#
…

2021-11-20

00:27squags has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
00:32squags has joined #asahi-dev
00:32squags has quit []
00:33squags has joined #asahi-dev
01:06squags has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
01:07squags has joined #asahi-dev
01:41squags has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
01:56yuyichao_ has quit [Ping timeout: 482 seconds]
02:20Emantor has quit [Quit: ZNC - http://znc.in]
02:20Emantor has joined #asahi-dev
02:24nico_32 has quit [Remote host closed the connection]
02:27nico_32 has joined #asahi-dev
02:50yuyichao_ has joined #asahi-dev
03:18phiologe has joined #asahi-dev
03:20slicey has joined #asahi-dev
03:21PhilippvK has quit [Ping timeout: 480 seconds]
03:41rkt has joined #asahi-dev
04:03kov has quit [Quit: Coyote finally caught me]
04:03kov has joined #asahi-dev
04:59leah has joined #asahi-dev
06:33slicey has quit [Quit: zzz]
07:32squags has joined #asahi-dev
08:01gabuscus has quit [Read error: Connection reset by peer]
08:05gabuscus has joined #asahi-dev
09:09___nick___ has joined #asahi-dev
09:53slicey has joined #asahi-dev
09:59___nick___ has quit []
10:00___nick___ has joined #asahi-dev
10:01___nick___ has quit []
10:02___nick___ has joined #asahi-dev
10:04Dcow has joined #asahi-dev
10:42the_lanetly_052__ has joined #asahi-dev
10:46svenhuh, looks like it's possible to disable this nvme mode apple calls "linear-sq"
10:48povikthat's good news?
10:48svenno idea yet :D
10:48svenwe'll probably still need a separate (aka copy/pasted) driver and at that point it doesn't really matter i guess
10:49rkt has left #asahi-dev [#asahi-dev]
10:50svenhah, and macOS even supports this mode afaict
10:50svennvme-linear-sq-disable as a bootarg should switch to it i guess
10:51svennvme-linear-sq-disable=1 actually
10:52aleasto has joined #asahi-dev
10:56slicey has quit [Quit: zzz]
10:56svenlooks like that also disables the weird NVMMU
10:57j_eysven: might be nice as a first approach if it removes enough code?
10:57svenyeah, possibly
10:58svenstill need to play with it some more but i think this more or less converts it to a regular NVMe (just not over PCIe)
11:10svenugh. guess it's time to make my nvme tracer support normal nvme as well
11:10svenbut in this mode the NVMMU is completely unused and macOS seems to just use the normal cq/sq
11:11svenit also creates a smaller adminqueue for some reason
11:13kettenisinteresting
11:14kettenismakes some sense since Intel Macs uses "normal" NVMe
11:15kettenisbut on those NVMe was provided by Apples own implementation in the T2 chip
11:16svenyeah, and T2 emulates a slightly-differently-quirked NVMe-over-PCIe for the x86 chip afaik
11:16kettenisI bet we'd need the same quirks for the M1 in this mode
11:18svenso maybe T2 doesn't really emulate another NVMe interface at all but just passes through the ANS interface with "linear sq" disabled
11:25svenmacOS also limits the ioqueue to 0x24 instead of 0x40 entries
11:26kettenismy assumption has always been that the underlying hardware was very similar
11:26kettenisjust different firmware
11:26kettenisbut even that might not be the case
11:29kettenisthe risk of using this other mode is that Apple introduces bugs into it or removes it altogether in the future
11:29kettenisbecause macos doesn't use it
11:30sveni also don't think it makes the linux driver much simpler since the maintainers want us to copy/paste the pci driver anyway
11:32ketteniswell, this might change their minds
11:32sveniirc they already didn't like that we had to add platform support
11:34kettenisfair enough
11:35ketteniswe probably want the NVMMU anyway
11:38svenand it looks like in this "linear sq" disabled mode it needs the same quirks as the T2 macs.
11:38svenor, well, at least the sq entries seem to be 128 bytes just like they are on the T2
11:55ketteniseh, sven, regarding the mailbox
11:55kettenisthe register to start the coprocessor isn't exposed in your binding
11:57kettenishow is that supposed to work?
