Skip to contentSkip to conversation

Fix Several Test Failures

PATCH Superseded Andrew Hamilton 3 participants 6 messages Aug 26, 2025

Subsystems (2)

v2 of 7 Superseded Latest is v7 →

lore
Share & tools
lore
#0 lore raw

v2 cover Fix Several Test Failures

Correct several test failures observed while running GRUB 'make check' on Debian 13. The test failures fixed here are:

1. ZFS ZSTD - Test was failing zfs_test due to lack of support in zfs-fuse for zstd compression. This specific test case was split into a new test file allowing skipping the test if zstd is not supported to avoid masking other possible (future) zfs test failures.
2. Correct test failure in erofs_test in newer Linux distros caused by a change in mkfs.erofs to limit the maximum file system label length to 15 characters. Reduce the label used for testing to work within this limit.
3. Correct test failure in ext234_test in newer Linux distros caused by a change in e2fsprogs that removed the ability to set "version 0" ext2 file systems using the "-r" flag. Instead a new flag is added to support this. Update the test to support the old method in older distros and the new method in newer distros.

Changes since v1:
- Remove unrelated changes to .gitignore

Andrew Hamilton (3):
tests: Split ZFS ZSTD Test Into New File
tests: Avoid Test Failure in erofs for label length
tests: Support Changed mkfs.ext2 Behavior for -r Flag

 .gitignore                   |  1 +
 Makefile.util.def            |  6 ++++++
 tests/util/grub-fs-tester.in | 13 ++++++++++---
 tests/zfs_test.in            |  1 -
 tests/zfs_zstd_test.in       | 30 ++++++++++++++++++++++++++++++
 5 files changed, 47 insertions(+), 4 deletions(-)
 create mode 100644 tests/zfs_zstd_test.in
Signature · 7 lines
2.39.5


_______________________________________________
Grub-devel mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/grub-devel
#1 lore raw

PATCH v2 1/3 tests: Split ZFS ZSTD Test Into New File

Split ZFS ZSTD test into its own test script. Add a check to the new test script to see if the zfs utility installed on the host supports "zstd" compression before running the test and skip the test if not. It seems at least some zfs-fuse binaries do not support zstd compression and the current test will fail in that case. Splitting into a new file will avoid masking other test failures due to missing zstd support.

---
 .gitignore             |  1 +
 Makefile.util.def      |  6 ++++++
 tests/zfs_test.in      |  1 -
 tests/zfs_zstd_test.in | 30 ++++++++++++++++++++++++++++++
 4 files changed, 37 insertions(+), 1 deletion(-)
 create mode 100644 tests/zfs_zstd_test.in
.gitignore +1 -0
index 524f2e6d0..67ad2d26d 100644
@@ -283,3 +283,4 @@ widthspec.bin
 /xfs_test
 /xzcompress_test
 /zfs_test
+/zfs_zstd_test
Makefile.util.def +6 -0
index 038253b37..703e8c08d 100644
@@ -911,6 +911,12 @@ script = {
   common = tests/zfs_test.in;
 };
 
+script = {
+  testcase = native;
+  name = zfs_zstd_test;
+  common = tests/zfs_zstd_test.in;
+};
+
 script = {
   testcase = native;
   name = cpio_test;
tests/zfs_test.in +0 -1
index 0d0a57f7d..58cc25b22 100644
@@ -19,7 +19,6 @@ fi
 "@builddir@/grub-fs-tester" zfs_lzjb
 "@builddir@/grub-fs-tester" zfs_gzip
 "@builddir@/grub-fs-tester" zfs_zle
-"@builddir@/grub-fs-tester" zfs_zstd
 "@builddir@/grub-fs-tester" zfs_raidz3
 "@builddir@/grub-fs-tester" zfs_raidz2
 "@builddir@/grub-fs-tester" zfs_raidz
tests/zfs_zstd_test.in new file +30 -0
new file mode 100644
index 000000000..1b8a20212
@@ -0,0 +1,30 @@
+#!@BUILD_SHEBANG@
+
+set -e
+
+if [ "x$EUID" = "x" ] ; then
+  EUID=`id -u`
+fi
+
+if [ "$EUID" != 0 ] ; then
+   exit 99
+fi
+
+if ! which zpool >/dev/null 2>&1; then
+   echo "zpool not installed; cannot test zfs."
+   exit 99
+fi
+
+if ! which zfs >/dev/null 2>&1; then
+   echo "zfs not installed; cannot test zfs."
+   exit 99
+fi
+
+# If OpenZFS is not installed (only zfs-fuse for example) then
+# skip ZSTD compression testing.
+if ! zfs get 2>&1 | grep -F "compression" | grep -F "zstd"; then
+   echo "zfs zstd compression not supported; cannot test zfs zstd."
+   exit 77
+fi
+
+"@builddir@/grub-fs-tester" zfs_zstd
Signature · 7 lines
2.39.5


