Skip to content

Fix broken Wi-Fi on 2015–2017 MacBook Pros by shipping the missing NVRAM file - #7487

Closed
csvenke wants to merge 1 commit into
omacom:quattrofrom
csvenke:feat/macbookpro-wifi-fix
Closed

csvenke wants to merge 1 commit into
omacom:quattrofrom
csvenke:feat/macbookpro-wifi-fix

Conversation

@csvenke

@csvenke csvenke commented Aug 19, 2026

Copy link
Copy Markdown

Issue

  • 2015–2017 MacBook Pros (and any Apple machine with a BCM43602, PCI 14e4:43ba/43bb/43bc) ship with brcmfmac43602-pcie.bin but no board NVRAM (brcmfmac43602-pcie.txt) — linux-firmware has never carried it.
  • Without the NVRAM, the firmware's country detection breaks (kernel bug 193121): only 2.4 GHz networks appear, signal strength collapses, and the WPA 4-way handshake times out (deauth reason=15).
  • wpa_supplicant misreports that timeout as "pre-shared key may be incorrect" and NetworkManager as "psk mismatch" — so affected users chase their passphrase instead of the real cause.
  • This is a second, distinct firmware defect on the same chip as Apply the Broadcom Wi-Fi quirk to Macs without a T2 #6652. That quirk (feature_disable=0x82000) fixes the WPA-offload failure against WPA2/WPA3 transition APs; it was active on the machine below the whole time and was not sufficient. Both defects surface as "wrong password", which is why this one hides behind it.

Fix

Vendor the known-good NVRAM (Apple Boot Camp driver lineage, from bug 193121) and install it for Apple machines carrying a BCM43602:

  • default/firmware/apple/brcmfmac43602-pcie.txt — the NVRAM with ccode=0/regrev=1, which bypasses the broken country detection, plus a provenance header.
  • install/hardware/apple/fix-brcmfmac-nvram.sh — install-time leaf, gated on Apple DMI vendor + 14e4:(43ba|43bb|43bc) (narrower than the supplicant quirk's list: this NVRAM is 43602-family only, and BCM4360 Macs run the out-of-tree wl driver and are excluded for free). Substitutes the live interface MAC into the macaddr= line; strips it when no MAC is discoverable (the OTP fallback is proven). Never overwrites an existing file.
  • migrations/<epoch>.sh — the same repair for existing installs: idempotent, env-overridable paths, never clobbers, sets reboot-required only when it actually installed — it does not reload brcmfmac, which would drop the Wi-Fi connection carrying the update.
  • test/shell.d/brcmfmac-nvram-test.sh — 13 assertions covering gates, exclusions, MAC substitution/strip, idempotency, no-clobber, and the escalation contract.

Evidence

MacBookPro14,3 (2017, BCM43602), before → after installing the NVRAM and reloading brcmfmac:

before after
bands visible 2.4 GHz only 2.4 + 5 GHz
signal (same spot) −92 dBm −72 dBm
RX bitrate 13 Mbit/s 216 Mbit/s
WPA handshake timeout, reason=15 completes in <1s

Independent confirmations on 13,2 and 14,3: mbp-2016-linux#172, mbp-2016-linux#208.

Provenance & licensing

The NVRAM is Broadcom-proprietary, extracted from Apple's Windows driver package — which is exactly why linux-firmware can't ship it and bug 193121 has been open for years. Vendoring it here is the same trade-off, made at the distro layer; if you'd rather not carry the file in-repo, the alternatives are fetch-at-install (the canonical bugzilla source is now behind an anti-bot wall, and the gists are a supply-chain risk) or detect-and-instruct (weak). Flagging it explicitly so the call is yours.

linux-firmware ships brcmfmac43602-pcie.bin but not the board NVRAM,
and without it the 43602 firmware's country detection breaks: only
2.4GHz networks appear, signal collapses, and the WPA 4-way handshake
times out — which NetworkManager misreports as a wrong password
(kernel bug 193121).

