Failure on MTD sub-page test

[Danesh Daroui]

Hi Artem,

I ran the tests this time without mouting the file system and now all tests except OOB has been passed. The result for OOB test is:

mtd_oobtest: MTD device: 0
mtd_oobtest: MTD device size 1073741824, eraseblock size 524288, page size 4096, count of eraseblocks 2048, pages per eraseblock 128, OOB size 128
mtd_oobtest: scanning for bad eraseblocks
mtd_oobtest: scanned 2048 eraseblocks, 0 are bad
mtd_oobtest: test 1 of 5
mtd_oobtest: erasing whole device
mtd_oobtest: erased 2048 eraseblocks
mtd_oobtest: writing OOBs of whole device
mtd_oobtest: written up to eraseblock 0
mtd_oobtest: written up to eraseblock 256
mtd_oobtest: written up to eraseblock 512
mtd_oobtest: written up to eraseblock 768
mtd_oobtest: written up to eraseblock 1024
mtd_oobtest: written up to eraseblock 1280
mtd_oobtest: written up to eraseblock 1536
mtd_oobtest: written up to eraseblock 1792
mtd_oobtest: written 2048 eraseblocks
mtd_oobtest: verifying all eraseblocks
mtd_oobtest: verified up to eraseblock 0
mtd_oobtest: verified up to eraseblock 256
mtd_oobtest: verified up to eraseblock 512
mtd_oobtest: verified up to eraseblock 768
mtd_oobtest: verified up to eraseblock 1024
mtd_oobtest: verified up to eraseblock 1280
mtd_oobtest: verified up to eraseblock 1536
mtd_oobtest: verified up to eraseblock 1792
mtd_oobtest: verified 2048 eraseblocks
mtd_oobtest: test 2 of 5
mtd_oobtest: erasing whole device
mtd_oobtest: erased 2048 eraseblocks
mtd_oobtest: writing OOBs of whole device
mtd_oobtest: written up to eraseblock 0
mtd_oobtest: written up to eraseblock 256
mtd_oobtest: written up to eraseblock 512
mtd_oobtest: written up to eraseblock 768
mtd_oobtest: written up to eraseblock 1024
mtd_oobtest: written up to eraseblock 1280
mtd_oobtest: written up to eraseblock 1536
mtd_oobtest: written up to eraseblock 1792
mtd_oobtest: written 2048 eraseblocks
mtd_oobtest: verifying all eraseblocks
mtd_oobtest: verified up to eraseblock 0
mtd_oobtest: error: verify failed at 0x5300000
mtd_oobtest: verified up to eraseblock 256
mtd_oobtest: verified up to eraseblock 512
mtd_oobtest: verified up to eraseblock 768
mtd_oobtest: verified up to eraseblock 1024
mtd_oobtest: verified up to eraseblock 1280
mtd_oobtest: verified up to eraseblock 1536
mtd_oobtest: verified up to eraseblock 1792
mtd_oobtest: verified 2048 eraseblocks
mtd_oobtest: test 3 of 5
mtd_oobtest: erasing whole device
mtd_oobtest: erased 2048 eraseblocks
mtd_oobtest: writing OOBs of whole device
mtd_oobtest: written up to eraseblock 0
mtd_oobtest: written up to eraseblock 256
mtd_oobtest: written up to eraseblock 512
mtd_oobtest: written up to eraseblock 768
mtd_oobtest: written up to eraseblock 1024
mtd_oobtest: written up to eraseblock 1280
mtd_oobtest: written up to eraseblock 1536
mtd_oobtest: written up to eraseblock 1792
mtd_oobtest: written 2048 eraseblocks
mtd_oobtest: verifying all eraseblocks
mtd_oobtest: verified up to eraseblock 0
mtd_oobtest: verified up to eraseblock 256
mtd_oobtest: verified up to eraseblock 512
mtd_oobtest: verified up to eraseblock 768
mtd_oobtest: verified up to eraseblock 1024
mtd_oobtest: verified up to eraseblock 1280
mtd_oobtest: verified up to eraseblock 1536
mtd_oobtest: verified up to eraseblock 1792
mtd_oobtest: verified 2048 eraseblocks
mtd_oobtest: test 4 of 5
mtd_oobtest: erasing whole device
mtd_oobtest: erased 2048 eraseblocks
mtd_oobtest: attempting to start write past end of OOB
mtd_oobtest: an error is expected...
mtd_oobtest: error occurred as expected
mtd_oobtest: attempting to start read past end of OOB
mtd_oobtest: an error is expected...
mtd_oobtest: error occurred as expected
mtd_oobtest: attempting to write past end of device
mtd_oobtest: an error is expected...
mtd_oobtest: error occurred as expected
mtd_oobtest: attempting to read past end of device
mtd_oobtest: an error is expected...
mtd_oobtest: error occurred as expected
mtd_oobtest: attempting to write past end of device
mtd_oobtest: an error is expected...
mtd_oobtest: error occurred as expected
mtd_oobtest: attempting to read past end of device
mtd_oobtest: an error is expected...
mtd_oobtest: error occurred as expected
mtd_oobtest: test 5 of 5
mtd_oobtest: erasing whole device
mtd_oobtest: erased 2048 eraseblocks
mtd_oobtest: writing OOBs of whole device
mtd_oobtest: written up to eraseblock 0
mtd_oobtest: written up to eraseblock 0
mtd_oobtest: written up to eraseblock 256
mtd_oobtest: written up to eraseblock 256
mtd_oobtest: written up to eraseblock 512
mtd_oobtest: written up to eraseblock 512
mtd_oobtest: written up to eraseblock 768
mtd_oobtest: written up to eraseblock 768
mtd_oobtest: written up to eraseblock 1024
mtd_oobtest: written up to eraseblock 1024
mtd_oobtest: written up to eraseblock 1280
mtd_oobtest: written up to eraseblock 1280
mtd_oobtest: written up to eraseblock 1536
mtd_oobtest: written up to eraseblock 1536
mtd_oobtest: written up to eraseblock 1792
mtd_oobtest: written up to eraseblock 1792
mtd_oobtest: written 2047 eraseblocks
mtd_oobtest: verifying all eraseblocks
mtd_oobtest: verified up to eraseblock 0
mtd_oobtest: verified up to eraseblock 256
mtd_oobtest: verified up to eraseblock 512
mtd_oobtest: verified up to eraseblock 768
mtd_oobtest: verified up to eraseblock 1024
mtd_oobtest: verified up to eraseblock 1280
mtd_oobtest: verified up to eraseblock 1536
mtd_oobtest: verified up to eraseblock 1792
mtd_oobtest: verified 2047 eraseblocks
mtd_oobtest: finished with 1 errors

