From 61c1917f47f73c968e92d04d15370b1dc3ec4592 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 17 Dec 2009 05:40:33 +0100 Subject: perf events, x86/stacktrace: Make stack walking optional The current print_context_stack helper that does the stack walking job is good for usual stacktraces as it walks through all the stack and reports even addresses that look unreliable, which is nice when we don't have frame pointers for example. But we have users like perf that only require reliable stacktraces, and those may want a more adapted stack walker, so lets make this function a callback in stacktrace_ops that users can tune for their needs. Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Arnaldo Carvalho de Melo Cc: Paul Mackerras LKML-Reference: <1261024834-5336-1-git-send-regression-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- arch/x86/include/asm/stacktrace.h | 18 ++++++++++++++++++ arch/x86/kernel/cpu/perf_event.c | 1 + arch/x86/kernel/dumpstack.c | 9 +++++---- arch/x86/kernel/dumpstack.h | 6 ------ arch/x86/kernel/dumpstack_32.c | 2 +- arch/x86/kernel/dumpstack_64.c | 4 ++-- arch/x86/kernel/stacktrace.c | 18 ++++++++++-------- arch/x86/oprofile/backtrace.c | 9 +++++---- 8 files changed, 42 insertions(+), 25 deletions(-) (limited to 'arch') diff --git a/arch/x86/include/asm/stacktrace.h b/arch/x86/include/asm/stacktrace.h index cf86a5e7381..6c75151a3cc 100644 --- a/arch/x86/include/asm/stacktrace.h +++ b/arch/x86/include/asm/stacktrace.h @@ -5,6 +5,23 @@ extern int kstack_depth_to_print; int x86_is_stack_id(int id, char *name); +struct thread_info; +struct stacktrace_ops; + +typedef unsigned long (*walk_stack_t)(struct thread_info *tinfo, + unsigned long *stack, + unsigned long bp, + const struct stacktrace_ops *ops, + void *data, + unsigned long *end, + int *graph); + +extern unsigned long +print_context_stack(struct thread_info *tinfo, + unsigned long *stack, unsigned long bp, + const struct stacktrace_ops *ops, void *data, + unsigned long *end, int *graph); + /* Generic stack tracer with callbacks */ struct stacktrace_ops { @@ -14,6 +31,7 @@ struct stacktrace_ops { void (*address)(void *data, unsigned long address, int reliable); /* On negative return stop dumping */ int (*stack)(void *data, char *name); + walk_stack_t walk_stack; }; void dump_trace(struct task_struct *tsk, struct pt_regs *regs, diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 45506d5dd8d..d3802ee5a41 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -2336,6 +2336,7 @@ static const struct stacktrace_ops backtrace_ops = { .warning_symbol = backtrace_warning_symbol, .stack = backtrace_stack, .address = backtrace_address, + .walk_stack = print_context_stack, }; #include "../dumpstack.h" diff --git a/arch/x86/kernel/dumpstack.c b/arch/x86/kernel/dumpstack.c index 0a0aa1cec8f..8aaa119b7ca 100644 --- a/arch/x86/kernel/dumpstack.c +++ b/arch/x86/kernel/dumpstack.c @@ -141,10 +141,11 @@ static void print_trace_address(void *data, unsigned long addr, int reliable) } static const struct stacktrace_ops print_trace_ops = { - .warning = print_trace_warning, - .warning_symbol = print_trace_warning_symbol, - .stack = print_trace_stack, - .address = print_trace_address, + .warning = print_trace_warning, + .warning_symbol = print_trace_warning_symbol, + .stack = print_trace_stack, + .address = print_trace_address, + .walk_stack = print_context_stack, }; void diff --git a/arch/x86/kernel/dumpstack.h b/arch/x86/kernel/dumpstack.h index 81086c227ab..4fd1420faff 100644 --- a/arch/x86/kernel/dumpstack.h +++ b/arch/x86/kernel/dumpstack.h @@ -14,12 +14,6 @@ #define get_bp(bp) asm("movq %%rbp, %0" : "=r" (bp) :) #endif -extern unsigned long -print_context_stack(struct thread_info *tinfo, - unsigned long *stack, unsigned long bp, - const struct stacktrace_ops *ops, void *data, - unsigned long *end, int *graph); - extern void show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs, unsigned long *stack, unsigned long bp, char *log_lvl); diff --git a/arch/x86/kernel/dumpstack_32.c b/arch/x86/kernel/dumpstack_32.c index e0ed4c7abb6..ae775ca47b2 100644 --- a/arch/x86/kernel/dumpstack_32.c +++ b/arch/x86/kernel/dumpstack_32.c @@ -58,7 +58,7 @@ void dump_trace(struct task_struct *task, struct pt_regs *regs, context = (struct thread_info *) ((unsigned long)stack & (~(THREAD_SIZE - 1))); - bp = print_context_stack(context, stack, bp, ops, data, NULL, &graph); + bp = ops->walk_stack(context, stack, bp, ops, data, NULL, &graph); stack = (unsigned long *)context->previous_esp; if (!stack) diff --git a/arch/x86/kernel/dumpstack_64.c b/arch/x86/kernel/dumpstack_64.c index b13af53883a..0ad9597073f 100644 --- a/arch/x86/kernel/dumpstack_64.c +++ b/arch/x86/kernel/dumpstack_64.c @@ -188,8 +188,8 @@ void dump_trace(struct task_struct *task, struct pt_regs *regs, if (ops->stack(data, id) < 0) break; - bp = print_context_stack(tinfo, stack, bp, ops, - data, estack_end, &graph); + bp = ops->walk_stack(tinfo, stack, bp, ops, + data, estack_end, &graph); ops->stack(data, ""); /* * We link to the next stack via the diff --git a/arch/x86/kernel/stacktrace.c b/arch/x86/kernel/stacktrace.c index c3eb207181f..922eefbb3f6 100644 --- a/arch/x86/kernel/stacktrace.c +++ b/arch/x86/kernel/stacktrace.c @@ -53,17 +53,19 @@ save_stack_address_nosched(void *data, unsigned long addr, int reliable) } static const struct stacktrace_ops save_stack_ops = { - .warning = save_stack_warning, - .warning_symbol = save_stack_warning_symbol, - .stack = save_stack_stack, - .address = save_stack_address, + .warning = save_stack_warning, + .warning_symbol = save_stack_warning_symbol, + .stack = save_stack_stack, + .address = save_stack_address, + .walk_stack = print_context_stack, }; static const struct stacktrace_ops save_stack_ops_nosched = { - .warning = save_stack_warning, - .warning_symbol = save_stack_warning_symbol, - .stack = save_stack_stack, - .address = save_stack_address_nosched, + .warning = save_stack_warning, + .warning_symbol = save_stack_warning_symbol, + .stack = save_stack_stack, + .address = save_stack_address_nosched, + .walk_stack = print_context_stack, }; /* diff --git a/arch/x86/oprofile/backtrace.c b/arch/x86/oprofile/backtrace.c index 044897be021..3855096c59b 100644 --- a/arch/x86/oprofile/backtrace.c +++ b/arch/x86/oprofile/backtrace.c @@ -41,10 +41,11 @@ static void backtrace_address(void *data, unsigned long addr, int reliable) } static struct stacktrace_ops backtrace_ops = { - .warning = backtrace_warning, - .warning_symbol = backtrace_warning_symbol, - .stack = backtrace_stack, - .address = backtrace_address, + .warning = backtrace_warning, + .warning_symbol = backtrace_warning_symbol, + .stack = backtrace_stack, + .address = backtrace_address, + .walk_stack = print_context_stack, }; struct frame_head { -- cgit v1.2.3 From 06d65bda75341485d32f33da474b0664819ad497 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 17 Dec 2009 05:40:34 +0100 Subject: perf events, x86/stacktrace: Fix performance/softlockup by providing a special frame pointer-only stack walker It's just wasteful for stacktrace users like perf to walk through every entries on the stack whereas these only accept reliable ones, ie: that the frame pointer validates. Since perf requires pure reliable stacktraces, it needs a stack walker based on frame pointers-only to optimize the stacktrace processing. This might solve some near-lockup scenarios that can be triggered by call-graph tracing timer events. Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Arnaldo Carvalho de Melo Cc: Paul Mackerras LKML-Reference: <1261024834-5336-2-git-send-regression-fweisbec@gmail.com> [ v2: fix for modular builds and small detail tidyup ] Signed-off-by: Ingo Molnar --- arch/x86/include/asm/stacktrace.h | 6 ++++++ arch/x86/kernel/cpu/perf_event.c | 2 +- arch/x86/kernel/dumpstack.c | 28 ++++++++++++++++++++++++++-- 3 files changed, 33 insertions(+), 3 deletions(-) (limited to 'arch') diff --git a/arch/x86/include/asm/stacktrace.h b/arch/x86/include/asm/stacktrace.h index 6c75151a3cc..35e89122a42 100644 --- a/arch/x86/include/asm/stacktrace.h +++ b/arch/x86/include/asm/stacktrace.h @@ -22,6 +22,12 @@ print_context_stack(struct thread_info *tinfo, const struct stacktrace_ops *ops, void *data, unsigned long *end, int *graph); +extern unsigned long +print_context_stack_bp(struct thread_info *tinfo, + unsigned long *stack, unsigned long bp, + const struct stacktrace_ops *ops, void *data, + unsigned long *end, int *graph); + /* Generic stack tracer with callbacks */ struct stacktrace_ops { diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index d3802ee5a41..c223b7e895d 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -2336,7 +2336,7 @@ static const struct stacktrace_ops backtrace_ops = { .warning_symbol = backtrace_warning_symbol, .stack = backtrace_stack, .address = backtrace_address, - .walk_stack = print_context_stack, + .walk_stack = print_context_stack_bp, }; #include "../dumpstack.h" diff --git a/arch/x86/kernel/dumpstack.c b/arch/x86/kernel/dumpstack.c index 8aaa119b7ca..c56bc287303 100644 --- a/arch/x86/kernel/dumpstack.c +++ b/arch/x86/kernel/dumpstack.c @@ -109,6 +109,30 @@ print_context_stack(struct thread_info *tinfo, } return bp; } +EXPORT_SYMBOL_GPL(print_context_stack); + +unsigned long +print_context_stack_bp(struct thread_info *tinfo, + unsigned long *stack, unsigned long bp, + const struct stacktrace_ops *ops, void *data, + unsigned long *end, int *graph) +{ + struct stack_frame *frame = (struct stack_frame *)bp; + unsigned long *ret_addr = &frame->return_address; + + while (valid_stack_ptr(tinfo, ret_addr, sizeof(*ret_addr), end)) { + unsigned long addr = *ret_addr; + + if (__kernel_text_address(addr)) { + ops->address(data, addr, 1); + frame = frame->next_frame; + ret_addr = &frame->return_address; + print_ftrace_graph_addr(addr, data, ops, tinfo, graph); + } + } + return (unsigned long)frame; +} +EXPORT_SYMBOL_GPL(print_context_stack_bp); static void @@ -143,8 +167,8 @@ static void print_trace_address(void *data, unsigned long addr, int reliable) static const struct stacktrace_ops print_trace_ops = { .warning = print_trace_warning, .warning_symbol = print_trace_warning_symbol, - .stack = print_trace_stack, - .address = print_trace_address, + .stack = print_trace_stack, + .address = print_trace_address, .walk_stack = print_context_stack, }; -- cgit v1.2.3 From 99e8c5a3b875a34d894a711c9a3669858d6adf45 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 17 Dec 2009 01:33:54 +0100 Subject: hw-breakpoints: Fix hardware breakpoints -> perf events dependency The kbuild's select command doesn't propagate through the config dependencies. Hence the current rules of hardware breakpoint's config can't ensure perf can never be disabled under us. We have: config X86 selects HAVE_HW_BREAKPOINTS config HAVE_HW_BREAKPOINTS select PERF_EVENTS config PERF_EVENTS [...] x86 will select the breakpoints but that won't propagate to perf events. The user can still disable the latter, but it is necessary for the breakpoints. What we need is: - x86 selects HAVE_HW_BREAKPOINTS and PERF_EVENTS - HAVE_HW_BREAKPOINTS depends on PERF_EVENTS so that we ensure PERF_EVENTS is enabled and frozen for x86. This fixes the following kind of build errors: In file included from arch/x86/kernel/hw_breakpoint.c:31: include/linux/hw_breakpoint.h: In function 'hw_breakpoint_addr': include/linux/hw_breakpoint.h:39: error: 'struct perf_event' has no member named 'attr' v2: Select also ANON_INODES from x86, required for perf Reported-by: Cyrill Gorcunov Reported-by: Michal Marek Reported-by: Andrew Randrianasulu Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Arnaldo Carvalho de Melo Cc: Paul Mackerras Cc: Randy Dunlap Cc: K.Prasad LKML-Reference: <1261010034-7786-1-git-send-regression-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- arch/Kconfig | 4 +--- arch/x86/Kconfig | 2 ++ 2 files changed, 3 insertions(+), 3 deletions(-) (limited to 'arch') diff --git a/arch/Kconfig b/arch/Kconfig index d82875820a1..9d055b4f058 100644 --- a/arch/Kconfig +++ b/arch/Kconfig @@ -135,9 +135,7 @@ config HAVE_DEFAULT_NO_SPIN_MUTEXES config HAVE_HW_BREAKPOINT bool - depends on HAVE_PERF_EVENTS - select ANON_INODES - select PERF_EVENTS + depends on PERF_EVENTS config HAVE_USER_RETURN_NOTIFIER bool diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 3b2a5aca4ed..55298e89157 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -50,6 +50,8 @@ config X86 select HAVE_KERNEL_BZIP2 select HAVE_KERNEL_LZMA select HAVE_HW_BREAKPOINT + select PERF_EVENTS + select ANON_INODES select HAVE_ARCH_KMEMCHECK select HAVE_USER_RETURN_NOTIFIER -- cgit v1.2.3