Skip to content

device-lynx: Bluetooth on the APM UART - #98

Open
FlorencioJP wants to merge 1 commit into
derekross:mainfrom
FlorencioJP:lynx-bluetooth
Open

FlorencioJP wants to merge 1 commit into
derekross:mainfrom
FlorencioJP:lynx-bluetooth

Conversation

@FlorencioJP

@FlorencioJP FlorencioJP commented Oct 27, 2026 •

Copy link
Copy Markdown
Contributor

Userspace side of linux-omarchy-mobile#98 (merged 2026-11-03): the Pixel 7a's WCN6855 BT core works over the APM USI1 UART, powered by its qcom,wcn6855-pmu node. One commit on today's main.

  • kernel/configs/omamobile.config: CONFIG_BT_BNEP with its two filters, for PAN tethering, next to the Bluetooth lines (it can't affect another phone). hci_qca, btqca and pwrseq-qcom-wcn already come from defconfig: merge_config of defconfig plus this fragment gives CONFIG_BT_HCIUART_QCA=y, CONFIG_BT_QCA=m, CONFIG_POWER_SEQUENCING_QCOM_WCN=m, so nothing else to add.
  • packages/omamobile-device-lynx 0.2.3 (0.2.2 is Release channels, release-key signing and the bundle manifest (decisions 4, 9, 12) #313's): omamobile-bt-addr sets the address the bootloader publishes in /chosen/config/bt_addr when udev sees the controller (DRIVERS=="hci_uart_qca"), through omamobile-bt-addr@.service as on cheetah. Its own script: the WCN6855 is unconfigured until the address is set, so it answers neither btmgmt info nor a power cycle; the script retries btmgmt public-addr while the index is not registered yet after a rebind. test/all passes in the ALARM container.
  • device/lynx/NOTES.md: the 2026-10-29 Bluetooth entry with the PMU boot evidence, the module order, the rebind race and the gph2-5 bench note for Wi-Fi work.
  • docs/DEVICES.md: Bluetooth testing on lynx (scans with its own address; kernel device-lynx: Bluetooth on the APM UART #98 is merged but kernel-2026.11.01 predates it; this package's address service has not run on the phone yet; pairing untested).

The address logic is the one that has run on the test phone since the 29th (from a local package); this package's copy of it is not yet installed there. Context: #36.

https://claude.ai/code/session_014epmuRTsr4rKgN3czrBeke

@derekross

Copy link
Copy Markdown
Owner

Thank you, @FlorencioJP. The NOTES table is the right record: driver and firmware, chip version, the baud finding with its control, the clocks against stock, the missing address and the __init trap, each with its source. tools/pr-check passes on it here.

A few things before it can merge:

  1. It conflicts with main, so CI hasn't run. The branch is 16 commits behind, and the lynx row in docs/DEVICES.md has changed since. latentoperator's device-lynx 0.2.0: the device package, the Arch root and the Wi-Fi plan #88 is about to change that row again (its Bluetooth cell says "no (same chip as Wi-Fi)"). Please rebase onto main once device-lynx 0.2.0: the device package, the Arch root and the Wi-Fi plan #88 is in, and change only the Bluetooth cell.
  2. "works" claims more than the matrix allows. The legend's "works" means verified on the phone and on by default. Here the device tree is in an open kernel PR (linux-omarchy-mobile device-lynx: Bluetooth on the APM UART #98), which doesn't build on the published kernel yet, and the modules and the address service come from a local package. Pairing and PAN aren't tested yet. Something like testing (scans; in review, [kernel #98](https://github.com/derekross/linux-omarchy-mobile/pull/98); the address is set by a local service; pairing untested; [notes](../device/lynx/NOTES.md)) would be accurate. I'd merge this after kernel device-lynx: Bluetooth on the APM UART #98, so the link points at merged work.
  3. An SPDX header on device/lynx/kernel.config: # SPDX-License-Identifier: MIT as the first line, as in husky's and raven's fragments.
  4. WIFI-PLAN.md after Bluetooth. Once device-lynx 0.2.0: the device package, the Arch root and the Wi-Fi plan #88 is in, the plan says Bluetooth "waits on the same question", and its section 5 says gph2-5 reads output low at every Linux boot. With your device tree, a regulator owns gph2-5, raises it a few seconds into boot and drops it whenever Bluetooth goes down. So please add a line there with your bench note: unbind serial0-0 before a Wi-Fi harness takes the pin, and record whether each Wi-Fi run's kernel had the Bluetooth node. You and @latentoperator can decide which PR carries it.

Smaller things:

  • NOTES is newest first, so the Bluetooth section goes at the top rather than after the stock facts.
  • Please add the uname -r the phone ran and which tree it came from (the kernel PR needs the same).
  • The masked 94:45:60:xx:xx:xx is only the vendor prefix, but <bt_addr> says the same without it.
  • The body says the address comes from /chosen/config/bt_addr, while your comment on Pixel 7a (lynx, gs201 IPOP) port: stock facts and first light #36 says the service on the phone derives a locally administered one from machine-id. Please say in the notes which one the phone uses now.
  • The address service and the module setup belong in the lynx device package from device-lynx 0.2.0: the device package, the Arch root and the Wi-Fi plan #88, in a follow-up, so Bluetooth can be on by default.

FlorencioJP pushed a commit to FlorencioJP/linux-omarchy-mobile that referenced this pull request Oct 28, 2026
The Pixel 7a carries the WCN6855 BT core on the APM USI1 UART (ttySAC18
in Google's kernel), as cheetah carries its BCM4389. Enable cmu_apm,
usi_apm1, serial_apm1 and pinctrl_peric0 and describe the controller as
qcom,wcn6855-bt: BT_EN on gpp16-2, and the 0.95 V buck shared with Wi-Fi
on gph2-5 as a fixed regulator raised 7 ms before BT_EN, as stock btpower
does. WLAN_EN (gph2-4) stays low.

