Re: WD MyCloud Ex, Armada-370, pxa3xx-nand, marvell-nfc

From: Miquel Raynal <miquel.raynal@bootlin.com>
To: David Schaefer <david.tillmann.schaefer@gmail.com>
Cc: Richard Weinberger <richard@nod.at>,
	linux-mtd <linux-mtd@lists.infradead.org>
Subject: Re: WD MyCloud Ex, Armada-370, pxa3xx-nand, marvell-nfc
Date: Wed, 16 Dec 2020 08:51:35 +0100	[thread overview]
Message-ID: <20201216085135.367f1f61@xps13> (raw)
In-Reply-To: <CAK+FTGgZV2CnTBT6i+1G=fLzpNcCBFhY04a_hfSeV_kmoDcaSA@mail.gmail.com>

Hi David,

David Schaefer <david.tillmann.schaefer@gmail.com> wrote on Tue, 15 Dec
2020 23:04:24 +0100:

> Hi Miquel,
> 
> Am Di., 15. Dez. 2020 um 22:20 Uhr schrieb David Schaefer
> <david.tillmann.schaefer@gmail.com>:
> >
> > Hi Miquèl
> >  
> > > Maybe I should ask you why you are using mtdblock and using dd on top
> > > of it? If you want to tackle a NAND issue, you should probably  address
> > > the mtd device directly and use the tools provided by the mtd-utils
> > > userspace test suite (nandpagetest seems a good fit).  
> > What I am trying to accomplish is running a debian with a kernel more
> > than 4.15 on a WD MyCloud Ex2. I follow a written down approach that
> > has been led to success by others. It involves
> > dd if=/dev/zero of=/dev/mtdblock1
> > dd if=uImage of=/dev/mtdblock1
> > dd if=/dev/zero of=/dev/mtdblock2
> > dd if=uRamdisk of=/dev/mtdblock2
> >
> > I could use more appropriate ways to deal with these devices if I knew of any.
> >  
> > > You might want to try a much more recent kernel then. Probably that
> > > using a 5.9 or 5.10 is a good target for now.  
> > I am compiling 5.10 right now and am looking forward to the results. I
> > have had 5.4 running already with no mtdblock devices showing up.
> >  
> > > I really don't think using -X theirs blindly is wise, you should
> > > definitely try to understand the conflict and merge it by hand.  
> > I tried but didn't feel to be in the position to make the call.
> > Anyhow, since I have seen the commits are all part of kernels 4.17,
> > 4.20, 5.0, and 5.10 this will not be an issue now when I run 5.10.
> >  
> > > > - marvell-nfc f10d0000.nand: Timeout waiting for RB signal  
> > > This one is a real error but got fixed by another commit.  
> > > > - print_req_error: I/O error, dev mtdblock6, sector 0
> > > > print_req_error: I/O error, dev mtdblock6, sector 0
> > > > print_req_error: I/O error, dev mtdblock6, sector 0
> > > > print_req_error: I/O error, dev mtdblock5, sector 0
> > > > print_req_error: I/O error, dev mtdblock5, sector 0
> > > > print_req_error: I/O error, dev mtdblock5, sector 0
> > > > print_req_error: I/O error, dev mtdblock4, sector 0
> > > > print_req_error: I/O error, dev mtdblock4, sector 0
> > > > print_req_error: I/O error, dev mtdblock4, sector 0
> > > > print_req_error: I/O error, dev mtdblock3, sector 0  
> > > This one seem related but I don't know from where it comes.  
> > > > dd: writing '/dev/mtdblock1': Input/output error  
> > I will check if I see these issues with 5.10 when it's built.  
> 
> running 5.10 does not show any /dev/mtdblock devices. There are errors
> in the boot showing up as
> marvell-nfc soc:internal-regs:nand@d0000: invalid resource
> marvell-nfc: probe of soc:internal-regs:nand@d0000 failed with error -22

Perhaps the issues comes from your device tree, as this lines says. You
should check the bindings and verify they are in sync.

Bindings should be backward compatible but who knows...

Thanks,
Miquèl

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