#
…

2021-06-16

00:11Gues_____ has joined #asahi
00:13Gues_____ is now known as r0ni
00:42tomtastic has joined #asahi
00:44tomtastic_ has quit [Ping timeout: 480 seconds]
00:49vup2 has joined #asahi
00:49vup2 has quit []
01:07vup2 has joined #asahi
01:07vup2 has quit []
01:09klltkr has quit [Ping timeout: 480 seconds]
03:00phiologe has joined #asahi
03:04PhilippvK has quit [Ping timeout: 480 seconds]
03:13skipwich has quit [Ping timeout: 480 seconds]
03:25marvin24 has joined #asahi
03:28marvin24_ has quit [Ping timeout: 480 seconds]
03:58Izumoo has quit [Quit: Konversation terminated!]
04:44JTL1 has joined #asahi
04:45JTL1 has quit []
04:49quarkyalice_ has joined #asahi
04:54quarkyalice has quit [Ping timeout: 480 seconds]
05:27VinDuv has joined #asahi
05:30quarkyalice_ has quit [Quit: Leaving]
05:30quarkyalice has joined #asahi
05:51arekm has quit [Quit: leaving]
05:51arekm has joined #asahi
06:25VinDuv has quit [Quit: Leaving.]
07:40mini has quit [Quit: ZNC closing...]
07:42mini has joined #asahi
08:11aleasto has joined #asahi
08:20linuxgemini has quit [Quit: Ping timeout (120 seconds)]
08:20linuxgemini has joined #asahi
08:39bgb_ has joined #asahi
08:56bisko has joined #asahi
09:00bgb_ has quit [Ping timeout: 480 seconds]
09:10bgb_ has joined #asahi
09:27BlitzWorks has joined #asahi
09:35bgb_ has quit [Ping timeout: 480 seconds]
09:40robinp_ has joined #asahi
09:46bgb_ has joined #asahi
09:48robinp has quit [Ping timeout: 480 seconds]
09:49marcanalways fun to bisect a GPU driver bug down to a memory coherency change
09:49* marcanthrows it over to the mesa folks to look at
09:49marcanat least my menus aren't flashing any more
09:55marcanat least the repro is trivial
09:55marcanbut it was a layer 2 bisect
09:55j_eylayer 2?
09:56marcanbisecting to a commit that exposes the bug but isn't the bug, then having to bisect again adding that
10:01bisko has quit [Read error: Connection reset by peer]
10:11bgb_ has quit [Ping timeout: 480 seconds]
10:15arndj_ey: that looks like a scary bug on the i.MX, I hope it's an SoC specific issue rather than a fundamental issue in the way we think of the barriers
11:01EmantorI wouldn't even be remotely surprised if this is an i.MX8Q* issue…
11:09bisko has joined #asahi
11:13jkc_ has joined #asahi
11:30Raqbit has quit [Quit: The Lounge - https://thelounge.chat]
11:31Raqbit has joined #asahi
11:33bisko has quit [Ping timeout: 480 seconds]
11:37bisko has joined #asahi
11:40pastly-antispam has joined #asahi
11:41choozy has joined #asahi
11:42jkc has quit [Quit: WeeChat 3.0.1]
11:59radex2 has joined #asahi
12:00radex1 has quit [Ping timeout: 480 seconds]
12:15choozy has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
12:36jato has joined #asahi
13:24impromptu-cauterize has joined #asahi
13:25impromptu-cauterize has left #asahi [#asahi]
13:38Izumoo has joined #asahi
13:40bgb_ has joined #asahi
13:49bgb_ has quit [Ping timeout: 480 seconds]
14:27bgb_ has joined #asahi
15:12skipwich has joined #asahi
15:12bisko has quit [Read error: Connection reset by peer]
15:13bisko has joined #asahi
15:14yuyichao has quit [Ping timeout: 480 seconds]
15:27skipwich has quit [Quit: DISCONNECT]
15:27skipwich has joined #asahi
16:08bisko has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
16:09bisko has joined #asahi
16:09jkc_ is now known as jkc
16:11bisko has quit []
16:34VinDuv has joined #asahi
16:47aleasto has quit [Remote host closed the connection]
16:47aleasto has joined #asahi
17:04marcanit's great when you work for a few hours on a bunch of code and the end result is everything works exactly the way it did before and no new features have been added :P
17:04marcan(...yet)
17:05hell__I think this is called "refactoring"
17:05marcanasync MMIO tracing now uses the stacked tracer concept I mentioned above (sync/hooks are still TODO)
17:06quarkyaliceyeah I've been there
17:06marcanone thing this does do is fix the "clobbering the vuart" problem
17:06quarkyalicerefactoring can be quite useful for cleaning up and creating more readable code
17:06marcanpagetable updates are now automatically done based on whatever tracers you configure, and it keeps track of dirty sections and updates as necessary
17:06marcanand also flushes the TLBs
17:07marcanquarkyalice: or in this case preparing for new features later (and fixing problems)
17:07quarkyalicethere's been times when I've changed one line of code and broken an entire project by creating a cascade of problems that need fixing :P
17:08marcanI mean I basically changed the entire way MMIO tracing works
17:08marcan... it's just that there is a default tracer now that just so happens to implement the same exact behavior the old code had
17:08marcan(and which is configured the same way)
17:10hell__I... once caused build failures by replacing license headers. (some uses of "dead code" macros that used the line number to declare undefined functions ended up having the same line number in the same compilation unit)
17:10* hell__likes to use reproducible builds to verify the bulk of changes when refactoring stuff, especially when testing is impractical (e.g. missing the necessary hardware, or way too many variables to account for with testing)
17:13j_eymarcan: is it re-doing the whole page tables in pt_update?