Vendor the known-good NVRAM (Apple Boot Camp driver lineage, with
ccode=0/regrev=1 to bypass the broken detection) and install it for
Apple machines carrying a BCM43602 (14e4:43ba/43bb/43bc), at install
time and via migration for existing installs. The live interface MAC
is substituted in; the line is stripped when no MAC is discoverable.
An existing NVRAM is never overwritten, and the migration marks
reboot-required instead of reloading brcmfmac out from under the
connection carrying the update.

Verified on MacBookPro14,3: 5GHz networks appear, stronger signal and
handshake completes instantly. Complements the firmware-supplicant quirk from omacom#6652,
which was active on the affected machine but not sufficient on its own.
@csvenke csvenke changed the title Fix broken Wi-Fi on 2015–2017 MacBook Pros by shipping the missing NVRAM Fix broken Wi-Fi on 2015–2017 MacBook Pros by shipping the missing NVRAM file Aug 19, 2026
@bernardcosta

Copy link
Copy Markdown

Can confirm this fixes it on a MacBookPro14,3 (15" 2017 Touch Bar, BCM43602 14e4:43ba, Omarchy 4.0.1, kernel 7.1.9-arch1-2). Out of the box I had exactly the symptoms described here: no clm_blob available in dmesg on every boot, only 2.4GHz networks visible, terrible signal. In a room with weak coverage the scan list was completely empty, which first sent me down the broadcom-wl-dkms rabbit hole (wrong driver for this chip, and it blacklists brcmfmac — don't).

Installed the NVRAM file from this PR into /usr/lib/firmware/brcm/, reloaded the driver, and the 5GHz band appeared (channels 36–165, VHT80). Getting 160+ Mbps near the router now. I'm in Germany, so ccode=0 also works fine in an ETSI regulatory domain.

Two small notes:

  • Testing without a reboot: modprobe -r brcmfmac fails with "module in use" until you remove brcmfmac_wcc first. The migration's reboot-required approach avoids this.
  • The file declares 2 antenna chains, but the 15" board is 3x3 (HT MCS caps at 0–15 instead of 0–23). Still a massive improvement over no 5GHz at all.

@omarchybot

Copy link
Copy Markdown
Collaborator

Reviewed on a disposable worker VM, with an independent second opinion from Codex at xhigh reasoning. Three things still need a decision from you or from the maintainer, and the confirmation report above — welcome as it is — does not settle any of them.

Tests

test/shell.d/brcmfmac-nvram-test.sh (15 assertions) and ./test/cli (116 assertions) both pass on the worker. Nothing was exercised on real hardware: the coordinator has no BCM43602, so everything below is read out of the source and the vendored file rather than measured on a Mac.

Still open

1. The destination sits inside linux-firmware's own directory. install/hardware/apple/fix-brcmfmac-nvram.sh:24 and migrations/1786961462.sh:10 write /usr/lib/firmware/brcm/brcmfmac43602-pcie.txt, a path no package owns. The day linux-firmware-broadcom ships that filename, pacman refuses the upgrade for everyone who ran this. test/shell.d/unowned-system-paths-test.sh exists to catch exactly that, and it passes green on this branch: it reads the destination off the command, and here the destination is held in $dest — the escape hatch its own header documents at line 11. /usr/lib/firmware/updates/brcm/ is the kernel's override directory, searched ahead of /usr/lib/firmware, and installing there gets the same effect with no future conflict.

2. One board's calibration is installed on boards that are not that board. The vendored file declares devid=0x43ba, boardtype=0x073e, boardrev=0x1101 and rxchain=3/txchain=3/aa2g=3/aa5g=3, and the gate at fix-brcmfmac-nvram.sh:27 installs it byte-identically on 43bb and 43bc too. brcmfmac-nvram-test.sh:113 loops for wifi_id in 43ba 43bb 43bc and asserts all three receive the same file, so the suite pins this as intended behaviour rather than catching it. The antenna note in the report above is this same defect seen from the other end: 3 is the bitmask 0b011, two chains, so on a 3x3 board the file caps HT at MCS 15 instead of 23 — measurable underperformance on a 43ba machine the file otherwise works on. A confirmation on one MacBookPro14,3 cannot speak for the single-band variants, and those are where the exposure is largest, because nobody reporting here has one.

