From 1a960b402a51d80abf54e3f8e4972374ffe5f22d Mon Sep 17 00:00:00 2001 From: Jason Yeh Date: Wed, 23 Jul 2008 23:05:53 +0200 Subject: [PATCH 01/21] Oprofile Multiplexing Patch This patch introduces multiplexing support for the Oprofile kernel module. It basically adds a new function pointer in oprofile_operator allowing each architecture to supply its callback to switch between different sets of event when the timer expires. Userspace tools can modify the time slice through /dev/oprofile/time_slice. It also modifies the number of counters exposed to the userspace through /dev/oprofile. For example, the number of counters for AMD CPUs are changed to 32 and multiplexed in the sets of 4. Signed-off-by: Jason Yeh Signed-off-by: Robert Richter Cc: oprofile-list Signed-off-by: Ingo Molnar --- arch/x86/oprofile/nmi_int.c | 100 +++++++++++++++++++++++++++--- arch/x86/oprofile/op_counter.h | 3 +- arch/x86/oprofile/op_model_amd.c | 76 ++++++++++++++--------- arch/x86/oprofile/op_model_p4.c | 4 ++ arch/x86/oprofile/op_model_ppro.c | 2 + arch/x86/oprofile/op_x86_model.h | 3 + drivers/oprofile/oprof.c | 58 ++++++++++++++++- drivers/oprofile/oprof.h | 4 +- drivers/oprofile/oprofile_files.c | 39 +++++++++++- include/linux/oprofile.h | 3 + 10 files changed, 247 insertions(+), 45 deletions(-) diff --git a/arch/x86/oprofile/nmi_int.c b/arch/x86/oprofile/nmi_int.c index 287513a0981..2a65fe7680a 100644 --- a/arch/x86/oprofile/nmi_int.c +++ b/arch/x86/oprofile/nmi_int.c @@ -23,12 +23,18 @@ #include "op_counter.h" #include "op_x86_model.h" +DEFINE_PER_CPU(int, switch_index); + static struct op_x86_model_spec const *model; static DEFINE_PER_CPU(struct op_msrs, cpu_msrs); static DEFINE_PER_CPU(unsigned long, saved_lvtpc); static int nmi_start(void); static void nmi_stop(void); +static void nmi_cpu_save_mpx_registers(struct op_msrs *msrs); +static void nmi_cpu_restore_mpx_registers(struct op_msrs *msrs); +static void nmi_cpu_stop(void *dummy); +static void nmi_cpu_start(void *dummy); /* 0 == registered but off, 1 == registered and on */ static int nmi_enabled = 0; @@ -81,6 +87,47 @@ static void exit_sysfs(void) #define exit_sysfs() do { } while (0) #endif /* CONFIG_PM */ +static void nmi_cpu_switch(void *dummy) +{ + int cpu = smp_processor_id(); + int si = per_cpu(switch_index, cpu); + struct op_msrs *msrs = &per_cpu(cpu_msrs, cpu); + + nmi_cpu_stop(NULL); + nmi_cpu_save_mpx_registers(msrs); + + /* move to next set */ + si += model->num_hardware_counters; + if ((si > model->num_counters) || (counter_config[si].count == 0)) + per_cpu(switch_index, smp_processor_id()) = 0; + else + per_cpu(switch_index, smp_processor_id()) = si; + + nmi_cpu_restore_mpx_registers(msrs); + model->setup_ctrs(msrs); + nmi_cpu_start(NULL); +} + +/* + * Quick check to see if multiplexing is necessary. + * The check should be sufficient since counters are used + * in ordre. + */ +static int nmi_multiplex_on(void) +{ + return counter_config[model->num_hardware_counters].count ? 0 : -EINVAL; +} + +static int nmi_switch_event(void) +{ + if (nmi_multiplex_on() < 0) + return -EINVAL; + + on_each_cpu(nmi_cpu_switch, NULL, 0, 1); + + return 0; +} + static int profile_exceptions_notify(struct notifier_block *self, unsigned long val, void *data) { @@ -144,11 +191,10 @@ static void free_msrs(void) static int allocate_msrs(void) { - int success = 1; + int i, success = 1; size_t controls_size = sizeof(struct op_msr) * model->num_controls; size_t counters_size = sizeof(struct op_msr) * model->num_counters; - int i; for_each_possible_cpu(i) { per_cpu(cpu_msrs, i).counters = kmalloc(counters_size, GFP_KERNEL); @@ -156,8 +202,8 @@ static int allocate_msrs(void) success = 0; break; } - per_cpu(cpu_msrs, i).controls = kmalloc(controls_size, - GFP_KERNEL); + per_cpu(cpu_msrs, i).controls = + kmalloc(controls_size, GFP_KERNEL); if (!per_cpu(cpu_msrs, i).controls) { success = 0; break; @@ -201,7 +247,8 @@ static int nmi_setup(void) return err; } - /* We need to serialize save and setup for HT because the subset + /* + * We need to serialize save and setup for HT because the subset * of msrs are distinct for save and setup operations */ @@ -217,7 +264,6 @@ static int nmi_setup(void) per_cpu(cpu_msrs, 0).controls, sizeof(struct op_msr) * model->num_controls); } - } on_each_cpu(nmi_save_registers, NULL, 1); on_each_cpu(nmi_cpu_setup, NULL, 1); @@ -225,7 +271,41 @@ static int nmi_setup(void) return 0; } -static void nmi_restore_registers(struct op_msrs *msrs) +static void nmi_cpu_save_mpx_registers(struct op_msrs *msrs) +{ + unsigned int si = __get_cpu_var(switch_index); + unsigned int const nr_ctrs = model->num_hardware_counters; + struct op_msr *counters = &msrs->counters[si]; + unsigned int i; + + for (i = 0; i < nr_ctrs; ++i) { + int offset = i + si; + if (counters[offset].addr) { + rdmsr(counters[offset].addr, + counters[offset].multiplex.low, + counters[offset].multiplex.high); + } + } +} + +static void nmi_cpu_restore_mpx_registers(struct op_msrs *msrs) +{ + unsigned int si = __get_cpu_var(switch_index); + unsigned int const nr_ctrs = model->num_hardware_counters; + struct op_msr *counters = &msrs->counters[si]; + unsigned int i; + + for (i = 0; i < nr_ctrs; ++i) { + int offset = i + si; + if (counters[offset].addr) { + wrmsr(counters[offset].addr, + counters[offset].multiplex.low, + counters[offset].multiplex.high); + } + } +} + +static void nmi_cpu_restore_registers(struct op_msrs *msrs) { unsigned int const nr_ctrs = model->num_counters; unsigned int const nr_ctrls = model->num_controls; @@ -265,7 +345,8 @@ static void nmi_cpu_shutdown(void *dummy) apic_write(APIC_LVTERR, v | APIC_LVT_MASKED); apic_write(APIC_LVTPC, per_cpu(saved_lvtpc, cpu)); apic_write(APIC_LVTERR, v); - nmi_restore_registers(msrs); + nmi_cpu_restore_registers(msrs); + __get_cpu_var(switch_index) = 0; } static void nmi_shutdown(void) @@ -328,6 +409,7 @@ static int nmi_create_files(struct super_block *sb, struct dentry *root) oprofilefs_create_ulong(sb, dir, "unit_mask", &counter_config[i].unit_mask); oprofilefs_create_ulong(sb, dir, "kernel", &counter_config[i].kernel); oprofilefs_create_ulong(sb, dir, "user", &counter_config[i].user); + counter_config[i].save_count_low = 0; } return 0; @@ -469,12 +551,14 @@ int __init op_nmi_init(struct oprofile_operations *ops) } /* default values, can be overwritten by model */ + __get_cpu_var(switch_index) = 0; ops->create_files = nmi_create_files; ops->setup = nmi_setup; ops->shutdown = nmi_shutdown; ops->start = nmi_start; ops->stop = nmi_stop; ops->cpu_type = cpu_type; + ops->switch_events = nmi_switch_event; if (model->init) ret = model->init(ops); diff --git a/arch/x86/oprofile/op_counter.h b/arch/x86/oprofile/op_counter.h index 2880b15c467..786d6e01cf7 100644 --- a/arch/x86/oprofile/op_counter.h +++ b/arch/x86/oprofile/op_counter.h @@ -10,13 +10,14 @@ #ifndef OP_COUNTER_H #define OP_COUNTER_H -#define OP_MAX_COUNTER 8 +#define OP_MAX_COUNTER 32 /* Per-perfctr configuration as set via * oprofilefs. */ struct op_counter_config { unsigned long count; + unsigned long save_count_low; unsigned long enabled; unsigned long event; unsigned long kernel; diff --git a/arch/x86/oprofile/op_model_amd.c b/arch/x86/oprofile/op_model_amd.c index d9faf607b3a..bbf2b68bcc5 100644 --- a/arch/x86/oprofile/op_model_amd.c +++ b/arch/x86/oprofile/op_model_amd.c @@ -15,6 +15,7 @@ #include #include #include +#include #include #include @@ -23,8 +24,10 @@ #include "op_x86_model.h" #include "op_counter.h" -#define NUM_COUNTERS 4 -#define NUM_CONTROLS 4 +#define NUM_COUNTERS 32 +#define NUM_HARDWARE_COUNTERS 4 +#define NUM_CONTROLS 32 +#define NUM_HARDWARE_CONTROLS 4 #define CTR_IS_RESERVED(msrs, c) (msrs->counters[(c)].addr ? 1 : 0) #define CTR_READ(l, h, msrs, c) do {rdmsr(msrs->counters[(c)].addr, (l), (h)); } while (0) @@ -48,6 +51,7 @@ #define CTRL_SET_GUEST_ONLY(val, h) (val |= ((h & 1) << 8)) static unsigned long reset_value[NUM_COUNTERS]; +DECLARE_PER_CPU(int, switch_index); #ifdef CONFIG_OPROFILE_IBS @@ -130,15 +134,17 @@ static void op_amd_fill_in_addresses(struct op_msrs * const msrs) int i; for (i = 0; i < NUM_COUNTERS; i++) { - if (reserve_perfctr_nmi(MSR_K7_PERFCTR0 + i)) - msrs->counters[i].addr = MSR_K7_PERFCTR0 + i; + int hw_counter = i % NUM_HARDWARE_COUNTERS; + if (reserve_perfctr_nmi(MSR_K7_PERFCTR0 + hw_counter)) + msrs->counters[i].addr = MSR_K7_PERFCTR0 + hw_counter; else msrs->counters[i].addr = 0; } for (i = 0; i < NUM_CONTROLS; i++) { - if (reserve_evntsel_nmi(MSR_K7_EVNTSEL0 + i)) - msrs->controls[i].addr = MSR_K7_EVNTSEL0 + i; + int hw_control = i % NUM_HARDWARE_CONTROLS; + if (reserve_evntsel_nmi(MSR_K7_EVNTSEL0 + hw_control)) + msrs->controls[i].addr = MSR_K7_EVNTSEL0 + hw_control; else msrs->controls[i].addr = 0; } @@ -150,8 +156,16 @@ static void op_amd_setup_ctrs(struct op_msrs const * const msrs) unsigned int low, high; int i; + for (i = 0; i < NUM_HARDWARE_CONTROLS; ++i) { + int offset = i + __get_cpu_var(switch_index); + if (counter_config[offset].enabled) + reset_value[offset] = counter_config[offset].count; + else + reset_value[offset] = 0; + } + /* clear all counters */ - for (i = 0 ; i < NUM_CONTROLS; ++i) { + for (i = 0 ; i < NUM_HARDWARE_CONTROLS; ++i) { if (unlikely(!CTRL_IS_RESERVED(msrs, i))) continue; CTRL_READ(low, high, msrs, i); @@ -161,34 +175,31 @@ static void op_amd_setup_ctrs(struct op_msrs const * const msrs) } /* avoid a false detection of ctr overflows in NMI handler */ - for (i = 0; i < NUM_COUNTERS; ++i) { + for (i = 0; i < NUM_HARDWARE_COUNTERS; ++i) { if (unlikely(!CTR_IS_RESERVED(msrs, i))) continue; CTR_WRITE(1, msrs, i); } /* enable active counters */ - for (i = 0; i < NUM_COUNTERS; ++i) { - if ((counter_config[i].enabled) && (CTR_IS_RESERVED(msrs, i))) { - reset_value[i] = counter_config[i].count; - - CTR_WRITE(counter_config[i].count, msrs, i); + for (i = 0; i < NUM_HARDWARE_COUNTERS; ++i) { + int offset = i + __get_cpu_var(switch_index); + if ((counter_config[offset].enabled) && (CTR_IS_RESERVED(msrs, i))) { + CTR_WRITE(counter_config[offset].count, msrs, i); CTRL_READ(low, high, msrs, i); CTRL_CLEAR_LO(low); CTRL_CLEAR_HI(high); CTRL_SET_ENABLE(low); - CTRL_SET_USR(low, counter_config[i].user); - CTRL_SET_KERN(low, counter_config[i].kernel); - CTRL_SET_UM(low, counter_config[i].unit_mask); - CTRL_SET_EVENT_LOW(low, counter_config[i].event); - CTRL_SET_EVENT_HIGH(high, counter_config[i].event); + CTRL_SET_USR(low, counter_config[offset].user); + CTRL_SET_KERN(low, counter_config[offset].kernel); + CTRL_SET_UM(low, counter_config[offset].unit_mask); + CTRL_SET_EVENT_LOW(low, counter_config[offset].event); + CTRL_SET_EVENT_HIGH(high, counter_config[offset].event); CTRL_SET_HOST_ONLY(high, 0); CTRL_SET_GUEST_ONLY(high, 0); CTRL_WRITE(low, high, msrs, i); - } else { - reset_value[i] = 0; } } } @@ -276,13 +287,14 @@ static int op_amd_check_ctrs(struct pt_regs * const regs, unsigned int low, high; int i; - for (i = 0 ; i < NUM_COUNTERS; ++i) { - if (!reset_value[i]) + for (i = 0 ; i < NUM_HARDWARE_COUNTERS ; ++i) { + int offset = i + __get_cpu_var(switch_index); + if (!reset_value[offset]) continue; CTR_READ(low, high, msrs, i); if (CTR_OVERFLOWED(low)) { - oprofile_add_sample(regs, i); - CTR_WRITE(reset_value[i], msrs, i); + oprofile_add_sample(regs, offset); + CTR_WRITE(reset_value[offset], msrs, i); } } @@ -298,8 +310,10 @@ static void op_amd_start(struct op_msrs const * const msrs) { unsigned int low, high; int i; - for (i = 0 ; i < NUM_COUNTERS ; ++i) { - if (reset_value[i]) { + + for (i = 0 ; i < NUM_HARDWARE_COUNTERS ; ++i) { + int offset = i + __get_cpu_var(switch_index); + if (reset_value[offset]) { CTRL_READ(low, high, msrs, i); CTRL_SET_ACTIVE(low); CTRL_WRITE(low, high, msrs, i); @@ -329,8 +343,8 @@ static void op_amd_stop(struct op_msrs const * const msrs) /* Subtle: stop on all counters to avoid race with * setting our pm callback */ - for (i = 0 ; i < NUM_COUNTERS ; ++i) { - if (!reset_value[i]) + for (i = 0 ; i < NUM_HARDWARE_COUNTERS ; ++i) { + if (!reset_value[i + per_cpu(switch_index, smp_processor_id())]) continue; CTRL_READ(low, high, msrs, i); CTRL_SET_INACTIVE(low); @@ -356,11 +370,11 @@ static void op_amd_shutdown(struct op_msrs const * const msrs) { int i; - for (i = 0 ; i < NUM_COUNTERS ; ++i) { + for (i = 0 ; i < NUM_HARDWARE_COUNTERS ; ++i) { if (CTR_IS_RESERVED(msrs, i)) release_perfctr_nmi(MSR_K7_PERFCTR0 + i); } - for (i = 0 ; i < NUM_CONTROLS ; ++i) { + for (i = 0 ; i < NUM_HARDWARE_COUNTERS ; ++i) { if (CTRL_IS_RESERVED(msrs, i)) release_evntsel_nmi(MSR_K7_EVNTSEL0 + i); } @@ -534,6 +548,8 @@ struct op_x86_model_spec const op_amd_spec = { .exit = op_amd_exit, .num_counters = NUM_COUNTERS, .num_controls = NUM_CONTROLS, + .num_hardware_counters = NUM_HARDWARE_COUNTERS, + .num_hardware_controls = NUM_HARDWARE_CONTROLS, .fill_in_addresses = &op_amd_fill_in_addresses, .setup_ctrs = &op_amd_setup_ctrs, .check_ctrs = &op_amd_check_ctrs, diff --git a/arch/x86/oprofile/op_model_p4.c b/arch/x86/oprofile/op_model_p4.c index 56b4757a1f4..e641545d479 100644 --- a/arch/x86/oprofile/op_model_p4.c +++ b/arch/x86/oprofile/op_model_p4.c @@ -701,6 +701,8 @@ static void p4_shutdown(struct op_msrs const * const msrs) struct op_x86_model_spec const op_p4_ht2_spec = { .num_counters = NUM_COUNTERS_HT2, .num_controls = NUM_CONTROLS_HT2, + .num_hardware_counters = NUM_COUNTERS_HT2, + .num_hardware_controls = NUM_CONTROLS_HT2, .fill_in_addresses = &p4_fill_in_addresses, .setup_ctrs = &p4_setup_ctrs, .check_ctrs = &p4_check_ctrs, @@ -713,6 +715,8 @@ struct op_x86_model_spec const op_p4_ht2_spec = { struct op_x86_model_spec const op_p4_spec = { .num_counters = NUM_COUNTERS_NON_HT, .num_controls = NUM_CONTROLS_NON_HT, + .num_hardware_counters = NUM_COUNTERS_NON_HT, + .num_hardware_controls = NUM_CONTROLS_NON_HT, .fill_in_addresses = &p4_fill_in_addresses, .setup_ctrs = &p4_setup_ctrs, .check_ctrs = &p4_check_ctrs, diff --git a/arch/x86/oprofile/op_model_ppro.c b/arch/x86/oprofile/op_model_ppro.c index eff431f6c57..e5811aa480e 100644 --- a/arch/x86/oprofile/op_model_ppro.c +++ b/arch/x86/oprofile/op_model_ppro.c @@ -183,6 +183,8 @@ static void ppro_shutdown(struct op_msrs const * const msrs) struct op_x86_model_spec const op_ppro_spec = { .num_counters = NUM_COUNTERS, .num_controls = NUM_CONTROLS, + .num_hardware_counters = NUM_COUNTERS, + .num_hardware_controls = NUM_CONTROLS, .fill_in_addresses = &ppro_fill_in_addresses, .setup_ctrs = &ppro_setup_ctrs, .check_ctrs = &ppro_check_ctrs, diff --git a/arch/x86/oprofile/op_x86_model.h b/arch/x86/oprofile/op_x86_model.h index 05a0261ba0c..e07ba107637 100644 --- a/arch/x86/oprofile/op_x86_model.h +++ b/arch/x86/oprofile/op_x86_model.h @@ -19,6 +19,7 @@ struct op_saved_msr { struct op_msr { unsigned long addr; struct op_saved_msr saved; + struct op_saved_msr multiplex; }; struct op_msrs { @@ -34,6 +35,8 @@ struct pt_regs; struct op_x86_model_spec { int (*init)(struct oprofile_operations *ops); void (*exit)(void); + unsigned int const num_hardware_counters; + unsigned int const num_hardware_controls; unsigned int const num_counters; unsigned int const num_controls; void (*fill_in_addresses)(struct op_msrs * const msrs); diff --git a/drivers/oprofile/oprof.c b/drivers/oprofile/oprof.c