[PATCH net-next v5 0/2] vsock/test: check for null-ptr-deref when transport changes

[PATCH net-next v5 0/2] vsock/test: check for null-ptr-deref when transport changes

[Luigi Leonardi]

This series introduces a new test that checks for a null pointer 
dereference that may happen when there is a transport change[1]. This 
bug was fixed in [2].

Note that this test *cannot* fail, it hangs if it triggers a kernel
oops. The intended use-case is to run it and then check if there is any 
oops in the dmesg.

This test is based on Hyunwoo Kim's[3] and Michal's python 
reproducers[4].

[1]https://lore.kernel.org/netdev/Z2LvdTTQR7dBmPb5@v4bel-B760M-AORUS-ELITE-AX/
[2]https://lore.kernel.org/netdev/20250110083511.30419-1-sgarzare@redhat.com/
[3]https://lore.kernel.org/netdev/Z2LvdTTQR7dBmPb5@v4bel-B760M-AORUS-ELITE-AX/#t
[4]https://lore.kernel.org/netdev/2b3062e3-bdaa-4c94-a3c0-2930595b9670@rbox.co/

Signed-off-by: Luigi Leonardi <leonardi@redhat.com>
---
Changes in v5:
- Addressed Stefano's comments:
    - Use a macro for G2H transport detection
    - Improved commits and comments text
    - Rebased on latest net-next
- Link to v4: 
https://lore.kernel.org/r/20250624-test_vsock-v4-1-087c9c8e25a2@redhat.com

Changes in v4:
- Addressed Stefano's comments:
    - Minor style changes
    - Use `get_transports()` to print a warning when a G2H transport is 
    loaded
    - Removed check on second connect: Because the first connect is 
    interrupted, the socket is in an unspecified state (see man connect) 
    . This can cause strange and unexpected behaviors (connect returning 
    success on a non-existing CID).

- Link to v3: 
https://lore.kernel.org/r/20250611-test_vsock-v3-1-8414a2d4df62@redhat.com

Sorry, this took waaay longer than expected.

Changes in v3:
Addressed Stefano's and Michal's comments:
    - Added the splat text to the commit commessage.
    - Introduced commit hash that fixes the bug.
    - Not using perror anymore on pthread_* functions.
    - Listener is just created once.

- Link to v2:
https://lore.kernel.org/r/20250314-test_vsock-v2-1-3c0a1d878a6d@redhat.com

Changes in v2:
- Addressed Stefano's comments:
    - Timeout is now using current_nsec()
    - Check for return values
    - Style issues
- Added Hyunwoo Kim to Suggested-by
- Link to v1: 
https://lore.kernel.org/r/20250306-test_vsock-v1-0-0320b5accf92@redhat.com

---
Luigi Leonardi (2):
      vsock/test: Add macros to identify transports
      vsock/test: Add test for null ptr deref when transport changes

 tools/testing/vsock/Makefile     |   1 +
 tools/testing/vsock/util.h       |   4 +
 tools/testing/vsock/vsock_test.c | 170 +++++++++++++++++++++++++++++++++++++++
 3 files changed, 175 insertions(+)
---
base-commit: 647496422ba9d2784fb8e15b3fda7fe801b1f2ff
change-id: 20250306-test_vsock-3e77a9c7a245

Best regards,
-- 
Luigi Leonardi <leonardi@redhat.com>

[PATCH net-next v5 1/2] vsock/test: Add macros to identify transports

[Luigi Leonardi]

Add three new macros: TRANSPORTS_G2H, TRANSPORTS_H2G and
TRANSPORTS_LOCAL.
They can be used to identify the type of the transport(s) loaded when
using the `get_transports()` function.

