linuxppc-dev.lists.ozlabs.org archive mirror
 help / color / mirror / Atom feed
* [PATCH] Performance Stats: Kernel patch
@ 2007-05-08 16:26 Maxim Uvarov
  2007-05-08 19:32 ` Andrew Morton
  2007-05-08 23:32 ` Linas Vepstas
  0 siblings, 2 replies; 22+ messages in thread
From: Maxim Uvarov @ 2007-05-08 16:26 UTC (permalink / raw)
  To: Andrew Morton; +Cc: linuxppc-dev, wli, pavel, dada1


Patch makes available to the user the following
task and process performance statistics:
	* Involuntary Context Switches (task_struct->nivcsw)
	* Voluntary Context Switches (task_struct->nvcsw)
	* Number of system calls (added new counter
	  thread_info->sysall_count)
	           
Statistics information is available from:
        1. taskstats interface (Documentation/accounting/)
	2. /proc/PID/status (task only).

This data is useful for detecting hyperactivity
patterns between processes.

Signed-off-by: Maxim Uvarov <muvarov@ru.mvista.com>

Signed-off-by: Maxim Uvarov <muvarov@ru.mvista.com>                                                                                  
---

 Documentation/accounting/getdelays.c          |   20 ++++++++++++++++++--
 Documentation/accounting/taskstats-struct.txt |    7 +++++++
 arch/i386/kernel/asm-offsets.c                |    1 +
 arch/i386/kernel/entry.S                      |    1 +
 arch/powerpc/kernel/asm-offsets.c             |    2 ++
 arch/powerpc/kernel/entry_32.S                |    5 +++++
 arch/powerpc/kernel/entry_64.S                |    5 +++++
 arch/x86_64/kernel/asm-offsets.c              |    1 +
 arch/x86_64/kernel/entry.S                    |    1 +
 fs/proc/array.c                               |   14 ++++++++++++++
 include/asm-i386/thread_info.h                |    1 +
 include/asm-powerpc/thread_info.h             |    1 +
 include/asm-x86_64/thread_info.h              |    1 +
 include/linux/taskstats.h                     |    6 +++++-
 kernel/fork.c                                 |    3 +++
 kernel/taskstats.c                            |    6 ++++++
 16 files changed, 72 insertions(+), 3 deletions(-)

diff --git a/Documentation/accounting/getdelays.c b/Documentation/accounting/getdelays.c
index e9126e7..1be7d65 100644
--- a/Documentation/accounting/getdelays.c
+++ b/Documentation/accounting/getdelays.c
@@ -49,6 +49,7 @@ char name[100];
 int dbg;
 int print_delays;
 int print_io_accounting;
+int print_task_stats;
 __u64 stime, utime;
 
 #define PRINTF(fmt, arg...) {			\
@@ -187,7 +188,7 @@ void print_delayacct(struct taskstats *t)
 	       "IO    %15s%15s\n"
 	       "      %15llu%15llu\n"
 	       "MEM   %15s%15s\n"
-	       "      %15llu%15llu\n\n",
+	       "      %15llu%15llu\n"
 	       "count", "real total", "virtual total", "delay total",
 	       t->cpu_count, t->cpu_run_real_total, t->cpu_run_virtual_total,
 	       t->cpu_delay_total,
@@ -196,6 +197,15 @@ void print_delayacct(struct taskstats *t)
 	       "count", "delay total", t->swapin_count, t->swapin_delay_total);
 }
 
+void print_taskstats(struct taskstats *t)
+{
+	printf("\n\nTask  %15s%15s%15s\n"
+	       "      %15lu%15lu%15lu\n",
+	       "syscalls", "voluntary", "nonvoluntary",
+	       t->syscall_counter, t->nvcsw, t->nivcsw);
+
+}
+
 void print_ioacct(struct taskstats *t)
 {
 	printf("%s: read=%llu, write=%llu, cancelled_write=%llu\n",
@@ -227,7 +237,7 @@ int main(int argc, char *argv[])
 	struct msgtemplate msg;
 
 	while (1) {
-		c = getopt(argc, argv, "diw:r:m:t:p:v:l");
+		c = getopt(argc, argv, "qdiw:r:m:t:p:v:l");
 		if (c < 0)
 			break;
 
@@ -240,6 +250,10 @@ int main(int argc, char *argv[])
 			printf("printing IO accounting\n");
 			print_io_accounting = 1;
 			break;
+		case 'q':
+			printf("printing task/process stasistics:\n");
+			print_task_stats = 1;
+			break;
 		case 'w':
 			strncpy(logfile, optarg, MAX_FILENAME);
 			printf("write to file %s\n", logfile);
@@ -381,6 +395,8 @@ int main(int argc, char *argv[])
 							print_delayacct((struct taskstats *) NLA_DATA(na));
 						if (print_io_accounting)
 							print_ioacct((struct taskstats *) NLA_DATA(na));
+						if (print_task_stats)
+							print_taskstats((struct taskstats *) NLA_DATA(na));
 						if (fd) {
 							if (write(fd, NLA_DATA(na), na->nla_len) < 0) {
 								err(1,"write error\n");
diff --git a/Documentation/accounting/taskstats-struct.txt b/Documentation/accounting/taskstats-struct.txt
index 661c797..5dac173 100644
--- a/Documentation/accounting/taskstats-struct.txt
+++ b/Documentation/accounting/taskstats-struct.txt
@@ -22,6 +22,8 @@ There are three different groups of fields in the struct taskstats:
     /* Extended accounting fields end */
     Their values are collected if CONFIG_TASK_XACCT is set.
 
+4) Per-task and per-thread statistics
+
 Future extension should add fields to the end of the taskstats struct, and
 should not change the relative position of each field within the struct.
 
@@ -158,4 +160,9 @@ struct taskstats {
 
 	/* Extended accounting fields end */
 
+4) Per-task and per-thread statiscits
+	__u32	syscall_counter;	/* Syscall counter */
+	__u32	nvcsw;			/* Context voluntary switch counter */
+	__u32	nivcsw;			/* Context involuntary switch counter */
+
 }
diff --git a/arch/i386/kernel/asm-offsets.c b/arch/i386/kernel/asm-offsets.c
index 1b2f3cd..4ad49d2 100644
--- a/arch/i386/kernel/asm-offsets.c
+++ b/arch/i386/kernel/asm-offsets.c
@@ -56,6 +56,7 @@ void foo(void)
 	OFFSET(TI_addr_limit, thread_info, addr_limit);
 	OFFSET(TI_restart_block, thread_info, restart_block);
 	OFFSET(TI_sysenter_return, thread_info, sysenter_return);
+	OFFSET(TI_syscall_count, thread_info, syscall_count);
 	BLANK();
 
 	OFFSET(GDS_size, Xgt_desc_struct, size);
diff --git a/arch/i386/kernel/entry.S b/arch/i386/kernel/entry.S
index 5e47683..26f0cc0 100644
--- a/arch/i386/kernel/entry.S
+++ b/arch/i386/kernel/entry.S
@@ -331,6 +331,7 @@ sysenter_past_esp:
 	CFI_ADJUST_CFA_OFFSET 4
 	SAVE_ALL
 	GET_THREAD_INFO(%ebp)
+	incl	TI_syscall_count(%ebp)	# Increment syscalls counter
 
 	/* Note, _TIF_SECCOMP is bit number 8, and so it needs testw and not testb */
 	testw $(_TIF_SYSCALL_EMU|_TIF_SYSCALL_TRACE|_TIF_SECCOMP|_TIF_SYSCALL_AUDIT),TI_flags(%ebp)
diff --git a/arch/powerpc/kernel/asm-offsets.c b/arch/powerpc/kernel/asm-offsets.c
index 030d300..b640039 100644
--- a/arch/powerpc/kernel/asm-offsets.c
+++ b/arch/powerpc/kernel/asm-offsets.c
@@ -94,6 +94,8 @@ int main(void)
 	DEFINE(TI_LOCAL_FLAGS, offsetof(struct thread_info, local_flags));
 	DEFINE(TI_PREEMPT, offsetof(struct thread_info, preempt_count));
 	DEFINE(TI_TASK, offsetof(struct thread_info, task));
+	DEFINE(TI_SYSCALL_COUNT, offsetof(struct thread_info, syscall_count));
+
 #ifdef CONFIG_PPC32
 	DEFINE(TI_EXECDOMAIN, offsetof(struct thread_info, exec_domain));
 	DEFINE(TI_CPU, offsetof(struct thread_info, cpu));
diff --git a/arch/powerpc/kernel/entry_32.S b/arch/powerpc/kernel/entry_32.S
index c03e829..329c2f8 100644
--- a/arch/powerpc/kernel/entry_32.S
+++ b/arch/powerpc/kernel/entry_32.S
@@ -202,6 +202,11 @@ _GLOBAL(DoSyscall)
 	bl	do_show_syscall
 #endif /* SHOW_SYSCALLS */
 	rlwinm	r10,r1,0,0,(31-THREAD_SHIFT)	/* current_thread_info() */
+#ifdef CONFIG_THREAD_PERF_STAT_SYSC
+	lwz r11,TI_SYSC_CNT(r10)
+	addi r11,r11,1
+	stw r11,TI_SYSC_CNT(r10)
+#endif /* CONFIG_THREAD_PERF_STAT_SYSC */
 	lwz	r11,TI_FLAGS(r10)
 	andi.	r11,r11,_TIF_SYSCALL_T_OR_A
 	bne-	syscall_dotrace
diff --git a/arch/powerpc/kernel/entry_64.S b/arch/powerpc/kernel/entry_64.S
index 2551c08..5907f76 100644
--- a/arch/powerpc/kernel/entry_64.S
+++ b/arch/powerpc/kernel/entry_64.S
@@ -115,6 +115,11 @@ END_FW_FTR_SECTION_IFSET(FW_FEATURE_ISERIES)
 	addi	r9,r1,STACK_FRAME_OVERHEAD
 #endif
 	clrrdi	r11,r1,THREAD_SHIFT
+#ifdef CONFIG_TASKSTATS
+	ld r10,TI_SYSCALL_COUNT(r11)
+	addi r10,r10,1
+	std r10,TI_SYSCALL_COUNT(r11)
+#endif
 	ld	r10,TI_FLAGS(r11)
 	andi.	r11,r10,_TIF_SYSCALL_T_OR_A
 	bne-	syscall_dotrace
diff --git a/arch/x86_64/kernel/asm-offsets.c b/arch/x86_64/kernel/asm-offsets.c
index 96687e2..da57356 100644
--- a/arch/x86_64/kernel/asm-offsets.c
+++ b/arch/x86_64/kernel/asm-offsets.c
@@ -35,6 +35,7 @@ int main(void)
 	ENTRY(addr_limit);
 	ENTRY(preempt_count);
 	ENTRY(status);
+	ENTRY(syscall_count);
 	BLANK();
 #undef ENTRY
 #define ENTRY(entry) DEFINE(pda_ ## entry, offsetof(struct x8664_pda, entry))
diff --git a/arch/x86_64/kernel/entry.S b/arch/x86_64/kernel/entry.S
index 9f5dac6..9fd97df 100644
--- a/arch/x86_64/kernel/entry.S
+++ b/arch/x86_64/kernel/entry.S
@@ -229,6 +229,7 @@ ENTRY(system_call)
 	movq  %rcx,RIP-ARGOFFSET(%rsp)
 	CFI_REL_OFFSET rip,RIP-ARGOFFSET
 	GET_THREAD_INFO(%rcx)
+	addq $1, threadinfo_syscall_count(%rcx)   # Increment syscalls counter
 	testl $(_TIF_SYSCALL_TRACE|_TIF_SYSCALL_AUDIT|_TIF_SECCOMP),threadinfo_flags(%rcx)
 	jnz tracesys
 	cmpq $__NR_syscall_max,%rax
diff --git a/fs/proc/array.c b/fs/proc/array.c
index 70e4fab..30f6fd5 100644
--- a/fs/proc/array.c
+++ b/fs/proc/array.c
@@ -290,6 +290,19 @@ static inline char *task_cap(struct task_struct *p, char *buffer)
 			    cap_t(p->cap_permitted),
 			    cap_t(p->cap_effective));
 }
+static inline char *task_perf(struct task_struct *p, char *buffer)
+{
+	/* Syscall counter adds 1 line overhead on each syscall execution
+	 * in entry.S, so probably it is the leave this stuff under ifdefs.
+	 */
+#ifdef CONFIG_TASKSTATS
+	buffer += sprintf(buffer, "Syscalls:\t%lu\n", p->thread_info->syscall_count);
+#endif
+	return buffer + sprintf(buffer, "Nvcsw:\t%lu\n"
+                           "Nivcsw:\t%lu\n",
+                           p->nvcsw,
+                           p->nivcsw);
+}
 
 int proc_pid_status(struct task_struct *task, char * buffer)
 {
@@ -309,6 +322,7 @@ int proc_pid_status(struct task_struct *task, char * buffer)
 #if defined(CONFIG_S390)
 	buffer = task_show_regs(task, buffer);
 #endif
+	buffer = task_perf(task, buffer);
 	return buffer - orig;
 }
 
diff --git a/include/asm-i386/thread_info.h b/include/asm-i386/thread_info.h
index 4b187bb..bccfd6a 100644
--- a/include/asm-i386/thread_info.h
+++ b/include/asm-i386/thread_info.h
@@ -33,6 +33,7 @@ struct thread_info {
 	int			preempt_count;	/* 0 => preemptable, <0 => BUG */
 
 
+	unsigned long		syscall_count;	/* Syscall counter */
 	mm_segment_t		addr_limit;	/* thread address space:
 					 	   0-0xBFFFFFFF for user-thead
 						   0-0xFFFFFFFF for kernel-thread
diff --git a/include/asm-powerpc/thread_info.h b/include/asm-powerpc/thread_info.h
index 3f32ca8..5306ac2 100644
--- a/include/asm-powerpc/thread_info.h
+++ b/include/asm-powerpc/thread_info.h
@@ -35,6 +35,7 @@ struct thread_info {
 	int		cpu;			/* cpu we're on */
 	int		preempt_count;		/* 0 => preemptable,
 						   <0 => BUG */
+	unsigned long	syscall_count;		/* Syscall counter */
 	struct restart_block restart_block;
 	unsigned long	local_flags;		/* private flags for thread */
 
diff --git a/include/asm-x86_64/thread_info.h b/include/asm-x86_64/thread_info.h
index 74a6c74..e53022d 100644
--- a/include/asm-x86_64/thread_info.h
+++ b/include/asm-x86_64/thread_info.h
@@ -31,6 +31,7 @@ struct thread_info {
 	__u32			cpu;		/* current CPU */
 	int 			preempt_count;	/* 0 => preemptable, <0 => BUG */
 
+	unsigned long		syscall_count;	/* Syscall counter */
 	mm_segment_t		addr_limit;	
 	struct restart_block    restart_block;
 };
diff --git a/include/linux/taskstats.h b/include/linux/taskstats.h
index 3fced47..e3341b6 100644
--- a/include/linux/taskstats.h
+++ b/include/linux/taskstats.h
@@ -31,7 +31,7 @@
  */
 
