@@ -63,7 +63,7 @@ _require_scratch
# interested in contains our project name in the first field. For "df"
# it contains our project directory in the last field.
# But if the device name is too long, the "df" output is broke into two
-# lines, the fourth field is not correct, so take $(NF-2) of "df"
+# lines, the second field is not correct, so take $(NF-4) of "df".
_filter_quota_rpt() {
awk '
BEGIN {
@@ -89,7 +89,7 @@ _filter_quota_rpt() {
bsize = byte_size($4);
} else if ($NF ~ proj_dir) {
# this is the "df" output
- bsize = byte_size($(NF-2));
+ bsize = byte_size($(NF-4));
} else {
next;
}
Since xfstests commit 5568077041bb ("generic/386: Fix _filter_quota_rpt"), this case outputs things. And it fails on ext4 as below: -------------------------- QA output created by 386 Silence is golden. hard limit 524283904 bytes, expected 524288000 hard limit 500 bytes, expected 524288000 -------------------------- the df command doesn't get expected output, origial output as below: -------------------------- for df: Filesystem 1K-blocks Used Available Use% Pathname /dev/sda11 20511312 45108 20466204 0% /mnt/xfstests/scratch /dev/sda11 512000 4 511996 0% /mnt/xfstests/scratch/test for df -h command: Filesystem Size Used Avail Use% Pathname /dev/sda11 19.6G 44.1M 19.5G 0% /mnt/xfstests/scratch /dev/sda11 500M 4K 500.0M 0% /mnt/xfstests/scratch/test -------------------------- It fails because ext4 project quota hides on common inode instead of superblock inode, so it has used 1 block. Also, the "df -h" fails because byte_size() only can parse integer value correctly. So, I think we can get size/1k-blocks instead of Avail/Available field in this case. Signed-off-by: Yang Xu <xuyang2018.jy@cn.fujitsu.com> --- tests/generic/386 | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)