Suggested-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Luigi Leonardi <leonardi@redhat.com>
---
 tools/testing/vsock/util.h | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/tools/testing/vsock/util.h b/tools/testing/vsock/util.h
index 71895192cc02313bf52784e2f77aa3b0c28a0c94..fdd4649fe2d49f57c93c4aa5dfbb37b710c65918 100644
--- a/tools/testing/vsock/util.h
+++ b/tools/testing/vsock/util.h
@@ -33,6 +33,10 @@ static const char * const transport_ksyms[] = {
 static_assert(ARRAY_SIZE(transport_ksyms) == TRANSPORT_NUM);
 static_assert(BITS_PER_TYPE(int) >= TRANSPORT_NUM);
 
+#define TRANSPORTS_G2H   (TRANSPORT_VIRTIO | TRANSPORT_VMCI | TRANSPORT_HYPERV)
+#define TRANSPORTS_H2G   (TRANSPORT_VHOST | TRANSPORT_VMCI)
+#define TRANSPORTS_LOCAL (TRANSPORT_LOOPBACK)
+
 /* Tests can either run as the client or the server */
 enum test_mode {
 	TEST_MODE_UNSET,

-- 
2.50.0

Re: [PATCH net-next v5 1/2] vsock/test: Add macros to identify transports

[Stefano Garzarella]

On Mon, Jun 30, 2025 at 06:33:03PM +0200, Luigi Leonardi wrote:
>Add three new macros: TRANSPORTS_G2H, TRANSPORTS_H2G and
>TRANSPORTS_LOCAL.
>They can be used to identify the type of the transport(s) loaded when
>using the `get_transports()` function.
>
>Suggested-by: Stefano Garzarella <sgarzare@redhat.com>
>Signed-off-by: Luigi Leonardi <leonardi@redhat.com>
>---
> tools/testing/vsock/util.h | 4 ++++
> 1 file changed, 4 insertions(+)

Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>

>
>diff --git a/tools/testing/vsock/util.h b/tools/testing/vsock/util.h
>index 71895192cc02313bf52784e2f77aa3b0c28a0c94..fdd4649fe2d49f57c93c4aa5dfbb37b710c65918 100644
>--- a/tools/testing/vsock/util.h
>+++ b/tools/testing/vsock/util.h
>@@ -33,6 +33,10 @@ static const char * const transport_ksyms[] = {
> static_assert(ARRAY_SIZE(transport_ksyms) == TRANSPORT_NUM);
> static_assert(BITS_PER_TYPE(int) >= TRANSPORT_NUM);
>
>+#define TRANSPORTS_G2H   (TRANSPORT_VIRTIO | TRANSPORT_VMCI | TRANSPORT_HYPERV)
>+#define TRANSPORTS_H2G   (TRANSPORT_VHOST | TRANSPORT_VMCI)
>+#define TRANSPORTS_LOCAL (TRANSPORT_LOOPBACK)
>+
> /* Tests can either run as the client or the server */
> enum test_mode {
> 	TEST_MODE_UNSET,
>
>-- 
>2.50.0
>

[PATCH net-next v5 2/2] vsock/test: Add test for null ptr deref when transport changes

[Luigi Leonardi]

Add a new test to ensure that when the transport changes a null pointer
dereference does not occur. The bug was reported upstream [1] and fixed
with commit 2cb7c756f605 ("vsock/virtio: discard packets if the
transport changes").

KASAN: null-ptr-deref in range [0x0000000000000060-0x0000000000000067]
CPU: 2 UID: 0 PID: 463 Comm: kworker/2:3 Not tainted
Workqueue: vsock-loopback vsock_loopback_work
RIP: 0010:vsock_stream_has_data+0x44/0x70
Call Trace:
 virtio_transport_do_close+0x68/0x1a0
 virtio_transport_recv_pkt+0x1045/0x2ae4
 vsock_loopback_work+0x27d/0x3f0
 process_one_work+0x846/0x1420
 worker_thread+0x5b3/0xf80
 kthread+0x35a/0x700
 ret_from_fork+0x2d/0x70
 ret_from_fork_asm+0x1a/0x30

Note that this test may not fail in a kernel without the fix, but it may
hang on the client side if it triggers a kernel oops.

This works by creating a socket, trying to connect to a server, and then
executing a second connect operation on the same socket but to a
different CID (0). This triggers a transport change. If the connect
operation is interrupted by a signal, this could cause a null-ptr-deref.

Since this bug is non-deterministic, we need to try several times. It
is reasonable to assume that the bug will show up within the timeout
period.

