#openwrt-devel on 2022-10-31 — irc logs at oftc.irclog.whitequark.org

2022-10-18 04:57 rmilecki changed the topic of #openwrt-devel to: Forum: https://forum.openwrt.org/ | Wiki: http://openwrt.org/ | Release: https://downloads.openwrt.org/releases/22.03.2/ | Notes: https://openwrt.org/releases/22.03/notes-22.03.2 | Logs: https://oftc.irclog.whitequark.org/openwrt-devel

00:10 Tapper has quit [Quit: Tapper]

00:11 romany0 has quit [Quit: The Lounge - https://thelounge.github.io]

00:13 romany0 has joined #openwrt-devel

00:14 goliath has joined #openwrt-devel

00:42 <soxrok2212_> Mangix: yes

00:42 <soxrok2212_> i have 802.11s + batman with 5 vlans

00:48 danitool has quit [Remote host closed the connection]

00:48 <Mangix> soxrok2212_: fun question, why does LuCI only allow WPA3 for 802.11s?

00:49 <soxrok2212_> dunno, probably because its the best security practice moving forward?

00:49 <soxrok2212_> theres no reason _not_ to choose wpa3 for it that I can think of

00:49 <Mangix> erm, what about older devices that would like to connect?

00:53 <soxrok2212_> um, WDS lol

00:54 <Mangix> ?

00:57 <soxrok2212_> perhaps there's a minimum hardware requirement? sometimes its good to drop older devices at some point...

00:58 <Mangix> that's just strangde

00:58 <Mangix> *strange

00:59 <\x> theres WDS and ad hoc

00:59 <Mangix> yeah WDS seems to be the only real alternative

00:59 <\x> infact most "mesh" is just WDS/AP+client relayed

01:00 <\x> the mesh products out there arent really using 11s hehe

01:00 <Mangix> thought they were

01:00 <Mangix> what's relayd?

01:00 <slh> iirc google wifi is the only thing trying 802.11s

01:01 <\x> Mangix: pseudo bridge

01:01 <\x> since you cant really bridge a normal client connection

01:01 <\x> wds client is fine

01:01 <\x> but a normal client isnt

01:01 <\x> that one needs relayd/kmod-trelay

01:01 <Mangix> that's uhhhh, fascinating

01:02 <Mangix> so I assume x:AP + y:WDS-client doesn't work

01:02 <\x> wds/4addr needs to be enabled on both sides

01:03 <Mangix> let's say I have 1 AP. Instead of using AP mode, what's the downside of using AP/WDS ?

01:04 <slh> not really any

01:04 <slh> normal clients can connect to AP/WDS (there might be some very buggy ones refusing, but I haven't seen any so far)

01:06 <soxrok2212_> whats strange?

01:06 <Mangix> not allowing WPA2 for 802.11s

01:06 <\x> its not strange

01:06 <\x> its by spec

01:06 <\x> just like wifi 6e

01:07 <\x> infact 11s has a loophole

01:07 <\x> it allows no encryption

01:07 <soxrok2212_> i cant find the 802.11s spec :(

01:07 <Mangix> even if I have an ac device

01:07 <Mangix> WPA3 is required?

01:07 <\x> unlike 6e where you can only do owe and wpa3-sae

01:07 <\x> yes

01:07 <Mangix> wow

01:08 <\x> 11s shouldve never allowed no encryption

01:08 <\x> should have done owe atleast

01:08 <\x> its still open but its like the wpa3 open

01:08 <soxrok2212_> does anything even support owe?

01:08 <\x> yup

01:08 <\x> my phone does

01:08 <soxrok2212_> which phone?

01:08 <\x> most android 10 and newer do

01:08 <soxrok2212_> ah

01:09 <soxrok2212_> btw you have a link to 802.11s spec

01:09 <\x> idk where to get that hehe]

01:10 <\x> but yeah so far accepted security suites are none and wpa3 sae

01:10 <\x> its not like normal client devices can use 11s anyway its mostly only APs and mesh points

01:10 <soxrok2212_> yeah

01:10 <\x> so theyll likely have sae support

01:11 <\x> older way to do this is adhoc + whatever L2 routing shit on top

01:11 <\x> 11s normally comes with olsr as default

01:11 <soxrok2212_> batman works on 11s

01:11 <soxrok2212_> works wonderfully

01:11 <\x> did you have to disable olsr?

01:11 <\x> batman works on anything I think

01:11 <\x> even on cables

01:12 <Mangix> I wonder if there's a relationship between WDS and https://hashcat.net/forum/thread-7717.html

01:12 <soxrok2212_> yep i disabled forward traffic on 11s

01:12 <\x> yeah

01:12 <soxrok2212_> batman worked beautifully over the wire too

01:13 <soxrok2212_> Mangix: wds and pmkid?

