Bug 12946

Summary: [2.6 M3 RC1] e2fsprogs ptest failed
Product: [Build System, Metadata & Runtime] OE-Core Reporter: Manish Das <manish.das>
Component: coreAssignee: Changqing Li <changqing.li>
Status: RESOLVED FIXED QA Contact:
Severity: normal    
Priority: Medium+ CC: anuj.mittal, meta.mr.watcher, meta.watcher, ross.burton
Version: 2.6   
Target Milestone: 2.6 M4   
Hardware: x86   
OS: Multiple   
Whiteboard:
OS type for building Yocto: --- Type of Regression: ---
Verified: Documentation change: No (bug/feature does not impact docs)

Description Manish Das 2018-10-03 09:51:06 UTC 
Performed diff and found that e2fsprogs tests that passed in 2.6 M2 rc1 are failing on 2.6 M3 rc1

--------------------------------------------------------------------------------------
Package Name:  e2fsprogs
--------------------------------------------------------------------------------------
-PASS: d_bad_ostype: handle bad (negative) os_type: ok

+FAIL: d_bad_ostype: handle bad (negative) os_type: failed

-PASS: d_dumpe2fs_group_only: dumpe2fs group only mode: ok

+FAIL: d_dumpe2fs_group_only: dumpe2fs group only mode: failed

-PASS: d_fallocate: fallocate sparse files and big files: ok

+FAIL: d_fallocate: fallocate sparse files and big files: failed

-PASS: d_fallocate_bigalloc: fallocate sparse files and big files with bigalloc: ok

+FAIL: d_fallocate_bigalloc: fallocate sparse files and big files with bigalloc: failed

-PASS: d_fallocate_blkmap: fallocate sparse files and big files on a blockmap fs: ok

+FAIL: d_fallocate_blkmap: fallocate sparse files and big files on a blockmap fs: failed

-PASS: d_inline_dump: debugfs dump inline data test: ok

+FAIL: d_inline_dump: debugfs dump inline data test: failed

-PASS: d_loaddump: debugfs load/dump test: ok

+FAIL: d_loaddump: debugfs load/dump test: failed

-PASS: d_punch: punch sparse files and big files: ok

+FAIL: d_punch: punch sparse files and big files: failed

-PASS: d_punch_bigalloc: punch sparse files and big files with bigalloc: ok

+FAIL: d_punch_bigalloc: punch sparse files and big files with bigalloc: failed

-PASS: d_special_files: create special files in debugfs: ok

+FAIL: d_special_files: create special files in debugfs: failed

-PASS: d_xattr_edits: edit extended attributes in debugfs: ok

+FAIL: d_xattr_edits: edit extended attributes in debugfs: failed

-PASS: d_xattr_sorting: sort extended attributes in debugfs: ok

+FAIL: d_xattr_sorting: sort extended attributes in debugfs: failed

-PASS: e_icount_normal: inode counting structure optimized for low counts: ok

+FAIL: e_icount_normal: inode counting structure optimized for low counts: failed

-PASS: e_icount_opt: inode counting abstraction optimized for counting: ok

+FAIL: e_icount_opt: inode counting abstraction optimized for counting: failed

-PASS: f_badcluster: test alignment problems with bigalloc clusters: ok

+FAIL: f_badcluster: test alignment problems with bigalloc clusters: failed

-PASS: f_bigalloc_badinode: delete bad inode handling for bigalloc filesystems: ok

+FAIL: f_bigalloc_badinode: delete bad inode handling for bigalloc filesystems: failed

-PASS: f_bigalloc_orphan_list: orphaned list handling with bigalloc file systems: ok

+FAIL: f_bigalloc_orphan_list: orphaned list handling with bigalloc file systems: failed

-PASS: f_collapse_extent_tree: extent tree can be collapsed one level: ok

+FAIL: f_collapse_extent_tree: extent tree can be collapsed one level: failed

-PASS: f_compress_extent_tree_level: compress an extent tree level: ok

+FAIL: f_compress_extent_tree_level: compress an extent tree level: failed

-PASS: f_convert_bmap: convert blockmap file to extents file: ok

+FAIL: f_convert_bmap: convert blockmap file to extents file: failed

-PASS: f_convert_bmap_and_extent: convert blockmap to extents files: ok

+FAIL: f_convert_bmap_and_extent: convert blockmap to extents files: failed

-PASS: f_convert_bmap_sparse: convert sparse blockmap file to extents file: ok

+FAIL: f_convert_bmap_sparse: convert sparse blockmap file to extents file: failed

