[PATCH] src/fssum: skip subvolumes when building a sum

[PATCH] src/fssum: skip subvolumes when building a sum

[Josef Bacik]

With the snapshot/subvolume support added to fsstress I've been seeing
random failures with our send/receive related tests.  This is because
fssum is summing the path with the subvolumes for our test fs'es that
are generated by fsstress.  But with send/receive it skips subvolumes,
which makes the sums mismatch.  Fix this by skipping directories that do
not match our st_dev, which is how we differentiate subvolumes in btrfs.

Signed-off-by: Josef Bacik <josef@toxicpanda.com>
---
 src/fssum.c | 11 +++++++++++
 1 file changed, 11 insertions(+)

diff --git a/src/fssum.c b/src/fssum.c
index 6ba0a95c..a243839a 100644
--- a/src/fssum.c
+++ b/src/fssum.c
@@ -517,6 +517,12 @@ sum(int dirfd, int level, sum_t *dircs, char *path_prefix, char *path_in)
 	int excl;
 	sum_file_data_t sum_file_data = flags[FLAG_STRUCTURE] ?
 			sum_file_data_strict : sum_file_data_permissive;
+	struct stat64 dir_st;
+
+	if (fstat64(dirfd, &dir_st)) {
+		perror("fstat");
+		exit(-1);
+	}
 
 	d = fdopendir(dirfd);
 	if (!d) {
@@ -570,6 +576,11 @@ sum(int dirfd, int level, sum_t *dircs, char *path_prefix, char *path_in)
 				path_prefix, path, strerror(errno));
 			exit(-1);
 		}
+
+		/* We are crossing into a different subvol, skip this subtree. */
+		if (st.st_dev != dir_st.st_dev)
+			goto next;
+
 		sum_add_u64(&meta, level);
 		sum_add(&meta, namelist[i], strlen(namelist[i]));
 		if (!S_ISDIR(st.st_mode))
-- 
2.23.0

Re: [PATCH] src/fssum: skip subvolumes when building a sum

[Filipe Manana]

On Tue, Dec 17, 2019 at 3:25 PM Josef Bacik <josef@toxicpanda.com> wrote:
>
> With the snapshot/subvolume support added to fsstress I've been seeing
> random failures with our send/receive related tests.  This is because
> fssum is summing the path with the subvolumes for our test fs'es that
> are generated by fsstress.  But with send/receive it skips subvolumes,
> which makes the sums mismatch.  Fix this by skipping directories that do
> not match our st_dev, which is how we differentiate subvolumes in btrfs.
>
> Signed-off-by: Josef Bacik <josef@toxicpanda.com>

Looks good, thanks.

Reviewed-by: Filipe Manana <fdmanana@suse.com>

> ---
>  src/fssum.c | 11 +++++++++++
>  1 file changed, 11 insertions(+)
>
> diff --git a/src/fssum.c b/src/fssum.c
> index 6ba0a95c..a243839a 100644
> --- a/src/fssum.c
> +++ b/src/fssum.c
> @@ -517,6 +517,12 @@ sum(int dirfd, int level, sum_t *dircs, char *path_prefix, char *path_in)
>         int excl;
>         sum_file_data_t sum_file_data = flags[FLAG_STRUCTURE] ?
>                         sum_file_data_strict : sum_file_data_permissive;
> +       struct stat64 dir_st;
> +
> +       if (fstat64(dirfd, &dir_st)) {
> +               perror("fstat");
> +               exit(-1);
> +       }
>
>         d = fdopendir(dirfd);
>         if (!d) {
> @@ -570,6 +576,11 @@ sum(int dirfd, int level, sum_t *dircs, char *path_prefix, char *path_in)
>                                 path_prefix, path, strerror(errno));
>                         exit(-1);
>                 }
> +
> +               /* We are crossing into a different subvol, skip this subtree. */
> +               if (st.st_dev != dir_st.st_dev)
> +                       goto next;
> +
>                 sum_add_u64(&meta, level);
>                 sum_add(&meta, namelist[i], strlen(namelist[i]));
>                 if (!S_ISDIR(st.st_mode))
> --
> 2.23.0
>


-- 
Filipe David Manana,

“Whether you think you can, or you think you can't — you're right.”