On Tue, May 20, 2025 at 04:29:18PM +0200, Konrad Dybcio wrote: > On 5/19/25 11:35 AM, Johan Hovold wrote: > > On Sat, May 17, 2025 at 07:27:49PM +0200, Konrad Dybcio wrote:
> >> Dmitry Baryshkov (1): > >> arm64: dts: qcom: sc8280xp-lenovo-thinkpad-x13s: enable sensors DSP > > > > At first I was worried that missing firmware could cause issues here > > (e.g. drivers not reaching sync state as with venus), but Lenovo has > > indeed released the SLPI firmware already. > >> Konrad Dybcio (4): > >> dt-bindings: remoteproc: qcom,sm8350-pas: Add SC8280XP > >> arm64: dts: qcom: sc8280xp: Fix node order > >> arm64: dts: qcom: sc8280xp: Add SLPI > > > >> arm64: dts: qcom: sc8280xp-crd: Enable SLPI > > > > Without firmware this results in errors like: > > > > remoteproc remoteproc0: slpi is available > > remoteproc remoteproc0: Direct firmware load for > > qcom/sc8280xp/qcslpi8280.mbn failed with error -2 > > remoteproc remoteproc0: powering up slpi > > remoteproc remoteproc0: Direct firmware load for > > qcom/sc8280xp/qcslpi8280.mbn failed with error -2 > > remoteproc remoteproc0: request_firmware failed: -2 > > > > but enabling for the CRD reference design and requiring users (read: > > developers) to copy it from Windows should be OK. > > We shouldn't expect non-developers to have the CRD on hand, right? ;) That was my point: requiring user of the CRD to install firmware manually is fine (e.g. as they would all be developers and currently all firmware for the CRD needs to be copied from Windows anyway). But we should not enable things like venus (that can prevent the interconnect driver from reaching the sync state) on the X13s which has lots of regular users until the corresponding firmware has been released and is in linux-firmware. Johan
