ChanServ changed the topic of #asahi-dev to: Asahi Linux: porting Linux to Apple Silicon macs | Non-development talk: #asahi | General development | GitHub: https://alx.sh/g | Wiki: https://alx.sh/w | Logs: https://alx.sh/l/asahi-dev
Z750 has quit [Quit: bye]
pb17 has quit [Ping timeout: 480 seconds]
ddxtanx has quit [Remote host closed the connection]
gnuiyl has quit [Read error: Connection reset by peer]
gnuiyl has joined #asahi-dev
mischa85 has joined #asahi-dev
ravikant_ has joined #asahi-dev
mischa85 has quit [Ping timeout: 480 seconds]
<sven>
mischa85: when you tried to bring up the secondary cores did you see anything on the serial line?
<sven>
if they’re all headed to the entry point you should see m1n1 (or parts of it) and then something might crash because we’re running the entry point again
tamar-away is now known as tamar
chadmed has quit [Read error: Connection reset by peer]
chadmed has joined #asahi-dev
pb17 has quit [Ping timeout: 480 seconds]
yuyichao_ has joined #asahi-dev
pb17 has joined #asahi-dev
roxfan has quit [Ping timeout: 480 seconds]
pb17 has quit [Ping timeout: 480 seconds]
ravikant_ has quit [Ping timeout: 480 seconds]
pb17 has joined #asahi-dev
ravikant_ has joined #asahi-dev
ethantwardy has quit [Quit: WeeChat 4.4.2]
ethantwardy has joined #asahi-dev
roxfan has joined #asahi-dev
mischa85 has joined #asahi-dev
ravikant_ has quit [Ping timeout: 480 seconds]
mischa85 has quit [Ping timeout: 480 seconds]
psykose has quit [Remote host closed the connection]
psykose has joined #asahi-dev
mripard has joined #asahi-dev
ravikant_ has joined #asahi-dev
rrendec has joined #asahi-dev
ddxtanx has quit [Ping timeout: 480 seconds]
pb17 has quit [Ping timeout: 480 seconds]
john-cabaj has joined #asahi-dev
pb17 has joined #asahi-dev
mischa85 has joined #asahi-dev
mischa85 has quit [Ping timeout: 480 seconds]
pb17 has quit [Ping timeout: 480 seconds]
<nickchan>
so about pmu event descriptions
<nickchan>
there is a bunch of tools/perf/pmu-events/arch/arm64/**/core-imp-def.json in Linux for different cores
<nickchan>
can definitely add the event names at least they are already in the kernel
<nickchan>
descrption for those events can be found in the apple silicon cpu optimization pdf but of course that's copyrighted
<nickchan>
and there is probably more than one way to write the description for some of the events, that is, possibly involves creativity and therefore copyright
<nickchan>
like there's probably only one way (that fits with the kernel style) to write the description of RETIRE_UOP but the same cannot be said for like, FLUSH_RESTART_OTHER_NONSPEC
<nickchan>
all short descriptions, less than 20 words, most less than 10 words
<chaos_princess>
more of a shitpost than a real idea, but what about going for a firmware-style approach, where you have a script that can be fed the pdf and then create the perf jsons?
pb17 has joined #asahi-dev
<nickchan>
chaos_princess: that already exists
opticron has quit [Read error: Connection reset by peer]
<chaos_princess>
by direct link i meant sth like applecdn.com/the-pdf-in-question.pdf that you can hit w/o auth
<alyssa>
chaos_princess: 5/10, needs more LLMs
<alyssa>
have we considered getting the pdfs from the blockchain and then deep learning them with GPT
<chaos_princess>
alyssa: so, before that idea i was typing sth like "just use a llm to rewrite it and check by hand, after all llm output is uncopyrightable", but felt too dirty to post that before you said that
<alyssa>
how about NFTs
<alyssa>
i once read a comment on reddit that mentioned NFTs so i'm an expert on them and we should use them here
<chaos_princess>
you deserve to live in america for saying that :P
<alyssa>
=D
<alyssa>
(for the benefit of everyone here but chaos_princess, all of these messages are shitposts)
<alyssa>
(for chaos_princess , no no no this is 100% srs)
<chaos_princess>
for benefit of nobody i guess my messages are schroedinger's shitposts where if they are or not depends on what is funnier
<nickchan>
in the license agreement it does say the pdf can be used to create a licensed product ... "a Licensee software package that is designed to run on or be interoperable with Apple products."
<nickchan>
and all of this boils down a laywer's interpetation of adding perf event descriptions that are either a pharse or a sentence is "using" or "distributing"
<chaos_princess>
where is the license text? i can't seem to find it
<nickchan>
chaos_princess: if you follow the authed link with an apple id you will find the license text
<chaos_princess>
except, i get straight to the download page, since i think i already accepted it on my apple id
<nickchan>
chaos_princess: yeah right then you are "supposed" to have read and understodd it :-)
ravikant_ has quit []
kenzie7001031 has joined #asahi-dev
chadmed has quit [Quit: Konversation terminated!]
