2020-02-13 06:39:59

by Zhang Yi

[permalink] [raw]
Subject: [PATCH v3 0/2] jbd2: fix an oops problem

Changes since v2:
- Back to use "mapping && !sb_is_blkdev_sb(mapping->host->i_sb)" to
distinguish metadata buffers, and add more comments.
- Add 'Reviewed-by' to the first patch.

Changes since v1:
- Switch to clear b_modified just after set_buffer_freed() instead of
reuse codes at the end of journal_unmap_buffer().
- Switch to distinguish metadata buffers through the page mapping dev.

Thanks,
Yi.

--------------
Original description:

We encountered a jbd2 oops problem on an aarch64 machine with 4K block
size and 64K page size when doing stress tests.

Unable to handle kernel NULL pointer dereference at virtual address 0000000000000008
...
user pgtable: 64k pages, 42-bit VAs, pgdp = (____ptrval____)
...
pc : jbd2_journal_put_journal_head+0x7c/0x284
lr : jbd2_journal_put_journal_head+0x3c/0x284
...
Call trace:
jbd2_journal_put_journal_head+0x7c/0x284
__jbd2_journal_refile_buffer+0x164/0x188
jbd2_journal_commit_transaction+0x12a0/0x1a50
kjournald2+0xd0/0x260
kthread+0x134/0x138
ret_from_fork+0x10/0x1c
Code: 51000400 b9000ac0 35000760 f9402274 (b9400a80)
---[ end trace 8fa99273d06aeb63 ]---

These patch set can fix this issue, the first patch is just a cleanup
patch, and the second one describe the root cause and fix it.


zhangyi (F) (2):
jbd2: move the clearing of b_modified flag to the
journal_unmap_buffer()
jbd2: do not clear the BH_Mapped flag when forgetting a metadata
buffer

fs/jbd2/commit.c | 46 +++++++++++++++++++++++--------------------
fs/jbd2/transaction.c | 10 ++++++----
2 files changed, 31 insertions(+), 25 deletions(-)

--
2.17.2


2020-02-13 06:39:59

by Zhang Yi

[permalink] [raw]
Subject: [PATCH v3 2/2] jbd2: do not clear the BH_Mapped flag when forgetting a metadata buffer

Commit 904cdbd41d74 ("jbd2: clear dirty flag when revoking a buffer from
an older transaction") set the BH_Freed flag when forgetting a metadata
buffer which belongs to the committing transaction, it indicate the
committing process clear dirty bits when it is done with the buffer. But
it also clear the BH_Mapped flag at the same time, which may trigger
below NULL pointer oops when block_size < PAGE_SIZE.

rmdir 1 kjournald2 mkdir 2
jbd2_journal_commit_transaction
commit transaction N
jbd2_journal_forget
set_buffer_freed(bh1)
jbd2_journal_commit_transaction
commit transaction N+1
...
clear_buffer_mapped(bh1)
ext4_getblk(bh2 ummapped)
...
grow_dev_page
init_page_buffers
bh1->b_private=NULL
bh2->b_private=NULL
jbd2_journal_put_journal_head(jh1)
__journal_remove_journal_head(hb1)
jh1 is NULL and trigger oops

*) Dir entry block bh1 and bh2 belongs to one page, and the bh2 has
already been unmapped.

For the metadata buffer we forgetting, we should always keep the mapped
flag and clear the dirty flags is enough, so this patch pick out the
these buffers and keep their BH_Mapped flag.

Fixes: 904cdbd41d74 ("jbd2: clear dirty flag when revoking a buffer from an older transaction")
Signed-off-by: zhangyi (F) <[email protected]>
---
fs/jbd2/commit.c | 25 +++++++++++++++++++++----
1 file changed, 21 insertions(+), 4 deletions(-)

diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c
index 6396fe70085b..27373f5792a4 100644
--- a/fs/jbd2/commit.c
+++ b/fs/jbd2/commit.c
@@ -985,12 +985,29 @@ void jbd2_journal_commit_transaction(journal_t *journal)
* pagesize and it is attached to the last partial page.
*/
if (buffer_freed(bh) && !jh->b_next_transaction) {
+ struct address_space *mapping;
+
clear_buffer_freed(bh);
clear_buffer_jbddirty(bh);
- clear_buffer_mapped(bh);
- clear_buffer_new(bh);
- clear_buffer_req(bh);
- bh->b_bdev = NULL;
+
+ /*
+ * Block device buffers need to stay mapped all the
+ * time, so it is enough to clear buffer_jbddirty and
+ * buffer_freed bits. For the file mapping buffers (i.e.
+ * journalled data) we need to unmap buffer and clear
+ * more bits. We also need to be careful about the check
+ * because the data page mapping can get cleared under
+ * out hands, which alse need not to clear more bits
+ * because the page and buffers will be freed and can
+ * never be reused once we are done with them.
+ */
+ mapping = READ_ONCE(bh->b_page->mapping);
+ if (mapping && !sb_is_blkdev_sb(mapping->host->i_sb)) {
+ clear_buffer_mapped(bh);
+ clear_buffer_new(bh);
+ clear_buffer_req(bh);
+ bh->b_bdev = NULL;
+ }
}

if (buffer_jbddirty(bh)) {
--
2.17.2

2020-02-13 06:39:59

by Zhang Yi

[permalink] [raw]
Subject: [PATCH v3 1/2] jbd2: move the clearing of b_modified flag to the journal_unmap_buffer()

There is no need to delay the clearing of b_modified flag to the
transaction committing time when unmapping the journalled buffer, so
just move it to the journal_unmap_buffer().

Signed-off-by: zhangyi (F) <[email protected]>
Reviewed-by: Jan Kara <[email protected]>
---
fs/jbd2/commit.c | 43 +++++++++++++++----------------------------
fs/jbd2/transaction.c | 10 ++++++----
2 files changed, 21 insertions(+), 32 deletions(-)

diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c
index 2494095e0340..6396fe70085b 100644
--- a/fs/jbd2/commit.c
+++ b/fs/jbd2/commit.c
@@ -976,34 +976,21 @@ void jbd2_journal_commit_transaction(journal_t *journal)
* it. */

/*
- * A buffer which has been freed while still being journaled by
- * a previous transaction.
- */
- if (buffer_freed(bh)) {
- /*
- * If the running transaction is the one containing
- * "add to orphan" operation (b_next_transaction !=
- * NULL), we have to wait for that transaction to
- * commit before we can really get rid of the buffer.
- * So just clear b_modified to not confuse transaction
- * credit accounting and refile the buffer to
- * BJ_Forget of the running transaction. If the just
- * committed transaction contains "add to orphan"
- * operation, we can completely invalidate the buffer
- * now. We are rather through in that since the
- * buffer may be still accessible when blocksize <
- * pagesize and it is attached to the last partial
- * page.
- */
- jh->b_modified = 0;
- if (!jh->b_next_transaction) {
- clear_buffer_freed(bh);
- clear_buffer_jbddirty(bh);
- clear_buffer_mapped(bh);
- clear_buffer_new(bh);
- clear_buffer_req(bh);
- bh->b_bdev = NULL;
- }
+ * A buffer which has been freed while still being journaled
+ * by a previous transaction, refile the buffer to BJ_Forget of
+ * the running transaction. If the just committed transaction
+ * contains "add to orphan" operation, we can completely
+ * invalidate the buffer now. We are rather through in that
+ * since the buffer may be still accessible when blocksize <
+ * pagesize and it is attached to the last partial page.
+ */
+ if (buffer_freed(bh) && !jh->b_next_transaction) {
+ clear_buffer_freed(bh);
+ clear_buffer_jbddirty(bh);
+ clear_buffer_mapped(bh);
+ clear_buffer_new(bh);
+ clear_buffer_req(bh);
+ bh->b_bdev = NULL;
}

if (buffer_jbddirty(bh)) {
diff --git a/fs/jbd2/transaction.c b/fs/jbd2/transaction.c
index e77a5a0b4e46..2dd848a743ed 100644
--- a/fs/jbd2/transaction.c
+++ b/fs/jbd2/transaction.c
@@ -2329,14 +2329,16 @@ static int journal_unmap_buffer(journal_t *journal, struct buffer_head *bh,
return -EBUSY;
}
/*
- * OK, buffer won't be reachable after truncate. We just set
- * j_next_transaction to the running transaction (if there is
- * one) and mark buffer as freed so that commit code knows it
- * should clear dirty bits when it is done with the buffer.
+ * OK, buffer won't be reachable after truncate. We just clear
+ * b_modified to not confuse transaction credit accounting, and
+ * set j_next_transaction to the running transaction (if there
+ * is one) and mark buffer as freed so that commit code knows
+ * it should clear dirty bits when it is done with the buffer.
*/
set_buffer_freed(bh);
if (journal->j_running_transaction && buffer_jbddirty(bh))
jh->b_next_transaction = journal->j_running_transaction;
+ jh->b_modified = 0;
spin_unlock(&journal->j_list_lock);
spin_unlock(&jh->b_state_lock);
write_unlock(&journal->j_state_lock);
--
2.17.2

2020-02-17 09:38:22

by Jan Kara

[permalink] [raw]
Subject: Re: [PATCH v3 2/2] jbd2: do not clear the BH_Mapped flag when forgetting a metadata buffer

On Thu 13-02-20 14:38:21, zhangyi (F) wrote:
> Commit 904cdbd41d74 ("jbd2: clear dirty flag when revoking a buffer from
> an older transaction") set the BH_Freed flag when forgetting a metadata
> buffer which belongs to the committing transaction, it indicate the
> committing process clear dirty bits when it is done with the buffer. But
> it also clear the BH_Mapped flag at the same time, which may trigger
> below NULL pointer oops when block_size < PAGE_SIZE.
>
> rmdir 1 kjournald2 mkdir 2
> jbd2_journal_commit_transaction
> commit transaction N
> jbd2_journal_forget
> set_buffer_freed(bh1)
> jbd2_journal_commit_transaction
> commit transaction N+1
> ...
> clear_buffer_mapped(bh1)
> ext4_getblk(bh2 ummapped)
> ...
> grow_dev_page
> init_page_buffers
> bh1->b_private=NULL
> bh2->b_private=NULL
> jbd2_journal_put_journal_head(jh1)
> __journal_remove_journal_head(hb1)
> jh1 is NULL and trigger oops
>
> *) Dir entry block bh1 and bh2 belongs to one page, and the bh2 has
> already been unmapped.
>
> For the metadata buffer we forgetting, we should always keep the mapped
> flag and clear the dirty flags is enough, so this patch pick out the
> these buffers and keep their BH_Mapped flag.
>
> Fixes: 904cdbd41d74 ("jbd2: clear dirty flag when revoking a buffer from an older transaction")
> Signed-off-by: zhangyi (F) <[email protected]>
> ---
> fs/jbd2/commit.c | 25 +++++++++++++++++++++----
> 1 file changed, 21 insertions(+), 4 deletions(-)

Thanks! The patch looks good. I have just some comment reformulation
suggestion below, otherwise feel free to add:

Reviewed-by: Jan Kara <[email protected]>

> diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c
> index 6396fe70085b..27373f5792a4 100644
> --- a/fs/jbd2/commit.c
> +++ b/fs/jbd2/commit.c
> @@ -985,12 +985,29 @@ void jbd2_journal_commit_transaction(journal_t *journal)
> * pagesize and it is attached to the last partial page.
> */
> if (buffer_freed(bh) && !jh->b_next_transaction) {
> + struct address_space *mapping;
> +
> clear_buffer_freed(bh);
> clear_buffer_jbddirty(bh);
> - clear_buffer_mapped(bh);
> - clear_buffer_new(bh);
> - clear_buffer_req(bh);
> - bh->b_bdev = NULL;
> +
> + /*
> + * Block device buffers need to stay mapped all the
> + * time, so it is enough to clear buffer_jbddirty and
> + * buffer_freed bits. For the file mapping buffers (i.e.
> + * journalled data) we need to unmap buffer and clear
> + * more bits. We also need to be careful about the check
> + * because the data page mapping can get cleared under
> + * out hands, which alse need not to clear more bits
^^^ our ^^^^ Maybe I'd rephrase this like:

... under our hands. Note that if mapping == NULL, we don't need to make
buffer unmapped because the page is already detached from the mapping and
buffers cannot get reused.

> + * because the page and buffers will be freed and can
> + * never be reused once we are done with them.
> + */
> + mapping = READ_ONCE(bh->b_page->mapping);
> + if (mapping && !sb_is_blkdev_sb(mapping->host->i_sb)) {
> + clear_buffer_mapped(bh);
> + clear_buffer_new(bh);
> + clear_buffer_req(bh);
> + bh->b_bdev = NULL;
> + }
> }

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

2020-02-17 12:11:42

by Zhang Yi

[permalink] [raw]
Subject: Re: [PATCH v3 2/2] jbd2: do not clear the BH_Mapped flag when forgetting a metadata buffer

Hi,

On 2020/2/17 17:36, Jan Kara wrote:
> On Thu 13-02-20 14:38:21, zhangyi (F) wrote:
>> Commit 904cdbd41d74 ("jbd2: clear dirty flag when revoking a buffer from
>> an older transaction") set the BH_Freed flag when forgetting a metadata
>> buffer which belongs to the committing transaction, it indicate the
>> committing process clear dirty bits when it is done with the buffer. But
>> it also clear the BH_Mapped flag at the same time, which may trigger
>> below NULL pointer oops when block_size < PAGE_SIZE.
>>
>> rmdir 1 kjournald2 mkdir 2
>> jbd2_journal_commit_transaction
>> commit transaction N
>> jbd2_journal_forget
>> set_buffer_freed(bh1)
>> jbd2_journal_commit_transaction
>> commit transaction N+1
>> ...
>> clear_buffer_mapped(bh1)
>> ext4_getblk(bh2 ummapped)
>> ...
>> grow_dev_page
>> init_page_buffers
>> bh1->b_private=NULL
>> bh2->b_private=NULL
>> jbd2_journal_put_journal_head(jh1)
>> __journal_remove_journal_head(hb1)
>> jh1 is NULL and trigger oops
>>
>> *) Dir entry block bh1 and bh2 belongs to one page, and the bh2 has
>> already been unmapped.
>>
>> For the metadata buffer we forgetting, we should always keep the mapped
>> flag and clear the dirty flags is enough, so this patch pick out the
>> these buffers and keep their BH_Mapped flag.
>>
>> Fixes: 904cdbd41d74 ("jbd2: clear dirty flag when revoking a buffer from an older transaction")
>> Signed-off-by: zhangyi (F) <[email protected]>
>> ---
>> fs/jbd2/commit.c | 25 +++++++++++++++++++++----
>> 1 file changed, 21 insertions(+), 4 deletions(-)
>
> Thanks! The patch looks good. I have just some comment reformulation
> suggestion below, otherwise feel free to add:
>
> Reviewed-by: Jan Kara <[email protected]>
>
>> diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c
>> index 6396fe70085b..27373f5792a4 100644
>> --- a/fs/jbd2/commit.c
>> +++ b/fs/jbd2/commit.c
>> @@ -985,12 +985,29 @@ void jbd2_journal_commit_transaction(journal_t *journal)
>> * pagesize and it is attached to the last partial page.
>> */
>> if (buffer_freed(bh) && !jh->b_next_transaction) {
>> + struct address_space *mapping;
>> +
>> clear_buffer_freed(bh);
>> clear_buffer_jbddirty(bh);
>> - clear_buffer_mapped(bh);
>> - clear_buffer_new(bh);
>> - clear_buffer_req(bh);
>> - bh->b_bdev = NULL;
>> +
>> + /*
>> + * Block device buffers need to stay mapped all the
>> + * time, so it is enough to clear buffer_jbddirty and
>> + * buffer_freed bits. For the file mapping buffers (i.e.
>> + * journalled data) we need to unmap buffer and clear
>> + * more bits. We also need to be careful about the check
>> + * because the data page mapping can get cleared under
>> + * out hands, which alse need not to clear more bits
> ^^^ our ^^^^ Maybe I'd rephrase this like:
>
> ... under our hands. Note that if mapping == NULL, we don't need to make
> buffer unmapped because the page is already detached from the mapping and
> buffers cannot get reused.
>
Thanks for your suggestion, Ted has already pushed this patch to upstream,
I could write an appending patch to fix this.

Thanks,
Yi.

2020-02-18 05:18:35

by Ritesh Harjani

[permalink] [raw]
Subject: Re: [PATCH v3 2/2] jbd2: do not clear the BH_Mapped flag when forgetting a metadata buffer



On 2/13/20 12:08 PM, zhangyi (F) wrote:
> Commit 904cdbd41d74 ("jbd2: clear dirty flag when revoking a buffer from
> an older transaction") set the BH_Freed flag when forgetting a metadata
> buffer which belongs to the committing transaction, it indicate the
> committing process clear dirty bits when it is done with the buffer. But
> it also clear the BH_Mapped flag at the same time, which may trigger
> below NULL pointer oops when block_size < PAGE_SIZE.
>
> rmdir 1 kjournald2 mkdir 2
> jbd2_journal_commit_transaction
> commit transaction N
> jbd2_journal_forget
> set_buffer_freed(bh1)
> jbd2_journal_commit_transaction
> commit transaction N+1
> ...
> clear_buffer_mapped(bh1)
> ext4_getblk(bh2 ummapped)
> ...
> grow_dev_page
> init_page_buffers
> bh1->b_private=NULL
> bh2->b_private=NULL
> jbd2_journal_put_journal_head(jh1)
> __journal_remove_journal_head(hb1)
> jh1 is NULL and trigger oops
>
> *) Dir entry block bh1 and bh2 belongs to one page, and the bh2 has
> already been unmapped.
>
> For the metadata buffer we forgetting, we should always keep the mapped
> flag and clear the dirty flags is enough, so this patch pick out the
> these buffers and keep their BH_Mapped flag.
>
> Fixes: 904cdbd41d74 ("jbd2: clear dirty flag when revoking a buffer from an older transaction")
> Signed-off-by: zhangyi (F) <[email protected]>

This should be a stable candidate I guess.

-ritesh

> ---
> fs/jbd2/commit.c | 25 +++++++++++++++++++++----
> 1 file changed, 21 insertions(+), 4 deletions(-)
>
> diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c
> index 6396fe70085b..27373f5792a4 100644
> --- a/fs/jbd2/commit.c
> +++ b/fs/jbd2/commit.c
> @@ -985,12 +985,29 @@ void jbd2_journal_commit_transaction(journal_t *journal)
> * pagesize and it is attached to the last partial page.
> */
> if (buffer_freed(bh) && !jh->b_next_transaction) {
> + struct address_space *mapping;
> +
> clear_buffer_freed(bh);
> clear_buffer_jbddirty(bh);
> - clear_buffer_mapped(bh);
> - clear_buffer_new(bh);
> - clear_buffer_req(bh);
> - bh->b_bdev = NULL;
> +
> + /*
> + * Block device buffers need to stay mapped all the
> + * time, so it is enough to clear buffer_jbddirty and
> + * buffer_freed bits. For the file mapping buffers (i.e.
> + * journalled data) we need to unmap buffer and clear
> + * more bits. We also need to be careful about the check
> + * because the data page mapping can get cleared under
> + * out hands, which alse need not to clear more bits
> + * because the page and buffers will be freed and can
> + * never be reused once we are done with them.
> + */
> + mapping = READ_ONCE(bh->b_page->mapping);
> + if (mapping && !sb_is_blkdev_sb(mapping->host->i_sb)) {
> + clear_buffer_mapped(bh);
> + clear_buffer_new(bh);
> + clear_buffer_req(bh);
> + bh->b_bdev = NULL;
> + }
> }
>
> if (buffer_jbddirty(bh)) {
>

2020-02-18 16:47:53

by Theodore Ts'o

[permalink] [raw]
Subject: Re: [PATCH v3 2/2] jbd2: do not clear the BH_Mapped flag when forgetting a metadata buffer

On Tue, Feb 18, 2020 at 10:48:13AM +0530, Ritesh Harjani wrote:
> > Fixes: 904cdbd41d74 ("jbd2: clear dirty flag when revoking a buffer from an older transaction")
> > Signed-off-by: zhangyi (F) <[email protected]>
>
> This should be a stable candidate I guess.

I took care of that before sending a pull request to Linus. :-)

- Ted

2020-02-18 17:05:27

by Theodore Ts'o

[permalink] [raw]
Subject: Re: [PATCH v3 2/2] jbd2: do not clear the BH_Mapped flag when forgetting a metadata buffer

On Mon, Feb 17, 2020 at 06:38:23PM +0800, zhangyi (F) wrote:
> >> + /*
> >> + * Block device buffers need to stay mapped all the
> >> + * time, so it is enough to clear buffer_jbddirty and
> >> + * buffer_freed bits. For the file mapping buffers (i.e.
> >> + * journalled data) we need to unmap buffer and clear
> >> + * more bits. We also need to be careful about the check
> >> + * because the data page mapping can get cleared under
> >> + * out hands, which alse need not to clear more bits
> > ^^^ our ^^^^ Maybe I'd rephrase this like:
> >
> > ... under our hands. Note that if mapping == NULL, we don't need to make
> > buffer unmapped because the page is already detached from the mapping and
> > buffers cannot get reused.
> >
> Thanks for your suggestion, Ted has already pushed this patch to upstream,
> I could write an appending patch to fix this.

Feel free to send a patch to fix up the comment if you like.

Thanks,

- Ted