Re: com(4) dialout device behavior change?

["John D. Baker" <jdbaker%consolidated.net@localhost>]

On Sun, 14 Apr 2019, John D. Baker wrote:

> As a sanity check, I booted -current GENERIC rather than the machine's
> custom kernel and the serial port works fine (although I'm still using
> the rewired adapter with RTS/CTS looped).  Rebooting with the custom
> kernel and it sends (because CTS is asserted), but doesn't receive.
> 
> So it appears my kernel paring has elided something that is now
> necessary.

Turns out there's something wrong with the:

  com* at acpi?

attachment.

This is normally commented out in GENERIC and the

  comN at isa?

attachments are used.

Since I have systems that reported UAR resources at ACPI, my habit has
been to enable the "com* at acpi?" attachment and disable the "comN at isa?"
attachments.

I reverted these in my custom kernel config and my serial ports work
normally again (still need to revert the RTS/CTS loop in the adapter
to know for sure).

So one of the ACPI subsystem updates since netbsd-8 was branched may
have brought this on.  Will have to check the logs and bisect.

-- 
|/"\ John D. Baker, KN5UKS               NetBSD     Darwin/MacOS X
|\ / jdbaker[snail]consolidated[flyspeck]net  OpenBSD            FreeBSD
| X  No HTML/proprietary data in email.   BSD just sits there and works!
|/ \ GPGkeyID:  D703 4A7E 479F 63F8 D3F4  BD99 9572 8F23 E4AD 1645