If there is a G2H transport loaded in the system, the bug is not
triggered and this test will always pass. This is because
`vsock_assign_transport`, when using CID 0, like in this case, sets
vsk->transport to `transport_g2h` that is not NULL if a G2H transport is
available.

[1]https://lore.kernel.org/netdev/Z2LvdTTQR7dBmPb5@v4bel-B760M-AORUS-ELITE-AX/

Suggested-by: Hyunwoo Kim <v4bel@theori.io>
Suggested-by: Michal Luczaj <mhal@rbox.co>
Signed-off-by: Luigi Leonardi <leonardi@redhat.com>
---
 tools/testing/vsock/Makefile     |   1 +
 tools/testing/vsock/vsock_test.c | 170 +++++++++++++++++++++++++++++++++++++++
 2 files changed, 171 insertions(+)

diff --git a/tools/testing/vsock/Makefile b/tools/testing/vsock/Makefile
index 6e0b4e95e230500f99bb9c74350701a037ecd198..88211fd132d23ecdfd56ab0815580a237889e7f2 100644
--- a/tools/testing/vsock/Makefile
+++ b/tools/testing/vsock/Makefile
@@ -5,6 +5,7 @@ vsock_test: vsock_test.o vsock_test_zerocopy.o timeout.o control.o util.o msg_ze
 vsock_diag_test: vsock_diag_test.o timeout.o control.o util.o
 vsock_perf: vsock_perf.o msg_zerocopy_common.o
 
+vsock_test: LDLIBS = -lpthread
 vsock_uring_test: LDLIBS = -luring
 vsock_uring_test: control.o util.o vsock_uring_test.o timeout.o msg_zerocopy_common.o
 
diff --git a/tools/testing/vsock/vsock_test.c b/tools/testing/vsock/vsock_test.c
index eb6f54378667ac7ed324f4823e988ec9846e41a3..be6ce764f69480c0f9c3e2288fc19cd2e74be148 100644
--- a/tools/testing/vsock/vsock_test.c
+++ b/tools/testing/vsock/vsock_test.c
@@ -22,6 +22,8 @@
 #include <signal.h>
 #include <sys/ioctl.h>
 #include <linux/time64.h>
+#include <pthread.h>
+#include <fcntl.h>
 
 #include "vsock_test_zerocopy.h"
 #include "timeout.h"
@@ -1867,6 +1869,169 @@ static void test_stream_connect_retry_server(const struct test_opts *opts)
 	close(fd);
 }
 
