, then we can
  * just skip over this block. */
 if (pass != PASS_REVOKE) {
 
-----邮件原件-----

发件人: Jan Kara <[email protected]>
发送时间: 2020年9月11日 18:06
收件人: changfengnan <[email protected]>
抄送: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; 常凤楠 <[email protected]>
主题: Re: [PATCH] jbd2: avoid transaction reuse after reformatting

Quoted 8 lines · changfengnan

From: Fengnan Chang <[email protected]>

When format ext4 with lazy_journal_init=1, the previous transaction is still on disk, it is possible that the previous transaction will be used again during jbd2 recovery.Because the seed is changed, the CRC check will fail.

Thanks for the patch! I have a few functional comments below, also on a coding style side your patch contains lines a lot longer than 80 characters. Although that is not a strict requirement anymore it is still prefered that lines are wrapped to fit 80 columns to make code more easily readable...

fs/jbd2/recovery.c struct recovery_info intnr_replays; · line 35

Quoted 13 lines
---
 fs/jbd2/recovery.c | 60
+++++++++++++++++++++++++++++++++++++++++-----
 1 file changed, 54 insertions(+), 6 deletions(-)
fs/jbd2/recovery.c index renamed from fs/jbd2/recovery.c +0 -0

a4967b27ffb6..8a6bc322a06d 100644

fs/jbd2/recovery.c +2 -0
@@ -33,6 +33,8 @@ struct recovery_info  intnr_replays;
 intnr_revokes;  intnr_revoke_hits;
+unsigned long  ri_commit_block;
+__be64  last_trans_commit_time;

Is there a need for these in recovery_info? They are needed only in PASS_SCAN AFAICT so they can be as well just local variables in do_one_pass()...

Quoted 7 lines
fs/jbd2/recovery.c +2 -0
 };
 
 enum passtype {PASS_SCAN, PASS_REVOKE, PASS_REPLAY}; @@ -412,7
+414,27 @@ static int jbd2_block_tag_csum_verify(journal_t *j,
+journal_block_tag_t *tag,
 else
 return tag->t_checksum == cpu_to_be16(csum32);  }

fs/jbd2/recovery.c struct recovery_info intnr_replays;

Quoted 10 lines
fs/jbd2/recovery.c +10 -0
+/*
+ * We check the commit time and compare it with the commit time of
+ * the previous transaction, if it's smaller than previous,
+ * We think it's not belong to same journal.
+ */
+static bool is_same_journal(journal_t *journal, struct buffer_head
+*bh, unsigned long blocknr, __u64 last_commit_sec) { unsigned long
+commit_block = blocknr + count_tags(journal, bh) + 1; struct
+buffer_head *nbh; struct commit_header *cbh; __u64commit_sec; int
+err = jread(&nbh, journal, commit_block);

So there are two flaws in this logic. First, the computed commit_block
may be beyond end of the journal. You need to wrap the computed value
using
wrap() macro. Second, you seem to assume that each transaction has
only a single descriptor block. That is definitely not true. There can
be arbitrary number (well, up to a maximum transaction size which
depends on the total size of the journal) of descriptor blocks in a
transaction before a commit block. So to locate commit block, you need
to read the block, check its type (it can be another descriptor block,
revoke block, or commit
block) and then decide accordingly what to do. And you have to be rather careful when going through the journal like this because you cannot check checksums and blocks can be containing arbitrary garbage. Probably the cleanest way how to implement this would be to use the scanning loop in do_one_pass(), just indicate in some state variable that "we are now searching for next commit block to determine whether a checksum failure is due to lazy init or storage error". In this state we will continue PASS_SCAN, just not checking checksums anymore until we find a commit block or reach end of journal and then based on what we found decide whether to report checksum error or just silently end PASS_SCAN pass...

Quoted 46 lines
fs/jbd2/recovery.c +7 -0
+if (err)
+return true;
+
+cbh = (struct commit_header *)nbh->b_data; commit_sec =
+be64_to_cpu(cbh->h_commit_sec);
+
+return commit_sec >= last_commit_sec; }
 static int do_one_pass(journal_t *journal,  struct recovery_info

*info, enum passtype pass) { @@ -514,18 +536,29 @@ static int do_one_pass(journal_t *journal,
switch(blocktype) {
case JBD2_DESCRIPTOR_BLOCK:
/* Verify checksum first */

fs/jbd2/recovery.c +14 -5
+if (pass == PASS_SCAN)
+info->ri_commit_block = 0;
+
 if (jbd2_journal_has_csum_v2or3(journal))
 descr_csum_size =
 sizeof(struct jbd2_journal_block_tail);  if (descr_csum_size > 0 &&
     !jbd2_descriptor_block_csum_verify(journal,
        bh->b_data)) {
-printk(KERN_ERR "JBD2: Invalid checksum "
-       "recovering block %lu in log\n",
+if (is_same_journal(journal, bh, next_log_block-1,
+info->last_trans_commit_time)) { printk(KERN_ERR "JBD2: Invalid
+checksum recovering block %lu in log\n",
        next_log_block);
-err = -EFSBADCRC;
-brelse(bh);
-goto failed;
+err = -EFSBADCRC;
+brelse(bh);
+goto failed;
+} else {
+/*it's not belong to same journal, just end this recovery with
+success*/ jbd_debug(1, "JBD2: Invalid checksum found in block %lu in log, but not same journal %d\n",
+       next_log_block, next_commit_ID); err = 0; brelse(bh); goto
+done; }
 }
 
 /* If it is a valid descriptor block, replay it @@ -688,6 +721,17

@@ static int do_one_pass(journal_t *journal,
* are present verify them in PASS_SCAN; else not
* much to do other than move on to the next sequence
* number. */

fs/jbd2/recovery.c

Quoted 8 lines
fs/jbd2/recovery.c +8 -0
+if (pass == PASS_SCAN) {
+struct commit_header *cbh =
+(struct commit_header *)bh->b_data; if (info->ri_commit_block) {
+jbd_debug(1, "invalid commit block found in %lu, stop here.\n",
+next_log_block); brelse(bh); goto done; }
+info->ri_commit_block = next_log_block; last_trans_commit_time =
+info->be64_to_cpu(cbh->h_commit_sec);
+}

I don't think you need to update ri_commit_block and last_trans_commit_time if we are not checking checksums (lazy journal init cannot happen without journal checksumming being used). So I'd just move this into the

                        if (pass == PASS_SCAN &&
                            jbd2_has_feature_checksum(journal)) {

branch below. Also we could validate the commit block the same way
(using
timestamp) as the descriptor block. Then you will not need the 'ri_commit_block' logic and things will be nicely uniform...

fs/jbd2/recovery.c

Quoted 6 lines

if (pass == PASS_SCAN &&
jbd2_has_feature_checksum(journal)) { int chksum_err,
chksum_seen; @@ -761,7 +805,11 @@ static int do_one_pass(journal_t *journal, brelse(bh); continue; }

fs/jbd2/recovery.c +1 -1
-
+if (pass != PASS_SCAN && info->ri_commit_block) {

I don't understand this. We get here only if "pass == PASS_REVOKE" so "pass != PASS_SCAN" is guaranteed to be true. But also info->ri_commit_block will likely be != 0 from the previous PASS_SCAN pass. So this will always skip revoke block handling?

fs/jbd2/recovery.c

Quoted 5 lines
fs/jbd2/recovery.c +2 -0
+jbd_debug(1, "invalid revoke block found in %lu, stop here.\n",
+next_log_block); brelse(bh); goto done; }
 err = scan_revoke_records(journal, bh,
   next_commit_ID, info);
 brelse(bh);

Honza
--
Jan Kara <[email protected]>
SUSE Labs, CR

________________________________

Quoted 12 lines

CONFIDENTIALITY NOTICE: This electronic message is intended to be viewed only by the individual or entity to whom it is addressed. It may contain information that is privileged, confidential and exempt from disclosure under applicable law. Any dissemination, distribution or copying of this communication is strictly prohibited without our prior permission. If the reader of this message is not the intended recipient, or the employee or agent responsible for delivering the message to the intended recipient, or if you have received this communication in error, please notify us immediately by return e-mail and delete the original message and any copies of it from your computer system. For further information about Hikvision company. please see our website at www.hikvision.com<http://www.hikvision.com>

