[PATCH] selftests/fs/mount-notify: fix the indentation in the fsmount test

[PATCH] selftests/fs/mount-notify: fix the indentation in the fsmount test

[Sasha Levin]

A few lines in the fsmount test are indented with spaces instead of a
tab. That puts them deeper than the ASSERT_*() in front of them, which
ends in a for loop with an empty body, so clang warns:

  mount-notify_test.c:338:18: warning: for loop has empty body [-Wempty-body]

Indent them with tabs.

Signed-off-by: Sasha Levin <sashal@kernel.org>
---
Found with clang 21; KernelCI's clang builds show the same warnings.
Amir's pending fanotify series moves this file to filesystems/fanotify/,
and its rename still applies on top of this.

 .../filesystems/mount-notify/mount-notify_test.c          | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/tools/testing/selftests/filesystems/mount-notify/mount-notify_test.c b/tools/testing/selftests/filesystems/mount-notify/mount-notify_test.c
index 6381af6a40e30..6c95c275b1629 100644
--- a/tools/testing/selftests/filesystems/mount-notify/mount-notify_test.c
+++ b/tools/testing/selftests/filesystems/mount-notify/mount-notify_test.c
@@ -337,18 +337,18 @@ TEST_F(fanotify, fsmount)
 	fs = fsopen("tmpfs", 0);
 	ASSERT_GE(fs, 0);
 
-        ret = fsconfig(fs, FSCONFIG_CMD_CREATE, 0, 0, 0);
+	ret = fsconfig(fs, FSCONFIG_CMD_CREATE, 0, 0, 0);
 	ASSERT_EQ(ret, 0);
 
-        mnt = fsmount(fs, 0, 0);
+	mnt = fsmount(fs, 0, 0);
 	ASSERT_GE(mnt, 0);
 
-        close(fs);
+	close(fs);
 
 	ret = move_mount(mnt, "", AT_FDCWD, "/a", MOVE_MOUNT_F_EMPTY_PATH);
 	ASSERT_EQ(ret, 0);
 
-        close(mnt);
+	close(mnt);
 
 	mnts[1] = expect_notify_mask(_metadata, self, FAN_MNT_ATTACH);
 	ASSERT_NE(mnts[0], mnts[1]);
-- 
2.53.0

[PATCH] selftests/filesystems: remove the unused list_del()

[Sasha Levin]

Nothing uses list_del() in utils.c, and clang warns about it in each
test that builds utils.c with -Wall:

  ../utils.c:106:20: warning: unused function 'list_del' [-Wunused-function]

Remove it.

Signed-off-by: Sasha Levin <sashal@kernel.org>
---
Found with clang 21; KernelCI's clang builds show the same warning.

 tools/testing/selftests/filesystems/utils.c | 10 ----------
 1 file changed, 10 deletions(-)

diff --git a/tools/testing/selftests/filesystems/utils.c b/tools/testing/selftests/filesystems/utils.c
index d73d7d8171db6..4321f1a552587 100644
--- a/tools/testing/selftests/filesystems/utils.c
+++ b/tools/testing/selftests/filesystems/utils.c
@@ -103,16 +103,6 @@ static inline void list_add_tail(struct list *head, struct list *list)
 	__list_add(list, head->prev, head);
 }
 
-static inline void list_del(struct list *list)
-{
-	struct list *next, *prev;
-
-	next = list->next;
-	prev = list->prev;
-	next->prev = prev;
-	prev->next = next;
-}
-
 static ssize_t read_nointr(int fd, void *buf, size_t count)
 {
 	ssize_t ret;
-- 
2.53.0

Re: (subset) [PATCH] selftests/filesystems: remove the unused list_del()

[Christian Brauner]

On Tue, 03 Nov 2026 11:54:15 -0500, Sasha Levin wrote:
> Nothing uses list_del() in utils.c, and clang warns about it in each
> test that builds utils.c with -Wall:
> 
>   ../utils.c:106:20: warning: unused function 'list_del' [-Wunused-function]
> 
> Remove it.
> 
> [...]

