gstreamer1.0-plugins-bad: handle padded buffers in wl_shm buffer creation

[tdarote]

Use GstBuffer metadata when constructing wl_shm buffers so stride and size reflect the actual buffer layout. This fixes incorrect rendering with DMA-backed buffers that include padding or custom strides.

[quaresmajose]

This patch is for gstreamer1.0-plugins-bad and not wayland so please rename the PR and commit to gstreamer1.0-plugins-bad: ...

[tdarote]

This patch is for gstreamer1.0-plugins-bad and not wayland so please rename the PR and commit to gstreamer1.0-plugins-bad: ...

done

[lumag]

Why is it being submitted here rather than OE-Core?

[tdarote]

Why is it being submitted here rather than OE-Core?

This patch is intended to live in meta-qcom. We are following the same practice used for other Qualcomm‑specific enablement patches, where required functionality is carried in the BSP layer to support current platforms.
A similar change has already been submitted upstream for review. Once the upstream patch is accepted and becomes available via an upstream GStreamer release (or is otherwise picked up by OE-Core), this patch will be dropped from meta-qcom.
We do not plan to move this patch directly into OE-Core; instead, OE-Core should consume the upstreamed change, after which maintaining a BSP‑local patch will no longer be required.

[lumag]

This patch is intended to live in meta-qcom. We are following the same practice used for other Qualcomm‑specific enablement patches, where required functionality is carried in the BSP layer to support current platforms.

The change is not Qualcomm-specific. It can affect other platforms. Please submit it to OE-Core instead.

[ricardosalveti]

Qualcomm‑specific enablement patches

How is this qualcomm specific?

[github-actions]

Test run workflow

Test jobs for commit 12aa8ca

Test	dragonboard-820c	qcs615-adp-air	qcs6490	qcs8300	qcs9100	qcs9100-rb8	qrb2210-rb1
boot	✅ pass	✅ pass	✅ pass	✅ pass	✅ pass	✅ pass	✅ pass

All jobs summary

Job ID	Device	State	Health
181258	qcs9100	Finished	Complete
181284	qcs6490	Finished	Complete
181259	qcs8300	Finished	Complete
181283	qcs6490	Finished	Complete
181261	qcs8300	Finished	Complete
181278	qcs9100-rb8	Finished	Complete
181277	dragonboard-820c	Finished	Incomplete
181267	qcs9100-rb8	Finished	Complete
181262	qcs6490	Finished	Complete
181265	qcs9100-rb8	Finished	Complete
181271	qcs8300	Finished	Complete
181269	qrb2210-rb1	Finished	Complete
181260	qcs9100	Finished	Complete
181270	qcs8300	Finished	Complete
181281	qrb2210-rb1	Finished	Complete
181276	qcs615-adp-air	Finished	Complete
181266	qcs9100	Finished	Complete
181264	dragonboard-820c	Finished	Complete
181280	qcs615-adp-air	Running	Unknown
181274	qcs6490	Finished	Complete
181282	qcs615-adp-air	Submitted	Unknown
181263	qcs9100	Finished	Complete
181273	qcs615-adp-air	Running	Unknown
181279	qrb2210-rb1	Finished	Complete
181268	dragonboard-820c	Finished	Complete
181275	dragonboard-820c	Finished	Complete
181285	qrb2210-rb1	Finished	Complete

[test-reporting-app]

Test Results

 24 files   -    40   24 suites   - 262   51m 27s ⏱️ - 2h 43m 54s
 13 tests  -    46    9 ✅  -    48  0 💤 ± 0  4 ❌ +2 
193 runs   - 2 499  189 ✅  - 2 474  0 💤  - 25  4 ❌ ±0 

For more details on these failures, see this check.

Results for commit 12aa8ca. ± Comparison against base commit 05ac1c3.

This pull request removes 48 and adds 2 tests. Note that renamed tests count towards both.

0_BT_FW_KMD_Service ‑ BT_FW_KMD_Service
0_core_auth ‑ core_auth
0_hotplug ‑ hotplug
10_AudioRecord ‑ AudioRecord
10_Ethernet ‑ Ethernet
11_DSP_AudioPD ‑ DSP_AudioPD
11_Ethernet ‑ Ethernet
12_DSP_AudioPD ‑ DSP_AudioPD
12_fastrpc_test ‑ fastrpc_test
13_fastrpc_test ‑ fastrpc_test
…

lava ‑ auto-login-action
lava ‑ fastboot-boot

[tdarote]

This patch is intended to live in meta-qcom. We are following the same practice used for other Qualcomm‑specific enablement patches, where required functionality is carried in the BSP layer to support current platforms.

The change is not Qualcomm-specific. It can affect other platforms. Please submit it to OE-Core instead.

if I will go for creating PR on OE core getting below issue as :
An owner of this repository has disabled the ability to open pull requests.

[lumag]

@tdarote why are you trying to submit a PR? Have you read the README inside OE-Core?

[tdarote]

@tdarote why are you trying to submit a PR? Have you read the README inside OE-Core?

ok , got it, we need to send mail to openembedded-core@lists.openembedded.org

[lumag]

@tdarote did you succeed in sending the email?

[tdarote]

@tdarote did you succeed in sending the email?

OE-Core contributions need to be sent via the mailing list. We have an internal review process for upstream submissions, so I’ve created an OSR ticket for this change. Once it’s approved, I’ll send the patch to openembedded-core@lists.openembedded.org.
OSR ticket : https://jira-dc4.qualcomm.com/jira/browse/OSR-21610

Since the upstream timeline is not guaranteed, would it be acceptable to keep this change here as a temporary downstream patch until it is available in OE-Core?

[lumag]

@tdarote you just need to join the contributors.yocto list. I think it's covered already by OSR-12812.

Landing the patch here will create additional troubles if the patch is later accepted to Yocto.

[tdarote]

Hi @lumag,

The upstream change has now been merged:
https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/11370

I’ve also sent a request to oe-core, but we don’t yet have visibility into their merge timeline. Given that, would it be okay to carry a backported patch in meta-qcom until it lands in oe-core?

[ricardosalveti]

It is already merged in master-next (https://git.openembedded.org/openembedded-core/commit/?h=master-next&id=cd081145a764c916ecc8190cc66be3b4ad557a5a), so it should land in master over the next few days.

So no need to land in meta-qcom for now. I believe it is also needed for wrynose, so that would have to be proposed to wrynose as well once it lands in master.