From 0d4dd797564cddc1f71ab0b239e9ea50ddd40b2a Mon Sep 17 00:00:00 2001 From: Baruch Siach Date: Sat, 25 Jan 2014 23:50:23 +0200 Subject: perf/doc: Remove mention of non-existent set_perf_event_pending() from design.txt set_perf_event_pending() was removed in e360adbe ("irq_work: Add generic hardirq context callbacks"). Signed-off-by: Baruch Siach Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Link: http://lkml.kernel.org/r/4c54761865d40210be0628cb84701afc5d57b5d8.1390686193.git.baruch@tkos.co.il Signed-off-by: Ingo Molnar --- tools/perf/design.txt | 1 - 1 file changed, 1 deletion(-) (limited to 'tools/perf') diff --git a/tools/perf/design.txt b/tools/perf/design.txt index 67e5d0c..63a0e6f 100644 --- a/tools/perf/design.txt +++ b/tools/perf/design.txt @@ -454,7 +454,6 @@ So to start with, in order to add HAVE_PERF_EVENTS to your Kconfig, you will need at least this: - asm/perf_event.h - a basic stub will suffice at first - support for atomic64 types (and associated helper functions) - - set_perf_event_pending() implemented If your architecture does have hardware capabilities, you can override the weak stub hw_perf_event_init() to register hardware counters. -- cgit v1.1 From 950b8354716eb1f9c0b39777d379efa5f4125c04 Mon Sep 17 00:00:00 2001 From: Avi Kivity Date: Wed, 22 Jan 2014 21:58:46 +0200 Subject: perf tools: Demangle kernel and kernel module symbols too Some kernels contain C++ code, and thus their symbols need to be demangled. This allows 'perf kvm top' to generate readable output. Signed-off-by: Avi Kivity Cc: Ingo Molnar Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/26f71bf5bf7ee1408e3f1a803556d5df18223ef1.1390420726.git.avi@cloudius-systems.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/symbol-elf.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools/perf') diff --git a/tools/perf/util/symbol-elf.c b/tools/perf/util/symbol-elf.c index 7594567..8f12f0f 100644 --- a/tools/perf/util/symbol-elf.c +++ b/tools/perf/util/symbol-elf.c @@ -922,6 +922,7 @@ int dso__load_sym(struct dso *dso, struct map *map, (u64)shdr.sh_offset); sym.st_value -= shdr.sh_addr - shdr.sh_offset; } +new_symbol: /* * We need to figure out if the object was created from C++ sources * DWARF DW_compile_unit has this, but we don't always have access @@ -933,7 +934,6 @@ int dso__load_sym(struct dso *dso, struct map *map, if (demangled != NULL) elf_name = demangled; } -new_symbol: f = symbol__new(sym.st_value, sym.st_size, GELF_ST_BIND(sym.st_info), elf_name); free(demangled); -- cgit v1.1 From f428ebd184c82a7914b2aa7e9f868918aaf7ea78 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 24 Jan 2014 16:40:02 +0100 Subject: perf tools: Fix AAAAARGH64 memory barriers Someone got the load and store barriers mixed up for AAAAARGH64. Turn them the right side up. Reported-by: Will Deacon Signed-off-by: Peter Zijlstra Fixes: a94d342b9cb0 ("tools/perf: Add required memory barriers") Cc: Ingo Molnar Cc: Will Deacon Link: http://lkml.kernel.org/r/20140124154002.GF31570@twins.programming.kicks-ass.net Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/perf.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'tools/perf') diff --git a/tools/perf/perf.h b/tools/perf/perf.h index 7daa806..e84fa26 100644 --- a/tools/perf/perf.h +++ b/tools/perf/perf.h @@ -100,8 +100,8 @@ #ifdef __aarch64__ #define mb() asm volatile("dmb ish" ::: "memory") -#define wmb() asm volatile("dmb ishld" ::: "memory") -#define rmb() asm volatile("dmb ishst" ::: "memory") +#define wmb() asm volatile("dmb ishst" ::: "memory") +#define rmb() asm volatile("dmb ishld" ::: "memory") #define cpu_relax() asm volatile("yield" ::: "memory") #endif -- cgit v1.1 From 6a02652df511029127406cf8fa89cdf5e987f963 Mon Sep 17 00:00:00 2001 From: Francesco Fusco Date: Mon, 27 Jan 2014 14:39:13 +0100 Subject: perf tools: Fix include for non x86 architectures Commit 71ae8aac ("lib: introduce arch optimized hash library") added an include to for setting up an architecture specific fast hash. Since perf includes directly the non-uapi kernel header, it cannot find on non-x86 and thus prevents perf to be compiled on every architecture other than x86. The problem is the inclusion of in hash.h that results in the following error originating from util/evlist.c: fatal error: asm/hash.h: No such file or directory This commit simply adds an empty stub/file to fix the compile issue on non-x86 architectures. As perf does not use any of these new functions, it fixes the compilation and therefore seems to be the most appropriate solution to go with. Signed-off-by: Francesco Fusco Link: http://lkml.kernel.org/r/2cf8143aad65a6aa6fe30325ef8a65847141afa2.1390829373.git.ffusco@redhat.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/include/asm/hash.h | 6 ++++++ 1 file changed, 6 insertions(+) create mode 100644 tools/perf/util/include/asm/hash.h (limited to 'tools/perf') diff --git a/tools/perf/util/include/asm/hash.h b/tools/perf/util/include/asm/hash.h new file mode 100644 index 0000000..d82b170b --- /dev/null +++ b/tools/perf/util/include/asm/hash.h @@ -0,0 +1,6 @@ +#ifndef __ASM_GENERIC_HASH_H +#define __ASM_GENERIC_HASH_H + +/* Stub */ + +#endif /* __ASM_GENERIC_HASH_H */ -- cgit v1.1 From 9176753d1ed56951a6ee2a0f0a3f367904e35567 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Wed, 29 Jan 2014 16:14:36 +0200 Subject: perf symbols: Fix symbol annotation for relocated kernel Kernel maps map memory addresses to file offsets. For symbol annotation, objdump needs the object VMA addresses. For an unrelocated kernel, that is the same as the memory address. The addresses passed to objdump for symbol annotation did not take into account kernel relocation. This patch fixes that. Reported-by: Linus Torvalds Signed-off-by: Adrian Hunter Tested-by: Jiri Olsa Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Jiri Olsa Cc: Mike Galbraith Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1391004884-10334-2-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/map.c | 5 +++-- tools/perf/util/map.h | 1 + tools/perf/util/symbol-elf.c | 2 ++ 3 files changed, 6 insertions(+), 2 deletions(-) (limited to 'tools/perf') diff --git a/tools/perf/util/map.c b/tools/perf/util/map.c