[PATCH v3 5/5]Added bitmask flag and user struct argument size as input for extensibility

[Mathura_Kumar <academic1mathura@gmail.com>]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset=yes, Size: 17604 bytes --]

Signed-off-by: Mathura_Kumar <academic1mathura@gmail.com>
---
 Documentation/userspace-api/ipc.rst   |  24 ++--
 include/linux/compat.h                |   4 +-
 include/linux/syscalls.h              |   4 +-
 ipc/mqueue.c                          | 179 +++++++++++++++-----------
 kernel/sys_ni.c                       |   2 +
 tools/testing/selftests/ipc/mq_peek.c |  54 ++++++--
 6 files changed, 163 insertions(+), 104 deletions(-)

diff --git a/Documentation/userspace-api/ipc.rst b/Documentation/userspace-api/ipc.rst
index 14a5f0ce7230..3fc46621b5c6 100644
--- a/Documentation/userspace-api/ipc.rst
+++ b/Documentation/userspace-api/ipc.rst
@@ -54,13 +54,10 @@ SYNOPSIS
 
         ssize_t mq_timedreceive2(mqd_t mqdes,
                                  struct mq_timedreceive2_args *uargs,
-                                 unsigned int flags,
+                                 size_t usize, unsigned int flags,
                                  unsigned long index,
                                  const struct timespec *abs_timeout);
 
-Note: No glibc wrapper exists for this syscall. Callers must invoke it
-directly using syscall(2).
-
 DESCRIPTION
         mq_timedreceive2() receives or peeks at a message from the
         message queue referred to by the descriptor mqdes.
@@ -78,14 +75,19 @@ DESCRIPTION
                 If not NULL, the priority of the received message is
                 stored here.
 
+        The usize argument take size of struct to make mq_timedreceive2_args
+        extensible in future if required.
+
         The flags argument controls receive behavior. The following
         flag is defined:
 
         ``MQ_PEEK``
                 Copy the message into msg_ptr without removing it from
-                the queue. The queue is not modified. If this flag is
-                not set, behavior is identical to mq_timedreceive() and
-                the message is consumed.
+                the queue. The queue is not modified.
+
+        ``MQ_RECV``
+                Copy the message into msg_ptr and consume a single message by
+                invoking the existing do_mq_timedreceive() handler.
 
         The index argument selects which message to operate on within
         the priority-ordered queue. index 0 refers to the highest
@@ -116,7 +118,7 @@ ERRORS
 
         ``EINVAL``
                 flags contains an unknown value, or index is nonzero
-                and MQ_PEEK is not set.
+                and contradictory flag was passed.
 
         ``EMSGSIZE``
                 msg_len is less than the mq_msgsize attribute of
@@ -137,7 +139,7 @@ consolidated into struct mq_timedreceive2_args rather than passed
 as individual syscall arguments. Due to limited six arguments,
 The original mq_timedreceive() consumes all six slots,
 leaving no room for extension. Consolidating the buffer parameters
-into a struct recovers two argument slots for flags and index while
+into a struct recovers three argument slots for flags, size and index while
 keeping the interface clean and forward-compatible.
 
