* [PATCH] [PATCH V5] mqueue: introduce new do_mq_timedreceive2() [ mq_peek syscall] for non-destructive receive and inspection,fix minor issue,prepared doc.
@ 2026-03-06 7:50 Mathura_Kumar
2026-03-06 9:38 ` Geert Uytterhoeven
` (3 more replies)
0 siblings, 4 replies; 6+ messages in thread
From: Mathura_Kumar @ 2026-03-06 7:50 UTC (permalink / raw)
To: brauner; +Cc: academic1mathura, linux-arch, linux-kernel, viro
Change from v3 to v5:
-Styles issue fixed
-Doc was prepared/modified in Documentation/userspace-api/ipc.rst,
-Corrected file references
-I am squashing previous four commit in this latest one for convenience in applying
and merging
-risc-v related macro handling error fixed
-some struct definition removed from ipc/mqueue.c for better compiler control flow
Required Attention:-As per mail from bot
kernel test robot noticed the following build errors:
[auto build test ERROR on shuah-kselftest/next]
[also build test ERROR on shuah-kselftest/fixes next-20260304]
[cannot apply to arnd-asm-generic/master linus/master tip/x86/asm v6.16-rc1]
Short Description:
POSIX message queues currently lack a mechanism to read
a message without removing it from the queue. This is a
long-standing limitation,when we require inspection of queue state
without altering it.
Modifying existing mq_receive() semantics via additional
flags was considered. However, altering behavior of an
existing syscall risks breaking backward compatibility
for applications relying on current semantics. Since
mq_receive() guarantees message removal, changing this
contract is not safe.
To preserve ABI stability, this patch introduces a new
system call that performs a non-destructive receive
operation (peek). The existing behavior remains unchanged.
Design considerations:
Two approaches for copying message data to userspace
were evaluated:
1) Refcount-based message lifecycle handling
- This can help us Avoids intermediate temp kernel copy
- Extends message lifetime
-But this may increase writer starvation under heavy load and
add unnecessary complication on priority management and
delay more time to free space in inode due refcount may prevent
2) Temporary kernel buffer copy
- Copies message into a bounded kernel buffer
- Reduces time message remains locked
- Improves fairness under write-heavy workloads
- Simpler lifetime management
My implementation adopts the temporary buffer approach
to minimize starvation and reduce locking complexity.
The design allows future transition if refcounting is
deemed preferable.
Architecture support: ALL
Testing:
- 15+ functional test cases
- Multi-threaded producer/consumer scenarios
- concurrent pop and peek
- Edge cases: empty queue, FIFO
invalid flags, signal interruption etc.
Use Case:
1) Observability in distributed systems
e.g Monitoring tools can inspect queue contents without interfering with
normal processing
2) Check pointing system can have a look into queue without consuming and can store messages
3) Resource-aware processing
4) Selective Consumption for Specialized Workers
Signed-off-by: Mathura_Kumar <academic1mathura@gmail.com>
---
Documentation/userspace-api/index.rst | 1 +
Documentation/userspace-api/ipc.rst | 273 +++++++
arch/alpha/kernel/syscalls/syscall.tbl | 1 +
arch/arm/tools/syscall.tbl | 2 +
arch/arm64/tools/syscall_32.tbl | 2 +
arch/m68k/kernel/syscalls/syscall.tbl | 3 +
arch/microblaze/kernel/syscalls/syscall.tbl | 3 +
arch/mips/kernel/syscalls/syscall_n32.tbl | 2 +
arch/mips/kernel/syscalls/syscall_n64.tbl | 1 +
arch/mips/kernel/syscalls/syscall_o32.tbl | 2 +
arch/powerpc/kernel/syscalls/syscall.tbl | 3 +
arch/s390/kernel/syscalls/syscall.tbl | 1 +
arch/sh/kernel/syscalls/syscall.tbl | 2 +
arch/sparc/kernel/syscalls/syscall.tbl | 3 +
arch/x86/entry/syscalls/syscall_32.tbl | 2 +
arch/x86/entry/syscalls/syscall_64.tbl | 2 +
arch/xtensa/kernel/syscalls/syscall.tbl | 2 +
include/linux/syscalls.h | 10 +
include/uapi/asm-generic/unistd.h | 12 +-
ipc/mqueue.c | 184 +++++
ipc/msg.c | 2 +-
ipc/msgutil.c | 48 +-
ipc/util.h | 3 +-
kernel/sys_ni.c | 2 +
scripts/checksyscalls.sh | 1 +
scripts/syscall.tbl | 3 +
tools/testing/selftests/ipc/.gitignore | 1 +
tools/testing/selftests/ipc/Makefile | 5 +-
tools/testing/selftests/ipc/mq_peek.c | 779 ++++++++++++++++++++
29 files changed, 1323 insertions(+), 32 deletions(-)
create mode 100644 Documentation/userspace-api/ipc.rst
create mode 100644 tools/testing/selftests/ipc/mq_peek.c
diff --git a/Documentation/userspace-api/index.rst b/Documentation/userspace-api/index.rst
index a68b1bea57a8..bf747a91e74a 100644
--- a/Documentation/userspace-api/index.rst
+++ b/Documentation/userspace-api/index.rst
@@ -22,6 +22,7 @@ System calls
ioctl/index
mseal
rseq
+ ipc
Security-related interfaces
===========================
diff --git a/Documentation/userspace-api/ipc.rst b/Documentation/userspace-api/ipc.rst
new file mode 100644
index 000000000000..8feb98e6b02c
--- /dev/null
+++ b/Documentation/userspace-api/ipc.rst
@@ -0,0 +1,273 @@
+mq_timedreceive2 system call
+=============================
+
+This document describes the mq_timedreceive2() system call. It provides
+an overview of the feature, its motivation, interface specification,
+design, and test specification.
+
+Contents
+--------
+
+ 1) Overview
+ 2) Motivation
+ 3) Functional Specification
+ 4) Design
+ 5) Implementation Notes
+ 6) Test Specification
+
+1) Overview
+-----------
+
+POSIX message queues on Linux provide mq_receive() and mq_timedreceive()
+for consuming messages from a queue. Both interfaces require the caller
+to pass the message buffer, length, and priority pointer as individual
+arguments to the system call. This imposes a fixed calling convention
+that cannot be extended without breaking the ABI.
+
+mq_timedreceive2() introduces a new system call entry point that accepts
+message buffer parameters via a struct argument rather than as individual
+syscall arguments. This frees the remaining syscall argument slots for
+new functionality flags and a message index, enabling non-destructive
+peek and indexed access semantics that are not possible with the
+original interface.
+
+Two variants are provided:
+
+ mq_timedreceive2() - primary variant, 64-bit time (Y2038-safe)
+ mq_timedreceive2_time32() - 32-bit time variant for legacy and compat
+
+2) Motivation
+-------------
+
+The original mq_timedreceive() interface exhausts all usable syscall
+argument slots with mqdes, msg_ptr, msg_len, msg_prio, and abs_timeout.
+Adding new behavior such as peek-without-consume or indexed message
+access would require either a new syscall with a completely different
+signature, or overloading existing arguments in ways that harm clarity
+and correctness.
+
+By consolidating the message buffer parameters into a struct, this
+syscall recovers argument slots for:
+
+ flags - controls peek vs consume and other future behavior
+ index - selects a specific message position in the priority queue
+
+This design follows the precedent of other Linux syscalls that use
+struct arguments to allow forward-compatible extensibility, such as
+clone3() and openat2().
+
+2.1 Non-destructive peek
+~~~~~~~~~~~~~~~~~~~~~~~~
+
+When the appropriate flag is set, mq_timedreceive2() copies the
+message into the caller's buffer without removing it from the queue.
+The queue state and priority ordering are fully preserved. This allows
+applications to inspect a message before deciding whether to consume
+it, eliminating the need to re-queue messages and the race conditions
+that re-queuing introduces in multi-consumer scenarios.
+
+2.2 Indexed access
+~~~~~~~~~~~~~~~~~~
+
+The index argument allows the caller to address a specific position
+within the priority-ordered queue rather than always receiving the
+highest priority message. This is useful for queue inspection,
+debugging, and monitoring tools that need visibility into queue
+contents without disturbing normal consumers.
+
+3) Functional Specification
+---------------------------
+
+NAME
+ mq_timedreceive2 - receive or peek at a message from a
+ POSIX message queue
+
+SYNOPSIS
+
+.. code-block:: c
+
+ #include <mqueue.h>
+
+ struct mq_timedreceive2_args {
+ size_t msg_len;
+ unsigned int *msg_prio;
+ char *msg_ptr;
+ };
+
+ ssize_t mq_timedreceive2(mqd_t mqdes,
+ struct mq_timedreceive2_args *uargs,
+ 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.
+
+ The uargs structure provides the message buffer parameters:
+
+ ``msg_ptr``
+ Userspace buffer to receive the message body.
+
+ ``msg_len``
+ Size of msg_ptr in bytes. Must be greater than or equal
+ to the mq_msgsize attribute of the queue.
+
+ ``msg_prio``
+ If not NULL, the priority of the received message is
+ stored here.
+
+ 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 index argument selects which message to operate on within
+ the priority-ordered queue. index 0 refers to the highest
+ priority message. When MQ_PEEK is not set, index is ignored
+ but must be 0.
+
+ The abs_timeout argument specifies an absolute timeout. When
+ MQ_PEEK is set, abs_timeout is ignored since peek is a
+ non-blocking snapshot operation. When MQ_PEEK is not set,
+ abs_timeout behaves identically to mq_timedreceive().
+
+RETURN VALUE
+ On success, returns the number of bytes copied into msg_ptr.
+ On failure, returns -1 and sets errno.
+
+ERRORS
+ ``EAGAIN``
+ Queue is empty and MQ_PEEK is set. Peek is always
+ non-blocking and returns immediately on empty queue.
+
+ ``EBADF``
+ mqdes is not a valid message queue descriptor open
+ for reading.
+
+ ``EFAULT``
+ uargs, msg_ptr, msg_prio, or abs_timeout points to
+ an invalid address.
+
+ ``EINVAL``
+ flags contains an unknown value, or index is nonzero
+ and MQ_PEEK is not set.
+
+ ``EMSGSIZE``
+ msg_len is less than the mq_msgsize attribute of
+ the queue.
+
+ ``ETIMEDOUT``
+ Pop path only. The call timed out before a message
+ became available. Never returned on peek path.
+
+4) Design
+---------
+
+4.1 Struct-based argument passing
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+The message buffer parameters (msg_ptr, msg_len, msg_prio) are
+consolidated into struct mq_timedreceive2_args rather than passed
+as individual syscall arguments. Linux syscalls are limited to 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
+keeping the interface clean and forward-compatible.
+
+Future extensions can be made by adding new flag bits without
+requiring a new syscall number, following the same extensibility
+model as openat2() and clone3().
+
+4.2 Compat handling
+~~~~~~~~~~~~~~~~~~~~
+
+32-bit userspace on 64-bit kernels requires a separate compat
+variant because pointer sizes and alignment differ between ABIs.
+mq_timedreceive2_time32() accepts struct compat_mq_timedreceive2_args
+which uses compat_uptr_t and compat_size_t in place of native pointer
+and size_t types, and struct old_timespec32 for the timeout.
+
+.. code-block:: c
+
+ struct compat_mq_timedreceive2_args {
+ compat_size_t msg_len;
+ compat_uptr_t msg_prio;
+ compat_uptr_t msg_ptr;
+ };
+
+The compat entry point performs the necessary conversions before
+calling the shared do_mq_timedreceive2() implementation.
+
+4.3 Peek implementation
+~~~~~~~~~~~~~~~~~~~~~~~~
+
+When MQ_PEEK is set, the implementation locates the target message
+in the priority tree but does not remove it. Two locks are taken:
+the first confirms a message exists before any allocation is
+attempted, avoiding allocation on empty queues. The second protects
+the kernel temporary buffer copy operation. The message is copied to
+userspace and both locks are released with the queue unmodified.
+
+4.4 Index argument
+~~~~~~~~~~~~~~~~~~~
+
+The priority tree is walked to the node at position index within
+the priority ordering. index 0 is always the highest priority
+message, consistent with what an unconditional mq_timedreceive()
+would return.
+
+5) Implementation Notes
+-----------------------
+
+The implementation lives in ipc/mqueue.c. The two syscall entry
+points mq_timedreceive2 and mq_timedreceive2_time32 are thin wrappers
+that validate and convert arguments before calling the shared internal
+function do_mq_timedreceive2(). This mirrors the existing pattern used
+by mq_timedreceive and mq_timedreceive_time32.
+
+Syscall numbers are assigned for all supported architectures. Refer
+to the respective syscall table files under arch/ for per-architecture
+numbers.
+
+6) Test Specification
+---------------------
+
+Tests for mq_timedreceive2() should cover the following:
+
+1) Basic receive: verify that without MQ_PEEK the message is consumed
+ and queue depth decreases by one. Verify message body and priority
+ are correct.
+
+2) Peek semantics: verify that with MQ_PEEK the message body and
+ priority are returned correctly and the queue depth is unchanged
+ after the call.
+
+3) Repeated peek: verify that calling mq_timedreceive2() with MQ_PEEK
+ multiple times on the same queue returns the same message each time
+ without modifying the queue.
+
+4) Index argument: verify that index 0 returns the highest priority
+ message. Verify that out-of-range index returns ENOENT.
+
+5) Empty queue: verify that peek on an empty queue returns EAGAIN
+ immediately without blocking.
