Skip to contentSkip to conversation

kernel-fit-image.bbclass: Do not include kernel property in DTBO config subnodes

PATCH In review Marek Vasut 4 participants 19 messages Jun 30 – Aug 6

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 kernel-fit-image.bbclass: Do not include kernel property in DTBO config subnodes

The OE 5.0 implementation of fitImage generation did not include the kernel property in DTBO configuration node subnodes, while the current OE 6.0 implementation does. The previous implementation was correct, because kernel together with only a DTBO overlay can not form bootable result. The kernel property should only be part of the configuration subnode which includes both kernel and DTB. Fix this.

Fixes 05d0c7342d76 ("kernel-fit-image.bbclass: add a new FIT image implementation")
(cherry picked from commit 06ed34005957a6afb88270603df5e545941546b0)

---
Cc: Adrian Freihofer <[email protected]>
Cc: Richard Purdie <[email protected]>
Cc: Yoann Congal <[email protected]>

---
 meta/lib/oe/fitimage.py | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
meta/lib/oe/fitimage.py +6 -1
index 881d0eae0ab..d7e21171ab9 100644
@@ -425,8 +425,13 @@ class ItsNodeRootKernel(ItsNode):
         opt_props = {}
         conf_desc = []
         sign_entries = []
+        include_kernel = 1
 
-        if self._kernel:
+        # DTBO configuration sections do not include kernel property
+        if dtb and os.path.splitext(dtb.name)[1] == ".dtbo":
+            include_kernel = 0
+
+        if self._kernel and include_kernel:
             conf_desc.append("Linux kernel")
             opt_props["kernel"] = self._kernel.name
             if self._sign_enable:
Signature · 1 line
2.53.0
#1 lore raw

Hi Marek

I guess, before this patch can be applied, this one (or a back-port of it) is required:
https://git.openembedded.org/openembedded-core/commit/?id=85e0408a81bdca0411a8625061e5e343c05f92d9
Otherwise, the oe-selftest fail.

Regards,
Adrian

Quoted 44 lines · Marek Vasut

The OE 5.0 implementation of fitImage generation did not include the kernel property in DTBO configuration node subnodes, while the current
OE 6.0 implementation does. The previous implementation was correct, because kernel together with only a DTBO overlay can not form bootable
result. The kernel property should only be part of the configuration subnode which includes both kernel and DTB. Fix this.

Fixes 05d0c7342d76 ("kernel-fit-image.bbclass: add a new FIT image

implementation")

Signed-off-by Mathieu Dubois-Briand

<[email protected]>

(cherry picked from commit 06ed34005957a6afb88270603df5e545941546b0)

---
Cc: Adrian Freihofer <[email protected]>
Cc: Richard Purdie <[email protected]>
Cc: Yoann Congal <[email protected]>
---
 meta/lib/oe/fitimage.py | 7 ++++++-

 1 file changed, 6 insertions(+), 1 deletion(-)
meta/lib/oe/fitimage.py +0 -0
index 881d0eae0ab..d7e21171ab9 100644
@@ -425,8 +425,13 @@ class ItsNodeRootKernel(ItsNode):
          opt_props = {}

         conf_desc = []
         sign_entries = []
+        include_kernel = 1

-        if self._kernel:
+        # DTBO configuration sections do not include kernel property +        if dtb and os.path.splitext(dtb.name)[1] == ".dtbo":
+            include_kernel = 0
+
+        if self._kernel and include_kernel:
             conf_desc.append("Linux kernel") opt_props["kernel"] = self._kernel.name
             if self._sign_enable:

#3 lore raw
Quoted 8 lines · Marek Vasut

On 6/30/26 12:51 PM, Freihofer, Adrian wrote:

Hello Adrian,

Quoted 4 lines

I guess, before this patch can be applied, this one (or a back-port of it) is required:
https://git.openembedded.org/openembedded-core/commit/?id=85e0408a81bdca0411a8625061e5e343c05f92d9
Otherwise, the oe-selftest fail.

Quoted 5 lines

Thank you for staying vigilant.

Yoann, can you please pick the aforementioned commit too ?

Thank you for your help !

Hello,

I missed that discussion, sorry. But was reminded by Ryan (thanks).

If I understood correclty, the set of patches we need for wrynose is:
* kernel-fit-image.bbclass: Do not include kernel property in DTBO config subnodes
* kernel-fit-image.bbclass: Fix operation with KERNEL_DTBVENDORED = "1"
* oe-selftest: fitimage: Do not expect kernel property in DTBO config subnodes
* kernel-fit-image: Add KERNEL_DTBVENDORED support for FIT_CONF_DEFAULT_DTB (this one is not merged on master yet)

Marek, Adrian, Ryan: does that sound correct?

If yes, Marek, when the 4th patch merge, can you send an updated series with all 4 backport patches?

Thanks!

Signature · 2 lines
Yoann Congal
Smile ECS
#4 lore raw
Quoted 30 lines · Yoann Congal via lists.openembedded.org
Quoted 9 lines · Marek Vasut

On 6/30/26 12:51 PM, Freihofer, Adrian wrote:

Hello Adrian,

Quoted 5 lines

I guess, before this patch can be applied, this one (or a back- port of
it) is required:
https://git.openembedded.org/openembedded-core/commit/?id=85e0408a81bdca0411a8625061e5e343c05f92d9
Otherwise, the oe-selftest fail.