 
-#define TASKSTATS_VERSION	3
+#define TASKSTATS_VERSION	4
 #define TS_COMM_LEN		32	/* should be >= TASK_COMM_LEN
 					 * in linux/sched.h */
 
@@ -141,6 +141,10 @@ struct taskstats {
 	__u64	write_syscalls;		/* write syscalls */
 	/* Extended accounting fields end */
 
+	__u32  syscall_counter;   /* Syscall counter */
+	__u32  nvcsw;
+	__u32  nivcsw;
+
 #define TASKSTATS_HAS_IO_ACCOUNTING
 	/* Per-task storage I/O accounting starts */
 	__u64	read_bytes;		/* bytes of read I/O */
diff --git a/kernel/fork.c b/kernel/fork.c
index fc723e5..5213738 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -1042,6 +1042,9 @@ static struct task_struct *copy_process(unsigned long clone_flags,
 	p->wchar = 0;		/* I/O counter: bytes written */
 	p->syscr = 0;		/* I/O counter: read syscalls */
 	p->syscw = 0;		/* I/O counter: write syscalls */
+#ifdef CONFIG_TASKSTATS
+	p->thread_info->syscall_count = 0; /* Syscall counter: total numbers of syscalls */
+#endif
 	task_io_accounting_init(p);
 	acct_clear_integrals(p);
 
diff --git a/kernel/taskstats.c b/kernel/taskstats.c
index 4c3476f..d7bf33f 100644
--- a/kernel/taskstats.c
+++ b/kernel/taskstats.c
@@ -196,6 +196,9 @@ static int fill_pid(pid_t pid, struct task_struct *tsk,
 
 	/* fill in basic acct fields */
 	stats->version = TASKSTATS_VERSION;
+	stats->syscall_counter = tsk->thread_info->syscall_count;
+	stats->nvcsw = tsk->nvcsw;
+	stats->nivcsw = tsk->nivcsw;
 	bacct_add_tsk(stats, tsk);
 
 	/* fill in extended acct fields */
@@ -242,6 +245,9 @@ static int fill_tgid(pid_t tgid, struct task_struct *first,
 		 */
 		delayacct_add_tsk(stats, tsk);
 
+		stats->syscall_counter += tsk->thread_info->syscall_count;
+		stats->nvcsw += tsk->nvcsw;
+		stats->nivcsw += tsk->nivcsw;
 	} while_each_thread(first, tsk);
 
 	unlock_task_sighand(first, &flags);

^ permalink raw reply related	[flat|nested] 22+ messages in thread
* Re: [PATCH] Performance Stats: Kernel patch
  2007-05-08 16:26 [PATCH] Performance Stats: Kernel patch Maxim Uvarov
@ 2007-05-08 19:32 ` Andrew Morton
  2007-05-10 11:11   ` Maxim Uvarov
  2007-05-08 23:32 ` Linas Vepstas
  1 sibling, 1 reply; 22+ messages in thread
From: Andrew Morton @ 2007-05-08 19:32 UTC (permalink / raw)
  To: Maxim Uvarov; +Cc: linuxppc-dev, wli, pavel, dada1

On Tue, 08 May 2007 16:26:51 +0000
Maxim Uvarov <muvarov@ru.mvista.com> wrote:

> From: Maxim Uvarov <muvarov@ru.mvista.com>
> To: Andrew Morton <akpm@linux-foundation.org>
> Cc: pavel@ucw.cz
> Cc: linuxppc-dev@ozlabs.org
> Cc: wli@holomorphy.com
> Cc: dada1@cosmosbay.com
> Cc: pavel@ucw.cz

Strange that linux-ppc was the only mailing list you copied?

> Subject: [PATCH] Performance Stats: Kernel patch
> Date: Tue, 08 May 2007 16:26:51 +0000
> User-Agent: StGIT/0.12.1
> 
> 
> Patch makes available to the user the following
> task and process performance statistics:
> 	* Involuntary Context Switches (task_struct->nivcsw)
> 	* Voluntary Context Switches (task_struct->nvcsw)
> 	* Number of system calls (added new counter
> 	  thread_info->sysall_count)
> 	           
> Statistics information is available from:
>         1. taskstats interface (Documentation/accounting/)
> 	2. /proc/PID/status (task only).
> 
> This data is useful for detecting hyperactivity
> patterns between processes.

syscall entry is a super-hotpath and there will be a developer revolt
if we go adding an incl to that path, sorry.

(Did that incl also cover the int 80 entry?)

^ permalink raw reply	[flat|nested] 22+ messages in thread
* Re: [PATCH] Performance Stats: Kernel patch
  2007-05-08 16:26 [PATCH] Performance Stats: Kernel patch Maxim Uvarov
  2007-05-08 19:32 ` Andrew Morton
@ 2007-05-08 23:32 ` Linas Vepstas
  2007-05-10 10:22   ` Maxim Uvarov
  1 sibling, 1 reply; 22+ messages in thread
From: Linas Vepstas @ 2007-05-08 23:32 UTC (permalink / raw)
  To: Maxim Uvarov; +Cc: linuxppc-dev, Andrew Morton, pavel, wli, dada1


Hi,

On Tue, May 08, 2007 at 04:26:51PM +0000, Maxim Uvarov wrote:
> 
> Patch makes available to the user the following
> task and process performance statistics:

[...]

> diff --git a/arch/i386/kernel/entry.S b/arch/i386/kernel/entry.S
> index 5e47683..26f0cc0 100644
> --- a/arch/i386/kernel/entry.S
> +++ b/arch/i386/kernel/entry.S
> @@ -331,6 +331,7 @@ sysenter_past_esp:
>  	CFI_ADJUST_CFA_OFFSET 4
>  	SAVE_ALL
>  	GET_THREAD_INFO(%ebp)
> +	incl	TI_syscall_count(%ebp)	# Increment syscalls counter

Other arches have this protected with #ifdef CONFIG_TASKSTATS
why not here?

> diff --git a/arch/powerpc/kernel/entry_32.S b/arch/powerpc/kernel/entry_32.S
> index c03e829..329c2f8 100644
> --- a/arch/powerpc/kernel/entry_32.S
> +++ b/arch/powerpc/kernel/entry_32.S
> @@ -202,6 +202,11 @@ _GLOBAL(DoSyscall)
>  	bl	do_show_syscall
>  #endif /* SHOW_SYSCALLS */
>  	rlwinm	r10,r1,0,0,(31-THREAD_SHIFT)	/* current_thread_info() */
> +#ifdef CONFIG_THREAD_PERF_STAT_SYSC
> +	lwz r11,TI_SYSC_CNT(r10)
> +	addi r11,r11,1
> +	stw r11,TI_SYSC_CNT(r10)
> +#endif /* CONFIG_THREAD_PERF_STAT_SYSC */

Why not CONFIG_TASKSTATS, as in entry_64.S ?

Actually, grep shows that CONFIG_THREAD_PERF_STAT_SYSC is not defined
anywhere.

> diff --git a/arch/x86_64/kernel/entry.S b/arch/x86_64/kernel/entry.S
> index 9f5dac6..9fd97df 100644
> --- a/arch/x86_64/kernel/entry.S
> +++ b/arch/x86_64/kernel/entry.S
> @@ -229,6 +229,7 @@ ENTRY(system_call)
>  	movq  %rcx,RIP-ARGOFFSET(%rsp)
>  	CFI_REL_OFFSET rip,RIP-ARGOFFSET
>  	GET_THREAD_INFO(%rcx)
> +	addq $1, threadinfo_syscall_count(%rcx)   # Increment syscalls counter

again, #ifdef CONFIG_TASKSTATS,

--linas

^ permalink raw reply	[flat|nested] 22+ messages in thread
* Re: [PATCH] Performance Stats: Kernel patch
  2007-05-08 23:32 ` Linas Vepstas
@ 2007-05-10 10:22   ` Maxim Uvarov
  2007-05-10 16:47     ` Linas Vepstas
  0 siblings, 1 reply; 22+ messages in thread
From: Maxim Uvarov @ 2007-05-10 10:22 UTC (permalink / raw)
  To: Linas Vepstas; +Cc: linuxppc-dev, Andrew Morton, pavel, wli, dada1

Linas Vepstas wrote:

>Hi,
>
>On Tue, May 08, 2007 at 04:26:51PM +0000, Maxim Uvarov wrote:
>  
>
>>Patch makes available to the user the following
>>task and process performance statistics:
>>    
>>
>
>[...]
>
>  
>
>>diff --git a/arch/i386/kernel/entry.S b/arch/i386/kernel/entry.S
>>index 5e47683..26f0cc0 100644
>>--- a/arch/i386/kernel/entry.S
>>+++ b/arch/i386/kernel/entry.S
>>@@ -331,6 +331,7 @@ sysenter_past_esp:
>> 	CFI_ADJUST_CFA_OFFSET 4
>> 	SAVE_ALL
>> 	GET_THREAD_INFO(%ebp)
>>+	incl	TI_syscall_count(%ebp)	# Increment syscalls counter
>>    
>>
>
>Other arches have this protected with #ifdef CONFIG_TASKSTATS
>why not here?
>
>  
>
Thank you. It should be protected.

>>diff --git a/arch/powerpc/kernel/entry_32.S b/arch/powerpc/kernel/entry_32.S
>>index c03e829..329c2f8 100644
>>--- a/arch/powerpc/kernel/entry_32.S
>>+++ b/arch/powerpc/kernel/entry_32.S
>>@@ -202,6 +202,11 @@ _GLOBAL(DoSyscall)
>> 	bl	do_show_syscall
>> #endif /* SHOW_SYSCALLS */
>> 	rlwinm	r10,r1,0,0,(31-THREAD_SHIFT)	/* current_thread_info() */
>>+#ifdef CONFIG_THREAD_PERF_STAT_SYSC
>>+	lwz r11,TI_SYSC_CNT(r10)
>>+	addi r11,r11,1
>>+	stw r11,TI_SYSC_CNT(r10)
>>+#endif /* CONFIG_THREAD_PERF_STAT_SYSC */
>>    
>>
>
>Why not CONFIG_TASKSTATS, as in entry_64.S ?
>
>Actually, grep shows that CONFIG_THREAD_PERF_STAT_SYSC is not defined
>anywhere.
>
>  
>
This is mistake, thank you. I have to be more considerate.

>>diff --git a/arch/x86_64/kernel/entry.S b/arch/x86_64/kernel/entry.S
>>index 9f5dac6..9fd97df 100644
>>--- a/arch/x86_64/kernel/entry.S
>>+++ b/arch/x86_64/kernel/entry.S
>>@@ -229,6 +229,7 @@ ENTRY(system_call)
>> 	movq  %rcx,RIP-ARGOFFSET(%rsp)
>> 	CFI_REL_OFFSET rip,RIP-ARGOFFSET
>> 	GET_THREAD_INFO(%rcx)
>>+	addq $1, threadinfo_syscall_count(%rcx)   # Increment syscalls counter
>>    
>>
>
>again, #ifdef CONFIG_TASKSTATS,
>
>--linas
>  
>

In general I had the following idea. Defend with CONFIG_TASKSTATS all
syscall increment counters in  arch/*/entry.S  code.  It will allow to
avoid this overhead if task stats interface is not supported by kernel.
The overhead is too small but,  it is not needed to increment counters 
without interface
to get them.

In first version of patch I have used only /proc interface. And it is 
available now.
So I don't really sure if  defines are needed in  entry.S code. I tried 
to do
some tests and  it is impossible to catch the performance difference (small
overhead).

^ permalink raw reply	[flat|nested] 22+ messages in thread
* Re: [PATCH] Performance Stats: Kernel patch
  2007-05-08 19:32 ` Andrew Morton
@ 2007-05-10 11:11   ` Maxim Uvarov
  2007-05-10 18:12     ` Andrew Morton
  0 siblings, 1 reply; 22+ messages in thread
From: Maxim Uvarov @ 2007-05-10 11:11 UTC (permalink / raw)
  To: Andrew Morton; +Cc: linuxppc-dev, wli, pavel, dada1

Andrew Morton wrote:

>On Tue, 08 May 2007 16:26:51 +0000
>Maxim Uvarov <muvarov@ru.mvista.com> wrote:
>
>  
>
>>From: Maxim Uvarov <muvarov@ru.mvista.com>
>>To: Andrew Morton <akpm@linux-foundation.org>
>>Cc: pavel@ucw.cz
>>Cc: linuxppc-dev@ozlabs.org
>>Cc: wli@holomorphy.com
>>Cc: dada1@cosmosbay.com
>>Cc: pavel@ucw.cz
>>    
>>
>
>Strange that linux-ppc was the only mailing list you copied?
>
>  
>
Very strange. Possible stgit mail have limitation of CC.

>>Subject: [PATCH] Performance Stats: Kernel patch
>>Date: Tue, 08 May 2007 16:26:51 +0000
>>User-Agent: StGIT/0.12.1
>>
>>
>>Patch makes available to the user the following
>>task and process performance statistics:
>>	* Involuntary Context Switches (task_struct->nivcsw)
>>	* Voluntary Context Switches (task_struct->nvcsw)
>>	* Number of system calls (added new counter
>>	  thread_info->sysall_count)
>>	           
>>Statistics information is available from:
>>        1. taskstats interface (Documentation/accounting/)
>>	2. /proc/PID/status (task only).
>>
>>This data is useful for detecting hyperactivity
>>patterns between processes.
>>    
>>
>
>syscall entry is a super-hotpath and there will be a developer revolt
>if we go adding an incl to that path, sorry.
>
>  
>
Yes, I know that it is very critical part of code. May be protection with
ifdef CONFIG_TASKSTATS  in all entry.S can be solution?
In that case we can be sure that nothing wrong can not happen,
and if this functionality is needed we can turn on this.

>(Did that incl also cover the int 80 entry?)
>  
>
Actually not. But counter increments each time when syscall  calls.
But it is in very save place, just after
        SAVE_ALL
        GET_THREAD_INFO(%ebp)
So it is very hard to break something.

(I'm going to update patch.)

^ permalink raw reply	[flat|nested] 22+ messages in thread
* [PATCH] Performance Stats: Kernel patch
@ 2007-05-10 11:42 Maxim Uvarov
  2007-05-10 12:03 ` Pavel Machek
  2007-05-10 18:25 ` Andrew Morton
  0 siblings, 2 replies; 22+ messages in thread
From: Maxim Uvarov @ 2007-05-10 11:42 UTC (permalink / raw)
  To: Andrew Morton; +Cc: linuxppc-dev, pavel, wli, dada1


Patch makes available to the user the following
task and process performance statistics:
	* Involuntary Context Switches (task_struct->nivcsw)
	* Voluntary Context Switches (task_struct->nvcsw)
	* Number of system calls (added new counter
	  thread_info->sysall_count)
	           
Statistics information is available from:
        1. taskstats interface (Documentation/accounting/)
	2. /proc/PID/status (task only).

This data is useful for detecting hyperactivity
patterns between processes.

