#
…

2021-03-03

00:06crabbedhaloablut has quit [Ping timeout: 265 seconds]
00:06crabbedhaloablut has joined #asahi
00:17hothotleg[m] has quit [Ping timeout: 240 seconds]
00:17Tokamak has joined #asahi
00:18foxlet has quit [Ping timeout: 244 seconds]
00:18hothotleg[m] has joined #asahi
00:19AONeiLL[m] has quit [Ping timeout: 240 seconds]
00:19Alice[m] has quit [Ping timeout: 240 seconds]
00:19Bennett[m] has quit [Ping timeout: 244 seconds]
00:19TellowKrinkle[m] has quit [Ping timeout: 244 seconds]
00:19daniel[m]2 has quit [Ping timeout: 244 seconds]
00:19peterkovar[m] has quit [Ping timeout: 258 seconds]
00:19Bastian[m] has quit [Ping timeout: 258 seconds]
00:20irradiated_lemmi has quit [Ping timeout: 244 seconds]
00:22AONeiLL[m] has joined #asahi
00:23TellowKrinkle[m] has joined #asahi
00:23irradiated_lemmi has joined #asahi
00:23daniel[m]2 has joined #asahi
00:24Bennett[m] has joined #asahi
00:24peterkovar[m] has joined #asahi
00:25foxlet has joined #asahi
00:26Bastian[m] has joined #asahi
00:27Alice[m] has joined #asahi
00:48aquijoule__ has joined #asahi
00:51aquijoule_ has quit [Ping timeout: 264 seconds]
00:53vimal has quit [Ping timeout: 240 seconds]
01:12vimal has joined #asahi
01:51leah2 has quit [Ping timeout: 260 seconds]
01:53leah2 has joined #asahi
02:17anarsoul has joined #asahi
02:29anarsoul has quit [Quit: ZNC 1.7.5 - https://znc.in]
02:29mbrennwa[m] has quit [*.net *.split]
02:29josiahmendes[m] has quit [*.net *.split]
02:29dpatterbee[m] has quit [*.net *.split]
02:30anarsoul has joined #asahi
02:33mbrennwa[m] has joined #asahi
02:33dpatterbee[m] has joined #asahi
02:33josiahmendes[m] has joined #asahi
03:04delroth has quit [Ping timeout: 264 seconds]
03:04manawyrm has quit [Ping timeout: 272 seconds]
03:05Swant has quit [Ping timeout: 606 seconds]
03:14Swant has joined #asahi
03:14manawyrm has joined #asahi
03:15delroth has joined #asahi
03:29nicolas17 has quit [Quit: Konversation terminated!]
03:53phiologe has joined #asahi
03:53PhilippvK has quit [Ping timeout: 260 seconds]
04:04jobbe has quit [Ping timeout: 246 seconds]
04:05aodaki[m] has joined #asahi
04:06balrog has quit [Ping timeout: 264 seconds]
04:17jobbe has joined #asahi
04:17marvin24_ has joined #asahi
04:17marvin24 has quit [Ping timeout: 260 seconds]
04:26balrog has joined #asahi
04:40aodaki[m] is now known as AkihikoOdaki[m]
04:59suskun has joined #asahi
05:10odmir has quit [Remote host closed the connection]
05:11Jasper[m] has quit [Ping timeout: 240 seconds]
05:11Jasper[m] has joined #asahi
05:39_whitelogger has joined #asahi
05:59suskun has quit [Quit: My MacBook Air has gone to sleep. ZZZzzz…]
06:04suskun has joined #asahi
06:18Emantor has quit [Remote host closed the connection]
06:19Emantor has joined #asahi
06:24Tokamak has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
06:29VinDuv has joined #asahi
06:34jeff_miw has joined #asahi
06:36Emantor has quit [Remote host closed the connection]
06:38thresh has quit [Ping timeout: 265 seconds]
06:41thresh has joined #asahi
06:43Tokamak has joined #asahi
06:45Emantor has joined #asahi
06:45jeff_miw has quit [Quit: Connection closed]
06:55kubes has quit [Quit: Leaving]
06:55suskun has quit [Quit: My MacBook Air has gone to sleep. ZZZzzz…]
07:35VinDuv has quit [Quit: Leaving.]
