perf_counter: add more context information

Put in counts to tell which ips belong to what context.

  -----
   | |  hv
   | --
nr | |  kernel
   | --
   | |  user
  -----
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Orig-LKML-Reference: <20090402091319.493101305@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

arch/x86/kernel/cpu/perf_counter.c

@@ -1088,6 +1088,7 @@ perf_callchain_kernel(struct pt_regs *regs, struct perf_callchain_entry *entry)
 {
 	unsigned long bp;
 	char *stack;
+	int nr = entry->nr;
 
 	callchain_store(entry, instruction_pointer(regs));
 
@@ -1099,6 +1100,8 @@ perf_callchain_kernel(struct pt_regs *regs, struct perf_callchain_entry *entry)
 #endif
 
 	dump_trace(NULL, regs, (void *)stack, bp, &backtrace_ops, entry);
+
+	entry->kernel = entry->nr - nr;
 }
 
 
@@ -1128,6 +1131,7 @@ perf_callchain_user(struct pt_regs *regs, struct perf_callchain_entry *entry)
 {
 	struct stack_frame frame;
 	const void __user *fp;
+	int nr = entry->nr;
 
 	regs = (struct pt_regs *)current->thread.sp0 - 1;
 	fp   = (void __user *)regs->bp;
@@ -1147,6 +1151,8 @@ perf_callchain_user(struct pt_regs *regs, struct perf_callchain_entry *entry)
 		callchain_store(entry, frame.return_address);
 		fp = frame.next_fp;
 	}
+
+	entry->user = entry->nr - nr;
 }
 
 static void
@@ -1182,6 +1188,9 @@ struct perf_callchain_entry *perf_callchain(struct pt_regs *regs)
 		entry = &__get_cpu_var(irq_entry);
 
 	entry->nr = 0;
+	entry->hv = 0;
+	entry->kernel = 0;
+	entry->user = 0;
 
 	perf_do_callchain(regs, entry);
 

include/linux/perf_counter.h

@@ -513,10 +513,10 @@ extern void perf_counter_mmap(unsigned long addr, unsigned long len,
 extern void perf_counter_munmap(unsigned long addr, unsigned long len,
 				unsigned long pgoff, struct file *file);
 
-#define MAX_STACK_DEPTH		255
+#define MAX_STACK_DEPTH		254
 
 struct perf_callchain_entry {
-	u64	nr;
+	u32	nr, hv, kernel, user;
 	u64	ip[MAX_STACK_DEPTH];
 };
 

kernel/perf_counter.c

@@ -1819,7 +1819,7 @@ void perf_counter_output(struct perf_counter *counter,
 		callchain = perf_callchain(regs);
 
 		if (callchain) {
-			callchain_size = (1 + callchain->nr) * sizeof(u64);
+			callchain_size = (2 + callchain->nr) * sizeof(u64);
 
 			header.type |= __PERF_EVENT_CALLCHAIN;
 			header.size += callchain_size;