问题描述
在push_back上调用erase或std::list时，我的某些代码随机崩溃，该代码处于长期运行服务中。我删除了不重要的部分，只保留了最相关的部分。
std::list<RxCCPReport> pendingSync;
size_t AnchorConnection::HandleCCPTxReport(const char *rxBytes,size_t length) {
if (!pendingSync.empty())
{
std::list<RxCCPReport>::iterator iter = pendingSync.begin();
while (iter != pendingSync.end())
{
if (condition1)
{
iter = pendingSync.erase(iter);
}
else
{
if (condition2)
{
++iter;
}
else if (condition3)
{
++iter;
}
else
{
iter = pendingSync.erase(iter);
}
}
}
}
return length; }
size_t AnchorConnection::HandleCCPRxReport(const char *rxBytes,size_t length) {
RxCCPReport rxCCP;
rxCCP.anc = anc;
rxCCP.rxTime = ccpRxTime;
rxCCP.seqNum = pCsMsg->seqNum;
pendingSync.push_back(rxCCP);
return length;
}
两个函数在同一线程上运行；代码有时会在pendingSync.erase(iter)或pendingSync.push_back(rxCCP)崩溃，每周大约1次。
回溯：
- 在擦除时崩溃了
#0 std::__detail::_List_node_base::_M_unhook (this=0x7fffcc024040) at ../../../../../libstdc++-v3/src/c++98/list.cc:142
__next_node = 0x0
__prev_node = 0x7fffcc024040
#1 0x00000000004db14a in std::list<RxCCPReport,std::allocator<RxCCPReport> >::_M_erase (this=0x7fffdc011970,__position=...) at /usr/include/c++/4.8.2/bits/stl_list.h:1570
__n = 0x7fffe5ffad80
#2 0x00000000004daddb in std::list<RxCCPReport,std::allocator<RxCCPReport> >::erase (this=0x7fffdc011970,__position=...) at /usr/include/c++/4.8.2/bits/list.tcc:112
__ret = <error reading variable __ret (Cannot access memory at address 0x10)>
#3 0x00000000004e3d8b in AnchorConnection::HandleCCPTxReport (this=0x7fffdc009240,rxBytes=0x7fffc4022abb "0rl\206%M/\004\205O\263\002\216\356\003\002f",length=12)
at LE/Network/rtls_anchor_conn.cpp:731
rxCCP = {rxTime = 3.1943677384753104,seqNum = 114 'r',anc = }
cSeqRxCCP = 114 'r'
cs = 0x7fffdc016430
ccpTx = 203157767788
ccpTxTime = 3.1794272052158457
pCsMsg = 0x7fffc4022abb
cSeqTxCCP = 114 'r'
- 再次在push_back崩溃
> #0 0x00007ffff6dff387 in __GI_raise (sig=sig@entry=6) at ../nptl/sysdeps/unix/sysv/linux/raise.c:55
resultvar = 0
pid = 5068
selftid = 5418
#1 0x00007ffff6e00a78 in __GI_abort () at abort.c:90
save_stage = 2
act = {__sigaction_handler = {sa_handler = 0x7ffff6ddb76d,sa_sigaction = 0x7ffff6ddb76d},sa_mask = {__val = {5,140737336646961,3,140737043728478,2,140737336643303,1,140737336652017,140737043728452,12,140737336652021,140737043729264,140737043731024}},sa_flags = 16,sa_restorer = 0x7fffe57f8d70}
sigs = {__val = {32,0 <repeats 15 times>}}
#2 0x00007ffff6e41ed7 in __libc_message (do_abort=do_abort@entry=2,fmt=fmt@entry=0x7ffff6f54350 "*** Error in `%s': %s: 0x%s ***\n")
at ../sysdeps/unix/sysv/linux/libc_fatal.c:196
ap = {{gp_offset = 40,fp_offset = 48,overflow_arg_area = 0x7fffe57f9260,reg_save_area = 0x7fffe57f9170}}
ap_copy = {{gp_offset = 16,reg_save_area = 0x7fffe57f9170}}
fd = 26
on_2 = <optimized out>
list = <optimized out>
nlist = <optimized out>
cp = <optimized out>
written = <optimized out>