+#define TRANSPORT_CHANGE_TIMEOUT 2 /* seconds */
+
+static void *test_stream_transport_change_thread(void *vargp)
+{
+	pid_t *pid = (pid_t *)vargp;
+	int ret;
+
+	ret = pthread_setcanceltype(PTHREAD_CANCEL_ASYNCHRONOUS, NULL);
+	if (ret) {
+		fprintf(stderr, "pthread_setcanceltype: %d\n", ret);
+		exit(EXIT_FAILURE);
+	}
+
+	while (true) {
+		if (kill(*pid, SIGUSR1) < 0) {
+			perror("kill");
+			exit(EXIT_FAILURE);
+		}
+	}
+	return NULL;
+}
+
+static void test_transport_change_signal_handler(int signal)
+{
+	/* We need a custom handler for SIGUSR1 as the default one terminates the process. */
+}
+
+static void test_stream_transport_change_client(const struct test_opts *opts)
+{
+	__sighandler_t old_handler;
+	pid_t pid = getpid();
+	pthread_t thread_id;
+	time_t tout;
+	int ret, tr;
+
+	tr = get_transports();
+
+	/* Print a warning if there is a G2H transport loaded.
+	 * This is on a best effort basis because VMCI can be either G2H and H2G, and there is
+	 * no easy way to understand it.
+	 * The bug we are testing only appears when G2H transports are not loaded.
+	 * This is because `vsock_assign_transport`, when using CID 0, assigns a G2H transport
+	 * to vsk->transport. If none is available it is set to NULL, causing the null-ptr-deref.
+	 */
+	if (tr & TRANSPORTS_G2H)
+		fprintf(stderr, "G2H Transport detected. This test will not fail.\n");
+
+	old_handler = signal(SIGUSR1, test_transport_change_signal_handler);
+	if (old_handler == SIG_ERR) {
+		perror("signal");
+		exit(EXIT_FAILURE);
+	}
+
+	ret = pthread_create(&thread_id, NULL, test_stream_transport_change_thread, &pid);
+	if (ret) {
+		fprintf(stderr, "pthread_create: %d\n", ret);
+		exit(EXIT_FAILURE);
+	}
+
+	control_expectln("LISTENING");
+
+	tout = current_nsec() + TRANSPORT_CHANGE_TIMEOUT * NSEC_PER_SEC;
+	do {
+		struct sockaddr_vm sa = {
+			.svm_family = AF_VSOCK,
+			.svm_cid = opts->peer_cid,
+			.svm_port = opts->peer_port,
+		};
+		int s;
+
+		s = socket(AF_VSOCK, SOCK_STREAM, 0);
+		if (s < 0) {
+			perror("socket");
+			exit(EXIT_FAILURE);
+		}
+
+		ret = connect(s, (struct sockaddr *)&sa, sizeof(sa));
+		/* The connect can fail due to signals coming from the thread,
+		 * or because the receiver connection queue is full.
+		 * Ignoring also the latter case because there is no way
+		 * of synchronizing client's connect and server's accept when
+		 * connect(s) are constantly being interrupted by signals.
+		 */
+		if (ret == -1 && (errno != EINTR && errno != ECONNRESET)) {
+			perror("connect");
+			exit(EXIT_FAILURE);
+		}
+
+		/* Set CID to 0 cause a transport change. */
+		sa.svm_cid = 0;
+
+		/* Ignore return value since it can fail or not.
+		 * If the previous connect is interrupted while the
+		 * connection request is already sent, the second
+		 * connect() will wait for the response.
+		 */
+		connect(s, (struct sockaddr *)&sa, sizeof(sa));
+
+		close(s);
+
+		control_writeulong(CONTROL_CONTINUE);
+
+	} while (current_nsec() < tout);
+
+	control_writeulong(CONTROL_DONE);
+
+	ret = pthread_cancel(thread_id);
+	if (ret) {
+		fprintf(stderr, "pthread_cancel: %d\n", ret);
+		exit(EXIT_FAILURE);
+	}
+
+	ret = pthread_join(thread_id, NULL);
+	if (ret) {
+		fprintf(stderr, "pthread_join: %d\n", ret);
+		exit(EXIT_FAILURE);
+	}
+
+	if (signal(SIGUSR1, old_handler) == SIG_ERR) {
+		perror("signal");
+		exit(EXIT_FAILURE);
+	}
+}
+
+static void test_stream_transport_change_server(const struct test_opts *opts)
+{
+	int s = vsock_stream_listen(VMADDR_CID_ANY, opts->peer_port);
+
+	/* Set the socket to be nonblocking because connects that have been interrupted
+	 * (EINTR) can fill the receiver's accept queue anyway, leading to connect failure.
+	 * As of today (6.15) in such situation there is no way to understand, from the
+	 * client side, if the connection has been queued in the server or not.
+	 */
+	if (fcntl(s, F_SETFL, fcntl(s, F_GETFL, 0) | O_NONBLOCK) < 0) {
+		perror("fcntl");
+		exit(EXIT_FAILURE);
+	}
+	control_writeln("LISTENING");
+
+	while (control_readulong() == CONTROL_CONTINUE) {
+		/* Must accept the connection, otherwise the `listen`
+		 * queue will fill up and new connections will fail.
+		 * There can be more than one queued connection,
+		 * clear them all.
+		 */
+		while (true) {
+			int client = accept(s, NULL, NULL);
+
+			if (client < 0) {
+				if (errno == EAGAIN)
+					break;
+
+				perror("accept");
+				exit(EXIT_FAILURE);
+			}
+
+			close(client);
+		}
+	}
+
+	close(s);
+}
+
 static void test_stream_linger_client(const struct test_opts *opts)
 {
 	int fd;
@@ -2106,6 +2271,11 @@ static struct test_case test_cases[] = {
 		.run_client = test_stream_nolinger_client,
 		.run_server = test_stream_nolinger_server,
 	},
+	{
+		.name = "SOCK_STREAM transport change null-ptr-deref",
+		.run_client = test_stream_transport_change_client,
+		.run_server = test_stream_transport_change_server,
+	},
 	{},
 };
 

-- 
2.50.0

Re: [PATCH net-next v5 2/2] vsock/test: Add test for null ptr deref when transport changes

[Stefano Garzarella]

On Mon, Jun 30, 2025 at 06:33:04PM +0200, Luigi Leonardi wrote:
>Add a new test to ensure that when the transport changes a null pointer
>dereference does not occur. The bug was reported upstream [1] and fixed
>with commit 2cb7c756f605 ("vsock/virtio: discard packets if the
>transport changes").
>
>KASAN: null-ptr-deref in range [0x0000000000000060-0x0000000000000067]
>CPU: 2 UID: 0 PID: 463 Comm: kworker/2:3 Not tainted
>Workqueue: vsock-loopback vsock_loopback_work
>RIP: 0010:vsock_stream_has_data+0x44/0x70
>Call Trace:
> virtio_transport_do_close+0x68/0x1a0
> virtio_transport_recv_pkt+0x1045/0x2ae4
> vsock_loopback_work+0x27d/0x3f0
> process_one_work+0x846/0x1420
> worker_thread+0x5b3/0xf80
> kthread+0x35a/0x700
> ret_from_fork+0x2d/0x70
> ret_from_fork_asm+0x1a/0x30
>
>Note that this test may not fail in a kernel without the fix, but it may
>hang on the client side if it triggers a kernel oops.
>
>This works by creating a socket, trying to connect to a server, and then
>executing a second connect operation on the same socket but to a
>different CID (0). This triggers a transport change. If the connect
>operation is interrupted by a signal, this could cause a null-ptr-deref.
>
>Since this bug is non-deterministic, we need to try several times. It
>is reasonable to assume that the bug will show up within the timeout
>period.
>
>If there is a G2H transport loaded in the system, the bug is not
>triggered and this test will always pass. This is because
>`vsock_assign_transport`, when using CID 0, like in this case, sets
>vsk->transport to `transport_g2h` that is not NULL if a G2H transport is
>available.
>
>[1]https://lore.kernel.org/netdev/Z2LvdTTQR7dBmPb5@v4bel-B760M-AORUS-ELITE-AX/
>
>Suggested-by: Hyunwoo Kim <v4bel@theori.io>
>Suggested-by: Michal Luczaj <mhal@rbox.co>
>Signed-off-by: Luigi Leonardi <leonardi@redhat.com>
>---
> tools/testing/vsock/Makefile     |   1 +
> tools/testing/vsock/vsock_test.c | 170 +++++++++++++++++++++++++++++++++++++++
> 2 files changed, 171 insertions(+)

Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>

Thanks,
Stefano

>
>diff --git a/tools/testing/vsock/Makefile b/tools/testing/vsock/Makefile
>index 6e0b4e95e230500f99bb9c74350701a037ecd198..88211fd132d23ecdfd56ab0815580a237889e7f2 100644
>--- a/tools/testing/vsock/Makefile
>+++ b/tools/testing/vsock/Makefile
>@@ -5,6 +5,7 @@ vsock_test: vsock_test.o vsock_test_zerocopy.o timeout.o control.o util.o msg_ze
> vsock_diag_test: vsock_diag_test.o timeout.o control.o util.o
> vsock_perf: vsock_perf.o msg_zerocopy_common.o
>
>+vsock_test: LDLIBS = -lpthread
> vsock_uring_test: LDLIBS = -luring
> vsock_uring_test: control.o util.o vsock_uring_test.o timeout.o msg_zerocopy_common.o
>
>diff --git a/tools/testing/vsock/vsock_test.c b/tools/testing/vsock/vsock_test.c
>index eb6f54378667ac7ed324f4823e988ec9846e41a3..be6ce764f69480c0f9c3e2288fc19cd2e74be148 100644
>--- a/tools/testing/vsock/vsock_test.c
>+++ b/tools/testing/vsock/vsock_test.c
>@@ -22,6 +22,8 @@
> #include <signal.h>
> #include <sys/ioctl.h>
> #include <linux/time64.h>
>+#include <pthread.h>
>+#include <fcntl.h>
>
> #include "vsock_test_zerocopy.h"
> #include "timeout.h"
>@@ -1867,6 +1869,169 @@ static void test_stream_connect_retry_server(const struct test_opts *opts)
> 	close(fd);
> }
>
>+#define TRANSPORT_CHANGE_TIMEOUT 2 /* seconds */
>+
>+static void *test_stream_transport_change_thread(void *vargp)
>+{
>+	pid_t *pid = (pid_t *)vargp;
>+	int ret;
>+
>+	ret = pthread_setcanceltype(PTHREAD_CANCEL_ASYNCHRONOUS, NULL);
>+	if (ret) {
>+		fprintf(stderr, "pthread_setcanceltype: %d\n", ret);
>+		exit(EXIT_FAILURE);
>+	}
>+
>+	while (true) {
>+		if (kill(*pid, SIGUSR1) < 0) {
>+			perror("kill");
>+			exit(EXIT_FAILURE);
>+		}
>+	}
>+	return NULL;
>+}
>+
>+static void test_transport_change_signal_handler(int signal)
>+{
>+	/* We need a custom handler for SIGUSR1 as the default one terminates the process. */
>+}
>+
>+static void test_stream_transport_change_client(const struct test_opts *opts)
>+{
>+	__sighandler_t old_handler;
>+	pid_t pid = getpid();
>+	pthread_t thread_id;
>+	time_t tout;
>+	int ret, tr;
>+
>+	tr = get_transports();
>+
>+	/* Print a warning if there is a G2H transport loaded.
>+	 * This is on a best effort basis because VMCI can be either G2H and H2G, and there is
>+	 * no easy way to understand it.
>+	 * The bug we are testing only appears when G2H transports are not loaded.
>+	 * This is because `vsock_assign_transport`, when using CID 0, assigns a G2H transport
>+	 * to vsk->transport. If none is available it is set to NULL, causing the null-ptr-deref.
>+	 */
>+	if (tr & TRANSPORTS_G2H)
>+		fprintf(stderr, "G2H Transport detected. This test will not fail.\n");
>+
>+	old_handler = signal(SIGUSR1, test_transport_change_signal_handler);
>+	if (old_handler == SIG_ERR) {
>+		perror("signal");
>+		exit(EXIT_FAILURE);
>+	}
>+
>+	ret = pthread_create(&thread_id, NULL, test_stream_transport_change_thread, &pid);
>+	if (ret) {
>+		fprintf(stderr, "pthread_create: %d\n", ret);
>+		exit(EXIT_FAILURE);
>+	}
>+
>+	control_expectln("LISTENING");
>+
>+	tout = current_nsec() + TRANSPORT_CHANGE_TIMEOUT * NSEC_PER_SEC;
>+	do {
>+		struct sockaddr_vm sa = {
>+			.svm_family = AF_VSOCK,
>+			.svm_cid = opts->peer_cid,
>+			.svm_port = opts->peer_port,
>+		};
>+		int s;
>+
>+		s = socket(AF_VSOCK, SOCK_STREAM, 0);
>+		if (s < 0) {
>+			perror("socket");
>+			exit(EXIT_FAILURE);
>+		}
>+
>+		ret = connect(s, (struct sockaddr *)&sa, sizeof(sa));
>+		/* The connect can fail due to signals coming from the thread,
>+		 * or because the receiver connection queue is full.
>+		 * Ignoring also the latter case because there is no way
