public inbox for isar-users@googlegroups.com
 help / color / mirror / Atom feed
* [RFC PATCH 0/2] Implementation of UBI
@ 2019-01-29 12:10 claudius.heine.ext
  2019-01-29 12:10 ` [RFC PATCH 1/2] meta/classes: added ubi-img class claudius.heine.ext
  2019-01-29 12:10 ` [RFC PATCH 2/2] meta-isar: added qemusabrelite example to demonstrate UBI use claudius.heine.ext
  0 siblings, 2 replies; 3+ messages in thread
From: claudius.heine.ext @ 2019-01-29 12:10 UTC (permalink / raw)
  To: isar-users; +Cc: Claudius Heine

From: Claudius Heine <ch@denx.de>

Hi,

I send this patch as an RFC, because I would like to have feedback about
it first, before finalizing it with documentation and more.

Here a few remarks about what I would like to have feedback on.

# meta/classes: added ubi-img class

    I would like to know what you think about the template system
employed here. For example usage see the second patch.

    I plan to add some documentation about this into the user manual as
well.

# meta-isar: added qemusabrelite example to demonstrate UBI use

    My primary goal here is to provide some reference and ci test case
for the UBI image creation. Not a real goal for me is to provide a
working BSP for sabrelite running on qemu, but that would be nice to
have.

    Background for this is that qemu supports a `--mtdblock` parameter
for sabrelite, so it might be possible to use the resulting ubi image to
boot a system using this. The problem is however that this doesn't seem
to work. I am not sure if what ubinize outputs can even be directly used
for this parameter or how to generate a valid mtd file to be used for
that. Normally ubiformat is used to write ubi images to mtd partitions
on real hardware, so I guess some emulation needs to happen in order to
write correct mtd image files.

   Another issue is that in order to boot from NAND directy in i.MX6
platforms, the NAND has to contain some propritary infomation that is
written by the kobs [1]. So for me the efford needed to provide a
working qemu or hardware sabrelite support seems a bit to high for this
patchset.

   So instead I think I will just rename the qemusabrelite stuff to
nand-ubi-demo to provide just some reference about how to use this
instead of falsely claim support for some hardware. IMO testing if the
image can be created is more of a goal here than testing if it can be
booted as well.

   Are the other possiblities I don't see here?

Thanks,
Claudius

[1] https://github.com/NXPmicro/imx-kobs 

Claudius Heine (2):
  meta/classes: added ubi-img class
  meta-isar: added qemusabrelite example to demonstrate UBI use

 meta-isar/classes/ubi-ubifs-img.bbclass       |  9 +++
 meta-isar/conf/machine/qemusabrelite.conf     | 11 +++
 .../multiconfig/qemusabrelite-buster.conf     | 13 ++++
 .../images/files/ubinize.cfg.tmpl             | 35 +++++++++
 .../recipes-core/images/isar-image-ubi.bb     | 49 +++++++++++++
 meta/classes/ubi-img.bbclass                  | 71 +++++++++++++++++++
 6 files changed, 188 insertions(+)
 create mode 100644 meta-isar/classes/ubi-ubifs-img.bbclass
 create mode 100644 meta-isar/conf/machine/qemusabrelite.conf
 create mode 100644 meta-isar/conf/multiconfig/qemusabrelite-buster.conf
 create mode 100644 meta-isar/recipes-core/images/files/ubinize.cfg.tmpl
 create mode 100644 meta-isar/recipes-core/images/isar-image-ubi.bb
 create mode 100644 meta/classes/ubi-img.bbclass

-- 
2.20.1


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

end of thread, other threads:[~2019-01-29 12:10 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-01-29 12:10 [RFC PATCH 0/2] Implementation of UBI claudius.heine.ext
2019-01-29 12:10 ` [RFC PATCH 1/2] meta/classes: added ubi-img class claudius.heine.ext
2019-01-29 12:10 ` [RFC PATCH 2/2] meta-isar: added qemusabrelite example to demonstrate UBI use claudius.heine.ext

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