Addition of new tests for extsize hints

PATCH Superseded Nirjhar Roy 3 participants 15 messages Nov 21, 2024 – Nov 22, 2024

Share & tools

Nirjhar Roy

author

2024-11-21 05:11 UTC #0 lore raw

v3 cover Addition of new tests for extsize hints

This test checks the behaviour of xfs/ext4 filesystems when extsize hint is set on files with inode size as 0, non-empty
files with allocated and delalloc extents and so on.
This test set also tests one of the scenarios where setting extsize hints on files with delayed allocation incorrectly succeeds and there is a recent patch series[1] that got merged[2] which fixes this.

Currently this test only runs in xfs but there is a patch series[3] in review that adds support for the extsize hints for ext4 as well.

I have tested it on ppc64le (with page size 64k) and x86_64 (with page size 4k). The block sizes I have tested with are 2k, 4k, 8k, 16k, 32k, 64k with extsize being twice of the block size or 4 * block size(if the default extsize is 2 * blocksize). I have also tested with configurations where I added extszinherit mkfs option on the root. I had to enable CONFIG_TRANSPARENT_HUGEPAGE to enable block size greater than 4k on x86_64.

Changes since [v2]
- Renamed _test_fsx_xflags_field() to _test_fsxattr_xflag().
- Changed the definition of _test_fsx_xflags_field (now renamed to _test_fsxattr_xflag) to accept xflag names (like extsize, cowextsize) instead of single characters (like 'e') and modified the callers of this function accordingly (in xfs/207 and generic/366)
- Some changes in the comments in the setup() function of tests/generic/366.
- Added another helper function _require_scratch_extsize().
- Replaced the usage of _require_scratch() with _require_scratch_extsize() in tests/generic/366.
- Fixed some indentation issues in the test scripts(xfs/207, generic/366) and the commit messages.

Changes since [v1]
- Making the definition of _test_xfs_xflags_field() more compact and renamed _test_xfs_xflags_field() to _test_fsx_xflags_field() since I moved it to common/rc.
- Removed the explicit import of common/xfs from the test script.
- Renamed the test file from generic/365 to generic/366 since generic/365 was taken.
- Made the following functions in tests/generic/366 accept a second parameter (extsize)
a) check_extsz_xflag_across_remount()
b) check_extsz_and_xflag()
c) read_file_extsize()
d) filter_extsz
This will help filter out any extsize that the test wants. This is required because, this version no longer assumes any particular fixed value of the default extsize.
- Removed the check for xflags in test_data_allocated() and test_data_delayed() test
- Added an extra line in output that denotes that we are checking that the extent size hasn't changed if the xfs_io command fails (in test_data_allocated() and test_data_delayed() test).
- Changed the new extsize to 4*blocksize if default extsize size is 2*blocksize to make sure that the new extsize is not the same as the default extsize so that we can observe it changing.

[1] https://lore.kernel.org/all/[email protected]/
[2] kernel commit - 2a492ff66673
[3] https://lore.kernel.org/linux-ext4/[email protected]/
[v1] https://lore.kernel.org/all/[email protected]/
[v2] https://lore.kernel.org/all/[email protected]/

Nirjhar Roy (3):
common/rc,xfs/207: Add a common helper function to check xflag bits common/rc: Add a new _require_scratch_extsize helper function
generic: Addition of new tests for extsize hints

 common/rc             |  24 ++++++
 tests/generic/366     | 175 ++++++++++++++++++++++++++++++++++++++++++
 tests/generic/366.out |  26 +++++++
 tests/xfs/207         |  15 +---
 4 files changed, 229 insertions(+), 11 deletions(-)
 create mode 100755 tests/generic/366
 create mode 100644 tests/generic/366.out

Signature · 1 line

2.43.5

Nirjhar Roy

author

2024-11-21 05:12 UTC #1 lore raw

PATCH v3 1/3 common/rc,xfs/207: Add a common helper function to check xflag bits

This patch defines a common helper function to test whether any of fsxattr xflags field is set or not. We will use this helper in an upcoming patch for checking extsize (e) flag.

Reviewed-by Ritesh Harjani (IBM) <[email protected]>

Reviewed-by Ojaswin Mujoo <[email protected]>

Signed-off-by Nirjhar Roy <[email protected]>

---
 common/rc     |  7 +++++++
 tests/xfs/207 | 15 ++++-----------
 2 files changed, 11 insertions(+), 11 deletions(-)

common/rc +7 -0

index 2af26f23..cccc98f5 100644
@@ -41,6 +41,13 @@ _md5_checksum()
 	md5sum $1 | cut -d ' ' -f1
 }
 
+# Check whether a fsxattr xflags name ($2) field is set on a given file ($1).
+# e.g, fsxattr.xflags =  0x80000800 [extsize, has-xattr]
+_test_fsxattr_xflag()
+{
+	grep -q "fsxattr.xflags.*\[.*$2.*\]" <($XFS_IO_PROG -c "stat -v" "$1")
+}
+
 # Write a byte into a range of a file
 _pwrite_byte() {
 	local pattern="$1"

tests/xfs/207 +4 ~~-11~~

index bbe21307..394e7e55 100755
@@ -21,15 +21,6 @@ _require_cp_reflink
 _require_xfs_io_command "fiemap"
 _require_xfs_io_command "cowextsize"
 
-# Takes the fsxattr.xflags line,
-# i.e. fsxattr.xflags = 0x0 [--------------C-]
-# and tests whether a flag character is set
-test_xflag()
-{
-    local flg=$1
-    grep -q "\[.*${flg}.*\]" && echo "$flg flag set" || echo "$flg flag unset"
-}
-
 echo "Format and mount"
 _scratch_mkfs > $seqres.full 2>&1
 _scratch_mount >> $seqres.full 2>&1

@@ -65,14 +56,16 @@ echo "Set cowextsize and check flag"
 $XFS_IO_PROG -c "cowextsize 1048576" $testdir/file3 | _filter_scratch
 _scratch_cycle_mount
 
-$XFS_IO_PROG -c "stat" $testdir/file3 | grep 'fsxattr.xflags' | test_xflag "C"
+_test_fsxattr_xflag "$testdir/file3" "cowextsize" && echo "C flag set" || \
+	echo "C flag unset"
 $XFS_IO_PROG -c "cowextsize" $testdir/file3 | _filter_scratch
 
 echo "Unset cowextsize and check flag"
 $XFS_IO_PROG -c "cowextsize 0" $testdir/file3 | _filter_scratch
 _scratch_cycle_mount
 
-$XFS_IO_PROG -c "stat" $testdir/file3 | grep 'fsxattr.xflags' | test_xflag "C"
+_test_fsxattr_xflag "$testdir/file3" "cowextsize" && echo "C flag set" || \
+	echo "C flag unset"
 $XFS_IO_PROG -c "cowextsize" $testdir/file3 | _filter_scratch
 
 status=0

Signature · 1 line

2.43.5

Darrick J. Wong

Reviewed-by on v3 1/3

2024-11-21 17:20 UTC #2 lore raw

Re: [PATCH v3 1/3] common/rc,xfs/207: Add a common helper function to check xflag bits

Quoted 7 lines · Nirjhar Roy

This patch defines a common helper function to test whether any of fsxattr xflags field is set or not. We will use this helper in an upcoming patch for checking extsize (e) flag.
Reviewed-by Ritesh Harjani (IBM) <[email protected]>
Reviewed-by Ojaswin Mujoo <[email protected]>
Signed-off-by Nirjhar Roy <[email protected]>

Looks good to me now,

Reviewed-by "Darrick J. Wong" <[email protected]>

--D

Quoted 66 lines

---
 common/rc     |  7 +++++++
 tests/xfs/207 | 15 ++++-----------
 2 files changed, 11 insertions(+), 11 deletions(-)

common/rc +7 -0

index 2af26f23..cccc98f5 100644
@@ -41,6 +41,13 @@ _md5_checksum()
 	md5sum $1 | cut -d ' ' -f1
 }
 
+# Check whether a fsxattr xflags name ($2) field is set on a given file ($1).
+# e.g, fsxattr.xflags =  0x80000800 [extsize, has-xattr]
+_test_fsxattr_xflag()
+{
+	grep -q "fsxattr.xflags.*\[.*$2.*\]" <($XFS_IO_PROG -c "stat -v" "$1")
+}
+
 # Write a byte into a range of a file
 _pwrite_byte() {
 	local pattern="$1"

tests/xfs/207 +4 ~~-11~~

index bbe21307..394e7e55 100755
@@ -21,15 +21,6 @@ _require_cp_reflink
 _require_xfs_io_command "fiemap"
 _require_xfs_io_command "cowextsize"
 
