[PATCH] oprofile: Pentium IV support

The below patch implements a P4 driver for OProfile, mostly written
by Graydon Hoare.

arch/i386/kernel/nmi.c

@@ -63,8 +63,6 @@ extern void show_registers(struct pt_regs *regs);
    CRU_ESCR0 (with any non-null event selector) through a complemented
    max threshold. [IA32-Vol3, Section 14.9.9] */
 #define MSR_P4_IQ_COUNTER0	0x30C
-#define MSR_P4_IQ_CCCR0		0x36C
-#define MSR_P4_CRU_ESCR0	0x3B8
 #define P4_NMI_CRU_ESCR0	(P4_ESCR_EVENT_SELECT(0x3F)|P4_ESCR_OS|P4_ESCR_USR)
 #define P4_NMI_IQ_CCCR0	\
 	(P4_CCCR_OVF_PMI|P4_CCCR_THRESHOLD(15)|P4_CCCR_COMPLEMENT|	\

arch/i386/oprofile/Makefile

@@ -7,4 +7,4 @@ DRIVER_OBJS = $(addprefix ../../../drivers/oprofile/, \
 
 oprofile-y				:= $(DRIVER_OBJS) init.o timer_int.o
 oprofile-$(CONFIG_X86_LOCAL_APIC) 	+= nmi_int.o op_model_athlon.o \
-					   op_model_ppro.o
+					   op_model_ppro.o op_model_p4.o

arch/i386/oprofile/nmi_int.c

@@ -214,12 +214,62 @@ struct oprofile_operations nmi_ops = {
 	.stop		= nmi_stop
 };
  
+
+#if !defined(CONFIG_X86_64)
+
+static int __init p4_init(enum oprofile_cpu * cpu)
+{
+	__u8 cpu_model = current_cpu_data.x86_model;
+
+	if (cpu_model > 3)
+		return 0;
+
+#ifndef CONFIG_SMP
+	*cpu = OPROFILE_CPU_P4;
+	model = &op_p4_spec;
+	return 1;
+#else
+	switch (smp_num_siblings) {
+		case 1:
+			*cpu = OPROFILE_CPU_P4;
+			model = &op_p4_spec;
+			return 1;
+
+		case 2:
+			*cpu = OPROFILE_CPU_P4_HT2;
+			model = &op_p4_ht2_spec;
+			return 1;
+	}
+#endif
+
+	printk(KERN_INFO "oprofile: P4 HyperThreading detected with > 2 threads\n");
+	printk(KERN_INFO "oprofile: Reverting to timer mode.\n");
+	return 0;
+}
+
+
+static int __init ppro_init(enum oprofile_cpu * cpu)
+{
+	__u8 cpu_model = current_cpu_data.x86_model;
+
+	if (cpu_model > 5) {
+		*cpu = OPROFILE_CPU_PIII;
+	} else if (cpu_model > 2) {
+		*cpu = OPROFILE_CPU_PII;
+	} else {
+		*cpu = OPROFILE_CPU_PPRO;
+	}
+
+	model = &op_ppro_spec;
+	return 1;
+}
+
+#endif /* !CONFIG_X86_64 */
  
 int __init nmi_init(struct oprofile_operations ** ops, enum oprofile_cpu * cpu)
 {
 	__u8 vendor = current_cpu_data.x86_vendor;
 	__u8 family = current_cpu_data.x86;
-	__u8 cpu_model = current_cpu_data.x86_model;
  
 	if (!cpu_has_apic)
 		return 0;
@@ -233,23 +283,26 @@ int __init nmi_init(struct oprofile_operations ** ops, enum oprofile_cpu * cpu)
 			*cpu = OPROFILE_CPU_ATHLON;
 			break;
  
-#ifndef CONFIG_X86_64
+#if !defined(CONFIG_X86_64)
 		case X86_VENDOR_INTEL:
-			/* Less than a P6-class processor */
-			if (family != 6)
-				return 0;
-	 
-			if (cpu_model > 5) {
-				*cpu = OPROFILE_CPU_PIII;
-			} else if (cpu_model > 2) {
-				*cpu = OPROFILE_CPU_PII;
-			} else {
-				*cpu = OPROFILE_CPU_PPRO;
+			switch (family) {
+				/* Pentium IV */
+				case 0xf:
+					if (!p4_init(cpu))
+						return 0;
+					break;
+
+				/* A P6-class processor */
+				case 6:
+					if (!ppro_init(cpu))
+						return 0;
+					break;
+
+				default:
+					return 0;
 			}
- 
-			model = &op_ppro_spec;
 			break;
-#endif
+#endif /* !CONFIG_X86_64 */
 
 		default:
 			return 0;

arch/i386/oprofile/op_counter.h

@@ -10,7 +10,7 @@
 #ifndef OP_COUNTER_H
 #define OP_COUNTER_H
  
-#define OP_MAX_COUNTER 4
+#define OP_MAX_COUNTER 8
  
 /* Per-perfctr configuration as set via
  * oprofilefs.

arch/i386/oprofile/op_x86_model.h

@@ -11,8 +11,8 @@
 #ifndef OP_X86_MODEL_H
 #define OP_X86_MODEL_H
 
-/* will need re-working for Pentium IV */
-#define MAX_MSR 4
+/* Pentium IV needs all these */
+#define MAX_MSR 63
  
 struct op_saved_msr {
 	unsigned int high;
@@ -47,6 +47,8 @@ struct op_x86_model_spec {
 };
 
 extern struct op_x86_model_spec const op_ppro_spec;
+extern struct op_x86_model_spec const op_p4_spec;
+extern struct op_x86_model_spec const op_p4_ht2_spec;
 extern struct op_x86_model_spec const op_athlon_spec;
 
 #endif /* OP_X86_MODEL_H */

include/asm-i386/msr.h

@@ -93,6 +93,90 @@
 #define MSR_IA32_MC0_ADDR		0x402
 #define MSR_IA32_MC0_MISC		0x403
 