3. The MAC baked into the NVRAM may be a temporary one. fix-brcmfmac-nvram.sh:35 and 1786961462.sh:30 read /sys/bus/pci/devices/<bdf>/net/*/address, which under NetworkManager's default scan randomisation is a random locally-administered address whenever the interface is unassociated — and it is then written permanently into firmware NVRAM. Nothing in the repo overrides that default. perm_address, beside it in the same directory, is the permanent address and is the usual source for this; it reads all-zero on some drivers, so falling back to address when it does would be the complete form.

On the confirmation report

Thank you for it — before-and-after on real hardware is the part nobody else can supply, and the broadcom-wl-dkms warning will save the next person an evening. Both of your notes, answered. The brcmfmac_wcc unload order is a testing detail rather than a defect here: the migration deliberately does not reload the module (1786961462.sh:49) and sets reboot-required instead, so nobody following the shipped path hits it. And ccode=0 behaving well in an ETSI domain is a genuine data point but not a general regulatory clearance — it shows channels appearing and throughput arriving, not per-channel no-IR/DFS state or EIRP limits. Omarchy does persist a country from the timezone in install/hardware/set-wireless-regdom.sh, and brcmfmac forwards a non-world cfg80211 country through to the firmware, so the pinned ccode=0/regrev=1 acts as a fallback rather than the last word — worth saying out loud, since the PR does not currently rest on it.

Second opinion

Codex at xhigh reasoning agreed that all three findings remain open and that a single field report answers none of them; its independence is not currently guaranteed, since it can read this session's own working files, so read that as agreement rather than a separate discovery. It did contribute two things this review had not reasoned out: the arithmetic behind the antenna point — that 3 and 7 are bitmasks of enabled RF paths, and that Linux treats HT MCS 8-15 as two-stream and 16-23 as three-stream — and the argument that the observation belongs inside finding 2 rather than beside it as a fourth, since the root defect in both is one board dump generalised to hardware with a different RF topology. It also found the set-wireless-regdom.sh mitigation noted above.

Where this stands

Waiting on both of us. The maintainer has a call no amount of review settles: whether Omarchy vendors a Broadcom-proprietary blob at all. You flagged that yourself in the description and were right to. Waiting on you are findings 1 and 2 — the destination path, and either narrowing the gate to 43ba or shipping per-variant data — and finding 3, which is close to a one-line change. Nothing was pushed to your branch: the choice inside finding 2 is about hardware you have and this reviewer does not, and it is yours to make rather than to receive as a commit.

PR #7671 is an open, competing implementation of the same fix, built on a different NVRAM dump. Choosing between the two, or asking for one merged approach, is the maintainer's call; there is nothing here for you to act on.

@rand0mdud3

rand0mdud3 commented Sep 11, 2026 •

Copy link
Copy Markdown

A/B test on MacBookPro11,5 (BCM43602 14e4:43ba, same chip, different board): baseline has no NVRAM (no clm_blob in journal) yet 5GHz already works here (5180MHz, -30dBm, 400/400Mbit/s, 14x5GHz nets visible), so this tests cross-board safety, not efficacy.

Ran the actual leaf: installs a 7131-byte file with macaddr correctly substituted for this NIC's real (non-random) address. Reload recipe that works: nmcli disconnect + ip link down + modprobe -r brcmfmac_wcc, then brcmfmac (wcc must go first; plain modprobe -r brcmfmac fails 'in use'). Note reconnect needs explicit nmcli dev connect -- disconnect blocks autoconnect by design.

With NVRAM loaded: identical link (5180, 400/400), scans no worse (20x5GHz, within rescan variance), clm_blob still absent (expected -- NVRAM is not CLM). Reverted afterwards; baseline restored exactly. No degradation on this 3x3 board -- with one caveat: at VHT 400Mbit/s the HT MCS-cap concern from review is not exercised, so this does not settle that half of the cross-board question, only that nothing regresses at these rates.

