Skip to content

extract-vendor-firmware: the Pixel 8 Pro's Wi-Fi firmware (-d husky) - #214

Merged
derekross merged 2 commits into
derekross:mainfrom
greenbrettmichael:husky-wifi-firmware
Nov 1, 2026
Merged

derekross merged 2 commits into
derekross:mainfrom
greenbrettmichael:husky-wifi-firmware

Conversation

@greenbrettmichael

@greenbrettmichael greenbrettmichael commented Oct 31, 2026 •

Copy link
Copy Markdown
Contributor

tools/extract-vendor-firmware -d husky stages the Pixel 8 Pro's Wi-Fi firmware from the owner's own image. These are the four files the BCM4398 needs under derekross/linux-omarchy-mobile#156 (draft):

From /vendor/firmware To /usr/lib/firmware/updates/brcm
fw_bcmdhd.bin_4398_d0 brcmfmac4398d0-pcie.google,zuma-husky.bin
bcmdhd.cal_<region> ….txt
bcmdhd_clm.blob_4398_d0 ….clm_blob
bcmdhd_txcap.blob_4398_d0 ….txcap_blob

The NVRAM depends on the phone. Stock picks the CLM and TX capability files by chip revision. It picks the NVRAM by the hardware SKU (androidboot.hardware.sku): Google's bcmdhd maps the SKU to a region in dhd_custom_google.c and loads bcmdhd.cal_<region> when the vendor partition has that file. The first husky's stock log shows nvram path=bcmdhd.cal_MMW.

-s SKU uses the same table. husky's vendor has MMW, ROW and NA NVRAMs, so any other SKU, or no -s at all, gets the module's own bcmdhd.cal_USI_G5SN_9957_V17, which is the same file as _ROW. That's also what stock falls back to. The regions differ only in three energy-detect thresholds. The tool says which NVRAM it took, and it refuses -s for the other phones.

The SKU is easy to read:

  • on our image: grep hardware.sku /proc/bootconfig;
  • on Android: adb shell getprop ro.boot.hardware.sku.

Not included for husky: the sensor registry and the Bluetooth patch, since there's no AoC or Bluetooth driver yet. The Bluetooth check is skipped too. The other phones' maps are unchanged.

device/husky/kernel.config gains CONFIG_PHY_ZUMA_PCIE=m. #156's Wi-Fi needs this zuma-only PHY, and I'd set it by hand in my build. On a kernel without #156, olddefconfig drops the unknown symbol (checked on omamobile with #153).

packages/omamobile-device-husky/README.md gains a "Firmware the phone needs" section listing the four files, like tegu's.

Tested on the first husky's OTA vendor.img (CP3A.261005.005, from payload.bin):

  • -s G1MNW stages the four files the phone runs Wi-Fi with today (same SHA-256);
  • no -s stages the default NVRAM;
  • -s G9BQD stages _NA;
  • -d tegu -s … is refused.

tools/pr-check, packages/omamobile-device-husky/test/all and tests/phone-provision pass.

🤖 Generated with Claude Code

greenbrettmichael and others added 2 commits October 31, 2026 12:18
husky's Wi-Fi is a BCM4398 D0, and brcmfmac asks for
brcmfmac4398d0-pcie.google,zuma-husky.{bin,txt,clm_blob,txcap_blob}
(kernel derekross#156). -d husky copies stock's fw_bcmdhd.bin_4398_d0,
bcmdhd_clm.blob_4398_d0 and bcmdhd_txcap.blob_4398_d0 under those names,
and checks the firmware is a 4398d0 one.

The NVRAM depends on the phone. Google's bcmdhd maps the hardware SKU
(androidboot.hardware.sku) to a region with a table in
dhd_custom_google.c, and loads bcmdhd.cal_<region> when the vendor
partition has one: on the first husky (G1MNW) stock loaded
bcmdhd.cal_MMW. -s SKU uses the same table. husky's vendor has NVRAMs
for MMW, ROW and NA, so any other SKU, or none, gets the module's own
bcmdhd.cal_USI_G5SN_9957_V17 (the same bytes as _ROW), as on stock. The
regions differ only in three energy-detect thresholds. -s is refused for
the other phones.

husky has no AoC or Bluetooth driver yet, so it skips the sensor
registry and the Bluetooth patch and check. The package README lists
the four files.

Tested on the first husky's OTA vendor.img (CP3A.261005.005): -s G1MNW
gives the four files the phone runs Wi-Fi with today (the same
SHA-256), no -s gives the default NVRAM, -s G9BQD gives _NA, and
-d tegu -s ... is refused.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Signed-off-by: Brett-Michael Green <brett@bmgvisualtech.com>
kernel derekross#156's Wi-Fi needs CONFIG_PHY_ZUMA_PCIE, a zuma-only PHY, so it
goes in husky's fragment. On a kernel without derekross#156, olddefconfig drops
the unknown symbol and nothing else changes (checked on omamobile with
derekross#153).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Signed-off-by: Brett-Michael Green <brett@bmgvisualtech.com>
@derekross
derekross merged commit c392a0c into derekross:main Nov 1, 2026
2 checks passed
@derekross

Copy link
Copy Markdown
Owner

Thanks, Brett, this is a careful addition. I checked your SKU table against Google's dhd_custom_google.c (bcm4398, android-gs-shusky-6.1-android16-dp) and it matches entry for entry; I also like that the tool says which NVRAM it took and refuses -s on the other phones. Merged.

Two notes for a follow-up, neither blocking:

greenbrettmichael added a commit to greenbrettmichael/omarchy-mobile that referenced this pull request Nov 1, 2026
A NOTES entry for the BCM4398 under kernel derekross#156 (draft): the chip and
its NCI backplane, the pins, zuma's PCIe PHY, how stock picks the
firmware files (derekross#214 does the same), what brcmfmac needed, the reload
fix, throughput, what isn't done, and the BAR read that hung the phone.

DEVICES and the ROADMAP move husky's Wi-Fi from no to testing: it is
not in the kernel the image ships until derekross#156 and the PRs under it land.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Signed-off-by: Brett-Michael Green <brett@bmgvisualtech.com>
greenbrettmichael added a commit to greenbrettmichael/omarchy-mobile that referenced this pull request Nov 1, 2026
A NOTES entry for the BCM4398 under kernel derekross#156 (draft): the chip and
its NCI backplane, the pins, zuma's PCIe PHY, how stock picks the
firmware files (derekross#214 does the same), what brcmfmac needed, the reload
fix, throughput, what isn't done, and the BAR read that hung the phone.

DEVICES and the ROADMAP move husky's Wi-Fi from no to testing: it is
not in the kernel the image ships until derekross#156 and the PRs under it land.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Signed-off-by: Brett-Michael Green <brett@bmgvisualtech.com>
robertzas pushed a commit to robertzas/omarchy-mobile that referenced this pull request Nov 14, 2026
…erekross#214)

Adds husky to extract-vendor-firmware: the BCM4398 firmware, CLM blob and the per-SKU NVRAM (MMW, ROW, NA as in Google's dhd_custom_google.c; other SKUs take the default), named for kernel derekross#156's brcmfmac4398d0-pcie. -s is refused on other devices; the cheetah, panther, raven and tegu paths are unchanged. CONFIG_PHY_ZUMA_PCIE=m in device/husky/kernel.config for bring-up (it moves to the shared fragment when kernel derekross#156 merges).
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