index 3b97513..39cd2d0 100644 --- a/tools/perf/util/map.c +++ b/tools/perf/util/map.c @@ -39,6 +39,7 @@ void map__init(struct map *map, enum map_type type, map->start = start; map->end = end; map->pgoff = pgoff; + map->reloc = 0; map->dso = dso; map->map_ip = map__map_ip; map->unmap_ip = map__unmap_ip; @@ -288,7 +289,7 @@ u64 map__rip_2objdump(struct map *map, u64 rip) if (map->dso->rel) return rip - map->pgoff; - return map->unmap_ip(map, rip); + return map->unmap_ip(map, rip) - map->reloc; } /** @@ -311,7 +312,7 @@ u64 map__objdump_2mem(struct map *map, u64 ip) if (map->dso->rel) return map->unmap_ip(map, ip + map->pgoff); - return ip; + return ip + map->reloc; } void map_groups__init(struct map_groups *mg) diff --git a/tools/perf/util/map.h b/tools/perf/util/map.h index 18068c6..257e513 100644 --- a/tools/perf/util/map.h +++ b/tools/perf/util/map.h @@ -36,6 +36,7 @@ struct map { bool erange_warned; u32 priv; u64 pgoff; + u64 reloc; u32 maj, min; /* only valid for MMAP2 record */ u64 ino; /* only valid for MMAP2 record */ u64 ino_generation;/* only valid for MMAP2 record */ diff --git a/tools/perf/util/symbol-elf.c b/tools/perf/util/symbol-elf.c index 8f12f0f..3e9f336 100644 --- a/tools/perf/util/symbol-elf.c +++ b/tools/perf/util/symbol-elf.c @@ -751,6 +751,8 @@ int dso__load_sym(struct dso *dso, struct map *map, if (strcmp(elf_name, kmap->ref_reloc_sym->name)) continue; kmap->ref_reloc_sym->unrelocated_addr = sym.st_value; + map->reloc = kmap->ref_reloc_sym->addr - + kmap->ref_reloc_sym->unrelocated_addr; break; } } -- cgit v1.1 From 29b596b57426831fce92cd0ebb01c77627616fdf Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Wed, 29 Jan 2014 16:14:37 +0200 Subject: perf tools: Add kallsyms__get_function_start() Separate out the logic used to find the start address of the reference symbol used to track kernel relocation. kallsyms__get_function_start() is used in subsequent patches. Signed-off-by: Adrian Hunter Tested-by: Jiri Olsa Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Jiri Olsa Cc: Mike Galbraith Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1391004884-10334-3-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/event.c | 18 +++++++++++++++--- tools/perf/util/event.h | 3 +++ 2 files changed, 18 insertions(+), 3 deletions(-) (limited to 'tools/perf') diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index 1fc1c2f..17476df 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -470,6 +470,17 @@ static int find_symbol_cb(void *arg, const char *name, char type, return 1; } +u64 kallsyms__get_function_start(const char *kallsyms_filename, + const char *symbol_name) +{ + struct process_symbol_args args = { .name = symbol_name, }; + + if (kallsyms__parse(kallsyms_filename, &args, find_symbol_cb) <= 0) + return 0; + + return args.start; +} + int perf_event__synthesize_kernel_mmap(struct perf_tool *tool, perf_event__handler_t process, struct machine *machine, @@ -480,13 +491,13 @@ int perf_event__synthesize_kernel_mmap(struct perf_tool *tool, char path[PATH_MAX]; char name_buff[PATH_MAX]; struct map *map; + u64 start; int err; /* * We should get this from /sys/kernel/sections/.text, but till that is * available use this, and after it is use this as a fallback for older * kernels. */ - struct process_symbol_args args = { .name = symbol_name, }; union perf_event *event = zalloc((sizeof(event->mmap) + machine->id_hdr_size)); if (event == NULL) { @@ -513,7 +524,8 @@ int perf_event__synthesize_kernel_mmap(struct perf_tool *tool, } } - if (kallsyms__parse(filename, &args, find_symbol_cb) <= 0) { + start = kallsyms__get_function_start(filename, symbol_name); + if (!start) { free(event); return -ENOENT; } @@ -525,7 +537,7 @@ int perf_event__synthesize_kernel_mmap(struct perf_tool *tool, event->mmap.header.type = PERF_RECORD_MMAP; event->mmap.header.size = (sizeof(event->mmap) - (sizeof(event->mmap.filename) - size) + machine->id_hdr_size); - event->mmap.pgoff = args.start; + event->mmap.pgoff = start; event->mmap.start = map->start; event->mmap.len = map->end - event->mmap.start; event->mmap.pid = machine->pid; diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h index faf6e21..66a0c03 100644 --- a/tools/perf/util/event.h +++ b/tools/perf/util/event.h @@ -279,4 +279,7 @@ size_t perf_event__fprintf_mmap2(union perf_event *event, FILE *fp); size_t perf_event__fprintf_task(union perf_event *event, FILE *fp); size_t perf_event__fprintf(union perf_event *event, FILE *fp); +u64 kallsyms__get_function_start(const char *kallsyms_filename, + const char *symbol_name); + #endif /* __PERF_RECORD_H */ -- cgit v1.1 From 15a0a8706c32bd38bff9ebf7c6ef24f32d1ea921 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Wed, 29 Jan 2014 16:14:38 +0200 Subject: perf machine: Add machine__get_kallsyms_filename() Separate out the logic used to make the kallsyms full path name for a machine. It will be reused in a subsequent patch. Signed-off-by: Adrian Hunter Tested-by: Jiri Olsa Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Jiri Olsa Cc: Mike Galbraith Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1391004884-10334-4-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/machine.c | 19 +++++++++++-------- 1 file changed, 11 insertions(+), 8 deletions(-) (limited to 'tools/perf') diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c index ded7459..290c2e6 100644 --- a/tools/perf/util/machine.c +++ b/tools/perf/util/machine.c @@ -496,19 +496,22 @@ static int symbol__in_kernel(void *arg, const char *name, return 1; } +static void machine__get_kallsyms_filename(struct machine *machine, char *buf, + size_t bufsz) +{ + if (machine__is_default_guest(machine)) + scnprintf(buf, bufsz, "%s", symbol_conf.default_guest_kallsyms); + else + 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 *filename; - char path[PATH_MAX]; + char filename[PATH_MAX]; struct process_args args; - if (machine__is_default_guest(machine)) - filename = (char *)symbol_conf.default_guest_kallsyms; - else { - sprintf(path, "%s/proc/kallsyms", machine->root_dir); - filename = path; - } + machine__get_kallsyms_filename(machine, filename, PATH_MAX); if (symbol__restricted_filename(filename, "/proc/kallsyms")) return 