最近碰到的一个关于memcpy的奇葩问题

最近写代码，碰到一个奇葩问题，memcpy函数用起来，编译居然提示我stackoverflow，简直是对写C的码农的最大羞辱，WTF！

UINT8 numBuffers = 0;
UINT16 cpLength = 0;
//  TPM2B_DIGEST *bufferList;
if(length%(MAX_DIGEST_BUFFER) != 0)
numBuffers = length/(MAX_DIGEST_BUFFER) + 1;
else
numBuffers = length/(MAX_DIGEST_BUFFER);

TPM2B_DIGEST *bufferList[numBuffers];
for(UINT8 i = 0; i < numBuffers; i++)
{
(bufferList)[i] = (TPM2B_DIGEST *)calloc(1,sizeof(TPM2B_DIGEST));
if(i < numBuffers-1)
{
//          memcpy(bufferList[i]->t.buffer, &buffer[cpLength], MAX_DIGEST_BUFFER);
for( UINT16 m = 0; m < MAX_DIGEST_BUFFER; m++)
{
bufferList[i]->t.buffer[m] = buffer[m + cpLength];
}
cpLength = i * MAX_DIGEST_BUFFER;
}
if(i == numBuffers-1 )
{
for(UINT16 j= 0; j < (length-cpLength); j++)
{
bufferList[i]->t.buffer[j] = buffer[cpLength + j];
}
}

}

就是上面注释掉的

memcpy(bufferList[i]->t.buffer, &buffer[cpLength], MAX_DIGEST_BUFFER);

编译的时候，结果如下：

In file included from /usr/include/string.h:639:0,

                 from tpm2_sign.cpp:61:

In function ‘void* memcpy(void*, const void*, size_t)’,

    inlined from ‘int computeDataHash(BYTE*, UINT16, TPMI_ALG_HASH, TPM2B_DIGEST*)’ at tpm2_sign.cpp:669:73:

/usr/include/x86_64-linux-gnu/bits/string3.h:53:71: error: call to void* __builtin___memcpy_chk(void*, const void*, long unsigned int, long unsigned int) will always overflow destination buffer [-Werror]

   return __builtin___memcpy_chk (__dest, __src, __len, __bos0 (__dest));

自查半天，没觉得有什么错，网上查了半天，居然查到一个“GCC Bugzilla – Bug 37060”，翻到最后跟我说是“fixed in GCC 4.7.0”

链接在此   GCC Bugzilla -Bug 37060

我gcc -v 查自己的gcc版本， 妈蛋， 都4.9.2了啊，我勒个去。

只好折衷，选个别的办法解决问题，一位位的复制吧，不用memcpy了。

for( UINT16 m = 0; m < MAX_DIGEST_BUFFER; m++)
{
bufferList[i]->t.buffer[m] = buffer[m + cpLength];
}

尼玛，现在还没弄明白，到底是个神马情况，有时间好好扒一扒这家伙的老底，奶奶个腿儿的，先记下来！