--
Jan Kara <[email protected]>
SUSE Labs, CR

________________________________

CONFIDENTIALITY NOTICE: This electronic message is intended to be viewed only by the individual or entity to whom it is addressed. It may contain information that is privileged, confidential and exempt from disclosure under applicable law. Any dissemination, distribution or copying of this communication is strictly prohibited without our prior permission. If the reader of this message is not the intended recipient, or the employee or agent responsible for delivering the message to the intended recipient, or if you have received this communication in error, please notify us immediately by return e-mail and delete the original message and any copies of it from your computer system. For further information about Hikvision company. please see our website at www.hikvision.com<http://www.hikvision.com>

fs/jbd2/recovery.c +47 -8
index a4967b27ffb6..f7702e14077f 100644
@@ -417,7 +417,7 @@ static int do_one_pass(journal_t *journal,
 			struct recovery_info *info, enum passtype pass)
 {
 	unsigned int		first_commit_ID, next_commit_ID;
-	unsigned long		next_log_block;
+	unsigned long		next_log_block, ri_commit_block = 0;
 	int			err, success = 0;
 	journal_superblock_t *	sb;
 	journal_header_t *	tmp;
@@ -428,7 +428,9 @@ static int do_one_pass(journal_t *journal,
 	__u32			crc32_sum = ~0; /* Transactional Checksums */
 	int			descr_csum_size = 0;
 	int			block_error = 0;
-
+	bool		need_check_commit_time = false;
+	__be64		last_trans_commit_time;
+	
 	/*
 	 * First thing is to establish what we expect to find in the log
 	 * (in terms of transaction IDs), and where (in terms of log
@@ -514,18 +516,18 @@ static int do_one_pass(journal_t *journal,
 		switch(blocktype) {
 		case JBD2_DESCRIPTOR_BLOCK:
 			/* Verify checksum first */
+			if(pass == PASS_SCAN) 
+				ri_commit_block = 0;
+
 			if (jbd2_journal_has_csum_v2or3(journal))
 				descr_csum_size =
 					sizeof(struct jbd2_journal_block_tail);
 			if (descr_csum_size > 0 &&
 			    !jbd2_descriptor_block_csum_verify(journal,
 							       bh->b_data)) {
-				printk(KERN_ERR "JBD2: Invalid checksum "
-				       "recovering block %lu in log\n",
-				       next_log_block);
-				err = -EFSBADCRC;
-				brelse(bh);
-				goto failed;
+				need_check_commit_time = true;
+				jbd_debug(1, "invalid descriptor block found in %lu, continue recovery first.\n",next_log_block);
+				
 			}
 
 			/* If it is a valid descriptor block, replay it
@@ -535,6 +537,7 @@ static int do_one_pass(journal_t *journal,
 			if (pass != PASS_REPLAY) {
 				if (pass == PASS_SCAN &&
 				    jbd2_has_feature_checksum(journal) &&
+				    !need_check_commit_time &&
 				    !info->end_transaction) {
 					if (calc_chksums(journal, bh,
 							&next_log_block,
@@ -688,6 +691,36 @@ static int do_one_pass(journal_t *journal,
 			 * are present verify them in PASS_SCAN; else not
 			 * much to do other than move on to the next sequence
 			 * number. */
+			if(pass == PASS_SCAN) {
+				struct commit_header *cbh =
+					(struct commit_header *)bh->b_data;
+				if(need_check_commit_time) {
+					__be64 commit_time = be64_to_cpu(cbh->h_commit_sec);
+					if(commit_time >= last_trans_commit_time) {
+						printk(KERN_ERR "JBD2: Invalid checksum found in log, %d\n",
+						next_commit_ID);
+						err = -EFSBADCRC;
+						brelse(bh);
+						goto failed;
+					}
+					else
+					{
+						/*it's not belong to same journal, just end this recovery with success*/
+						jbd_debug(1, "JBD2: Invalid checksum found in block in log, but not same journal %d\n",
+						next_commit_ID);
+						err = 0;
+						brelse(bh);
+						goto done;
+					}
+				}
+				if(ri_commit_block) {
+					jbd_debug(1, "invalid commit block found in %lu, stop here.\n",next_log_block);
+					brelse(bh);
+					goto done;
+				}
+				ri_commit_block = next_log_block;
+				last_trans_commit_time = be64_to_cpu(cbh->h_commit_sec);
+			}
 			if (pass == PASS_SCAN &&
 			    jbd2_has_feature_checksum(journal)) {
 				int chksum_err, chksum_seen;
@@ -755,6 +788,12 @@ static int do_one_pass(journal_t *journal,
 			continue;
 
 		case JBD2_REVOKE_BLOCK:
+			if (pass == PASS_SCAN && 
+				ri_commit_block) {
+				jbd_debug(1, "invalid revoke block found in %lu, stop here.\n",next_log_block);
+				brelse(bh);
+				goto done;
+			}
 			/* If we aren't in the REVOKE pass, then we can
 			 * just skip over this block. */
 			if (pass != PASS_REVOKE) {
#7 lore raw

Hello,

Quoted 1 line · 常凤楠

Sorry about my mailer, the patch is in the attachment.

Thanks for the patch. Functionally the patch looks mostly OK now. The only concern I have is that it handles checksum failures only in JBD2_DESCRIPTOR_BLOCK. This is the most likely case but it could also happen that JBD2_REVOKE_BLOCK or JBD2_COMMIT_BLOCK is the first one you see with mismatching checksum. So I think you need to handle these cases as well. I think your ri_commit_block logic below is an attempt to deal with these cases (but it's difficult to be sure because of complete lack of comments) but it is not reliable. A valid transaction can begin both with a descriptor or with a revoke block.

A few other comments mostly about coding style below:

fs/jbd2/recovery.c +3 -2
index a4967b27ffb6..f7702e14077f 100644
@@ -417,7 +417,7 @@ static int do_one_pass(journal_t *journal,
 			struct recovery_info *info, enum passtype pass)
 {
 	unsigned int		first_commit_ID, next_commit_ID;
-	unsigned long		next_log_block;
+	unsigned long		next_log_block, ri_commit_block = 0;
 	int			err, success = 0;
 	journal_superblock_t *	sb;
 	journal_header_t *	tmp;
@@ -428,7 +428,9 @@ static int do_one_pass(journal_t *journal,
 	__u32			crc32_sum = ~0; /* Transactional Checksums */
 	int			descr_csum_size = 0;
 	int			block_error = 0;
-
+	bool		need_check_commit_time = false;
+	__be64		last_trans_commit_time;
 

All variable names in this function seem to be indented by one more column. Please keep the indentation.

+

This empty line has whitespace on it. Please delete.

/*
* First thing is to establish what we expect to find in the log
* (in terms of transaction IDs), and where (in terms of log
@@ -514,18 +516,18 @@ static int do_one_pass(journal_t *journal,
switch(blocktype) {
case JBD2_DESCRIPTOR_BLOCK:
/* Verify checksum first */
+ if(pass == PASS_SCAN)
^ Coding style requires space before opening (.
You have this problem at multiple places.

+ ri_commit_block = 0;
+
if (jbd2_journal_has_csum_v2or3(journal))
descr_csum_size =
sizeof(struct jbd2_journal_block_tail);
if (descr_csum_size > 0 &&
!jbd2_descriptor_block_csum_verify(journal,
bh->b_data)) {
- printk(KERN_ERR "JBD2: Invalid checksum "
- "recovering block %lu in log\n",
- next_log_block);
- err = -EFSBADCRC;
- brelse(bh);
- goto failed;
+ need_check_commit_time = true;
+ jbd_debug(1, "invalid descriptor block found in %lu, continue recovery first.\n",next_log_block);
+
}

/* If it is a valid descriptor block, replay it
@@ -535,6 +537,7 @@ static int do_one_pass(journal_t *journal,
if (pass != PASS_REPLAY) {
if (pass == PASS_SCAN &&
jbd2_has_feature_checksum(journal) &&
+ !need_check_commit_time &&
!info->end_transaction) {
if (calc_chksums(journal, bh,
&next_log_block,
@@ -688,6 +691,36 @@ static int do_one_pass(journal_t *journal,
* are present verify them in PASS_SCAN; else not
* much to do other than move on to the next sequence
* number. */
+ if(pass == PASS_SCAN) {
+ struct commit_header *cbh =
+ (struct commit_header *)bh->b_data;
+ if(need_check_commit_time) {
+ __be64 commit_time = be64_to_cpu(cbh->h_commit_sec);
+ if(commit_time >= last_trans_commit_time) {
+ printk(KERN_ERR "JBD2: Invalid checksum found in log, %d\n",
+ next_commit_ID);
+ err = -EFSBADCRC;
+ brelse(bh);
+ goto failed;
+ }
+ else
+ {
Coding style requires to put opening { on the same line as 'else'. Like:

					else {

+ /*it's not belong to same journal, just end this recovery with success*/
+ jbd_debug(1, "JBD2: Invalid checksum found in block in log, but not same journal %d\n",
+ next_commit_ID);
+ err = 0;
+ brelse(bh);
+ goto done;
+ }
+ }
+ if(ri_commit_block) {
+ jbd_debug(1, "invalid commit block found in %lu, stop here.\n",next_log_block);
+ brelse(bh);
+ goto done;
+ }
+ ri_commit_block = next_log_block;

Why does the ri_commit_block logic exist? I don't see it bringing any benefit...

+ last_trans_commit_time = be64_to_cpu(cbh->h_commit_sec);
+ }
if (pass == PASS_SCAN &&
jbd2_has_feature_checksum(journal)) {
int chksum_err, chksum_seen;
@@ -755,6 +788,12 @@ static int do_one_pass(journal_t *journal,
continue;

case JBD2_REVOKE_BLOCK:
+ if (pass == PASS_SCAN &&
+ ri_commit_block) {
+ jbd_debug(1, "invalid revoke block found in %lu, stop here.\n",next_log_block);
+ brelse(bh);
+ goto done;
+ }

This is wrong. A valid transaction can start with a revoke block...

/* If we aren't in the REVOKE pass, then we can
* just skip over this block. */
if (pass != PASS_REVOKE) {

Honza

Signature · 2 lines
Jan Kara <[email protected]>
SUSE Labs, CR
author
#8 lore raw

The attachment is new patch.
I have fix the logic in JBD2_REVOKE_BLOCK and JBD2_COMMIT_BLOCK case.
If the revoke block is the first block after valid transaction, I set the flag like descriptor block ,and check it in commit block. If the commit block is the first block after valid transaction, I use ri_commit_block to judge whether this commit block is next to another commit block, if so it is illegal. I did't use time to judge commit block, because of the possibility of time calibration, I think use ri_commit_block is more reliable.

-----邮件原件-----
发件人: Jan Kara <[email protected]>
发送时间: 2020年9月18日 21:03
收件人: 常凤楠 <[email protected]>
抄送: Jan Kara <[email protected]>; changfengnan <[email protected]>; [email protected]; [email protected]; [email protected]; [email protected]; [email protected] 主题: Re: 答复: 答复: [PATCH] jbd2: avoid transaction reuse after reformatting

Hello,

Quoted 1 line · 常凤楠

Sorry about my mailer, the patch is in the attachment.

Thanks for the patch. Functionally the patch looks mostly OK now. The only concern I have is that it handles checksum failures only in JBD2_DESCRIPTOR_BLOCK. This is the most likely case but it could also happen that JBD2_REVOKE_BLOCK or JBD2_COMMIT_BLOCK is the first one you see with mismatching checksum. So I think you need to handle these cases as well. I think your ri_commit_block logic below is an attempt to deal with these cases (but it's difficult to be sure because of complete lack of comments) but it is not reliable. A valid transaction can begin both with a descriptor or with a revoke block.

A few other comments mostly about coding style below:

fs/jbd2/recovery.c +3 -2
@@ -417,7 +417,7 @@ static int do_one_pass(journal_t *journal,
 struct recovery_info *info, enum passtype pass)  {
 unsigned intfirst_commit_ID, next_commit_ID;
-unsigned longnext_log_block;
+unsigned longnext_log_block, ri_commit_block = 0;
 interr, success = 0;
 journal_superblock_t *sb;
 journal_header_t *tmp;
@@ -428,7 +428,9 @@ static int do_one_pass(journal_t *journal,
 __u32crc32_sum = ~0; /* Transactional Checksums */
 intdescr_csum_size = 0;
 intblock_error = 0;
-
+boolneed_check_commit_time = false;
+__be64last_trans_commit_time;
 

All variable names in this function seem to be indented by one more column. Please keep the indentation.