08:24Yuzu has quit [Ping timeout: 272 seconds]
08:24Yuzu has joined #asahi
08:24korzq4 has joined #asahi
08:24korzq4 has joined #asahi
08:24korzq4 has quit [Changing host]
08:24korzq has quit [Ping timeout: 240 seconds]
08:24korzq4 is now known as korzq
08:36raster has joined #asahi
09:00enverb[m] has quit [Quit: Idle for 30+ days]
09:00affieuk[m] has quit [Quit: Idle for 30+ days]
09:00flokk[m] has quit [Quit: Idle for 30+ days]
09:02Tokamak has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
09:07jordan_ has quit [Quit: WeeChat 3.0]
09:07jordan has joined #asahi
09:07jordan is now known as Guest63876
09:11ephe_meral has joined #asahi
09:11odmir has joined #asahi
09:15odmir has quit [Ping timeout: 245 seconds]
10:20macc24 has quit [Ping timeout: 272 seconds]
10:22macc24 has joined #asahi
10:33amw has quit [Ping timeout: 240 seconds]
10:36Guest63876 has quit [Quit: WeeChat 3.0]
10:36jordan has joined #asahi
10:37jordan is now known as Guest20635
10:48jaalsa has joined #asahi
11:05ephe_meral has quit [Ping timeout: 260 seconds]
11:10suskun has joined #asahi
11:15brandas_ has joined #asahi
11:15brandas has quit [Read error: Connection reset by peer]
11:18glytchbinary has joined #asahi
11:22klaus has joined #asahi
11:23vimal has quit [Quit: Leaving]
11:25klauson the task list here: https://github.com/AsahiLinux/docs/wiki/Tasks , at what point is this project currently working on ? is it storage driver & filesystem ?
11:28suskun has quit [Quit: My MacBook Air has gone to sleep. ZZZzzz…]
11:29dlnx_`` is now known as dlnx``
11:44vimal has joined #asahi
12:00marcanklaus: I'm working on getting the basic stuff checked there upstreamed first
12:00marcanmaz: current status: back at agonizing over memory ordering
12:01marcanI now have a theory that the atomic_lse.h implementation on arm64 is incorrect vs the atomics documentation
12:01marcanplease prove me wrong :D
12:02marcanatomic_t.txt says "RMW operations that have a return value are fully ordered"
12:02marcanthose ops are implemented using e.g. ldsetal
12:08marcanhowever, the *al instructions are defined as having load-acquire semantics for the load, and store-release semantics for the store, AFAICT
12:08marcanthat is weaker, isn't it?
12:09marcannothing stops the store-release operation from moving after subsequent operations, and vice versa for the load
12:09marcanwhich means those are not fully ordered as atomic_t.txt says
12:09marcanI tried writing a litmus test for this, but I think RMEM doesn't support the LSE atomic memory ops...
12:11mazmarcan: I'd be surprised, given that atomic.txt, the LSE atomics and the ARMv8 memory model have the same author...
12:12mazmarcan: but hey, bugs everywhere at the moment! :D
12:13mazmarcan: I'm pretty sure herd7 does support LSE atomics.
12:14marcanah yeah, that one does work
12:17marcanand herd7 says I'm right :-)
12:18marcantests positive; adding a dmb ish between the ldsetal and the ldr makes it negative
12:18marcanthe docs say
12:18marcanFully ordered primitives are ordered against everything prior and everything
12:18marcansubsequent. Therefore a fully ordered primitive is like having an smp_mb()
12:18marcanbefore and an smp_mb() after the primitive.
12:18marcanso this is clearly inconsistent: adding an smp_mb() is stronger
12:18klaus@marcan thanks for telling
12:27ifthenelse has quit [Quit: Ping timeout (120 seconds)]
12:27ifthenelse has joined #asahi
12:35marcanmaz: fire of an email to the list and CC Will I guess?
12:39flokk[m] has joined #asahi
12:42luca020400is this after deacon changed the default barriers for io reads or unrelated?
12:43marcanthis has nothing to do with io reads, it's purely an atomics thing
12:43luca020400missed this part, was thinking about memory model ops
12:44marcanbasically I convinced myself that the IO op ordering stuff was fine, only to now fall into another pit of ordering stuff related to atomics
12:44ephe_meral has joined #asahi
12:44luca020400well have fun.
12:45Guest20635 has quit [Quit: WeeChat 3.0]
12:47tomtastic has quit [Quit: ZNC - https://znc.in]
12:48tomtastic has joined #asahi
13:29suskun has joined #asahi
13:45suskun has quit [Quit: My MacBook Air has gone to sleep. ZZZzzz…]
14:00klausis there a way to follow on acceptation of this project's submitted elements to upstream ?
14:01j`eyklaus: easiest way is to check Linus' tree, or you can try check the linux-arm-kernel mailing list
14:02klausis there a way to follow on acceptation of this project's submitted elements to upstream ?
14:02klauswhat would be a convenient page to check on linux tree ?
14:04klausfor example, that would be a valid place : https://github.com/torvalds/linux/pulls ?
14:05j`eythere isn't such a place
14:06klaushow do one check out evolution of things happenings around linux kernel tree then ?