One correction to the thread discussion: /sys/class/net/wlp4s0/perm_address does not exist at all on brcmfmac here (not all-zero, absent), so a perm_address-first fallback would need to handle ENOENT, not just zero content -- and would land back on address anyway, which is what the leaf already reads.

@rand0mdud3

rand0mdud3 commented Sep 11, 2026 •

Copy link
Copy Markdown

Addendum -- HT/VHT capability check for the cross-board question, same MacBookPro11,5. With the NVRAM loaded and the driver reloaded, iw phy advertises byte-identical capabilities to baseline: HT MCS 0-23 and VHT MCS 0-9 on all 3 streams, link still 400/400. So the file's rxchain/txchain=3 mask does not shrink the advertised stream support on this 3x3 board -- the predicted MCS cap does not materialize at the advertisement level. Caveat: advertisement is not runtime enforcement; forcing HT MCS 16-23 would be the final proof, but that needs a 3SS AP and band manipulation, and with identical advertisements plus identical link rates there is no signal pointing at hidden harm.

@makoni

makoni commented Sep 15, 2026

Copy link
Copy Markdown

Another MacBookPro14,3 data point, and one result that bears on the cross-board question @rand0mdud3 raised.

First, confirming @bernardcosta: this PR's file does fix 5 GHz here. MacBookPro14,3 (15-inch 2017, T1), Omarchy 4.0.3, kernel 7.2.3-arch1-3, BCM43602 14e4:43ba, Apple subsystem 106b:0173. Stock install sees no 5 GHz band at all (iw phy has no Band 2); with the file installed, 29 channels appear and the machine associates on 5240 MHz.

On this board the file does halve advertised spatial streams. @rand0mdud3's MacBookPro11,5 result was that rxchain/txchain=3 left capabilities byte-identical to baseline — HT MCS 0-23 and VHT MCS 0-9 on all three streams. That does not reproduce here. Swapping between this PR's attachment 285753 and #7671's attachment 290569 on the same machine, same session, changes what the driver advertises:

290569 (boardtype=0x61b, txchain=7) 285753 (boardtype=0x073e, txchain=3)
HT MCS 0-23 0-15
VHT caps 0x0c025820 0x0c015820
VHT RX/TX streams 3 2
Raw iw phy info, Band 2, both calibrations
===== 290569 (PR #7671) =====
aa2g=7  aa5g=7  boardrev=0x1421  boardtype=0x61b  ccode=00  regrev=245  rxchain=7  txchain=7

		HT TX/RX MCS rate indexes supported: 0-23
		VHT Capabilities (0x0c025820):
		VHT RX MCS set:
			1 streams: MCS 0-9
			2 streams: MCS 0-9
			3 streams: MCS 0-9
			4 streams: not supported
		VHT TX MCS set:
			1 streams: MCS 0-9
			2 streams: MCS 0-9
			3 streams: MCS 0-9
			4 streams: not supported

===== 285753 (this PR) =====
aa2g=3  aa5g=3  boardrev=0x1101  boardtype=0x073e  ccode=0  regrev=1  rxchain=3  txchain=3

		HT TX/RX MCS rate indexes supported: 0-15
		VHT Capabilities (0x0c015820):
		VHT RX MCS set:
			1 streams: MCS 0-9
			2 streams: MCS 0-9
			3 streams: not supported
		VHT TX MCS set:
			1 streams: MCS 0-9
			2 streams: MCS 0-9
			3 streams: not supported

So the effect looks board-dependent rather than absent: the 11,5 board keeps its three streams under this mask, the 14,3 board does not. Worth noting since #7671 gates specifically on 14,2/14,3.

Signal readings are not comparable between the two files. The RSSI offset is itself calibration data, and the difference here is large and reproducible — 14.5 dB on an unchanged machine, position and AP:

290569 (run 1):  -63.8 dBm     (mean of 5)
285753:          -78.4 dBm
290569 (run 2):  -64.0 dBm

I mention it because "signal got worse" is an easy and wrong conclusion to draw when comparing these two files; the word rssi does not appear anywhere in this thread or #7671.