+

This empty line has whitespace on it. Please delete.

/*
* First thing is to establish what we expect to find in the log
* (in terms of transaction IDs), and where (in terms of log @@ -514,18 +516,18 @@ static int do_one_pass(journal_t *journal,
switch(blocktype) {
case JBD2_DESCRIPTOR_BLOCK:
/* Verify checksum first */
+if(pass == PASS_SCAN)
^ Coding style requires space before opening (.
You have this problem at multiple places.

+ri_commit_block = 0;
+
if (jbd2_journal_has_csum_v2or3(journal))
descr_csum_size =
sizeof(struct jbd2_journal_block_tail);
if (descr_csum_size > 0 &&

     !jbd2_descriptor_block_csum_verify(journal,
        bh->b_data)) {

-printk(KERN_ERR "JBD2: Invalid checksum "
- "recovering block %lu in log\n",
- next_log_block);
-err = -EFSBADCRC;
-brelse(bh);
-goto failed;
+need_check_commit_time = true;
+jbd_debug(1, "invalid descriptor block found in %lu, continue +recovery first.\n",next_log_block);
+
}

/* If it is a valid descriptor block, replay it @@ -535,6 +537,7 @@ static int do_one_pass(journal_t *journal, if (pass != PASS_REPLAY) {
if (pass == PASS_SCAN &&

     jbd2_has_feature_checksum(journal) &&
+    !need_check_commit_time &&
     !info->end_transaction) {

if (calc_chksums(journal, bh,
&next_log_block,
@@ -688,6 +691,36 @@ static int do_one_pass(journal_t *journal,
* are present verify them in PASS_SCAN; else not
* much to do other than move on to the next sequence
* number. */
+if(pass == PASS_SCAN) {
+struct commit_header *cbh =
+(struct commit_header *)bh->b_data;
+if(need_check_commit_time) {
+__be64 commit_time = be64_to_cpu(cbh->h_commit_sec);
+if(commit_time >= last_trans_commit_time) {
+printk(KERN_ERR "JBD2: Invalid checksum found in log, %d\n", +next_commit_ID);
+err = -EFSBADCRC;
+brelse(bh);
+goto failed;
+}
+else
+{
Coding style requires to put opening { on the same line as 'else'. Like:
else {
+/*it's not belong to same journal, just end this recovery with success*/ +jbd_debug(1, "JBD2: Invalid checksum found in block in log, but not same journal %d\n", +next_commit_ID);
+err = 0;
+brelse(bh);
+goto done;
+}
+}
+if(ri_commit_block) {
+jbd_debug(1, "invalid commit block found in %lu, stop here.\n",next_log_block); +brelse(bh);
+goto done;
+}
+ri_commit_block = next_log_block;

Why does the ri_commit_block logic exist? I don't see it bringing any benefit...

+last_trans_commit_time = be64_to_cpu(cbh->h_commit_sec); +}
if (pass == PASS_SCAN &&

     jbd2_has_feature_checksum(journal)) {

int chksum_err, chksum_seen;
@@ -755,6 +788,12 @@ static int do_one_pass(journal_t *journal,
continue;

case JBD2_REVOKE_BLOCK:
+if (pass == PASS_SCAN &&
+ri_commit_block) {
+jbd_debug(1, "invalid revoke block found in %lu, stop here.\n",next_log_block); +brelse(bh);
+goto done;
+}

This is wrong. A valid transaction can start with a revoke block...

/* If we aren't in the REVOKE pass, then we can
* just skip over this block. */
if (pass != PASS_REVOKE) {

Honza
--
Jan Kara <[email protected]>
SUSE Labs, CR

________________________________

CONFIDENTIALITY NOTICE: This electronic message is intended to be viewed only by the individual or entity to whom it is addressed. It may contain information that is privileged, confidential and exempt from disclosure under applicable law. Any dissemination, distribution or copying of this communication is strictly prohibited without our prior permission. If the reader of this message is not the intended recipient, or the employee or agent responsible for delivering the message to the intended recipient, or if you have received this communication in error, please notify us immediately by return e-mail and delete the original message and any copies of it from your computer system. For further information about Hikvision company. please see our website at www.hikvision.com<http://www.hikvision.com>

fs/jbd2/recovery.c +65 -10

123 lines not on this page · raw message

#9 lore raw
Quoted 1 line · 常凤楠

The attachment is new patch.