14:07j`eyby following the mailing lists
14:07mazmarcan: yup, interesting litmus test.
14:07jn__fetching the linux-next tree can also be useful
14:07j`eybut I guess there's sites like phoronix that do high level overviews too
14:09dhewggit fetch && git log --grep apple --grep marcan origin/master ?
14:09klausphoronix won't report often on asahi linux, probably at milestones but that would be like following evolution from afar, i'd like to get much more fresh news, often
14:09j`eyfollow marcan on twitter :P
14:10klausi don't use any social networks :)
14:10dhewgcheck the mailing lists then
14:10klauseven using them gives me allergies
14:10jn__i also suggest following the branches at https://github.com/AsahiLinux/linux
14:11klausjn__ that last link, yes of course i look at it often, but i was talking about inclusion into upstream
14:11dhewgthat even lets you poll via atom feed
14:11klausdhewg that looks more like it, many thanks for the link :)
14:15jn__klaus: ah right, sorry
14:16klausno problem, thanks for the help anyway
14:33marcanklaus: fwiw nothing is merged yet, but the first series should go into 5.13 if things don't drag on too much.
14:33j`eyI think some of the prelim FIQ stuff might go in 5.12?
14:33marcanIf only I didn't keep running into "this can't possibly be a bug, can it" bugs :-)
14:34marcanj`ey: I think some cleanup, but not any patches by me
14:34marcanbut yes, it qualifies as some stuff indirectly prompted by the M1 series
14:35klausmarcan yes that what you were saying on your patreon post i think, that is why i was interested in checking how weel or fast inclusion process was progressing
14:36j`eyindirectly prompted, directly helpful :P
14:36marcanwell yes :)
14:36marcanI'm going to start writing that progress update post for jan/feb too, that's long overdue
14:36marcanI wonder what level of technical detail I should go into...
14:51klausgeneral overview, not that much technical i'd say. more technical details could be given to contributors that are actively participating, committing or working with code.. basically devs that'd need more details. once in a while do one of those workflow video for the rest. that way you could focus more on coding, less about explaining ..
14:59korzq has quit [Quit: Ping timeout (120 seconds)]
14:59korzq has joined #asahi
14:59korzq has quit [Changing host]
14:59korzq has joined #asahi
15:28ephe_meral has quit [Ping timeout: 264 seconds]
15:32bisko has joined #asahi
15:33ephe_meral has joined #asahi
15:50dottedmagmarcan: It's cans of worms all the way down.
15:50dottedmagre: "can't possibly be a bug"
16:15mazj`ey: no much will go in 5.12, only small cleanups. the core FIQ stuff definitely is 5.13 material.
16:16j`eyoh ok, guess I misunderstood
16:16odmir has joined #asahi
16:17mazj`ey: I'm planning to take the first two patches of that series for -rc2 (because they are obvious fixes), but the rest needs to simmer in -next for a bit.
16:18maz(and it gives me more time to sort out the KVM changes!)
16:45marcanmaz: so my current plan is I'll send v3 tomorrow, using explicit barriers in the IPI code based on my current understanding of the architecture and that apparent atomic bug, and let's see how that goes
16:45mazmarcan: SGTM!
16:45marcangood thing I can tell myself this was all worth it to have a better understanding of the memory model for everything that's to come :-)
16:47mazmarcan: definitely. it is a maddening part of the architecture... someone even said that it is an adversarial architecture, and I can't really say otherwise.
16:48marcanI swear though, the ARM ARM... I can't help but think surely there *must* be a better way of describing all this
16:49mazmarcan: the architecture is shifting to a "rule system" to describe all that, but I'm not sure it is any better.
16:49marcanlike for my latest question I'm pondering about atomics ordering vs MMIO... I'm basically jumping around 8 places in the doc and trying to piece together a coherent story, involving things like "oh, the word 'coherent' here probably means this only applies to the shareability domain it is mapped as... and then I guess there are no guarantees against other domains"
16:49marcanbut I'm nowhere near certain I'm interpreting things right
16:50marcanthings get very sticky once you involve shareability domains and device memory; AIUI the litmus tests don't cover any of that and just assume IS
16:51marcanbut yeah, I think this in combination with the atomics issue means it's probably time to stop here, throw a few barriers in (reasonable ones, still weaker and better thought out than I was using before) and if someone wants to suggest a better way of doing it and prove it with docs-lawyering they can do so :)
16:52mazyup. now imagine trying to get both HW and SW people to agree on the meaning of these things, with different interpretations of the English text. Fun. Not.
16:52BaughnLogBot has quit [Ping timeout: 245 seconds]
16:53marcanfigures :)
16:53marcanI will say Will's talks on this subject are great
16:53mazhe's great fun to watch.
