Add /proc/syscalls to display percpu syscall count.
We need a less resource-intensive way to count syscall per cpu
for system problem location.
There is a similar utility syscount in the BCC project, but syscount
has a high performance cost.
The following is a comparison on the same machine, using UnixBench
System Call Overhead:
┏━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┳━━━━━━━━┓
┃ Change ┃ Unixbench Score ┃ Loss ┃
┡━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━╇━━━━━━━━┩
│ no change │ 1072.6 │ --- │
│ syscall count │ 982.5 │ 8.40% │
│ bpf syscount │ 614.2 │ 42.74% │
└───────────────┴─────────────────┴────────┘
UnixBench System Call Use sys_gettid to test, this system call only reads
one variable, so the performance penalty seems large. When tested with
fork, the test scores were almost the same.
So the conclusion is that it does not have a significant impact on system
call performance.
This function depends on CONFIG_FTRACE_SYSCALLS because the system call
number is stored in syscall_metadata.
Signed-off-by: Zhang Yuchen <zhangyuchen.lcr@xxxxxxxxxxxxx>
---
Documentation/filesystems/proc.rst | 28 +++++++++
arch/arm64/include/asm/syscall_wrapper.h | 2 +-
arch/s390/include/asm/syscall_wrapper.h | 4 +-
arch/x86/include/asm/syscall_wrapper.h | 2 +-
fs/proc/Kconfig | 7 +++
fs/proc/Makefile | 1 +
fs/proc/syscall.c | 79 ++++++++++++++++++++++++
include/linux/syscalls.h | 51 +++++++++++++--
8 files changed, 165 insertions(+), 9 deletions(-)
create mode 100644 fs/proc/syscall.c
diff --git a/Documentation/filesystems/proc.rst b/Documentation/filesystems/proc.rst
index 1bc91fb8c321..80394a98a192 100644
--- a/Documentation/filesystems/proc.rst
+++ b/Documentation/filesystems/proc.rst
@@ -686,6 +686,7 @@ files are there, and which are missing.
fs File system parameters, currently nfs/exports (2.4)
ide Directory containing info about the IDE subsystem
interrupts Interrupt usage
+ syscalls Syscall count for each cpu
iomem Memory map (2.4)
ioports I/O port usage
irq Masks for irq to cpu affinity (2.4)(smp?)
@@ -1225,6 +1226,33 @@ Provides counts of softirq handlers serviced since boot time, for each CPU.
HRTIMER: 0 0 0 0
RCU: 1678 1769 2178 2250
+syscalls
+~~~~~~~~
+
+Provides counts of syscall since boot time, for each cpu.
+
+::
+
+ > cat /proc/syscalls
+ CPU0 CPU1 CPU2 CPU3
+ 0: 3743 3099 3770 3242 sys_read
+ 1: 222 559 822 522 sys_write
+ 2: 0 0 0 0 sys_open
+ 3: 6481 18754 12077 7349 sys_close
+ 4: 11362 11120 11343 10665 sys_newstat
+ 5: 5224 13880 8578 5971 sys_newfstat
+ 6: 1228 1269 1459 1508 sys_newlstat
+ 7: 90 43 64 67 sys_poll
+ 8: 1635 1000 2071 1161 sys_lseek
+ .... omit the middle line ....
+ 441: 0 0 0 0 sys_epoll_pwait2
+ 442: 0 0 0 0 sys_mount_setattr
+ 443: 0 0 0 0 sys_quotactl_fd
+ 447: 0 0 0 0 sys_memfd_secret
+ 448: 0 0 0 0 sys_process_mrelease
+ 449: 0 0 0 0 sys_futex_waitv
+ 450: 0 0 0 0 sys_set_mempolicy_home_node
+
1.3 Networking info in /proc/net
--------------------------------
diff --git a/arch/arm64/include/asm/syscall_wrapper.h b/arch/arm64/include/asm/syscall_wrapper.h
index b383b4802a7b..d9ec21df4c44 100644
--- a/arch/arm64/include/asm/syscall_wrapper.h
+++ b/arch/arm64/include/asm/syscall_wrapper.h
@@ -66,7 +66,7 @@ struct pt_regs;
} \
static inline long __do_sys##name(__MAP(x,__SC_DECL,__VA_ARGS__))
-#define SYSCALL_DEFINE0(sname) \
+#define __SYSCALL_DEFINE0(sname) \
SYSCALL_METADATA(_##sname, 0); \
asmlinkage long __arm64_sys_##sname(const struct pt_regs *__unused); \
ALLOW_ERROR_INJECTION(__arm64_sys_##sname, ERRNO); \
diff --git a/arch/s390/include/asm/syscall_wrapper.h b/arch/s390/include/asm/syscall_wrapper.h
index fde7e6b1df48..f8d7d9010de2 100644
--- a/arch/s390/include/asm/syscall_wrapper.h
+++ b/arch/s390/include/asm/syscall_wrapper.h
@@ -77,7 +77,7 @@
ALLOW_ERROR_INJECTION(__s390_compat_sys_##sname, ERRNO); \
long __s390_compat_sys_##sname(void)
-#define SYSCALL_DEFINE0(sname) \
+#define __SYSCALL_DEFINE0(sname) \
SYSCALL_METADATA(_##sname, 0); \
long __s390x_sys_##sname(void); \
ALLOW_ERROR_INJECTION(__s390x_sys_##sname, ERRNO); \
@@ -128,7 +128,7 @@
#define __S390_SYS_STUBx(x, fullname, name, ...)
