Re: [PATCH 2/5] arm64: dts: qcom: sc8180x: Add Microsoft Surface Pro X
From: Val Packett
Date: Sun Nov 15 2026 - 03:20:44 EST
On 11/15/26 12:16 AM, Haipeng Lin wrote:
On 11/4/26 2:04 AM, Val Packett wrote:Huuuh, okay..
From: Maximilian Luz <luzmaximilian@xxxxxxxxx>[...]
Introduce a device tree for the Surface Pro X tablet in all 3 variants
(SQ1 with cellular, SQ1 Wi-Fi only, SQ2 with cellular).
I believe there is also a Wi-Fi only SQ2 variant. Looking at the WoA
drivers for 8180_CAS (Caspar, Wi-Fi only)
https://github.com/WOA-Project/Qualcomm-Reference-Drivers/tree/master/Surface/8180_CAS/200.0.31.0
This variant is the only that has both SurfaceSMF_SQ1.cab and
SurfaceSMF_SQ2.cab. Only 8180_CAS MPSS INF maps the no-modem subsystem
ID (SUBSYS_MSNM8180 -> qcmpss8180_nm.mbn), the 8180_CAR (Carina) only
lists the cellular MSWW/MSJP/MSCN IDs.
My SQ2 Wi-Fi boots with the caspar dts in this patch fine but the
gold cluster is pinned at SQ1's top clocks because the 3.072 / 3.1488
GHz OPPs are only in carina.dts.
diff --git a/arch/arm64/boot/dts/qcom/sc8180x-microsoft-carina.dts b/arch/arm64/boot/dts/qcom/sc8180x-microsoft-carina.dts[...]
new file mode 100644
index 000000000000..5d155c69907a
--- /dev/null
+++ b/arch/arm64/boot/dts/qcom/sc8180x-microsoft-carina.dts
@@ -0,0 +1,50 @@
+// SPDX-License-Identifier: BSD-3-Clause
+
+/dts-v1/;
+
+#include "sc8180x-microsoft-surface-pro-x.dtsi"
+
+/ {
+ model = "Microsoft Surface Pro X (SQ2)";
+ compatible = "microsoft,carina", "qcom,sc8180x";
+};
+
+&cpu4_opp_table {
+ opp-3072000000 {
+ opp-hz = /bits/ 64 <3072000000>;
+ opp-peak-kBps = <8368000 51609600>;
+ };
+
+ opp-3148800000 {
+ opp-hz = /bits/ 64 <3148800000>;
+ opp-peak-kBps = <8368000 51609600>;
+ };
+};
+
Would it make sense to move opp-3072000000 and opp-3148800000 into the
shared sc8180x-microsoft-surface-pro-x.dtsi? qcom-cpufreq-hw disables
all DT OPPs first and only re-enables freqs present in the hardware LUT
so this would work for SQ1 and caspar.dtb will cover both the SQ1 and
SQ2 Wi-Fi models out of the box and carina.dts won't need its own
&cpu4_opp_table override.
I'm pretty sure mine *did* complain very loudly about having these OPPs and might've even had broken scaling.
I'll retest again, if it works I'll move it to the shared dtsi (or even the SoC's?)
Yeahh, forgot to include, will be in v2diff --git a/arch/arm64/boot/dts/qcom/sc8180x-microsoft-surface-pro-x.dtsi b/arch/arm64/boot/dts/qcom/sc8180x-microsoft-surface-pro-x.dtsi[...]
new file mode 100644
index 000000000000..17fb01c6b164
--- /dev/null
+++ b/arch/arm64/boot/dts/qcom/sc8180x-microsoft-surface-pro-x.dtsi
+&smb2351_2 {Confirmed on hardware that labeling pmic@2 and pmic@a in
+ status = "disabled";
+};
+
+&smb2351_a {
+ status = "disabled";
+};
sc8180x-pmics.dtsi fixes the dtc build failure here.
[...]> +&uart13 {Of course; it should've been the _2 in the first place but I made a mistake during last-minute cleanup and left out the wrong ones…
+ pinctrl-0 = <&uart13_state>;[...]
+ pinctrl-names = "default";
+
+ status = "okay";
On linux-surface, ad1bb617a15a from Maximilian Luz had
&qupv3_id_2 {
status = "okay";
};
Would it be helpful to add here? I confirmed on hardware that adding it
allows uart13 (WCN3998 Bluetooth) and uart15 (SAM EC) to populate and
probe properly.
Thanks,
Haipeng
BTW, please send me the output of `sudo fwupdtool hwids`, these are needed to add support to dtbloader/systemd-stub/stubble DT auto-selection.
(If anyone with the cellular variants reads this: you too :))
Thanks,
~val