+
+6) Memory behavior: verify that both small and large message copies
+ work correctly without corruption.
+
+7) Invalid arguments: verify EBADF for invalid mqdes, EFAULT for bad
+ pointers in uargs, EINVAL for unknown flags, EMSGSIZE for msg_len
+ smaller than queue mq_msgsize.
+
+8) Concurrent access: verify that simultaneous peek from multiple
+ threads returns consistent results and does not corrupt queue
+ state. Verify that a concurrent mq_receive() and mq_peek() do
+ not race or oops.
diff --git a/arch/alpha/kernel/syscalls/syscall.tbl b/arch/alpha/kernel/syscalls/syscall.tbl
index f31b7afffc34..0ff42d241419 100644
--- a/arch/alpha/kernel/syscalls/syscall.tbl
+++ b/arch/alpha/kernel/syscalls/syscall.tbl
@@ -511,3 +511,4 @@
579 common file_setattr sys_file_setattr
580 common listns sys_listns
581 common rseq_slice_yield sys_rseq_slice_yield
+582 common mq_timedreceive2 sys_mq_timedreceive2
diff --git a/arch/arm/tools/syscall.tbl b/arch/arm/tools/syscall.tbl
index 94351e22bfcf..bcec28be852c 100644
--- a/arch/arm/tools/syscall.tbl
+++ b/arch/arm/tools/syscall.tbl
@@ -486,3 +486,5 @@
469 common file_setattr sys_file_setattr
470 common listns sys_listns
471 common rseq_slice_yield sys_rseq_slice_yield
+472 common mq_timedreceive2 sys_mq_timedreceive2_time32
+473 common mq_timedreceive2_time64 sys_mq_timedreceive2
diff --git a/arch/arm64/tools/syscall_32.tbl b/arch/arm64/tools/syscall_32.tbl
index 62d93d88e0fe..02603704231e 100644
--- a/arch/arm64/tools/syscall_32.tbl
+++ b/arch/arm64/tools/syscall_32.tbl
@@ -483,3 +483,5 @@
469 common file_setattr sys_file_setattr
470 common listns sys_listns
471 common rseq_slice_yield sys_rseq_slice_yield
+472 common mq_timedreceive2_time64 sys_mq_timedreceive2
+473 common mq_timedreceive2 sys_mq_timedreceive2_time32
diff --git a/arch/m68k/kernel/syscalls/syscall.tbl b/arch/m68k/kernel/syscalls/syscall.tbl
index 248934257101..b67f23edda11 100644
--- a/arch/m68k/kernel/syscalls/syscall.tbl
+++ b/arch/m68k/kernel/syscalls/syscall.tbl
@@ -471,3 +471,6 @@
469 common file_setattr sys_file_setattr
470 common listns sys_listns
471 common rseq_slice_yield sys_rseq_slice_yield
+472 common mq_timedreceive2 sys_mq_timedreceive2_time32
+473 common mq_timedreceive2_time64 sys_mq_timedreceive2
+
diff --git a/arch/microblaze/kernel/syscalls/syscall.tbl b/arch/microblaze/kernel/syscalls/syscall.tbl
index 223d26303627..89fb60006a99 100644
--- a/arch/microblaze/kernel/syscalls/syscall.tbl
+++ b/arch/microblaze/kernel/syscalls/syscall.tbl
@@ -477,3 +477,6 @@
469 common file_setattr sys_file_setattr
470 common listns sys_listns
471 common rseq_slice_yield sys_rseq_slice_yield
+472 common mq_timedreceive2 sys_mq_timedreceive2_time32
+473 common mq_timedreceive2_time64 sys_mq_timedreceive2
+
diff --git a/arch/mips/kernel/syscalls/syscall_n32.tbl b/arch/mips/kernel/syscalls/syscall_n32.tbl
index 7430714e2b8f..fee830be67a6 100644
--- a/arch/mips/kernel/syscalls/syscall_n32.tbl
+++ b/arch/mips/kernel/syscalls/syscall_n32.tbl
@@ -410,3 +410,5 @@
469 n32 file_setattr sys_file_setattr
470 n32 listns sys_listns
471 n32 rseq_slice_yield sys_rseq_slice_yield
+472 n32 mq_timedreceive2 sys_mq_timedreceive2_time32
+473 n32 mq_timedreceive2_time64 sys_mq_timedreceive2
diff --git a/arch/mips/kernel/syscalls/syscall_n64.tbl b/arch/mips/kernel/syscalls/syscall_n64.tbl
index 630aab9e5425..75de6ee2df94 100644
--- a/arch/mips/kernel/syscalls/syscall_n64.tbl
+++ b/arch/mips/kernel/syscalls/syscall_n64.tbl
@@ -386,3 +386,4 @@
469 n64 file_setattr sys_file_setattr
470 n64 listns sys_listns
471 n64 rseq_slice_yield sys_rseq_slice_yield
+472 n64 mq_timedreceive2 sys_mq_timedreceive2
diff --git a/arch/mips/kernel/syscalls/syscall_o32.tbl b/arch/mips/kernel/syscalls/syscall_o32.tbl
index 128653112284..8694a2d2f084 100644
--- a/arch/mips/kernel/syscalls/syscall_o32.tbl
+++ b/arch/mips/kernel/syscalls/syscall_o32.tbl
@@ -459,3 +459,5 @@
469 o32 file_setattr sys_file_setattr
470 o32 listns sys_listns
471 o32 rseq_slice_yield sys_rseq_slice_yield
+472 o32 mq_timedreceive2 sys_mq_timedreceive2_time32
+473 o32 mq_timedreceive2_time64 sys_mq_timedreceive2
diff --git a/arch/powerpc/kernel/syscalls/syscall.tbl b/arch/powerpc/kernel/syscalls/syscall.tbl
index 4fcc7c58a105..fd90a2f500b6 100644
--- a/arch/powerpc/kernel/syscalls/syscall.tbl
+++ b/arch/powerpc/kernel/syscalls/syscall.tbl
@@ -562,3 +562,6 @@
469 common file_setattr sys_file_setattr
470 common listns sys_listns
471 nospu rseq_slice_yield sys_rseq_slice_yield
+472 32 mq_timedreceive2 sys_mq_timedreceive2_time32
+473 64 mq_timedreceive2 sys_mq_timedreceive2
+474 32 mq_timedreceive2_time64 sys_mq_timedreceive2 sys_mq_timedreceive2
diff --git a/arch/s390/kernel/syscalls/syscall.tbl b/arch/s390/kernel/syscalls/syscall.tbl
index 09a7ef04d979..34df682061d9 100644
--- a/arch/s390/kernel/syscalls/syscall.tbl
+++ b/arch/s390/kernel/syscalls/syscall.tbl
@@ -398,3 +398,4 @@
469 common file_setattr sys_file_setattr
470 common listns sys_listns
471 common rseq_slice_yield sys_rseq_slice_yield
+472 common mq_timedreceive2 sys_mq_timedreceive2
diff --git a/arch/sh/kernel/syscalls/syscall.tbl b/arch/sh/kernel/syscalls/syscall.tbl
index 70b315cbe710..1796ccd23d10 100644
--- a/arch/sh/kernel/syscalls/syscall.tbl
+++ b/arch/sh/kernel/syscalls/syscall.tbl
@@ -475,3 +475,5 @@
469 common file_setattr sys_file_setattr
470 common listns sys_listns
471 common rseq_slice_yield sys_rseq_slice_yield
+472 common mq_timedreceive2 sys_mq_timedreceive2_time32
+473 common mq_timedreceive2_time64 sys_mq_timedreceive2
diff --git a/arch/sparc/kernel/syscalls/syscall.tbl b/arch/sparc/kernel/syscalls/syscall.tbl
index 7e71bf7fcd14..5cfde6e0e6ba 100644
--- a/arch/sparc/kernel/syscalls/syscall.tbl
+++ b/arch/sparc/kernel/syscalls/syscall.tbl
@@ -517,3 +517,6 @@
469 common file_setattr sys_file_setattr
470 common listns sys_listns
471 common rseq_slice_yield sys_rseq_slice_yield
+472 32 mq_timedreceive2 sys_mq_timedreceive2_time32
+473 64 mq_timedreceive2 sys_mq_timedreceive2
+474 32 mq_timedreceive2_time64 sys_mq_timedreceive2 sys_mq_timedreceive2
diff --git a/arch/x86/entry/syscalls/syscall_32.tbl b/arch/x86/entry/syscalls/syscall_32.tbl
index f832ebd2d79b..d6701f763203 100644
--- a/arch/x86/entry/syscalls/syscall_32.tbl
+++ b/arch/x86/entry/syscalls/syscall_32.tbl
@@ -477,3 +477,5 @@
469 i386 file_setattr sys_file_setattr
470 i386 listns sys_listns
471 i386 rseq_slice_yield sys_rseq_slice_yield
+472 i386 mq_timedreceive2 sys_mq_timedreceive2_time32
+473 i386 mq_timedreceive2_time64 sys_mq_timedreceive2
diff --git a/arch/x86/entry/syscalls/syscall_64.tbl b/arch/x86/entry/syscalls/syscall_64.tbl
index 524155d655da..539654953280 100644
--- a/arch/x86/entry/syscalls/syscall_64.tbl
+++ b/arch/x86/entry/syscalls/syscall_64.tbl
@@ -441,3 +441,5 @@
547 x32 pwritev2 compat_sys_pwritev64v2
# This is the end of the legacy x32 range. Numbers 548 and above are
# not special and are not to be used for x32-specific syscalls.
+# Adding new syscalls from here to keep marginal gap intact in 472 to 511 for future expansion
+548 common mq_timedreceive2 sys_mq_timedreceive2
diff --git a/arch/xtensa/kernel/syscalls/syscall.tbl b/arch/xtensa/kernel/syscalls/syscall.tbl
index a9bca4e484de..9f69ef99e617 100644
--- a/arch/xtensa/kernel/syscalls/syscall.tbl
+++ b/arch/xtensa/kernel/syscalls/syscall.tbl
@@ -442,3 +442,5 @@
469 common file_setattr sys_file_setattr
470 common listns sys_listns
471 common rseq_slice_yield sys_rseq_slice_yield
+472 common mq_timedreceive2 sys_mq_timedreceive2_time32
+473 common mq_timedreceive2_time64 sys_mq_timedreceive2
diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h
index 02bd6ddb6278..307cfd6f154c 100644
--- a/include/linux/syscalls.h
+++ b/include/linux/syscalls.h
@@ -93,6 +93,7 @@ struct file_attr;
#include <linux/key.h>
#include <linux/personality.h>
#include <trace/syscall.h>
+#include <linux/mqueue.h>
#ifdef CONFIG_ARCH_HAS_SYSCALL_WRAPPER
/*
@@ -746,6 +747,15 @@ asmlinkage long sys_mq_timedsend_time32(mqd_t mqdes,
const char __user *u_msg_ptr,
unsigned int msg_len, unsigned int msg_prio,
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);
+asmlinkage long
+sys_mq_timedreceive2_time32(mqd_t mqdes,
+ struct mq_timedreceive2_args __user *uargs,
+ unsigned int flags, unsigned long index,
+ struct old_timespec32 __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/include/uapi/asm-generic/unistd.h b/include/uapi/asm-generic/unistd.h
index a627acc8fb5f..3e1c580a8b1f 100644
--- a/include/uapi/asm-generic/unistd.h
+++ b/include/uapi/asm-generic/unistd.h
@@ -863,9 +863,17 @@ __SYSCALL(__NR_listns, sys_listns)
#define __NR_rseq_slice_yield 471
__SYSCALL(__NR_rseq_slice_yield, sys_rseq_slice_yield)
-#undef __NR_syscalls
-#define __NR_syscalls 472
+#if defined(__SYSCALL_COMPAT) || __BITS_PER_LONG == 32
+#define __NR_mq_timedreceive2_time64 472
+__SYSCALL(__NR_mq_timedreceive2_time64, sys_mq_timedreceive2)
+#endif
+#if defined(__ARCH_WANT_TIME32_SYSCALLS) || __BITS_PER_LONG != 32
+#define __NR_mq_timedreceive2 473
+__SC_3264(__NR_mq_timedreceive2, sys_mq_timedreceive2_time32, sys_mq_timedreceive2)
+#endif
+#undef __NR_syscalls
+#define __NR_syscalls 474
/*
* 32 bit systems traditionally used different
* syscalls for off_t and loff_t arguments, while
diff --git a/ipc/mqueue.c b/ipc/mqueue.c
index 4798b375972b..ff5089495204 100644
--- a/ipc/mqueue.c
+++ b/ipc/mqueue.c
@@ -53,6 +53,7 @@ struct mqueue_fs_context {
#define SEND 0
#define RECV 1
+#define MQ_PEEK 2
#define STATE_NONE 0
#define STATE_READY 1
@@ -63,6 +64,12 @@ struct posix_msg_tree_node {
int priority;
};
+struct mq_timedreceive2_args {
+ size_t msg_len;
+ unsigned int *msg_prio;
+ char *msg_ptr;
+};
+
/*
* Locking:
*
@@ -1230,6 +1237,120 @@ static int do_mq_timedreceive(mqd_t mqdes, char __user *u_msg_ptr,
return ret;
}
+static struct msg_msg *mq_peek_index(struct mqueue_inode_info *info, int index)
+{
+ struct rb_node *node;
+ struct posix_msg_tree_node *leaf;
+ struct msg_msg *msg;
+
+ int count = 0;
+
+ /* Start from highest priority */
+ node = rb_last(&info->msg_tree);
+ while (node) {
+ leaf = rb_entry(node, struct posix_msg_tree_node, rb_node);
+ list_for_each_entry(msg, &leaf->msg_list, m_list) {
+ if (count == index)
+ return msg;
+ count++;
+ }
+
+ node = rb_prev(node);
+ }
+
+ return NULL;
+}
+
+static int do_mq_timedreceive2(mqd_t mqdes,
+ struct mq_timedreceive2_args __user *uargs,
+ unsigned int flags, unsigned long index,
+ struct timespec64 *ts)
+{
+ struct mq_timedreceive2_args args;
+ ssize_t ret;
+ struct msg_msg *msg_ptr, *k_msg_buffer;
+ long k_m_type;
+ size_t k_m_ts;
+ struct inode *inode;
+ struct mqueue_inode_info *info;
+
+ if (copy_from_user(&args, uargs, sizeof(args)))
+ return -EFAULT;
+
+ 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 (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;
+
+ 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);
+
+ k_msg_buffer = alloc_msg(k_m_ts);
+ if (!k_msg_buffer)
+ return -ENOMEM;
+
+ /*
+ * Two spin locks are necessary here. We avoid 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_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)) {
+ 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 ret;
+}
+
SYSCALL_DEFINE5(mq_timedsend, mqd_t, mqdes, const char __user *, u_msg_ptr,
size_t, msg_len, unsigned int, msg_prio,
const struct __kernel_timespec __user *, u_abs_timeout)
@@ -1258,6 +1379,23 @@ 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)
+{
+ struct timespec64 ts, *p = NULL;
+
+ if (u_abs_timeout) {
+ int res = prepare_timeout(u_abs_timeout, &ts);
+
+ if (res)
+ return res;
+ p = &ts;
+ }
+ return do_mq_timedreceive2(mqdes, uargs, flags, index, p);
+}
+
/*
* Notes: the case when user wants us to deregister (with NULL as pointer)
* and he isn't currently owner of notification, will be silently discarded.
