Repository navigation
device-lynx 0.2.0: the device package, the Arch root and the Wi-Fi plan - #88
Conversation
The board file's UFS nodes bring the link up at HS-G4 on two lanes, as on stock, and the storage guard leaves only slot B's boot images and userdata writable. Record the part, the LUN layout and the guard's result. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Add omamobile-device-lynx, a copy of felix's first package without the lid drop-in: the storage guard rule, the USB address, the power-key drop-in, the mirrorlist and reboot-to. It marks no slot successful and starts no hardware. tools/build-packages builds it after felix, and tests/all runs its tests. With it, tools/install-phone put a root image built for lynx on userdata, and the phone booted into Arch and Hyprland on simpledrm. Record the install, the missing RTC and what it does to the keyring, and the lock screen without touch in NOTES. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The bootloader gives slot B three retries after set-active and takes one on every boot that is not marked. Pushing new images four times in a row used them up and the phone fell back to slot A's Android. omamobile-device-lynx 0.2.0 ships the Pixel 7 Pro's boot-success unit and omamobile-bootctl unchanged. lynx's devinfo is sdd1 with the same 3.15 layout. The test checks both copies match the cheetah package and runs a bootctl dry run against a synthetic devinfo. NOTES records the touchscreen, the PMIC RTC, and the slot retries. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The QCA6490 enumerates and reads host memory, but its boot ROM stalls before checking the first firmware hash. Notes cover the power pins, the image bisect, what is ruled out, and the stock differences left to try. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The descriptor dropped BRINGUP_PROBE and gained the rollback-index comment; the package ships a byte-identical copy, which its test checks. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Record the last verified integration state and distinguish historical first-light behavior from the current installation. Android slot A cannot boot with Arch userdata, so physical buttons are the recovery route. Assisted-by: LLM
The QCA6490's boot ROM reads the ELF header and the hash segment of an image byte-identical to stock's, then stops at stage 0x0a of 0xef0a0002, which the ROM's own source names put in pbl_auth.c. Every host-side difference two testers could act on was tried one change per boot, all negative, so the maintainer and both testers agreed on derekross#36 to park it and write it up where the next person starts. WIFI-PLAN.md holds the facts, where it stops, the ruled-out list, the stock-side read of WLAN_SECURE_BOOTN (0x0 under stock's cnss2, not the 0x11111111 seen on mainline after the stall), both stop conditions as stop conditions, the open leads and the state to resume from. NOTES.md gets the short entry and points at it; the 2026-10-24 "next" line is marked superseded. The published kernel has no PCIe node for lynx; the Wi-Fi branch stays local and the modules stay blacklisted on the phone. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Boot: an Arch root on userdata from slot B, with the device package marking the slot successful. Storage and Touch: WIP on unpublished kernel branches, linked to the notes that record them, as the Pixel 6 Pro row does. Wi-Fi and Bluetooth: still no, now pointing at the plan that says why. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
FlorencioJP re-read the register set after a stall on 2026-10-27: the WLAON and TLMM neighbours match the stock column exactly, so the window lands, and the three registers that differ are the three the ROM's authentication stage consults. Their caveat on our stock column stands: cnss2 on lynx never drops the buck, so "pre-BHI after a cycle" is after a good PBL run. The open leads now say what a cold stock read takes on this build and what each outcome would mean. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
The maintainer's reading of the stock trace, checked against the 2026-10-27 dmesg here and FlorencioJP's full-Android log: the buck comes up once, the probe asserts WLAN_EN, enumerates and powers the link off without a BHI, and the first image download follows the second assert, 7.7 s later. Google's minimal diagnostic on the stock kernel stalls like mainline when the BHI follows the probe cycle by 13 ms, so the wait is the variable. The open leads are reordered on that: the stock-shaped power sequence with the wait first (FlorencioJP's phone), the cold stock read only if it fails. Both phones' pin states at power-up are in a table, and the hang difference (stock's pre-BHI reads against a read right after ath11k's SoC reset) is noted beside the stop condition, which stays as written. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
FlorencioJP read the stock-side register list at the stall: 132 of 147 match the warm pre-BHI column, the power sequencer included. Beyond the stall's own trace, one more register differs from a warm chip, WLAON_MEM_SVS_CFG_REG, 0x1 warm and 0x0 at the stall, so it joins the cold-read list. The JTAG id reads the same on both phones, which puts both register windows in the same place. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
|
Thank you, @latentoperator. This is a clean first package and a very good plan. The package is the Pixel 7 Pro's files byte for byte where it says so ( Two things before it merges:
And some smaller things. None of them holds the merge, though the first is worth doing in this PR:
On merge order: this doesn't depend on #91, and #91 doesn't depend on it, since the storage guard's allowlist for lynx is already on main. If #91 merges first, the Storage cell can say so in this PR. It will be squashed under your title, which already has the right prefix and version. |
docs/DEVICES.md conflicted on the Pixel 7a row (derekross#95 updated it). Keep this branch's row; the Storage cell now links kernel derekross#91, merged, and the Touch cell kernel derekross#93, in review.