0; -- cgit v1.1 From 5512cf24bed2de56f1ef44b6cc9a0a9b15499cea Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Wed, 29 Jan 2014 16:14:39 +0200 Subject: perf machine: Set up ref_reloc_sym in machine__create_kernel_maps() The ref_reloc_sym is always needed for the kernel map in order to check for relocation. Consequently set it up when the kernel map is created. Otherwise it was only being set up by 'perf record'. Signed-off-by: Adrian Hunter Tested-by: Jiri Olsa Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Jiri Olsa Cc: Mike Galbraith Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1391004884-10334-5-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/machine.c | 23 +++++++++++++++++++++++ tools/perf/util/machine.h | 2 ++ 2 files changed, 25 insertions(+) (limited to 'tools/perf') diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c index 290c2e6..c872991 100644 --- a/tools/perf/util/machine.c +++ b/tools/perf/util/machine.c @@ -832,9 +832,25 @@ 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; + } + if (!addr) + return -1; if (kernel == NULL || __machine__create_kernel_maps(machine, kernel) < 0) @@ -853,6 +869,13 @@ int machine__create_kernel_maps(struct machine *machine) * Now that we have all the maps created, just set the ->end of them: */ map_groups__fixup_end(&machine->kmaps); + + if (maps__set_kallsyms_ref_reloc_sym(machine->vmlinux_maps, name, + addr)) { + machine__destroy_kernel_maps(machine); + return -1; + } + return 0; } diff --git a/tools/perf/util/machine.h b/tools/perf/util/machine.h index 4771330..f77e91e 100644 --- a/tools/perf/util/machine.h +++ b/tools/perf/util/machine.h @@ -18,6 +18,8 @@ union perf_event; #define HOST_KERNEL_ID (-1) #define DEFAULT_GUEST_KERNEL_ID (0) +extern const char *ref_reloc_sym_names[]; + struct machine { struct rb_node rb_node; pid_t pid; -- cgit v1.1 From 0ae617bedde062003fd70e566e9a2601e273ea0e Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Wed, 29 Jan 2014 16:14:40 +0200 Subject: perf record: Get ref_reloc_sym from kernel map Now that ref_reloc_sym is set up when the kernel map is created, 'perf record' does not need to pass the symbol names to perf_event__synthesize_kernel_mmap() which can read the values needed from ref_reloc_sym directly. Signed-off-by: Adrian Hunter Tested-by: Jiri Olsa Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Jiri Olsa Cc: Mike Galbraith Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1391004884-10334-6-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-record.c | 10 ++-------- tools/perf/util/event.c | 26 ++++++-------------------- tools/perf/util/event.h | 3 +-- 3 files changed, 9 insertions(+), 30 deletions(-) (limited to 'tools/perf') diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c index 3c394bf..af47531 100644 --- a/tools/perf/builtin-record.c +++ b/tools/perf/builtin-record.c @@ -287,10 +287,7 @@ static void perf_event__synthesize_guest_os(struct machine *machine, void *data) * have no _text sometimes. */ err = perf_event__synthesize_kernel_mmap(tool, process_synthesized_event, - machine, "_text"); - if (err < 0) - err = perf_event__synthesize_kernel_mmap(tool, process_synthesized_event, - machine, "_stext"); + machine); if (err < 0) pr_err("Couldn't record guest kernel [%d]'s reference" " relocation symbol.\n", machine->pid); @@ -457,10 +454,7 @@ static int __cmd_record(struct record *rec, int argc, const char **argv) } err = perf_event__synthesize_kernel_mmap(tool, process_synthesized_event, - machine, "_text"); - if (err < 0) - err = perf_event__synthesize_kernel_mmap(tool, process_synthesized_event, - machine, "_stext"); + machine); if (err < 0) pr_err("Couldn't record kernel reference relocation symbol\n" "Symbol resolution may be skewed if relocation was used (e.g. kexec).\n" diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index 17476df..b0f3ca8 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -483,15 +483,13 @@ u64 kallsyms__get_function_start(const char *kallsyms_filename, int perf_event__synthesize_kernel_mmap(struct perf_tool *tool, perf_event__handler_t process, - struct machine *machine, - const char *symbol_name) + struct machine *machine) { size_t size; - const char *filename, *mmap_name; - char path[PATH_MAX]; + const char *mmap_name; char name_buff[PATH_MAX]; struct map *map; - u64 start; + struct kmap *kmap; int err; /* * We should get this from /sys/kernel/sections/.text, but till that is @@ -513,31 +511,19 @@ int perf_event__synthesize_kernel_mmap(struct perf_tool *tool, * see kernel/perf_event.c __perf_event_mmap */ event->header.misc = PERF_RECORD_MISC_KERNEL; - filename = "/proc/kallsyms"; } else { event->header.misc = PERF_RECORD_MISC_GUEST_KERNEL; - if (machine__is_default_guest(machine)) - filename = (char *) symbol_conf.default_guest_kallsyms; - else { - sprintf(path, "%s/proc/kallsyms", machine->root_dir); - filename = path; - } - } - - start = kallsyms__get_function_start(filename, symbol_name); - if (!start) { - free(event); - return -ENOENT; } map = machine->vmlinux_maps[MAP__FUNCTION]; + kmap = map__kmap(map); size = snprintf(event->mmap.filename, sizeof(event->mmap.filename), - "%s%s", mmap_name, symbol_name) + 1; + "%s%s", mmap_name, kmap->ref_reloc_sym->name) + 1; size = PERF_ALIGN(size, sizeof(u64)); event->mmap.header.type = PERF_RECORD_MMAP; event->mmap.header.size = (sizeof(event->mmap) - (sizeof(event->mmap.filename) - size) + machine->id_hdr_size); - event->mmap.pgoff = start; + event->mmap.pgoff = kmap->ref_reloc_sym->addr; event->mmap.start = map->start; event->mmap.len = map->end - event->mmap.start; event->mmap.pid = machine->pid; diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h index 66a0c03..851fa06 100644 --- a/tools/perf/util/event.h +++ b/tools/perf/util/event.h @@ -214,8 +214,7 @@ int perf_event__synthesize_threads(struct perf_tool *tool, struct machine *machine, bool mmap_data); int perf_event__synthesize_kernel_mmap(struct perf_tool *tool, perf_event__handler_t process, - struct machine *machine, - const char *symbol_name); + struct machine *machine); int perf_event__synthesize_modules(struct perf_tool *tool, perf_event__handler_t process, -- cgit v1.1 From