Applied to the vfs-7.5.misc branch of the vfs/vfs.git tree.
Patches in the vfs-7.5.misc branch should appear in linux-next soon.

Please report any outstanding bugs that were missed during review in a
new review to the original patch series allowing us to drop it.

It's encouraged to provide Acked-bys and Reviewed-bys even though the
patch has now been applied. If possible patch trailers will be updated.

Note that commit hashes shown below are subject to change due to rebase,
trailer updates or similar. If in doubt, please check the listed branch.

tree:   https://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs.git
branch: vfs-7.5.misc

[1/1] selftests/filesystems: remove the unused list_del()
      https://git.kernel.org/vfs/vfs/c/65c9cbad74a7

[PATCH] selftests/statmount: initialize the byte written to the child

[Sasha Levin]

test_listmount_ns() writes pval to the child only to let it exit, but
never sets it, and clang 21 warns:

  statmount_test_ns.c:334:35: warning: variable 'pval' is uninitialized when passed as a const pointer argument here [-Wuninitialized-const-pointer]

Initialize it to 0.

Signed-off-by: Sasha Levin <sashal@kernel.org>
---
Found with clang 21; KernelCI's clang builds show the same warning.

 .../testing/selftests/filesystems/statmount/statmount_test_ns.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/testing/selftests/filesystems/statmount/statmount_test_ns.c b/tools/testing/selftests/filesystems/statmount/statmount_test_ns.c
index e500905e4c078..e7261d4f3be4e 100644
--- a/tools/testing/selftests/filesystems/statmount/statmount_test_ns.c
+++ b/tools/testing/selftests/filesystems/statmount/statmount_test_ns.c
@@ -275,7 +275,7 @@ static int validate_external_listmount(pid_t pid, uint64_t child_nr_mounts)
 static void test_listmount_ns(void)
 {
 	uint64_t nr_mounts;
-	char pval;
+	char pval = 0;
 	int child_ready_pipe[2];
 	int parent_ready_pipe[2];
 	pid_t pid;
-- 
2.53.0

Re: (subset) [PATCH] selftests/statmount: initialize the byte written to the child

[Christian Brauner]

On Tue, 03 Nov 2026 11:54:16 -0500, Sasha Levin wrote:
> test_listmount_ns() writes pval to the child only to let it exit, but
> never sets it, and clang 21 warns:
> 
>   statmount_test_ns.c:334:35: warning: variable 'pval' is uninitialized when passed as a const pointer argument here [-Wuninitialized-const-pointer]
> 
> Initialize it to 0.
> 
> [...]

Applied to the vfs-7.5.misc branch of the vfs/vfs.git tree.
Patches in the vfs-7.5.misc branch should appear in linux-next soon.

Please report any outstanding bugs that were missed during review in a
new review to the original patch series allowing us to drop it.

It's encouraged to provide Acked-bys and Reviewed-bys even though the
patch has now been applied. If possible patch trailers will be updated.

Note that commit hashes shown below are subject to change due to rebase,
trailer updates or similar. If in doubt, please check the listed branch.

tree:   https://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs.git
branch: vfs-7.5.misc

[1/1] selftests/statmount: initialize the byte written to the child
      https://git.kernel.org/vfs/vfs/c/678e669241c8

[PATCH] selftests/namespaces: initialize the byte the children write

[Sasha Levin]

The children in listns_pagination_bug and stress_test write c to the
parent only to say they are ready, but never set it, and clang 21
warns:

  stress_test.c:137:22: warning: variable 'c' is uninitialized when passed as a const pointer argument here [-Wuninitialized-const-pointer]

Initialize it to 'X', as the parent does with its byte.