index 2c645170f06..b2fa5df64a6 100644 --- a/drivers/oprofile/oprof.c +++ b/drivers/oprofile/oprof.c @@ -12,6 +12,8 @@ #include #include #include +#include +#include #include #include "oprof.h" @@ -19,13 +21,18 @@ #include "cpu_buffer.h" #include "buffer_sync.h" #include "oprofile_stats.h" + +static unsigned long is_setup; +static void switch_worker(struct work_struct *work); +static DECLARE_DELAYED_WORK(switch_work, switch_worker); +static DEFINE_MUTEX(start_mutex); struct oprofile_operations oprofile_ops; +unsigned long timeout_jiffies; unsigned long oprofile_started; unsigned long backtrace_depth; -static unsigned long is_setup; -static DEFINE_MUTEX(start_mutex); +/* Multiplexing defaults at 1 msec*/ /* timer 0 - use performance monitoring hardware if available @@ -87,6 +94,16 @@ out: return err; } +static void start_switch_worker(void) +{ + schedule_delayed_work(&switch_work, timeout_jiffies); +} + +static void switch_worker(struct work_struct *work) +{ + if (!oprofile_ops.switch_events()) + start_switch_worker(); +} /* Actually start profiling (echo 1>/dev/oprofile/enable) */ int oprofile_start(void) @@ -94,7 +111,6 @@ int oprofile_start(void) int err = -EINVAL; mutex_lock(&start_mutex); - if (!is_setup) goto out; @@ -108,6 +124,9 @@ int oprofile_start(void) if ((err = oprofile_ops.start())) goto out; + if (oprofile_ops.switch_events) + start_switch_worker(); + oprofile_started = 1; out: mutex_unlock(&start_mutex); @@ -123,6 +142,7 @@ void oprofile_stop(void) goto out; oprofile_ops.stop(); oprofile_started = 0; + cancel_delayed_work_sync(&switch_work); /* wake up the daemon to read what remains */ wake_up_buffer_waiter(); out: @@ -155,6 +175,32 @@ post_sync: mutex_unlock(&start_mutex); } +/* User inputs in ms, converts to jiffies */ +int oprofile_set_timeout(unsigned long val_msec) +{ + int err = 0; + + mutex_lock(&start_mutex); + + if (oprofile_started) { + err = -EBUSY; + goto out; + } + + if (!oprofile_ops.switch_events) { + err = -EINVAL; + goto out; + } + + timeout_jiffies = msecs_to_jiffies(val_msec); + if (timeout_jiffies == MAX_JIFFY_OFFSET) + timeout_jiffies = msecs_to_jiffies(1); + +out: + mutex_unlock(&start_mutex); + return err; + +} int oprofile_set_backtrace(unsigned long val) { @@ -179,10 +225,16 @@ out: return err; } +static void __init oprofile_switch_timer_init(void) +{ + timeout_jiffies = msecs_to_jiffies(1); +} + static int __init oprofile_init(void) { int err; + oprofile_switch_timer_init(); err = oprofile_arch_init(&oprofile_ops); if (err < 0 || timer) { diff --git a/drivers/oprofile/oprof.h b/drivers/oprofile/oprof.h index 18323650806..c4406a7366b 100644 --- a/drivers/oprofile/oprof.h +++ b/drivers/oprofile/oprof.h @@ -27,7 +27,8 @@ extern unsigned long fs_buffer_watershed; extern struct oprofile_operations oprofile_ops; extern unsigned long oprofile_started; extern unsigned long backtrace_depth; - +extern unsigned long timeout_jiffies; + struct super_block; struct dentry; @@ -35,5 +36,6 @@ void oprofile_create_files(struct super_block * sb, struct dentry * root); void oprofile_timer_init(struct oprofile_operations * ops); int oprofile_set_backtrace(unsigned long depth); +int oprofile_set_timeout(unsigned long time); #endif /* OPROF_H */ diff --git a/drivers/oprofile/oprofile_files.c b/drivers/oprofile/oprofile_files.c index ef953ba5ab6..cc4f5a1f8ef 100644 --- a/drivers/oprofile/oprofile_files.c +++ b/drivers/oprofile/oprofile_files.c @@ -9,6 +9,7 @@ #include #include +#include #include "event_buffer.h" #include "oprofile_stats.h" @@ -18,6 +19,40 @@ unsigned long fs_buffer_size = 131072; unsigned long fs_cpu_buffer_size = 8192; unsigned long fs_buffer_watershed = 32768; /* FIXME: tune */ +static ssize_t timeout_read(struct file *file, char __user *buf, + size_t count, loff_t *offset) +{ + return oprofilefs_ulong_to_user(jiffies_to_msecs(timeout_jiffies), + buf, count, offset); +} + + +static ssize_t timeout_write(struct file *file, char const __user *buf, + size_t count, loff_t *offset) +{ + unsigned long val; + int retval; + + if (*offset) + return -EINVAL; + + retval = oprofilefs_ulong_from_user(&val, buf, count); + if (retval) + return retval; + + retval = oprofile_set_timeout(val); + + if (retval) + return retval; + return count; +} + +static const struct file_operations timeout_fops = { + .read = timeout_read, + .write = timeout_write, +}; + + static ssize_t depth_read(struct file * file, char __user * buf, size_t count, loff_t * offset) { return oprofilefs_ulong_to_user(backtrace_depth, buf, count, offset); @@ -85,11 +120,10 @@ static ssize_t enable_write(struct file * file, char const __user * buf, size_t if (*offset) return -EINVAL; - retval = oprofilefs_ulong_from_user(&val, buf, count); if (retval) return retval; - + if (val) retval = oprofile_start(); else @@ -129,6 +163,7 @@ void oprofile_create_files(struct super_block * sb, struct dentry * root) oprofilefs_create_file(sb, root, "cpu_type", &cpu_type_fops); oprofilefs_create_file(sb, root, "backtrace_depth", &depth_fops); oprofilefs_create_file(sb, root, "pointer_size", &pointer_size_fops); + oprofilefs_create_file(sb, root, "timeout_ms", &timeout_fops); oprofile_create_stats_files(sb, root); if (oprofile_ops.create_files) oprofile_ops.create_files(sb, root); diff --git a/include/linux/oprofile.h b/include/linux/oprofile.h index bcb8f725427..687f2f4c36a 100644 --- a/include/linux/oprofile.h +++ b/include/linux/oprofile.h @@ -67,6 +67,9 @@ struct oprofile_operations { /* Initiate a stack backtrace. Optional. */ void (*backtrace)(struct pt_regs * const regs, unsigned int depth); + + /* Multiplex between different events. Optional. */ + int (*switch_events)(void); /* CPU identification string. */ char * cpu_type; }; From 7e7b43892b87b6be259479ef4de14029dcb4012f Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Thu, 24 Jul 2008 16:00:16 +0200 Subject: [PATCH 02/21] x86/oprofile: fix on_each_cpu build error Signed-off-by: Robert Richter Cc: oprofile-list Cc: Jason Yeh Signed-off-by: Ingo Molnar --- arch/x86/oprofile/nmi_int.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/x86/oprofile/nmi_int.c b/arch/x86/oprofile/nmi_int.c index 2a65fe7680a..fb4902bc6f1 100644 --- a/arch/x86/oprofile/nmi_int.c +++ b/arch/x86/oprofile/nmi_int.c @@ -123,7 +123,7 @@ static int nmi_switch_event(void) if (nmi_multiplex_on() < 0) return -EINVAL; - on_each_cpu(nmi_cpu_switch, NULL, 0, 1); + on_each_cpu(nmi_cpu_switch, NULL, 1); return 0; } From 45f197ade73ba95681b9803680c75352fc0a1c0a Mon Sep 17 00:00:00 2001 From: Andrea Righi Date: Sat, 20 Sep 2008 12:58:40 +0200 Subject: [PATCH 03/21] x86, oprofile: BUG: using smp_processor_id() in preemptible code Add __raw access before setting per cpu variable switch_index, to avoid the following BUG: [ 449.166827] BUG: using smp_processor_id() in preemptible [00000000] code: modprobe/6998 [ 449.166848] caller is op_nmi_init+0xf0/0x2b0 [oprofile] [ 449.166855] Pid: 6998, comm: modprobe Not tainted 2.6.27-rc5-mm1 #29 [ 449.166860] Call Trace: [ 449.166872] [] debug_smp_processor_id+0xd7/0xe0 [ 449.166887] [] op_nmi_init+0xf0/0x2b0 [oprofile] [ 449.166902] [] oprofile_init+0x0/0x60 [oprofile] [ 449.166915] [] oprofile_arch_init+0x9/0x30 [oprofile] [ 449.166928] [] oprofile_init+0x1e/0x60 [oprofile] [ 449.166937] [] _stext+0x3b/0x160 [ 449.166946] [] __mutex_unlock_slowpath+0xe5/0x190 [ 449.166955] [] trace_hardirqs_on_caller+0xca/0x140 [ 449.166965] [] sys_init_module+0xdc/0x210 [ 449.166972] [] system_call_fastpath+0x16/0x1b Signed-off-by: Andrea Righi Acked-by: Robert Richter Signed-off-by: Ingo Molnar --- arch/x86/oprofile/nmi_int.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/x86/oprofile/nmi_int.c b/arch/x86/oprofile/nmi_int.c index fb4902bc6f1..4108d02c529 100644 --- a/arch/x86/oprofile/nmi_int.c +++ b/arch/x86/oprofile/nmi_int.c @@ -551,7 +551,7 @@ int __init op_nmi_init(struct oprofile_operations *ops) } /* default values, can be overwritten by model */ - __get_cpu_var(switch_index) = 0; + __raw_get_cpu_var(switch_index) = 0; ops->create_files = nmi_create_files; ops->setup = nmi_setup; ops->shutdown = nmi_shutdown; From 4c168eaf7ea39f25a45a3d8c7eebc3fedb633a1d Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Wed, 24 Sep 2008 11:08:52 +0200 Subject: [PATCH 04/21] Revert "Oprofile Multiplexing Patch" Reverting commit 1a960b402a51d80abf54e3f8e4972374ffe5f22d for the main branch. Multiplexing will be tracked on a separate feature branch. Conflicts: arch/x86/oprofile/nmi_int.c --- arch/x86/oprofile/nmi_int.c | 100 +++--------------------------- arch/x86/oprofile/op_counter.h | 3 +- arch/x86/oprofile/op_model_amd.c | 76 +++++++++-------------- arch/x86/oprofile/op_model_p4.c | 4 -- arch/x86/oprofile/op_model_ppro.c | 2 - arch/x86/oprofile/op_x86_model.h | 3 - drivers/oprofile/oprof.c | 58 +---------------- drivers/oprofile/oprof.h | 4 +- drivers/oprofile/oprofile_files.c | 39 +----------- include/linux/oprofile.h | 3 - 10 files changed, 45 insertions(+), 247 deletions(-) diff --git a/arch/x86/oprofile/nmi_int.c b/arch/x86/oprofile/nmi_int.c index 4108d02c529..287513a0981 100644 --- a/arch/x86/oprofile/nmi_int.c +++ b/arch/x86/oprofile/nmi_int.c @@ -23,18 +23,12 @@ #include "op_counter.h" #include "op_x86_model.h" -DEFINE_PER_CPU(int, switch_index); - static struct op_x86_model_spec const *model; static DEFINE_PER_CPU(struct op_msrs, cpu_msrs); static DEFINE_PER_CPU(unsigned long, saved_lvtpc); static int nmi_start(void); static void nmi_stop(void); -static void nmi_cpu_save_mpx_registers(struct op_msrs *msrs); -static void nmi_cpu_restore_mpx_registers(struct op_msrs *msrs); -static void nmi_cpu_stop(void *dummy); -static void nmi_cpu_start(void *dummy); /* 0 == registered but off, 1 == registered and on */ static int nmi_enabled = 0; @@ -87,47 +81,6 @@ static void exit_sysfs(void) #define exit_sysfs() do { } while (0) #endif /* CONFIG_PM */ -static void nmi_cpu_switch(void *dummy) -{ - int cpu = smp_processor_id(); - int si = per_cpu(switch_index, cpu); - struct op_msrs *msrs = &per_cpu(cpu_msrs, cpu); - - nmi_cpu_stop(NULL); - nmi_cpu_save_mpx_registers(msrs); - - /* move to next set */ - si += model->num_hardware_counters; - if ((si > model->num_counters) || (counter_config[si].count == 0)) - per_cpu(switch_index, smp_processor_id()) = 0; - else - per_cpu(switch_index, smp_processor_id()) = si; - - nmi_cpu_restore_mpx_registers(msrs); - model->setup_ctrs(msrs); - nmi_cpu_start(NULL); -} - -/* - * Quick check to see if multiplexing is necessary. - * The check should be sufficient since counters are used - * in ordre. - */ -static int nmi_multiplex_on(void) -{ - return counter_config[model->num_hardware_counters].count ? 0 : -EINVAL; -} - -static int nmi_switch_event(void) -{ - if (nmi_multiplex_on() < 0) - return -EINVAL; - - on_each_cpu(nmi_cpu_switch, NULL, 1); - - return 0; -} - static int profile_exceptions_notify(struct notifier_block *self, unsigned long val, void *data) { @@ -191,10 +144,11 @@ static void free_msrs(void) static int allocate_msrs(void) { - int i, success = 1; + int success = 1; size_t controls_size = sizeof(struct op_msr) * model->num_controls; size_t counters_size = sizeof(struct op_msr) * model->num_counters; + int i; for_each_possible_cpu(i) { per_cpu(cpu_msrs, i).counters = kmalloc(counters_size, GFP_KERNEL); @@ -202,8 +156,8 @@ static int allocate_msrs(void) success = 0; break; } - per_cpu(cpu_msrs, i).controls = - kmalloc(controls_size, GFP_KERNEL); + per_cpu(cpu_msrs, i).controls = kmalloc(controls_size, + GFP_KERNEL); if (!per_cpu(cpu_msrs, i).controls) { success = 0; break; @@ -247,8 +201,7 @@ static int nmi_setup(void) return err; } - /* - * We need to serialize save and setup for HT because the subset + /* We need to serialize save and setup for HT because the subset * of msrs are distinct for save and setup operations */ @@ -264,6 +217,7 @@ static int nmi_setup(void) per_cpu(cpu_msrs, 0).controls, sizeof(struct op_msr) * model->num_controls); } + } on_each_cpu(nmi_save_registers, NULL, 1); on_each_cpu(nmi_cpu_setup, NULL, 1); @@ -271,41 +225,7 @@ static int nmi_setup(void) return 0; } -static void nmi_cpu_save_mpx_registers(struct op_msrs *msrs) -{ - unsigned int si = __get_cpu_var(switch_index); - unsigned int const nr_ctrs = model->num_hardware_counters; - struct op_msr *counters = &msrs->counters[si]; - unsigned int i; - - for (i = 0; i < nr_ctrs; ++i) { - int offset = i + si; - if (counters[offset].addr) { - rdmsr(counters[offset].addr, - counters[offset].multiplex.low, - counters[offset].multiplex.high); - } - } -} - -static void nmi_cpu_restore_mpx_registers(struct op_msrs *msrs) -{ - unsigned int si = __get_cpu_var(switch_index); - unsigned int const nr_ctrs = model->num_hardware_counters; - struct op_msr *counters = &msrs->counters[si]; - unsigned int i; - - for (i = 0; i < nr_ctrs; ++i) { - int offset = i + si; - if (counters[offset].addr) { - wrmsr(counters[offset].addr, - counters[offset].multiplex.low, - counters[offset].multiplex.high); - } - } -} - -static void nmi_cpu_restore_registers(struct op_msrs *msrs) +static void nmi_restore_registers(struct op_msrs *msrs) { unsigned int const nr_ctrs = model->num_counters; unsigned int const nr_ctrls = model->num_controls; @@ -345,8 +265,7 @@ static void nmi_cpu_shutdown(void *dummy) apic_write(APIC_LVTERR, v | APIC_LVT_MASKED); apic_write(APIC_LVTPC, per_cpu(saved_lvtpc, cpu)); apic_write(APIC_LVTERR, v); - nmi_cpu_restore_registers(msrs); - __get_cpu_var(switch_index) = 0; + nmi_restore_registers(msrs); } static void nmi_shutdown(void) @@ -409,7 +328,6 @@ static int nmi_create_files(struct super_block *sb, struct dentry *root) oprofilefs_create_ulong(sb, dir, "unit_mask", &counter_config[i].unit_mask); oprofilefs_create_ulong(sb, dir, "kernel", &counter_config[i].kernel); oprofilefs_create_ulong(sb, dir, "user", &counter_config[i].user); - counter_config[i].save_count_low = 0; } return 0; @@ -551,14 +469,12 @@ int __init op_nmi_init(struct oprofile_operations *ops) } /* default values, can be overwritten by model */ - __raw_get_cpu_var(switch_index) = 0; ops->create_files = nmi_create_files; ops->setup = nmi_setup; ops->shutdown = nmi_shutdown; ops->start = nmi_start; ops->stop = nmi_stop; ops->cpu_type = cpu_type; - ops->switch_events = nmi_switch_event; if (model->init) ret = model->init(ops); diff --git a/arch/x86/oprofile/op_counter.h b/arch/x86/oprofile/op_counter.h index 786d6e01cf7..2880b15c467 100644 --- a/arch/x86/oprofile/op_counter.h +++ b/arch/x86/oprofile/op_counter.h @@ -10,14 +10,13 @@ #ifndef OP_COUNTER_H #define OP_COUNTER_H -#define OP_MAX_COUNTER 32 +#define OP_MAX_COUNTER 8 /* Per-perfctr configuration as set via * oprofilefs. */ struct op_counter_config { unsigned long count; - unsigned long save_count_low; unsigned long enabled; unsigned long event; unsigned long kernel; diff --git a/arch/x86/oprofile/op_model_amd.c b/arch/x86/oprofile/op_model_amd.c index bbf2b68bcc5..d9faf607b3a 100644 --- a/arch/x86/oprofile/op_model_amd.c +++ b/arch/x86/oprofile/op_model_amd.c @@ -15,7 +15,6 @@ #include #include #include -#include #include #include @@ -24,10 +23,8 @@ #include "op_x86_model.h" #include "op_counter.h" -#define NUM_COUNTERS 32 -#define NUM_HARDWARE_COUNTERS 4 -#define NUM_CONTROLS 32 -#define NUM_HARDWARE_CONTROLS 4 +#define NUM_COUNTERS 4 +#define NUM_CONTROLS 4 #define CTR_IS_RESERVED(msrs, c) (msrs->counters[(c)].addr ? 