-# Takes the fsxattr.xflags line,
-# i.e. fsxattr.xflags = 0x0 [--------------C-]
-# and tests whether a flag character is set
-test_xflag()
-{
-    local flg=$1
-    grep -q "\[.*${flg}.*\]" && echo "$flg flag set" || echo "$flg flag unset"
-}
-
 echo "Format and mount"
 _scratch_mkfs > $seqres.full 2>&1
 _scratch_mount >> $seqres.full 2>&1

@@ -65,14 +56,16 @@ echo "Set cowextsize and check flag"
 $XFS_IO_PROG -c "cowextsize 1048576" $testdir/file3 | _filter_scratch
 _scratch_cycle_mount
 
-$XFS_IO_PROG -c "stat" $testdir/file3 | grep 'fsxattr.xflags' | test_xflag "C"
+_test_fsxattr_xflag "$testdir/file3" "cowextsize" && echo "C flag set" || \
+	echo "C flag unset"
 $XFS_IO_PROG -c "cowextsize" $testdir/file3 | _filter_scratch
 
 echo "Unset cowextsize and check flag"
 $XFS_IO_PROG -c "cowextsize 0" $testdir/file3 | _filter_scratch
 _scratch_cycle_mount
 
-$XFS_IO_PROG -c "stat" $testdir/file3 | grep 'fsxattr.xflags' | test_xflag "C"
+_test_fsxattr_xflag "$testdir/file3" "cowextsize" && echo "C flag set" || \
+	echo "C flag unset"
 $XFS_IO_PROG -c "cowextsize" $testdir/file3 | _filter_scratch
 
 status=0

Signature · 1 line

2.43.5

Nirjhar Roy

author

2024-11-21 18:31 UTC #3 lore raw

Re: [PATCH v3 1/3] common/rc,xfs/207: Add a common helper function to check xflag bits

Quoted 5 lines · Darrick J. Wong

Quoted 4 lines · Nirjhar Roy
This patch defines a common helper function to test whether any of fsxattr xflags field is set or not. We will use this helper in an upcoming patch for checking extsize (e) flag.

Quoted 7 lines

Quoted 3 lines
Reviewed-by Ritesh Harjani (IBM) <[email protected]>
Reviewed-by Ojaswin Mujoo <[email protected]>
Signed-off-by Nirjhar Roy <[email protected]>
Looks good to me now,
Reviewed-by "Darrick J. Wong" <[email protected]>
--D

Thank you.

--NR

Quoted 67 lines

---
  common/rc     |  7 +++++++
  tests/xfs/207 | 15 ++++-----------
  2 files changed, 11 insertions(+), 11 deletions(-)

common/rc +7 -0

index 2af26f23..cccc98f5 100644
@@ -41,6 +41,13 @@ _md5_checksum()
  	md5sum $1 | cut -d ' ' -f1
  }
 
+# Check whether a fsxattr xflags name ($2) field is set on a given file ($1).
+# e.g, fsxattr.xflags =  0x80000800 [extsize, has-xattr]
+_test_fsxattr_xflag()
+{
+	grep -q "fsxattr.xflags.*\[.*$2.*\]" <($XFS_IO_PROG -c "stat -v" "$1")
+}
+
  # Write a byte into a range of a file
  _pwrite_byte() {
  	local pattern="$1"

tests/xfs/207 +4 ~~-11~~

index bbe21307..394e7e55 100755
@@ -21,15 +21,6 @@ _require_cp_reflink
  _require_xfs_io_command "fiemap"
  _require_xfs_io_command "cowextsize"
 
-# Takes the fsxattr.xflags line,
-# i.e. fsxattr.xflags = 0x0 [--------------C-]
-# and tests whether a flag character is set
-test_xflag()
-{
-    local flg=$1
-    grep -q "\[.*${flg}.*\]" && echo "$flg flag set" || echo "$flg flag unset"
-}
-
  echo "Format and mount"
  _scratch_mkfs > $seqres.full 2>&1
  _scratch_mount >> $seqres.full 2>&1

@@ -65,14 +56,16 @@ echo "Set cowextsize and check flag"
  $XFS_IO_PROG -c "cowextsize 1048576" $testdir/file3 | _filter_scratch
  _scratch_cycle_mount
 
-$XFS_IO_PROG -c "stat" $testdir/file3 | grep 'fsxattr.xflags' | test_xflag "C"
+_test_fsxattr_xflag "$testdir/file3" "cowextsize" && echo "C flag set" || \
+	echo "C flag unset"
  $XFS_IO_PROG -c "cowextsize" $testdir/file3 | _filter_scratch
 
  echo "Unset cowextsize and check flag"
  $XFS_IO_PROG -c "cowextsize 0" $testdir/file3 | _filter_scratch
  _scratch_cycle_mount
 
-$XFS_IO_PROG -c "stat" $testdir/file3 | grep 'fsxattr.xflags' | test_xflag "C"
+_test_fsxattr_xflag "$testdir/file3" "cowextsize" && echo "C flag set" || \
+	echo "C flag unset"
  $XFS_IO_PROG -c "cowextsize" $testdir/file3 | _filter_scratch
 
  status=0

Signature · 1 line

2.43.5

Signature · 4 lines

---
Nirjhar Roy
Linux Kernel Developer
IBM, Bangalore

Nirjhar Roy

author

2024-11-21 05:12 UTC #4 lore raw

PATCH v3 2/3 common/rc: Add a new _require_scratch_extsize helper function

In reply to Nirjhar Roy

_require_scratch_extsize helper function will be used in the the next patch to make the test run only on filesystems with extsize support.

Reviewed-by Ojaswin Mujoo <[email protected]>

Signed-off-by Nirjhar Roy <[email protected]>

---
 common/rc | 17 +++++++++++++++++
 1 file changed, 17 insertions(+)

common/rc +17 -0

index cccc98f5..995979e9 100644
@@ -48,6 +48,23 @@ _test_fsxattr_xflag()
 	grep -q "fsxattr.xflags.*\[.*$2.*\]" <($XFS_IO_PROG -c "stat -v" "$1")
 }
 
+# This test requires extsize support on the  filesystem
+_require_scratch_extsize()
+{
+	_require_scratch
+	_scratch_mkfs > /dev/null
+	_scratch_mount
+	local filename=$SCRATCH_MNT/$RANDOM
+	local blksz=$(_get_block_size $SCRATCH_MNT)
+	local extsz=$(( blksz*2 ))
+	local res=$($XFS_IO_PROG -c "open -f $filename" -c "extsize $extsz" \
+		-c "extsize")
+	_scratch_unmount
+	grep -q "\[$extsz\] $filename" <(echo $res) || \
+		_notrun "this test requires extsize support on the filesystem"
+}
+
+
 # Write a byte into a range of a file
 _pwrite_byte() {
 	local pattern="$1"

Signature · 1 line

2.43.5

Ritesh Harjani (IBM)

2024-11-21 08:00 UTC #5 lore raw

Re: [PATCH v3 2/3] common/rc: Add a new _require_scratch_extsize helper function

Quoted 15 lines · Nirjhar Roy

_require_scratch_extsize helper function will be used in the the next patch to make the test run only on filesystems with extsize support.
Reviewed-by Ojaswin Mujoo <[email protected]>
Signed-off-by Nirjhar Roy <[email protected]>
---
 common/rc | 17 +++++++++++++++++
 1 file changed, 17 insertions(+)
common/rc +0 -0
index cccc98f5..995979e9 100644

common/rc _test_fsxattr_xflag() · line 54

Quoted 7 lines

common/rc +4 -0

@@ -48,6 +48,23 @@ _test_fsxattr_xflag()
 	grep -q "fsxattr.xflags.*\[.*$2.*\]" <($XFS_IO_PROG -c "stat -v" "$1")
 }
 
+# This test requires extsize support on the  filesystem
+_require_scratch_extsize()
+{
+	_require_scratch

_require_xfs_io_command "extsize"

^^^ Don't we need this too?

common/rc _test_fsxattr_xflag() · line 64

Quoted 10 lines

common/rc +10 -0

+	_scratch_mkfs > /dev/null
+	_scratch_mount
+	local filename=$SCRATCH_MNT/$RANDOM
+	local blksz=$(_get_block_size $SCRATCH_MNT)
+	local extsz=$(( blksz*2 ))
+	local res=$($XFS_IO_PROG -c "open -f $filename" -c "extsize $extsz" \
+		-c "extsize")
+	_scratch_unmount
+	grep -q "\[$extsz\] $filename" <(echo $res) || \
+		_notrun "this test requires extsize support on the filesystem"

Why grep when we can simply just check the return value of previous xfs_io command?

common/rc _test_fsxattr_xflag() · line 67

Quoted 3 lines

common/rc +3 -0
+}
+
+

^^ Extra newline.

