Messages in this thread | | | Subject | Re: [PATCH V1 06/19] perf tools: Fix kernel_start for PTI on x86 | From | Adrian Hunter <> | Date | Wed, 16 May 2018 16:45:41 +0300 |
| |
On 16/05/18 16:00, Arnaldo Carvalho de Melo wrote: > Em Tue, May 15, 2018 at 03:43:20PM +0300, Adrian Hunter escreveu: >> On x86_64, PTI entry trampolines are less than the start of kernel text, >> but still above 2^63. So leave kernel_start = 1ULL << 63 for x86_64. >> >> Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> >> --- >> tools/perf/util/machine.c | 16 +++++++++++++++- >> tools/perf/util/machine.h | 2 ++ >> 2 files changed, 17 insertions(+), 1 deletion(-) >> >> diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c >> index 72a351613d85..2f39a4d014de 100644 >> --- a/tools/perf/util/machine.c >> +++ b/tools/perf/util/machine.c >> @@ -2296,6 +2296,15 @@ int machine__set_current_tid(struct machine *machine, int cpu, pid_t pid, >> return 0; >> } >> >> +/* >> + * Compares the raw arch string. N.B. see instead perf_env__arch() if a >> + * normalized arch is needed. >> + */ >> +bool machine__is(struct machine *machine, const char *arch) >> +{ >> + return machine && machine->env && !strcmp(machine->env->arch, arch); >> +} >> + > > Broke 'perf top': > > gdb perf > bt > bt > #0 __strcmp_sse2_unaligned () at ../sysdeps/x86_64/multiarch/strcmp-sse2-unaligned.S:31 > #1 0x0000000000528b1e in machine__is (machine=0xc55548, arch=0x7d9058 "x86_64") at util/machine.c:2305 > #2 0x0000000000528bbb in machine__get_kernel_start (machine=0xc55548) at util/machine.c:2329 > #3 0x00000000004df894 in machine__kernel_start (machine=0xc55548) at util/machine.h:90 > #4 0x00000000004df8e0 in machine__kernel_ip (machine=0xc55548, ip=140352692842292) at util/machine.h:96 > #5 0x00000000004e3a9c in thread__find_map (thread=0x21f9de0, cpumode=2 '\002', addr=140352692842292, al=0x7fffffff8160) > at util/event.c:1550 > #6 0x00000000004e3c51 in machine__resolve (machine=0xc55548, al=0x7fffffff8160, sample=0x7fffffff8360) at util/event.c:1591 > #7 0x000000000045a274 in perf_event__process_sample (tool=0x7fffffffa720, event=0x7fffea662878, evsel=0xc120e0, > sample=0x7fffffff8360, machine=0xc55548) at builtin-top.c:734 > #8 0x000000000045a91c in perf_top__mmap_read_idx (top=0x7fffffffa720, idx=3) at builtin-top.c:872 > #9 0x000000000045aa6b in perf_top__mmap_read (top=0x7fffffffa720) at builtin-top.c:898 > #10 0x000000000045b460 in __cmd_top (top=0x7fffffffa720) at builtin-top.c:1167 > #11 0x000000000045c973 in cmd_top (argc=0, argv=0x7fffffffe0e0) at builtin-top.c:1509 > #12 0x00000000004ce403 in run_builtin (p=0xac5670 <commands+336>, argc=1, argv=0x7fffffffe0e0) at perf.c:303 > #13 0x00000000004ce670 in handle_internal_command (argc=1, argv=0x7fffffffe0e0) at perf.c:355 > #14 0x00000000004ce7c2 in run_argv (argcp=0x7fffffffdf3c, argv=0x7fffffffdf30) at perf.c:399 > #15 0x00000000004ceb90 in main (argc=1, argv=0x7fffffffe0e0) at perf.c:543 > (gdb) fr 1 > #1 0x0000000000528b1e in machine__is (machine=0xc55548, arch=0x7d9058 "x86_64") at util/machine.c:2305 > 2305 return machine && machine->env && !strcmp(machine->env->arch, arch); > (gdb) list > 2300 * Compares the raw arch string. N.B. see instead perf_env__arch() if a > 2301 * normalized arch is needed. > 2302 */ > 2303 bool machine__is(struct machine *machine, const char *arch) > 2304 { > 2305 return machine && machine->env && !strcmp(machine->env->arch, arch); > 2306 } > 2307 > 2308 int machine__get_kernel_start(struct machine *machine) > 2309 { > (gdb) p machine > $1 = (struct machine *) 0xc55548 > (gdb) p machine->env > $2 = (struct perf_env *) 0xc06400 <perf_env> > (gdb) p machine-env->arch > No symbol "env" in current context. > (gdb) p machine->env->arch > $3 = 0x0 > (gdb) >
If there is no perf_data then perf_session__new() uses perf_env but it seems perf_env.arch is not initialized. Would it be OK to initialize perf_env.arch and perf_env.nr_cpus_avail?
| |