From owner-freebsd-current Sat Aug 11 17:53:57 2001 Delivered-To: freebsd-current@freebsd.org Received: from zircon.seattle.wa.us (sense-sea-CovadSub-0-228.oz.net [216.39.147.228]) by hub.freebsd.org (Postfix) with SMTP id 9E90237B401 for ; Sat, 11 Aug 2001 17:53:53 -0700 (PDT) (envelope-from joe@zircon.seattle.wa.us) Received: (qmail 44092 invoked by uid 1001); 12 Aug 2001 00:57:02 -0000 From: Joe Kelsey MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-ID: <15221.54366.195608.585938@zircon.zircon.seattle.wa.us> Date: Sat, 11 Aug 2001 17:57:02 -0700 To: current@freebsd.org Subject: Re: FreeBSD's aggressive keyboard probe/attach In-Reply-To: <200108120026.f7C0Qvb03574@mass.dis.org> References: <3B75C7B1.FF2E739E@mindspring.com> <200108120026.f7C0Qvb03574@mass.dis.org> X-Mailer: VM 6.92 under Emacs 20.7.1 Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG Mike Smith writes: > > > The Belkin switch is hardly the "best" out there; it's also the > > > only KVM switch that I'm aware of that had problems with FreeBSD, > > > and even then only because their QC is so poor that they managed > > > to ship the product with the ROMs _reversed_. > > > > > > Not that it needs stating, but I've used FreeBSD with a wide > > > range of KVM switches over the years; I'm writing this through an > > > 8-port Cybex, for example. > > > > This switch emulates the keyboard controller to the point that > > it looks as if a keyboard is present at all times to the connected > > system, even if it's not. > > "This" switch being what? > > Note that I have a selection of rotary-switch KVMs here; they all work > with FreeBSD. They sure as shit don't emulate a keyboard controller. Then you obviously know absolutely nothing about the debate. The debate is over the modern, electronic KVM's which misbehave *massively* under FreeBSD. Obviously, a purely mechanical switch cannot misbehave in the way specified--for one thing, you need to take special action to allow reboot on the machines which are not actively connected to the keyboard, otherwise they will stop in the BIOS... > > People who run *Free*BSD are going to run it on cheaper equipment; > > evidence the amount of effort that has gone into ATA/IDE support. > > People who run FreeBSD don't actually typically complain about KVMs, > interestingly enough. "People" don't complain about KVM's on FreeBSD because the developers are obviously interested in doing things exactly one way and not interested in hearing about misbehaviors. The keyboard/mouse code is quite opaque and needs massive work, but no one seems interested in doing the work. I really like FreeBSD, but some of the "gotchas" are quite annoying. The KVM switch that I *used* to use was a real pain-in-the-ass. I could not find the magic incantation to stop the "psmintr" errors which occurred every time you switch systems. Finally, I read the source and discovered the error location and the flag that disables it. Then, I had to figure out how to set the flag, again a task that is not spelled out anywhere in any documentation. Finally, I settled on boot -c, go to visual mode, set the flag. After further experimentation, I finally figured out how to use loader.conf to accomplish the same thing. However, it is not at all obvious that this is the correct way to do it. There are absolutely no examples of setting driver flags in any man page or sample loader.conf script. How is someone who cannot read the source to know how to do this? Why do I keep needing to do it anyway, when Solaris has no problem dealing with both the mouse and the switch? Why does FreeBAD cause so much trouble that other OS's seem to be able to deal with? Even now that I have decomissioned the old Solaris x86 box, I still get the psmintr errors on a standard Logitech wireless mouse. Why? Should I be stepping thorugh the driver code to figure it out? surely others have seen the errors and can deal with them. Why is the stupid error message there anyway if it makes not difference at all? Setting the flag makes the mouse work, so obviously the error message is what causes the misbehavior, rather than anything else. If the driver works with the flag set in all cases, why is the flag even an option? /Joe To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message