Quoted 5 lines

common/rc +0 -0

 # Write a byte into a range of a file
 _pwrite_byte() {
 	local pattern="$1"

Signature · 1 line

2.43.5

Nirjhar Roy

author

2024-11-21 18:33 UTC #6 lore raw

Re: [PATCH v3 2/3] common/rc: Add a new _require_scratch_extsize helper function

Quoted 20 lines · Ritesh Harjani (IBM)

Quoted 18 lines · Nirjhar Roy
_require_scratch_extsize helper function will be used in the the next patch to make the test run only on filesystems with extsize support.
Reviewed-by Ojaswin Mujoo <[email protected]>
Signed-off-by Nirjhar Roy <[email protected]>
---
  common/rc | 17 +++++++++++++++++
  1 file changed, 17 insertions(+)
common/rc +0 -0
index cccc98f5..995979e9 100644
@@ -48,6 +48,23 @@ _test_fsxattr_xflag()
  	grep -q "fsxattr.xflags.*\[.*$2.*\]" <($XFS_IO_PROG -c "stat -v" "$1")
  }
 

Quoted 7 lines

common/rc _test_fsxattr_xflag() · line 54
Quoted 4 lines
common/rc +4 -0
+# This test requires extsize support on the  filesystem
+_require_scratch_extsize()
+{
+	_require_scratch
_require_xfs_io_command "extsize"
^^^ Don't we need this too?

Yes, good point. I will add this in the next revision.

Quoted 12 lines

common/rc _test_fsxattr_xflag() · line 64

Quoted 11 lines

common/rc +10 -0

+	_scratch_mkfs > /dev/null
+	_scratch_mount
+	local filename=$SCRATCH_MNT/$RANDOM
+	local blksz=$(_get_block_size $SCRATCH_MNT)
+	local extsz=$(( blksz*2 ))
+	local res=$($XFS_IO_PROG -c "open -f $filename" -c "extsize $extsz" \
+		-c "extsize")
+	_scratch_unmount
+	grep -q "\[$extsz\] $filename" <(echo $res) || \
+		_notrun "this test requires extsize support on the filesystem"

Why grep when we can simply just check the return value of previous xfs_io command?

No, I don't think we can rely on the return value of xfs_io. For ex, let's look at the following set of commands which are ran on an ext4 system:

root@AMARPC: /mnt1/test$ xfs_io -V
xfs_io version 5.13.0
root@AMARPC: /mnt1/test$ touch new
root@AMARPC: /mnt1/test$ xfs_io -c "extsize 8k" new
foreign file active, extsize command is for XFS filesystems only
root@AMARPC: /mnt1/test$ echo "$?"
0
This incorrect return value might have been fixed in some later versions of xfs_io but there are still versions where we can't solely rely on the return value.

Quoted 5 lines

common/rc _test_fsxattr_xflag() · line 67
Quoted 4 lines
common/rc +3 -0
+}
+
+
^^ Extra newline.

Noted. I will fix this.

--NR

Quoted 6 lines

common/rc +0 -0

  # Write a byte into a range of a file
  _pwrite_byte() {
  	local pattern="$1"

Signature · 1 line

2.43.5

Signature · 4 lines

---
Nirjhar Roy
Linux Kernel Developer
IBM, Bangalore

Ritesh Harjani (IBM)

2024-11-21 18:56 UTC #7 lore raw

Re: [PATCH v3 2/3] common/rc: Add a new _require_scratch_extsize helper function

Quoted 44 lines · Nirjhar Roy

Quoted 20 lines · Ritesh Harjani (IBM)
Quoted 18 lines · Nirjhar Roy
_require_scratch_extsize helper function will be used in the the next patch to make the test run only on filesystems with extsize support.
Reviewed-by Ojaswin Mujoo <[email protected]>
Signed-off-by Nirjhar Roy <[email protected]>
---
  common/rc | 17 +++++++++++++++++
  1 file changed, 17 insertions(+)
common/rc +0 -0
index cccc98f5..995979e9 100644
@@ -48,6 +48,23 @@ _test_fsxattr_xflag()
  	grep -q "fsxattr.xflags.*\[.*$2.*\]" <($XFS_IO_PROG -c "stat -v" "$1")
  }
 
Quoted 7 lines
common/rc _test_fsxattr_xflag() · line 54
Quoted 4 lines
common/rc +4 -0
+# This test requires extsize support on the  filesystem
+_require_scratch_extsize()
+{
+	_require_scratch
_require_xfs_io_command "extsize"
^^^ Don't we need this too?
Yes, good point. I will add this in the next revision.
Quoted 12 lines
common/rc _test_fsxattr_xflag() · line 64
Quoted 11 lines
common/rc +10 -0
+	_scratch_mkfs > /dev/null
+	_scratch_mount
+	local filename=$SCRATCH_MNT/$RANDOM
+	local blksz=$(_get_block_size $SCRATCH_MNT)
+	local extsz=$(( blksz*2 ))
+	local res=$($XFS_IO_PROG -c "open -f $filename" -c "extsize $extsz" \
+		-c "extsize")
+	_scratch_unmount
+	grep -q "\[$extsz\] $filename" <(echo $res) || \
+		_notrun "this test requires extsize support on the filesystem"
Why grep when we can simply just check the return value of previous xfs_io command?
No, I don't think we can rely on the return value of xfs_io. For ex, let's look at the following set of commands which are ran on an ext4 system:

Quoted 10 lines

root@AMARPC: /mnt1/test$ xfs_io -V
xfs_io version 5.13.0
root@AMARPC: /mnt1/test$ touch new
root@AMARPC: /mnt1/test$ xfs_io -c "extsize 8k" new
foreign file active, extsize command is for XFS filesystems only root@AMARPC: /mnt1/test$ echo "$?"
0
This incorrect return value might have been fixed in some later versions of xfs_io but there are still versions where we can't solely rely on the return value.

Ok. That's bad, we then have to rely on grep.
Sure, thanks for checking and confirming that.

-ritesh

Darrick J. Wong

2024-11-22 16:04 UTC #8 lore raw

Re: [PATCH v3 2/3] common/rc: Add a new _require_scratch_extsize helper function

Quoted 57 lines · Ritesh Harjani

Quoted 44 lines · Nirjhar Roy
Quoted 20 lines · Ritesh Harjani (IBM)
Quoted 18 lines · Nirjhar Roy
_require_scratch_extsize helper function will be used in the the next patch to make the test run only on filesystems with extsize support.
Reviewed-by Ojaswin Mujoo <[email protected]>
Signed-off-by Nirjhar Roy <[email protected]>
---
  common/rc | 17 +++++++++++++++++
  1 file changed, 17 insertions(+)
common/rc +0 -0
index cccc98f5..995979e9 100644
@@ -48,6 +48,23 @@ _test_fsxattr_xflag()
  	grep -q "fsxattr.xflags.*\[.*$2.*\]" <($XFS_IO_PROG -c "stat -v" "$1")
  }
 
Quoted 7 lines
common/rc _test_fsxattr_xflag() · line 54
Quoted 4 lines
common/rc +4 -0
+# This test requires extsize support on the  filesystem
+_require_scratch_extsize()
+{
+	_require_scratch
_require_xfs_io_command "extsize"
^^^ Don't we need this too?
Yes, good point. I will add this in the next revision.
Quoted 12 lines
common/rc _test_fsxattr_xflag() · line 64
Quoted 11 lines
common/rc +10 -0
+	_scratch_mkfs > /dev/null
+	_scratch_mount
+	local filename=$SCRATCH_MNT/$RANDOM
+	local blksz=$(_get_block_size $SCRATCH_MNT)
+	local extsz=$(( blksz*2 ))
+	local res=$($XFS_IO_PROG -c "open -f $filename" -c "extsize $extsz" \
+		-c "extsize")
+	_scratch_unmount
+	grep -q "\[$extsz\] $filename" <(echo $res) || \
+		_notrun "this test requires extsize support on the filesystem"
Why grep when we can simply just check the return value of previous xfs_io command?
No, I don't think we can rely on the return value of xfs_io. For ex, let's look at the following set of commands which are ran on an ext4 system:
Quoted 10 lines
root@AMARPC: /mnt1/test$ xfs_io -V
xfs_io version 5.13.0
root@AMARPC: /mnt1/test$ touch new
root@AMARPC: /mnt1/test$ xfs_io -c "extsize 8k" new
foreign file active, extsize command is for XFS filesystems only root@AMARPC: /mnt1/test$ echo "$?"
0
This incorrect return value might have been fixed in some later versions of xfs_io but there are still versions where we can't solely rely on the return value.

Quoted 2 lines

Ok. That's bad, we then have to rely on grep.
Sure, thanks for checking and confirming that.

You all should add CMD_FOREIGN_OK to the extsize command in xfs_io, assuming that you've not already done that in your dev workspace.

