From vandys@glare.cisco.com Sun Aug 28 19:06:11 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.11.56]) by amri.cisco.com (8.3/8.3) with ESMTP id TAA02062; Sun, 28 Aug 1994 19:06:10 -0700 Received: from avalencia-gw.cisco.com (avalencia-gw.cisco.com [198.92.29.34]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id UAA06372 for vsta; Sun, 28 Aug 1994 20:10:43 -0700 Date: Sun, 28 Aug 1994 20:10:43 -0700 From: Andrew Valencia Message-Id: <199408290310.UAA06372@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host avalencia-gw.cisco.com didn't use HELO protocol Subject: VSTa: networking! Apparently-To: vsta@cisco.com This is being typed from my i486 running VSTa! It is KA9Q networking talking over a n ethernet driver provided by David Johnson. It's a little shaky as yet, but what a blast! Reagards... Andy From vandys@glare.cisco.com Sun Aug 28 19:32:33 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.11.56]) by amri.cisco.com (8.3/8.3) with ESMTP id TAA02074; Sun, 28 Aug 1994 19:32:31 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id UAA06623 for ; Sun, 28 Aug 1994 20:39:26 -0700 Message-Id: <199408290339.UAA06623@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: vsta@amri.cisco.com Subject: VSta networking, some more details Date: Sun, 28 Aug 1994 20:39:26 -0700 From: Andrew Valencia I've already had some requests for more details. The ethernet driver (for an ne2000) runs as a normal device driver task. It offers a directory to mount (under namer net/ne) which is initially empty. If you mount net/ne under, say, /dev/eth, and then open() /dev/eth/800, all IP packets (type 0x800) for your MAC address (or broadcast) will start arriving as read() completions. /dev/eth/0 receives everything. Because each client connection can only do one I/O operation at a time, KA9Q open()'s two connections; one for reading, another for writing. KA9Q software itself run as a user task. I started with the Linux port, but have hacked it extensively. The code no longer polls or does any spinning. Instead, I created a thread for each source of data/events (keyboard, LAN interface, timers, local TCP clients) and have the threads block for their own data. When they get something, they take a global semaphore which I added to the KA9Q code, queue their data, and then run the usual KA9Q processing loop. When finished, they release the semaphore and go back to waiting for data. This replaces the select() overhead, and also allows me to avoid the CPU drainage associated with all the timeout-polling which was done previously. It seems to work OK in practice, but the code's still shaky, so it's a little hard to tell! I need to get the local TCP client support running, so an inetd can start watching for incoming network stuff. This will be a VSTa server, which will be mountable as a filesystem. The directory will contain TCP port numbers, as well as a clone entry for just picking up a free port number. A telnetd is the obvious service to bring up first, after which I can finally start connecting to VSTa directly from my Xterminal. After telnet, I will bring up a remote filesystem service. This would be a general service for running VSTa messages over a remote connection. Using this, VSTa contributors can then mount the central RCS repository and check their work into the source tree directly. So... telnet, remote filesystem, and then V1.4 will finally be out! Regards, Andy From vandys@glare.cisco.com Mon Aug 29 16:34:43 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id QAA00066; Mon, 29 Aug 1994 16:34:41 -0700 Received: from wariat.org (wariat.org [192.147.147.1]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id RAA14252 for ; Mon, 29 Aug 1994 17:42:52 -0700 Received: from fmsystm by wariat.org with uucp(/\==/\ Smail3.1.28.1 #28.5) id ; Mon, 29 Aug 94 20:41 EDT for cisco.com!vsta Received: from nightfly by fms.com with uucp (Smail3.1.28.1 #1) id m0qfHEg-000116C; Mon, 29 Aug 94 20:39 EDT Received: by nightfly.telemax.com (4.1/SMI-4.1) id AA00482; Mon, 29 Aug 94 18:56:55 EDT From: matt@nightfly.telemax.com (Matt Emerson) Message-Id: <9408292256.AA00482@nightfly.telemax.com> Subject: Re: SPARC port To: mmceniry@uahcs2.cs.uah.edu (Michael McEniry) Date: Mon, 29 Aug 1994 18:56:54 -0400 (EDT) Cc: vsta@cisco.com In-Reply-To: <199408251440.JAA06264@uahcs2.cs.uah.edu> from "Michael McEniry" at Aug 25, 94 09:40:44 am Content-Type: text Content-Length: 329 > What's the status on the port to SPARC platforms? I might > be getting a (cheap/used/old *sigh*) IPC in the very near > future and would like to play--er, do useful and productive > things with VSTa on it. well, i've started working on it, but it will be some while before you can do anything "useful and productive." -matt From vandys@glare.cisco.com Tue Aug 30 00:00:27 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id AAA00087; Tue, 30 Aug 1994 00:00:26 -0700 Received: from relay1.Hawaii.Edu (relay1.Hawaii.Edu [128.171.41.53]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id BAA28133 for ; Tue, 30 Aug 1994 01:08:33 -0700 Received: from uhunix.uhcc.Hawaii.Edu ([128.171.44.6]) by relay1.Hawaii.Edu with SMTP id <11362>; Mon, 29 Aug 1994 22:08:22 -1000 Received: by uhunix.uhcc.Hawaii.Edu id <184427>; Mon, 29 Aug 1994 22:08:12 -1000 Subject: libc/printf.c From: Tim Newsham To: vsta@cisco.com Date: Mon, 29 Aug 1994 22:08:06 -1000 X-Mailer: ELM [version 2.3 PL11] Message-Id: <94Aug29.220812hst.184427@uhunix.uhcc.Hawaii.Edu> been reading through some code and saw somethng that made my stomach turn: static __fprintf(FILE *fp, char *fmt, int *argptr) { char buf[BUFSIZ], *p, c; __doprnt(buf, fmt, argptr); p = buf; while (c = *p++) { putc(c, fp); } return(0); } this is called by fprintf() as well as printf(). fixed sized buffer used on the stack. Could be the source of many core dumps and security violations (ie. old fingerd bug). Tim N. From vandys@glare.cisco.com Tue Aug 30 03:06:14 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id DAA00122; Tue, 30 Aug 1994 03:06:12 -0700 Received: from relay1.Hawaii.Edu (relay1.Hawaii.Edu [128.171.41.53]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id EAA07907 for ; Tue, 30 Aug 1994 04:14:26 -0700 Received: from uhunix.uhcc.Hawaii.Edu ([128.171.44.6]) by relay1.Hawaii.Edu with SMTP id <11411>; Tue, 30 Aug 1994 01:14:15 -1000 Received: by uhunix.uhcc.Hawaii.Edu id <184430>; Tue, 30 Aug 1994 01:09:27 -1000 Subject: scanf.c From: Tim Newsham To: vsta@cisco.com Date: Tue, 30 Aug 1994 01:09:23 -1000 X-Mailer: ELM [version 2.3 PL11] Message-Id: <94Aug30.010927hst.184430@uhunix.uhcc.Hawaii.Edu> in scanf.c: bug? switch((scale<<4) | size) { [...] case (INT<<3) | LONG: **(long **)ptr = lcval; break; all the other cases were << 4 like the switch(). Tim N. From vandys@glare.cisco.com Tue Aug 30 03:28:48 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id DAA00126; Tue, 30 Aug 1994 03:28:47 -0700 Received: from relay1.Hawaii.Edu (relay1.Hawaii.Edu [128.171.41.53]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id EAA08392 for ; Tue, 30 Aug 1994 04:37:01 -0700 Received: from uhunix.uhcc.Hawaii.Edu ([128.171.44.6]) by relay1.Hawaii.Edu with SMTP id <11344>; Tue, 30 Aug 1994 01:36:55 -1000 Received: by uhunix.uhcc.Hawaii.Edu id <184427>; Tue, 30 Aug 1994 01:36:48 -1000 Subject: signal handling From: Tim Newsham To: vsta@cisco.com Date: Tue, 30 Aug 1994 01:36:36 -1000 X-Mailer: ELM [version 2.3 PL11] Message-Id: <94Aug30.013648hst.184427@uhunix.uhcc.Hawaii.Edu> yet another post... maybe I should have wrote the questions down and asked them all at once.... It doesnt appear the signal() library call does anything at the present time. Is this indeed the case? Tim N. From vandys@glare.cisco.com Tue Aug 30 06:20:22 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.11.56]) by amri.cisco.com (8.3/8.3) with ESMTP id GAA00158; Tue, 30 Aug 1994 06:20:21 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id HAA07245; Tue, 30 Aug 1994 07:28:35 -0700 Message-Id: <199408301428.HAA07245@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: Tim Newsham Cc: vsta@cisco.com Subject: Re: libc/printf.c In-Reply-To: Your message of "Mon, 29 Aug 1994 22:08:06 -1000." <94Aug29.220812hst.184427@uhunix.uhcc.Hawaii.Edu> Date: Tue, 30 Aug 1994 07:28:34 -0700 From: Andrew Valencia [Tim Newsham writes:] >static >__fprintf(FILE *fp, char *fmt, int *argptr) > char buf[BUFSIZ], *p, c; >... >this is called by fprintf() as well as printf(). fixed sized >buffer used on the stack. Could be the source of many core dumps >and security violations (ie. old fingerd bug). Well, now that we have a telnet server I guess this matters more. I'm open to patches; the obvious one just caps _doscan() at a limit. > switch((scale<<4) | size) { >[...] > case (INT<<3) | LONG: >all the other cases were << 4 like the switch(). This is code imported from BSD, and the current scanf() code is very different. It sure looks wrong to me, and yet it seems to work. I just traced through the code, and the assignment case for the longword *is* reached. >yet another post... maybe I should have wrote the questions down >and asked them all at once.... Fixed. :-) >It doesnt appear the signal() library call does anything >at the present time. Is this indeed the case? Yup. In bringing up networking I fixed the event handling code, so now we have a basis for doing the rest of signal handling. I'd expect this to follow after v1.4, unless somebody else wants to jump in. Thanks, Andy From vandys@glare.cisco.com Tue Aug 30 08:15:33 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id IAA00177; Tue, 30 Aug 1994 08:15:31 -0700 Received: from post.demon.co.uk (post.demon.co.uk [158.152.1.72]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id JAA19738; Tue, 30 Aug 1994 09:23:43 -0700 