Keyboard shortcuts

Press ← or → to navigate between chapters

Press S or / to search in the book

Press ? to show this help

Press Esc to hide this help

First boot of mainline Linux on mica (2026-11-02)

Redacted excerpt (ADR 0005). Raw logs: dumps/firstboot/2026-11-02-ap.log (AP console over SuzyQ) and dumps/firstboot/2026-11-02-2-lastkmsg.txt (dropbox SYSTEM_LAST_KMSG on the next stock boot). Procedure: first-boot runbook. Attempt 2026-11-02-2.

Result: Gate 2 passes

From the ramoops readback (dropbox entry stamped 2026-11-02 20:19:42, Googlebook clock):

[    0.000000] Linux version 7.3.0-ga804c5742bff (the owner@the build host) (clang version 23.1.3, LLD 23.1.3) #1 SMP PREEMPT_DYNAMIC Mon Nov  2 17:04:02 EST 2026
[    0.000000] Machine model: Dell XPS Googlebook
[    0.039540] CPU: All CPU(s) started at EL2
[   61.158726] reboot: Restarting system
  • Image: mica-usb.img sha256 15a107e5…, KERN-A on a USB 3 stick in the right port, booted with Boot from external drive. torvalds a804c57, defconfig + kernel/mica-firstboot.config, x1e80100-google-mica-el2.dtb.
  • Coreboot table: board-id 00000003; no sku-id and no fw-config node.
  • /init ran to the end: every GBOOK: line printed, then reboot -f after 60 s. No panic, no oops.

Channels

ChannelSeenNote
AP console (-if01, ttyMSM0)yes, from 0.67 searlycon: qcom_geni0 MMIO:0x894000 and the bootconsole registered, but nothing reached the SuzyQ before ttyMSM0 probed. Inference: the firmware does not leave QUP2 SE5 set up for 115200 (ticket 04’s open point), so the banner is not on the AP console
ramoops → dropboxyes, full logramoops: using 0x100000@0x86584000; read on the next stock boot
Panelfb0 upsimpledrm bound simple-framebuffer.0 from the coreboot table (no DRM_COREBOOTDRM, as expected)
EC consolepower states onlywarm reset at the reboot

For Phase 3 (from the same log)

  • arm-smmu 3da0000.iommu probe fails with -110.
  • USB (a200000, a400000, a600000, a800000): dwc3-qcom deferred, DWC3 core not registered.
  • PCIe 1bf8000 and 1c08000: root port parse deferred (-517); vdda/vddpe-3v3 fall back to dummy regulators.
  • arm-scmi: channel setup fails for protocol 0x10.
  • reserved-memory: ta@d8600000 fails to reserve; depthcharge’s ramoops and several firmware-added region@… nodes overlap the DT’s own reservations (warnings only).
  • Firmware log: depthcharge disables the NVMe controller and all USB controllers before the jump (“Exiting depthcharge with code 4”).