Quoted 5 lines

Thank you for staying vigilant.

Yoann, can you please pick the aforementioned commit too ?

Thank you for your help !

Hello,

I missed that discussion, sorry. But was reminded by Ryan (thanks).

If I understood correclty, the set of patches we need for wrynose is:
* kernel-fit-image.bbclass: Do not include kernel property in DTBO config subnodes
* kernel-fit-image.bbclass: Fix operation with KERNEL_DTBVENDORED = "1"
* oe-selftest: fitimage: Do not expect kernel property in DTBO config subnodes
* kernel-fit-image: Add KERNEL_DTBVENDORED support for FIT_CONF_DEFAULT_DTB (this one is not merged on master yet)

Quoted 2 lines

Marek, Adrian, Ryan: does that sound correct?

Yes, that's correct. One little detail is the order of the patches. If you pick the oe-selftest commit first, all commits are compile and test-able. That's my suggestion:

* oe-selftest: fitimage: Do not expect kernel property in DTBO config
subnodes
* kernel-fit-image.bbclass: Do not include kernel property in DTBO
config subnodes
* kernel-fit-image.bbclass: Fix operation with KERNEL_DTBVENDORED =
"1"
* kernel-fit-image: Add KERNEL_DTBVENDORED support for
FIT_CONF_DEFAULT_DTB (this one is not merged on master yet)

same order as on master branch.

Regards,
Adrian

Quoted 16 lines

If yes, Marek, when the 4th patch merge, can you send an updated series
with all 4 backport patches?

Thanks!

-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group. View/Reply Online (#240033):
https://lists.openembedded.org/g/openembedded-core/message/240033
Mute This Topic: https://lists.openembedded.org/mt/120043543/4454582
Group Owner: [email protected] Unsubscribe:
https://lists.openembedded.org/g/openembedded-core/unsub [
[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-

#5 lore raw

On 7/2/26 1:38 PM, [email protected] wrote:

Hello everyone,

Quoted 28 lines
Quoted 25 lines
Quoted 10 lines
Quoted 5 lines

I guess, before this patch can be applied, this one (or a back- port of
it) is required:
https://git.openembedded.org/openembedded-core/commit/?id=85e0408a81bdca0411a8625061e5e343c05f92d9
Otherwise, the oe-selftest fail.

Thank you for staying vigilant.

Yoann, can you please pick the aforementioned commit too ?

Thank you for your help !

Hello,

I missed that discussion, sorry. But was reminded by Ryan (thanks).

If I understood correclty, the set of patches we need for wrynose is:
* kernel-fit-image.bbclass: Do not include kernel property in DTBO config subnodes
* kernel-fit-image.bbclass: Fix operation with KERNEL_DTBVENDORED = "1"
* oe-selftest: fitimage: Do not expect kernel property in DTBO config subnodes
* kernel-fit-image: Add KERNEL_DTBVENDORED support for FIT_CONF_DEFAULT_DTB (this one is not merged on master yet)

Quoted 2 lines

Marek, Adrian, Ryan: does that sound correct?

Quoted 6 lines

Yes, that's correct. One little detail is the order of the patches. If you pick the oe-selftest commit first, all commits are compile and test-able. That's my suggestion:

* oe-selftest: fitimage: Do not expect kernel property in DTBO config
subnodes

This is [1].

Quoted 2 lines

* kernel-fit-image.bbclass: Do not include kernel property in DTBO
config subnodes

This is [2].

Quoted 2 lines

* kernel-fit-image.bbclass: Fix operation with KERNEL_DTBVENDORED =
"1"

This is [3].

Quoted 2 lines

* kernel-fit-image: Add KERNEL_DTBVENDORED support for
FIT_CONF_DEFAULT_DTB (this one is not merged on master yet)

This is still not even in master-next .

Quoted 1 line

same order as on master branch.

Thank you for your help !

[1] https://patchwork.yoctoproject.org/project/oe-core/patch/[email protected]/
[2] https://patchwork.yoctoproject.org/project/oe-core/patch/[email protected]/
[3] https://patchwork.yoctoproject.org/project/oe-core/patch/[email protected]/

Signature · 2 lines
Best regards,
Marek Vasut
#6 lore raw

On 7/4/26 11:55 PM, Marek Vasut wrote:

Hello one more time,

Quoted 52 lines
Quoted 28 lines
Quoted 25 lines
Quoted 10 lines
Quoted 5 lines

I guess, before this patch can be applied, this one (or a back-
port of
it) is required:
https://git.openembedded.org/openembedded-core/commit/? id=85e0408a81bdca0411a8625061e5e343c05f92d9 Otherwise, the oe-selftest fail.

Thank you for staying vigilant.

Yoann, can you please pick the aforementioned commit too ?

Thank you for your help !

Hello,

I missed that discussion, sorry. But was reminded by Ryan (thanks).

If I understood correclty, the set of patches we need for wrynose is:
* kernel-fit-image.bbclass: Do not include kernel property in DTBO config subnodes
* kernel-fit-image.bbclass: Fix operation with KERNEL_DTBVENDORED = "1"
* oe-selftest: fitimage: Do not expect kernel property in DTBO config subnodes
* kernel-fit-image: Add KERNEL_DTBVENDORED support for FIT_CONF_DEFAULT_DTB (this one is not merged on master yet)

Quoted 2 lines

Marek, Adrian, Ryan: does that sound correct?

