#asahi-dev · 2025-09-21 17:01
nickchanjannau: openbsd's install's instructions uses uefi-only as well#asahi-dev · 2025-09-21 16:57
jannaunot on that branch, no rust adt code in there#asahi-dev · 2025-09-21 16:56
jannauI pulled the atc fuses changes as well under the assumption that running upstream linux is more likely on uefi-only#asahi-dev · 2025-09-21 16:56
nickchanthere's still a regression in adt_is_compatible()#asahi-dev · 2025-09-21 16:55
jannauI've created https://github.com/AsahiLinux/m1n1/tree/releases/1.5.1 for tagging m1n1 v1.5.1#asahi-dev · 2025-09-21 15:15
jannaualso not really something to worry about for its use cases#asahi-dev · 2025-09-21 15:15
jannautps6598x_check_mode() during resume might cause issues when *vdmtool races with suspend#asahi-dev · 2025-09-21 15:04
svenso probably nothing to worry about if you're using this tool anyway#asahi-dev · 2025-09-21 15:04
svenand that can only happen through user interaction via typec_ops i think#asahi-dev · 2025-09-21 15:03
sveni think the only possible race is when we try to send a command and the kernel driver also sends one concurrently#asahi-dev · 2025-09-21 14:52
jannauI've created https://github.com/AsahiLinux/tuxvdmtool/issues/1#asahi-dev · 2025-09-21 14:49
jannauit exists at least since 2007 (`git log --follow` doesn't at that point)#asahi-dev · 2025-09-21 14:46
svenjust reading the i2c-dev.c code and it looks like it should work at least#asahi-dev · 2025-09-21 14:46
sven*adjusting#asahi-dev · 2025-09-21 14:46
svenhm.. maybe it was added without adjust the documentation?#asahi-dev · 2025-09-21 14:45
jannauhttps://docs.kernel.org/i2c/dev-interface.html doesn't mention I2C_SLAVE_FORCE#asahi-dev · 2025-09-21 14:30
svenlol, i wasn't aware that essentially also does raw i2c#asahi-dev · 2025-09-21 14:30
jannauat this point it probably makes sense to integrate macvdmtool as separate backend as well#asahi-dev · 2025-09-21 14:29
jannauI guess I'll port tuxvdmtool to i2c-client then#asahi-dev · 2025-09-21 14:28
jannauit works in macvdmtool#asahi-dev · 2025-09-21 14:28
svenwhat could go wrong :>#asahi-dev · 2025-09-21 14:28
jannauyes: I2C_SLAVE_FORCE ?* Use this slave address, even if it is already in use by a driver! */#asahi-dev · 2025-09-21 14:21
svenno idea if it would work, but there's not much activity on that bus usually #asahi-dev · 2025-09-21 14:20
jannauI'm not planning to upstream the sysfs interface#asahi-dev · 2025-09-21 14:20
sven*mode#asahi-dev · 2025-09-21 14:20
sveni2ctransfer has a -f mdoe that comes with a bunch of warnings but "Force access to the device even if it is already busy"#asahi-dev · 2025-09-21 14:20
nickchan{0x5ac8012, 0x824606}; <- this gives "Debug USB"#asahi-dev · 2025-09-21 14:20
jannauI did not check the linux source#asahi-dev · 2025-09-21 14:19
jannauI did a quick search if i2c-client works if a driver claims the chip id already. no clear answer there but it sounded more like it won't work#asahi-dev · 2025-09-21 14:18
nickchanand that gives swd but it's not particlarly useful on production devices#asahi-dev · 2025-09-21 14:18
nickchanthere's also a vdm commands that gives a USB PID of 0x1881 i think#asahi-dev · 2025-09-21 14:17
svenalso, at this point it may make sense to split stuff into cd321x.c since core.c is getting quite huge#asahi-dev · 2025-09-21 14:17
nickchanhowever t2 mac cannot serve as a host with macvdmtool#asahi-dev · 2025-09-21 14:17
svenhm, i guess we can't do userspace i2c if there's already a driver probed?#asahi-dev · 2025-09-21 14:17
nickchan(for macvdmtool)#asahi-dev · 2025-09-21 14:17
nickchanit works as both a target and host#asahi-dev · 2025-09-21 14:16
nickchanjannau: works on m2 ipad at least#asahi-dev · 2025-09-21 14:15
jannauI think modern iphones and ipads should work as well. not sure if locked devices produce uart output#asahi-dev · 2025-09-21 14:10
nickchanit can be used to control a t2 target however only reboot and dfu will work#asahi-dev · 2025-09-21 14:02
jannauideas for a better driver "API" than sysfs like in https://github.com/AsahiLinux/linux/commit/786523ac62f0aeec37bf9c6b991e8bf2fadc590d would be welcome as well#asahi-dev · 2025-09-21 14:00
jannauIf someone can look into why the samsung uart driver resets the port within 5-20 seconds on host that would be welcome#asahi-dev · 2025-09-21 13:56
jannauI've pushed asahi-6.16.8-1 to github with support for https://github.com/AsahiLinux/tuxvdmtool#asahi-dev · 2025-09-21 11:46
jannaukettenis: is there a reason why you dropped the sleep for the pci reads in https://github.com/AsahiLinux/u-boot/commit/52593774ec11c8ff57661109a764d24c27aa9561 ? granted udelay(1) is not much and its primary use was probably the timeout itself#asahi-dev · 2025-09-21 10:28
nickchanlinux driver causes the kernel boot to be frozen on t2 (actually frozen not just the screen); guess I should try m1n1's touchbar bad apple#asahi-dev · 2025-09-21 07:59
nickchan0x87a = 2170 and 0x3c = 60; so at least ADP_SCREEN_SIZE is the same and this is not the case on A11#asahi-dev · 2025-09-21 07:59
nickchan>>> read32(u.adt["/arm-io/disp0"].get_reg(2)[0] + 0xc) 0x87a003c#asahi-dev · 2025-09-21 07:54
nickchant2's display play reg layout in adt is similar to m1 dfr too#asahi-dev · 2025-09-21 07:53
nickchani probably should investigate #asahi-dev · 2025-09-21 07:53
nickchan(in adt)#asahi-dev · 2025-09-21 07:53
nickchanthis is suspicious the reg layout of a9 and a10's display pipe is more similar to m1's dfr display pipe than a11's display pipe