00:32
<
alyssa >
j`ey: on parle seulement l'anglais, merci beaucoup :-p
01:06
yuyichao_ has quit [Ping timeout: 480 seconds]
02:00
DragoonAethis has quit [Quit: hej-hej!]
02:00
DragoonAethis has joined #asahi-dev
02:20
Emantor has joined #asahi-dev
03:16
yuyichao has joined #asahi-dev
03:26
PhilippvK has joined #asahi-dev
03:29
phiologe has quit [Ping timeout: 480 seconds]
03:48
kov has quit [Quit: Coyote finally caught me]
03:51
kov has joined #asahi-dev
04:12
axboe_ has quit [Quit: leaving]
06:17
<
Glanzmann >
povik: mini and air target are online.
06:43
sirn- has joined #asahi-dev
06:49
sirn has quit [Ping timeout: 480 seconds]
06:49
sirn has joined #asahi-dev
06:56
sirn- has quit [Ping timeout: 480 seconds]
07:08
<
rkjnsn >
marcan, yeah, 300-600 IOPS isn't great, but it's a lot better than 46 for cases where one wants to ensure internal consistency but doesn't care about losing a second or two of writes. It's almost certainly the mode I'll end up using for fsync if we provide it as an option.
07:27
willemml has joined #asahi-dev
07:40
<
marcan >
rkjnsn: that's partially a filesystem thing; according to axboe linux doesn't really do barriers these days
07:41
<
rkjnsn >
Right. I mean if there was an option to translate fsyncs into drive barriers instead of full flushes.
07:41
<
marcan >
I don't think drive barriers are a thing
07:41
<
marcan >
the barrier flush thing in macos is probably just something it constructs on top of the cow/journal of the filesystems
07:42
<
marcan >
and it's still slow because it still has to flush once in a while, but can probably enforce barrier semantics across multiple writes through the metadata without flushing every time
07:43
<
marcan >
no --fsync?
07:45
<
marcan >
not sure what test sven was doing but I'm pretty sure you aren't flushing each write that way if you're getting that performance
07:45
<
j`ey >
alyssa: desole mais tu viens de parler en francais aussi!
07:46
<
rkjnsn >
Without some way to pass the barrier on to the ANS, how would it enforce ordering to persistent storage without requiring individual flushes?
07:46
<
Glanzmann >
marcan: This was with axboes patch, I just wanted to show the file system overhead.
07:46
<
marcan >
oh, well then yeah
07:52
<
rkjnsn >
Maybe it's worth seeing if there's a similar speedup for F_BARRIERFSYNC vs F_BARRIERFSYNC on an HFS+ partition, which presumably can't be doing any CoW shenanigans try to enforce ordering with fewer flushes.
08:22
<
marcan >
alright, so things for the next asahi
08:23
<
marcan >
am I forgetting anything other than what was already in spmi/work and jannau's spi fixes?
08:23
<
marcan >
sven: nvme stuff? I think RTKit is already in order in the spmi branch, right?
08:23
<
Glanzmann >
marcan: Two patches from axboe:
08:24
<
marcan >
povik: any new audio updates?
08:25
<
marcan >
ah right, that bug
08:25
<
Glanzmann >
Yep, without the first patch with cpufreq 'nvme' hangs. Second one for performance.
08:25
<
jannau >
rtkit should be covered by spmi/work and works for nvme, smc, and dcp
08:26
<
povik >
yes, some fixes at the very least
08:26
<
povik >
marcan: you doing this right now?
08:27
<
povik >
what's the base? latest next?
08:27
<
j`ey >
the 'defer nvme flush' commit has the interval defaulting to 1s, there was talk of that actually defaulting to 0s (aka always flush)
08:27
<
j`ey >
povik: next-20220217
08:27
willemml has quit [Quit: willemml]
08:28
<
marcan >
povik: I pushed asahi-soc/next, that
08:28
<
marcan >
(which is what j`ey said)
08:28
<
marcan >
jannau: ah, thanks, missed thsoe
08:29
<
rkjnsn >
As a random person on the internet, I would feel better defaulting to always flushing.
08:30
<
povik >
then one or two audio fixes will be there already :-)
08:30
<
povik >
let me prepare a 'for-marcan' then
08:30
<
rkjnsn >
(I also think abxoe mentioned that flushing by default would be needed for it to be accepted upstream.)
