From ba21594cddee0a3af582971656702b1c4509d8f5 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 14 Jan 2010 12:23:10 -0200 Subject: [PATCH] perf tools: Cross platform perf.data analysis support MIME-Version: 1.0 Content-Type: text/plain; charset=utf8 Content-Transfer-Encoding: 8bit There are still some problems related to loading vmlinux files, but those are unrelated to the feature implemented in this patch, so will get fixed in the next patches, but here are some results: 1. collect perf.data file on a Fedora 12 machine, x86_64, 64-bit userland 2. transfer it to a Debian Testing machine, PARISC64, 32-bit userland acme@parisc:~/git/linux-2.6-tip$ perf buildid-list | head -5 74f9930ee94475b6b3238caf3725a50d59cb994b [kernel.kallsyms] 55fdd56670453ea66c011158c4b9d30179c1d049 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/ipt_MASQUERADE.ko 41adff63c730890480980d5d8ba513f1c216a858 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/iptable_nat.ko 90a33def1077bb8e97b8a78546dc96c2de62df46 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/nf_nat.ko 984c7bea90ce1376d5c8e7ef43a781801286e62d /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/tun.ko acme@parisc:~/git/linux-2.6-tip$ perf buildid-list | tail -5 22492f3753c6a67de5c7ccbd6b863390c92c0723 /usr/lib64/libXt.so.6.0.0 353802bb7e1b895ba43507cc678f951e778e4c6f /usr/lib64/libMagickCore.so.2.0.0 d10c2897558595efe7be8b0584cf7e6398bc776c /usr/lib64/libfprint.so.0.0.0 a83ecfb519a788774a84d5ddde633c9ba56c03ab /home/acme/bin/perf d3ca765a8ecf257d263801d7ad8c49c189082317 /usr/lib64/libdwarf.so.0.0 acme@parisc:~/git/linux-2.6-tip$ acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm The file [kernel.kallsyms] cannot be used, trying to use /proc/kallsyms... ^^^^ The problem related to vmlinux handling, it shouldn't be trying this ^^^^ rather alien /proc/kallsyms at all... /lib64/libpthread-2.10.2.so with build id 5c68f7afeb33309c78037e374b0deee84dd441f6 not found, continuing without symbols /lib64/libc-2.10.2.so with build id eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 not found, continuing without symbols /home/acme/bin/perf with build id a83ecfb519a788774a84d5ddde633c9ba56c03ab not found, continuing without symbols /usr/sbin/openvpn with build id f2037a091ef36b591187a858d75e203690ea9409 not found, continuing without symbols Failed to open /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/e1000e/e1000e.ko, continuing without symbols Failed to open /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/wireless/iwlwifi/iwlcore.ko, continuing without symbols # Samples: 293085637 # # Overhead Command # ........ ............... # 61.70% find 23.50% perf 5.86% swapper 3.12% sshd 2.39% init 0.87% bash 0.86% sleep 0.59% dbus-daemon 0.25% hald 0.24% NetworkManager 0.19% hald-addon-rfki 0.15% openvpn 0.07% phy0 0.07% events/0 0.05% iwl3945 0.05% events/1 0.03% kondemand/0 acme@parisc:~/git/linux-2.6-tip$ Which matches what we get when running the same command for the same perf.data file on the F12, x86_64, source machine: [root@doppio linux-2.6-tip]# perf report --sort comm # Samples: 293085637 # # Overhead Command # ........ ............... # 61.70% find 23.50% perf 5.86% swapper 3.12% sshd 2.39% init 0.87% bash 0.86% sleep 0.59% dbus-daemon 0.25% hald 0.24% NetworkManager 0.19% hald-addon-rfki 0.15% openvpn 0.07% phy0 0.07% events/0 0.05% iwl3945 0.05% events/1 0.03% kondemand/0 [root@doppio linux-2.6-tip]# The other modes work as well, modulo the problem with vmlinux: acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm,dso 2> /dev/null | head -15 # Samples: 293085637 # # Overhead Command Shared Object # ........ ............... ................................. # 35.11% find ffffffff81002b5a 18.25% perf ffffffff8102235f 16.17% find libc-2.10.2.so 9.07% find find 5.80% swapper ffffffff8102235f 3.95% perf libc-2.10.2.so 2.33% init ffffffff810091b9 1.65% sshd libcrypto.so.0.9.8k 1.35% find [e1000e] 0.68% sleep libc-2.10.2.so acme@parisc:~/git/linux-2.6-tip$ And the lack of the right buildids: acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm,dso,symbol 2> /dev/null | head -15 # Samples: 293085637 # # Overhead Command Shared Object Symbol # ........ ............... ................................. ...... # 35.11% find ffffffff81002b5a [k] 0xffffffff81002b5a 18.25% perf ffffffff8102235f [k] 0xffffffff8102235f 16.17% find libc-2.10.2.so [.] 0x00000000045782 9.07% find find [.] 0x0000000000fb0e 5.80% swapper ffffffff8102235f [k] 0xffffffff8102235f 3.95% perf libc-2.10.2.so [.] 0x0000000007f398 2.33% init ffffffff810091b9 [k] 0xffffffff810091b9 1.65% sshd libcrypto.so.0.9.8k [.] 0x00000000105440 1.35% find [e1000e] [k] 0x00000000010948 0.68% sleep libc-2.10.2.so [.] 0x0000000011ad5b acme@parisc:~/git/linux-2.6-tip$ But if we: acme@parisc:~/git/linux-2.6-tip$ ls ~/.debug ls: cannot access /home/acme/.debug: No such file or directory acme@parisc:~/git/linux-2.6-tip$ mkdir -p ~/.debug/lib64/libc-2.10.2.so/ acme@parisc:~/git/linux-2.6-tip$ scp doppio:.debug/lib64/libc-2.10.2.so/* ~/.debug/lib64/libc-2.10.2.so/ acme@doppio's password: eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 100% 1783KB 714.7KB/s 00:02 acme@parisc:~/git/linux-2.6-tip$ mkdir -p ~/.debug/.build-id/eb acme@parisc:~/git/linux-2.6-tip$ ln -s ../../lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 ~/.debug/.build-id/eb/4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 acme@parisc:~/git/linux-2.6-tip$ perf report --dsos libc-2.10.2.so 2> /dev/null # dso: libc-2.10.2.so # Samples: 64281170 # # Overhead Command Symbol # ........ ............... ...... # 14.98% perf [.] __GI_strcmp 12.30% find [.] __GI_memmove 9.25% find [.] _int_malloc 7.60% find [.] _IO_vfprintf_internal 6.10% find [.] _IO_new_file_xsputn 6.02% find [.] __GI_close 3.08% find [.] _IO_file_overflow_internal 3.08% find [.] malloc_consolidate 3.08% find [.] _int_free 3.08% find [.] __strchrnul 3.08% find [.] __getdents64 3.08% find [.] __write_nocancel 3.08% sleep [.] __GI__dl_addr 3.08% sshd [.] __libc_select 3.08% find [.] _IO_new_file_write 3.07% find [.] _IO_new_do_write 3.06% find [.] __GI___errno_location 3.05% find [.] __GI___libc_malloc 3.04% perf [.] __GI_memcpy 1.71% find [.] __fprintf_chk 1.29% bash [.] __gconv_transform_utf8_internal 0.79% dbus-daemon [.] __GI_strlen # # (For a higher level overview, try: perf report --sort comm,dso) # acme@parisc:~/git/linux-2.6-tip$ Which matches what we get on the source, F12, x86_64 machine: [root@doppio linux-2.6-tip]# perf report --dsos libc-2.10.2.so # dso: libc-2.10.2.so # Samples: 64281170 # # Overhead Command Symbol # ........ ............... ...... # 14.98% perf [.] __GI_strcmp 12.30% find [.] __GI_memmove 9.25% find [.] _int_malloc 7.60% find [.] _IO_vfprintf_internal 6.10% find [.] _IO_new_file_xsputn 6.02% find [.] __GI_close 3.08% find [.] _IO_file_overflow_internal 3.08% find [.] malloc_consolidate 3.08% find [.] _int_free 3.08% find [.] __strchrnul 3.08% find [.] __getdents64 3.08% find [.] __write_nocancel 3.08% sleep [.] __GI__dl_addr 3.08% sshd [.] __libc_select 3.08% find [.] _IO_new_file_write 3.07% find [.] _IO_new_do_write 3.06% find [.] __GI___errno_location 3.05% find [.] __GI___libc_malloc 3.04% perf [.] __GI_memcpy 1.71% find [.] __fprintf_chk 1.29% bash [.] __gconv_transform_utf8_internal 0.79% dbus-daemon [.] __GI_strlen # # (For a higher level overview, try: perf report --sort comm,dso) # [root@doppio linux-2.6-tip]# So I think this is really, really nice in that it demonstrates the portability of perf.data files and the use of build-ids accross such aliens worlds :-) There are some things to fix tho, like the bitmap on the header, but things are looking good. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1263478990-8200-2-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar --- tools/perf/util/header.c | 63 +++++++++++++++++----- tools/perf/util/header.h | 2 + tools/perf/util/session.c | 108 ++++++++++++++++++++++++++++++++++---- tools/perf/util/session.h | 7 ++- 4 files changed, 157 insertions(+), 23 deletions(-) diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c index ec96321eb9e4..b31e0ae4b8db 100644 --- a/tools/perf/util/header.c +++ b/tools/perf/util/header.c @@ -1,8 +1,10 @@ #include +#include #include #include #include #include +#include #include "util.h" #include "header.h" @@ -464,8 +466,21 @@ static int do_read(int fd, void *buf, size_t size) return 0; } +static int perf_header__getbuffer64(struct perf_header *self, + int fd, void *buf, size_t size) +{ + if (do_read(fd, buf, size)) + return -1; + + if (self->needs_swap) + mem_bswap_64(buf, size); + + return 0; +} + int perf_header__process_sections(struct perf_header *self, int fd, int (*process)(struct perf_file_section *self, + struct perf_header *ph, int feat, int fd)) { struct perf_file_section *feat_sec; @@ -486,7 +501,7 @@ int perf_header__process_sections(struct perf_header *self, int fd, lseek(fd, self->data_offset + self->data_size, SEEK_SET); - if (do_read(fd, feat_sec, sec_size)) + if (perf_header__getbuffer64(self, fd, feat_sec, sec_size)) goto out_free; err = 0; @@ -494,7 +509,7 @@ int perf_header__process_sections(struct perf_header *self, int fd, if (perf_header__has_feat(self, feat)) { struct perf_file_section *sec = &feat_sec[idx++]; - err = process(sec, feat, fd); + err = process(sec, self, feat, fd); if (err < 0) break; } @@ -511,10 +526,20 @@ int perf_file_header__read(struct perf_file_header *self, lseek(fd, 0, SEEK_SET); if (do_read(fd, self, sizeof(*self)) || - self->magic != PERF_MAGIC || - self->attr_size != sizeof(struct perf_file_attr)) + memcmp(&self->magic, __perf_magic, sizeof(self->magic))) return -1; + if (self->attr_size != sizeof(struct perf_file_attr)) { + u64 attr_size = bswap_64(self->attr_size); + + if (attr_size != sizeof(struct perf_file_attr)) + return -1; + + mem_bswap_64(self, offsetof(struct perf_file_header, + adds_features)); + ph->needs_swap = true; + } + if (self->size != sizeof(*self)) { /* Support the previous format */ if (self->size == offsetof(typeof(*self), adds_features)) @@ -524,16 +549,28 @@ int perf_file_header__read(struct perf_file_header *self, } memcpy(&ph->adds_features, &self->adds_features, - sizeof(self->adds_features)); + sizeof(ph->adds_features)); + /* + * FIXME: hack that assumes that if we need swap the perf.data file + * may be coming from an arch with a different word-size, ergo different + * DEFINE_BITMAP format, investigate more later, but for now its mostly + * safe to assume that we have a build-id section. Trace files probably + * have several other issues in this realm anyway... + */ + if (ph->needs_swap) { + memset(&ph->adds_features, 0, sizeof(ph->adds_features)); + perf_header__set_feat(ph, HEADER_BUILD_ID); + } ph->event_offset = self->event_types.offset; - ph->event_size = self->event_types.size; - ph->data_offset = self->data.offset; + ph->event_size = self->event_types.size; + ph->data_offset = self->data.offset; ph->data_size = self->data.size; return 0; } static int perf_file_section__process(struct perf_file_section *self, + struct perf_header *ph, int feat, int fd) { if (lseek(fd, self->offset, SEEK_SET) < 0) { @@ -548,7 +585,7 @@ static int perf_file_section__process(struct perf_file_section *self, break; case HEADER_BUILD_ID: - if (perf_header__read_build_ids(fd, self->offset, self->size)) + if (perf_header__read_build_ids(ph, fd, self->offset, self->size)) pr_debug("Failed to read buildids, continuing...\n"); break; default: @@ -560,7 +597,7 @@ static int perf_file_section__process(struct perf_file_section *self, int perf_header__read(struct perf_header *self, int fd) { - struct perf_file_header f_header; + struct perf_file_header f_header; struct perf_file_attr f_attr; u64 f_id; int nr_attrs, nr_ids, i, j; @@ -577,8 +614,9 @@ int perf_header__read(struct perf_header *self, int fd) struct perf_header_attr *attr; off_t tmp; - if (do_read(fd, &f_attr, sizeof(f_attr))) + if (perf_header__getbuffer64(self, fd, &f_attr, sizeof(f_attr))) goto out_errno; + tmp = lseek(fd, 0, SEEK_CUR); attr = perf_header_attr__new(&f_attr.attr); @@ -589,7 +627,7 @@ int perf_header__read(struct perf_header *self, int fd) lseek(fd, f_attr.ids.offset, SEEK_SET); for (j = 0; j < nr_ids; j++) { - if (do_read(fd, &f_id, sizeof(f_id))) + if (perf_header__getbuffer64(self, fd, &f_id, sizeof(f_id))) goto out_errno; if (perf_header_attr__add_id(attr, f_id) < 0) { @@ -610,7 +648,8 @@ int perf_header__read(struct perf_header *self, int fd) events = malloc(f_header.event_types.size); if (events == NULL) return -ENOMEM; - if (do_read(fd, events, f_header.event_types.size)) + if (perf_header__getbuffer64(self, fd, events, + f_header.event_types.size)) goto out_errno; event_count = f_header.event_types.size / sizeof(struct perf_trace_event_type); } diff --git a/tools/perf/util/header.h b/tools/perf/util/header.h index 2b69aab67e35..ccc8540feccd 100644 --- a/tools/perf/util/header.h +++ b/tools/perf/util/header.h @@ -52,6 +52,7 @@ struct perf_header { u64 data_size; u64 event_offset; u64 event_size; + bool needs_swap; DECLARE_BITMAP(adds_features, HEADER_FEAT_BITS); }; @@ -80,6 +81,7 @@ bool perf_header__has_feat(const struct perf_header *self, int feat); int perf_header__process_sections(struct perf_header *self, int fd, int (*process)(struct perf_file_section *self, + struct perf_header *ph, int feat, int fd)); #endif /* __PERF_HEADER_H */ diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index e3ccdb46d6c4..604e14f6a6f9 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -1,5 +1,6 @@ #include +#include #include #include @@ -201,21 +202,88 @@ void event__print_totals(void) event__name[i], event__total[i]); } +void mem_bswap_64(void *src, int byte_size) +{ + u64 *m = src; + + while (byte_size > 0) { + *m = bswap_64(*m); + byte_size -= sizeof(u64); + ++m; + } +} + +static void event__all64_swap(event_t *self) +{ + struct perf_event_header *hdr = &self->header; + mem_bswap_64(hdr + 1, self->header.size - sizeof(*hdr)); +} + +static void event__comm_swap(event_t *self) +{ + self->comm.pid = bswap_32(self->comm.pid); + self->comm.tid = bswap_32(self->comm.tid); +} + +static void event__mmap_swap(event_t *self) +{ + self->mmap.pid = bswap_32(self->mmap.pid); + self->mmap.tid = bswap_32(self->mmap.tid); + self->mmap.start = bswap_64(self->mmap.start); + self->mmap.len = bswap_64(self->mmap.len); + self->mmap.pgoff = bswap_64(self->mmap.pgoff); +} + +static void event__task_swap(event_t *self) +{ + self->fork.pid = bswap_32(self->fork.pid); + self->fork.tid = bswap_32(self->fork.tid); + self->fork.ppid = bswap_32(self->fork.ppid); + self->fork.ptid = bswap_32(self->fork.ptid); + self->fork.time = bswap_64(self->fork.time); +} + +static void event__read_swap(event_t *self) +{ + self->read.pid = bswap_32(self->read.pid); + self->read.tid = bswap_32(self->read.tid); + self->read.value = bswap_64(self->read.value); + self->read.time_enabled = bswap_64(self->read.time_enabled); + self->read.time_running = bswap_64(self->read.time_running); + self->read.id = bswap_64(self->read.id); +} + +typedef void (*event__swap_op)(event_t *self); + +static event__swap_op event__swap_ops[] = { + [PERF_RECORD_MMAP] = event__mmap_swap, + [PERF_RECORD_COMM] = event__comm_swap, + [PERF_RECORD_FORK] = event__task_swap, + [PERF_RECORD_EXIT] = event__task_swap, + [PERF_RECORD_LOST] = event__all64_swap, + [PERF_RECORD_READ] = event__read_swap, + [PERF_RECORD_SAMPLE] = event__all64_swap, + [PERF_RECORD_MAX] = NULL, +}; + static int perf_session__process_event(struct perf_session *self, event_t *event, struct perf_event_ops *ops, - unsigned long offset, unsigned long head) + u64 offset, u64 head) { trace_event(event); if (event->header.type < PERF_RECORD_MAX) { - dump_printf("%#lx [%#x]: PERF_RECORD_%s", + dump_printf("%#Lx [%#x]: PERF_RECORD_%s", offset + head, event->header.size, event__name[event->header.type]); ++event__total[0]; ++event__total[event->header.type]; } + if (self->header.needs_swap && event__swap_ops[event->header.type]) + event__swap_ops[event->header.type](event); + switch (event->header.type) { case PERF_RECORD_SAMPLE: return ops->sample(event, self); @@ -241,7 +309,15 @@ static int perf_session__process_event(struct perf_session *self, } } -int perf_header__read_build_ids(int input, u64 offset, u64 size) +void perf_event_header__bswap(struct perf_event_header *self) +{ + self->type = bswap_32(self->type); + self->misc = bswap_16(self->misc); + self->size = bswap_16(self->size); +} + +int perf_header__read_build_ids(struct perf_header *self, + int input, u64 offset, u64 size) { struct build_id_event bev; char filename[PATH_MAX]; @@ -256,6 +332,9 @@ int perf_header__read_build_ids(int input, u64 offset, u64 size) if (read(input, &bev, sizeof(bev)) != sizeof(bev)) goto out; + if (self->needs_swap) + perf_event_header__bswap(&bev.header); + len = bev.header.size - sizeof(bev); if (read(input, filename, len) != len) goto out; @@ -292,9 +371,9 @@ static struct thread *perf_session__register_idle_thread(struct perf_session *se int perf_session__process_events(struct perf_session *self, struct perf_event_ops *ops) { - int err; - unsigned long head, shift; - unsigned long offset = 0; + int err, mmap_prot, mmap_flags; + u64 head, shift; + u64 offset = 0; size_t page_size; event_t *event; uint32_t size; @@ -330,9 +409,16 @@ out_getcwd_err: offset += shift; head -= shift; + mmap_prot = PROT_READ; + mmap_flags = MAP_SHARED; + + if (self->header.needs_swap) { + mmap_prot |= PROT_WRITE; + mmap_flags = MAP_PRIVATE; + } remap: - buf = mmap(NULL, page_size * self->mmap_window, PROT_READ, - MAP_SHARED, self->fd, offset); + buf = mmap(NULL, page_size * self->mmap_window, mmap_prot, + mmap_flags, self->fd, offset); if (buf == MAP_FAILED) { pr_err("failed to mmap file\n"); err = -errno; @@ -342,6 +428,8 @@ remap: more: event = (event_t *)(buf + head); + if (self->header.needs_swap) + perf_event_header__bswap(&event->header); size = event->header.size; if (size == 0) size = 8; @@ -361,12 +449,12 @@ more: size = event->header.size; - dump_printf("\n%#lx [%#x]: event: %d\n", + dump_printf("\n%#Lx [%#x]: event: %d\n", offset + head, event->header.size, event->header.type); if (size == 0 || perf_session__process_event(self, event, ops, offset, head) < 0) { - dump_printf("%#lx [%#x]: skipping unknown header type: %d\n", + dump_printf("%#Lx [%#x]: skipping unknown header type: %d\n", offset + head, event->header.size, event->header.type); /* diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index d4a9d20f8d44..36d1a80c0b6c 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -51,6 +51,8 @@ struct perf_event_ops { struct perf_session *perf_session__new(const char *filename, int mode, bool force); void perf_session__delete(struct perf_session *self); +void perf_event_header__bswap(struct perf_event_header *self); + int perf_session__process_events(struct perf_session *self, struct perf_event_ops *event_ops); @@ -61,7 +63,8 @@ struct symbol **perf_session__resolve_callchain(struct perf_session *self, bool perf_session__has_traces(struct perf_session *self, const char *msg); -int perf_header__read_build_ids(int input, u64 offset, u64 file_size); +int perf_header__read_build_ids(struct perf_header *self, int input, + u64 offset, u64 file_size); int perf_session__set_kallsyms_ref_reloc_sym(struct perf_session *self, const char *symbol_name, @@ -69,4 +72,6 @@ int perf_session__set_kallsyms_ref_reloc_sym(struct perf_session *self, void perf_session__reloc_vmlinux_maps(struct perf_session *self, u64 unrelocated_addr); +void mem_bswap_64(void *src, int byte_size); + #endif /* __PERF_SESSION_H */ -- 2.39.5