01:14 <Mangix> The forum posts talks about extra information included in EAPOL frames

01:14 <Mangix> No idea about WDS exposing that.

01:15 <\x> it says it needs roaming functions enabled

01:15 <\x> so if you dont offer FT/PSK maybe youre good against this

01:16 <Mangix> Maybe I'll run a monitor session later on and run it through Wireshark

01:17 * Mangix is mad his Brother printer does not work with 802.11w, even optional

01:18 <soxrok2212_> so it really has issues parsing that field element in teh beacon frames? lol

01:18 <Mangix> it's probably some old embedded chip

01:19 <Mangix> with garbo firmware/drivers

01:19 <soxrok2212_> tbh im surprised wpa3 isnt broken yet, though rumor was atom found somethng when he found pmkid

01:19 <\x> arent wpa3 attacks like uhhh

01:19 <\x> downgrade attacks

01:19 <soxrok2212_> yep

01:19 <\x> and like dos/exhaustion

01:19 <Mangix> soxrok2212_: I just noticed you're on that forum...

01:19 <soxrok2212_> yep

01:20 <soxrok2212_> i have a great rep on there xD

01:21 <Mangix> I still remember reporting that SHA256 PSK did not work with hashcat. Then they added support.

01:21 <Mangix> OTOH it was short lived...

01:25 <Mangix> alright back to 802.11s. So how does this work. Google Wifi has an internal 802.11s network and exposes a separate AP?

01:26 <slh> probably, yes (don't own any of them)

01:26 <\x> try 802.11s yourself and youll see

01:26 <soxrok2212_> yeah, sometimes the new mesh devices also sport a 3rd "backhaul" radio

01:26 <\x> if they follow the standartd youll see it on wpa_cli scan_result

01:26 <Mangix> because I doubt they restrict security to WPA3

01:26 <\x> it will be reported as [MESH]

01:27 <\x> well only mesh nodes will connect to it

01:27 <Mangix> right

01:27 <\x> so it wont be a problem for like legacy support

01:27 <Mangix> so clients will connect to a separate interface

01:27 <\x> yup

01:27 <\x> the clients connect to an AP

01:28 <Mangix> hrm WDS sounds simpler

01:28 <soxrok2212_> easy to do vlans over 802.11s w/ batman

01:28 <\x> wds saves you on beacons atleast if youre using the same radio

01:28 <slh> if you don't need the dynamic aspects of meshing, WDS/ 4addr is much simpler, yes

01:29 <\x> like on a single radio you can do WDS/AP compared to MESH+AP

01:29 <\x> and you can have full L2

01:29 <\x> to pass vlans most of them just use gre

01:29 <soxrok2212_> MESH ID: XXXXXX

01:29 <soxrok2212_> VHT capabilities:

01:29 <soxrok2212_> VHT Capabilities (0x339b79f6):

01:29 <soxrok2212_> Max MPDU length: 11454

01:30 <\x> linksys does for velop series

01:30 <\x> the guest network is on a gre tunnel

01:30 <Mangix> hmmmmm someone had a patch for 802.11i for openwrt. where did it go...

01:32 <soxrok2212_> dangole: whats the consensus on the partition table for the redmi ax6000? how does the community decide on that? imo we should keep it the way it currently is. users that are smart enough to flash the fw in the first place should be smart enough to fix it/revert if they brick

01:33 <Mangix> ah found it. it was stintel. wasn't 802.11i though...

01:35 <\x> oh man soxrok2212_ you gave up on ath11k right?

01:35 <\x> maaan

01:36 <soxrok2212_> yep

01:36 <\x> that thing shit hangs up on regdb change lmao

01:36 <\x> that 2.7 firmware is something

01:36 <soxrok2212_> i couldn't get it to keep working reliably, gave up on it. mtk is WAY more stable

01:36 <\x> you change country and youll be forced to reboot

01:36 <\x> https://i.imgur.com/LWySduS.png

01:36 <\x> cake perf here atleast

01:37 <\x> https://i.imgur.com/nPn2DMW.png

01:37 <\x> crash on regdb change ahaha

01:37 <Mangix> soxrok2212_: I haven't tested. hcxdumptool work well on mtk?