--D

Quoted 2 lines

-ritesh

Nirjhar Roy

author

2024-11-22 18:07 UTC #9 lore raw

Re: [PATCH v3 2/3] common/rc: Add a new _require_scratch_extsize helper function

Quoted 56 lines · Darrick J. Wong

Quoted 55 lines · Ritesh Harjani
Quoted 53 lines · Nirjhar Roy
Quoted 20 lines · Ritesh Harjani (IBM)
Quoted 18 lines · Nirjhar Roy
_require_scratch_extsize helper function will be used in the the next patch to make the test run only on filesystems with extsize support.
Reviewed-by Ojaswin Mujoo <[email protected]>
Signed-off-by Nirjhar Roy <[email protected]>
---
   common/rc | 17 +++++++++++++++++
   1 file changed, 17 insertions(+)
common/rc +0 -0
index cccc98f5..995979e9 100644
@@ -48,6 +48,23 @@ _test_fsxattr_xflag()
   	grep -q "fsxattr.xflags.*\[.*$2.*\]" <($XFS_IO_PROG -c "stat -v" "$1")
   }
 
Quoted 7 lines
common/rc _test_fsxattr_xflag() · line 54
Quoted 4 lines
common/rc +4 -0
+# This test requires extsize support on the  filesystem
+_require_scratch_extsize()
+{
+	_require_scratch
_require_xfs_io_command "extsize"
^^^ Don't we need this too?
Yes, good point. I will add this in the next revision.
Quoted 11 lines
common/rc _test_fsxattr_xflag() · line 64
Quoted 10 lines
common/rc +10 -0
+	_scratch_mkfs > /dev/null
+	_scratch_mount
+	local filename=$SCRATCH_MNT/$RANDOM
+	local blksz=$(_get_block_size $SCRATCH_MNT)
+	local extsz=$(( blksz*2 ))
+	local res=$($XFS_IO_PROG -c "open -f $filename" -c "extsize $extsz" \
+		-c "extsize")
+	_scratch_unmount
+	grep -q "\[$extsz\] $filename" <(echo $res) || \
+		_notrun "this test requires extsize support on the filesystem"
Why grep when we can simply just check the return value of previous xfs_io command?
No, I don't think we can rely on the return value of xfs_io. For ex, let's look at the following set of commands which are ran on an ext4 system:
root@AMARPC: /mnt1/test$ xfs_io -V
xfs_io version 5.13.0
root@AMARPC: /mnt1/test$ touch new
root@AMARPC: /mnt1/test$ xfs_io -c "extsize 8k" new
foreign file active, extsize command is for XFS filesystems only root@AMARPC: /mnt1/test$ echo "$?"
0
This incorrect return value might have been fixed in some later versions of xfs_io but there are still versions where we can't solely rely on the return value.

Quoted 6 lines

Quoted 2 lines
Ok. That's bad, we then have to rely on grep.
Sure, thanks for checking and confirming that.
You all should add CMD_FOREIGN_OK to the extsize command in xfs_io, assuming that you've not already done that in your dev workspace.
--D

Yes, I have tested with that as well. I have applied the following patch to xfsprogs and tested with the modified xfs_io.

io/open.c +0 -0

index 15850b55..6407b7e8 100644
@@ -980,7 +980,7 @@ open_init(void)

        extsize_cmd.args = _("[-D | -R] [extsize]"); extsize_cmd.argmin = 0;
        extsize_cmd.argmax = -1;
-       extsize_cmd.flags = CMD_NOMAP_OK;
+       extsize_cmd.flags = CMD_NOMAP_OK | CMD_FOREIGN_OK; extsize_cmd.oneline =
                _("get/set preferred extent size (in bytes) for the open file"); extsize_cmd.help = extsize_help;

The return values are similar.

root@AMARPC: /mnt1/scratch$ touch new
root@AMARPC: /mnt1/scratch$ /home/ubuntu/xfsprogs-dev/io/xfs_io -c "extsize 8k" new root@AMARPC: /mnt1/scratch$ echo "$?"
0
root@AMARPC: /mnt1/scratch$ /home/ubuntu/xfsprogs-dev/io/xfs_io -c "extsize" new
[0] new

This is the reason I am not relying on the return value, instead I am checking if only the extsize gets changed, we will assume that the extsize support is there, else the test will _notrun.

Also,

root@AMARPC: /mnt1/scratch$ strace -f /mnt1/scratch$ /home/ubuntu/xfsprogs-dev/io/xfs_io -c "extsize 16k" new

...

ioctl(3, FS_IOC_FSGETXATTR, {fsx_xflags=0, fsx_extsize=0, fsx_nextents=0, fsx_projid=0, fsx_cowextsize=0}) = 0 ioctl(3, FS_IOC_FSSETXATTR, {fsx_xflags=FS_XFLAG_EXTSIZE, fsx_extsize=16384, fsx_projid=0, fsx_cowextsize=0}) = 0 exit_group(0)

Looking at the existing code for ext4_fileattr_set(), We validate the flags but I think we silently don't validate(and ignore) the xflags. Like, we have

int err = -EOPNOTSUPP;
if (flags & ~EXT4_FL_USER_VISIBLE)
goto out;

BUT we do NOT have something like

int err = -EOPNOTSUPP;
if (fa->fsx_flags & ~EXT4_VALID_XFLAGS) // where EXT4_VALID_XFLAGS should be an || of all the supported xflags in ext4. goto out;

I am not sure what other filesystems do, but if we check whether the extsize got changed, then we can correctly determine extsize support.

Does that make sense?

--NR

Quoted 3 lines

Quoted 3 lines
-ritesh

Signature · 4 lines

---
Nirjhar Roy
Linux Kernel Developer
IBM, Bangalore

Darrick J. Wong

2024-11-22 18:44 UTC #10 lore raw

Re: [PATCH v3 2/3] common/rc: Add a new _require_scratch_extsize helper function

Quoted 126 lines · Nirjhar Roy

Quoted 55 lines · Darrick J. Wong
Quoted 54 lines · Ritesh Harjani
Quoted 52 lines · Nirjhar Roy
Quoted 25 lines · Ritesh Harjani (IBM)
Quoted 15 lines · Nirjhar Roy
_require_scratch_extsize helper function will be used in the the next patch to make the test run only on filesystems with extsize support.
Reviewed-by Ojaswin Mujoo <[email protected]>
Signed-off-by Nirjhar Roy <[email protected]>
---
   common/rc | 17 +++++++++++++++++
   1 file changed, 17 insertions(+)
common/rc +0 -0
index cccc98f5..995979e9 100644
common/rc _test_fsxattr_xflag() · line 53
Quoted 6 lines
common/rc +4 -0
@@ -48,6 +48,23 @@ _test_fsxattr_xflag()
   	grep -q "fsxattr.xflags.*\[.*$2.*\]" <($XFS_IO_PROG -c "stat -v" "$1")
   }
+# This test requires extsize support on the  filesystem
+_require_scratch_extsize()
+{
+	_require_scratch
_require_xfs_io_command "extsize"
Quoted 1 line
^^^ Don't we need this too?
Yes, good point. I will add this in the next revision.
Quoted 11 lines
common/rc _test_fsxattr_xflag() · line 63
Quoted 10 lines
common/rc +10 -0
+	_scratch_mkfs > /dev/null
+	_scratch_mount
+	local filename=$SCRATCH_MNT/$RANDOM
+	local blksz=$(_get_block_size $SCRATCH_MNT)
+	local extsz=$(( blksz*2 ))
+	local res=$($XFS_IO_PROG -c "open -f $filename" -c "extsize $extsz" \
+		-c "extsize")
+	_scratch_unmount
+	grep -q "\[$extsz\] $filename" <(echo $res) || \
+		_notrun "this test requires extsize support on the filesystem"
Why grep when we can simply just check the return value of previous xfs_io command?
No, I don't think we can rely on the return value of xfs_io. For ex, let's look at the following set of commands which are ran on an ext4 system:
root@AMARPC: /mnt1/test$ xfs_io -V
xfs_io version 5.13.0
root@AMARPC: /mnt1/test$ touch new
root@AMARPC: /mnt1/test$ xfs_io -c "extsize 8k" new
foreign file active, extsize command is for XFS filesystems only root@AMARPC: /mnt1/test$ echo "$?"
0
This incorrect return value might have been fixed in some later versions of xfs_io but there are still versions where we can't solely rely on the return value.
Quoted 6 lines
Quoted 2 lines
Ok. That's bad, we then have to rely on grep.
Sure, thanks for checking and confirming that.
You all should add CMD_FOREIGN_OK to the extsize command in xfs_io, assuming that you've not already done that in your dev workspace.
--D
Yes, I have tested with that as well. I have applied the following patch to xfsprogs and tested with the modified xfs_io.
io/open.c +1 -1
index 15850b55..6407b7e8 100644
@@ -980,7 +980,7 @@ open_init(void)
         extsize_cmd.args = _("[-D | -R] [extsize]");
         extsize_cmd.argmin = 0;
         extsize_cmd.argmax = -1;
-       extsize_cmd.flags = CMD_NOMAP_OK;
+       extsize_cmd.flags = CMD_NOMAP_OK | CMD_FOREIGN_OK;
        extsize_cmd.oneline =
                _("get/set preferred extent size (in bytes) for the open file");
        extsize_cmd.help = extsize_help;
The return values are similar.
root@AMARPC: /mnt1/scratch$ touch new
root@AMARPC: /mnt1/scratch$ /home/ubuntu/xfsprogs-dev/io/xfs_io -c "extsize 8k" new
root@AMARPC: /mnt1/scratch$ echo "$?"
0
root@AMARPC: /mnt1/scratch$ /home/ubuntu/xfsprogs-dev/io/xfs_io -c "extsize" new
[0] new
This is the reason I am not relying on the return value, instead I am checking if only the extsize gets changed, we will assume that the extsize support is there, else the test will _notrun.
Also,
root@AMARPC: /mnt1/scratch$ strace -f /mnt1/scratch$ /home/ubuntu/xfsprogs-dev/io/xfs_io -c "extsize 16k" new
...
...
ioctl(3, FS_IOC_FSGETXATTR, {fsx_xflags=0, fsx_extsize=0, fsx_nextents=0, fsx_projid=0, fsx_cowextsize=0}) = 0
ioctl(3, FS_IOC_FSSETXATTR, {fsx_xflags=FS_XFLAG_EXTSIZE, fsx_extsize=16384, fsx_projid=0, fsx_cowextsize=0}) = 0
exit_group(0)
Looking at the existing code for ext4_fileattr_set(), We validate the flags but I think we silently don't validate(and ignore) the xflags. Like, we have
int err = -EOPNOTSUPP;
if (flags & ~EXT4_FL_USER_VISIBLE)
        goto out;
BUT we do NOT have something like
int err = -EOPNOTSUPP;
if (fa->fsx_flags & ~EXT4_VALID_XFLAGS) // where EXT4_VALID_XFLAGS should be an || of all the supported xflags in ext4.
        goto out;