Thanks!

Quoted 8 lines

I have fix the logic in JBD2_REVOKE_BLOCK and JBD2_COMMIT_BLOCK case. If the revoke block is the first block after valid transaction, I set the flag like descriptor block ,and check it in commit block. If the commit block is the first block after valid transaction, I use ri_commit_block to judge whether this commit block is next to another commit block, if so it is illegal. I did't use time to judge commit block, because of the possibility of time calibration, I think use ri_commit_block is more reliable.

If the time is unreliable, your logic with ri_commit_block will detect only a small minority of the cases. The majority of cases will fail with checksum error. So I still don't think the ri_commit_block logic is really worth the additional code. Other than that the current version looks OK to me. So please do next submission with full changelog, Signed-off-by line, and the coding style issues fixed (you can use scripts/checkpatch.pl for patch verification). Thanks!

Honza

Quoted 148 lines

-----邮件原件-----
发件人: Jan Kara <[email protected]>
发送时间: 2020年9月18日 21:03
收件人: 常凤楠 <[email protected]>
抄送: Jan Kara <[email protected]>; changfengnan <[email protected]>; [email protected]; [email protected]; [email protected]; [email protected]; [email protected] 主题: Re: 答复: 答复: [PATCH] jbd2: avoid transaction reuse after reformatting

Hello,

Quoted 1 line · 常凤楠

Sorry about my mailer, the patch is in the attachment.

Thanks for the patch. Functionally the patch looks mostly OK now. The only concern I have is that it handles checksum failures only in JBD2_DESCRIPTOR_BLOCK. This is the most likely case but it could also happen that JBD2_REVOKE_BLOCK or JBD2_COMMIT_BLOCK is the first one you see with mismatching checksum. So I think you need to handle these cases as well. I think your ri_commit_block logic below is an attempt to deal with these cases (but it's difficult to be sure because of complete lack of comments) but it is not reliable. A valid transaction can begin both with a descriptor or with a revoke block.

A few other comments mostly about coding style below:

fs/jbd2/recovery.c +3 -2
@@ -417,7 +417,7 @@ static int do_one_pass(journal_t *journal,
 struct recovery_info *info, enum passtype pass)  {
 unsigned intfirst_commit_ID, next_commit_ID;
-unsigned longnext_log_block;
+unsigned longnext_log_block, ri_commit_block = 0;
 interr, success = 0;
 journal_superblock_t *sb;
 journal_header_t *tmp;
@@ -428,7 +428,9 @@ static int do_one_pass(journal_t *journal,
 __u32crc32_sum = ~0; /* Transactional Checksums */
 intdescr_csum_size = 0;
 intblock_error = 0;
-
+boolneed_check_commit_time = false;
+__be64last_trans_commit_time;
 

All variable names in this function seem to be indented by one more column. Please keep the indentation.

+

This empty line has whitespace on it. Please delete.

/*
* First thing is to establish what we expect to find in the log
* (in terms of transaction IDs), and where (in terms of log @@ -514,18 +516,18 @@ static int do_one_pass(journal_t *journal,
switch(blocktype) {
case JBD2_DESCRIPTOR_BLOCK:
/* Verify checksum first */
+if(pass == PASS_SCAN)
^ Coding style requires space before opening (.
You have this problem at multiple places.

fs/jbd2/recovery.c +35 -6
+ri_commit_block = 0;
+
 if (jbd2_journal_has_csum_v2or3(journal))
 descr_csum_size =
 sizeof(struct jbd2_journal_block_tail);
 if (descr_csum_size > 0 &&
     !jbd2_descriptor_block_csum_verify(journal,
        bh->b_data)) {
-printk(KERN_ERR "JBD2: Invalid checksum "
-       "recovering block %lu in log\n",
-       next_log_block);
-err = -EFSBADCRC;
-brelse(bh);
-goto failed;
+need_check_commit_time = true;
+jbd_debug(1, "invalid descriptor block found in %lu, continue
+recovery first.\n",next_log_block);
+
 }
 