@@ -1459,6 +1597,12 @@ struct compat_mq_attr {
compat_long_t __reserved[4]; /* ignored for input, zeroed for output */
};
+struct compat_mq_timedreceive2_args {
+ compat_size_t msg_len;
+ compat_uptr_t msg_prio;
+ compat_uptr_t msg_ptr;
+};
+
static inline int get_compat_mq_attr(struct mq_attr *attr,
const struct compat_mq_attr __user *uattr)
{
@@ -1490,6 +1634,22 @@ static inline int put_compat_mq_attr(const struct mq_attr *attr,
return 0;
}
+static inline int get_compat_mq_args(struct mq_timedreceive2_args *args,
+ struct compat_mq_timedreceive2_args __user *uargs)
+{
+ struct compat_mq_timedreceive2_args v;
+
+ if (copy_from_user(&v, uargs, sizeof(*uargs)))
+ return -EFAULT;
+
+ memset(args, 0, sizeof(*args));
+ args->msg_len = (size_t)compat_ptr(v.msg_len);
+ args->msg_prio = (unsigned int *)compat_ptr(v.msg_prio);
+ args->msg_ptr = (char *)compat_ptr(v.msg_ptr);
+
+ return 0;
+}
+
COMPAT_SYSCALL_DEFINE4(mq_open, const char __user *, u_name,
int, oflag, compat_mode_t, mode,
struct compat_mq_attr __user *, u_attr)
@@ -1583,6 +1743,30 @@ SYSCALL_DEFINE5(mq_timedreceive_time32, mqd_t, mqdes,
}
return do_mq_timedreceive(mqdes, u_msg_ptr, msg_len, u_msg_prio, p);
}
+
+SYSCALL_DEFINE5(mq_timedreceive2_time32, 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)
+{
+ struct mq_timedreceive2_args args, *pargs = NULL;
+
+ pargs = &args;
+
+ if (get_compat_mq_args(pargs, uargs))
+ return -EFAULT;
+
+ struct timespec64 ts, *p = NULL;
+
+ if (u_abs_timeout) {
+ int res = compat_prepare_timeout(u_abs_timeout, &ts);
+
+ if (res)
+ return res;
+ p = &ts;
+ }
+ return do_mq_timedreceive2(mqdes, pargs, flags, index, p);
+}
#endif
static const struct inode_operations mqueue_dir_inode_operations = {
diff --git a/ipc/msg.c b/ipc/msg.c
index 62996b97f0ac..6392b11dd7f7 100644
--- a/ipc/msg.c
+++ b/ipc/msg.c
@@ -1156,7 +1156,7 @@ static long do_msgrcv(int msqid, void __user *buf, size_t bufsz, long msgtyp, in
* not update queue parameters.
*/
if (msgflg & MSG_COPY) {
- msg = copy_msg(msg, copy);
+ msg = copy_msg(msg, copy, msg->m_ts);
goto out_unlock0;
}
diff --git a/ipc/msgutil.c b/ipc/msgutil.c
index e28f0cecb2ec..c5536ce47cc2 100644
--- a/ipc/msgutil.c
+++ b/ipc/msgutil.c
@@ -51,7 +51,7 @@ static int __init init_msg_buckets(void)
}
subsys_initcall(init_msg_buckets);
-static struct msg_msg *alloc_msg(size_t len)
+struct msg_msg *alloc_msg(size_t len)
{
struct msg_msg *msg;
struct msg_msgseg **pseg;
@@ -122,39 +122,33 @@ struct msg_msg *load_msg(const void __user *src, size_t len)
free_msg(msg);
return ERR_PTR(err);
}
-#ifdef CONFIG_CHECKPOINT_RESTORE
-struct msg_msg *copy_msg(struct msg_msg *src, struct msg_msg *dst)
+
+struct msg_msg *copy_msg(struct msg_msg *src, struct msg_msg *dst, size_t len)
{
- struct msg_msgseg *dst_pseg, *src_pseg;
- size_t len = src->m_ts;
- size_t alen;
+ struct msg_msgseg *src_seg, *dst_seg;
+ size_t remaining, chunk;
- if (src->m_ts > dst->m_ts)
+ if (len > src->m_ts)
return ERR_PTR(-EINVAL);
-
- alen = min(len, DATALEN_MSG);
- memcpy(dst + 1, src + 1, alen);
-
- for (dst_pseg = dst->next, src_pseg = src->next;
- src_pseg != NULL;
- dst_pseg = dst_pseg->next, src_pseg = src_pseg->next) {
-
- len -= alen;
- alen = min(len, DATALEN_SEG);
- memcpy(dst_pseg + 1, src_pseg + 1, alen);
+ chunk = min(len, DATALEN_MSG);
+ memcpy(dst + 1, src + 1, chunk);
+ remaining = len - chunk;
+ src_seg = src->next;
+ dst_seg = dst->next;
+ while (remaining > 0 && src_seg && dst_seg) {
+ chunk = min(remaining, DATALEN_SEG);
+ memcpy(dst_seg + 1, src_seg + 1, chunk);
+ remaining -= chunk;
+ src_seg = src_seg->next;
+ dst_seg = dst_seg->next;
}
-
+ if (remaining != 0)
+ return ERR_PTR(-EINVAL);
dst->m_type = src->m_type;
- dst->m_ts = src->m_ts;
-
+ dst->m_ts = src->m_ts;
return dst;
}
-#else
-struct msg_msg *copy_msg(struct msg_msg *src, struct msg_msg *dst)
-{
- return ERR_PTR(-ENOSYS);
-}
-#endif
+
int store_msg(void __user *dest, struct msg_msg *msg, size_t len)
{
size_t alen;
diff --git a/ipc/util.h b/ipc/util.h
index a55d6cebe6d3..374abeee79b3 100644
--- a/ipc/util.h
+++ b/ipc/util.h
@@ -197,8 +197,9 @@ int ipc_parse_version(int *cmd);
extern void free_msg(struct msg_msg *msg);
extern struct msg_msg *load_msg(const void __user *src, size_t len);
-extern struct msg_msg *copy_msg(struct msg_msg *src, struct msg_msg *dst);
+extern struct msg_msg *copy_msg(struct msg_msg *src, struct msg_msg *dst, size_t len);
extern int store_msg(void __user *dest, struct msg_msg *msg, size_t len);
+extern struct msg_msg *alloc_msg(size_t len);
static inline int ipc_checkid(struct kern_ipc_perm *ipcp, int id)
{
diff --git a/kernel/sys_ni.c b/kernel/sys_ni.c
index add3032da16f..e6adc8c4220e 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);
+COND_SYSCALL(mq_timedreceive2);
+COND_SYSCALL(mq_timedreceive2_time32);
COND_SYSCALL(uretprobe);
COND_SYSCALL(uprobe);
diff --git a/scripts/checksyscalls.sh b/scripts/checksyscalls.sh
index 1e5d2eeb726d..605a65cfa144 100755
--- a/scripts/checksyscalls.sh
+++ b/scripts/checksyscalls.sh
@@ -104,6 +104,7 @@ cat << EOF
#define __IGNORE_recvmmsg_time64
#define __IGNORE_mq_timedsend_time64
#define __IGNORE_mq_timedreceive_time64
+#define __IGNORE_mq_timedreceive2_time64
#define __IGNORE_semtimedop_time64
#define __IGNORE_rt_sigtimedwait_time64
#define __IGNORE_futex_time64
diff --git a/scripts/syscall.tbl b/scripts/syscall.tbl
index 7a42b32b6577..57b711f93756 100644
--- a/scripts/syscall.tbl
+++ b/scripts/syscall.tbl
@@ -412,3 +412,6 @@
469 common file_setattr sys_file_setattr
470 common listns sys_listns
471 common rseq_slice_yield sys_rseq_slice_yield
+472 32 mq_timedreceive2_time64 sys_mq_timedreceive2
+473 64 mq_timedreceive2 sys_mq_timedreceive2
+473 time32 mq_timedreceive2 sys_mq_timedreceive2_time32
diff --git a/tools/testing/selftests/ipc/.gitignore b/tools/testing/selftests/ipc/.gitignore
index 9ed280e4c704..fe609d98ecb2 100644
--- a/tools/testing/selftests/ipc/.gitignore
+++ b/tools/testing/selftests/ipc/.gitignore
@@ -1,3 +1,4 @@
# SPDX-License-Identifier: GPL-2.0-only
msgque_test
msgque
+mq_peek
diff --git a/tools/testing/selftests/ipc/Makefile b/tools/testing/selftests/ipc/Makefile
index 50e9c299fc4a..fa9d4f500783 100644
--- a/tools/testing/selftests/ipc/Makefile
+++ b/tools/testing/selftests/ipc/Makefile
@@ -1,6 +1,8 @@
# SPDX-License-Identifier: GPL-2.0
uname_M := $(shell uname -m 2>/dev/null || echo not)
ARCH ?= $(shell echo $(uname_M) | sed -e s/i.86/i386/)
+CC = gcc
+CFLAGS = -static -O2
ifeq ($(ARCH),i386)
ARCH := x86
CFLAGS := -DCONFIG_X86_32 -D__i386__
@@ -12,7 +14,8 @@ endif
CFLAGS += $(KHDR_INCLUDES)
-TEST_GEN_PROGS := msgque
+TEST_GEN_PROGS := msgque mq_peek
+LDLIBS += -lrt -lpthread
include ../lib.mk
diff --git a/tools/testing/selftests/ipc/mq_peek.c b/tools/testing/selftests/ipc/mq_peek.c
new file mode 100644
index 000000000000..5803b1aa8767
--- /dev/null
+++ b/tools/testing/selftests/ipc/mq_peek.c
@@ -0,0 +1,779 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * mq_peek.c - Selftest for mq_timedreceive2(MQ_PEEK)
+ *
+ * Tests the peek-without-dequeue extension to POSIX message queues:
+ * - Correct priority-rank indexed access (index 0 = highest priority)
+ * - FIFO ordering within same priority level
+ * - Non-destructive semantics (queue is never modified on peek)
+ * - All error paths (-EBADF, -EMSGSIZE, -EFAULT, -EAGAIN, -ENOENT)
+ * - Large (multi-segment) message payload copy
+ * - Concurrent peek + receive / peek + send races
+ */
+
+#include "linux/array_size.h"
+#define _GNU_SOURCE
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <stdint.h>
+#include <unistd.h>
+#include <errno.h>
+#include <pthread.h>
+#include <sys/syscall.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <mqueue.h>
+#include <time.h>
+
+#include "kselftest.h"
+
+#ifndef __NR_mq_timedreceive2
+#if defined(__x86_64__)
+#define __NR_mq_timedreceive2 548
+#elif defined(__aarch64__)
+#define __NR_mq_timedreceive2 473
+#elif defined(__riscv)
+#define __NR_mq_timedreceive2 473
+#else
+#define __NR_mq_timedreceive2 548
+#endif
+#endif
+
+#define MQ_PEEK 2U
+
+struct mq_timedreceive2_args {
+ size_t msg_len;
+ unsigned int *msg_prio;
+ char *msg_ptr;
+};
+
+static long mq_timedreceive2(mqd_t mqdes, struct mq_timedreceive2_args *uargs,
+ unsigned int flags, unsigned long index,
+ const struct timespec *timeout)
+{
+ return syscall(__NR_mq_timedreceive2, (long)mqdes, uargs, (long)flags,
+ index, timeout);
+}
+
+#define MQ_NAME_PREFIX "/mq_peek_test_"
+#define MAX_MSG_SIZE 128
+#define LARGE_MSG_SIZE 4064
+#define MQ_MAXMSG 16
+
+#define PRIO_HIGH 9
+#define PRIO_MED 5
+#define PRIO_LOW 1
+
+static mqd_t open_queue(const char *suffix, long msgsize)
+{
+ char name[64];
+ struct mq_attr attr = {
+ .mq_flags = 0,
+ .mq_maxmsg = MQ_MAXMSG,
+ .mq_msgsize = msgsize,
+ .mq_curmsgs = 0,
+ };
+ mqd_t mqd;
+
+ snprintf(name, sizeof(name), "%s%s", MQ_NAME_PREFIX, suffix);
+ mq_unlink(name);
+
+ mqd = mq_open(name, O_NONBLOCK | O_RDWR | O_CREAT | O_EXCL, 0600, &attr);
+ if (mqd == (mqd_t)-1) {
+ ksft_test_result_fail("mq_open(%s): %m\n", name);
+ ksft_exit_fail();