Quoted 4 lines

I am not sure what other filesystems do, but if we check whether the extsize got changed, then we can correctly determine extsize support.
Does that make sense?

You don't have to check fsx_flags if you don't use fileattr_fill_xflags. ext4 doesn't use that.

--D

Quoted 13 lines

--NR
Quoted 3 lines
Quoted 3 lines
-ritesh
Signature · 4 lines
---
Nirjhar Roy
Linux Kernel Developer
IBM, Bangalore

Nirjhar Roy

author

2024-11-22 19:06 UTC #11 lore raw

Re: [PATCH v3 2/3] common/rc: Add a new _require_scratch_extsize helper function

Quoted 125 lines · Darrick J. Wong

Quoted 124 lines · Nirjhar Roy
Quoted 55 lines · Darrick J. Wong
Quoted 54 lines · Ritesh Harjani
Quoted 52 lines · Nirjhar Roy
Quoted 25 lines · Ritesh Harjani (IBM)
Quoted 15 lines · Nirjhar Roy
_require_scratch_extsize helper function will be used in the the next patch to make the test run only on filesystems with extsize support.
Reviewed-by Ojaswin Mujoo <[email protected]>
Signed-off-by Nirjhar Roy <[email protected]>
---
    common/rc | 17 +++++++++++++++++
    1 file changed, 17 insertions(+)
common/rc +0 -0
index cccc98f5..995979e9 100644
common/rc _test_fsxattr_xflag() · line 53
Quoted 6 lines
common/rc +4 -0
@@ -48,6 +48,23 @@ _test_fsxattr_xflag()
    	grep -q "fsxattr.xflags.*\[.*$2.*\]" <($XFS_IO_PROG -c "stat -v" "$1")
    }
+# This test requires extsize support on the  filesystem
+_require_scratch_extsize()
+{
+	_require_scratch
_require_xfs_io_command "extsize"
Quoted 1 line
^^^ Don't we need this too?
Yes, good point. I will add this in the next revision.
Quoted 11 lines
common/rc _test_fsxattr_xflag() · line 63
Quoted 10 lines
common/rc +10 -0
+	_scratch_mkfs > /dev/null
+	_scratch_mount
+	local filename=$SCRATCH_MNT/$RANDOM
+	local blksz=$(_get_block_size $SCRATCH_MNT)
+	local extsz=$(( blksz*2 ))
+	local res=$($XFS_IO_PROG -c "open -f $filename" -c "extsize $extsz" \
+		-c "extsize")
+	_scratch_unmount
+	grep -q "\[$extsz\] $filename" <(echo $res) || \
+		_notrun "this test requires extsize support on the filesystem"
Why grep when we can simply just check the return value of previous xfs_io command?
No, I don't think we can rely on the return value of xfs_io. For ex, let's look at the following set of commands which are ran on an ext4 system:
root@AMARPC: /mnt1/test$ xfs_io -V
xfs_io version 5.13.0
root@AMARPC: /mnt1/test$ touch new
root@AMARPC: /mnt1/test$ xfs_io -c "extsize 8k" new
foreign file active, extsize command is for XFS filesystems only root@AMARPC: /mnt1/test$ echo "$?"
0
This incorrect return value might have been fixed in some later versions of xfs_io but there are still versions where we can't solely rely on the return value.
Quoted 6 lines
Quoted 2 lines
Ok. That's bad, we then have to rely on grep.
Sure, thanks for checking and confirming that.
You all should add CMD_FOREIGN_OK to the extsize command in xfs_io, assuming that you've not already done that in your dev workspace.
--D
Yes, I have tested with that as well. I have applied the following patch to xfsprogs and tested with the modified xfs_io.
io/open.c +1 -1
index 15850b55..6407b7e8 100644
@@ -980,7 +980,7 @@ open_init(void)
         extsize_cmd.args = _("[-D | -R] [extsize]");
         extsize_cmd.argmin = 0;
         extsize_cmd.argmax = -1;
-       extsize_cmd.flags = CMD_NOMAP_OK;
+       extsize_cmd.flags = CMD_NOMAP_OK | CMD_FOREIGN_OK;
         extsize_cmd.oneline =
                 _("get/set preferred extent size (in bytes) for the open
 file");
        extsize_cmd.help = extsize_help;
The return values are similar.
root@AMARPC: /mnt1/scratch$ touch new
root@AMARPC: /mnt1/scratch$ /home/ubuntu/xfsprogs-dev/io/xfs_io -c "extsize 8k" new
root@AMARPC: /mnt1/scratch$ echo "$?"
0
root@AMARPC: /mnt1/scratch$ /home/ubuntu/xfsprogs-dev/io/xfs_io -c "extsize" new
[0] new
This is the reason I am not relying on the return value, instead I am checking if only the extsize gets changed, we will assume that the extsize support is there, else the test will _notrun.
Also,
root@AMARPC: /mnt1/scratch$ strace -f /mnt1/scratch$ /home/ubuntu/xfsprogs-dev/io/xfs_io -c "extsize 16k" new
...
...
ioctl(3, FS_IOC_FSGETXATTR, {fsx_xflags=0, fsx_extsize=0, fsx_nextents=0, fsx_projid=0, fsx_cowextsize=0}) = 0
ioctl(3, FS_IOC_FSSETXATTR, {fsx_xflags=FS_XFLAG_EXTSIZE, fsx_extsize=16384, fsx_projid=0, fsx_cowextsize=0}) = 0
exit_group(0)
Looking at the existing code for ext4_fileattr_set(), We validate the flags but I think we silently don't validate(and ignore) the xflags. Like, we have
int err = -EOPNOTSUPP;
if (flags & ~EXT4_FL_USER_VISIBLE)
        goto out;
BUT we do NOT have something like
int err = -EOPNOTSUPP;
if (fa->fsx_flags & ~EXT4_VALID_XFLAGS) // where EXT4_VALID_XFLAGS should be an || of all the supported xflags in ext4.
        goto out;

Quoted 8 lines

Quoted 4 lines
I am not sure what other filesystems do, but if we check whether the extsize got changed, then we can correctly determine extsize support.
Does that make sense?
You don't have to check fsx_flags if you don't use fileattr_fill_xflags. ext4 doesn't use that.
--D

Okay got it. Thank you. How does the overall logic of the _require_scratch_extsize() look?