 Future extensions can be made by adding new flag bits without
@@ -220,3 +222,7 @@ Tests for mq_timedreceive2() should cover the following:
    threads returns consistent results and does not corrupt queue
    state. Verify that a concurrent mq_receive() and mq_peek() do
    not race.
+
+9) Contradictory Operation: verify that simultaneous inconsistent flag work as
+   expected.
+
diff --git a/include/linux/compat.h b/include/linux/compat.h
index 9f5ca26e76d8..837a59559fa5 100644
--- a/include/linux/compat.h
+++ b/include/linux/compat.h
@@ -803,8 +803,8 @@ asmlinkage long compat_sys_pwritev64v2(unsigned long fd,
 		unsigned long vlen, loff_t pos, rwf_t flags);
 #endif
 asmlinkage long compat_sys_mq_timedreceive2(mqd_t mqdes, struct compat_mq_timedreceive2_args __user *uargs,
-											unsigned int flags, unsigned long index,
-											struct old_timespec32 __user *abs_timeout);
+											size_t usize, unsigned int flags, const unsigned long index,
+											const struct old_timespec32 __user *abs_timeout);
 /*
  * Deprecated system calls which are still defined in
  * include/uapi/asm-generic/unistd.h and wanted by >= 1 arch
diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h
index 993e570c90ab..e5b544464a44 100644
--- a/include/linux/syscalls.h
+++ b/include/linux/syscalls.h
@@ -750,8 +750,8 @@ asmlinkage long sys_mq_timedsend_time32(mqd_t mqdes,
 			const struct old_timespec32 __user *u_abs_timeout);
 asmlinkage long
 sys_mq_timedreceive2(mqd_t mqdes, struct mq_timedreceive2_args __user *uargs,
-		     unsigned int flags, unsigned long index,
-		     struct __kernel_timespec __user *abs_timeout);
+					size_t usize, unsigned int flags, const unsigned long index,
+					const struct __kernel_timespec __user *abs_timeout);
 asmlinkage long sys_msgget(key_t key, int msgflg);
 asmlinkage long sys_old_msgctl(int msqid, int cmd, struct msqid_ds __user *buf);
 asmlinkage long sys_msgctl(int msqid, int cmd, struct msqid_ds __user *buf);
diff --git a/ipc/mqueue.c b/ipc/mqueue.c
index 78dc414967a2..2d4c285e4aa1 100644
--- a/ipc/mqueue.c
+++ b/ipc/mqueue.c
@@ -38,6 +38,7 @@
 #include <linux/sched/wake_q.h>
 #include <linux/sched/signal.h>
 #include <linux/sched/user.h>
+#include <linux/uaccess.h>
 
 #include <net/sock.h>
 #include "util.h"
@@ -53,7 +54,10 @@ struct mqueue_fs_context {
 
 #define SEND		0
 #define RECV		1
-#define MQ_PEEK     2
+
+#define MQ_PEEK     0x02
+#define MQ_RECV     0x04
+#define MQ_VALID_FLAGS (MQ_PEEK | MQ_RECV)
 
 #define STATE_NONE	0
 #define STATE_READY	1
@@ -1266,78 +1270,85 @@ static int do_mq_timedreceive2(mqd_t mqdes, struct mq_timedreceive2_args *args,
 	struct inode *inode;
 	struct mqueue_inode_info *info;
 
-	if (!(flags & MQ_PEEK)) {
-		return do_mq_timedreceive(mqdes, args->msg_ptr, args->msg_len,
-					  args->msg_prio, ts);
-	}
-	audit_mq_sendrecv(mqdes, args->msg_len, 0, ts);
-	CLASS(fd, f)(mqdes);
-	if (fd_empty(f))
-		return -EBADF;
-
-	inode = file_inode(fd_file(f));
-	if (unlikely(fd_file(f)->f_op != &mqueue_file_operations))
-		return -EBADF;
-	info = MQUEUE_I(inode);
-	audit_file(fd_file(f));
+	if (flags & (~MQ_VALID_FLAGS))
+		return -EINVAL;
+	if ((flags & MQ_RECV) && (flags & MQ_PEEK))
+		return -EINVAL;
+	if (flags & MQ_PEEK) {
+		audit_mq_sendrecv(mqdes, args->msg_len, 0, ts);
+		CLASS(fd, f)(mqdes);
+		if (fd_empty(f))
+			return -EBADF;
+
+		inode = file_inode(fd_file(f));
+		if (unlikely(fd_file(f)->f_op != &mqueue_file_operations))
+			return -EBADF;
+		info = MQUEUE_I(inode);
+		audit_file(fd_file(f));
 
-	if (unlikely(!(fd_file(f)->f_mode & FMODE_READ)))
-		return -EBADF;
+		if (unlikely(!(fd_file(f)->f_mode & FMODE_READ)))
+			return -EBADF;
 
