> Date: Sun, 31 Mar 2024 10:47:53 +0200 > From: Landry Breuil <lan...@openbsd.org> > > Le Sun, Mar 31, 2024 at 09:30:05AM +0200, Landry Breuil a écrit : > > hi, > > > > istr this has been discussed/fixed at some point and it used to work > > last year, but the t495s i have here on -current hangs at ttyflags -a in > > /etc/rc, commenting it again allows boot to succeed. > > > > dmesg attached with -current. i dont boot that machine often enough, so > > the regression window is .. large.. guess i'll try bisecting. > > > > last known working: #1463: Wed Nov 22 21:13:03 MST 2023. > > after bisecting a bit, i'm puzzled because it seems ttyflags -a hangs > only happen when a spurious com0 is found in dmesg: > > com0 at isa0 port 0x3f8/8 irq 4: ns16550a, 16 byte fifo > com0: probed fifo depth: 0 bytes > > but that device isnt present in the working boots from various kernel > versions (tried kernels from end of december to 1 feb so far) > > it's enough to test with boot -s and ttyflags -a, i think i triggered it > once with a kernel from #1587: Sat Dec 30 22:44:51 MST 2023, next boots > on the same kernel were okay.. > > I've tried differentiating cold boots vs reboots, but that didn't help.
Bleah. Those ISA-style probes are not very sophisticated and if vendors are stupid enough to put something at the same address as the legacy COM ports, it may be detected as a phantom port. Maybe we should stop doing the ISA probes on systems where ACPI tells us there are no legacy devices. Although I'm not sure if that would help your system. Can you send me the files from /var/db/acpi?