Quoted 13 lines

Quoted 13 lines
--NR
Quoted 2 lines
Quoted 2 lines
-ritesh
Signature · 4 lines
---
Nirjhar Roy
Linux Kernel Developer
IBM, Bangalore

Signature · 4 lines

---
Nirjhar Roy
Linux Kernel Developer
IBM, Bangalore

Nirjhar Roy

author

2024-11-21 05:13 UTC #12 lore raw

PATCH v3 3/3 generic: Addition of new tests for extsize hints

In reply to Nirjhar Roy

This commit adds new tests that checks the behaviour of xfs/ext4 filesystems when extsize hint is set on file with inode size as 0, non-empty files with allocated and delalloc extents and so on. Although currently this test is placed under tests/generic, it only runs on xfs and there is an ongoing patch series[1] to enable extsize hints for ext4 as well.

[1] https://lore.kernel.org/linux-ext4/[email protected]/

Reviewed-by Ritesh Harjani (IBM) <[email protected]>

Reviewed-by Ojaswin Mujoo <[email protected]>

Signed-off-by Nirjhar Roy <[email protected]>

---
 tests/generic/366     | 175 ++++++++++++++++++++++++++++++++++++++++++
 tests/generic/366.out |  26 +++++++
 2 files changed, 201 insertions(+)
 create mode 100755 tests/generic/366
 create mode 100644 tests/generic/366.out

tests/generic/366 new file +175 -0

new file mode 100755
index 00000000..25d23f42
@@ -0,0 +1,175 @@
+#! /bin/bash
+# SPDX-License-Identifier: GPL-2.0
+# Copyright (c) 2024 Nirjhar Roy ([email protected]).  All Rights Reserved.
+#
+# FS QA Test 366
+#
+# This test verifies that extent allocation hint setting works correctly on
+# files with no extents allocated and non-empty files which are truncated.
+# It also checks that the  extent hints setting fails with non-empty file
+# i.e, with any file with allocated extents or delayed allocation. We also
+# check if the extsize value and the xflag bit actually got reflected after
+# setting/re-setting the extsize value.
+
+. ./common/config
+. ./common/filter
+. ./common/preamble
+
+_begin_fstest ioctl quick
+
+_fixed_by_kernel_commit "2a492ff66673 \
+                        xfs: Check for delayed allocations before setting \
+			extsize"
+
+_require_scratch_extsize
+
+FILE_DATA_SIZE=1M
+
+get_default_extsize()
+{
+    if [ -z $1 ] || [ ! -d $1 ]; then
+        echo "Missing mount point argument for get_default_extsize"
+        exit 1
+    fi
+    $XFS_IO_PROG -c "extsize" "$1" | sed 's/^\[\([0-9]\+\)\].*/\1/'
+}
+
+filter_extsz()
+{
+    sed "s/\[$1\]/\[EXTSIZE\]/g"
+}
+
+setup()
+{
+    _scratch_mkfs >> "$seqres.full"  2>&1
+    _scratch_mount >> "$seqres.full" 2>&1
+    BLKSZ=`_get_block_size $SCRATCH_MNT`
+    DEFAULT_EXTSIZE=`get_default_extsize $SCRATCH_MNT`
+    EXTSIZE=$(( BLKSZ*2 ))
+    # Make sure the new extsize is not the same as the default
+    # extsize so that we can observe it changing
+    [[ "$DEFAULT_EXTSIZE" -eq "$EXTSIZE" ]] && EXTSIZE=$(( BLKSZ*4 ))
+}
+
+read_file_extsize()
+{
+    $XFS_IO_PROG -c "extsize" $1 | _filter_scratch | filter_extsz $2
+}
+
+check_extsz_and_xflag()
+{
+    local filename=$1
+    local extsize=$2
+    read_file_extsize $filename $extsize
+    _test_fsxattr_xflag $filename "extsize" && echo "e flag set" || \
+	    echo "e flag unset"
+}
+
+check_extsz_xflag_across_remount()
+{
+    local filename=$1
+    local extsize=$2
+    _scratch_cycle_mount
+    check_extsz_and_xflag $filename $extsize
+}
+
+# Extsize flag should be cleared when extsize is reset, so this function
+# checks that this behavior is followed.
+reset_extsz_and_recheck_extsz_xflag()
+{
+    local filename=$1
+    echo "Re-setting extsize hint to 0"
+    $XFS_IO_PROG -c "extsize 0" $filename
+    check_extsz_xflag_across_remount $filename "0"
+}
+
+check_extsz_xflag_before_and_after_reset()
+{
+    local filename=$1
+    local extsize=$2
+    check_extsz_xflag_across_remount $filename $extsize
+    reset_extsz_and_recheck_extsz_xflag $filename
+}
+
+test_empty_file()
+{
+    echo "TEST: Set extsize on empty file"
+    local filename=$1
+    $XFS_IO_PROG \
+        -c "open -f $filename" \
+        -c "extsize $EXTSIZE" \
+
+    check_extsz_xflag_before_and_after_reset $filename $EXTSIZE
+    echo
+}
+
+test_data_delayed()
+{
+    echo "TEST: Set extsize on non-empty file with delayed allocation"
+    local filename=$1
+    $XFS_IO_PROG \
+        -c "open -f $filename" \
+        -c "pwrite -q  0 $FILE_DATA_SIZE" \
+        -c "extsize $EXTSIZE" | _filter_scratch
+
+    echo "test for default extsize setting if any"
+    read_file_extsize $filename $DEFAULT_EXTSIZE
+    echo
+}
+
+test_data_allocated()
+{
+    echo "TEST: Set extsize on non-empty file with allocated extents"
+    local filename=$1
+    $XFS_IO_PROG \
+        -c "open -f $filename" \
+        -c "pwrite -qW  0 $FILE_DATA_SIZE" \
+        -c "extsize $EXTSIZE" | _filter_scratch
+
+    echo "test for default extsize setting if any"
+    read_file_extsize $filename $DEFAULT_EXTSIZE
+    echo
+}
+
+test_truncate_allocated()
+{
+    echo "TEST: Set extsize after truncating a file with allocated extents"
+    local filename=$1
+    $XFS_IO_PROG \
+        -c "open -f $filename" \
+        -c "pwrite -qW  0 $FILE_DATA_SIZE" \
+        -c "truncate 0" \
+        -c "extsize $EXTSIZE" \
+
+    check_extsz_xflag_across_remount $filename $EXTSIZE
+    echo
+}
+
+test_truncate_delayed()
+{
+    echo "TEST: Set extsize after truncating a file with delayed allocation"
+    local filename=$1
+    $XFS_IO_PROG \
+        -c "open -f $filename" \
+        -c "pwrite -q  0 $FILE_DATA_SIZE" \
+        -c "truncate 0" \
+        -c "extsize $EXTSIZE" \
+
+    check_extsz_xflag_across_remount $filename $EXTSIZE
+    echo
+}
+
+setup
+echo -e "EXTSIZE = $EXTSIZE DEFAULT_EXTSIZE = $DEFAULT_EXTSIZE \
+	BLOCKSIZE = $BLKSZ\n" >> "$seqres.full"
+
+NEW_FILE_NAME_PREFIX=$SCRATCH_MNT/new-file-
+
+test_empty_file "$NEW_FILE_NAME_PREFIX"00
+test_data_delayed "$NEW_FILE_NAME_PREFIX"01
+test_data_allocated "$NEW_FILE_NAME_PREFIX"02
+test_truncate_allocated "$NEW_FILE_NAME_PREFIX"03
+test_truncate_delayed "$NEW_FILE_NAME_PREFIX"04
+
+status=0
+exit

tests/generic/366.out new file +26 -0

new file mode 100644
index 00000000..cdd2f5fa
@@ -0,0 +1,26 @@
+QA output created by 366
+TEST: Set extsize on empty file
+[EXTSIZE] SCRATCH_MNT/new-file-00
+e flag set
+Re-setting extsize hint to 0
+[EXTSIZE] SCRATCH_MNT/new-file-00
+e flag unset
+
+TEST: Set extsize on non-empty file with delayed allocation
+xfs_io: FS_IOC_FSSETXATTR SCRATCH_MNT/new-file-01: Invalid argument
+test for default extsize setting if any
+[EXTSIZE] SCRATCH_MNT/new-file-01
+
+TEST: Set extsize on non-empty file with allocated extents
+xfs_io: FS_IOC_FSSETXATTR SCRATCH_MNT/new-file-02: Invalid argument
+test for default extsize setting if any
+[EXTSIZE] SCRATCH_MNT/new-file-02
+
+TEST: Set extsize after truncating a file with allocated extents
+[EXTSIZE] SCRATCH_MNT/new-file-03
+e flag set
+
+TEST: Set extsize after truncating a file with delayed allocation
+[EXTSIZE] SCRATCH_MNT/new-file-04
+e flag set
+

Signature · 1 line

2.43.5

Darrick J. Wong

Reviewed-by on v3 3/3

2024-11-21 17:27 UTC #13 lore raw

Re: [PATCH v3 3/3] generic: Addition of new tests for extsize hints

Quoted 7 lines · Nirjhar Roy

This commit adds new tests that checks the behaviour of xfs/ext4 filesystems when extsize hint is set on file with inode size as 0, non-empty files with allocated and delalloc extents and so on. Although currently this test is placed under tests/generic, it only runs on xfs and there is an ongoing patch series[1] to enable extsize hints for ext4 as well.