a00d28cb72d3629c6481fe21ba6c6b4f96caed49 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Wed, 29 Jan 2014 16:14:41 +0200 Subject: perf symbols: Prevent the use of kcore if the kernel has moved Use of kcore is predicated upon it matching the recorded data. If the kernel has been relocated at boot time (i.e. since the data was recorded) then do not use kcore. Note that it is possible to make a copy of kcore at the time the data is recorded using 'perf buildid-cache'. Then the perf tools will use the copy because it does match the data. Signed-off-by: Adrian Hunter Tested-by: Jiri Olsa Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Jiri Olsa Cc: Mike Galbraith Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1391004884-10334-7-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/symbol.c | 25 +++++++++++++++++++++---- 1 file changed, 21 insertions(+), 4 deletions(-) (limited to 'tools/perf') diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index 39ce9ad..4ac1f87 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -976,6 +976,23 @@ static int validate_kcore_modules(const char *kallsyms_filename, return 0; } +static int validate_kcore_addresses(const char *kallsyms_filename, + struct map *map) +{ + struct kmap *kmap = map__kmap(map); + + if (kmap->ref_reloc_sym && kmap->ref_reloc_sym->name) { + u64 start; + + start = kallsyms__get_function_start(kallsyms_filename, + kmap->ref_reloc_sym->name); + if (start != kmap->ref_reloc_sym->addr) + return -EINVAL; + } + + return validate_kcore_modules(kallsyms_filename, map); +} + struct kcore_mapfn_data { struct dso *dso; enum map_type type; @@ -1019,8 +1036,8 @@ static int dso__load_kcore(struct dso *dso, struct map *map, kallsyms_filename)) return -EINVAL; - /* All modules must be present at their original addresses */ - if (validate_kcore_modules(kallsyms_filename, map)) + /* Modules and kernel must be present at their original addresses */ + if (validate_kcore_addresses(kallsyms_filename, map)) return -EINVAL; md.dso = dso; @@ -1424,7 +1441,7 @@ static int find_matching_kcore(struct map *map, char *dir, size_t dir_sz) continue; scnprintf(kallsyms_filename, sizeof(kallsyms_filename), "%s/%s/kallsyms", dir, dent->d_name); - if (!validate_kcore_modules(kallsyms_filename, map)) { + if (!validate_kcore_addresses(kallsyms_filename, map)) { strlcpy(dir, kallsyms_filename, dir_sz); ret = 0; break; @@ -1479,7 +1496,7 @@ static char *dso__find_kallsyms(struct dso *dso, struct map *map) if (fd != -1) { close(fd); /* If module maps match go with /proc/kallsyms */ - if (!validate_kcore_modules("/proc/kallsyms", map)) + if (!validate_kcore_addresses("/proc/kallsyms", map)) goto proc_kallsyms; } -- cgit v1.1 From c080f72753def150993144d755379941f8b14683 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Wed, 29 Jan 2014 16:14:42 +0200 Subject: perf tests: No need to set up ref_reloc_sym Now that ref_reloc_sym is set up by machine__create_kernel_maps(), the "vmlinux symtab matches kallsyms" test does have to do it. Signed-off-by: Adrian Hunter Tested-by: Jiri Olsa Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Jiri Olsa Cc: Mike Galbraith Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1391004884-10334-8-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/tests/vmlinux-kallsyms.c | 10 ---------- 1 file changed, 10 deletions(-) (limited to 'tools/perf') diff --git a/tools/perf/tests/vmlinux-kallsyms.c b/tools/perf/tests/vmlinux-kallsyms.c index 2bd13ed..3d90880 100644 --- a/tools/perf/tests/vmlinux-kallsyms.c +++ b/tools/perf/tests/vmlinux-kallsyms.c @@ -26,7 +26,6 @@ int test__vmlinux_matches_kallsyms(void) struct map *kallsyms_map, *vmlinux_map; struct machine kallsyms, vmlinux; enum map_type type = MAP__FUNCTION; - struct ref_reloc_sym ref_reloc_sym = { .name = "_stext", }; u64 mem_start, mem_end; /* @@ -70,14 +69,6 @@ int test__vmlinux_matches_kallsyms(void) */ kallsyms_map = machine__kernel_map(&kallsyms, type); - sym = map__find_symbol_by_name(kallsyms_map, ref_reloc_sym.name, NULL); - if (sym == NULL) { - pr_debug("dso__find_symbol_by_name "); - goto out; - } - - ref_reloc_sym.addr = UM(sym->start); - /* * Step 5: * @@ -89,7 +80,6 @@ int test__vmlinux_matches_kallsyms(void) } vmlinux_map = machine__kernel_map(&vmlinux, type); - map__kmap(vmlinux_map)->ref_reloc_sym = &ref_reloc_sym; /* * Step 6: -- cgit v1.1 From d9b62aba87a82939c73f451a166c7a21342350d6 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Wed, 29 Jan 2014 16:14:43 +0200 Subject: perf tools: Adjust kallsyms for relocated kernel If the kernel is relocated at boot time, kallsyms will not match data recorded previously. That does not matter for modules because they are corrected anyway. It also does not matter if vmlinux is being used for symbols. But if perf tools has only kallsyms then the symbols will not match. Fix by applying the delta gained by comparing the old and current addresses of the relocation reference symbol. Signed-off-by: Adrian Hunter Tested-by: Jiri Olsa Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Jiri Olsa Cc: Mike Galbraith Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1391004884-10334-9-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/symbol.c | 40 ++++++++++++++++++++++++++++++++++++++-- 1 file changed, 38 insertions(+), 2 deletions(-) (limited to 'tools/perf') diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index 4ac1f87..a9d758a 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -627,7 +627,7 @@ static int dso__split_kallsyms_for_kcore(struct dso *dso, struct map *map, * kernel range is broken in several maps, named [kernel].N, as we don't have * the original ELF section names vmlinux have. */ -static int dso__split_kallsyms(struct dso *dso, struct map *map, +static int dso__split_kallsyms(struct dso *dso, struct map *map, u64 delta, symbol_filter_t filter) { struct map_groups *kmaps = map__kmap(map)->kmaps; @@ -692,6 +692,12 @@ static int dso__split_kallsyms(struct dso *dso, struct map *map, char dso_name[PATH_MAX]; struct dso *ndso; + if (delta) { + /* Kernel was relocated at boot time */ + pos->start -= delta; + pos->end -= delta; + } + if (count == 0) { curr_map = map; goto filter_symbol; @@ -721,6 +727,10 @@ static