acoul:

same behavior with ath79 OpenWrt 19.07.4, r11208-ce6496d796 ath79 with the addition of the ethernet not working probably due to [[https://bugs.openwrt.org/index.php?do=details&task_id=296|this bug]]

with ar71xx OpenWrt 19.07.4, r11208-ce6496d796 atheros wifi is detected and loded just fine. ethernet is also working just fine

/sys/firmware

is empty (no wlan calibration data)

russell:

It was working for me as of r15104-451c1eb8c2

Also works for me as of r15139-33c27ccf4a

acoul:

Hi Russell,

can you post a dmesg from your working device ? do you have wlan calibration data under the /sys/firmware directory ?

is your device an AR7240 or an AR7241 & what ubiquiti firmware version you had prior to flashing openwrt on your device ?

about the calibration data on this device, I found [[https://community.ui.com/questions/Bullet-M5-Board-is-not-CALIBRATED/56f08eda-42e2-49f3-8ca1-ff5e8040803b|this information]]

acoul:

I flashed openwrt-18.06.9-ar71xx-generic-ubnt-bullet-m-squashfs-factory.bin and everything worked fine

I then flashed XM.v5.5.10.24241.141001.1649.bin & then openwrt-ath79-generic-ubnt_bullet-m-ar7240-squashfs-factory.bin (OpenWrt SNAPSHOT, r15139-33c27ccf4a), but still, ath9k does not detect the Qualcomm Atheros AR5008 Wireless chip

I wonder if this issue has to do with [[https://cateee.net/lkddb/web-lkddb/ATH9K_PCI_NO_EEPROM.html|this option]] which is currently disabled

I traced the origins of this behavior on initial Gabor Juhos [[https://frickel.cloud/firmware/openwrt/-/commit/b445943cdf967696f5c61b4e19aa3b984c8ffb15|pci ath9k fiup]] works.

russell:

I don't know what you'd expect under /sys/firmware. All I see is some references to the device tree. The calibration data is in the art partition, afaik. I see data when I do this:

dd if=/dev/mtd8 | hexdump -C

I flashed this so long ago, I don't recall what version of Ubiquiti's software was on it, but it would have been 5.5.something. Here's the hash of my u-boot:

# dd if=/dev/mtd0 bs=64k | sha256sum
4+0 records in
4+0 records out
4371527ea4d5ff526884ed387fbc084c73722b76488a0d3dc02a55c297e658e8 -


Here's my bootlog (note, this is a custom build, it has batman-adv installed).

U-Boot 1.1.4.2 (Jul 13 2010 - 20:06:30)