Signed-off-by: Maxim Uvarov <muvarov@ru.mvista.com>

Signed-off-by: Maxim Uvarov <muvarov@ru.mvista.com>                                                                                  
---

 Documentation/accounting/getdelays.c          |   20 ++++++++++++++++++--
 Documentation/accounting/taskstats-struct.txt |    7 +++++++
 arch/i386/kernel/asm-offsets.c                |    1 +
 arch/i386/kernel/entry.S                      |    3 +++
 arch/powerpc/kernel/asm-offsets.c             |    2 ++
 arch/powerpc/kernel/entry_32.S                |    5 +++++
 arch/powerpc/kernel/entry_64.S                |    5 +++++
 arch/x86_64/kernel/asm-offsets.c              |    1 +
 arch/x86_64/kernel/entry.S                    |    3 +++
 fs/proc/array.c                               |   14 ++++++++++++++
 include/asm-i386/thread_info.h                |    1 +
 include/asm-powerpc/thread_info.h             |    1 +
 include/asm-x86_64/thread_info.h              |    1 +
 include/linux/taskstats.h                     |    6 +++++-
 kernel/fork.c                                 |    3 +++
 kernel/taskstats.c                            |    6 ++++++
 16 files changed, 76 insertions(+), 3 deletions(-)

diff --git a/Documentation/accounting/getdelays.c b/Documentation/accounting/getdelays.c
index e9126e7..1be7d65 100644
--- a/Documentation/accounting/getdelays.c
+++ b/Documentation/accounting/getdelays.c
@@ -49,6 +49,7 @@ char name[100];
 int dbg;
 int print_delays;
 int print_io_accounting;
+int print_task_stats;
 __u64 stime, utime;
 
 #define PRINTF(fmt, arg...) {			\
@@ -187,7 +188,7 @@ void print_delayacct(struct taskstats *t)
 	       "IO    %15s%15s\n"
 	       "      %15llu%15llu\n"
 	       "MEM   %15s%15s\n"
-	       "      %15llu%15llu\n\n",
+	       "      %15llu%15llu\n"
 	       "count", "real total", "virtual total", "delay total",
 	       t->cpu_count, t->cpu_run_real_total, t->cpu_run_virtual_total,
 	       t->cpu_delay_total,
@@ -196,6 +197,15 @@ void print_delayacct(struct taskstats *t)
 	       "count", "delay total", t->swapin_count, t->swapin_delay_total);
 }
 
+void print_taskstats(struct taskstats *t)
+{
+	printf("\n\nTask  %15s%15s%15s\n"
+	       "      %15lu%15lu%15lu\n",
+	       "syscalls", "voluntary", "nonvoluntary",
+	       t->syscall_counter, t->nvcsw, t->nivcsw);
+
+}
+
 void print_ioacct(struct taskstats *t)
 {
 	printf("%s: read=%llu, write=%llu, cancelled_write=%llu\n",
@@ -227,7 +237,7 @@ int main(int argc, char *argv[])
 	struct msgtemplate msg;
 
 	while (1) {
-		c = getopt(argc, argv, "diw:r:m:t:p:v:l");
+		c = getopt(argc, argv, "qdiw:r:m:t:p:v:l");
 		if (c < 0)
 			break;
 
@@ -240,6 +250,10 @@ int main(int argc, char *argv[])
 			printf("printing IO accounting\n");
 			print_io_accounting = 1;
 			break;
+		case 'q':
+			printf("printing task/process stasistics:\n");
+			print_task_stats = 1;
+			break;
 		case 'w':
 			strncpy(logfile, optarg, MAX_FILENAME);
 			printf("write to file %s\n", logfile);
@@ -381,6 +395,8 @@ int main(int argc, char *argv[])
 							print_delayacct((struct taskstats *) NLA_DATA(na));
 						if (print_io_accounting)
 							print_ioacct((struct taskstats *) NLA_DATA(na));
+						if (print_task_stats)
+							print_taskstats((struct taskstats *) NLA_DATA(na));
 						if (fd) {
 							if (write(fd, NLA_DATA(na), na->nla_len) < 0) {
 								err(1,"write error\n");
diff --git a/Documentation/accounting/taskstats-struct.txt b/Documentation/accounting/taskstats-struct.txt
index 661c797..5dac173 100644
--- a/Documentation/accounting/taskstats-struct.txt
+++ b/Documentation/accounting/taskstats-struct.txt
@@ -22,6 +22,8 @@ There are three different groups of fields in the struct taskstats:
     /* Extended accounting fields end */
     Their values are collected if CONFIG_TASK_XACCT is set.
 
+4) Per-task and per-thread statistics
+
 Future extension should add fields to the end of the taskstats struct, and
 should not change the relative position of each field within the struct.
 
@@ -158,4 +160,9 @@ struct taskstats {
 
 	/* Extended accounting fields end */
 
+4) Per-task and per-thread statiscits
+	__u32	syscall_counter;	/* Syscall counter */
+	__u32	nvcsw;			/* Context voluntary switch counter */
+	__u32	nivcsw;			/* Context involuntary switch counter */
+
 }
diff --git a/arch/i386/kernel/asm-offsets.c b/arch/i386/kernel/asm-offsets.c
index 1b2f3cd..4ad49d2 100644
--- a/arch/i386/kernel/asm-offsets.c
+++ b/arch/i386/kernel/asm-offsets.c
@@ -56,6 +56,7 @@ void foo(void)
 	OFFSET(TI_addr_limit, thread_info, addr_limit);
 	OFFSET(TI_restart_block, thread_info, restart_block);
 	OFFSET(TI_sysenter_return, thread_info, sysenter_return);
+	OFFSET(TI_syscall_count, thread_info, syscall_count);
 	BLANK();
 
 	OFFSET(GDS_size, Xgt_desc_struct, size);
diff --git a/arch/i386/kernel/entry.S b/arch/i386/kernel/entry.S
index 5e47683..836961f 100644
--- a/arch/i386/kernel/entry.S
+++ b/arch/i386/kernel/entry.S
@@ -332,6 +332,9 @@ sysenter_past_esp:
 	SAVE_ALL
 	GET_THREAD_INFO(%ebp)
 
+#ifdef CONFIG_TASKSTATS
+	incl	TI_syscall_count(%ebp)	# Increment syscalls counter
+#endif
 	/* Note, _TIF_SECCOMP is bit number 8, and so it needs testw and not testb */
 	testw $(_TIF_SYSCALL_EMU|_TIF_SYSCALL_TRACE|_TIF_SECCOMP|_TIF_SYSCALL_AUDIT),TI_flags(%ebp)
 	jnz syscall_trace_entry
diff --git a/arch/powerpc/kernel/asm-offsets.c b/arch/powerpc/kernel/asm-offsets.c
index 030d300..b640039 100644
--- a/arch/powerpc/kernel/asm-offsets.c
+++ b/arch/powerpc/kernel/asm-offsets.c
@@ -94,6 +94,8 @@ int main(void)
 	DEFINE(TI_LOCAL_FLAGS, offsetof(struct thread_info, local_flags));
 	DEFINE(TI_PREEMPT, offsetof(struct thread_info, preempt_count));
 	DEFINE(TI_TASK, offsetof(struct thread_info, task));
+	DEFINE(TI_SYSCALL_COUNT, offsetof(struct thread_info, syscall_count));
+
 #ifdef CONFIG_PPC32
 	DEFINE(TI_EXECDOMAIN, offsetof(struct thread_info, exec_domain));
 	DEFINE(TI_CPU, offsetof(struct thread_info, cpu));
diff --git a/arch/powerpc/kernel/entry_32.S b/arch/powerpc/kernel/entry_32.S
index c03e829..5d919e4 100644
--- a/arch/powerpc/kernel/entry_32.S
+++ b/arch/powerpc/kernel/entry_32.S
@@ -202,6 +202,11 @@ _GLOBAL(DoSyscall)
 	bl	do_show_syscall
 #endif /* SHOW_SYSCALLS */
 	rlwinm	r10,r1,0,0,(31-THREAD_SHIFT)	/* current_thread_info() */
+#ifdef CONFIG_TASKSTATS 
+	lwz r11,TI_SYSC_CNT(r10)
+	addi r11,r11,1
+	stw r11,TI_SYSC_CNT(r10)
+#endif
 	lwz	r11,TI_FLAGS(r10)
 	andi.	r11,r11,_TIF_SYSCALL_T_OR_A
 	bne-	syscall_dotrace
diff --git a/arch/powerpc/kernel/entry_64.S b/arch/powerpc/kernel/entry_64.S
index 2551c08..5907f76 100644
--- a/arch/powerpc/kernel/entry_64.S
+++ b/arch/powerpc/kernel/entry_64.S
@@ -115,6 +115,11 @@ END_FW_FTR_SECTION_IFSET(FW_FEATURE_ISERIES)
 	addi	r9,r1,STACK_FRAME_OVERHEAD
 #endif
 	clrrdi	r11,r1,THREAD_SHIFT
+#ifdef CONFIG_TASKSTATS
+	ld r10,TI_SYSCALL_COUNT(r11)
+	addi r10,r10,1
+	std r10,TI_SYSCALL_COUNT(r11)
+#endif
 	ld	r10,TI_FLAGS(r11)
 	andi.	r11,r10,_TIF_SYSCALL_T_OR_A
 	bne-	syscall_dotrace
diff --git a/arch/x86_64/kernel/asm-offsets.c b/arch/x86_64/kernel/asm-offsets.c
index 96687e2..da57356 100644
--- a/arch/x86_64/kernel/asm-offsets.c
+++ b/arch/x86_64/kernel/asm-offsets.c
@@ -35,6 +35,7 @@ int main(void)
 	ENTRY(addr_limit);
 	ENTRY(preempt_count);
 	ENTRY(status);
+	ENTRY(syscall_count);
 	BLANK();
 #undef ENTRY
 #define ENTRY(entry) DEFINE(pda_ ## entry, offsetof(struct x8664_pda, entry))
diff --git a/arch/x86_64/kernel/entry.S b/arch/x86_64/kernel/entry.S
index 9f5dac6..af40ead 100644
--- a/arch/x86_64/kernel/entry.S
+++ b/arch/x86_64/kernel/entry.S
@@ -229,6 +229,9 @@ ENTRY(system_call)
 	movq  %rcx,RIP-ARGOFFSET(%rsp)
 	CFI_REL_OFFSET rip,RIP-ARGOFFSET
 	GET_THREAD_INFO(%rcx)
+#ifdef CONFIG_TASKSTATS
+	addq $1, threadinfo_syscall_count(%rcx)   # Increment syscalls counter
+#endif
 	testl $(_TIF_SYSCALL_TRACE|_TIF_SYSCALL_AUDIT|_TIF_SECCOMP),threadinfo_flags(%rcx)
 	jnz tracesys
 	cmpq $__NR_syscall_max,%rax
diff --git a/fs/proc/array.c b/fs/proc/array.c
index 70e4fab..30f6fd5 100644
--- a/fs/proc/array.c
+++ b/fs/proc/array.c
@@ -290,6 +290,19 @@ static inline char *task_cap(struct task_struct *p, char *buffer)
 			    cap_t(p->cap_permitted),
 			    cap_t(p->cap_effective));
 }
+static inline char *task_perf(struct task_struct *p, char *buffer)
+{
+	/* Syscall counter adds 1 line overhead on each syscall execution
+	 * in entry.S, so probably it is the leave this stuff under ifdefs.
+	 */
+#ifdef CONFIG_TASKSTATS
+	buffer += sprintf(buffer, "Syscalls:\t%lu\n", p->thread_info->syscall_count);
+#endif
+	return buffer + sprintf(buffer, "Nvcsw:\t%lu\n"
+                           "Nivcsw:\t%lu\n",
+                           p->nvcsw,
+                           p->nivcsw);
+}
 
 int proc_pid_status(struct task_struct *task, char * buffer)
 {
@@ -309,6 +322,7 @@ int proc_pid_status(struct task_struct *task, char * buffer)
 #if defined(CONFIG_S390)
 	buffer = task_show_regs(task, buffer);
 #endif
+	buffer = task_perf(task, buffer);
 	return buffer - orig;
 }
 
diff --git a/include/asm-i386/thread_info.h b/include/asm-i386/thread_info.h
index 4b187bb..bccfd6a 100644
--- a/include/asm-i386/thread_info.h
+++ b/include/asm-i386/thread_info.h
@@ -33,6 +33,7 @@ struct thread_info {
 	int			preempt_count;	/* 0 => preemptable, <0 => BUG */
 
 
+	unsigned long		syscall_count;	/* Syscall counter */
 	mm_segment_t		addr_limit;	/* thread address space:
 					 	   0-0xBFFFFFFF for user-thead
 						   0-0xFFFFFFFF for kernel-thread
diff --git a/include/asm-powerpc/thread_info.h b/include/asm-powerpc/thread_info.h
index 3f32ca8..5306ac2 100644
--- a/include/asm-powerpc/thread_info.h
+++ b/include/asm-powerpc/thread_info.h
@@ -35,6 +35,7 @@ struct thread_info {
 	int		cpu;			/* cpu we're on */
 	int		preempt_count;		/* 0 => preemptable,
 						   <0 => BUG */
+	unsigned long	syscall_count;		/* Syscall counter */
 	struct restart_block restart_block;
 	unsigned long	local_flags;		/* private flags for thread */
 
diff --git a/include/asm-x86_64/thread_info.h b/include/asm-x86_64/thread_info.h
index 74a6c74..e53022d 100644
--- a/include/asm-x86_64/thread_info.h
+++ b/include/asm-x86_64/thread_info.h
@@ -31,6 +31,7 @@ struct thread_info {
 	__u32			cpu;		/* current CPU */
 	int 			preempt_count;	/* 0 => preemptable, <0 => BUG */
 
+	unsigned long		syscall_count;	/* Syscall counter */
 	mm_segment_t		addr_limit;	
 	struct restart_block    restart_block;
 };
diff --git a/include/linux/taskstats.h b/include/linux/taskstats.h
index 3fced47..e3341b6 100644
--- a/include/linux/taskstats.h
+++ b/include/linux/taskstats.h
@@ -31,7 +31,7 @@
  */
 
 
-#define TASKSTATS_VERSION	3
+#define TASKSTATS_VERSION	4
 #define TS_COMM_LEN		32	/* should be >= TASK_COMM_LEN
 					 * in linux/sched.h */
 
@@ -141,6 +141,10 @@ struct taskstats {
 	__u64	write_syscalls;		/* write syscalls */
 	/* Extended accounting fields end */
 
+	__u32  syscall_counter;   /* Syscall counter */
+	__u32  nvcsw;
+	__u32  nivcsw;
+
 #define TASKSTATS_HAS_IO_ACCOUNTING
 	/* Per-task storage I/O accounting starts */
 	__u64	read_bytes;		/* bytes of read I/O */
diff --git a/kernel/fork.c b/kernel/fork.c
index fc723e5..5213738 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -1042,6 +1042,9 @@ static struct task_struct *copy_process(unsigned long clone_flags,
 	p->wchar = 0;		/* I/O counter: bytes written */
 	p->syscr = 0;		/* I/O counter: read syscalls */
 	p->syscw = 0;		/* I/O counter: write syscalls */
+#ifdef CONFIG_TASKSTATS
+	p->thread_info->syscall_count = 0; /* Syscall counter: total numbers of syscalls */
+#endif
 	task_io_accounting_init(p);
 	acct_clear_integrals(p);
 
diff --git a/kernel/taskstats.c b/kernel/taskstats.c
index 4c3476f..d7bf33f 100644
--- a/kernel/taskstats.c
+++ b/kernel/taskstats.c
@@ -196,6 +196,9 @@ static int fill_pid(pid_t pid, struct task_struct *tsk,
 
 	/* fill in basic acct fields */
 	stats->version = TASKSTATS_VERSION;
+	stats->syscall_counter = tsk->thread_info->syscall_count;
+	stats->nvcsw = tsk->nvcsw;
+	stats->nivcsw = tsk->nivcsw;
 	bacct_add_tsk(stats, tsk);
 
 	/* fill in extended acct fields */
@@ -242,6 +245,9 @@ static int fill_tgid(pid_t tgid, struct task_struct *first,
 		 */
 		delayacct_add_tsk(stats, tsk);
 
+		stats->syscall_counter += tsk->thread_info->syscall_count;
+		stats->nvcsw += tsk->nvcsw;
+		stats->nivcsw += tsk->nivcsw;
 	} while_each_thread(first, tsk);
 
 	unlock_task_sighand(first, &flags);

^ permalink raw reply related	[flat|nested] 22+ messages in thread
* Re: [PATCH] Performance Stats: Kernel patch
  2007-05-10 11:42 Maxim Uvarov
@ 2007-05-10 12:03 ` Pavel Machek
  2007-05-10 18:25 ` Andrew Morton
  1 sibling, 0 replies; 22+ messages in thread
From: Pavel Machek @ 2007-05-10 12:03 UTC (permalink / raw)
  To: Maxim Uvarov; +Cc: linuxppc-dev, Andrew Morton, wli, dada1

Hi!

