数组结构铸造

Question

我有这三种结构,

typedef struct serial_header {
    int zigbeeMsgType;
    int seqNumber;
    int commandIdentifier;
    int dest;
    int src;
}serial_header_t;


typedef struct serial_packet {
  serial_header_t header;
  int data[];
} serial_packet_t;

最后一个是

typedef struct readAttributePacket
{
     int                     u8SourceEndPointId;
     int                     u8DestinationEndPointId;
     int                     u16ClusterId;
     int                     bDirectionIsServerToClient;
     int                     u8NumberOfAttributesInRequest;
     int                     bIsManufacturerSpecific;
     int                     u16ManufacturerCode;
     int                     pu16AttributeRequestList[];
}readAttributePacket_t;

我对此代码感到不安,我只想将驻留在serial_packet_t中的data []数组转换为readAttributePacket_t结构.我认为数据[]应该是

data[]={0x01,0x01,0x04,0x02,0x00,0x02,0x00,0x00,0x00,0x00,0x00,0x00,0x01};

我需要将这些数据转换为readAttributePacket_t结构.但是这下面的代码显示错误.

void main()
{
int a[]=  {0x32,0x00,0x31,0x69,0x69,0x00,0x00,0x01,0x01,0x04,0x02,0x00,0x02,0x00,0x00,0x00,0x00,0x00,0x00,0x01};
int i;
readAttributePacket_t *p;
serial_packet_t *data;

data = (serial_packet_t*)&a;

for(i=0;i<20;i++){
    printf(" %02x \n",a[i]);
}
p = (readAttributePacket_t *)&data->data;
printf("\nu8SourceEndPointId:%x \nu8DestinationEndPointId:%x \nu16ClusterId:%04x \nbDirectionIsServerToClient:%x \nu8NumberOfAttributesInRequest:%x \nbIsManufacturerSpecific:%x \nu16ManufacturerCode:%04x",p->u8SourceEndPointId,
    p->u8DestinationEndPointId,
    p->u16ClusterId,
    p->bDirectionIsServerToClient,
    p->u8NumberOfAttributesInRequest,
    p->bIsManufacturerSpecific,
    p->u16ManufacturerCode);


getch();
}

输出应该是这样的

u8SourceEndPointId=01
u8DestinationEndPointId=01
u16ClusterId=0402
bDirectionIsServerToClient=00
u8NumberOfAttributesInRequest=02
bIsManufacturerSpecific=00
u16ManufacturerCode=0000

我怎么能把pu16AttributeRequestList []数组放到readAttributePacket_t结构中,应该这样,

pu16AttributeRequestList[0]=0000
pu16AttributeRequestList[1]=0001

Answer 1

您不能只是将数组转换为结构,因为它们只是不兼容的类型.由于内存对齐约束,编译器需要在结构的字段之间插入填充,因此成员不会位于您可能期望的内存地址.解决方案:

1. 便携但手动较慢/较难(首选):手动将结构的字段复制到数组.
2. 写入较短但特定于GCC:使用__attribute__((packed))关键字使GCC不在struct字段之间引入填充.