With this node Bluetooth owns gph2-5: hci_qca disables its supplies on
power-off, so the buck drops whenever Bluetooth goes down (stock btpower
never drops it), including once at every boot while the controller has
no address yet. A shared qcom,wcn6855-pmu node is the way out once Wi-Fi
runs. No device-wake or host-wake line is described, so Bluetooth cannot
wake the phone.

max-speed is 3 Mbaud: at 3.2 Mbaud every TLV segment of the firmware
download answers with "Frame reassembly failed (-84)" on this UART; 1, 2
and 3 Mbaud work.

Tested on one phone (chip revision 0x400c0210; untested on 0x400c1211)
with hci_uart/btqca and linux-firmware's qca/hpbtfw21.tlv + hpnv21.bin,
on 7.3.0-rc6-omamobile-lynx-wifi1 built from the integration tree at
1cb4bd3, which carries the HSI2 pin controller that defines gph2:
firmware loaded 3 s after boot, controller powered, scan sees the
neighbours. The controller reports no public address; userspace sets the
one the bootloader publishes in /chosen/config/bt_addr. Facts and log
excerpts: derekross/omarchy-mobile#98 and device/lynx/NOTES.md there.

These nodes must be enabled in the boot DTB: enabling cmu_apm through a
runtime overlay oopses, since gs201_cmu_probe is __init.

Assisted-by: LLM
Signed-off-by: Jean-Paul Florencio <jpflorencio@version-tracker.app>
@FlorencioJP

Copy link
Copy Markdown
Contributor Author

Thanks. Pushed as a second commit for now (bbb1f16): Bluetooth section first in NOTES, with the kernel the phone ran (7.3.0-rc6-omamobile-lynx-wifi1 from the integration tree at 1cb4bd36f3be) and which address the phone uses now (the bootloader's /chosen/config/bt_addr, set by a local service; machine-id only as fallback); the gph2-5 bench note (unbind serial0-0 first, record whether the kernel had the Bluetooth node); SPDX header on kernel.config; testing (...) in DEVICES.md with your wording.

Left for the rebase once #88 is in: squash into one commit signed by Jean-Paul Florencio, change only the Bluetooth cell of the lynx row, and add the WIFI-PLAN.md line (or latentoperator can carry it on #88, whichever lands first). The tools/phone-smoke report comes with the next boot of the phone. The address service and the module setup will follow in the lynx device package once #88 is merged.

@derekross

Copy link
Copy Markdown
Owner

Thank you, @FlorencioJP. The second commit covers what I asked for in the notes and the matrix. The Bluetooth section now comes first, with the kernel the phone ran. It says which address the phone uses now: the bootloader's, with machine-id only as a fallback. The gph2-5 bench note is in it, the address prefix is gone, the SPDX line is on kernel.config, and the Bluetooth cell says testing in the words we agreed. tools/pr-check passes on the new head.

What's left is mostly the rebase you already planned. Some of it got bigger now that #88 is merged:

  1. It still conflicts with main, now in device/lynx/NOTES.md as well as docs/DEVICES.md, so CI hasn't run. In the lynx row, change only the Bluetooth cell, which now reads no (same chip as Wi-Fi; [plan](../device/lynx/WIFI-PLAN.md)). In NOTES, put the Bluetooth section above ## 2026-10-27: Wi-Fi parked.

  2. Two lines from device-lynx 0.2.0: the device package, the Arch root and the Wi-Fi plan #88 that your section makes out of date. In NOTES, "Wi-Fi parked" says "Bluetooth, on the same chip, waits on the same question". The Wi-Fi bring-up section says BT_EN is gpp16-2 "on the PERIC1 pin controller (0x10840000), which our tree does not describe yet". Our tree calls the controller at 0x10840000 pinctrl_peric0, and gpp16 is described there, which is the block your kernel PR uses. The facts table in WIFI-PLAN.md section 2 repeats the second line. Please correct both while you're there. latentoperator may want to look at the wording.

  3. The WIFI-PLAN.md line. device-lynx 0.2.0: the device package, the Arch root and the Wi-Fi plan #88 has landed, so it can't carry the bench note any more, and it belongs here. The note in NOTES is right. Please add the same two points to the plan, in section 7 where the next Wi-Fi runs are listed (or section 6, as a stop condition, if you prefer): unbind serial0-0 before a harness takes gph2-5, and record whether that run's kernel had the Bluetooth node.

  4. The matrix cell after kernel device-lynx: Bluetooth on the APM UART #98. Since this merges after the kernel PR, the cell's "in review" will be wrong by then. Please use the wording the touch cell next to it uses: testing (scans; in the published kernel, [kernel #98](…); the address is set by a local service; pairing untested; [notes](…)).

  5. The squash. Your plan of one commit authored and signed by you is right. The first commit's author address is claude@…, and if it stays in the branch, GitHub's squash adds it as a Co-authored-by: line on main. Squashing it yourself avoids that.

Smaller things:

  • kernel.config is only read by bin/omamobile kernel with DEVICE=lynx. CI's kernel check builds without device fragments, so CONFIG_BT_BNEP is untested until someone builds for lynx. That's fine for a module, I just wanted you to know.
  • The Pixel 7a entry under the matrix lists what is in the published kernel. Once kernel device-lynx: Bluetooth on the APM UART #98 is in, Bluetooth could be added there next to UFS and touch.

