From d37a369790774af66a4aee61a188384d21b17a43 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 10 Jul 2017 16:08:07 -0300 Subject: perf evsel: Fix attr.exclude_kernel setting for default cycles:p To allow probing the max attr.precise_ip setting for non-root users we unconditionally set attr.exclude_kernel, which makes the detection work but should be done only for !root, fix it. Cc: Adrian Hunter Cc: Andy Lutomirski Cc: David Ahern Cc: Jiri Olsa Cc: Namhyung Kim Cc: Wang Nan Fixes: 97365e81366f ("perf evsel: Set attr.exclude_kernel when probing max attr.precise_ip") Link: http://lkml.kernel.org/n/tip-bl6bbxzxloonzvm4nvt7oqgj@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/evsel.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c index 87b431886670..f2a1876d268e 100644 --- a/tools/perf/util/evsel.c +++ b/tools/perf/util/evsel.c @@ -273,7 +273,7 @@ struct perf_evsel *perf_evsel__new_cycles(void) struct perf_event_attr attr = { .type = PERF_TYPE_HARDWARE, .config = PERF_COUNT_HW_CPU_CYCLES, - .exclude_kernel = 1, + .exclude_kernel = geteuid() != 0, }; struct perf_evsel *evsel; -- cgit v1.2.3-58-ga151 From ede5626d303b721dd02246a3850380943c24e380 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 10 Jul 2017 16:19:25 -0300 Subject: perf evsel: State in the default event name if attr.exclude_kernel is set When no event is specified perf will use the "cycles" hardware event with the highest precision available in the processor, and excluding kernel events for non-root users, so make that clear in the event name by setting the "u" event modifier, i.e. "cycles:upp". E.g.: The default for root: # perf record usleep 1 # perf evlist -v cycles:ppp: ..., precise_ip: 3, exclude_kernel: 0, ... # And for !root: $ perf record usleep 1 $ perf evlist -v cycles:uppp: ... , precise_ip: 3, exclude_kernel: 1, ... $ Cc: Adrian Hunter Cc: David Ahern Cc: Jiri Olsa Cc: Namhyung Kim Cc: Wang Nan Link: http://lkml.kernel.org/n/tip-lf29zcdl422i9knrgde0uwy3@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/evsel.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c index f2a1876d268e..413f74df08de 100644 --- a/tools/perf/util/evsel.c +++ b/tools/perf/util/evsel.c @@ -298,8 +298,10 @@ struct perf_evsel *perf_evsel__new_cycles(void) goto out; /* use asprintf() because free(evsel) assumes name is allocated */ - if (asprintf(&evsel->name, "cycles%.*s", - attr.precise_ip ? attr.precise_ip + 1 : 0, ":ppp") < 0) + if (asprintf(&evsel->name, "cycles%s%s%.*s", + (attr.precise_ip || attr.exclude_kernel) ? ":" : "", + attr.exclude_kernel ? "u" : "", + attr.precise_ip ? attr.precise_ip + 1 : 0, "ppp") < 0) goto error_free; out: return evsel; -- cgit v1.2.3-58-ga151 From 80f62589fa52f530cffc50e78c0b5a2ae572d61e Mon Sep 17 00:00:00 2001 From: Jin Yao Date: Thu, 8 Jun 2017 14:01:44 +0800 Subject: perf annotate: Fix broken arrow at row 0 connecting jmp instruction to its target MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When the jump instruction is displayed at the row 0 in annotate view, the arrow is broken. An example: 16.86 │ ┌──je 82 0.01 │ movsd (%rsp),%xmm0 │ movsd 0x8(%rsp),%xmm4 │ movsd 0x8(%rsp),%xmm1 │ movsd (%rsp),%xmm3 │ divsd %xmm4,%xmm0 │ divsd %xmm3,%xmm1 │ movsd (%rsp),%xmm2 │ addsd %xmm1,%xmm0 │ addsd %xmm2,%xmm0 │ movsd %xmm0,(%rsp) │82: sub $0x1,%ebx 83.03 │ ↑ jne 38 │ add $0x10,%rsp │ xor %eax,%eax │ pop %rbx │ ← retq The patch increments the row number before checking with 0. Signed-off-by: Yao Jin Tested-by: Arnaldo Carvalho de Melo Cc: Alexander Shishkin Cc: Andi Kleen Cc: Jiri Olsa Cc: Kan Liang Cc: Peter Zijlstra Cc: stable@vger.kernel.org Fixes: 944e1abed9e1 ("perf ui browser: Add method to draw up/down arrow line") Link: http://lkml.kernel.org/r/1496901704-30275-1-git-send-email-yao.jin@linux.intel.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/ui/browser.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/ui/browser.c b/tools/perf/ui/browser.c index a4d3762cd825..83874b0e266c 100644 --- a/tools/perf/ui/browser.c +++ b/tools/perf/ui/browser.c @@ -704,7 +704,7 @@ static void __ui_browser__line_arrow_down(struct ui_browser *browser, ui_browser__gotorc(browser, row, column + 1); SLsmg_draw_hline(2); - if (row++ == 0) + if (++row == 0) goto out; } else row = 0; -- cgit v1.2.3-58-ga151 From 4b1303d0b01440f224cf81493b7e8e43d9b4965e Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Tue, 11 Jul 2017 16:21:40 -0300 Subject: perf symbols: Accept zero as the kernel base address Which is the case in S/390, where symbols were not being resolved because machine__get_kernel_start was only setting machine->kernel_start when the just successfully loaded kernel symtab had its map->start set to !0, when it was left at (1ULL << 63) assuming a partitioning of the address space for user/kernel, which is not the case in S/390 nor in Sparc. So just check if map__load() was successfull and set machine->kernel_start to zero, fixing kernel symbol resolution on S/390. Test performed by Thomas: ---- I like this patch. I have done a new build and removed all my debug output to start from scratch. Without your patch I get this: # Samples: 4 of event 'cpu-clock' # Event count (approx.): 1000000 # # Children Self Command Shared Object Symbol # ........ ........ ....... ................ ........................ 75.00% 0.00% true [unknown] [k] 0x00000000004bedda | ---0x4bedda | |--50.00%--0x42693a | | | --25.00%--0x2a72e0 | 0x2af0ca | 0x3d1003fe4c0 | --25.00%--0x4272bc 0x26fa84 and with your patch (I just rebuilt the perf tool, nothing else and used the same perf.data file as input): # Samples: 4 of event 'cpu-clock' # Event count (approx.): 1000000 # # Children Self Command Shared Object Symbol # ........ ........ ....... .......................... .................................. 75.00% 0.00% true [kernel.vmlinux] [k] pgm_check_handler | ---pgm_check_handler do_dat_exception handle_mm_fault __handle_mm_fault filemap_map_pages | |--25.00%--rcu_read_lock_held | rcu_lockdep_current_cpu_online | 0x3d1003ff4c0 | --25.00%--lock_release Looks good to me.... ---- Reported-and-Tested-by: Thomas-Mich Richter Cc: Adrian Hunter Cc: David Ahern Cc: Hendrik Brueckner Cc: Jiri Olsa Cc: Namhyung Kim Cc: Wang Nan Cc: Zvonko Kosic Link: http://lkml.kernel.org/n/tip-dk0n1uzmbe0tbthrpfqlx6bz@git.kernel.org Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/machine.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c index 5de2b86b9880..2e9eb6aa3ce2 100644 --- a/tools/perf/util/machine.c +++ b/tools/perf/util/machine.c @@ -2209,7 +2209,7 @@ int machine__get_kernel_start(struct machine *machine) machine->kernel_start = 1ULL << 63; if (map) { err = map__load(map); - if (map->start) + if (!err) machine->kernel_start = map->start; } return err; -- cgit v1.2.3-58-ga151 From 545722cb0fc993226a01844fb27cf832459eb1c0 Mon Sep 17 00:00:00 2001 From: Edward Cree Date: Fri, 21 Jul 2017 14:36:57 +0100 Subject: selftests/bpf: subtraction bounds test There is a bug in the verifier's handling of BPF_SUB: [a,b] - [c,d] yields was [a-c, b-d] rather than the correct [a-d, b-c]. So here is a test which, with the bogus handling, will produce ranges of [0,0] and thus allowed accesses; whereas the correct handling will give a range of [-255, 255] (and hence the right-shift will give a range of [0, 255]) and the accesses will be rejected. Signed-off-by: Edward Cree Acked-by: Daniel Borkmann Signed-off-by: David S. Miller --- tools/testing/selftests/bpf/test_verifier.c | 28 ++++++++++++++++++++++++++++ 1 file changed, 28 insertions(+) (limited to 'tools') diff --git a/tools/testing/selftests/bpf/test_verifier.c b/tools/testing/selftests/bpf/test_verifier.c index af7d173910f4..addea82f76c9 100644 --- a/tools/testing/selftests/bpf/test_verifier.c +++ b/tools/testing/selftests/bpf/test_verifier.c @@ -5980,6 +5980,34 @@ static struct bpf_test tests[] = { .result = REJECT, .result_unpriv = REJECT, }, + { + "subtraction bounds (map value)", + .insns = { + BPF_ST_MEM(BPF_DW, BPF_REG_10, -8, 0), + BPF_MOV64_REG(BPF_REG_2, BPF_REG_10), + BPF_ALU64_IMM(BPF_ADD, BPF_REG_2, -8), + BPF_LD_MAP_FD(BPF_REG_1, 0), + BPF_RAW_INSN(BPF_JMP | BPF_CALL, 0, 0, 0, + BPF_FUNC_map_lookup_elem), + BPF_JMP_IMM(BPF_JEQ, BPF_REG_0, 0, 9), + BPF_LDX_MEM(BPF_B, BPF_REG_1, BPF_REG_0, 0), + BPF_JMP_IMM(BPF_JGT, BPF_REG_1, 0xff, 7), + BPF_LDX_MEM(BPF_B, BPF_REG_3, BPF_REG_0, 1), + BPF_JMP_IMM(BPF_JGT, BPF_REG_3, 0xff, 5), + BPF_ALU64_REG(BPF_SUB, BPF_REG_1, BPF_REG_3), + BPF_ALU64_IMM(BPF_RSH, BPF_REG_1, 56), + BPF_ALU64_REG(BPF_ADD, BPF_REG_0, BPF_REG_1), + BPF_LDX_MEM(BPF_B, BPF_REG_0, BPF_REG_0, 0), + BPF_EXIT_INSN(), + BPF_MOV64_IMM(BPF_REG_0, 0), + BPF_EXIT_INSN(), + }, + .fixup_map1 = { 3 }, + .errstr_unpriv = "R0 pointer arithmetic prohibited", + .errstr = "R0 min value is negative, either use unsigned index or do a if (index >=0) check.", + .result = REJECT, + .result_unpriv = REJECT, + }, }; static int probe_filter_length(const struct bpf_insn *fp) -- cgit v1.2.3-58-ga151 From efcb521943a8df5210f16f312037c2edc3e1449f Mon Sep 17 00:00:00 2001 From: Lin Ma Date: Tue, 25 Jul 2017 19:05:53 +0800 Subject: tools/kvm_stat: use variables instead of hard paths in help output Using variables instead of hard paths makes the requirements information more accurate. Signed-off-by: Lin Ma Signed-off-by: Paolo Bonzini --- tools/kvm/kvm_stat/kvm_stat | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'tools') diff --git a/tools/kvm/kvm_stat/kvm_stat b/tools/kvm/kvm_stat/kvm_stat index dd8f00cfb8b4..5704044b1c0b 100755 --- a/tools/kvm/kvm_stat/kvm_stat +++ b/tools/kvm/kvm_stat/kvm_stat @@ -1413,8 +1413,8 @@ performance. Requirements: - Access to: - /sys/kernel/debug/kvm - /sys/kernel/debug/trace/events/* + %s + %s/events/* /proc/pid/task - /proc/sys/kernel/perf_event_paranoid < 1 if user has no CAP_SYS_ADMIN and perf events are used. @@ -1434,7 +1434,7 @@ Interactive Commands: s set update interval x toggle reporting of stats for individual child trace events Press any other key to refresh statistics immediately. -""" +""" % (PATH_DEBUGFS_KVM, PATH_DEBUGFS_TRACING) class PlainHelpFormatter(optparse.IndentedHelpFormatter): def format_description(self, description): -- cgit v1.2.3-58-ga151 From 67fbcd62f54d4503e3dc63b68af1c6757b74e050 Mon Sep 17 00:00:00 2001 From: Lin Ma Date: Tue, 25 Jul 2017 19:05:54 +0800 Subject: tools/kvm_stat: add '-f help' to get the available event list Signed-off-by: Lin Ma Signed-off-by: Paolo Bonzini --- tools/kvm/kvm_stat/kvm_stat | 16 ++++++++++++++-- 1 file changed, 14 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/kvm/kvm_stat/kvm_stat b/tools/kvm/kvm_stat/kvm_stat index 5704044b1c0b..32283d88701a 100755 --- a/tools/kvm/kvm_stat/kvm_stat +++ b/tools/kvm/kvm_stat/kvm_stat @@ -474,7 +474,7 @@ class Provider(object): @staticmethod def is_field_wanted(fields_filter, field): """Indicate whether field is valid according to fields_filter.""" - if not fields_filter: + if not fields_filter or fields_filter == "help": return True return re.match(fields_filter, field) is not None @@ -1496,7 +1496,8 @@ Press any other key to refresh statistics immediately. action='store', default=DEFAULT_REGEX, dest='fields', - help='fields to display (regex)', + help='''fields to display (regex) + "-f help" for a list of available events''', ) optparser.add_option('-p', '--pid', action='store', @@ -1559,6 +1560,17 @@ def main(): stats = Stats(options) + if options.fields == "help": + event_list = "\n" + s = stats.get() + for key in s.keys(): + if key.find('(') != -1: + key = key[0:key.find('(')] + if event_list.find('\n' + key + '\n') == -1: + event_list += key + '\n' + sys.stdout.write(event_list) + return "" + if options.log: log(stats) elif not options.once: -- cgit v1.2.3-58-ga151 From d777b2ddbecf509bc61ee4f0fe0d3b5a909d698a Mon Sep 17 00:00:00 2001 From: Jakub Kicinski Date: Tue, 25 Jul 2017 15:16:12 -0700 Subject: bpf: don't zero out the info struct in bpf_obj_get_info_by_fd() The buffer passed to bpf_obj_get_info_by_fd() should be initialized to zeros. Kernel will enforce that to guarantee we can safely extend info structures in the future. Making the bpf_obj_get_info_by_fd() call in libbpf perform the zeroing is problematic, however, since some members of the info structures may need to be initialized by the callers (for instance pointers to buffers to which kernel is to dump translated and jited images). Remove the zeroing and fix up the in-tree callers before any kernel has been released with this code. As Daniel points out this seems to be the intended operation anyway, since commit 95b9afd3987f ("bpf: Test for bpf ID") is itself setting the buffer pointers before calling bpf_obj_get_info_by_fd(). Signed-off-by: Jakub Kicinski Acked-by: Daniel Borkmann Signed-off-by: David S. Miller --- tools/lib/bpf/bpf.c | 1 - tools/testing/selftests/bpf/test_progs.c | 8 ++++++-- 2 files changed, 6 insertions(+), 3 deletions(-) (limited to 'tools') diff --git a/tools/lib/bpf/bpf.c b/tools/lib/bpf/bpf.c index 412a7c82995a..256f571f2ab5 100644 --- a/tools/lib/bpf/bpf.c +++ b/tools/lib/bpf/bpf.c @@ -314,7 +314,6 @@ int bpf_obj_get_info_by_fd(int prog_fd, void *info, __u32 *info_len) int err; bzero(&attr, sizeof(attr)); - bzero(info, *info_len); attr.info.bpf_fd = prog_fd; attr.info.info_len = *info_len; attr.info.info = ptr_to_u64(info); diff --git a/tools/testing/selftests/bpf/test_progs.c b/tools/testing/selftests/bpf/test_progs.c index 5855cd3d3d45..1f7dd35551b9 100644 --- a/tools/testing/selftests/bpf/test_progs.c +++ b/tools/testing/selftests/bpf/test_progs.c @@ -340,6 +340,7 @@ static void test_bpf_obj_id(void) /* Check getting prog info */ info_len = sizeof(struct bpf_prog_info) * 2; + bzero(&prog_infos[i], info_len); prog_infos[i].jited_prog_insns = ptr_to_u64(jited_insns); prog_infos[i].jited_prog_len = sizeof(jited_insns); prog_infos[i].xlated_prog_insns = ptr_to_u64(xlated_insns); @@ -369,6 +370,7 @@ static void test_bpf_obj_id(void) /* Check getting map info */ info_len = sizeof(struct bpf_map_info) * 2; + bzero(&map_infos[i], info_len); err = bpf_obj_get_info_by_fd(map_fds[i], &map_infos[i], &info_len); if (CHECK(err || @@ -394,7 +396,7 @@ static void test_bpf_obj_id(void) nr_id_found = 0; next_id = 0; while (!bpf_prog_get_next_id(next_id, &next_id)) { - struct bpf_prog_info prog_info; + struct bpf_prog_info prog_info = {}; int prog_fd; info_len = sizeof(prog_info); @@ -418,6 +420,8 @@ static void test_bpf_obj_id(void) nr_id_found++; err = bpf_obj_get_info_by_fd(prog_fd, &prog_info, &info_len); + prog_infos[i].jited_prog_insns = 0; + prog_infos[i].xlated_prog_insns = 0; CHECK(err || info_len != sizeof(struct bpf_prog_info) || memcmp(&prog_info, &prog_infos[i], info_len), "get-prog-info(next_id->fd)", @@ -436,7 +440,7 @@ static void test_bpf_obj_id(void) nr_id_found = 0; next_id = 0; while (!bpf_map_get_next_id(next_id, &next_id)) { - struct bpf_map_info map_info; + struct bpf_map_info map_info = {}; int map_fd; info_len = sizeof(map_info); -- cgit v1.2.3-58-ga151