#3 0x00007ffff6e483e4 in malloc_printerr (action=<optimized out>,str=0x7ffff6f54548 "malloc(): memory corruption (fast)",ptr=<optimized out>,ar_ptr=<optimized out>)
at malloc.c:4967
buf = "00007fffd8001570"
cp = <optimized out>
ar_ptr = <optimized out>
ptr = <optimized out>
str = 0x7ffff6f54548 "malloc(): memory corruption (fast)"
action = <optimized out>
#4 0x00007ffff6e4bb20 in _int_malloc (av=av@entry=0x7fffd0000020,bytes=bytes@entry=40) at malloc.c:3383
p = 0x7fffd8001560
fb = <optimized out>
pp = <optimized out>
nb = 48
idx = <optimized out>
bin = <optimized out>
victim = 0x7fffd8001560
size = <optimized out>
victim_index = <optimized out>
remainder = <optimized out>
remainder_size = <optimized out>
block = <optimized out>
bit = <optimized out>
map = <optimized out>
fwd = <optimized out>
bck = <optimized out>
errstr = <optimized out>
#5 0x00007ffff6e4e6fc in __GI___libc_malloc (bytes=40) at malloc.c:2905
ar_ptr = 0x7fffd0000020
victim = <optimized out>
hook = <optimized out>
#6 0x00007ffff7929ecd in operator new (sz=40) at ../../../../libstdc++-v3/libsupc++/new_op.cc:51
p = <optimized out>
#7 0x00000000004dc6f8 in __gnu_cxx::new_allocator<std::_List_node<RxCCPReport> >::allocate (this=0x7fffdc011930,__n=1) at /usr/include/c++/4.8.2/ext/new_allocator.h:104
No locals.
#8 0x00000000004dc006 in std::_List_base<RxCCPReport,std::allocator<RxCCPReport> >::_M_get_node (this=0x7fffdc011930) at /usr/include/c++/4.8.2/bits/stl_list.h:334
No locals.
#9 0x00000000004db699 in std::list<RxCCPReport,std::allocator<RxCCPReport> >::_M_create_node<RxCCPReport const&> (this=0x7fffdc011930)
at /usr/include/c++/4.8.2/bits/stl_list.h:502
__p = 0x0
#10 0x00000000004db1c2 in std::list<RxCCPReport,std::allocator<RxCCPReport> >::_M_insert<RxCCPReport const&> (this=0x7fffdc011930,__position=...)
at /usr/include/c++/4.8.2/bits/stl_list.h:1561
__tmp = 0x0
#11 0x00000000004dae68 in std::list<RxCCPReport,std::allocator<RxCCPReport> >::push_back (this=0x7fffdc011930,__x=...) at /usr/include/c++/4.8.2/bits/stl_list.h:1016
No locals.
#12 0x00000000004e475b in AnchorConnection::HandleCCPRxReport (this=0x7fffdc01be90,rxBytes=0x7fffd004d7fb "1{\212\211D\021\006\061+\b\250\256i\227@ԺF(ú\002\005\003o\200@",length=88) at LE/Network/rtls_anchor_conn.cpp:936
rxCCP = {rxTime = 4.3416058279497198,seqNum = 123 '{',anc = 0x7fffdc016320}
prevSeqNum = 122 'z'
master = 0x7fffdc0101a0
cs = 0x7fffdc016450
masterID = 588618078431250826
logNum = 45794088
pCsMsg = 0x7fffe57f9500
ccpRx = 277418192552
ccpRxTime = 4.3416058279497198
txTime = "0000000000\000\334\377\177\000\000\177\236",<incomplete sequence \345>
temp = "\000\000\000\000"
diagMsg2 = "0F000005036F8040001000000412C8A607120000086000BC07390AD806130000040000FC0114000005F9FF0700731500000EA8AE699740D4BA740800EC69974025000010A604C3FB4C07A3F7520743FE260538072E100002EF022700280491FE1F6D2E100"...
csmsg = {type = 49 '1',masterID = "\212\211D\021\006\061+\b",csRxTime = "\250\256i\227@",{firstPath16 = 47828,firstPath = "Ժ"},extLen = 70 'F',{
ext = "(ú\002\005\003o\200@\000\022Ȧ\a`\000\274\a9\n\330\006\000\000\374\001\371\377\a\000s\250\256i\227@Ժt\b\000\354i\227@\000\246\004\303\373L\a\243\367R\aC\376&\005\070\a\357\002\221\376\037m?\021",{{logNum32 = 45794088,logNum = "(ú\002"},diagnostics = "\005\003o\200@\000\022Ȧ\a`\000\274\a9\n\330\006\000\000\374\001\371\377\a\000s\250\256i\227@Ժt\b\000\354i\227@\000\246\004\303\373L\a\243\367R\aC\376&\005\070\a\357\002\221\376\037m?\021"}}}
FP = 747
rxTime = "409769AEA8\000\345\377\177\000\000Н",<incomplete sequence \345>
我快疯了，任何建议都赞赏
通过套接字接收数据的代码如下：
#define FRAME_HEADER_LEN (6)
#define FRAME_DATA_IDX (3)
virtual void Run()
{
unsigned int length = 0;
std::string rxBytes;
bool msgProcess;
uint8_t fcode;
unsigned int flen = 0;
while (shutdown_select_anc(socket) != -1)
{
int count = socket->ReceiveData();
if (count <= 0) // Something failed. Socket is probably closed.
{
break;
}
length += count;
rxBytes.append(socket->buf_,count);
/*
* the data coming from the anchor is framed:
*<STX><LENlsb><LENmsb><DATA:<FC><XX>.....><CRClsb><CRCmsb><ETX>
* STX = 0x2
* LEN is the length of data message(16 bits)
* CRC is the 16 - bit CRC of the data bytes
* ETX = 0x3
* FC = is the function code (API code)
*/
if (length <= FRAME_HEADER_LEN) //minimum length is 7 bytes (header + 1 byte function code)
{
continue;
}
do {
msgProcess = false;
crc_err_t pckt_crc = CRC_ERROR;
//search through the received data to find the start of frame
while (length > FRAME_HEADER_LEN)
{
if ((rxBytes.c_str())[0] == 0x2) //this is the start - check if length and end match
{
uint8_t hi = (rxBytes.c_str())[2];
uint8_t lo = (rxBytes.c_str())[1];
flen = lo + ((uint16_t)hi << 8);
if ((flen + FRAME_HEADER_LEN) > length) //we don't have all the frame yet
{
break;
}
if ((rxBytes.c_str())[flen + 5] == 0x3) //end of frame present
{
//received whole frame: check frame checksum here
pckt_crc = check_crc16((uint8_t *)&(rxBytes.c_str())[3],(flen + 2));
break;
}
else //no end of frame in the expected place - start byte was not real start
{
rxBytes.erase(0,1);
length -= 1;
}
}
else //remove the byte and check if next is a start of frame
{
rxBytes.erase(0,1);
length -= 1;
}
}
//check if we have a frame
if (((flen + FRAME_HEADER_LEN) > length) || (length <= FRAME_HEADER_LEN))
{
break;
}
if (pckt_crc != CRC_OKAY)
{
//frame packet fully received but it's CRC bad
DBG_PRINTF(DBG_DEFAULT,"ACHTUNG ACHTUNG : incompatible software without CRC or network CRC error\n");
//let incompatible software proceed at the moment
}
fcode = (rxBytes.c_str())[FRAME_DATA_IDX]; //get function code
typename std::map<uint8_t,std::pair<size_t,handler> >::iterator it = handlers.find(fcode);
if (it != handlers.end()) //we have found a handler for this function code
{
if (flen >= it->second.first)
{
handler f = it->second.second;
size_t n = (static_cast<Child*>(this)->*f)(&((rxBytes.c_str())[FRAME_DATA_IDX]),flen );
if (n > 0)
{
msgProcess = true;
rxBytes.erase(0,n + FRAME_HEADER_LEN);
length -= (unsigned int)(n + FRAME_HEADER_LEN);
}
else
{
//Unexpected Message
rxBytes.clear();
length = 0;
}
}
}
else
{
rxBytes.clear();
length = 0;
}
} while((length > 0) && msgProcess);
}
socket->Close();
}
解决方法
在push_back情况下，
str = 0x7ffff6f54548“ malloc（）：内存损坏（快速）”
建议您在其他地方存在内存损坏，很可能是在致电HandleCCPRxReport之前不久。
依赖报错 idea导入项目后依赖报错，解决方案：https://blog....
错误1：gradle项目控制台输出为乱码 # 解决方案：https://bl...