+ }
+
+ mq_unlink(name);
+ return mqd;
+}
+
+static void send_msg(mqd_t mqd, unsigned int prio, const char *text, size_t len)
+{
+ if (mq_send(mqd, text, len, prio) != 0) {
+ ksft_test_result_fail("mq_send(prio=%u): %m\n", prio);
+ ksft_exit_fail();
+ }
+}
+
+static long peek(mqd_t mqd, unsigned long index, char *buf, size_t bufsz,
+ unsigned int *prio)
+{
+ struct mq_timedreceive2_args args = {
+ .msg_len = bufsz,
+ .msg_prio = prio,
+ .msg_ptr = buf,
+ };
+ return mq_timedreceive2(mqd, &args, MQ_PEEK, index, NULL);
+}
+
+static long queue_depth(mqd_t mqd)
+{
+ struct mq_attr attr;
+
+ if (mq_getattr(mqd, &attr) != 0)
+ return -1;
+ return attr.mq_curmsgs;
+}
+
+static void test_peek_empty_queue(void)
+{
+ mqd_t mqd = open_queue("empty", MAX_MSG_SIZE);
+ char buf[MAX_MSG_SIZE];
+ unsigned int prio;
+ long ret;
+
+ ret = peek(mqd, 0, buf, sizeof(buf), &prio);
+ if (ret == -1 && errno == EAGAIN)
+ ksft_test_result_pass("peek on empty queue [EAGAIN]\n");
+ else
+ ksft_test_result_fail("peek on empty queue: expected EAGAIN, got ret=%ld errno=%d (%m)\n",
+ ret, errno);
+
+ mq_close(mqd);
+}
+
+static void test_peek_invalid_fd(void)
+{
+ char buf[MAX_MSG_SIZE];
+ unsigned int prio;
+ long ret;
+
+ ret = peek((mqd_t)-1, 0, buf, sizeof(buf), &prio);
+ if (ret == -1 && errno == EBADF)
+ ksft_test_result_pass("peek invalid fd [ EBADF]\n");
+ else
+ ksft_test_result_fail("peek invalid fd: expected EBADF, got ret=%ld errno=%d\n",
+ ret, errno);
+}
+
+static void test_peek_non_mqueue_fd(void)
+{
+ int pipefd[2];
+ char buf[MAX_MSG_SIZE];
+ unsigned int prio;
+ long ret;
+
+ if (pipe(pipefd) != 0) {
+ ksft_test_result_skip("pipe() failed, skipping non-mqueue-fd test\n");
+ return;
+ }
+
+ ret = peek((mqd_t)pipefd[0], 0, buf, sizeof(buf), &prio);
+ if (ret == -1 && errno == EBADF)
+ ksft_test_result_pass("peek on pipe fd [EBADF]\n");
+ else
+ ksft_test_result_fail("peek non-mqueue fd: expected EBADF, got ret=%ld errno=%d\n",
+ ret, errno);
+
+ close(pipefd[0]);
+ close(pipefd[1]);
+}
+
+static void test_peek_writeonly_fd(void)
+{
+ char name[] = "/ksft_mq_peek_wo";
+ struct mq_attr attr = { .mq_maxmsg = 4, .mq_msgsize = MAX_MSG_SIZE };
+ mqd_t rw, wo;
+ char buf[MAX_MSG_SIZE];
+ unsigned int prio;
+ long ret;
+
+ mq_unlink(name);
+ rw = mq_open(name, O_RDWR | O_CREAT, 0600, &attr);
+ if (rw == (mqd_t)-1) {
+ ksft_test_result_skip("mq_open RW failed: %m\n");
+ return;
+ }
+ wo = mq_open(name, O_WRONLY);
+ mq_unlink(name);
+
+ if (wo == (mqd_t)-1) {
+ ksft_test_result_skip("mq_open WO failed: %m\n");
+ mq_close(rw);
+ return;
+ }
+
+ send_msg(rw, PRIO_HIGH, "x", 1);
+
+ ret = peek(wo, 0, buf, sizeof(buf), &prio);
+ if (ret == -1 && errno == EBADF)
+ ksft_test_result_pass("peek on O_WRONLY fd [EBADF]\n");
+ else
+ ksft_test_result_fail("peek WO fd: expected EBADF, got ret=%ld errno=%d\n",
+ ret, errno);
+
+ mq_close(wo);
+ mq_close(rw);
+}
+
+static void test_peek_buffer_too_small(void)
+{
+ mqd_t mqd = open_queue("small", MAX_MSG_SIZE);
+ char tiny[1]; /* deliberately too small */
+ unsigned int prio;
+ struct mq_timedreceive2_args args = {
+ .msg_len = sizeof(tiny),
+ .msg_prio = &prio,
+ .msg_ptr = tiny,
+ };
+ long ret;
+
+ send_msg(mqd, PRIO_HIGH, "hello", 5);
+
+ ret = mq_timedreceive2(mqd, &args, MQ_PEEK, 0, NULL);
+ if (ret == -1 && errno == EMSGSIZE)
+ ksft_test_result_pass("peek with small buf [EMSGSIZE]\n");
+ else
+ ksft_test_result_fail("peek small buf: expected EMSGSIZE, got ret=%ld errno=%d\n",
+ ret, errno);
+
+ mq_close(mqd);
+}
+
+static void test_peek_bad_msg_ptr(void)
+{
+ mqd_t mqd = open_queue("bad_ptr", MAX_MSG_SIZE);
+ unsigned int prio;
+
+ struct mq_timedreceive2_args args = {
+ .msg_len = MAX_MSG_SIZE,
+ .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);
+ if (ret == -1 && errno == EFAULT)
+ ksft_test_result_pass("peek bad msg_ptr [EFAULT]\n");
+ else
+ ksft_test_result_fail("peek bad msg_ptr: expected EFAULT, got ret=%ld errno=%d\n",
+ ret, errno);
+
+ mq_close(mqd);
+}
+
+
+static void test_peek_index_out_of_range(void)
+{
+ mqd_t mqd = open_queue("oob", MAX_MSG_SIZE);
+ char buf[MAX_MSG_SIZE];
+ unsigned int prio;
+ long ret;
+
+ send_msg(mqd, PRIO_MED, "one", 3);
+ ret = peek(mqd, 1, buf, sizeof(buf), &prio);
+ if (ret == -1 && errno == ENOENT)
+ ksft_test_result_pass("peek OOB index [ENOENT]\n");
+ else
+ ksft_test_result_fail("peek OOB: expected ENOENT, got ret=%ld errno=%d\n",
+ ret, errno);
+
+ mq_close(mqd);
+}
+
+static void test_peek_basic_data(void)
+{
+ mqd_t mqd = open_queue("basic", MAX_MSG_SIZE);
+ const char *payload = "peek-test-payload";
+ char buf[MAX_MSG_SIZE];
+ unsigned int prio = 0;
+ long ret;
+
+ send_msg(mqd, PRIO_HIGH, payload, strlen(payload));
+
+ memset(buf, 0, sizeof(buf));
+ ret = peek(mqd, 0, buf, sizeof(buf), &prio);
+
+ if (ret < 0) {
+ ksft_test_result_fail("basic peek failed: ret=%ld errno=%d (%m)\n", ret, errno);
+ goto out;
+ }
+ if ((size_t)ret != strlen(payload)) {
+ ksft_test_result_fail("basic peek: wrong size %ld (expected %zu)\n", ret, strlen(payload));
+ goto out;
+ }
+ if (memcmp(buf, payload, strlen(payload)) != 0) {
+ ksft_test_result_fail("basic peek: payload mismatch\n");
+ goto out;
+ }
+ if (prio != PRIO_HIGH) {
+ ksft_test_result_fail("basic peek: wrong prio %u (expected %d)\n", prio, PRIO_HIGH);
+ goto out;
+ }
+ ksft_test_result_pass("basic peek: correct data and priority\n");
+out:
+ mq_close(mqd);
+}
+
+
+static void test_peek_nondestructive(void)
+{
+ mqd_t mqd = open_queue("nodestr", MAX_MSG_SIZE);
+ char buf[MAX_MSG_SIZE];
+ unsigned int prio;
+ int i;
+
+ send_msg(mqd, PRIO_HIGH, "A", 1);
+ send_msg(mqd, PRIO_MED, "B", 1);
+ send_msg(mqd, PRIO_LOW, "C", 1);
+
+ if (queue_depth(mqd) != 3) {
+ ksft_test_result_fail("initial depth != 3\n");
+ mq_close(mqd);
+ return;
+ }
+
+ for (i = 0; i < 10; i++) {
+ peek(mqd, 0, buf, sizeof(buf), &prio);
+ peek(mqd, 1, buf, sizeof(buf), &prio);
+ peek(mqd, 2, buf, sizeof(buf), &prio);
+ }
+
+ if (queue_depth(mqd) == 3)
+ ksft_test_result_pass("peek is non-destructive (depth stays 3)\n");
+ else
+ ksft_test_result_fail("peek modified queue: depth=%ld (expected 3)\n", queue_depth(mqd));
+
+ mq_close(mqd);
+}
+
+static void test_peek_priority_order(void)
+{
+ mqd_t mqd = open_queue("prio_order", MAX_MSG_SIZE);
+ char buf[MAX_MSG_SIZE];
+ unsigned int prio;
+ long ret;
+ int pass = 1;
+
+ send_msg(mqd, PRIO_LOW, "low", 3);
+ send_msg(mqd, PRIO_HIGH, "high", 4);
+ send_msg(mqd, PRIO_MED, "med", 3);
+
+ /* index 0 must return highest priority */
+ ret = peek(mqd, 0, buf, sizeof(buf), &prio);
+ if (ret < 0 || prio != PRIO_HIGH) {
+ ksft_test_result_fail(
+ "prio_order index0: prio=%u ret=%ld errno=%d\n", prio,
+ ret, errno);
+ pass = 0;
+ }
+ if (pass && memcmp(buf, "high", 4) != 0) {
+ ksft_test_result_fail("prio_order index0: wrong payload\n");
+ pass = 0;
+ }
+
+ /* index 1 must return medium priority */
+ ret = peek(mqd, 1, buf, sizeof(buf), &prio);
+ if (pass && (ret < 0 || prio != PRIO_MED)) {
+ ksft_test_result_fail("prio_order index1: prio=%u ret=%ld\n",
+ prio, ret);
+ pass = 0;
+ }
+ if (pass && memcmp(buf, "med", 3) != 0) {
+ ksft_test_result_fail("prio_order index1: wrong payload\n");
+ pass = 0;
+ }
+
+ /* index 2 must return lowest priority */
+ ret = peek(mqd, 2, buf, sizeof(buf), &prio);
+ if (pass && (ret < 0 || prio != PRIO_LOW)) {
+ ksft_test_result_fail("prio_order index2: prio=%u ret=%ld\n",
+ prio, ret);
+ pass = 0;
+ }
+ if (pass && memcmp(buf, "low", 3) != 0) {
+ ksft_test_result_fail("prio_order index2: wrong payload\n");
+ pass = 0;
+ }
+
+ if (pass)
+ ksft_test_result_pass(
+ "priority ordering: index0=HIGH, index1=MED, index2=LOW\n");
+
+ mq_close(mqd);
+}
+
+
+static void test_peek_fifo_within_priority(void)
+{
+ mqd_t mqd = open_queue("fifo", MAX_MSG_SIZE);
+ char buf[MAX_MSG_SIZE];
+ unsigned int prio;
+ long ret;
+ int pass = 1;
+
+ send_msg(mqd, PRIO_HIGH, "first", 5);
+ send_msg(mqd, PRIO_HIGH, "second", 6);
+ send_msg(mqd, PRIO_HIGH, "third", 5);
+
+ memset(buf, 0, sizeof(buf));
+ ret = peek(mqd, 0, buf, sizeof(buf), &prio);
+ if (ret < 0 || memcmp(buf, "first", 5) != 0) {
+ ksft_test_result_fail(
+ "FIFO peek[0]: expected 'first', got '%.*s' ret=%ld\n",
+ (int)ret, buf, ret);
+ pass = 0;
+ }
+
+ if (pass) {
+ char rbuf[MAX_MSG_SIZE];
+ unsigned int rprio;
+
+ ssize_t r = mq_receive(mqd, rbuf, sizeof(rbuf), &rprio);
+ if (r < 0 || memcmp(rbuf, "first", 5) != 0) {
+ ksft_test_result_fail("mq_receive 'first' failed\n");
+ pass = 0;
+ }
+ }
+
+ if (pass) {
+ memset(buf, 0, sizeof(buf));
+ ret = peek(mqd, 0, buf, sizeof(buf), &prio);
+ if (ret < 0 || memcmp(buf, "second", 6) != 0) {
+ ksft_test_result_fail("FIFO peek after receive: expected 'second', got '%.*s'\n",
+ (int)ret, buf);
+ pass = 0;
+ }
+ }
+
+ if (pass)
+ ksft_test_result_pass("FIFO within same priority is correct\n");
+
+ mq_close(mqd);
+}
+
+static void test_peek_all_indices(void)
+{
+ const unsigned int prios[] = { 2, 7, 4, 9, 1, 6 };
+ const int N = ARRAY_SIZE(prios);
+ mqd_t mqd = open_queue("all_idx", MAX_MSG_SIZE);
+ char buf[MAX_MSG_SIZE];