1 : 0) #define CTR_READ(l, h, msrs, c) do {rdmsr(msrs->counters[(c)].addr, (l), (h)); } while (0) @@ -51,7 +48,6 @@ #define CTRL_SET_GUEST_ONLY(val, h) (val |= ((h & 1) << 8)) static unsigned long reset_value[NUM_COUNTERS]; -DECLARE_PER_CPU(int, switch_index); #ifdef CONFIG_OPROFILE_IBS @@ -134,17 +130,15 @@ static void op_amd_fill_in_addresses(struct op_msrs * const msrs) int i; for (i = 0; i < NUM_COUNTERS; i++) { - int hw_counter = i % NUM_HARDWARE_COUNTERS; - if (reserve_perfctr_nmi(MSR_K7_PERFCTR0 + hw_counter)) - msrs->counters[i].addr = MSR_K7_PERFCTR0 + hw_counter; + if (reserve_perfctr_nmi(MSR_K7_PERFCTR0 + i)) + msrs->counters[i].addr = MSR_K7_PERFCTR0 + i; else msrs->counters[i].addr = 0; } for (i = 0; i < NUM_CONTROLS; i++) { - int hw_control = i % NUM_HARDWARE_CONTROLS; - if (reserve_evntsel_nmi(MSR_K7_EVNTSEL0 + hw_control)) - msrs->controls[i].addr = MSR_K7_EVNTSEL0 + hw_control; + if (reserve_evntsel_nmi(MSR_K7_EVNTSEL0 + i)) + msrs->controls[i].addr = MSR_K7_EVNTSEL0 + i; else msrs->controls[i].addr = 0; } @@ -156,16 +150,8 @@ static void op_amd_setup_ctrs(struct op_msrs const * const msrs) unsigned int low, high; int i; - for (i = 0; i < NUM_HARDWARE_CONTROLS; ++i) { - int offset = i + __get_cpu_var(switch_index); - if (counter_config[offset].enabled) - reset_value[offset] = counter_config[offset].count; - else - reset_value[offset] = 0; - } - /* clear all counters */ - for (i = 0 ; i < NUM_HARDWARE_CONTROLS; ++i) { + for (i = 0 ; i < NUM_CONTROLS; ++i) { if (unlikely(!CTRL_IS_RESERVED(msrs, i))) continue; CTRL_READ(low, high, msrs, i); @@ -175,31 +161,34 @@ static void op_amd_setup_ctrs(struct op_msrs const * const msrs) } /* avoid a false detection of ctr overflows in NMI handler */ - for (i = 0; i < NUM_HARDWARE_COUNTERS; ++i) { + for (i = 0; i < NUM_COUNTERS; ++i) { if (unlikely(!CTR_IS_RESERVED(msrs, i))) continue; CTR_WRITE(1, msrs, i); } /* enable active counters */ - for (i = 0; i < NUM_HARDWARE_COUNTERS; ++i) { - int offset = i + __get_cpu_var(switch_index); - if ((counter_config[offset].enabled) && (CTR_IS_RESERVED(msrs, i))) { - CTR_WRITE(counter_config[offset].count, msrs, i); + for (i = 0; i < NUM_COUNTERS; ++i) { + if ((counter_config[i].enabled) && (CTR_IS_RESERVED(msrs, i))) { + reset_value[i] = counter_config[i].count; + + CTR_WRITE(counter_config[i].count, msrs, i); CTRL_READ(low, high, msrs, i); CTRL_CLEAR_LO(low); CTRL_CLEAR_HI(high); CTRL_SET_ENABLE(low); - CTRL_SET_USR(low, counter_config[offset].user); - CTRL_SET_KERN(low, counter_config[offset].kernel); - CTRL_SET_UM(low, counter_config[offset].unit_mask); - CTRL_SET_EVENT_LOW(low, counter_config[offset].event); - CTRL_SET_EVENT_HIGH(high, counter_config[offset].event); + CTRL_SET_USR(low, counter_config[i].user); + CTRL_SET_KERN(low, counter_config[i].kernel); + CTRL_SET_UM(low, counter_config[i].unit_mask); + CTRL_SET_EVENT_LOW(low, counter_config[i].event); + CTRL_SET_EVENT_HIGH(high, counter_config[i].event); CTRL_SET_HOST_ONLY(high, 0); CTRL_SET_GUEST_ONLY(high, 0); CTRL_WRITE(low, high, msrs, i); + } else { + reset_value[i] = 0; } } } @@ -287,14 +276,13 @@ static int op_amd_check_ctrs(struct pt_regs * const regs, unsigned int low, high; int i; - for (i = 0 ; i < NUM_HARDWARE_COUNTERS ; ++i) { - int offset = i + __get_cpu_var(switch_index); - if (!reset_value[offset]) + for (i = 0 ; i < NUM_COUNTERS; ++i) { + if (!reset_value[i]) continue; CTR_READ(low, high, msrs, i); if (CTR_OVERFLOWED(low)) { - oprofile_add_sample(regs, offset); - CTR_WRITE(reset_value[offset], msrs, i); + oprofile_add_sample(regs, i); + CTR_WRITE(reset_value[i], msrs, i); } } @@ -310,10 +298,8 @@ static void op_amd_start(struct op_msrs const * const msrs) { unsigned int low, high; int i; - - for (i = 0 ; i < NUM_HARDWARE_COUNTERS ; ++i) { - int offset = i + __get_cpu_var(switch_index); - if (reset_value[offset]) { + for (i = 0 ; i < NUM_COUNTERS ; ++i) { + if (reset_value[i]) { CTRL_READ(low, high, msrs, i); CTRL_SET_ACTIVE(low); CTRL_WRITE(low, high, msrs, i); @@ -343,8 +329,8 @@ static void op_amd_stop(struct op_msrs const * const msrs) /* Subtle: stop on all counters to avoid race with * setting our pm callback */ - for (i = 0 ; i < NUM_HARDWARE_COUNTERS ; ++i) { - if (!reset_value[i + per_cpu(switch_index, smp_processor_id())]) + for (i = 0 ; i < NUM_COUNTERS ; ++i) { + if (!reset_value[i]) continue; CTRL_READ(low, high, msrs, i); CTRL_SET_INACTIVE(low); @@ -370,11 +356,11 @@ static void op_amd_shutdown(struct op_msrs const * const msrs) { int i; - for (i = 0 ; i < NUM_HARDWARE_COUNTERS ; ++i) { + for (i = 0 ; i < NUM_COUNTERS ; ++i) { if (CTR_IS_RESERVED(msrs, i)) release_perfctr_nmi(MSR_K7_PERFCTR0 + i); } - for (i = 0 ; i < NUM_HARDWARE_COUNTERS ; ++i) { + for (i = 0 ; i < NUM_CONTROLS ; ++i) { if (CTRL_IS_RESERVED(msrs, i)) release_evntsel_nmi(MSR_K7_EVNTSEL0 + i); } @@ -548,8 +534,6 @@ struct op_x86_model_spec const op_amd_spec = { .exit = op_amd_exit, .num_counters = NUM_COUNTERS, .num_controls = NUM_CONTROLS, - .num_hardware_counters = NUM_HARDWARE_COUNTERS, - .num_hardware_controls = NUM_HARDWARE_CONTROLS, .fill_in_addresses = &op_amd_fill_in_addresses, .setup_ctrs = &op_amd_setup_ctrs, .check_ctrs = &op_amd_check_ctrs, diff --git a/arch/x86/oprofile/op_model_p4.c b/arch/x86/oprofile/op_model_p4.c index cacba61ffba..43ac5af338d 100644 --- a/arch/x86/oprofile/op_model_p4.c +++ b/arch/x86/oprofile/op_model_p4.c @@ -700,8 +700,6 @@ static void p4_shutdown(struct op_msrs const * const msrs) struct op_x86_model_spec const op_p4_ht2_spec = { .num_counters = NUM_COUNTERS_HT2, .num_controls = NUM_CONTROLS_HT2, - .num_hardware_counters = NUM_COUNTERS_HT2, - .num_hardware_controls = NUM_CONTROLS_HT2, .fill_in_addresses = &p4_fill_in_addresses, .setup_ctrs = &p4_setup_ctrs, .check_ctrs = &p4_check_ctrs, @@ -714,8 +712,6 @@ struct op_x86_model_spec const op_p4_ht2_spec = { struct op_x86_model_spec const op_p4_spec = { .num_counters = NUM_COUNTERS_NON_HT, .num_controls = NUM_CONTROLS_NON_HT, - .num_hardware_counters = NUM_COUNTERS_NON_HT, - .num_hardware_controls = NUM_CONTROLS_NON_HT, .fill_in_addresses = &p4_fill_in_addresses, .setup_ctrs = &p4_setup_ctrs, .check_ctrs = &p4_check_ctrs, diff --git a/arch/x86/oprofile/op_model_ppro.c b/arch/x86/oprofile/op_model_ppro.c index e5811aa480e..eff431f6c57 100644 --- a/arch/x86/oprofile/op_model_ppro.c +++ b/arch/x86/oprofile/op_model_ppro.c @@ -183,8 +183,6 @@ static void ppro_shutdown(struct op_msrs const * const msrs) struct op_x86_model_spec const op_ppro_spec = { .num_counters = NUM_COUNTERS, .num_controls = NUM_CONTROLS, - .num_hardware_counters = NUM_COUNTERS, - .num_hardware_controls = NUM_CONTROLS, .fill_in_addresses = &ppro_fill_in_addresses, .setup_ctrs = &ppro_setup_ctrs, .check_ctrs = &ppro_check_ctrs, diff --git a/arch/x86/oprofile/op_x86_model.h b/arch/x86/oprofile/op_x86_model.h index e07ba107637..05a0261ba0c 100644 --- a/arch/x86/oprofile/op_x86_model.h +++ b/arch/x86/oprofile/op_x86_model.h @@ -19,7 +19,6 @@ struct op_saved_msr { struct op_msr { unsigned long addr; struct op_saved_msr saved; - struct op_saved_msr multiplex; }; struct op_msrs { @@ -35,8 +34,6 @@ struct pt_regs; struct op_x86_model_spec { int (*init)(struct oprofile_operations *ops); void (*exit)(void); - unsigned int const num_hardware_counters; - unsigned int const num_hardware_controls; unsigned int const num_counters; unsigned int const num_controls; void (*fill_in_addresses)(struct op_msrs * const msrs); diff --git a/drivers/oprofile/oprof.c b/drivers/oprofile/oprof.c index b2fa5df64a6..2c645170f06 100644 --- a/drivers/oprofile/oprof.c +++ b/drivers/oprofile/oprof.c @@ -12,8 +12,6 @@ #include #include #include -#include -#include #include #include "oprof.h" @@ -21,18 +19,13 @@ #include "cpu_buffer.h" #include "buffer_sync.h" #include "oprofile_stats.h" - -static unsigned long is_setup; -static void switch_worker(struct work_struct *work); -static DECLARE_DELAYED_WORK(switch_work, switch_worker); -static DEFINE_MUTEX(start_mutex); struct oprofile_operations oprofile_ops; -unsigned long timeout_jiffies; unsigned long oprofile_started; unsigned long backtrace_depth; -/* Multiplexing defaults at 1 msec*/ +static unsigned long is_setup; +static DEFINE_MUTEX(start_mutex); /* timer 0 - use performance monitoring hardware if available @@ -94,16 +87,6 @@ out: return err; } -static void start_switch_worker(void) -{ - schedule_delayed_work(&switch_work, timeout_jiffies); -} - -static void switch_worker(struct work_struct *work) -{ - if (!oprofile_ops.switch_events()) - start_switch_worker(); -} /* Actually start profiling (echo 1>/dev/oprofile/enable) */ int oprofile_start(void) @@ -111,6 +94,7 @@ int oprofile_start(void) int err = -EINVAL; mutex_lock(&start_mutex); + if (!is_setup) goto out; @@ -124,9 +108,6 @@ int oprofile_start(void) if ((err = oprofile_ops.start())) goto out; - if (oprofile_ops.switch_events) - start_switch_worker(); - oprofile_started = 1; out: mutex_unlock(&start_mutex); @@ -142,7 +123,6 @@ void oprofile_stop(void) goto out; oprofile_ops.stop(); oprofile_started = 0; - cancel_delayed_work_sync(&switch_work); /* wake up the daemon to read what remains */ wake_up_buffer_waiter(); out: @@ -175,32 +155,6 @@ post_sync: mutex_unlock(&start_mutex); } -/* User inputs in ms, converts to jiffies */ -int oprofile_set_timeout(unsigned long val_msec) -{ - int err = 0; - - mutex_lock(&start_mutex); - - if (oprofile_started) { - err = -EBUSY; - goto out; - } - - if (!oprofile_ops.switch_events) { - err = -EINVAL; - goto out; - } - - timeout_jiffies = msecs_to_jiffies(val_msec); - if (timeout_jiffies == MAX_JIFFY_OFFSET) - timeout_jiffies = msecs_to_jiffies(1); - -out: - mutex_unlock(&start_mutex); - return err; - -} int oprofile_set_backtrace(unsigned long val) { @@ -225,16 +179,10 @@ out: return err; } -static void __init oprofile_switch_timer_init(void) -{ - timeout_jiffies = msecs_to_jiffies(1); -} - static int __init oprofile_init(void) { int err; - oprofile_switch_timer_init(); err = oprofile_arch_init(&oprofile_ops); if (err < 0 || timer) { diff --git a/drivers/oprofile/oprof.h b/drivers/oprofile/oprof.h index c4406a7366b..18323650806 100644 --- a/drivers/oprofile/oprof.h +++ b/drivers/oprofile/oprof.h @@ -27,8 +27,7 @@ extern unsigned long fs_buffer_watershed; extern struct oprofile_operations oprofile_ops; extern unsigned long oprofile_started; extern unsigned long backtrace_depth; -extern unsigned long timeout_jiffies; - + struct super_block; struct dentry; @@ -36,6 +35,5 @@ void oprofile_create_files(struct super_block * sb, struct dentry * root); void oprofile_timer_init(struct oprofile_operations * ops); int oprofile_set_backtrace(unsigned long depth); -int oprofile_set_timeout(unsigned long time); #endif /* OPROF_H */ diff --git a/drivers/oprofile/oprofile_files.c b/drivers/oprofile/oprofile_files.c index cc4f5a1f8ef..ef953ba5ab6 100644 --- a/drivers/oprofile/oprofile_files.c +++ b/drivers/oprofile/oprofile_files.c @@ -9,7 +9,6 @@ #include #include -#include #include "event_buffer.h" #include "oprofile_stats.h" @@ -19,40 +18,6 @@ unsigned long fs_buffer_size = 131072; unsigned long fs_cpu_buffer_size = 8192; unsigned long fs_buffer_watershed = 32768; /* FIXME: tune */ -static ssize_t timeout_read(struct file *file, char __user *buf, - size_t count, loff_t *offset) -{ - return oprofilefs_ulong_to_user(jiffies_to_msecs(timeout_jiffies), - buf, count, offset); -} - - -static ssize_t timeout_write(struct file *file, char const __user *buf, - size_t count, loff_t *offset) -{ - unsigned long val; - int retval; - - if (*offset) - return -EINVAL; - - retval = oprofilefs_ulong_from_user(&val, buf, count); - if (retval) - return retval; - - retval = oprofile_set_timeout(val); - - if (retval) - return retval; - return count; -} - -static const struct file_operations timeout_fops = { - .read = timeout_read, - .write = timeout_write, -}; - - static ssize_t depth_read(struct file * file, char __user * buf, size_t count, loff_t * offset) { return oprofilefs_ulong_to_user(backtrace_depth, buf, count, offset); @@ -120,10 +85,11 @@ static ssize_t enable_write(struct file * file, char const __user * buf, size_t if (*offset) return -EINVAL; + retval = oprofilefs_ulong_from_user(&val, buf, count); if (retval) return retval; - + if (val) retval = oprofile_start(); else @@ -163,7 +129,6 @@ void oprofile_create_files(struct super_block * sb, struct dentry * root) oprofilefs_create_file(sb, root, "cpu_type", &cpu_type_fops); oprofilefs_create_file(sb, root, "backtrace_depth", &depth_fops); oprofilefs_create_file(sb, root, "pointer_size", &pointer_size_fops); - oprofilefs_create_file(sb, root, "timeout_ms", &timeout_fops); oprofile_create_stats_files(sb, root); if (oprofile_ops.create_files) oprofile_ops.create_files(sb, root); diff --git a/include/linux/oprofile.h b/include/linux/oprofile.h index 687f2f4c36a..bcb8f725427 100644 --- a/include/linux/oprofile.h +++ b/include/linux/oprofile.h @@ -67,9 +67,6 @@ struct oprofile_operations { /* Initiate a stack backtrace. Optional. */ void (*backtrace)(struct pt_regs * const regs, unsigned int depth); - - /* Multiplex between different events. Optional. */ - int (*switch_events)(void); /* CPU identification string. */ char * cpu_type; }; From 5d4488027d9cf3161c71566dfabb116bf69ab4d9 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Mon, 18 Aug 2008 14:49:47 +0200 Subject: [PATCH 05/21] oprofile: drop const in num counters field allow to modify it at runtime Signed-off-by: Andi Kleen Signed-off-by: Robert Richter --- arch/x86/oprofile/op_x86_model.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/arch/x86/oprofile/op_x86_model.h b/arch/x86/oprofile/op_x86_model.h index 05a0261ba0c..3d3b85d3c25 100644 --- a/arch/x86/oprofile/op_x86_model.h +++ b/arch/x86/oprofile/op_x86_model.h @@ -34,8 +34,8 @@ struct pt_regs; struct op_x86_model_spec { int (*init)(struct oprofile_operations *ops); void (*exit)(void); - unsigned int const num_counters; - unsigned int const num_controls; + unsigned int num_counters; + unsigned int num_controls; void (*fill_in_addresses)(struct op_msrs * const msrs); void (*setup_ctrs)(struct op_msrs const * const msrs); int (*check_ctrs)(struct pt_regs * const regs, From f645f6406463a01869c50844befc76d528971690 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Wed, 20 Aug 2008 17:22:02 +0200 Subject: [PATCH 06/21] oprofile: Don't report Nehalem as core_2 This essentially reverts Linus' earlier 4b9f12a3779c548b68bc9af7d94030868ad3aa1b commit. Nehalem is not core_2, so it shouldn't be reported as such. However with the earlier arch perfmon patch it will fall back to arch perfmon mode now, so there is no need to fake it as core_2. The only drawback is that Linus will need to patch the arch perfmon support into his oprofile binary now, but I think he can do that. Signed-off-by: Andi Kleen Signed-off-by: Robert Richter --- arch/x86/oprofile/nmi_int.c | 3 --- 1 file changed, 3 deletions(-) diff --git a/arch/x86/oprofile/nmi_int.c b/arch/x86/oprofile/nmi_int.c index 57f6c908808..1059f3fe6b1 100644 --- a/arch/x86/oprofile/nmi_int.c +++ b/arch/x86/oprofile/nmi_int.c @@ -420,9 +420,6 @@ static int __init ppro_init(char **cpu_type) case 15: case 23: *cpu_type = "i386/core_2"; break; - case 26: - *cpu_type = "i386/core_2"; - break; default: /* Unknown */ return 0; From b99170288421c79f0c2efa8b33e26e65f4bb7fb8 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Mon, 18 Aug 2008 14:50:31 +0200 Subject: [PATCH 07/21] oprofile: Implement Intel architectural perfmon support Newer Intel CPUs (Core1+) have support for architectural events described in CPUID 0xA. See the IA32 SDM Vol3b.18 for details. The advantage of this is that it can be done without knowing about the specific CPU, because the CPU describes by itself what performance events are supported. This is only a fallback because only a limited set of 6 events are supported. This allows to do profiling on Nehalem and on Atom systems (later not tested) This patch implements support for that in oprofile's Intel Family 6 profiling module. It also has the advantage of supporting an arbitary number of events now as reported by the CPU. Also allow arbitary counter widths >32bit while we're at it. Requires a patched oprofile userland to support the new architecture. v2: update for latest oprofile tree remove force_arch_perfmon Signed-off-by: Andi Kleen Signed-off-by: Robert Richter --- arch/x86/oprofile/nmi_int.c | 23 +++++-- arch/x86/oprofile/op_model_ppro.c | 104 +++++++++++++++++++++++------- arch/x86/oprofile/op_x86_model.h | 3 + 3 files changed, 102 insertions(+), 28 deletions(-) diff --git a/arch/x86/oprofile/nmi_int.c b/arch/x86/oprofile/nmi_int.c index 1059f3fe6b1..12d6f85084f 100644 --- a/arch/x86/oprofile/nmi_int.c +++ b/arch/x86/oprofile/nmi_int.c @@ -429,6 +429,16 @@ static int __init ppro_init(char **cpu_type) return 1; } +static int __init arch_perfmon_init(char **cpu_type) +{ + if (!cpu_has_arch_perfmon) + return 0; + *cpu_type = "i386/arch_perfmon"; + model = &op_arch_perfmon_spec; + arch_perfmon_setup_counters(); + return 1; +} + /* in order to get sysfs right */ static int using_nmi; @@ -436,7 +446,7 @@ int __init op_nmi_init(struct oprofile_operations *ops) { __u8 vendor = boot_cpu_data.x86_vendor; __u8 family = boot_cpu_data.x86; - char *cpu_type; + char *cpu_type = NULL; int ret = 0; if (!cpu_has_apic) @@ -474,19 +484,20 @@ int __init op_nmi_init(struct oprofile_operations *ops) switch (family) { /* Pentium IV */ case 0xf: - if (!p4_init(&cpu_type)) - return -ENODEV; + p4_init(&cpu_type); break; /* A P6-class processor */ case 6: - if (!ppro_init(&cpu_type)) - return -ENODEV; + ppro_init(&cpu_type); break; default: - return -ENODEV; + break; } + + if (!cpu_type && !arch_perfmon_init(&cpu_type)) + return -ENODEV; break; default: diff --git a/arch/x86/oprofile/op_model_ppro.c b/arch/x86/oprofile/op_model_ppro.c index eff431f6c57..12e207a67f1 100644 --- a/arch/x86/oprofile/op_model_ppro.c +++ b/arch/x86/oprofile/op_model_ppro.c @@ -1,32 +1,34 @@ /* * @file op_model_ppro.h - * pentium pro / P6 model-specific MSR operations + * Family 6 perfmon and architectural perfmon MSR operations * * @remark Copyright 2002 OProfile authors + * @remark Copyright 2008 Intel Corporation * @remark Read the file COPYING * * @author John Levon * @author Philippe Elie * @author Graydon Hoare + * @author Andi Kleen */ #include +#include #include #include #include #include +#include #include "op_x86_model.h" #include "op_counter.h" -#define NUM_COUNTERS 2 -#define NUM_CONTROLS 2 +static int num_counters = 2; +static int counter_width = 32; #define CTR_IS_RESERVED(msrs, c) (msrs->counters[(c)].addr ? 1 : 0) #define CTR_READ(l, h, msrs, c) do {rdmsr(msrs->counters[(c)].addr, (l), (h)); } while (0) -#define CTR_32BIT_WRITE(l, msrs, c) \ - do {wrmsr(msrs->counters[(c)].addr, -(u32)(l), 0); } while (0) -#define CTR_OVERFLOWED(n) (!((n) & (1U<<31))) +#define CTR_OVERFLOWED(n) (!((n) & (1U<<(counter_width-1)))) #define CTRL_IS_RESERVED(msrs, c) (msrs->controls[(c)].addr ? 1 : 0) #define CTRL_READ(l, h, msrs, c) do {rdmsr((msrs->controls[(c)].addr), (l), (h)); } while (0) @@ -40,20 +42,20 @@ #define CTRL_SET_UM(val, m) (val |= (m << 8)) #define CTRL_SET_EVENT(val, e) (val |= e) -static unsigned long reset_value[NUM_COUNTERS]; +static u64 *reset_value; static void ppro_fill_in_addresses(struct op_msrs * const msrs) { int i; - for (i = 0; i < NUM_COUNTERS; i++) { + for (i = 0; i < num_counters; i++) { if (reserve_perfctr_nmi(MSR_P6_PERFCTR0 + i)) msrs->counters[i].addr = MSR_P6_PERFCTR0 + i; else msrs->counters[i].addr = 0; } - for (i = 0; i < NUM_CONTROLS; i++) { + for (i = 0; i < num_counters; i++) { if (reserve_evntsel_nmi(MSR_P6_EVNTSEL0 + i)) msrs->controls[i].addr = MSR_P6_EVNTSEL0 + i; else @@ -67,8 +69,22 @@ static void ppro_setup_ctrs(struct op_msrs const * const msrs) unsigned int low, high; int i; + if (!reset_value) { + reset_value = kmalloc(sizeof(unsigned) * num_counters, + GFP_ATOMIC); + if (!reset_value) + return; + } + + if (cpu_has_arch_perfmon) { + union cpuid10_eax eax; + eax.full = cpuid_eax(0xa); + if (counter_width < eax.split.bit_width) + counter_width = eax.split.bit_width; + } + /* clear all counters */ - for (i = 0 ; i < NUM_CONTROLS; ++i) { + for (i = 0 ; i < num_counters; ++i) { if (unlikely(!CTRL_IS_RESERVED(msrs, i))) continue; CTRL_READ(low, high, msrs, i); @@ -77,18 +93,18 @@ static void ppro_setup_ctrs(struct op_msrs const * const msrs) } /* avoid a false detection of ctr overflows in NMI handler */ - for (i = 0; i < NUM_COUNTERS; ++i) { + for (i = 0; i < num_counters; ++i) { if (unlikely(!CTR_IS_RESERVED(msrs, i))) continue; - CTR_32BIT_WRITE(1, msrs, i); + wrmsrl(msrs->counters[i].addr, -1LL); } /* enable active counters */ - for (i = 0; i < NUM_COUNTERS; ++i) { + for (i = 0; i < num_counters; ++i) { if ((counter_config[i].enabled) && (CTR_IS_RESERVED(msrs, i))) { reset_value[i] = counter_config[i].count; - CTR_32BIT_WRITE(counter_config[i].count, msrs, i); + wrmsrl(msrs->counters[i].addr, -reset_value[i]); CTRL_READ(low, high, msrs, i); CTRL_CLEAR(low); @@ -111,13 +127,13 @@ static int ppro_check_ctrs(struct pt_regs * const regs, unsigned int low, high; int i; - for (i = 0 ; i < NUM_COUNTERS; ++i) { + for (i = 0 ; i < num_counters; ++i) { if (!reset_value[i]) continue; CTR_READ(low, high, msrs, i); if (CTR_OVERFLOWED(low)) { oprofile_add_sample(regs, i); - CTR_32BIT_WRITE(reset_value[i], msrs, i); + wrmsrl(msrs->counters[i].addr, -reset_value[i]); } } @@ -141,7 +157,7 @@ static void ppro_start(struct op_msrs const * const msrs) unsigned int low, high; int i; - for (i = 0; i < NUM_COUNTERS; ++i) { + for (i = 0; i < num_counters; ++i) { if (reset_value[i]) { CTRL_READ(low, high, msrs, i); CTRL_SET_ACTIVE(low); @@ -156,7 +172,7 @@ static void ppro_stop(struct op_msrs const * const msrs) unsigned int low, high; int i; - for (i = 0; i < NUM_COUNTERS; ++i) { + for (i = 0; i < num_counters; ++i) { if (!reset_value[i]) continue; CTRL_READ(low, high, msrs, i); @@ -169,20 +185,24 @@ static void ppro_shutdown(struct op_msrs const * const msrs) { int i; - for (i = 0 ; i < NUM_COUNTERS ; ++i) { + for (i = 0 ; i < num_counters ; ++i) { if (CTR_IS_RESERVED(msrs, i)) release_perfctr_nmi(MSR_P6_PERFCTR0 + i); } - for (i = 0 ; i < NUM_CONTROLS ; ++i) { + for (i = 0 ; i < num_counters ; ++i) { if (CTRL_IS_RESERVED(msrs, i)) release_evntsel_nmi(MSR_P6_EVNTSEL0 + i); } + if (reset_value) { + kfree(reset_value); + reset_value = NULL; + } } struct op_x86_model_spec const op_ppro_spec = { - .num_counters = NUM_COUNTERS, - .num_controls = NUM_CONTROLS, + .num_counters = 2, + .num_controls = 2, .fill_in_addresses = &ppro_fill_in_addresses, .setup_ctrs = &ppro_setup_ctrs, .check_ctrs = &ppro_check_ctrs, @@ -190,3 +210,43 @@ struct op_x86_model_spec const op_ppro_spec = { .stop = &ppro_stop, .shutdown = &ppro_shutdown }; + +/* + * Architectural performance monitoring. + * + * Newer Intel CPUs (Core1+) have support for architectural + * events described in CPUID 0xA. See the IA32 SDM Vol3b.18 for details. + * The advantage of this is that it can be done without knowing about + * the specific CPU. + */ + +void arch_perfmon_setup_counters(void) +{ + union cpuid10_eax eax; + + eax.full = cpuid_eax(0xa); + + /* Workaround for BIOS bugs in 6/15. Taken from perfmon2 */ + if (eax.split.version_id == 0 && current_cpu_data.x86 == 6 && + current_cpu_data.x86_model == 15) { + eax.split.version_id = 2; + eax.split.num_counters = 2; + eax.split.bit_width = 40; + } + + num_counters = eax.split.num_counters; + + op_arch_perfmon_spec.num_counters = num_counters; + op_arch_perfmon_spec.num_controls = num_counters; +} + +struct op_x86_model_spec op_arch_perfmon_spec = { + /* num_counters/num_controls filled in at runtime */ + .fill_in_addresses = &ppro_fill_in_addresses, + /* user space does the cpuid check for available events */ + .setup_ctrs = &ppro_setup_ctrs, + .check_ctrs = &ppro_check_ctrs, + .start = &ppro_start, + .stop = &ppro_stop, + .shutdown = &ppro_shutdown +}; diff --git a/arch/x86/oprofile/op_x86_model.h b/arch/x86/oprofile/op_x86_model.h index 3d3b85d3c25..0b601893a4d 100644 --- a/arch/x86/oprofile/op_x86_model.h +++ b/arch/x86/oprofile/op_x86_model.h @@ -49,5 +49,8 @@ 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_amd_spec; +extern struct op_x86_model_spec op_arch_perfmon_spec; + +extern void arch_perfmon_setup_counters(void); #endif /* OP_X86_MODEL_H */ From 59512900baab03c5629f2ff5efad1d5d4e682ece Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Mon, 29 Sep 2008 22:23:33 +0200 Subject: [PATCH 08/21] oprofile: discover counters for op ppro too Discover number of counters for all family 6 models even when not in arch perfmon mode. Signed-off-by: Andi Kleen Signed-off-by: Robert Richter --- arch/x86/oprofile/op_model_ppro.c | 8 +++++--- arch/x86/oprofile/op_x86_model.h | 2 +- 2 files changed, 6 insertions(+), 4 deletions(-) diff --git a/arch/x86/oprofile/op_model_ppro.c b/arch/x86/oprofile/op_model_ppro.c index 12e207a67f1..f5a226823e9 100644 --- a/arch/x86/oprofile/op_model_ppro.c +++ b/arch/x86/oprofile/op_model_ppro.c @@ -200,9 +200,9 @@ static void ppro_shutdown(struct op_msrs const * const msrs) } -struct op_x86_model_spec const op_ppro_spec = { - .num_counters = 2, - .num_controls = 2, +struct op_x86_model_spec op_ppro_spec = { + .num_counters = 2, /* can be overriden */ + .num_controls = 2, /* dito */ .fill_in_addresses = &ppro_fill_in_addresses, .setup_ctrs = &ppro_setup_ctrs, .check_ctrs = &ppro_check_ctrs, @@ -238,6 +238,8 @@ void arch_perfmon_setup_counters(void) op_arch_perfmon_spec.num_counters = num_counters; op_arch_perfmon_spec.num_controls = num_counters; + op_ppro_spec.num_counters = num_counters; + op_ppro_spec.num_controls = num_counters; } struct op_x86_model_spec op_arch_perfmon_spec = { diff --git a/arch/x86/oprofile/op_x86_model.h b/arch/x86/oprofile/op_x86_model.h index 0b601893a4d..596de7a5559 100644 --- a/arch/x86/oprofile/op_x86_model.h +++ b/arch/x86/oprofile/op_x86_model.h @@ -45,7 +45,7 @@ struct op_x86_model_spec { void (*shutdown)(struct op_msrs const * const msrs); }; -extern struct op_x86_model_spec const op_ppro_spec; +extern struct op_x86_model_spec 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_amd_spec; From e2fee2761ad1df2d29b9d502a3cefc87a17b32ca Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Fri, 18 Jul 2008 17:36:20 +0200 Subject: [PATCH 09/21] OProfile: Rework oprofile_add_ibs_sample() function Code looks much more cleaner now. Signed-off-by: Robert Richter --- arch/x86/oprofile/op_model_amd.c | 5 ++-- drivers/oprofile/cpu_buffer.c | 48 ++++++++++---------------------- 2 files changed, 18 insertions(+), 35 deletions(-) diff --git a/arch/x86/oprofile/op_model_amd.c b/arch/x86/oprofile/op_model_amd.c index d9faf607b3a..54632e0d6bf 100644 --- a/arch/x86/oprofile/op_model_amd.c +++ b/arch/x86/oprofile/op_model_amd.c @@ -67,8 +67,9 @@ static unsigned long reset_value[NUM_COUNTERS]; /* The function interface needs to be fixed, something like add data. Should then be added to linux/oprofile.h. */ -extern void oprofile_add_ibs_sample(struct pt_regs *const regs, - unsigned int * const ibs_sample, u8 code); +extern void +oprofile_add_ibs_sample(struct pt_regs *const regs, + unsigned int * const ibs_sample, int ibs_code); struct ibs_fetch_sample { /* MSRC001_1031 IBS Fetch Linear Address Register */ diff --git a/drivers/oprofile/cpu_buffer.c b/drivers/oprofile/cpu_buffer.c index e1bd5a937f6..d6c68270b3d 100644 --- a/drivers/oprofile/cpu_buffer.c +++ b/drivers/oprofile/cpu_buffer.c @@ -257,21 +257,23 @@ void oprofile_add_sample(struct pt_regs * const regs, unsigned long event) #ifdef CONFIG_OPROFILE_IBS -#define MAX_IBS_SAMPLE_SIZE 14 -static int log_ibs_sample(struct oprofile_cpu_buffer *cpu_buf, - unsigned long pc, int is_kernel, unsigned int *ibs, int ibs_code) +#define MAX_IBS_SAMPLE_SIZE 14 + +void oprofile_add_ibs_sample(struct pt_regs *const regs, + unsigned int * const ibs_sample, int ibs_code) { + int is_kernel = !user_mode(regs); + struct oprofile_cpu_buffer *cpu_buf = &__get_cpu_var(cpu_buffer); struct task_struct *task; cpu_buf->sample_received++; if (nr_available_slots(cpu_buf) < MAX_IBS_SAMPLE_SIZE) { + /* we can't backtrace since we lost the source of this event */ cpu_buf->sample_lost_overflow++; - return 0; + return; } - is_kernel = !!is_kernel; - /* notice a switch from user->kernel or vice versa */ if (cpu_buf->last_is_kernel != is_kernel) { cpu_buf->last_is_kernel = is_kernel; @@ -281,7 +283,6 @@ static int log_ibs_sample(struct oprofile_cpu_buffer *cpu_buf, /* notice a task switch */ if (!is_kernel) { task = current; - if (cpu_buf->last_task != task) { cpu_buf->last_task = task; add_code(cpu_buf, (unsigned long)task); @@ -289,36 +290,17 @@ static int log_ibs_sample(struct oprofile_cpu_buffer *cpu_buf, } add_code(cpu_buf, ibs_code); - add_sample(cpu_buf, ibs[0], ibs[1]); - add_sample(cpu_buf, ibs[2], ibs[3]); - add_sample(cpu_buf, ibs[4], ibs[5]); + add_sample(cpu_buf, ibs_sample[0], ibs_sample[1]); + add_sample(cpu_buf, ibs_sample[2], ibs_sample[3]); + add_sample(cpu_buf, ibs_sample[4], ibs_sample[5]); if (ibs_code == IBS_OP_BEGIN) { - add_sample(cpu_buf, ibs[6], ibs[7]); - add_sample(cpu_buf, ibs[8], ibs[9]); - add_sample(cpu_buf, ibs[10], ibs[11]); + add_sample(cpu_buf, ibs_sample[6], ibs_sample[7]); + add_sample(cpu_buf, ibs_sample[8], ibs_sample[9]); + add_sample(cpu_buf, ibs_sample[10], ibs_sample[11]); } - return 1; -} - -void oprofile_add_ibs_sample(struct pt_regs *const regs, - unsigned int * const ibs_sample, u8 code) -{ - int is_kernel = !user_mode(regs); - unsigned long pc = profile_pc(regs); - - struct oprofile_cpu_buffer *cpu_buf = - &per_cpu(cpu_buffer, smp_processor_id()); - - if (!backtrace_depth) { - log_ibs_sample(cpu_buf, pc, is_kernel, ibs_sample, code); - return; - } - - /* if log_sample() fails we can't backtrace since we lost the source - * of this event */ - if (log_ibs_sample(cpu_buf, pc, is_kernel, ibs_sample, code)) + if (backtrace_depth) oprofile_ops.backtrace(regs, backtrace_depth); } From 2d55a478827f3eed2ee9701605fdeb9cac2d78dc Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Fri, 18 Jul 2008 17:56:05 +0200 Subject: [PATCH 10/21] OProfile: Rework string handling in setup_ibs_files() Signed-off-by: Robert Richter --- arch/x86/oprofile/op_model_amd.c | 23 +++++++++++------------ 1 file changed, 11 insertions(+), 12 deletions(-) diff --git a/arch/x86/oprofile/op_model_amd.c b/arch/x86/oprofile/op_model_amd.c index 54632e0d6bf..0657c56a66b 100644 --- a/arch/x86/oprofile/op_model_amd.c +++ b/arch/x86/oprofile/op_model_amd.c @@ -473,7 +473,6 @@ static int (*create_arch_files)(struct super_block * sb, struct dentry * root); static int setup_ibs_files(struct super_block * sb, struct dentry * root) { - char buf[12]; struct dentry *dir; int ret = 0; @@ -495,22 +494,22 @@ static int setup_ibs_files(struct super_block * sb, struct dentry * root) ibs_config.max_cnt_op = 250000; ibs_config.op_enabled = 0; ibs_config.dispatched_ops = 1; - snprintf(buf, sizeof(buf), "ibs_fetch"); - dir = oprofilefs_mkdir(sb, root, buf); + + dir = oprofilefs_mkdir(sb, root, "ibs_fetch"); + oprofilefs_create_ulong(sb, dir, "enable", + &ibs_config.fetch_enabled); + oprofilefs_create_ulong(sb, dir, "max_count", + &ibs_config.max_cnt_fetch); oprofilefs_create_ulong(sb, dir, "rand_enable", &ibs_config.rand_en); + + dir = oprofilefs_mkdir(sb, root, "ibs_uops"); oprofilefs_create_ulong(sb, dir, "enable", - &ibs_config.fetch_enabled); + &ibs_config.op_enabled); oprofilefs_create_ulong(sb, dir, "max_count", - &ibs_config.max_cnt_fetch); - snprintf(buf, sizeof(buf), "ibs_uops"); - dir = oprofilefs_mkdir(sb, root, buf); - oprofilefs_create_ulong(sb, dir, "enable", - &ibs_config.op_enabled); - oprofilefs_create_ulong(sb, dir, "max_count", - &ibs_config.max_cnt_op); + &ibs_config.max_cnt_op); oprofilefs_create_ulong(sb, dir, "dispatched_ops", - &ibs_config.dispatched_ops); + &ibs_config.dispatched_ops); return 0; } From ccd755c2d90dd9b9729ba5975f7c92bf206ddcf7 Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Tue, 29 Jul 2008 16:57:10 +0200 Subject: [PATCH 11/21] OProfile: Rename IBS sysfs dir into "ibs_op" The new name is now more close to those used in the spec. Signed-off-by: Robert Richter --- arch/x86/oprofile/op_model_amd.