00:24
parazyd has joined #openwrt-devel
00:45
minimal has quit [Quit: Leaving]
00:59
Daanct12 has joined #openwrt-devel
01:01
thejoker8814 is now known as Guest1225
01:01
thejoker8814 has joined #openwrt-devel
01:08
Guest1225 has quit [Ping timeout: 480 seconds]
01:16
skynet2 has quit [Quit: Leaving]
01:27
Znevna has quit [Ping timeout: 480 seconds]
01:29
Znevna has joined #openwrt-devel
01:53
danitool has quit [Ping timeout: 480 seconds]
02:00
tSYS has quit [Quit: *squeak*]
02:00
tSYS has joined #openwrt-devel
02:25
Mangix has quit [Read error: Connection reset by peer]
03:52
Daanct12 has quit [Remote host closed the connection]
04:14
Daanct12 has joined #openwrt-devel
05:23
Mangix has joined #openwrt-devel
05:41
rua has quit [Remote host closed the connection]
06:23
goliath has joined #openwrt-devel
06:25
rua has joined #openwrt-devel
06:38
Daanct12 has quit [Quit: WeeChat 4.2.2]
06:39
Daanct12 has joined #openwrt-devel
06:39
Daanct12 has quit []
06:41
Daanct12 has joined #openwrt-devel
06:51
Daaanct12 has joined #openwrt-devel
06:52
Daanct12 has quit [Ping timeout: 480 seconds]
07:51
robimarko has joined #openwrt-devel
08:15
<
robimarko >
But both of those are binutils gas assembler errors because they are using Ubuntu 18.04 with GCC13.2 but with the original binutils 2.30 that lack these instructions.
08:15
<
robimarko >
I am getting report for libdeflate and mold compilation breaking from 2 users.
08:16
<
robimarko >
So the question is, whether we should maybe start building binutils for host as well or just tell them to update their binutils
08:21
<
stintel >
Ubuntu 18.04 with gcc 13.2 sounds like a weird combination?
08:21
<
stintel >
or did you mean OpenWrt's gcc 13.2?
08:22
<
robimarko >
Sadly no, they replaced the Ubuntu provided GCC with GCC13
08:22
<
robimarko >
So its hacked together with binutils 2.30 which dont support a lot of things
08:22
<
stintel >
I'm leaning towards "if they can replace gcc they can replace binutils"
08:23
<
robimarko >
Same here, its these X-WRT devs/users that for whatever reason refuse to update 18.04
08:23
<
robimarko >
And now they are spamming random commits with compile errors
08:23
<
dwfreed >
sucks to be them
08:24
<
Znevna >
lucky this isn't x-wrt support? :P
08:24
<
dwfreed >
(I really wish Ubuntu hadn't extended LTS duration)
08:24
<
stintel >
they can always build in a container
08:25
<
robimarko >
Znevna: They do some stuff in OpenWrt as well
08:25
<
robimarko >
But yeah, I am heavily in favor of update your distro, binutils or use containers
08:28
<
Znevna >
yeah nobody is gonna merge that
08:48
noltari has joined #openwrt-devel
09:41
Daaanct12 has quit [Ping timeout: 480 seconds]
10:18
Daanct12 has joined #openwrt-devel
10:53
mark22k has joined #openwrt-devel
11:04
<
Ansuel >
robimarko ping ?
11:32
<
Ansuel >
nice the rmmod crash is fixed :D
11:33
<
Ansuel >
also i love how they totally forgot to add some logic to unregister the switchdev notifier...
11:33
<
Ansuel >
JUST MAGIC.
11:43
<
robimarko >
Ansuel: that is great news
11:48
<
Ansuel >
i'm having fun with edma_reset_hw printing warning now :D
11:48
<
Ansuel >
(having fun with insmod back in)
11:49
<
robimarko >
Hehe, we seem to have a weird way of fun by trying to fix badly broken things
12:28
fakuivan has quit [Remote host closed the connection]
12:28
fakuivan has joined #openwrt-devel
12:36
<
Ansuel >
robimarko [ 38.517523] (null): PHY already attached
12:36
<
Ansuel >
[ 38.521246] nss-dp 3a001000.dp1 lan1 (uninitialized): failed to connect to phy device
12:36
<
Ansuel >
[ 38.565457] nss-dp: probe of 3a001000.dp1 failed with error -14
12:36
<
Ansuel >
YHEEEE -.-
12:36
<
Ansuel >
this thing really works BY LUCK.
12:36
<
robimarko >
Just as intended :)
12:49
<
Ansuel >
robimarko... i'm impressed... with the correct logic and API rmmod and insmod correctly works
12:49
<
Ansuel >
and interface correctly works after a restart
12:50
<
robimarko >
Now that is surprising
12:51
<
Ansuel >
you will be surprised by the patch...
13:00
<
robimarko >
Is it large?
13:18
Daanct12 has quit [Quit: WeeChat 4.2.2]
13:53
rmilecki has joined #openwrt-devel
13:53
rmilecki has quit []
13:54
rmilecki has joined #openwrt-devel
13:54
rmilecki has quit []
13:54
rmilecki has joined #openwrt-devel
13:54
rmilecki has quit []
13:55
rmilecki has joined #openwrt-devel
14:01
mattsm has joined #openwrt-devel
14:10
mattsm has quit [Ping timeout: 480 seconds]
14:41
goliath has quit [Quit: SIGSEGV]
14:41
<
Ansuel >
robimarko man i really dream of rewriting the edma driver from scratch... it's not even that hard...