Quoted 6 lines

Yes, that's correct. One little detail is the order of the patches. If you pick the oe-selftest commit first, all commits are compile and test-able. That's my suggestion:

  * oe-selftest: fitimage: Do not expect kernel property in DTBO config subnodes

This is [1].

Quoted 2 lines

  * kernel-fit-image.bbclass: Do not include kernel property in DTBO config subnodes

This is [2].

Quoted 2 lines

  * kernel-fit-image.bbclass: Fix operation with KERNEL_DTBVENDORED = "1"

This is [3].

Quoted 2 lines

  * kernel-fit-image: Add KERNEL_DTBVENDORED support for FIT_CONF_DEFAULT_DTB (this one is not merged on master yet)

This is still not even in master-next .

Quoted 7 lines
Quoted 1 line

same order as on master branch.

Thank you for your help !

[1] https://patchwork.yoctoproject.org/project/oe-core/ patch/[email protected]/
[2] https://patchwork.yoctoproject.org/project/oe-core/ patch/[email protected]/
[3] https://patchwork.yoctoproject.org/project/oe-core/ patch/[email protected]/

Do you think those three backports can be picked up for OE wrynose ?

Thank you for your help !

Signature · 2 lines
Best regards,
Marek Vasut
#7 lore raw
Quoted 68 lines · Marek Vasut

On 7/4/26 11:55 PM, Marek Vasut wrote:

Hello one more time,

Quoted 55 lines
Quoted 29 lines
Quoted 26 lines
Quoted 11 lines
Quoted 6 lines

I guess, before this patch can be applied, this one (or a back- port of
it) is required:
https://git.openembedded.org/openembedded-core/commit/?
id=85e0408a81bdca0411a8625061e5e343c05f92d9 Otherwise, the oe-selftest fail.

Thank you for staying vigilant.

Yoann, can you please pick the aforementioned commit too ?

Thank you for your help !

Hello,

I missed that discussion, sorry. But was reminded by Ryan (thanks).

If I understood correclty, the set of patches we need for wrynose is:
* kernel-fit-image.bbclass: Do not include kernel property in DTBO config subnodes
* kernel-fit-image.bbclass: Fix operation with KERNEL_DTBVENDORED = "1"
* oe-selftest: fitimage: Do not expect kernel property in DTBO config subnodes
* kernel-fit-image: Add KERNEL_DTBVENDORED support for FIT_CONF_DEFAULT_DTB (this one is not merged on master yet)

Quoted 2 lines

Marek, Adrian, Ryan: does that sound correct?

Quoted 6 lines

Yes, that's correct. One little detail is the order of the patches. If you pick the oe-selftest commit first, all commits are compile and test-able. That's my suggestion:

  * oe-selftest: fitimage: Do not expect kernel property in DTBO config subnodes

This is [1].

Quoted 2 lines

  * kernel-fit-image.bbclass: Do not include kernel property in DTBO config subnodes

This is [2].

Quoted 2 lines

  * kernel-fit-image.bbclass: Fix operation with KERNEL_DTBVENDORED = "1"

This is [3].

Quoted 2 lines

  * kernel-fit-image: Add KERNEL_DTBVENDORED support for FIT_CONF_DEFAULT_DTB (this one is not merged on master yet)

This is still not even in master-next .

Quoted 1 line

same order as on master branch.

Quoted 8 lines

Thank you for your help !

[1] https://patchwork.yoctoproject.org/project/oe-core/
patch/[email protected]/
[2] https://patchwork.yoctoproject.org/project/oe-core/
patch/[email protected]/
[3] https://patchwork.yoctoproject.org/project/oe-core/
patch/[email protected]/

Quoted 3 lines

Do you think those three backports can be picked up for OE wrynose ?

Thank you for your help !

Sure! They even are in the branch I sent for review:
https://lore.kernel.org/all/[email protected]/

Following Paul's review, I will most likely hold it for one cycle in order to merge the test at the same time.
See: https://lore.kernel.org/all/[email protected]/

Regards,

Signature · 2 lines
Yoann Congal
Smile ECS
#8 lore raw

On 7/20/26 5:19 PM, Yoann Congal wrote:

Hello Yoann,

Quoted 7 lines
Quoted 7 lines
Quoted 6 lines

[1] https://patchwork.yoctoproject.org/project/oe-core/
patch/[email protected]/
[2] https://patchwork.yoctoproject.org/project/oe-core/
patch/[email protected]/
[3] https://patchwork.yoctoproject.org/project/oe-core/
patch/[email protected]/

Quoted 6 lines
Quoted 3 lines

Do you think those three backports can be picked up for OE wrynose ?

Thank you for your help !

Sure! They even are in the branch I sent for review:
https://lore.kernel.org/all/[email protected]/

Oh, that's nice.

Quoted 3 lines

Following Paul's review, I will most likely hold it for one cycle in order to merge the test at the same time.
See: https://lore.kernel.org/all/[email protected]/

Thank you !

