13:41 cdo256: I'm trying to understand the nouveau source in an attempt to understand why the drm master can't seem to talk to the hdmi/tmds chip.
13:41 cdo256: (the same issue I mentioned a couple of weeks back)
13:42 cdo256: Tangentally to this (for my own understanding), I'm trying to find out why nouveau runs into this mmio fault:
13:43 cdo256: bus: MMIO write of 8000002f FAULT at 10eb14 [ PRIVRING ]
13:44 cdo256: it appears to come from gm107_fan_pwm_set. Is it because of an incorrect assumption about the memory layout?
15:02 imirkin: cdo256: are you on a GM108?
15:02 imirkin: cdo256: basically it's saying that this mmio address doesn't exist
15:03 cdo256: GM204, I put a trace in that function but it seems to be a red herring as it's not showing in the log
15:04 imirkin: cdo256: oh, we can't control the fan on GM200+
15:04 imirkin: that's all hidden away now
15:04 imirkin: anyways, my comment doesn't really change though
15:04 imirkin: it's saying that mmio address can't accept that write.
15:06 cdo256: imirkin: by 'hidden away' do you mean the nvkm_trace isn't written to the log?
15:06 imirkin: by 'hidden away' i mean it requires "high-secure" firmware to affect fan settings
15:07 cdo256: ah right got it
15:08 ccr:plays MC Hammer - U Can't Touch This
15:08 imirkin: pretty much.
15:08 cdo256: should it be replaced by a stub/warning for GM200+ so the it's not spitting out an obscure mmio error
15:08 imirkin: is it?
15:09 cdo256: is it spitting out an mmio error? yes. is it obscure?, that's debatable
15:10 imirkin: ok, well i can't look at it right now
15:11 cdo256: that's fine. thank you for your help, you saved me a lot of time
15:11 imirkin: the fan stuff should not be hooked up for setting fans on gm200+
15:12 imirkin: maybe we can't even retrieve the settings? i forget
15:16 cdo256: gm200_therm has just init, fini, temp_get (from g84) and program_alarms so it shouldn't be hooked up
15:16 cdo256: I'll do some more poking around
15:33 cdo256: I think I used the term 'spitting out' wrong, causing confusion. I meant prints the error to the log once each boot, rather than lots of times.
15:35 imirkin: yea
15:36 imirkin: i mean, probably something in the init which triggers the write which the gpu doesn't like