> Patch makes available to the user the following
> task and process performance statistics:
> 	* Involuntary Context Switches (task_struct->nivcsw)
> 	* Voluntary Context Switches (task_struct->nvcsw)
> 	* Number of system calls (added new counter
> 	  thread_info->sysall_count)
> 	           
> Statistics information is available from:
>         1. taskstats interface (Documentation/accounting/)
> 	2. /proc/PID/status (task only).
> 
> This data is useful for detecting hyperactivity
> patterns between processes.

I was told nvcsw is standard name, but given how other fields are
named ("syscall_counter") I think we should use some more reasonable
name (like "voluntary_ctxt_switches" or something).
								Pavel

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

^ permalink raw reply	[flat|nested] 22+ messages in thread
* Re: [PATCH] Performance Stats: Kernel patch
  2007-05-10 12:39 Maxim Uvarov
@ 2007-05-10 12:38 ` Josh Boyer
  2007-05-10 18:23   ` Maxim Uvarov
  0 siblings, 1 reply; 22+ messages in thread
From: Josh Boyer @ 2007-05-10 12:38 UTC (permalink / raw)
  To: Maxim Uvarov; +Cc: linuxppc-dev, Andrew Morton, wli, pavel, dada1

On Thu, 2007-05-10 at 12:39 +0000, Maxim Uvarov wrote:
> Patch makes available to the user the following
> task and process performance statistics:
> 	* Involuntary Context Switches (task_struct->nivcsw)
> 	* Voluntary Context Switches (task_struct->nvcsw)
> 	* Number of system calls (added new counter
> 	  thread_info->sysall_count)
> 	           
> Statistics information is available from:
>         1. taskstats interface (Documentation/accounting/)
> 	2. /proc/PID/status (task only).
> 
> This data is useful for detecting hyperactivity
> patterns between processes.

This should be sent to the linux-kernel list I think...  Or at least
CCd.

josh

^ permalink raw reply	[flat|nested] 22+ messages in thread
* [PATCH] Performance Stats: Kernel patch
@ 2007-05-10 12:39 Maxim Uvarov
  2007-05-10 12:38 ` Josh Boyer
  0 siblings, 1 reply; 22+ messages in thread
From: Maxim Uvarov @ 2007-05-10 12:39 UTC (permalink / raw)
  To: Andrew Morton; +Cc: linuxppc-dev, pavel, wli, dada1


Patch makes available to the user the following
task and process performance statistics:
	* Involuntary Context Switches (task_struct->nivcsw)
	* Voluntary Context Switches (task_struct->nvcsw)
	* Number of system calls (added new counter
	  thread_info->sysall_count)
	           
Statistics information is available from:
        1. taskstats interface (Documentation/accounting/)
	2. /proc/PID/status (task only).

This data is useful for detecting hyperactivity
patterns between processes.

Signed-off-by: Maxim Uvarov <muvarov@ru.mvista.com>

Changing nvcsw to voluntary_ctxt_switches.


Signed-off-by: Maxim Uvarov <muvarov@ru.mvista.com>                                                                                  
---

 Documentation/accounting/getdelays.c          |   20 ++++++++++++++++++--
 Documentation/accounting/taskstats-struct.txt |    7 +++++++
 arch/i386/kernel/asm-offsets.c                |    1 +
 arch/i386/kernel/entry.S                      |    3 +++
 arch/powerpc/kernel/asm-offsets.c             |    2 ++
 arch/powerpc/kernel/entry_32.S                |    5 +++++
 arch/powerpc/kernel/entry_64.S                |    5 +++++
 arch/x86_64/kernel/asm-offsets.c              |    1 +
 arch/x86_64/kernel/entry.S                    |    3 +++
 fs/proc/array.c                               |   14 ++++++++++++++
 include/asm-i386/thread_info.h                |    1 +
 include/asm-powerpc/thread_info.h             |    1 +
 include/asm-x86_64/thread_info.h              |    1 +
 include/linux/taskstats.h                     |    6 +++++-
 kernel/fork.c                                 |    3 +++
 kernel/taskstats.c                            |    6 ++++++
 16 files changed, 76 insertions(+), 3 deletions(-)

diff --git a/Documentation/accounting/getdelays.c b/Documentation/accounting/getdelays.c
index e9126e7..1be7d65 100644
--- a/Documentation/accounting/getdelays.c
+++ b/Documentation/accounting/getdelays.c
@@ -49,6 +49,7 @@ char name[100];
 int dbg;
 int print_delays;
 int print_io_accounting;
+int print_task_stats;
 __u64 stime, utime;
 
 #define PRINTF(fmt, arg...) {			\
@@ -187,7 +188,7 @@ void print_delayacct(struct taskstats *t)
 	       "IO    %15s%15s\n"
 	       "      %15llu%15llu\n"
 	       "MEM   %15s%15s\n"
-	       "      %15llu%15llu\n\n",
+	       "      %15llu%15llu\n"
 	       "count", "real total", "virtual total", "delay total",
 	       t->cpu_count, t->cpu_run_real_total, t->cpu_run_virtual_total,
 	       t->cpu_delay_total,
@@ -196,6 +197,15 @@ void print_delayacct(struct taskstats *t)
 	       "count", "delay total", t->swapin_count, t->swapin_delay_total);
 }
 
+void print_taskstats(struct taskstats *t)
+{
+	printf("\n\nTask  %15s%15s%15s\n"
+	       "      %15lu%15lu%15lu\n",
+	       "syscalls", "voluntary", "nonvoluntary",
+	       t->syscall_counter, t->nvcsw, t->nivcsw);
+
+}
+
 void print_ioacct(struct taskstats *t)
 {
 	printf("%s: read=%llu, write=%llu, cancelled_write=%llu\n",
@@ -227,7 +237,7 @@ int main(int argc, char *argv[])
 	struct msgtemplate msg;
 
 	while (1) {
-		c = getopt(argc, argv, "diw:r:m:t:p:v:l");
+		c = getopt(argc, argv, "qdiw:r:m:t:p:v:l");
 		if (c < 0)
 			break;
 
@@ -240,6 +250,10 @@ int main(int argc, char *argv[])
 			printf("printing IO accounting\n");
 			print_io_accounting = 1;
 			break;
+		case 'q':
+			printf("printing task/process stasistics:\n");
+			print_task_stats = 1;
+			break;
 		case 'w':
 			strncpy(logfile, optarg, MAX_FILENAME);
 			printf("write to file %s\n", logfile);
@@ -381,6 +395,8 @@ int main(int argc, char *argv[])
 							print_delayacct((struct taskstats *) NLA_DATA(na));
 						if (print_io_accounting)
 							print_ioacct((struct taskstats *) NLA_DATA(na));
+						if (print_task_stats)
+							print_taskstats((struct taskstats *) NLA_DATA(na));
 						if (fd) {
 							if (write(fd, NLA_DATA(na), na->nla_len) < 0) {
 								err(1,"write error\n");
diff --git a/Documentation/accounting/taskstats-struct.txt b/Documentation/accounting/taskstats-struct.txt
index 661c797..5dac173 100644
--- a/Documentation/accounting/taskstats-struct.txt
+++ b/Documentation/accounting/taskstats-struct.txt
@@ -22,6 +22,8 @@ There are three different groups of fields in the struct taskstats:
     /* Extended accounting fields end */
     Their values are collected if CONFIG_TASK_XACCT is set.
 
+4) Per-task and per-thread statistics
+
 Future extension should add fields to the end of the taskstats struct, and
 should not change the relative position of each field within the struct.
 
@@ -158,4 +160,9 @@ struct taskstats {
 
 	/* Extended accounting fields end */
 
+4) Per-task and per-thread statiscits
+	__u32	syscall_counter;	/* Syscall counter */
+	__u32	nvcsw;			/* Context voluntary switch counter */
+	__u32	nivcsw;			/* Context involuntary switch counter */
+
 }
diff --git a/arch/i386/kernel/asm-offsets.c b/arch/i386/kernel/asm-offsets.c
index 1b2f3cd..4ad49d2 100644
--- a/arch/i386/kernel/asm-offsets.c
+++ b/arch/i386/kernel/asm-offsets.c
@@ -56,6 +56,7 @@ void foo(void)
 	OFFSET(TI_addr_limit, thread_info, addr_limit);
 	OFFSET(TI_restart_block, thread_info, restart_block);
 	OFFSET(TI_sysenter_return, thread_info, sysenter_return);
+	OFFSET(TI_syscall_count, thread_info, syscall_count);
 	BLANK();
 
 	OFFSET(GDS_size, Xgt_desc_struct, size);
diff --git a/arch/i386/kernel/entry.S b/arch/i386/kernel/entry.S
index 5e47683..836961f 100644
--- a/arch/i386/kernel/entry.S
+++ b/arch/i386/kernel/entry.S
@@ -332,6 +332,9 @@ sysenter_past_esp:
 	SAVE_ALL
 	GET_THREAD_INFO(%ebp)
 
+#ifdef CONFIG_TASKSTATS
+	incl	TI_syscall_count(%ebp)	# Increment syscalls counter
+#endif
 	/* Note, _TIF_SECCOMP is bit number 8, and so it needs testw and not testb */
 	testw $(_TIF_SYSCALL_EMU|_TIF_SYSCALL_TRACE|_TIF_SECCOMP|_TIF_SYSCALL_AUDIT),TI_flags(%ebp)
 	jnz syscall_trace_entry
diff --git a/arch/powerpc/kernel/asm-offsets.c b/arch/powerpc/kernel/asm-offsets.c
index 030d300..b640039 100644
--- a/arch/powerpc/kernel/asm-offsets.c
+++ b/arch/powerpc/kernel/asm-offsets.c
@@ -94,6 +94,8 @@ int main(void)
 	DEFINE(TI_LOCAL_FLAGS, offsetof(struct thread_info, local_flags));
 	DEFINE(TI_PREEMPT, offsetof(struct thread_info, preempt_count));
 	DEFINE(TI_TASK, offsetof(struct thread_info, task));
+	DEFINE(TI_SYSCALL_COUNT, offsetof(struct thread_info, syscall_count));
+
 #ifdef CONFIG_PPC32
 	DEFINE(TI_EXECDOMAIN, offsetof(struct thread_info, exec_domain));
 	DEFINE(TI_CPU, offsetof(struct thread_info, cpu));
diff --git a/arch/powerpc/kernel/entry_32.S b/arch/powerpc/kernel/entry_32.S
index c03e829..5d919e4 100644
--- a/arch/powerpc/kernel/entry_32.S
+++ b/arch/powerpc/kernel/entry_32.S
@@ -202,6 +202,11 @@ _GLOBAL(DoSyscall)
 	bl	do_show_syscall
 #endif /* SHOW_SYSCALLS */
 	rlwinm	r10,r1,0,0,(31-THREAD_SHIFT)	/* current_thread_info() */
+#ifdef CONFIG_TASKSTATS 
+	lwz r11,TI_SYSC_CNT(r10)
+	addi r11,r11,1
+	stw r11,TI_SYSC_CNT(r10)
+#endif
 	lwz	r11,TI_FLAGS(r10)
 	andi.	r11,r11,_TIF_SYSCALL_T_OR_A
 	bne-	syscall_dotrace
diff --git a/arch/powerpc/kernel/entry_64.S b/arch/powerpc/kernel/entry_64.S
index 2551c08..5907f76 100644
--- a/arch/powerpc/kernel/entry_64.S
+++ b/arch/powerpc/kernel/entry_64.S
@@ -115,6 +115,11 @@ END_FW_FTR_SECTION_IFSET(FW_FEATURE_ISERIES)
 	addi	r9,r1,STACK_FRAME_OVERHEAD
 #endif
 	clrrdi	r11,r1,THREAD_SHIFT
+#ifdef CONFIG_TASKSTATS
+	ld r10,TI_SYSCALL_COUNT(r11)
+	addi r10,r10,1
+	std r10,TI_SYSCALL_COUNT(r11)
+#endif
 	ld	r10,TI_FLAGS(r11)
 	andi.	r11,r10,_TIF_SYSCALL_T_OR_A
 	bne-	syscall_dotrace
diff --git a/arch/x86_64/kernel/asm-offsets.c b/arch/x86_64/kernel/asm-offsets.c
index 96687e2..da57356 100644
--- a/arch/x86_64/kernel/asm-offsets.c
+++ b/arch/x86_64/kernel/asm-offsets.c
@@ -35,6 +35,7 @@ int main(void)
 	ENTRY(addr_limit);
 	ENTRY(preempt_count);
 	ENTRY(status);
+	ENTRY(syscall_count);
 	BLANK();
 #undef ENTRY
 #define ENTRY(entry) DEFINE(pda_ ## entry, offsetof(struct x8664_pda, entry))
diff --git a/arch/x86_64/kernel/entry.S b/arch/x86_64/kernel/entry.S
index 9f5dac6..af40ead 100644
--- a/arch/x86_64/kernel/entry.S
+++ b/arch/x86_64/kernel/entry.S
@@ -229,6 +229,9 @@ ENTRY(system_call)
 	movq  %rcx,RIP-ARGOFFSET(%rsp)
 	CFI_REL_OFFSET rip,RIP-ARGOFFSET
 	GET_THREAD_INFO(%rcx)
+#ifdef CONFIG_TASKSTATS
+	addq $1, threadinfo_syscall_count(%rcx)   # Increment syscalls counter
+#endif
 	testl $(_TIF_SYSCALL_TRACE|_TIF_SYSCALL_AUDIT|_TIF_SECCOMP),threadinfo_flags(%rcx)
 	jnz tracesys
 	cmpq $__NR_syscall_max,%rax
diff --git a/fs/proc/array.c b/fs/proc/array.c
index 70e4fab..c805c08 100644
--- a/fs/proc/array.c
+++ b/fs/proc/array.c
@@ -290,6 +290,19 @@ static inline char *task_cap(struct task_struct *p, char *buffer)
 			    cap_t(p->cap_permitted),
 			    cap_t(p->cap_effective));
 }
+static inline char *task_perf(struct task_struct *p, char *buffer)
+{
+	/* Syscall counter adds 1 line overhead on each syscall execution
+	 * in entry.S, so probably it is the leave this stuff under ifdefs.
+	 */
+#ifdef CONFIG_TASKSTATS
+	buffer += sprintf(buffer, "Syscalls:\t%lu\n", p->thread_info->syscall_count);
+#endif
+	return buffer + sprintf(buffer, "voluntary_ctxt_switches:\t%lu\n"
+			    "nonvoluntary_ctxt_switches:\t%lu\n",
+			    p->nvcsw,
+			    p->nivcsw);
+}
 
 int proc_pid_status(struct task_struct *task, char * buffer)
 {
@@ -309,6 +322,7 @@ int proc_pid_status(struct task_struct *task, char * buffer)
 #if defined(CONFIG_S390)
 	buffer = task_show_regs(task, buffer);
 #endif
+	buffer = task_perf(task, buffer);
 	return buffer - orig;
 }
 
diff --git a/include/asm-i386/thread_info.h b/include/asm-i386/thread_info.h
index 4b187bb..bccfd6a 100644
--- a/include/asm-i386/thread_info.h
+++ b/include/asm-i386/thread_info.h
@@ -33,6 +33,7 @@ struct thread_info {
 	int			preempt_count;	/* 0 => preemptable, <0 => BUG */
 
