marcan changed the topic of #asahi to: Asahi Linux: porting Linux to Apple Silicon macs | General project discussion | GitHub: https://alx.sh/g | Wiki: https://alx.sh/w | Topics: #asahi-dev #asahi-re #asahi-gpu #asahi-offtopic | Keep things on topic | Logs: https://alx.sh/l/asahi
dstzd has quit [Remote host closed the connection]
dstzd has joined #asahi
dstzd has quit [Client Quit]
dstzd has joined #asahi
ransom has joined #asahi
klaus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
aratuk has joined #asahi
aratuk has quit [Ping timeout: 264 seconds]
wiloon has joined #asahi
Tokamak has quit [Ping timeout: 264 seconds]
Tokamak has joined #asahi
hspak has joined #asahi
aratuk has joined #asahi
wiloon has quit [Quit: Connection closed for inactivity]
<maximus64>
Spend 2hrs debugging why my USB-C PD chip isn't working. Hook up my scope and all i2c command doesn't get ack, I thought maybe ESD killed the chip so I deadbug another one but same result. Turn out FUSB302B chip I order is different than FUSB302 the one used in vdmtool. @@
* maximus64
facepalm
<maximus64>
quick look at the datasheet and it used different slave address
<maximus64>
ah look like FUSB302.h already have the slave address for my B11 chip
<maximus64>
yup that was it. Seem to be working now. yay :)
<rwhitby>
that reminds me, I wanted to review the PD doc to make sure the 32-bit values are being displayed in the "normal" bit order to make it easy to compare to the spec
<rwhitby>
ok, good, the bits are in the right order in the vdmtool output
<maximus64>
I guess normal is LE?
<rwhitby>
my normal is if the spec says bits 31:0 then it's displayed as 31:0, even though it is transmitted 0:31
Tokamak has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<maximus64>
ah I see
aratuk has quit [Read error: Connection reset by peer]
<rwhitby>
yeah, I wanted to make sure that the 32bit values that vdmtool prints out matches the 32bit values I see in "real" PD analysers ;-)
<rwhitby>
marcan: any additional PD testing stuff you want to hand off, just let me know. Otherwise I'm going to look into the extra I2C buses as that's something I can do without needing to go to 1.2V levels
<rwhitby>
I can compare those I2C responses against the TPS65987 EVM that I have
ky0ko has quit [Remote host closed the connection]
dstzd has joined #asahi
dstzd has quit [Remote host closed the connection]
dstzd has joined #asahi
dstzd has quit [Client Quit]
furkan has quit [Ping timeout: 264 seconds]
furkan has joined #asahi
aratuk has quit [Remote host closed the connection]
dstzd has joined #asahi
dstzd has quit [Remote host closed the connection]
raster has joined #asahi
dstzd has joined #asahi
HoneyTubes has quit [Ping timeout: 272 seconds]
HoneyTubes has joined #asahi
HoneyTubes has quit [Ping timeout: 240 seconds]
ransom has joined #asahi
ransom_ has quit [Ping timeout: 240 seconds]
HoneyTubes has joined #asahi
klaus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<rwhitby>
I suspect that table is the result of the OOLLLL interaction for OO = 0x00 to 0xFF (dword addresses)
<j`ey>
AH
<j`ey>
Ah
thresh has joined #asahi
klaus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<marcan>
yes, that first command
<marcan>
but it's just two blocks
<marcan>
you can read many registers at once
<marcan>
I double checked that I wasn't missing any registers or anything
<marcan>
rwhitby: if you find a way, somehow, of emulating power button presses or otherwise going into 1TR via PD, that would be useful
<marcan>
but if you don't no worries, plain old fast reboots are the most useful feature and we already have that
<rwhitby>
I doubt we'll find a way, as that would be a non-user interaction security breach.
<marcan>
true, good point.
<marcan>
I hadn't thought of that - it's still physical access, but could be done by devices and also emulate a keyboard and subvert everything.... yeah okay they obviously haven't implemented it for that reason
<marcan>
but now I had an idea
<Shiz>
going into 1TR over debug USB :)
<rwhitby>
marcan: nothing stopping you soldering a USB-controlled relay across the power button ;-)
<marcan>
well, more like what if I tape down the power button
<marcan>
rwhitby: I know but so far I've gotten away without tearing this thing apart!
furkan has quit [Ping timeout: 264 seconds]
<Shiz>
don't turn it on, etc.
<marcan>
as expected, then it goes into 1TR forever, even if you pick an OS
<marcan>
not useful
<rwhitby>
USB controlled finger-analogue
<marcan>
yeah... :)
<jix>
any servo motors or solenoids at hand?
