Re: Advice on adding com2 to (amd64) GENERIC; enabling easier IPMI SOL with SuperMicro boards
On Sat, Apr 6, 2013 at 1:35 AM, Ted Unangst wrote: > I guess you missed the subsequent put back yesterday. :) Guilty as charged. > [...] com2 renumbers any other pci attached com ports from the likes of puc. I suppose for those running tools such as conserver, this would mean changing the config lines that carry the 'baseport' values. In case it's helpful, I've added the following snippet for faq/current.html to warn unsuspecting serial users. Regards, Rogier Index: current.html === RCS file: /cvs/www/faq/current.html,v retrieving revision 1.373 diff -u -r1.373 current.html --- current.html 28 Mar 2013 21:49:08 - 1.373 +++ current.html 6 Apr 2013 09:01:26 - @@ -54,6 +54,7 @@ 2013/03/10 - fontconfig update 2013/03/11 - pf translation counter added 2013/03/25 - Perl update +2013/04/05 - amd64 adds com2 and com3 to GENERIC @@ -562,6 +563,15 @@ of this being committed and rely on such packages, you might like to wait for updated packages to become available to save the trouble of building them yourself. + + + +2013/04/05 - amd64 adds com2 and com3 to GENERIC +OpenBSD/amd64 GENERIC and GENERIC.MP kernels now include the com2 (COM3) +and com3 (disabled by default) devices that were commented out before. This +may cause the renumbering of serial ports on other devices such as puc(4). +Users of the conserver port may want to use the 'portbase' and 'devicesubst' +settings to easily adjust their configuration.
Re: Advice on adding com2 to (amd64) GENERIC; enabling easier IPMI SOL with SuperMicro boards
On Sat, Apr 06, 2013 at 01:16, Rogier Krieger wrote: > Out of curiosity, after seeing the commit and subsequent backing out of > this change, what'd be the expected issues with enabling com2 that require > more thought? I guess you missed the subsequent put back yesterday. :) The main issue was that adding com2 renumbers any other pci attached com ports from the likes of puc. In the end, we decided that wasn't such a big deal and even com3 was added (in a disabled state, because the ioport issue is somewhat complicated).
Re: Advice on adding com2 to (amd64) GENERIC; enabling easier IPMI SOL with SuperMicro boards
Out of curiosity, after seeing the commit and subsequent backing out of this change, what'd be the expected issues with enabling com2 that require more thought? Regards, Rogier On Sat, Mar 30, 2013 at 8:01 AM, Ted Unangst wrote: > On Sat, Mar 30, 2013 at 02:06, Rogier Krieger wrote: > > > The GENERIC kernel config has commented out com2 (at isa0, addr 0x3e8, > > irq 5) and I assume this is not without reason. I've been unable to > > find that reason in source changes, but perhaps someone here knows. On > > i386, it is present. > > I am guessing this is an oversight. i386 runs on the same machines, so > if com2 were causing trouble it would be disabled there too. > -- If you don't know where you're going, any road will get you there.
Re: Advice on adding com2 to (amd64) GENERIC; enabling easier IPMI SOL with SuperMicro boards
On Sat, Mar 30, 2013 at 02:06, Rogier Krieger wrote: > The GENERIC kernel config has commented out com2 (at isa0, addr 0x3e8, > irq 5) and I assume this is not without reason. I've been unable to > find that reason in source changes, but perhaps someone here knows. On > i386, it is present. I am guessing this is an oversight. i386 runs on the same machines, so if com2 were causing trouble it would be disabled there too.
Advice on adding com2 to (amd64) GENERIC; enabling easier IPMI SOL with SuperMicro boards
Dear list, in an attempt to save on serial cabling for our machines, I'm trying to see if IPMI Serial over Lan (SOL) works as advertised. For our Dell boxes, things seem to work, but our SuperMicro boards (X7SPA-HF and X8ST3-F) require extra work. The latter seem to insist on using com2 (i.e. COM3 in BIOS), which isn't present by default in GENERIC[.MP]. Obviously, adding this creates a bit of hassle and the risk of the com2 device being unavailable should I ever forget to add it back after upgrades. The GENERIC kernel config has commented out com2 (at isa0, addr 0x3e8, irq 5) and I assume this is not without reason. I've been unable to find that reason in source changes, but perhaps someone here knows. On i386, it is present. In summary, would the following be acceptable? Regards, Rogier Index: GENERIC === RCS file: /cvs/src/sys/arch/amd64/conf/GENERIC,v retrieving revision 1.338 diff -u -r1.338 GENERIC --- GENERIC 15 Mar 2013 09:10:52 - 1.338 +++ GENERIC 30 Mar 2013 01:04:54 - @@ -315,7 +315,7 @@ com0 at isa? port 0x3f8 irq 4# standard PC serial ports com1 at isa? port 0x2f8 irq 3 -#com2 at isa? port 0x3e8 irq 5 +com2 at isa? port 0x3e8 irq 5 #com3 at isa? port 0x2e8 irq 9# (conflicts with some video cards) com* at pcmcia? # PCMCIA modems/serial ports