 
+	unsigned long		syscall_count;	/* Syscall counter */
 	mm_segment_t		addr_limit;	/* thread address space:
 					 	   0-0xBFFFFFFF for user-thead
 						   0-0xFFFFFFFF for kernel-thread
diff --git a/include/asm-powerpc/thread_info.h b/include/asm-powerpc/thread_info.h
index 3f32ca8..5306ac2 100644
--- a/include/asm-powerpc/thread_info.h
+++ b/include/asm-powerpc/thread_info.h
@@ -35,6 +35,7 @@ struct thread_info {
 	int		cpu;			/* cpu we're on */
 	int		preempt_count;		/* 0 => preemptable,
 						   <0 => BUG */
+	unsigned long	syscall_count;		/* Syscall counter */
 	struct restart_block restart_block;
 	unsigned long	local_flags;		/* private flags for thread */
 
diff --git a/include/asm-x86_64/thread_info.h b/include/asm-x86_64/thread_info.h
index 74a6c74..e53022d 100644
--- a/include/asm-x86_64/thread_info.h
+++ b/include/asm-x86_64/thread_info.h
@@ -31,6 +31,7 @@ struct thread_info {
 	__u32			cpu;		/* current CPU */
 	int 			preempt_count;	/* 0 => preemptable, <0 => BUG */
 
+	unsigned long		syscall_count;	/* Syscall counter */
 	mm_segment_t		addr_limit;	
 	struct restart_block    restart_block;
 };
diff --git a/include/linux/taskstats.h b/include/linux/taskstats.h
index 3fced47..e3341b6 100644
--- a/include/linux/taskstats.h
+++ b/include/linux/taskstats.h
@@ -31,7 +31,7 @@
  */
 
 
-#define TASKSTATS_VERSION	3
+#define TASKSTATS_VERSION	4
 #define TS_COMM_LEN		32	/* should be >= TASK_COMM_LEN
 					 * in linux/sched.h */
 
@@ -141,6 +141,10 @@ struct taskstats {
 	__u64	write_syscalls;		/* write syscalls */
 	/* Extended accounting fields end */
 
+	__u32  syscall_counter;   /* Syscall counter */
+	__u32  nvcsw;
+	__u32  nivcsw;
+
 #define TASKSTATS_HAS_IO_ACCOUNTING
 	/* Per-task storage I/O accounting starts */
 	__u64	read_bytes;		/* bytes of read I/O */
diff --git a/kernel/fork.c b/kernel/fork.c
index fc723e5..5213738 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -1042,6 +1042,9 @@ static struct task_struct *copy_process(unsigned long clone_flags,
 	p->wchar = 0;		/* I/O counter: bytes written */
 	p->syscr = 0;		/* I/O counter: read syscalls */
 	p->syscw = 0;		/* I/O counter: write syscalls */
+#ifdef CONFIG_TASKSTATS
+	p->thread_info->syscall_count = 0; /* Syscall counter: total numbers of syscalls */
+#endif
 	task_io_accounting_init(p);
 	acct_clear_integrals(p);
 
diff --git a/kernel/taskstats.c b/kernel/taskstats.c
index 4c3476f..d7bf33f 100644
--- a/kernel/taskstats.c
+++ b/kernel/taskstats.c
@@ -196,6 +196,9 @@ static int fill_pid(pid_t pid, struct task_struct *tsk,
 
 	/* fill in basic acct fields */
 	stats->version = TASKSTATS_VERSION;
+	stats->syscall_counter = tsk->thread_info->syscall_count;
+	stats->nvcsw = tsk->nvcsw;
+	stats->nivcsw = tsk->nivcsw;
 	bacct_add_tsk(stats, tsk);
 
 	/* fill in extended acct fields */
@@ -242,6 +245,9 @@ static int fill_tgid(pid_t tgid, struct task_struct *first,
 		 */
 		delayacct_add_tsk(stats, tsk);
 
+		stats->syscall_counter += tsk->thread_info->syscall_count;
+		stats->nvcsw += tsk->nvcsw;
+		stats->nivcsw += tsk->nivcsw;
 	} while_each_thread(first, tsk);
 
 	unlock_task_sighand(first, &flags);

^ permalink raw reply related	[flat|nested] 22+ messages in thread
* Re: [PATCH] Performance Stats: Kernel patch
  2007-05-10 10:22   ` Maxim Uvarov
@ 2007-05-10 16:47     ` Linas Vepstas
  0 siblings, 0 replies; 22+ messages in thread
From: Linas Vepstas @ 2007-05-10 16:47 UTC (permalink / raw)
  To: Maxim Uvarov; +Cc: linuxppc-dev, Andrew Morton, pavel, wli, dada1

On Thu, May 10, 2007 at 10:22:54AM +0000, Maxim Uvarov wrote:
> Linas Vepstas wrote:
> >On Tue, May 08, 2007 at 04:26:51PM +0000, Maxim Uvarov wrote:
> >>+	addq $1, threadinfo_syscall_count(%rcx)   # Increment syscalls 
> >
> >again, #ifdef CONFIG_TASKSTATS,
> 
> So I don't really sure if  defines are needed in  entry.S code. I tried 
> to do
> some tests and  it is impossible to catch the performance difference (small
> overhead).

"A penny here, and a penny there, and pretty soon it adds up to real
money". Its a hot path, and its an easy fix. 

--linas

^ permalink raw reply	[flat|nested] 22+ messages in thread
* Re: [PATCH] Performance Stats: Kernel patch
  2007-05-10 11:11   ` Maxim Uvarov
@ 2007-05-10 18:12     ` Andrew Morton
  2007-05-11 16:51       ` Maxim Uvarov
  2007-05-11 17:23       ` Using kprobes [was " Linas Vepstas
  0 siblings, 2 replies; 22+ messages in thread
From: Andrew Morton @ 2007-05-10 18:12 UTC (permalink / raw)
  To: Maxim Uvarov; +Cc: linuxppc-dev, wli, pavel, dada1

On Thu, 10 May 2007 11:11:18 +0000 Maxim Uvarov <muvarov@ru.mvista.com> wrote:

> >syscall entry is a super-hotpath and there will be a developer revolt
> >if we go adding an incl to that path, sorry.
> >
> >  
> >
> Yes, I know that it is very critical part of code. May be protection with
> ifdef CONFIG_TASKSTATS  in all entry.S can be solution?
> In that case we can be sure that nothing wrong can not happen,
> and if this functionality is needed we can turn on this.

I very much hope that all distributors will enable taskstats, so that won't
help.

I don't think the syscall-counting feature has a future, sorry.  Perhaps
you could do something like hooking it up on-demand by insertion of a kprobe,
dunno.

> >(Did that incl also cover the int 80 entry?)
> >  
> >
> Actually not. But counter increments each time when syscall  calls.
> But it is in very save place, just after
>         SAVE_ALL
>         GET_THREAD_INFO(%ebp)
> So it is very hard to break something.
> 

What I meant was: the code appears to be counting sysenter-invoked syscalls,
but it isn't counting not int 80-invoked syscalls.

^ permalink raw reply	[flat|nested] 22+ messages in thread
* Re: [PATCH] Performance Stats: Kernel patch
  2007-05-10 12:38 ` Josh Boyer
@ 2007-05-10 18:23   ` Maxim Uvarov
  0 siblings, 0 replies; 22+ messages in thread
From: Maxim Uvarov @ 2007-05-10 18:23 UTC (permalink / raw)
  To: Josh Boyer; +Cc: linuxppc-dev, Andrew Morton, wli, pavel, dada1

Josh Boyer wrote:

>On Thu, 2007-05-10 at 12:39 +0000, Maxim Uvarov wrote:
>  
>
>>Patch makes available to the user the following
>>task and process performance statistics:
>>	* Involuntary Context Switches (task_struct->nivcsw)
>>	* Voluntary Context Switches (task_struct->nvcsw)
>>	* Number of system calls (added new counter
>>	  thread_info->sysall_count)
>>	           
>>Statistics information is available from:
>>        1. taskstats interface (Documentation/accounting/)
>>	2. /proc/PID/status (task only).
>>
>>This data is useful for detecting hyperactivity
>>patterns between processes.
>>    
>>
>
>This should be sent to the linux-kernel list I think...  Or at least
>CCd.
>
>  
>
Thanks,  It is interesting but I can not Cc message to  linux-kernel.
Only sending directly works (To:).

Link is here:
http://lkml.org/lkml/2007/5/10/289

>josh
>
>  
>

^ permalink raw reply	[flat|nested] 22+ messages in thread
* Re: [PATCH] Performance Stats: Kernel patch
  2007-05-10 11:42 Maxim Uvarov
  2007-05-10 12:03 ` Pavel Machek
@ 2007-05-10 18:25 ` Andrew Morton
  1 sibling, 0 replies; 22+ messages in thread
From: Andrew Morton @ 2007-05-10 18:25 UTC (permalink / raw)
  To: Maxim Uvarov; +Cc: linuxppc-dev, pavel, wli, dada1

On Thu, 10 May 2007 11:42:22 +0000 Maxim Uvarov <muvarov@ru.mvista.com> wrote:

> From: Maxim Uvarov <muvarov@ru.mvista.com>
> To: Andrew Morton <akpm@linux-foundation.org>
> Cc: <wli@holomorphy.com>, <dada1@cosmosbay.com>, <pavel@ucw.cz>, <linas@austin.ibm.com>
> Cc: linuxppc-dev@ozlabs.org
> Cc: <wli@holomorphy.com>, <dada1@cosmosbay.com>, <pavel@ucw.cz>, <linas@austin.ibm.com>

The addressees are still odd.

> Subject: [PATCH] Performance Stats: Kernel patch
> Date: Thu, 10 May 2007 11:42:22 +0000
> User-Agent: StGIT/0.12.1
> 
> 
> Patch makes available to the user the following
> task and process performance statistics:
> 	* Involuntary Context Switches (task_struct->nivcsw)
> 	* Voluntary Context Switches (task_struct->nvcsw)
> 	* Number of system calls (added new counter
> 	  thread_info->sysall_count)

The patch looks good, thanks.  Except, as I said, I don't think we can
justify the syscall counter.  Yes, it's useful information and yes the
overhead is low.

But it adds overhead to every syscall on every machine in the world, which
is a lot of overhead.  And the number of machines which actually _use_ that
information will be small.

But hey, I might be wrong - perhaps lots of people have a burning need for
this information.  But we're not in a position to find that out because you
keep on mucking up the email recipient list :(


The patch is confused about the type of the syscall counter.  In some
places it is "unsigned long" but it gets sent to userspace as __u32.

Considering that this counter can be overflowed in a matter of seconds, you
definitely want it to be 64-bit on 64-bit machines.  And you probably want
it to be 64-bit on 32-bit machines as well.  More overhead there though.

Anyway, I suggest you fix that up and rereresend the patch, getting the
headers correct and we'll ask a wider audience about the
value-versus-overhead issue.

I suspect we'll end up only supporting the context-switch counters.  Given
that this requires an uprev to the taskstats interface, we should look at
whether there are other things which we should be exporting but which were
forgotten.  I'll have a think about that sometime.

^ permalink raw reply	[flat|nested] 22+ messages in thread
* Re: [PATCH] Performance Stats: Kernel patch
  2007-05-10 18:12     ` Andrew Morton
@ 2007-05-11 16:51       ` Maxim Uvarov
  2007-05-11 17:23       ` Using kprobes [was " Linas Vepstas
  1 sibling, 0 replies; 22+ messages in thread
From: Maxim Uvarov @ 2007-05-11 16:51 UTC (permalink / raw)
  To: Andrew Morton; +Cc: linuxppc-dev, wli, pavel, dada1

Andrew Morton wrote:

>
>What I meant was: the code appears to be counting sysenter-invoked syscalls,
>but it isn't counting not int 80-invoked syscalls.
>
>  
>
You are right. Seems it should be added too.

^ permalink raw reply	[flat|nested] 22+ messages in thread
* [PATCH] Performance Stats: Kernel patch
@ 2007-05-11 17:13 Maxim Uvarov
  2007-05-12 10:39 ` Andrea Righi
  0 siblings, 1 reply; 22+ messages in thread
From: Maxim Uvarov @ 2007-05-11 17:13 UTC (permalink / raw)
  To: LKML; +Cc: msyrchin, wli, dada1, linuxppc-dev, pavel


Hello,

Thanks all for very useful comments. 
Please review this version.


Best regards,
Maxim.

Patch makes available to the user the following
task and process performance statistics:
	* Involuntary Context Switches (task_struct->nivcsw)
	* Voluntary Context Switches (task_struct->nvcsw)
	* Number of system calls (added new counter
	  thread_info->sysall_count)
	           
Statistics information is available from:
        1. taskstats interface (Documentation/accounting/)
	2. /proc/PID/status (task only).

This data is useful for detecting hyperactivity
patterns between processes.

