UBIFS corruptions on NOR-Flash

Roshni Shah roshni.shah at timesys.com
Wed Sep 23 11:28:37 PDT 2015


Hi Han,

Thank you very much for looking into this.

I do not have CONFIG_CMA enabled. I did not know it was required. I
can try a test with that enabled.

I tried the older version of the fsl_quadspi driver yesterday and with
that, I was able to mount the UBIFS.

http://git.freescale.com/git/cgit.cgi/imx/linux-2.6-imx.git/tree/drivers/mtd/spi-nor/fsl-quadspi.c?h=imx_3.14.28_1.0.0_ga&id=7a998cb82c4e6862eff95ff75d253717e8f5598d

Is the CONFIG_CMA requirement applicable to later versions of the driver?

Thanks,
Roshni

On Wed, Sep 23, 2015 at 2:21 PM, Han Xu <xhnjupt at gmail.com> wrote:
> Have you enabled the CONFIG_CMA in kernel setting?
>
> On Fri, Sep 18, 2015 at 4:50 PM, Richard Weinberger <richard at nod.at> wrote:
>> Am 18.09.2015 um 23:36 schrieb Roshni Shah:
>>> Thanks. I tried 4.1 Kernel ran UBI tests and mounting the UBIFS on
>>> Vybrid TWR. I see a different error, however corruptions & UBI test
>>> failures still persist.
>>>
>>> Flash chip: s25fl128s (same as above).
>>>
>>> # uname -r
>>> 4.1.5
>>>
>>> Following are the results:
>>
>> [...]
>>
>>> ubi0 warning: ubi_eba_read_leb: CRC error: calculated 0x8f7714b9, must
>>> be 0x31a3fc87
>>> ubi0 warning: vol_cdev_write: volume 0 on UBI device 0 is corrupted
>>> ubi0 warning: ubi_eba_read_leb: CRC error: calculated 0xd1882123, must
>>
>> So, reading data using UBI does not work as expected.
>> One thing MTD drivers sometimes get wrong is that UBI is using vmalloc'ed buffers
>> and drivers cannot deal with that.
>>
>> Adding Han Xu, maybe he has an idea.
>>
>> Thanks,
>> //richard
>>
>> ______________________________________________________
>> Linux MTD discussion mailing list
>> http://lists.infradead.org/mailman/listinfo/linux-mtd/



More information about the linux-mtd mailing list