OpenHarmony使用strace

简介

Strace是一个强大的进程行为跟踪工具，当怀疑某个进程运行异常的时候，比如死锁，CPU高或者莫名卡顿等，可以使用strace查看该进程发起的系统调用和接收到的信号，从而确定问题所在。当然也可以利用strace去学习了解进程的运行流程，调用库以及参数和统计。那么如何在OpenHarmony里编译使用此工具就是我们这篇文章要分享的话题。

编译构建

• 下载：从https://github.com/strace下载最新源码

  //下载地址
  wget https://github.com/strace/strace/releases/download/v6.6/strace-6.6.tar.xz

• 安装编译环境

  //编译工具链（32位）
  sudo apt-get install gcc-arm-linux-gnueabi
  //ohos的源码下载后，其实也有对应的工具链，位置如下
  prebuilts/gcc/linux-x86/arm（32位）/gcc-linaro-7.5.0-arm-linux-gnueabi/bin

• 编译

  //设置环境变量
  export CC=arm-linux-gnueabi-gcc
  export CFLAGS=-static
  export CPPFLAGS=-static
  export LDFLAGS='-static -pthread'
  //配置
   ./configure CC=arm-linux-gnueabi-gcc CXX=arm-linux-gnueabi-cpp  LD=arm-linux-gnueabi-ld RANLIB=arm-linux-gnueabi-ranlib --host=arm-linux --target=arm-linux --enable-mpers=no 
  //编译
  make
  //strip
  arm-linux-gnueabi-strip strace

• 运行测试

  //传到板子上
  hdc_std.exe shell mount -o rw,remount /
  hdc_std.exe file send strace /system/bin
  hdc_std.exe shell chmod +x /system/bin/strace
  
  //运行测试
  hdc_std.exe shell
  # strace --help
  Usage: strace [-ACdffhiqqrtttTvVwxxyyzZ] [-I N] [-b execve] [-e EXPR]...
                [-a COLUMN] [-o FILE] [-s STRSIZE] [-X FORMAT] [-P PATH]...
                [-p PID]... [--seccomp-bpf]
                { -p PID | [-DDD] [-E VAR=VAL]... [-u USERNAME] PROG [ARGS] }
     or: strace -c[dfwzZ] [-I N] [-b execve] [-e EXPR]... [-O OVERHEAD]
                [-S SORTBY] [-P PATH]... [-p PID]... [--seccomp-bpf]
                { -p PID | [-DDD] [-E VAR=VAL]... [-u USERNAME] PROG [ARGS] }
  
  General:
    -e EXPR        a qualifying expression: OPTION=[!]all or OPTION=[!]VAL1[,VAL2]...
       options:    trace, abbrev, verbose, raw, signal, read, write, fault,
                   inject, status, kvm
  
  Startup:
    -E VAR=VAL, --env=VAR=VAL
                   put VAR=VAL in the environment for command
    -E VAR, --env=VAR
                   remove VAR from the environment for command
    -p PID, --attach=PID
                   trace process with process id PID, may be repeated
    -u USERNAME, --user=USERNAME
                   run command as USERNAME handling setuid and/or setgid
  
  Tracing:
    -b execve, --detach-on=execve
                   detach on execve syscall
    -D             run tracer process as a grandchild, not as a parent
    -DD            run tracer process in a separate process group
    -DDD           run tracer process in a separate session
    -f             follow forks
    -ff            follow forks with output into separate files
    -I INTERRUPTIBLE
       1:          no signals are blocked
       2:          fatal signals are blocked while decoding syscall (default)
       3:          fatal signals are always blocked (default if '-o FILE PROG')
       4:          fatal signals and SIGTSTP (^Z) are always blocked
                   (useful to make 'strace -o FILE PROG' not stop on ^Z)
  
  Filtering:
    -e trace=[!]{[?]SYSCALL[@64|@32|@x32]|[?]/REGEX|GROUP|all|none},
    --trace=[!]{[?]SYSCALL[@64|@32|@x32]|[?]/REGEX|GROUP|all|none}
                   trace only specified syscalls.
       groups:     %creds, %desc, %file, %fstat, %fstatfs %ipc, %lstat,
                   %memory, %net, %process, %pure, %signal, %stat, %%stat,
                   %statfs, %%statfs
    -e signal=SET, --signal=SET
                   trace only the specified set of signals
                   print only the signals from SET
    -e status=SET, --status=SET
                   print only system calls with the return statuses in SET
       statuses:   successful, failed, unfinished, unavailable, detached
    -P PATH, --trace-path=PATH
                   trace accesses to PATH
    -z             print only syscalls that returned without an error code
    -Z             print only syscalls that returned with an error code
  
  Output format:
    -a COLUMN, --columns=COLUMN
                   alignment COLUMN for printing syscall results (default 40)
    -e abbrev=SET, --abbrev=SET
                   abbreviate output for the syscalls in SET
    -e verbose=SET, --verbose=SET
                   dereference structures for the syscall in SET
    -e raw=SET, --raw=SET
                   print undecoded arguments for the syscalls in SET
    -e read=SET, --read=SET
                   dump the data read from the file descriptors in SET
    -e write=SET, --write=SET
                   dump the data written to the file descriptors in SET
    -e kvm=vcpu, --kvm=vcpu
                   print exit reason of kvm vcpu
    -i, --instruction-pointer
                   print instruction pointer at time of syscall
    -o FILE, --output=FILE
                   send trace output to FILE instead of stderr
    -A, --output-append-mode
                   open the file provided in the -o option in append mode
    -q             suppress messages about attaching, detaching, etc.
    -qq            suppress messages about process exit status as well.
    -r             print relative timestamp
    -s STRSIZE, --string-limit=STRSIZE
                   limit length of print strings to STRSIZE chars (default 32)
    -t             print absolute timestamp
    -tt            print absolute timestamp with usecs
    -ttt           print absolute UNIX time with usecs
    -T             print time spent in each syscall
    -v, --no-abbrev
                   verbose mode: print entities unabbreviated
    -x             print non-ascii strings in hex
    -xx            print all strings in hex
    -X FORMAT      set the FORMAT for printing of named constants and flags
       formats:    raw, abbrev, verbose
    -y             print paths associated with file descriptor arguments
    -yy            print protocol specific information associated with socket
                   file descriptors
  
  Statistics:
    -c, --summary-only
                   count time, calls, and errors for each syscall and report
                   summary
    -C, --summary  like -c, but also print the regular output
    -O OVERHEAD    set overhead for tracing syscalls to OVERHEAD usecs
    -S SORTBY, --summary-sort-by=SORTBY
                   sort syscall counts by: time, calls, errors, name, nothing
                   (default time)
    -w             summarise syscall latency (default is system time)
  
  Tampering:
    -e inject=SET[:error=ERRNO|:retval=VALUE][:signal=SIG][:syscall=SYSCALL]
              [:delay_enter=DELAY][:delay_exit=DELAY][:when=WHEN],
    --inject=SET[:error=ERRNO|:retval=VALUE][:signal=SIG][:syscall=SYSCALL]
             [:delay_enter=DELAY][:delay_exit=DELAY][:when=WHEN]
                   perform syscall tampering for the syscalls in SET
       delay:      milliseconds or NUMBER{s|ms|us|ns}
       when:       FIRST, FIRST+, or FIRST+STEP
    -e fault=SET[:error=ERRNO][:when=WHEN], --fault=SET[:error=ERRNO][:when=WHEN]
                   synonym for -e inject with default ERRNO set to ENOSYS.
  Miscellaneous:
    -d, --debug    enable debug output to stderr
    -h, --help     print help message
    --seccomp-bpf  enable seccomp-bpf filtering
    -V, --version  print version

小结

• 简单说说strace的典型场景：

  • 进程统计：针对某个运行高cpu消耗的进程，进行监控统计，什么操作耗时多？后台在做什么事？统计出来高消耗的系统操作

  # top
  Tasks: 243 total,   2 running, 237 sleeping,   0 stopped,   4 zombie
    Mem:  1992724K total,  1501188K used,   491536K free,   1830912 buffers
   Swap:  1048572K total,         0 used,  1048572K free,   683260K cached
  400%cpu 125%user   1%nice  70%sys 202%idle   0%iow   0%irq   2%sirq   0%host
    PID USER         PR  NI VIRT  RES  SHR S[%CPU] %MEM     TIME+ ARGS
    555 graphics     -2  -8 150M  62M  45M S 72.4   3.1 895:19.80 render_service
    539 composer_ho+ -2  -8  17M 4.9M 3.8M S 23.2   0.2 282:37.59 composer_host           9 composer_host -8 1
    247 logd         20   0  16M 4.9M 2.3M S 12.2   0.2 167:45.62 hilogd
  14516 root          0 -20    0    0    0 I  3.3   0.0   0:07.05 [kworker/u9:1-kbase_pm_poweroff_wait]
  14587 root         25   5  14M 3.0M 2.6M R  2.9   0.1   0:00.12 top
    178 root         RT   0    0    0    0 S  2.3   0.0  36:26.03 [sugov:0]
  
  # strace -cp 555
  strace: Process 555 attached
  strace: Process 555 detached
  % time     seconds  usecs/call     calls    errors syscall
  ------ ----------- ----------- --------- --------- ----------------
   22.34    0.456649          22     20705           clock_gettime64
   20.30    0.415126          66      6201           ioctl
   16.03    0.327786          45      7236           writev
   10.87    0.222298          31      7006           close
    7.24    0.148065          18      7855           gettid
    5.90    0.120715          58      2058       168 futex
    3.65    0.074596          22      3296           dup
    2.76    0.056384          68       825           munmap
    2.27    0.046345          55       828           write
    1.93    0.039523          27      1442           poll
    1.64    0.033546          40       824           mmap2
    1.49    0.030400          36       828           fcntl64
    1.40    0.028689          34       824           prctl
    0.80    0.016341          36       447       206 read
    0.50    0.010236          24       412           _llseek
    0.46    0.009408          45       207           madvise
    0.36    0.007341          35       206           epoll_pwait
    0.05    0.001017        1017         1           restart_syscall
  ------ ----------- ----------- --------- --------- ----------------
  100.00    2.044465                 61201       374 total

  • 跟进分析：通过「T」选项可以获取操作实际消耗的时间，通过「e」选项可以跟踪某个操作

  # strace -T -e clock_gettime64 -p 555
  strace: Process 555 attached
  clock_gettime64(CLOCK_MONOTONIC, {tv_sec=89616, tv_nsec=627412894}) = 0 <0.000037>
  clock_gettime64(CLOCK_MONOTONIC, {tv_sec=89616, tv_nsec=627777769}) = 0 <0.000149>
  clock_gettime64(CLOCK_REALTIME, {tv_sec=1702433139, tv_nsec=94341493}) = 0 <0.000047>
  clock_gettime64(CLOCK_MONOTONIC, {tv_sec=89616, tv_nsec=628287311}) = 0 <0.000042>
  clock_gettime64(CLOCK_MONOTONIC, {tv_sec=89616, tv_nsec=628442478}) = 0 <0.000038>
  clock_gettime64(CLOCK_MONOTONIC, {tv_sec=89616, tv_nsec=628586853}) = 0 <0.000039>
  clock_gettime64(CLOCK_MONOTONIC, {tv_sec=89616, tv_nsec=628774978}) = 0 <0.000037>
  clock_gettime64(CLOCK_MONOTONIC, {tv_sec=89616, tv_nsec=629033978}) = 0 <0.000039>
  clock_gettime64(CLOCK_MONOTONIC, {tv_sec=89616, tv_nsec=629318936}) = 0 <0.000041>
  clock_gettime64(CLOCK_MONOTONIC, {tv_sec=89616, tv_nsec=629491020}) = 0 <0.000039>

  这就是一个最简单的查看性能问题的流程，先统计，再看具体操作，然后查看代码，锁定问题。剩下的就是祝好运咯。当然如果发现输出比较少，还有个工具是ltrace