IO working abnormally under small probability

[zhaozq]

   I use  MK10DX128F on my products for more than 5  years. But every year in recent years ，about 0.01% chip  individual GPIO input work abnormally.

Details as following:
1. The abnormal IO Input low level Voltage(0V), my software shows high level.（abnormal），at the same time other IO may be normal.
2. When the IO input is high(3.3V), my software shows a high level.（uncertain）
3. When the low level is input again, the software displays a low level and returns to normal.（normal）

     When abnormal, IO will return to normal when changing level. I was not sure when and which chip，which IO would work abnormally, but the problem persisted. I checked my software for no abnormalities.When GPIO work abnormally, register cannot be read by the emulator due to mechanical structure limitation.   

     Please help me analyze the above question,and is IO chip internal has Schmitt trigger？

     Thanks very much.

[vicentegomez]

Would be possible that you tell me more about how are you testing the device?

can you please try with a simple test code that only set the gpio as input and check what happens, on a failure part can you check if there is a low impedance on any failed pin? 

just for debugging purpose, what happed if you make a swap test, I mean put the failed device on a known good board and put the known good device on the failed board