In our linux system, after we start our program with valgrind,it will be fail.
The reason is  instruction 0xEF9F0002(svc #0x9f0002) whichi is the sys_cacheflush for ARM Linux) Unrecognised. And I have been check the code instruction is really "svc #0x9f0002".The main log is as bellow:
(170918_11:50:46.092)[JVM][Pre Mem]CDCA VM allocated memory success, size = 18567168
(170918_11:50:46.092)disInstr(arm): unhandled instruction: 0xEF9F0002
(170918_11:50:46.092)                 cond=14(0xE) 27:20=249(0xF9) 4:4=0 3:0=2(0x2)
(170918_11:50:46.092)==132== valgrind: Unrecognised instruction at address 0x4ecbfb8.
(170918_11:50:46.092)==132==    at 0x4ECBFB8: CVMflushCache (flushcache_arch.S:65)
(170918_11:50:46.092)==132== Your program just tried to execute an instruction that Valgrind
(170918_11:50:46.092)==132== did not recognise.  There are two possible reasons for this.
(170918_11:50:46.092)==132== 1. Your program has a bug and erroneously jumped to a non-code
(170918_11:50:46.170)you just saw a
(170918_11:50:46.170)==132==    warning about a bad jump, it's probably your program's fault.
(170918_11:50:46.170)==132== 2. The instruction is legitimate but Valgrind doesn't handle it,
(170918_11:50:46.170)==132==    i.e. it's Valgrind's fault.  If you think this is the case or
(170918_11:50:46.170)==132==    you are not sure, please let us know and we'll try to fix it.
(170918_11:50:46.170)==132== Either way, Valgrind will now raise a SIGILL signal which will
(170918_11:50:46.170)==132== probably kill your program.

Looking forward to your help.