I have a normal amd64 (2 CPUs, 8G) running -current, from sources updated 21 February around 1805Z. The only odd thing is a thinkpad-style external USB keyboard, with IBM branding -- imagine a regular USB keyboard but physically a Thinkpad one with the eraser/pointer and a trackpad, plus an IBM special key and volume control buttons. It has two USB ports. All the hardware, including the keyboard, seems to be ok under GENERIC. The keyboard works, both to control booting in /boot and when NetBSD is up. When I boot xen without the keyboard plugged in, it boots up fine, complaining that xendomains isn't set in rc.conf (true) because I have no domUs. With the keyboard plugged in, it drops into ddb and hits so many subsequent faults even taking a video can't sort it out. I can't backtrace because -- the keyboad is apparently not really attached. If I boot w/o the keyboard and then plug it it (so that ddb.onpanic is no longer set), I get some sort of fault/panic and it claims to be dumping to disk, but I couldn't get a crash dump saved. I think I saw a line about handing over full speed device to uhci from ehci, but it was too fast to really read. The dmesg lines from GENERIC for booting with the keyboard are (not sure if this handover is about the keyboard, but I think so): [ 2.826758] ehci0: handing over full speed device on port 3 to uhci1 [ 5.081785] uhub9 at uhub5 port 1: Lite-On Tech (0x04b3) USB 1.1 2 port downstream low-power hub (0x3016), class 9/0, rev 1.10/2.01, addr 2 [ 5.101755] uhub9: 4 ports with 2 removable, bus powered [ 6.490356] uhidev0 at uhub9 port 3 configuration 1 interface 0 [ 6.490356] uhidev0: Lite-On Tech (0x04b3) IBM USB Keyboard with UltraNav (0x3018), rev 1.10/1.16, addr 3, iclass 3/1 [ 6.500337] ukbd0 at uhidev0: 8 Variable keys, 6 Array codes [ 6.909947] wskbd0 at ukbd0: console keyboard, using wsdisplay0 [ 6.909947] uhidev1 at uhub9 port 3 configuration 1 interface 1 [ 6.909947] uhidev1: Lite-On Tech (0x04b3) IBM USB Keyboard iwith UltraNav (0x3018), rev 1.10/1.16, addr 3, iclass 3/0 [ 6.929924] uhidev1: 3 report ids [ 6.929924] uhid0 at uhidev1 reportid 2: input=1, output=0, feature=0 [ 6.929924] uhid1 at uhidev1 reportid 3: input=3, output=0, feature=0 [ 7.968881] uhidev2 at uhub9 port 4 configuration 1 interface 0 [ 7.968881] uhidev2: Synaptics Inc. (0x06cb) Composite TouchPad / TrackPoint (0x0009), rev 1.10/0.20, addr 4, iclass 3/1 [ 7.978858] ums0 at uhidev2: 3 buttons [ 7.978858] wsmouse0 at ums0 mux 0 [ 7.978858] uhidev3 at uhub9 port 4 configuration 1 interface 1 [ 7.978858] uhidev3: Synaptics Inc. (0x06cb) Composite TouchPad / TrackPoint (0x0009), rev 1.10/0.20, addr 4, iclass 3/1 [ 7.998826] ums1 at uhidev3: 3 buttons [ 7.998826] wsmouse1 at ums1 mux 0 so it seems something is messed up under Xen. There seems to be slightly too many devices, but I think uhid0/uhid1 are the extra buttons. I wonder if there is some SMP issue not yet sorted. I'll work on a real backtrace, but clues welcome.
Attachment:
signature.asc
Description: PGP signature