00:04 <mirko> jow: do you have a feasible idea of being able to name switch ports in a human readable way (via LuCI)? Given my freshly supported 24-port running OpenWrt, LAN1-LAN24 really isn't that helpful..

04:52 <KGB-2> https://tests.reproducible-builds.org/openwrt/openwrt_mpc85xx.html has been updated. (100.0% images and 99.9% packages reproducible in our current test framework.)

05:31 <russell--> mirko: how is that not helpful?

06:21 <neggles> are the ports not marked 1-24? :P

06:22 <mirko> russell--: at the risk of you trolling: i can name firewall rules - giving it a descriptive name. e.g. which device - let's say, a phone's name - is allowed (not) to access which service. semantics. i would like not to look and trace cables every time to remember which device is connected to which switch port, but have semantically descriptive names right where i'm configuring the VLANs of each port.

06:32 <neggles> hm, do we not have "port description" as a thing

06:42 <KanjiMonster> mirko: so you want to allow the user (in this case you ;) to rename or apply labels to interfaces (or annotate them by other means)?

06:44 <KanjiMonster> so it would then display e.g. "LAN3 (teapot)", where 'teapot' is a user description set somwhere

08:52 <jow> mirko: no immediate idea

10:01 <danitool> Hi Ansuel, I'm compiling the current snapshot with your proposed patch for fixing bmips RAC:

10:01 <danitool> https://github.com/openwrt/openwrt/pull/15354/commits/02343a79d2d8e11197a194d1c1325f42d43b150a

10:02 <danitool> however there are two issues with it, first: the documentation patch doesn't work, i just deleted it, not needed for testing

10:03 <danitool> second issue, a compilation problem

10:03 <danitool> arch/mips/bmips/setup.c: In function 'device_tree_init':

10:03 <danitool> arch/mips/bmips/setup.c:527:52: error: comparison between pointer and integer [-Werror]

10:03 <danitool> 527 | bmips_cbr_addr && addr != bmips_cbr_addr) {

10:44 <KGB-2> https://tests.reproducible-builds.org/openwrt/openwrt_bcm47xx.html has been updated. (100.0% images and 100.0% packages reproducible in our current test framework.)

11:11 <mirko> KanjiMonster: correct

11:12 <mirko> KanjiMonster: which teapot model with ethernet jack do you have?

11:20 <wigyori> internet of teapots

11:20 <hell__> but does it return HTTP 418 when appropriate? :D

11:21 <KanjiMonster> that's what I was thinking of with that example ;D

11:34 <Ansuel> danitool can you disable werror just for testing?

11:38 <danitool> Ansuel: sure, KERNEL_WERROR = n, this build option right?

11:38 <Ansuel> yes

11:38 <Ansuel> or from menu config

11:38 <Ansuel> uncheck compile kernel with warning as error

11:50 <danitool> I unchecked that option still the same error

11:50 <danitool> maybe has no effect, when compiling it still says: cc1: all warnings being treated as errors

11:56 <Ansuel> danitool did you clean the kernel ? strange tho

12:47 <danitool> Ansuel: yes, I did. I never used this option so no idea what's going on

12:51 <Ansuel> danitool pushed a new version that should fix the warning

12:51 <danitool> ok

13:21 <danitool> Ansuel: https://bpa.st/LURQ

13:21 <Ansuel> mhhh seems to be working?

13:22 <Ansuel> where is that warning tho

13:22 <Ansuel> oh ok... register CBR ff400000 differ from DT CBR a0000000. Ignoring DT CBR.

13:22 <danitool> yes, it works

13:22 <Ansuel> you tweaked the CBR in DT?

13:23 <danitool> no, I didnt, I applied your patch as is

13:25 <Ansuel> actually i swapped the 2 thing interesting...

13:26 <Ansuel> bootloader is setting a0000000

13:26 <Ansuel> and dt is ff400000

13:27 <Ansuel> we also need the mips-broken-cbr-reg

13:27 <Ansuel> wth the bootloader set the cbr reg to that addr ??? o.O

13:29 <Ansuel> danitool i pushed another version can you try ?

13:29 <danitool> sure

13:30 <Ansuel> also i assume an iperf should be enough to test if the thing works correctly?

13:40 <danitool> Ansuel: https://bpa.st/BJ2A

13:41 <danitool> no apparent issues

13:41 <Ansuel> i need one last test tho

13:41 <Ansuel> can you help me?

13:41 <danitool> sure

13:46 <Ansuel> new version pushed... you should be spammed with CALL SYNC DMA...

13:46 <Ansuel> if everything is working correctly we should have 0 and not 1

13:53 <danitool> Ansuel: I can't compile again: https://bpa.st/VSKA

13:54 <Ansuel> my bad sorry for wasting you time... should be fixed now

14:04 <danitool> Ansuel: It enters into an infinte error loop: https://bpa.st/P5DQ

14:04 <Ansuel> it's expected

14:05 <Ansuel> CALL SYNC DMA 0 ok nice the thing is working and it's not disabled

14:06 <Ansuel> Ok I pushed the final version of it (with the hacky debug log removed)

14:06 <danitool> ok, thank you

14:06 <Ansuel> if you can I would ask you to keep that around and check if the perf are the same or if you notice crash.... but in theory we are doing the same than the previous patch just more upstream friendly

14:07 <Ansuel> and leave some comments in the pr so we can continue it there

14:08 <danitool> ok, I'll also test it in the dual core BCM6368

14:11 <Ansuel> the dual core one i'm more interested in ehehhe... still didn't find anywhere if cbr change based on the core executed

14:11 <Ansuel> but the current workaround is to hardcode it so.... i guess it doesn't ?

15:16 <jow> so if I am going to bump a package now I have to dirclean & make to get the newest zstd stuff?

15:16 <jow> otherwise the mirror chekcsums will be wrong?

15:18 <Ansuel> yes correct... still have to investigate how to correctly make that

15:18 <Ansuel> make tools/zstd/compile is enough

15:19 <jow> alright, thanks

17:22 <csharper2005> WPA3 Enterprise downgrades to a less secure WPA2 after the upgrade from 22.03 to 23.5 https://github.com/openwrt/openwrt/issues/15347

17:22 <csharper2005> Would be nice to get WPA3-Enterprise top secret security standard again, out-a-box :)

17:24 <csharper2005> My knowledge, alas, allows me to do this only with the dirty hack hostapd.sh lol