_______________________________________________
Grub-devel mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/grub-devel
#2 lore raw

Re: [PATCH v2 1/3] tests: Split ZFS ZSTD Test Into New File

Quoted 64 lines · On Mon, 25 Aug 2025 19:28:24 -0500 Andrew Hamilton <[email protected]> wrote:

Split ZFS ZSTD test into its own test script. Add a check to the new test script to see if the zfs utility installed on the host supports "zstd" compression before running the test and skip the test if not. It seems at least some zfs-fuse binaries do not support zstd compression and the current test will fail in that case. Splitting into a new file will avoid masking other test failures due to missing zstd support.

---
 .gitignore             |  1 +
 Makefile.util.def      |  6 ++++++
 tests/zfs_test.in      |  1 -
 tests/zfs_zstd_test.in | 30 ++++++++++++++++++++++++++++++
 4 files changed, 37 insertions(+), 1 deletion(-)
 create mode 100644 tests/zfs_zstd_test.in
.gitignore +1 -0
index 524f2e6d0..67ad2d26d 100644
@@ -283,3 +283,4 @@ widthspec.bin
 /xfs_test
 /xzcompress_test
 /zfs_test
+/zfs_zstd_test
Makefile.util.def +6 -0
index 038253b37..703e8c08d 100644
@@ -911,6 +911,12 @@ script = {
   common = tests/zfs_test.in;
 };
 
+script = {
+  testcase = native;
+  name = zfs_zstd_test;
+  common = tests/zfs_zstd_test.in;
+};
+
 script = {
   testcase = native;
   name = cpio_test;
tests/zfs_test.in +0 -1
index 0d0a57f7d..58cc25b22 100644
@@ -19,7 +19,6 @@ fi
 "@builddir@/grub-fs-tester" zfs_lzjb
 "@builddir@/grub-fs-tester" zfs_gzip
 "@builddir@/grub-fs-tester" zfs_zle
-"@builddir@/grub-fs-tester" zfs_zstd
 "@builddir@/grub-fs-tester" zfs_raidz3
 "@builddir@/grub-fs-tester" zfs_raidz2
 "@builddir@/grub-fs-tester" zfs_raidz
tests/zfs_zstd_test.in new file +3 -0
new file mode 100644
index 000000000..1b8a20212
@@ -0,0 +1,30 @@
+#!@BUILD_SHEBANG@
+
+set -e

tests/zfs_zstd_test.in line 13

Quoted 10 lines
tests/zfs_zstd_test.in new file +10 -0
+
+if [ "x$EUID" = "x" ] ; then
+  EUID=`id -u`
+fi
+
+if [ "$EUID" != 0 ] ; then
+   exit 99
+fi
+
+if ! which zpool >/dev/null 2>&1; then

Use builtin command -v instead of which.

Quoted 4 lines
tests/zfs_zstd_test.in new file +4 -0
+   echo "zpool not installed; cannot test zfs."
+   exit 99
+fi
+

tests/zfs_zstd_test.in line 27

Quoted 10 lines
tests/zfs_zstd_test.in new file +10 -0
+if ! which zfs >/dev/null 2>&1; then
+   echo "zfs not installed; cannot test zfs."
+   exit 99
+fi
+
+# If OpenZFS is not installed (only zfs-fuse for example) then
+# skip ZSTD compression testing.
+if ! zfs get 2>&1 | grep -F "compression" | grep -F "zstd"; then
+   echo "zfs zstd compression not supported; cannot test zfs zstd."
+   exit 77

This should exit 99 for hard error, which means that the test failed due to some reason that was not an actual failure (eg. because the test couldn't be run because of some environmental reason).

