[PATCH] statx: add STATX_ATTR_MULTIGRAIN_TIMESTAMP

[PATCH] statx: add STATX_ATTR_MULTIGRAIN_TIMESTAMP

[Jeff Layton]

Userland has no way to tell whether a file's timestamps are multigrain,
i.e. whether successive changes after a query are guaranteed to yield
distinct timestamps.

Add STATX_ATTR_MULTIGRAIN_TIMESTAMP and set it in vfs_getattr_nosec()
when is_mgtime() is true (IOP_MGTIME is set from FS_MGTIME at inode
init). Always advertise it in attributes_mask.

Suggested-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jeff Layton <jlayton@kernel.org>
---
HCH suggested this when I posted the fstest for multigrain timestamps. I
think it makes sense. If this looks good, I'll plan to send a manpage
update for statx(2).
---
 fs/stat.c                 | 6 +++++-
 include/uapi/linux/stat.h | 1 +
 2 files changed, 6 insertions(+), 1 deletion(-)

diff --git a/fs/stat.c b/fs/stat.c
index a2e63c6c45b7..67b9f1784b23 100644
--- a/fs/stat.c
+++ b/fs/stat.c
@@ -206,9 +206,13 @@ int vfs_getattr_nosec(const struct path *path, struct kstat *stat,
 	if (IS_VERITY(inode))
 		stat->attributes |= STATX_ATTR_VERITY;
 
+	if (is_mgtime(inode))
+		stat->attributes |= STATX_ATTR_MULTIGRAIN_TIMESTAMP;
+
 	stat->attributes_mask |= (STATX_ATTR_AUTOMOUNT |
 				  STATX_ATTR_DAX |
-				  STATX_ATTR_VERITY);
+				  STATX_ATTR_VERITY |
+				  STATX_ATTR_MULTIGRAIN_TIMESTAMP);
 
 	idmap = mnt_idmap(path->mnt);
 	if (inode->i_op->getattr) {
diff --git a/include/uapi/linux/stat.h b/include/uapi/linux/stat.h
index 1686861aae20..b978860590b1 100644
--- a/include/uapi/linux/stat.h
+++ b/include/uapi/linux/stat.h
@@ -255,6 +255,7 @@ struct statx {
 #define STATX_ATTR_VERITY		0x00100000 /* [I] Verity protected file */
 #define STATX_ATTR_DAX			0x00200000 /* File is currently in DAX state */
 #define STATX_ATTR_WRITE_ATOMIC		0x00400000 /* File supports atomic write operations */
+#define STATX_ATTR_MULTIGRAIN_TIMESTAMP	0x00800000 /* File uses multigrain timestamps */
 
 
 #endif /* _UAPI_LINUX_STAT_H */

---
base-commit: c339f4010dea8b0d754f84ace70cd1b089878ef8
change-id: 20261029-mgtime-statx-26a229f9c3b4

Best regards,
-- 
Jeff Layton <jlayton@kernel.org>

Re: [PATCH] statx: add STATX_ATTR_MULTIGRAIN_TIMESTAMP

[Jan Kara]

On Thu 29-10-26 12:28:16, Jeff Layton wrote:
> Userland has no way to tell whether a file's timestamps are multigrain,
> i.e. whether successive changes after a query are guaranteed to yield
> distinct timestamps.
> 
> Add STATX_ATTR_MULTIGRAIN_TIMESTAMP and set it in vfs_getattr_nosec()
> when is_mgtime() is true (IOP_MGTIME is set from FS_MGTIME at inode
> init). Always advertise it in attributes_mask.
> 
> Suggested-by: Christoph Hellwig <hch@lst.de>
> Signed-off-by: Jeff Layton <jlayton@kernel.org>

Makes sense to me. Feel free to add:

Reviewed-by: Jan Kara <jack@suse.cz>

								Honza

> ---
> HCH suggested this when I posted the fstest for multigrain timestamps. I
> think it makes sense. If this looks good, I'll plan to send a manpage
> update for statx(2).
> ---
>  fs/stat.c                 | 6 +++++-
>  include/uapi/linux/stat.h | 1 +
>  2 files changed, 6 insertions(+), 1 deletion(-)
> 
> diff --git a/fs/stat.c b/fs/stat.c
> index a2e63c6c45b7..67b9f1784b23 100644
> --- a/fs/stat.c
> +++ b/fs/stat.c
> @@ -206,9 +206,13 @@ int vfs_getattr_nosec(const struct path *path, struct kstat *stat,
>  	if (IS_VERITY(inode))
>  		stat->attributes |= STATX_ATTR_VERITY;
>  
> +	if (is_mgtime(inode))
> +		stat->attributes |= STATX_ATTR_MULTIGRAIN_TIMESTAMP;
> +
>  	stat->attributes_mask |= (STATX_ATTR_AUTOMOUNT |
>  				  STATX_ATTR_DAX |
> -				  STATX_ATTR_VERITY);
> +				  STATX_ATTR_VERITY |
> +				  STATX_ATTR_MULTIGRAIN_TIMESTAMP);
>  
>  	idmap = mnt_idmap(path->mnt);
>  	if (inode->i_op->getattr) {
> diff --git a/include/uapi/linux/stat.h b/include/uapi/linux/stat.h
> index 1686861aae20..b978860590b1 100644
> --- a/include/uapi/linux/stat.h
> +++ b/include/uapi/linux/stat.h
> @@ -255,6 +255,7 @@ struct statx {
>  #define STATX_ATTR_VERITY		0x00100000 /* [I] Verity protected file */
>  #define STATX_ATTR_DAX			0x00200000 /* File is currently in DAX state */
>  #define STATX_ATTR_WRITE_ATOMIC		0x00400000 /* File supports atomic write operations */
> +#define STATX_ATTR_MULTIGRAIN_TIMESTAMP	0x00800000 /* File uses multigrain timestamps */
>  
>  
>  #endif /* _UAPI_LINUX_STAT_H */
> 
> ---
> base-commit: c339f4010dea8b0d754f84ace70cd1b089878ef8
> change-id: 20261029-mgtime-statx-26a229f9c3b4
> 
> Best regards,
> -- 
> Jeff Layton <jlayton@kernel.org>
> 
-- 
Jan Kara <jack@suse.com>
SUSE Labs, CR

Re: [PATCH] statx: add STATX_ATTR_MULTIGRAIN_TIMESTAMP

[Christoph Hellwig]

Looks good:

Reviewed-by: Christoph Hellwig <hch@lst.de>