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/x86/oprofile/op_model_amd.c b/arch/x86/oprofile/op_model_amd.c index 0657c56a66b..23ce63f2762 100644 --- a/arch/x86/oprofile/op_model_amd.c +++ b/arch/x86/oprofile/op_model_amd.c @@ -503,7 +503,7 @@ static int setup_ibs_files(struct super_block * sb, struct dentry * root) oprofilefs_create_ulong(sb, dir, "rand_enable", &ibs_config.rand_en); - dir = oprofilefs_mkdir(sb, root, "ibs_uops"); + dir = oprofilefs_mkdir(sb, root, "ibs_op"); oprofilefs_create_ulong(sb, dir, "enable", &ibs_config.op_enabled); oprofilefs_create_ulong(sb, dir, "max_count", From c92960fccb9f32a1d6110f6dcfe483ed96c62beb Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Fri, 5 Sep 2008 17:12:36 +0200 Subject: [PATCH 12/21] oprofile: whitespace fixes Signed-off-by: Robert Richter --- arch/x86/oprofile/op_model_amd.c | 20 +++++++++---------- arch/x86/oprofile/op_model_p4.c | 32 +++++++++++++++---------------- arch/x86/oprofile/op_model_ppro.c | 16 ++++++++-------- drivers/oprofile/buffer_sync.c | 1 - drivers/oprofile/oprof.c | 24 +++++++++++------------ 5 files changed, 46 insertions(+), 47 deletions(-) diff --git a/arch/x86/oprofile/op_model_amd.c b/arch/x86/oprofile/op_model_amd.c index 23ce63f2762..b9a810b3326 100644 --- a/arch/x86/oprofile/op_model_amd.c +++ b/arch/x86/oprofile/op_model_amd.c @@ -530,14 +530,14 @@ static void op_amd_exit(void) #endif struct op_x86_model_spec const op_amd_spec = { - .init = op_amd_init, - .exit = op_amd_exit, - .num_counters = NUM_COUNTERS, - .num_controls = NUM_CONTROLS, - .fill_in_addresses = &op_amd_fill_in_addresses, - .setup_ctrs = &op_amd_setup_ctrs, - .check_ctrs = &op_amd_check_ctrs, - .start = &op_amd_start, - .stop = &op_amd_stop, - .shutdown = &op_amd_shutdown + .init = op_amd_init, + .exit = op_amd_exit, + .num_counters = NUM_COUNTERS, + .num_controls = NUM_CONTROLS, + .fill_in_addresses = &op_amd_fill_in_addresses, + .setup_ctrs = &op_amd_setup_ctrs, + .check_ctrs = &op_amd_check_ctrs, + .start = &op_amd_start, + .stop = &op_amd_stop, + .shutdown = &op_amd_shutdown }; diff --git a/arch/x86/oprofile/op_model_p4.c b/arch/x86/oprofile/op_model_p4.c index 43ac5af338d..4c4a51c90bc 100644 --- a/arch/x86/oprofile/op_model_p4.c +++ b/arch/x86/oprofile/op_model_p4.c @@ -698,24 +698,24 @@ static void p4_shutdown(struct op_msrs const * const msrs) #ifdef CONFIG_SMP struct op_x86_model_spec const op_p4_ht2_spec = { - .num_counters = NUM_COUNTERS_HT2, - .num_controls = NUM_CONTROLS_HT2, - .fill_in_addresses = &p4_fill_in_addresses, - .setup_ctrs = &p4_setup_ctrs, - .check_ctrs = &p4_check_ctrs, - .start = &p4_start, - .stop = &p4_stop, - .shutdown = &p4_shutdown + .num_counters = NUM_COUNTERS_HT2, + .num_controls = NUM_CONTROLS_HT2, + .fill_in_addresses = &p4_fill_in_addresses, + .setup_ctrs = &p4_setup_ctrs, + .check_ctrs = &p4_check_ctrs, + .start = &p4_start, + .stop = &p4_stop, + .shutdown = &p4_shutdown }; #endif struct op_x86_model_spec const op_p4_spec = { - .num_counters = NUM_COUNTERS_NON_HT, - .num_controls = NUM_CONTROLS_NON_HT, - .fill_in_addresses = &p4_fill_in_addresses, - .setup_ctrs = &p4_setup_ctrs, - .check_ctrs = &p4_check_ctrs, - .start = &p4_start, - .stop = &p4_stop, - .shutdown = &p4_shutdown + .num_counters = NUM_COUNTERS_NON_HT, + .num_controls = NUM_CONTROLS_NON_HT, + .fill_in_addresses = &p4_fill_in_addresses, + .setup_ctrs = &p4_setup_ctrs, + .check_ctrs = &p4_check_ctrs, + .start = &p4_start, + .stop = &p4_stop, + .shutdown = &p4_shutdown }; diff --git a/arch/x86/oprofile/op_model_ppro.c b/arch/x86/oprofile/op_model_ppro.c index eff431f6c57..c665bac4a14 100644 --- a/arch/x86/oprofile/op_model_ppro.c +++ b/arch/x86/oprofile/op_model_ppro.c @@ -181,12 +181,12 @@ static void ppro_shutdown(struct op_msrs const * const msrs) struct op_x86_model_spec const op_ppro_spec = { - .num_counters = NUM_COUNTERS, - .num_controls = NUM_CONTROLS, - .fill_in_addresses = &ppro_fill_in_addresses, - .setup_ctrs = &ppro_setup_ctrs, - .check_ctrs = &ppro_check_ctrs, - .start = &ppro_start, - .stop = &ppro_stop, - .shutdown = &ppro_shutdown + .num_counters = NUM_COUNTERS, + .num_controls = NUM_CONTROLS, + .fill_in_addresses = &ppro_fill_in_addresses, + .setup_ctrs = &ppro_setup_ctrs, + .check_ctrs = &ppro_check_ctrs, + .start = &ppro_start, + .stop = &ppro_stop, + .shutdown = &ppro_shutdown }; diff --git a/drivers/oprofile/buffer_sync.c b/drivers/oprofile/buffer_sync.c index ed982273fb8..564577307a5 100644 --- a/drivers/oprofile/buffer_sync.c +++ b/drivers/oprofile/buffer_sync.c @@ -41,7 +41,6 @@ static cpumask_t marked_cpus = CPU_MASK_NONE; static DEFINE_SPINLOCK(task_mortuary); static void process_task_mortuary(void); - /* Take ownership of the task struct and place it on the * list for processing. Only after two full buffer syncs * does the task eventually get freed, because by then diff --git a/drivers/oprofile/oprof.c b/drivers/oprofile/oprof.c index 2c645170f06..50062cea292 100644 --- a/drivers/oprofile/oprof.c +++ b/drivers/oprofile/oprof.c @@ -19,7 +19,7 @@ #include "cpu_buffer.h" #include "buffer_sync.h" #include "oprofile_stats.h" - + struct oprofile_operations oprofile_ops; unsigned long oprofile_started; @@ -36,7 +36,7 @@ static int timer = 0; int oprofile_setup(void) { int err; - + mutex_lock(&start_mutex); if ((err = alloc_cpu_buffers())) @@ -44,10 +44,10 @@ int oprofile_setup(void) if ((err = alloc_event_buffer())) goto out1; - + if (oprofile_ops.setup && (err = oprofile_ops.setup())) goto out2; - + /* Note even though this starts part of the * profiling overhead, it's necessary to prevent * us missing task deaths and eventually oopsing @@ -74,7 +74,7 @@ post_sync: is_setup = 1; mutex_unlock(&start_mutex); return 0; - + out3: if (oprofile_ops.shutdown) oprofile_ops.shutdown(); @@ -92,17 +92,17 @@ out: int oprofile_start(void) { int err = -EINVAL; - + mutex_lock(&start_mutex); if (!is_setup) goto out; - err = 0; - + err = 0; + if (oprofile_started) goto out; - + oprofile_reset_stats(); if ((err = oprofile_ops.start())) @@ -114,7 +114,7 @@ out: return err; } - + /* echo 0>/dev/oprofile/enable */ void oprofile_stop(void) { @@ -204,13 +204,13 @@ static void __exit oprofile_exit(void) oprofile_arch_exit(); } - + module_init(oprofile_init); module_exit(oprofile_exit); module_param_named(timer, timer, int, 0644); MODULE_PARM_DESC(timer, "force use of timer interrupt"); - + MODULE_LICENSE("GPL"); MODULE_AUTHOR("John Levon "); MODULE_DESCRIPTION("OProfile system profiler"); From 25ad2913cae9c9e3ed28075caeb2eefccd636f4f Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Fri, 5 Sep 2008 17:12:36 +0200 Subject: [PATCH 13/21] oprofile: more whitespace fixes Signed-off-by: Robert Richter --- arch/alpha/oprofile/common.c | 2 +- arch/ia64/oprofile/init.c | 4 +- arch/ia64/oprofile/perfmon.c | 4 +- arch/m32r/oprofile/init.c | 2 +- arch/mips/oprofile/common.c | 2 +- arch/mips/oprofile/op_impl.h | 2 +- arch/mips/oprofile/op_model_rm9000.c | 2 +- arch/parisc/oprofile/init.c | 2 +- arch/powerpc/oprofile/op_model_cell.c | 2 +- arch/sparc/oprofile/init.c | 2 +- arch/sparc64/oprofile/init.c | 2 +- arch/x86/oprofile/backtrace.c | 2 +- arch/x86/oprofile/op_model_amd.c | 6 +-- arch/x86/oprofile/op_x86_model.h | 4 +- drivers/oprofile/cpu_buffer.c | 22 ++++---- drivers/oprofile/cpu_buffer.h | 6 +-- drivers/oprofile/event_buffer.c | 10 ++-- drivers/oprofile/oprof.h | 4 +- drivers/oprofile/oprofile_files.c | 16 +++--- drivers/oprofile/oprofile_stats.c | 10 ++-- drivers/oprofile/oprofile_stats.h | 2 +- drivers/oprofile/oprofilefs.c | 72 +++++++++++++-------------- drivers/oprofile/timer_int.c | 2 +- 23 files changed, 91 insertions(+), 91 deletions(-) diff --git a/arch/alpha/oprofile/common.c b/arch/alpha/oprofile/common.c index 7c3d5ec6ec6..bd8ac533a50 100644 --- a/arch/alpha/oprofile/common.c +++ b/arch/alpha/oprofile/common.c @@ -106,7 +106,7 @@ op_axp_stop(void) } static int -op_axp_create_files(struct super_block * sb, struct dentry * root) +op_axp_create_files(struct super_block *sb, struct dentry *root) { int i; diff --git a/arch/ia64/oprofile/init.c b/arch/ia64/oprofile/init.c index 125a602a660..31b545c3546 100644 --- a/arch/ia64/oprofile/init.c +++ b/arch/ia64/oprofile/init.c @@ -12,11 +12,11 @@ #include #include -extern int perfmon_init(struct oprofile_operations * ops); +extern int perfmon_init(struct oprofile_operations *ops); extern void perfmon_exit(void); extern void ia64_backtrace(struct pt_regs * const regs, unsigned int depth); -int __init oprofile_arch_init(struct oprofile_operations * ops) +int __init oprofile_arch_init(struct oprofile_operations *ops) { int ret = -ENODEV; diff --git a/arch/ia64/oprofile/perfmon.c b/arch/ia64/oprofile/perfmon.c index bc41dd32fec..192d3e8e1f6 100644 --- a/arch/ia64/oprofile/perfmon.c +++ b/arch/ia64/oprofile/perfmon.c @@ -56,7 +56,7 @@ static pfm_buffer_fmt_t oprofile_fmt = { }; -static char * get_cpu_type(void) +static char *get_cpu_type(void) { __u8 family = local_cpu_data->family; @@ -75,7 +75,7 @@ static char * get_cpu_type(void) static int using_perfmon; -int perfmon_init(struct oprofile_operations * ops) +int perfmon_init(struct oprofile_operations *ops) { int ret = pfm_register_buffer_fmt(&oprofile_fmt); if (ret) diff --git a/arch/m32r/oprofile/init.c b/arch/m32r/oprofile/init.c index b7773e45c43..fa56860f425 100644 --- a/arch/m32r/oprofile/init.c +++ b/arch/m32r/oprofile/init.c @@ -12,7 +12,7 @@ #include #include -int __init oprofile_arch_init(struct oprofile_operations * ops) +int __init oprofile_arch_init(struct oprofile_operations *ops) { return -ENODEV; } diff --git a/arch/mips/oprofile/common.c b/arch/mips/oprofile/common.c index dd2fbd6645c..3bf3354547f 100644 --- a/arch/mips/oprofile/common.c +++ b/arch/mips/oprofile/common.c @@ -32,7 +32,7 @@ static int op_mips_setup(void) return 0; } -static int op_mips_create_files(struct super_block * sb, struct dentry * root) +static int op_mips_create_files(struct super_block *sb, struct dentry *root) { int i; diff --git a/arch/mips/oprofile/op_impl.h b/arch/mips/oprofile/op_impl.h index 2bfc17c3010..f04b54fb37d 100644 --- a/arch/mips/oprofile/op_impl.h +++ b/arch/mips/oprofile/op_impl.h @@ -27,7 +27,7 @@ struct op_counter_config { /* Per-architecture configury and hooks. */ struct op_mips_model { void (*reg_setup) (struct op_counter_config *); - void (*cpu_setup) (void * dummy); + void (*cpu_setup) (void *dummy); int (*init)(void); void (*exit)(void); void (*cpu_start)(void *args); diff --git a/arch/mips/oprofile/op_model_rm9000.c b/arch/mips/oprofile/op_model_rm9000.c index a45d3202894..3aa81384966 100644 --- a/arch/mips/oprofile/op_model_rm9000.c +++ b/arch/mips/oprofile/op_model_rm9000.c @@ -80,7 +80,7 @@ static void rm9000_cpu_stop(void *args) write_c0_perfcontrol(0); } -static irqreturn_t rm9000_perfcount_handler(int irq, void * dev_id) +static irqreturn_t rm9000_perfcount_handler(int irq, void *dev_id) { unsigned int control = read_c0_perfcontrol(); struct pt_regs *regs = get_irq_regs(); diff --git a/arch/parisc/oprofile/init.c b/arch/parisc/oprofile/init.c index 113f5139f55..026cba2af07 100644 --- a/arch/parisc/oprofile/init.c +++ b/arch/parisc/oprofile/init.c @@ -12,7 +12,7 @@ #include #include -int __init oprofile_arch_init(struct oprofile_operations * ops) +int __init oprofile_arch_init(struct oprofile_operations *ops) { return -ENODEV; } diff --git a/arch/powerpc/oprofile/op_model_cell.c b/arch/powerpc/oprofile/op_model_cell.c index 5ff4de3eb3b..35141a8bc3d 100644 --- a/arch/powerpc/oprofile/op_model_cell.c +++ b/arch/powerpc/oprofile/op_model_cell.c @@ -404,7 +404,7 @@ set_count_mode(u32 kernel, u32 user) } } -static inline void enable_ctr(u32 cpu, u32 ctr, u32 * pm07_cntrl) +static inline void enable_ctr(u32 cpu, u32 ctr, u32 *pm07_cntrl) { pm07_cntrl[ctr] |= CBE_PM_CTR_ENABLE; diff --git a/arch/sparc/oprofile/init.c b/arch/sparc/oprofile/init.c index 9ab815b95b5..17bb6035069 100644 --- a/arch/sparc/oprofile/init.c +++ b/arch/sparc/oprofile/init.c @@ -12,7 +12,7 @@ #include #include -int __init oprofile_arch_init(struct oprofile_operations * ops) +int __init oprofile_arch_init(struct oprofile_operations *ops) { return -ENODEV; } diff --git a/arch/sparc64/oprofile/init.c b/arch/sparc64/oprofile/init.c index 9ab815b95b5..17bb6035069 100644 --- a/arch/sparc64/oprofile/init.c +++ b/arch/sparc64/oprofile/init.c @@ -12,7 +12,7 @@ #include #include -int __init oprofile_arch_init(struct oprofile_operations * ops) +int __init oprofile_arch_init(struct oprofile_operations *ops) { return -ENODEV; } diff --git a/arch/x86/oprofile/backtrace.c b/arch/x86/oprofile/backtrace.c index e2095cba409..36e324139f7 100644 --- a/arch/x86/oprofile/backtrace.c +++ b/arch/x86/oprofile/backtrace.c @@ -53,7 +53,7 @@ struct frame_head { } __attribute__((packed)); static struct frame_head * -dump_user_backtrace(struct frame_head * head) +dump_user_backtrace(struct frame_head *head) { struct frame_head bufhead[2]; diff --git a/arch/x86/oprofile/op_model_amd.c b/arch/x86/oprofile/op_model_amd.c index b9a810b3326..a2e83afbe3e 100644 --- a/arch/x86/oprofile/op_model_amd.c +++ b/arch/x86/oprofile/op_model_amd.c @@ -69,7 +69,7 @@ static unsigned long