AnsweredAssumed Answered

eMMC file system corruption and data loss issues

Question asked by Chen Kun on Dec 20, 2019
Latest reply on Dec 23, 2019 by Yiping Wang


The processor we use is P2020,file system and data stored in eMMC. Now we have encountered a serious problem:File system and data stored on eMMC are often corrupted.

The specific performance is as follows: we have database files (sqlite) and other R/W files stored on the emmc partition, and the superblock damage occurs during use, causing the partition mount failure. Sometimes it can be repaired with e2fsck, sometimes it can only be re-partitioned to be used again.
When the superblock is damaged, it can be successfully mounted after using the reboot/reset system, but using e2fsck, you can see that the partition table information (superblock / bitmap, etc) is abnormal.
Our kernel version is Linux-3.8.13. When the kernel is updated to Linux-4.14.140, the system is stuck and the emmc device cannot be found during the test. You can only repartition after uboot erases.
When power loss during emmc writing, the superblock damage phenomenon is easier to reproduce. We have enabled the power loss protection mechanism inside emmc, but the problem has not improved.
Do you have encountered similar problems and have corresponding solutions?
Looking forward to your reply.