+ char expected_payload[MAX_MSG_SIZE];
+ unsigned int prio;
+ long ret;
+ int i, pass = 1;
+ unsigned int sorted[6];
+
+ for (i = 0; i < N; i++) {
+ snprintf(expected_payload, sizeof(expected_payload),
+ "msg_prio_%u", prios[i]);
+ send_msg(mqd, prios[i], expected_payload,
+ strlen(expected_payload));
+ sorted[i] = prios[i];
+ }
+
+ for (i = 0; i < N - 1; i++) {
+ int j;
+
+ for (j = i + 1; j < N; j++) {
+ if (sorted[j] > sorted[i]) {
+ unsigned int tmp = sorted[i];
+ sorted[i] = sorted[j];
+ sorted[j] = tmp;
+ }
+ }
+ }
+
+ for (i = 0; i < N && pass; i++) {
+ memset(buf, 0, sizeof(buf));
+ ret = peek(mqd, (unsigned long)i, buf, sizeof(buf), &prio);
+ if (ret < 0) {
+ ksft_test_result_fail("all_indices peek[%d] failed: ret=%ld errno=%d\n",
+ i, ret, errno);
+ pass = 0;
+ break;
+ }
+ if (prio != sorted[i]) {
+ ksft_test_result_fail("all_indices peek[%d]: prio=%u expected=%u\n",
+ i, prio, sorted[i]);
+ pass = 0;
+ }
+
+ snprintf(expected_payload, sizeof(expected_payload),
+ "msg_prio_%u", sorted[i]);
+ if (memcmp(buf, expected_payload, strlen(expected_payload))) {
+ ksft_test_result_fail("all_indices peek[%d]: payload mismatch\n", i);
+ pass = 0;
+ }
+ }
+
+ if (pass && queue_depth(mqd) != N) {
+ ksft_test_result_fail("all_indices: depth=%ld expected=%d after peek\n",
+ queue_depth(mqd), N);
+ pass = 0;
+ }
+
+ if (pass) {
+ ret = peek(mqd, (unsigned long)N, buf, sizeof(buf), &prio);
+ if (!(ret == -1 && errno == ENOENT)) {
+ ksft_test_result_fail("all_indices OOB[%d]: expected ENOENT, got ret=%ld errno=%d\n",
+ N, ret, errno);
+ pass = 0;
+ }
+ }
+
+ if (pass)
+ ksft_test_result_pass("all-indices: correct prio order + OOB ENOENT\n");
+
+ mq_close(mqd);
+}
+
+static void test_peek_large_message(void)
+{
+ mqd_t mqd = open_queue("large", LARGE_MSG_SIZE);
+ char *send_buf, *recv_buf;
+ unsigned int prio = 0;
+ long ret;
+ int pass = 1;
+
+ send_buf = malloc(LARGE_MSG_SIZE);
+ recv_buf = calloc(1, LARGE_MSG_SIZE);
+ if (!send_buf || !recv_buf) {
+ ksft_test_result_skip("OOM allocating large message buffers\n");
+ goto out;
+ }
+
+ for (int i = 0; i < LARGE_MSG_SIZE; i++)
+ send_buf[i] = (char)(i & 0xFF);
+
+ send_msg(mqd, PRIO_HIGH, send_buf, LARGE_MSG_SIZE);
+
+ ret = peek(mqd, 0, recv_buf, LARGE_MSG_SIZE, &prio);
+ if (ret != LARGE_MSG_SIZE) {
+ ksft_test_result_fail("large msg peek: ret=%ld expected=%d\n",
+ ret, LARGE_MSG_SIZE);
+ pass = 0;
+ }
+ if (pass && memcmp(send_buf, recv_buf, LARGE_MSG_SIZE) != 0) {
+ ksft_test_result_fail("large msg peek: payload mismatch\n");
+ pass = 0;
+ }
+ if (pass && prio != PRIO_HIGH) {
+ ksft_test_result_fail("large msg peek: prio=%u expected=%d\n", prio, PRIO_HIGH);
+ pass = 0;
+ }
+ if (pass && queue_depth(mqd) != 1) {
+ ksft_test_result_fail("large msg peek: queue modified (depth=%ld)\n",queue_depth(mqd));
+ pass = 0;
+ }
+ if (pass)
+ ksft_test_result_pass("large (%d B) multi-segment peek: correct\n",LARGE_MSG_SIZE);
+out:
+ free(send_buf);
+ free(recv_buf);
+ mq_close(mqd);
+}
+
+static void test_no_peek_flag_is_receive(void)
+{
+ mqd_t mqd = open_queue("nopeek", MAX_MSG_SIZE);
+ char buf[MAX_MSG_SIZE];
+ unsigned int prio = 0;
+ struct mq_timedreceive2_args args = {
+ .msg_len = sizeof(buf),
+ .msg_prio = &prio,
+ .msg_ptr = buf,
+ };
+ long ret;
+
+ send_msg(mqd, PRIO_HIGH, "consume-me", 10);
+
+ ret = mq_timedreceive2(mqd, &args, 0, 0, NULL);
+ if (ret < 0) {
+ ksft_test_result_fail("no-peek receive failed: ret=%ld errno=%d\n", ret, errno);
+ mq_close(mqd);
+ return;
+ }
+ if (queue_depth(mqd) != 0)
+ ksft_test_result_fail("no-peek: queue still has messages (depth=%ld)\n", queue_depth(mqd));
+ else
+ ksft_test_result_pass("without MQ_PEEK the message is consumed normally\n");
+
+ mq_close(mqd);
+}
+
+struct race_ctx {
+ mqd_t mqd;
+ int errors;
+};
+
+static void *receiver_thread(void *arg)
+{
+ struct race_ctx *ctx = arg;
+ char buf[MAX_MSG_SIZE];
+ unsigned int prio;
+ ssize_t r;
+
+ while ((r = mq_receive(ctx->mqd, buf, sizeof(buf), &prio)) > 0)
+ ;
+
+ return NULL;
+}
+
+static void test_peek_concurrent_receive(void)
+{
+ struct race_ctx ctx;
+ pthread_t tid;
+ char buf[MAX_MSG_SIZE];
+ unsigned int prio;
+ long ret;
+ int i;
+
+ ctx.mqd = open_queue("concurrent", MAX_MSG_SIZE);
+ ctx.errors = 0;
+
+ for (i = 0; i < MQ_MAXMSG; i++) {
+ char payload[32];
+
+ snprintf(payload, sizeof(payload), "msg%d", i);
+ send_msg(ctx.mqd, (unsigned int)(i % 5) + 1, payload,
+ strlen(payload));
+ }
+
+ if (pthread_create(&tid, NULL, receiver_thread, &ctx) != 0) {
+ ksft_test_result_skip("pthread_create failed\n");
+ mq_close(ctx.mqd);
+ return;
+ }
+
+ /*
+ * Peek repeatedly.The queue is being drained concurrently.
+ */
+ for (i = 0; i < 200; i++) {
+ ret = peek(ctx.mqd, (unsigned long)(i % 4), buf, sizeof(buf),
+ &prio);
+ if (ret < 0 && errno != EAGAIN && errno != ENOENT) {
+ ctx.errors++;
+ }
+ }
+
+ pthread_join(tid, NULL);
+
+ if (ctx.errors == 0)
+ ksft_test_result_pass("concurrent peek+receive: no unexpected errors\n");
+ else
+ ksft_test_result_fail("concurrent peek+receive: %d unexpected errors\n", ctx.errors);
+
+ mq_close(ctx.mqd);
+}
+
+static void test_peek_null_prio_ptr(void)
+{
+ mqd_t mqd = open_queue("null_prio", 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_MED, "no-prio-needed", 14);
+
+ ret = mq_timedreceive2(mqd, &args, MQ_PEEK, 0, NULL);
+ if (ret >= 0)
+ ksft_test_result_pass("peek with NULL msg_prio ptr: OK\n");
+ else
+ ksft_test_result_fail("peek NULL msg_prio: ret=%ld errno=%d (%m)\n", ret, errno);
+
+ mq_close(mqd);
+}
+
+static void test_peek_priority_matches_receive(void)
+{
+ mqd_t mqd = open_queue("prio_match", MAX_MSG_SIZE);
+ char peek_buf[MAX_MSG_SIZE], recv_buf[MAX_MSG_SIZE];
+ unsigned int peek_prio = 0, recv_prio = 0;
+ long peek_ret;
+ ssize_t recv_ret;
+ int pass = 1;
+
+ send_msg(mqd, PRIO_MED, "consistent-prio", 15);
+
+ peek_ret = peek(mqd, 0, peek_buf, sizeof(peek_buf), &peek_prio);
+ if (peek_ret < 0) {
+ ksft_test_result_fail("peek failed: %m\n");
+ mq_close(mqd);
+ return;
+ }
+
+ recv_ret = mq_receive(mqd, recv_buf, sizeof(recv_buf), &recv_prio);
+ if (recv_ret < 0) {
+ ksft_test_result_fail("mq_receive failed: %m\n");
+ mq_close(mqd);
+ return;
+ }
+
+ if (peek_prio != recv_prio) {
+ ksft_test_result_fail("prio mismatch: peek=%u receive=%u\n",
+ peek_prio, recv_prio);
+ pass = 0;
+ }
+ if (pass && peek_ret != recv_ret) {
+ ksft_test_result_fail("size mismatch: peek=%ld receive=%zd\n",
+ peek_ret, recv_ret);
+ pass = 0;
+ }
+ if (pass && memcmp(peek_buf, recv_buf, (size_t)recv_ret) != 0) {
+ ksft_test_result_fail("payload mismatch between peek and receive\n");
+ pass = 0;
+ }
+ if (pass)
+ ksft_test_result_pass("peeked priority/payload matches mq_receive output\n");
+
+ mq_close(mqd);
+}
+
+static const struct {
+ const char *name;
+ void (*fn)(void);
+} tests[] = {
+ { "empty queue → EAGAIN", test_peek_empty_queue },
+ { "invalid fd → EBADF", test_peek_invalid_fd },
+ { "non-mqueue fd → EBADF", test_peek_non_mqueue_fd },
+ { "O_WRONLY fd → EBADF", test_peek_writeonly_fd },
+ { "buffer too small → EMSGSIZE", test_peek_buffer_too_small },
+ { "bad msg_ptr → EFAULT", test_peek_bad_msg_ptr },
+ { "OOB index → ENOENT", test_peek_index_out_of_range },
+ { "basic data+prio correctness", test_peek_basic_data },
+ { "non-destructive semantics", test_peek_nondestructive },
+ { "priority ordering across indices", test_peek_priority_order },
+ { "FIFO within same priority", test_peek_fifo_within_priority },
+ { "all distinct priority indices", test_peek_all_indices },
+ { "large multi-segment message", test_peek_large_message },
+ { "no MQ_PEEK → normal receive", test_no_peek_flag_is_receive },
+ { "concurrent peek + receive", test_peek_concurrent_receive },
+ { "NULL msg_prio ptr", test_peek_null_prio_ptr },
+ { "peeked prio matches mq_receive",
+ test_peek_priority_matches_receive },
+};
+
+int main(void)
+{
+ unsigned int i;
+ long sc_ret;
+
+ ksft_print_header();
+ ksft_set_plan(ARRAY_SIZE(tests));
+
+ {
+ struct mq_timedreceive2_args probe_args = { 0 };
+ sc_ret = mq_timedreceive2((mqd_t)-1, &probe_args, 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",
+ __NR_mq_timedreceive2);
+ }
+
+ for (i = 0; i < ARRAY_SIZE(tests); i++) {
+ ksft_print_msg("--- [%02u] %s ---\n", i + 1, tests[i].name);
+ tests[i].fn();
+ }
+
+ return ksft_get_fail_cnt() ? 1 : 0;
+}
--
2.43.0
^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH] [PATCH V5] mqueue: introduce new do_mq_timedreceive2() [ mq_peek syscall] for non-destructive receive and inspection,fix minor issue,prepared doc.
2026-03-06 7:50 [PATCH] [PATCH V5] mqueue: introduce new do_mq_timedreceive2() [ mq_peek syscall] for non-destructive receive and inspection,fix minor issue,prepared doc Mathura_Kumar
@ 2026-03-06 9:38 ` Geert Uytterhoeven
2026-03-06 10:30 ` Arnd Bergmann
2026-03-06 21:56 ` Randy Dunlap
` (2 subsequent siblings)
3 siblings, 1 reply; 6+ messages in thread
From: Geert Uytterhoeven @ 2026-03-06 9:38 UTC (permalink / raw)
To: Mathura_Kumar; +Cc: brauner, linux-arch, linux-kernel, viro, Arnd Bergmann
Hi Mathura,
CC Arnd
Thanks for your patch!