The firmware offers PSCI system suspend, so the kernel offers deep sleep, and the phone reads `s2idle [deep]`. On gs201 deep never wakes, so the desktop's Suspend entry could leave the phone in a sleep only a held power key ends. Ship the Pixel 7 Pro's drop-in (MemorySleepMode=s2idle, no hibernation), byte for byte, with a cmp in the test like the other shared files. Suspend on lynx is still untested. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Lynx's allowlist is cheetah's, so it joins the loop with the same list. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
DEVICES: the Wanted line now matches the matrix row (UFS in the published kernel, touch in review, Wi-Fi parked). BRINGUP-M1: the status line from the 2026-10-27 reinstall, with the mem_sleep reading. NOTES: the 2026-10-24 BT_EN test was the pin only, with no Bluetooth firmware. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Step 6 (FlorencioJP, 8 s and 30 s of WLAN_EN low with the buck held, on mainline) is negative, so the 7.7 s line's [V] is marked as an inference that failed. Google's calibration handler sleeps 5 s before the first load, both stock boots here show it, and the stock-kernel diagnostic skipped it: no run has had Google's host stack and the wait together, which makes the fs_ready boot the next item. Bluetooth is ruled out by FlorencioJP's reference 2. 0xef150600 shows up before the first BHI on a chip that never passed, so section 5's warm/cold reading is withdrawn. The stock-kernel diagnostic moves SysMMU translating to the ruled-out table, on its own. The 2026-10-24 BT_EN test is described as what it was, the pin only. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Thanks @derekross. Pushed:
The |
Bytes 0x30-0x37 from omamobile-bootctl dump next to fastboot's slot-successful, slot-unbootable and slot-retry-count, in three states: slot B marked, after --set-active=b, and after one boot with the marker masked. Byte 0 is the retry count and byte 1 the flags, as on the Pixel 7 Pro and Pixel 6 Pro. The first marks on 2026-10-24 and 2026-10-27 were made by hand, outside the storage guard's rules; that is recorded too. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
@derekross, the
(fastboot columns are successful / unbootable / retry-count.) Same meaning as on the Pixel 7 Pro and 6 Pro: byte 0 is the retry count, byte 1 the flags (0x04 active, 0x02 successful). After the test the unit was unmasked and slot b is marked again ( |
The two test phones carry different revisions of the chip: stock cnss logs major/minor 0x12/0x11 here and 0x2/0x10 on FlorencioJP's, with the Bluetooth SOC version to match. Both stall the same way and read the same JTAG id. FlorencioJP's full-Android logs also show the calibration handler's 5 s wait before the first load, as this phone's do. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Thank you, @latentoperator. All of it is in, and the The rest is done as well: the sleep drop-in, byte for byte with a A few things for a later plan update, none of them for this PR:
|
The Pixel 7a's touch is in the published kernel (kernel #93) and its device package is merged (#88); the Pixel 4 XL's notes and results merged with #52, so its cells drop the "in review" qualifier; the Pixel 7's board device tree, touch driver and package are in review. Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