-	if (unlikely(args->msg_len < info->attr.mq_msgsize))
-		return -EMSGSIZE;
-	if (index >= (unsigned long)info->attr.mq_maxmsg)
-		return -ENOENT;
+		if (unlikely(args->msg_len < info->attr.mq_msgsize))
+			return -EMSGSIZE;
+		if (index >= (unsigned long)info->attr.mq_maxmsg)
+			return -ENOENT;
 
-	spin_lock(&info->lock);
-	if (info->attr.mq_curmsgs == 0) {
-		spin_unlock(&info->lock);
-		return -EAGAIN;
-	}
-	msg_ptr = mq_peek_index(info, index);
-	if (!msg_ptr) {
+		spin_lock(&info->lock);
+		if (info->attr.mq_curmsgs == 0) {
+			spin_unlock(&info->lock);
+			return -EAGAIN;
+		}
+		msg_ptr = mq_peek_index(info, index);
+		if (!msg_ptr) {
+			spin_unlock(&info->lock);
+			return -ENOENT;
+		}
+		k_m_type = msg_ptr->m_type;
+		k_m_ts = msg_ptr->m_ts;
 		spin_unlock(&info->lock);
-		return -ENOENT;
-	}
-	k_m_type = msg_ptr->m_type;
-	k_m_ts = msg_ptr->m_ts;
-	spin_unlock(&info->lock);
 
-	k_msg_buffer = alloc_msg(k_m_ts);
-	if (!k_msg_buffer)
-		return -ENOMEM;
+		k_msg_buffer = alloc_msg(k_m_ts);
+		if (!k_msg_buffer)
+			return -ENOMEM;
 
 	/*
 	 * Two spin locks are necessary here. We are avoiding atomic memory
 	 * allocation and premature allocation before confirming that
 	 * a message actually exists to peek.
 	 */
-	spin_lock(&info->lock);
-	msg_ptr = mq_peek_index(info, index);
-	if (!msg_ptr || msg_ptr->m_type != k_m_type ||
-	    msg_ptr->m_ts != k_m_ts) {
-		spin_unlock(&info->lock);
-		free_msg(k_msg_buffer);
-		return -EAGAIN;
-	}
-	if (IS_ERR(copy_msg(msg_ptr, k_msg_buffer, k_m_ts))) {
+		spin_lock(&info->lock);
+		msg_ptr = mq_peek_index(info, index);
+		if (!msg_ptr || msg_ptr->m_type != k_m_type ||
+			msg_ptr->m_ts != k_m_ts) {
+			spin_unlock(&info->lock);
+			free_msg(k_msg_buffer);
+			return -EAGAIN;
+		}
+		if (IS_ERR(copy_msg(msg_ptr, k_msg_buffer, k_m_ts))) {
+			spin_unlock(&info->lock);
+			free_msg(k_msg_buffer);
+			return -EINVAL;
+		}
 		spin_unlock(&info->lock);
-		free_msg(k_msg_buffer);
-		return -EINVAL;
-	}
-	spin_unlock(&info->lock);
 
-	ret = k_msg_buffer->m_ts;
-	if (args->msg_prio && put_user(k_m_type, args->msg_prio)) {
+		ret = k_msg_buffer->m_ts;
+		if (args->msg_prio && put_user(k_m_type, args->msg_prio)) {
+			free_msg(k_msg_buffer);
+			return -EFAULT;
+		}
+		if (store_msg(args->msg_ptr, k_msg_buffer, k_m_ts)) {
+			free_msg(k_msg_buffer);
+			return -EFAULT;
+		}
 		free_msg(k_msg_buffer);
-		return -EFAULT;
+			return ret;
 	}
-	if (store_msg(args->msg_ptr, k_msg_buffer, k_m_ts)) {
-		free_msg(k_msg_buffer);
-		return -EFAULT;
+	if (flags & MQ_RECV) {
+		return do_mq_timedreceive(mqdes, args->msg_ptr, args->msg_len, args->msg_prio, ts);
 	}
-	free_msg(k_msg_buffer);
-	return ret;
+
+	return -EINVAL;
 }
 
