IPNC DMVA2 5MP --> ./system_server issue

Ivan,

It seems that filesystem is not mounted at /mnt/nand/ .

Try to mout it first.
Look at rcS /etc/init.d/rcS. There is a line :   mount -t jffs2 /dev/mtdblock4 /mnt/nand,  which mounts filesystem.

Regards,
Marko.

Thank you very much Marko. 

I have tried that and it confirms to me that it is mounted  "mount: according to mtab, /dev/mtdblock4 is mounted on /mnt/nand"

More over, i put this files in mnt/nand folder in my NFS machine and if i start using this NFS filesystem and then call "unmount /mnt/nand" the files will appear (as now nand is read from NFS) and system_server will run properly. I think that the problem is I made some nand_erase at some point and now I don´t know how to restore that files into nand.

:(

Hi Ivan,

I'm not sure about DMVA2, but models without video analitics (MT5),
recreate configuration files automatically, if they are not found at the /mnt/nand.
The line : "FileMng: Error Can't create system file",  seems to me that system_server can not recreate file at that location.

Regards,
Marko.

It is exactly that, when mnt/nand is mounted it won´t be able to create those files, now i don´t know why is it mounting it at startup if you can´t write there, maybe wrong kernel or filesistem not for DMVA2.....

Ivan,

If everything worked for you at first place, that reflash everything again, maybe you erased part of your  kernel or filesystem.
You can see your MTD partitions sizes in a console log, during kernel boot, and change partiotion sizes in:
Open_src_IPNC_DM36x_2.6.0\Application\src\lsp_ipnc_dm36x_2.6.0\ti-davinci\arch\arm\mach-davinci\board-dm368-ipnc.c (static struct mtd_partition nand_partitions[] = { )
Are you sure you are using right kernel, the one for evm has larger filesystem partition(512MB), so mtd4 falls out of 128RAM.

You can check mtd4  with :  df -h

Regards,
Marko.

I get:

Filesystem Size Used Available Use% Mounted on

/dev/mtdblock4 100.0M 100.0M 0 100% /mnt/nand

Could it be that there is a jffs2 image there and I deleted the files that were inside it? Do you know how could I reflash that 100M jffs2?

Thank you very very much!

Wow, I did "flash_eraseall -j /dev/mtd4" and I got everything "Skipping bad block at 0x........... ". Does it mean that NAND is broken? I guess not becouse kernel and filesys are being loaded properly from NAND....

I got the same issue with you. I think that come from broken nand. Bad block can make the cramfs file was not burned correctly or make some function relate to filesystem not work. I have just wondered that how can we step over it except replace a new nand memory. Please share to me if you got some solutions.

I did so using NFS filesystem and unmounting that NAND partition, when you can write there then the IPNC will boot properly.