lynx's package (derekross#88) ships the same 20-usb-gadget.network as the others, with the laptop as usb0's Gateway= at metric 0, a route that outlives an unplug. It gets the same fix: the gateway is a [Route] at metric 2000, after Wi-Fi (600) and mobile data (700), and the same test. Not tested on a Pixel 7a. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The plan merged in #88 as it stood before the fs_ready diagnostic boot (run 17, 2026-10-27, #36). Record its result and what follows from it: - status line: the boot with Google's host stack, Google's own trigger and full Android's 5 s wait ran and stalled the same way, so that pair is ruled out and the variants are stopped; the cold stock read is the one measurement left; - section 7 item 1: the run as reported on #36 (Google's kernel and vendor_dlkm modules, 68 modules, Bluetooth off, fs_ready at 10.8 s, IN_COLD_BOOT_CAL, the power-up 5 s later, BHI from 16.0 s, 0xfad00000 on every dump, -110 at 76.8 s), and the module count that matches it; - section 7 item 2: the module comparison as background, with the caveat about first-stage loads, and the CNSS profile's 16 modules on top of the USB profile's 52, in load order; - section 8: the ramoops console survives only a warm return, so the diagnostic's init writes warm to /sys/kernel/reboot/mode first; - section 5 and the pin table: gph2-5 reads output low at every Linux boot only on a kernel without the Bluetooth node (linux-omarchy-mobile #98); with it a regulator owns the pin, so unbind the serdev first and record which kernel a run used. Assisted-by: LLM Signed-off-by: Jean-Paul Florencio <jpflorencio@version-tracker.app> Co-authored-by: Jean-Paul Florencio <jpflorencio@version-tracker.app>
The Pixel 9a's SoC and board device tree, 120 Hz, the touchscreen and its sleep fix, NFC's ST54L mode and poll stop, and the panel's local high brightness mode are merged (kernel #80/#81, #88, #82, #131, #71, #126; #155). The Pixel 8 Pro's buttons are in the published kernel (kernel #148). The OnePlus 13's device package is merged (#111), mustang has stage 0 notes (#198), and kodiak joins the Pixel 11 row (#201). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The plan merged in derekross#88 as it stood before the fs_ready diagnostic boot (run 17, 2026-10-27, derekross#36). Record its result and what follows from it: - status line: the boot with Google's host stack, Google's own trigger and full Android's 5 s wait ran and stalled the same way, so that pair is ruled out and the variants are stopped; the cold stock read is the one measurement left; - section 7 item 1: the run as reported on derekross#36 (Google's kernel and vendor_dlkm modules, 68 modules, Bluetooth off, fs_ready at 10.8 s, IN_COLD_BOOT_CAL, the power-up 5 s later, BHI from 16.0 s, 0xfad00000 on every dump, -110 at 76.8 s), and the module count that matches it; - section 7 item 2: the module comparison as background, with the caveat about first-stage loads, and the CNSS profile's 16 modules on top of the USB profile's 52, in load order; - section 8: the ramoops console survives only a warm return, so the diagnostic's init writes warm to /sys/kernel/reboot/mode first; - section 5 and the pin table: gph2-5 reads output low at every Linux boot only on a kernel without the Bluetooth node (linux-omarchy-mobile derekross#98); with it a regulator owns the pin, so unbind the serdev first and record which kernel a run used. Assisted-by: LLM Signed-off-by: Jean-Paul Florencio <jpflorencio@version-tracker.app> Co-authored-by: Jean-Paul Florencio <jpflorencio@version-tracker.app>
The Pixel 9a's SoC and board device tree, 120 Hz, the touchscreen and its sleep fix, NFC's ST54L mode and poll stop, and the panel's local high brightness mode are merged (kernel derekross#80/derekross#81, derekross#88, derekross#82, derekross#131, derekross#71, derekross#126; derekross#155). The Pixel 8 Pro's buttons are in the published kernel (kernel derekross#148). The OnePlus 13's device package is merged (derekross#111), mustang has stage 0 notes (derekross#198), and kodiak joins the Pixel 11 row (derekross#201). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
What and why
The Pixel 7a's device package, the notes from the Arch root, and the Wi-Fi plan asked for on #36. Tracking issue: #36. Rebased onto today's main; nothing stacked.
