02:39:27 using a mlx5en card for the first time with freebsd. it's negotiating a 1G connection instead of 10G. is there something i'd need to tweak to make it negotiate the higher speed by default? 02:47:35 i noticed if i set my switch to 10G instead of autonegotiate it goes to 10G. i guess that's fine, but still curious why it'd go for the lower rate. is that indicative of a problem? 07:04:59 https://bsd.to/miCk 07:05:00 Title: dpaste/miCk (Plain Text) 07:05:08 not sure whats goin on here 07:05:42 similar thing to ungoogled-chromium works fine... is there some alias to egrep buried somewhere? not my doing I think 07:07:04 concrete_houses: the link was removed some time ago 07:08:06 at least from GNU grep. pretty sure our grep still has it 07:08:17 you have 'g' defined as an alias or shell function 07:08:40 what shell are you using? 07:31:22 hi, my server (freebsd 13.2) logged 6 times "kernel: ." in /var/log/messages and then booted. never seen this before, anybody knows what this message means? 07:37:15 shdw: booted? or rebooted? 07:37:56 the . is probably spurious. 07:38:09 was any message shown when it rebooted? 07:40:55 these were the messages: https://pastebin.com/YKdNi1T5 07:40:56 Title: Jul 14 10:09:58 XXXXX kernel: .Jul 14 10:09:58 XXXXX syslogd: last message rep - Pastebin.com 07:44:51 something triggered a clean shutdown 07:46:49 strange 07:48:35 does last | head show a recent shutdown time? 07:49:02 (or just a boot time?) 07:51:00 uptime shows boot time 07:51:12 ist shows a shutdown time 07:52:16 something executed a shutdown command then, I think 07:52:44 its a vm, perhaps someone accidantly pressed the shutdown button 07:53:31 didn't thought of this, thanks 07:55:58 checking, something caused init to run rc.shutdown 07:56:55 if the hypervisor signalled an ACPI shutdown event (i.e. simulating pushing the power button) that would explain it 08:06:15 yeah, if there's no "shutdown by ... at ..." line in the logs, the most likely cause is shutdown being triggered by the kernel in response to an ACPI signal 08:06:53 ACPI power button event causes the kernel to do shutdown_nice without needing userland intervention 08:07:24 shutdown_nice in turn signals init to run rc.shutdown 08:08:13 there's a flag to avoid signaling init, but i don't know of anyone who uses it 09:14:11 AllanJude: might be working by coincidence, as readonly check uses vdev_prop_table[-1].pd_attr, so whatever is there in the memory before vdev_prop_table[] array... https://github.com/openzfs/zfs/pull/15062 09:14:13 Title: Don't do a readonly check for vdev user properties by yuripv · Pull Request #15062 · openzfs/zfs · GitHub 10:40:29 fascinating, a classic array-out-of-bounds even in these days of clever tooling 15:29:44 Hi, anyone with a ThinkPad x270?, I'm struggling with audio 15:33:53 cat /dev/sndstat shows 3 devices, pcm0: (play/rec), pcm1: (rec) default and pcm2: (play) 15:34:21 hw.snd.default_unit: 1 15:35:08 when default_unit is 1 I can record with mic, but cannot play audio. The other way around I can play audio but cannot record... 15:37:00 Can you pick different /dev/dsp* in software? 15:39:33 VVD: the software is Firefox 15:39:38 it might be possible to rewire the associations to put the devices you want to use together on the same dsp* 15:40:03 RhodiumToad: yes, I think so, but I don't know how... 15:40:06 what's the sysctl dev.hdaa output? 15:40:49 oh, hold on 15:41:04 What was the command to generate a pastebin from console? 15:41:06 RhodiumToad: ok 15:41:20 you can play audio with which unit? 0? 15:42:03 sysctl dev.hdaa dev.hdac | nc termbin.com 9999 15:42:17 RhodiumToad: yes, with 0 I can play audio 15:43:02 if pcm0 and pcm1 are on the same hdac chip, it's probably fixable 15:43:22 (pcm2 is obviously a separate chip, but that's just the digital outputs to HDMI) 15:43:26 https://termbin.com/lies 15:43:34 lies 15:44:27 yuripv: sorry 15:44:29 https://termbin.com/nkd2 15:45:57 oh, i just liked the url it gave you 15:46:37 martinrame: and sysctl dev.pcm and cat /dev/sndstat please 15:50:38 also, have you put anything in loader.conf or sysctl.conf about hdaa? 15:53:03 ok, I think i can see how to fix it 16:02:44 * RhodiumToad waits... 