Received: from humbug.demon.co.uk by post.demon.co.uk id aa02010; 30 Aug 94 17:00 GMT-60:00 Received: by humbug.demon.co.uk (Linux Smail3.1.28.1 #1) id m0qfVak-000384C; Tue, 30 Aug 94 16:59 BST Message-Id: From: Dave Hudson Subject: Re: libc/printf.c To: Andy Valencia Date: Tue, 30 Aug 1994 16:59:22 +0100 (BST) Cc: VSTa mailing list X-Mailer: ELM [version 2.4 PL23] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 1103 Andrew Valencia wrote: > > [Tim Newsham writes:] > > >static > >__fprintf(FILE *fp, char *fmt, int *argptr) > > char buf[BUFSIZ], *p, c; > >... > >this is called by fprintf() as well as printf(). fixed sized > >buffer used on the stack. Could be the source of many core dumps > >and security violations (ie. old fingerd bug). > > Well, now that we have a telnet server I guess this matters more. I'm open > to patches; the obvious one just caps _doscan() at a limit. I'll take a look at this one - I've already done a lot of changes to add all of the vprintf() family. > > switch((scale<<4) | size) { > >[...] > > case (INT<<3) | LONG: > >all the other cases were << 4 like the switch(). > > This is code imported from BSD, and the current scanf() code is very > different. It sure looks wrong to me, and yet it seems to work. I just > traced through the code, and the assignment case for the longword *is* > reached. If you look at the definition INT is 0 anyway - I guess this is a bug, but 0 << 3 is still 0 :-) Regards, Dave From vandys@glare.cisco.com Thu Sep 1 00:26:46 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id AAA00444; Thu, 1 Sep 1994 00:26:45 -0700 Received: from vespucci.iquest.com (root@vespucci.iquest.com [199.170.120.42]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id BAA21591 for ; Thu, 1 Sep 1994 01:35:08 -0700 Received: by vespucci.iquest.com (Linux Smail3.1.28.1 #14) id m0qg6BR-0003CJC; Thu, 1 Sep 94 02:03 CDT Message-Id: From: chris@iquest.com (Happy Guy) Subject: First glances. To: vsta@cisco.com Date: Thu, 1 Sep 1994 02:03:40 -0500 (CDT) X-Mailer: ELM [version 2.4 PL23] Content-Type: text Content-Length: 1119 Well after much searching I finally decided to put a new OS on my home pc. After looking at lots of them and asking lots of people I finally (with help) decided on VSTa. Admittedly I didn't know much about it when I installed and I don't know a great deal now. It was a breeze to get and install and after fixing a stupid error on my part (put everything in vista and not vsta which made for a nice dying boot process) it ran smoothly. A little exploring here and there and after some more help finally did a make in libc. I like what I have seen so far just a few points and questions. 1) Documentation. Uhm well need I say more? If I new more about what everything was/did I would be willing to donate some time to helping flesh this out. 2) What compiles/doesn't? Is there a list? How hard is it to port a small utility? Maybe a small faq on what need to be done could be "whipped" up. 3) How far off are certain projects? net server? MADO (which I would love to hear more about)? Will slip or ppp come with the net server? Am I just being an impatient little brat? Regardless, keep up the excellent work. chris From vandys@glare.cisco.com Thu Sep 1 06:01:55 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.11.56]) by amri.cisco.com (8.3/8.3) with ESMTP id GAA00614; Thu, 1 Sep 1994 06:01:54 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id HAA13811; Thu, 1 Sep 1994 07:10:20 -0700 Message-Id: <199409011410.HAA13811@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: chris@iquest.com (Happy Guy) Cc: vsta@cisco.com Subject: Re: First glances. In-Reply-To: Your message of "Thu, 01 Sep 1994 02:03:40 CDT." Date: Thu, 01 Sep 1994 07:10:20 -0700 From: Andrew Valencia [chris@iquest.com (Happy Guy) writes:] >1) Documentation. Uhm well need I say more? If I knew more about what >everything was/did I would be willing to donate some time to helping flesh >this out. Well, I did some man pages for the messaging engine. I agree that there's LOTS of room for improvement! >2) What compiles/doesn't? Is there a list? How hard is it to port a small >utility? Maybe a small faq on what need to be done could be "whipped" up. Ports are becoming increasingly easy, as missing pieces are filled in. At this point I'd say that small utilities come across with no fuss, and things the size of "vim" come across with a little work. The biggest problem I had with "vim" (the VI eMulator, BTW) was that they used some memory after they had free()'ed it. RCS has been a real headache. They use all sorts of whacky fringe stuff. Parts work, but each piece of functionality comes only after a fight. >3) How far off are certain projects? net server? MADO (which I would love to >hear more about)? Will slip or ppp come with the net server? I'll leave the MADO status to Gavin and/or Dave. Networking via LAN is up (he types from VSTa). It looks like the Linux port of KA9Q I started from doesn't have PPP, but it does have SLIP and I've already converted the code to full POSIX (). I'm currently working on a /net/tcp filesystem to offer KA9Q's services to other VSTa processes. When I have telnet's into VSTa running I'll perhaps bring up SLIP before doing remote filesystem support. Or would you like to do this part? Welcome to VSTa! Andy From vandys@glare.cisco.com Thu Sep 1 07:07:54 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id HAA00625; Thu, 1 Sep 1994 07:07:53 -0700 Received: from ebt.com (spock.ebt.com [192.111.115.3]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id IAA08228 for ; Thu, 1 Sep 1994 08:16:19 -0700 Received: from ebt-inc.ebt.com by ebt.com (4.1/SMI-4.1) id AA07006; Thu, 1 Sep 94 11:17:14 EDT Received: by ebt-inc.ebt.com (5.0/SMI-SVR4) id AA29043; Thu, 1 Sep 1994 11:17:14 +0500 Date: Thu, 1 Sep 1994 11:17:14 +0500 From: gtn@ebt.com (Gavin Nicol) Message-Id: <9409011517.AA29043@ebt-inc.ebt.com> To: vsta@cisco.com Subject: Re: First glances. Content-Length: 324 Well, MADO is coming along slowly. Dave is working on BitBlt, so we can expect his usual miracles :-) I will be making a presentation later this year comparing *BSD, Linux, VSTa, and the Hurd. I *really* want to have MADO up by then. Once I get my current (work) project out of the way. I'll swing into full MADO hacking! From vandys@glare.cisco.com Thu Sep 1 20:45:29 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id UAA00686; Thu, 1 Sep 1994 20:45:27 -0700 Received: from relay1.Hawaii.Edu (relay1.Hawaii.Edu [128.171.41.53]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id VAA10333 for ; Thu, 1 Sep 1994 21:53:57 -0700 Received: from uhunix.uhcc.Hawaii.Edu ([128.171.44.6]) by relay1.Hawaii.Edu with SMTP id <11400>; Thu, 1 Sep 1994 18:53:55 -1000 Received: by uhunix.uhcc.Hawaii.Edu id <184408>; Thu, 1 Sep 1994 18:53:29 -1000 Subject: questions.. From: Tim Newsham To: vsta@cisco.com Date: Thu, 1 Sep 1994 18:53:20 -1000 X-Mailer: ELM [version 2.3 PL11] Message-Id: <94Sep1.185329hst.184408@uhunix.uhcc.Hawaii.Edu> Ok.. in mach/init.c it sets up the page tables. The level1 stuff is: text (0), data (4megs), cr3 (8megs... what is this for?) util (12 megs), free (16 megs) 1:1 mapping of memory .. more free ... utext (512 * 4megs) ustack (1023 * 4megs) is this all correct? Please point out any errors. Ok. The big question is... data is moving between when main() starts up and when the mmu is turned on. Before turning on the MMU data is the page following text. After turning on the MMU it is at 4megs. What exactly does GCC put into the data segment? Is this data just avoided up until the MMU is started? smaller questions.. What is the 8-12meg mapping for? It points back to the top level (level 1) table. What effect does this have? The first page of text is invalid according to the MMU. It appears that text starts at 4K rather than 0K. Is this correct? Does the linker usually load to this address or is this requested specially? Tim N. From vandys@glare.cisco.com Fri Sep 2 05:37:19 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id FAA00747; Fri, 2 Sep 1994 05:37:18 -0700 Received: from post.demon.co.uk (post.demon.co.uk [158.152.1.72]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id GAA28800 for ; Fri, 2 Sep 1994 06:45:47 -0700 Received: from humbug.demon.co.uk by post.demon.co.uk id aa14262; 2 Sep 94 13:59 GMT-60:00 Received: by humbug.demon.co.uk (Linux Smail3.1.28.1 #1) id m0qgW3F-00039IC; Fri, 2 Sep 94 11:40 BST Message-Id: From: Dave Hudson Subject: Hardware/performance comments To: VSTa mailing list Date: Fri, 2 Sep 1994 11:40:56 +0100 (BST) X-Mailer: ELM [version 2.4 PL23] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 1667 Hi All, I've recently been running some cache tests under VSTa on a range of PCs and thought that the results might be interesting. I've been running a compilation job, with a couple of recursive directory listings going on in the background. To make the tests fair I've been using the same hard drive and setting up configurations that are as close as possible (all DX/2-66, 256k second level write back cache, Cirrus Logic 5428 video, same Seagate HDD). In case anyone had any doubts about how hard VSTa works the CPU and memory subsystems I recently found a fault in a 2nd level cache (lockups within 30 secs to 5mins of starting the tests) that was not evident under DOS, Windows or OS/2. (The fault was actually verified when I looked at some signal timings). What I really wanted to see was how much performance was gained by second level caching, so I set up some tests running various combinations of CPU and 2nd level caches. The relative performance benefits on 3 types of PC were roughly the same. The results showed that a DX/2 with its CPU cache disabled hardly gains anything from a second level cache (about 5%). The CPU (1st level) cache gives between 400% and 600% over none at all (about what I'd expected). The one that surprised me was that VSTa gains 30% from having the second level and CPU caches enabled over just the CPU cache. This is much higher than the figures I've seen for DOS and Windows (between 15 and 20%). I guess this 2nd level caching benefit is mainly down to the nature of the task switching and memory management in VSTa. Has anyone seen any figures for the same sort of thing on other OS's? Regards, Dave From vandys@glare.cisco.com Fri Sep 2 07:35:43 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.11.56]) by amri.cisco.com (8.3/8.3) with ESMTP id HAA00757; Fri, 2 Sep 1994 07:35:42 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id IAA04295; Fri, 2 Sep 1994 08:44:14 -0700 Message-Id: <199409021544.IAA04295@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: Tim Newsham Cc: vsta@cisco.com Subject: Re: questions.. In-Reply-To: Your message of "Thu, 01 Sep 1994 18:53:20 -1000." <94Sep1.185329hst.184408@uhunix.uhcc.Hawaii.Edu> Date: Fri, 02 Sep 1994 08:44:14 -0700 From: Andrew Valencia [Tim Newsham writes:] >Ok. The big question is... data is moving between when main() >starts up and when the mmu is turned on. Before turning on the >MMU data is the page following text. After turning on the MMU >it is at 4megs. What exactly does GCC put into the data segment? >Is this data just avoided up until the MMU is started? boot.exe running from DOS sets up page tables and enables the MMU (including paging) before it jumps into VSTa's startup code. So data accesses are fine from the initial instruction in the VSTa kernel. See vsta/boot/ptes.c, especially the treatment of the first (l1[0]) and second (l1[1]) entries in the first level PTEs. These page tables (and, in fact, the GDT itself) are *replaced* by VSTa with ones it creates. But the mapping puts data at 4 meg, and VSTa leaves it there when it replaces the page tables with its own. >smaller questions.. What is the 8-12meg mapping for? It points >back to the top level (level 1) table. What effect does this >have? It makes it easy to reference your level 1 PTEs, since it auto-magically creates a virtual address for all of the first level PTEs (which, note, take up a max of 4 megs). See how vtop() works, especially the use of l2ptmap. >The first page of text is invalid according to the MMU. It appears >that text starts at 4K rather than 0K. Is this correct? Does >the linker usually load to this address or is this requested >specially? The linker does this, it was like this for the DOS djgpp C compiler, and I left it this way since it catches NULL pointer dereferences. Regards, Andy From vandys@glare.cisco.com Tue Sep 6 01:51:44 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id BAA01251; Tue, 6 Sep 1994 01:51:42 -0700 Received: from dolly.par.univie.ac.at (dolly.par.univie.ac.at [131.130.70.42]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with ESMTP id DAA05023 for ; Tue, 6 Sep 1994 03:00:08 -0700 Received: from tracy.par.univie.ac.at (tracy [131.130.70.13]) by dolly.par.univie.ac.at (8.6.9/8.6.9) with ESMTP id LAA08744 for ; Tue, 6 Sep 1994 11:59:15 +0200 From: Robert Mayer - Student Received: (robert@localhost) by tracy.par.univie.ac.at (8.6.9/8.6.9) id LAA10754 for vsta@cisco.com; Tue, 6 Sep 1994 11:59:46 +0200 Date: Tue, 6 Sep 1994 11:59:46 +0200 Message-Id: <199409060959.LAA10754@tracy.par.univie.ac.at> To: vsta@cisco.com Subject: file attributes Hi all, yesterday I thought of something: OS/2 has these nice "Extended Attributes", where one can attach a string=value pair to a file. Wouldn't it be nice to have such a thing under VSTa too? The first thing that came to my mind was to extend VSTa's rstat/wstat mechanism so that the valid stat-fields are not limited to these defined at server-compile-time, and one could add new ones by simply wstat()ing them. On the other hand, one could also extend the message protocol by e.g. FS_GETATTR/FS_SETATTR, and handle attributes and stat-fields seperately. Thoughts anyone? Regards, Robert. From vandys@glare.cisco.com Tue Sep 6 03:44:57 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id DAA01276; Tue, 6 Sep 1994 03:44:54 -0700 Received: from post.demon.co.uk (post.demon.co.uk [158.152.1.72]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id EAA13096 for ; Tue, 6 Sep 1994 04:53:46 -0700 Received: from humbug.demon.co.uk by post.demon.co.uk id aa14662; 6 Sep 94 12:17 GMT-60:00 Received: by humbug.demon.co.uk (Linux Smail3.1.28.1 #1) id m0qhyYG-00038lC; Tue, 6 Sep 94 12:19 BST Message-Id: From: Dave Hudson Subject: Re: file attributes To: Robert Mayer - Student Date: Tue, 6 Sep 1994 12:18:59 +0100 (BST) Cc: VSTa mailing list In-Reply-To: <199409060959.LAA10754@tracy.par.univie.ac.at> from "Robert Mayer - Student" at Sep 6, 94 11:59:46 am X-Mailer: ELM [version 2.4 