You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
J815: N1 ROM boot answers response=0x6 and stays at stage 0 (26.6.2 and 27.0 FTAB, both board words) #2575
On this kernel everything else comes up: display, GPU, NVMe, keyboard, trackpad and sound.
Problem: the N1 probes, but the ROM boot (rom_boot = 1) never gets past stage 0:
pci-host-generic 1cb0000000.pcie: PCI host bridge to bus 0000:00
OF: /pcie@1cb0000000: no iommu-map translation for id 0x0 on (null)
pci 0000:01:00.0: [106b:1900] type 00 class 0xff0000 PCIe Endpoint
pci 0000:01:00.0: 2.000 Gb/s available PCIe bandwidth, limited by 2.5 GT/s PCIe x1 link at 0000:00:00.0
pcieport 0000:00:00.0: of_irq_parse_pci: failed with rc=-22
apple_centauri 0000:01:00.0: of_irq_parse_pci: failed with rc=-22
apple_centauri 0000:01:00.0: ACIPC endpoint reachable: stage=0x0, 40-bit DMA mapping=0x00000000ffff8000
apple_centauri 0000:01:00.0: ROM boot: apple/centauri/j815-ftab.bin, 31109444 bytes, board word 0x268142
apple_centauri 0000:01:00.0: ROM firmware result: response=0x6 stage=0x0 mailbox=0xa5a0 poll=-110
After that, rom_boot = 2 and rom_boot = 3 return -EBUSY.
What I tried: one ROM boot per cold boot, four boots in total. All four combinations gave the same response=0x6 stage=0x0 poll=-110:
FTAB (Wireless1,* of j815ap, Firmware/t2026macG1/Release/ftab.bin)
board word
26.6.2 (25G83), 34108536 bytes
0x268142 (table)
26.6.2 (25G83), 34108536 bytes
0x248142 (J813)
27.0 (26A428), 31109444 bytes
0x268142 (table)
27.0 (26A428), 31109444 bytes
0x248142 (J813)
So the ROM reads the descriptor and answers, and the MSI write lands in the mailbox page. But the image is never accepted or executed.
#1530 says it was tested on a J815 with a 27.0 stub. My stub is 26.6.2 while the system firmware is 27.2.
Questions:
What does ROM response 0x6 mean?
Does the N1 ROM boot depend on the stub (os-fw) version, or on something the stage 2 must hand over for the PCIe port / N1, as m1n1/iBoot do for J813 on 26.6.2?
I can test patches or extra debug output on this machine.
Same symptom on a J813 (Mac17,3), system firmware 27.2 (mBoot-20457.41.3) and a 27.2 (26B123) stub: response=0x6 stage=0x0 poll=-110 with the raw Firmware/t2026macG1/Release/ftab.bin of both 26B123 and 25G83.
The cause was the FTAB, not the firmware versions. The N1 ROM wants the board's Image4 manifest in the FTAB's ticket slot, as sithglan described in aurora-silicon/linux-aurora#18 and #26 (and m1n1#65 for the J716s):
take centauri.<board>ap.im4m from the same build as the FTAB (in the IPSW under Firmware/Manifests/restore/macOS Customer/centauri/, or /System/Volumes/Preboot/<vg>/boot/<hash>/usr/standalone/firmware/Centauri/); for the J813 it's 7,772 bytes, the production one (not .dev/.insecure)
append it after the FTAB's last entry
write its offset and length as little-endian u32s at FTAB header +0x10 and +0x14 (zero in the raw file)
With the 26B123 FTAB + centauri.j813ap.im4m stitched that way, on a fresh boot:
apple_centauri 0000:01:00.0: ROM boot: apple/centauri/j813-ftab.bin, 31213200 bytes, board word 0x248142
apple_centauri 0000:01:00.0: ROM firmware result: response=0x1 stage=0x4 mailbox=0xa5a0 poll=0
then rom_boot 2 and 3 reach "Alpha Wi-Fi firmware running", wlp1s0f0 registers with the factory MAC, and iw scan finishes all 38 channels. btcentauri binds once the firmware runs.
Two more things I hit:
One ROM attempt per cold boot. After a rejected image, later attempts (after a module reload) keep answering 0x6, even with a good file, until a reboot.
25G83's raw FTAB is 34,108,536 bytes; with its 7,772-byte manifest that's within a byte of the 34,116,307-byte j813-ftab.bin in centauri: boot the N1 firmware automatically on J813 #1945, so that working setup most likely had it stitched too.
For the J815, centauri.j815ap.im4m is in the same directory in the 27.x builds. The 27.0 one should match your 27.0 FTAB.
@GameOverStudios71, niteria's J813 result above most likely applies to your J815 as well: response=0x6 stage=0x0 is what the N1 ROM answers when the FTAB's ticket slot is empty (offset and length at header +0x10/+0x14 both zero, as in the raw ftab.bin).
Could you try once with the 27.0 FTAB plus centauri.j815ap.im4m from the same 27.0 build, stitched as described above, and post the ROM boot: / ROM firmware result: lines? Two things to keep in mind:
Do it on a cold boot. After one rejected image the ROM keeps answering 0x6 to later attempts until a reboot, even with a good file.
Use the production manifest, not a .dev or .insecure one.
On our side, the driver should say "FTAB has no ticket" instead of a bare 0x6 when the slot is empty; that is queued as a separate change.
Activity
niteria commentedon Nov 20, 2026
Same symptom on a J813 (Mac17,3), system firmware 27.2 (mBoot-20457.41.3) and a 27.2 (26B123) stub:
response=0x6 stage=0x0 poll=-110with the rawFirmware/t2026macG1/Release/ftab.binof both 26B123 and 25G83.The cause was the FTAB, not the firmware versions. The N1 ROM wants the board's Image4 manifest in the FTAB's ticket slot, as sithglan described in aurora-silicon/linux-aurora#18 and #26 (and m1n1#65 for the J716s):
centauri.<board>ap.im4mfrom the same build as the FTAB (in the IPSW underFirmware/Manifests/restore/macOS Customer/centauri/, or/System/Volumes/Preboot/<vg>/boot/<hash>/usr/standalone/firmware/Centauri/); for the J813 it's 7,772 bytes, the production one (not.dev/.insecure)With the 26B123 FTAB +
centauri.j813ap.im4mstitched that way, on a fresh boot:then
rom_boot2 and 3 reach "Alpha Wi-Fi firmware running",wlp1s0f0registers with the factory MAC, andiw scanfinishes all 38 channels. btcentauri binds once the firmware runs.Two more things I hit:
j813-ftab.binin centauri: boot the N1 firmware automatically on J813 #1945, so that working setup most likely had it stitched too.For the J815,
centauri.j815ap.im4mis in the same directory in the 27.x builds. The 27.0 one should match your 27.0 FTAB.🤖 Generated with Claude Code
RyanTheTide commentedon Nov 20, 2026
@GameOverStudios71, niteria's J813 result above most likely applies to your J815 as well:
response=0x6 stage=0x0is what the N1 ROM answers when the FTAB's ticket slot is empty (offset and length at header +0x10/+0x14 both zero, as in the rawftab.bin).Could you try once with the 27.0 FTAB plus
centauri.j815ap.im4mfrom the same 27.0 build, stitched as described above, and post theROM boot:/ROM firmware result:lines? Two things to keep in mind:.devor.insecureone.On our side, the driver should say "FTAB has no ticket" instead of a bare 0x6 when the slot is empty; that is queued as a separate change.
Assisted-by: LLM (Claude:Opus 5.5)