+/* Pentium IV performance counter MSRs */
+#define MSR_P4_BPU_PERFCTR0 		0x300
+#define MSR_P4_BPU_PERFCTR1 		0x301
+#define MSR_P4_BPU_PERFCTR2 		0x302
+#define MSR_P4_BPU_PERFCTR3 		0x303
+#define MSR_P4_MS_PERFCTR0 		0x304
+#define MSR_P4_MS_PERFCTR1 		0x305
+#define MSR_P4_MS_PERFCTR2 		0x306
+#define MSR_P4_MS_PERFCTR3 		0x307
+#define MSR_P4_FLAME_PERFCTR0 		0x308
+#define MSR_P4_FLAME_PERFCTR1 		0x309
+#define MSR_P4_FLAME_PERFCTR2 		0x30a
+#define MSR_P4_FLAME_PERFCTR3 		0x30b
+#define MSR_P4_IQ_PERFCTR0 		0x30c
+#define MSR_P4_IQ_PERFCTR1 		0x30d
+#define MSR_P4_IQ_PERFCTR2 		0x30e
+#define MSR_P4_IQ_PERFCTR3 		0x30f
+#define MSR_P4_IQ_PERFCTR4 		0x310
+#define MSR_P4_IQ_PERFCTR5 		0x311
+#define MSR_P4_BPU_CCCR0 		0x360
+#define MSR_P4_BPU_CCCR1 		0x361
+#define MSR_P4_BPU_CCCR2 		0x362
+#define MSR_P4_BPU_CCCR3 		0x363
+#define MSR_P4_MS_CCCR0 		0x364
+#define MSR_P4_MS_CCCR1 		0x365
+#define MSR_P4_MS_CCCR2 		0x366
+#define MSR_P4_MS_CCCR3 		0x367
+#define MSR_P4_FLAME_CCCR0 		0x368
+#define MSR_P4_FLAME_CCCR1 		0x369
+#define MSR_P4_FLAME_CCCR2 		0x36a
+#define MSR_P4_FLAME_CCCR3 		0x36b
+#define MSR_P4_IQ_CCCR0 		0x36c
+#define MSR_P4_IQ_CCCR1 		0x36d
+#define MSR_P4_IQ_CCCR2 		0x36e
+#define MSR_P4_IQ_CCCR3 		0x36f
+#define MSR_P4_IQ_CCCR4 		0x370
+#define MSR_P4_IQ_CCCR5 		0x371
+#define MSR_P4_ALF_ESCR0 		0x3ca
+#define MSR_P4_ALF_ESCR1 		0x3cb
+#define MSR_P4_BPU_ESCR0 		0x3b2
+#define MSR_P4_BPU_ESCR1 		0x3b3
+#define MSR_P4_BSU_ESCR0 		0x3a0
+#define MSR_P4_BSU_ESCR1 		0x3a1
+#define MSR_P4_CRU_ESCR0 		0x3b8
+#define MSR_P4_CRU_ESCR1 		0x3b9
+#define MSR_P4_CRU_ESCR2 		0x3cc
+#define MSR_P4_CRU_ESCR3 		0x3cd
+#define MSR_P4_CRU_ESCR4 		0x3e0
+#define MSR_P4_CRU_ESCR5 		0x3e1
+#define MSR_P4_DAC_ESCR0 		0x3a8
+#define MSR_P4_DAC_ESCR1 		0x3a9
+#define MSR_P4_FIRM_ESCR0 		0x3a4
+#define MSR_P4_FIRM_ESCR1 		0x3a5
+#define MSR_P4_FLAME_ESCR0 		0x3a6
+#define MSR_P4_FLAME_ESCR1 		0x3a7
+#define MSR_P4_FSB_ESCR0 		0x3a2
+#define MSR_P4_FSB_ESCR1 		0x3a3
+#define MSR_P4_IQ_ESCR0 		0x3ba
+#define MSR_P4_IQ_ESCR1 		0x3bb
+#define MSR_P4_IS_ESCR0 		0x3b4
+#define MSR_P4_IS_ESCR1 		0x3b5
+#define MSR_P4_ITLB_ESCR0 		0x3b6
+#define MSR_P4_ITLB_ESCR1 		0x3b7
+#define MSR_P4_IX_ESCR0 		0x3c8
+#define MSR_P4_IX_ESCR1 		0x3c9
+#define MSR_P4_MOB_ESCR0 		0x3aa
+#define MSR_P4_MOB_ESCR1 		0x3ab
+#define MSR_P4_MS_ESCR0 		0x3c0
+#define MSR_P4_MS_ESCR1 		0x3c1
+#define MSR_P4_PMH_ESCR0 		0x3ac
+#define MSR_P4_PMH_ESCR1 		0x3ad
+#define MSR_P4_RAT_ESCR0 		0x3bc
+#define MSR_P4_RAT_ESCR1 		0x3bd
+#define MSR_P4_SAAT_ESCR0 		0x3ae
+#define MSR_P4_SAAT_ESCR1 		0x3af
+#define MSR_P4_SSU_ESCR0 		0x3be
+#define MSR_P4_SSU_ESCR1 		0x3bf    /* guess: not defined in manual */
+#define MSR_P4_TBPU_ESCR0 		0x3c2
+#define MSR_P4_TBPU_ESCR1 		0x3c3
+#define MSR_P4_TC_ESCR0 		0x3c4
+#define MSR_P4_TC_ESCR1 		0x3c5
+#define MSR_P4_U2L_ESCR0 		0x3b0
+#define MSR_P4_U2L_ESCR1 		0x3b1
+
 /* AMD Defined MSRs */
 #define MSR_K6_EFER			0xC0000080
 #define MSR_K6_STAR			0xC0000081

include/linux/oprofile.h

@@ -21,12 +21,22 @@ struct super_block;
 struct dentry;
 struct file_operations;
  
+/* This is duplicated from user-space so
+ * must be kept in sync :(
+ */
 enum oprofile_cpu {
 	OPROFILE_CPU_PPRO,
 	OPROFILE_CPU_PII,
 	OPROFILE_CPU_PIII,
 	OPROFILE_CPU_ATHLON,
-	OPROFILE_CPU_TIMER
+	OPROFILE_CPU_TIMER,
+	OPROFILE_UNUSED1, /* 2.4's RTC mode */
+	OPROFILE_CPU_P4,
+	OPROFILE_CPU_IA64,
+	OPROFILE_CPU_IA64_1,
+	OPROFILE_CPU_IA64_2,
+	OPROFILE_CPU_HAMMER,
+	OPROFILE_CPU_P4_HT2
 };
 
 /* Operations structure to be filled in */