[PATCH 1/2] [v2] btrfs: use monotonic time for transaction handling

[PATCH 1/2] [v2] btrfs: use monotonic time for transaction handling

[Arnd Bergmann]

The transaction times were changed to ktime_get_real_seconds to avoid
the y2038 overflow, but they still have a minor problem when they go
backwards or jump due to settimeofday() or leap seconds.

This changes the transaction handling to instead use ktime_get_seconds(),
which returns a CLOCK_MONOTONIC timestamp that has neither of those
problems.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
---
v2: Rebased on top of an earlier patch by Allen Pais, which
    also obsoleted the original patch 2/3
---
 fs/btrfs/disk-io.c     | 2 +-
 fs/btrfs/transaction.c | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index f3224e23d5fa..bf0717f2824d 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -1819,7 +1819,7 @@ static int transaction_kthread(void *arg)
 			goto sleep;
 		}
 
-		now = ktime_get_real_seconds();
+		now = ktime_get_seconds();
 		if (cur->state < TRANS_STATE_BLOCKED &&
 		    !test_bit(BTRFS_FS_NEED_ASYNC_COMMIT, &fs_info->flags) &&
 		    (now < cur->start_time ||
diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c
index 56c8bab0b816..ebe50dfb8947 100644
--- a/fs/btrfs/transaction.c
+++ b/fs/btrfs/transaction.c
@@ -241,7 +241,7 @@ static noinline int join_transaction(struct btrfs_fs_info *fs_info,
 	refcount_set(&cur_trans->use_count, 2);
 	atomic_set(&cur_trans->pending_ordered, 0);
 	cur_trans->flags = 0;
-	cur_trans->start_time = ktime_get_real_seconds();
+	cur_trans->start_time = ktime_get_seconds();
 
 	memset(&cur_trans->delayed_refs, 0, sizeof(cur_trans->delayed_refs));
 
-- 
2.9.0

[PATCH 2/2] [v2] btrfs: use timespec64 for i_otime

[Arnd Bergmann]

While the regular inode timestamps all use timespec64 now, the
i_otime field is btrfs specific and still needs to be converted
to correctly represent times beyond 2038.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
---
v2: no changes
---
 fs/btrfs/btrfs_inode.h | 2 +-
 fs/btrfs/inode.c       | 4 ++--
 2 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h
index 7e075343daa5..1343ac57b438 100644
--- a/fs/btrfs/btrfs_inode.h
+++ b/fs/btrfs/btrfs_inode.h
@@ -178,7 +178,7 @@ struct btrfs_inode {
 	struct btrfs_delayed_node *delayed_node;
 
 	/* File creation time. */
-	struct timespec i_otime;
+	struct timespec64 i_otime;
 
 	/* Hook into fs_info->delayed_iputs */
 	struct list_head delayed_iput;
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index e9482f0db9d0..22dcc8afd38f 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -5745,7 +5745,7 @@ static struct inode *new_simple_dir(struct super_block *s,
 	inode->i_mtime = current_time(inode);
 	inode->i_atime = inode->i_mtime;
 	inode->i_ctime = inode->i_mtime;
-	BTRFS_I(inode)->i_otime = timespec64_to_timespec(inode->i_mtime);
+	BTRFS_I(inode)->i_otime = inode->i_mtime;
 
 	return inode;
 }
@@ -6349,7 +6349,7 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans,
 	inode->i_mtime = current_time(inode);
 	inode->i_atime = inode->i_mtime;
 	inode->i_ctime = inode->i_mtime;
-	BTRFS_I(inode)->i_otime = timespec64_to_timespec(inode->i_mtime);
+	BTRFS_I(inode)->i_otime = inode->i_mtime;
 
 	inode_item = btrfs_item_ptr(path->nodes[0], path->slots[0],
 				  struct btrfs_inode_item);
-- 
2.9.0

Re: [PATCH 2/2] [v2] btrfs: use timespec64 for i_otime

[Nikolay Borisov]

On 21.06.2018 19:04, Arnd Bergmann wrote:
> While the regular inode timestamps all use timespec64 now, the
> i_otime field is btrfs specific and still needs to be converted
> to correctly represent times beyond 2038.
> 
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>