PL23] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 1329 Robert Mayer - Student wrote: > > Hi all, > yesterday I thought of something: > OS/2 has these nice "Extended Attributes", where one can attach a > string=value pair to a file. Wouldn't it be nice to have such a thing > under VSTa too? > The first thing that came to my mind was to extend VSTa's rstat/wstat > mechanism so that the valid stat-fields are not limited to these > defined at server-compile-time, and one could add new ones by simply > wstat()ing them. > On the other hand, one could also extend the message protocol by e.g. > FS_GETATTR/FS_SETATTR, and handle attributes and stat-fields > seperately. > Thoughts anyone? I discussed something similar with Andy a while back, but decided in the end that extending the message format wasn't the best way of doing this sort of thing. The problem I hit when I thought up my extension was that the decision about whether a field was accessed via one mechanism or the other became too abitrary. As an example an extended attribute on say an HPFS filesystem might be the file owner, but on a Linux ext2fs it's there as a standard attribute. It wouldn't seem to make much sense (and it would certainly confuse ls) to have potentially two ways to get the same information. I like the idea of allowing dynamic additions to the stat fields though :-) Regards, Dave From vandys@glare.cisco.com Wed Sep 14 06:36:11 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id GAA00382; Wed, 14 Sep 1994 06:36:09 -0700 Received: from post.demon.co.uk (post.demon.co.uk [158.152.1.72]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id HAA05047 for ; Wed, 14 Sep 1994 07:46:59 -0700 Received: by post.demon.co.uk id al26374; 14 Sep 94 15:37 GMT-60:00 Received: from humbug.demon.co.uk by post.demon.co.uk id aa20558; 14 Sep 94 15:20 GMT-60:00 Received: by humbug.demon.co.uk (Linux Smail3.1.28.1 #1) id m0qkv8G-000384C; Wed, 14 Sep 94 15:16 BST Message-Id: From: Dave Hudson Subject: What video devices are people using? To: VSTa mailing list Date: Wed, 14 Sep 1994 15:16:20 +0100 (BST) X-Mailer: ELM [version 2.4 PL23] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 975 Hi All, I'm now heavily into the coding on the bitblt graphics server and I need to make a decision about where to put my next major amount of effort. As it stands the hardware specific code is running on generic VGAs, Cirrus 542X and Cirrus 6440 (LCD only). I'll be adding support for many other SVGAs in the coming weeks and also coding an S3 variant. It would be a good time to put other hooks in there while I'm at the current stage of development, so could anyone mail me if they're running anything other than the following (eg EGA or Hercules, etc): Generic VGA Cirrus 542X Cirrus 543X Cirrus 6440 Trident TVGA8900B, C, CL and D Paradise/Western Digital PVGA1A, WD90CXX ATI Wonder, XL Tseng ET4000 Similarly, if anyone really wants to see mono support at first release please say so now as I'm currently only handling 4, 8, 15, 16 and 24 bits per pixel in hardware (Gavin's higher level library code may do 1bpp software ops though). Regards, Dave From vandys@glare.cisco.com Wed Sep 14 07:30:21 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id HAA00387; Wed, 14 Sep 1994 07:30:20 -0700 Received: from ebt.com (spock.ebt.com [192.111.115.3]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id IAA07535 for ; Wed, 14 Sep 1994 08:41:10 -0700 Received: from ebt-inc.ebt.com by ebt.com (4.1/SMI-4.1) id AA11813; Wed, 14 Sep 94 11:42:09 EDT Received: by ebt-inc.ebt.com (5.0/SMI-SVR4) id AA08017; Wed, 14 Sep 1994 11:42:10 +0500 Date: Wed, 14 Sep 1994 11:42:10 +0500 From: gtn@ebt.com (Gavin Nicol) Message-Id: <9409141542.AA08017@ebt-inc.ebt.com> To: dave@humbug.demon.co.uk Cc: vsta@cisco.com In-Reply-To: (message from Dave Hudson on Wed, 14 Sep 1994 15:16:20 +0100 (BST)) Subject: Re: What video devices are people using? Content-Length: 174 I should note that the higher level library that Dave refers to is the library that bitblt will use to do drawing etc. I plan to support 1, 8, 16, and 24 bpp at this level. From vandys@glare.cisco.com Wed Sep 14 16:35:16 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id QAA00452; Wed, 14 Sep 1994 16:35:15 -0700 Received: from relay1.Hawaii.Edu (relay1.Hawaii.Edu [128.171.41.53]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id RAA08636 for ; Wed, 14 Sep 1994 17:46:08 -0700 Received: from uhunix.uhcc.Hawaii.Edu ([128.171.44.6]) by relay1.Hawaii.Edu with SMTP id <11359>; Wed, 14 Sep 1994 14:46:06 -1000 Received: by uhunix.uhcc.Hawaii.Edu id <184452>; Wed, 14 Sep 1994 14:45:43 -1000 Subject: small problem in boot loader From: Tim Newsham To: vsta@cisco.com Date: Wed, 14 Sep 1994 14:45:16 -1000 X-Mailer: ELM [version 2.3 PL11] Message-Id: <94Sep14.144543hst.184452@uhunix.uhcc.Hawaii.Edu> There's a small problem in the boot loader that only shows up if the last booted server is smaller than a single page (probably never happens in normal circumstances). The boot program doesnt round up the page after loading the last boot program. This means the free_pfn will be the same as the starting page of the last boot server and the OS will not load that server. Adding a round_pbase() after the loop in which all the servers are loaded fixes this. (ran across this when loading some tiny test programs) Tim N. From vandys@glare.cisco.com Wed Sep 14 17:07:33 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id RAA00467; Wed, 14 Sep 1994 17:07:29 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id SAA05149; Wed, 14 Sep 1994 18:17:56 -0700 Message-Id: <199409150117.SAA05149@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: Tim Newsham Cc: vsta@cisco.com Subject: Re: small problem in boot loader In-Reply-To: Your message of "Wed, 14 Sep 1994 14:45:16 -1000." <94Sep14.144543hst.184452@uhunix.uhcc.Hawaii.Edu> Date: Wed, 14 Sep 1994 18:17:56 -0700 From: Andrew Valencia [Tim Newsham writes:] >There's a small problem in the boot loader that only shows up >if the last booted server is smaller than a single page (probably >never happens in normal circumstances). The boot program doesnt >round up the page after loading the last boot program. This means >the free_pfn will be the same as the starting page of the last >boot server and the OS will not load that server. Adding >a round_pbase() after the loop in which all the servers are loaded >fixes this. This will always be a no-op with an i386 a.out; sizes are rounded up to page values, I believe. However, it's worth doing; I'll add your patch to the base system. In case you all are wondering, Tim is making great progress on a port to the 68030 Amiga! "Live from Interop in Atlanta" Andy From vandys@glare.cisco.com Wed Sep 14 21:00:30 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id VAA00504; Wed, 14 Sep 1994 21:00:29 -0700 Received: from relay1.Hawaii.Edu (relay1.Hawaii.Edu [128.171.41.53]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id WAA16319 for ; Wed, 14 Sep 1994 22:11:23 -0700 Received: from uhunix.uhcc.Hawaii.Edu ([128.171.44.6]) by relay1.Hawaii.Edu with SMTP id <11338>; Wed, 14 Sep 1994 19:11:11 -1000 Received: by uhunix.uhcc.Hawaii.Edu id <184450>; Wed, 14 Sep 1994 19:10:48 -1000 Subject: console vs. debugger console From: Tim Newsham To: vsta@cisco.com Date: Wed, 14 Sep 1994 19:10:40 -1000 X-Mailer: ELM [version 2.3 PL11] Message-Id: <94Sep14.191048hst.184450@uhunix.uhcc.Hawaii.Edu> Question: How is the console handled when the kernel is compiled for debugging? Are two screens maintained and switched between? Tim N. From vandys@glare.cisco.com Thu Sep 15 04:37:43 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id EAA00650; Thu, 15 Sep 1994 04:37:42 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id FAA25737; Thu, 15 Sep 1994 05:48:37 -0700 Message-Id: <199409151248.FAA25737@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: Tim Newsham Cc: vsta@cisco.com Subject: Re: console vs. debugger console In-Reply-To: Your message of "Wed, 14 Sep 1994 19:10:40 -1000." <94Sep14.191048hst.184450@uhunix.uhcc.Hawaii.Edu> Date: Thu, 15 Sep 1994 05:48:37 -0700 From: Andrew Valencia [Tim Newsham writes:] > How is the console handled when the kernel is compiled for >debugging? Are two screens maintained and switched between? The kernel debugger uses a spinloop watching for keystrokes and writes directly to screen memory. Or, if kdb is configured for serial port operation, it spins on the UART instead. The cons2 server causes the kernel debugger to be entered when a ^Z is typed; it saves the screen before calling the debugger, and restores it on return. Andy From vandys@glare.cisco.com Fri Sep 16 03:04:24 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id DAA00729; Fri, 16 Sep 1994 03:04:23 -0700 Received: from post.demon.co.uk (post.demon.co.uk [158.152.1.72]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id EAA01142 for ; Fri, 16 Sep 1994 04:15:15 -0700 Received: from humbug.demon.co.uk by post.demon.co.uk id aa10364; 16 Sep 94 11:37 GMT-60:00 Received: by humbug.demon.co.uk (Linux Smail3.1.28.1 #1) id m0qlaiJ-0003E4C; Fri, 16 Sep 94 11:40 BST Message-Id: From: Dave Hudson Subject: Re: What video devices are people using To: VSTa mailing list Date: Fri, 16 Sep 1994 11:40:19 +0100 (BST) X-Mailer: ELM [version 2.4 PL23] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 323 Hi All, Thanks for the replies. So far I've had about half a dozen asking for S3 support. I think I mentioned it but missed it off my list. Anyway S3 is a priority as that's what I use on my main development system :-) I've also had an offer of help with a Compaq AVGA/Qvision driver. Any more? Regards, Dave From vandys@glare.cisco.com Sat Sep 17 15:17:44 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id