08:31
<
Glanzmann >
rkjnsn: Yes, he did.
08:31
<
mps >
(I prefer consistency and correctness over performance)
08:32
<
marcan >
I think I'm going to make the default deferring flushes
*this time* just so people test that codepath
08:33
<
marcan >
if you're concerned about consistency/safety you probably shouldn't be running this kernel anyway for a myriad of other reasons :-)
08:33
<
marcan >
but it will not be the default down the road, at least not on desktops
08:33
<
j`ey >
that sounds reasonable
08:33
<
Glanzmann >
marcan: \o/
08:33
<
mps >
marcan: yes, for current status it is not important
08:35
<
marcan >
done with the merges so far, fixing build issues due to conflicts now
08:35
<
marcan >
I think that's just povik's changes left :)
08:36
<
povik >
it will take me half an hour let's say
08:36
<
povik >
there are changes that i want to keep out and i need to test the result
08:38
<
marcan >
ah, just a fixup, sure
08:38
<
Glanzmann >
Yes, he did. Small performance improvement.
08:38
<
j`ey >
marcan: btw are either of your m1 max/pro the lower core models?
08:39
<
Glanzmann >
j`ey: You should no him better than that.
08:39
<
marcan >
j`ey: noipe
08:39
<
marcan >
and yeah, I need to fix m1n1 for that
08:39
<
j`ey >
Glanzmann: he was trying to cover a range of models :)
08:40
<
marcan >
I can simulate it with the HV though, easily
08:40
<
marcan >
just nuke nodes there
08:40
<
marcan >
we already do it for !SMP mode anyway
08:42
<
marcan >
for now I pushed updated bits/*, in case anyone wants to take a look
08:50
<
mps >
which branch? asahi-soc/next?
08:51
* mps
goes to prepare tea and after that build and test
08:51
<
jannau >
mps: just the bits branches for now
08:51
<
jannau >
still waiting for audio fixes
08:51
<
jannau >
asahi-soc/next is just the base
08:52
<
mps >
jannau: aha, I see. thanks
08:52
<
Glanzmann >
jannau: Are your dtsi and spi patched for the device tree for u-boot in?
08:52
<
j`ey >
marcan: 13bff188e3817dbede4fbb458d3e48847058a42e is not needed anymore, fixed in next
08:53
<
j`ey >
marcan: see 147ab5376f18045da9f22a8262185707745bbf77
08:53
<
marcan >
ah, thought so but I didn't check :)
08:54
<
marcan >
dropped it
08:57
<
marcan >
they are already squashed if they're fixups
08:57
<
jannau >
Glanzmann: the changes are squashed in the main commits
09:06
<
jannau >
looks good otherwise
09:07
<
marcan >
thanks, I'll throw it in misc
09:08
<
j`ey >
marcan: I think macsmc_rtc_get_time is leaking p_off
09:08
<
j`ey >
nvmem_cell_read() states buffer should be freed by the consumer with a kfree().
09:09
<
Glanzmann >
jannau: I see, thanks. I did not know.
09:10
<
marcan >
j`ey: indeed, good catch
09:10
* marcan
is not a fan of APIs that return newly allocated buffers...
09:11
<
j`ey >
also theres a double newline before macsmc_rtc_probe if youre editing that file
09:15
<
marcan >
fwiw, the asahi diffstat:
09:15
<
marcan >
165 files changed, 18045 insertions(+), 559 deletions(-)
09:16
<
j`ey >
marcan: I suppose that means that nvmem_cell_get_u8 is leaking too, in macsmc-reboot
09:17
psykose has quit [Remote host closed the connection]
09:17
<
j`ey >
I mean, if youre about to reboot or powroff anyway...
09:18
psykose has joined #asahi-dev
09:19
<
jannau >
18k insertions is less than I would have guessed
09:19
<
marcan >
well, it'll leak every time you read that sysfs file :)
09:20
<
j`ey >
bad indentation
09:21
<
j`ey >
seems to be spaces, not tabs
09:21
<
marcan >
yeah, copypaste fail
09:26
<
j`ey >
marcan: not sure for the reason of having an 'enum pm_setting'
09:26
<
j`ey >
you just use them as ints in ac_power_mode_map[]
09:26
<
marcan >
no real reason, it could be #defines
09:27
<
j`ey >
#defines seems to be the usual way, but maybe this subsystem does it like that?
