04:51Venemo: Ristovski: no, I don't think DC is used at all when you have no outputs connected. that said it would be still nice to know that this setup works correctly.
04:51Venemo: johnny0: thanks
22:15gnarface: is this the right place to ask about fan control issues for the amdgpu driver?
22:16gnarface: i'm using the current debian stable kernel and a recent update seems to have stopped my fans from kicking in (RX 5500)
22:16gnarface: i know it's not a hardware issue because if i poke at /sys/... i can turn them back on again
22:17gnarface: didn't want to change kernels, so i was hoping someone could help me figure out what exactly broke it so i could just revert the patch and rebuild
22:19gnarface: oh, and i know it's not the normal "zero RPM" thing, they used to kick on well before 100C, i'm sure of that
23:02Frogging101: gnarface: Looks like the last update was on May 22, does that sound right?
23:02Frogging101: https://metadata.ftp-master.debian.org/changelogs//main/l/linux-signed-amd64/linux-signed-amd64_6.1.140+1_changelog
23:02Frogging101: linked from https://packages.debian.org/stable/kernel/linux-image-amd64
23:03gnarface: yes, seems about right, though i'm not 100% sure
23:04gnarface: i went from 6.1.129-1 to 6.1.140-1 and i think it was still working as of 6.1.129-1
23:05gnarface: i also had 6.1.124-1, 6.1.112-1, 6.1.94-1, and 6.1.85-1 and i was pretty sure it was working throughout all of those too
23:05Frogging101: I see some suspend related commits in the changelog I linked
23:06gnarface: i have not been using suspend at all
23:07gnarface: i know for sure it used to work at some point during these 6.1.x kernel revisions, then just noticed after the most recent one the fans weren't actually spinning, and sensors was reading the speed as "65535"
23:07gnarface: so something appears to have changed in what it gets set to, but it gets set to something invalid
23:08Frogging101: hmm
23:08gnarface: it still goes between that and 0 though, when it cools back down on its own
23:08gnarface: and i can still poke /sys/class/drm/card0/device/hwmon/hwmon0/pwm1_enable and /sys/class/drm/card0/device/hwmon/hwmon0/pwm1 to control them manually, and that seems to still work as described
23:09gnarface: i wonder if i should just use fancontrol and pwmconfig anyway?
23:09gnarface: i wasn't super happy about how it would wait until 60C to turn the fan on by default to begin with
23:10Frogging101: Perhaps
23:10Frogging101: Not sure what changed though. Not that I'm an expert, I'm just reading the changelog and trying to look for anything that sounds related.
23:11Venemo: 6.1 is super old
23:11gnarface: eh, the card is old too
23:11gnarface: aside from this, it's been stable
23:12Frogging101: Yeah, it's the LTS kernel that debian stable is using
23:12Venemo: debian stable is a lie. it isn't stable
23:13gnarface: so, it's been the same firmware-amd-graphics package version throughout though, so it seems like the change had to be made in the kernel source, but it was news to me that this could even break from a kernel update like that, i thought these fans were... BIOS controlled by default
23:13Venemo: can you boot to the old kernel and verify that it works correctly there?
23:13Frogging101: <Venemo> debian stable is a lie. it isn't stable
23:14Frogging101: The version numbers are stable :)
23:14Venemo: debian isn't shipping stable graphics drivers, so it isn't stable. it is that simple.
23:14gnarface: i could in theory boot the 6.1.129-1 one (linux-image-6.1.0-32-amd64)
23:14gnarface: or any of the others, but i was leery about going back to pre-xfs fixes (i'm using xfs)
23:15gnarface: i tried to downgrade to linux-image-6.1.0-36-amd64 but someone seems to have deleted the corresponding headers package from debian's repo... (sigh)
23:16gnarface: Venemo: aside from this one fan issue, it's been very stable for me. it's only a old RX 5500, do you think there's really anything in a newer driver version that could possibly improve anything for such an old device?
23:16Frogging101: Yes
23:16Frogging101: Drivers improve all the time for all devices.
23:17gnarface: so you think i should just install the backports kernel?
23:17gnarface: (and corresponding firmware, and mesa...)
23:17Frogging101: Yeah, I would, if it works.
23:17gnarface: well, it just seemed like a bigger risk to stability to me, so i had not yet
23:18gnarface: and like i said, it didn't seem like the firmware blob had anything to do with this fan issue
23:18gnarface: but, if that's the advice you're giving, i'll take it into consideration...
23:19Venemo: gnarface: both the kernel and mesa get fixes and new features all the time. so if you lock yourself to a version that is 3 years old, you are basically rejecting all our work we did to improve it in the past 3 years
23:20gnarface: Venemo: i hear where you're coming from, i'll take that into account as i decide what to do. it's a 6 year old card though, so i just didn't think it would make any difference because of that.
23:21gnarface: if you're saying you know otherwise, i'll take your word for it
23:21Frogging101: A lot of stuff in the drivers isn't device specific, and device specific stuff gets updated too.
23:21gnarface: Venemo: (as a former life-long nvidia customer, all i would expect from a newer driver version for a 6-year old card would be a bunch of sandbags)