Glenn

tests/zfs_zstd_test.in line 30

Quoted 3 lines
tests/zfs_zstd_test.in new file +3 -0
+fi
+
+"@builddir@/grub-fs-tester" zfs_zstd

_______________________________________________
Grub-devel mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/grub-devel

#3 lore raw

Re: [PATCH v2 1/3] tests: Split ZFS ZSTD Test Into New File

Thank you for the review.

Quoted 3 lines

This should exit 99 for hard error, which means that the test failed due to some reason that was not an actual failure (eg. because the test couldn't be run because of some environmental reason).

Ok, I didn't realize this was the expected behavior in a case like this, I will make this change tonight in a v3.

Thanks!
Andrew

Quoted 100 lines · Glenn Washburn
Quoted 64 lines · On Mon, 25 Aug 2025 19:28:24 -0500 Andrew Hamilton <[email protected]> wrote:

Split ZFS ZSTD test into its own test script. Add a check to the new test script to see if the zfs utility installed on the host supports "zstd" compression before running the test and skip the test if not. It seems at least some zfs-fuse binaries do not support zstd compression and the current test will fail in that case. Splitting into a new file will avoid masking other test failures due to missing zstd support.

---
 .gitignore             |  1 +
 Makefile.util.def      |  6 ++++++
 tests/zfs_test.in      |  1 -
 tests/zfs_zstd_test.in | 30 ++++++++++++++++++++++++++++++
 4 files changed, 37 insertions(+), 1 deletion(-)
 create mode 100644 tests/zfs_zstd_test.in
.gitignore +1 -0
index 524f2e6d0..67ad2d26d 100644
@@ -283,3 +283,4 @@ widthspec.bin
 /xfs_test
 /xzcompress_test
 /zfs_test
+/zfs_zstd_test
Makefile.util.def +6 -0
index 038253b37..703e8c08d 100644
@@ -911,6 +911,12 @@ script = {
   common = tests/zfs_test.in;
 };
 
+script = {
+  testcase = native;
+  name = zfs_zstd_test;
+  common = tests/zfs_zstd_test.in;
+};
+
 script = {
   testcase = native;
   name = cpio_test;
tests/zfs_test.in +0 -1
index 0d0a57f7d..58cc25b22 100644
@@ -19,7 +19,6 @@ fi
 "@builddir@/grub-fs-tester" zfs_lzjb
 "@builddir@/grub-fs-tester" zfs_gzip
 "@builddir@/grub-fs-tester" zfs_zle
-"@builddir@/grub-fs-tester" zfs_zstd
 "@builddir@/grub-fs-tester" zfs_raidz3
 "@builddir@/grub-fs-tester" zfs_raidz2
 "@builddir@/grub-fs-tester" zfs_raidz
tests/zfs_zstd_test.in new file +3 -0
new file mode 100644
index 000000000..1b8a20212
@@ -0,0 +1,30 @@
+#!@BUILD_SHEBANG@
+
+set -e

tests/zfs_zstd_test.in line 13

Quoted 10 lines
tests/zfs_zstd_test.in new file +10 -0
+
+if [ "x$EUID" = "x" ] ; then
+  EUID=`id -u`
+fi
+
+if [ "$EUID" != 0 ] ; then
+   exit 99
+fi
+
+if ! which zpool >/dev/null 2>&1; then

Use builtin command -v instead of which.

Quoted 4 lines
tests/zfs_zstd_test.in new file +4 -0
+   echo "zpool not installed; cannot test zfs."
+   exit 99
+fi
+

tests/zfs_zstd_test.in line 27

Quoted 10 lines
tests/zfs_zstd_test.in new file +10 -0
+if ! which zfs >/dev/null 2>&1; then
+   echo "zfs not installed; cannot test zfs."
+   exit 99
+fi
+
+# If OpenZFS is not installed (only zfs-fuse for example) then
+# skip ZSTD compression testing.
+if ! zfs get 2>&1 | grep -F "compression" | grep -F "zstd"; then
+   echo "zfs zstd compression not supported; cannot test zfs zstd."
+   exit 77

This should exit 99 for hard error, which means that the test failed due to some reason that was not an actual failure (eg. because the test couldn't be run because of some environmental reason).

