#
…

2021-11-03

00:00X-Scale has quit [Ping timeout: 480 seconds]
00:33robinp has quit [Ping timeout: 480 seconds]
00:34robinp[m] has quit [Ping timeout: 480 seconds]
00:35rohin2 has quit [Ping timeout: 480 seconds]
02:04X-Scale has joined #asahi-dev
02:08X-Scale` has quit [Ping timeout: 480 seconds]
02:20yuyichao_ has quit [Ping timeout: 480 seconds]
02:43yuyichao_ has joined #asahi-dev
03:38kov has quit [Quit: Coyote finally caught me]
03:47aratuk has joined #asahi-dev
03:59phiologe has joined #asahi-dev
04:03PhilippvK has quit [Ping timeout: 480 seconds]
04:18aratuk has quit [Quit: Leaving...]
04:27useableporthole has joined #asahi-dev
05:05chadmed has quit [Quit: Konversation terminated!]
05:14jmr2 has joined #asahi-dev
05:15skoobasteeve has quit [Ping timeout: 480 seconds]
05:15jmr2 has quit []
05:17skoobasteeve has joined #asahi-dev
05:21jmr2 has joined #asahi-dev
05:21jmr2 has quit []
06:02chadmed has joined #asahi-dev
07:44marcanI guess I need to review that one :p
07:48svenmaybe point him to alyssa's version which is already cleaned up a bit
07:49dhewgmac via cmdline is prolly not acceptable anyway
07:49dhewgvia module arg I mean
07:51svenyeah, alyssa’s versions takes it from the device tree
07:56j_eywhy is that v3, I see no other versions
07:58svenProbably the first time they submitted something to the ML. Takes a while to get used to it and eventually accept that the only sane way is git send-mail
07:59j_eyoh.. 3 patches in the series, v3, maybe some confusion there
08:28egavinc has joined #asahi-dev
08:29marcanalso threw some shade at stan :p
08:31j_eypoor stan
08:34j_eymarcan: 286 firmware files... is that for all the different macs?
08:46egavinc has quit [Quit: Leaving]
08:49j_eyso much less gunk in alyssas series for brcfmac
08:52egavinc has joined #asahi-dev
08:54bps2 has joined #asahi-dev
09:01gabuscus has quit [Ping timeout: 480 seconds]
09:04dhewgwhere's that series?
09:06dhewgah thx, yet another alias, can't have enough of those!
09:13chadmedwow that brcmfmac patch really isnt very good
09:17j_eymarcan: are you planning on pushing your linux branch from the stream?
09:18DarkShadow44 has quit [Quit: ZNC - https://znc.in]
09:18marcanj_ey: yes, once I commit it properly
09:18marcan286 files is for all possible board/antenna/vendor/etc combinations
09:18marcanbut a lot are copies/symlinks
09:18DarkShadow44 has joined #asahi-dev
09:19j_eybut does it contain all the old intel mac firmware too?
09:26chadmedohhh its directly ripped from corellium's fork that explains it lmao
09:35mazmarcan: do max/pro have a similar cluster layout as m1, where MPIDR_EL1.Aff2 indicates whether the CPU is E or P?
09:36mazI end-up having do know that for the PMU...
09:36maz(stupid big-little strikes again)
10:03ihalip has joined #asahi-dev
10:24marcanmaz: yes
10:25marcanMPIDRs are 00000[01] 000001 01010[0-3] 0102[0-3]
10:26marcaner, minus the second one
10:27marcan(plus the next commit in t6000/bringup-work for the DART fix)
10:35svenoh... ouch. so it just took the paddr straight from the PTE when walking the table
10:36svener. wait. the other way around. it put the paddr straight to the PTE when installing a table
10:36svenand then when walking the table later everything exploded
10:36mazmarcan: looking good! if you want to split the fast IPI from the t6k bring up, that could trickle into upstream quickly (5.17).
10:37marcanyeah, I'll try to break it up into the compatible/fastipi handling first, and then add t6k / groups
10:37marcansven: yeah
10:37marcanit's clearly a bug
10:37marcanhow did that ever work with 52-bit addrs?
10:38svenyeah, probably makes sense to send that separately from the t6000 dart stuff
10:39sveni think "iommu/io-pgtable:" is the iommu style commit prefix and make sure to cc robin murphy if get_maintainers.pl doesn't include him automatically
10:40marcanmaz: I was wondering about the division in the mask/unmask functions; alternatively I could make hwirq sparse and use bitfields, but at that point I really need to move to irq_domain_create_tree
10:40marcanhowever, with ~3000 hwirqs at this point even without holes (but with 2 groups), maybe that makes sense already?
10:40marcansven: thanks, and yeah, the commits are BS right now :-)
10:40marcanwhole thing needs to be redone and reworded
10:41marcan(also comments are wrong etc)
10:41svenyeah, just making sure you don't need a v2 just because the commit prefix is slightly off :-)
10:41j_eysven: now you're the eager one :P
10:41svenj_ey: true :D
10:42mazmarcan: +1 for a tree. it would even make sense today, where we only use a handful of interrupts, and yet allocate a fixed-size domain.
10:43marcantree it is then, and then I'll just split the hwirq numbers into 16/16 bits and avoid the division
10:47j_eymarcan: how many interrupts did AICv1 have?
10:48marcanup to 1024 I think? 700-odd in practice IIRC
10:49j_eyok, so quite a big jump
10:49marcanwell, AICv2 has up to 4096 per block
10:50marcanbut the current implementation has 1806 implemented, times 2 blocks
10:50marcanonly one of which is in use right now
10:50marcanI strongly suspect the second block is where the second die's IRQs will show up on the rumored 2-die systems
10:50marcanthough currently macOS ignores it completely (doesn't even set the masks)
10:51marcanit does work though, I can sw_gen IRQs in the second block and unmask them and they get delivered
10:51marcanso I implemented support for it, maybe I'll get lucky and the 2-die systems will need no AIC patches ;)
10:57aleasto has joined #asahi-dev
10:58kettenisone can hope...