On order: kernel #98 first (it waits for me to publish the HSI2 pin controller, and then for your PMU version), then this one, rebased, with the cell above. Thanks for keeping both PRs in step. The notes table is a model of how a bring-up finding should be recorded.

derekross added a commit that referenced this pull request Oct 28, 2026
…wenty-sixth rounds (#127)

* docs/review: 19 lessons and 3 corrections from the twenty-fifth round

The confirmed candidates from round 25 (L-260 to L-278): the 9a slot B
marking as a devinfo question, cross-repository links, squash-merged
stacks and version files, stale hand-built prebuilts, and before-merge
Pixel 7 Pro runs on omamobile-int. C-16 to C-18 correct our own earlier
comments on kernel #98, kernel #80 and #74; L-241 is corrected in place.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* docs/review: 15 lessons from the twenty-sixth round

L-279 to L-293: the 9a's slot B hold and what made devinfo writable
(#125), per-device flags checked by building both command lines
(#101), env defaults in the shared build library and CI waiting for
approval (#79), GLib retry paths and radio stops checked at the
controller (#99), silent re-pushes and stacks with several tops (the
tegu stack), and two of our own: merge only after every check has
finished (#126), and an empty diff is checked for empty inputs
(kernel #123).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
derekross pushed a commit that referenced this pull request Oct 29, 2026
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>
derekross added a commit that referenced this pull request Oct 29, 2026
Twenty-six lessons (L-318 to L-343) from the round 29 omamobile-reviewer
runs, and three corrections: C-20, our kernel #109 advice that the
Pixel 7's SPI controller DMAs the touch report (it polls; gs201's spi_0
has no dmas); C-21, asking for FT3658=m in panther's device
kernel.config against RELEASES; C-22, a Bluetooth address in kernel
#98's thread that the first review missed. Five candidates were dropped
as already covered.

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
derekross added a commit that referenced this pull request Oct 30, 2026
The matrix and the Wanted list against what has merged since #140, with
each pull request's state checked on GitHub today:

- Pixel 7 (panther): the GPU and haptics nodes (kernel #129, #130) were
  closed because they landed unchanged through kernel #134, merged with
  the touchscreen node and s2idle by default; device-panther 0.1.1 (#145)
  carries the GPU's always-on rule, the feedbackd rule and the sleep.conf
  drop-in. GPU and Haptics stay "testing" and now cite #134, #145 and the
  2026-10-28 notes. Touch stays WIP: kernel #134's body records touches
  driving Hyprland and the PIN screen, but the driver fixes (kernel #133)
  and the kernel option with the touch notes (#149) are still in review.
  Suspend: "no" with the s2idle default named, since the notes say it
  hasn't run.
- Pixel 6 Pro (raven): Haptics "no" -> "works": HAPTICS-TORCH-PLAN H5
  passed (2026-10-28/29), cs40l25-vibra and the node merged as kernel
  #114, feedbackd's rule in device-raven 0.1.27 (#113). Display names
  its open kernel PR, #116.
- Pixel 7a (lynx): Bluetooth "no" -> WIP: the merged WIFI-PLAN (#133)
  records it scanning through its own UART; kernel #98 and #98 are in
  review. The Wanted line says so too.
- Pixel 7 Pro (cheetah) detail: Boot names the AVB footers a chained
  vbmeta_b needs (NOTES 2026-10-28 and 10-29, #104); Suspend describes
  the merged tier 1 auto-sleep (#64, off by default behind the
  omamobile-sleep toggle, P0-P5 on 2026-10-26) in place of the old
  "until S8" text. Nothing from the open #171, #172, #174-#176.
- Wanted: coral's #52 is merged, so its line no longer says "in review";
  flame gets a line for its native install in review (#55, #56, #58,
  #59, split out of #17); the OnePlus 13 (dodge) is listed with #111 and
  kernel #115, both open.

The Pixel 9a row and the Pixel 9 family line are left to #165. Every
status word still parses with the website's parseCell.

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
@derekross

Copy link
Copy Markdown
Owner

One correction from my side before your rebase: the BNEP options can't break another phone, so they belong in kernel/configs/omamobile.config next to the Bluetooth lines, not in a lynx-only fragment, which the release kernel never reads. Sorry I didn't catch that earlier. Everything else stands once kernel #98 is in.

derekross added a commit that referenced this pull request Oct 31, 2026
… rounds

L-348 to L-361: shared drivers judged against every int-line user (#150),
board pins against every stock user (#151), ownership flags under the
consumer's lock (#152), /chosen children and Google's panel constants
(#154), same-day notes PRs on one device (#208/#209), merge simulations for
DEVICES rows, ours included (#165), stacked PRs that merged main (#115/#117),
titles behind renumbered versions, a contributor merge that dropped the PR
(#99), smoke package lists as provenance and "owner-authorized" writes
(#200), promised fixes and named tags checked on the remote (#153/#157/#158),
and file swaps into old packages (#193). C-23 (usbc-deadman's device scope,
kernel #107) and C-24 (BNEP in a lynx-only fragment, #98).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
derekross pushed a commit to derekross/linux-omarchy-mobile that referenced this pull request Nov 3, 2026
The Pixel 7a carries the WCN6855 BT core on the APM USI1 UART (ttySAC18
in Google's kernel), as cheetah carries its BCM4389. Enable cmu_apm,
usi_apm1, serial_apm1, pinctrl_peric0 and pinctrl_hsi2, and describe
the controller as qcom,wcn6855-bt powered by a qcom,wcn6855-pmu node
(pwrseq-qcom-wcn): BT_EN on gpp16-2, and the 0.95 V buck shared with
Wi-Fi on gph2-5 as the sequencer's input, a fixed regulator with the
7 ms ramp stock btpower waits before BT_EN. The sequencer then holds
BT_EN high and waits 50 ms. No wlan-enable-gpios is given: WLAN_EN
(gph2-4) stays low until a PCIe node takes the sequencer's wlan target.

