From d755330c5e0658d8056242b5b81e2f44ed7a96d8 Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Sun, 15 Jun 2014 10:22:15 +0200 Subject: perf tools: Fix segfault in cumulative.callchain report When cumulative callchain mode is on, we could get samples with with no actual hits. This breaks the assumption of the annotation code, that each sample has annotation counts allocated and leads to segfault. Fixing this by additional checks for annotation stats. Acked-by: Namhyung Kim Acked-by: Arnaldo Carvalho de Melo Cc: Arnaldo Carvalho de Melo Cc: Corey Ashford Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1402821332-12419-1-git-send-email-jolsa@kernel.org Signed-off-by: Jiri Olsa diff --git a/tools/perf/ui/browsers/hists.c b/tools/perf/ui/browsers/hists.c index 52c03fb..04a229a 100644 --- a/tools/perf/ui/browsers/hists.c +++ b/tools/perf/ui/browsers/hists.c @@ -17,6 +17,7 @@ #include "../util.h" #include "../ui.h" #include "map.h" +#include "annotate.h" struct hist_browser { struct ui_browser b; @@ -1593,13 +1594,18 @@ static int perf_evsel__hists_browse(struct perf_evsel *evsel, int nr_events, bi->to.sym->name) > 0) annotate_t = nr_options++; } else { - if (browser->selection != NULL && browser->selection->sym != NULL && - !browser->selection->map->dso->annotate_warned && - asprintf(&options[nr_options], "Annotate %s", - browser->selection->sym->name) > 0) - annotate = nr_options++; + !browser->selection->map->dso->annotate_warned) { + struct annotation *notes; + + notes = symbol__annotation(browser->selection->sym); + + if (notes->src && + asprintf(&options[nr_options], "Annotate %s", + browser->selection->sym->name) > 0) + annotate = nr_options++; + } } if (thread != NULL && @@ -1656,6 +1662,7 @@ retry_popup_menu: if (choice == annotate || choice == annotate_t || choice == annotate_f) { struct hist_entry *he; + struct annotation *notes; int err; do_annotate: if (!objdump_path && perf_session_env__lookup_objdump(env)) @@ -1679,6 +1686,10 @@ do_annotate: he->ms.map = he->branch_info->to.map; } + notes = symbol__annotation(he->ms.sym); + if (!notes->src) + continue; + /* * Don't let this be freed, say, by hists__decay_entry. */ -- cgit v0.10.2 From a93f0e551af9e194db38bfe16001e17a3a1d189a Mon Sep 17 00:00:00 2001 From: Simon Que Date: Mon, 16 Jun 2014 11:32:09 -0700 Subject: perf symbols: Get kernel start address by symbol name The function machine__get_kernel_start_addr() was taking the first symbol of kallsyms as the start address. This is incorrect in certain cases where the first symbol is something at 0, while the actual kernel functions begin at a later point (e.g. 0x80200000). This patch fixes machine__get_kernel_start_addr() to search for the symbol "_text" or "_stext", which marks the beginning of kernel mapping. This was already being done in machine__create_kernel_maps(). Thus, this patch is just a refactor, to move that code into machine__get_kernel_start_addr(). Signed-off-by: Simon Que Link: http://lkml.kernel.org/r/1402943529-13244-1-git-send-email-sque@chromium.org Signed-off-by: Jiri Olsa diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c index 0e5fea9..c73e1fc 100644 --- a/tools/perf/util/machine.c +++ b/tools/perf/util/machine.c @@ -496,18 +496,6 @@ struct process_args { u64 start; }; -static int symbol__in_kernel(void *arg, const char *name, - char type __maybe_unused, u64 start) -{ - struct process_args *args = arg; - - if (strchr(name, '[')) - return 0; - - args->start = start; - return 1; -} - static void machine__get_kallsyms_filename(struct machine *machine, char *buf, size_t bufsz) { @@ -517,27 +505,41 @@ static void machine__get_kallsyms_filename(struct machine *machine, char *buf, scnprintf(buf, bufsz, "%s/proc/kallsyms", machine->root_dir); } -/* Figure out the start address of kernel map from /proc/kallsyms */ -static u64 machine__get_kernel_start_addr(struct machine *machine) +const char *ref_reloc_sym_names[] = {"_text", "_stext", NULL}; + +/* Figure out the start address of kernel map from /proc/kallsyms. + * Returns the name of the start symbol in *symbol_name. Pass in NULL as + * symbol_name if it's not that important. + */ +static u64 machine__get_kernel_start_addr(struct machine *machine, + const char **symbol_name) { char filename[PATH_MAX]; - struct process_args args; + int i; + const char *name; + u64 addr = 0; machine__get_kallsyms_filename(machine, filename, PATH_MAX); if (symbol__restricted_filename(filename, "/proc/kallsyms")) return 0; - if (kallsyms__parse(filename, &args, symbol__in_kernel) <= 0) - return 0; + for (i = 0; (name = ref_reloc_sym_names[i]) != NULL; i++) { + addr = kallsyms__get_function_start(filename, name); + if (addr) + break; + } + + if (symbol_name) + *symbol_name = name; - return args.start; + return addr; } int __machine__create_kernel_maps(struct machine *machine, struct dso *kernel) { enum map_type type; - u64 start = machine__get_kernel_start_addr(machine); + u64 start = machine__get_kernel_start_addr(machine, NULL); for (type = 0; type < MAP__NR_TYPES; ++type) { struct kmap *kmap; @@ -852,23 +854,11 @@ static int machine__create_modules(struct machine *machine) return 0; } -const char *ref_reloc_sym_names[] = {"_text", "_stext", NULL}; - int machine__create_kernel_maps(struct machine *machine) { struct dso *kernel = machine__get_kernel(machine); - char filename[PATH_MAX]; const char *name; - u64 addr = 0; - int i; - - machine__get_kallsyms_filename(machine, filename, PATH_MAX); - - for (i = 0; (name = ref_reloc_sym_names[i]) != NULL; i++) { - addr = kallsyms__get_function_start(filename, name); - if (addr) - break; - } + u64 addr = machine__get_kernel_start_addr(machine, &name); if (!addr) return -1; -- cgit v0.10.2 From b292d7a10487aee6e74b1c18b8d95b92f40d4a4f Mon Sep 17 00:00:00 2001 From: HATAYAMA Daisuke Date: Wed, 25 Jun 2014 10:09:07 +0900 Subject: perf/x86/intel: ignore CondChgd bit to avoid false NMI handling MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Currently, any NMI is falsely handled by a NMI handler of NMI watchdog if CondChgd bit in MSR_CORE_PERF_GLOBAL_STATUS MSR is set. For example, we use external NMI to make system panic to get crash dump, but in this case, the external NMI is falsely handled do to the issue. This commit deals with the issue simply by ignoring CondChgd bit. Here is explanation in detail. On x86 NMI watchdog uses performance monitoring feature to periodically signal NMI each time performance counter gets overflowed. intel_pmu_handle_irq() is called as a NMI_LOCAL handler from a NMI handler of NMI watchdog, perf_event_nmi_handler(). It identifies an owner of a given NMI by looking at overflow status bits in MSR_CORE_PERF_GLOBAL_STATUS MSR. If some of the bits are set, then it handles the given NMI as its own NMI. The problem is that the intel_pmu_handle_irq() doesn't distinguish CondChgd bit from other bits. Unlike the other status bits, CondChgd bit doesn't represent overflow status for performance counters. Thus, CondChgd bit cannot be thought of as a mark indicating a given NMI is NMI watchdog's. As a result, if CondChgd bit is set, any NMI is falsely handled by the NMI handler of NMI watchdog. Also, if type of the falsely handled NMI is either NMI_UNKNOWN, NMI_SERR or NMI_IO_CHECK, the corresponding action is never performed until CondChgd bit is cleared. I noticed this behavior on systems with Ivy Bridge processors: Intel Xeon CPU E5-2630 v2 and Intel Xeon CPU E7-8890 v2. On both systems, CondChgd bit in MSR_CORE_PERF_GLOBAL_STATUS MSR has already been set in the beginning at boot. Then the CondChgd bit is immediately cleared by next wrmsr to MSR_CORE_PERF_GLOBAL_CTRL MSR and appears to remain 0. On the other hand, on older processors such as Nehalem, Xeon E7540, CondChgd bit is not set in the beginning at boot. I'm not sure about exact behavior of CondChgd bit, in particular when this bit is set. Although I read Intel System Programmer's Manual to figure out that, the descriptions I found are: In 18.9.1: "The MSR_PERF_GLOBAL_STATUS MSR also provides a ¡sticky bit¢ to indicate changes to the state of performancmonitoring hardware" In Table 35-2 IA-32 Architectural MSRs 63 CondChg: status bits of this register has changed. These are different from the bahviour I see on the actual system as I explained above. At least, I think ignoring CondChgd bit should be enough for NMI watchdog perspective. Signed-off-by: HATAYAMA Daisuke Acked-by: Don Zickus Signed-off-by: Peter Zijlstra Cc: Cc: Arnaldo Carvalho de Melo Cc: Linus Torvalds Cc: linux-kernel@vger.kernel.org Link: http://lkml.kernel.org/r/20140625.103503.409316067.d.hatayama@jp.fujitsu.com Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c index adb02aa..07846d7 100644 --- a/arch/x86/kernel/cpu/perf_event_intel.c +++ b/arch/x86/kernel/cpu/perf_event_intel.c @@ -1382,6 +1382,15 @@ again: intel_pmu_lbr_read(); /* + * CondChgd bit 63 doesn't mean any overflow status. Ignore + * and clear the bit. + */ + if (__test_and_clear_bit(63, (unsigned long *)&status)) { + if (!status) + goto done; + } + + /* * PEBS overflow sets bit 62 in the global status register */ if (__test_and_clear_bit(62, (unsigned long *)&status)) { -- cgit v0.10.2 From 1f9a7268c67f0290837aada443d28fd953ddca90 Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Tue, 24 Jun 2014 10:20:25 +0200 Subject: perf: Do not allow optimized switch for non-cloned events The context check in perf_event_context_sched_out allows non-cloned context to be part of the optimized schedule out switch. This could move non-cloned context into another workload child. Once this child exits, the context is closed and leaves all original (parent) events in closed state. Any other new cloned event will have closed state and not measure anything. And probably causing other odd bugs. Signed-off-by: Jiri Olsa Signed-off-by: Peter Zijlstra Cc: Cc: Arnaldo Carvalho de Melo Cc: Paul Mackerras Cc: Frederic Weisbecker Cc: Namhyung Kim Cc: Paul Mackerras Cc: Corey Ashford Cc: David Ahern Cc: Jiri Olsa Cc: Linus Torvalds Link: http://lkml.kernel.org/r/1403598026-2310-2-git-send-email-jolsa@kernel.org Signed-off-by: Ingo Molnar diff --git a/kernel/events/core.c b/kernel/events/core.c index a33d9a2b..b0c95f0 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -2320,7 +2320,7 @@ static void perf_event_context_sched_out(struct task_struct *task, int ctxn, next_parent = rcu_dereference(next_ctx->parent_ctx); /* If neither context have a parent context; they cannot be clones. */ - if (!parent && !next_parent) + if (!parent || !next_parent) goto unlock; if (next_parent == ctx || next_ctx == parent || next_parent == parent) { -- cgit v0.10.2