[ 0.000000] Linux version 5.4.81 (openwrt@hawg) (gcc version 8.4.0 (OpenWrt GCC 8.4.0 r15142-0efbfef4f6)) #0 Sat Dec 5 09:26:39 2020
[ 0.000000] printk: bootconsole [early0] enabled
[ 0.000000] CPU0 revision is: 00019374 (MIPS 24Kc)
[ 0.000000] MIPS: machine is Ubiquiti Bullet M (XM AR7240)
[ 0.000000] SoC: Atheros AR7240 rev 2
[ 0.000000] Initrd not found or empty - disabling initrd
[ 0.000000] Primary instruction cache 64kB, VIPT, 4-way, linesize 32 bytes.
[ 0.000000] Primary data cache 32kB, 4-way, VIPT, cache aliases, linesize 32 bytes
[ 0.000000] Zone ranges:
[ 0.000000] Normal [mem 0x0000000000000000-0x0000000001ffffff]
[ 0.000000] Movable zone start for each node
[ 0.000000] Early memory node ranges
[ 0.000000] node 0: [mem 0x0000000000000000-0x0000000001ffffff]
[ 0.000000] Initmem setup node 0 [mem 0x0000000000000000-0x0000000001ffffff]
[ 0.000000] On node 0 totalpages: 8192
[ 0.000000] Normal zone: 72 pages used for memmap
[ 0.000000] Normal zone: 0 pages reserved
[ 0.000000] Normal zone: 8192 pages, LIFO batch:0
[ 0.000000] pcpu-alloc: s0 r0 d32768 u32768 alloc=132768
[ 0.000000] pcpu-alloc: [0] 0
[ 0.000000] Built 1 zonelists, mobility grouping on. Total pages: 8120
[ 0.000000] Kernel command line: console=ttyS0,115200 rootfstype=squashfs,jffs2
[ 0.000000] Dentry cache hash table entries: 4096 (order: 2, 16384 bytes, linear)
[ 0.000000] Inode-cache hash table entries: 2048 (order: 1, 8192 bytes, linear)
[ 0.000000] Writing ErrCtl register=00000000
[ 0.000000] Readback ErrCtl register=00000000
[ 0.000000] mem auto-init: stack:off, heap alloc:off, heap free:off
[ 0.000000] Memory: 24960K/32768K available (4751K kernel code, 181K rwdata, 1096K rodata, 1196K init, 195K bss, 7808K reserved, 0K cma-reserved)
[ 0.000000] SLUB: HWalign=32, Order=0-3, MinObjects=0, CPUs=1, Nodes=1
[ 0.000000] NR_IRQS: 51
[ 0.000000] random: get_random_bytes called from start_kernel+0x32c/0x518 with crng_init=0
[ 0.000000] CPU clock: 390.000 MHz
[ 0.000000] clocksource: MIPS: mask: 0xffffffff max_cycles: 0xffffffff, max_idle_ns: 9801335621 ns
[ 0.000012] sched_clock: 32 bits at 195MHz, resolution 5ns, wraps every 11012737021ns
[ 0.007887] Calibrating delay loop... 259.58 BogoMIPS (lpj=519168)
[ 0.046048] pid_max: default: 32768 minimum: 301
[ 0.050892] Mount-cache hash table entries: 1024 (order: 0, 4096 bytes, linear)
[ 0.058187] Mountpoint-cache hash table entries: 1024 (order: 0, 4096 bytes, linear)
[ 0.072312] clocksource: jiffies: mask: 0xffffffff max_cycles: 0xffffffff, max_idle_ns: 7645041785100000 ns
[ 0.082068] futex hash table entries: 256 (order: -1, 3072 bytes, linear)
[ 0.089019] pinctrl core: initialized pinctrl subsystem
[ 0.095663] NET: Registered protocol family 16
[ 0.142525] workqueue: max_active 576 requested for napi_workq is out of range, clamping between 1 and 512
[ 0.157052] clocksource: Switched to clocksource MIPS
[ 0.163935] NET: Registered protocol family 2
[ 0.169620] tcp_listen_portaddr_hash hash table entries: 512 (order: 0, 4096 bytes, linear)
[ 0.178032] TCP established hash table entries: 1024 (order: 0, 4096 bytes, linear)
[ 0.185685] TCP bind hash table entries: 1024 (order: 0, 4096 bytes, linear)
[ 0.192728] TCP: Hash tables configured (established 1024 bind 1024)
[ 0.199295] UDP hash table entries: 256 (order: 0, 4096 bytes, linear)
[ 0.205876] UDP-Lite hash table entries: 256 (order: 0, 4096 bytes, linear)
[ 0.213310] NET: Registered protocol family 1
[ 0.217719] PCI: CLS 0 bytes, default 32
[ 0.227143] workingset: timestamp_bits=14 max_order=13 bucket_order=0
[ 0.245125] squashfs: version 4.0 (2009/01/31) Phillip Lougher
[ 0.250940] jffs2: version 2.2 (NAND) (SUMMARY) (LZMA) (RTIME) (CMODE_PRIORITY) (c) 2001-2006 Red Hat, Inc.
[ 0.282576] pinctrl-single 18040028.pinmux: 64 pins, size 8
[ 0.289665] Serial: 8250/16550 driver, 1 ports, IRQ sharing disabled
[ 0.297170] printk: console [ttyS0] disabled
[ 0.301518] 18020000.uart: ttyS0 at MMIO 0x18020000 (irq = 9, base_baud = 12187500) is a 16550A
[ 0.310207] printk: console [ttyS0] enabled
[ 0.318604] printk: bootconsole [early0] disabled