Signature · 2 lines
Best regards,
Marek Vasut
#9 lore raw

On 7/20/26 5:31 PM, Marek Vasut wrote:

Hello Yoann,

Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 6 lines

[1] https://patchwork.yoctoproject.org/project/oe-core/
patch/[email protected]/
[2] https://patchwork.yoctoproject.org/project/oe-core/
patch/[email protected]/
[3] https://patchwork.yoctoproject.org/project/oe-core/
patch/[email protected]/

Quoted 12 lines
Quoted 6 lines
Quoted 3 lines

Do you think those three backports can be picked up for OE wrynose ?

Thank you for your help !

Sure! They even are in the branch I sent for review:
https://lore.kernel.org/all/[email protected]/

Oh, that's nice.

Quoted 3 lines

Following Paul's review, I will most likely hold it for one cycle in order to merge the test at the same time.
See: https://lore.kernel.org/all/ [email protected]/

Are there any news on getting the fitImage fixes into the wrynose branch, is this still on your radar ?

Thank you, and I'm sorry for the push.

Signature · 2 lines
Best regards,
Marek Vasut
#10 lore raw
Quoted 26 lines · Marek Vasut

On 7/20/26 5:31 PM, Marek Vasut wrote:

Hello Yoann,

Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 6 lines

[1] https://patchwork.yoctoproject.org/project/oe-core/
patch/[email protected]/
[2] https://patchwork.yoctoproject.org/project/oe-core/
patch/[email protected]/
[3] https://patchwork.yoctoproject.org/project/oe-core/
patch/[email protected]/

Quoted 14 lines
Quoted 6 lines
Quoted 3 lines

Do you think those three backports can be picked up for OE wrynose ?

Thank you for your help !

Sure! They even are in the branch I sent for review:
https://lore.kernel.org/all/[email protected]/

Oh, that's nice.

Quoted 5 lines

Following Paul's review, I will most likely hold it for one cycle in order to merge
the test at the same time.
See: https://lore.kernel.org/all/
[email protected]/

Quoted 4 lines

Are there any news on getting the fitImage fixes into the wrynose branch, is this still on your radar ?

Thank you, and I'm sorry for the push.

Hello,

I don't mind the push, it's fair :)

This series is still on my radar but blocked on a wrynose test patch:
Basically this:
https://lore.kernel.org/openembedded-core/[email protected]/
... but backported to wrynose.
Ryan was working on it but did not sent the patch, I don't know what happened.

Regards,

Signature · 2 lines
Yoann Congal
Smile ECS
#11 lore raw

On 7/30/26 9:36 AM, Yoann Congal wrote:

Hello Yoann,

Quoted 29 lines
Quoted 22 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 6 lines

[1] https://patchwork.yoctoproject.org/project/oe-core/
patch/[email protected]/
[2] https://patchwork.yoctoproject.org/project/oe-core/
patch/[email protected]/
[3] https://patchwork.yoctoproject.org/project/oe-core/
patch/[email protected]/

Quoted 14 lines
Quoted 6 lines
Quoted 3 lines

Do you think those three backports can be picked up for OE wrynose ?

Thank you for your help !

Sure! They even are in the branch I sent for review:
https://lore.kernel.org/all/[email protected]/

Oh, that's nice.

Quoted 5 lines

Following Paul's review, I will most likely hold it for one cycle in order to merge
the test at the same time.
See: https://lore.kernel.org/all/
[email protected]/

Quoted 4 lines

Are there any news on getting the fitImage fixes into the wrynose branch, is this still on your radar ?

Thank you, and I'm sorry for the push.

Hello,

Quoted 8 lines

I don't mind the push, it's fair :)

This series is still on my radar but blocked on a wrynose test patch:
Basically this:
https://lore.kernel.org/openembedded-core/[email protected]/
... but backported to wrynose.
Ryan was working on it but did not sent the patch, I don't know what happened.

Do you think the fitImage fixes could be backported without the test patch, or would that make some test suite fail? Please pardon my ignorance.

The fitimage fixes landed in oe-core master 6 weeks ago and they fix fitImage usability issues which block me from updating to wrynose in multiple projects, that's why I keep monitoring this thread so, but again, please do not read this as a push.

Signature · 2 lines
Best regards,
Marek Vasut
#12 lore raw

Hello Marek, hello Yoann

Quoted 46 lines · Marek Vasut

On 7/30/26 9:36 AM, Yoann Congal wrote:

Hello Yoann,

Quoted 33 lines
Quoted 24 lines
Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
Quoted 6 lines

[1] https://patchwork.yoctoproject.org/project/oe-core/
patch/[email protected]/
[2] https://patchwork.yoctoproject.org/project/oe-core/
patch/[email protected]/
[3] https://patchwork.yoctoproject.org/project/oe-core/
patch/[email protected]/

Do you think those three backports can be picked up for OE wrynose ?

Quoted 13 lines
Quoted 4 lines
Quoted 1 line

Thank you for your help !

Sure! They even are in the branch I sent for review:
https://lore.kernel.org/all/[email protected]/

Oh, that's nice.

Quoted 6 lines

Following Paul's review, I will most likely hold it for one cycle in
order to merge
the test at the same time.
See: https://lore.kernel.org/all/
[email protected]/

Quoted 4 lines

Are there any news on getting the fitImage fixes into the wrynose branch, is this still on your radar ?

Thank you, and I'm sorry for the push.

Hello,

I don't mind the push, it's fair :)

Quoted 8 lines

This series is still on my radar but blocked on a wrynose test patch:
Basically this:
https://lore.kernel.org/openembedded-core/[email protected]/
... but backported to wrynose.
Ryan was working on it but did not sent the patch, I don't know what
happened.