With this node Bluetooth owns gph2-5 and drops it on power-off: off is
BT_EN low and then the regulators, so the buck goes down whenever
Bluetooth goes down (stock btpower never drops it), including once at
every boot while the controller has no address yet. The five PMIC rails
behind the chip's I/O, 1.3 V and 1.9 V inputs are an always-on stand-in
regulator: stock's device tree names no supply on its cnss2 or btpower
node, so which PMIC regulators feed them is unknown. No device-wake or
host-wake line is described, so Bluetooth cannot wake the phone.

max-speed is 3 Mbaud: at 3.2 Mbaud every TLV segment of the firmware
download answers with "Frame reassembly failed (-84)" on this UART; 1, 2
and 3 Mbaud work.

Tested on one phone (chip revision 0x400c0210; untested on 0x400c1211)
with hci_uart/btqca, pwrseq-qcom-wcn and linux-firmware's
qca/hpbtfw21.tlv + hpnv21.bin, on 7.3.0-rc6-omamobile-lynx-wifi1 with
this file's DTB: the sequencer holds the buck (five users), BT_EN and
gph2-5 read high, firmware loaded 3 s after boot, controller powered,
a 20 s scan sees the neighbours, and unbinding the serdev driver drops
both pins and disables the buck. The controller reports no public
address; userspace sets the one the bootloader publishes in
/chosen/config/bt_addr. Facts and log excerpts: derekross/omarchy-mobile#98
and device/lynx/NOTES.md there.

These nodes must be enabled in the boot DTB: enabling cmu_apm through a
runtime overlay oopses, since gs201_cmu_probe is __init.

Assisted-by: LLM
Signed-off-by: Jean-Paul Florencio <jpflorencio@version-tracker.app>
@FlorencioJP
FlorencioJP force-pushed the lynx-bluetooth branch 2 times, most recently from e82e8de to a53e1b2 Compare November 5, 2026 22:33
@FlorencioJP

Copy link
Copy Markdown
Contributor Author

@derekross, thanks for merging kernel #98. Rebased onto today's main as one commit, a53e1b2, with your correction: the three BNEP options sit in kernel/configs/omamobile.config next to the Bluetooth lines (after #304's HID and uinput block), and the lynx fragment is gone. I also checked what the fragment leaves the lynx kernel with: merge_config of defconfig plus omamobile.config already gives CONFIG_BT_HCIUART_QCA=y, CONFIG_BT_QCA=m and CONFIG_POWER_SEQUENCING_QCOM_WCN=m, so BNEP was the only missing option.

Since the lynx device package is upstream now, the address service moved into it, as device-lynx 0.2.3 (#313 took 0.2.2): omamobile-bt-addr@.service and an 80-omamobile-bt-addr.rules keyed on hci_uart_qca, as on cheetah, with its own script because the WCN6855 is unconfigured until the address is set (so btmgmt info and the power cycle of cheetah's script don't apply). test/all passes in the ALARM container and tools/pr-check is clean. The PMU boot is in device/lynx/NOTES.md as the 2026-10-29 entry.

DEVICES.md says testing, in your words: the kernel side is merged but kernel-2026.11.01 predates it, and this package's copy of the address service hasn't run on the phone yet (the local package's has, since the 29th). I'll move the cell once the phone boots the next published kernel with this package, and again after a pairing test.

@derekross

Copy link
Copy Markdown
Owner

Thank you, @FlorencioJP. The rebase is clean, the BNEP options sit where I asked, and the address service is in the device package. The notes are as careful as before, and both #88 sentences are fixed in NOTES. CI is green, and the package test passes here too.

Three things before I merge:

  1. A run of this package's own copy. The body says it plainly: the address logic that has run since the 29th is your local copy, and this file hasn't run yet. Since it deliberately differs from the 7 Pro's script, I'd like one boot with device-lynx built from this branch (tools/phone-install-package packages/omamobile-device-lynx from a checkout of it, then pacman -Q omamobile-device-lynx). Please post journalctl -u 'omamobile-bt-addr@*' -b with the address as <bt_addr>, btmgmt info showing the controller configured, and the tools/phone-smoke report with its header line. If it's easy, one unbind and bind of serial0-0 would show the "Invalid Index" retry doing its job.
  2. The Claude-Session: line. The commit message ends with a link to your assistant session. We squash with the commit messages, so it would end up on main. Please drop it. Assisted-by: LLM already says what it needs to.
  3. The version. device-lynx 0.2.3: persist the DSIM transmit enable #438 also bumps device-lynx to 0.2.3, and device-lynx 0.2.2: accept devinfo 3.11 for boot-success #436 to 0.2.4. Yours is closest to ready, so I expect it to keep 0.2.3 and the others to move up. I'll sort that out at merge.

One small leftover from October: device/lynx/WIFI-PLAN.md line 43 still says BT_EN is on "PERIC1 … which our tree does not describe". Could you fix it the way you fixed NOTES?

For the record, adding BNEP to the shared fragment also gives the 7 Pro a bnep module, which bluetoothd most likely loads at start. That's fine and was my request. If you like, the comment could say it's for every phone.

derekross added a commit that referenced this pull request Nov 7, 2026
The matrix re-read against what merged and ran on hardware since 2026-11-02:
- Pixel 7: Bluetooth and the torch move to testing (kernel #196, #197, #426); fingerprint moves to WIP (kernel #207; the package side, #458, is in review); AVB footers and reboot=warm noted.
- Pixel 8a: a new row, from first light and UFS behind the guard (kernel #192, #206; #414).
- Pixel 6 Pro: GPU scaling and cooling (kernel #189) and UFS suspend (kernel #188) merged.
- Pixel 9a: GPU levels merged through kernel #96; GNSS's moving-car session (#114) and the link driver (kernel #185).
- Pixel 7a: Bluetooth's device tree merged (kernel #98).
The Pixel 8 Pro's row is left to #472.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@FlorencioJP

Copy link
Copy Markdown
Contributor Author

@derekross, done, as 436db68 on today's main: the Claude-Session: line is gone from the message (only Assisted-by: LLM and my sign-off remain), the WIFI-PLAN.md facts row now says BT_EN is on pinctrl_peric0, described since kernel #98, as NOTES does, and the BNEP comment says it is for every phone, with lynx as its first user. The version stays 0.2.3, as you expect.

