From 5c167b8585c8d91206b395d57011ead7711e322f Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Wed, 17 Dec 2008 09:02:19 +0100 Subject: x86, perfcounters: rename intel_arch_perfmon.h => perf_counter.h Impact: rename include file We'll be providing an asm/perf_counter.h to the generic perfcounter code, so use the already existing x86 file for this purpose and rename it. Signed-off-by: Ingo Molnar --- arch/x86/oprofile/op_model_ppro.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86/oprofile') diff --git a/arch/x86/oprofile/op_model_ppro.c b/arch/x86/oprofile/op_model_ppro.c index e9f80c744cf..07c914555a5 100644 --- a/arch/x86/oprofile/op_model_ppro.c +++ b/arch/x86/oprofile/op_model_ppro.c @@ -18,7 +18,7 @@ #include #include #include -#include +#include #include "op_x86_model.h" #include "op_counter.h" -- cgit v1.2.3 From 5b75af0a02fcf3b8899f38ff6f22164c5d8e2fdd Mon Sep 17 00:00:00 2001 From: Mike Galbraith Date: Wed, 4 Feb 2009 17:11:34 +0100 Subject: perfcounters: fix "perf counters kill oprofile" bug With oprofile as a module, and unloaded by profiling script, both oprofile and kerneltop work fine.. unless you leave kerneltop running when you start profiling, then you may see badness. Signed-off-by: Mike Galbraith Signed-off-by: Ingo Molnar --- arch/x86/oprofile/nmi_int.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) (limited to 'arch/x86/oprofile') diff --git a/arch/x86/oprofile/nmi_int.c b/arch/x86/oprofile/nmi_int.c index 202864ad49a..c638685136e 100644 --- a/arch/x86/oprofile/nmi_int.c +++ b/arch/x86/oprofile/nmi_int.c @@ -40,8 +40,9 @@ static int profile_exceptions_notify(struct notifier_block *self, switch (val) { case DIE_NMI: - if (model->check_ctrs(args->regs, &per_cpu(cpu_msrs, cpu))) - ret = NOTIFY_STOP; + case DIE_NMI_IPI: + model->check_ctrs(args->regs, &per_cpu(cpu_msrs, cpu)); + ret = NOTIFY_STOP; break; default: break; @@ -134,7 +135,7 @@ static void nmi_cpu_setup(void *dummy) static struct notifier_block profile_exceptions_nb = { .notifier_call = profile_exceptions_notify, .next = NULL, - .priority = 0 + .priority = 2 }; static int nmi_setup(void) -- cgit v1.2.3 From 82aa9a1829199233f9bdaf26e2ee271114f4701e Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Thu, 5 Feb 2009 15:23:08 +0100 Subject: perfcounters: fix "perf counters kills oprofile" bug, v2 Impact: fix kernel crash Both oprofile and perfcounters register an NMI die handler, but only one can handle the NMI. Conveniently, oprofile unregisters it's notifier when not actively in use, so setting it's notifier priority higher than perfcounter's allows oprofile to borrow the NMI for the duration of it's run. Tested/works both as module and built-in. While testing, I found that if kerneltop was generating NMIs at very high frequency, the kernel may panic when oprofile registered it's handler. This turned out to be because oprofile registers it's handler before reset_value has been allocated, so if an NMI comes in while it's still setting up, kabOom. Rather than try more invasive changes, I followed the lead of other places in op_model_ppro.c, and simply returned in that highly unlikely event. (debug warnings attached) Signed-off-by: Mike Galbraith Signed-off-by: Ingo Molnar --- arch/x86/oprofile/op_model_ppro.c | 8 ++++++++ 1 file changed, 8 insertions(+) (limited to 'arch/x86/oprofile') diff --git a/arch/x86/oprofile/op_model_ppro.c b/arch/x86/oprofile/op_model_ppro.c index 07c914555a5..85eb6268374 100644 --- a/arch/x86/oprofile/op_model_ppro.c +++ b/arch/x86/oprofile/op_model_ppro.c @@ -126,6 +126,13 @@ static int ppro_check_ctrs(struct pt_regs * const regs, u64 val; int i; + /* + * This can happen if perf counters are in use when + * we steal the die notifier NMI. + */ + if (unlikely(!reset_value)) + goto out; + for (i = 0 ; i < num_counters; ++i) { if (!reset_value[i]) continue; @@ -136,6 +143,7 @@ static int ppro_check_ctrs(struct pt_regs * const regs, } } +out: /* Only P6 based Pentium M need to re-unmask the apic vector but it * doesn't hurt other P6 variant */ apic_write(APIC_LVTPC, apic_read(APIC_LVTPC) & ~APIC_LVT_MASKED); -- cgit v1.2.3 From 1dcdb5a9e7c235e6e80f1f4d5b8247b3e5347e48 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Mon, 27 Apr 2009 17:44:11 +0200 Subject: oprofile: re-add force_arch_perfmon option This re-adds the force_arch_perfmon option that was in the original arch perfmon patchkit. Originally this was rejected in favour of a generalized perfmon=name option, but it turned out implementing the later in a reliable way is hard (and it would have been easy to crash the kernel if a user gets it wrong) But now Atom and Core i7 support being readded a user would need to update their oprofile userland to beyond 0.9.4 to use oprofile again on Atom or Core i7. To avoid this problem readd the force_arch_perfmon option. Signed-off-by: Andi Kleen Signed-off-by: Robert Richter --- arch/x86/oprofile/nmi_int.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'arch/x86/oprofile') diff --git a/arch/x86/oprofile/nmi_int.c b/arch/x86/oprofile/nmi_int.c index 202864ad49a..e5171c99e15 100644 --- a/arch/x86/oprofile/nmi_int.c +++ b/arch/x86/oprofile/nmi_int.c @@ -389,10 +389,16 @@ static int __init p4_init(char **cpu_type) return 0; } +int force_arch_perfmon; +module_param(force_arch_perfmon, int, 0); + static int __init ppro_init(char **cpu_type) { __u8 cpu_model = boot_cpu_data.x86_model; + if (force_arch_perfmon && cpu_has_arch_perfmon) + return 0; + switch (cpu_model) { case 0 ... 2: *cpu_type = "i386/ppro"; -- cgit v1.2.3 From 1f3d7b60691993d8d368d8dd7d5d85871d41e8f5 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Mon, 27 Apr 2009 17:44:12 +0200 Subject: oprofile: remove undocumented oprofile.p4force option There are no new P4s and the oprofile code knows about all existing ones, so we don't really need the p4force option anymore. Remove it. Signed-off-by: Andi Kleen Signed-off-by: Robert Richter --- arch/x86/oprofile/nmi_int.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) (limited to 'arch/x86/oprofile') diff --git a/arch/x86/oprofile/nmi_int.c b/arch/x86/oprofile/nmi_int.c index e5171c99e15..f472c0c48a3 100644 --- a/arch/x86/oprofile/nmi_int.c +++ b/arch/x86/oprofile/nmi_int.c @@ -356,14 +356,11 @@ static void exit_sysfs(void) #define exit_sysfs() do { } while (0) #endif /* CONFIG_PM */ -static int p4force; -module_param(p4force, int, 0); - static int __init p4_init(char **cpu_type) { __u8 cpu_model = boot_cpu_data.x86_model; - if (!p4force && (cpu_model > 6 || cpu_model == 5)) + if (cpu_model > 6 || cpu_model == 5) return 0; #ifndef CONFIG_SMP -- cgit v1.2.3 From 6adf406f0a0eaf37251018d15f51e93f5b538ee6 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Mon, 27 Apr 2009 17:44:13 +0200 Subject: oprofile: add support for Core i7 and Atom The registers are about the same as other Family 6 CPUs so we only need to add detection. I'm not completely happy with calling Nehalem Core i7 because there will be undoubtedly other Nehalem based CPUs in the future with different marketing names, but it's the best we got for now. Requires updated oprofile userland for the new event files. If you don't want to update right now you can also use oprofile.force_arch_perfmon=1 (added in the next patch) with 0.9.4 Signed-off-by: Andi Kleen Signed-off-by: Robert Richter --- arch/x86/oprofile/nmi_int.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'arch/x86/oprofile') diff --git a/arch/x86/oprofile/nmi_int.c b/arch/x86/oprofile/nmi_int.c index f472c0c48a3..3308147182a 100644 --- a/arch/x86/oprofile/nmi_int.c +++ b/arch/x86/oprofile/nmi_int.c @@ -417,6 +417,13 @@ static int __init ppro_init(char **cpu_type) case 15: case 23: *cpu_type = "i386/core_2"; break; + case 26: + arch_perfmon_setup_counters(); + *cpu_type = "i386/core_i7"; + break; + case 28: + *cpu_type = "i386/atom"; + break; default: /* Unknown */ return 0; -- cgit v1.2.3 From 7e4e0bd50e80df2fe5501f48f872448376cdd997 Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Wed, 6 May 2009 12:10:23 +0200 Subject: oprofile: introduce module_param oprofile.cpu_type This patch removes module_param oprofile.force_arch_perfmon and introduces oprofile.cpu_type=archperfmon instead. This new parameter can be reused for other models and architectures. Currently only archperfmon is supported. Cc: Andi Kleen Signed-off-by: Robert Richter --- arch/x86/oprofile/nmi_int.c | 13 +++++++++++-- 1 file changed, 11 insertions(+), 2 deletions(-) (limited to 'arch/x86/oprofile') diff --git a/arch/x86/oprofile/nmi_int.c b/arch/x86/oprofile/nmi_int.c index 3308147182a..3b285e656e2 100644 --- a/arch/x86/oprofile/nmi_int.c +++ b/arch/x86/oprofile/nmi_int.c @@ -386,8 +386,17 @@ static int __init p4_init(char **cpu_type) return 0; } -int force_arch_perfmon; -module_param(force_arch_perfmon, int, 0); +static int force_arch_perfmon; +static int force_cpu_type(const char *str, struct kernel_param *kp) +{ + if (!strcmp(str, "archperfmon")) { + force_arch_perfmon = 1; + printk(KERN_INFO "oprofile: forcing architectural perfmon\n"); + } + + return 0; +} +module_param_call(cpu_type, force_cpu_type, NULL, NULL, 0); static int __init ppro_init(char **cpu_type) { -- cgit v1.2.3