[ 0.338068] spi-nor spi0.0: mx25l6405d (8192 Kbytes)
[ 0.343196] 6 fixed-partitions partitions found on MTD device spi0.0
[ 0.349591] Creating 6 MTD partitions on "spi0.0":
[ 0.354443] 0x000000000000-0x000000040000 : "u-boot"
[ 0.360890] 0x000000040000-0x000000050000 : "u-boot-env"
[ 0.367675] 0x000000050000-0x0000007a0000 : "firmware"
[ 0.379190] 2 uimage-fw partitions found on MTD device firmware
[ 0.385208] Creating 2 MTD partitions on "firmware":
[ 0.390219] 0x000000000000-0x0000001f0000 : "kernel"
[ 0.396536] 0x0000001f0000-0x000000750000 : "rootfs"
[ 0.402945] mtd: device 4 (rootfs) set to be root filesystem
[ 0.411384] 1 squashfs-split partitions found on MTD device rootfs
[ 0.417698] 0x000000590000-0x000000750000 : "rootfs_data"
[ 0.424555] 0x0000007a0000-0x0000007b0000 : "board_config"
[ 0.431563] 0x0000007b0000-0x0000007f0000 : "cfg"
[ 0.437853] 0x0000007f0000-0x000000800000 : "art"
[ 0.447096] libphy: Fixed MDIO Bus: probed
[ 0.799191] libphy: ag71xx_mdio: probed
[ 0.804145] libphy: ar8xxx-mdio: probed
[ 0.811514] switch0: Atheros AR724X/AR933X built-in rev. 2 switch registered on mdio.0
[ 0.851248] ag71xx 19000000.eth: connected to PHY at fixed-0:00 [uid=00000000, driver=Generic PHY]
[ 0.861366] eth0: Atheros AG71xx at 0xb9000000, irq 4, mode: mii
[ 0.867711] i2c /dev entries driver
[ 0.874570] NET: Registered protocol family 10
[ 0.887018] Segment Routing with IPv6
[ 0.890965] NET: Registered protocol family 17
[ 0.895537] 8021q: 802.1Q VLAN Support v1.8
[ 0.900992] PCI host bridge /ahb/apb/pcie-controller@180c0000 ranges:
[ 0.907577] MEM 0x0000000010000000..0x0000000013ffffff
[ 0.912850] IO 0x0000000000000000..0x0000000000000000
[ 0.918384] PCI host bridge to bus 0000:00
[ 0.922562] pci_bus 0000:00: root bus resource [mem 0x10000000-0x13ffffff]
[ 0.929481] pci_bus 0000:00: root bus resource [io 0x0000]
[ 0.935102] pci_bus 0000:00: root bus resource [??? 0x00000000 flags 0x0]
[ 0.941934] pci_bus 0000:00: No busn resource found for root bus, will use [bus 00-ff]
[ 0.949947] pci 0000:00:00.0: [168c:002a] type 00 class 0x028000
[ 0.956069] pci 0000:00:00.0: reg 0x10: [mem 0x00000000-0x0000ffff 64bit]
[ 0.963075] pci 0000:00:00.0: supports D1
[ 0.967124] pci 0000:00:00.0: PME# supported from D0 D1 D3hot
[ 0.974477] pci_bus 0000:00: busn_res: [bus 00-ff] end is updated to 00
[ 0.981194] pci 0000:00:00.0: BAR 0: assigned [mem 0x10000000-0x1000ffff 64bit]
[ 1.003297] VFS: Mounted root (squashfs filesystem) readonly on device 31:4.
[ 1.020442] Freeing unused kernel memory: 1196K
[ 1.025035] This architecture does not have kernel memory protection.
[ 1.031504] Run /sbin/init as init process
[ 1.145066] random: fast init done
[ 2.257914] init: Console is alive
[ 2.261815] init: - watchdog -
[ 3.840862] kmodloader: loading kernel modules from /etc/modules-boot.d/
[ 4.212726] kmodloader: done loading kernel modules from /etc/modules-boot.d/*
[ 4.231665] init: - preinit -
[ 6.615968] random: jshn: uninitialized urandom read (4 bytes read)
[ 7.083695] random: jshn: uninitialized urandom read (4 bytes read)
[ 7.229904] random: jshn: uninitialized urandom read (4 bytes read)
[ 7.638095] urandom_read: 3 callbacks suppressed
[ 7.638107] random: jshn: uninitialized urandom read (4 bytes read)
[ 8.041590] eth0: link up (100Mbps/Full duplex)
[ 8.048098] IPv6: ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
[ 8.074665] random: procd: uninitialized urandom read (4 bytes read)
[ 12.457838] mount_root: jffs2 not ready yet, using temporary tmpfs overlay
[ 12.471368] urandom-seed: Seed file not found (/etc/urandom.seed)
[ 12.632908] eth0: link down
[ 12.655777] procd: - early -
[ 12.658927] procd: - watchdog -
[ 13.351570] procd: - watchdog -
[ 13.514313] procd: - ubus -
[ 13.568258] random: ubusd: uninitialized urandom read (4 bytes read)
[ 13.575371] random: ubusd: uninitialized urandom read (4 bytes read)
[ 13.583423] procd: - init -
[ 15.037949] kmodloader: loading kernel modules from /etc/modules.d/*
[ 15.389252] Loading modules backported from Linux version v5.8-0-gbcf876870b95