01:38 <soxrok2212_> it works great on my mt7610u chip. best card ive ever owned, even better than my trusty awus036h

01:38 <\x> i havent tried that hxcdumptool thing but I have an MT7921U, injection and monitoring works well

01:39 <soxrok2212_> mtk has been amazing lately. filogic is good too

01:39 <\x> i cant verify 6GHz though but it should work imo

01:39 <\x> cant verify since I only have one

01:39 <soxrok2212_> https://openwrt.org/docs/guide-user/perf_and_log/benchmark.openssl

01:39 <soxrok2212_> did an openssl bench on filogic 830/mt7986 too. killer

01:39 <Mangix> filogic?

01:39 <soxrok2212_> yeah

01:39 <\x> whats the coremark of that MT7986 soxrok2212_

01:40 <\x> Mangix: filogic is mediatek

01:40 <soxrok2212_> \x: ugh youre saying i gotta install coremark

01:40 <\x> opkg upadte ; opkg install coremark

01:40 <Mangix> thought there were no devices yet

01:40 <soxrok2212_> its a custom build, dont think coremark is available yet for filogic targets

01:41 philipp64 has joined #openwrt-devel

01:41 <\x> its just normal aarch64 cortexa53

01:41 <\x> it should work soxrok2212_

01:41 <soxrok2212_> i also think i broke something in my most recent build as opkg cant find most packages xD

01:41 <\x> if your opkg isnt configured then

01:41 <\x> https://downloads.openwrt.org/snapshots/packages/aarch64_cortex-a53/packages/coremark_2022-07-27-eefc986e-1_aarch64_cortex-a53.ipk

01:41 <\x> opkg install that

01:41 <soxrok2212_> sec

01:41 <Mangix> remove the s *cough*

01:43 <soxrok2212_> i dont even know how to understand the results

01:43 <\x> theres a coiremark 1.0 score

01:44 <soxrok2212_> CoreMark Size : 666

01:44 <soxrok2212_> Total ticks : 15795

01:44 <soxrok2212_> Total time (secs): 15.795000

01:44 <soxrok2212_> Iterations/Sec : 6964.229186

01:45 <soxrok2212_> Iterations : 110000

01:45 <soxrok2212_> that what you're looking for/

01:45 isak has quit [Remote host closed the connection]

01:45 <soxrok2212_> ?*

01:45 <\x> 7k

01:45 <\x> Iterations/Sec : 6964.229186

01:45 <soxrok2212_> hows that stack up against say mt7622

01:46 <\x> MT7622B is 4650 coremarks

01:46 <soxrok2212_> dang

01:46 <\x> so in theory you should be able to sqm cake like 700Mbps

01:46 <soxrok2212_> big improvement

01:46 <soxrok2212_> 4c at 2ghz vs 2 at 1.35

01:46 <\x> ipq6018 @ 1.6GHz does 5200

01:47 <soxrok2212_> yeah so sqm cuts at like 450mbit/s but it doesnt peg CPU

01:47 <\x> codel isnt an issue on these things though

01:47 <\x> all of these will gigabit codel

01:47 <\x> <soxrok2212_> yeah so sqm cuts at like 450mbit/s but it doesnt peg CPU

01:47 <\x> theres driver efficiency shenanigans too

01:47 <\x> for ethernet

01:47 <Mangix> Total ticks : 14018

01:47 <Mangix> CoreMark Size : 666

01:47 <Mangix> Total time (secs): 14.018000

01:47 <Mangix> Iterations : 30000

01:47 <Mangix> Iterations/Sec : 2140.105579

01:48 * Mangix cries

01:48 <soxrok2212_> what chip is that?

01:48 <\x> that sounds like ipq40xx Mangix

01:48 <\x> stock ipq40xx?

01:48 <Mangix> mt7621

01:48 <\x> ah yeah

01:48 <soxrok2212_> ah

01:48 <\x> you can oc it

01:48 <\x> a little you can get like 2500+

01:48 <Mangix> yeah...no

01:48 <Mangix> last time I tried the results were bad

01:49 <soxrok2212_> does openwrt have a standard process for OC chips

01:49 <Mangix> squashfs errors and whatnot

01:49 <\x> soxrok2212_ nope

01:49 <\x> Mangix: if you have nand you have to oc via kernel patch

01:49 <Mangix> soxrok2212_: yeah. find a device that is supported by breed bootloader and use that

01:49 <\x> for spi nor you can just use breed

01:49 <soxrok2212_> ah. not worth it lol

01:50 <\x> https://github.com/gw826943555/openwrt/commit/e1856e8fc65c75864adf295de3cea167badefaf9

01:50 <Mangix> I think it is. for mips devices especially

01:50 <\x> heres the thing for mt7621a ^^ with nand

01:50 <Mangix> my archer c7v2s are running at 1ghz

01:51 <Mangix> \x: sounds bogus. overclocking without the bootloader?