<marcan>
hm, I don't think I have any servos lying around
<marcan>
solenoids... doesn't come to mind either
<marcan>
haven't really done *that* kind of project since I moved away from spain... I probably should
<Shiz>
always time to freshen up knowledge
ephe_meral has quit [Ping timeout: 246 seconds]
<rwhitby>
marcan: PD action 0x0103 is really a disconnect, not a PD reset. There's no hard reset message occuring on the bus, the M1 releases it's pull down (this is called an ErrorRecovery in the Type-C spec).
<rwhitby>
did the name "PD reset" come from somewhere, or just something you coined?
<marcan>
but I think the mux mappings reset at least
<marcan>
it could be the TI chip resets
dstzd has quit [Client Quit]
dstzd has joined #asahi
wiloon has joined #asahi
furkan has joined #asahi
aratuk has joined #asahi
aratuk has quit [Ping timeout: 256 seconds]
<justMaku>
afaik, those M1 Mac minis are easy to open unlike intel ones, so soldering something on top of the power button shouldn’t be too hard
<justMaku>
You could even maybe go with clip-on connectors
<dagb>
marcan: you could always use an old cdrom to act as your sw-controlled button pusher. there is precedence for such setups. :-)
<marcan>
I... don't think I have a single motorized tray-loader lying around
<jix>
marcan: I'm pretty sure you have a few servo galvos, ever tested how strong those are? :D (probably a terrible idea)
<marcan>
they are definitely not this strong :p
<marcan>
oh wait, I do have a tray loader
<marcan>
in my PS2
<marcan>
... but given it's so weak it doesn't even work properly, that also sounds like a terrible idea
<marcan>
:p
<dagb>
:-)
<dottedmag>
no kids around to borrow a servo from Lego?
<GaveUp>
would there be time to dev with anklebiters running around? :P
<Fanfwe>
on intel macs, there's always been a probe point for a signal called SMC_ONOFF which, once grounded, turned the computer on. Schematics for the new boards are not available, but I'm pretty sure we can find a similar probe point. Then it's a matter of using a small circuit mosfet to trigger the button automatically
<dhewg>
lego works without kids too ;) search for "savemii robot" on yt
<dhewg>
no wait, that was fischer, nvm me
bitcoinr has joined #asahi
bitcoinr is now known as ray
ray is now known as rayyy
rayyy is now known as rayonx
rayonx is now known as rayyy
rayyy has quit [Quit: Exit game!]
aratuk has joined #asahi
aratuk has quit [Ping timeout: 240 seconds]
tarik02[m] has joined #asahi
xvpower has quit [Read error: Connection reset by peer]
konrad__ has quit [Remote host closed the connection]
konrad__ has joined #asahi
konrad__ has quit [Remote host closed the connection]
konrad__ has joined #asahi
konrad__ has quit [Remote host closed the connection]
konrad__ has joined #asahi
konrad__ has quit [Remote host closed the connection]
konrad__ has joined #asahi
ephe_meral has joined #asahi
konrad__ has quit [Remote host closed the connection]
konrad__ has joined #asahi
konrad__ has quit [Remote host closed the connection]
konrad__ has joined #asahi
konrad__ has quit [Remote host closed the connection]
konrad__ has joined #asahi
konrad__ has quit [Remote host closed the connection]
konrad__ has joined #asahi
konrad__ has quit [Remote host closed the connection]
konrad__ has joined #asahi
konrad__ has quit [Remote host closed the connection]
konrad__ has joined #asahi
dstzd has joined #asahi
dstzd has quit [Read error: Connection reset by peer]
konrad___ has joined #asahi
konrad__ has quit [Ping timeout: 264 seconds]
konrad___ has quit [Remote host closed the connection]
konrad__ has joined #asahi
konrad__ has quit [Remote host closed the connection]
konrad__ has joined #asahi
<opticron>
konrad__, STAHP
konrad__ has quit [Remote host closed the connection]
konrad__ has joined #asahi
dstzd has joined #asahi
dstzd has quit [Client Quit]
modwizcode has joined #asahi
konrad__ has quit [Remote host closed the connection]
konrad__ has joined #asahi
konrad__ has quit []
<davidrysk[m]>
konra____: can you fix your IRC client?
squiffy has joined #asahi
<marcan>
should I pay the dev ransom and find out?
dstzd has joined #asahi
<marcan>
though I suspect talking about it would violate the developer agreement on paper at least :p
<Shiz>
depends on how badly you can wait 2 days :D
<marcan>
is it 2 days?
<justMaku>
you should be able to download the beta with free developer account
<Shiz>
19:35:59 davidrysk[m] : within a day or two if there are no blocker bugs I'd expect it to go public
<Shiz>
dunno for sure but estimate as per above
<Shiz>
:p
<j`ey>
marcan: you said you were thinking about streaming thursday, still consdiering it?
<marcan>
well it'll certainly be interesting if we did finally get kmutil
<davidrysk[m]>
marcan: I'm doing some testing on my end (see pm)
<davidrysk[m]>
marcan: they did
<davidrysk[m]>
I didn't test it but strings shows what I expect it to show :)
<Shiz>
:eyes:
<marcan>
well then.