Reviewed-by: Nikolay Borisov <nborisov@suse.com>

> ---
> v2: no changes
> ---
>  fs/btrfs/btrfs_inode.h | 2 +-
>  fs/btrfs/inode.c       | 4 ++--
>  2 files changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h
> index 7e075343daa5..1343ac57b438 100644
> --- a/fs/btrfs/btrfs_inode.h
> +++ b/fs/btrfs/btrfs_inode.h
> @@ -178,7 +178,7 @@ struct btrfs_inode {
>  	struct btrfs_delayed_node *delayed_node;
>  
>  	/* File creation time. */
> -	struct timespec i_otime;
> +	struct timespec64 i_otime;
>  
>  	/* Hook into fs_info->delayed_iputs */
>  	struct list_head delayed_iput;
> diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
> index e9482f0db9d0..22dcc8afd38f 100644
> --- a/fs/btrfs/inode.c
> +++ b/fs/btrfs/inode.c
> @@ -5745,7 +5745,7 @@ static struct inode *new_simple_dir(struct super_block *s,
>  	inode->i_mtime = current_time(inode);
>  	inode->i_atime = inode->i_mtime;
>  	inode->i_ctime = inode->i_mtime;
> -	BTRFS_I(inode)->i_otime = timespec64_to_timespec(inode->i_mtime);
> +	BTRFS_I(inode)->i_otime = inode->i_mtime;
>  
>  	return inode;
>  }
> @@ -6349,7 +6349,7 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans,
>  	inode->i_mtime = current_time(inode);
>  	inode->i_atime = inode->i_mtime;
>  	inode->i_ctime = inode->i_mtime;
> -	BTRFS_I(inode)->i_otime = timespec64_to_timespec(inode->i_mtime);
> +	BTRFS_I(inode)->i_otime = inode->i_mtime;
>  
>  	inode_item = btrfs_item_ptr(path->nodes[0], path->slots[0],
>  				  struct btrfs_inode_item);
> 

Re: [PATCH 1/2] [v2] btrfs: use monotonic time for transaction handling

[Nikolay Borisov]

On 21.06.2018 19:04, Arnd Bergmann wrote:
> The transaction times were changed to ktime_get_real_seconds to avoid
> the y2038 overflow, but they still have a minor problem when they go
> backwards or jump due to settimeofday() or leap seconds.
> 
> This changes the transaction handling to instead use ktime_get_seconds(),
> which returns a CLOCK_MONOTONIC timestamp that has neither of those
> problems.
> 
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>

Reviewed-by: Nikolay Borisov <nborisov@suse.com>

> ---
> v2: Rebased on top of an earlier patch by Allen Pais, which
>     also obsoleted the original patch 2/3
> ---
>  fs/btrfs/disk-io.c     | 2 +-
>  fs/btrfs/transaction.c | 2 +-
>  2 files changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
> index f3224e23d5fa..bf0717f2824d 100644
> --- a/fs/btrfs/disk-io.c
> +++ b/fs/btrfs/disk-io.c
> @@ -1819,7 +1819,7 @@ static int transaction_kthread(void *arg)
>  			goto sleep;
>  		}
>  
> -		now = ktime_get_real_seconds();
> +		now = ktime_get_seconds();
>  		if (cur->state < TRANS_STATE_BLOCKED &&
>  		    !test_bit(BTRFS_FS_NEED_ASYNC_COMMIT, &fs_info->flags) &&
>  		    (now < cur->start_time ||
> diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c
> index 56c8bab0b816..ebe50dfb8947 100644
> --- a/fs/btrfs/transaction.c
> +++ b/fs/btrfs/transaction.c
> @@ -241,7 +241,7 @@ static noinline int join_transaction(struct btrfs_fs_info *fs_info,
>  	refcount_set(&cur_trans->use_count, 2);
>  	atomic_set(&cur_trans->pending_ordered, 0);
>  	cur_trans->flags = 0;
> -	cur_trans->start_time = ktime_get_real_seconds();
> +	cur_trans->start_time = ktime_get_seconds();
>  
>  	memset(&cur_trans->delayed_refs, 0, sizeof(cur_trans->delayed_refs));
>  
> 

Re: [PATCH 1/2] [v2] btrfs: use monotonic time for transaction handling

[David Sterba]

On Thu, Jun 21, 2018 at 06:04:05PM +0200, Arnd Bergmann wrote:
> The transaction times were changed to ktime_get_real_seconds to avoid
> the y2038 overflow, but they still have a minor problem when they go
> backwards or jump due to settimeofday() or leap seconds.
> 
> This changes the transaction handling to instead use ktime_get_seconds(),
> which returns a CLOCK_MONOTONIC timestamp that has neither of those
> problems.
> 
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
> ---
> v2: Rebased on top of an earlier patch by Allen Pais, which
>     also obsoleted the original patch 2/3

1 and 2 added to patch queue, thanks.