11:58svenimho that register belongs to whatever node handles the rtkit stuff as well. the thunderbolt co-processors e.g. don't even have that register
12:02kettenisso that's why you added that 2nd range to the nvme node
12:02svenwe also need to poke that register again if we need to reset ANS for whatever reason
12:02sveniirc SEP also doesn't have it
12:03kettenisand SMC doesn't need it
12:04kettenis(trying to convert u-boot to the new mailbox binding)
12:08svenhm.. i wonder if i can increase the nvmmu and queue size beyond the 0x40 macOS configures
12:09kettenisdoubt that's a good idea
12:11svenprobably not. just trying to understand how this HW actually works before I clean up this driver
12:20kettenispushing things until they break is always fun!
12:29bps has joined #asahi-dev
12:42marcansven: I never thought T2 emulates NVMe
12:42marcanalways assumed it was ANS passed through
12:42marcanbut linear SQ being an option makes sense in that context, yeah
12:43marcanit doesn't really make any sense for T2 to emulate another layer on top of ANS, it's already NVMe
12:43sveni just assumed it translated the normal SQ to this linear SQ stuff and passed through most of the rest
12:44marcanI just figured whatever ANS firmware they used for it did things differently
12:52svenok, looks like 0x40 is the actual max queue size
12:54Dcow has quit [Quit: My Mac Mini has gone to sleep. ZZZzzz…]
12:54svenit's possibly to decrease the size of admin queue though. might make sense to see if a smaller admin queue and a bigger io queue results in better performance once everything else works
12:56svenugh, no nevermind. that won't work because the NVMMU entries for that queue will not increase
13:06bps has quit [Ping timeout: 480 seconds]
13:09bps has joined #asahi-dev
13:14gladiac is now known as Guest6367
13:14gladiac has joined #asahi-dev
13:17aleasto has quit [Remote host closed the connection]
13:20Guest6367 has quit [Ping timeout: 480 seconds]
13:22nskl has quit [Remote host closed the connection]
13:23nskl has joined #asahi-dev
14:27bps has quit [Ping timeout: 480 seconds]
14:42slicey has joined #asahi-dev
14:43slicey has quit []
15:13nskl has quit [Quit: WeeChat 3.3]
15:13nsklaus has joined #asahi-dev
15:13nsklaus has quit []
15:16nsklaus has joined #asahi-dev
16:15the_lanetly_052___ has joined #asahi-dev
16:21the_lanetly_052__ has quit [Ping timeout: 480 seconds]
16:31gladiac is now known as Guest6373
16:31gladiac has joined #asahi-dev
16:37Guest6373 has quit [Ping timeout: 480 seconds]
17:21X-Scale has joined #asahi-dev
17:26X-Scale` has quit [Ping timeout: 480 seconds]
18:02nsklaus has quit [Quit: Textual IRC Client: www.textualapp.com]
18:03nsklaus has joined #asahi-dev
18:19nsklaus has quit [Ping timeout: 480 seconds]
18:26nsklaus has joined #asahi-dev
18:53mini has quit [Ping timeout: 480 seconds]
18:59the_lanetly_052___ has quit [Remote host closed the connection]
19:04nsklaus has quit [Quit: WeeChat 3.3]
19:10nsklaus has joined #asahi-dev
19:17nsklaus has quit [Quit: WeeChat 3.3]
19:17nsklaus has joined #asahi-dev
19:18nsklaus has quit []
19:24mini has joined #asahi-dev
19:33yuyichao_ has quit [Ping timeout: 480 seconds]
19:47nsklaus has joined #asahi-dev
19:53Dcow has joined #asahi-dev
20:01Dcow has quit [Ping timeout: 480 seconds]
20:15nsklaus has quit [Quit: WeeChat 3.3]
20:15nsklaus has joined #asahi-dev
20:15nsklaus has quit []
20:15nsklaus has joined #asahi-dev
21:03faiz_abbas[m] has joined #asahi-dev
21:07___nick___ has quit [Ping timeout: 480 seconds]
21:11X-Scale` has joined #asahi-dev
21:14X-Scale has quit [Ping timeout: 480 seconds]
21:54nsklaus has quit [Quit: WeeChat 3.3]
21:54nsklaus has joined #asahi-dev
22:24nsklaus has quit [Quit: WeeChat 3.3]
23:08flying_sausages has quit [Ping timeout: 480 seconds]
23:42nsklaus has joined #asahi-dev
23:46flying_sausages has joined #asahi-dev
…