The boot with this package's own copy of the address service comes next, when the phone is back on the bench: tools/phone-install-package packages/omamobile-device-lynx from this checkout, pacman -Q, the omamobile-bt-addr@* journal with the address as <bt_addr>, btmgmt info configured, the tools/phone-smoke report with its header, and one unbind and bind of serial0-0 for the retry.

@derekross

Copy link
Copy Markdown
Owner

Thank you, @FlorencioJP. All three are in: the session line is gone, WIFI-PLAN now says pinctrl_peric0, and the BNEP comment reads well. I'll look for the boot with this package's own copy when the phone is back on the bench.

A heads-up for that push: today's DEVICES pass (#474) touched the lynx Bluetooth cell and the Pixel 7a line in the Wanted list, so the branch now conflicts there. When you rebase, keep your two lines as they are. They already say what mine says, and mine ("package side in review") stops being true once this merges.

W4HO-ham added a commit to W4HO-ham/omarchy-mobile that referenced this pull request Nov 8, 2026
…ixed, 300 of 300 twice

The dead storage after a wake was the UFS pads (reference clock and reset) coming back as inputs; the host now re-applies its pin state at resume. The silent no-wake was C2 during suspend-to-idle with the PMU's wake-up masks never programmed; the C2 states are kept out of s2idle. 300 RTC-woken cycles passed WFI-only (derekross#96) and again with C2 enabled at runtime (derekross#98). The series scripts (sleep-series.sh, sleep-series-loop.sh) record the modem unit's state and put it back, cap the automatic recoveries, save the ramoops console after a self-reboot, and never reboot a palmtop that rebooted on its own. The record lists the silent no-wakes and the dead-storage wakes separately, with the conditions for unattended runs.
W4HO-ham added a commit to W4HO-ham/omarchy-mobile that referenced this pull request Nov 9, 2026
…ixed, 300 of 300 twice

The dead storage after a wake was the UFS pads (reference clock and reset) coming back as inputs; the host now re-applies its pin state at resume. The silent no-wake was C2 during suspend-to-idle with the PMU's wake-up masks never programmed; the C2 states are kept out of s2idle. 300 RTC-woken cycles passed WFI-only (derekross#96) and again with C2 enabled at runtime (derekross#98). The series scripts (sleep-series.sh, sleep-series-loop.sh) record the modem unit's state and put it back, cap the automatic recoveries, save the ramoops console after a self-reboot, and never reboot a palmtop that rebooted on its own. The record lists the silent no-wakes and the dead-storage wakes separately, with the conditions for unattended runs.
@FlorencioJP

Copy link
Copy Markdown
Contributor Author

Rebased once more onto main after #474, as 5b5413c: same content as 436db68, only the two lynx lines of DEVICES.md re-applied on top of today's row, as you asked. Merges clean on the current main; tools/pr-check passes. The boot with this package's own copy still waits for the phone to be back on the bench.

@derekross

Copy link
Copy Markdown
Owner

Thank you, @FlorencioJP. I compared the new head with 436db68 and it is the same commit, with only your two lynx lines moved onto today's row. tools/pr-check passes, CI is green and it merges cleanly on today's main.

So the one thing left is the boot with this package's own copy of the address service, whenever the phone is back on the bench: pacman -Q omamobile-device-lynx, journalctl -u 'omamobile-bt-addr@*' -b with the address as <bt_addr>, btmgmt info showing the controller configured, the tools/phone-smoke report, and one unbind and bind of serial0-0 if it's easy.

One small thing for that push: the Bluetooth cell says the kernel side is "not yet in a published kernel", and it is now. kernel-2026.11.08 contains 5df2fd34a4fd ("gs201-lynx: enable Bluetooth on the APM UART"), so the cell can say "in the published kernel". The BNEP options come with the first kernel built after this merges, so PAN stays untested until then.

On the version: #438, #484 and #403 also write device-lynx 0.2.3 at the moment. Yours keeps it if it lands first; I'll check the number again when I merge.

@FlorencioJP

Copy link
Copy Markdown
Contributor Author

Booted with this branch's own package today, as asked. Built on the phone with tools/phone-install-package packages/omamobile-device-lynx from a checkout of the branch, then one reboot.

$ pacman -Q omamobile-device-lynx
omamobile-device-lynx 0.2.3-1

$ journalctl -u 'omamobile-bt-addr@*' -b        # address as <bt_addr>
[    3.845803] systemd[1]: Starting Set the Bluetooth address of hci0 from the bootloader...
[    3.895609] omamobile-bt-addr[702]: hci0 Set Public Address complete, options:
[    3.897995] systemd[1]: omamobile-bt-addr@hci0.service: Deactivated successfully.
[    3.898877] systemd[1]: Finished Set the Bluetooth address of hci0 from the bootloader.

$ btmgmt info
Index list with 1 item
hci0:	Primary controller
	addr <bt_addr> version 12 manufacturer 29 class 0x0c0000
	supported settings: powered connectable fast-connectable discoverable bondable link-security ssp br/edr le advertising secure-conn debug-keys privacy configuration static-addr phy-configuration wide-band-speech cis-central cis-peripheral iso-broadcaster sync-receiver ll-privacy past-sender past-receiver
	current settings: powered bondable ssp br/edr le secure-conn wide-band-speech cis-central cis-peripheral iso-broadcaster sync-receiver ll-privacy past-sender past-receiver
	name pixel-omarchy

The address the controller carries is the bootloader's /chosen/config/bt_addr (compared on the phone; it differs from the machine-id fallback file). bluetoothctl show: Powered yes.

One unbind and bind of serial0-0 on hci_uart_qca, 284 s into the boot:

[  284.611568] systemd[1]: Starting Set the Bluetooth address of hci0 from the bootloader...
[  285.681757] omamobile-bt-addr[3895]: hci0 Set Public Address complete, options:
[  285.683302] systemd[1]: omamobile-bt-addr@hci0.service: Deactivated successfully.

1.07 s from start to the address, against 0.05 s at boot: that is the "Invalid Index" retry doing its work (the script prints only the last answer). btmgmt info and bluetoothctl show give the controller powered with the same address afterwards.

tools/phone-smoke report

phone-smoke: lynx

Kernel 7.3.0-rc6-omamobile-lynx-wifi1, tools at 5b5413c, 2026-11-12T09:58Z.

Check Result Detail
boot (required) ok kernel 7.3.0-rc6-omamobile-lynx-wifi1, slot b, up 350s
storage guard (required) ok 55 partitions; only userdata and slot B's boot images writable
root (required) ok ext4 on userdata, 1.8G free
kernel health (required) ok no oops or BUG; 0 WARN, 10 error lines in dmesg (3 known, 7 new)
new dmesg errors not ok OF: overlay: WARNING: memory leak will occur if overlay removed, property: /soc@N/usi@N/status x1; OF: overlay: WARNING: memory leak will occur if overlay removed, property: /soc@N/usi@N/spi@N/status x1; OF: overlay: WARNING: memory leak will occur if overlay removed, property: /soc@N/usi@N/spi@N/num-cs x1
systemd units not ok failed: ufw.service; user: xdg-document-portal.service
USB gadget (required) ok bound to 11210000.usb
display ok card0-Unknown-1 connected, 1080x2400
GPU ok render node on panthor
session (required) ok Hyprland and the Omarchy shell running
lock ok PAM file present, the shell can lock (passwordPam true, locked true)
touch ok Goodix Berlin Capacitive TouchScreen
buttons skip no key input device found by name
Wi-Fi skip no wireless interface
Bluetooth ok hci0, powered true
battery skip no battery power supply
sound cards skip none
sensors skip no sensor proxy
cameras skip no libcamera tools

Packages: hyprland 0.56.2-5.1, libcamera 0.7.2-4.21, mesa 1:26.2.4-1.4, omacam 0.3.3-1, omamobile-apps 0.2.0-1, omamobile-device-lynx 0.2.3-1, omamobile-firstboot 0.1.2-1, omamobile-keyring 20261026.lynx1-1, omamobile-lynx-bluetooth 0.1.2-1, omamobile-lynx-hardware 0.1.2-1, omamobile-mirror 0.1.1-1, omamobile-omarchy 0.1.12-1, omamobile-owl-widgets 0.1.1-1, omamobile-shell 0.2.29-1, omarchy 4.0.6-1, quickshell 0.3.2.r7.g2b4501a-1

Two things to say plainly:

  • The kernel is not kernel-2026.11.08: the phone runs my local 7.3.0-rc6 build of the tree (the Wi-Fi bench kernel), and the Bluetooth modules come from my local omamobile-lynx-bluetooth package. For this boot I masked that package's own address service and udev rule, so only omamobile-bt-addr@hci0 ran. Moving the phone to the published kernel is a separate step I'd rather not mix with this one.
  • The build's check() failed first on this phone: cmp was missing (no diffutils on it), and test/all compares files with the cheetah package. tegu's PKGBUILD declares checkdepends=(diffutils); lynx didn't. Added as 8fa0be5, together with the DEVICES cell you asked for ("in the published kernel", and "booted on the phone" for the package). Same commit otherwise; tools/pr-check passes.

@derekross

Copy link
Copy Markdown
Owner

Thank you, @FlorencioJP. That is the boot I was waiting for, and it shows what it needed to: omamobile-device-lynx built from the branch, omamobile-bt-addr@hci0 setting the bootloader's address 3.8 s into the boot, the controller configured and powered, and after the unbind and bind of serial0-0 the unit taking 1.07 s where it took 0.05 s at boot, which is two "Invalid Index" answers half a second apart. Thank you also for saying plainly that the kernel was your local build and that the modules came from your own package with its address service masked. I compared the new head with 5b5413c: only the checkdepends line and the DEVICES cell changed, so the package you ran is the package in the PR. diffutils in checkdepends is a good catch; six of our other device packages have the same gap, and I'll fix those.

Two small things and it goes in:

  1. The boot belongs in NOTES. Right now it's only in your comment, and the cell's "booted on the phone" links to the 10-29 entry. A few lines there are enough: the date, uname -r and that it was a local build, the package version and that it was built on the phone from this branch, that your omamobile-lynx-bluetooth package supplied the modules with its own address service and rule masked, and the two timings. The cell could then say "booted on the phone with a local kernel build", so nobody reads it as the published kernel.
  2. The version. reboot-to, omamobile-modem: a deliberate restart doesn't cost the next boot its modem #403 landed first and took 0.2.3, so this one moves up. I'll set the number myself when I merge (the version file, the two DEVICES lines and the Address row in NOTES); you don't need to chase it.

Two questions, neither blocking:

  • Where does the firmware come from on a fresh image? NOTES says qca/hpbtfw21.tlv and qca/hpnv21.bin are from linux-firmware-atheros, and our image installs only linux-firmware-other. If you installed it by hand, it should go into device-lynx's depends, so an updated phone gets it too.
  • What does your omamobile-lynx-bluetooth package still do on the phone, now that the address service is here? If it's only module loading, a line in NOTES saying so would tell the next 7a owner what the public package and the published kernel need.