On Fri, 6 Mar 2026 at 08:50, Mathura_Kumar <academic1mathura@gmail.com> wrote:
> Change from v3 to v5:
>
> -Styles issue fixed
> -Doc was prepared/modified in Documentation/userspace-api/ipc.rst,
> -Corrected file references
> -I am squashing previous four commit in this latest one for convenience in applying
> and merging
> -risc-v related macro handling error fixed
> -some struct definition removed from ipc/mqueue.c for better compiler control flow
>
> Required Attention:-As per mail from bot
What does this mean?
>
> kernel test robot noticed the following build errors:
> [auto build test ERROR on shuah-kselftest/next]
> [also build test ERROR on shuah-kselftest/fixes next-20260304]
> [cannot apply to arnd-asm-generic/master linus/master tip/x86/asm v6.16-rc1]
>
> Short Description:
> POSIX message queues currently lack a mechanism to read
> a message without removing it from the queue. This is a
> long-standing limitation,when we require inspection of queue state
> without altering it.
>
> Modifying existing mq_receive() semantics via additional
> flags was considered. However, altering behavior of an
> existing syscall risks breaking backward compatibility
> for applications relying on current semantics. Since
> mq_receive() guarantees message removal, changing this
> contract is not safe.
>
> To preserve ABI stability, this patch introduces a new
> system call that performs a non-destructive receive
> operation (peek). The existing behavior remains unchanged.
>
> Design considerations:
>
> Two approaches for copying message data to userspace
> were evaluated:
>
> 1) Refcount-based message lifecycle handling
> - This can help us Avoids intermediate temp kernel copy
> - Extends message lifetime
> -But this may increase writer starvation under heavy load and
> add unnecessary complication on priority management and
> delay more time to free space in inode due refcount may prevent
>
> 2) Temporary kernel buffer copy
> - Copies message into a bounded kernel buffer
> - Reduces time message remains locked
> - Improves fairness under write-heavy workloads
> - Simpler lifetime management
>
> My implementation adopts the temporary buffer approach
> to minimize starvation and reduce locking complexity.
> The design allows future transition if refcounting is
> deemed preferable.
>
> Architecture support: ALL
>
> Testing:
> - 15+ functional test cases
> - Multi-threaded producer/consumer scenarios
> - concurrent pop and peek
> - Edge cases: empty queue, FIFO
> invalid flags, signal interruption etc.
>
> Use Case:
>
> 1) Observability in distributed systems
> e.g Monitoring tools can inspect queue contents without interfering with
> normal processing
>
> 2) Check pointing system can have a look into queue without consuming and can store messages
>
> 3) Resource-aware processing
>
> 4) Selective Consumption for Specialized Workers
>
> Signed-off-by: Mathura_Kumar <academic1mathura@gmail.com>
> --- a/Documentation/userspace-api/index.rst
> +++ b/Documentation/userspace-api/index.rst
> --- /dev/null
> +++ b/Documentation/userspace-api/ipc.rst
> @@ -0,0 +1,273 @@
> +1) Overview
> +-----------
> +
> +POSIX message queues on Linux provide mq_receive() and mq_timedreceive()
> +for consuming messages from a queue. Both interfaces require the caller
> +to pass the message buffer, length, and priority pointer as individual
> +arguments to the system call. This imposes a fixed calling convention
> +that cannot be extended without breaking the ABI.
> +
> +mq_timedreceive2() introduces a new system call entry point that accepts
> +message buffer parameters via a struct argument rather than as individual
> +syscall arguments. This frees the remaining syscall argument slots for
> +new functionality flags and a message index, enabling non-destructive
> +peek and indexed access semantics that are not possible with the
> +original interface.
> +
> +Two variants are provided:
> +
> + mq_timedreceive2() - primary variant, 64-bit time (Y2038-safe)
> + mq_timedreceive2_time32() - 32-bit time variant for legacy and compat
What is the rationale behind adding a new syscall that is not Y2038-safe?
> --- a/arch/arm/tools/syscall.tbl
> +++ b/arch/arm/tools/syscall.tbl
> @@ -486,3 +486,5 @@
> 469 common file_setattr sys_file_setattr
> 470 common listns sys_listns
> 471 common rseq_slice_yield sys_rseq_slice_yield
> +472 common mq_timedreceive2 sys_mq_timedreceive2_time32
> +473 common mq_timedreceive2_time64 sys_mq_timedreceive2
> diff --git a/arch/arm64/tools/syscall_32.tbl b/arch/arm64/tools/syscall_32.tbl
> index 62d93d88e0fe..02603704231e 100644
> --- a/arch/arm64/tools/syscall_32.tbl
> +++ b/arch/arm64/tools/syscall_32.tbl
> @@ -483,3 +483,5 @@
> 469 common file_setattr sys_file_setattr
> 470 common listns sys_listns
> 471 common rseq_slice_yield sys_rseq_slice_yield
> +472 common mq_timedreceive2_time64 sys_mq_timedreceive2
> +473 common mq_timedreceive2 sys_mq_timedreceive2_time32
> diff --git a/arch/m68k/kernel/syscalls/syscall.tbl b/arch/m68k/kernel/syscalls/syscall.tbl
> index 248934257101..b67f23edda11 100644
> --- a/arch/m68k/kernel/syscalls/syscall.tbl
> +++ b/arch/m68k/kernel/syscalls/syscall.tbl
> @@ -471,3 +471,6 @@
> 469 common file_setattr sys_file_setattr
> 470 common listns sys_listns
> 471 common rseq_slice_yield sys_rseq_slice_yield
> +472 common mq_timedreceive2 sys_mq_timedreceive2_time32
> +473 common mq_timedreceive2_time64 sys_mq_timedreceive2
> +
> diff --git a/arch/microblaze/kernel/syscalls/syscall.tbl b/arch/microblaze/kernel/syscalls/syscall.tbl
> index 223d26303627..89fb60006a99 100644
> --- a/arch/microblaze/kernel/syscalls/syscall.tbl
> +++ b/arch/microblaze/kernel/syscalls/syscall.tbl
> @@ -477,3 +477,6 @@
> 469 common file_setattr sys_file_setattr
> 470 common listns sys_listns
> 471 common rseq_slice_yield sys_rseq_slice_yield
> +472 common mq_timedreceive2 sys_mq_timedreceive2_time32
> +473 common mq_timedreceive2_time64 sys_mq_timedreceive2
> +
> diff --git a/arch/mips/kernel/syscalls/syscall_n32.tbl b/arch/mips/kernel/syscalls/syscall_n32.tbl
> index 7430714e2b8f..fee830be67a6 100644
> --- a/arch/mips/kernel/syscalls/syscall_n32.tbl
> +++ b/arch/mips/kernel/syscalls/syscall_n32.tbl
> @@ -410,3 +410,5 @@
> 469 n32 file_setattr sys_file_setattr
> 470 n32 listns sys_listns
> 471 n32 rseq_slice_yield sys_rseq_slice_yield
> +472 n32 mq_timedreceive2 sys_mq_timedreceive2_time32
> +473 n32 mq_timedreceive2_time64 sys_mq_timedreceive2
So these receive two variants...
> diff --git a/arch/mips/kernel/syscalls/syscall_n64.tbl b/arch/mips/kernel/syscalls/syscall_n64.tbl
> index 630aab9e5425..75de6ee2df94 100644
> --- a/arch/mips/kernel/syscalls/syscall_n64.tbl
> +++ b/arch/mips/kernel/syscalls/syscall_n64.tbl
> @@ -386,3 +386,4 @@
> 469 n64 file_setattr sys_file_setattr
> 470 n64 listns sys_listns
> 471 n64 rseq_slice_yield sys_rseq_slice_yield
> +472 n64 mq_timedreceive2 sys_mq_timedreceive2
MIPS64 receives only one?
> diff --git a/arch/mips/kernel/syscalls/syscall_o32.tbl b/arch/mips/kernel/syscalls/syscall_o32.tbl
> index 128653112284..8694a2d2f084 100644
> --- a/arch/mips/kernel/syscalls/syscall_o32.tbl
> +++ b/arch/mips/kernel/syscalls/syscall_o32.tbl
> @@ -459,3 +459,5 @@
> 469 o32 file_setattr sys_file_setattr
> 470 o32 listns sys_listns
> 471 o32 rseq_slice_yield sys_rseq_slice_yield
> +472 o32 mq_timedreceive2 sys_mq_timedreceive2_time32
> +473 o32 mq_timedreceive2_time64 sys_mq_timedreceive2
> diff --git a/arch/powerpc/kernel/syscalls/syscall.tbl b/arch/powerpc/kernel/syscalls/syscall.tbl
> index 4fcc7c58a105..fd90a2f500b6 100644
> --- a/arch/powerpc/kernel/syscalls/syscall.tbl
> +++ b/arch/powerpc/kernel/syscalls/syscall.tbl
> @@ -562,3 +562,6 @@
> 469 common file_setattr sys_file_setattr
> 470 common listns sys_listns
> 471 nospu rseq_slice_yield sys_rseq_slice_yield
> +472 32 mq_timedreceive2 sys_mq_timedreceive2_time32
> +473 64 mq_timedreceive2 sys_mq_timedreceive2
> +474 32 mq_timedreceive2_time64 sys_mq_timedreceive2 sys_mq_timedreceive2
PowerPC receives three variants?
Oh, the first two are for 32-bit vs. 64-bit, so they should use the
same syscall number.
More of this below...
Furthermore, this breaks the "new syscalls use the same number on
most architectures"-rule. Next free slot is now 472, 473, or 474,
depending on architecture.
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org
In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] [PATCH V5] mqueue: introduce new do_mq_timedreceive2() [ mq_peek syscall] for non-destructive receive and inspection,fix minor issue,prepared doc.
2026-03-06 9:38 ` Geert Uytterhoeven
@ 2026-03-06 10:30 ` Arnd Bergmann
0 siblings, 0 replies; 6+ messages in thread
From: Arnd Bergmann @ 2026-03-06 10:30 UTC (permalink / raw)
To: Geert Uytterhoeven, Mathura_Kumar
Cc: Christian Brauner, Linux-Arch, linux-kernel, Alexander Viro
On Fri, Mar 6, 2026, at 10:38, Geert Uytterhoeven wrote:
> CC Arnd
Thanks!
> On Fri, 6 Mar 2026 at 08:50, Mathura_Kumar <academic1mathura@gmail.com> wrote:
>
>> +1) Overview
>> +-----------
>> +
>> +POSIX message queues on Linux provide mq_receive() and mq_timedreceive()
>> +for consuming messages from a queue. Both interfaces require the caller
>> +to pass the message buffer, length, and priority pointer as individual
>> +arguments to the system call. This imposes a fixed calling convention
>> +that cannot be extended without breaking the ABI.
>> +
>> +mq_timedreceive2() introduces a new system call entry point that accepts
>> +message buffer parameters via a struct argument rather than as individual
>> +syscall arguments. This frees the remaining syscall argument slots for
>> +new functionality flags and a message index, enabling non-destructive
>> +peek and indexed access semantics that are not possible with the
>> +original interface.
>> +
>> +Two variants are provided:
>> +
>> + mq_timedreceive2() - primary variant, 64-bit time (Y2038-safe)
>> + mq_timedreceive2_time32() - 32-bit time variant for legacy and compat
>
> What is the rationale behind adding a new syscall that is not Y2038-safe?
Indeed, that would need a very strong reason. New userspace already has to
support old kernels without the syscalls, and it would seem easy enough
to just not support it on time32 libc implementations.
>> index 4fcc7c58a105..fd90a2f500b6 100644
>> --- a/arch/powerpc/kernel/syscalls/syscall.tbl
>> +++ b/arch/powerpc/kernel/syscalls/syscall.tbl
>> @@ -562,3 +562,6 @@
>> 469 common file_setattr sys_file_setattr
>> 470 common listns sys_listns
>> 471 nospu rseq_slice_yield sys_rseq_slice_yield
>> +472 32 mq_timedreceive2 sys_mq_timedreceive2_time32
>> +473 64 mq_timedreceive2 sys_mq_timedreceive2
>> +474 32 mq_timedreceive2_time64 sys_mq_timedreceive2 sys_mq_timedreceive2
>
> PowerPC receives three variants?
> Oh, the first two are for 32-bit vs. 64-bit, so they should use the
> same syscall number.
Probably just a typo there (same on sparc), I assume the intention was
to use the same number as the generic scripts/syscall.tbl version.
> Furthermore, this breaks the "new syscalls use the same number on
> most architectures"-rule. Next free slot is now 472, 473, or 474,
> depending on architecture.
Right. I also noticed a few more mistakes with the numbers:
- parisc is missing the syscall entirely
- arm32 compat mode has the numbers reversed compared to native mode
- x86 only gets a version in the 'x32' range, but is missing the
native x86-64 version.
- The testcase hardcodes the x32 syscall number for all architectures
other than a couple of hardcoded ones
Also, I would point out few more general issues:
- compat handling is is busted for the normal syscalls, and only
seems to work by accident on the time32 one.
- the time32 syscall passes a kernel structure through a __user
pointer
- this is the first patch (or public email) ever sent by the "author"
- the changelog text and subject line make no sense whatsoever
- four submissions of the series in three days, with the latest
one labeled 'V5'.