[ 15.396555] Backport generated by backports.git v5.8-1-0-g79400d9e
[ 15.630160] batman_adv: B.A.T.M.A.N. advanced 2020.4-openwrt-1 (compatibility version 15) loaded
[ 15.680215] ath9k 0000:00:00.0: Direct firmware load for ath9k-eeprom-pci-0000:00:00.0.bin failed with error -2
[ 15.690438] ath9k 0000:00:00.0: Falling back to sysfs fallback for: ath9k-eeprom-pci-0000:00:00.0.bin
[ 16.055260] urngd: v1.0.2 started.
[ 16.384366] ath: phy0: Ignoring endianness difference in EEPROM magic bytes.
[ 16.393042] ath: EEPROM regdomain sanitized
[ 16.393049] ath: EEPROM regdomain: 0x64
[ 16.393055] ath: EEPROM indicates we should expect a direct regpair map
[ 16.393081] ath: Country alpha2 being used: 00
[ 16.393086] ath: Regpair used: 0x64
[ 16.410269] ieee80211 phy0: Selected rate control algorithm 'minstrel_ht'
[ 16.413742] ieee80211 phy0: Atheros AR9280 Rev:2 mem=0xb0000000, irq=13
[ 16.506281] kmodloader: done loading kernel modules from /etc/modules.d/*
[ 16.653430] random: crng init done
[ 34.125963] jffs2_scan_eraseblock(): End of filesystem marker found at 0x0
[ 34.145113] jffs2_build_filesystem(): unlocking the mtd device...
[ 34.145123] done.
[ 34.153302] jffs2_build_filesystem(): erasing all blocks after the end marker...
[ 39.627668] eth0: link up (100Mbps/Full duplex)
[ 39.643261] br-pub: port 1(eth0) entered blocking state
[ 39.648591] br-pub: port 1(eth0) entered disabled state
[ 39.686951] device eth0 entered promiscuous mode
[ 39.711684] br-pub: port 1(eth0) entered blocking state
[ 39.717017] br-pub: port 1(eth0) entered forwarding state
[ 40.644449] IPv6: ADDRCONF(NETDEV_CHANGE): br-pub: link becomes ready
[ 41.592881] 8021q: adding VLAN 0 to HW filter on device bat0
[ 41.627962] br-pub: port 2(bat0) entered blocking state
[ 41.633297] br-pub: port 2(bat0) entered disabled state
[ 41.639039] device bat0 entered promiscuous mode
[ 41.643948] br-pub: port 2(bat0) entered blocking state
[ 41.649246] br-pub: port 2(bat0) entered forwarding state
[ 42.117161] batman_adv: bat0: No IGMP Querier present - multicast optimizations disabled
[ 42.125318] batman_adv: bat0: No MLD Querier present - multicast optimizations disabled
[ 42.803746] ath: EEPROM regdomain: 0x8348
[ 42.803758] ath: EEPROM indicates we should expect a country code
[ 42.803770] ath: doing EEPROM country->regdmn map search
[ 42.803780] ath: country maps to regdmn code: 0x3a
[ 42.803787] ath: Country alpha2 being used: US
[ 42.803792] ath: Regpair used: 0x3a
[ 42.803803] ath: regdomain 0x8348 dynamically updated by user
[ 46.685558] br-pub: port 3(wlan0) entered blocking state
[ 46.691095] br-pub: port 3(wlan0) entered disabled state
[ 46.696945] device wlan0 entered promiscuous mode
[ 46.742160] IPv6: ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready
[ 46.748834] br-pub: port 3(wlan0) entered blocking state
[ 46.754229] br-pub: port 3(wlan0) entered forwarding state
[ 48.471185] IPv6: ADDRCONF(NETDEV_CHANGE): mesh0: link becomes ready
[ 49.002802] batman_adv: bat0: Adding interface: mesh0
[ 49.007982] batman_adv: bat0: Interface activated: mesh0
[ 51.015360] done.
[ 51.017543] jffs2: notice: (1400) jffs2_build_xattr_subsystem: complete building xattr subsystem, 0 of xdatum (0 unchecked, 0 orphan) and 0 of xref (0 dead, 0 orphan) found.
[ 51.596099] overlayfs: upper fs does not support tmpfile.
[ 66.540065] device wlan0 left promiscuous mode
[ 66.544773] br-pub: port 3(wlan0) entered disabled state
[ 66.964531] batman_adv: bat0: Interface deactivated: mesh0
[ 67.516085] batman_adv: bat0: Removing interface: mesh0
[ 73.052820] br-pub: port 3(wlan0) entered blocking state
[ 73.058388] br-pub: port 3(wlan0) entered disabled state
[ 73.064266] device wlan0 entered promiscuous mode
[ 73.219758] IPv6: ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready
[ 73.226424] br-pub: port 3(wlan0) entered blocking state
[ 73.231819] br-pub: port 3(wlan0) entered forwarding state
[ 74.993646] IPv6: ADDRCONF(NETDEV_CHANGE): mesh0: link becomes ready
[ 75.802603] batman_adv: bat0: Adding interface: mesh0
[ 75.808045] batman_adv: bat0: Interface activated: mesh0

acoul:

Hello Russell & thank you for the feedback

it looks like you are using a custom build on your bullet-M. is it possible to post the contents of the following:

zcat /proc/config.gz | grep ^CON

for the above to work, image has to be build with the kernel option "IKCONFIG_PROC" enabled - wish OpenWrt builds made this option a default

and/or possibly a link of your image that I may try to flash & test on my bullet-M ?

flashing today's ath79/snapshot on my bullet-M, the problem of the ath9k not loading remains

enabling CONFIG_PACKAGE_kmod-owl-loader though, resulted on a successful detection & load of ath9k

I am attaching a dmesg diff of my (trunk) dmesg against yours & also a dmesg with the CONFIG_PACKAGE_kmod-owl-loader enabled

finally, any hints, references, pointers onto how the wlan calibration data is been loaded, utilized & used would be quite welcomed

russell:

My custom build is for a dumb-ap configuration. My diffconfig is as follows:

CONFIG_TARGET_ath79=y
CONFIG_TARGET_ath79_generic=y
CONFIG_TARGET_ath79_generic_DEVICE_ubnt_bullet-m-ar7240=y
CONFIG_DEVEL=y
CONFIG_BUSYBOX_CUSTOM=y
CONFIG_ALFRED_NEEDS_lua=y
CONFIG_BATMAN_ADV_BATMAN_V=y
CONFIG_BATMAN_ADV_BLA=y
CONFIG_BATMAN_ADV_DAT=y
CONFIG_BATMAN_ADV_MCAST=y
CONFIG_BUILD_LOG=y

CONFIG_PACKAGE_procps-ng=y
CONFIG_PACKAGE_procps-ng-free=y
CONFIG_PACKAGE_procps-ng-kill=y
CONFIG_PACKAGE_procps-ng-pgrep=y
CONFIG_PACKAGE_procps-ng-pkill=y
CONFIG_PACKAGE_procps-ng-pmap=y
CONFIG_PACKAGE_procps-ng-ps=y
CONFIG_PACKAGE_procps-ng-pwdx=y
CONFIG_PACKAGE_procps-ng-skill=y
CONFIG_PACKAGE_procps-ng-slabtop=y
CONFIG_PACKAGE_procps-ng-snice=y
CONFIG_PACKAGE_procps-ng-tload=y
CONFIG_PACKAGE_procps-ng-top=y
CONFIG_PACKAGE_procps-ng-uptime=y
CONFIG_PACKAGE_procps-ng-vmstat=y
CONFIG_PACKAGE_procps-ng-w=y
CONFIG_PACKAGE_procps-ng-watch=y
CONFIG_PACKAGE_rsync=y
CONFIG_PACKAGE_snmpd=y
CONFIG_PACKAGE_tcpdump=y
CONFIG_PACKAGE_terminfo=y
CONFIG_PACKAGE_wpad-basic=y
CONFIG_PACKAGE_wpad-basic-wolfssl=m
CONFIG_PACKAGE_zlib=y


acoul:

Hey Russell,

I made a diff of our Bullet-M dmesg

it looks like that our Bullet0M hw differs a bit

spi-nor spi0.0: mx25l6405d (8192 Kbytes)
spi-nor spi0.0: m25p64 (8192 Kbytes)

also on my Bullet, the following section is totally missing from all ath79/ar71xx available images:

On node 0 totalpages: 8192
Normal zone: 72 pages used for memmap
Normal zone: 0 pages reserved
Normal zone: 8192 pages, LIFO batch:0
pcpu-alloc: s0 r0 d32768 u32768 alloc=1*32768
pcpu-alloc: [0] 0

acoul:

Greetings Russel,

from all the above, do you have any idea as to why on my device, ath9k fails to load & to why enabling CONFIG_PACKAGE_kmod-owl-loader, ath9k successfully loads ?

is it because, for some reason, ath9k can't find the art info & owl-loader works around this ? the only info I found about the owl-loader [[https://git.openwrt.org/?p=openwrt/openwrt.git;a=commitdiff;h=e9401a2335cc442bd06209ff7becdd60e940c3de|is here]] but still this is not too enlightening for me.

// The calibration data is in the art partition, afaik. I see data when I do this: //

dd if=/dev/mtd8 | hexdump -C

same in my device, there is data on /dev/mtd8

here is the output of /proc/mtd

dev: size erasesize name
mtd0: 00040000 00010000 "u-boot"
mtd1: 00010000 00010000 "u-boot-env"
mtd2: 00750000 00010000 "firmware"
mtd3: 00180000 00010000 "kernel"
mtd4: 005d0000 00010000 "rootfs"
mtd5: 003b0000 00010000 "rootfs_data"
mtd6: 00010000 00010000 "board_config"
mtd7: 00040000 00010000 "cfg"
mtd8: 00010000 00010000 "art"

any suggestions as to how further debug this issue ?

thank you for the feedback

russell:

https://forum.openwrt.org/t/openwrt-18-06-1-dont-show-wifi-wireless-network-and-dont-work/23482/26

Note your pci id matches (googling on it might lead to more insight). Ignore the sidetrack in the forum thread mentioning broadcom, that's a red herring.

russell:

Below is a hexdump -C of what I have in my /lib/firmware/ath9k-eeprom-pci-0000:00:00.0.bin

Note the pci id at address 0x1a-0x1d (168c:002a). You might look at yours and compare. It looks like that eeprom bin is extracted from offset 0x1000 of the /dev/mtd8 partition.

This thread implies it might have been damaged by 18.06:
[[https://forum.openwrt.org/t/wifi-ar922x-not-detected-after-upgrade-18-06/18338]]

acoul:

Hi Russel,

thank you for the feedback

//Note the pci id at address 0x1a-0x1d (168c:002a). You might look at yours and compare. It looks like that eeprom bin is extracted from offset 0x1000 of the /dev/mtd8 partition. //

dd if=/dev/mtd8 of=/tmp/eeprom.bin bs=4k skip=1 count=1

here is the result on my Bullet-M5 with owl-loader enabled

sha256sum /tmp/eeprom.bin /lib/firmware/ath9k-eeprom-pci-0000:00:00.0.bin
60503673c3ecd3e0a9d9579f2b95a614e9f79fb5e7ad601086f409d3e33dcce4 /tmp/eeprom.bin
60503673c3ecd3e0a9d9579f2b95a614e9f79fb5e7ad601086f409d3e33dcce4 /lib/firmware/ath9k-eeprom-pci-0000:00:00.0.bin

I think you really nailed the issue by referring to the //pci id address//.

BTW, is your Bullet-M a 2.4GHz version ? mine is a 5GHz

let me summarize the issue so far:

on my Bullet-M5, ar71xx is working fine. under ath79/trunk though, ath9k can't load without the [[https://git.openwrt.org/?p=openwrt/openwrt.git;a=commitdiff;h=e9401a2335cc442bd06209ff7becdd60e940c3de|owl-loader helper]] which apparently does the following "magic:

dmesg | grep -i pci'|'awl'|'ath

SoC: Atheros AR7240 rev 2
PCI: CLS 0 bytes, default 32
switch0: Atheros AR724X/AR933X built-in rev. 2 switch registered on mdio.0
eth0: Atheros AG71xx at 0xb9000000, irq 4, mode: mii
PCI host bridge /ahb/apb/pcie-controller@180c0000 ranges:
PCI host bridge to bus 0000:00
pci_bus 0000:00: root bus resource [mem 0x10000000-0x13ffffff]
pci_bus 0000:00: root bus resource [io 0x0000]
pci_bus 0000:00: root bus resource [??? 0x00000000 flags 0x0]
pci_bus 0000:00: No busn resource found for root bus, will use [bus 00-ff]
pci 0000:00:00.0: [168c:ff1c] type 00 class 0x020000
pci 0000:00:00.0: reg 0x10: [mem 0x00000000-0x0000ffff 64bit]
pci 0000:00:00.0: supports D1
pci 0000:00:00.0: PME# supported from D0 D1 D3hot
pci_bus 0000:00: busn_res: [bus 00-ff] end is updated to 00
pci 0000:00:00.0: BAR 0: assigned [mem 0x10000000-0x1000ffff 64bit]
ath9k_pci_owl_loader 0000:00:00.0: enabling device (0000 -> 0002)
ath9k_pci_owl_loader 0000:00:00.0: fixup device configuration
pci 0000:00:00.0: [168c:002a] type 00 class 0x028000
pci 0000:00:00.0: reg 0x10: [mem 0x10000000-0x1000ffff 64bit]
pci 0000:00:00.0: supports D1
pci 0000:00:00.0: PME# supported from D0 D1 D3hot
pci 0000:00:00.0: BAR 0: assigned [mem 0x10000000-0x1000ffff 64bit]
ath9k 0000:00:00.0: enabling device (0000 -> 0002)
ath: phy0: Ignoring endianness difference in EEPROM magic bytes.
ath: EEPROM regdomain sanitized
ath: EEPROM regdomain: 0x64
ath: EEPROM indicates we should expect a direct regpair map
ath: Country alpha2 being used: 00
ath: Regpair used: 0x64
ieee80211 phy0: Atheros AR9280 Rev:2 mem=0xb0000000, irq=12
ath: EEPROM regdomain: 0x812c
ath: EEPROM indicates we should expect a country code
ath: doing EEPROM country->regdmn map search
ath: country maps to regdmn code: 0x37
ath: Country alpha2 being used: GR
ath: Regpair used: 0x37
ath: regdomain 0x812c dynamically updated by user

specifically the "magic" is done here:

dmesg | grep "type 00 class"

pci 0000:00:00.0: [168c:ff1c] type 00 class 0x020000
pci 0000:00:00.0: [168c:002a] type 00 class 0x028000

I don't know if this ath79 issue is related on the peculiarity [[https://git.openwrt.org/?p=openwrt/openwrt.git;a=commitdiff;h=f6385f30bd2ef202e526d964c5eebcf2b04d0675|described here]] (probably not)

in my ar71xx Bullet-M5:

dmesg | grep -i pci'|'awl'|'ath
SoC: Atheros AR7240 rev 2
registering PCI controller with io_map_base unset
PCI host bridge to bus 0000:00
pci_bus 0000:00: root bus resource [mem 0x10000000-0x13ffffff]
pci_bus 0000:00: root bus resource [io 0x0000]
pci_bus 0000:00: root bus resource [??? 0x00000000 flags 0x0]
pci_bus 0000:00: No busn resource found for root bus, will use [bus 00-ff]
pci 0000:00:00.0: [168c:ff1c] type 00 class 0x020000
pci 0000:00:00.0: fixup device configuration
pci 0000:00:00.0: reg 0x10: [mem 0x00000000-0x0000ffff 64bit]
pci 0000:00:00.0: supports D1
pci 0000:00:00.0: PME# supported from D0 D1 D3hot
pci_bus 0000:00: busn_res: [bus 00-ff] end is updated to 00
pci 0000:00:00.0: BAR 0: assigned [mem 0x10000000-0x1000ffff 64bit]
pci 0000:00:00.0: using irq 40 for pin 1
PCI: CLS 0 bytes, default 32
eth0: Atheros AG71xx at 0xb9000000, irq 4, mode:MII
PCI: Enabling device 0000:00:00.0 (0000 -> 0002)
ath: phy0: Ignoring endianness difference in EEPROM magic bytes.
ath: EEPROM regdomain: 0x0
ath: EEPROM indicates default country code should be used
ath: doing EEPROM country->regdmn map search
ath: country maps to regdmn code: 0x3a
ath: Country alpha2 being used: US
ath: Regpair used: 0x3a
ieee80211 phy0: Atheros AR9280 Rev:2 mem=0xb0000000, irq=40

I wonder if this is the spot where the ar71xx "magic" happens, which is possibly related to this [[https://git.openwrt.org/?p=openwrt/openwrt.git;a=commitdiff;h=b445943cdf967696f5c61b4e19aa3b984c8ffb15|ancient trick:]]
pci 0000:00:00.0: [168c:ff1c] type 00 class 0x020000
pci 0000:00:00.0: fixup device configuration
PCI: Enabling device 0000:00:00.0 (0000 -> 0002)

which is similar to this ath79/owl trickery:
pci 0000:00:00.0: [168c:ff1c] type 00 class 0x020000
ath9k_pci_owl_loader 0000:00:00.0: enabling device (0000 -> 0002)
ath9k_pci_owl_loader 0000:00:00.0: fixup device configuration

cheers

acoul:

digging further into this issue, I bumped onto [[https://patchwork.kernel.org/project/linux-wireless/patch/20190705175313.19482-1-chunkeey@gmail.com/|this patch]]

if the EEPROM content is not valid, a value of 0xFF1C returns
when read from the register". So, they will show up on the system's
pci bus. However in that state, ath9k can't load, since it relies
on having the correct pci-id, otherwise it doesn't know what chip it
actually is

this current BUG report looks like a duplicate of [[https://bugs.openwrt.org/index.php?do=details&task_id=1734|#1734 BUG]]

which states:

Please try to add the kmod-owl-loader (target/linux/generic/files/drivers/misc/owl-loader.c)
to your image. the owl-loader is supposed to do the fixup.

so I guess, if there are no objections, this BUG can shamefully {again} close as invalid ?

still Russel, the fact that your Bullet-M, under ath79, doesn't need owl-loder fixup is quite puzzling to me.

here is some [[https://ath9k-devel.ath9k.narkive.com/w7UrUj8N/rfc-v3-ath9k-add-devicetree-support-to-ath9k#post5|extra historical]] & a more [[https://linuxlists.cc/l/9/linux-wireless/t/3193913/(rfc_patch)_ath9k:_integrate_ar92xx_pci_fixup_code|recent]] reading on this topic.

acoul:

Hey Russell,

can you post the output of the "lspci" with & without ath9k loaded on your bullet-M ?

in my case, owl-loader is a prerequisite in order for the ath9k to successfully load.

here is the "lspci" output in my case:

no ath9k:
00:00.0 Ethernet controller: Qualcomm Atheros AR5008 Wireless Network Adapter (rev 01)

owl-loader/ath9k loaded:
00:00.0 Network controller: Qualcomm Atheros AR928X Wireless Network Adapter (PCI-Express) (rev 01)