Re: mtd-utils io_paral test fails on NAND

[Otto Blom <oblom586@xxxxxxxxx>]

Hi Richard !

nandpagetest ran for 12h+ w/o failing, so it looks like a driver
problem like you suggested.

All the times I've seen if fail (around 10) the failed page came back
with all zeros.. Which is a bit strange as I would think "random" /
duplicated page would be more likely. Maybe the page read fails
silently and the controller returns all zeros

Thanks for you help & insights in this !

/Otto

On Mon, Nov 25, 2019 at 2:27 PM Richard Weinberger <richard@xxxxxx> wrote:
>
> Otto,
>
> ----- Ursprüngliche Mail -----
> > Von: "Otto Blom" <oblom586@xxxxxxxxx>
> > An: "Richard Weinberger" <richard.weinberger@xxxxxxxxx>
> > CC: "linux-mtd" <linux-mtd@xxxxxxxxxxxxxxxxxxx>
> > Gesendet: Montag, 25. November 2019 18:22:10
> > Betreff: Re: mtd-utils io_paral test fails on NAND
>
> > Hi Richard !
> >
> > Interestingly I had flash_readtest running in a loop the whole weekend
> > without any errors. Although this test does not verify the data read
> > so it could point towards a controller / driver problem. I'm in the
> > process of upgrading the SDK from 2017.3 to 2019.2 to see if that
> > makes any difference
>
> pagetest shoud also verify the data.
>
> If it fails only on top of UBI another possible error source is
> DMA. Both UBI and UBIFS use vmalloced() buffers, a driver must not DMA
> from/to these and use bounce buffers. If the driver is buggy in this
> area everything looks shiny until you start using UBI...
>
> Is the bad data always zeros? This also does not sound like an UBI
> LEB->PEB mismatch.
>
> Thanks,
> //richard

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