From 3d90a00763b51e1db344a7430c966be723b67a29 Mon Sep 17 00:00:00 2001 From: Matt Fleming Date: Mon, 27 Sep 2010 20:45:08 +0100 Subject: oprofile: Abstract the perf-events backend Move the perf-events backend from arch/arm/oprofile into drivers/oprofile so that the code can be shared between architectures. This allows each architecture to maintain only a single copy of the PMU accessor functions instead of one for both perf and OProfile. It also becomes possible for other architectures to delete much of their OProfile code in favour of the common code now available in drivers/oprofile/oprofile_perf.c. Signed-off-by: Matt Fleming Tested-by: Will Deacon Signed-off-by: Robert Richter --- drivers/oprofile/oprofile_perf.c | 326 +++++++++++++++++++++++++++++++++++++++ 1 file changed, 326 insertions(+) create mode 100644 drivers/oprofile/oprofile_perf.c (limited to 'drivers/oprofile/oprofile_perf.c') diff --git a/drivers/oprofile/oprofile_perf.c b/drivers/oprofile/oprofile_perf.c new file mode 100644 index 00000000000..ebb40cb8747 --- /dev/null +++ b/drivers/oprofile/oprofile_perf.c @@ -0,0 +1,326 @@ +/* + * Copyright 2010 ARM Ltd. + * + * Perf-events backend for OProfile. + */ +#include +#include +#include + +/* + * Per performance monitor configuration as set via oprofilefs. + */ +struct op_counter_config { + unsigned long count; + unsigned long enabled; + unsigned long event; + unsigned long unit_mask; + unsigned long kernel; + unsigned long user; + struct perf_event_attr attr; +}; + +static int oprofile_perf_enabled; +static DEFINE_MUTEX(oprofile_perf_mutex); + +static struct op_counter_config *counter_config; +static struct perf_event **perf_events[nr_cpumask_bits]; +static int num_counters; + +/* + * Overflow callback for oprofile. + */ +static void op_overflow_handler(struct perf_event *event, int unused, + struct perf_sample_data *data, struct pt_regs *regs) +{ + int id; + u32 cpu = smp_processor_id(); + + for (id = 0; id < num_counters; ++id) + if (perf_events[cpu][id] == event) + break; + + if (id != num_counters) + oprofile_add_sample(regs, id); + else + pr_warning("oprofile: ignoring spurious overflow " + "on cpu %u\n", cpu); +} + +/* + * Called by oprofile_perf_setup to create perf attributes to mirror the oprofile + * settings in counter_config. Attributes are created as `pinned' events and + * so are permanently scheduled on the PMU. + */ +static void op_perf_setup(void) +{ + int i; + u32 size = sizeof(struct perf_event_attr); + struct perf_event_attr *attr; + + for (i = 0; i < num_counters; ++i) { + attr = &counter_config[i].attr; + memset(attr, 0, size); + attr->type = PERF_TYPE_RAW; + attr->size = size; + attr->config = counter_config[i].event; + attr->sample_period = counter_config[i].count; + attr->pinned = 1; + } +} + +static int op_create_counter(int cpu, int event) +{ + int ret = 0; + struct perf_event *pevent; + + if (!counter_config[event].enabled || (perf_events[cpu][event] != NULL)) + return ret; + + pevent = perf_event_create_kernel_counter(&counter_config[event].attr, + cpu, -1, + op_overflow_handler); + + if (IS_ERR(pevent)) { + ret = PTR_ERR(pevent); + } else if (pevent->state != PERF_EVENT_STATE_ACTIVE) { + pr_warning("oprofile: failed to enable event %d " + "on CPU %d\n", event, cpu); + ret = -EBUSY; + } else { + perf_events[cpu][event] = pevent; + } + + return ret; +} + +static void op_destroy_counter(int cpu, int event) +{ + struct perf_event *pevent = perf_events[cpu][event]; + + if (pevent) { + perf_event_release_kernel(pevent); + perf_events[cpu][event] = NULL; + } +} + +/* + * Called by oprofile_perf_start to create active perf events based on the + * perviously configured attributes. + */ +static int op_perf_start(void) +{ + int cpu, event, ret = 0; + + for_each_online_cpu(cpu) { + for (event = 0; event < num_counters; ++event) { + ret = op_create_counter(cpu, event); + if (ret) + goto out; + } + } + +out: + return ret; +} + +/* + * Called by oprofile_perf_stop at the end of a profiling run. + */ +static void op_perf_stop(void) +{ + int cpu, event; + + for_each_online_cpu(cpu) + for (event = 0; event < num_counters; ++event) + op_destroy_counter(cpu, event); +} + +static int oprofile_perf_create_files(struct super_block *sb, struct dentry *root) +{ + unsigned int i; + + for (i = 0; i < num_counters; i++) { + struct dentry *dir; + char buf[4]; + + snprintf(buf, sizeof buf, "%d", i); + dir = oprofilefs_mkdir(sb, root, buf); + oprofilefs_create_ulong(sb, dir, "enabled", &counter_config[i].enabled); + oprofilefs_create_ulong(sb, dir, "event", &counter_config[i].event); + oprofilefs_create_ulong(sb, dir, "count", &counter_config[i].count); + 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); + } + + return 0; +} + +static int oprofile_perf_setup(void) +{ + spin_lock(&oprofilefs_lock); + op_perf_setup(); + spin_unlock(&oprofilefs_lock); + return 0; +} + +static int oprofile_perf_start(void) +{ + int ret = -EBUSY; + + mutex_lock(&oprofile_perf_mutex); + if (!oprofile_perf_enabled) { + ret = 0; + op_perf_start(); + oprofile_perf_enabled = 1; + } + mutex_unlock(&oprofile_perf_mutex); + return ret; +} + +static void oprofile_perf_stop(void) +{ + mutex_lock(&oprofile_perf_mutex); + if (oprofile_perf_enabled) + op_perf_stop(); + oprofile_perf_enabled = 0; + mutex_unlock(&oprofile_perf_mutex); +} + +#ifdef CONFIG_PM +static int oprofile_perf_suspend(struct platform_device *dev, pm_message_t state) +{ + mutex_lock(&oprofile_perf_mutex); + if (oprofile_perf_enabled) + op_perf_stop(); + mutex_unlock(&oprofile_perf_mutex); + return 0; +} + +static int oprofile_perf_resume(struct platform_device *dev) +{ + mutex_lock(&oprofile_perf_mutex); + if (oprofile_perf_enabled && op_perf_start()) + oprofile_perf_enabled = 0; + mutex_unlock(&oprofile_perf_mutex); + return 0; +} + +static struct platform_driver oprofile_driver = { + .driver = { + .name = "oprofile-perf", + }, + .resume = oprofile_perf_resume, + .suspend = oprofile_perf_suspend, +}; + +static struct platform_device *oprofile_pdev; + +static int __init init_driverfs(void) +{ + int ret; + + ret = platform_driver_register(&oprofile_driver); + if (ret) + goto out; + + oprofile_pdev = platform_device_register_simple( + oprofile_driver.driver.name, 0, NULL, 0); + if (IS_ERR(oprofile_pdev)) { + ret = PTR_ERR(oprofile_pdev); + platform_driver_unregister(&oprofile_driver); + } + +out: + return ret; +} + +static void __exit exit_driverfs(void) +{ + platform_device_unregister(oprofile_pdev); + platform_driver_unregister(&oprofile_driver); +} +#else +static int __init init_driverfs(void) { return 0; } +#define exit_driverfs() do { } while (0) +#endif /* CONFIG_PM */ + +int __init oprofile_perf_init(struct oprofile_operations *ops) +{ + int cpu, ret = 0; + + memset(&perf_events, 0, sizeof(perf_events)); + + num_counters = perf_num_counters(); + if (num_counters <= 0) { + pr_info("oprofile: no performance counters\n"); + ret = -ENODEV; + goto out; + } + + counter_config = kcalloc(num_counters, + sizeof(struct op_counter_config), GFP_KERNEL); + + if (!counter_config) { + pr_info("oprofile: failed to allocate %d " + "counters\n", num_counters); + ret = -ENOMEM; + goto out; + } + + ret = init_driverfs(); + if (ret) + goto out; + + for_each_possible_cpu(cpu) { + perf_events[cpu] = kcalloc(num_counters, + sizeof(struct perf_event *), GFP_KERNEL); + if (!perf_events[cpu]) { + pr_info("oprofile: failed to allocate %d perf events " + "for cpu %d\n", num_counters, cpu); + ret = -ENOMEM; + goto out; + } + } + + ops->create_files = oprofile_perf_create_files; + ops->setup = oprofile_perf_setup; + ops->start = oprofile_perf_start; + ops->stop = oprofile_perf_stop; + ops->shutdown = oprofile_perf_stop; + ops->cpu_type = op_name_from_perf_id(); + + if (!ops->cpu_type) + ret = -ENODEV; + else + pr_info("oprofile: using %s\n", ops->cpu_type); + +out: + if (ret) { + for_each_possible_cpu(cpu) + kfree(perf_events[cpu]); + kfree(counter_config); + } + + return ret; +} + +void __exit oprofile_perf_exit(void) +{ + int cpu, id; + struct perf_event *event; + + for_each_possible_cpu(cpu) { + for (id = 0; id < num_counters; ++id) { + event = perf_events[cpu][id]; + if (event) + perf_event_release_kernel(event); + } + + kfree(perf_events[cpu]); + } + + kfree(counter_config); + exit_driverfs(); +} -- cgit v1.2.3-70-g09d2 From 81771974ae49bf79aab60c42eac7a6d730a9ef2b Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Wed, 29 Sep 2010 16:52:25 +0200 Subject: oprofile, ARM: Release resources on failure This patch fixes a resource leak on failure, where the oprofilefs and some counters may not released properly. Signed-off-by: Robert Richter Acked-by: Will Deacon Cc: linux-arm-kernel@lists.infradead.org Cc: # .35.x LKML-Reference: <20100929145225.GJ13563@erda.amd.com> Signed-off-by: Ingo Molnar --- drivers/oprofile/oprofile_perf.c | 1 + 1 file changed, 1 insertion(+) (limited to 'drivers/oprofile/oprofile_perf.c') diff --git a/drivers/oprofile/oprofile_perf.c b/drivers/oprofile/oprofile_perf.c index ebb40cb8747..f3d3df229a4 100644 --- a/drivers/oprofile/oprofile_perf.c +++ b/drivers/oprofile/oprofile_perf.c @@ -84,6 +84,7 @@ static int op_create_counter(int cpu, int event) if (IS_ERR(pevent)) { ret = PTR_ERR(pevent); } else if (pevent->state != PERF_EVENT_STATE_ACTIVE) { + perf_event_release_kernel(pevent); pr_warning("oprofile: failed to enable event %d " "on CPU %d\n", event, cpu); ret = -EBUSY; -- cgit v1.2.3-70-g09d2 From 9c91283a19c2d998a83f50f113f8585709c15caf Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Fri, 27 Aug 2010 14:32:41 +0200 Subject: oprofile, ARM: Remove some goto statements This patch removes some unnecessary goto statements. Acked-by: Will Deacon Signed-off-by: Robert Richter --- drivers/oprofile/oprofile_perf.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) (limited to 'drivers/oprofile/oprofile_perf.c') diff --git a/drivers/oprofile/oprofile_perf.c b/drivers/oprofile/oprofile_perf.c index f3d3df229a4..6853634c468 100644 --- a/drivers/oprofile/oprofile_perf.c +++ b/drivers/oprofile/oprofile_perf.c @@ -117,11 +117,10 @@ static int op_perf_start(void) for (event = 0; event < num_counters; ++event) { ret = op_create_counter(cpu, event); if (ret) - goto out; + return ret; } } -out: return ret; } @@ -224,7 +223,7 @@ static int __init init_driverfs(void) ret = platform_driver_register(&oprofile_driver); if (ret) - goto out; + return ret; oprofile_pdev = platform_device_register_simple( oprofile_driver.driver.name, 0, NULL, 0); @@ -233,7 +232,6 @@ static int __init init_driverfs(void) platform_driver_unregister(&oprofile_driver); } -out: return ret; } -- cgit v1.2.3-70-g09d2 From 2bcb2b641a8f1524f7a9801eb9e52a00b8f18c6e Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Wed, 29 Sep 2010 14:43:29 +0200 Subject: oprofile, ARM: Rework op_create_counter() This patch simplifies op_create_counter(). Removing if/else if paths and return code variable by direct returning from function. Acked-by: Will Deacon Signed-off-by: Robert Richter --- drivers/oprofile/oprofile_perf.c | 20 ++++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) (limited to 'drivers/oprofile/oprofile_perf.c') diff --git a/drivers/oprofile/oprofile_perf.c b/drivers/oprofile/oprofile_perf.c index 6853634c468..a34137f2e26 100644 --- a/drivers/oprofile/oprofile_perf.c +++ b/drivers/oprofile/oprofile_perf.c @@ -71,28 +71,28 @@ static void op_perf_setup(void) static int op_create_counter(int cpu, int event) { - int ret = 0; struct perf_event *pevent; - if (!counter_config[event].enabled || (perf_events[cpu][event] != NULL)) - return ret; + if (!counter_config[event].enabled || perf_events[cpu][event]) + return 0; pevent = perf_event_create_kernel_counter(&counter_config[event].attr, cpu, -1, op_overflow_handler); - if (IS_ERR(pevent)) { - ret = PTR_ERR(pevent); - } else if (pevent->state != PERF_EVENT_STATE_ACTIVE) { + if (IS_ERR(pevent)) + return PTR_ERR(pevent); + + if (pevent->state != PERF_EVENT_STATE_ACTIVE) { perf_event_release_kernel(pevent); pr_warning("oprofile: failed to enable event %d " "on CPU %d\n", event, cpu); - ret = -EBUSY; - } else { - perf_events[cpu][event] = pevent; + return -EBUSY; } - return ret; + perf_events[cpu][event] = pevent; + + return 0; } static void op_destroy_counter(int cpu, int event) -- cgit v1.2.3-70-g09d2 From e9677b3ce207a07fad5746b6f7ddc70cae79de0a Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Wed, 29 Sep 2010 15:42:30 +0200 Subject: oprofile, ARM: Use oprofile_arch_exit() to cleanup on failure There is duplicate cleanup code in the init and exit functions. Now, oprofile_arch_exit() is also used if oprofile_arch_init() fails. Acked-by: Will Deacon Signed-off-by: Robert Richter --- drivers/oprofile/oprofile_perf.c | 54 +++++++++++++++++++--------------------- 1 file changed, 26 insertions(+), 28 deletions(-) (limited to 'drivers/oprofile/oprofile_perf.c') diff --git a/drivers/oprofile/oprofile_perf.c b/drivers/oprofile/oprofile_perf.c index a34137f2e26..b17235a24a4 100644 --- a/drivers/oprofile/oprofile_perf.c +++ b/drivers/oprofile/oprofile_perf.c @@ -245,10 +245,33 @@ static int __init init_driverfs(void) { return 0; } #define exit_driverfs() do { } while (0) #endif /* CONFIG_PM */ +void oprofile_perf_exit(void) +{ + int cpu, id; + struct perf_event *event; + + for_each_possible_cpu(cpu) { + for (id = 0; id < num_counters; ++id) { + event = perf_events[cpu][id]; + if (event) + perf_event_release_kernel(event); + } + + kfree(perf_events[cpu]); + } + + kfree(counter_config); + exit_driverfs(); +} + int __init oprofile_perf_init(struct oprofile_operations *ops) { int cpu, ret = 0; + ret = init_driverfs(); + if (ret) + return ret; + memset(&perf_events, 0, sizeof(perf_events)); num_counters = perf_num_counters(); @@ -265,13 +288,10 @@ int __init oprofile_perf_init(struct oprofile_operations *ops) pr_info("oprofile: failed to allocate %d " "counters\n", num_counters); ret = -ENOMEM; + num_counters = 0; goto out; } - ret = init_driverfs(); - if (ret) - goto out; - for_each_possible_cpu(cpu) { perf_events[cpu] = kcalloc(num_counters, sizeof(struct perf_event *), GFP_KERNEL); @@ -296,30 +316,8 @@ int __init oprofile_perf_init(struct oprofile_operations *ops) pr_info("oprofile: using %s\n", ops->cpu_type); out: - if (ret) { - for_each_possible_cpu(cpu) - kfree(perf_events[cpu]); - kfree(counter_config); - } + if (ret) + oprofile_perf_exit(); return ret; } - -void __exit oprofile_perf_exit(void) -{ - int cpu, id; - struct perf_event *event; - - for_each_possible_cpu(cpu) { - for (id = 0; id < num_counters; ++id) { - event = perf_events[cpu][id]; - if (event) - perf_event_release_kernel(event); - } - - kfree(perf_events[cpu]); - } - - kfree(counter_config); - exit_driverfs(); -} -- cgit v1.2.3-70-g09d2 From 277dd984172b063497d2ff6cfa7f2355f13a292d Mon Sep 17 00:00:00 2001 From: Anand Gadiyar Date: Thu, 14 Oct 2010 11:31:43 -0400 Subject: oprofile: include platform_device.h to fix build break oprofile_perf.c needs to include platform_device.h Otherwise we get the following build break. CC arch/arm/oprofile/../../../drivers/oprofile/oprofile_perf.o arch/arm/oprofile/../../../drivers/oprofile/oprofile_perf.c:192: warning: 'struct platform_device' declared inside parameter list arch/arm/oprofile/../../../drivers/oprofile/oprofile_perf.c:192: warning: its scope is only this definition or declaration, which is probably not what you want arch/arm/oprofile/../../../drivers/oprofile/oprofile_perf.c:201: warning: 'struct platform_device' declared inside parameter list arch/arm/oprofile/../../../drivers/oprofile/oprofile_perf.c:210: error: variable 'oprofile_driver' has initializer but incomplete type arch/arm/oprofile/../../../drivers/oprofile/oprofile_perf.c:211: error: unknown field 'driver' specified in initializer arch/arm/oprofile/../../../drivers/oprofile/oprofile_perf.c:211: error: extra brace group at end of initializer arch/arm/oprofile/../../../drivers/oprofile/oprofile_perf.c:211: error: (near initialization for 'oprofile_driver') arch/arm/oprofile/../../../drivers/oprofile/oprofile_perf.c:213: warning: excess elements in struct initializer arch/arm/oprofile/../../../drivers/oprofile/oprofile_perf.c:213: warning: (near initialization for 'oprofile_driver') arch/arm/oprofile/../../../drivers/oprofile/oprofile_perf.c:214: error: unknown field 'resume' specified in initializer arch/arm/oprofile/../../../drivers/oprofile/oprofile_perf.c:214: warning: excess elements in struct initializer arch/arm/oprofile/../../../drivers/oprofile/oprofile_perf.c:214: warning: (near initialization for 'oprofile_driver') arch/arm/oprofile/../../../drivers/oprofile/oprofile_perf.c:215: error: unknown field 'suspend' specified in initializer arch/arm/oprofile/../../../drivers/oprofile/oprofile_perf.c:215: warning: excess elements in struct initializer arch/arm/oprofile/../../../drivers/oprofile/oprofile_perf.c:215: warning: (near initialization for 'oprofile_driver') arch/arm/oprofile/../../../drivers/oprofile/oprofile_perf.c: In function 'init_driverfs': Signed-off-by: Anand Gadiyar Cc: Matt Fleming Cc: Will Deacon Signed-off-by: Robert Richter --- drivers/oprofile/oprofile_perf.c | 1 + 1 file changed, 1 insertion(+) (limited to 'drivers/oprofile/oprofile_perf.c') diff --git a/drivers/oprofile/oprofile_perf.c b/drivers/oprofile/oprofile_perf.c index 79c0005134a..e707a72a379 100644 --- a/drivers/oprofile/oprofile_perf.c +++ b/drivers/oprofile/oprofile_perf.c @@ -4,6 +4,7 @@ * Perf-events backend for OProfile. */ #include +#include #include #include -- cgit v1.2.3-70-g09d2 From b3b3a9b63f2deacfd59137e3781211d21a568ca9 Mon Sep 17 00:00:00 2001 From: Anand Gadiyar Date: Thu, 14 Oct 2010 11:31:42 -0400 Subject: oprofile: fix linker errors Commit e9677b3ce (oprofile, ARM: Use oprofile_arch_exit() to cleanup on failure) caused oprofile_perf_exit to be called in the cleanup path of oprofile_perf_init. The __exit tag for oprofile_perf_exit should therefore be dropped. The same has to be done for exit_driverfs as well, as this function is called from oprofile_perf_exit. Else, we get the following two linker errors. LD .tmp_vmlinux1 `oprofile_perf_exit' referenced in section `.init.text' of arch/arm/oprofile/built-in.o: defined in discarded section `.exit.text' of arch/arm/oprofile/built-in.o make: *** [.tmp_vmlinux1] Error 1 LD .tmp_vmlinux1 `exit_driverfs' referenced in section `.text' of arch/arm/oprofile/built-in.o: defined in discarded section `.exit.text' of arch/arm/oprofile/built-in.o make: *** [.tmp_vmlinux1] Error 1 Signed-off-by: Anand Gadiyar Cc: Will Deacon Signed-off-by: Robert Richter --- drivers/oprofile/oprofile_perf.c | 2 +- include/linux/oprofile.h | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'drivers/oprofile/oprofile_perf.c') diff --git a/drivers/oprofile/oprofile_perf.c b/drivers/oprofile/oprofile_perf.c index e707a72a379..36ec67e3d91 100644 --- a/drivers/oprofile/oprofile_perf.c +++ b/drivers/oprofile/oprofile_perf.c @@ -236,7 +236,7 @@ static int __init init_driverfs(void) return ret; } -static void __exit exit_driverfs(void) +static void exit_driverfs(void) { platform_device_unregister(oprofile_pdev); platform_driver_unregister(&oprofile_driver); diff --git a/include/linux/oprofile.h b/include/linux/oprofile.h index d67a8330b41..32fb81212fd 100644 --- a/include/linux/oprofile.h +++ b/include/linux/oprofile.h @@ -188,7 +188,7 @@ int oprofile_write_commit(struct op_entry *entry); #ifdef CONFIG_PERF_EVENTS int __init oprofile_perf_init(struct oprofile_operations *ops); -void __exit oprofile_perf_exit(void); +void oprofile_perf_exit(void); char *op_name_from_perf_id(void); #endif /* CONFIG_PERF_EVENTS */ -- cgit v1.2.3-70-g09d2 From cd254f295248c98b62ea824f361e10d80a081fe7 Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Fri, 15 Oct 2010 11:28:07 +0200 Subject: oprofile: make !CONFIG_PM function stubs static inline Make !CONFIG_PM function stubs static inline and remove section attribute. Signed-off-by: Robert Richter --- drivers/oprofile/oprofile_perf.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) (limited to 'drivers/oprofile/oprofile_perf.c') diff --git a/drivers/oprofile/oprofile_perf.c b/drivers/oprofile/oprofile_perf.c index 36ec67e3d91..9046f7b2ed7 100644 --- a/drivers/oprofile/oprofile_perf.c +++ b/drivers/oprofile/oprofile_perf.c @@ -190,6 +190,7 @@ static void oprofile_perf_stop(void) } #ifdef CONFIG_PM + static int oprofile_perf_suspend(struct platform_device *dev, pm_message_t state) { mutex_lock(&oprofile_perf_mutex); @@ -241,9 +242,12 @@ static void exit_driverfs(void) platform_device_unregister(oprofile_pdev); platform_driver_unregister(&oprofile_driver); } + #else -static int __init init_driverfs(void) { return 0; } -#define exit_driverfs() do { } while (0) + +static inline int init_driverfs(void) { return 0; } +static inline void exit_driverfs(void) { } + #endif /* CONFIG_PM */ void oprofile_perf_exit(void) -- cgit v1.2.3-70-g09d2