16:16:16 meh, don't ask the question if you're not going to stick around for the answer 16:24:10 RhodiumToad: sorry, I'm here 16:25:10 sysctl dev.pcm --> https://termbin.com/3936 16:25:49 cat /dev/sndstat -->https://termbin.com/hobc 16:27:22 In /boot/loader.conf I don't see anything related to audio 16:27:56 In sysctl.conf I added hw.snd.default_unit=0 16:30:30 if you do sysctl dev.hdac.0.pindump=1 it should dump some stuff to the console, which you can see with dmesg 16:30:36 can you paste that too 16:30:40 let me see 16:31:16 dev.hdac.0.pindump: 0 -> 0 16:31:49 Or do you ask for the output of dmesg? 16:33:23 the output of dmesg, yes - at least the part with hdaa: 16:34:23 https://termbin.com/6gitm 16:37:32 ok, try this: sysctl dev.hdaa.0.nid18_config="as=4" dev.hdaa.0.nid24_config="as=3" 16:37:41 then sysctl dev.hdaa.0.reconfig=1 16:38:40 then sysctl hw.snd.default_unit=0 16:38:50 then see if playback and mic work 16:39:33 if it does, you can put hint.hdaa.0.nid18_config="as=4" and hint.hdaa.0.nid24_config="as=3" in /boot/loader.conf 16:40:08 RhodiumToad: playback works, but recording not. 16:40:16 hm. 16:40:34 does recording still work with hw.snd.default_unit=1 ? 16:40:49 and can you do the pindump / dmesg thing again 16:41:28 after service sndiod restart recording works with hw.snd.default_unit=1 16:41:58 https://termbin.com/ll6s 16:42:57 oh sorry, I think I got them backwards 16:43:12 sysctl dev.hdaa.0.nid18_config="as=3" dev.hdaa.0.nid24_config="as=4" 16:43:17 sysctl dev.hdaa.0.reconfig=1 16:43:30 then try again with default_unit=0 16:43:40 YEEEESSSS!!!! 16:44:09 ok. so put hint.hdaa.0.nid18_config="as=3" and hint.hdaa.0.nid24_config="as=4" in /boot/loader.conf 16:44:17 and then it should still work after a reboot 16:45:34 RhodiumToad: nothing else in /etc/sysctl.conf or /etc/rc.conf? 16:45:55 shouldn't need to 16:46:20 Ok, this is the last /boot/loader.conf https://termbin.com/mkf2 16:46:58 those need to be two separate lines at the end there 16:47:15 ok 16:47:28 and they need to be hint.* not dev.* 16:48:10 https://termbin.com/n3k6 16:48:30 that's better 16:48:49 Great, let's reboot. 16:48:57 the idea here is that the as= parameters tell the driver how to wire up the various audio components 16:52:26 RhodiumToad: I'm back. Sadly dind't work after rebot 16:52:28 reboot 16:53:12 oh jeez, whoever wrote these options should be shot 16:53:33 in loader.conf, change the nidNN_config to nidNN.config 16:54:14 RhodiumToad: replace the _ with a . ? 16:54:28 (the sysctls have _ there, the hints have . instead) 16:54:39 RhodiumToad: ok, changed, let's reboot again. 16:54:40 yes, replace the _ with . 16:57:00 RhodiumToad: yes!, now it works. Thank you very much! 16:57:33 excellent 16:58:04 what we basically did is swap the order in which the mic jack and the internal mic are assigned 16:58:30 (I assume there's a mic jack somewhere, the hardware thinks there is one) 16:58:59 previously it had the mic jack on pcm0 and the mic on pcm1, now it has them the other way 17:01:01 RhodiumToad: Mmm, a couple of weeks ago I connected some Bluetooth headphones 17:01:34 But I don't remember if before that playing/recording worked as expected. 17:01:53 This isn't my main machine. 17:03:33 RhodiumToad: do you have the same laptop? 17:04:12 nope 17:05:09 Hopefully someone here has one, to see if it works without tweaking anything. 17:07:34 bluetooth needs virtual_oss, as I recall, so it'll be a separate pcm device 22:17:52 freebsd repos are building under 13.2 now? 22:18:18 oh nm.. misread something 22:18:42 that's the current version 22:18:59 13.1 is 22:19:18 13.1 is the oldest supported version, which is what packages are built on. 22:30:03 huh, an ircII user. 22:42:51 * kenrap looks up what the ircII client is 22:51:43 I love how lurking here sometimes gives me exposure to past-time Unix and technology