Quoted 8 lines

Do you think the fitImage fixes could be backported without the test patch, or would that make some test suite fail? Please pardon my ignorance.

The fitimage fixes landed in oe-core master 6 weeks ago and they fix fitImage usability issues which block me from updating to wrynose in multiple projects, that's why I keep monitoring this thread so, but again, please do not read this as a push.

I think going with the four patches from
https://lore.kernel.org/all/[email protected]/
without a fifth patch which adds test coverage for this fairly isolated detail is fine.

Important is: We need 4 not only 3 patches:

Adrian Freihofer (1):
oe-selftest: fitimage: Do not expect kernel property in DTBO config
subnodes

Marek Vasut (2):
kernel-fit-image.bbclass: Do not include kernel property in DTBO
config subnodes
kernel-fit-image.bbclass: Fix operation with KERNEL_DTBVENDORED = "1"

Ryan Eatmon (1):
kernel-fit-image: Add KERNEL_DTBVENDORED support for
FIT_CONF_DEFAULT_DTB

And also important is: We need time for testing on master before back- porting to stable. Six weeks is probably enough. But the exact same pattern ended up with a broken stable branch maybe one year ago. And also this time there was a state on master which was at least incomplete (three instead of four patches). But still, I think this is ready for stable now.

Regards,
Adrian

#13 lore raw
Quoted 52 lines · adrian.freihofer

Hello Marek, hello Yoann

Quoted 49 lines · Marek Vasut

On 7/30/26 9:36 AM, Yoann Congal wrote:

Hello Yoann,

Quoted 33 lines
Quoted 24 lines
Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
Quoted 6 lines

[1] https://patchwork.yoctoproject.org/project/oe-core/
patch/[email protected]/
[2] https://patchwork.yoctoproject.org/project/oe-core/
patch/[email protected]/
[3] https://patchwork.yoctoproject.org/project/oe-core/
patch/[email protected]/

Do you think those three backports can be picked up for OE wrynose ?

Quoted 13 lines
Quoted 4 lines
Quoted 1 line

Thank you for your help !

Sure! They even are in the branch I sent for review:
https://lore.kernel.org/all/[email protected]/

Oh, that's nice.

Quoted 6 lines

Following Paul's review, I will most likely hold it for one cycle in
order to merge
the test at the same time.
See: https://lore.kernel.org/all/
[email protected]/

Quoted 4 lines

Are there any news on getting the fitImage fixes into the wrynose branch, is this still on your radar ?

Thank you, and I'm sorry for the push.

Hello,

I don't mind the push, it's fair :)

Quoted 8 lines

This series is still on my radar but blocked on a wrynose test patch:
Basically this:
https://lore.kernel.org/openembedded-core/[email protected]/
... but backported to wrynose.
Ryan was working on it but did not sent the patch, I don't know what
happened.

Do you think the fitImage fixes could be backported without the test patch, or would that make some test suite fail? Please pardon my ignorance.

The series of patches (without test) did not break the autotbuilder tests.

Quoted 9 lines
Quoted 4 lines

The fitimage fixes landed in oe-core master 6 weeks ago and they fix fitImage usability issues which block me from updating to wrynose in multiple projects, that's why I keep monitoring this thread so, but again, please do not read this as a push.

I think going with the four patches from
https://lore.kernel.org/all/[email protected]/
without a fifth patch which adds test coverage for this fairly isolated detail is fine.

This is where I disagree. Last time I reviewed this series it looked good but I had no way to check it actually worked once backported to wrynose. That is why I asked Ryan to backport the test from master to wrynose.

Quoted 7 lines

Important is: We need 4 not only 3 patches:

Adrian Freihofer (1):
oe-selftest: fitimage: Do not expect kernel property in DTBO config
subnodes

Quoted 8 lines

Marek Vasut (2):
kernel-fit-image.bbclass: Do not include kernel property in DTBO
config subnodes
kernel-fit-image.bbclass: Fix operation with KERNEL_DTBVENDORED = "1"

Ryan Eatmon (1):
kernel-fit-image: Add KERNEL_DTBVENDORED support for
FIT_CONF_DEFAULT_DTB

Yes. This is the patches I have for this work.

Quoted 6 lines

And also important is: We need time for testing on master before back- porting to stable. Six weeks is probably enough. But the exact same pattern ended up with a broken stable branch maybe one year ago. And also this time there was a state on master which was at least incomplete (three instead of four patches). But still, I think this is ready for stable now.

I agree on the series needing testing. One more argument for waiting for a wrynose test.
See https://lore.kernel.org/all/[email protected]/
(from the same thread you linked)

Regards,

Signature · 2 lines
Yoann Congal
Smile ECS
#14 lore raw

On 7/30/26 2:39 PM, Yoann Congal wrote:

Hello everyone,

Quoted 10 lines
Quoted 9 lines
Quoted 4 lines

The fitimage fixes landed in oe-core master 6 weeks ago and they fix fitImage usability issues which block me from updating to wrynose in multiple projects, that's why I keep monitoring this thread so, but again, please do not read this as a push.

I think going with the four patches from
https://lore.kernel.org/all/[email protected]/
without a fifth patch which adds test coverage for this fairly isolated detail is fine.

