TI中文支持网Part Number:CC2652POther Parts Discussed in Thread:Z-STACK
- I have 30 end devices power on at the same time and request to join the same network, given the air congestion, it's acceptable that some of them fail to jion. So the devices join again 3 seconds after failure until successful join.
- the problem is that some devices will never join successfully, even if I turn off most devices to reduce congestion. From the sniff log, I know the end device send a association request and get a response from the coordinator which didn't send a network key to end devices later. Also, if I reboot the coordinator, the device can successfully join the network. So I think the problem is on the coordinator side.
- An additional point is that it's ok to join if the 30 end devices join one by one, not at the same time. Has anyone encountered the same problem? Please help analyze it, thanks.
Alex Zhang：
您好，已经跟进您的问题
,
Alex Zhang：
您正在评估什么 SDK？您能否提供嗅探器日志并识别一些未加入的设备？您可以尝试增加 STACK_TASK_STACK_SIZE 和 APP_TASK_STACK_SIZE 来补偿此流量，或监视堆是否溢出，但最终 TI 不会针对此类使用情况测试 Z-Stack 解决方案，并建议您错开设备加入。如果您可以检测到问题已发生，您还可以从其应用程序代码触发协调器的重置。
（What SDK are you evaluating? Can you provide the sniffer log and identify some devices which do not join? You can try to increase the STACK_TASK_STACK_SIZE and APP_TASK_STACK_SIZE to compensate for this traffic, or monitor the heap for overflow, but ultimately TI does not test the Z-Stack solution against this type of usage and recommends that you stagger device joining. You can also trigger a reset of the coordinator from its application code if you can detect that the issue has occurred.）
,
sinjin guo：
hi，Alex感谢回复，经过进一步测试，发现部分硬件的射频信号强度有异常，现在需要等这个问题解决以后再重新测试验证，到时我再追加回复。
,
Alex Zhang：
好的，期待您的回复
,
sinjin guo：
hi, alex，经查是协议栈zd_app.c的函数有问题。
uint32_t ZDApp_ProcessSecEvent( uint8_t task_id, uint32_t events ){ if ( events & ZDO_NEW_DEVICE ) {…………………….
while (pNewDevice ) {pNewDevice->timeDelta -= timeDelta; //此处可能出现较小数减去较大数的情况！！！！！！！！！ pNewDevice = pNewDevice->next;
}……………………………………. }
函数ZDApp_ProcessSecEvent()里头while循环，相数相减可能出现异常（如短时间内大量设备加网的情况），导致pNewDevice->timeDelta等于接近于65535的值，单位为ms，约等于65秒的时间。
此数值又会影响函数ZDO_JoinIndicationCB() 的变量 timeToFire，导致65秒的时间内，向子设备发送association response后，一直不发送network key给子设备，一直得等65秒timeout后才恢复正常。
ZStatus_t ZDO_JoinIndicationCB(uint16_t ShortAddress, uint8_t *ExtendedAddress, uint8_t CapabilityFlags, uint8_t type)
{
……………………………
timeToFire = OsalPortTimers_getTimerTimeout( ZDAppTaskID, ZDO_NEW_DEVICE );
pNewDevice->timeDelta = ZDAPP_NEW_DEVICE_TIME – timeToFire;
// Start the timer only if there is no pending timer if ( pNewDevice->timeDelta == ZDAPP_NEW_DEVICE_TIME ) { OsalPortTimers_startTimer( ZDAppTaskID, ZDO_NEW_DEVICE, ZDAPP_NEW_DEVICE_TIME ); }
}
以上，请你们的工程师研究研究再修复。