Quoted 5 lines

Glenn

tests/zfs_zstd_test.in line 30

Quoted 3 lines
tests/zfs_zstd_test.in new file +3 -0
+fi
+
+"@builddir@/grub-fs-tester" zfs_zstd

_______________________________________________
Grub-devel mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/grub-devel

#4 lore raw

PATCH v2 3/3 tests: Support Changed mkfs.ext2 Behavior for -r Flag

In reply to Andrew Hamilton

Correct nuisance ext234_test failure on newer Linux distros.

Recently, the mkfs.ext2 utility removed support for the -r flag to specify old (version 0) formats of ext2. A new flag was added to allow the same behavior. Support both ways of specifying version 0 ext2 file systems when testing ext2 in GRUB.

---
 tests/util/grub-fs-tester.in | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
tests/util/grub-fs-tester.in +5 -1
index 23f72dd35..c4af16d51 100644
@@ -874,7 +874,11 @@ for LOGSECSIZE in $(range "$MINLOGSECSIZE" "$MAXLOGSECSIZE" 1); do
 		xnilfs2)
 		    "mkfs.nilfs2" -L "$FSLABEL" -b $BLKSIZE  -q "${MOUNTDEVICE}" ;;
 		xext2_old)
-		    MKE2FS_DEVICE_SECTSIZE=$SECSIZE "mkfs.ext2" -r 0 -b $BLKSIZE -L "$FSLABEL" -q "${MOUNTDEVICE}"
+		    if "mkfs.ext2" -r 0 2>&1 | grep -F "the -r option has been removed"; then
+			MKE2FS_DEVICE_SECTSIZE=$SECSIZE "mkfs.ext2" -E revision=0 -b $BLKSIZE -L "$FSLABEL" -q "${MOUNTDEVICE}"
+		    else
+			MKE2FS_DEVICE_SECTSIZE=$SECSIZE "mkfs.ext2" -r 0 -b $BLKSIZE -L "$FSLABEL" -q "${MOUNTDEVICE}"
+		    fi
 		    MOUNTFS=ext2
 		    ;;
 		xext4_metabg)
Signature · 7 lines
2.39.5


_______________________________________________
Grub-devel mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/grub-devel
#5 lore raw

PATCH v2 2/3 tests: Avoid Test Failure in erofs for label length

In reply to Andrew Hamilton

Recently, mkfs.erofs began to enforce that the file system label is 15 characters or less (exluding NULL terminator). This causes the current erofs test in GRUB to fail. Reduce the test label used to fit in this limit allowing the test to work as expected.

---
 tests/util/grub-fs-tester.in | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)
tests/util/grub-fs-tester.in +5 -2
index cac58dafa..23f72dd35 100644
@@ -386,9 +386,12 @@ for LOGSECSIZE in $(range "$MINLOGSECSIZE" "$MAXLOGSECSIZE" 1); do
 		    FSLABEL="g;/_é䏌䐓䏕䎛䎾䏴кит u"
 		    #FSLABEL="g;/_é莭莽😁кит u"
 		    ;;
-		# FS LIMITATION: reiserfs, extN, jfs and erofs label is at most 16 UTF-8 characters
-		x"reiserfs_old" | x"reiserfs" | x"ext"* | x"lvm"* | x"luks"* | x"mdraid"* | x"jfs" | x"jfs_caseins" | x"erofs_"*)
+		# FS LIMITATION: reiserfs, extN, and jfs label is at most 16 UTF-8 characters
+		x"reiserfs_old" | x"reiserfs" | x"ext"* | x"lvm"* | x"luks"* | x"mdraid"* | x"jfs" | x"jfs_caseins")
 		    FSLABEL="g;/éт 莭😁";;
+		# FS LIMITATION: erofs label is at most 15 UTF-8 characters
+		x"erofs_"*)
+		    FSLABEL="g;/é 莭😁";;
 		# FS LIMITATION: No underscore, space, semicolon, slash or international characters in UFS* in label. Limited to 32 UTF-8 characters
 		x"ufs1" | x"ufs1_sun" | x"ufs2")
 		    FSLABEL="grubtest""ieurrucnenreeiurueurewf";;
Signature · 7 lines
2.39.5


_______________________________________________
Grub-devel mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/grub-devel