16:54BaughnLogBot has joined #asahi
17:01delroth has quit [*.net *.split]
17:01br377 has quit [*.net *.split]
17:01inglor has quit [*.net *.split]
17:01KindOne has quit [*.net *.split]
17:01baryluk has quit [*.net *.split]
17:01bdju has quit [*.net *.split]
17:01m42uko has quit [*.net *.split]
17:01opticron has quit [*.net *.split]
17:01xarkes has quit [*.net *.split]
17:01nspaced has quit [*.net *.split]
17:01Jangal has quit [*.net *.split]
17:01jaXvi has quit [*.net *.split]
17:03Tokamak has joined #asahi
17:06delroth has joined #asahi
17:06bdju has joined #asahi
17:06br377 has joined #asahi
17:06baryluk has joined #asahi
17:06xarkes has joined #asahi
17:06opticron has joined #asahi
17:06m42uko has joined #asahi
17:06nspaced has joined #asahi
17:06jaXvi has joined #asahi
17:06inglor has joined #asahi
17:06Jangal has joined #asahi
17:08KindOne has joined #asahi
17:24VinDuv has joined #asahi
17:30hthh has quit [Remote host closed the connection]
17:44ephe_meral has quit [Ping timeout: 256 seconds]
17:48choozy has joined #asahi
17:57dibas has quit [Read error: Connection reset by peer]
17:58dibas has joined #asahi
17:58dibas has joined #asahi
17:58dibas has quit [Changing host]
18:13Bublik has joined #asahi
18:23marcanmaz: ok, so I have two options, pick your poison.
18:23marcan1) atomic_fetch_or_release(); atomic_read_acquire();
18:24marcanthat one relies on guarantees that ARM makes beyond Linux atomic ops (there is ordering between release-acquire pairs in the ARM ARM)
18:24marcan2) just atomic_or(); smp_mb__after_atomic(); atomic_read();
18:24marcanthat one is obviously arch-independent
18:25marcannote that atomic_fetch_or_acquire() does not cut it, as I need to order the *write* side of that, and _acquire applies to the *read* side.
18:26mazI'm vaguely partial to (1), but that's only me. I bet that performance wise, they are strictly equivalent.
18:27marcanI was leaning towards (1) too
18:27marcanatomic_fetch_or(); atomic_read() is what would've worked if I hadn't run into that issue :)
18:40odmir_ has joined #asahi
18:42odmir has quit [Ping timeout: 245 seconds]
19:19klaus has quit [Quit: Textual IRC Client: www.textualapp.com]
19:19aquijoule__ has quit [Ping timeout: 256 seconds]
20:24richbridger has joined #asahi
20:31raster has quit [Quit: Gettin' stinky!]
20:45nicolas17 has joined #asahi
21:16Raqbit has quit [Quit: Ping timeout (120 seconds)]
21:16Raqbit has joined #asahi
21:22CDFH has quit [Ping timeout: 276 seconds]
21:22CDFH_ has joined #asahi
21:28Calchan has joined #asahi
21:33bgianf has quit [*.net *.split]
21:33Thomas__ has quit [*.net *.split]
21:33segher has quit [*.net *.split]
21:33Mary_ has quit [*.net *.split]
21:33theracermaster has quit [*.net *.split]
21:33minicom has quit [*.net *.split]
21:33massoud has quit [*.net *.split]
21:33dgb has quit [*.net *.split]
21:33j`ey has quit [*.net *.split]
21:33HotSwap has quit [*.net *.split]
21:33xMopx has quit [*.net *.split]
21:33maximus64 has quit [*.net *.split]
21:33MoxZ_ has quit [*.net *.split]
21:33kakobrekla has quit [*.net *.split]
21:34segher has joined #asahi
21:34xMopx has joined #asahi
21:34MoxZ_ has joined #asahi
21:34massoud has joined #asahi
21:34maximus64 has joined #asahi
21:34theracermaster has joined #asahi
21:34Mary_ has joined #asahi
21:34HotSwap has joined #asahi
21:34dgb has joined #asahi
21:34bgianf has joined #asahi
21:34kakobrekla has joined #asahi
21:34j`ey has joined #asahi
21:34Thomas__ has joined #asahi
21:40Tokamak has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
22:06ketteniswhat? the M1 only supports 8-bit ASIDs?
22:06Tokamak has joined #asahi
22:08VinDuv has quit [Quit: Leaving.]
22:16choozy has quit [Remote host closed the connection]
22:33ketteniscertainly explains why I was seeing crashes with OpenBSD as our kernel assumed 16-bit ASIDs were available
22:56modmuss50 has left #asahi [#asahi]
23:40jn__ has quit [Remote host closed the connection]
23:41jn__ has joined #asahi
…