#asahi-dev · 2021-03-16 21:26
svencan you try adding a large udelay before the first usb call and see if that fixes anything? :>#asahi-dev · 2021-03-16 21:26
svencan you try adding a large udelay before the first usb call and see if that fixes anything? :>#asahi-dev · 2021-03-16 21:25
svenmaybe there's some timing issue with e.g. the phy bringup#asahi-dev · 2021-03-16 21:25
svenmaybe there's some timing issue with e.g. the phy bringup#asahi-dev · 2021-03-16 21:25
svenmaybe there's some timing issue with e.g. the phy bringup#asahi-dev · 2021-03-16 21:25
svenjeffmiw: btw. something that came to my mind: i always chainload to a usb-enabled m1n1 while you probably load one stored on the mac using kmutil#asahi-dev · 2021-03-16 21:25
svenjeffmiw: btw. something that came to my mind: i always chainload to a usb-enabled m1n1 while you probably load one stored on the mac using kmutil#asahi-dev · 2021-03-16 21:25
svenjeffmiw: btw. something that came to my mind: i always chainload to a usb-enabled m1n1 while you probably load one stored on the mac using kmutil#asahi-dev · 2021-03-15 21:11
sventhat's.. strange. i'm using a usb 3.1 c-to-c cable and everything works there, but i thought that all usb-c cables should support usb 2.0 mode#asahi-dev · 2021-03-15 19:26
svenjeffmiw: strange, let me know if i can help in any way#asahi-dev · 2021-03-15 07:53
svensure, right now it's just the dumbest thing possible ;)#asahi-dev · 2021-03-15 07:25
svenalso, just blindly doing those writes from the fake-mmiotrace is enough to bring up the usb phy :D#asahi-dev · 2021-03-15 07:24
svenhttps://github.com/AsahiLinux/m1n1/pull/38 does that on top of the usb stuff#asahi-dev · 2021-03-15 07:24
svenmarcan: yeah, probably a good idea#asahi-dev · 2021-03-14 23:20
svenanyway, bedtime for me. i'll update the fb console MR once the usb stuff is merged#asahi-dev · 2021-03-14 23:19
svenVersion: ImageMagick 7.0.11-3 Q16 arm 2021-03-06 https://imagemagick.org fwiw#asahi-dev · 2021-03-14 23:18
svenmarcan: huh, looks like your -interline-spacing -1 produces a rather strange font.b file on mac os. the individual chars don't seem to line up nicely anymore#asahi-dev · 2021-03-14 15:42
svennothing strange in dmesg either?#asahi-dev · 2021-03-14 15:24
svenseems to work fine even with a linux host here#asahi-dev · 2021-03-14 15:23
svenor maybe not#asahi-dev · 2021-03-14 15:22
svenhm, strange. i had to run shell.py twice before it work on a linux host#asahi-dev · 2021-03-14 15:19
sveni've only tested this with a mac os host so far but it should also work with linux#asahi-dev · 2021-03-14 15:19
svenjeffmiw: does lsusb show the device with the string descriptors?#asahi-dev · 2021-03-14 12:59
sventhe bottleneck now actually is the reboot + building the kernel itself instead of waiting for the upload to finish#asahi-dev · 2021-03-14 12:58
svenloading kernels with this is fun :>#asahi-dev · 2021-03-14 12:36
svenmarcan: https://f.svpe.de/749afd07925249e765a83887ad2281d363ba7e81a764dba9b75ecf3742c450de_phy_bringup.txt and that one is with the reg ranges extracted from the device tree#asahi-dev · 2021-03-14 12:29
svenmarcan: https://files.svenpeter.dev/12b640e63b2747a8a015db20ad8eca43ce0567639e9e649c62f3d96647bb6a5f_phy_bringup.txt#asahi-dev · 2021-03-14 12:16
svenand https://github.com/svenpeter42/m1n1/tree/usb-dwc3-serial-wip is with phy bringup (based on pongoos/kext disassembly :() and a quick uartproxy hack#asahi-dev · 2021-03-14 12:10
svenhttps://github.com/AsahiLinux/m1n1/pull/36 so that's essentially the usb code without the PHY bringup (because I can't really sign that off) and the uartproxy/console integration#asahi-dev · 2021-03-14 09:08
svenmarcan: sounds good. regarding the console: we'll also need a way to disable it for certain prints. (e.g. printing to usb from within the usb code might be bad a idea)#asahi-dev · 2021-03-13 19:59
sveni kinda like the "default to USB/UART and provide a command to switch at runtime" though#asahi-dev · 2021-03-13 19:58
sveni'm happy with all options#asahi-dev · 2021-03-13 19:58
svenyeah :-)#asahi-dev · 2021-03-13 19:57
svenalso fine with me#asahi-dev · 2021-03-13 19:57
svenhm, true#asahi-dev · 2021-03-13 19:57
svenj`ey: also fine with me, but how will you select it at runtime? right now there's no configuration file#asahi-dev · 2021-03-13 19:50
svenmarcan: so.. how do you want to handle uartproxy.c now? should i just patch it to always use usb or would you prefer to have some kind of device that gets passed so that you can switch between uart and usb at compile time?#asahi-dev · 2021-03-13 19:49
sven(after i removed the code that hexdumped every received bulk xfer over the serial uart :D)#asahi-dev · 2021-03-13 19:48
svenM1N1DEVICE=/dev/tty.usbmodemm1n1_uartproxy1 python3 proxyclient/linux.py Image.gz t8103-j274.dtb initramfs.cpio.gz is almost instantaneous now :-)#asahi-dev · 2021-03-13 19:34
svenit would break *so* many things#asahi-dev · 2021-03-13 19:34
sventhat sounds rather unlikely though#asahi-dev · 2021-03-13 19:33
svenif they were completely crazy they could share the TTBR between cores#asahi-dev · 2021-03-13 19:32
sveni didn't check the sctlr though#asahi-dev · 2021-03-13 19:32
svenhuh, i remember spinning up the secondary cpus and it looked like there was no MMU running given how slow writing to the framebuffer was#asahi-dev · 2021-03-13 18:01
svenand then you can test it :-)#asahi-dev · 2021-03-13 18:01
svenjust need to clean up the code and plug it into uartproxy.c#asahi-dev · 2021-03-13 18:00
svenjeffmiw: that probably won't be required. i have read/write working now#asahi-dev · 2021-03-13 16:34
svenmakes sense#asahi-dev · 2021-03-13 16:20
sven*barriers