I don't normally dismiss new submissions easily, but I think it's
safe to ignore this one.
Arnd
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] [PATCH V5] mqueue: introduce new do_mq_timedreceive2() [ mq_peek syscall] for non-destructive receive and inspection,fix minor issue,prepared doc.
2026-03-06 7:50 [PATCH] [PATCH V5] mqueue: introduce new do_mq_timedreceive2() [ mq_peek syscall] for non-destructive receive and inspection,fix minor issue,prepared doc Mathura_Kumar
2026-03-06 9:38 ` Geert Uytterhoeven
@ 2026-03-06 21:56 ` Randy Dunlap
2026-03-06 22:11 ` kernel test robot
2026-03-07 12:43 ` kernel test robot
3 siblings, 0 replies; 6+ messages in thread
From: Randy Dunlap @ 2026-03-06 21:56 UTC (permalink / raw)
To: Mathura_Kumar, brauner; +Cc: linux-arch, linux-kernel, viro
Warning: trailing whitespace in line 1763 of ipc/mqueue.c
On 3/5/26 11:50 PM, Mathura_Kumar wrote:
> +2) Motivation
> +-------------
> +
> +The original mq_timedreceive() interface exhausts all usable syscall
> +argument slots with mqdes, msg_ptr, msg_len, msg_prio, and abs_timeout.
> +Adding new behavior such as peek-without-consume or indexed message
> +access would require either a new syscall with a completely different
> +signature, or overloading existing arguments in ways that harm clarity
> +and correctness.
> +
> +By consolidating the message buffer parameters into a struct, this
> +syscall recovers argument slots for:
End the line above with "for::" so that the following lines don't run
together.
> +
> + flags - controls peek vs consume and other future behavior
> + index - selects a specific message position in the priority queue
> +
> +This design follows the precedent of other Linux syscalls that use
> +struct arguments to allow forward-compatible extensibility, such as
> +clone3() and openat2().
How much AI use is present in this patch?
thanks.
--
~Randy
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] [PATCH V5] mqueue: introduce new do_mq_timedreceive2() [ mq_peek syscall] for non-destructive receive and inspection,fix minor issue,prepared doc.
2026-03-06 7:50 [PATCH] [PATCH V5] mqueue: introduce new do_mq_timedreceive2() [ mq_peek syscall] for non-destructive receive and inspection,fix minor issue,prepared doc Mathura_Kumar
2026-03-06 9:38 ` Geert Uytterhoeven
2026-03-06 21:56 ` Randy Dunlap
@ 2026-03-06 22:11 ` kernel test robot
2026-03-07 12:43 ` kernel test robot
3 siblings, 0 replies; 6+ messages in thread
From: kernel test robot @ 2026-03-06 22:11 UTC (permalink / raw)
To: Mathura_Kumar, brauner
Cc: oe-kbuild-all, academic1mathura, linux-arch, linux-kernel, viro
Hi Mathura_Kumar,
kernel test robot noticed the following build errors:
[auto build test ERROR on shuah-kselftest/next]
[also build test ERROR on shuah-kselftest/fixes brauner-vfs/vfs.all linus/master v7.0-rc2 next-20260306]
[cannot apply to tip/x86/asm]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]
url: https://github.com/intel-lab-lkp/linux/commits/Mathura_Kumar/mqueue-introduce-new-do_mq_timedreceive2-mq_peek-syscall-for-non-destructive-receive-and-inspection-fix-minor-issue-prep/20260306-155240
base: https://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest.git next
patch link: https://lore.kernel.org/r/20260306075009.83723-1-academic1mathura%40gmail.com
patch subject: [PATCH] [PATCH V5] mqueue: introduce new do_mq_timedreceive2() [ mq_peek syscall] for non-destructive receive and inspection,fix minor issue,prepared doc.
config: powerpc-randconfig-r053-20260307 (https://download.01.org/0day-ci/archive/20260307/202603070636.7HA192z4-lkp@intel.com/config)
compiler: powerpc-linux-gcc (GCC) 8.5.0
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20260307/202603070636.7HA192z4-lkp@intel.com/reproduce)
If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Closes: https://lore.kernel.org/oe-kbuild-all/202603070636.7HA192z4-lkp@intel.com/
All errors (new ones prefixed by >>):
include/linux/syscalls.h:122:27: note: in expansion of macro '__SC_LONG'
#define __MAP4(m,t,a,...) m(t,a), __MAP3(m,__VA_ARGS__)
^
include/linux/syscalls.h:123:35: note: in expansion of macro '__MAP4'
#define __MAP5(m,t,a,...) m(t,a), __MAP4(m,__VA_ARGS__)
^~~~~~
include/linux/syscalls.h:125:22: note: in expansion of macro '__MAP5'
#define __MAP(n,...) __MAP##n(__VA_ARGS__)
^~~~~
arch/powerpc/include/asm/syscall_wrapper.h:27:29: note: in expansion of macro '__MAP'
static long __se_sys##name(__MAP(x,__SC_LONG,__VA_ARGS__)) \
^~~~~
include/linux/syscalls.h:237:2: note: in expansion of macro '__SYSCALL_DEFINEx'
__SYSCALL_DEFINEx(x, sname, __VA_ARGS__)
^~~~~~~~~~~~~~~~~
include/linux/syscalls.h:230:36: note: in expansion of macro 'SYSCALL_DEFINEx'
#define SYSCALL_DEFINE5(name, ...) SYSCALL_DEFINEx(5, _##name, __VA_ARGS__)
^~~~~~~~~~~~~~~
ipc/mqueue.c:1747:1: note: in expansion of macro 'SYSCALL_DEFINE5'
SYSCALL_DEFINE5(mq_timedreceive2_time32, mqd_t, mqdes,
^~~~~~~~~~~~~~~
In file included from ipc/mqueue.c:29:
ipc/mqueue.c: In function '__se_sys_mq_timedreceive2_time32':
include/linux/syscalls.h:133:25: error: passing argument 2 of '__do_sys_mq_timedreceive2_time32' from incompatible pointer type [-Werror=incompatible-pointer-types]
#define __SC_CAST(t, a) (__force t) a
include/linux/syscalls.h:122:27: note: in expansion of macro '__SC_CAST'
#define __MAP4(m,t,a,...) m(t,a), __MAP3(m,__VA_ARGS__)
^
include/linux/syscalls.h:123:35: note: in expansion of macro '__MAP4'
#define __MAP5(m,t,a,...) m(t,a), __MAP4(m,__VA_ARGS__)
^~~~~~
include/linux/syscalls.h:125:22: note: in expansion of macro '__MAP5'
#define __MAP(n,...) __MAP##n(__VA_ARGS__)
^~~~~
arch/powerpc/include/asm/syscall_wrapper.h:29:29: note: in expansion of macro '__MAP'
long ret = __do_sys##name(__MAP(x,__SC_CAST,__VA_ARGS__)); \
^~~~~
include/linux/syscalls.h:237:2: note: in expansion of macro '__SYSCALL_DEFINEx'
__SYSCALL_DEFINEx(x, sname, __VA_ARGS__)
^~~~~~~~~~~~~~~~~
include/linux/syscalls.h:230:36: note: in expansion of macro 'SYSCALL_DEFINEx'
#define SYSCALL_DEFINE5(name, ...) SYSCALL_DEFINEx(5, _##name, __VA_ARGS__)
^~~~~~~~~~~~~~~
ipc/mqueue.c:1747:1: note: in expansion of macro 'SYSCALL_DEFINE5'
SYSCALL_DEFINE5(mq_timedreceive2_time32, mqd_t, mqdes,
^~~~~~~~~~~~~~~
ipc/mqueue.c:1748:49: note: expected 'struct compat_mq_timedreceive2_args *' but argument is of type 'struct compat_mq_timedreceive2_args *'
struct compat_mq_timedreceive2_args __user *, uargs,
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~
include/linux/syscalls.h:127:27: note: in definition of macro '__SC_DECL'
#define __SC_DECL(t, a) t a
^
include/linux/syscalls.h:123:35: note: in expansion of macro '__MAP4'
#define __MAP5(m,t,a,...) m(t,a), __MAP4(m,__VA_ARGS__)
^~~~~~
include/linux/syscalls.h:125:22: note: in expansion of macro '__MAP5'
#define __MAP(n,...) __MAP##n(__VA_ARGS__)
^~~~~
arch/powerpc/include/asm/syscall_wrapper.h:22:36: note: in expansion of macro '__MAP'
static inline long __do_sys##name(__MAP(x,__SC_DECL,__VA_ARGS__)); \
^~~~~
include/linux/syscalls.h:237:2: note: in expansion of macro '__SYSCALL_DEFINEx'
__SYSCALL_DEFINEx(x, sname, __VA_ARGS__)
^~~~~~~~~~~~~~~~~
include/linux/syscalls.h:230:36: note: in expansion of macro 'SYSCALL_DEFINEx'
#define SYSCALL_DEFINE5(name, ...) SYSCALL_DEFINEx(5, _##name, __VA_ARGS__)
^~~~~~~~~~~~~~~
ipc/mqueue.c:1747:1: note: in expansion of macro 'SYSCALL_DEFINE5'
SYSCALL_DEFINE5(mq_timedreceive2_time32, mqd_t, mqdes,
^~~~~~~~~~~~~~~
ipc/mqueue.c: At top level:
ipc/mqueue.c:1748:10: warning: 'struct compat_mq_timedreceive2_args' declared inside parameter list will not be visible outside of this definition or declaration
struct compat_mq_timedreceive2_args __user *, uargs,
^~~~~~~~~~~~~~~~~~~~~~~~~~~~
include/linux/syscalls.h:127:25: note: in definition of macro '__SC_DECL'
#define __SC_DECL(t, a) t a
^
include/linux/syscalls.h:123:35: note: in expansion of macro '__MAP4'
#define __MAP5(m,t,a,...) m(t,a), __MAP4(m,__VA_ARGS__)
^~~~~~
include/linux/syscalls.h:125:22: note: in expansion of macro '__MAP5'
#define __MAP(n,...) __MAP##n(__VA_ARGS__)
^~~~~
arch/powerpc/include/asm/syscall_wrapper.h:34:36: note: in expansion of macro '__MAP'
static inline long __do_sys##name(__MAP(x,__SC_DECL,__VA_ARGS__))
^~~~~
include/linux/syscalls.h:237:2: note: in expansion of macro '__SYSCALL_DEFINEx'
__SYSCALL_DEFINEx(x, sname, __VA_ARGS__)
^~~~~~~~~~~~~~~~~
include/linux/syscalls.h:230:36: note: in expansion of macro 'SYSCALL_DEFINEx'
#define SYSCALL_DEFINE5(name, ...) SYSCALL_DEFINEx(5, _##name, __VA_ARGS__)
^~~~~~~~~~~~~~~
ipc/mqueue.c:1747:1: note: in expansion of macro 'SYSCALL_DEFINE5'
SYSCALL_DEFINE5(mq_timedreceive2_time32, mqd_t, mqdes,
^~~~~~~~~~~~~~~
In file included from include/linux/compat.h:34,
from include/net/compat.h:8,
from include/net/scm.h:13,
from include/linux/netlink.h:9,
from ipc/mqueue.c:28:
>> arch/powerpc/include/asm/syscall_wrapper.h:34:21: error: conflicting types for '__do_sys_mq_timedreceive2_time32'
static inline long __do_sys##name(__MAP(x,__SC_DECL,__VA_ARGS__))
^~~~~~~~
include/linux/syscalls.h:237:2: note: in expansion of macro '__SYSCALL_DEFINEx'
__SYSCALL_DEFINEx(x, sname, __VA_ARGS__)
^~~~~~~~~~~~~~~~~
include/linux/syscalls.h:230:36: note: in expansion of macro 'SYSCALL_DEFINEx'
#define SYSCALL_DEFINE5(name, ...) SYSCALL_DEFINEx(5, _##name, __VA_ARGS__)
^~~~~~~~~~~~~~~
ipc/mqueue.c:1747:1: note: in expansion of macro 'SYSCALL_DEFINE5'
SYSCALL_DEFINE5(mq_timedreceive2_time32, mqd_t, mqdes,
^~~~~~~~~~~~~~~
arch/powerpc/include/asm/syscall_wrapper.h:22:21: note: previous declaration of '__do_sys_mq_timedreceive2_time32' was here
static inline long __do_sys##name(__MAP(x,__SC_DECL,__VA_ARGS__)); \
^~~~~~~~
include/linux/syscalls.h:237:2: note: in expansion of macro '__SYSCALL_DEFINEx'
__SYSCALL_DEFINEx(x, sname, __VA_ARGS__)
^~~~~~~~~~~~~~~~~
include/linux/syscalls.h:230:36: note: in expansion of macro 'SYSCALL_DEFINEx'
#define SYSCALL_DEFINE5(name, ...) SYSCALL_DEFINEx(5, _##name, __VA_ARGS__)
^~~~~~~~~~~~~~~
ipc/mqueue.c:1747:1: note: in expansion of macro 'SYSCALL_DEFINE5'
SYSCALL_DEFINE5(mq_timedreceive2_time32, mqd_t, mqdes,
^~~~~~~~~~~~~~~
ipc/mqueue.c: In function '__do_sys_mq_timedreceive2_time32':
ipc/mqueue.c:1756:6: error: implicit declaration of function 'get_compat_mq_args'; did you mean 'get_compat_msghdr'? [-Werror=implicit-function-declaration]
if (get_compat_mq_args(pargs, uargs))
^~~~~~~~~~~~~~~~~~
get_compat_msghdr
In file included from include/linux/compat.h:34,