14:41
minimal has joined #openwrt-devel
14:45
<
robimarko >
Ansuel: Well, its not that large when we remove the unsued parts
14:47
<
Ansuel >
github.com/openwrt/openwrt/pull/15175
14:48
<
Ansuel >
lovely change :D
15:11
mcbridematt has quit [Read error: Connection reset by peer]
15:12
mcbridematt has joined #openwrt-devel
16:15
valku has joined #openwrt-devel
16:48
fakuivan has quit [Remote host closed the connection]
16:49
fakuivan has joined #openwrt-devel
18:27
cmonroe has joined #openwrt-devel
18:48
torv has quit [Remote host closed the connection]
18:48
torv has joined #openwrt-devel
19:10
parazyd has quit [Read error: No route to host]
19:15
parazyd has joined #openwrt-devel
19:24
<
robimarko >
I think its fine
19:31
valku has quit [Ping timeout: 480 seconds]
19:52
<
robimarko >
Yeah, most likely some ancient stock FW had those limits
20:36
<
mrnuke >
robimarko: I left a comment on the eap610-outdoor BDF PR. Do you have a preference which of the two methods should be used?
20:39
<
Ansuel >
robimarko btw any chance how can i test for the remoteproc restart?
20:40
<
Ansuel >
i tried with rmmod ath11k_ahb insmod ath11k_ahb
20:57
goliath has joined #openwrt-devel
20:58
<
robimarko >
Ansuel: After you remove ath11k_ahb then you can request the remoteproc state via sysfs
20:58
<
robimarko >
In /sys/class/remoteproc/remoteproc0#
20:59
<
Ansuel >
what entry ?
20:59
<
robimarko >
It should be state
20:59
<
robimarko >
You can echo to it as well
21:00
<
robimarko >
You can use start or stop to control the desired state
21:01
<
robimarko >
BTW, Ansuel any chance you can do your magic to update the regulatory info in BDF-s for ipq807x and ipq60xx?
21:01
<
robimarko >
We can also try removing the regulatory info and shipping regdb.bin from the latest QCA BDF-s
21:02
<
robimarko >
That should make it easier to update all at the same time
21:02
<
Ansuel >
do you think external regdb works by wiping the regdb entry in the bdf?
21:02
<
Ansuel >
from what i notice wcn bdf still have regdb entry inside
21:04
<
robimarko >
Gotta try it but it should since there will be no regulatory info in the BDF that is priority
21:04
<
Ansuel >
robimarko i can echo start and stop but still with insmod the thing starts to work correctly
21:04
<
robimarko >
It doesnt crash?
21:04
<
robimarko >
Well, then it fixed itself automagically
21:04
<
Ansuel >
maybe i should test to simulate a crash by ath11k
21:05
<
robimarko >
Or try removing the ath11k patch to disable coldboot
21:06
<
Ansuel >
where the hell are the ath11k proc
21:06
<
robimarko >
In mac80211 ath11k patches
21:07
<
robimarko >
Think its 902 for ipq807x
21:07
<
Mangix >
robimarko: lol at that ubuntu thing
21:07
<
Mangix >
to be fair, I use it for CI
21:08
<
robimarko >
Mangix: 18.04 with updated GCC and Python but not binutils?
21:08
<
Mangix >
although GitHub Actions has 20.04 as a minimum
21:08
<
Ansuel >
ath11k c000000.wifi: pdev 2 successfully recovered
21:09
<
robimarko >
Mangix: it was kind of miracle that nothing broke so far with 18.04
21:10
<
Mangix >
those compiler minimums are because of linker errors IIRC
21:10
<
robimarko >
Mangix: and libdeflate and mold builds?
21:10
<
Mangix >
no idea about those, or 18.04
21:10
<
Mangix >
mainline OpenWrt does not support 18.04 AFAIK
21:11
<
robimarko >
Yes, I have stated that multiple times
21:11
<
Ansuel >
thing is why they use 18.04...
21:11
<
robimarko >
Thing is that if you use GCC 12.1 or newer libedeflate will compile with AVX-VNNI instructions that require binutils 2.36
21:12
<
robimarko >
And mold also uses some instructions that require 2.33.1 I think
21:12
<
Mangix >
Ansuel: you'd be surprised how many legacy systems there are out there
21:12
<
robimarko >
They can just compile newer binutils
21:12
<
Ansuel >
also robimarko can you check the pr for nss-dp ?
21:12
<
robimarko >
I tried that in Ubuntu 18.04 docker and it works
21:13
<
robimarko >
Had to compile newer Python first though
21:13
* Mangix
is reminded of my suggestion for an XP user to switch to ReactOS
21:13
<
robimarko >
Ansuel: I gave it a quick look, nice catches
21:13
<
Ansuel >
Mangix i know GPL tar still suggest meme 18.04 or even 16.something but most of the time is just stupid suggestion...
21:13
<
robimarko >
I am just not sure that about the last patch
21:13
<
robimarko >
Dont see a point in keeping that call at all
21:13
<
Ansuel >
i'm trying to remove the patch that disable coldboot
21:15
<
Ansuel >
OK WOW it works?????
21:15
<
Ansuel >
i bet kernel 6.6 magically fixed it???
21:16
<
robimarko >
Wow, it coldboot really works on Qnap as well
21:16
<
robimarko >
Hehe, 6.6 FTW
21:18
<
robimarko >
Anyway, off to bed, another day in U-Boot land took its toll
22:01
<
mrnuke >
svanheule: Booted up EWS2910P-v3 here from initramfs -- no issues thusfar. The PoE chip is a RTL8238B, and there is what appears to be an MCU before the optoisolators
22:03
gladiac1 has quit []
23:22
Obi-Wan has quit [Remote host closed the connection]
23:32
torv has quit [Quit: torv]
23:35
torv has joined #openwrt-devel
23:56
gladiac has joined #openwrt-devel