[U-Boot] [PATCH v4 00/27] SPI-NAND support

[Miquel Raynal]

Hi Jagan,

Stefan Roese <sr at denx.de> wrote on Tue, 31 Jul 2018 07:50:54 +0200:

> Hi Jagan,
> 
> On 31.07.2018 07:36, Jagan Teki wrote:
> 
> <snip>
> 
> >>>> Acked--by: Jagan Teki <jagan at openedev.com>
> >>>>  
> >>>
> >>> Thanks!
> >>>  
> >>>> Can you rebase on master and send the needed patches or whole? Look
> >>>> like some changes been added in drivers/mtd/nand/Kconfig  
> >>>
> >>>
> >>> I'll wait a bit for Stefan review also and I think I missed something
> >>> in mtdparts: old partitions are not freed when creating new ones.  
> > > Is this resolved with v5?  

Absolutely not, but I will fix it in a next series during the
stabilization cycle. Maybe I will rework a bit mtdparts too, to trash
the thin layer that handles the partitions to use in a much more
consistent way the MTD core.

> > >>  
> >> I'm back from vacation and am starting to work on this SPI NAND
> >> support again. Right now, I'm facing a problem with 32 Bytes
> >> missing when written to NAND and read back. Most likely a problem
> >> with my SPI driver which supports a maximum of 32 Bytes per SPI
> >> message (I'm using adjust_op_size() to adjust the max xfer size).
> > > Is it with kirkwood?  
> 
> No. Its with a new platform I'm currently working on (MediaTek
> MT7688 MIPS) - custom board. The platform and board port will be
> posted in a few days / weeks (once its ready).
> 
> >> As for waiting for my review comments, I would suggest to pull
> >> this patchset (once updated onto TOT) soon, as the merge window
> >> closes just today. We can fix issues later in this release cycle.
> >> Otherwise we need to postpone this series to the next release, which
> >> is of course also on option.
> > > If the issue is generic and it changed the current behavior on mtd or  
> > any other related areas, it's better to hold.  
> 
> AFAICT its not generic. The SPI NAND seems to be working for Miquel
> and its a new infrastructure. So chances for breaking an existing
> board / platform are pretty low.

I totally agree on this.

> By pulling this patchset now, we
> enable the possibility for implementing and testing it on other
> platforms sooner. Tom also seems to be willing to do so.

Jagan, I just saw your answer to the cover letter of the v5,
thanks for applying it.

Regards,
Miquèl