Skip to content

Pixel 8a (akita): buttons, and suspend-to-idle by default - #238

Merged
derekross merged 2 commits into
derekross:omamobilefrom
aninmucek:omamobile-zuma-akita-input
Nov 11, 2026
Merged

derekross merged 2 commits into
derekross:omamobilefrom
aninmucek:omamobile-zuma-akita-input

Conversation

@aninmucek

@aninmucek aninmucek commented Nov 10, 2026 •

Copy link
Copy Markdown
Contributor

The Pixel 8a's buttons, and suspend-to-idle as the default sleep, on today's omamobile. These are the first two steps of akita's M3 (input). Touch follows once the supply question below is settled.

  • arm64: dts: exynos: google: zuma-akita: Add the buttons.
    • Volume down, volume up and power are wired as on the 8 Pro: gpa4-1, gpa6-2 and gpa4-0, active low, all wake sources.
    • Google's zuma-common.dtsi has them, and the stock phone's /proc/interrupts lists the same three gpio-keys lines.
    • So this is zuma-husky's gpio-keys node as it is, with zuma-pinctrl.dtsi's key pins.
  • arm64: dts: exynos: google: zuma-akita: Default to suspend-to-idle. mem_sleep_default=s2idle in the bootargs, as on the 7 Pro and the 7. This is your suggestion from Pixel 8a (akita): enable UFS, as on the Pixel 8 Pro #206's review: under Arch, /sys/power/mem_sleep read s2idle [deep] (BRINGUP-M2, M2.6).

Tested on a Pixel 8a from slot B (7.3.0-rc6, these two commits on 253208c11):

  • gpio-keys is an input device with all three keys.
  • One short press of each raised its interrupt twice (press and release).
  • A short Power press wakes the screen through the lock, and logind logs it as a short press without powering the phone off.
  • /sys/power/mem_sleep reads [s2idle] deep.
  • clk: samsung: exynos-clkout: NFC clock pad on the Pixel 8 Pro (zuma) #209's fix is confirmed on the 8a: this kernel has it, and the two exynos-clkout lines are gone from the boot log. tools/phone-smoke from omarchy-mobile's main reads buttons ok, with 2 known error lines and 0 new.

tools/kernel-pr-check --build OUT --merge origin/omamobile, with the dtschema validator, passes everything:

  • a clean merge onto 253208c11;
  • the sign-offs: 2 commits, each signed off by its author, none by an AI;
  • checkpatch --strict with no errors. Its one warning is the bootargs line's length (229 columns). It was already 204 before this, and a bootargs string can't be wrapped;
  • the 10 Google device trees build with W=1 and pass dtbs_check, with no new warnings.

A question for the next step (touch). The 8a's touchscreen is a Goodix GT9895, the 7a's chip. Its supplies are the S2MPG14's LDO25M (1.8 V) and LDO4M, and the stock phone runs LDO4M at 3.0 V. With LDO4M fixed at 3.0 V in the board file, the regulator probe fails: touch-avdd: failed to apply 3000000-3000000uV constraint: -EINVAL.

That's because the S2MPG14 LDO ops (5e4cda3a7, from the 9a work) read the voltage selector but deliberately don't set it. A test boot with the driver's whole range, no consumer and regulator_ignore_unused read what ABL leaves on this phone:

  • LDO4M: 3.3 V, off;
  • LDO25M: 1.8 V, off.

So with today's driver, touch would get 3.3 V where stock uses 3.0 V. The 7a runs the same chip at 3.0 V, and the S2MPG12's LDO ops there have set_voltage_sel.

Would you take a small change that adds set_voltage_sel (regulator_set_voltage_sel_regmap) to the S2MPG14 LDO ops? With fixed min = max constraints, the core would then set stock's 3.0 V before anything enables the rail. As far as I can see it changes nothing on the other boards:

  • zumapro-tegu gives these regulators no voltage constraints, so the core never applies one.
  • zuma-husky's fixed 3.3 V is what the core already accepts there, which suggests its ABL leaves 3.3 V too. On the 8a it does (above), but that's not checked on an 8 Pro.

Or is there a reason to keep the selectors firmware-owned that I'm missing? I won't power the touch chip until this is settled.

tools/phone-smoke report (this PR's kernel, omarchy-mobile main 6a9251f)

phone-smoke: akita

Kernel 7.3.0-rc6-omamobile-push1-g23120efb6488, tools at 6a9251f, 2026-11-10T16:07Z.

