github TheUnboundDeveloper/AM-Reaper v3.3.2-beta-RT-BE88U
Reaper v3.3.2-beta — RT-BE88U

latest release: v3.3.2-beta-RT-BE96U
pre-release4 hours ago

v3.3.2 — a flash releases the USB volumes first, the setup box ends in the reboot screen, the ZenWiFi BQ16 joins

  • A firmware flash releases the USB volumes before it ejects them (field, RT-BE86U). Every
    flash path - the GUI upload handlers, the upgrade service and the update-check script - ran
    ejusb -1 0 before anything else was stopped. Stock's unmount stops only the NAS services and
    the ASUS apps, so Entware's daemons (they stop from the user's services-stop, which only the
    normal reboot path runs), a dnsmasq writing its log under /opt and rtrafd's history store
    kept the volume busy: the eject failed for forty seconds, fell back to a lazy detach, and the
    router flashed with a filesystem that was never cleanly unmounted. The next boot mounted it
    only after dnsmasq had already failed to start. A new step, reaper_usb_release
    (rc/reaper_usbrel.c), runs just before each of those ejects (and before the factory reset's)
    and does what a reboot would have done first: the user's services-stop (once per flash),
    rtrafd (it saves on SIGTERM), then every process still holding a file, its working directory,
    its executable or a mapped library under /tmp/mnt - SIGTERM, five seconds, SIGKILL, names
    logged. It does nothing while no USB volume is mounted, and the GUI's own Eject does not use
    it. The watchdog leaves dnsmasq alone while it runs, so it cannot restart one that would
    reopen a log on the volume. Behaviour change: a dnsmasq holding a file on the USB volume is
    stopped for the length of the flash. Test test_usb_flash_release.py runs the sweep against
    real holder processes.
  • A boot clears a service request that survived the flash. The new image refused every
    notify_rc with "rc could not be informed (rc_last:restart_upgrade)" because rc_service still
    named a service from the previous image; nothing on a boot ever cleared it. init.c now unsets
    rc_service, rc_service_pid and last_rc_service beside the ASUS_STOP_COMMIT unset -
    nothing can be mid-service that early. (The same report's reboot loop is not explained by
    either fix and is still open.)
  • The first-boot setup box (field, GT-BE98). Chrome's saved-login autofill could put the
    router password into the Wi-Fi password field, because both sat in one form and a browser
    pairs a saved login with the first visible password field there. The Wi-Fi fields now live
    outside the credential form, and the Wi-Fi key and the new login password are marked as new
    passwords. Nothing is prepopulated any more, the router login name is lowercase only (folded
    as you type), and the button reads Apply & Reboot. The form used to send a wait of zero,
    so the hidden frame's reply sent the page back to itself within seconds while the router went
    down; it now sends the model's reboot time, and the page shows the same REBOOTING screen the
    GUI's Reboot shows - countdown, then it waits for the router to answer and goes to the login
    page.
  • ZenWiFi BQ16 (BE25000) joins the roster. Quad-band 2.4/5/5/6 GHz on the BCM4916, built
    from the ASUS 102_39256 GPL drop for every closed object and the stock 102_39256 image for the
    radio firmware (BCM6717 + BCM6726, both stamped for the BQ16). networkmap is pinned to the
    fleet's binary, so the client list does not hit the shared-memory mismatch the GT-BE98 once
    had. No Realtek switch on this board (Broadcom PHYs and a BCM53134), so no switch blob. Both
    variants build and pass verification; the System Information page maps its four radios like
    the GT-BE98's. It publishes only as a prerelease until a unit has booted an image.

Images & checksums (RT-BE88U)

Two flashable images: + AI Advisor (default) and Standard (noMCP, all AI components compiled out entirely). Flash the *_nand_squashfs.pkgtb via Administration > Firmware Upgrade.

This is a beta release. Its filename carries _BETA and the router reports the same string on the dashboard and the About page, so you can always tell which channel a flashed image came from. Stable releases carry no marker.

Variant File SHA-256
+ AI Advisor RT-BE88U_3006_102.8_Reaper_v3.3.2_BETA_nand_squashfs.pkgtb d21a70709b4fd450a3bd4e94afa5cb7fc7b67cb68f9b8c7bcf6a0a2f8464a1ce
Standard RT-BE88U_3006_102.8_Reaper_v3.3.2_BETA_noMCP_nand_squashfs.pkgtb 572997dc0f414951e3de54ab76471e73fb1e9908fcef7d5094844537ef80915d

Verify a download against the attached SHA256SUMS-RT-BE88U-Reaper_v3.3.2.txt.


Corresponding source & reproducibility

The RT-BE88U image for v3.3.2-beta is built from this repository at tag v3.3.2-beta-RT-BE88U: the pinned Asuswrt-Merlin base (3006.102.8-beta2, a7ebfa133a) plus the complete patch series. The tag freezes the exact source that produced it.

The auto-attached Source code (zip/tar.gz) asset below is this repository at tag v3.3.2-beta-RT-BE88U (patches + docs).

Don't miss a new AM-Reaper release

NewReleases is sending notifications on new releases.