int dso__split_kallsyms(struct dso *dso, struct map *map, curr_map->map_ip = curr_map->unmap_ip = identity__map_ip; map_groups__insert(kmaps, curr_map); ++kernel_range; + } else if (delta) { + /* Kernel was relocated at boot time */ + pos->start -= delta; + pos->end -= delta; } filter_symbol: if (filter && filter(curr_map, pos)) { @@ -1130,15 +1140,41 @@ out_err: return -EINVAL; } +/* + * If the kernel is relocated at boot time, kallsyms won't match. Compute the + * delta based on the relocation reference symbol. + */ +static int kallsyms__delta(struct map *map, const char *filename, u64 *delta) +{ + struct kmap *kmap = map__kmap(map); + u64 addr; + + if (!kmap->ref_reloc_sym || !kmap->ref_reloc_sym->name) + return 0; + + addr = kallsyms__get_function_start(filename, + kmap->ref_reloc_sym->name); + if (!addr) + return -1; + + *delta = addr - kmap->ref_reloc_sym->addr; + return 0; +} + int dso__load_kallsyms(struct dso *dso, const char *filename, struct map *map, symbol_filter_t filter) { + u64 delta = 0; + if (symbol__restricted_filename(filename, "/proc/kallsyms")) return -1; if (dso__load_all_kallsyms(dso, filename, map) < 0) return -1; + if (kallsyms__delta(map, filename, &delta)) + return -1; + symbols__fixup_duplicate(&dso->symbols[map->type]); symbols__fixup_end(&dso->symbols[map->type]); @@ -1150,7 +1186,7 @@ int dso__load_kallsyms(struct dso *dso, const char *filename, if (!dso__load_kcore(dso, map, filename)) return dso__split_kallsyms_for_kcore(dso, map, filter); else - return dso__split_kallsyms(dso, map, filter); + return dso__split_kallsyms(dso, map, delta, filter); } static int dso__load_perf_map(struct dso *dso, struct map *map, -- cgit v1.1 From d3b70220292c40d3b499797fd2f33f608fc35edb Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Wed, 29 Jan 2014 16:14:44 +0200 Subject: perf buildid-cache: Check relocation when checking for existing kcore perf buildid-cache does not make another copy of kcore if the buildid and modules match an existing copy. That does not take into account the possibility that the kernel has been relocated. Extend the check to check if the reference relocation symbol matches too, otherwise do make a copy. Signed-off-by: Adrian Hunter Tested-by: Jiri Olsa Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Jiri Olsa Cc: Mike Galbraith Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1391004884-10334-10-git-send-email-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-buildid-cache.c | 33 +++++++++++++++++++++++++++++---- 1 file changed, 29 insertions(+), 4 deletions(-) (limited to 'tools/perf') diff --git a/tools/perf/builtin-buildid-cache.c b/tools/perf/builtin-buildid-cache.c index cfede86..b22dbb1 100644 --- a/tools/perf/builtin-buildid-cache.c +++ b/tools/perf/builtin-buildid-cache.c @@ -63,11 +63,35 @@ static int build_id_cache__kcore_dir(char *dir, size_t sz) return 0; } +static bool same_kallsyms_reloc(const char *from_dir, char *to_dir) +{ + char from[PATH_MAX]; + char to[PATH_MAX]; + const char *name; + u64 addr1 = 0, addr2 = 0; + int i; + + scnprintf(from, sizeof(from), "%s/kallsyms", from_dir); + scnprintf(to, sizeof(to), "%s/kallsyms", to_dir); + + for (i = 0; (name = ref_reloc_sym_names[i]) != NULL; i++) { + addr1 = kallsyms__get_function_start(from, name); + if (addr1) + break; + } + + if (name) + addr2 = kallsyms__get_function_start(to, name); + + return addr1 == addr2; +} + static int build_id_cache__kcore_existing(const char *from_dir, char *to_dir, size_t to_dir_sz) { char from[PATH_MAX]; char to[PATH_MAX]; + char to_subdir[PATH_MAX]; struct dirent *dent; int ret = -1; DIR *d; @@ -86,10 +110,11 @@ static int build_id_cache__kcore_existing(const char *from_dir, char *to_dir, continue; scnprintf(to, sizeof(to), "%s/%s/modules", to_dir, dent->d_name); - if (!compare_proc_modules(from, to)) { - scnprintf(to, sizeof(to), "%s/%s", to_dir, - dent->d_name); - strlcpy(to_dir, to, to_dir_sz); + scnprintf(to_subdir, sizeof(to_subdir), "%s/%s", + to_dir, dent->d_name); + if (!compare_proc_modules(from, to) && + same_kallsyms_reloc(from_dir, to_subdir)) { + strlcpy(to_dir, to_subdir, to_dir_sz); ret = 0; break; } -- cgit v1.1 From 981a23792cd02631f8cd5dd65753208a44de5ae1 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Wed, 5 Feb 2014 05:18:58 +0000 Subject: perf probe: Do not add offset twice to uprobe address Fix perf-probe not to add offset value twice to uprobe probe address when post processing. The tevs[i].point.address struct member is the address of symbol+offset, but current perf-probe adjusts the point.address by adding the offset. As a result, the probe address becomes symbol+offset+offset. This may cause unexpected code corruption. Urgent fix is needed. Without this fix: --- # ./perf probe -x ./perf dso__load_vmlinux+4 # ./perf probe -l probe_perf:dso__load_vmlinux (on 0x000000000006d2b8) # nm ./perf.orig | grep dso__load_vmlinux\$ 000000000046d0a0 T dso__load_vmlinux --- You can see the given offset is 3 but the actual probed address is dso__load_vmlinux+8. With this fix: --- # ./perf probe -x ./perf dso__load_vmlinux+4 # ./perf probe -l probe_perf:dso__load_vmlinux (on 0x000000000006d2b4) --- Now the problem is fixed. Note: This bug is introduced by commit fb7345bbf7fad9bf72ef63a19c707970b9685812 Signed-off-by: Masami Hiramatsu Cc: "David A. Long" Cc: David Ahern Cc: Ingo Molnar Cc: Namhyung Kim Cc: Oleg Nesterov Cc: Srikar Dronamraju Cc: Steven Rostedt Cc: yrl.pp-manager.tt@hitachi.com Link: http://lkml.kernel.org/r/20140205051858.6519.27314.stgit@kbuild-fedora.yrl.intra.hitachi.co.jp Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/probe-event.