omamobile-device-lynx0.1.0, then 0.2.0. The Pixel Fold's package without the lid drop-in: storage guard rule, USB address, power-key drop-in, mirrorlist,reboot-to, the descriptor. 0.2.0 adds the Pixel 7 Pro'somamobile-boot-successunit andomamobile-bootctl, because slot B has three retries afterfastboot --set-active=band every unmarked boot burns one; the fourth boot fell to slot A, where Android now stops at "Cannot load Android system" (shareduserdata).devinfoissdd1, layout 3.15, the same as the Pixel 7 Pro. Registered intools/build-packagesandtests/all. The descriptor's comment changes from "not written yet" to what the package does.tools/install-phoneputting an Arch root onuserdata(needs--bootloader-slots-match: the bootloader reports only the current slot's version), the Goodix GT9895 touchscreen with 14-byte contact records and its two LDOs, the S2MPG12 RTC, the slot retries, and Wi-Fi as far as it got.fastboot --set-active=bfrom the bootloader menu, and never "Factory data reset" there.pbl_auth.c, with the image bisection and the iATU bisection), the ruled-out list from both phones, the stock-side read ofWLAN_SECURE_BOOTN(0x0under stock'scnss2, six reads over three states, not the0x11111111seen on mainline after the stall), both stop conditions as stop conditions, the open leads and the state to resume from. The published kernel has no PCIe node for lynx; the Wi-Fi branch stays local and the modules stay blacklisted on the phone. NOTES.md gets the short entry and points at it.The kernel side of UFS, touch and the sub-PMIC is on local branches on top of #44 (
omamobile-gs201-lynx-ufs,-touch); they come as PRs next, UFS first. The package does not depend on them to install, butomamobile-boot-successneeds UFS to finddevinfo, and the storage guard has nothing to guard without it.Device and build
lynx-17.0-15199429on both slotsuname -r):7.3.0-rc6-omamobile-push1-g38b5a04d67e0(the local Wi-Fi branch tip; it carries the unpublished UFS, touch and sub-PMIC commits on top of omamobile-audio 0.5.0: a UCM profile and PipeWire rule for the Pixel 9a #44)Checks (no phone needed)
tools/pr-check origin/mainpassestests/allpasses (186 ok; it now runspackages/omamobile-device-lynx/test/all)tests/packages: the package I touched,packages/omamobile-device-lynx/test/all, passes (20 ok); the package built withtools/build-packages omamobile-device-lynxon 2026-10-24versionbumped, and the new version is in the commit subjectCo-Authored-By:orAssisted-by:trailer, never aSigned-off-by:On the phone
tools/install-phonefrom this exact head (the 2026-10-27 reinstall used images built before the rebase; the package files are unchanged since). The root image on the phone carries 0.1.0; 0.2.0'somamobile-boot-successand unit were copied in and enabled by hand on 2026-10-24 and have marked the slot after every boot since. The[omamobile]mirrorlist entry against the published repository. Anything radio.reboot-towas built but onlyomamobile fastbootwas used.userdata, weakens the storage guard, or reads registers from userspace (AGENTS.md §1). The one write elsewhere is the Pixel 7 Pro's own boot-success marker indevinfo: it opens that disk and partition read-write for the oneomamobile-bootctlwrite and relocks them, unchanged from cheetah's package.Two things I hit, in case they are worth a look:
tools/phone-smokerun from the laptop came back empty here (the report below is--hereon the phone, with the output to a file; run throughphone-sshthe session also outlived the script until a timeout), andpackages/.gitignoreignoressrc/, so the package's two C sources neededgit add -f, as the other device packages' did.tools/phone-smokereportphone-smoke: lynx
Kernel
7.3.0-rc6-omamobile-push1-g38b5a04d67e0, tools atb17370c, 2026-10-27T16:54Z.Packages: hyprland 0.56.2-7, libcamera 0.7.2-4, mesa 1:26.2.5-1, omacam 0.3.3-1, omamobile-apps 0.2.0-1, omamobile-device-lynx 0.1.0-1, omamobile-firstboot 0.1.0-1, omamobile-keyring 20261013-1, omamobile-mirror 0.1.1-1, omamobile-omarchy 0.1.6-1, omamobile-owl-widgets 0.1.1-1, omamobile-shell 0.2.15-1, omarchy 4.0.6-1, quickshell 0.3.2.r7.g2b4501a-1
I used an AI assistant (Claude Code) for this; the commits carry a
Co-Authored-By:trailer.🤖 Generated with Claude Code