01:51 <\x> yeah Mangix

01:52 <\x> actually for nand mt7621a, breed disabled oc features

01:52 <\x> it was only kept for the spi-nor ones

01:52 <Mangix> why?

01:52 <\x> theres some issues with nand when ocing on the bootloader

01:52 <Mangix> weird

01:52 <\x> idk he just removed it

01:52 <Mangix> breed is still worth it though

01:52 <\x> back then old Mi R3G breed has oc support

01:53 <\x> later ones he removed it and said to just oc via kernel patch

01:53 <Mangix> sigh. my mind is telling me nooooooo!!!

01:54 <\x> I also use this one for ipq40xx

01:54 <\x> https://github.com/gw826943555/openwrt/commit/c273c73c909c7ecd019ed7c37e2fada6aa92f7c8

01:54 <\x> 717 -> 896MHz

01:54 <\x> makes it route and codel gigabit without any offloads hehe

01:56 <Mangix> 322-mt7621-fix-cpu-clk-add-clkdev.patch was upstreamed looks like

01:57 <Mangix> oh it wasn't...

02:00 goliath has quit [Quit: SIGSEGV]

02:00 <Mangix> yeah...patch needs rewriting

02:01 guidosarducci_ has joined #openwrt-devel

02:01 torv has quit [Remote host closed the connection]

02:01 torv has joined #openwrt-devel

02:03 torv has quit [Remote host closed the connection]

02:04 torv has joined #openwrt-devel

02:07 philipp64 has quit [Quit: philipp64]

02:07 guidosarducci has quit [Ping timeout: 480 seconds]

02:13 <Mangix> something tells me this is completely wrong

02:13 <Mangix> https://gist.github.com/neheb/e3876aee464312b2481b5be46c97d59c

02:14 torv has quit [Remote host closed the connection]

02:17 torv has joined #openwrt-devel

02:18 torv has quit [Remote host closed the connection]

02:22 torv has joined #openwrt-devel

02:24 torv has quit [Remote host closed the connection]

02:25 torv has joined #openwrt-devel

02:30 <soxrok2212_> anyone use the tp-link tl-sx1008?

02:31 <soxrok2212_> fan on this thing is mondo annoying

02:35 <Mangix> mondo?

02:35 <soxrok2212_> s/mondo/very

02:36 <Mangix> you running stock firmware?

02:36 <Mangix> box says low noise operation

02:39 <soxrok2212_> yeah its just a dumb switch

02:39 <soxrok2212_> very loud

03:01 dangole has quit [Ping timeout: 480 seconds]

03:21 <owrt-snap-builds> Build [#376](https://buildbot.openwrt.org/master/images/#builders/72/builds/376) of `imx/cortexa9` completed successfully.

03:40 philipp64 has joined #openwrt-devel

03:41 philipp64 has quit []

03:44 philipp64 has joined #openwrt-devel

03:45 philipp64 has quit []

03:47 philipp64 has joined #openwrt-devel

03:50 philipp64 has quit []

04:02 <Mangix> soxrok2212_: open it up and add a resistor on the fan

04:12 MaxSoniX has joined #openwrt-devel

04:24 felix has quit []

04:24 felix has joined #openwrt-devel

04:33 philipp64 has joined #openwrt-devel

04:43 philipp64 has quit [Quit: philipp64]

04:52 srslypascal has quit [Quit: Leaving]

04:54 philipp64 has joined #openwrt-devel

04:56 srslypascal has joined #openwrt-devel

05:08 philipp64 has quit [Quit: philipp64]

05:16 MaxSoniX has quit [Quit: Konversation terminated!]

05:17 embargo has joined #openwrt-devel

05:18 _embargo_ has quit [Read error: Connection reset by peer]

06:10 <KGB-1> https://tests.reproducible-builds.org/openwrt/openwrt_tegra.html has been updated. (100.0% images and 99.6% packages reproducible in our current test framework.)

06:33 <vulpes2[m]> robimarko: Yeah there was a networking regression for some people on the nanopi r2s, and I haven't been able to figure out the root cause yet. Vendor kernel works fine https://github.com/friendlyarm/kernel-rockchip/tree/nanopi-r2-v5.15.y but mainline can't seem to initialize the LAN interface (USB ethernet) no matter what I try.

06:34 <vulpes2[m]> I used builroot and tinkered with ATF and u-boot versions, no luck on that either.

06:35 <vulpes2[m]> At this point no openwrt releases will work on my board which is really strange. I've even tried multiple different MicroSD cards too just to be sure.

06:38 <vulpes2[m]> The funny thing is, I've got a nanopi neo3 as well which is pretty much identical to the r2s, and the same image works fine with an externally attached rtl8153 adapter on the same xhci host controller. (the rk3328 only has a single USB3 port as far as I can tell)

06:38 guidosarducci_ has quit [Remote host closed the connection]

06:40 guidosarducci has joined #openwrt-devel

06:44 <vulpes2[m]> A diff between the vendor kernel and matching mainline release revealed there are a couple more drivers that aren't present in mainline, notably a rockchip USB3 PHY driver.

06:46 rua has quit [Quit: Leaving.]

06:53 rua has joined #openwrt-devel

07:31 torv is now known as Guest11

07:31 torv has joined #openwrt-devel

07:33 Guest11 has quit [Ping timeout: 480 seconds]

07:52 cbeznea has joined #openwrt-devel

07:56 csrf has quit [Ping timeout: 480 seconds]

08:00 torv has quit [Ping timeout: 480 seconds]

08:02 torv has joined #openwrt-devel

08:14 torv has quit [Ping timeout: 480 seconds]

08:17 goliath has joined #openwrt-devel

08:22 torv has joined #openwrt-devel

08:27 snh has quit [Read error: Connection reset by peer]

08:30 Tapper has joined #openwrt-devel

08:31 torv has quit [Ping timeout: 480 seconds]

08:31 <Tapper> after flashing back to SNAPSHOT r20893 on my r7800 I have a uptime of 13h 35m 8s

08:32 <Tapper> So the bug making it reboot was introduced after that build.

08:40 snh has joined #openwrt-devel

09:00 danitool has joined #openwrt-devel

09:28 torv has joined #openwrt-devel

09:45 ahf has joined #openwrt-devel

09:50 torv has quit [Ping timeout: 480 seconds]

09:54 torv has joined #openwrt-devel

10:23 Tapper has quit [Ping timeout: 480 seconds]

10:28 mirko has joined #openwrt-devel

10:30 <mirko> is there a way for building a plain initramfs/initrd, without it being wrapped in a (legacy) uboot-image? I need it for embedding it into an FIT image along with a DTB and kernel image

10:44 robimarko has joined #openwrt-devel

10:56 <hauke> aiyion: where did I do this suggestion?

10:56 <aiyion> few days ago, one sec

10:57 <aiyion> https://github.com/openwrt/openwrt/issues/11104

11:02 torv has quit [Ping timeout: 480 seconds]

11:12 <mirko> is there a way for building a plain initramfs/initrd, *without* it being wrapped in a (legacy) uboot-image and/or appended to the actual kernel image?

11:15 torv has joined #openwrt-devel

11:21 bluew has quit [Ping timeout: 480 seconds]

11:28 guidosarducci has quit [Remote host closed the connection]

11:28 guidosarducci has joined #openwrt-devel

12:02 <stintel> robimarko: 4 day work week :)

12:10 dangole has joined #openwrt-devel

12:16 torv has quit [Remote host closed the connection]

12:16 torv has joined #openwrt-devel

12:17 <robimarko> stintel: That is nice

12:17 <robimarko> I took a day off today and tommorow is a public holiday so only 3 days to work

12:21 <stintel> took the day off today as well, it's a bank holiday in Ireland anyway

12:22 <stintel> so extra long weekend and only 3 days work also

12:23 <stintel> last day of the month though, so some administrative work to do

12:23 dedeckeh has joined #openwrt-devel

12:40 cbeznea has quit [Read error: Connection reset by peer]

12:42 cbeznea has joined #openwrt-devel

12:48 MaxSoniX has joined #openwrt-devel

12:55 zorun has quit [Ping timeout: 480 seconds]

13:03 <f00b4r0> stintel: you're in ireland? I thought you were in bulgaria? ;)

13:03 zorun has joined #openwrt-devel

13:04 <stintel> f00b4r0: I live in Bulgaria but I'm assigned to the Dublin office

13:04 <stintel> because our team lead is there I guess

13:13 srslypascal has quit [Ping timeout: 480 seconds]

13:20 valku has joined #openwrt-devel

13:22 valku has quit []

13:24 valku has joined #openwrt-devel

13:45 minimal has joined #openwrt-devel

13:52 nixuser has quit [Ping timeout: 480 seconds]

13:55 <Slimey> hiya stintel

13:56 <Slimey> have you tried to use those 3040 ath10k radios anywhere else besides they ap the came in? :P

14:03 robimarko has quit [Quit: Leaving]

14:05 valku has quit [Quit: valku]

14:17 Gaspare has joined #openwrt-devel

14:26 enyc has quit [Ping timeout: 480 seconds]

14:29 torv has quit [Remote host closed the connection]

14:31 MaxSoniX has quit [Quit: Konversation terminated!]

14:43 MAbeeTT2 has joined #openwrt-devel

14:45 valku has joined #openwrt-devel

14:46 dangole has quit [Ping timeout: 480 seconds]

14:46 guidosarducci has quit []

14:47 guidosarducci has joined #openwrt-devel

14:50 MAbeeTT has quit [Ping timeout: 480 seconds]

15:02 torv has joined #openwrt-devel

15:02 <vulpes2[m]> with mt7615, are those RF frontends typically controlled by the firmware, or are they hooked up to GPIOs on the main soc?

15:03 <vulpes2[m]> there are some skyworks chip that are probably LNAs

15:04 <f00b4r0> stintel: I see, so you travel a lot I guess :)

15:07 schwicht has joined #openwrt-devel

15:09 schwicht_ has quit [Read error: Connection reset by peer]

15:17 Salman has joined #openwrt-devel

15:17 <Salman> zet

15:18 Salman has quit [Remote host closed the connection]

15:24 <stintel> f00b4r0: not really, full-time remote

15:28 Salman has joined #openwrt-devel

15:30 Salman has quit [Remote host closed the connection]

15:34 Gaspare has quit [Quit: Gaspare]

15:47 isak has joined #openwrt-devel

16:03 Piraty_ has quit [Remote host closed the connection]

16:03 Piraty has joined #openwrt-devel

16:06 <vulpes2[m]> hauke: What is the preferred approach for the partition table setup when the stock firmware has separate kernel and rootfs partitions? Should I follow the factory setup or merge them together as one unified firmware partition?

16:13 <mrnuke_> svanheule: Yeah, upstream not making up their minds about how they want things modeled in dt is a pain in the. I remember the debate about spidev driver binding. As of yet, still unresolved

16:22 <svanheule> mrnuke_: thanks for testing the driver!

16:26 <svanheule> mrnuke_: I tried to model the offloading configuration on the software netdev trigger (rx, tx flags, link selection), so hopefully it can be useful to distill out some common interface over different switches

16:27 floof58 is now known as Guest39

16:27 floof58 has joined #openwrt-devel

16:31 Guest39 has quit [Ping timeout: 480 seconds]

16:38 <karlp> mrnuke_: rohm,dh2228fv is life.. what do you mean spidev is unresolved? :)

16:40 <mrnuke_> karlp: It's all over the place: https://github.com/torvalds/linux/blob/30a0b95b1335e12efef89dd78518ed3e4a71a763/drivers/spi/spidev.c#L711-L721

16:40 <mrnuke_> Do I have to go ahead and modify a kernel driver that otherwise works perfectly with _any_ userspace controlled SPI device just to add my device?

16:41 <karlp> I know, it's absoltuely insane, I have no comprehension of how this was the situation they thought made sense.

16:42 srslypascal has joined #openwrt-devel

16:42 <karlp> hence the "rohm,dh2228fv is life" just use one of the ones taht someone bothered, the kernel doesn't care anywhere...

16:42 <mrnuke_> 'echo spi0.0 > /sys/bus/spi/drivers/spidev/bind' does not work unless the SPI node has a compatible "compatible" string

16:43 <karlp> "hurhur, you should write a kernel driver for your device instead hur hur"

16:43 <mrnuke_> What happens when rohm,dh2228fv gets a proper kernel driver, then?

16:44 <karlp> like you said, it's still completely unresolved.

16:44 <karlp> the situation is bonkers.

16:44 <mrnuke_> Exactly!

16:48 <mrnuke_> karlp: Step I: get a 2228FV. Step II: write a proper IIO driver. Step III: Watch the world burn.

16:49 torv has quit [Remote host closed the connection]

16:50 dedeckeh has quit [Remote host closed the connection]

16:51 torv has joined #openwrt-devel

16:52 MatrixTravelerbot[m]1 has quit [Write error: connection closed]

16:52 will[m]1 has quit [Write error: connection closed]

16:52 hexagonwin[m] has quit [Write error: connection closed]

16:52 decke[m] has quit [Write error: connection closed]

16:52 gnustomp[m] has quit [Write error: connection closed]

16:52 nick[m]1234 has quit [Write error: connection closed]

16:52 lipnitsk has quit [Write error: connection closed]

16:52 fpsusername[m] has quit [Write error: connection closed]

16:52 bluse-blue[m] has quit [Write error: connection closed]

16:52 Movedtomkg20001mkg20001io[m] has quit [Write error: connection closed]

16:52 dfceaef[m] has quit [Write error: connection closed]

16:52 evils[m]1 has quit [Write error: connection closed]

16:52 whatevs111[m] has quit [Write error: connection closed]

16:52 domon has quit [Write error: connection closed]

16:52 MatMaul[m] has quit [Remote host closed the connection]

16:52 John[m]123456 has quit [Remote host closed the connection]

16:52 Jonny[m] has quit [Remote host closed the connection]

16:52 olmari has quit [Write error: connection closed]

16:52 pavlix has quit [Write error: connection closed]

16:52 Q__ has quit [Write error: connection closed]

16:52 ctdvqgg445[m] has quit [Write error: connection closed]

16:52 JuniorJPDJ has quit [Write error: connection closed]

16:52 t4h4[m] has quit [Write error: connection closed]

16:52 vulpes2[m] has quit [Write error: connection closed]

16:52 Kiste has quit [Write error: connection closed]

16:52 barhom has quit [Write error: connection closed]

16:52 fieryeagle954[m] has quit [Write error: connection closed]

16:52 znullptr[m] has quit [Write error: connection closed]

16:52 mkg20001 has quit [Write error: connection closed]

16:52 aparcar[m] has quit [Write error: connection closed]

16:52 oliv3r[m] has quit [Write error: connection closed]

16:52 tohojo has quit [Write error: connection closed]

16:52 schmars[m] has quit [Remote host closed the connection]

16:58 Gaspare has joined #openwrt-devel

17:03 aparcar[m] has joined #openwrt-devel

17:15 vulpes2[m] has joined #openwrt-devel

17:15 <vulpes2[m]> speaking of dt, is there any comprehensive documentation of downstream openwrt-specific properties?

17:16 <vulpes2[m]> iirc someone on lkml complained about downstream openwrt properties being included in a patch sent to the list

17:26 rsalvaterra has quit []

17:27 rsalvaterra has joined #openwrt-devel

17:29 srslypascal is now known as Guest58

17:29 srslypascal has joined #openwrt-devel

17:31 nixuser has joined #openwrt-devel

17:36 Guest58 has quit [Ping timeout: 480 seconds]

17:37 zorun_ has joined #openwrt-devel

17:39 zorun has quit [Ping timeout: 480 seconds]

17:57 Gaspare has quit [Quit: Gaspare]

18:01 Gaspare has joined #openwrt-devel

18:09 Borromini has joined #openwrt-devel

18:11 Tapper has joined #openwrt-devel

18:56 robimarko has joined #openwrt-devel

19:00 nixuser has quit [Ping timeout: 480 seconds]

19:18 Obi-Wan has quit [Remote host closed the connection]

19:29 Obi-Wan has joined #openwrt-devel

19:31 AtomiclyCursed has quit [Quit: ZNC 1.8.2 - https://znc.in]

19:32 AtomiclyCursed has joined #openwrt-devel

19:40 torv is now known as Guest88

19:40 torv has joined #openwrt-devel

19:46 <Mangix> Anyone know about Linux FIELD macros?

19:47 Guest88 has quit [Ping timeout: 480 seconds]

19:58 cbeznea has quit [Quit: Leaving.]

19:59 <nbd> Mangix: what do you want to know?

19:59 csrf has joined #openwrt-devel

20:01 valku has quit [Quit: valku]

20:19 <karlp> mrnuke_: (reject the iio driver as conflicts with real worl uses of the bonkers compatible property ;)

20:37 <Mangix> nbd: it was suggested that I use them here: https://gist.github.com/neheb/e3876aee464312b2481b5be46c97d59c#gistcomment-4353195 . No idea about any of this.

20:38 <robimarko> You should use them

20:38 <robimarko> As they allow getting rid of manual bit calculations

20:39 <robimarko> They are great for extracting or setting/clearing certain bits along with the BIT and GENMASK macros

20:40 <vulpes2[m]> I'm dealing with a stubborn device with no uart (u-boot had that disabled which is very annoying), and it keeps bootlooping on me