chadmed has joined #asahi-dev
<nickchan>
since i talked about legal stuff i probably should rename some of my linux repos... eventually
<nickchan>
"linux-apple" is not appropriate at all
<nickchan>
need a better name besides the obvious trademark issue
<nickchan>
unfortunately, i am not good at imagination
<alyssa>
linux-pomme
<nickchan>
"linux-sandcastle" would cause confusion, "linux-fruit" is too boring
<nickchan>
hmm
<alyssa>
or because it's for old iphones
<alyssa>
linux-pomme-de-terre
<chaos_princess>
linux-ananas :P
<nickchan>
alternatively i can be like apple and name it after some ape
<tobhe>
this is why in Debian we have a Bananas Team :)
<alyssa>
tobhe: not because you have to be bananas to use a Mac? ;P
ddxtanx has joined #asahi-dev
pb17 has quit [Ping timeout: 480 seconds]
<tobhe>
next time someone asks I'll go with that explanation
mripard_ has joined #asahi-dev
mripard has quit [Ping timeout: 480 seconds]
pb17 has joined #asahi-dev
mattia013 has quit [Remote host closed the connection]
mattia013 has joined #asahi-dev
mripard has joined #asahi-dev
mripard_ has quit [Ping timeout: 480 seconds]
___nick___ has joined #asahi-dev
mattia013_ has joined #asahi-dev
<alyssa>
=D
<sven>
lol
mattia013_ has quit [Read error: No route to host]
mattia013_ has joined #asahi-dev
___nick___ has quit []
___nick___ has joined #asahi-dev
___nick___ has quit []
___nick___ has joined #asahi-dev
mattia013_ has quit [Ping timeout: 480 seconds]
JayBeeFOSS has quit [Ping timeout: 480 seconds]
JayBeeFOSS has joined #asahi-dev
nela has quit [Quit: bye!]
nela has joined #asahi-dev
tanty has quit [Quit: Ciao!]
tanty has joined #asahi-dev
mattia013_ has joined #asahi-dev
pb17 has quit [Ping timeout: 480 seconds]
pb17 has joined #asahi-dev
tanty has quit [Quit: Ciao!]
tanty has joined #asahi-dev
___nick___ has quit [Remote host closed the connection]
tanty has quit [Quit: Ciao!]
tanty has joined #asahi-dev
tanty has quit [Quit: Ciao!]
tanty has joined #asahi-dev
catred has joined #asahi-dev
mattia013_ has quit [Ping timeout: 480 seconds]
pb17 has quit [Ping timeout: 480 seconds]
ciara has quit [Ping timeout: 480 seconds]
ciara has joined #asahi-dev
<jannau>
probably no unaligned access to PCI BAR memory on mac pro PCIe slots. intel dg2 with 16k fixes is not particularly happy (many firmware schedular job timeouts) and I get strange SErrors
<jannau>
for unaligned accesses to vram BAR from m1n1's hypervisor I get unaligned SErros as one would expect
<jannau>
I've pushed asahi-6.14.1-1 to github. Please test
<kettenis>
jannau: how are those BARs mapped? Normal-NC or Device-something?
<chaos_princess>
mapped as normal or device?
<nicolas17>
jannau: damn you got a j180? :D
<jannau>
should be mapped as device. xe driver uses iomap_wc()
<jannau>
err, mapped as normal memory
<chaos_princess>
and m1n1 too?
<kettenis>
bet m1n1 is using device-something
<kettenis>
this is another variant of the usual problem with GPUs on arm64
<kettenis>
the only thing that matches really x86 WC is Norma-NC
<kettenis>
but a lot of the interconnects found on arm64 SoCs don't support that
<jannau>
yes, m1n1 is using Device-nGnRE. I ignored that because it doesn't matter after linux is booted
<jannau>
but of course that's not true under the hypervisor
<kettenis>
at least Apple throws an SError instead of failing to produce the right results in random ways
<kettenis>
would be interesting to see whether Device-GRE works
<kettenis>
but IIRC you get an SError for Device-nGnRnE too, so I expect the hardware really only works with Device-nGnRE
<chadmed>
i would be surprised if Device-GRE worked. theres nothing in the apple ecosystem that would really require it
<chadmed>
_unless_ the rumours about server-grade apple silicon devices are real and this was a practice run
<chadmed>
no ones gonna be running AGX in a datacentre
<chaos_princess>
the pcc boxes are arguably "server-grade", but that uses agx gpus
<chaos_princess>
but we know that the j180 pcie is different... somehow, just not in what specific way
pb17 has joined #asahi-dev
Calandracas has quit [Ping timeout: 480 seconds]
Calandracas has joined #asahi-dev
<jannau>
Normal NC seems to work after mapping it as such in m1n1
<jannau>
so the xe issue I still see could be still page size related
<chaos_princess>
hell yes
<chaos_princess>
unaligned accesses too?
<jannau>
oh no. esr 0xbe000000 is "32-bit instruction trapped"
<jannau>
that is what xe triggers in linux on load
ciara has quit [Remote host closed the connection]
<jannau>
yes, unaligned reads and writes seem to work
ciara has joined #asahi-dev
gprimola has joined #asahi-dev
gprimola has quit [Quit: Konversation terminated!]
<kettenis>
jannau: 0xbe000000 >> 26 = 0x2f which is SError
<jannau>
with the ah, yes. it's coming back to me. and it's annoyingly async so PC doesn't make sense when it arrives
<jannau>
s/with the //
<kettenis>
the l2c_err_xxx registers might provide some hints