Throughput did not separate them, and I would not trust anyone's single-pass numbers here, including my own earlier ones. A/B/A, three 50 MB downloads per phase:

290569 (run 1):  71.4  78.3  77.7  ->  75.8 Mbit/s
285753:          78.8  79.0  83.1  ->  80.3 Mbit/s
290569 (run 2):  66.8  64.6  48.9  ->  60.1 Mbit/s

The spread between the two identical control runs (15.7 Mbit/s) is over three times the A-vs-B difference (4.5 Mbit/s), so this environment cannot resolve it. One caveat that probably caps both: my AP advertises VHT operation: channel width 1, center freq segment 1: 42, segment 2: 50, i.e. 160 MHz, which this card does not support (Supported Channel Width: neither 160 nor 80+80). The link falls back to HT40 rates on both files, so neither calibration is being exercised near its ceiling. Someone with an 80 MHz AP would get a more meaningful comparison.

@gunars

gunars commented Sep 17, 2026

Copy link
Copy Markdown

Measured data from a 106b:015a board (MacBookPro13,3), in case it is useful here.

This NVRAM file is from subsystem 106b:0157. On 106b:015a it does enable 5 GHz — and it also stops 2.4 GHz from associating. The responsible key is a single one: sromrev=11. It does not add 5 GHz, it switches which band works, by activating the foreign board's srom11 calibration set.

Measured in both directions, after a full reboot each time (modprobe -r brcmfmac_wcc segfaults on this kernel, so reloads were avoided):

NVRAM 2.4 GHz 5 GHz
full 0157 file, 144 keys CTRL-EVENT-ASSOC-REJECT status_code=16, zero entries AP-side associates
same file minus sromrev only (one line) CTRL-EVENT-CONNECTED, AP registration table gains the station does not associate
no file associates 0 channels offered

The falsifier throughout was the AP-side registration table, not the client's own status. nmcli is unusable as an instrument here: it returns Secrets were required, but not provided from its own secrets layer before any frame reaches the radio, which reads exactly like a radio failure. Association was driven by wpa_supplicant with NetworkManager off the interface.

boardrev is excluded with its real value, not merely by removal. macOS on the same machine reports via IO80211HardwareVersion:

boardid: 0x15a   boardvendor: 0x106b   boardrev: 0x1108

The file declares boardrev=0x1101. Setting it to the measured 0x1108 — one changed line — did not restore 2.4 GHz (same status_code=16, same zero entries AP-side), while 5 GHz kept associating.

Two further measured facts:

  • With no CLM blob the card does not see DFS channels at all. An AP on 5500 MHz was invisible at 23 dBm; the same AP on 5180 MHz appeared immediately. Any 5 GHz AP for these machines has to be on ch 36–48.
  • Apple EFI passes no device properties to this PCI device: with dump_apple_properties and log_buf_len=32M, 91 devices were dumped and 0000:03:00.0 received zero. macOS IORegistry contains no wifi-calibration-msf or wifi-antenna-sku-info either. OTP on this card reads blank.

I am not proposing a fix and not asserting a cause beyond the table above. Whether correct srom11 data for 106b:015a exists in retrievable form is unknown to me — the three sources I could measure were all empty.

@makoni

makoni commented Sep 17, 2026

Copy link
Copy Markdown

@gunars — a third board, and on this one sromrev=11 does not switch bands.

MacBookPro14,3, Apple subsystem 106b:0173 (same machine as my earlier comment, now on kernel 7.2.5-3-omarchy). This PR's file unmodified, 147 keys, sromrev=11 present; identified by the same keys as before, boardrev=0x1101 / boardtype=0x073e / txchain=3. Both bands work at the same time:

result
2.4 GHz associates — CTRL-EVENT-CONNECTED, 2457 MHz (ch 10), NetworkManager then completes activation
5 GHz associates, 5240 MHz, as reported earlier
phy channel list 14 on 2.4 GHz, 29 on 5 GHz

So the three boards in this thread now read: the file originates from 106b:0157, your 106b:015a loses 2.4 GHz with it, this 106b:0173 keeps it. That is the same shape as the other divergence here — the halved stream count I measured on this board did not reproduce on @rand0mdud3's 11,5 board either. Board-dependent rather than absent, in both cases.

On your instrument warning. Agreed that nmcli is not evidence, and I did not rely on it — but my falsifier is weaker than yours and I want that on the record. I do not control this AP, so I have no registration table. What I have is the supplicant's own CTRL-EVENT-CONNECTED, i.e. the 4-way handshake completed, which the AP cannot complete without accepting the station. Association was driven through NetworkManager; the evidence is taken from wpa_supplicant, not from NM's return code.

DFS: I can neither confirm nor refute. There is no AP on a DFS channel here to test against. For what it is worth, on this board those channels are present in the phy list rather than missing:

* 5260 MHz [52]  ... No IR | Radar detection | DFS state: usable (for 734 sec) | DFS CAC time: 60000 ms
  ... through ...
* 5720 MHz [144] ... No IR | Radar detection | DFS state: usable (for 734 sec) | DFS CAC time: 60000 ms

no clm_blob available (err=-2) is logged here too. So if these cards cannot hear DFS beacons without a CLM blob, on this board at least the mechanism is not cfg80211 dropping the channels.

One methodology note that may save you reboots. You avoided module reloads because modprobe -r brcmfmac_wcc segfaults on your kernel. On 7.2.5-3-omarchy the reload is clean:

modprobe -r brcmfmac_wcc && modprobe -r brcmfmac && modprobe brcmfmac

The interface returns and reassociates in 7 seconds, with nothing in dmesg beyond the expected bus is down during teardown and a normal firmware reload. Different kernel from yours, so this is a data point rather than a contradiction.

@rand0mdud3

Copy link
Copy Markdown

Fourth board for the matrix: MacBookPro11,5, Apple subsystem 106b:0152 (BCM43602 14e4:43ba). Both bands work with and without the file here (associated 5180 MHz, 12+ 5GHz nets visible stock), and with the file loaded the advertised caps stayed byte-identical (HT 0-23, VHT 3x0-9) with unchanged link rates -- so 0152 neither needs the file nor is harmed by it at VHT rates.

@monomyth

Copy link
Copy Markdown

@bjarneo @csvenke Following the duplicate closure of #7671, could we agree on the consolidation requirements before treating the two PRs as interchangeable?

At current heads eb8c83e8 here and d5493deb in #7671:

Consolidation is reasonable, but please retain the required-key/unique-MAC behavior, board-scoped calibration selection, existing-file protections and failure-path tests. The missing-MAC strip case is not just duplicate implementation style. Calibration redistribution/provenance remains a separate maintainer decision.

Our draft #10141 contains the shared #7671 work plus its 13,3 extension. I have marked that dependency as unresolved after the closure; I am not rebasing it onto this broader installer as-is or changing installed firmware. Can maintainers confirm these differences should be reconciled here, or reconsider keeping #7671 open until that is done?

@bjarneo

bjarneo commented Sep 22, 2026

Copy link
Copy Markdown
Member

@monomyth

Hey, sorry about this being part of the automated flow. Let us open the PRs again! Still tweaking the duplication runner so I appreciate the feedback.

@omarchybot

Copy link
Copy Markdown
Collaborator

Re-reviewed at eb8c83e8, which is unchanged since the last review here, against the field reports posted since and against #7671 (d5493deb), which fixes the same bug and is open again. This was a reading review of both branches and both threads; nothing was run on a worker this time, because no worker has a BCM43602 and the question now is which of the two fixes to carry, not whether this one's shell passes. The second opinion was Codex at medium effort.

Recommendation: #7671 is the better of the two to bring to the maintainer, and this one should not also be marked ready. The reasons are the field data posted on this thread since the last review:

  • This file breaks 2.4 GHz on a board the gate reaches. @gunars measured the unmodified file (attachment 285753, sromrev=11, from subsystem 106b:0157) on a MacBookPro13,3 (106b:015a): 5 GHz associates, 2.4 GHz is rejected with status_code=16 and never reaches the AP's registration table, and removing only sromrev swaps which band works. That machine associates on 2.4 GHz today with no file at all. fix-brcmfmac-nvram.sh:27-28 and migrations/1786961462.sh:15-16 install on any Apple machine with 43ba, 43bb or 43bc, so the 13,3 gets it. Install BCM43602 NVRAM so 2017 Touch Bar Macs see 5 GHz Wi-Fi #7671 gates on MacBookPro14,2/14,3 and names 13,3 as excluded.
  • The antenna concern from the last review is now measured, not predicted. @makoni swapped both files on one MacBookPro14,3 (106b:0173): with this file the driver advertises HT MCS 0-15 and two VHT streams, with Install BCM43602 NVRAM so 2017 Touch Bar Macs see 5 GHz Wi-Fi #7671's three streams and MCS 0-23. @rand0mdud3's 11,5 kept three streams, so the effect is board-dependent, which is the point: the gate spreads one board's calibration across boards that respond to it differently. Neither report found a throughput difference their environment could resolve.
  • A MacBookPro11,5 does not need it. @rand0mdud3 found 5 GHz working at 400/400 Mbit/s with no NVRAM on 106b:0152; the file did no harm there, but the gate's premise that every Apple BCM43602 without the file is broken does not hold.
  • Stripping macaddr= is reported to crash the firmware. When no address is found, both scripts delete the line (fix-brcmfmac-nvram.sh:46-51, 1786961462.sh:39-44). Install BCM43602 NVRAM so 2017 Touch Bar Macs see 5 GHz Wi-Fi #7671 reports that on a 14,3 the firmware then times out on cur_etheraddr and the dongle fails to come up, and keeps the key with a per-machine address instead. That is a report from the other thread, not something checked here.

Two findings from the last review still stand at this head: the destination /usr/lib/firmware/brcm/brcmfmac43602-pcie.txt is the path a future linux-firmware-broadcom would ship (#7671 writes under /usr/lib/firmware/updates/brcm/ and also recognises a packaged .txt.zst), and the MAC comes from net/*/address, which can be a randomised scan address. On the second, @rand0mdud3 is right that perm_address does not exist on brcmfmac; #7671 reads the wiphy's ieee80211/phy*/macaddress first, which is where the permanent address lives. Codex added one failure path this review had not considered: sudo install writes straight to the final path, so a write truncated by a full disk leaves a file that the [[ -e $dest ]] check at 1786961462.sh:23 accepts on the retry, the migration is recorded as done, and no reboot is requested.

Codex, briefed on both pull requests without this review's answer, also chose #7671, for the same reasons: the board-scoped gate, the override directory and the MAC handling. Its independence is not guaranteed, since it can read this session's working files. It also named what remains open on #7671 — a live-address fallback when the wiphy address is unavailable, and an override that a later packaged board file cannot replace without a cleanup migration — which belongs on that pull request's own review.

Nothing was pushed to this branch. This waits on the maintainer, who has two decisions no review settles: whether Omarchy ships a Broadcom-proprietary calibration file at all, which you flagged yourself in the description,, and which of #7487 and #7671 to carry. There is nothing for you to act on until they decide.

@omarchybot

Copy link
Copy Markdown
Collaborator

Thank you for this, @csvenke. #7671 fixes BCM43602 Wi-Fi on these Macs without the board-dependent regressions field reports showed for this approach, so the work continues there.

Closed at the maintainer's request. The review above has the details; it was done by Claude Opus 5.5 with Codex Medium as a second opinion, whose agreement is not independent.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

bug Something isn't working compatibility Make Omarchy work better on everything!

Projects

None yet

Development

Successfully merging this pull request may close these issues.

8 participants