diff --git a/device/tegu/MODEM-PLAN.md b/device/tegu/MODEM-PLAN.md index 10d1b8f11..dd96bfe5f 100644 --- a/device/tegu/MODEM-PLAN.md +++ b/device/tegu/MODEM-PLAN.md @@ -1850,7 +1850,7 @@ The driver and binding are drafted as - **The stop.** `systemctl reboot` with the modem up: "CP OFFLINE", then "modem stopped (success, the CP OFFLINE before the daemons went)", the latch removed; the next boot, with the unit masked, started nothing and left no latch. - **GNSS.** Four 150 s sessions through the Location switch's socket, indoors by a window: 108 of 113, 142 of 143, 105 of 107 and 142 of 144 GGA sentences with a fix, up to 24 satellites; each session's service stopped on its idle lease with the marker gone and the CP still ONLINE. - **The GNSS bus keeps zumapro's SPI port configuration.** The first two sessions ran with the bus on gs101's compatible alone (the controller's automatic chip select, as stock uses for this device): spi-s3c64xx then cut every 64-byte block into 60 and 4 bytes, and 12 of those reads timed out (`I/O Error ... len-4 ... res-(-110)`, then `len-60`) in 13 minutes. With zumapro's compatible kept (a 64-byte block in one piece, polled; the chip select held per message) the other two sessions and 10 minutes of the modem's link cycling logged none. The feedback delay is stock's 0 in both. -- **The handshake pins as the bootloader leaves them.** The tree's GNSS nodes give the two handshake lines no pin group (linux-omarchy-mobile#237), so they run on what the bootloader set. Read on 2026-11-10 from the pin controllers' debugfs `pins` files, one controller per read, on `7.3.0-rc6-omamobile-push1-dbg-gff4fdf58035e` (the published tree plus #224, this overlay applied, the modem unit masked, so before any GNSS driver had requested them): gnss2ap_spi, `pin 4 (gpa5-4)` on `154e0000.pinctrl`, `CON(0x0) DAT(0x0) PUD(0x1) DRV(0x0)`, an input with the pull-down, which is stock's pull; ap2gnss_spi, `pin 12 (gpp2-3)` on `10840000.pinctrl`, `CON(0x0) DAT(0x0) PUD(0x1) DRV(0x0) CON_PDN(0x2) PUD_PDN(0x1)`, an input with a pull-down, where stock's group has an output with no pull at drive 0. The driver makes gpp2-3 an output when it loads, and nothing in the tree sets its pull, so the four sessions above most likely ran with that pull-down under a driven line. Not read: the two pins with the drivers loaded. +- **The handshake pins as the bootloader leaves them.** The tree's GNSS nodes give the two handshake lines no pin group (linux-omarchy-mobile#237), so they run on what the bootloader set. Read on 2026-11-10 from the pin controllers' debugfs `pins` files, one controller per read, on `7.3.0-rc6-omamobile-push1-dbg-gff4fdf58035e` (the published tree plus #224, this overlay applied, the modem unit masked, so before any GNSS driver had requested them): gnss2ap_spi, `pin 4 (gpa5-4)` on `154e0000.pinctrl`, `CON(0x0) DAT(0x0) PUD(0x1) DRV(0x0)`, an input with the pull-down, which is stock's pull (`gnss_spi_irq` under `&pinctrl_2`: an interrupt, pull-down, drive 0, the pin kept as it was in power-down); ap2gnss_spi, `pin 12 (gpp2-3)` on `10840000.pinctrl`, `CON(0x0) DAT(0x0) PUD(0x1) DRV(0x0) CON_PDN(0x2) PUD_PDN(0x1)`, an input with a pull-down, where stock's group has an output with no pull at drive 0 (`gnss_spi_int` under `&pinctrl_11`). Both groups are in Google's [`zuma-cp-s5400-phone-gnss.dtsi`](https://android.googlesource.com/kernel/devices/google/zuma/+/42395e8adcb880aa336f0da6bb2a19ed53929c40/dts/zuma-cp-s5400-phone-gnss.dtsi), which `gnssif`'s `pinctrl-0` names. The driver makes gpp2-3 an output when it loads, and nothing in the tree sets its pull, so the four sessions above most likely ran with that pull-down under a driven line. Not read: the two pins with the drivers loaded. - **The guards.** With the nodes in the device tree udev loaded `gnss_spi` and `gnssif` on a boot without the modem; `omamobile-device-tegu`'s guards now blacklist them with CPIF's modules, so only the modem start loads them. Checked on a boot with the unit masked and the package with the new lines installed: the `cpif`, `/gnss` and SPI device nodes there, none of the five modules loaded, no latch; on the boots with the modem the starter loaded all five by name. - **A call, data and the NV check.** On the fourth boot by the unit: an incoming call, accepted through the bridge, active for 8 s, hung up, the release confirmed 272 ms after the hang-up (omarchy-mobile#461), the CP ONLINE after; `omamobile-sit-data` brought `rmnet0` up and three of three pings went out over it; `omamobile-sit status` on each boot read SIM ready, home LTE, IMS voice and SMS, data bearer up. Three `systemctl reboot`s with the modem up were each confirmed. After all of it `tools/nv-backup -d tegu --check` read efs, efs_backup, modem_userdata, persist, modem_a, modem_b and metadata SAME as the 2026-10-28 backup (devinfo different, as expected). Texts, later the same evening on a sixth boot of this kernel (the modem up at 20:37 MST, the texts at about 20:40): one sent from the palmtop with `omamobile-sit sms send` (`AIMS_SEND_SMS`, error 0) arrived on the owner's phone, and his reply reached the palmtop 22 s after the send ("SMS stored and acknowledged", then the bridge's `SMS/0 received`). - **What a build from the tree has, and when.** Of the four, kernel #236 and #237 and the PHY's config line (#274) were merged on 2026-11-10, and this overlay with the GNSS link driver's config line (#594) on 2026-11-11 (UTC). So a 9a kernel built from the tree from then on has everything the in-tree start needs: the `cpif` node from this overlay, `phy_zuma_pcie` and the two GNSS modules. Only kernel #224 is open, and the modem and GNSS don't need it: it is the UFS pin state, a sleep fix.