reset_value[NUM_COUNTERS]; data. Should then be added to linux/oprofile.h. */ extern void oprofile_add_ibs_sample(struct pt_regs *const regs, - unsigned int * const ibs_sample, int ibs_code); + unsigned int *const ibs_sample, int ibs_code); struct ibs_fetch_sample { /* MSRC001_1031 IBS Fetch Linear Address Register */ @@ -469,9 +469,9 @@ static void clear_ibs_nmi(void) on_each_cpu(apic_clear_ibs_nmi_per_cpu, NULL, 1); } -static int (*create_arch_files)(struct super_block * sb, struct dentry * root); +static int (*create_arch_files)(struct super_block *sb, struct dentry *root); -static int setup_ibs_files(struct super_block * sb, struct dentry * root) +static int setup_ibs_files(struct super_block *sb, struct dentry *root) { struct dentry *dir; int ret = 0; diff --git a/arch/x86/oprofile/op_x86_model.h b/arch/x86/oprofile/op_x86_model.h index 05a0261ba0c..24ccdebf3ac 100644 --- a/arch/x86/oprofile/op_x86_model.h +++ b/arch/x86/oprofile/op_x86_model.h @@ -22,8 +22,8 @@ struct op_msr { }; struct op_msrs { - struct op_msr * counters; - struct op_msr * controls; + struct op_msr *counters; + struct op_msr *controls; }; struct pt_regs; diff --git a/drivers/oprofile/cpu_buffer.c b/drivers/oprofile/cpu_buffer.c index d6c68270b3d..b47ce038490 100644 --- a/drivers/oprofile/cpu_buffer.c +++ b/drivers/oprofile/cpu_buffer.c @@ -112,7 +112,7 @@ void end_cpu_work(void) } /* Resets the cpu buffer to a sane state. */ -void cpu_buffer_reset(struct oprofile_cpu_buffer * cpu_buf) +void cpu_buffer_reset(struct oprofile_cpu_buffer *cpu_buf) { /* reset these to invalid values; the next sample * collected will populate the buffer with proper @@ -123,7 +123,7 @@ void cpu_buffer_reset(struct oprofile_cpu_buffer * cpu_buf) } /* compute number of available slots in cpu_buffer queue */ -static unsigned long nr_available_slots(struct oprofile_cpu_buffer const * b) +static unsigned long nr_available_slots(struct oprofile_cpu_buffer const *b) { unsigned long head = b->head_pos; unsigned long tail = b->tail_pos; @@ -134,7 +134,7 @@ static unsigned long nr_available_slots(struct oprofile_cpu_buffer const * b) return tail + (b->buffer_size - head) - 1; } -static void increment_head(struct oprofile_cpu_buffer * b) +static void increment_head(struct oprofile_cpu_buffer *b) { unsigned long new_head = b->head_pos + 1; @@ -149,17 +149,17 @@ static void increment_head(struct oprofile_cpu_buffer * b) } static inline void -add_sample(struct oprofile_cpu_buffer * cpu_buf, +add_sample(struct oprofile_cpu_buffer *cpu_buf, unsigned long pc, unsigned long event) { - struct op_sample * entry = &cpu_buf->buffer[cpu_buf->head_pos]; + struct op_sample *entry = &cpu_buf->buffer[cpu_buf->head_pos]; entry->eip = pc; entry->event = event; increment_head(cpu_buf); } static inline void -add_code(struct oprofile_cpu_buffer * buffer, unsigned long value) +add_code(struct oprofile_cpu_buffer *buffer, unsigned long value) { add_sample(buffer, ESCAPE_CODE, value); } @@ -173,10 +173,10 @@ add_code(struct oprofile_cpu_buffer * buffer, unsigned long value) * pc. We tag this in the buffer by generating kernel enter/exit * events whenever is_kernel changes */ -static int log_sample(struct oprofile_cpu_buffer * cpu_buf, unsigned long pc, +static int log_sample(struct oprofile_cpu_buffer *cpu_buf, unsigned long pc, int is_kernel, unsigned long event) { - struct task_struct * task; + struct task_struct *task; cpu_buf->sample_received++; @@ -222,7 +222,7 @@ static int oprofile_begin_trace(struct oprofile_cpu_buffer *cpu_buf) return 1; } -static void oprofile_end_trace(struct oprofile_cpu_buffer * cpu_buf) +static void oprofile_end_trace(struct oprofile_cpu_buffer *cpu_buf) { cpu_buf->tracing = 0; } @@ -260,7 +260,7 @@ void oprofile_add_sample(struct pt_regs * const regs, unsigned long event) #define MAX_IBS_SAMPLE_SIZE 14 void oprofile_add_ibs_sample(struct pt_regs *const regs, - unsigned int * const ibs_sample, int ibs_code) + unsigned int *const ibs_sample, int ibs_code) { int is_kernel = !user_mode(regs); struct oprofile_cpu_buffer *cpu_buf = &__get_cpu_var(cpu_buffer); @@ -345,7 +345,7 @@ void oprofile_add_trace(unsigned long pc) */ static void wq_sync_buffer(struct work_struct *work) { - struct oprofile_cpu_buffer * b = + struct oprofile_cpu_buffer *b = container_of(work, struct oprofile_cpu_buffer, work.work); if (b->cpu != smp_processor_id()) { printk(KERN_DEBUG "WQ on CPU%d, prefer CPU%d\n", diff --git a/drivers/oprofile/cpu_buffer.h b/drivers/oprofile/cpu_buffer.h index 9c44d004da6..9bc6bb20b6d 100644 --- a/drivers/oprofile/cpu_buffer.h +++ b/drivers/oprofile/cpu_buffer.h @@ -36,10 +36,10 @@ struct oprofile_cpu_buffer { volatile unsigned long head_pos; volatile unsigned long tail_pos; unsigned long buffer_size; - struct task_struct * last_task; + struct task_struct *last_task; int last_is_kernel; int tracing; - struct op_sample * buffer; + struct op_sample *buffer; unsigned long sample_received; unsigned long sample_lost_overflow; unsigned long backtrace_aborted; @@ -50,7 +50,7 @@ struct oprofile_cpu_buffer { DECLARE_PER_CPU(struct oprofile_cpu_buffer, cpu_buffer); -void cpu_buffer_reset(struct oprofile_cpu_buffer * cpu_buf); +void cpu_buffer_reset(struct oprofile_cpu_buffer *cpu_buf); /* transient events for the CPU buffer -> event buffer */ #define CPU_IS_KERNEL 1 diff --git a/drivers/oprofile/event_buffer.c b/drivers/oprofile/event_buffer.c index 8d692a5c8e7..c9329f4e090 100644 --- a/drivers/oprofile/event_buffer.c +++ b/drivers/oprofile/event_buffer.c @@ -28,7 +28,7 @@ DEFINE_MUTEX(buffer_mutex); static unsigned long buffer_opened; static DECLARE_WAIT_QUEUE_HEAD(buffer_wait); -static unsigned long * event_buffer; +static unsigned long *event_buffer; static unsigned long buffer_size; static unsigned long buffer_watershed; static size_t buffer_pos; @@ -98,7 +98,7 @@ void free_event_buffer(void) } -static int event_buffer_open(struct inode * inode, struct file * file) +static int event_buffer_open(struct inode *inode, struct file *file) { int err = -EPERM; @@ -134,7 +134,7 @@ out: } -static int event_buffer_release(struct inode * inode, struct file * file) +static int event_buffer_release(struct inode *inode, struct file *file) { oprofile_stop(); oprofile_shutdown(); @@ -146,8 +146,8 @@ static int event_buffer_release(struct inode * inode, struct file * file) } -static ssize_t event_buffer_read(struct file * file, char __user * buf, - size_t count, loff_t * offset) +static ssize_t event_buffer_read(struct file *file, char __user *buf, + size_t count, loff_t *offset) { int retval = -EINVAL; size_t const max = buffer_size * sizeof(unsigned long); diff --git a/drivers/oprofile/oprof.h b/drivers/oprofile/oprof.h index 18323650806..7a44ddba0be 100644 --- a/drivers/oprofile/oprof.h +++ b/drivers/oprofile/oprof.h @@ -31,8 +31,8 @@ extern unsigned long backtrace_depth; struct super_block; struct dentry; -void oprofile_create_files(struct super_block * sb, struct dentry * root); -void oprofile_timer_init(struct oprofile_operations * ops); +void oprofile_create_files(struct super_block *sb, struct dentry *root); +void oprofile_timer_init(struct oprofile_operations *ops); int oprofile_set_backtrace(unsigned long depth); diff --git a/drivers/oprofile/oprofile_files.c b/drivers/oprofile/oprofile_files.c index ef953ba5ab6..241804abbb5 100644 --- a/drivers/oprofile/oprofile_files.c +++ b/drivers/oprofile/oprofile_files.c @@ -18,13 +18,13 @@ unsigned long fs_buffer_size = 131072; unsigned long fs_cpu_buffer_size = 8192; unsigned long fs_buffer_watershed = 32768; /* FIXME: tune */ -static ssize_t depth_read(struct file * file, char __user * buf, size_t count, loff_t * offset) +static ssize_t depth_read(struct file *file, char __user *buf, size_t count, loff_t *offset) { return oprofilefs_ulong_to_user(backtrace_depth, buf, count, offset); } -static ssize_t depth_write(struct file * file, char const __user * buf, size_t count, loff_t * offset) +static ssize_t depth_write(struct file *file, char const __user *buf, size_t count, loff_t *offset) { unsigned long val; int retval; @@ -50,7 +50,7 @@ static const struct file_operations depth_fops = { }; -static ssize_t pointer_size_read(struct file * file, char __user * buf, size_t count, loff_t * offset) +static ssize_t pointer_size_read(struct file *file, char __user *buf, size_t count, loff_t *offset) { return oprofilefs_ulong_to_user(sizeof(void *), buf, count, offset); } @@ -61,7 +61,7 @@ static const struct file_operations pointer_size_fops = { }; -static ssize_t cpu_type_read(struct file * file, char __user * buf, size_t count, loff_t * offset) +static ssize_t cpu_type_read(struct file *file, char __user *buf, size_t count, loff_t *offset) { return oprofilefs_str_to_user(oprofile_ops.cpu_type, buf, count, offset); } @@ -72,13 +72,13 @@ static const struct file_operations cpu_type_fops = { }; -static ssize_t enable_read(struct file * file, char __user * buf, size_t count, loff_t * offset) +static ssize_t enable_read(struct file *file, char __user *buf, size_t count, loff_t *offset) { return oprofilefs_ulong_to_user(oprofile_started, buf, count, offset); } -static ssize_t enable_write(struct file * file, char const __user * buf, size_t count, loff_t * offset) +static ssize_t enable_write(struct file *file, char const __user *buf, size_t count, loff_t *offset) { unsigned long val; int retval; @@ -107,7 +107,7 @@ static const struct file_operations enable_fops = { }; -static ssize_t dump_write(struct file * file, char const __user * buf, size_t count, loff_t * offset) +static ssize_t dump_write(struct file *file, char const __user *buf, size_t count, loff_t *offset) { wake_up_buffer_waiter(); return count; @@ -118,7 +118,7 @@ static const struct file_operations dump_fops = { .write = dump_write, }; -void oprofile_create_files(struct super_block * sb, struct dentry * root) +void oprofile_create_files(struct super_block *sb, struct dentry *root) { oprofilefs_create_file(sb, root, "enable", &enable_fops); oprofilefs_create_file_perm(sb, root, "dump", &dump_fops, 0666); diff --git a/drivers/oprofile/oprofile_stats.c b/drivers/oprofile/oprofile_stats.c index f99b28e7b79..e0c45498d17 100644 --- a/drivers/oprofile/oprofile_stats.c +++ b/drivers/oprofile/oprofile_stats.c @@ -19,7 +19,7 @@ struct oprofile_stat_struct oprofile_stats; void oprofile_reset_stats(void) { - struct oprofile_cpu_buffer * cpu_buf; + struct oprofile_cpu_buffer *cpu_buf; int i; for_each_possible_cpu(i) { @@ -36,11 +36,11 @@ void oprofile_reset_stats(void) } -void oprofile_create_stats_files(struct super_block * sb, struct dentry * root) +void oprofile_create_stats_files(struct super_block *sb, struct dentry *root) { - struct oprofile_cpu_buffer * cpu_buf; - struct dentry * cpudir; - struct dentry * dir; + struct oprofile_cpu_buffer *cpu_buf; + struct dentry *cpudir; + struct dentry *dir; char buf[10]; int i; diff --git a/drivers/oprofile/oprofile_stats.h b/drivers/oprofile/oprofile_stats.h index 6d755a633f1..54e59c29b43 100644 --- a/drivers/oprofile/oprofile_stats.h +++ b/drivers/oprofile/oprofile_stats.h @@ -28,6 +28,6 @@ struct super_block; struct dentry; /* create the stats/ dir */ -void oprofile_create_stats_files(struct super_block * sb, struct dentry * root); +void oprofile_create_stats_files(struct super_block *sb, struct dentry *root); #endif /* OPROFILE_STATS_H */ diff --git a/drivers/oprofile/oprofilefs.c b/drivers/oprofile/oprofilefs.c index 8543cb26cf3..a275a3aa5f0 100644 --- a/drivers/oprofile/oprofilefs.c +++ b/drivers/oprofile/oprofilefs.c @@ -23,9 +23,9 @@ DEFINE_SPINLOCK(oprofilefs_lock); -static struct inode * oprofilefs_get_inode(struct super_block * sb, int mode) +static struct inode *oprofilefs_get_inode(struct super_block *sb, int mode) { - struct inode * inode = new_inode(sb); + struct inode *inode = new_inode(sb); if (inode) { inode->i_mode = mode; @@ -44,7 +44,7 @@ static struct super_operations s_ops = { }; -ssize_t oprofilefs_str_to_user(char const * str, char __user * buf, size_t count, loff_t * offset) +ssize_t oprofilefs_str_to_user(char const *str, char __user *buf, size_t count, loff_t *offset) { return simple_read_from_buffer(buf, count, offset, str, strlen(str)); } @@ -52,7 +52,7 @@ ssize_t oprofilefs_str_to_user(char const * str, char __user * buf, size_t count #define TMPBUFSIZE 50 -ssize_t oprofilefs_ulong_to_user(unsigned long val, char __user * buf, size_t count, loff_t * offset) +ssize_t oprofilefs_ulong_to_user(unsigned long val, char __user *buf, size_t count, loff_t *offset) { char tmpbuf[TMPBUFSIZE]; size_t maxlen = snprintf(tmpbuf, TMPBUFSIZE, "%lu\n", val); @@ -62,7 +62,7 @@ ssize_t oprofilefs_ulong_to_user(unsigned long val, char __user * buf, size_t co } -int oprofilefs_ulong_from_user(unsigned long * val, char const __user * buf, size_t count) +int oprofilefs_ulong_from_user(unsigned long *val, char const __user *buf, size_t count) { char tmpbuf[TMPBUFSIZE]; unsigned long flags; @@ -85,16 +85,16 @@ int oprofilefs_ulong_from_user(unsigned long * val, char const __user * buf, siz } -static ssize_t ulong_read_file(struct file * file, char __user * buf, size_t count, loff_t * offset) +static ssize_t ulong_read_file(struct file *file, char __user *buf, size_t count, loff_t *offset) { - unsigned long * val = file->private_data; + unsigned long *val = file->private_data; return oprofilefs_ulong_to_user(*val, buf, count, offset); } -static ssize_t ulong_write_file(struct file * file, char const __user * buf, size_t count, loff_t * offset) +static ssize_t ulong_write_file(struct file *file, char const __user *buf, size_t count, loff_t *offset) { - unsigned long * value = file->private_data; + unsigned long *value = file->private_data; int retval; if (*offset) @@ -108,7 +108,7 @@ static ssize_t ulong_write_file(struct file * file, char const __user * buf, siz } -static int default_open(struct inode * inode, struct file * filp) +static int default_open(struct inode *inode, struct file *filp) { if (inode->i_private) filp->private_data = inode->i_private; @@ -129,12 +129,12 @@ static const struct file_operations ulong_ro_fops = { }; -static struct dentry * __oprofilefs_create_file(struct super_block * sb, - struct dentry * root, char const * name, const struct file_operations * fops, +static struct dentry *__oprofilefs_create_file(struct super_block *sb, + struct dentry *root, char const *name, const struct file_operations *fops, int perm) { - struct dentry * dentry; - struct inode * inode; + struct dentry *dentry; + struct inode *inode; dentry = d_alloc_name(root, name); if (!dentry) @@ -150,10 +150,10 @@ static struct dentry * __oprofilefs_create_file(struct super_block * sb, } -int oprofilefs_create_ulong(struct super_block * sb, struct dentry * root, - char const * name, unsigned long * val) +int oprofilefs_create_ulong(struct super_block *sb, struct dentry *root, + char const *name, unsigned long *val) { - struct dentry * d = __oprofilefs_create_file(sb, root, name, + struct dentry *d = __oprofilefs_create_file(sb, root, name, &ulong_fops, 0644); if (!d) return -EFAULT; @@ -163,10 +163,10 @@ int oprofilefs_create_ulong(struct super_block * sb, struct dentry * root, } -int oprofilefs_create_ro_ulong(struct super_block * sb, struct dentry * root, - char const * name, unsigned long * val) +int oprofilefs_create_ro_ulong(struct super_block *sb, struct dentry *root, + char const *name, unsigned long *val) { - struct dentry * d = __oprofilefs_create_file(sb, root, name, + struct dentry *d = __oprofilefs_create_file(sb, root, name, &ulong_ro_fops, 0444); if (!d) return -EFAULT; @@ -176,9 +176,9 @@ int oprofilefs_create_ro_ulong(struct super_block * sb, struct dentry * root, } -static ssize_t atomic_read_file(struct file * file, char __user * buf, size_t count, loff_t * offset) +static ssize_t atomic_read_file(struct file *file, char __user *buf, size_t count, loff_t *offset) { - atomic_t * val = file->private_data; + atomic_t *val = file->private_data; return oprofilefs_ulong_to_user(atomic_read(val), buf, count, offset); } @@ -189,10 +189,10 @@ static const struct file_operations atomic_ro_fops = { }; -int oprofilefs_create_ro_atomic(struct super_block * sb, struct dentry * root, - char const * name, atomic_t * val) +int oprofilefs_create_ro_atomic(struct super_block *sb, struct dentry *root, + char const *name, atomic_t *val) { - struct dentry * d = __oprofilefs_create_file(sb, root, name, + struct dentry *d = __oprofilefs_create_file(sb, root, name, &atomic_ro_fops, 0444); if (!d) return -EFAULT; @@ -202,8 +202,8 @@ int oprofilefs_create_ro_atomic(struct super_block * sb, struct dentry * root, } -int oprofilefs_create_file(struct super_block * sb, struct dentry * root, - char const * name, const struct file_operations * fops) +int oprofilefs_create_file(struct super_block *sb, struct dentry *root, + char const *name, const struct file_operations *fops) { if (!