Signed-off-by: Sasha Levin <sashal@kernel.org>
---
Found with clang 21. The namespaces tests aren't in TARGETS in
tools/testing/selftests/Makefile, so CI doesn't build or run them. Is
that on purpose?

 tools/testing/selftests/namespaces/listns_pagination_bug.c | 2 +-
 tools/testing/selftests/namespaces/stress_test.c           | 4 ++--
 2 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/tools/testing/selftests/namespaces/listns_pagination_bug.c b/tools/testing/selftests/namespaces/listns_pagination_bug.c
index da7d33f963972..6452825e88506 100644
--- a/tools/testing/selftests/namespaces/listns_pagination_bug.c
+++ b/tools/testing/selftests/namespaces/listns_pagination_bug.c
@@ -46,7 +46,7 @@ TEST(pagination_with_type_filter)
 		ASSERT_GE(pids[i], 0);
 
 		if (pids[i] == 0) {
-			char c;
+			char c = 'X';
 			close(sv[0]);
 
 			if (setup_userns() < 0) {
diff --git a/tools/testing/selftests/namespaces/stress_test.c b/tools/testing/selftests/namespaces/stress_test.c
index dd7df7d6cb277..6a27b98c5bc2a 100644
--- a/tools/testing/selftests/namespaces/stress_test.c
+++ b/tools/testing/selftests/namespaces/stress_test.c
@@ -124,7 +124,7 @@ TEST(many_concurrent_namespaces)
 
 		if (pids[i] == 0) {
 			/* Child: create user namespace and wait for parent signal */
-			char c;
+			char c = 'X';
 
 			close(sv[0]);
 
@@ -366,7 +366,7 @@ TEST(listns_pagination_stress)
 		ASSERT_GE(pids[i], 0);
 
 		if (pids[i] == 0) {
-			char c;
+			char c = 'X';
 			close(sv[0]);
 
 			if (setup_userns() < 0) {
-- 
2.53.0

Re: (subset) [PATCH] selftests/namespaces: initialize the byte the children write

[Christian Brauner]

On Tue, 03 Nov 2026 11:54:17 -0500, Sasha Levin wrote:
> The children in listns_pagination_bug and stress_test write c to the
> parent only to say they are ready, but never set it, and clang 21
> warns:
> 
>   stress_test.c:137:22: warning: variable 'c' is uninitialized when passed as a const pointer argument here [-Wuninitialized-const-pointer]
> 
> Initialize it to 'X', as the parent does with its byte.
> 
> [...]

Applied to the vfs-7.5.misc branch of the vfs/vfs.git tree.
Patches in the vfs-7.5.misc branch should appear in linux-next soon.

Please report any outstanding bugs that were missed during review in a
new review to the original patch series allowing us to drop it.

It's encouraged to provide Acked-bys and Reviewed-bys even though the
patch has now been applied. If possible patch trailers will be updated.

Note that commit hashes shown below are subject to change due to rebase,
trailer updates or similar. If in doubt, please check the listed branch.

tree:   https://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs.git
branch: vfs-7.5.misc

[1/1] selftests/namespaces: initialize the byte the children write
      https://git.kernel.org/vfs/vfs/c/b2971bf97205

Re: [PATCH] selftests/namespaces: initialize the byte the children write

[Christian Brauner]

On Tue, Nov 03, 2026 at 11:54:17AM -0500, Sasha Levin wrote:
> Found with clang 21. The namespaces tests aren't in TARGETS in
> tools/testing/selftests/Makefile, so CI doesn't build or run them. Is
> that on purpose?

That's an accident. Please send a patch adding it. Thanks for spotting it.

[PATCH] selftests: add namespaces to TARGETS

[Sasha Levin]

The namespaces tests are not in TARGETS, so "make kselftest" and the CI
systems that use the top-level Makefile never build or run them. Add
them.

