From owner-freebsd-hackers Fri Oct 10 14:44:49 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id OAA12191 for hackers-outgoing; Fri, 10 Oct 1997 14:44:49 -0700 (PDT) (envelope-from owner-freebsd-hackers) Received: from usr08.primenet.com (tlambert@usr08.primenet.com [206.165.6.208]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id OAA12186; Fri, 10 Oct 1997 14:44:44 -0700 (PDT) (envelope-from tlambert@usr08.primenet.com) Received: (from tlambert@localhost) by usr08.primenet.com (8.8.5/8.8.5) id OAA14799; Fri, 10 Oct 1997 14:44:35 -0700 (MST) From: Terry Lambert Message-Id: <199710102144.OAA14799@usr08.primenet.com> Subject: Re: LINUX emulation and uname(3). To: richard@pegasus.com (Richard Foulk) Date: Fri, 10 Oct 1997 21:44:34 +0000 (GMT) Cc: freebsd-emulation@freefall.FreeBSD.org, freebsd-hackers@freefall.FreeBSD.org In-Reply-To: <199710101754.HAA17182@pegasus.com> from "Richard Foulk" at Oct 10, 97 07:54:17 am X-Mailer: ELM [version 2.4 PL23] Content-Type: text Sender: owner-freebsd-hackers@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk > This is absurd. > > When you're emulating another OS you want to get as close as possible > to acting just like that OS. Period. Agreed. > The goal is to run software. If the emulation is incomplete then > chances are some things won't run -- which means the emulator fails it's > primary task. So, when do the patches for vgalib, user LDT, and Linux vm86() call support arrive? There are many more programs that care about these than uname(). Frankly, given the speed of the version rotor, I'd be very surprised if anything keyed on "release" or "version" at all... > If you find the `Linux' report to be that hideous then make the output > configurable. The value of "sysname" is exactly what was at issue here. The result that the company now produces FreeBSD specific license keys is a highly desirable outcome. I would even say "optimal", except I believe they should port their product to run natively on FreeBSD. Even so, now they are aware of FreeBSD, and we may see it mentioned in their ads. > But the default action should be as close to what Linux produces as possible. Fine. What's the exact version of Linux we are emulating? > If Linux software doesn't run, for any reason, then the emulator has failed. Agreed. But there are more serious failures than this. This was a license failure, not a product failure, in any case. As I said before, "sysname" is a silly thing to use for your license on an OS where its value can be changed at will (which is all of them). I much prefer floating instance licenses, in any case. Hmmm. Maybe I should make my loadable license spoofer for Solaris (which I wrote because it was easier than transferring the license) available? It vfork's and sends down the desired license data with the PID before exec'ing the licensed program. It handles both uname and hostid data. Any call with a spoofed PID (flag for PPID for child processes) gets the spoofed return. Pretty darn trivial "emulator failure"... Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.