Openembedded Core Discussions
 help / color / mirror / Atom feed
* UBI filesystem creation fails silently with 4k erase size
@ 2014-07-14 12:07 Mike Looijmans
  2014-07-14 13:06 ` Mike Looijmans
  0 siblings, 1 reply; 3+ messages in thread
From: Mike Looijmans @ 2014-07-14 12:07 UTC (permalink / raw)
  To: Patches and discussions about the oe-core layer

I have a NOR chip that reports a 4k erase size, so in my machine.conf I entered:

MKUBIFS_ARGS = "-m 1 -e 3968 -c 7040"
UBINIZE_ARGS = "-m 1 -p 4096"
UBI_VOLNAME = "qspi-rootfs"


When I create an image with "ubi" in the IMAGE_FSTYPES, the build runs and 
does not report any error, however it did NOT create the ubi filesystem image.

If I change the parameters for a 64k flash using:

MKUBIFS_ARGS = "-m 1 -e 65408 -c 440"

the ubi filesystem is created (both ubifs and ubi files).


Where can I find what went wrong? I'd expect big read error messages telling 
me that it failed to create an image for my device, but I cannot find such 
message anywhere.

So my questions are:

1) Why is there no error message when image creation fails?
2) Where is the log/output from the mkfs.ubi commands?
2) What could be wrong with my mkubifs parameters?

Mike.


Met vriendelijke groet / kind regards,

Mike Looijmans

TOPIC Embedded Systems
Eindhovenseweg 32-C, NL-5683 KH Best
Postbus 440, NL-5680 AK Best
Telefoon: (+31) (0) 499 33 69 79
Telefax:  (+31) (0) 499 33 69 70
E-mail: mike.looijmans@topic.nl
Website: www.topic.nl

Please consider the environment before printing this e-mail

Topic zoekt FPGA experts
http://topic.nl/vacatures/word-jij-onze-nieuwe-fpga-expert/



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

* Re: UBI filesystem creation fails silently with 4k erase size
  2014-07-14 12:07 UBI filesystem creation fails silently with 4k erase size Mike Looijmans
@ 2014-07-14 13:06 ` Mike Looijmans
  2014-07-15 15:15   ` Denys Dmytriyenko
  0 siblings, 1 reply; 3+ messages in thread
From: Mike Looijmans @ 2014-07-14 13:06 UTC (permalink / raw)
  To: Patches and discussions about the oe-core layer

On 07/14/2014 02:07 PM, Mike Looijmans wrote:
> I have a NOR chip that reports a 4k erase size, so in my machine.conf I entered:
>
> MKUBIFS_ARGS = "-m 1 -e 3968 -c 7040"
> UBINIZE_ARGS = "-m 1 -p 4096"
> UBI_VOLNAME = "qspi-rootfs"
>
>
> When I create an image with "ubi" in the IMAGE_FSTYPES, the build runs and
> does not report any error, however it did NOT create the ubi filesystem image.
>
> If I change the parameters for a 64k flash using:
>
> MKUBIFS_ARGS = "-m 1 -e 65408 -c 440"
>
> the ubi filesystem is created (both ubifs and ubi files).
>
>
> Where can I find what went wrong? I'd expect big read error messages telling
> me that it failed to create an image for my device, but I cannot find such
> message anywhere.
>
> So my questions are:
>
> 1) Why is there no error message when image creation fails?
> 2) Where is the log/output from the mkfs.ubi commands?
> 2) What could be wrong with my mkubifs parameters?
>

The answer to that last question is that UBI itself refuses to work with a 
sector size below 15kB. Which results in UBI now failing to work on most NOR 
chips on the market because the kernel recently uses 4k erasing for all serial 
flash chips that support it.

So just the question remains why OE doesn't tell me about mkfs.ubifs failure?

Mike.


Met vriendelijke groet / kind regards,

Mike Looijmans

TOPIC Embedded Systems
Eindhovenseweg 32-C, NL-5683 KH Best
Postbus 440, NL-5680 AK Best
Telefoon: (+31) (0) 499 33 69 79
Telefax:  (+31) (0) 499 33 69 70
E-mail: mike.looijmans@topic.nl
Website: www.topic.nl

Please consider the environment before printing this e-mail

Topic zoekt FPGA experts
http://topic.nl/vacatures/word-jij-onze-nieuwe-fpga-expert/



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

* Re: UBI filesystem creation fails silently with 4k erase size
  2014-07-14 13:06 ` Mike Looijmans
@ 2014-07-15 15:15   ` Denys Dmytriyenko
  0 siblings, 0 replies; 3+ messages in thread
From: Denys Dmytriyenko @ 2014-07-15 15:15 UTC (permalink / raw)
  To: Mike Looijmans; +Cc: Patches and discussions about the oe-core layer

On Mon, Jul 14, 2014 at 03:06:31PM +0200, Mike Looijmans wrote:
> On 07/14/2014 02:07 PM, Mike Looijmans wrote:
> >I have a NOR chip that reports a 4k erase size, so in my machine.conf I entered:
> >
> >MKUBIFS_ARGS = "-m 1 -e 3968 -c 7040"
> >UBINIZE_ARGS = "-m 1 -p 4096"
> >UBI_VOLNAME = "qspi-rootfs"
> >
> >
> >When I create an image with "ubi" in the IMAGE_FSTYPES, the build runs and
> >does not report any error, however it did NOT create the ubi filesystem image.
> >
> >If I change the parameters for a 64k flash using:
> >
> >MKUBIFS_ARGS = "-m 1 -e 65408 -c 440"
> >
> >the ubi filesystem is created (both ubifs and ubi files).
> >
> >
> >Where can I find what went wrong? I'd expect big read error messages telling
> >me that it failed to create an image for my device, but I cannot find such
> >message anywhere.
> >
> >So my questions are:
> >
> >1) Why is there no error message when image creation fails?
> >2) Where is the log/output from the mkfs.ubi commands?
> >2) What could be wrong with my mkubifs parameters?
> >
> 
> The answer to that last question is that UBI itself refuses to work
> with a sector size below 15kB. Which results in UBI now failing to
> work on most NOR chips on the market because the kernel recently
> uses 4k erasing for all serial flash chips that support it.
> 
> So just the question remains why OE doesn't tell me about mkfs.ubifs failure?

Do you have the following fix?

http://cgit.openembedded.org/openembedded-core/commit/?id=f8125a1e9b6893a12355d55d4df584a8d97f0bff

-- 
Denys


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

end of thread, other threads:[~2014-07-15 15:15 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2014-07-14 12:07 UBI filesystem creation fails silently with 4k erase size Mike Looijmans
2014-07-14 13:06 ` Mike Looijmans
2014-07-15 15:15   ` Denys Dmytriyenko

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