From a831100aeefbe6d9f3e47a3e2712f82c042f1f5c Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Tue, 10 May 2016 16:34:53 -0300 Subject: perf core: Generalize max_stack sysctl handler So that it can be used for other stack related knobs, such as the upcoming one to tweak the max number of of contexts per stack sample. In all those cases we can only change the value if there are no perf sessions collecting stacks, so they need to grab that mutex, etc. Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Peter Zijlstra Link: http://lkml.kernel.org/n/tip-8t3fk94wuzp8m2z1n4gc0s17@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- kernel/events/callchain.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'kernel/events') diff --git a/kernel/events/callchain.c b/kernel/events/callchain.c index b9325e7..7fc8993 100644 --- a/kernel/events/callchain.c +++ b/kernel/events/callchain.c @@ -228,7 +228,8 @@ exit_put: int perf_event_max_stack_handler(struct ctl_table *table, int write, void __user *buffer, size_t *lenp, loff_t *ppos) { - int new_value = sysctl_perf_event_max_stack, ret; + int *value = table->data; + int new_value = *value, ret; struct ctl_table new_table = *table; new_table.data = &new_value; @@ -240,7 +241,7 @@ int perf_event_max_stack_handler(struct ctl_table *table, int write, if (atomic_read(&nr_callchain_events)) ret = -EBUSY; else - sysctl_perf_event_max_stack = new_value; + *value = new_value; mutex_unlock(&callchain_mutex); -- cgit v1.1 From cfbcf468454ab4b20f0b4b62da51920b99fdb19e Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 28 Apr 2016 12:30:53 -0300 Subject: perf core: Pass max stack as a perf_callchain_entry context This makes perf_callchain_{user,kernel}() receive the max stack as context for the perf_callchain_entry, instead of accessing the global sysctl_perf_event_max_stack. Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Alexei Starovoitov Cc: Brendan Gregg Cc: David Ahern Cc: Frederic Weisbecker Cc: He Kuang Cc: Jiri Olsa Cc: Linus Torvalds Cc: Masami Hiramatsu Cc: Milian Wolff Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Stephane Eranian Cc: Thomas Gleixner Cc: Vince Weaver Cc: Wang Nan Cc: Zefan Li Link: http://lkml.kernel.org/n/tip-kolmn1yo40p7jhswxwrc7rrd@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- kernel/events/callchain.c | 20 ++++++++++++-------- 1 file changed, 12 insertions(+), 8 deletions(-) (limited to 'kernel/events') diff --git a/kernel/events/callchain.c b/kernel/events/callchain.c index 7fc8993..af95ad9 100644 --- a/kernel/events/callchain.c +++ b/kernel/events/callchain.c @@ -32,12 +32,12 @@ static DEFINE_MUTEX(callchain_mutex); static struct callchain_cpus_entries *callchain_cpus_entries; -__weak void perf_callchain_kernel(struct perf_callchain_entry *entry, +__weak void perf_callchain_kernel(struct perf_callchain_entry_ctx *entry, struct pt_regs *regs) { } -__weak void perf_callchain_user(struct perf_callchain_entry *entry, +__weak void perf_callchain_user(struct perf_callchain_entry_ctx *entry, struct pt_regs *regs) { } @@ -176,14 +176,15 @@ perf_callchain(struct perf_event *event, struct pt_regs *regs) if (!kernel && !user) return NULL; - return get_perf_callchain(regs, 0, kernel, user, crosstask, true); + return get_perf_callchain(regs, 0, kernel, user, sysctl_perf_event_max_stack, crosstask, true); } struct perf_callchain_entry * get_perf_callchain(struct pt_regs *regs, u32 init_nr, bool kernel, bool user, - bool crosstask, bool add_mark) + u32 max_stack, bool crosstask, bool add_mark) { struct perf_callchain_entry *entry; + struct perf_callchain_entry_ctx ctx; int rctx; entry = get_callchain_entry(&rctx); @@ -193,12 +194,15 @@ get_perf_callchain(struct pt_regs *regs, u32 init_nr, bool kernel, bool user, if (!entry) goto exit_put; + ctx.entry = entry; + ctx.max_stack = max_stack; + entry->nr = init_nr; if (kernel && !user_mode(regs)) { if (add_mark) - perf_callchain_store(entry, PERF_CONTEXT_KERNEL); - perf_callchain_kernel(entry, regs); + perf_callchain_store(&ctx, PERF_CONTEXT_KERNEL); + perf_callchain_kernel(&ctx, regs); } if (user) { @@ -214,8 +218,8 @@ get_perf_callchain(struct pt_regs *regs, u32 init_nr, bool kernel, bool user, goto exit_put; if (add_mark) - perf_callchain_store(entry, PERF_CONTEXT_USER); - perf_callchain_user(entry, regs); + perf_callchain_store(&ctx, PERF_CONTEXT_USER); + perf_callchain_user(&ctx, regs); } } -- cgit v1.1 From 3b1fff08038bd0792b1aa1e9703b2dd0512a3fd0 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Tue, 10 May 2016 18:08:32 -0300 Subject: perf core: Add a 'nr' field to perf_event_callchain_context We will use it to count how many addresses are in the entry->ip[] array, excluding PERF_CONTEXT_{KERNEL,USER,etc} entries, so that we can really return the number of entries specified by the user via the relevant sysctl, kernel.perf_event_max_contexts, or via the per event perf_event_attr.sample_max_stack knob. This way we keep the perf_sample->ip_callchain->nr meaning, that is the number of entries, be it real addresses or PERF_CONTEXT_ entries, while honouring the max_stack knobs, i.e. the end result will be max_stack entries if we have at least that many entries in a given stack trace. Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Peter Zijlstra Link: http://lkml.kernel.org/n/tip-s8teto51tdqvlfhefndtat9r@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- kernel/events/callchain.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'kernel/events') diff --git a/kernel/events/callchain.c b/kernel/events/callchain.c index af95ad9..8774ff8 100644 --- a/kernel/events/callchain.c +++ b/kernel/events/callchain.c @@ -196,8 +196,7 @@ get_perf_callchain(struct pt_regs *regs, u32 init_nr, bool kernel, bool user, ctx.entry = entry; ctx.max_stack = max_stack; - - entry->nr = init_nr; + ctx.nr = entry->nr = init_nr; if (kernel && !user_mode(regs)) { if (add_mark) -- cgit v1.1 From 3e4de4ec4cfea40994b47a79767610153edbf45b Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 12 May 2016 13:01:50 -0300 Subject: perf core: Add perf_callchain_store_context() helper We need have different helpers to account how many contexts we have in the sample and for real addresses, so do it now as a prep patch, to ease review. Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Namhyung Kim Cc: Peter Zijlstra Link: http://lkml.kernel.org/n/tip-q964tnyuqrxw5gld18vizs3c@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- kernel/events/callchain.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel/events') diff --git a/kernel/events/callchain.c b/kernel/events/callchain.c index 8774ff8..ca64573 100644 --- a/kernel/events/callchain.c +++ b/kernel/events/callchain.c @@ -200,7 +200,7 @@ get_perf_callchain(struct pt_regs *regs, u32 init_nr, bool kernel, bool user, if (kernel && !user_mode(regs)) { if (add_mark) - perf_callchain_store(&ctx, PERF_CONTEXT_KERNEL); + perf_callchain_store_context(&ctx, PERF_CONTEXT_KERNEL); perf_callchain_kernel(&ctx, regs); } @@ -217,7 +217,7 @@ get_perf_callchain(struct pt_regs *regs, u32 init_nr, bool kernel, bool user, goto exit_put; if (add_mark) - perf_callchain_store(&ctx, PERF_CONTEXT_USER); + perf_callchain_store_context(&ctx, PERF_CONTEXT_USER); perf_callchain_user(&ctx, regs); } } -- cgit v1.1 From c85b03349640b34f3545503c8429fc43005e9a92 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 12 May 2016 13:06:21 -0300 Subject: perf core: Separate accounting of contexts and real addresses in a stack trace The perf_sample->ip_callchain->nr value includes all the entries in the ip_callchain->ip[] array, real addresses and PERF_CONTEXT_{KERNEL,USER,etc}, while what the user expects is that what is in the kernel.perf_event_max_stack sysctl or in the upcoming per event perf_event_attr.sample_max_stack knob be honoured in terms of IP addresses in the stack trace. So allocate a bunch of extra entries for contexts, and do the accounting via perf_callchain_entry_ctx struct members. A new sysctl, kernel.perf_event_max_contexts_per_stack is also introduced for investigating possible bugs in the callchain implementation by some arch. Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Alexei Starovoitov Cc: Brendan Gregg Cc: David Ahern Cc: Frederic Weisbecker Cc: He Kuang Cc: Jiri Olsa Cc: Masami Hiramatsu Cc: Milian Wolff Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Stephane Eranian Cc: Thomas Gleixner Cc: Vince Weaver Cc: Wang Nan Cc: Zefan Li Link: http://lkml.kernel.org/n/tip-3b4wnqk340c4sg4gwkfdi9yk@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- kernel/events/callchain.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) (limited to 'kernel/events') diff --git a/kernel/events/callchain.c b/kernel/events/callchain.c index ca64573..179ef46 100644 --- a/kernel/events/callchain.c +++ b/kernel/events/callchain.c @@ -19,11 +19,13 @@ struct callchain_cpus_entries { }; int sysctl_perf_event_max_stack __read_mostly = PERF_MAX_STACK_DEPTH; +int sysctl_perf_event_max_contexts_per_stack __read_mostly = PERF_MAX_CONTEXTS_PER_STACK; static inline size_t perf_callchain_entry__sizeof(void) { return (sizeof(struct perf_callchain_entry) + - sizeof(__u64) * sysctl_perf_event_max_stack); + sizeof(__u64) * (sysctl_perf_event_max_stack + + sysctl_perf_event_max_contexts_per_stack)); } static DEFINE_PER_CPU(int, callchain_recursion[PERF_NR_CONTEXTS]); @@ -197,6 +199,8 @@ get_perf_callchain(struct pt_regs *regs, u32 init_nr, bool kernel, bool user, ctx.entry = entry; ctx.max_stack = max_stack; ctx.nr = entry->nr = init_nr; + ctx.contexts = 0; + ctx.contexts_maxed = false; if (kernel && !user_mode(regs)) { if (add_mark) @@ -228,6 +232,10 @@ exit_put: return entry; } +/* + * Used for sysctl_perf_event_max_stack and + * sysctl_perf_event_max_contexts_per_stack. + */ int perf_event_max_stack_handler(struct ctl_table *table, int write, void __user *buffer, size_t *lenp, loff_t *ppos) { -- cgit v1.1