Created attachment 190911 [details] journalctl dump When trying to load the nandsim driver sudo modprobe nandsim id_bytes=0xad,0xde,0x14,0xa7,0x42,0x4a System hangs and remains unresponsive. Snippet of interest from log nand: device found, Manufacturer ID: 0xad, Chip ID: 0xde nand: Hynix NAND 8GiB 3,3V 8-bit nand: 8192 MiB, MLC, erase size: 4096 KiB, page size: 16384, OOB size: 224 nand: No oob scheme defined for oobsize 224 kernel BUG at drivers/mtd/nand/nand_base.c:3988! invalid opcode: 0000 [#1] PREEMPT SMP My guess is that the NAND (0xad,0xde,0x14,0xa7,0x42,0x4a) used by NextThingCo (Allwinner R8M) currently isn't supported by nandsim. Cheers soderstrom
Tried this in a virtual machine with kernel 4.2.2-1-ARCH. Here the system does not hang. But the error is the same. Thus I think the priority can be lowered to normal. Going to try with same kernel version as before in a virtual machine to see if I can reproduce it then.
Running this in a virtual machine with the same kernel (4.2.2-1-ARCH) does not cause the system to hang. So I think the system hang is related to my hardware, going to try on my laptop tonight when I get home. For now I will start by looking at the nandsim code to see what is needed for patching in support for NextThingCo C.H.I.P NANDs (Allwinner R8M).
Did not find anything useful in regards to the problem at hand in my initial scan of the sources, going to put this on hold for now. Anyways, if someone else takes a look, some facts for you. # Kernel bootup messages [ 1.170000] nand: device found, Manufacturer ID: 0xad, Chip ID: 0xde [ 1.180000] nand: chip id data: 0xad, 0xde, 0x14, 0xa7, 0x42, 0x4a; 0xad, 0xde # mtdinfo - linux mtdinfo /dev/mtd0 Name: env Type: mlc-nand Eraseblock size: 4194304 bytes, 4.0 MiB Amount of eraseblocks: 1 (4194304 bytes, 4.0 MiB) Minimum input/output unit size: 16384 bytes Sub-page size: 16384 bytes OOB size: 1664 bytes Character device major/minor: 90:6 Bad blocks are allowed: true Device is writable: true # mtdinfo - uboot sunxi# mtdinfo Page Size 16384 b ( 0x 4000) OOB Size 1664 b ( 0x 680) Erase Size 2097154 b ( 0x 200002) subpagesize 16384 b ( 0x 4000) Options 0x40000000 bit options 0x 70000 ------------------------------ So these are the properties of the nand to simulate. Cheers soderstrom
There is some difference in reported spare area size: nandsim: 224 bytes mtdinfo: 1664 bytes and it looks like chip ID should be overridden. Could you please provide your exact NAND part number?
The nand I am trying to simulate is the one used in the Allwinner R8 SoC CHIP BOM: https://github.com/NextThingCo/CHIP-Hardware/raw/master/2015-08-11-CHIPv0_21-BOM.pdf Allwinner R8 Datasheet: https://github.com/NextThingCo/CHIP-Hardware/raw/master/CHIPv0_21-BOM-Datasheets/Allwinner%20R8%20Datasheet%20V1.2.pdf So I am unable to find the NAND part number, since it is not a standalone part. Or is there a way to check this from userspace?
Okay, here is the problem: Looks like R8M has NAND with unsupported coding of 4th ID byte. Currently mainline MTD can't decode it correctly (another Hynix NAND generation?) and we need to add this particular chip into nand_flash_ids table. As I can see from logs, your HW board runs custom kernel (https://linux-sunxi.org) and this commit https://github.com/bbrezillon/linux-sunxi/commit/8176c0934c125f076c03e6e95f726bbd781cc26a is exactly what we need here. So there are two ways to solve the issue: - integrate changes (see commit above) into mainline - add another ID decoding scheme for Hynix NANDs regarding to their generation Let me see what we can do for the 2nd case.