 SYSCALL_DEFINE5(mq_timedsend, mqd_t, mqdes, const char __user *, u_msg_ptr,
@@ -1368,17 +1379,22 @@ SYSCALL_DEFINE5(mq_timedreceive, mqd_t, mqdes, char __user *, u_msg_ptr,
 	return do_mq_timedreceive(mqdes, u_msg_ptr, msg_len, u_msg_prio, p);
 }
 
-SYSCALL_DEFINE5(mq_timedreceive2, mqd_t, mqdes,
-		struct mq_timedreceive2_args __user *, uargs, unsigned int,
-		flags, const unsigned long, index,
-		const struct __kernel_timespec __user *, u_abs_timeout)
+SYSCALL_DEFINE6(mq_timedreceive2, mqd_t, mqdes,
+				struct mq_timedreceive2_args __user *, uargs, size_t, usize,
+				unsigned int, flags, const unsigned long, index,
+				const struct __kernel_timespec __user *, u_abs_timeout)
 {
-	struct mq_timedreceive2_args args;
+	struct mq_timedreceive2_args k_args;
 	struct timespec64 ts, *p = NULL;
+	int err;
 
-	if (copy_from_user(&args, uargs, sizeof(args)))
-		return -EFAULT;
-
+	if (unlikely(usize < sizeof(struct mq_timedreceive2_args)))
+		return -EINVAL;
+	if (unlikely(usize > PAGE_SIZE))
+		return -E2BIG;
+	err = copy_struct_from_user(&k_args, sizeof(k_args), uargs, usize);
+	if (err)
+		return err;
 	if (u_abs_timeout) {
 		int res = prepare_timeout(u_abs_timeout, &ts);
 
@@ -1386,7 +1402,7 @@ SYSCALL_DEFINE5(mq_timedreceive2, mqd_t, mqdes,
 			return res;
 		p = &ts;
 	}
-	return do_mq_timedreceive2(mqdes, &args, flags, index, p);
+	return do_mq_timedreceive2(mqdes, &k_args, flags, index, p);
 }
 
 /*
@@ -1633,12 +1649,18 @@ static inline int put_compat_mq_attr(const struct mq_attr *attr,
 }
 
 static inline int get_compat_mq_args(struct mq_timedreceive2_args *args,
-									struct compat_mq_timedreceive2_args __user *uargs)
+									struct compat_mq_timedreceive2_args __user *uargs, size_t usize)
 {
 	struct compat_mq_timedreceive2_args v;
+	int err;
 
-	if (copy_from_user(&v, uargs, sizeof(v)))
-		return -EFAULT;
+	if (unlikely(usize < sizeof(struct compat_mq_timedreceive2_args)))
+		return -EINVAL;
+	if (unlikely(usize > PAGE_SIZE))
+		return -E2BIG;
+	err = copy_struct_from_user(&v, sizeof(v), uargs, usize);
+	if (err)
+		return err;
 
 	memset(args, 0, sizeof(*args));
 	args->msg_len = (size_t)v.msg_len;
@@ -1700,17 +1722,18 @@ COMPAT_SYSCALL_DEFINE3(mq_getsetattr, mqd_t, mqdes,
 	return 0;
 }
 
-COMPAT_SYSCALL_DEFINE5(mq_timedreceive2, mqd_t, mqdes,
-		       struct compat_mq_timedreceive2_args __user *, uargs,
-		       unsigned int, flags, const unsigned long, index,
-		       const struct old_timespec32 __user *, u_abs_timeout)
+COMPAT_SYSCALL_DEFINE6(mq_timedreceive2, mqd_t, mqdes,
+					struct compat_mq_timedreceive2_args __user *, uargs, size_t, usize,
+					unsigned int, flags, const unsigned long, index,
+					const struct old_timespec32 __user *, u_abs_timeout)
 {
-	struct mq_timedreceive2_args args;
+	struct mq_timedreceive2_args k_args;
 	struct timespec64 ts, *p = NULL;
+	int err;
 