One thing for later, not for this PR: the rule's KERNEL=="hci[0-9]*" also matches the per-connection hci0:N devices, so once you pair something each connection will start the address unit and leave a failed one. The 7 Pro's rule has the same pattern today and #423 fixes it; I'll make sure lynx's copy is included.

robertzas pushed a commit to robertzas/omarchy-mobile that referenced this pull request Nov 14, 2026
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>
robertzas pushed a commit to robertzas/omarchy-mobile that referenced this pull request Nov 14, 2026
Twenty-six lessons (L-318 to L-343) from the round 29 omamobile-reviewer
runs, and three corrections: C-20, our kernel derekross#109 advice that the
Pixel 7's SPI controller DMAs the touch report (it polls; gs201's spi_0
has no dmas); C-21, asking for FT3658=m in panther's device
kernel.config against RELEASES; C-22, a Bluetooth address in kernel
derekross#98's thread that the first review missed. Five candidates were dropped
as already covered.

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
robertzas pushed a commit to robertzas/omarchy-mobile that referenced this pull request Nov 14, 2026
The matrix and the Wanted list against what has merged since derekross#140, with
each pull request's state checked on GitHub today:

- Pixel 7 (panther): the GPU and haptics nodes (kernel derekross#129, derekross#130) were
  closed because they landed unchanged through kernel derekross#134, merged with
  the touchscreen node and s2idle by default; device-panther 0.1.1 (derekross#145)
  carries the GPU's always-on rule, the feedbackd rule and the sleep.conf
  drop-in. GPU and Haptics stay "testing" and now cite derekross#134, derekross#145 and the
  2026-10-28 notes. Touch stays WIP: kernel derekross#134's body records touches
  driving Hyprland and the PIN screen, but the driver fixes (kernel derekross#133)
  and the kernel option with the touch notes (derekross#149) are still in review.
  Suspend: "no" with the s2idle default named, since the notes say it
  hasn't run.
- Pixel 6 Pro (raven): Haptics "no" -> "works": HAPTICS-TORCH-PLAN H5
  passed (2026-10-28/29), cs40l25-vibra and the node merged as kernel
  derekross#114, feedbackd's rule in device-raven 0.1.27 (derekross#113). Display names
  its open kernel PR, derekross#116.
- Pixel 7a (lynx): Bluetooth "no" -> WIP: the merged WIFI-PLAN (derekross#133)
  records it scanning through its own UART; kernel derekross#98 and derekross#98 are in
  review. The Wanted line says so too.
- Pixel 7 Pro (cheetah) detail: Boot names the AVB footers a chained
  vbmeta_b needs (NOTES 2026-10-28 and 10-29, derekross#104); Suspend describes
  the merged tier 1 auto-sleep (derekross#64, off by default behind the
  omamobile-sleep toggle, P0-P5 on 2026-10-26) in place of the old
  "until S8" text. Nothing from the open derekross#171, derekross#172, derekross#174-derekross#176.
- Wanted: coral's derekross#52 is merged, so its line no longer says "in review";
  flame gets a line for its native install in review (derekross#55, derekross#56, derekross#58,
  derekross#59, split out of derekross#17); the OnePlus 13 (dodge) is listed with derekross#111 and
  kernel derekross#115, both open.

The Pixel 9a row and the Pixel 9 family line are left to derekross#165. Every
status word still parses with the website's parseCell.

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
robertzas pushed a commit to robertzas/omarchy-mobile that referenced this pull request Nov 14, 2026
… rounds

L-348 to L-361: shared drivers judged against every int-line user (derekross#150),
board pins against every stock user (derekross#151), ownership flags under the
consumer's lock (derekross#152), /chosen children and Google's panel constants
(derekross#154), same-day notes PRs on one device (derekross#208/derekross#209), merge simulations for
DEVICES rows, ours included (derekross#165), stacked PRs that merged main (derekross#115/derekross#117),
titles behind renumbered versions, a contributor merge that dropped the PR
(derekross#99), smoke package lists as provenance and "owner-authorized" writes
(derekross#200), promised fixes and named tags checked on the remote (derekross#153/derekross#157/derekross#158),
and file swaps into old packages (derekross#193). C-23 (usbc-deadman's device scope,
kernel derekross#107) and C-24 (BNEP in a lynx-only fragment, derekross#98).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
robertzas pushed a commit to robertzas/omarchy-mobile that referenced this pull request Nov 14, 2026
The matrix re-read against what merged and ran on hardware since 2026-11-02:
- Pixel 7: Bluetooth and the torch move to testing (kernel derekross#196, derekross#197, derekross#426); fingerprint moves to WIP (kernel derekross#207; the package side, derekross#458, is in review); AVB footers and reboot=warm noted.
- Pixel 8a: a new row, from first light and UFS behind the guard (kernel derekross#192, derekross#206; derekross#414).
- Pixel 6 Pro: GPU scaling and cooling (kernel derekross#189) and UFS suspend (kernel derekross#188) merged.
- Pixel 9a: GPU levels merged through kernel derekross#96; GNSS's moving-car session (derekross#114) and the link driver (kernel derekross#185).
- Pixel 7a: Bluetooth's device tree merged (kernel derekross#98).
The Pixel 8 Pro's row is left to derekross#472.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@FlorencioJP
FlorencioJP force-pushed the lynx-bluetooth branch 2 times, most recently from 14c97c3 to e8f9f40 Compare November 14, 2026 13:06
@FlorencioJP

Copy link
Copy Markdown
Contributor Author

Thanks, @derekross. Both are in, head e8f9f405, rebased onto main at 6649b76 (#436 landed in the same NOTES and version, so the branch had stopped merging):

  1. The boot is in NOTES, under the Bluetooth entry: the local kernel (7.3.0-rc6-omamobile-lynx-wifi1), omamobile-device-lynx built on the phone from this branch with tools/phone-install-package (pacman -Q said 0.2.3-1), my local omamobile-lynx-bluetooth 0.1.2 supplying the modules with its address service and rule masked, and the two timings (0.05 s at boot, 1.07 s after the rebind). The DEVICES cell now says "booted on the phone with a local kernel build".
  2. The version: the rebase forced a number, since main is at 0.2.4 now. I took 0.2.5 in version, the two DEVICES lines, the Address row; change it at merge as you see fit.