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools/perf') diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c index a8a9b6c..d8b048c 100644 --- a/tools/perf/util/probe-event.c +++ b/tools/perf/util/probe-event.c @@ -336,8 +336,8 @@ static int add_exec_to_probe_trace_events(struct probe_trace_event *tevs, return ret; for (i = 0; i < ntevs && ret >= 0; i++) { + /* point.address is the addres of point.symbol + point.offset */ offset = tevs[i].point.address - stext; - offset += tevs[i].point.offset; tevs[i].point.offset = 0; zfree(&tevs[i].point.symbol); ret = e_snprintf(buf, 32, "0x%lx", offset); -- cgit v1.1 From f67697bd079f4bbcbe7d6d26765a06b18afe0630 Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Tue, 4 Feb 2014 15:37:48 +0100 Subject: perf tools: Handle PERF_RECORD_HEADER_EVENT_TYPE properly We removed event types from data file in following commits: 6065210 perf tools: Remove event types framework completely 44b3c57 perf tools: Remove event types from perf data file We no longer need this information, because we can get it directly from tracepoints. But we still need to handle PERF_RECORD_HEADER_EVENT_TYPE event for the sake of old perf data files created in pipe mode like: $ perf.3.4 record -o - foo >perf.data $ perf.312 report -i - < perf.data Reported-by: Stephane Eranian Signed-off-by: Jiri Olsa 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/1391524668-12546-1-git-send-email-jolsa@redhat.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/session.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'tools/perf') diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 0b39a48..5da6ce7 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -1008,6 +1008,12 @@ static int perf_session__process_user_event(struct perf_session *session, union if (err == 0) perf_session__set_id_hdr_size(session); return err; + case PERF_RECORD_HEADER_EVENT_TYPE: + /* + * Depreceated, but we need to handle it for sake + * of old data files create in pipe mode. + */ + return 0; case PERF_RECORD_HEADER_TRACING_DATA: /* setup for reading amidst mmap */ lseek(fd, file_offset, SEEK_SET); -- cgit v1.1 From 88fee52e58ca14d8465b614774ed0bf08e1a7790 Mon Sep 17 00:00:00 2001 From: Vince Weaver Date: Mon, 30 Dec 2013 15:39:45 -0500 Subject: perf list: Fix checking for supported events on older kernels "perf list" listing of hardware events doesn't work on older ARM devices. The change enabling event detection: commit b41f1cec91c37eeea6fdb15effbfa24ea0a5536b Author: Namhyung Kim Date: Tue Aug 27 11:41:53 2013 +0900 perf list: Skip unsupported events uses the following code in tools/perf/util/parse-events.c: struct perf_event_attr attr = { .type = type, .config = config, .disabled = 1, .exclude_kernel = 1, }; On ARM machines pre-dating the Cortex-A15 this doesn't work, as these machines don't support .exclude_kernel. So starting with 3.12 "perf list" does not report any hardware events at all on older machines (seen on Rasp-Pi, Pandaboard, Beagleboard, etc). This version of the patch makes changes suggested by Namhyung Kim to check for EACCESS and retry (instead of just dropping the exclude_kernel) so we can properly handle machines where /proc/sys/kernel/perf_event_paranoid is set to 2. Reported-by: Chad Paradis Signed-off-by: Vince Weaver Acked-by: Namhyung Kim Cc: Chad Paradis Cc: Ingo Molnar Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/alpine.DEB.2.10.1312301536150.28814@vincent-weaver-1.um.maine.edu Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/parse-events.c | 17 +++++++++++++++-- 1 file changed, 15 insertions(+), 2 deletions(-) (limited to 'tools/perf') diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index d248fca..1e15df1 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -1091,12 +1091,12 @@ int is_valid_tracepoint(const char *event_string) static bool is_event_supported(u8 type, unsigned config) { bool ret = true; + int open_return; struct perf_evsel *evsel; struct perf_event_attr attr = { .type = type, .config = config, .disabled = 1, - .exclude_kernel = 1, }; struct { struct thread_map map; @@ -1108,7 +1108,20 @@ static bool is_event_supported(u8 type, unsigned config) evsel = perf_evsel__new(&attr); if (evsel) { - ret = perf_evsel__open(evsel, NULL, &tmap.map) >= 0; + open_return = perf_evsel__open(evsel, NULL, &tmap.map); + ret = open_return >= 0; + + if (open_return == -EACCES) { + /* + * This happens if the paranoid value + * /proc/sys/kernel/perf_event_paranoid is set to 2 + * Re-run with exclude_kernel set; we don't do that + * by default as some ARM machines do not support it. + * + */ + evsel->attr.exclude_kernel = 1; + ret = perf_evsel__open(evsel, NULL, &tmap.map) >= 0; + } perf_evsel__delete(evsel); } -- cgit v1.1 From 79d26a6a19ace19faabf8d8d27d3430be2e26d34 Mon Sep 17 00:00:00 2001 From: Ben Hutchings Date: Thu, 6 Feb 2014 01:00:35 +0000 Subject: perf trace: Add fallback definition of EFD_SEMAPHORE glibc 2.17 is missing this on sparc, despite the fact that it's not architecture-specific. Signed-off-by: Ben Hutchings Fixes: 49af9e93adfa ('perf trace: Beautify eventfd2 'flags' arg') Cc: Cc: Ingo Molnar Cc: Paul Mackerras Cc: Peter Zijlstra Cc: stable@vger.kernel.org Link: http://lkml.kernel.org/r/1391648435.3003.100.camel@deadeye.wl.decadent.org.uk Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-trace.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'tools/perf') diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index 896f270..619d11c 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -37,6 +37,10 @@ # define MADV_UNMERGEABLE 13 #endif +#ifndef EFD_SEMAPHORE +# define EFD_SEMAPHORE 1 +#endif + struct tp_field { int offset; union { -- cgit v1.1 From 844ae5b46c08dbc7ba695b543c023f9cf3bbf9ff Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 10 Feb 2014 14:09:48 -0300 Subject: perf trace: Fix ioctl 'request' beautifier build problems on !(i386 || x86_64) arches Supporting decoding the ioctl 'request' parameter needs more work to properly support more architectures, the current approach doesn't work on at least powerpc and sparc, as reported by Ben Hutchings in http://lkml.kernel.org/r/1391593985.3003.48.camel@deadeye.wl.decadent.org.uk . Work around that by making it to be ifdefed for the architectures known to work with the current, limited approach, i386 and x86_64 till better code is written. Reported-by: Ben Hutchings Acked-by: Ben Hutchings Cc: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Mike Galbraith Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Stephane Eranian Cc: # 3.13 Fixes: 78645cf3ed32 ("perf trace: Initial beautifier for ioctl's 'cmd' arg") Link: http://lkml.kernel.org/n/tip-ss04k11insqlu329xh5g02q0@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-trace.c | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+) (limited to 'tools/perf') diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index 619d11c..6aa6fb6 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -283,6 +283,11 @@ static size_t syscall_arg__scnprintf_strarray(char *bf, size_t size, #define SCA_STRARRAY syscall_arg__scnprintf_strarray +#if defined(__i386__) || defined(__x86_64__) +/* + * FIXME: Make this available to all arches as soon as the ioctl beautifier + * gets rewritten to support all arches. + */ static size_t syscall_arg__scnprintf_strhexarray(char *bf, size_t size, struct syscall_arg *arg) { @@ -290,6 +295,7 @@ static size_t syscall_arg__scnprintf_strhexarray(char *bf, size_t size, } #define SCA_STRHEXARRAY syscall_arg__scnprintf_strhexarray +#endif /* defined(__i386__) || defined(__x86_64__) */ static size_t syscall_arg__scnprintf_fd(char *bf, size_t size, struct syscall_arg *arg); @@ -843,6 +849,10 @@ static size_t syscall_arg__scnprintf_signum(char *bf, size_t size, struct syscal #define SCA_SIGNUM syscall_arg__scnprintf_signum +#if defined(__i386__) || defined(__x86_64__) +/* + * FIXME: Make this available to all arches. + */ #define TCGETS 0x5401 static const char *tioctls[] = { @@ -864,6 +874,7 @@ static const char *tioctls[] = { }; static DEFINE_STRARRAY_OFFSET(tioctls, 0x5401); +#endif /* defined(__i386__) || defined(__x86_64__) */ #define STRARRAY(arg, name, array) \ .arg_scnprintf = { [arg] = SCA_STRARRAY, }, \ @@ -945,9 +956,16 @@ static struct syscall_fmt { { .name = "getrlimit", .errmsg = true, STRARRAY(0, resource, rlimit_resources), }, { .name = "ioctl", .errmsg = true, .arg_scnprintf = { [0] = SCA_FD, /* fd */ +#if defined(__i386__) || defined(__x86_64__) +/* + * FIXME: Make this available to all arches. + */ [1] = SCA_STRHEXARRAY, /* cmd */ [2] = SCA_HEX, /* arg */ }, .arg_parm = { [1] = &strarray__tioctls, /* cmd */ }, }, +#else + [2] = SCA_HEX, /* arg */ }, }, +#endif { .name = "kill", .errmsg = true, .arg_scnprintf = { [1] = SCA_SIGNUM, /* sig */ }, }, { .name = "linkat", .errmsg = true, -- cgit v1.1 From 48c65bda95d692076de7e5eae3188ddae8635dca Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Thu, 20 Feb 2014 10:32:53 +0900 Subject: perf annotate: Check availability of annotate when processing samples The TUI of perf report and top support annotation, but stdio and GTK don't. So it should be checked before calling hist_entry__inc_addr_ samples() to avoid wasting resources that will never be used. perf annotate need it regardless of UI and sort keys, so the check of whether to allocate resources should be on the tools that have annotate as an option in the TUI, 'report' and 'top', not on the function called by all of them. It caused perf annotate on ppc64 to produce zero output, since the buckets were not being allocated. Reported-by: Anton Blanchard Signed-off-by: Namhyung Kim Cc: Anton Blanchard Cc: Ingo Molnar Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1392859976-32760-1-git-send-email-namhyung@kernel.org [ Renamed (report,top)__needs_annotate() to ui__has_annotation() ] Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-report.c | 40 ++++++++++++++++++++++++---------------- tools/perf/builtin-top.c | 6 ++++-- tools/perf/util/annotate.c | 9 ++++++++- tools/perf/util/annotate.h | 2 ++ 4 files changed, 38 insertions(+), 19 deletions(-) (limited to 'tools/perf') diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 3c53ec2..02f985f 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -113,14 +113,16 @@ static int report__add_mem_hist_entry(struct perf_tool *tool, struct addr_locati if (!he) return -ENOMEM; - err = hist_entry__inc_addr_samples(he, evsel->idx, al->addr); - if (err) - goto out; + if (ui__has_annotation()) { + err = hist_entry__inc_addr_samples(he, evsel->idx, al->addr); + if (err) + goto out; - mx = he->mem_info; - err = addr_map_symbol__inc_samples(&mx->daddr, evsel->idx); - if (err) - goto out; + mx = he->mem_info; + err = addr_map_symbol__inc_samples(&mx->daddr, evsel->idx); + if (err) + goto out; + } evsel->hists.stats.total_period += cost; hists__inc_nr_events(&evsel->hists, PERF_RECORD_SAMPLE); @@ -164,14 +166,18 @@ static int report__add_branch_hist_entry(struct perf_tool *tool, struct addr_loc he = __hists__add_entry(&evsel->hists, al, parent, &bi[i], NULL, 1, 1, 0); if (he) { - bx = he->branch_info; - err = addr_map_symbol__inc_samples(&bx->from, evsel->idx); - if (err) - goto out; - - err = addr_map_symbol__inc_samples(&bx->to, evsel->idx); - if (err) - goto out; + if (ui__has_annotation()) { + bx = he->branch_info; + err = addr_map_symbol__inc_samples(&bx->from, + evsel->idx); + if (err) + goto out; + + err = addr_map_symbol__inc_samples(&bx->to, + evsel->idx); + if (err) + goto out; + } evsel->hists.stats.total_period += 1; hists__inc_nr_events(&evsel->hists, PERF_RECORD_SAMPLE); @@ -205,7 +211,9 @@ static int report__add_hist_entry(struct perf_tool *tool, struct perf_evsel *evs if (err) goto out; - err = hist_entry__inc_addr_samples(he, evsel->idx, al->addr); + if (ui__has_annotation()) + err = hist_entry__inc_addr_samples(he, evsel->idx, al->addr); + evsel->hists.stats.total_period += sample->period; hists__inc_nr_events(&evsel->hists, PERF_RECORD_SAMPLE); out: diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index 76cd510..5f989a7 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -176,7 +176,7 @@ static void perf_top__record_precise_ip(struct perf_top *top, { struct annotation *notes; struct symbol *sym; - int err; + int err = 0; if (he == NULL || he->ms.sym == NULL || ((top->sym_filter_entry == NULL || @@ -190,7 +190,9 @@ static void perf_top__record_precise_ip(struct perf_top *top, return; ip = he->ms.map->map_ip(he->ms.map, ip); - err = hist_entry__inc_addr_samples(he, counter, ip); + + if (ui__has_annotation()) + err = hist_entry__inc_addr_samples(he, counter, ip); pthread_mutex_unlock(¬es->lock); diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c index 469eb67..3aa555f 100644 --- a/tools/perf/util/annotate.c +++ b/tools/perf/util/annotate.c @@ -8,6 +8,8 @@ */ #include "util.h" +#include "ui/ui.h" +#include "sort.h" #include "build-id.h" #include "color.h" #include "cache.h" @@ -489,7 +491,7 @@ static int symbol__inc_addr_samples(struct symbol *sym, struct map *map, { struct annotation *notes; - if (sym == NULL || use_browser != 1 || !sort__has_sym) + if (sym == NULL) return 0; notes = symbol__annotation(sym); @@ -1399,3 +1401,8 @@ int hist_entry__annotate(struct hist_entry *he, size_t privsize) { return symbol__annotate(he->ms.sym, he->ms.map, privsize); } + +bool ui__has_annotation(void) +{ + return use_browser == 1 && sort__has_sym; +} diff --git a/tools/perf/util/annotate.h b/tools/perf/util/annotate.h index b2aef59..56ad4f5 100644 --- a/tools/perf/util/annotate.h +++ b/tools/perf/util/annotate.h @@ -151,6 +151,8 @@ void symbol__annotate_zero_histogram(struct symbol *sym, int evidx); void symbol__annotate_decay_histogram(struct symbol *sym, int evidx); void disasm__purge(struct list_head *head); +bool ui__has_annotation(void); + int symbol__tty_annotate(struct symbol *sym, struct map *map, struct perf_evsel *evsel, bool print_lines, bool full_paths, int min_pcnt, int max_lines); -- cgit v1.1 From 98e9f03bbf2cb21a60f94b8b700eb5d38470819d Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Thu, 20 Feb 2014 10:32:54 +0900 Subject: perf symbols: Destroy unused symsrcs Stephane reported that perf report and annotate failed to process data using lots of (> 500) shared libraries. It was because of the limit on number of open files (ulimit -n). Currently when perf loads a DSO, it'll look for normal and dynamic symbol tables. And if it fails to find out both tables, it'll iterate all of possible symtab types. But many of them are useless since they have no additional information and the problem is that it's not closing those files even though they're not used. Fix it. Reported-by: Stephane Eranian Signed-off-by: Namhyung Kim Cc: Cody P Schafer Cc: Ingo Molnar Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1392859976-32760-2-git-send-email-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/symbol.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'tools/perf') diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index a9d758a..e89afc0 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -1336,6 +1336,8 @@ int dso__load(struct dso *dso, struct map *map, symbol_filter_t filter) if (syms_ss && runtime_ss) break; + } else { + symsrc__destroy(ss); } } -- cgit v1.1 From 280e7c48c3b873e4987a63da276ecab25383f494 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Sat, 11 Jan 2014 11:42:51 -0800 Subject: perf tools: fix BFD detection on opensuse opensuse libbfd requires -lz -liberty to build. Add those to the BFD feature detection. Signed-off-by: Andi Kleen Acked-by: David Ahern Acked-by: Jiri Olsa Cc: Adrian Hunter Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Jiri Olsa Cc: Namhyung Kim Link: http://lkml.kernel.org/r/1389469379-13340-2-git-send-email-andi@firstfloor.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/config/Makefile | 2 +- tools/perf/config/feature-checks/Makefile | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'tools/perf') diff --git a/tools/perf/config/Makefile b/tools/perf/config/Makefile index c48d449..0331ea2 100644 --- a/tools/perf/config/Makefile +++ b/tools/perf/config/Makefile @@ -478,7 +478,7 @@ else endif ifeq ($(feature-libbfd), 1) - EXTLIBS += -lbfd + EXTLIBS += -lbfd -lz -liberty endif ifdef NO_DEMANGLE diff --git a/tools/perf/config/feature-checks/Makefile b/tools/perf/config/feature-checks/Makefile index 12e5513..523b7bc 100644 --- a/tools/perf/config/feature-checks/Makefile +++ b/tools/perf/config/feature-checks/Makefile @@ -121,7 +121,7 @@ test-libpython-version.bin: $(BUILD) $(FLAGS_PYTHON_EMBED) test-libbfd.bin: - $(BUILD) -DPACKAGE='"perf"' -lbfd -ldl + $(BUILD) -DPACKAGE='"perf"' -lbfd -lz -liberty -ldl test-liberty.bin: $(CC) -o $(OUTPUT)$@ test-libbfd.c -DPACKAGE='"perf"' -lbfd -ldl -liberty -- cgit v1.1 From b39c2a57a00a841f057a75b41df4c26173288b66 Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Wed, 26 Feb 2014 18:14:26 +0100 Subject: perf tools: Fix strict alias issue for find_first_bit MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When compiling perf tool code with gcc 4.4.7 I'm getting following error: CC util/session.o cc1: warnings being treated as errors util/session.c: In function ‘perf_session_deliver_event’: tools/perf/util/include/linux/bitops.h:109: error: dereferencing pointer ‘p’ does break strict-aliasing rules tools/perf/util/include/linux/bitops.h:101: error: dereferencing pointer ‘p’ does break strict-aliasing rules util/session.c:697: note: initialized from here tools/perf/util/include/linux/bitops.h:101: note: initialized from here make[1]: *** [util/session.o] Error 1 make: *** [util/session.o] Error 2 The aliased types here are u64 and unsigned long pointers, which is safe for the find_first_bit processing. This error shows up for me only for gcc 4.4 on 32bit x86, even for -Wstrict-aliasing=3, while newer gcc are quiet and scream here for -Wstrict-aliasing={2,1}. Looks like newer gcc changed the rules for strict alias warnings. The gcc documentation offers workaround for valid aliasing by using __may_alias__ attribute: http://gcc.gnu.org/onlinedocs/gcc-4.4.0/gcc/Type-Attributes.html Using this workaround for the find_first_bit function. Signed-off-by: Jiri Olsa 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/1393434867-20271-1-git-send-email-jolsa@redhat.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/include/linux/bitops.h | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'tools/perf') diff --git a/tools/perf/util/include/linux/bitops.h b/tools/perf/util/include/linux/bitops.h index 45cf10a..dadfa7e 100644 --- a/tools/perf/util/include/linux/bitops.h +++ b/tools/perf/util/include/linux/bitops.h @@ -87,13 +87,15 @@ static __always_inline unsigned long __ffs(unsigned long word) return num; } +typedef const unsigned long __attribute__((__may_alias__)) long_alias_t; + /* * Find the first set bit in a memory region. */ static inline unsigned long find_first_bit(const unsigned long *addr, unsigned long size) { - const unsigned long *p = addr; + long_alias_t *p = (long_alias_t *) addr; unsigned long result = 0; unsigned long tmp; -- cgit v1.1