Signed-off-by: Maxim Uvarov <muvarov@ru.mvista.com>


                                                          
---

 Documentation/accounting/getdelays.c          |   20 ++++++++++++++++++--
 Documentation/accounting/taskstats-struct.txt |    7 +++++++
 arch/i386/kernel/asm-offsets.c                |    1 +
 arch/i386/kernel/entry.S                      |    3 +++
 arch/powerpc/kernel/asm-offsets.c             |    2 ++
 arch/powerpc/kernel/entry_32.S                |    5 +++++
 arch/powerpc/kernel/entry_64.S                |    5 +++++
 arch/x86_64/kernel/asm-offsets.c              |    1 +
 arch/x86_64/kernel/entry.S                    |    3 +++
 fs/proc/array.c                               |   14 ++++++++++++++
 include/asm-i386/thread_info.h                |    1 +
 include/asm-powerpc/thread_info.h             |    1 +
 include/asm-x86_64/thread_info.h              |    1 +
 include/linux/taskstats.h                     |    6 +++++-
 kernel/fork.c                                 |    3 +++
 kernel/taskstats.c                            |    6 ++++++
 16 files changed, 76 insertions(+), 3 deletions(-)

diff --git a/Documentation/accounting/getdelays.c b/Documentation/accounting/getdelays.c
index e9126e7..1be7d65 100644
--- a/Documentation/accounting/getdelays.c
+++ b/Documentation/accounting/getdelays.c
@@ -49,6 +49,7 @@ char name[100];
 int dbg;
 int print_delays;
 int print_io_accounting;
+int print_task_stats;
 __u64 stime, utime;
 
 #define PRINTF(fmt, arg...) {			\
@@ -187,7 +188,7 @@ void print_delayacct(struct taskstats *t)
 	       "IO    %15s%15s\n"
 	       "      %15llu%15llu\n"
 	       "MEM   %15s%15s\n"
-	       "      %15llu%15llu\n\n",
+	       "      %15llu%15llu\n"
 	       "count", "real total", "virtual total", "delay total",
 	       t->cpu_count, t->cpu_run_real_total, t->cpu_run_virtual_total,
 	       t->cpu_delay_total,
@@ -196,6 +197,15 @@ void print_delayacct(struct taskstats *t)
 	       "count", "delay total", t->swapin_count, t->swapin_delay_total);
 }
 
+void print_taskstats(struct taskstats *t)
+{
+	printf("\n\nTask  %15s%15s%15s\n"
+	       "      %15lu%15lu%15lu\n",
+	       "syscalls", "voluntary", "nonvoluntary",
+	       t->syscall_counter, t->nvcsw, t->nivcsw);
+
+}
+
 void print_ioacct(struct taskstats *t)
 {
 	printf("%s: read=%llu, write=%llu, cancelled_write=%llu\n",
@@ -227,7 +237,7 @@ int main(int argc, char *argv[])
 	struct msgtemplate msg;
 
 	while (1) {
-		c = getopt(argc, argv, "diw:r:m:t:p:v:l");
+		c = getopt(argc, argv, "qdiw:r:m:t:p:v:l");
 		if (c < 0)
 			break;
 
@@ -240,6 +250,10 @@ int main(int argc, char *argv[])
 			printf("printing IO accounting\n");
 			print_io_accounting = 1;
 			break;
+		case 'q':
+			printf("printing task/process stasistics:\n");
+			print_task_stats = 1;
+			break;
 		case 'w':
 			strncpy(logfile, optarg, MAX_FILENAME);
 			printf("write to file %s\n", logfile);
@@ -381,6 +395,8 @@ int main(int argc, char *argv[])
 							print_delayacct((struct taskstats *) NLA_DATA(na));
 						if (print_io_accounting)
 							print_ioacct((struct taskstats *) NLA_DATA(na));
+						if (print_task_stats)
+							print_taskstats((struct taskstats *) NLA_DATA(na));
 						if (fd) {
 							if (write(fd, NLA_DATA(na), na->nla_len) < 0) {
 								err(1,"write error\n");
diff --git a/Documentation/accounting/taskstats-struct.txt b/Documentation/accounting/taskstats-struct.txt
index 661c797..606aef6 100644
--- a/Documentation/accounting/taskstats-struct.txt
+++ b/Documentation/accounting/taskstats-struct.txt
@@ -22,6 +22,8 @@ There are three different groups of fields in the struct taskstats:
     /* Extended accounting fields end */
     Their values are collected if CONFIG_TASK_XACCT is set.
 
+4) Per-task and per-thread statistics
+
 Future extension should add fields to the end of the taskstats struct, and
 should not change the relative position of each field within the struct.
 
@@ -158,4 +160,9 @@ struct taskstats {
 
 	/* Extended accounting fields end */
 
+4) Per-task and per-thread statiscits
+	__u64	syscall_counter;	/* Syscall counter */
+	__u64	nvcsw;			/* Context voluntary switch counter */
+	__u64	nivcsw;			/* Context involuntary switch counter */
+
 }
diff --git a/arch/i386/kernel/asm-offsets.c b/arch/i386/kernel/asm-offsets.c
index 1b2f3cd..4ad49d2 100644
--- a/arch/i386/kernel/asm-offsets.c
+++ b/arch/i386/kernel/asm-offsets.c
@@ -56,6 +56,7 @@ void foo(void)
 	OFFSET(TI_addr_limit, thread_info, addr_limit);
 	OFFSET(TI_restart_block, thread_info, restart_block);
 	OFFSET(TI_sysenter_return, thread_info, sysenter_return);
+	OFFSET(TI_syscall_count, thread_info, syscall_count);
 	BLANK();
 
 	OFFSET(GDS_size, Xgt_desc_struct, size);
diff --git a/arch/i386/kernel/entry.S b/arch/i386/kernel/entry.S
index 5e47683..836961f 100644
--- a/arch/i386/kernel/entry.S
+++ b/arch/i386/kernel/entry.S
@@ -332,6 +332,9 @@ sysenter_past_esp:
 	SAVE_ALL
 	GET_THREAD_INFO(%ebp)
 
+#ifdef CONFIG_TASKSTATS
+	incl	TI_syscall_count(%ebp)	# Increment syscalls counter
+#endif
 	/* Note, _TIF_SECCOMP is bit number 8, and so it needs testw and not testb */
 	testw $(_TIF_SYSCALL_EMU|_TIF_SYSCALL_TRACE|_TIF_SECCOMP|_TIF_SYSCALL_AUDIT),TI_flags(%ebp)
 	jnz syscall_trace_entry
diff --git a/arch/powerpc/kernel/asm-offsets.c b/arch/powerpc/kernel/asm-offsets.c
index 030d300..b640039 100644
--- a/arch/powerpc/kernel/asm-offsets.c
+++ b/arch/powerpc/kernel/asm-offsets.c
@@ -94,6 +94,8 @@ int main(void)
 	DEFINE(TI_LOCAL_FLAGS, offsetof(struct thread_info, local_flags));
 	DEFINE(TI_PREEMPT, offsetof(struct thread_info, preempt_count));
 	DEFINE(TI_TASK, offsetof(struct thread_info, task));
+	DEFINE(TI_SYSCALL_COUNT, offsetof(struct thread_info, syscall_count));
+
 #ifdef CONFIG_PPC32
 	DEFINE(TI_EXECDOMAIN, offsetof(struct thread_info, exec_domain));
 	DEFINE(TI_CPU, offsetof(struct thread_info, cpu));
diff --git a/arch/powerpc/kernel/entry_32.S b/arch/powerpc/kernel/entry_32.S
index c03e829..5d919e4 100644
--- a/arch/powerpc/kernel/entry_32.S
+++ b/arch/powerpc/kernel/entry_32.S
@@ -202,6 +202,11 @@ _GLOBAL(DoSyscall)
 	bl	do_show_syscall
 #endif /* SHOW_SYSCALLS */
 	rlwinm	r10,r1,0,0,(31-THREAD_SHIFT)	/* current_thread_info() */
+#ifdef CONFIG_TASKSTATS 
+	lwz r11,TI_SYSC_CNT(r10)
+	addi r11,r11,1
+	stw r11,TI_SYSC_CNT(r10)
+#endif
 	lwz	r11,TI_FLAGS(r10)
 	andi.	r11,r11,_TIF_SYSCALL_T_OR_A
 	bne-	syscall_dotrace
diff --git a/arch/powerpc/kernel/entry_64.S b/arch/powerpc/kernel/entry_64.S
index 2551c08..5907f76 100644
--- a/arch/powerpc/kernel/entry_64.S
+++ b/arch/powerpc/kernel/entry_64.S
@@ -115,6 +115,11 @@ END_FW_FTR_SECTION_IFSET(FW_FEATURE_ISERIES)
 	addi	r9,r1,STACK_FRAME_OVERHEAD
 #endif
 	clrrdi	r11,r1,THREAD_SHIFT
+#ifdef CONFIG_TASKSTATS
+	ld r10,TI_SYSCALL_COUNT(r11)
+	addi r10,r10,1
+	std r10,TI_SYSCALL_COUNT(r11)
+#endif
 	ld	r10,TI_FLAGS(r11)
 	andi.	r11,r10,_TIF_SYSCALL_T_OR_A
 	bne-	syscall_dotrace
diff --git a/arch/x86_64/kernel/asm-offsets.c b/arch/x86_64/kernel/asm-offsets.c
index 96687e2..da57356 100644
--- a/arch/x86_64/kernel/asm-offsets.c
+++ b/arch/x86_64/kernel/asm-offsets.c
@@ -35,6 +35,7 @@ int main(void)
 	ENTRY(addr_limit);
 	ENTRY(preempt_count);
 	ENTRY(status);
+	ENTRY(syscall_count);
 	BLANK();
 #undef ENTRY
 #define ENTRY(entry) DEFINE(pda_ ## entry, offsetof(struct x8664_pda, entry))
diff --git a/arch/x86_64/kernel/entry.S b/arch/x86_64/kernel/entry.S
index 9f5dac6..af40ead 100644
--- a/arch/x86_64/kernel/entry.S
+++ b/arch/x86_64/kernel/entry.S
@@ -229,6 +229,9 @@ ENTRY(system_call)
 	movq  %rcx,RIP-ARGOFFSET(%rsp)
 	CFI_REL_OFFSET rip,RIP-ARGOFFSET
 	GET_THREAD_INFO(%rcx)
+#ifdef CONFIG_TASKSTATS
+	addq $1, threadinfo_syscall_count(%rcx)   # Increment syscalls counter
+#endif
 	testl $(_TIF_SYSCALL_TRACE|_TIF_SYSCALL_AUDIT|_TIF_SECCOMP),threadinfo_flags(%rcx)
 	jnz tracesys
 	cmpq $__NR_syscall_max,%rax
diff --git a/fs/proc/array.c b/fs/proc/array.c
index 70e4fab..c805c08 100644
--- a/fs/proc/array.c
+++ b/fs/proc/array.c
@@ -290,6 +290,19 @@ static inline char *task_cap(struct task_struct *p, char *buffer)
 			    cap_t(p->cap_permitted),
 			    cap_t(p->cap_effective));
 }
+static inline char *task_perf(struct task_struct *p, char *buffer)
+{
+	/* Syscall counter adds 1 line overhead on each syscall execution
+	 * in entry.S, so probably it is the leave this stuff under ifdefs.
+	 */
+#ifdef CONFIG_TASKSTATS
+	buffer += sprintf(buffer, "Syscalls:\t%lu\n", p->thread_info->syscall_count);
+#endif
+	return buffer + sprintf(buffer, "voluntary_ctxt_switches:\t%lu\n"
+			    "nonvoluntary_ctxt_switches:\t%lu\n",
+			    p->nvcsw,
+			    p->nivcsw);
+}
 
 int proc_pid_status(struct task_struct *task, char * buffer)
 {
@@ -309,6 +322,7 @@ int proc_pid_status(struct task_struct *task, char * buffer)
 #if defined(CONFIG_S390)
 	buffer = task_show_regs(task, buffer);
 #endif
+	buffer = task_perf(task, buffer);
 	return buffer - orig;
 }
 
diff --git a/include/asm-i386/thread_info.h b/include/asm-i386/thread_info.h
index 4b187bb..bccfd6a 100644
--- a/include/asm-i386/thread_info.h
+++ b/include/asm-i386/thread_info.h
@@ -33,6 +33,7 @@ struct thread_info {
 	int			preempt_count;	/* 0 => preemptable, <0 => BUG */
 
 
+	unsigned long		syscall_count;	/* Syscall counter */
 	mm_segment_t		addr_limit;	/* thread address space:
 					 	   0-0xBFFFFFFF for user-thead
 						   0-0xFFFFFFFF for kernel-thread
diff --git a/include/asm-powerpc/thread_info.h b/include/asm-powerpc/thread_info.h
index 3f32ca8..5306ac2 100644
--- a/include/asm-powerpc/thread_info.h
+++ b/include/asm-powerpc/thread_info.h
@@ -35,6 +35,7 @@ struct thread_info {
 	int		cpu;			/* cpu we're on */
 	int		preempt_count;		/* 0 => preemptable,
 						   <0 => BUG */
+	unsigned long	syscall_count;		/* Syscall counter */
 	struct restart_block restart_block;
 	unsigned long	local_flags;		/* private flags for thread */
 
diff --git a/include/asm-x86_64/thread_info.h b/include/asm-x86_64/thread_info.h
index 74a6c74..e53022d 100644
--- a/include/asm-x86_64/thread_info.h
+++ b/include/asm-x86_64/thread_info.h
@@ -31,6 +31,7 @@ struct thread_info {
 	__u32			cpu;		/* current CPU */
 	int 			preempt_count;	/* 0 => preemptable, <0 => BUG */
 
+	unsigned long		syscall_count;	/* Syscall counter */
 	mm_segment_t		addr_limit;	
 	struct restart_block    restart_block;
 };
diff --git a/include/linux/taskstats.h b/include/linux/taskstats.h
index 3fced47..98dfde7 100644
--- a/include/linux/taskstats.h
+++ b/include/linux/taskstats.h
@@ -31,7 +31,7 @@
  */
 
 
-#define TASKSTATS_VERSION	3
+#define TASKSTATS_VERSION	4
 #define TS_COMM_LEN		32	/* should be >= TASK_COMM_LEN
 					 * in linux/sched.h */
 
@@ -141,6 +141,10 @@ struct taskstats {
 	__u64	write_syscalls;		/* write syscalls */
 	/* Extended accounting fields end */
 
+	__u64  syscall_counter;		/* Syscall counter */
+	__u64  nvcsw;
+	__u64  nivcsw;
+
 #define TASKSTATS_HAS_IO_ACCOUNTING
 	/* Per-task storage I/O accounting starts */
 	__u64	read_bytes;		/* bytes of read I/O */
diff --git a/kernel/fork.c b/kernel/fork.c
index fc723e5..5213738 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -1042,6 +1042,9 @@ static struct task_struct *copy_process(unsigned long clone_flags,
 	p->wchar = 0;		/* I/O counter: bytes written */
 	p->syscr = 0;		/* I/O counter: read syscalls */
 	p->syscw = 0;		/* I/O counter: write syscalls */
+#ifdef CONFIG_TASKSTATS
+	p->thread_info->syscall_count = 0; /* Syscall counter: total numbers of syscalls */
+#endif
 	task_io_accounting_init(p);
 	acct_clear_integrals(p);
 
diff --git a/kernel/taskstats.c b/kernel/taskstats.c
index 4c3476f..d7bf33f 100644
--- a/kernel/taskstats.c
+++ b/kernel/taskstats.c
@@ -196,6 +196,9 @@ static int fill_pid(pid_t pid, struct task_struct *tsk,
 
 	/* fill in basic acct fields */
 	stats->version = TASKSTATS_VERSION;
+	stats->syscall_counter = tsk->thread_info->syscall_count;
+	stats->nvcsw = tsk->nvcsw;
+	stats->nivcsw = tsk->nivcsw;
 	bacct_add_tsk(stats, tsk);
 
 	/* fill in extended acct fields */
@@ -242,6 +245,9 @@ static int fill_tgid(pid_t tgid, struct task_struct *first,
 		 */
 		delayacct_add_tsk(stats, tsk);
 
+		stats->syscall_counter += tsk->thread_info->syscall_count;
+		stats->nvcsw += tsk->nvcsw;
+		stats->nivcsw += tsk->nivcsw;
 	} while_each_thread(first, tsk);
 
 	unlock_task_sighand(first, &flags);

^ permalink raw reply related	[flat|nested] 22+ messages in thread
* Using kprobes [was Re: [PATCH] Performance Stats: Kernel patch
  2007-05-10 18:12     ` Andrew Morton
  2007-05-11 16:51       ` Maxim Uvarov
@ 2007-05-11 17:23       ` Linas Vepstas
  2007-05-12  9:15         ` Christoph Hellwig
  2007-05-21 15:28         ` Maxim Uvarov
  1 sibling, 2 replies; 22+ messages in thread
From: Linas Vepstas @ 2007-05-11 17:23 UTC (permalink / raw)
  To: Andrew Morton; +Cc: Maxim Uvarov, dada1, pavel, wli, linuxppc-dev

On Thu, May 10, 2007 at 11:12:42AM -0700, Andrew Morton wrote:
> 
> I don't think the syscall-counting feature has a future, sorry.  Perhaps
> you could do something like hooking it up on-demand by insertion of a kprobe,
> dunno.

