iq-9075-evk: build sdcc feature DTs as overlays by mchunara007 · Pull Request #1952 · qualcomm-linux/meta-qcom

mchunara007 · 2026-04-15T12:09:24Z

LeMans EVK board supports either eMMC or SD-card, but only one can be
active at a time due to shared host controller. The storage medium is
selected via a dip switch.

Enable metadata based selection for eMMC and SD card as overlays to support both the medium based on the DIP switch selection on the EVK board.

lumag

Commit message should start by the description of the problem the commit tries to solve.

lumag · 2026-04-16T21:06:55Z

@@ -32,6 +32,18 @@ FIT_DTB_COMPATIBLE[lemans-evk+lemans-evk-ifp-mezzanine] = " \
    qcom,qcs9075-iot-subtype1 \
    qcom,qcs9075v2-iot-subtype1 \
 "
+FIT_DTB_COMPATIBLE[lemans-evk+lemans-evk-sd-card] = " \


So, what happens with the default setting?

By default, the UEFI would populate 'sdcard' metadata. only in case of eMMC attached, 'emmc' will get populated. So by default this would select the lemans-evk-sd-card overlay

Then why do we need the -sd-card at all?? Why can't it be a default here too, being overriden only if emmc is attached.

s/attached/selected/

If SD-card-specific properties are added unconditionally in the base DT, certain boolean flags such as no-mmc would conflict with the eMMC use case. These properties also cannot be removed later through the eMMC overlay because of DT overlay limitations. On the other hand, if such flags are omitted, the current driver probe flow attempts SD, MMC, and SDIO detection, which leads to failure logs for unsupported modes. Such logs are not acceptable for certification requirements. Similarly, on the Monaco board, where eMMC is the default configuration, eMMC-specific properties such as supports-cqe for command queue support would conflict with SD-card operation.

I didn't suggest adding uSD properties to the base DT. However, corresponding dtso can be selected by default.

Ok, thanks for clarification. Updated the commit to load the SD card as default case and overlay eMMC only when eMMC is detected and metadata is populated with the string accordingly.

This aligns with the default DIP switch position on the boards.

lumag · 2026-04-16T21:07:40Z

+    qcom,qcs9075-iot-subtype1-sdcard \
+    qcom,qcs9075v2-iot-subtype1-sdcard \
+"
+FIT_DTB_COMPATIBLE[lemans-evk+lemans-evk-ifp-mezzanine+lemans-evk-emmc] = " \


Can there be a non-ifp board with eMMC?

For LeMans EVK, the eMMC is present on ifp mezz board, so non-ifp board with eMMC is not possible.

Is it explained in the commit message? No.

Updated the commit message to explain the eMMC being IFP specific case.

mchunara007 · 2026-04-17T11:16:22Z

Commit message should start by the description of the problem the commit tries to solve.

Updated the commit message with problem statement.

ricardosalveti · 2026-05-05T01:12:44Z

Please rebase to fix the conflicts.

LeMans EVK board supports either eMMC or SD-card, but only one can be active at a time due to shared host controller. The storage medium is selected via a dip switch. Add the sd card and eMMC device tree overlay for the LeMans EVK to KERNEL_DEVICETREE in the iq-9075-evk machine configuration to ensure device tree coverage for LeMans sdcc storage usecase. Signed-off-by: Monish Chunara <monish.chunara@oss.qualcomm.com>

LeMans EVK can use either SD card or eMMC depending on the board configuration selected through the DIP switch. However, eMMC support is available only when the IFP mezzanine board is present, so a non-IFP eMMC configuration is not a valid hardware combination. Build the SDCC-related DTs as overlays and use metadata-based selection to pick the appropriate storage configuration for the detected board variant. This keeps the valid SD-card and IFP+eMMC combinations explicitly modeled without introducing unsupported configurations. Signed-off-by: Monish Chunara <monish.chunara@oss.qualcomm.com>

mchunara007 · 2026-05-12T10:46:27Z

Rebased and conflicts resolved.

ricardosalveti · 2026-05-12T19:53:04Z

["ERROR: linux-yocto-6.18.24+git-r0 do_generate_qcom_fitimage: FIT_DTB_COMPATIBLE[lemans-evk] is not set for 'lemans-evk.dtb'.\n", 'ERROR: Logfile of failure stored in: /work/build/tmp/work/qcom_armv8a-qcom-linux/linux-yocto/6.18.24+git/temp/log.do_generate_qcom_fitimage.292270\n', "ERROR: Task (/work/build/../oe-core/meta/recipes-kernel/linux/linux-yocto_6.18.bb:do_generate_qcom_fitimage) failed with exit code '1'\n"]

ricardosalveti · 2026-05-12T19:54:19Z

You changed the default but included a dtbo that is only available in our own kernel trees.

mchunara007 · 2026-05-27T07:39:29Z

You changed the default but included a dtbo that is only available in our own kernel trees.

Can you please help to understand which tree the DT changes should be merged in. Is it the linux-next where we expect the DT files to be present? On qcom-next and qcom-6.18.y branch the DT changes are merged and available.

lumag · 2026-05-28T04:09:49Z

You changed the default but included a dtbo that is only available in our own kernel trees.

Can you please help to understand which tree the DT changes should be merged in. Is it the linux-next where we expect the DT files to be present? On qcom-next and qcom-6.18.y branch the DT changes are merged and available.

See LINUX_QCOM_KERNEL_DEVICETREE.

ricardosalveti · 2026-05-29T21:10:14Z

With #2305 you should be able to separate what is qcom specific in the fit-dtb-compatible as well.

mchunara007 requested review from anujm1, koenkooi, lumag, ndechesne, ricardosalveti, sbanerjeeqc and vkraleti as code owners April 15, 2026 12:09

lumag requested changes Apr 15, 2026

View reviewed changes

ricardosalveti reviewed Apr 15, 2026

View reviewed changes

Comment thread conf/machine/iq-9075-evk.conf

mchunara007 force-pushed the lemans_fitdtb_sdcc branch from 88b810e to f8ee98d Compare April 16, 2026 18:40

lumag reviewed Apr 16, 2026

View reviewed changes

mchunara007 force-pushed the lemans_fitdtb_sdcc branch from f8ee98d to 281a5b2 Compare April 17, 2026 11:15

mchunara007 force-pushed the lemans_fitdtb_sdcc branch from 281a5b2 to a70cdf7 Compare April 17, 2026 11:32

mchunara007 force-pushed the lemans_fitdtb_sdcc branch from a70cdf7 to f943755 Compare April 30, 2026 15:56

mchunara007 added 2 commits May 12, 2026 16:02

mchunara007 force-pushed the lemans_fitdtb_sdcc branch from f943755 to 6849a74 Compare May 12, 2026 10:44

lumag approved these changes May 12, 2026

View reviewed changes

This comment was marked as off-topic.

Sign in to view

Conversation

mchunara007 commented Apr 15, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

lumag left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

mchunara007 commented Apr 17, 2026

Uh oh!

ricardosalveti commented May 5, 2026

Uh oh!

mchunara007 commented May 12, 2026

Uh oh!

ricardosalveti commented May 12, 2026

Uh oh!

ricardosalveti commented May 12, 2026

Uh oh!

This comment was marked as off-topic.

Uh oh!

mchunara007 commented May 27, 2026

Uh oh!

lumag commented May 28, 2026

Uh oh!

ricardosalveti commented May 29, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

mchunara007 commented Apr 15, 2026 •

edited

Loading