[PATCH v3] mtd: cfi_cmdset_0002: Change write buffer to check correct value

[IKEGAMI Tokunori]

Hi Jocke-san,

Thanks for the comments.
Noted it.
I am happy with it;-)

Regards,
Ikegami

> -----Original Message-----
> From: Joakim Tjernlund [mailto:Joakim.Tjernlund at infinera.com]
> Sent: Wednesday, May 16, 2018 2:17 AM
> To: boris.brezillon at bootlin.com; IKEGAMI Tokunori
> Cc: linux-mtd at lists.infradead.org; PACKHAM Chris;
> cyrille.pitchen at wedev4u.fr; dwmw2 at infradead.org;
> computersforpeace at gmail.com; boris.brezillon at free-electrons.com;
> richard at nod.at; marek.vasut at gmail.com
> Subject: Re: [PATCH v3] mtd: cfi_cmdset_0002: Change write buffer to
> check correct value
> 
> On Tue, 2018-05-15 at 16:49 +0000, IKEGAMI Tokunori wrote:
> >
> > Hi Jocke-san,
> 
> Hi Ikegami-san,
> 
> >
> > > Then I got really curious, dos the same work for your erase problem?
> > > Replace chip_ready(...) with chip_good(map, adr, map_word_ff(map) ?
> > > Don't you have any status bits when erase/write fails?
> >
> >   Yes this also you are right.
> >   I have just tested the effect as below.
> >   The flash erase error also does not occur by checking chip_good
> instead of chip_ready.
> >   This has been tested by the same reproduction method.
> >   It looks that the flash erase error is also able to be resolved by
> chip_good.
> >   So I will do update the patch series to change use chip_good instead
> of chip_ready.
> >   But to make sure I would like remain the changes to retry the error.
> >   Also I will combine the changes for the flash write error also as
> same series.
> 
> Nice :) I just had a look at chip_good() and saw that it is actually both
> toggle and DQ7, belt and suspenders :)
> I thought chip_good() was only DQ7, anyhow, I guess this makes the test
> very robust but a bit slow.
> 
> >
> > Regards,
> > Ikegami
> >