Quoted 6 lines

[1] https://lore.kernel.org/linux-ext4/[email protected]/
Reviewed-by Ritesh Harjani (IBM) <[email protected]>
Reviewed-by Ojaswin Mujoo <[email protected]>
Suggested-by Ojaswin Mujoo <[email protected]>
Signed-off-by Nirjhar Roy <[email protected]>

Looks ok, though you might want to change generic/366 to a higher number because that's now taken.

Reviewed-by "Darrick J. Wong" <[email protected]>

--D

Quoted 224 lines

---
 tests/generic/366     | 175 ++++++++++++++++++++++++++++++++++++++++++
 tests/generic/366.out |  26 +++++++
 2 files changed, 201 insertions(+)
 create mode 100755 tests/generic/366
 create mode 100644 tests/generic/366.out

tests/generic/366 new file +175 -0

new file mode 100755
index 00000000..25d23f42
@@ -0,0 +1,175 @@
+#! /bin/bash
+# SPDX-License-Identifier: GPL-2.0
+# Copyright (c) 2024 Nirjhar Roy ([email protected]).  All Rights Reserved.
+#
+# FS QA Test 366
+#
+# This test verifies that extent allocation hint setting works correctly on
+# files with no extents allocated and non-empty files which are truncated.
+# It also checks that the  extent hints setting fails with non-empty file
+# i.e, with any file with allocated extents or delayed allocation. We also
+# check if the extsize value and the xflag bit actually got reflected after
+# setting/re-setting the extsize value.
+
+. ./common/config
+. ./common/filter
+. ./common/preamble
+
+_begin_fstest ioctl quick
+
+_fixed_by_kernel_commit "2a492ff66673 \
+                        xfs: Check for delayed allocations before setting \
+			extsize"
+
+_require_scratch_extsize
+
+FILE_DATA_SIZE=1M
+
+get_default_extsize()
+{
+    if [ -z $1 ] || [ ! -d $1 ]; then
+        echo "Missing mount point argument for get_default_extsize"
+        exit 1
+    fi
+    $XFS_IO_PROG -c "extsize" "$1" | sed 's/^\[\([0-9]\+\)\].*/\1/'
+}
+
+filter_extsz()
+{
+    sed "s/\[$1\]/\[EXTSIZE\]/g"
+}
+
+setup()
+{
+    _scratch_mkfs >> "$seqres.full"  2>&1
+    _scratch_mount >> "$seqres.full" 2>&1
+    BLKSZ=`_get_block_size $SCRATCH_MNT`
+    DEFAULT_EXTSIZE=`get_default_extsize $SCRATCH_MNT`
+    EXTSIZE=$(( BLKSZ*2 ))
+    # Make sure the new extsize is not the same as the default
+    # extsize so that we can observe it changing
+    [[ "$DEFAULT_EXTSIZE" -eq "$EXTSIZE" ]] && EXTSIZE=$(( BLKSZ*4 ))
+}
+
+read_file_extsize()
+{
+    $XFS_IO_PROG -c "extsize" $1 | _filter_scratch | filter_extsz $2
+}
+
+check_extsz_and_xflag()
+{
+    local filename=$1
+    local extsize=$2
+    read_file_extsize $filename $extsize
+    _test_fsxattr_xflag $filename "extsize" && echo "e flag set" || \
+	    echo "e flag unset"
+}
+
+check_extsz_xflag_across_remount()
+{
+    local filename=$1
+    local extsize=$2
+    _scratch_cycle_mount
+    check_extsz_and_xflag $filename $extsize
+}
+
+# Extsize flag should be cleared when extsize is reset, so this function
+# checks that this behavior is followed.
+reset_extsz_and_recheck_extsz_xflag()
+{
+    local filename=$1
+    echo "Re-setting extsize hint to 0"
+    $XFS_IO_PROG -c "extsize 0" $filename
+    check_extsz_xflag_across_remount $filename "0"
+}
+
+check_extsz_xflag_before_and_after_reset()
+{
+    local filename=$1
+    local extsize=$2
+    check_extsz_xflag_across_remount $filename $extsize
+    reset_extsz_and_recheck_extsz_xflag $filename
+}
+
+test_empty_file()
+{
+    echo "TEST: Set extsize on empty file"
+    local filename=$1
+    $XFS_IO_PROG \
+        -c "open -f $filename" \
+        -c "extsize $EXTSIZE" \
+
+    check_extsz_xflag_before_and_after_reset $filename $EXTSIZE
+    echo
+}
+
+test_data_delayed()
+{
+    echo "TEST: Set extsize on non-empty file with delayed allocation"
+    local filename=$1
+    $XFS_IO_PROG \
+        -c "open -f $filename" \
+        -c "pwrite -q  0 $FILE_DATA_SIZE" \
+        -c "extsize $EXTSIZE" | _filter_scratch
+
+    echo "test for default extsize setting if any"
+    read_file_extsize $filename $DEFAULT_EXTSIZE
+    echo
+}
+
+test_data_allocated()
+{
+    echo "TEST: Set extsize on non-empty file with allocated extents"
+    local filename=$1
+    $XFS_IO_PROG \
+        -c "open -f $filename" \
+        -c "pwrite -qW  0 $FILE_DATA_SIZE" \
+        -c "extsize $EXTSIZE" | _filter_scratch
+
+    echo "test for default extsize setting if any"
+    read_file_extsize $filename $DEFAULT_EXTSIZE
+    echo
+}
+
+test_truncate_allocated()
+{
+    echo "TEST: Set extsize after truncating a file with allocated extents"
+    local filename=$1
+    $XFS_IO_PROG \
+        -c "open -f $filename" \
+        -c "pwrite -qW  0 $FILE_DATA_SIZE" \
+        -c "truncate 0" \
+        -c "extsize $EXTSIZE" \
+
+    check_extsz_xflag_across_remount $filename $EXTSIZE
+    echo
+}
+
+test_truncate_delayed()
+{
+    echo "TEST: Set extsize after truncating a file with delayed allocation"
+    local filename=$1
+    $XFS_IO_PROG \
+        -c "open -f $filename" \
+        -c "pwrite -q  0 $FILE_DATA_SIZE" \
+        -c "truncate 0" \
+        -c "extsize $EXTSIZE" \
+
+    check_extsz_xflag_across_remount $filename $EXTSIZE
+    echo
+}
+
+setup
+echo -e "EXTSIZE = $EXTSIZE DEFAULT_EXTSIZE = $DEFAULT_EXTSIZE \
+	BLOCKSIZE = $BLKSZ\n" >> "$seqres.full"
+
+NEW_FILE_NAME_PREFIX=$SCRATCH_MNT/new-file-
+
+test_empty_file "$NEW_FILE_NAME_PREFIX"00
+test_data_delayed "$NEW_FILE_NAME_PREFIX"01
+test_data_allocated "$NEW_FILE_NAME_PREFIX"02
+test_truncate_allocated "$NEW_FILE_NAME_PREFIX"03
+test_truncate_delayed "$NEW_FILE_NAME_PREFIX"04
+
+status=0
+exit

tests/generic/366.out new file +26 -0

new file mode 100644
index 00000000..cdd2f5fa
@@ -0,0 +1,26 @@
+QA output created by 366
+TEST: Set extsize on empty file
+[EXTSIZE] SCRATCH_MNT/new-file-00
+e flag set
+Re-setting extsize hint to 0
+[EXTSIZE] SCRATCH_MNT/new-file-00
+e flag unset
+
+TEST: Set extsize on non-empty file with delayed allocation
+xfs_io: FS_IOC_FSSETXATTR SCRATCH_MNT/new-file-01: Invalid argument
+test for default extsize setting if any
+[EXTSIZE] SCRATCH_MNT/new-file-01
+
+TEST: Set extsize on non-empty file with allocated extents
+xfs_io: FS_IOC_FSSETXATTR SCRATCH_MNT/new-file-02: Invalid argument
+test for default extsize setting if any
+[EXTSIZE] SCRATCH_MNT/new-file-02
+
+TEST: Set extsize after truncating a file with allocated extents
+[EXTSIZE] SCRATCH_MNT/new-file-03
+e flag set
+
+TEST: Set extsize after truncating a file with delayed allocation
+[EXTSIZE] SCRATCH_MNT/new-file-04
+e flag set
+

Signature · 1 line

2.43.5

Nirjhar Roy

author

2024-11-21 18:34 UTC #14 lore raw

Re: [PATCH v3 3/3] generic: Addition of new tests for extsize hints

Quoted 14 lines · Darrick J. Wong

Quoted 7 lines · Nirjhar Roy
This commit adds new tests that checks the behaviour of xfs/ext4 filesystems when extsize hint is set on file with inode size as 0, non-empty files with allocated and delalloc extents and so on. Although currently this test is placed under tests/generic, it only runs on xfs and there is an ongoing patch series[1] to enable extsize hints for ext4 as well.
Quoted 6 lines
[1] https://lore.kernel.org/linux-ext4/[email protected]/
Reviewed-by Ritesh Harjani (IBM) <[email protected]>
Reviewed-by Ojaswin Mujoo <[email protected]>
Suggested-by Ojaswin Mujoo <[email protected]>
Signed-off-by Nirjhar Roy <[email protected]>

Quoted 6 lines

Looks ok, though you might want to change generic/366 to a higher number because that's now taken.
Reviewed-by "Darrick J. Wong" <[email protected]>
--D

Okay. I will update the test sequence number in the next revision.