This is an interesting point. I've started fiddling with (a wrapper
around) kprobes that allows me to pick any subroutine symbol in System.map,
and then get an event whenever that subroutine gets called. Its pretty
slick, and allows me to gather data on certain unusual events in the
kernel.  (I'm not using this for performance monitoring, I'm trying to 
do RAS).

It makes a lot of sense to me to have a generic kprobe extension, where 
you could give it a list of subroutine names, and it'll collect stats
on the number of times that the routine was called. Some user-space thingy
could poll for those stats, or you could put them in /sys or wherever.

Its more complicted than just instriumenting syscalls, but a lot more
useful, I would think ... 

--linas

^ permalink raw reply	[flat|nested] 22+ messages in thread
* Re: Using kprobes [was Re: [PATCH] Performance Stats: Kernel patch
  2007-05-11 17:23       ` Using kprobes [was " Linas Vepstas
@ 2007-05-12  9:15         ` Christoph Hellwig
  2007-05-21 15:28         ` Maxim Uvarov
  1 sibling, 0 replies; 22+ messages in thread
From: Christoph Hellwig @ 2007-05-12  9:15 UTC (permalink / raw)
  To: Linas Vepstas
  Cc: wli, dada1, linuxppc-dev, pavel, Maxim Uvarov, Andrew Morton

On Fri, May 11, 2007 at 12:23:04PM -0500, Linas Vepstas wrote:
> On Thu, May 10, 2007 at 11:12:42AM -0700, Andrew Morton wrote:
> > 
> > I don't think the syscall-counting feature has a future, sorry.  Perhaps
> > you could do something like hooking it up on-demand by insertion of a kprobe,
> > dunno.
> 
> This is an interesting point. I've started fiddling with (a wrapper
> around) kprobes that allows me to pick any subroutine symbol in System.map,
> and then get an event whenever that subroutine gets called. Its pretty
> slick, and allows me to gather data on certain unusual events in the
> kernel.  (I'm not using this for performance monitoring, I'm trying to 
> do RAS).
> 
> It makes a lot of sense to me to have a generic kprobe extension, where 
> you could give it a list of subroutine names, and it'll collect stats
> on the number of times that the routine was called. Some user-space thingy
> could poll for those stats, or you could put them in /sys or wherever.
> 
> Its more complicted than just instriumenting syscalls, but a lot more
> useful, I would think ... 

Yes, this is exactly the useful kind of kprobes useage we should have in
the tree.  

> 
> --linas
> 
> _______________________________________________
> Linuxppc-dev mailing list
> Linuxppc-dev@ozlabs.org
> https://ozlabs.org/mailman/listinfo/linuxppc-dev
---end quoted text---

^ permalink raw reply	[flat|nested] 22+ messages in thread
* Re: [PATCH] Performance Stats: Kernel patch
  2007-05-11 17:13 Maxim Uvarov
@ 2007-05-12 10:39 ` Andrea Righi
  0 siblings, 0 replies; 22+ messages in thread
From: Andrea Righi @ 2007-05-12 10:39 UTC (permalink / raw)
  To: Maxim Uvarov; +Cc: msyrchin, LKML, wli, dada1, linuxppc-dev, pavel

Maxim Uvarov wrote:
> diff --git a/Documentation/accounting/taskstats-struct.txt b/Documentation/accounting/taskstats-struct.txt
> index 661c797..606aef6 100644
> --- a/Documentation/accounting/taskstats-struct.txt
> +++ b/Documentation/accounting/taskstats-struct.txt
> @@ -22,6 +22,8 @@ There are three different groups of fields in the struct taskstats:
>      /* Extended accounting fields end */
>      Their values are collected if CONFIG_TASK_XACCT is set.
>  
> +4) Per-task and per-thread statistics
> +
>  Future extension should add fields to the end of the taskstats struct, and
>  should not change the relative position of each field within the struct.
>  
> @@ -158,4 +160,9 @@ struct taskstats {
>  
>  	/* Extended accounting fields end */
>  
> +4) Per-task and per-thread statiscits

...a small typo in the documentation: s/statiscits/statistics/

-Andrea

^ permalink raw reply	[flat|nested] 22+ messages in thread
* Re: Using kprobes [was Re: [PATCH] Performance Stats: Kernel patch
  2007-05-11 17:23       ` Using kprobes [was " Linas Vepstas
  2007-05-12  9:15         ` Christoph Hellwig
@ 2007-05-21 15:28         ` Maxim Uvarov
  1 sibling, 0 replies; 22+ messages in thread
From: Maxim Uvarov @ 2007-05-21 15:28 UTC (permalink / raw)
  To: Linas Vepstas; +Cc: linuxppc-dev, Andrew Morton, pavel, wli, dada1

Linas Vepstas wrote:

>On Thu, May 10, 2007 at 11:12:42AM -0700, Andrew Morton wrote:
>  
>
>>I don't think the syscall-counting feature has a future, sorry.  Perhaps
>>you could do something like hooking it up on-demand by insertion of a kprobe,
>>dunno.
>>    
>>
>
>This is an interesting point. I've started fiddling with (a wrapper
>around) kprobes that allows me to pick any subroutine symbol in System.map,
>and then get an event whenever that subroutine gets called. Its pretty
>slick, and allows me to gather data on certain unusual events in the
>kernel.  (I'm not using this for performance monitoring, I'm trying to 
>do RAS).
>
>It makes a lot of sense to me to have a generic kprobe extension, where 
>you could give it a list of subroutine names, and it'll collect stats
>on the number of times that the routine was called. Some user-space thingy
>could poll for those stats, or you could put them in /sys or wherever.
>
>Its more complicted than just instriumenting syscalls, but a lot more
>useful, I would think ... 
>
>--linas
>
>  
>
Thanks Linas, but do you think that kprobes is safe for syscall counting?
I think that the overhead will be much bigger in that case. In my case it is
1 line asm instruction and it is defended with #ifdefs.

Kprobes is very good thing. But is not clear for me how to count 
per-thread statistics.
In my case I increased struct thread_info. In case with kprobes it is 
easy to count all
do_fork execution, for example, but not for each thread. And in any 
case, place for keeping this
counters is needed.

Maxim.

^ permalink raw reply	[flat|nested] 22+ messages in thread
* [PATCH] Performance Stats: Kernel patch
@ 2007-05-22 17:19 Maxim Uvarov
  2007-05-22 18:48 ` Dave Jones
  2007-05-22 20:08 ` Andrew Morton
  0 siblings, 2 replies; 22+ messages in thread
From: Maxim Uvarov @ 2007-05-22 17:19 UTC (permalink / raw)
  To: LKML; +Cc: linuxppc-dev

Hello Andrew,

Sorry for bothering you. I know you are very busy but could 
you please tell me what is situation of this patch? You wrote 
me you'll discuss  it with someone about it and say can it be 
accepted or not and in which form.  As I understand the 
situation all problems now are in syscall counter.  What 
about other counters?

I changed headers and re-resent patch as you asked. Are there any new?

If it could be accepted I can do fast syscalls  counting  too.


Thank you very much,
Maxim.

Patch makes available to the user the following
task and process performance statistics:
	* Involuntary Context Switches (task_struct->nivcsw)
	* Voluntary Context Switches (task_struct->nvcsw)
	* Number of system calls (added new counter
	  thread_info->sysall_count)
	           
Statistics information is available from:
        1. taskstats interface (Documentation/accounting/)
	2. /proc/PID/status (task only).

This data is useful for detecting hyperactivity
patterns between processes.

Signed-off-by: Maxim Uvarov <muvarov@ru.mvista.com>
                                                      
---

 Documentation/accounting/getdelays.c          |   20 ++++++++++++++++++--
 Documentation/accounting/taskstats-struct.txt |    7 +++++++
 arch/i386/kernel/asm-offsets.c                |    1 +
 arch/i386/kernel/entry.S                      |    3 +++
 arch/powerpc/kernel/asm-offsets.c             |    2 ++
 arch/powerpc/kernel/entry_32.S                |    5 +++++
 arch/powerpc/kernel/entry_64.S                |    5 +++++
 arch/x86_64/kernel/asm-offsets.c              |    1 +
 arch/x86_64/kernel/entry.S                    |    3 +++
 fs/proc/array.c                               |   14 ++++++++++++++
 include/asm-i386/thread_info.h                |    1 +
 include/asm-powerpc/thread_info.h             |    1 +
 include/asm-x86_64/thread_info.h              |    1 +
 include/linux/taskstats.h                     |    6 +++++-
 kernel/fork.c                                 |    3 +++
 kernel/taskstats.c                            |    6 ++++++
 16 files changed, 76 insertions(+), 3 deletions(-)

diff --git a/Documentation/accounting/getdelays.c b/Documentation/accounting/getdelays.c
index e9126e7..1be7d65 100644
--- a/Documentation/accounting/getdelays.c
+++ b/Documentation/accounting/getdelays.c
@@ -49,6 +49,7 @@ char name[100];
 int dbg;
 int print_delays;
 int print_io_accounting;
+int print_task_stats;
 __u64 stime, utime;
 
 #define PRINTF(fmt, arg...) {			\
@@ -187,7 +188,7 @@ void print_delayacct(struct taskstats *t)
 	       "IO    %15s%15s\n"
 	       "      %15llu%15llu\n"
 	       "MEM   %15s%15s\n"
-	       "      %15llu%15llu\n\n",
+	       "      %15llu%15llu\n"
 	       "count", "real total", "virtual total", "delay total",
 	       t->cpu_count, t->cpu_run_real_total, t->cpu_run_virtual_total,
 	       t->cpu_delay_total,
@@ -196,6 +197,15 @@ void print_delayacct(struct taskstats *t)
 	       "count", "delay total", t->swapin_count, t->swapin_delay_total);
 }
 
+void print_taskstats(struct taskstats *t)
+{
+	printf("\n\nTask  %15s%15s%15s\n"
+	       "      %15lu%15lu%15lu\n",
+	       "syscalls", "voluntary", "nonvoluntary",
+	       t->syscall_counter, t->nvcsw, t->nivcsw);
+
+}
+
 void print_ioacct(struct taskstats *t)
 {
 	printf("%s: read=%llu, write=%llu, cancelled_write=%llu\n",
@@ -227,7 +237,7 @@ int main(int argc, char *argv[])
 	struct msgtemplate msg;
 
 	while (1) {
-		c = getopt(argc, argv, "diw:r:m:t:p:v:l");
+		c = getopt(argc, argv, "qdiw:r:m:t:p:v:l");
 		if (c < 0)
 			break;
 
@@ -240,6 +250,10 @@ int main(int argc, char *argv[])
 			printf("printing IO accounting\n");
 			print_io_accounting = 1;
 			break;
+		case 'q':
+			printf("printing task/process stasistics:\n");
+			print_task_stats = 1;
+			break;
 		case 'w':
 			strncpy(logfile, optarg, MAX_FILENAME);
 			printf("write to file %s\n", logfile);
@@ -381,6 +395,8 @@ int main(int argc, char *argv[])
 							print_delayacct((struct taskstats *) NLA_DATA(na));
 						if (print_io_accounting)
 							print_ioacct((struct taskstats *) NLA_DATA(na));
+						if (print_task_stats)
+							print_taskstats((struct taskstats *) NLA_DATA(na));
 						if (fd) {
 							if (write(fd, NLA_DATA(na), na->nla_len) < 0) {
 								err(1,"write error\n");
diff --git a/Documentation/accounting/taskstats-struct.txt b/Documentation/accounting/taskstats-struct.txt
index 661c797..606aef6 100644
--- a/Documentation/accounting/taskstats-struct.txt
+++ b/Documentation/accounting/taskstats-struct.txt
@@ -22,6 +22,8 @@ There are three different groups of fields in the struct taskstats:
     /* Extended accounting fields end */
     Their values are collected if CONFIG_TASK_XACCT is set.
 
+4) Per-task and per-thread statistics
+
 Future extension should add fields to the end of the taskstats struct, and
 should not change the relative position of each field within the struct.
 
