01:05 <KGB-2> https://tests.reproducible-builds.org/openwrt/openwrt_kirkwood.html has been updated. (100.0% images and 100.0% packages reproducible in our current test framework.)

01:32 <Joral> Are there any tools that can automate (if only partially) the creation of a dts file for a new board? I have a Cisco ISA570 that I accidentally wiped the entire nand on so I don't even have the factory firmware to use for guidance. I was able to get bootable over uart with the help of kwboot but I still have a long way to go to be up and running completely.

01:34 <slh> no

01:47 <Joral> This board has two Marvell 88E6171R switch chips and an 88E1318 transceiver somehow connected to the 88F6282s two PHYs. Is there at least a way to map out how those are connected because at the moment I don't even have networking in uboot.

01:50 <dwfreed> honestly the easiest way is going to find a copy of the factory firmware that fell off the back of a truck

01:52 <Joral> If anyone can find me a source for EOL'd Cisco firmwares I would be more than happy to entertain that option but I have been searching for 6 months to no avail.

01:57 <dwfreed> might just have to buy another one

02:03 <Joral> With only $30 in it to begin with I think I will have to take the loss before I dump more money into 10 year old hardware, I was just hoping that I could get the performance of a high end home router for cheap.

06:55 <linusw> Perhaps we could get started with merging my IXP4xx patches now that I have addressed the review comments?

16:30 <jeff___m> Hey, I am having trouble with an ath11k device - the Compex WLE3002HX. I am using ath11k on a build that has kernel 5.15. The card is recognized, can scan and join networks and create an AP - however, signal is extremely weak, TX power is incorrect (30dbm reported when card only does 24) and 2.4GHz capability isn't recognized. Can anyone offer any pointers?

16:33 <jeff___m> This is the QCN6024/QCN9024 variant

16:42 <PaulFertser> jeff___m: looks like calibration data ("eeprom") isn't loaded.

16:42 <PaulFertser> jeff___m: also, are you using OpenWrt, does it mean the wireless drivers are from backports?

16:45 <PaulFertser> jeff___m: you've forgotten to pastebin full dmesg and "iw list" outputs.

16:47 <jeff___m> Standby for dmesg and iw list outputs, coming right up

16:48 <jeff___m> I'm unsure if drivers are coming from backports. I merged a fork that has my build target (imx8mm) with the current upstream master.

16:49 <KGB-2> https://tests.reproducible-builds.org/openwrt/openwrt_omap.html has been updated. (11.1% images and 100.0% packages reproducible in our current test framework.)

16:49 <PaulFertser> By default OpenWrt is using backports for all wireless drivers but you might have selected them to be statically linked in via kernel_menuconfig or something.

16:52 <jeff___m> dmesg: https://pastebin.com/aWVRysrJ

16:53 <jeff___m> iw list: https://pastebin.com/aFAQLieV

16:54 <PaulFertser> 303. [ 7.046618] Loading modules backported from Linux version v6.5-0-g2dde18cd1d8f

16:55 <PaulFertser> jeff___m: yeah, weird. Do you have a chance of testing this card on some x86 hardware with general purpose modern GNU/Linux distro?

16:55 <PaulFertser> To collect data from there for comparison.

17:02 <jeff___m> Hm, not immediately - I just need to get a Mini PCIe adapter. Well actually I could flash the host board with Ubuntu

17:02 <jeff___m> How could I tell if EEPROM isn't loaded?

17:03 <jeff___m> But wouldn't be x86 still

17:10 <PaulFertser> jeff___m: this card has integrated eeprom so should be loaded. Really strange how ath11k behaves there.

17:46 <jeff___m> Weird so even in Ubuntu it's still returning similarly low signal strengths..

17:49 <PaulFertser> Probably something's not good with the antennas?

17:49 <PaulFertser> And what about the bands, same single band?

17:50 <jeff___m> Running Ubuntu, with the atk11k-pci module loaded: https://pastebin.com/CbuPA2Dz

17:50 <jeff___m> (Thats iw list)

17:51 <jeff___m> Yeah same, single band in iw list

17:54 <PaulFertser> Please stay on channel there're people here who have much more clue about ath11k than I ever will.

17:54 <jeff___m> Will do!

18:02 <stintel> ugh, procps depends on rust/host and rust/host fails to build with some extremely vague error

18:03 <stintel> https://gist.github.com/stintel/573c899cdc96613992a91d707d7bbec3

18:03 <jakllsch> f'ing rust

18:04 <stintel> it's probably because I have my openwrt git clone on a btrfs subvolume to exclude it from snapshots of /home

18:05 <stintel> including it would cause ENOSPC fast

18:05 <stintel> but utterly stupid that it doesn't fall back to copy and rm

18:08 <stintel> here is with RUST_BACKTRACE=full: https://gist.github.com/stintel/faac2294516d2828a6725fe10ff61e9f

18:08 <stintel> does anyone see there what the values of &tempfile and dest_path are?

18:54 <jetm> Hi, I'd really appreciate your review for "armsr: armv8: set configs using arch/arm64/configs/defconfig https://github.com/openwrt/openwrt/pull/13684"

20:00 <jeff___m> Any ath11k experts floating around?

20:01 <robimarko> jeff__m: What is the issue?

20:03 <jeff___m> I have a QCN9074 device (Compex WLE3002HX).. only 5GHz is being recognized when 2.4 should also appear, the band TX powers are incorrect, and the device when scanning 5GHz networks is incredibly insensitive (shows -99dBm). Also when creating an AP, signal is extremely weak.

20:03 <jeff___m> dmesg: https://pastebin.com/aWVRysrJ

20:03 <jeff___m> iw list: https://pastebin.com/aFAQLieV

20:03 <robimarko> That screams BDF

20:04 <robimarko> Yes, BDF is not correct

20:04 <robimarko> board_id 0xff is not correct

20:04 <robimarko> Your vendor did not fuse the board-id so the proper BDF can be matched

20:05 <robimarko> So you end up with the default one which is obviously incorrect

20:08 <jeff___m> Can I overwrite it with the correct board id or BDF somehow?

20:09 <jeff___m> Google shows me others have apparently had issues with the wrong BDF/board id of 0xff on ath11k boards so you may be right on the money

20:17 <robimarko> Well, you can add a ath11k variant in the DTS if you have it pernamently plugged into one board

20:18 <robimarko> Then you can find the correct BDF and package it

20:22 <PaulFertser> Where one is supposed to find a BDF for a minipcie card like that?

20:28 <robimarko> Well, usually vendor ships some botched up driver and BDF for cases when they messed things up

20:29 <jeff___m> Someone else with the same issue - he mentions a variant of my board with the problem..

20:29 <jeff___m> https://yhbt.net/lore/all/1734146.k9CKHTOI9D@ripper/T/

20:30 <robimarko> That is the same issue

20:30 <robimarko> First Pineapple series had the same unfused issue

20:30 <robimarko> They replaced those for free

20:32 <jeff___m> I don't follow the part where he says "I could just overwrite it using qcom,board_id (or qcom,ath11k-calibration-variant)". Before I go yell at Compex and not get any response, does that mean anything to you guys as far as overwriting the board id to see if the board starts functioning

20:34 <PaulFertser> Damn that is cursed...

20:34 <robimarko> Like I explained, you can set a qcom,ath11k-calibration-variant in the DTS and package the proper BDF with it

20:34 <robimarko> But, that will only work if you dont move the card around as its just a hack

20:38 <jeff___m> Yeah, I understand. I'm planning on using this card with this specific piece of hardware so the hack is fine, I just can't figure out how to identify the correct BDF and where I can actually change that in the DTS (and how to find out what i need to set the correct board id to)

20:38 <Slimey> heh been away for a while, we bought a new house

20:46 <stintel> congratz and wb :)

20:53 <robimarko> jeff__m: Only the ODM can tell you what is the correct one

21:29 <KGB-2> https://tests.reproducible-builds.org/openwrt/openwrt_lantiq.html has been updated. (95.9% images and 100.0% packages reproducible in our current test framework.)