Quoted 4 lines

This is where I disagree. Last time I reviewed this series it looked good but I had no way to check it actually worked once backported to wrynose. That is why I asked Ryan to backport the test from master to wrynose.

Is there a git tree/branch with those patches ? If so, I can at least test this locally and confirm whether or not anything is broken on my end ?

Quoted 24 lines
Quoted 6 lines

Important is: We need 4 not only 3 patches:

Adrian Freihofer (1):
oe-selftest: fitimage: Do not expect kernel property in DTBO config
subnodes

Quoted 8 lines

Marek Vasut (2):
kernel-fit-image.bbclass: Do not include kernel property in DTBO
config subnodes
kernel-fit-image.bbclass: Fix operation with KERNEL_DTBVENDORED = "1"

Ryan Eatmon (1):
kernel-fit-image: Add KERNEL_DTBVENDORED support for
FIT_CONF_DEFAULT_DTB

Yes. This is the patches I have for this work.

Quoted 6 lines

And also important is: We need time for testing on master before back- porting to stable. Six weeks is probably enough. But the exact same pattern ended up with a broken stable branch maybe one year ago. And also this time there was a state on master which was at least incomplete (three instead of four patches). But still, I think this is ready for stable now.

Quoted 4 lines

I agree on the series needing testing. One more argument for waiting for a wrynose test.
See https://lore.kernel.org/all/[email protected]/
(from the same thread you linked)

I ran this now and got 8 commits which cleanly apply on the wrynose branch, maybe that is all that is missing for the test part ?

$ git log --oneline --reverse --grep=fitimage origin/wrynose..origin/master -- meta/lib/oeqa/selftest/cases/fitimage.py
e46d8473fb3 oeqa/selftest/fitimage: fix missing whitespace around assignment
0344e652a40 oe-selftest: fitimage: replace _gen_random_file with _gen_elf64_dummy
5592be53348 oe-selftest: fitimage: replace bbb-dtbs-as-ext with test-dtbs-as-ext
befc103a95d oe-selftest: fitimage: replace MACHINE==qemux86-64 guards with KERNEL_SETUP_BIN
4b08cf6048b oe-selftest: fitimage: add machine settings table and skip helpers
f17f4263c07 oe-selftest: fitimage: stop hardcoding MACHINE and DISTRO in tests
85e0408a81b oe-selftest: fitimage: Do not expect kernel property in DTBO config subnodes
de2e11e6362 oe-selftest: fitimage: Add tests for KERNEL_DTBVENDORED
Signature · 2 lines
Best regards,
Marek Vasut
#15 lore raw
Quoted 84 lines · Marek Vasut via lists.openembedded.org

On 7/30/26 2:39 PM, Yoann Congal wrote:

Hello everyone,

Quoted 14 lines
Quoted 8 lines
Quoted 7 lines

The fitimage fixes landed in oe-core master 6 weeks ago and they fix
fitImage usability issues which block me from updating to wrynose in
multiple projects, that's why I keep monitoring this thread so, but
again, please do not read this as a push.

Quoted 5 lines

I think going with the four patches from
https://lore.kernel.org/all/[email protected]/
without a fifth patch which adds test coverage for this fairly isolated
detail is fine.

Quoted 7 lines

This is where I disagree. Last time I reviewed this series it looked
good but I had no way to check it actually worked once backported to
wrynose. That is why I asked Ryan to backport the test from master to
wrynose.

Is there a git tree/branch with those patches ? If so, I can at least test this locally and confirm whether or not anything is broken on my end ?

Quoted 31 lines
Quoted 14 lines

Important is: We need 4 not only 3 patches:

Adrian Freihofer (1):
   oe-selftest: fitimage: Do not expect kernel property in DTBO config
     subnodes

Marek Vasut (2):
   kernel-fit-image.bbclass: Do not include kernel property in DTBO
     config subnodes
   kernel-fit-image.bbclass: Fix operation with KERNEL_DTBVENDORED = "1"

Quoted 3 lines

Ryan Eatmon (1):
   kernel-fit-image: Add KERNEL_DTBVENDORED support for FIT_CONF_DEFAULT_DTB

Yes. This is the patches I have for this work.

Quoted 10 lines

And also important is: We need time for testing on master before back-
porting to stable. Six weeks is probably enough. But the exact same
pattern ended up with a broken stable branch maybe one year ago. And
also this time there was a state on master which was at least incomplete (three instead of four patches). But still, I think this is
ready for stable now.

Quoted 6 lines

I agree on the series needing testing. One more argument for waiting for
a wrynose test.
See
https://lore.kernel.org/all/[email protected]/
(from the same thread you linked)

I ran this now and got 8 commits which cleanly apply on the wrynose branch, maybe that is all that is missing for the test part ?

$ git log --oneline --reverse --grep=fitimage 
origin/wrynose..origin/master --
meta/lib/oeqa/selftest/cases/fitimage.py
e46d8473fb3 oeqa/selftest/fitimage: fix missing whitespace around
assignment
0344e652a40 oe-selftest: fitimage: replace _gen_random_file with 
_gen_elf64_dummy
5592be53348 oe-selftest: fitimage: replace bbb-dtbs-as-ext with 
test-dtbs-as-ext
befc103a95d oe-selftest: fitimage: replace MACHINE==qemux86-64 guards
with KERNEL_SETUP_BIN
4b08cf6048b oe-selftest: fitimage: add machine settings table and
skip 
Quoted 9 lines
helpers
f17f4263c07 oe-selftest: fitimage: stop hardcoding MACHINE and DISTRO
in 
tests
85e0408a81b oe-selftest: fitimage: Do not expect kernel property in
DTBO 
config subnodes
de2e11e6362 oe-selftest: fitimage: Add tests for KERNEL_DTBVENDORED

It's this series which makes back-porting the tests hard, I agree:
https://patchwork.yoctoproject.org/project/oe-core/list/?series=47907&state=%2A&archive=both

Question is what's simpler and more appropriate: Back-porting major changes for "only" test code or adding test coverage to the old test code on wrynose. Technically, I would have sent the oe-selftest rework earlier if I had it available. So, back-porting=cherry-picking would make sense and be straight forward. Process wise, I would probably stick to just adding test coverage for the changes on wrynose. That should be easy doable, it's just not cherry-pickable.

Regards
Adrian

Quoted 11 lines

-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group. View/Reply Online (#242350):
https://lists.openembedded.org/g/openembedded-core/message/242350
Mute This Topic: https://lists.openembedded.org/mt/120043543/4454582
Group Owner: [email protected] Unsubscribe:
https://lists.openembedded.org/g/openembedded-core/unsub [
[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-

#16 lore raw
Quoted 28 lines · adrian.freihofer
Quoted 26 lines · Marek Vasut via lists.openembedded.org

On 7/30/26 2:39 PM, Yoann Congal wrote:

Hello everyone,

Quoted 21 lines
Quoted 8 lines
Quoted 7 lines

The fitimage fixes landed in oe-core master 6 weeks ago and they fix
fitImage usability issues which block me from updating to wrynose in
multiple projects, that's why I keep monitoring this thread so, but
again, please do not read this as a push.

Quoted 5 lines

I think going with the four patches from
https://lore.kernel.org/all/[email protected]/
without a fifth patch which adds test coverage for this fairly isolated
detail is fine.

This is where I disagree. Last time I reviewed this series it looked
good but I had no way to check it actually worked once backported to
wrynose. That is why I asked Ryan to backport the test from master to
wrynose.

Quoted 3 lines
Quoted 3 lines

Is there a git tree/branch with those patches ? If so, I can at least test this locally and confirm whether or not anything is broken on my end ?

I pushed that here:
https://git.openembedded.org/openembedded-core-contrib/log/?h=ycongal/wrynose/kernel-fit

Quoted 68 lines
Quoted 55 lines
Quoted 32 lines
Quoted 15 lines

Important is: We need 4 not only 3 patches:

Adrian Freihofer (1):
   oe-selftest: fitimage: Do not expect kernel property in DTBO config
     subnodes

Marek Vasut (2):
   kernel-fit-image.bbclass: Do not include kernel property in DTBO
     config subnodes
   kernel-fit-image.bbclass: Fix operation with KERNEL_DTBVENDORED = "1"

Quoted 3 lines

Ryan Eatmon (1):
   kernel-fit-image: Add KERNEL_DTBVENDORED support for FIT_CONF_DEFAULT_DTB

Yes. This is the patches I have for this work.

Quoted 10 lines

And also important is: We need time for testing on master before back-
porting to stable. Six weeks is probably enough. But the exact same
pattern ended up with a broken stable branch maybe one year ago. And
also this time there was a state on master which was at least incomplete (three instead of four patches). But still, I think this is
ready for stable now.

Quoted 6 lines

I agree on the series needing testing. One more argument for waiting for
a wrynose test.
See
https://lore.kernel.org/all/[email protected]/
(from the same thread you linked)

I ran this now and got 8 commits which cleanly apply on the wrynose branch, maybe that is all that is missing for the test part ?

$ git log --oneline --reverse --grep=fitimage 
origin/wrynose..origin/master --
meta/lib/oeqa/selftest/cases/fitimage.py
e46d8473fb3 oeqa/selftest/fitimage: fix missing whitespace around
assignment
0344e652a40 oe-selftest: fitimage: replace _gen_random_file with 
_gen_elf64_dummy
5592be53348 oe-selftest: fitimage: replace bbb-dtbs-as-ext with 
test-dtbs-as-ext
befc103a95d oe-selftest: fitimage: replace MACHINE==qemux86-64 guards
with KERNEL_SETUP_BIN
4b08cf6048b oe-selftest: fitimage: add machine settings table and
skip 
Quoted 9 lines
helpers
f17f4263c07 oe-selftest: fitimage: stop hardcoding MACHINE and DISTRO
in 
tests
85e0408a81b oe-selftest: fitimage: Do not expect kernel property in
DTBO 
config subnodes
de2e11e6362 oe-selftest: fitimage: Add tests for KERNEL_DTBVENDORED

It's this series which makes back-porting the tests hard, I agree:
https://patchwork.yoctoproject.org/project/oe-core/list/?series=47907&state=%2A&archive=both

Quoted 7 lines

Question is what's simpler and more appropriate: Back-porting major changes for "only" test code or adding test coverage to the old test code on wrynose. Technically, I would have sent the oe-selftest rework earlier if I had it available. So, back-porting=cherry-picking would make sense and be straight forward. Process wise, I would probably stick to just adding test coverage for the changes on wrynose. That should be easy doable, it's just not cherry-pickable.