The interesting thing is that the file system is forced to switch to read-only mode at startup due to checksum error and some corrupted data that cannot be verified. Now my question is, can failing OOB test explain this failure? I mean, if OOB test fails (like our case) then it is very likely that the device experience data integrity problem at start up since OOB or Out Of Band area contains meta data info for the file system and is the first part that is read and verified during start up so if it fails, it can force the file system into read-only mode to prevent further data corruption. Am I right?

Regards,

Danesh Daroui


-----Original Message-----
From: linux-mtd [mailto:linux-mtd-bounces at lists.infradead.org] On Behalf Of Danesh Daroui
Sent: den 14 oktober 2016 16:14
To: dedekind1 at gmail.com; linux-mtd at lists.infradead.org
Subject: RE: Failure on MTD sub-page test

Hi Artem,

Thanks for your reply. I can unmount the device and run the tests, but then how can we make sure whether the problem is in UBI/UBIFS layers or only MTD layer? Anyway, if we update the kernel, then both MTD drivers and UBI/UBIFS will be updated, am I right?

Regards,

Danesh Daroui

 
-----Original Message-----
From: Artem Bityutskiy [mailto:dedekind1 at gmail.com]
Sent: den 14 oktober 2016 14:41
To: Danesh Daroui <Danesh.Daroui at ascom.com>; linux-mtd at lists.infradead.org
Subject: Re: Failure on MTD sub-page test

On Tue, 2016-10-11 at 11:11 +0000, Danesh Daroui wrote:
> Hello,
> 
> We are using UBIFS which is shipped with the Kernel 2.6.39.4 on a NAND 
> Flash memory. When I run "mtd_debug info /dev/mtd0" following info are
> shown:
> 
> mtd.type = MTD_NANDFLASH
> mtd.flags = MTD_CAP_NANDFLASH
> mtd.size = 1073741824 (1G)
> mtd.erasesize = 524288 (512K)
> mtd.writesize = 4096 (4K)
> mtd.oobsize = 128
> regions = 0
> 
> Now the problem is that when I run mtd tests, only read and NAND ECC 
> tests are passed. Then all other tests fail on exactly page 21. For 
> instance this is the output when I run page test:

If mtd tests do not work, you need to focus on your MTD drivers and fix them.

> =================================================
> mtd_subpagetest: MTD device: 0
> mtd_subpagetest: MTD device size 1073741824, eraseblock size 524288, 
> page size 4096, subpage size 1024, count of eraseblocks 2048, pages
... snip ...
> mtd_subpagetest: verified up to eraseblock 1280
> mtd_subpagetest: verified up to eraseblock 1536
> mtd_subpagetest: verified up to eraseblock 1792 UBIFS error (pid
> 4243): ubifs_read_node: bad node type (255 but expected 1) UBIFS error 
> (pid 4243): ubifs_read_node: bad node at LEB 30:98088, LEB mapping 
> status 1 UBIFS error (pid 4243): do_readpage: cannot read page 21 of 
> inode 791, error -22

This is a bit confusing - why messages go together. Do you run mtd tests and also have the same mtd device mounted at the same time?

The tests hammer the MTD device directly, they do not go through UBI/UBIFS. So make sure UBI/UBIFS are not using the MTD device that you are testing.

Artem.

______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/