-#define SYSCALL_DEFINE0(sname) \
+#define __SYSCALL_DEFINE0(sname) \
SYSCALL_METADATA(_##sname, 0); \
long __s390x_sys_##sname(void); \
ALLOW_ERROR_INJECTION(__s390x_sys_##sname, ERRNO); \
diff --git a/arch/x86/include/asm/syscall_wrapper.h b/arch/x86/include/asm/syscall_wrapper.h
index 59358d1bf880..1f16436c13bd 100644
--- a/arch/x86/include/asm/syscall_wrapper.h
+++ b/arch/x86/include/asm/syscall_wrapper.h
@@ -246,7 +246,7 @@ extern long __ia32_sys_ni_syscall(const struct pt_regs *regs);
* SYSCALL_DEFINEx() -- which is essential for the COND_SYSCALL() and SYS_NI()
* macros to work correctly.
*/
-#define SYSCALL_DEFINE0(sname) \
+#define __SYSCALL_DEFINE0(sname) \
SYSCALL_METADATA(_##sname, 0); \
static long __do_sys_##sname(const struct pt_regs *__unused); \
__X64_SYS_STUB0(sname) \
diff --git a/fs/proc/Kconfig b/fs/proc/Kconfig
index c930001056f9..9e5fa75ebd2a 100644
--- a/fs/proc/Kconfig
+++ b/fs/proc/Kconfig
@@ -100,6 +100,13 @@ config PROC_CHILDREN
Say Y if you are running any user-space software which takes benefit from
this interface. For example, rkt is such a piece of software.
+config PROC_SYSCALLS
+ bool "Include /proc/syscalls file" if EXPERT
+ depends on PROC_FS && FTRACE_SYSCALLS
+ default n
+ help
+ Provides a file that shows the number of syscall on each cpu.
+
config PROC_PID_ARCH_STATUS
def_bool n
depends on PROC_FS
diff --git a/fs/proc/Makefile b/fs/proc/Makefile
index bd08616ed8ba..f381a7aa90ae 100644
--- a/fs/proc/Makefile
+++ b/fs/proc/Makefile
@@ -31,6 +31,7 @@ proc-$(CONFIG_PROC_SYSCTL) += proc_sysctl.o
proc-$(CONFIG_NET) += proc_net.o
proc-$(CONFIG_PROC_KCORE) += kcore.o
proc-$(CONFIG_PROC_VMCORE) += vmcore.o
+proc-$(CONFIG_PROC_SYSCALLS) += syscall.o
proc-$(CONFIG_PRINTK) += kmsg.o
proc-$(CONFIG_PROC_PAGE_MONITOR) += page.o
proc-$(CONFIG_BOOT_CONFIG) += bootconfig.o
diff --git a/fs/proc/syscall.c b/fs/proc/syscall.c
new file mode 100644
index 000000000000..88196b16f430
--- /dev/null
+++ b/fs/proc/syscall.c
@@ -0,0 +1,79 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <linux/fs.h>
+#include <linux/init.h>
+#include <linux/proc_fs.h>
+#include <linux/seq_file.h>
+
+#include <asm/syscall.h>
+
+DEFINE_PER_CPU(u64 [NR_syscalls], __per_cpu_syscall_count);
+
+extern const char *get_syscall_name(int syscall_nr);
+
+int show_syscalls(struct seq_file *p, void *v)
+{
+ int i = *(loff_t *)v, j;
+ static int prec;
+ const char *syscall_name = get_syscall_name(i);
+
+ if (i > NR_syscalls)
+ return 0;
+
+ /* print header and calculate the width of the first column */
+ if (i == 0) {
+ for (prec = 3, j = 1000; prec < 10 && j <= NR_syscalls; ++prec)
+ j *= 10;
+ seq_printf(p, "%*s", prec + 8, "");
+ for_each_online_cpu(j)
+ seq_printf(p, "CPU%-8d", j);
+ seq_putc(p, '\n');
+ }
+
+ if (syscall_name == NULL)
+ return 0;
+
+ seq_printf(p, "%*d: ", prec, i);
+ for_each_online_cpu(j)
+ seq_printf(p, "%10llu ",
+ per_cpu(__per_cpu_syscall_count, j)[i]);
+ seq_printf(p, " %s", syscall_name);
+ seq_putc(p, '\n');
+
+ return 0;
+}
+
+/*
+ * /proc/syscalls
+ */
+static void *int_seq_start(struct seq_file *f, loff_t *pos)
+{
+ return (*pos <= NR_syscalls) ? pos : NULL;
+}
+
+static void *int_seq_next(struct seq_file *f, void *v, loff_t *pos)
+{
+ (*pos)++;
+ if (*pos > NR_syscalls)
+ return NULL;
+ return pos;
+}
+
+static void int_seq_stop(struct seq_file *f, void *v)
+{
+ /* Nothing to do */
+}
+
+static const struct seq_operations int_seq_ops = {
+ .start = int_seq_start,
+ .next = int_seq_next,
+ .stop = int_seq_stop,
+ .show = show_syscalls
+};
+
+static int __init proc_syscall_init(void)
+{
+ proc_create_seq("syscalls", 0, NULL, &int_seq_ops);
+ return 0;
+}
+
+fs_initcall(proc_syscall_init);
diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h
index a34b0f9a9972..a3d50b8d39d8 100644
--- a/include/linux/syscalls.h
+++ b/include/linux/syscalls.h
@@ -86,6 +86,7 @@ enum landlock_rule_type;
#include <linux/key.h>
#include <linux/personality.h>
#include <trace/syscall.h>
+#include <asm/syscall.h>
#ifdef CONFIG_ARCH_HAS_SYSCALL_WRAPPER
/*
@@ -206,8 +207,8 @@ static inline int is_syscall_trace_event(struct trace_event_call *tp_event)
}
#endif
-#ifndef SYSCALL_DEFINE0
-#define SYSCALL_DEFINE0(sname) \
+#ifndef __SYSCALL_DEFINE0
+#define __SYSCALL_DEFINE0(sname) \
SYSCALL_METADATA(_##sname, 0); \
asmlinkage long sys_##sname(void); \
ALLOW_ERROR_INJECTION(sys_##sname, ERRNO); \
@@ -223,9 +224,49 @@ static inline int is_syscall_trace_event(struct trace_event_call *tp_event)
#define SYSCALL_DEFINE_MAXARGS 6
-#define SYSCALL_DEFINEx(x, sname, ...) \
- SYSCALL_METADATA(sname, x, __VA_ARGS__) \
- __SYSCALL_DEFINEx(x, sname, __VA_ARGS__)
+#ifdef CONFIG_PROC_SYSCALLS
+DECLARE_PER_CPU(u64[], __per_cpu_syscall_count);
+
+#define SYSCALL_COUNT_DECLAREx(sname, x, ...) \
+ static inline long __count_sys##sname(__MAP(x, __SC_DECL, __VA_ARGS__));
+
+#define __SYSCALL_COUNT(syscall_nr) \
+ this_cpu_inc(__per_cpu_syscall_count[(syscall_nr)])
+
+#define SYSCALL_COUNT_FUNCx(sname, x, ...) \
+ { \
+ __SYSCALL_COUNT(__syscall_meta_##sname.syscall_nr); \
+ return __count_sys##sname(__MAP(x, __SC_CAST, __VA_ARGS__)); \
+ } \
+ static inline long __count_sys##sname(__MAP(x, __SC_DECL, __VA_ARGS__))
+
+#define SYSCALL_COUNT_DECLARE0(sname) \
+ static inline long __count_sys_##sname(void);
+
+#define SYSCALL_COUNT_FUNC0(sname) \
+ { \
+ __SYSCALL_COUNT(__syscall_meta__##sname.syscall_nr); \
+ return __count_sys_##sname(); \
+ } \
+ static inline long __count_sys_##sname(void)
+
+#else
+#define SYSCALL_COUNT_DECLAREx(sname, x, ...)
+#define SYSCALL_COUNT_FUNCx(sname, x, ...)
+#define SYSCALL_COUNT_DECLARE0(sname)
+#define SYSCALL_COUNT_FUNC0(sname)
+#endif
+
+#define SYSCALL_DEFINEx(x, sname, ...) \
+ SYSCALL_METADATA(sname, x, __VA_ARGS__) \
+ SYSCALL_COUNT_DECLAREx(sname, x, __VA_ARGS__) \
+ __SYSCALL_DEFINEx(x, sname, __VA_ARGS__) \
+ SYSCALL_COUNT_FUNCx(sname, x, __VA_ARGS__)
+
+#define SYSCALL_DEFINE0(sname) \
+ SYSCALL_COUNT_DECLARE0(sname) \
+ __SYSCALL_DEFINE0(sname) \
+ SYSCALL_COUNT_FUNC0(sname)
#define __PROTECT(...) asmlinkage_protect(__VA_ARGS__)
--
2.30.2
