jffs2: No space for garbage collection. Aborting GC thread

[Olle Blomgren]

Hi, 
 
I'm running load tests on an embedded device using NAND flash and jffs2. After 3-4 weeks of running the system under heavy load, the following occasionally occurs:
1)
CRC calc of a node fails.
2)
The daemon experiencing the write failure aborts. This is by design, nothing strange.
3)
JFFS2 reports ENOSPC and claims that the garbage collection is broken and aborts.
 
My Linux kernel version is 4.4.27.

A system log follows:
 
kernel 04 kern.warning [4927999.241539] jffs2: Data CRC 961e0790 != calculated CRC 2e43ceec for node at 020a38e0
pacsd 0b user.err 11:20:57.793040 (0xa55800) Failed to clear table: Credential
kernel 06 kern.info [4927999.843684] potentially unexpected fatal signal 5.
kernel 04 kern.warning [4927999.843741] 
kernel 04 kern.warning [4927999.843741] Code: 8c658bd8  240210aa  0000000c <14e00003> 00000000  03e00008  00000000  7c03e83b  00602021 
kernel 04 kern.warning [4927999.843864] CPU: 0 PID: 984 Comm: GMain Tainted: G           O    4.4.27 #1
kernel 04 kern.warning [4927999.843911] task: 8d710248 ti: 8b15a000 task.ti: 8b15a000
kernel 04 kern.warning [4927999.843941] $ 0   : 00000000 10800108 00000000 76931880
kernel 04 kern.warning [4927999.843990] $ 4   : 000003d7 000003d8 00000005 00000000
kernel 04 kern.warning [4927999.844037] $ 8   : 7692995c 77698000 77696d24 f0000000
kernel 04 kern.warning [4927999.844085] $12   : 00000000 00000001 00000000 76929878
kernel 04 kern.warning [4927999.844134] $16   : 00000000 00000000 77116050 00000004
kernel 04 kern.warning [4927999.844180] $20   : 77116030 00000000 00000004 00000006
kernel 04 kern.warning [4927999.844230] $24   : 00000726 76fb4010                  
kernel 04 kern.warning [4927999.844278] $28   : 76fd1020 76929a20 00000002 77029170
kernel 04 kern.warning [4927999.844329] Hi    : 00000328
kernel 04 kern.warning [4927999.844359] Lo    : 00001deb
kernel 04 kern.warning [4927999.844400] epc   : 76fb4048 0x76fb4048
kernel 04 kern.warning [4927999.844439] ra    : 77029170 0x77029170
kernel 04 kern.warning [4927999.844465] Status: 01006313    USER EXL IE 
kernel 04 kern.warning [4927999.844508] Cause : 10800020 (ExcCode 08)
kernel 04 kern.warning [4927999.844541] PrId  : 01019550 (MIPS 34Kc)
kernel 04 kern.warning [4928006.637086] jffs2: Data CRC 961e0790 != calculated CRC 2e43ceec for node at 020a38e0
kernel 04 kern.warning [4928006.637144] jffs2: read_cache_page() returned error: -5
kernel 03 kern.err [4928006.637166] jffs2: Error garbage collecting node at 020a38e0!
kernel 05 kern.notice [4928006.637577] jffs2: No space for garbage collection. Aborting GC thread
kernel 04 kern.warning [4928006.964636] jffs2: Data CRC 961e0790 != calculated CRC 2e43ceec for node at 020a38e0
kernel 04 kern.warning [4928006.964694] jffs2: read_cache_page() returned error: -5
kernel 03 kern.err [4928006.964717] jffs2: Error garbage collecting node at 020a38e0!
kernel 04 kern.warning [4928006.967060] jffs2: Data CRC 961e0790 != calculated CRC 2e43ceec for node at 020a38e0
kernel 04 kern.warning [4928006.967109] jffs2: read_cache_page() returned error: -5
kernel 03 kern.err [4928006.967132] jffs2: Error garbage collecting node at 020a38e0!
                                                               :
                                              ...goes on forever...
 
Questions:
Has anyone seen this before?
Does anyone have any wild guesses as to under what circumstances JFFS2 would run out of memory?
 
Thanks!
/Olle