-PASS: f_create_symlinks: create fast, inlinedata, and regular symlinks: ok

+FAIL: f_create_symlinks: create fast, inlinedata, and regular symlinks: failed

-PASS: f_desc_size_bad: bad superblock s_desc_size: ok

+FAIL: f_desc_size_bad: bad superblock s_desc_size: failed

-PASS: f_detect_xfs: detect xfs filesystem: ok

+FAIL: f_detect_xfs: detect xfs filesystem: failed

-PASS: f_dup4: find all directory pathnames: ok

+FAIL: f_dup4: find all directory pathnames: failed

-PASS: f_dup_resize: blocks claimed by the resize inode and another inode: ok

+FAIL: f_dup_resize: blocks claimed by the resize inode and another inode: failed

-PASS: f_extent_htree: htree extent compression: ok

+FAIL: f_extent_htree: htree extent compression: failed

-PASS: f_extent_oobounds: ok

+FAIL: f_extent_oobounds: failed

-PASS: f_extent_too_deep: extent tree is deeper than it needs to be: ok

+FAIL: f_extent_too_deep: extent tree is deeper than it needs to be: failed

-PASS: f_imagic_fs: imagic filesystem with imagic inodes: ok

+FAIL: f_imagic_fs: imagic filesystem with imagic inodes: failed

-PASS: f_jnl_32bit: on-disk 32-bit journal format: ok

+FAIL: f_jnl_32bit: on-disk 32-bit journal format: failed

-PASS: f_jnl_64bit: on-disk 64-bit journal format: ok

+FAIL: f_jnl_64bit: on-disk 64-bit journal format: failed

-PASS: f_large_dir: optimize 3 level htree directories: ok

+FAIL: f_large_dir: optimize 3 level htree directories: failed

---------------------------------------------------------------------------------------
Total number of test cases passing in Previous release but failing in Current release:  34
---------------------------------------------------------------------------------------
Comment 1 Anuj Mittal 2018-10-04 07:39:52 UTC 
Happens because a missing quote in test script. Patch sent to list.
Comment 2 Ross Burton 2018-10-04 11:09:13 UTC 
With that fix and e2fsprogs-ptest in a core-image-minimal:

root@bob:~# ptest-runner e2fsprogs
START: ptest-runner
2018-10-04T11:23
BEGIN: /usr/lib/e2fsprogs/ptest
sed: invalid option -- 'u'
BusyBox v1.29.2 (2018-10-04 10:05:58 UTC) multi-call binary.

(looks like we need to depend on sed, or patch out -u)

debugfs 1.44.3 (10-July-2018)
grep: ../lib/config.h: No such file or directory
/usr/lib/e2fsprogs/ptest/test/f_detect_junk/script: line 3: [: : integer expression expected

([ error is because the file wasn't present)

f_large_dir: 5000/65010 links, 16s @ 312/s
f_large_dir: 10000/65010 links, 39s @ 128/s
f_large_dir: 15000/65010 links, 61s @ 81/s
f_large_dir: 20000/65010 links, 87s @ 57/s
f_large_dir: 25000/65010 links, 109s @ 45/s
f_large_dir: 30000/65010 links, 133s @ 37/s
f_large_dir: 35000/65010 links, 152s @ 32/s
f_large_dir: 40000/65010 links, 176s @ 28/s
f_large_dir: 45000/65010 links, 200s @ 25/s
f_large_dir: 50000/65010 links, 228s @ 21/s
f_large_dir: 55000/65010 links, 251s @ 19/s
f_large_dir: 60000/65010 links, 277s @ 18/s

ERROR: Exit status is 9
TIMEOUT: /usr/lib/e2fsprogs/ptest

The f_large_dir test takes so long, that ptest times out.

I'm wondering if the e2fsprogs test is actually useful. Should we just remove it?