2021-07-10
00:08quarkyalice has quit [Ping timeout: 480 seconds]
00:50quarkyalice has joined #asahi-gpu
00:58quarkyalice has quit [Ping timeout: 483 seconds]
01:20Emantor has quit [Quit: ZNC - http://znc.in]
01:20Emantor has joined #asahi-gpu
01:32quarkyalice has joined #asahi-gpu
01:40quarkyalice has quit [Ping timeout: 480 seconds]
01:54yuyichao has joined #asahi-gpu
02:01chadmed has joined #asahi-gpu
02:23PhilippvK has joined #asahi-gpu
02:26phiologe has quit [Ping timeout: 480 seconds]
02:45chadmed has quit [Remote host closed the connection]
02:52chadmed has joined #asahi-gpu
03:00al3xtjames has quit [Quit: The Lounge - https://thelounge.chat]
03:03al3xtjames has joined #asahi-gpu
03:51chadmed has quit [Ping timeout: 480 seconds]
04:27nafod has quit [Read error: Connection reset by peer]
04:27nafod has joined #asahi-gpu
04:42quarkyalice has joined #asahi-gpu
04:43quarkyalice_ has quit [Quit: Leaving]
04:47quarkyalice_ has joined #asahi-gpu
04:47quarkyalice has quit []
04:59al3xtjames has quit [Quit: The Lounge - https://thelounge.chat]
05:00al3xtjames has joined #asahi-gpu
05:23al3xtjames has quit [Quit: The Lounge - https://thelounge.chat]
05:24al3xtjames has joined #asahi-gpu
05:28al3xtjames has quit []
05:30chadmed has joined #asahi-gpu
05:33al3xtjames has joined #asahi-gpu
05:33al3xtjames has quit []
05:36al3xtjames has joined #asahi-gpu
05:51chadmed has quit [Remote host closed the connection]
05:57chadmed has joined #asahi-gpu
06:36chadmed has quit [Ping timeout: 480 seconds]
06:59chadmed has joined #asahi-gpu
07:39aleasto has joined #asahi-gpu
11:14choozy has joined #asahi-gpu
13:32alyssa has joined #asahi-gpu
13:32alyssa has left #asahi-gpu [#asahi-gpu]
13:37bloom has joined #asahi-gpu
13:50jixwould probably watch if you stream... might be too distracted to actually follow what's going on today though
13:53chadmednow im conflicted as to whether to fall asleep to such a stream or stay up and learn something
13:56choozy has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
14:40jixI've heard that even recent-ish discrete desktop GPUs use tiling (need to dig out the blog post which had some code to actually demonstrate that)
15:58jixi wonder, are the handles at the end of the header and at the end of each entry, or are they actually at the start of each entry? (might be obvious with more context, but isn't to me)
16:33choozy has joined #asahi-gpu
16:35phiologe has joined #asahi-gpu
16:39PhilippvK has quit [Ping timeout: 480 seconds]
16:55jixI'm a bit confused by having DIRTY_ANY defined to be a specific bit, or do you set it whenever you set another bit? (might have missed that)
17:17xerpi[m]Regarding this random hang/HW issue: are the CPU and GPU cache coherent? If so, only LLC or also L1/L2? Maybe you need to flush some shared stuff/descriptors up to LLC
17:17jixis the windowserver log msg just forwarding errors from kernel, or is it having an additional error when this happens?
17:22jix(I was asking because I was wondering whether the windowserver error has anything to do with how you get stuff on screen in the end (not knowing how you do that/how that works on osx), but just everything starting to fault makes sense, especially after seeing OBS hang before)
17:27xerpi[m]Let's hope this was not the issue and that the GPU also snoops the L1/L2, otherwise it will get hairy, you would also need to invalidate BOs when reading data the GPU has written..
17:29xerpi[m]Would be nice if the GPU reported some kind of error code. Maybe macOS kernel doesn't expose it to userspace, but would mmio tracing help?
17:32jixis there any apple code talking to the GPU that would need to be synchronized with what your mesa driver is doing? if it's not some cache issue, it seems like a race condition to me, just from the random timing and how it is affected by debug printing etc..
17:34quarkyalice has joined #asahi-gpu
17:38xerpi[m]Btw this is interesting: https://developer.apple.com/documentation/metal/mtlstoragemode/managed
17:41xerpi[m]Hmm but maybe you don't need to have CPU and GPU caches coherent for some buffers, so this "managed" mode can help reduce traffic
17:41jixsounds to me like they might have forgotten to update that this to say that it also includes macos on apple silicon
17:42svenfwiw, so far all dma transactions i've seen between the cpu and any peripheral have been cache coherent
18:04quarkyalice__ has joined #asahi-gpu
18:05quarkyalice has quit [Remote host closed the connection]
18:17choozy has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
19:23quarkyalice__ has quit [Ping timeout: 480 seconds]
20:12jixme neither, but I think it should allow you to do selective system wide tracing of stuff... e.g. the IOKit calls and scheduling/preemption of threads/processes
20:14jixat least that's what I remember from when it was introduced and I was still using mac os, never did more than trying a few example dtrace scripts though... and that was ~14 years ago, so I have no idea how well it is supported today
20:22jixthere's also the instruments GUI app that's based on dtrace, and looking at https://web.archive.org/web/20200620075030/https://developer.apple.com/documentation/metal/using_metal_system_trace_in_instruments_to_profile_your_app that seems useful, if you can add rules to also trigger on the calls the mesa driver makes
21:19quarkyalice_ has quit [Quit: Leaving]
21:20quarkyalice has joined #asahi-gpu
22:50aleasto has quit [Quit: Konversation terminated!]