mail archive of the barebox mailing list
 help / color / mirror / Atom feed
* mtd_peb_write broken on dataflash
@ 2018-09-20 18:27 Ladislav Michl
  2018-09-21  7:11 ` Sascha Hauer
  0 siblings, 1 reply; 3+ messages in thread
From: Ladislav Michl @ 2018-09-20 18:27 UTC (permalink / raw)
  To: barebox

As mtd_peb_write checks for alignment, it fails writing to dataflash:
barebox$ state -s
ERROR: state: Failed to write PEB 0, -22
ERROR: state: Failed to write circular to 0 length 48, -22
WARNING: state: Failed to write state backend bucket, -22
ERROR: state: Failed to write PEB 1, -22
ERROR: state: Failed to write circular to 0 length 48, -22
WARNING: state: Failed to write state backend bucket, -22
ERROR: state: Failed to write PEB 2, -22
ERROR: state: Failed to write circular to 0 length 48, -22
WARNING: state: Failed to write state backend bucket, -22
ERROR: state: Failed to write state to at least 1 buckets. Successfully written to 0 buckets
ERROR: state: Failed to write packed state, -5
state: I/O error

Patch bellow is a naive fix just to show where problem is.

There's also problem erasing dataflash:
barebox$ erase /dev/dataflash0.barebox-state
erase: Invalid argument

I guess we'll run into similar issue for MTD_RAM, so question is whenever
we want such a strict check here at all. Opinions?

Thank you.

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

diff --git a/drivers/mtd/peb.c b/drivers/mtd/peb.c
index c35b63f2f..6035fbc45 100644
--- a/drivers/mtd/peb.c
+++ b/drivers/mtd/peb.c
@@ -376,7 +376,7 @@ int mtd_peb_write(struct mtd_info *mtd, const void *buf, int pnum, int offset,
 		return -EINVAL;
 	if (len <= 0)
 		return -EINVAL;
-	if (len % (mtd->writesize >> mtd->subpage_sft))
+	if (mtd->type != MTD_DATAFLASH && len % (mtd->writesize >> mtd->subpage_sft))
 		return -EINVAL;
 	if (mtd_peb_is_bad(mtd, pnum))
 		return -EINVAL;
-- 
2.19.0


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: mtd_peb_write broken on dataflash
  2018-09-20 18:27 mtd_peb_write broken on dataflash Ladislav Michl
@ 2018-09-21  7:11 ` Sascha Hauer
  2018-09-21  9:46   ` Ladislav Michl
  0 siblings, 1 reply; 3+ messages in thread
From: Sascha Hauer @ 2018-09-21  7:11 UTC (permalink / raw)
  To: Ladislav Michl; +Cc: barebox

Hi Ladis,

On Thu, Sep 20, 2018 at 08:27:01PM +0200, Ladislav Michl wrote:
> As mtd_peb_write checks for alignment, it fails writing to dataflash:
> barebox$ state -s
> ERROR: state: Failed to write PEB 0, -22
> ERROR: state: Failed to write circular to 0 length 48, -22
> WARNING: state: Failed to write state backend bucket, -22
> ERROR: state: Failed to write PEB 1, -22
> ERROR: state: Failed to write circular to 0 length 48, -22
> WARNING: state: Failed to write state backend bucket, -22
> ERROR: state: Failed to write PEB 2, -22
> ERROR: state: Failed to write circular to 0 length 48, -22

How I read the code the length should be writesize aligned...

> diff --git a/drivers/mtd/peb.c b/drivers/mtd/peb.c
> index c35b63f2f..6035fbc45 100644
> --- a/drivers/mtd/peb.c
> +++ b/drivers/mtd/peb.c
> @@ -376,7 +376,7 @@ int mtd_peb_write(struct mtd_info *mtd, const void *buf, int pnum, int offset,
>  		return -EINVAL;
>  	if (len <= 0)
>  		return -EINVAL;
> -	if (len % (mtd->writesize >> mtd->subpage_sft))
> +	if (mtd->type != MTD_DATAFLASH && len % (mtd->writesize >> mtd->subpage_sft))

So how big is mtd->writesize here and is this value correct?

In state_storage_mtd_buckets_init() we have:

        if (circular)
                writesize = meminfo->writesize;
        else
                writesize = meminfo->erasesize;

How big are meminfo->writesize and meminfo->erasesize here and
which path is chosen?

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: mtd_peb_write broken on dataflash
  2018-09-21  7:11 ` Sascha Hauer
@ 2018-09-21  9:46   ` Ladislav Michl
  0 siblings, 0 replies; 3+ messages in thread
From: Ladislav Michl @ 2018-09-21  9:46 UTC (permalink / raw)
  To: Sascha Hauer; +Cc: barebox

On Fri, Sep 21, 2018 at 09:11:11AM +0200, Sascha Hauer wrote:
> Hi Ladis,
> 
> On Thu, Sep 20, 2018 at 08:27:01PM +0200, Ladislav Michl wrote:
> > As mtd_peb_write checks for alignment, it fails writing to dataflash:
> > barebox$ state -s
> > ERROR: state: Failed to write PEB 0, -22
> > ERROR: state: Failed to write circular to 0 length 48, -22
> > WARNING: state: Failed to write state backend bucket, -22
> > ERROR: state: Failed to write PEB 1, -22
> > ERROR: state: Failed to write circular to 0 length 48, -22
> > WARNING: state: Failed to write state backend bucket, -22
> > ERROR: state: Failed to write PEB 2, -22
> > ERROR: state: Failed to write circular to 0 length 48, -22
> 
> How I read the code the length should be writesize aligned...

It is perfectly valid to write less to dataflash... But see bellow.

> > diff --git a/drivers/mtd/peb.c b/drivers/mtd/peb.c
> > index c35b63f2f..6035fbc45 100644
> > --- a/drivers/mtd/peb.c
> > +++ b/drivers/mtd/peb.c
> > @@ -376,7 +376,7 @@ int mtd_peb_write(struct mtd_info *mtd, const void *buf, int pnum, int offset,
> >  		return -EINVAL;
> >  	if (len <= 0)
> >  		return -EINVAL;
> > -	if (len % (mtd->writesize >> mtd->subpage_sft))
> > +	if (mtd->type != MTD_DATAFLASH && len % (mtd->writesize >> mtd->subpage_sft))
> 
> So how big is mtd->writesize here and is this value correct?

Yes, it is correct:
dataflash@00: AT45DB041x (528 KBytes) pagesize 264 bytes (OTP)

> In state_storage_mtd_buckets_init() we have:
> 
>         if (circular)
>                 writesize = meminfo->writesize;
>         else
>                 writesize = meminfo->erasesize;
> 
> How big are meminfo->writesize and meminfo->erasesize here and
> which path is chosen?

writesize: 264
erasesize: 264
subpage_sft: 0
circular: true

But later we are trying to write 48 bytes:
dataflash0.barebox-state: write 48 bytes to PEB 0:0

	ladis

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2018-09-21  9:47 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-09-20 18:27 mtd_peb_write broken on dataflash Ladislav Michl
2018-09-21  7:11 ` Sascha Hauer
2018-09-21  9:46   ` Ladislav Michl

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox