AM3517 - nand sub-page - ubifs

Hi,

some additionnal infos.

After executing following instructions:

I can re-attach the ubi device later. But if I create the volume or writes some data, with these additionnal instr after formating - attaching:

As work around, I can avoid usage of subpages with " -O 2048" for format / attach.

I would be fine to have some clear statement about sub-page support.

Regards

Hi,

Some additional info :

If you are using jffs2 file system , make sure it is created without -p option when using mkfs.jffs2 command.

Refer : http://processors.wiki.ti.com/index.php/AM35x-OMAP35x-PSP_04.02.00.07_UserGuide#Creating_JFFS2_filesystem

Also , make sure that the image size is upward aligned to NAND page size which is 2KB (i.e. 0x800). 

Regards,

Hrishikesh

Hi David,

As part of PSP testing we never tried UBIFS, but our Android team has validated UBIFS filesystem with PSP04.02.00.07 release. I cross-verified it again before responding to this thread, and it is working for them.

Can you please make sure that you are following steps mentioned in the below User Guide -

http://processors.wiki.ti.com/index.php/TI-Android-GingerBread-2.3.4-DevKit-2.1_UserGuide#Fastboot_and_UBI_rootfs

It may possible that, your mtd-utils is not in sync.

Thanks,

Vaibhav

Hi,

Well, UBIFS works ... but without sub-pages support.  My initial post is a confusing, sorry.

Checked the guide, and it seems that they skip the usage of sub-pages too.  MTD Test module "mtd_subpagetest", gives some errors when testing, so, make me believe that sub-page support is broken.

So the remaining question for me is  .... can I use sub-pages or not ?

Thanks

David

David,

Sorry for delayed response,

I have conformed sub-pages support with Android team here and as I mentioned before we have not validated sub-pages support. So this feature falls down to non-supported category for us as of now.

Looking at current list of items, I am quite not sure when can we start looking into this issue. I have taken this issue in my TODO list as of now. I would suggest continue using sub-page configured to actual NAND page size, as this is working for you.

Thanks,

Vaibhav

I ran into the same problem - NAND sub-page access is broken in 2.6.37. I applied patch http://arago-project.org/git/people/?p=sriram/ti-psp-omap.git;a=blobdiff;f=drivers/mtd/nand/omap2.c;h=b77f6445f0b788ac194358fffe613f7867c145c6;hp=e6f14cb69bcab68be9e5814eaa4438f73343411d;hb=1f62a9d1143cffcdc5dbf5b433fb905fa5f78831;hpb=de7109fd12c7b9d4080256bdfcd8893364a8decc, but this did not change anything for me.

Reproducing this misbehavior is easy:

- compile a 2.6.37 kernel with CONFIG_MTD_TESTS=m

- boot the kernel

- load the module mod_subpagetest.ko

You will see:

mtd_subpagetest: MTD device: 5

mtd_subpagetest: MTD device size 36700160, eraseblock size 131072, page size 2048, subpage size 512, count of eraseblocks 0, pages per eraseblock 0, OOB size 64

mtd_subpagetest: scanning for bad eraseblocks

mtd_subpagetest: scanned 280 eraseblocks, 0 are bad

mtd_subpagetest: erasing whole device

[...]

mtd_subpagetest: erased 280 eraseblocks

mtd_subpagetest: writing whole device

mtd_subpagetest: written up to eraseblock 0

mtd_subpagetest: written up to eraseblock 256

mtd_subpagetest: written 280 eraseblocks

mtd_subpagetest: verifying all eraseblocks

UNCORRECTED_ERROR default

ECC UNCORRECTED_ERROR B

mtd_subpagetest: error: read failed at 0x0

mtd_subpagetest: error -74 occurred

Sub-page access is a fundamental platform feature. Since this has been a known issue for months - is it going to be fixed at all? Or will there be at least a statement by TI that sub-page access is not supported in Linux?

Concerned,

Thilo

It has been a whole week without a response now. Is anyone at Texas Instruments even actually reading this?

Hi Thilo, could you pls re-post yr issue into the Linux forum, with the steps to create the issue from the base SDK - thanks.

Just to make sure - this is the link - http://e2e.ti.com/support/embedded/linux/default.aspx