--NR

Quoted 224 lines

---
  tests/generic/366     | 175 ++++++++++++++++++++++++++++++++++++++++++
  tests/generic/366.out |  26 +++++++
  2 files changed, 201 insertions(+)

create mode 100755 tests/generic/366
create mode 100644 tests/generic/366.out

tests/generic/366 new file +175 -0

new file mode 100755
index 00000000..25d23f42
@@ -0,0 +1,175 @@
+#! /bin/bash
+# SPDX-License-Identifier: GPL-2.0
+# Copyright (c) 2024 Nirjhar Roy ([email protected]).  All Rights Reserved.
+#
+# FS QA Test 366
+#
+# This test verifies that extent allocation hint setting works correctly on
+# files with no extents allocated and non-empty files which are truncated.
+# It also checks that the  extent hints setting fails with non-empty file
+# i.e, with any file with allocated extents or delayed allocation. We also
+# check if the extsize value and the xflag bit actually got reflected after
+# setting/re-setting the extsize value.
+
+. ./common/config
+. ./common/filter
+. ./common/preamble
+
+_begin_fstest ioctl quick
+
+_fixed_by_kernel_commit "2a492ff66673 \
+                        xfs: Check for delayed allocations before setting \
+			extsize"
+
+_require_scratch_extsize
+
+FILE_DATA_SIZE=1M
+
+get_default_extsize()
+{
+    if [ -z $1 ] || [ ! -d $1 ]; then
+        echo "Missing mount point argument for get_default_extsize"
+        exit 1
+    fi
+    $XFS_IO_PROG -c "extsize" "$1" | sed 's/^\[\([0-9]\+\)\].*/\1/'
+}
+
+filter_extsz()
+{
+    sed "s/\[$1\]/\[EXTSIZE\]/g"
+}
+
+setup()
+{
+    _scratch_mkfs >> "$seqres.full"  2>&1
+    _scratch_mount >> "$seqres.full" 2>&1
+    BLKSZ=`_get_block_size $SCRATCH_MNT`
+    DEFAULT_EXTSIZE=`get_default_extsize $SCRATCH_MNT`
+    EXTSIZE=$(( BLKSZ*2 ))
+    # Make sure the new extsize is not the same as the default
+    # extsize so that we can observe it changing
+    [[ "$DEFAULT_EXTSIZE" -eq "$EXTSIZE" ]] && EXTSIZE=$(( BLKSZ*4 ))
+}
+
+read_file_extsize()
+{
+    $XFS_IO_PROG -c "extsize" $1 | _filter_scratch | filter_extsz $2
+}
+
+check_extsz_and_xflag()
+{
+    local filename=$1
+    local extsize=$2
+    read_file_extsize $filename $extsize
+    _test_fsxattr_xflag $filename "extsize" && echo "e flag set" || \
+	    echo "e flag unset"
+}
+
+check_extsz_xflag_across_remount()
+{
+    local filename=$1
+    local extsize=$2
+    _scratch_cycle_mount
+    check_extsz_and_xflag $filename $extsize
+}
+
+# Extsize flag should be cleared when extsize is reset, so this function
+# checks that this behavior is followed.
+reset_extsz_and_recheck_extsz_xflag()
+{
+    local filename=$1
+    echo "Re-setting extsize hint to 0"
+    $XFS_IO_PROG -c "extsize 0" $filename
+    check_extsz_xflag_across_remount $filename "0"
+}
+
+check_extsz_xflag_before_and_after_reset()
+{
+    local filename=$1
+    local extsize=$2
+    check_extsz_xflag_across_remount $filename $extsize
+    reset_extsz_and_recheck_extsz_xflag $filename
+}
+
+test_empty_file()
+{
+    echo "TEST: Set extsize on empty file"
+    local filename=$1
+    $XFS_IO_PROG \
+        -c "open -f $filename" \
+        -c "extsize $EXTSIZE" \
+
+    check_extsz_xflag_before_and_after_reset $filename $EXTSIZE
+    echo
+}
+
+test_data_delayed()
+{
+    echo "TEST: Set extsize on non-empty file with delayed allocation"
+    local filename=$1
+    $XFS_IO_PROG \
+        -c "open -f $filename" \
+        -c "pwrite -q  0 $FILE_DATA_SIZE" \
+        -c "extsize $EXTSIZE" | _filter_scratch
+
+    echo "test for default extsize setting if any"
+    read_file_extsize $filename $DEFAULT_EXTSIZE
+    echo
+}
+
+test_data_allocated()
+{
+    echo "TEST: Set extsize on non-empty file with allocated extents"
+    local filename=$1
+    $XFS_IO_PROG \
+        -c "open -f $filename" \
+        -c "pwrite -qW  0 $FILE_DATA_SIZE" \
+        -c "extsize $EXTSIZE" | _filter_scratch
+
+    echo "test for default extsize setting if any"
+    read_file_extsize $filename $DEFAULT_EXTSIZE
+    echo
+}
+
+test_truncate_allocated()
+{
+    echo "TEST: Set extsize after truncating a file with allocated extents"
+    local filename=$1
+    $XFS_IO_PROG \
+        -c "open -f $filename" \
+        -c "pwrite -qW  0 $FILE_DATA_SIZE" \
+        -c "truncate 0" \
+        -c "extsize $EXTSIZE" \
+
+    check_extsz_xflag_across_remount $filename $EXTSIZE
+    echo
+}
+
+test_truncate_delayed()
+{
+    echo "TEST: Set extsize after truncating a file with delayed allocation"
+    local filename=$1
+    $XFS_IO_PROG \
+        -c "open -f $filename" \
+        -c "pwrite -q  0 $FILE_DATA_SIZE" \
+        -c "truncate 0" \
+        -c "extsize $EXTSIZE" \
+
+    check_extsz_xflag_across_remount $filename $EXTSIZE
+    echo
+}
+
+setup
+echo -e "EXTSIZE = $EXTSIZE DEFAULT_EXTSIZE = $DEFAULT_EXTSIZE \
+	BLOCKSIZE = $BLKSZ\n" >> "$seqres.full"
+
+NEW_FILE_NAME_PREFIX=$SCRATCH_MNT/new-file-
+
+test_empty_file "$NEW_FILE_NAME_PREFIX"00
+test_data_delayed "$NEW_FILE_NAME_PREFIX"01
+test_data_allocated "$NEW_FILE_NAME_PREFIX"02
+test_truncate_allocated "$NEW_FILE_NAME_PREFIX"03
+test_truncate_delayed "$NEW_FILE_NAME_PREFIX"04
+
+status=0
+exit

tests/generic/366.out new file +26 -0

new file mode 100644
index 00000000..cdd2f5fa
@@ -0,0 +1,26 @@
+QA output created by 366
+TEST: Set extsize on empty file
+[EXTSIZE] SCRATCH_MNT/new-file-00
+e flag set
+Re-setting extsize hint to 0
+[EXTSIZE] SCRATCH_MNT/new-file-00
+e flag unset
+
+TEST: Set extsize on non-empty file with delayed allocation
+xfs_io: FS_IOC_FSSETXATTR SCRATCH_MNT/new-file-01: Invalid argument
+test for default extsize setting if any
+[EXTSIZE] SCRATCH_MNT/new-file-01
+
+TEST: Set extsize on non-empty file with allocated extents
+xfs_io: FS_IOC_FSSETXATTR SCRATCH_MNT/new-file-02: Invalid argument
+test for default extsize setting if any
+[EXTSIZE] SCRATCH_MNT/new-file-02
+
+TEST: Set extsize after truncating a file with allocated extents
+[EXTSIZE] SCRATCH_MNT/new-file-03
+e flag set
+
+TEST: Set extsize after truncating a file with delayed allocation
+[EXTSIZE] SCRATCH_MNT/new-file-04
+e flag set
+

Signature · 1 line

2.43.5

Signature · 4 lines

---
Nirjhar Roy
Linux Kernel Developer
IBM, Bangalore