from include/net/compat.h:8,
from include/net/scm.h:13,
from include/linux/netlink.h:9,
from ipc/mqueue.c:28:
ipc/mqueue.c: At top level:
arch/powerpc/include/asm/syscall_wrapper.h:22:21: warning: '__do_sys_mq_timedreceive2_time32' used but never defined
static inline long __do_sys##name(__MAP(x,__SC_DECL,__VA_ARGS__)); \
^~~~~~~~
include/linux/syscalls.h:237:2: note: in expansion of macro '__SYSCALL_DEFINEx'
__SYSCALL_DEFINEx(x, sname, __VA_ARGS__)
^~~~~~~~~~~~~~~~~
include/linux/syscalls.h:230:36: note: in expansion of macro 'SYSCALL_DEFINEx'
#define SYSCALL_DEFINE5(name, ...) SYSCALL_DEFINEx(5, _##name, __VA_ARGS__)
^~~~~~~~~~~~~~~
ipc/mqueue.c:1747:1: note: in expansion of macro 'SYSCALL_DEFINE5'
SYSCALL_DEFINE5(mq_timedreceive2_time32, mqd_t, mqdes,
^~~~~~~~~~~~~~~
cc1: some warnings being treated as errors
vim +/__do_sys_mq_timedreceive2_time32 +34 arch/powerpc/include/asm/syscall_wrapper.h
7e92e01b724526 Rohan McLure 2022-09-21 12
7e92e01b724526 Rohan McLure 2022-09-21 13 #define SC_POWERPC_REGS_TO_ARGS(x, ...) \
7e92e01b724526 Rohan McLure 2022-09-21 14 __MAP(x,__SC_ARGS \
7e92e01b724526 Rohan McLure 2022-09-21 15 ,,regs->gpr[3],,regs->gpr[4],,regs->gpr[5] \
7e92e01b724526 Rohan McLure 2022-09-21 16 ,,regs->gpr[6],,regs->gpr[7],,regs->gpr[8])
7e92e01b724526 Rohan McLure 2022-09-21 17
7e92e01b724526 Rohan McLure 2022-09-21 18 #define __SYSCALL_DEFINEx(x, name, ...) \
94746890202cf1 Michael Ellerman 2022-10-06 19 long sys##name(const struct pt_regs *regs); \
94746890202cf1 Michael Ellerman 2022-10-06 20 ALLOW_ERROR_INJECTION(sys##name, ERRNO); \
7e92e01b724526 Rohan McLure 2022-09-21 21 static long __se_sys##name(__MAP(x,__SC_LONG,__VA_ARGS__)); \
7e92e01b724526 Rohan McLure 2022-09-21 22 static inline long __do_sys##name(__MAP(x,__SC_DECL,__VA_ARGS__)); \
94746890202cf1 Michael Ellerman 2022-10-06 23 long sys##name(const struct pt_regs *regs) \
7e92e01b724526 Rohan McLure 2022-09-21 24 { \
7e92e01b724526 Rohan McLure 2022-09-21 25 return __se_sys##name(SC_POWERPC_REGS_TO_ARGS(x,__VA_ARGS__)); \
7e92e01b724526 Rohan McLure 2022-09-21 26 } \
7e92e01b724526 Rohan McLure 2022-09-21 27 static long __se_sys##name(__MAP(x,__SC_LONG,__VA_ARGS__)) \
7e92e01b724526 Rohan McLure 2022-09-21 28 { \
7e92e01b724526 Rohan McLure 2022-09-21 29 long ret = __do_sys##name(__MAP(x,__SC_CAST,__VA_ARGS__)); \
7e92e01b724526 Rohan McLure 2022-09-21 30 __MAP(x,__SC_TEST,__VA_ARGS__); \
7e92e01b724526 Rohan McLure 2022-09-21 31 __PROTECT(x, ret,__MAP(x,__SC_ARGS,__VA_ARGS__)); \
7e92e01b724526 Rohan McLure 2022-09-21 32 return ret; \
7e92e01b724526 Rohan McLure 2022-09-21 33 } \
7e92e01b724526 Rohan McLure 2022-09-21 @34 static inline long __do_sys##name(__MAP(x,__SC_DECL,__VA_ARGS__))
7e92e01b724526 Rohan McLure 2022-09-21 35
--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] [PATCH V5] mqueue: introduce new do_mq_timedreceive2() [ mq_peek syscall] for non-destructive receive and inspection,fix minor issue,prepared doc.
2026-03-06 7:50 [PATCH] [PATCH V5] mqueue: introduce new do_mq_timedreceive2() [ mq_peek syscall] for non-destructive receive and inspection,fix minor issue,prepared doc Mathura_Kumar
` (2 preceding siblings ...)
2026-03-06 22:11 ` kernel test robot
@ 2026-03-07 12:43 ` kernel test robot
3 siblings, 0 replies; 6+ messages in thread
From: kernel test robot @ 2026-03-07 12:43 UTC (permalink / raw)
To: Mathura_Kumar, brauner
Cc: llvm, oe-kbuild-all, academic1mathura, linux-arch, linux-kernel,
viro
Hi Mathura_Kumar,
kernel test robot noticed the following build errors:
[auto build test ERROR on shuah-kselftest/next]
[also build test ERROR on shuah-kselftest/fixes brauner-vfs/vfs.all linus/master v7.0-rc2 next-20260306]
[cannot apply to tip/x86/asm]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]
url: https://github.com/intel-lab-lkp/linux/commits/Mathura_Kumar/mqueue-introduce-new-do_mq_timedreceive2-mq_peek-syscall-for-non-destructive-receive-and-inspection-fix-minor-issue-prep/20260306-155240
base: https://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest.git next
patch link: https://lore.kernel.org/r/20260306075009.83723-1-academic1mathura%40gmail.com
patch subject: [PATCH] [PATCH V5] mqueue: introduce new do_mq_timedreceive2() [ mq_peek syscall] for non-destructive receive and inspection,fix minor issue,prepared doc.
config: x86_64-kexec (https://download.01.org/0day-ci/archive/20260307/202603072056.0e7wfE3v-lkp@intel.com/config)
compiler: clang version 20.1.8 (https://github.com/llvm/llvm-project 87f0227cb60147a26a1eeb4fb06e3b505e9c7261)
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20260307/202603072056.0e7wfE3v-lkp@intel.com/reproduce)
If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Closes: https://lore.kernel.org/oe-kbuild-all/202603072056.0e7wfE3v-lkp@intel.com/
All errors (new ones prefixed by >>):
ipc/mqueue.c:1748:10: warning: declaration of 'struct compat_mq_timedreceive2_args' will not be visible outside of this function [-Wvisibility]
1748 | struct compat_mq_timedreceive2_args __user *, uargs,
| ^
ipc/mqueue.c:1748:10: warning: declaration of 'struct compat_mq_timedreceive2_args' will not be visible outside of this function [-Wvisibility]
ipc/mqueue.c:1748:10: warning: declaration of 'struct compat_mq_timedreceive2_args' will not be visible outside of this function [-Wvisibility]
>> ipc/mqueue.c:1747:1: error: incompatible pointer types passing 'struct compat_mq_timedreceive2_args *' to parameter of type 'struct compat_mq_timedreceive2_args *' [-Werror,-Wincompatible-pointer-types]
1747 | SYSCALL_DEFINE5(mq_timedreceive2_time32, mqd_t, mqdes,
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1748 | struct compat_mq_timedreceive2_args __user *, uargs,
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1749 | unsigned int, flags, const unsigned long, index,
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1750 | const struct old_timespec32 __user *, u_abs_timeout)
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
include/linux/syscalls.h:230:36: note: expanded from macro 'SYSCALL_DEFINE5'
230 | #define SYSCALL_DEFINE5(name, ...) SYSCALL_DEFINEx(5, _##name, __VA_ARGS__)
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
include/linux/syscalls.h:237:2: note: expanded from macro 'SYSCALL_DEFINEx'
237 | __SYSCALL_DEFINEx(x, sname, __VA_ARGS__)
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
arch/x86/include/asm/syscall_wrapper.h:230:37: note: expanded from macro '__SYSCALL_DEFINEx'
230 | long ret = __do_sys##name(__MAP(x,__SC_CAST,__VA_ARGS__));\
| ~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~
note: (skipping 1 expansions in backtrace; use -fmacro-backtrace-limit=0 to see all)
include/linux/syscalls.h:123:42: note: expanded from macro '__MAP5'
123 | #define __MAP5(m,t,a,...) m(t,a), __MAP4(m,__VA_ARGS__)
| ~~~~~~~^~~~~~~~~~~~~~
include/linux/syscalls.h:122:27: note: expanded from macro '__MAP4'
122 | #define __MAP4(m,t,a,...) m(t,a), __MAP3(m,__VA_ARGS__)
| ^~~~~~
include/linux/syscalls.h:133:25: note: expanded from macro '__SC_CAST'
133 | #define __SC_CAST(t, a) (__force t) a
| ^~~~~~~~~~~~~
ipc/mqueue.c:1748:49: note: passing argument to parameter 'uargs' here
1748 | struct compat_mq_timedreceive2_args __user *, uargs,
| ^
ipc/mqueue.c:1748:10: warning: declaration of 'struct compat_mq_timedreceive2_args' will not be visible outside of this function [-Wvisibility]
1748 | struct compat_mq_timedreceive2_args __user *, uargs,
| ^
>> ipc/mqueue.c:1747:1: error: conflicting types for '__do_sys_mq_timedreceive2_time32'
1747 | SYSCALL_DEFINE5(mq_timedreceive2_time32, mqd_t, mqdes,
| ^
include/linux/syscalls.h:230:36: note: expanded from macro 'SYSCALL_DEFINE5'
230 | #define SYSCALL_DEFINE5(name, ...) SYSCALL_DEFINEx(5, _##name, __VA_ARGS__)
| ^
include/linux/syscalls.h:237:2: note: expanded from macro 'SYSCALL_DEFINEx'
237 | __SYSCALL_DEFINEx(x, sname, __VA_ARGS__)
| ^
arch/x86/include/asm/syscall_wrapper.h:235:21: note: expanded from macro '__SYSCALL_DEFINEx'
235 | static inline long __do_sys##name(__MAP(x,__SC_DECL,__VA_ARGS__))
| ^
<scratch space>:71:1: note: expanded from here
71 | __do_sys_mq_timedreceive2_time32
| ^
ipc/mqueue.c:1747:1: note: previous declaration is here
include/linux/syscalls.h:230:36: note: expanded from macro 'SYSCALL_DEFINE5'
230 | #define SYSCALL_DEFINE5(name, ...) SYSCALL_DEFINEx(5, _##name, __VA_ARGS__)
| ^
include/linux/syscalls.h:237:2: note: expanded from macro 'SYSCALL_DEFINEx'
237 | __SYSCALL_DEFINEx(x, sname, __VA_ARGS__)
| ^
arch/x86/include/asm/syscall_wrapper.h:225:21: note: expanded from macro '__SYSCALL_DEFINEx'
225 | static inline long __do_sys##name(__MAP(x,__SC_DECL,__VA_ARGS__));\
| ^
<scratch space>:44:1: note: expanded from here
44 | __do_sys_mq_timedreceive2_time32
| ^
>> ipc/mqueue.c:1756:6: error: call to undeclared function 'get_compat_mq_args'; ISO C99 and later do not support implicit function declarations [-Wimplicit-function-declaration]
1756 | if (get_compat_mq_args(pargs, uargs))
| ^
ipc/mqueue.c:1756:6: note: did you mean 'get_compat_msghdr'?
include/net/compat.h:51:5: note: 'get_compat_msghdr' declared here
51 | int get_compat_msghdr(struct msghdr *, struct compat_msghdr __user *,
| ^
4 warnings and 3 errors generated.
vim +1747 ipc/mqueue.c
1746
> 1747 SYSCALL_DEFINE5(mq_timedreceive2_time32, mqd_t, mqdes,
> 1748 struct compat_mq_timedreceive2_args __user *, uargs,
1749 unsigned int, flags, const unsigned long, index,
1750 const struct old_timespec32 __user *, u_abs_timeout)
1751 {
1752 struct mq_timedreceive2_args args, *pargs = NULL;
1753
1754 pargs = &args;
1755
> 1756 if (get_compat_mq_args(pargs, uargs))
1757 return -EFAULT;
1758
1759 struct timespec64 ts, *p = NULL;
1760
1761 if (u_abs_timeout) {
1762 int res = compat_prepare_timeout(u_abs_timeout, &ts);
1763
1764 if (res)
1765 return res;
1766 p = &ts;
1767 }
1768 return do_mq_timedreceive2(mqdes, pargs, flags, index, p);
1769 }
1770 #endif
1771
--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2026-03-07 12:44 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-03-06 7:50 [PATCH] [PATCH V5] mqueue: introduce new do_mq_timedreceive2() [ mq_peek syscall] for non-destructive receive and inspection,fix minor issue,prepared doc Mathura_Kumar
2026-03-06 9:38 ` Geert Uytterhoeven
2026-03-06 10:30 ` Arnd Bergmann
2026-03-06 21:56 ` Randy Dunlap
2026-03-06 22:11 ` kernel test robot
2026-03-07 12:43 ` kernel test robot
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox