09:10 <csharper2005> Can anyone help with this (wiki) - https://forum.openwrt.org/t/add-new-brand-mercusys-to-toh/159850/4? tmomas can do this, but they last seen on May 11 :(

09:12 <PaulFertser> csharper2005: do you know how to do it, where to edit the list?

09:12 <PaulFertser> (if it's even possible via wiki web interface)

09:18 <csharper2005> PaulFertser: unfortunately, no. I guess it's not exactly easy. tmomas wrote me when I asked to add Etisalat brand: "You could ask on IRC if someone knows how to add a new brand. If nobody knows, please come back to me."

09:19 <PaulFertser> csharper2005: ok, I'll see what I can do

09:20 <csharper2005> PaulFertser: thanks!

09:48 <Christophe[m]1> Hey everyone I can't get my head around that entropy issue on the bpi-r3 I mentioned two days ago. Shouldn't mtk_rng and urandom-seed deal with this? Kernel 6.1 includes even more rng improvements by Jason Donenfeld https://lore.kernel.org/lkml/20221003174431.1189919-1-Jason@zx2c4.com/

10:20 <PaulFertser> csharper2005: hm, the brand name is all caps?

10:21 <PaulFertser> csharper2005: guess I now understood almost everything that's needed to create a new brand there on the wiki.

10:24 <PaulFertser> csharper2005: please create the hwdata page now, it's in the selector.

10:30 <PaulFertser> csharper2005: also if everything works nicely please coordinate with the other forum members who need to add a new brand, I can do it ~easily now, I hope.

10:40 <csharper2005> PaulFertser: It works. I've managed to create the new device page with new brand - https://openwrt.org/toh/mercusys/mr90x_v1 Thanks a lot!

10:42 <PaulFertser> csharper2005: but first step should be adding hwdata

10:42 <csharper2005> PaulFertser: I guess Mercusys would be more traditional spelling than MERCUSYS. Is it possible to change?

10:42 <PaulFertser> csharper2005: I checked their website, they have all caps everywhere, so went with that.

10:43 <csharper2005> PaulFertser: Yes. I'll add hwdata soon.

10:43 <csharper2005> PaulFertser: Ok :)

10:44 <PaulFertser> csharper2005: I'm not using web forum so can't offer the others here to add other brands, please free to share the word

10:48 <csharper2005> PaulFertser: Ok, thanks. Should I add your contact here - https://openwrt.org/meta/create_new_device_page? "Is your Brand missing from the dropdown?"

10:48 <PaulFertser> csharper2005: hm, I guess that makes sense

10:49 <PaulFertser> csharper2005: please ping me after you add hwdata, I need to check https://openwrt.org/toh/hwdata/mercusys/start and https://openwrt.org/toh/mercusys/start

10:50 <PaulFertser> csharper2005: btw, thank you very much for all you're doing for proper upstream OpenWrt support

10:51 <PaulFertser> (instead of wasting time on 4pda forks)

10:54 <csharper2005> PaulFertser: I can't edit https://openwrt.org/meta/create_new_device_page and update the contacts. Ok, I'll keep you informed. And thanks for the kind words.

10:54 <PaulFertser> This needs a note too https://openwrt.org/meta/create_new_dataentry_page

10:57 <PaulFertser> Heh, "contact a wiki admin" hyper link points to https://openwrt.org/contact which has no information about any wiki admins.

11:28 <PaulFertser> djfe: do you need any other brands added?

11:40 <djfe> maybe, need to check my posts in the forum 😅

11:42 <csharper2005> PaulFertser: I added hwdata. It seems that https://openwrt.org/toh/hwdata/mercusys/start and https://openwrt.org/toh/mercusys/start are ok.

11:44 <PaulFertser> csharper2005: cool, thank you

11:49 <PaulFertser> Guess I need to add https://forum.openwrt.org/t/table-of-hardware-add-tozed-brand/163418 TOZED

11:50 <PaulFertser> csharper2005: do you think wiki admin is needed for https://forum.openwrt.org/t/move-beeline-devices-to-toh/165115

11:55 <csharper2005> PaulFertser: Yeah. I don't know the way how regular wiki users can move device pages from inbox to the main ToH. The mentioned pages has been moved to the ToH. I don't know who did this.

11:57 <PaulFertser> csharper2005: ok, I can do that when needed

11:58 <PaulFertser> Please someone update that TOZED thread, I added the brand.

13:14 <stintel> so I managed to boot OpenWrt on the Xirrus XD2-240... I can bring up both ethernet ports, and they work fine, until I unplug, or ip link set dev lanX down, afterwards there is no traffic possible. any ideas what this might be?

13:14 <stintel> a few times, doing link down, link up several times after each other, resulted in traffic again

16:18 <csharper2005> PaulFertser: I updated TOZED thread and completed the device page https://openwrt.org/toh/mercusys/mr90x_v1. Much better now! Why don't you use the forum?

16:28 <PaulFertser> csharper2005: thank you! I do not like web interfaces and I can't understand how a web forum can be any better than one's favourite fine-tuned MUA (or NNTP client).

16:29 <PaulFertser> A mailing list plus good archive (like lore.kernel.org) is better in every regard.

16:31 <PaulFertser> Those 8 antennas look intimidating.

16:33 <PaulFertser> csharper2005: I wonder why telnet shell is needed when ssh is enabled?

16:35 <PaulFertser> csharper2005: in buttons table you have "EXAMPLE Reset"

16:36 <PaulFertser> csharper2005: "The R3 (TX line) " but the photo says it's RX

16:37 <PaulFertser> EXAMPLE is also mentioned in UART parameters.

16:38 <PaulFertser> csharper2005: I think it's better to say "bridge" rather than "connect" on the picture.

16:40 <PaulFertser> csharper2005: R7+R6 look like a voltage divider probably to connect 5 V UART but then it means that line is Rx, not TX as the picture says.

16:41 <PaulFertser> Please double-check there's no error in the markings, as it really looks incorrect currently.

16:44 <csharper2005> PaulFertser: with up to 0.5 W TX power even more intimidating :)

16:45 <csharper2005> PaulFertser: thanks for the review. I'll try to find a way to check the UART section

16:55 <csharper2005> PaulFertser: ssh has limited access to the files (read-only or access denied). telnet gives full root access to the device.

17:08 <djfe> https://forum.openwrt.org/t/delete-duplicate-toh-entry-om5p-acv2/160654

17:08 <djfe> PaulFertser:

17:08 <djfe> https://forum.openwrt.org/t/unifi-6-lr-missing-from-ax-toh/163615/6

17:09 <djfe> 1gbit/s columns on device page have been broken since the change, what to do?

17:09 <djfe> https://forum.openwrt.org/t/table-of-hardware-add-new-fields-for-2-5-gbit-ports-and-outdoor-devices/144274/21

17:09 <djfe> An idea for toh entries, but should be discussed first

17:09 <djfe> https://forum.openwrt.org/t/toh-enhance-modem-column/160393

17:11 <djfe> and (no forum thread, yet) as already mentioned: the list of device pages in inbox. some of them have been released already (example: d-link dap-x1860)

17:20 <csharper2005> https://openwrt.org/toh/asus/ax4200 https://openwrt.org/toh/asus/tuf_ax4200 possible dublicates

17:37 <djfe> duplicates and they are empty still 😅

17:45 <csharper2005> djfe: and not in inbox. lol

17:52 <KGB-2> https://tests.reproducible-builds.org/openwrt/openwrt_lantiq.html has been updated. (4.0% images and 99.8% packages reproducible in our current test framework.)

17:52 <PaulFertser> csharper2005: but you say you're using ssh to write to init script and then to upload the OpenWrt image.

17:53 <PaulFertser> djfe: need to discuss that all in more details. And keep in mind only tmomas really knows how it's all supposed to work. I was doing quite some reading today to understand even that tiny part to add a new brand :)

17:55 <djfe> yeah just wanted to list, what has piled up in these months (haven't really worked all that much on the wiki, yet. else it would be more)

17:55 <djfe> feel free to pick any or none to address

17:55 <djfe> overall: I have no new brand to add

17:56 <csharper2005> PaulFertser: correct. And? Fortunately, Mercusys keep chmod 777 on this init script.

17:56 <djfe> thanks for adding mercusys. as it belongs to tp-link, it might not be the last two routers that find their way into openwrt. even if most of their routers have had less than 8mib flash 😅

18:03 <PaulFertser> djfe: thank you, I'll see what I can do.

18:04 <PaulFertser> csharper2005: so when you log in as root via ssh it's actually not the real root UID=0 user?

18:06 <csharper2005> PaulFertser: that's right. UID greater than 1000

18:07 <PaulFertser> csharper2005: hah, that explains it :) weird fellows

18:10 <PaulFertser> csharper2005: I thought you had serial connected to the target so it should be easy to check which one is target's Tx.

18:13 <PaulFertser> When you stop seeing messages from the board it means you disconnected Tx :)

19:01 <csharper2005> PaulFertser: From the one side - the bridges and serial pins are soldered. From the other side - uart cable is not connected, device is in operation and the device case latches are devilishly. Hence, this task requires serious selfmotivation. :)

19:02 <csharper2005> * devilishly strong

19:10 <PaulFertser> csharper2005: hm, ok. Why does your picture contradict the text? Probably you re-checked before writing the text?

19:13 <PaulFertser> If it's similar to tplink it should probably have same order of pins: https://openwrt.org/_detail/media/tplink/archerc7/tp-link_archer_c7_v2_serial.png?id=toh%3Atp-link%3Aarcher_c7

19:30 <csharper2005> PaulFertser: What do you suggest? The text has been written ~a month later at the forum. It is more likely that the picture and information in PR is correct. So, I think we should correct the text and keep this information until double check is done.

19:31 <csharper2005> From my point of view this situation has no risks. Such as hw damage and so on.

19:34 <PaulFertser> csharper2005: hw damage is unlikely but it always super annoying to me when tx and rx are mixed up.

19:34 <csharper2005> PaulFertser: I updated the device page. Now the text does not contradict the picture.

19:34 <PaulFertser> csharper2005: from the photo my guess would be that it matches tplink layout so the picture is probably incorrect, but well...

19:49 <csharper2005> PaulFertser: I agree with your idea about 5 V voltage devider. Ok, I'll delete the info about uart from the page.

20:56 <scr4tchy> Hey guys. Got a build question you'll likely find easy.. I modified the Makefile of the base package `package/firmware/ath11k-firmware/Makefile` to pull a different board file (target `Download/qcn9074-board`).

20:56 <scr4tchy> The correct file as per its sha256sum lands in `./staging_dir/target-xxx/root-ipq807x/lib/firmware/ath11k/QCN9074/hw1.0/board-2.bin` and in ./build_dir/target-xxx/ath11-firmware-2023-07-28-0064e2a/.pkgdir/lib/ath11k-firmware/lib/firmware/.../board-2.bin...` but `./build_dir/target-xxx/root-ipq806x/lib/firmware/.../board-2.bin` (and the sysupgrade image) gets the sha of the old image. I've been trying to figure it out fo

20:56 <scr4tchy> The package correct package version is said to be installed on the device after flashing the image but the sha is the old one too.

21:15 <dwfreed> pastebin the modified Makefile and the diff?

21:15 <dwfreed> also your second message got cut off at "figure it out fo"

21:16 <dwfreed> (IRC has line length limits, which that message exceeded)

21:30 <scr4tchy> dwfreed: https://github.com/scr4tchy/openwrt/commit/4151bdb5bcca8649ee2675fbb0e916da064f47ca#diff-02dcc52a7253ae5f793410753ddb9016443b1172886ab683343ee43713b7e64dR53 (only the makefile here is relevant, the wpa_supplicant change is completely unrelated)

21:30 <scr4tchy> I've been trying to figure it out for a few days to no-avail =D

21:33 <scr4tchy> The board-2.bin files on disk after a successful build; 971 is the correct file and bd7 is the old file -> https://imgur.com/a/1msFOhH .. Obviously this does not show the content of the sysupgrade image but both flash + binwalk show the old file bundled in there

21:34 <dwfreed> start with a completely clean build?

21:35 <scr4tchy> I've forced deleted ./tmp ./dl ./staging_dir ./build_dir ./bin & make clean before starting the builds, yeah. I can do a re-clone, it's true I haven't tried a fresh clone

21:37 <scr4tchy> I been scratching my head, i.e. is it due to maybe it pulling in a pre-built package from the web, or something. I don't have a `base` feed pointing to either my fork or my local folder with the commit.. but given the flashed image says the custom version is installed, it doesn't add up

21:39 <scr4tchy> Made a fresh clone as suggested, running ```cp ax9000.config .config && ./scripts/feeds update -a && ./scripts/feeds install -a && make -j$(nproc) defconfig clean download world | tee logs.txt```, will see how it goes

21:41 <dwfreed> 'git clean -dfx'

21:41 <dwfreed> will delete everything that's not in the repo

21:42 <dwfreed> equivalent of a fresh clone without actually having to reclone

21:42 <scr4tchy> haa thank you for that :+1:

21:51 <PaulFertser> 'git clean -fdX' for good measure

21:53 <dwfreed> PaulFertser: -X will only remove ignored files; -x will remove everything untracked or ignored

21:55 <PaulFertser> dwfreed: hm, I was sure one isn't a subset of the other

21:56 <PaulFertser> But the current man page says otherwise, thank you for correcting me.

22:17 <scr4tchy> This is from the fresh clone build - we still get the new file (971) in a few places, but the old file (b7fa) appears in the root (and root-orig/rootfs-overlay, whatever that is), which is absolute insanity..

22:18 <scr4tchy> I am running `find -type f -exec sha256sum {} \; | grep "b7faf05221b6ae0eef7db54e8d6f710ef76199223a55c269882a4cc4a2e98f2e"` to find all files that have the old sha, in case it can catch it being downloaded from somewhere, but I tried that a few days ago but it yielded nothing but the files in the screenshot above

22:18 <scr4tchy> Yep ok just two files (any names) in the whole repository with that sha (as per screenshot above)

22:20 <scr4tchy> Just can't quite figure out where it's finding them. `rg --no-ignore "kvalo/ath11k-firmware"` also yields nothing (which is the old download path for that board file)

22:23 <scr4tchy> All *board-2.bin* files: https://imgur.com/a/AdbakKZ

22:25 <scr4tchy> It's almost as if it's pulling the old package from some internet cache, and installing the old package in the root... but looking at `./build_dir/target-aarch64_cortex-a53_musl/root-ipq807x/usr/lib/opkg/info/ath11k-firmware-ipq8074.control` for example, it does contain the correct package version as per my Makefile: `Package: ath11k-firmware-ipq8074 | Version: 2023-07-28-006a4e2a-2`

22:26 <scr4tchy> I meant `./build_dir/target-aarch64_cortex-a53_musl/root-ipq807x/usr/lib/opkg/info/ath11k-firmware-qcn9074.control`, both opkgs are from the same Makefile

22:28 <scr4tchy> And `./staging_dir/packages/ipq807x/ath11k-firmware-qcn9074_2023-07-28-006a4e2a-2_aarch64_cortex-a53.ipk` contains the right file within it too - as verified by `xz d` the .ipk and then `tar xvf` the resulting `data.tar.gz`