Those cherry-picks do not look particularly good, please avoid them if we can...

If you can easily send a single patch that test the series on wrynose, that does sound better.

Thanks!

Signature · 2 lines
Yoann Congal
Smile ECS
#17 lore raw

On 7/30/26 3:49 PM, Yoann Congal wrote:

Hello everyone,

Quoted 22 lines
Quoted 22 lines
Quoted 22 lines
Quoted 21 lines
Quoted 8 lines
Quoted 7 lines

The fitimage fixes landed in oe-core master 6 weeks ago and they fix
fitImage usability issues which block me from updating to wrynose in
multiple projects, that's why I keep monitoring this thread so, but
again, please do not read this as a push.

Quoted 5 lines

I think going with the four patches from
https://lore.kernel.org/all/[email protected]/
without a fifth patch which adds test coverage for this fairly isolated
detail is fine.

This is where I disagree. Last time I reviewed this series it looked
good but I had no way to check it actually worked once backported to
wrynose. That is why I asked Ryan to backport the test from master to
wrynose.

Quoted 6 lines
Quoted 3 lines
Quoted 3 lines

Is there a git tree/branch with those patches ? If so, I can at least test this locally and confirm whether or not anything is broken on my end ?

I pushed that here:
https://git.openembedded.org/openembedded-core-contrib/log/?h=ycongal/wrynose/kernel-fit

This works for my use cases.

[...]

Quoted 37 lines
Quoted 29 lines
Quoted 16 lines

I ran this now and got 8 commits which cleanly apply on the wrynose branch, maybe that is all that is missing for the test part ?

$ git log --oneline --reverse --grep=fitimage
origin/wrynose..origin/master --
meta/lib/oeqa/selftest/cases/fitimage.py
e46d8473fb3 oeqa/selftest/fitimage: fix missing whitespace around
assignment
0344e652a40 oe-selftest: fitimage: replace _gen_random_file with
_gen_elf64_dummy
5592be53348 oe-selftest: fitimage: replace bbb-dtbs-as-ext with
test-dtbs-as-ext
befc103a95d oe-selftest: fitimage: replace MACHINE==qemux86-64 guards
with KERNEL_SETUP_BIN
4b08cf6048b oe-selftest: fitimage: add machine settings table and
skip
Quoted 9 lines
helpers
f17f4263c07 oe-selftest: fitimage: stop hardcoding MACHINE and DISTRO
in
tests
85e0408a81b oe-selftest: fitimage: Do not expect kernel property in
DTBO
config subnodes
de2e11e6362 oe-selftest: fitimage: Add tests for KERNEL_DTBVENDORED

It's this series which makes back-porting the tests hard, I agree:
https://patchwork.yoctoproject.org/project/oe-core/list/?series=47907&state=%2A&archive=both

Quoted 7 lines

Question is what's simpler and more appropriate: Back-porting major changes for "only" test code or adding test coverage to the old test code on wrynose. Technically, I would have sent the oe-selftest rework earlier if I had it available. So, back-porting=cherry-picking would make sense and be straight forward. Process wise, I would probably stick to just adding test coverage for the changes on wrynose. That should be easy doable, it's just not cherry-pickable.

Quoted 5 lines

Those cherry-picks do not look particularly good, please avoid them if we can...

If you can easily send a single patch that test the series on wrynose, that does sound better.

I submitted the backported test, but the test required adjustments to not depend on the other commits. I think the adjustments are harmless and the test still works as intended:

[wrynose][PATCH] oe-selftest: fitimage: Add tests for KERNEL_DTBVENDORED

Signature · 2 lines
Best regards,
Marek Vasut
#18 lore raw

On 7/31/26 4:34 AM, Marek Vasut wrote:

Hello again,

+CC Paul , it would be good to finally land those fitImage bugfix backports in OE wrynose , without them I cannot use OE wrynose with fitImages .

Quoted 30 lines
Quoted 22 lines
Quoted 22 lines
Quoted 22 lines
Quoted 21 lines
Quoted 8 lines
Quoted 7 lines

The fitimage fixes landed in oe-core master 6 weeks ago and they fix
fitImage usability issues which block me from updating to wrynose in
multiple projects, that's why I keep monitoring this thread so, but
again, please do not read this as a push.

Quoted 5 lines

I think going with the four patches from
https://lore.kernel.org/all/ [email protected]/ without a fifth patch which adds test coverage for this fairly isolated
detail is fine.

This is where I disagree. Last time I reviewed this series it looked
good but I had no way to check it actually worked once backported to
wrynose. That is why I asked Ryan to backport the test from master to
wrynose.

Quoted 6 lines
Quoted 3 lines
Quoted 3 lines

Is there a git tree/branch with those patches ? If so, I can at least test this locally and confirm whether or not anything is broken on my end ?

I pushed that here:
https://git.openembedded.org/openembedded-core-contrib/log/?h=ycongal/ wrynose/kernel-fit

This works for my use cases.

[...]

Quoted 3 lines

I submitted the backported test, but the test required adjustments to not depend on the other commits. I think the adjustments are harmless and the test still works as intended:

[wrynose][PATCH] oe-selftest: fitimage: Add tests for KERNEL_DTBVENDORED

Signature · 2 lines
Best regards,
Marek Vasut