-	if (get_compat_mq_args(&args, uargs))
-		return -EFAULT;
-
+	err = get_compat_mq_args(&k_args, uargs, usize);
+	if (err)
+		return err;
 	if (u_abs_timeout) {
 		int res = compat_prepare_timeout(u_abs_timeout, &ts);
 
@@ -1718,7 +1741,7 @@ COMPAT_SYSCALL_DEFINE5(mq_timedreceive2, mqd_t, mqdes,
 			return res;
 		p = &ts;
 	}
-	return do_mq_timedreceive2(mqdes, &args, flags, index, p);
+	return do_mq_timedreceive2(mqdes, &k_args, flags, index, p);
 }
 
 #endif
diff --git a/kernel/sys_ni.c b/kernel/sys_ni.c
index 658d6b8274b3..af956bc4e6fe 100644
--- a/kernel/sys_ni.c
+++ b/kernel/sys_ni.c
@@ -391,6 +391,8 @@ COND_SYSCALL(setuid16);
 /* restartable sequence */
 COND_SYSCALL(rseq);
 COND_SYSCALL(rseq_slice_yield);
+
+/* ipc */
 COND_SYSCALL(mq_timedreceive2);
 
 COND_SYSCALL(uretprobe);
diff --git a/tools/testing/selftests/ipc/mq_peek.c b/tools/testing/selftests/ipc/mq_peek.c
index d08596ae6e1c..0ec3f2fca82c 100644
--- a/tools/testing/selftests/ipc/mq_peek.c
+++ b/tools/testing/selftests/ipc/mq_peek.c
@@ -40,7 +40,8 @@
 #endif
 #endif
 
-#define MQ_PEEK 2U
+#define MQ_PEEK 0x02
+#define MQ_RECV 0x04
 
 #ifndef ARRAY_SIZE
 #define ARRAY_SIZE(arr) (sizeof(arr) / sizeof((arr)[0]))
@@ -52,11 +53,13 @@ struct mq_timedreceive2_args {
 	char *msg_ptr;
 };
 
-static long mq_timedreceive2(mqd_t mqdes, struct mq_timedreceive2_args *uargs,
-			     unsigned int flags, unsigned long index,
+#define u_struct_size sizeof(struct mq_timedreceive2_args)
+
+static long mq_timedreceive2(mqd_t mqdes, struct mq_timedreceive2_args *uargs, size_t usize,
+			     unsigned int flags, const unsigned long index,
 			     const struct timespec *timeout)
 {
-	return syscall(__NR_mq_timedreceive2, (long)mqdes, uargs, (long)flags,
+	return syscall(__NR_mq_timedreceive2, mqdes, uargs, usize, flags,
 		       index, timeout);
 }
 
@@ -109,7 +112,7 @@ static long peek(mqd_t mqd, unsigned long index, char *buf, size_t bufsz,
 		.msg_prio = prio,
 		.msg_ptr = buf,
 	};
-	return mq_timedreceive2(mqd, &args, MQ_PEEK, index, NULL);
+	return mq_timedreceive2(mqd, &args, u_struct_size, MQ_PEEK, index, NULL);
 }
 
 static long queue_depth(mqd_t mqd)
@@ -226,7 +229,7 @@ static void test_peek_buffer_too_small(void)
 
 	send_msg(mqd, PRIO_HIGH, "hello", 5);
 