17:13marcanj_ey: no, only the parts that changed
17:13marcanthat's what that dirty_maps thing is for
17:13marcanand if it's empty (which is a cheap check) it does nothing
17:14marcanso now I automatically call pt_update after every exception before returning to the guest
17:14j_eyoh right, i see. and if it's tracemode.OFF... it just maps_hw
17:14marcanso now the base MMIO map is actually a "HW" OFF tracer that underlies the entire address range
17:15marcan(originally I wasn't going to do that but it turns out that's the easiest way, code-wise, instead of trying to fill in holes in the RangeMap with map_hw)
17:21j_eymarcan: just to clarify, think im having a bit of a brainfart, hv_handle_dabort is only meant to be for SPTEs right?
17:22marcanyes; that path triggers when you forget to invalidate the TLBs :-)
17:22marcanotherwise it should never happen since the CPU should have taken care of it
17:23marcanwell, and also for actual missing maps (which means the guest accessed nonexistent hardware or you forgot to map a range)
17:23j_eyI was trying to convince myself that hw_map() when TraceMade.OFF doesnt call into C/python
17:23marcanhw_map(), modulo unaligned sub-page maps, creates real page table mappings
17:23marcanthat are therefore handled by the MMU as stage 2 translation
17:25marcan(I did make the python side fall back to sw_map for weird alignment or sub-page chunks at the start and end, because it makes the rest of the code easier to write instead of having to deal with the page-alignment requirement of the underlying hv_map() in hw mode)
17:25marcan(so in those unaligned cases it will go through C)
17:26marcanbasically the HW alignment is 16K and the SW alignment is 4b (due to the extra fake pagetable level I maintain, which only works for SPTEs)
17:27marcannote that this isn't massively different from what I was doing earlier; the commit replaces a hw_map of the entire MMIO range with setting up the HW tracer
17:27marcanthe main difference is before we were stacking and mutating page tables (map everything hw, replace some bits with sw maps)
17:28marcanwhile now since it gets coalesced and updated rangewise, it'll hw_map a bit, sw_map a bit, hw_map a bit, etc in address order, without overlaps
17:28marcan(which in this case does ~double the number of proxy calls but whatever)
17:28marcanj_ey: oh that will go through C
17:28marcanbut that's the point
17:28marcanthat's asking for a mmiotrace of everything
17:29marcanso that'll map_sw everything with async tracing turned on
17:29marcanso again that's no different from the map_sw it replaces :-)
17:30marcanyes, that one
17:31marcanwith the trace_all script you end up with this page table: https://mrcn.st/p/suiSn7HJ
17:32marcanwhich is PrintTracer everything except the exceptions in that script, and the PMU hack, and VUART
17:35marcanso next up will be building the scaffolding for easily writing higher level tracers with RegMaps, and finishing the sync/hook support (which is broken/not finished right now)
17:36j_eyI appreciate the tests for regmaps :)
17:45marcanthere's no way to make that kind of code work properly without tests :)
17:45marcan(and I still ran into a broken untested codepath at least once while doing the hv thing :p)
17:45marcan(was just a typo in one of the wrappers though, not the core code)
18:57robinp has joined #asahi
19:01Misthios has quit [Quit: Misthios]
19:01Misthios has joined #asahi
19:02robinp_ has quit [Ping timeout: 480 seconds]
19:11klltkr has joined #asahi
20:05nsklaus has quit [Quit: WeeChat 3.1]
20:17VinDuv has quit [Quit: Leaving.]
20:25nsklaus has joined #asahi
20:47skali has joined #asahi
20:59skali is now known as skali_
20:59skali_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
21:10klltkr has quit [Quit: Textual IRC Client: www.textualapp.com]
21:57angustrau[m] has joined #asahi
23:23aleasto has quit [Quit: Konversation terminated!]
…