Repository navigation
[26.04_linux-nvidia-bos] NVIDIA: SAUCE: pinctrl: mt8901: sync wake EINT event mask - #636
kbutala-nv wants to merge 1 commit into
Conversation
BaseOS Kernel ReviewWarning
|
PR Validation ReportPatchscan ✅ No Missing FixesAll cherry-picked commits checked — no missing upstream fixes found. PR Lint ✅ All checks passedDetailsChecking 1 commits... Cherry-pick digest: ┌──────────────┬──────────────────────────────────────────────────────────────────┬────────────┬─────────┬───────────────────────────┐ │ Local │ Referenced upstream / Patch subject │ Patch-ID │ Subject │ SoB chain │ ├──────────────┼──────────────────────────────────────────────────────────────────┼────────────┼─────────┼───────────────────────────┤ │ cd0924659926 │ [SAUCE] pinctrl: mt8901: sync wake eint event mask │ N/A │ N/A │ kbutala │ └──────────────┴──────────────────────────────────────────────────────────────────┴────────────┴─────────┴───────────────────────────┘ Lint: all checks passed. |
Please add the following Please also add Launchpad #2170130 as the first line of the PR description. |
MT8901 EINT has a separate event mask that controls whether an EINT can be forwarded as a wake event. The current driver clears that mask only for the static SoC list, which covers the ACPI event line. Firmware-described devices can be marked wake-capable independently, so PM arms their IRQs through the normal irq_set_wake path. Clear the hardware event mask when the IRQ core enables such an EINT as a wake source so keyboard, touchpad, and other firmware-declared EINT wake sources can reach the SPM without adding board-specific EINT numbers. Restore the hardware event mask when wake is disabled, otherwise a pin that was armed once would remain an SPM wake event until reboot. Keep pins from the SoC static event-unmask list unmasked so permanently required wake events such as the ACPI event line remain armed. Make event-mask register programming opt-in for SoCs that declare support for it, and re-apply both the static list and dynamically armed wake_mask pins on resume in case the event register state is lost across system sleep. Fixes: 927eeda ("NVIDIA: SAUCE: pinctrl: mt8901: Enable ACPI wake-event EINT support") Signed-off-by: Kaushal Rajeev Butala <kbutala@nvidia.com>
369dbea to
cd09246
Compare
Done! |
|
|
|
Merged, closing PR. |
LP: https://bugs.launchpad.net/bugs/2170130
MT8901 EINT has a separate event mask that controls whether an EINT can be forwarded as a wake event. The current driver clears that mask only for the static SoC list, which covers the ACPI event line.
Firmware-described devices can be marked wake-capable independently, so PM arms their IRQs through the normal irq_set_wake path. Clear the hardware event mask when the IRQ core enables such an EINT as a wake source so keyboard, touchpad, and other firmware-declared EINT wake sources can reach the SPM without adding board-specific EINT numbers.
Restore the hardware event mask when wake is disabled, otherwise a pin that was armed once would remain an SPM wake event until reboot. Keep pins from the SoC static event-unmask list unmasked so permanently required wake events such as the ACPI event line remain armed.
Make event-mask register programming opt-in for SoCs that declare support for it, and re-apply both the static list and dynamically armed wake_mask pins on resume in case the event register state is lost across system sleep.
Fixes: 927eeda ("NVIDIA: SAUCE: pinctrl: mt8901: Enable ACPI wake-event EINT support")
This is needed for internal bug 6839927