@@ -158,4 +160,9 @@ struct taskstats {
 
 	/* Extended accounting fields end */
 
+4) Per-task and per-thread statiscits
+	__u64	syscall_counter;	/* Syscall counter */
+	__u64	nvcsw;			/* Context voluntary switch counter */
+	__u64	nivcsw;			/* Context involuntary switch counter */
+
 }
diff --git a/arch/i386/kernel/asm-offsets.c b/arch/i386/kernel/asm-offsets.c
index 1b2f3cd..4ad49d2 100644
--- a/arch/i386/kernel/asm-offsets.c
+++ b/arch/i386/kernel/asm-offsets.c
@@ -56,6 +56,7 @@ void foo(void)
 	OFFSET(TI_addr_limit, thread_info, addr_limit);
 	OFFSET(TI_restart_block, thread_info, restart_block);
 	OFFSET(TI_sysenter_return, thread_info, sysenter_return);
+	OFFSET(TI_syscall_count, thread_info, syscall_count);
 	BLANK();
 
 	OFFSET(GDS_size, Xgt_desc_struct, size);
diff --git a/arch/i386/kernel/entry.S b/arch/i386/kernel/entry.S
index 5e47683..836961f 100644
--- a/arch/i386/kernel/entry.S
+++ b/arch/i386/kernel/entry.S
@@ -332,6 +332,9 @@ sysenter_past_esp:
 	SAVE_ALL
 	GET_THREAD_INFO(%ebp)
 
+#ifdef CONFIG_TASKSTATS
+	incl	TI_syscall_count(%ebp)	# Increment syscalls counter
+#endif
 	/* Note, _TIF_SECCOMP is bit number 8, and so it needs testw and not testb */
 	testw $(_TIF_SYSCALL_EMU|_TIF_SYSCALL_TRACE|_TIF_SECCOMP|_TIF_SYSCALL_AUDIT),TI_flags(%ebp)
 	jnz syscall_trace_entry
diff --git a/arch/powerpc/kernel/asm-offsets.c b/arch/powerpc/kernel/asm-offsets.c
index 030d300..b640039 100644
--- a/arch/powerpc/kernel/asm-offsets.c
+++ b/arch/powerpc/kernel/asm-offsets.c
@@ -94,6 +94,8 @@ int main(void)
 	DEFINE(TI_LOCAL_FLAGS, offsetof(struct thread_info, local_flags));
 	DEFINE(TI_PREEMPT, offsetof(struct thread_info, preempt_count));
 	DEFINE(TI_TASK, offsetof(struct thread_info, task));
+	DEFINE(TI_SYSCALL_COUNT, offsetof(struct thread_info, syscall_count));
+
 #ifdef CONFIG_PPC32
 	DEFINE(TI_EXECDOMAIN, offsetof(struct thread_info, exec_domain));
 	DEFINE(TI_CPU, offsetof(struct thread_info, cpu));
diff --git a/arch/powerpc/kernel/entry_32.S b/arch/powerpc/kernel/entry_32.S
index c03e829..5d919e4 100644
--- a/arch/powerpc/kernel/entry_32.S
+++ b/arch/powerpc/kernel/entry_32.S
@@ -202,6 +202,11 @@ _GLOBAL(DoSyscall)
 	bl	do_show_syscall
 #endif /* SHOW_SYSCALLS */
 	rlwinm	r10,r1,0,0,(31-THREAD_SHIFT)	/* current_thread_info() */
+#ifdef CONFIG_TASKSTATS 
+	lwz r11,TI_SYSC_CNT(r10)
+	addi r11,r11,1
+	stw r11,TI_SYSC_CNT(r10)
+#endif
 	lwz	r11,TI_FLAGS(r10)
 	andi.	r11,r11,_TIF_SYSCALL_T_OR_A
 	bne-	syscall_dotrace
diff --git a/arch/powerpc/kernel/entry_64.S b/arch/powerpc/kernel/entry_64.S
index 2551c08..5907f76 100644
--- a/arch/powerpc/kernel/entry_64.S
+++ b/arch/powerpc/kernel/entry_64.S
@@ -115,6 +115,11 @@ END_FW_FTR_SECTION_IFSET(FW_FEATURE_ISERIES)
 	addi	r9,r1,STACK_FRAME_OVERHEAD
 #endif
 	clrrdi	r11,r1,THREAD_SHIFT
+#ifdef CONFIG_TASKSTATS
+	ld r10,TI_SYSCALL_COUNT(r11)
+	addi r10,r10,1
+	std r10,TI_SYSCALL_COUNT(r11)
+#endif
 	ld	r10,TI_FLAGS(r11)
 	andi.	r11,r10,_TIF_SYSCALL_T_OR_A
 	bne-	syscall_dotrace
diff --git a/arch/x86_64/kernel/asm-offsets.c b/arch/x86_64/kernel/asm-offsets.c
index 96687e2..da57356 100644
--- a/arch/x86_64/kernel/asm-offsets.c
+++ b/arch/x86_64/kernel/asm-offsets.c
@@ -35,6 +35,7 @@ int main(void)
 	ENTRY(addr_limit);
 	ENTRY(preempt_count);
 	ENTRY(status);
+	ENTRY(syscall_count);
 	BLANK();
 #undef ENTRY
 #define ENTRY(entry) DEFINE(pda_ ## entry, offsetof(struct x8664_pda, entry))
diff --git a/arch/x86_64/kernel/entry.S b/arch/x86_64/kernel/entry.S
index 9f5dac6..af40ead 100644
--- a/arch/x86_64/kernel/entry.S
+++ b/arch/x86_64/kernel/entry.S
@@ -229,6 +229,9 @@ ENTRY(system_call)
 	movq  %rcx,RIP-ARGOFFSET(%rsp)
 	CFI_REL_OFFSET rip,RIP-ARGOFFSET
 	GET_THREAD_INFO(%rcx)
+#ifdef CONFIG_TASKSTATS
+	addq $1, threadinfo_syscall_count(%rcx)   # Increment syscalls counter
+#endif
 	testl $(_TIF_SYSCALL_TRACE|_TIF_SYSCALL_AUDIT|_TIF_SECCOMP),threadinfo_flags(%rcx)
 	jnz tracesys
 	cmpq $__NR_syscall_max,%rax
diff --git a/fs/proc/array.c b/fs/proc/array.c
index 70e4fab..c805c08 100644
--- a/fs/proc/array.c
+++ b/fs/proc/array.c
@@ -290,6 +290,19 @@ static inline char *task_cap(struct task_struct *p, char *buffer)
 			    cap_t(p->cap_permitted),
 			    cap_t(p->cap_effective));
 }
+static inline char *task_perf(struct task_struct *p, char *buffer)
+{
+	/* Syscall counter adds 1 line overhead on each syscall execution
+	 * in entry.S, so probably it is the leave this stuff under ifdefs.
+	 */
+#ifdef CONFIG_TASKSTATS
+	buffer += sprintf(buffer, "Syscalls:\t%lu\n", p->thread_info->syscall_count);
+#endif
+	return buffer + sprintf(buffer, "voluntary_ctxt_switches:\t%lu\n"
+			    "nonvoluntary_ctxt_switches:\t%lu\n",
+			    p->nvcsw,
+			    p->nivcsw);
+}
 
 int proc_pid_status(struct task_struct *task, char * buffer)
 {
@@ -309,6 +322,7 @@ int proc_pid_status(struct task_struct *task, char * buffer)
 #if defined(CONFIG_S390)
 	buffer = task_show_regs(task, buffer);
 #endif
+	buffer = task_perf(task, buffer);
 	return buffer - orig;
 }
 
diff --git a/include/asm-i386/thread_info.h b/include/asm-i386/thread_info.h
index 4b187bb..bccfd6a 100644
--- a/include/asm-i386/thread_info.h
+++ b/include/asm-i386/thread_info.h
@@ -33,6 +33,7 @@ struct thread_info {
 	int			preempt_count;	/* 0 => preemptable, <0 => BUG */
 
 
+	unsigned long		syscall_count;	/* Syscall counter */
 	mm_segment_t		addr_limit;	/* thread address space:
 					 	   0-0xBFFFFFFF for user-thead
 						   0-0xFFFFFFFF for kernel-thread
diff --git a/include/asm-powerpc/thread_info.h b/include/asm-powerpc/thread_info.h
index 3f32ca8..5306ac2 100644
--- a/include/asm-powerpc/thread_info.h
+++ b/include/asm-powerpc/thread_info.h
@@ -35,6 +35,7 @@ struct thread_info {
 	int		cpu;			/* cpu we're on */
 	int		preempt_count;		/* 0 => preemptable,
 						   <0 => BUG */
+	unsigned long	syscall_count;		/* Syscall counter */
 	struct restart_block restart_block;
 	unsigned long	local_flags;		/* private flags for thread */
 
diff --git a/include/asm-x86_64/thread_info.h b/include/asm-x86_64/thread_info.h
index 74a6c74..e53022d 100644
--- a/include/asm-x86_64/thread_info.h
+++ b/include/asm-x86_64/thread_info.h
@@ -31,6 +31,7 @@ struct thread_info {
 	__u32			cpu;		/* current CPU */
 	int 			preempt_count;	/* 0 => preemptable, <0 => BUG */
 
+	unsigned long		syscall_count;	/* Syscall counter */
 	mm_segment_t		addr_limit;	
 	struct restart_block    restart_block;
 };
diff --git a/include/linux/taskstats.h b/include/linux/taskstats.h
index 3fced47..98dfde7 100644
--- a/include/linux/taskstats.h
+++ b/include/linux/taskstats.h
@@ -31,7 +31,7 @@
  */
 
 
-#define TASKSTATS_VERSION	3
+#define TASKSTATS_VERSION	4
 #define TS_COMM_LEN		32	/* should be >= TASK_COMM_LEN
 					 * in linux/sched.h */
 
@@ -141,6 +141,10 @@ struct taskstats {
 	__u64	write_syscalls;		/* write syscalls */
 	/* Extended accounting fields end */
 
+	__u64  syscall_counter;		/* Syscall counter */
+	__u64  nvcsw;
+	__u64  nivcsw;
+
 #define TASKSTATS_HAS_IO_ACCOUNTING
 	/* Per-task storage I/O accounting starts */
 	__u64	read_bytes;		/* bytes of read I/O */
diff --git a/kernel/fork.c b/kernel/fork.c
index fc723e5..5213738 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -1042,6 +1042,9 @@ static struct task_struct *copy_process(unsigned long clone_flags,
 	p->wchar = 0;		/* I/O counter: bytes written */
 	p->syscr = 0;		/* I/O counter: read syscalls */
 	p->syscw = 0;		/* I/O counter: write syscalls */
+#ifdef CONFIG_TASKSTATS
+	p->thread_info->syscall_count = 0; /* Syscall counter: total numbers of syscalls */
+#endif
 	task_io_accounting_init(p);
 	acct_clear_integrals(p);
 
diff --git a/kernel/taskstats.c b/kernel/taskstats.c
index 4c3476f..d7bf33f 100644
--- a/kernel/taskstats.c
+++ b/kernel/taskstats.c
@@ -196,6 +196,9 @@ static int fill_pid(pid_t pid, struct task_struct *tsk,
 
 	/* fill in basic acct fields */
 	stats->version = TASKSTATS_VERSION;
+	stats->syscall_counter = tsk->thread_info->syscall_count;
+	stats->nvcsw = tsk->nvcsw;
+	stats->nivcsw = tsk->nivcsw;
 	bacct_add_tsk(stats, tsk);
 
 	/* fill in extended acct fields */
@@ -242,6 +245,9 @@ static int fill_tgid(pid_t tgid, struct task_struct *first,
 		 */
 		delayacct_add_tsk(stats, tsk);
 
+		stats->syscall_counter += tsk->thread_info->syscall_count;
+		stats->nvcsw += tsk->nvcsw;
+		stats->nivcsw += tsk->nivcsw;
 	} while_each_thread(first, tsk);
 
 	unlock_task_sighand(first, &flags);

^ permalink raw reply related	[flat|nested] 22+ messages in thread
* Re: [PATCH] Performance Stats: Kernel patch
  2007-05-22 17:19 Maxim Uvarov
@ 2007-05-22 18:48 ` Dave Jones
  2007-05-22 20:08 ` Andrew Morton
  1 sibling, 0 replies; 22+ messages in thread
From: Dave Jones @ 2007-05-22 18:48 UTC (permalink / raw)
  To: Maxim Uvarov; +Cc: linuxppc-dev, LKML

On Tue, May 22, 2007 at 05:19:52PM +0000, Maxim Uvarov wrote:

 > diff --git a/include/asm-i386/thread_info.h b/include/asm-i386/thread_info.h
 > index 4b187bb..bccfd6a 100644
 > --- a/include/asm-i386/thread_info.h
 > +++ b/include/asm-i386/thread_info.h
 > @@ -33,6 +33,7 @@ struct thread_info {
 >  	int			preempt_count;	/* 0 => preemptable, <0 => BUG */
 >  
 >  
 > +	unsigned long		syscall_count;	/* Syscall counter */
 >  	mm_segment_t		addr_limit;	/* thread address space:
 >  					 	   0-0xBFFFFFFF for user-thead
 >  						   0-0xFFFFFFFF for kernel-thread

It seems a bit unkind to bloat up the thread_info for every process
of every user when the common case will be people that don't care about
this feature at all.

Especially when the same information could be got from ptrace.

	Dave

-- 
http://www.codemonkey.org.uk

^ permalink raw reply	[flat|nested] 22+ messages in thread
* Re: [PATCH] Performance Stats: Kernel patch
  2007-05-22 17:19 Maxim Uvarov
  2007-05-22 18:48 ` Dave Jones
@ 2007-05-22 20:08 ` Andrew Morton
  1 sibling, 0 replies; 22+ messages in thread
From: Andrew Morton @ 2007-05-22 20:08 UTC (permalink / raw)
  To: Maxim Uvarov; +Cc: linuxppc-dev, LKML

On Tue, 22 May 2007 17:19:52 +0000
Maxim Uvarov <muvarov@ru.mvista.com> wrote:

> Sorry for bothering you. I know you are very busy but could 
> you please tell me what is situation of this patch?

I'd like to add the context-switch accounting to the taskstats payload.

As we'd then need to uprev the taskstats payload and version it makes sense
to have a look around, see if there's anything else which should be in
there but got missed.



I don't think we can accept the number-of-syscalls accounting feature.  It
adds a memory increment into the kernel's number-one hotpath.  Something
which people like to obsessively microbenchmark.

And as I said earlier, a 32-bit counter can be overflowed in mere seconds,
so that needs to become 64-bit, in which case we add a memory increment and
a test-n-branch to that hottest path.


There _is_ some cumulative overhead here, and I don't see how the value of
the syscall counter can justify it.

^ permalink raw reply	[flat|nested] 22+ messages in thread
end of thread, other threads:[~2007-05-22 20:21 UTC | newest]

Thread overview: 22+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2007-05-08 16:26 [PATCH] Performance Stats: Kernel patch Maxim Uvarov
2007-05-08 19:32 ` Andrew Morton
2007-05-10 11:11   ` Maxim Uvarov
2007-05-10 18:12     ` Andrew Morton
2007-05-11 16:51       ` Maxim Uvarov
2007-05-11 17:23       ` Using kprobes [was " Linas Vepstas
2007-05-12  9:15         ` Christoph Hellwig
2007-05-21 15:28         ` Maxim Uvarov
2007-05-08 23:32 ` Linas Vepstas
2007-05-10 10:22   ` Maxim Uvarov
2007-05-10 16:47     ` Linas Vepstas
  -- strict thread matches above, loose matches on Subject: below --
2007-05-10 11:42 Maxim Uvarov
2007-05-10 12:03 ` Pavel Machek
2007-05-10 18:25 ` Andrew Morton
2007-05-10 12:39 Maxim Uvarov
2007-05-10 12:38 ` Josh Boyer
2007-05-10 18:23   ` Maxim Uvarov
2007-05-11 17:13 Maxim Uvarov
2007-05-12 10:39 ` Andrea Righi
2007-05-22 17:19 Maxim Uvarov
2007-05-22 18:48 ` Dave Jones
2007-05-22 20:08 ` Andrew Morton

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).