11:00thunfisch is now known as Guest4804
11:00thunfisch has joined #asahi-dev
11:00Guest4804 has quit [Read error: Permission denied]
11:04kettenismarcan: any progress on the pmgr bindings?
11:09marcankettenis: was kinda hoping for more review comments, since other than the spurious |s in the DT bindings I think everything else is kosher?
11:10marcanKrzysztof didn't comment on the v2 of the serial runtime-PM patch yet
11:11riker77 has quit [Ping timeout: 480 seconds]
11:11kettenisthe discussion about the compatibles confused me a bit, but do I understand correctly that you don't plan to remove the apple,pmgr one from the parent node?
11:12kettenisif the number of child nodes gets large, I may want to use that in OpenBSD to avoid cluttering dmesg
11:12kettenisattach a single driver to the parent node instead of separate drivers to the child nodes
11:13kettenis(we print a line for every driver that attaches)
11:18marcanyeah, apple,pmgr can stay
11:22riker77 has joined #asahi-dev
11:26aleasto has quit [Quit: Konversation terminated!]
11:27aleasto has joined #asahi-dev
12:53ihalip has quit [Remote host closed the connection]
13:06kettenis has quit [Read error: Connection reset by peer]
13:06kettenis has joined #asahi-dev
13:23kov has joined #asahi-dev
13:41aleasto has quit [Remote host closed the connection]
13:42aleasto has joined #asahi-dev
13:42aleasto has quit [Remote host closed the connection]
13:45aleasto has joined #asahi-dev
13:58yuyichao_ has quit [Ping timeout: 480 seconds]
14:09aleasto has quit [Remote host closed the connection]
14:09aleasto has joined #asahi-dev
14:09aleasto has quit []
14:10yuyichao_ has joined #asahi-dev
14:10aleasto has joined #asahi-dev
14:11aleasto has quit []
14:11aleasto has joined #asahi-dev
14:15aleasto has quit [Remote host closed the connection]
14:16aleasto has joined #asahi-dev
14:16aleasto has quit [Remote host closed the connection]
14:16aleasto has joined #asahi-dev
14:17aleasto has quit [Remote host closed the connection]
14:25aleasto has joined #asahi-dev
14:34mazhopefully, this will give someone an incentive to work out the semantics of the counters...
15:17aleasto has quit [Remote host closed the connection]
15:26aleasto has joined #asahi-dev
15:44conradev has quit [Quit: -]
15:46conradev has joined #asahi-dev
15:50bps2 has quit [Ping timeout: 480 seconds]
16:01kovsven, does that mean 2 M1 chips for the next mini? so 16 cpu and gpu cores?
16:02svenYour guess is as good as mine
16:04kovI'm placing my bet on them doing an A15-based M2 only for the mini to keep that higher end for the iMac Pro
16:19bps2 has joined #asahi-dev
16:19jannaumaz: xnu source code has/had counter defines in osfmk/arm64/kpc.c http://www.jannau.net/kpc.c
16:19jannauinstructions are probably 'CPMU_INST_A64 0x7a'
16:23xaiki has quit [Read error: Connection reset by peer]
16:23jannaujust the defines assuming m1 is not the early cpu PMU: https://paste.debian.net/1218089/
16:35aleasto has quit [Quit: Konversation terminated!]
16:36aleasto has joined #asahi-dev
16:38mazjannau: 7a never seems to count anything here.
16:40mazjannau: on the other hand, 78 (which is supposed to be AArch32) does count something.
16:40j_eyjannau: didnt you copy the wrong side of the define?
16:40mazjust not the same thing as instructions
16:45hir0pro has joined #asahi-dev
16:47mazj_ey: the other side isn't any better.
16:49aleasto has quit [Remote host closed the connection]
16:50aleasto has joined #asahi-dev
17:37bps2 has quit [Ping timeout: 480 seconds]
17:40hir0pro has quit [Quit: hir0pro]
18:26marcankov: 2 M1 Max chips
18:26marcanM1 cannot do multi-die
18:26marcana bunch of the changes in M1 Pro/Max are for this
18:26marcan2x M1 makes no sense, that's already a chip called the M1 Pro.
18:29krbtgti'll note the intel mac mini still exists and would prob call for M1 Pro/Max in its replacement
18:29krbtgtthe plain mini has tons of dead air to fill
18:31TheLinkisn't there supposed to come a redesigned mini (+ max/pro)?
18:36krbtgtthatt's rumoured for prob a spring even
19:29jacoxon has joined #asahi-dev
19:44jacoxon has quit []
19:45jacoxon has joined #asahi-dev
19:55jacoxon has quit []
19:58yrlf has quit [Quit: The Lounge - https://thelounge.chat]
19:59yrlf has joined #asahi-dev
20:20aleasto has quit [Remote host closed the connection]
20:56jacoxon has joined #asahi-dev
21:03yuyichao has joined #asahi-dev
21:06yuyichao_ has quit [Ping timeout: 480 seconds]
21:08yuyichao_ has joined #asahi-dev
21:10jacoxon has quit []
21:10jacoxon has joined #asahi-dev
21:12yuyichao has quit [Ping timeout: 480 seconds]
21:32gladiac is now known as Guest4849
21:32gladiac has joined #asahi-dev
21:38egavinc has quit [Quit: Leaving]
21:38Guest4849 has quit [Ping timeout: 480 seconds]
22:47bps2 has joined #asahi-dev
23:05jacoxon has quit []
…