Signed-off-by: Sasha Levin <sashal@kernel.org>
---
Tested on vfs-7.5.misc with gcc 16 and clang 21 on x86_64: all 13
tests pass in a VM with the namespaces config fragment, and they also
build for i386 and arm64.

 tools/testing/selftests/Makefile | 1 +
 1 file changed, 1 insertion(+)

diff --git a/tools/testing/selftests/Makefile b/tools/testing/selftests/Makefile
index 07c3719f8b04b..f8fb771d3b0e8 100644
--- a/tools/testing/selftests/Makefile
+++ b/tools/testing/selftests/Makefile
@@ -85,6 +85,7 @@ TARGETS += mount_setattr
 TARGETS += move_mount_set_group
 TARGETS += mqueue
 TARGETS += mseal_system_mappings
+TARGETS += namespaces
 TARGETS += nci
 TARGETS += net
 TARGETS += net/af_unix

base-commit: 8f92edb375bc5b54b211686231d489912a91b32a
-- 
2.53.0

[PATCH nf-next] selftests: netfilter: nf_queue: drop the no-op loop initializer

[Sasha Levin]

The out-of-order verdict loop starts with "for (ooo_count; ...", and
clang warns about the unused expression:

  nf_queue.c:354:10: warning: expression result unused [-Wunused-value]

Leave the initializer empty.

Signed-off-by: Sasha Levin <sashal@kernel.org>
---
KernelCI's clang builds show this warning.

 tools/testing/selftests/net/netfilter/nf_queue.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/testing/selftests/net/netfilter/nf_queue.c b/tools/testing/selftests/net/netfilter/nf_queue.c
index cca6c3cfa9e46..bc0b2f601a0a4 100644
--- a/tools/testing/selftests/net/netfilter/nf_queue.c
+++ b/tools/testing/selftests/net/netfilter/nf_queue.c
@@ -351,7 +351,7 @@ static int mainloop(void)
 		if (opts.out_of_order) {
 			ooo_ids[ooo_count] = id;
 			if (ooo_count >= 15) {
-				for (ooo_count; ooo_count >= 0; ooo_count--) {
+				for (; ooo_count >= 0; ooo_count--) {
 					nlh = nfq_build_verdict(buf, ooo_ids[ooo_count],
 								opts.queue_num, opts.verdict);
 					if (mnl_socket_sendto(nl, nlh, nlh->nlmsg_len) < 0) {
-- 
2.53.0

Re: [PATCH nf-next] selftests: netfilter: nf_queue: drop the no-op loop initializer

[Hangbin Liu]

On Tue, Nov 03, 2026 at 11:54:18AM -0500, Sasha Levin wrote:
> The out-of-order verdict loop starts with "for (ooo_count; ...", and
> clang warns about the unused expression:
> 
>   nf_queue.c:354:10: warning: expression result unused [-Wunused-value]
> 
> Leave the initializer empty.
> 
> Signed-off-by: Sasha Levin <sashal@kernel.org>
> ---
> KernelCI's clang builds show this warning.
> 
>  tools/testing/selftests/net/netfilter/nf_queue.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/tools/testing/selftests/net/netfilter/nf_queue.c b/tools/testing/selftests/net/netfilter/nf_queue.c
> index cca6c3cfa9e46..bc0b2f601a0a4 100644
> --- a/tools/testing/selftests/net/netfilter/nf_queue.c
> +++ b/tools/testing/selftests/net/netfilter/nf_queue.c
> @@ -351,7 +351,7 @@ static int mainloop(void)
>  		if (opts.out_of_order) {
>  			ooo_ids[ooo_count] = id;
>  			if (ooo_count >= 15) {
> -				for (ooo_count; ooo_count >= 0; ooo_count--) {
> +				for (; ooo_count >= 0; ooo_count--) {
>  					nlh = nfq_build_verdict(buf, ooo_ids[ooo_count],
>  								opts.queue_num, opts.verdict);
>  					if (mnl_socket_sendto(nl, nlh, nlh->nlmsg_len) < 0) {
> -- 
> 2.53.0
> 

Reviewed-by: Hangbin Liu <liuhangbin@kylinos.cn>

Re: [PATCH nf-next] selftests: netfilter: nf_queue: drop the no-op loop initializer

