From: Tomas Racek <tracek@xxxxxxxxxx> It is just a resend of this patch from "Tomas Racek <tracek@xxxxxxxxxx>". Recently we run xfstests on ext4 with 'bigalloc' feature enabled, and come across some failure due to poor adaption for ext4 bigalloc. One if the failed cases is shared/298. I find this patch on internet [1] and it works in my case. I have no idea why this patch have not been merged. Maybe this buddy didn't send this patch at that time, or it was rejected for some reason but I can't find any discussion on internet. [1] https://lkml.org/lkml/2013/6/18/329 The original commit log: Count with cluster size instead of block size if bigalloc is used. Signed-off-by: Tomas Racek <tracek@xxxxxxxxxx> --- tests/shared/298 | 12 +++++++++--- 1 file changed, 9 insertions(+), 3 deletions(-) diff --git a/tests/shared/298 b/tests/shared/298 index 5d6c6cc..86b7cdc 100755 --- a/tests/shared/298 +++ b/tests/shared/298 @@ -60,15 +60,21 @@ get_free_sectors() { case $FSTYP in ext4) + cluster_size=$($DUMPE2FS_PROG $img_file 2>&1 | sed -n 's/Cluster size: *\(.*\)/\1/p') + if [ -n "$cluster_size" ]; then + blocks_per_cluster=`expr $cluster_size / $block_size` + else + blocks_per_cluster=1 + fi $UMOUNT_PROG $loop_mnt $DUMPE2FS_PROG $img_file 2>&1 | grep " Free blocks" | cut -d ":" -f2- | \ tr ',' '\n' | $SED_PROG 's/^ //' | \ - $AWK_PROG -v spb=$sectors_per_block 'BEGIN{FS="-"}; + $AWK_PROG -v spb=$sectors_per_block -v bpc=$blocks_per_cluster 'BEGIN{FS="-"}; NF { if($2 != "") # range of blocks - print spb * $1, spb * ($2 + 1) - 1; + print spb * $1, spb * ($2 + bpc) - 1; else # just single block - print spb * $1, spb * ($1 + 1) - 1; + print spb * $1, spb * ($1 + bpc) - 1; }' ;; xfs) -- 1.8.3.1
