[PATCH] generic: test that truncate(2) on fsverity is blocked

[PATCH] generic: test that truncate(2) on fsverity is blocked

[Christoph Hellwig]

truncate(2) doesn't go through open and needs special code to block
it on fsverity files.  btrfs was missing this check.

Note that truncate(2) in current core utils uses openat and ftruncate,
so add our own little test program for it.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 src/Makefile          |  2 +-
 src/truncate.c        | 17 +++++++++++++++++
 tests/generic/788     | 37 +++++++++++++++++++++++++++++++++++++
 tests/generic/788.out |  4 ++++
 4 files changed, 59 insertions(+), 1 deletion(-)
 create mode 100644 src/truncate.c
 create mode 100755 tests/generic/788
 create mode 100644 tests/generic/788.out

diff --git a/src/Makefile b/src/Makefile
index d0a4106e6be8..577d816ae859 100644
--- a/src/Makefile
+++ b/src/Makefile
@@ -21,7 +21,7 @@ TARGETS = dirstress fill fill2 getpagesize holes lstat64 \
 	t_mmap_writev_overlap checkpoint_journal mmap-rw-fault allocstale \
 	t_mmap_cow_memory_failure fake-dump-rootino dio-buf-fault rewinddir-test \
 	readdir-while-renames dio-append-buf-fault dio-write-fsync-same-fd \
-	dio-writeback-race unlink-fsync
+	dio-writeback-race unlink-fsync truncate
 
 LINUX_TARGETS = xfsctl bstat t_mtab getdevicesize preallo_rw_pattern_reader \
 	preallo_rw_pattern_writer ftrunc trunc fs_perms testx looptest \
diff --git a/src/truncate.c b/src/truncate.c
new file mode 100644
index 000000000000..4c106d26f29e
--- /dev/null
+++ b/src/truncate.c
@@ -0,0 +1,17 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2026 Christoph Hellwig.
+ */
+#include <stdio.h>
+#include <stdlib.h>
+#include <unistd.h>
+
+int main(int argc, char **argv)
+{
+	if (truncate(argv[1], 0) < 0) {
+		perror("truncate");
+		exit(1);
+	}
+
+	fprintf(stderr, "truncate succeeded\n");
+}
diff --git a/tests/generic/788 b/tests/generic/788
new file mode 100755
index 000000000000..02959037e7c3
--- /dev/null
+++ b/tests/generic/788
@@ -0,0 +1,37 @@
+#! /bin/bash
+# SPDX-License-Identifier: GPL-2.0
+# Copyright (c) 2026 Christoph Hellwig.
+#
+# FS QA Test 788
+#
+# Test that truncate(2) is blocked on fsverity files.
+#
+. ./common/preamble
+_begin_fstest auto quick verity
+
+_cleanup()
+{
+	cd /
+	_restore_fsverity_signatures
+	rm -f $tmp.*
+}
+
+. ./common/filter
+. ./common/verity
+
+_require_test
+_require_scratch_verity
+_disable_fsverity_signatures
+
+_scratch_mkfs_verity &>> $seqres.full
+_scratch_mount
+
+fsv_file=$SCRATCH_MNT/file.fsv
+
+_fsv_scratch_begin_subtest "Using truncate(2) on a verity file fails"
+_fsv_create_enable_file $fsv_file
+$here/src/truncate $fsv_file
+
+# success, all done
+status=0
+exit
diff --git a/tests/generic/788.out b/tests/generic/788.out
new file mode 100644
index 000000000000..11074b71f3eb
--- /dev/null
+++ b/tests/generic/788.out
@@ -0,0 +1,4 @@
+QA output created by 788
+
+# Using truncate(2) on a verity file fails
+truncate: Operation not permitted
-- 
2.47.3

Re: [PATCH] generic: test that truncate(2) on fsverity is blocked

[Zorro Lang]

On Mon, Jan 19, 2026 at 07:24:04AM +0100, Christoph Hellwig wrote:
> truncate(2) doesn't go through open and needs special code to block
> it on fsverity files.  btrfs was missing this check.
> 
> Note that truncate(2) in current core utils uses openat and ftruncate,
> so add our own little test program for it.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>  src/Makefile          |  2 +-
>  src/truncate.c        | 17 +++++++++++++++++
>  tests/generic/788     | 37 +++++++++++++++++++++++++++++++++++++
>  tests/generic/788.out |  4 ++++
>  4 files changed, 59 insertions(+), 1 deletion(-)
>  create mode 100644 src/truncate.c
>  create mode 100755 tests/generic/788
>  create mode 100644 tests/generic/788.out
> 
> diff --git a/src/Makefile b/src/Makefile
> index d0a4106e6be8..577d816ae859 100644
> --- a/src/Makefile
> +++ b/src/Makefile
> @@ -21,7 +21,7 @@ TARGETS = dirstress fill fill2 getpagesize holes lstat64 \
>  	t_mmap_writev_overlap checkpoint_journal mmap-rw-fault allocstale \
>  	t_mmap_cow_memory_failure fake-dump-rootino dio-buf-fault rewinddir-test \
>  	readdir-while-renames dio-append-buf-fault dio-write-fsync-same-fd \
> -	dio-writeback-race unlink-fsync
> +	dio-writeback-race unlink-fsync truncate

Need to add this into .gitignore too :)

>  
>  LINUX_TARGETS = xfsctl bstat t_mtab getdevicesize preallo_rw_pattern_reader \
>  	preallo_rw_pattern_writer ftrunc trunc fs_perms testx looptest \
> diff --git a/src/truncate.c b/src/truncate.c
> new file mode 100644
> index 000000000000..4c106d26f29e
> --- /dev/null
> +++ b/src/truncate.c
> @@ -0,0 +1,17 @@
> +// SPDX-License-Identifier: GPL-2.0
> +/*
> + * Copyright (c) 2026 Christoph Hellwig.
> + */
> +#include <stdio.h>
> +#include <stdlib.h>
> +#include <unistd.h>
> +
> +int main(int argc, char **argv)
> +{
> +	if (truncate(argv[1], 0) < 0) {
> +		perror("truncate");
> +		exit(1);
> +	}
> +
> +	fprintf(stderr, "truncate succeeded\n");
> +}
> diff --git a/tests/generic/788 b/tests/generic/788
> new file mode 100755
> index 000000000000..02959037e7c3
> --- /dev/null
> +++ b/tests/generic/788
> @@ -0,0 +1,37 @@
> +#! /bin/bash
> +# SPDX-License-Identifier: GPL-2.0
> +# Copyright (c) 2026 Christoph Hellwig.
> +#
> +# FS QA Test 788
> +#
> +# Test that truncate(2) is blocked on fsverity files.
> +#
> +. ./common/preamble
> +_begin_fstest auto quick verity
> +
> +_cleanup()
> +{
> +	cd /
> +	_restore_fsverity_signatures
> +	rm -f $tmp.*
> +}
> +
> +. ./common/filter
> +. ./common/verity
> +
> +_require_test

Does this case need TEST_DEV and TEST_DIR?

> +_require_scratch_verity
> +_disable_fsverity_signatures
> +
> +_scratch_mkfs_verity &>> $seqres.full
> +_scratch_mount
> +
> +fsv_file=$SCRATCH_MNT/file.fsv
> +
> +_fsv_scratch_begin_subtest "Using truncate(2) on a verity file fails"
> +_fsv_create_enable_file $fsv_file
> +$here/src/truncate $fsv_file

_require_test_program "truncate"

> +
> +# success, all done
> +status=0
> +exit
> diff --git a/tests/generic/788.out b/tests/generic/788.out
> new file mode 100644
> index 000000000000..11074b71f3eb
> --- /dev/null
> +++ b/tests/generic/788.out
> @@ -0,0 +1,4 @@
> +QA output created by 788
> +
> +# Using truncate(2) on a verity file fails
> +truncate: Operation not permitted
> -- 
> 2.47.3
> 
> 

Re: [PATCH] generic: test that truncate(2) on fsverity is blocked

[Christoph Hellwig]

On Wed, Jan 21, 2026 at 11:08:33PM +0800, Zorro Lang wrote:
> > +	dio-writeback-race unlink-fsync truncate
> 
> Need to add this into .gitignore too :)

True.

> > +_require_test
> 
> Does this case need TEST_DEV and TEST_DIR?

No, just copied and pasted.

> > +$here/src/truncate $fsv_file
> 
> _require_test_program "truncate"

Ok.

Re: [PATCH] generic: test that truncate(2) on fsverity is blocked

[Zorro Lang]

On Thu, Jan 22, 2026 at 06:59:08AM +0100, Christoph Hellwig wrote:
> On Wed, Jan 21, 2026 at 11:08:33PM +0800, Zorro Lang wrote:
> > > +	dio-writeback-race unlink-fsync truncate
> > 
> > Need to add this into .gitignore too :)
> 
> True.
> 
> > > +_require_test
> > 
> > Does this case need TEST_DEV and TEST_DIR?
> 
> No, just copied and pasted.
> 
> > > +$here/src/truncate $fsv_file
> > 
> > _require_test_program "truncate"
> 
> Ok.

Thanks Christoph, BTW, as you said "btrfs was missing this check", so is there
a known fix/commit related with this test case, if there's, feel free to mark
that with `_fixed_by_kernel_commit` :)

> 

Re: [PATCH] generic: test that truncate(2) on fsverity is blocked

[Christoph Hellwig]

On Thu, Jan 22, 2026 at 03:23:58PM +0800, Zorro Lang wrote:
> On Thu, Jan 22, 2026 at 06:59:08AM +0100, Christoph Hellwig wrote:
> > On Wed, Jan 21, 2026 at 11:08:33PM +0800, Zorro Lang wrote:
> > > > +	dio-writeback-race unlink-fsync truncate
> > > 
> > > Need to add this into .gitignore too :)
> > 
> > True.
> > 
> > > > +_require_test
> > > 
> > > Does this case need TEST_DEV and TEST_DIR?
> > 
> > No, just copied and pasted.
> > 
> > > > +$here/src/truncate $fsv_file
> > > 
> > > _require_test_program "truncate"
> > 
> > Ok.
> 
> Thanks Christoph, BTW, as you said "btrfs was missing this check", so is there
> a known fix/commit related with this test case, if there's, feel free to mark
> that with `_fixed_by_kernel_commit` :)

I sent a fix, but it hasn't been applied anywhere yet.