20:40 barhom has joined #openwrt-devel

20:40 bluse-blue[m] has joined #openwrt-devel

20:40 ctdvqgg445[m] has joined #openwrt-devel

20:40 decke[m] has joined #openwrt-devel

20:40 dfceaef[m] has joined #openwrt-devel

20:40 domon has joined #openwrt-devel

20:40 evils[m]1 has joined #openwrt-devel

20:40 fieryeagle954[m] has joined #openwrt-devel

20:40 fpsusername[m] has joined #openwrt-devel

20:40 gnustomp[m] has joined #openwrt-devel

20:40 hexagonwin[m] has joined #openwrt-devel

20:40 John[m]123456 has joined #openwrt-devel

20:40 Jonny[m] has joined #openwrt-devel

20:40 JuniorJPDJ has joined #openwrt-devel

20:40 Kiste has joined #openwrt-devel

20:40 Q__ has joined #openwrt-devel

20:40 lipnitsk has joined #openwrt-devel

20:40 MatMaul[m] has joined #openwrt-devel

20:40 Movedtomkg20001mkg20001io[m] has joined #openwrt-devel

20:40 mkg20001 has joined #openwrt-devel

20:40 nick[m]12 has joined #openwrt-devel

20:40 oliv3r[m] has joined #openwrt-devel

20:40 olmari has joined #openwrt-devel

20:40 pavlix has joined #openwrt-devel

20:40 schmars[m] has joined #openwrt-devel

20:40 t4h4[m] has joined #openwrt-devel

20:40 tohojo has joined #openwrt-devel

20:40 MatrixTravelerbot[m]1 has joined #openwrt-devel

20:40 whatevs111[m] has joined #openwrt-devel

20:40 will[m] has joined #openwrt-devel

20:40 znullptr[m] has joined #openwrt-devel

20:40 <nbd> Mangix: so let's say you have a register field: #define RANDOM_FIELD 0x2f0 - if you want to overwrite that field in an u32 val with something else, you can do: val &= ~RANDOM_FIELD; val |= FIELD_PREP(RANDOM_FIELD, new_val);

20:41 <nbd> if you want to extract a value, it's: field_val = FIELD_GET(RANDOM_FIELD, val);

20:41 <nbd> so it's really quite simple

20:41 <mrnuke_> karlp: sad part is, I wouldn't be surprised if they did exactly that

20:42 <vulpes2[m]> is there a way that I can tell the build system to use `lzma` instead of `xz` for the squashfs rootfs?

20:42 <vulpes2[m]> I know they are basically identical but the bootloader may disagree

20:52 <karlp> vulpes2[m]: the bootloader doesn't know anythingabout the rootfs though?

20:53 <vulpes2[m]> theoretically it shouldn't care

20:53 <vulpes2[m]> but maybe it does

20:54 <vulpes2[m]> without console logs I can't tell

20:54 <karlp> you might be running into the kernel decrompression needing more space than your bootloader allows, that's much more likely.

20:54 <karlp> normally fixed by the "lzmaloader" second level stuff

20:55 <karlp> but by the time the squashfs is involved, you're way past having linux console, not bootloader console.

20:55 <vulpes2[m]> my image was actually smaller when decompressed than the stock image

20:55 <vulpes2[m]> and the bootloader definitely has functional lzma support

20:56 <karlp> not all "lzma support" is equal in my experience :)

20:56 <karlp> as to your original question? not sure, at least in some targets, you need to edite the image makefiles to change the chains of commands, there's probably an "xz" line that you want to remove or change.

20:57 <vulpes2[m]> I ran binwalk on the stock kernel uimage and the one I got, and there was another difference

20:57 <vulpes2[m]> "properties" is 0x5D on stock and 0x6D from the image I buint

20:57 <vulpes2[m]> *built

21:05 aiyion has quit [Remote host closed the connection]

21:05 aiyion has joined #openwrt-devel

21:07 <robimarko> vulpes2[m]: can you pass earlycon to kernel bootargs?

21:08 <robimarko> That will give you the missing kernel log if it crashes before the serial adapter is initialized

21:23 robimarko has quit [Quit: Leaving]

21:28 Borromini has quit [Quit: leaving]

21:34 philipp64 has joined #openwrt-devel

22:17 danitool_ has joined #openwrt-devel

22:21 danitool has quit [Ping timeout: 480 seconds]

22:21 Gaspare has quit [Quit: Gaspare]

22:22 <KGB-2> https://tests.reproducible-builds.org/openwrt/openwrt_x86.html has been updated. (100.0% images and 99.6% packages reproducible in our current test framework.)

22:38 floof58 is now known as Guest106

22:38 floof58 has joined #openwrt-devel

22:42 Guest106 has quit [Ping timeout: 480 seconds]

22:44 philipp64 has quit [Quit: philipp64]

22:45 philipp64 has joined #openwrt-devel

22:47 GNUmoon has quit [Remote host closed the connection]

22:47 guidosarducci has quit [Read error: Connection reset by peer]

22:47 guidosarducci has joined #openwrt-devel

22:47 GNUmoon has joined #openwrt-devel

23:49 Tapper has quit [Quit: Tapper]