The two questions:

  • Firmware. Not by hand: linux-firmware-atheros came in as a dependency of my local omamobile-lynx-bluetooth package. A fresh image would not have it, so linux-firmware-atheros is now in device-lynx's depends.
  • What the local package still does. It carries out-of-tree copies of hci_uart, btqca, pwrseq-qcom-wcn and bnep for the local kernel (in updates/), a modules-load.d file for hci_uart and bnep, a softdep so the sequencer loads first, and the firmware dependency; its address service and rule are masked. On the published kernel the modules are in-tree, so I expect only the firmware dependency to matter, but I have not booted that kernel on this phone yet. NOTES says so.

Noted on the hci0:N match; thanks for folding lynx into #423.

@derekross

Copy link
Copy Markdown
Owner

Thank you, @FlorencioJP. Both points are in, and well done: the 11-12 boot is in NOTES with the local kernel, the package built on the phone, your own module package with its address service masked and the two timings, and the cell says "booted on the phone with a local kernel build". Thank you also for chasing the firmware: linux-firmware-atheros in depends is the right fix. I checked that it carries qca/hpbtfw21.tlv and qca/hpnv21.bin, and our image installs only linux-firmware-other, so a fresh 7a would otherwise have had the driver and no firmware.

Main has moved under the branch again (#751's Wi-Fi result landed in the same lines), so it needs one more rebase. Here is what to keep in each conflict:

  1. docs/DEVICES.md, the 7a's matrix row: take main's row and replace only the Bluetooth cell with yours. The Wi-Fi cell stays main's ("WIP (one guarded AoC-first scan, …)").
  2. docs/DEVICES.md, the 7a's Wanted row: take main's sentence, with its Wi-Fi clause ("one guarded AoC-first Wi-Fi connection/HTTPS test passed on a local experimental kernel …"), and end it with your Bluetooth clause ("and Bluetooth scans with its own address (merged, kernel device-lynx: Bluetooth on the APM UART #98, device-lynx 0.2.5, device-lynx: Bluetooth on the APM UART #98)").
  3. device/lynx/WIFI-PLAN.md, the power-pins row: take main's line. It already says "BT_EN is gpp16-2 on PERIC0 (0x10840000), described by the published gs201-pinctrl.dtsi", so your change there isn't needed any more.

NOTES, the package and the kernel config merge by themselves. Please take 0.2.6: #438 has merged first as 0.2.5 (it was ready and is inert on the published kernel), and it changed the same "install stages…" pass line in packages/omamobile-device-lynx/test/all, so keep both sets of checks there and one pass line that names both.

When you next build the package on the phone from the rebased branch, one line would be good to have: pacman -Qi linux-firmware-atheros showing omamobile-device-lynx under "Required By". It isn't a condition.

Once CI is green on the rebased head I'll merge.

Userspace side of linux-omarchy-mobile#98: the Pixel 7a's WCN6855 BT core
works on the APM USI1 UART, powered by its qcom,wcn6855-pmu node.

- kernel/configs/omamobile.config: CONFIG_BT_BNEP with its two filters, for
  PAN tethering. It can't affect another phone, so it sits next to the
  Bluetooth lines rather than in a lynx fragment the release kernel never
  reads. hci_qca, btqca and pwrseq-qcom-wcn already come from defconfig
  (checked with merge_config on defconfig plus this fragment).
- packages/omamobile-device-lynx 0.2.6: omamobile-bt-addr sets the address
  the bootloader publishes in /chosen/config/bt_addr when udev sees the
  controller, through omamobile-bt-addr@.service as on cheetah. Its own
  script: the WCN6855 is unconfigured until the address is set, so it
  answers neither `btmgmt info` nor a power cycle; the script retries
  `btmgmt public-addr` while the index is not yet registered after a rebind.
  The package depends on linux-firmware-atheros, which holds the chip's
  firmware files and is not in the image otherwise.
- device/lynx/NOTES.md: the 2026-10-29 Bluetooth entry, with the PMU boot
  evidence (sequencer held by hci_uart, the buck's users, both pins, the
  drop on unbind), the module order, the rebind race, and the gph2-5 bench
  note for Wi-Fi work, and the 2026-11-12 boot asked for at review (local
  kernel build, package built on the phone, the two timings); two stale
  sentences in the Wi-Fi entries fixed, and the same BT_EN line in
  WIFI-PLAN.md.
- docs/DEVICES.md: Bluetooth testing on lynx (scans with its own address; the
  kernel side is merged and published; booted on the phone with a local
  kernel build; pairing untested).

Assisted-by: LLM
Signed-off-by: Jean-Paul Florencio <jpflorencio@version-tracker.app>
@FlorencioJP

Copy link
Copy Markdown
Contributor Author

Rebased onto main (17f9a3c), head 41321a7, version 0.2.6. Resolved as you listed: DEVICES.md keeps main's Wi-Fi cell and Wanted clause with only the Bluetooth cell and clause from this PR, WIFI-PLAN.md takes main's power-pins line, and test/all keeps #438's display-gate checks next to the three bt-addr checks under one pass line that names both. Policy and Tests are green on this head. The pacman -Qi linux-firmware-atheros line will follow when I next build the package on the phone.

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