[Fernando Fernandez Mancera]

On 11/3/26 5:54 PM, Sasha Levin wrote:
> The out-of-order verdict loop starts with "for (ooo_count; ...", and
> clang warns about the unused expression:
> 
>    nf_queue.c:354:10: warning: expression result unused [-Wunused-value]
> 
> Leave the initializer empty.
> 
> Signed-off-by: Sasha Levin <sashal@kernel.org>
> ---
> KernelCI's clang builds show this warning.
> 
>   tools/testing/selftests/net/netfilter/nf_queue.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/tools/testing/selftests/net/netfilter/nf_queue.c b/tools/testing/selftests/net/netfilter/nf_queue.c
> index cca6c3cfa9e46..bc0b2f601a0a4 100644
> --- a/tools/testing/selftests/net/netfilter/nf_queue.c
> +++ b/tools/testing/selftests/net/netfilter/nf_queue.c
> @@ -351,7 +351,7 @@ static int mainloop(void)
>   		if (opts.out_of_order) {
>   			ooo_ids[ooo_count] = id;
>   			if (ooo_count >= 15) {
> -				for (ooo_count; ooo_count >= 0; ooo_count--) {
> +				for (; ooo_count >= 0; ooo_count--) {
>   					nlh = nfq_build_verdict(buf, ooo_ids[ooo_count],
>   								opts.queue_num, opts.verdict);
>   					if (mnl_socket_sendto(nl, nlh, nlh->nlmsg_len) < 0) {

Reviewed-by: Fernando Fernandez Mancera <fmancera@suse.de>

Thanks!

[PATCH net-next] selftests: net: tcp_ao: make the netstat column pointer const

[Sasha Levin]

lookup_get_column() stores the result of strchr() on a const string in
a plain char pointer. Since glibc 2.43, strchr() returns a const char
pointer for a const argument when C23 features are enabled, as they
are with _GNU_SOURCE, so gcc and clang warn:

  lib/proc.c:61:16: warning: assignment discards 'const' qualifier from pointer target type [-Wdiscarded-qualifiers]

The pointer is only used to compute the length, so make it const.

Signed-off-by: Sasha Levin <sashal@kernel.org>
---
Found with gcc 16 and clang 21 on glibc 2.43; KernelCI's gcc 15 builds
show the same warning.

 tools/testing/selftests/net/tcp_ao/lib/proc.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/testing/selftests/net/tcp_ao/lib/proc.c b/tools/testing/selftests/net/tcp_ao/lib/proc.c
index 8b984fa042869..50062b527a393 100644
--- a/tools/testing/selftests/net/tcp_ao/lib/proc.c
+++ b/tools/testing/selftests/net/tcp_ao/lib/proc.c
@@ -56,7 +56,7 @@ static struct netstat *lookup_get(struct netstat *ns,
 
 static struct netstat *lookup_get_column(struct netstat *ns, const char *line)
 {
-	char *column;
+	const char *column;
 
 	column = strchr(line, ':');
 	if (!column)
-- 
2.53.0

Re: [PATCH net-next] selftests: net: tcp_ao: make the netstat column pointer const

[Hangbin Liu]

