Date: Fri, 15 Apr 2011 12:41:10 -0400 From: John Baldwin <jhb@freebsd.org> To: freebsd-stable@freebsd.org Cc: Pawel Tyll <ptyll@nitronet.pl>, Przemyslaw Frasunek <przemyslaw@frasunek.com>, Jeremy Chadwick <freebsd@jdc.parodius.com> Subject: Re: Missing serial port after enabling serial console in loader.conf Message-ID: <201104151241.10291.jhb@freebsd.org> In-Reply-To: <20110412230822.GA33610@icarus.home.lan> References: <4DA4A96F.9000507@frasunek.com> <1474571124.20110412231306@nitronet.pl> <20110412230822.GA33610@icarus.home.lan>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tuesday, April 12, 2011 7:08:22 pm Jeremy Chadwick wrote: > On Tue, Apr 12, 2011 at 11:13:06PM +0200, Pawel Tyll wrote: > > >> P.S. -- What's "BMC" stand for? > > > Baseband Management Controller. It's the separate NIC (sometimes it's > > > piggy-backed on a NIC) used for remote management of motherboards. > > > Usually includes IPMI support. Depending on the motherboard, it may > > > even include support for keyboard/video/mouse redirection and > > > serial-over-lan and all that other remote management stuff. > > > > Actually, the board in reference has built-in BMC; it's IPMI managed > > and when properly configured it allows user to ssh to its IP and > > access things like SOL, chassis power control and other IPMI things > > without any IPMI clients. As to video redirection, there's addon for > > that, built-in BMC doesn't have anything that fancy. And on point, I > > was having similar problems with these boards some time ago, but it > > seems it has been fixed somewhere between 8.1 and 8.2. > > If I remember correctly from following commits, there were changes to > the boot loader pertaining to serial port initialisation. I think I > remember something about certain control bits being set/reset when they > should have been masked out (left alone). > > As I understand it, the serial console capabilities on FreeBSD are > tested predominantly with motherboards that have on-board serial ports. > That is to say, systems with classic/legacy DB9 ports on the board > itself and not via IPMI or out-of-band management modules (re: BMC). > So, one should not be too surprised that there may be "oddities" seen > with those. > > I imagine that UART enumeration/initialisation might differ on such > modules, so this sort of problem doesn't surprise me much. Marcel > Moolenaar (author of uart(4)) probably has better insights. It may be the issue with ACPI enumerating the ports backwards which 8 handles by matching hinted devices based on resources. That has nothing to do with either sio(4) or uart(4) if so. The BMC port usually uses the resources from COM2 by default which our default hints assign to sio1/uart1. -- John Baldwin
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201104151241.10291.jhb>