*shrug*
09:27
<
marcan >
no idea tbh, I usually favor enums for this kind of thing but I haven't checked
09:41
<
marcan >
that supersedes bits/070-audio, right?
09:42
<
marcan >
except the dt bits
09:42
<
povik >
yes, also the dt bits
09:50
<
marcan >
sound/soc/apple/mca.c: In function ‘apple_mca_probe’:
09:50
<
marcan >
sound/soc/apple/mca.c:917:45: warning: format ‘%d’ expects argument of type ‘int’, but argument 4 has type ‘long int’ [-Wformat=]
09:50
<
marcan >
sound/soc/apple/mca.c:951:47: warning: too many arguments for format [-Wformat-extra-args]
09:50
<
marcan >
sound/soc/apple/mca.c:951:47: warning: format ‘%d’ expects argument of type ‘int’, but argument 4 has type ‘struct device_node *’ [-Wformat=]
09:50
<
marcan >
that looks mildly broken :)
09:58
<
marcan >
also heh, panics early on t6000... I need to look at this
09:58
<
marcan >
(not the audio)
10:00
<
j`ey >
marcan: copy paste fail again in smc :P
10:00
<
j`ey >
kfree(buf); should be p_off
10:01
<
marcan >
already fixed :p
10:02
<
j`ey >
thanks to git push -f, it never happened!
10:03
<
povik >
clang didn't seem to kick me for those
10:12
<
marcan >
edecc06b4d34e9
10:12
<
marcan >
rebuilding, but I'm pretty sure that's broken
10:14
<
j`ey >
you have CONFIG_SPARSEMEM_EXTREME?
10:14
<
j`ey >
oh, I do too lol
10:14
<
j`ey >
sounds too EXTREME for regular use
10:15
<
marcan >
I love people fixing compiler warnings without reading the code
10:17
<
marcan >
yeah, reverting that (and the subsequent fixup) fixed it)
10:24
<
j`ey >
at least you found it quickly
10:27
<
marcan >
angry email sent :p
10:32
<
j`ey >
how can that not be breaking more things.
10:32
<
marcan >
j`ey: very few machines don't have memory at 0
10:32
<
marcan >
pushed asahi, barely smoke tested, have fun :-)
10:32
<
j`ey >
I thought a decent amount of arm machines didnt
10:32
<
marcan >
t8103 does boot and doesn't; I think it needs to be significantly above 0 :)
10:33
* marcan
off to shower and, er, lunch?
11:08
<
sven >
marcan: yeah, rtkit in there + axboe's patches are all fine
11:09
<
sven >
the flush thing is a bigger hack than it needs to be but it's good enough for now
11:09
<
sven >
(and axboe wanted to clean that up anyway)
11:09
<
mps >
povik: I booted latest asahi, audio card is not detected on MBP
11:10
<
mps >
and I don't see any new option related to audio .config
11:10
<
mps >
maybe DTS is changed
11:11
<
marcan >
CONFIG_SND_SOC_APPLE_SILICON
11:12
<
Glanzmann >
Thanks.
11:12
<
mps >
marcan: already had it set
11:14
<
mps >
Glanzmann: I didn't noticed this when compiling
11:14
<
marcan >
are you missing ARCH_APPLE?
11:15
<
Glanzmann >
marcan: Yes, I did. Forgot that my build box is now an amd64 ...
11:15
<
Glanzmann >
You found it faster than I, because that was the last thing that I checked.
11:15
<
mps >
marcan: ah, you asked Glanzmann and not me, sorry
11:15
<
marcan >
the message literally told you...
11:16
<
Glanzmann >
marcan: Yes, it did, but I started testing kernel options from the bottom up, now top down.
11:17
<
marcan >
but that sounds like SND_SOC_APPLE_MCA needs a dep on ARCH_APPLE too to prevent that broken dep chain
11:17
<
marcan >
Glanzmann: it's the only thing in the message that doesn't have a [y] next to it and isn't COMPILE_TEST
11:17
<
marcan >
it was the only option
11:17
<
mps >
I have few aarch64 machines (some are powerfull) and I always build on them everything for aarch64
11:20
<
Glanzmann >
marcan: I see.