Check Result Detail
boot (required) ok kernel 7.3.0-rc6-omamobile-push1-g23120efb6488, slot b, up 194s
storage guard (required) ok 62 partitions; only userdata and slot B's boot images writable
root (required) ok ext4 on userdata, 98G free
kernel health (required) ok no oops or BUG; 0 WARN, 2 error lines in dmesg (2 known, 0 new)
systemd units ok none failed
USB gadget (required) ok bound to 11210000.usb
display ok card0-Unknown-1 connected, 1080x2400
GPU not ok no render node
session (required) ok Hyprland and the Omarchy shell running
lock ok PAM file present, the shell can lock (passwordPam true, locked true)
touch not ok no multitouch input device
buttons ok gpio-keys
Wi-Fi skip no wireless interface
Bluetooth skip no controller
battery skip no battery power supply
sound cards skip none
sensors skip no sensor proxy
cameras skip no libcamera tools

Packages: hyprland 0.57.0-3.2, libcamera 0.7.2-4.23, mesa 1:26.2.4-1.4, omacam 0.3.3-1, omamessages 0.2.2-1, omamobile-apps 0.2.2-1, omamobile-device-akita 0.1.1-1, omamobile-firstboot 0.1.3-1, omamobile-keyring 20261026-1, omamobile-mirror 0.1.1-1, omamobile-omarchy 0.1.38-1, omamobile-owl-widgets 0.1.2-1, omamobile-shell 0.2.108-1, omamobile-voice 0.1.0-1, omaphone 0.2.3-1, omarchy 4.0.6-1, omatel-agent 0.2.4-2, quickshell 0.3.2.r7.g2b4501a-1

🤖 Generated with Claude Code

The Pixel 8a's volume and power keys are wired as on the Pixel 8 Pro:
volume down on gpa4-1, volume up on gpa6-2 and power on gpa4-0, all
active low and wake sources (Google's zuma-common.dtsi; the stock
phone's /proc/interrupts lists the same three gpio-keys lines). Add the
same gpio-keys node as zuma-husky, with zuma-pinctrl.dtsi's key pins.

Without it the power key is not an input device, so nothing on the
phone can wake the screen once the lock has blanked it.

Tested on a Pixel 8a from slot B (7.3.0-rc6): gpio-keys registers as an
input device with the three keys; one short press of each raises its
interrupt twice (press and release), and a short Power press wakes the
screen through the session's lock without powering the phone off.

Assisted-by: Claude Opus 5.5
Signed-off-by: DarekM <149845949+aninmucek@users.noreply.github.com>
The firmware advertises PSCI SYSTEM_SUSPEND, so the kernel offers deep
sleep and makes it the default: /sys/power/mem_sleep reads
"s2idle [deep]" on the Pixel 8a. Linux has no zuma sleep sequence yet,
so a deep suspend would not wake. Pass mem_sleep_default=s2idle, as the
Pixel 7 Pro and the Pixel 7 do, so a suspend that names no state stays
in s2idle. The omamobile-device-akita package already asks systemd for
s2idle; this covers anything that writes "mem" directly.

Tested on a Pixel 8a from slot B (7.3.0-rc6): /proc/cmdline carries
mem_sleep_default=s2idle and /sys/power/mem_sleep reads "[s2idle] deep".

Suggested-by: Derek Ross <derekross@gmail.com>
Assisted-by: Claude Opus 5.5
Signed-off-by: DarekM <149845949+aninmucek@users.noreply.github.com>
@derekross

Copy link
Copy Markdown
Owner

Thanks, @aninmucek. Both commits are in good shape and I'll merge them.

What I checked: the change is confined to zuma-akita.dts, and in a build on today's omamobile the 7 Pro's, 8 Pro's and 9a's DTBs come out byte-identical to the base. The gpio-keys node is the 8 Pro's line for line with the existing pin groups, which is what the rule for adding a node enabled asks for, and you tested it on your 8a. The sleep default has the evidence its claim needs: /sys/power/mem_sleep now reads [s2idle] deep. It doesn't claim a suspend cycle, and none should run yet; BRINGUP-M2 keeps Suspend out until s2idle has been tried on Tensor G3. The gate passes on the merge (sign-offs, prefixes, checkpatch with the one unavoidable bootargs line length, W=1 on the ten device trees), and CI's dtbs_check passed on your head.

Two small things:

  • Please paste the tools/phone-smoke report itself into the description, not only its summary.
  • For later, not this PR: both volume keys have no internal pull and are wake sources, as on the 8 Pro. When you first try s2idle, check those two pins first if the phone wakes by itself.

Your touch question. Yes, I'd take set_voltage_sel for the S2MPG14 LDOs, as its own PR, with these three things in it:

  1. The selector table for LDO4M (minimum, step, number of steps) cited from Google's s2mpg14-regulator.c, and proven on your 8a before anything consumes the rail: one boot with the fixed 3.0 V constraint and no consumer, then the selector read back and the voltage the core reports. The setter was left out on the 9a because that table had never been checked on hardware.
  2. The other two boards shown unchanged. The 9a's nodes have no voltage constraints, so the core never calls the setter there; please show that from the DTB. The 8 Pro fixes LDO4M at 3.3 V, so the setter would write there only if ABL leaves something else. That needs a read from an 8 Pro owner, or the 8 Pro's constraint path shown not to reach the setter when the voltage already matches.
  3. The touch chip powered in a later boot than the one that first sets the voltage, with the rail measured or read back first. One change per boot.

Thanks for not powering the chip at 3.3 V while this was open. That was the right call.

@derekross
derekross merged commit 2b1c2bc into derekross:omamobile Nov 11, 2026
1 check passed
@aninmucek
aninmucek deleted the omamobile-zuma-akita-input branch November 11, 2026 00:17
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants