perf 是怎么计算调用栈的时间的？

在我真个malloc的执行过程中共调用了8次的syswrite的系统调用，其中有两次来自于__lib_write, 两次来自于__memmove_avx_unaligned，然后__memmove_avx_unaligned中有两次是来自于funca，有两次是来自于funb，所以在总共的8次调用中，有两次是来自于funcb，有两次来自于funca.

这里并不是一个时间的关系呢；

Samples: 8  of event 'syscalls:sys_enter_write, Event count (approx.): 8
  Children      Self  Command  Shared Object     Symbol                                                                                       ▒
-  100.00%   100.00%  malloc   malloc            [.] __libc_write                                                                             ▒
   - 25.00% 0xa3437                                                                                                                           ▒
        __libc_write                                                                                                                          ▒
   - 25.00% 0x860f20f983483fe1                                                                                                                ▒
        __memmove_avx_unaligned                                                                                                               ▒
        generic_start_main                                                                                                                    ▒
      - main                                                                                                                                  ▒
         - 12.50% funcb                                                                                                                       ▒
              func1                                                                                                                           ▒
              func2                                                                                                                           ▒
              func3                                                                                                                           ▒
              __libc_write                                                                                                                    ▒
         - 12.50% funca                                                                                                                       ▒
              func1                                                                                                                           ▒
              func2                                                                                                                           ▒
              func3                                                                                                                           ▒
              __libc_write                                                                                                                    ▒
   - 25.00% 0xa3031300a3437                                                                                                                   ▒
        __libc_write                                                                                                                          ▒
   - 12.50% 0xa303130353532                                                                                                                   ▒
        __libc_write                                                                                                                          ▒
   - 12.50% 0xa303130343132                                                                                                                   ▒
        __libc_write                                                                                                                          ▒
+   25.00%     0.00%  malloc   malloc            [.] func3                                                                                    ◆
+   25.00%     0.00%  malloc   [unknown]         [.] 0x00000000000a3437                                                                       ▒
+   25.00%     0.00%  malloc   [unknown]         [.] 0x000a3031300a3437                                                                       ▒
+   25.00%     0.00%  malloc   malloc            [.] func2                                                                                    ▒
+   25.00%     0.00%  malloc   malloc            [.] func1                                                                                    ▒
+   25.00%     0.00%  malloc   malloc            [.] main                                                                                     ▒
+   25.00%     0.00%  malloc   malloc            [.] generic_start_main                                                                       ▒
+   25.00%     0.00%  malloc   malloc            [.] __memmove_avx_unaligned                                                                  ▒
+   25.00%     0.00%  malloc   [unknown]         [.] 0x860f20f983483fe1                                                                       ▒
+   12.50%     0.00%  malloc   [unknown]         [.] 0x000a303130343132                                                                       ▒
+   12.50%     0.00%  malloc   [unknown]         [.] 0x000a303130353532                                                                       ▒
+   12.50%     0.00%  malloc   malloc            [.] funca                                                                                    ▒
+   12.50%     0.00%  malloc   malloc            [.] funcb                                                                                    ▒
                                                                                                                                              ▒
                                                                                                                                              ▒

为了能够更加精确地看到软件事件的发生，我们选择一个比较生僻的系统调用来观察 perf_event_open