__oprofilefs_create_file(sb, root, name, fops, 0644)) return -EFAULT; @@ -211,8 +211,8 @@ int oprofilefs_create_file(struct super_block * sb, struct dentry * root, } -int oprofilefs_create_file_perm(struct super_block * sb, struct dentry * root, - char const * name, const struct file_operations * fops, int perm) +int oprofilefs_create_file_perm(struct super_block *sb, struct dentry *root, + char const *name, const struct file_operations *fops, int perm) { if (!__oprofilefs_create_file(sb, root, name, fops, perm)) return -EFAULT; @@ -220,11 +220,11 @@ int oprofilefs_create_file_perm(struct super_block * sb, struct dentry * root, } -struct dentry * oprofilefs_mkdir(struct super_block * sb, - struct dentry * root, char const * name) +struct dentry *oprofilefs_mkdir(struct super_block *sb, + struct dentry *root, char const *name) { - struct dentry * dentry; - struct inode * inode; + struct dentry *dentry; + struct inode *inode; dentry = d_alloc_name(root, name); if (!dentry) @@ -241,10 +241,10 @@ struct dentry * oprofilefs_mkdir(struct super_block * sb, } -static int oprofilefs_fill_super(struct super_block * sb, void * data, int silent) +static int oprofilefs_fill_super(struct super_block *sb, void *data, int silent) { - struct inode * root_inode; - struct dentry * root_dentry; + struct inode *root_inode; + struct dentry *root_dentry; sb->s_blocksize = PAGE_CACHE_SIZE; sb->s_blocksize_bits = PAGE_CACHE_SHIFT; diff --git a/drivers/oprofile/timer_int.c b/drivers/oprofile/timer_int.c index 710a45f0d73..7258b141a51 100644 --- a/drivers/oprofile/timer_int.c +++ b/drivers/oprofile/timer_int.c @@ -35,7 +35,7 @@ static void timer_stop(void) } -void __init oprofile_timer_init(struct oprofile_operations * ops) +void __init oprofile_timer_init(struct oprofile_operations *ops) { ops->create_files = NULL; ops->setup = NULL; From 8655a3b8725d5598adc438ec94916568afcc5ec9 Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Mon, 28 Jul 2008 18:13:10 +0200 Subject: [PATCH 14/21] x86/oprofile: removing unused function parameter in add_ibs_begin() Signed-off-by: Robert Richter --- drivers/oprofile/buffer_sync.c | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-) diff --git a/drivers/oprofile/buffer_sync.c b/drivers/oprofile/buffer_sync.c index 564577307a5..33bfa60b0c6 100644 --- a/drivers/oprofile/buffer_sync.c +++ b/drivers/oprofile/buffer_sync.c @@ -340,7 +340,7 @@ static void add_trace_begin(void) * Add IBS fetch and op entries to event buffer */ static void add_ibs_begin(struct oprofile_cpu_buffer *cpu_buf, int code, - int in_kernel, struct mm_struct *mm) + struct mm_struct *mm) { unsigned long rip; int i, count; @@ -592,12 +592,10 @@ void sync_buffer(int cpu) #ifdef CONFIG_OPROFILE_IBS } else if (s->event == IBS_FETCH_BEGIN) { state = sb_bt_start; - add_ibs_begin(cpu_buf, - IBS_FETCH_CODE, in_kernel, mm); + add_ibs_begin(cpu_buf, IBS_FETCH_CODE, mm); } else if (s->event == IBS_OP_BEGIN) { state = sb_bt_start; - add_ibs_begin(cpu_buf, - IBS_OP_CODE, in_kernel, mm); + add_ibs_begin(cpu_buf, IBS_OP_CODE, mm); #endif } else { struct mm_struct *oldmm = mm; From 69046d430417c30f48867fc52e892c9050b3a29b Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Fri, 5 Sep 2008 12:17:40 +0200 Subject: [PATCH 15/21] x86/oprofile: reordering functions in nmi_int.c No functional changes. The intension is to remove static function declarations. Signed-off-by: Robert Richter --- arch/x86/oprofile/nmi_int.c | 147 +++++++++++++++++------------------- 1 file changed, 71 insertions(+), 76 deletions(-) diff --git a/arch/x86/oprofile/nmi_int.c b/arch/x86/oprofile/nmi_int.c index 57f6c908808..370d832f398 100644 --- a/arch/x86/oprofile/nmi_int.c +++ b/arch/x86/oprofile/nmi_int.c @@ -28,85 +28,9 @@ static struct op_x86_model_spec const *model; static DEFINE_PER_CPU(struct op_msrs, cpu_msrs); static DEFINE_PER_CPU(unsigned long, saved_lvtpc); -static int nmi_start(void); -static void nmi_stop(void); -static void nmi_cpu_start(void *dummy); -static void nmi_cpu_stop(void *dummy); - /* 0 == registered but off, 1 == registered and on */ static int nmi_enabled = 0; -#ifdef CONFIG_SMP -static int oprofile_cpu_notifier(struct notifier_block *b, unsigned long action, - void *data) -{ - int cpu = (unsigned long)data; - switch (action) { - case CPU_DOWN_FAILED: - case CPU_ONLINE: - smp_call_function_single(cpu, nmi_cpu_start, NULL, 0); - break; - case CPU_DOWN_PREPARE: - smp_call_function_single(cpu, nmi_cpu_stop, NULL, 1); - break; - } - return NOTIFY_DONE; -} - -static struct notifier_block oprofile_cpu_nb = { - .notifier_call = oprofile_cpu_notifier -}; -#endif - -#ifdef CONFIG_PM - -static int nmi_suspend(struct sys_device *dev, pm_message_t state) -{ - /* Only one CPU left, just stop that one */ - if (nmi_enabled == 1) - nmi_cpu_stop(NULL); - return 0; -} - -static int nmi_resume(struct sys_device *dev) -{ - if (nmi_enabled == 1) - nmi_cpu_start(NULL); - return 0; -} - -static struct sysdev_class oprofile_sysclass = { - .name = "oprofile", - .resume = nmi_resume, - .suspend = nmi_suspend, -}; - -static struct sys_device device_oprofile = { - .id = 0, - .cls = &oprofile_sysclass, -}; - -static int __init init_sysfs(void) -{ - int error; - - error = sysdev_class_register(&oprofile_sysclass); - if (!error) - error = sysdev_register(&device_oprofile); - return error; -} - -static void exit_sysfs(void) -{ - sysdev_unregister(&device_oprofile); - sysdev_class_unregister(&oprofile_sysclass); -} - -#else -#define init_sysfs() do { } while (0) -#define exit_sysfs() do { } while (0) -#endif /* CONFIG_PM */ - static int profile_exceptions_notify(struct notifier_block *self, unsigned long val, void *data) { @@ -361,6 +285,77 @@ static int nmi_create_files(struct super_block *sb, struct dentry *root) return 0; } +#ifdef CONFIG_SMP +static int oprofile_cpu_notifier(struct notifier_block *b, unsigned long action, + void *data) +{ + int cpu = (unsigned long)data; + switch (action) { + case CPU_DOWN_FAILED: + case CPU_ONLINE: + smp_call_function_single(cpu, nmi_cpu_start, NULL, 0); + break; + case CPU_DOWN_PREPARE: + smp_call_function_single(cpu, nmi_cpu_stop, NULL, 1); + break; + } + return NOTIFY_DONE; +} + +static struct notifier_block oprofile_cpu_nb = { + .notifier_call = oprofile_cpu_notifier +}; +#endif + +#ifdef CONFIG_PM + +static int nmi_suspend(struct sys_device *dev, pm_message_t state) +{ + /* Only one CPU left, just stop that one */ + if (nmi_enabled == 1) + nmi_cpu_stop(NULL); + return 0; +} + +static int nmi_resume(struct sys_device *dev) +{ + if (nmi_enabled == 1) + nmi_cpu_start(NULL); + return 0; +} + +static struct sysdev_class oprofile_sysclass = { + .name = "oprofile", + .resume = nmi_resume, + .suspend = nmi_suspend, +}; + +static struct sys_device device_oprofile = { + .id = 0, + .cls = &oprofile_sysclass, +}; + +static int __init init_sysfs(void) +{ + int error; + + error = sysdev_class_register(&oprofile_sysclass); + if (!error) + error = sysdev_register(&device_oprofile); + return error; +} + +static void exit_sysfs(void) +{ + sysdev_unregister(&device_oprofile); + sysdev_class_unregister(&oprofile_sysclass); +} + +#else +#define init_sysfs() do { } while (0) +#define exit_sysfs() do { } while (0) +#endif /* CONFIG_PM */ + static int p4force; module_param(p4force, int, 0); From 5f87dfb79f829339508a5d989b8252eb30842587 Mon Sep 17 00:00:00 2001 From: Suravee Suthikulpanit Date: Wed, 15 Oct 2008 08:15:51 -0500 Subject: [PATCH 16/21] x86/oprofile: add the logic for enabling additional IBS bits This patch adds the logic for enabling additional IBS control bits : * IBS-Fetch IbsRandEn bit (bit 57) * IBS-Op IbsOpCntCtl bit (bit 19) Signed-off-by: Suravee Suthikulpanit Signed-off-by: Robert Richter --- arch/x86/oprofile/op_model_amd.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/arch/x86/oprofile/op_model_amd.c b/arch/x86/oprofile/op_model_amd.c index a2e83afbe3e..509513760a6 100644 --- a/arch/x86/oprofile/op_model_amd.c +++ b/arch/x86/oprofile/op_model_amd.c @@ -310,12 +310,15 @@ static void op_amd_start(struct op_msrs const * const msrs) #ifdef CONFIG_OPROFILE_IBS if (ibs_allowed && ibs_config.fetch_enabled) { low = (ibs_config.max_cnt_fetch >> 4) & 0xFFFF; - high = IBS_FETCH_HIGH_ENABLE; + high = ((ibs_config.rand_en & 0x1) << 25) /* bit 57 */ + + IBS_FETCH_HIGH_ENABLE; wrmsr(MSR_AMD64_IBSFETCHCTL, low, high); } if (ibs_allowed && ibs_config.op_enabled) { - low = ((ibs_config.max_cnt_op >> 4) & 0xFFFF) + IBS_OP_LOW_ENABLE; + low = ((ibs_config.max_cnt_op >> 4) & 0xFFFF) + + ((ibs_config.dispatched_ops & 0x1) << 19) /* bit 19 */ + + IBS_OP_LOW_ENABLE; high = 0; wrmsr(MSR_AMD64_IBSOPCTL, low, high); } From 6a18037d4165f691063b43816be3152e9006eb06 Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Thu, 16 Oct 2008 15:01:40 +0200 Subject: [PATCH 17/21] oprofile: fixing whitespaces in drivers/oprofile/* Signed-off-by: Robert Richter --- drivers/oprofile/buffer_sync.h | 4 ++-- drivers/oprofile/cpu_buffer.c | 16 ++++++++-------- drivers/oprofile/cpu_buffer.h | 6 +++--- drivers/oprofile/event_buffer.c | 24 ++++++++++++------------ drivers/oprofile/event_buffer.h | 10 +++++----- drivers/oprofile/oprof.c | 2 +- drivers/oprofile/oprof.h | 8 ++++---- drivers/oprofile/oprofile_files.c | 20 ++++++++++---------- drivers/oprofile/oprofile_stats.c | 14 +++++++------- drivers/oprofile/oprofile_stats.h | 8 ++++---- drivers/oprofile/oprofilefs.c | 6 +++--- drivers/oprofile/timer_int.c | 2 +- 12 files changed, 60 insertions(+), 60 deletions(-) diff --git a/drivers/oprofile/buffer_sync.h b/drivers/oprofile/buffer_sync.h index 08866f6a96a..3110732c183 100644 --- a/drivers/oprofile/buffer_sync.h +++ b/drivers/oprofile/buffer_sync.h @@ -9,13 +9,13 @@ #ifndef OPROFILE_BUFFER_SYNC_H #define OPROFILE_BUFFER_SYNC_H - + /* add the necessary profiling hooks */ int sync_start(void); /* remove the hooks */ void sync_stop(void); - + /* sync the given CPU's buffer */ void sync_buffer(int cpu); diff --git a/drivers/oprofile/cpu_buffer.c b/drivers/oprofile/cpu_buffer.c index b47ce038490..5a178065cfa 100644 --- a/drivers/oprofile/cpu_buffer.c +++ b/drivers/oprofile/cpu_buffer.c @@ -22,7 +22,7 @@ #include #include #include - + #include "event_buffer.h" #include "cpu_buffer.h" #include "buffer_sync.h" @@ -38,7 +38,7 @@ static int work_enabled; void free_cpu_buffers(void) { int i; - + for_each_online_cpu(i) { vfree(per_cpu(cpu_buffer, i).buffer); per_cpu(cpu_buffer, i).buffer = NULL; @@ -48,17 +48,17 @@ void free_cpu_buffers(void) int alloc_cpu_buffers(void) { int i; - + unsigned long buffer_size = fs_cpu_buffer_size; - + for_each_online_cpu(i) { struct oprofile_cpu_buffer *b = &per_cpu(cpu_buffer, i); - + b->buffer = vmalloc_node(sizeof(struct op_sample) * buffer_size, cpu_to_node(i)); if (!b->buffer) goto fail; - + b->last_task = NULL; b->last_is_kernel = -1; b->tracing = 0; @@ -150,7 +150,7 @@ static void increment_head(struct oprofile_cpu_buffer *b) static inline void add_sample(struct oprofile_cpu_buffer *cpu_buf, - unsigned long pc, unsigned long event) + unsigned long pc, unsigned long event) { struct op_sample *entry = &cpu_buf->buffer[cpu_buf->head_pos]; entry->eip = pc; @@ -205,7 +205,7 @@ static int log_sample(struct oprofile_cpu_buffer *cpu_buf, unsigned long pc, cpu_buf->last_task = task; add_code(cpu_buf, (unsigned long)task); } - + add_sample(cpu_buf, pc, event); return 1; } diff --git a/drivers/oprofile/cpu_buffer.h b/drivers/oprofile/cpu_buffer.h index 9bc6bb20b6d..d3cc26264db 100644 --- a/drivers/oprofile/cpu_buffer.h +++ b/drivers/oprofile/cpu_buffer.h @@ -15,9 +15,9 @@ #include #include #include - + struct task_struct; - + int alloc_cpu_buffers(void); void free_cpu_buffers(void); @@ -31,7 +31,7 @@ struct op_sample { unsigned long eip; unsigned long event; }; - + struct oprofile_cpu_buffer { volatile unsigned long head_pos; volatile unsigned long tail_pos; diff --git a/drivers/oprofile/event_buffer.c b/drivers/oprofile/event_buffer.c index c9329f4e090..d962ba0dd87 100644 --- a/drivers/oprofile/event_buffer.c +++ b/drivers/oprofile/event_buffer.c @@ -19,13 +19,13 @@ #include #include #include - + #include "oprof.h" #include "event_buffer.h" #include "oprofile_stats.h" DEFINE_MUTEX(buffer_mutex); - + static unsigned long buffer_opened; static DECLARE_WAIT_QUEUE_HEAD(buffer_wait); static unsigned long *event_buffer; @@ -66,7 +66,7 @@ void wake_up_buffer_waiter(void) mutex_unlock(&buffer_mutex); } - + int alloc_event_buffer(void) { int err = -ENOMEM; @@ -76,13 +76,13 @@ int alloc_event_buffer(void) buffer_size = fs_buffer_size; buffer_watershed = fs_buffer_watershed; spin_unlock_irqrestore(&oprofilefs_lock, flags); - + if (buffer_watershed >= buffer_size) return -EINVAL; - + event_buffer = vmalloc(sizeof(unsigned long) * buffer_size); if (!event_buffer) - goto out; + goto out; err = 0; out: @@ -97,7 +97,7 @@ void free_event_buffer(void) event_buffer = NULL; } - + static int event_buffer_open(struct inode *inode, struct file *file) { int err = -EPERM; @@ -116,14 +116,14 @@ static int event_buffer_open(struct inode *inode, struct file *file) file->private_data = dcookie_register(); if (!file->private_data) goto out; - + if ((err = oprofile_setup())) goto fail; /* NB: the actual start happens from userspace * echo 1 >/dev/oprofile/enable */ - + return 0; fail: @@ -172,18 +172,18 @@ static ssize_t event_buffer_read(struct file *file, char __user *buf, retval = -EFAULT; count = buffer_pos * sizeof(unsigned long); - + if (copy_to_user(buf, event_buffer, count)) goto out; retval = count; buffer_pos = 0; - + out: mutex_unlock(&buffer_mutex); return retval; } - + const struct file_operations event_buffer_fops = { .open = event_buffer_open, .release = event_buffer_release, diff --git a/drivers/oprofile/event_buffer.h b/drivers/oprofile/event_buffer.h index 5076ed1ebd8..00db2e66570 100644 --- a/drivers/oprofile/event_buffer.h +++ b/drivers/oprofile/event_buffer.h @@ -10,13 +10,13 @@ #ifndef EVENT_BUFFER_H #define EVENT_BUFFER_H -#include +#include #include - + int alloc_event_buffer(void); void free_event_buffer(void); - + /* wake up the process sleeping on the event file */ void wake_up_buffer_waiter(void); @@ -24,10 +24,10 @@ void wake_up_buffer_waiter(void); #define NO_COOKIE 0UL extern const struct file_operations event_buffer_fops; - + /* mutex between sync_cpu_buffers() and the * file reading code. */ extern struct mutex buffer_mutex; - + #endif /* EVENT_BUFFER_H */ diff --git a/drivers/oprofile/oprof.c b/drivers/oprofile/oprof.c index 50062cea292..cd375907f26 100644 --- a/drivers/oprofile/oprof.c +++ b/drivers/oprofile/oprof.c @@ -94,7 +94,7 @@ int oprofile_start(void) int err = -EINVAL; mutex_lock(&start_mutex); - + if (!is_setup) goto out; diff --git a/drivers/oprofile/oprof.h b/drivers/oprofile/oprof.h index 7a44ddba0be..5df0c21a608 100644 --- a/drivers/oprofile/oprof.h +++ b/drivers/oprofile/oprof.h @@ -11,7 +11,7 @@ #define OPROF_H int oprofile_setup(void); -void oprofile_shutdown(void); +void oprofile_shutdown(void); int oprofilefs_register(void); void oprofilefs_unregister(void); @@ -20,14 +20,14 @@ int oprofile_start(void); void oprofile_stop(void); struct oprofile_operations; - + extern unsigned long fs_buffer_size; extern unsigned long fs_cpu_buffer_size; extern unsigned long fs_buffer_watershed; extern struct oprofile_operations oprofile_ops; extern unsigned long oprofile_started; extern unsigned long backtrace_depth; - + struct super_block; struct dentry; @@ -35,5 +35,5 @@ void oprofile_create_files(struct super_block *sb, struct dentry *root); void oprofile_timer_init(struct oprofile_operations *ops); int oprofile_set_backtrace(unsigned long depth); - + #endif /* OPROF_H */ diff --git a/drivers/oprofile/oprofile_files.c b/drivers/oprofile/oprofile_files.c index 241804abbb5..cc106d503ac 100644 --- a/drivers/oprofile/oprofile_files.c +++ b/drivers/oprofile/oprofile_files.c @@ -13,7 +13,7 @@ #include "event_buffer.h" #include "oprofile_stats.h" #include "oprof.h" - + unsigned long fs_buffer_size = 131072; unsigned long fs_cpu_buffer_size = 8192; unsigned long fs_buffer_watershed = 32768; /* FIXME: tune */ @@ -49,7 +49,7 @@ static const struct file_operations depth_fops = { .write = depth_write }; - + static ssize_t pointer_size_read(struct file *file, char __user *buf, size_t count, loff_t *offset) { return oprofilefs_ulong_to_user(sizeof(void *), buf, count, offset); @@ -65,13 +65,13 @@ static ssize_t cpu_type_read(struct file *file, char __user *buf, size_t count, { return oprofilefs_str_to_user(oprofile_ops.cpu_type, buf, count, offset); } - - + + static const struct file_operations cpu_type_fops = { .read = cpu_type_read, }; - - + + static ssize_t enable_read(struct file *file, char __user *buf, size_t count, loff_t *offset) { return oprofilefs_ulong_to_user(oprofile_started, buf, count, offset); @@ -89,7 +89,7 @@ static ssize_t enable_write(struct file *file, char const __user *buf, size_t co retval = oprofilefs_ulong_from_user(&val, buf, count); if (retval) return retval; - + if (val) retval = oprofile_start(); else @@ -100,7 +100,7 @@ static ssize_t enable_write(struct file *file, char const __user *buf, size_t co return count; } - + static const struct file_operations enable_fops = { .read = enable_read, .write = enable_write, @@ -117,7 +117,7 @@ static ssize_t dump_write(struct file *file, char const __user *buf, size_t coun static const struct file_operations dump_fops = { .write = dump_write, }; - + void oprofile_create_files(struct super_block *sb, struct dentry *root) { oprofilefs_create_file(sb, root, "enable", &enable_fops); @@ -126,7 +126,7 @@ void oprofile_create_files(struct super_block *sb, struct dentry *root) oprofilefs_create_ulong(sb, root, "buffer_size", &fs_buffer_size); oprofilefs_create_ulong(sb, root, "buffer_watershed", &fs_buffer_watershed); oprofilefs_create_ulong(sb, root, "cpu_buffer_size", &fs_cpu_buffer_size); - oprofilefs_create_file(sb, root, "cpu_type", &cpu_type_fops); + oprofilefs_create_file(sb, root, "cpu_type", &cpu_type_fops); oprofilefs_create_file(sb, root, "backtrace_depth", &depth_fops); oprofilefs_create_file(sb, root, "pointer_size", &pointer_size_fops); oprofile_create_stats_files(sb, root); diff --git a/drivers/oprofile/oprofile_stats.c b/drivers/oprofile/oprofile_stats.c index e0c45498d17..e1f6ce03705 100644 --- a/drivers/oprofile/oprofile_stats.c +++ b/drivers/oprofile/oprofile_stats.c @@ -11,17 +11,17 @@ #include #include #include - + #include "oprofile_stats.h" #include "cpu_buffer.h" - + struct oprofile_stat_struct oprofile_stats; - + void oprofile_reset_stats(void) { struct oprofile_cpu_buffer *cpu_buf; int i; - + for_each_possible_cpu(i) { cpu_buf = &per_cpu(cpu_buffer, i); cpu_buf->sample_received = 0; @@ -29,7 +29,7 @@ void oprofile_reset_stats(void) cpu_buf->backtrace_aborted = 0; cpu_buf->sample_invalid_eip = 0; } - + atomic_set(&oprofile_stats.sample_lost_no_mm, 0); atomic_set(&oprofile_stats.sample_lost_no_mapping, 0); atomic_set(&oprofile_stats.event_lost_overflow, 0); @@ -52,7 +52,7 @@ void oprofile_create_stats_files(struct super_block *sb, struct dentry *root) cpu_buf = &per_cpu(cpu_buffer, i); snprintf(buf, 10, "cpu%d", i); cpudir = oprofilefs_mkdir(sb, dir, buf); - + /* Strictly speaking access to these ulongs is racy, * but we can't simply lock them, and they are * informational only. @@ -66,7 +66,7 @@ void oprofile_create_stats_files(struct super_block *sb, struct dentry *root) oprofilefs_create_ro_ulong(sb, cpudir, "sample_invalid_eip", &cpu_buf->sample_invalid_eip); } - + oprofilefs_create_ro_atomic(sb, dir, "sample_lost_no_mm", &oprofile_stats.sample_lost_no_mm); oprofilefs_create_ro_atomic(sb, dir, "sample_lost_no_mapping", diff --git a/drivers/oprofile/oprofile_stats.h b/drivers/oprofile/oprofile_stats.h index 54e59c29b43..3da0d08dc1f 100644 --- a/drivers/oprofile/oprofile_stats.h +++ b/drivers/oprofile/oprofile_stats.h @@ -11,7 +11,7 @@ #define OPROFILE_STATS_H #include - + struct oprofile_stat_struct { atomic_t sample_lost_no_mm; atomic_t sample_lost_no_mapping; @@ -20,13 +20,13 @@ struct oprofile_stat_struct { }; extern struct oprofile_stat_struct oprofile_stats; - + /* reset all stats to zero */ void oprofile_reset_stats(void); - + struct super_block; struct dentry; - + /* create the stats/ dir */ void oprofile_create_stats_files(struct super_block *sb, struct dentry *root); diff --git a/drivers/oprofile/oprofilefs.c b/drivers/oprofile/oprofilefs.c index a275a3aa5f0..ddc4c59f02d 100644 --- a/drivers/oprofile/oprofilefs.c +++ b/drivers/oprofile/oprofilefs.c @@ -181,13 +181,13 @@ static ssize_t atomic_read_file(struct file *file, char __user *buf, size_t coun atomic_t *val = file->private_data; return oprofilefs_ulong_to_user(atomic_read(val), buf, count, offset); } - + static const struct file_operations atomic_ro_fops = { .read = atomic_read_file, .open = default_open, }; - + int oprofilefs_create_ro_atomic(struct super_block *sb, struct dentry *root, char const *name, atomic_t *val) @@ -201,7 +201,7 @@ int oprofilefs_create_ro_atomic(struct super_block *sb, struct dentry *root, return 0; } - + int oprofilefs_create_file(struct super_block *sb, struct dentry *root, char const *name, const struct file_operations *fops) { diff --git a/drivers/oprofile/timer_int.c b/drivers/oprofile/timer_int.c index 7258b141a51..333f915568c 100644 --- a/drivers/oprofile/timer_int.c +++ b/drivers/oprofile/timer_int.c @@ -19,7 +19,7 @@ static int timer_notify(struct pt_regs *regs) { - oprofile_add_sample(regs, 0); + oprofile_add_sample(regs, 0); return 0; } From 818273ae46bd84251b4b8ea7ed3621e67c69d267 Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Thu, 16 Oct 2008 15:01:40 +0200 Subject: [PATCH 18/21] oprofile: fixing whitespaces in arch/x86/oprofile/* Signed-off-by: Robert Richter --- arch/x86/oprofile/backtrace.c | 3 +-- arch/x86/oprofile/op_counter.h | 18 +++++++++--------- arch/x86/oprofile/op_model_amd.c | 4 ++-- arch/x86/oprofile/op_x86_model.h | 4 ++-- 4 files changed, 14 insertions(+), 15 deletions(-) diff --git a/arch/x86/oprofile/backtrace.c b/arch/x86/oprofile/backtrace.c index e2095cba409..04df67f8a7b 100644 --- a/arch/x86/oprofile/backtrace.c +++ b/arch/x86/oprofile/backtrace.c @@ -52,8 +52,7 @@ struct frame_head { unsigned long ret; } __attribute__((packed)); -static struct frame_head * -dump_user_backtrace(struct frame_head * head) +static struct frame_head *dump_user_backtrace(struct frame_head *head) { struct frame_head bufhead[2]; diff --git a/arch/x86/oprofile/op_counter.h b/arch/x86/oprofile/op_counter.h index 2880b15c467..91b6a116165 100644 --- a/arch/x86/oprofile/op_counter.h +++ b/arch/x86/oprofile/op_counter.h @@ -6,22 +6,22 @@ * * @author John Levon */ - + #ifndef OP_COUNTER_H #define OP_COUNTER_H - + #define OP_MAX_COUNTER 8 - + /* Per-perfctr configuration as set via * oprofilefs. */ struct op_counter_config { - unsigned long count; - unsigned long enabled; - unsigned long event; - unsigned long kernel; - unsigned long user; - unsigned long unit_mask; + unsigned long count; + unsigned long enabled; + unsigned long event; + unsigned long kernel; + unsigned long user; + unsigned long unit_mask; }; extern struct op_counter_config counter_config[]; diff --git a/arch/x86/oprofile/op_model_amd.c b/arch/x86/oprofile/op_model_amd.c index d9faf607b3a..9cc0d4a4bdc 100644 --- a/arch/x86/oprofile/op_model_amd.c +++ b/arch/x86/oprofile/op_model_amd.c @@ -468,9 +468,9 @@ static void clear_ibs_nmi(void) on_each_cpu(apic_clear_ibs_nmi_per_cpu, NULL, 1); } -static int (*create_arch_files)(struct super_block * sb, struct dentry * root); +static int (*create_arch_files)(struct super_block *sb, struct dentry *root); -static int setup_ibs_files(struct super_block * sb, struct dentry * root) +static int setup_ibs_files(struct super_block *sb, struct dentry *root) { char buf[12]; struct dentry *dir; diff --git a/arch/x86/oprofile/op_x86_model.h b/arch/x86/oprofile/op_x86_model.h index 596de7a5559..825e79064d6 100644 --- a/arch/x86/oprofile/op_x86_model.h +++ b/arch/x86/oprofile/op_x86_model.h @@ -22,8 +22,8 @@ struct op_msr { }; struct op_msrs { - struct op_msr * counters; - struct op_msr * controls; + struct op_msr *counters; + struct op_msr *controls; }; struct pt_regs; From 0f019cc477b494dfc472f2a98eb64d02d4937741 Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Thu, 16 Oct 2008 15:01:40 +0200 Subject: [PATCH 19/21] oprofile: fixing whitespaces in arch/x86/oprofile/* Signed-off-by: Robert Richter --- arch/x86/oprofile/backtrace.c | 3 +-- arch/x86/oprofile/op_counter.h | 18 +++++++++--------- 2 files changed, 10 insertions(+), 11 deletions(-) diff --git a/arch/x86/oprofile/backtrace.c b/arch/x86/oprofile/backtrace.c index 36e324139f7..04df67f8a7b 100644 --- a/arch/x86/oprofile/backtrace.c +++ b/arch/x86/oprofile/backtrace.c @@ -52,8 +52,7 @@ struct frame_head { unsigned long ret; } __attribute__((packed)); -static struct frame_head * -dump_user_backtrace(struct frame_head *head) +static struct frame_head *dump_user_backtrace(struct frame_head *head) { struct frame_head bufhead[2]; diff --git a/arch/x86/oprofile/op_counter.h b/arch/x86/oprofile/op_counter.h index 2880b15c467..91b6a116165 100644 --- a/arch/x86/oprofile/op_counter.h +++ b/arch/x86/oprofile/op_counter.h @@ -6,22 +6,22 @@ * * @author John Levon */ - + #ifndef OP_COUNTER_H #define OP_COUNTER_H - + #define OP_MAX_COUNTER 8 - + /* Per-perfctr configuration as set via * oprofilefs. */ struct op_counter_config { - unsigned long count; - unsigned long enabled; - unsigned long event; - unsigned long kernel; - unsigned long user; - unsigned long unit_mask; + unsigned long count; + unsigned long enabled; + unsigned long event; + unsigned long kernel; + unsigned long user; + unsigned long unit_mask; }; extern struct op_counter_config counter_config[]; From 4bd9b9dc97e344670e9e5762399a07dcd5f15311 Mon Sep 17 00:00:00 2001 From: Chris J Arges Date: Wed, 15 Oct 2008 11:03:39 -0500 Subject: [PATCH 20/21] oprofile: hotplug cpu fix This patch addresses problems when hotplugging cpus while profiling. Instead of allocating only online cpus, all possible cpu buffers are allocated, which allows cpus to be onlined during operation. If a cpu is offlined before profiling is shutdown wq_sync_buffer checks for this condition then cancels this work and does not sync this buffer. Signed-off-by: Chris J Arges Signed-off-by: Robert Richter --- drivers/oprofile/cpu_buffer.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) diff --git a/drivers/oprofile/cpu_buffer.c b/drivers/oprofile/cpu_buffer.c index 5a178065cfa..67bcc1c95e6 100644 --- a/drivers/oprofile/cpu_buffer.c +++ b/drivers/oprofile/cpu_buffer.c @@ -39,7 +39,7 @@ void free_cpu_buffers(void) { int i; - for_each_online_cpu(i) { + for_each_possible_cpu(i) { vfree(per_cpu(cpu_buffer, i).buffer); per_cpu(cpu_buffer, i).buffer = NULL; } @@ -51,7 +51,7 @@ int alloc_cpu_buffers(void) unsigned long buffer_size = fs_cpu_buffer_size; - for_each_online_cpu(i) { + for_each_possible_cpu(i) { struct oprofile_cpu_buffer *b = &per_cpu(cpu_buffer, i); b->buffer = vmalloc_node(sizeof(struct op_sample) * buffer_size, @@ -350,6 +350,11 @@ static void wq_sync_buffer(struct work_struct *work) if (b->cpu != smp_processor_id()) { printk(KERN_DEBUG "WQ on CPU%d, prefer CPU%d\n", smp_processor_id(), b->cpu); + + if (!cpu_online(b->cpu)) { + cancel_delayed_work(&b->work); + return; + } } sync_buffer(b->cpu); From 9b1f261166f56d4b2c33fdf5aad64edd5e30b46f Mon Sep 17 00:00:00 2001 From: Barry Kasindorf Date: Tue, 15 Jul 2008 00:10:36 +0200 Subject: [PATCH 21/21] OProfile: Fix buffer synchronization for IBS The patch is needed since there is some IBS code in add_ibs_begin() that handles more than one sample per iteration. This requires calling get_slots() during each loop. This fixes the current problem, but a proper solution that reworks the cpu buffer synchronization is needed here in the future. Signed-off-by: Barry Kasindorf Signed-off-by: Robert Richter --- drivers/oprofile/buffer_sync.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/drivers/oprofile/buffer_sync.c b/drivers/oprofile/buffer_sync.c index 33bfa60b0c6..6c0c92a745d 100644 --- a/drivers/oprofile/buffer_sync.c +++ b/drivers/oprofile/buffer_sync.c @@ -564,9 +564,11 @@ void sync_buffer(int cpu) struct task_struct *new; unsigned long cookie = 0; int in_kernel = 1; - unsigned int i; sync_buffer_state state = sb_buffer_start; +#ifndef CONFIG_OPROFILE_IBS + unsigned int i; unsigned long available; +#endif mutex_lock(&buffer_mutex); @@ -574,9 +576,13 @@ void sync_buffer(int cpu) /* Remember, only we can modify tail_pos */ +#ifndef CONFIG_OPROFILE_IBS available = get_slots(cpu_buf); for (i = 0; i < available; ++i) { +#else + while (get_slots(cpu_buf)) { +#endif struct op_sample *s = &cpu_buf->buffer[cpu_buf->tail_pos]; if (is_code(s->eip)) {