11:20
<
marcan >
no, didn't realize that patch fixed a broken build
11:20
<
marcan >
just turn off that driver
11:21
<
Glanzmann >
mps: Debian bootstrap.sh needs 8 minutes on the air, 4 on the ryzen ...
11:22
<
povik >
shouldn't have touched the kconfigs
11:22
<
povik >
let's see what i did there
11:23
<
Glanzmann >
povik: I think it is unrelaed, maybe I had it just on in my usually broken config without any reason whatsoever.
11:35
<
kettenis >
marcan: I just created an "asahi" u-boot branch with all the diffs that we still need on top of u-boot master
11:36
<
kettenis >
but maybe it makes sense to have that under github.com/AsahiLinux?
11:38
<
j`ey >
did poweroff also rely on 12.x firmware, or just rtc?
11:39
<
kettenis >
just rtc I think
12:02
<
povik >
since the asahi branch now has NCO patches verbatim as they are on the list
13:35
___nick___ has joined #asahi-dev
13:37
___nick___ has quit []
13:39
___nick___ has joined #asahi-dev
13:40
___nick___ has quit []
13:42
___nick___ has joined #asahi-dev
14:35
joerg has joined #asahi-dev
14:35
joerg is now known as Guest293
14:36
<
Guest293 >
oh sorry
14:36
Guest293 has quit []
14:41
joerg1 has joined #asahi-dev
14:56
<
povik >
no, that's for the cs42l83 codec on t8103 macs
14:57
<
ah-[m] >
ah, t6000 is cs42l84
14:58
<
povik >
and the l92 even has a linux driver!
15:00
<
povik >
er, similar to cs42l84, not cs42l83
15:02
<
sven >
looks like i just can't reproduce those missed nvme interrupts anymore (without the additional lock around command submission) no matter what i try :/
15:13
<
jeffmiw >
j`ey: I'm with an old macOS 11.5.2 and neither rtc nor shutdown work with latest asahi branch, currently updating my system to latest macOS 12 to re-test
15:14
<
j`ey >
jeffmiw: shutdown worked for me in the end, also 11.x. make sure you had all the SPMI/MFD/NVMEM options
15:14
<
j`ey >
jeffmiw: but upgrading is needed anyway soo
15:14
<
j`ey >
(for rtc I mean, and dcp)
15:18
axboe has joined #asahi-dev
15:21
<
jeffmiw >
I had all the SPMI/MFD/NVMEM options, I saw some errors on the spmi nodes in the dt, which model are you testing on ? I'm on mba
15:22
<
j`ey >
jeffmiw: "bad cell count" errors are currently expected. MBA here too
15:23
<
jeffmiw >
ok so probably something else wrong in my setup then. I dig into it if I still see the issue after updating. For testing power off I tried halt or shutdown, how do you so it ?
15:24
<
j`ey >
I did 'poweroff -f' but that was from an initramfs setup with busybox
15:25
<
jeffmiw >
ok, I'll check this. Thanks. I also have a minimal initramfs with busybox.
15:26
<
mps >
poweroff also works for me
15:38
axboe has quit [Quit: leaving]
15:40
axboe has joined #asahi-dev
16:03
grgy has joined #asahi-dev
16:18
<
kettenis >
hmm, did I misunderstand that the CLUSTER_PSTATE register contains the actual power state of the cluster as well as the desired state?
18:52
willemml has joined #asahi-dev
19:22
willemml has quit [Quit: willemml]
19:59
bps has quit [Ping timeout: 480 seconds]
20:01
willemml has joined #asahi-dev
20:03
willemml has quit []
20:07
bps has joined #asahi-dev
20:10
boardwalk has joined #asahi-dev
20:18
Hiroshi has joined #asahi-dev
20:19
Hiroshi is now known as Guest306
20:19
axboe has quit [Quit: leaving]
20:20
Guest306 has quit []
20:21
axboe has joined #asahi-dev
20:54
axboe has quit [Quit: reboot]
20:56
axboe has joined #asahi-dev
21:07
___nick___ has quit [Ping timeout: 480 seconds]
21:17
bps has quit [Ping timeout: 480 seconds]
23:02
<
alyssa >
j`ey: blasphemie
23:03
<
j`ey >
alyssa: excusez-moi pour chaque faux pas que j'ai faire
23:25
Gaspare has joined #asahi-dev
23:40
willemml has joined #asahi-dev
23:41
willemml has quit []