<marcan>
that will be a fun stream.
rustylerp[m] has joined #asahi
aratuk has joined #asahi
aratuk has quit [Ping timeout: 260 seconds]
<marcan>
wondering if I should just yolo something quick with what I have of mini so far...
<marcan>
I mean it *should* work, right?
<marcan>
(famous last words)
<brentr123[m]>
worst comes the worst, you brick your computer ;)
<marcan>
unlikely it's *that* bad :p
<j`ey>
marcan: have you run it on qemu or whatever?
<marcan>
I tested the relocation code to make sure that worked at least, so I know I'm not doing anything dumb in the startup
<marcan>
the xnu bootinfo stuff/fb I have not tested, but there is some basic scaffolding in there... I mean it's not complicated code... should work right? :p
<justMaku>
marcan: wait, what, is that already running your code?
<justMaku>
damn, that's impressive
<marcan>
j`ey: the colors were wrong the first time.
<j`ey>
marcan: colours smholours
<brentr123[m]>
<justMaku "marcan: wait, what, is that alre"> It’s just a boot logo
<marcan>
because of *course* apple uses a 30bpp framebuffer. that's 10 bits per color. for the ultimate apple logo smoothness.
<j`ey>
brentr123[m]: 'just'
<marcan>
brentr123[m]: which my code puts on screen
<brentr123[m]>
Yes obviously sir
<browzing>
custom kernel booting :) :amaze:
<davidrysk[m]>
Apple moved to 10bpp some time ago (which is nice)
<davidrysk[m]>
wishlist item: true tone support
<davidrysk[m]>
:p
<marcan>
already thought of that
<marcan>
:p
<maximus64>
nice
<marcan>
https://hub.marcan.st/t/wacko_colors.png turns out this is what happens when you try to put a 24bpp logo on 30bpp framebuffer :-) (testing on mac mini via HDMI capture, doing the tweet with the Air was for effect)
<j`ey>
wacko
<jn__>
deep fried asahi apple :)
<maximus64>
that what happen when you drink too much asahi :)
<justMaku>
Really excited for tomorrow :-)
dstzd has joined #asahi
<marcan>
also, uh, apparently there's a watchdog
<marcan>
the thing reboots endlessly, every minute or half or so
<marcan>
... I guess that's going to be near the top of the list of hardware to reverse engineer
<marcan>
:p
<j`ey>
you were looking for a way to reboot without tocuhing it :D
<marcan>
lol :D
<marcan>
I already have that!
<marcan>
I was looking for a way to get into 1TR without touching it!
<eta>
marcan: those are indeed wacko colours
<eta>
to be fair to them their boot logo looks very nice
<maximus64>
Does the device tree have entry for the watchdog? you can try flip bits near there see if you can disable it
<JTL>
figure out how the kernel or whatever "resets" or configures the watchdog
<JTL>
imo
<marcan>
the wdt is in the devtree, yeah
<marcan>
I will implement the uartproxy first so I can experiment with the wdt through that
<JTL>
good
<marcan>
(I have enough time before it barks for it to be useful)
<marcan>
(and it reboots fast)
<j`ey>
marcan: just cos I dont want to miss it, when do you think you will stream roughly?
<davidrysk[m]>
marcan: how easy is it to write something that just jumps to unsigned code? I guess we need to get past the watchdog first
<j`ey>
(aka want to know if i should wake up early heh(
<marcan>
I'm going to guess 15:00 JST ish, maybe
<davidrysk[m]>
(basically a silly bootloader so that it's not necessary to keep rerunning kmutil)
<marcan>
but it'll probably go on for hours
<marcan>
with maybe a dinner break
<davidrysk[m]>
that's 1am for me :(
<marcan>
davidrysk[m]: as soon as I have the uartproxy I can do that in principle
<davidrysk[m]>
I'll probably get on around 9am
<marcan>
and then write a relocating mach-o loader to do it better
<davidrysk[m]>
maybe 8am
<marcan>
that's the plan :)
<marcan>
also this should be able to boot darwin via serial (and usb once that works)
* rwhitby
enjoys being in a similar timezone to JST
<j`ey>
ok that's 7h from now, so like 5am
<marcan>
so also useful for people who want to mess with xnu/darwin builds, once they finally publish code with M1 support
<marcan>
j`ey: you can sleep a bit more, it won't be over in 2 hours
<marcan>
and I may start lter
<marcan>
*later
<marcan>
we'll see
<j`ey>
start when i ping you in the morning, thx :P
<davidrysk[m]>
they are indicating that they will eventually publish code with m1 support
axt has left #asahi [#asahi]
<hthh_>
davidrysk[m]: how so?
<artemist>
It would be nice if tehre were a PR hook to make sure people weren't pasting in APSL code