On Tue, Nov 03, 2026 at 11:54:19AM -0500, Sasha Levin wrote:
> lookup_get_column() stores the result of strchr() on a const string in
> a plain char pointer. Since glibc 2.43, strchr() returns a const char
> pointer for a const argument when C23 features are enabled, as they
> are with _GNU_SOURCE, so gcc and clang warn:
> 
>   lib/proc.c:61:16: warning: assignment discards 'const' qualifier from pointer target type [-Wdiscarded-qualifiers]
> 
> The pointer is only used to compute the length, so make it const.
> 
> Signed-off-by: Sasha Levin <sashal@kernel.org>
> ---
> Found with gcc 16 and clang 21 on glibc 2.43; KernelCI's gcc 15 builds
> show the same warning.
> 
>  tools/testing/selftests/net/tcp_ao/lib/proc.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/tools/testing/selftests/net/tcp_ao/lib/proc.c b/tools/testing/selftests/net/tcp_ao/lib/proc.c
> index 8b984fa042869..50062b527a393 100644
> --- a/tools/testing/selftests/net/tcp_ao/lib/proc.c
> +++ b/tools/testing/selftests/net/tcp_ao/lib/proc.c
> @@ -56,7 +56,7 @@ static struct netstat *lookup_get(struct netstat *ns,
>  
>  static struct netstat *lookup_get_column(struct netstat *ns, const char *line)
>  {
> -	char *column;
> +	const char *column;
>  
>  	column = strchr(line, ':');
>  	if (!column)
> -- 
> 2.53.0
> 

Reviewed-by: Hangbin Liu <liuhangbin@kylinos.cn>

Re: [PATCH net-next] selftests: net: tcp_ao: make the netstat column pointer const

[Dmitry Safonov]

On Tue, 3 Nov 2026 at 16:54, Sasha Levin <sashal@kernel.org> wrote:
>
> lookup_get_column() stores the result of strchr() on a const string in
> a plain char pointer. Since glibc 2.43, strchr() returns a const char
> pointer for a const argument when C23 features are enabled, as they
> are with _GNU_SOURCE, so gcc and clang warn:
>
>   lib/proc.c:61:16: warning: assignment discards 'const' qualifier from pointer target type [-Wdiscarded-qualifiers]
>
> The pointer is only used to compute the length, so make it const.
>
> Signed-off-by: Sasha Levin <sashal@kernel.org>


Acked-by: Dmitry Safonov <0x7f454c46@gmail.com>

Thanks!

> ---
> Found with gcc 16 and clang 21 on glibc 2.43; KernelCI's gcc 15 builds
> show the same warning.
>
>  tools/testing/selftests/net/tcp_ao/lib/proc.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/tools/testing/selftests/net/tcp_ao/lib/proc.c b/tools/testing/selftests/net/tcp_ao/lib/proc.c
> index 8b984fa042869..50062b527a393 100644
> --- a/tools/testing/selftests/net/tcp_ao/lib/proc.c
> +++ b/tools/testing/selftests/net/tcp_ao/lib/proc.c
> @@ -56,7 +56,7 @@ static struct netstat *lookup_get(struct netstat *ns,
>
>  static struct netstat *lookup_get_column(struct netstat *ns, const char *line)
>  {
> -       char *column;
> +       const char *column;
>
>         column = strchr(line, ':');
>         if (!column)
> --
> 2.53.0
>

Re: [PATCH] selftests/fs/mount-notify: fix the indentation in the fsmount test

[Christian Brauner]

On Tue, 03 Nov 2026 11:54:14 -0500, Sasha Levin wrote:
> A few lines in the fsmount test are indented with spaces instead of a
> tab. That puts them deeper than the ASSERT_*() in front of them, which
> ends in a for loop with an empty body, so clang warns:
> 
>   mount-notify_test.c:338:18: warning: for loop has empty body [-Wempty-body]
> 
> Indent them with tabs.
> 
> [...]

Applied to the vfs.fixes branch of the vfs/vfs.git tree.
Patches in the vfs.fixes branch should appear in linux-next soon.

Please report any outstanding bugs that were missed during review in a
new review to the original patch series allowing us to drop it.

It's encouraged to provide Acked-bys and Reviewed-bys even though the
patch has now been applied. If possible patch trailers will be updated.

Note that commit hashes shown below are subject to change due to rebase,
trailer updates or similar. If in doubt, please check the listed branch.

tree:   https://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs.git
branch: vfs.fixes

[1/1] selftests/fs/mount-notify: fix the indentation in the fsmount test
      https://git.kernel.org/vfs/vfs/c/5e042e208dfe