>+		 * of synchronizing client's connect and server's accept when
>+		 * connect(s) are constantly being interrupted by signals.
>+		 */
>+		if (ret == -1 && (errno != EINTR && errno != ECONNRESET)) {
>+			perror("connect");
>+			exit(EXIT_FAILURE);
>+		}
>+
>+		/* Set CID to 0 cause a transport change. */
>+		sa.svm_cid = 0;
>+
>+		/* Ignore return value since it can fail or not.
>+		 * If the previous connect is interrupted while the
>+		 * connection request is already sent, the second
>+		 * connect() will wait for the response.
>+		 */
>+		connect(s, (struct sockaddr *)&sa, sizeof(sa));
>+
>+		close(s);
>+
>+		control_writeulong(CONTROL_CONTINUE);
>+
>+	} while (current_nsec() < tout);
>+
>+	control_writeulong(CONTROL_DONE);
>+
>+	ret = pthread_cancel(thread_id);
>+	if (ret) {
>+		fprintf(stderr, "pthread_cancel: %d\n", ret);
>+		exit(EXIT_FAILURE);
>+	}
>+
>+	ret = pthread_join(thread_id, NULL);
>+	if (ret) {
>+		fprintf(stderr, "pthread_join: %d\n", ret);
>+		exit(EXIT_FAILURE);
>+	}
>+
>+	if (signal(SIGUSR1, old_handler) == SIG_ERR) {
>+		perror("signal");
>+		exit(EXIT_FAILURE);
>+	}
>+}
>+
>+static void test_stream_transport_change_server(const struct test_opts *opts)
>+{
>+	int s = vsock_stream_listen(VMADDR_CID_ANY, opts->peer_port);
>+
>+	/* Set the socket to be nonblocking because connects that have been interrupted
>+	 * (EINTR) can fill the receiver's accept queue anyway, leading to connect failure.
>+	 * As of today (6.15) in such situation there is no way to understand, from the
>+	 * client side, if the connection has been queued in the server or not.
>+	 */
>+	if (fcntl(s, F_SETFL, fcntl(s, F_GETFL, 0) | O_NONBLOCK) < 0) {
>+		perror("fcntl");
>+		exit(EXIT_FAILURE);
>+	}
>+	control_writeln("LISTENING");
>+
>+	while (control_readulong() == CONTROL_CONTINUE) {
>+		/* Must accept the connection, otherwise the `listen`
>+		 * queue will fill up and new connections will fail.
>+		 * There can be more than one queued connection,
>+		 * clear them all.
>+		 */
>+		while (true) {
>+			int client = accept(s, NULL, NULL);
>+
>+			if (client < 0) {
>+				if (errno == EAGAIN)
>+					break;
>+
>+				perror("accept");
>+				exit(EXIT_FAILURE);
>+			}
>+
>+			close(client);
>+		}
>+	}
>+
>+	close(s);
>+}
>+
> static void test_stream_linger_client(const struct test_opts *opts)
> {
> 	int fd;
>@@ -2106,6 +2271,11 @@ static struct test_case test_cases[] = {
> 		.run_client = test_stream_nolinger_client,
> 		.run_server = test_stream_nolinger_server,
> 	},
>+	{
>+		.name = "SOCK_STREAM transport change null-ptr-deref",
>+		.run_client = test_stream_transport_change_client,
>+		.run_server = test_stream_transport_change_server,
>+	},
> 	{},
> };
>
>
>-- 
>2.50.0
>

Re: [PATCH net-next v5 0/2] vsock/test: check for null-ptr-deref when transport changes

[patchwork-bot+netdevbpf]

Hello:

This series was applied to netdev/net-next.git (main)
by Jakub Kicinski <kuba@kernel.org>:

On Mon, 30 Jun 2025 18:33:02 +0200 you wrote:
> This series introduces a new test that checks for a null pointer
> dereference that may happen when there is a transport change[1]. This
> bug was fixed in [2].
> 
> Note that this test *cannot* fail, it hangs if it triggers a kernel
> oops. The intended use-case is to run it and then check if there is any
> oops in the dmesg.
> 
> [...]

Here is the summary with links:
  - [net-next,v5,1/2] vsock/test: Add macros to identify transports
    https://git.kernel.org/netdev/net-next/c/e84b20b25d37
  - [net-next,v5,2/2] vsock/test: Add test for null ptr deref when transport changes
    https://git.kernel.org/netdev/net-next/c/3a764d93385c

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html