Unable to boot ISO

[Coldfirex]

Howdy, I am finally getting around to testing Haiku on real hardware (AMD 3300x, 16GB, nvidia 1650 super, etc). I am using the B3-TC2 ISO and a couple times it has booted to the installer and live desktop correctly. Most of the time, however, it fails at the Haiku icons. When this happens:

*None of the icons light up.

*Safe mode does not help much (1/10 times I got it to boot)

*Video fallback and disabling USB and nvidia driver didnt do anything

*My usb keyboard shuts off when it fails to boot

*It appears to be a hard lock and I have to power off the system to retry

*No change with APIC, SMP, ACPI, SMEP, etc disabled

*The onscreen debugging is always the same, but I dont see a printed error ​https://imgur.com/a/gjelywO and log output ​https://pastebin.com/7NqB3DNc.

Successful boot log ​https://pastebin.com/gcFyy9zK

[kallisti5]

"None of the icons light up."

Hm. I used to see something like this on my system. Are you booting our media via EFI or BIOS? (generally, bios boot menus will show "USB STICK XXX, EFI: USB STICK XXX". You want the one prefixed by EFI.

My Ryzen desktop won't boot via the older mbr method and shows a similar fault, but works fine via EFI.

[Coldfirex]

So all the tests were with EFI above. I tried the old style and actually had worse luck. Sometimes the bootloader options would never appear, sometimes I would only get a blinking cursor before the iso boot loader, sometimes the same result of no icons lit up, and sometimes even just nothing. Keyboard was always "locked". I reflashed by USB key with a fresh TC ISO to be safe as well.

Are there any alternatives to getting boot logs if no serial output at this point?

[Coldfirex]

After the recent USB changes I was able to boot off of a new anyboot ISO (hrev55376). Feel free to close.

[waddlesplash]

KERN: usb error xhci 0: ConfigureEndpoint() failed invalid max_burst_payload
KERN: usb error xhci 0: unable to configure endpoint: Invalid Argument

[waddlesplash]

Well, that error is still in there, and I don't see any considerable differences in your syslog. I'm not really convinced this is really XHCI related; it seems your failed-boot cut off long before XHCI would actually have initialized or done anything. Possibly some other change in recent weeks fixed this, but not the USB ones.