 /* If it is a valid descriptor block, replay it @@ -535,6 +537,7 @@ static int do_one_pass(journal_t *journal,
 if (pass != PASS_REPLAY) {
 if (pass == PASS_SCAN &&
     jbd2_has_feature_checksum(journal) &&
+    !need_check_commit_time &&
     !info->end_transaction) {
 if (calc_chksums(journal, bh,
 &next_log_block,
@@ -688,6 +691,36 @@ static int do_one_pass(journal_t *journal,
  * are present verify them in PASS_SCAN; else not
  * much to do other than move on to the next sequence
  * number. */
+if(pass == PASS_SCAN) {
+struct commit_header *cbh =
+(struct commit_header *)bh->b_data;
+if(need_check_commit_time) {
+__be64 commit_time = be64_to_cpu(cbh->h_commit_sec);
+if(commit_time >= last_trans_commit_time) {
+printk(KERN_ERR "JBD2: Invalid checksum found in log, %d\n",
+next_commit_ID);
+err = -EFSBADCRC;
+brelse(bh);
+goto failed;
+}
+else
+{
  Coding style requires to put opening { on the same line as 'else'. Like:
 else {
+/*it's not belong to same journal, just end this recovery with success*/
+jbd_debug(1, "JBD2: Invalid checksum found in block in log, but not same journal %d\n",
+next_commit_ID);
+err = 0;
+brelse(bh);
+goto done;
+}
+}
+if(ri_commit_block) {
+jbd_debug(1, "invalid commit block found in %lu, stop here.\n",next_log_block);
+brelse(bh);
+goto done;
+}
+ri_commit_block = next_log_block;
 

Why does the ri_commit_block logic exist? I don't see it bringing any benefit...

fs/jbd2/recovery.c +8 -0
+last_trans_commit_time = be64_to_cpu(cbh->h_commit_sec);
+}
 if (pass == PASS_SCAN &&
     jbd2_has_feature_checksum(journal)) {
 int chksum_err, chksum_seen;
@@ -755,6 +788,12 @@ static int do_one_pass(journal_t *journal,
 continue;
 
 case JBD2_REVOKE_BLOCK:
+if (pass == PASS_SCAN &&
+ri_commit_block) {
+jbd_debug(1, "invalid revoke block found in %lu, stop here.\n",next_log_block);
+brelse(bh);
+goto done;
+}
 

This is wrong. A valid transaction can start with a revoke block...

/* If we aren't in the REVOKE pass, then we can
* just skip over this block. */
if (pass != PASS_REVOKE) {

Honza
--
Jan Kara <[email protected]>
SUSE Labs, CR

________________________________

CONFIDENTIALITY NOTICE: This electronic message is intended to be viewed only by the individual or entity to whom it is addressed. It may contain information that is privileged, confidential and exempt from disclosure under applicable law. Any dissemination, distribution or copying of this communication is strictly prohibited without our prior permission. If the reader of this message is not the intended recipient, or the employee or agent responsible for delivering the message to the intended recipient, or if you have received this communication in error, please notify us immediately by return e-mail and delete the original message and any copies of it from your computer system. For further information about Hikvision company. please see our website at www.hikvision.com<http://www.hikvision.com>

Signature · 2 lines
Jan Kara <[email protected]>
SUSE Labs, CR
maintainer
#10 lore raw
Quoted 14 lines · Jan Kara
Quoted 1 line · 常凤楠

The attachment is new patch.

Thanks!

Quoted 8 lines

I have fix the logic in JBD2_REVOKE_BLOCK and JBD2_COMMIT_BLOCK case. If the revoke block is the first block after valid transaction, I set the flag like descriptor block ,and check it in commit block. If the commit block is the first block after valid transaction, I use ri_commit_block to judge whether this commit block is next to another commit block, if so it is illegal. I did't use time to judge commit block, because of the possibility of time calibration, I think use ri_commit_block is more reliable.

Quoted 7 lines

If the time is unreliable, your logic with ri_commit_block will detect only a small minority of the cases. The majority of cases will fail with checksum error. So I still don't think the ri_commit_block logic is really worth the additional code. Other than that the current version looks OK to me. So please do next submission with full changelog, Signed-off-by line, and the coding style issues fixed (you can use scripts/checkpatch.pl for patch verification). Thanks!

Hi changfengnan,

Have you had a chance to send a revised patch with the changes requested by Jan?

Thanks,

- Ted