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

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

[IKEGAMI Tokunori]

From: Tokunori Ikegami <ikegami@allied-telesis.co.jp>

For the word write it is checked if the chip has the correct value.
But it is not checked for the write buffer as only checked if ready.
To make sure for the write buffer change to check the value.

It is enough as this patch is only checking the last written word.
Since it is described by data sheets to check the operation status.

Signed-off-by: Tokunori Ikegami <ikegami@allied-telesis.co.jp>
Cc: Chris Packham <chris.packham@alliedtelesis.co.nz>
Cc: Brian Norris <computersforpeace@gmail.com>
Cc: David Woodhouse <dwmw2@infradead.org>
Cc: Boris Brezillon <boris.brezillon@free-electrons.com>
Cc: Marek Vasut <marek.vasut@gmail.com>
Cc: Richard Weinberger <richard@nod.at>
Cc: Cyrille Pitchen <cyrille.pitchen@wedev4u.fr>
Cc: linux-mtd@lists.infradead.org
---
No difference for the change itself from the original patch v1.
But just updated the patch subject to add version as v3.
Form the original patch v1 it is also changed to add Brian as Cc and the patch is sent to Boris-san as To.
  Note: This was changed from the second patch v2 and this v3 patch is not changed about this.

This fix is required to resolve the flash write error caused on our products.

This patch itself is not depended to other related patches to fix for the flash erase error.
The flash erase error is not able to still reproduce at this moment.
So I will do fix the flash erase error separately.

But the flash write error can be reproduced and this fix can resolve it.
So please review the patch again and if any problem please let me know.


 drivers/mtd/chips/cfi_cmdset_0002.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/mtd/chips/cfi_cmdset_0002.c b/drivers/mtd/chips/cfi_cmdset_0002.c
index 56aa6b75213d..5e9f2ca0a6c1 100644
--- a/drivers/mtd/chips/cfi_cmdset_0002.c
+++ b/drivers/mtd/chips/cfi_cmdset_0002.c
@@ -1879,7 +1879,7 @@ static int __xipram do_write_buffer(struct map_info *map, struct flchip *chip,
 		if (time_after(jiffies, timeo) && !chip_ready(map, adr))
 			break;
 
-		if (chip_ready(map, adr)) {
+		if (chip_ready(map, adr) && chip_good(map, adr, datum)) {
 			xip_enable(map, chip, adr);
 			goto op_done;
 		}
-- 
2.16.1

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

[Joakim Tjernlund]

On Fri, 2018-05-11 at 07:58 +0000, IKEGAMI Tokunori wrote:
> 
> From: Tokunori Ikegami <ikegami@allied-telesis.co.jp>
> 
> For the word write it is checked if the chip has the correct value.
> But it is not checked for the write buffer as only checked if ready.
> To make sure for the write buffer change to check the value.
> 
> It is enough as this patch is only checking the last written word.
> Since it is described by data sheets to check the operation status.
> 
> Signed-off-by: Tokunori Ikegami <ikegami@allied-telesis.co.jp>
> Cc: Chris Packham <chris.packham@alliedtelesis.co.nz>
> Cc: Brian Norris <computersforpeace@gmail.com>
> Cc: David Woodhouse <dwmw2@infradead.org>
> Cc: Boris Brezillon <boris.brezillon@free-electrons.com>
> Cc: Marek Vasut <marek.vasut@gmail.com>
> Cc: Richard Weinberger <richard@nod.at>
> Cc: Cyrille Pitchen <cyrille.pitchen@wedev4u.fr>
> Cc: linux-mtd@lists.infradead.org
> ---
> No difference for the change itself from the original patch v1.
> But just updated the patch subject to add version as v3.
> Form the original patch v1 it is also changed to add Brian as Cc and the patch is sent to Boris-san as To.
>   Note: This was changed from the second patch v2 and this v3 patch is not changed about this.
> 
> This fix is required to resolve the flash write error caused on our products.
> 
> This patch itself is not depended to other related patches to fix for the flash erase error.
> The flash erase error is not able to still reproduce at this moment.
> So I will do fix the flash erase error separately.
> 
> But the flash write error can be reproduced and this fix can resolve it.
> So please review the patch again and if any problem please let me know.
> 
> 
>  drivers/mtd/chips/cfi_cmdset_0002.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/mtd/chips/cfi_cmdset_0002.c b/drivers/mtd/chips/cfi_cmdset_0002.c
> index 56aa6b75213d..5e9f2ca0a6c1 100644
> --- a/drivers/mtd/chips/cfi_cmdset_0002.c
> +++ b/drivers/mtd/chips/cfi_cmdset_0002.c
> @@ -1879,7 +1879,7 @@ static int __xipram do_write_buffer(struct map_info *map, struct flchip *chip,
>                 if (time_after(jiffies, timeo) && !chip_ready(map, adr))
>                         break;
> 
> -               if (chip_ready(map, adr)) {
> +               if (chip_ready(map, adr) && chip_good(map, adr, datum)) {
>                         xip_enable(map, chip, adr);
>                         goto op_done;
>                 }

What goes on here, I think, is that the Toggle bit is not reliable and adding chip_good()
basically tests DQ7 too. 
If that is the case, one should only need chip_good(), have you tested that?

In general I think toggle bit is awkward and moving to DQ7 would simply things once
tests for other status bits are added.

 Jocke

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

[IKEGAMI Tokunori]

Hi Jocke-san,

Thank you so much for your comments.

> What goes on here, I think, is that the Toggle bit is not reliable and
> adding chip_good()
> basically tests DQ7 too.
> If that is the case, one should only need chip_good(), have you tested
> that?
> 
> In general I think toggle bit is awkward and moving to DQ7 would simply
> things once
> tests for other status bits are added.

  Yes you are right.
  Since I have just confirmed that the flash write issue can be resolved by checking chip_good() only.
  So I will do update the patch as so.

Regards,
Ikegami

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

[Joakim Tjernlund]

On Tue, 2018-05-15 at 09:08 +0000, IKEGAMI Tokunori wrote:
> CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you recognize the sender and know the content is safe.
> 
> 
> Hi Jocke-san,
> 
> Thank you so much for your comments.
> 
> > What goes on here, I think, is that the Toggle bit is not reliable and
> > adding chip_good()
> > basically tests DQ7 too.
> > If that is the case, one should only need chip_good(), have you tested
> > that?
> > 
> > In general I think toggle bit is awkward and moving to DQ7 would simply
> > things once
> > tests for other status bits are added.
> 
>   Yes you are right.
>   Since I have just confirmed that the flash write issue can be resolved by checking chip_good() only.
>   So I will do update the patch as so.

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?

 Jocke

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

[IKEGAMI Tokunori]

Hi Jocke-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.

Regards,
Ikegami

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

[Joakim Tjernlund]

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
> 

RE: [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@infinera.com]
> Sent: Wednesday, May 16, 2018 2:17 AM
> To: boris.brezillon@bootlin.com; IKEGAMI Tokunori
> Cc: linux-mtd@lists.infradead.org; PACKHAM Chris;
> cyrille.pitchen@wedev4u.fr; dwmw2@infradead.org;
> computersforpeace@gmail.com; boris.brezillon@free-electrons.com;
> richard@nod.at; marek.vasut@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
> >