-	ret = mq_timedreceive2(mqd, &args, MQ_PEEK, 0, NULL);
+	ret = mq_timedreceive2(mqd, &args, u_struct_size, MQ_PEEK, 0, NULL);
 	if (ret == -1 && errno == EMSGSIZE)
 		ksft_test_result_pass("peek with small buf [EMSGSIZE]\n");
 	else
@@ -246,11 +249,10 @@ static void test_peek_bad_msg_ptr(void)
 		.msg_prio = &prio,
 		.msg_ptr = (char *)0x1,
 	};
-
 	long ret;
 
 	send_msg(mqd, PRIO_HIGH, "payload", 7);
-	ret = mq_timedreceive2(mqd, &args, MQ_PEEK, 0, NULL);
+	ret = mq_timedreceive2(mqd, &args, u_struct_size, MQ_PEEK, 0, NULL);
 	if (ret == -1 && errno == EFAULT)
 		ksft_test_result_pass("peek bad msg_ptr [EFAULT]\n");
 	else
@@ -588,7 +590,7 @@ static void test_no_peek_flag_is_receive(void)
 
 	send_msg(mqd, PRIO_HIGH, "consume-me", 10);
 
-	ret = mq_timedreceive2(mqd, &args, 0, 0, NULL);
+	ret = mq_timedreceive2(mqd, &args, u_struct_size, MQ_RECV, 0, NULL);
 	if (ret < 0) {
 		ksft_test_result_fail("no-peek receive failed: ret=%ld errno=%d\n", ret, errno);
 		mq_close(mqd);
@@ -615,8 +617,7 @@ static void *receiver_thread(void *arg)
 	ssize_t r;
 
 	while ((r = mq_receive(ctx->mqd, buf, sizeof(buf), &prio)) > 0)
-		;
-
+			;
 	return NULL;
 }
 
@@ -680,7 +681,7 @@ static void test_peek_null_prio_ptr(void)
 
 	send_msg(mqd, PRIO_MED, "no-prio-needed", 14);
 
-	ret = mq_timedreceive2(mqd, &args, MQ_PEEK, 0, NULL);
+	ret = mq_timedreceive2(mqd, &args, u_struct_size, MQ_PEEK, 0, NULL);
 	if (ret >= 0)
 		ksft_test_result_pass("peek with NULL msg_prio ptr: OK\n");
 	else
@@ -734,6 +735,32 @@ static void test_peek_priority_matches_receive(void)
 	mq_close(mqd);
 }
 
+static void test_contradictory_operation(void)
+{
+	mqd_t mqd = open_queue("contra_op", MAX_MSG_SIZE);
+	char buf[MAX_MSG_SIZE];
+	struct mq_timedreceive2_args args = {
+		.msg_len = sizeof(buf),
+		.msg_prio = NULL,
+		.msg_ptr = buf,
+	};
+	long ret;
+
+	send_msg(mqd, PRIO_HIGH, "A", 1);
+	send_msg(mqd, PRIO_MED, "B", 1);
+	send_msg(mqd, PRIO_LOW, "C", 1);
+
+	ret = mq_timedreceive2(mqd, &args, u_struct_size, MQ_PEEK | MQ_RECV, 0, 0);
+
+	if (ret == -1 && errno == EINVAL)
+		ksft_test_result_pass("Handler rejected invalid/contradictory request [EINVAL]\n");
+	else
+		ksft_test_result_fail("Handler did not reject invalid request: ret=%ld errno=%d (%m)\n",
+							ret, errno);
+
+	mq_close(mqd);
+}
+
 static const struct {
 	const char *name;
 	void (*fn)(void);
@@ -756,6 +783,7 @@ static const struct {
 	{ "NULL msg_prio ptr", test_peek_null_prio_ptr },
 	{ "peeked prio matches mq_receive",
 	  test_peek_priority_matches_receive },
+	{"two_contradictory_operation_togther", test_contradictory_operation}
 };
 
 int main(void)
@@ -769,7 +797,7 @@ int main(void)
 	{
 		struct mq_timedreceive2_args probe_args = { 0 };
 
-		sc_ret = mq_timedreceive2((mqd_t)-1, &probe_args, MQ_PEEK, 0,
+		sc_ret = mq_timedreceive2((mqd_t)-1, &probe_args, u_struct_size, MQ_PEEK, 0,
 					  NULL);
 		if (sc_ret == -1 && errno == ENOSYS)
 			ksft_exit_skip("mq_timedreceive2 syscall not available (NR=%d ENOSYS) — is the kernel too old?\n",
-- 
2.43.0