PAA00874; Sat, 17 Sep 1994 15:17:43 -0700 Received: from relay1.Hawaii.Edu (relay1.Hawaii.Edu [128.171.41.53]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id QAA03560 for ; Sat, 17 Sep 1994 16:28:53 -0700 Received: from uhunix.uhcc.Hawaii.Edu ([128.171.44.6]) by relay1.Hawaii.Edu with SMTP id <11337>; Sat, 17 Sep 1994 13:28:41 -1000 Received: by uhunix.uhcc.Hawaii.Edu id <184465>; Sat, 17 Sep 1994 13:28:32 -1000 Subject: VSTa on the Amiga From: Tim Newsham To: amiga@netbsd.org, amiga-mach@appli.se, vsta@cisco.com Date: Sat, 17 Sep 1994 13:28:24 -1000 X-Mailer: ELM [version 2.3 PL11] Message-Id: <94Sep17.132832hst.184465@uhunix.uhcc.Hawaii.Edu> VSTa For Amiga Lives! I guess now is a good time to announce that I have VSTa running on my Amiga. I just got the keyboard driver working today and the screen driver earlier this week. I booted up and ran VSTa with three processes: the screen driver, the keyboard driver and a test process which does I/O through the two servers and everything seems to be working ok. What is VSTa? The name stands for Valencia's Simple Tasker. It is a small microkernel based operating system. It borrows ideas from QNX and Plan9. Under VSTa all device drivers and filesystems are user-level processes that interact with each other through message passing. VSTa presents the system to the user as a number of filesystems as Plan9 does. For more information: ftp.cisco.com:/vandys/vsta - ftp archive ftp.cygnus.com:/pub/embedded/vsta - ftp archive vsta@cisco.com - mailing list vsta-request@cisco.com - mailing list requests The amiga port somewhat of a kludge right now. It needs lots of cleaning up, some disk drivers and lots of love and care. It currently is written to run only on 68030's and possibly 020's with MMU's. There is no support for the '040 MMU. There are no disk drivers yet. Some of the features are in a "hack" state so that I could get things running and the boot loader has my system setup hardwired in. I invite any experienced amiga programmers to help out. I will continue to work on the port and hope to eventually have VSTa self hosting, running as my primary Operating System on my amiga. Tim Newsham newsham@uhunix.uhcc.hawaii.edu From vandys@glare.cisco.com Sat Sep 17 20:05:18 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id UAA00892; Sat, 17 Sep 1994 20:05:17 -0700 Received: from relay1.Hawaii.Edu (relay1.Hawaii.Edu [128.171.41.53]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id VAA08164 for ; Sat, 17 Sep 1994 21:16:28 -0700 Received: from uhunix.uhcc.Hawaii.Edu ([128.171.44.6]) by relay1.Hawaii.Edu with SMTP id <11367>; Sat, 17 Sep 1994 18:16:25 -1000 Received: by uhunix.uhcc.Hawaii.Edu id <184460>; Sat, 17 Sep 1994 18:16:15 -1000 Subject: ATOMIC_*() -- endian From: Tim Newsham To: vsta@cisco.com Date: Sat, 17 Sep 1994 18:16:05 -1000 X-Mailer: ELM [version 2.3 PL11] Message-Id: <94Sep17.181615hst.184460@uhunix.uhcc.Hawaii.Edu> the ATOMIC_INC() and ATOMIC_DEC() functions are currently dependant on the machine being little endian. There are both short (2 byte) and long (4 byte) values passed in. This is fine for little endian machines where the low bytes are stored first but on a big endian machine the type has to be known ahead of time. The prototypes specified void * previously which I changed to ushort *. There were a few values that were being passed in as long values: cpu ticks, number of threads and the statistics (if DEBUG is on) for malloc. Tim N. From vandys@glare.cisco.com Sun Sep 18 00:15:41 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id AAA00906; Sun, 18 Sep 1994 00:15:39 -0700 Received: from mail.swip.net (mail.swip.net [192.71.220.11]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with ESMTP id BAA13419 for ; Sun, 18 Sep 1994 01:26:51 -0700 Received: by mail.swip.net (8.6.8/3.01) id KAA27503; Sun, 18 Sep 1994 10:26:44 +0200 Received: from della.appli.se by liza.appli.se (4.1/SMI-4.1) id AA11769; Sun, 18 Sep 94 10:24:08 CDT Received: by della.appli.se (4.1/SMI-4.1) id AA12646; Sun, 18 Sep 94 10:24:29 CDT Date: Sun, 18 Sep 94 10:24:29 CDT From: niklas@appli.se (Niklas Hallqvist) Message-Id: <9409180824.AA12646@della.appli.se> To: newsham@uhunix.uhcc.Hawaii.Edu Cc: amiga@NetBSD.ORG, amiga-mach@appli.se, vsta@cisco.com In-Reply-To: <94Sep17.132832hst.184465@uhunix.uhcc.Hawaii.Edu> (message from Tim Newsham on Sat, 17 Sep 1994 13:28:24 -1000) Subject: Re: VSTa on the Amiga >>>>> "Tim" == Tim Newsham writes: Tim> VSTa For Amiga Lives! Nice! Does this mean you got the fault address code question answered? If so, I'm interested in the answer. Niklas From vandys@glare.cisco.com Mon Sep 19 11:00:03 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id LAA01154; Mon, 19 Sep 1994 11:00:00 -0700 Received: from relay1.UU.NET (relay1.UU.NET [192.48.96.5]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with ESMTP id MAA12545 for ; Mon, 19 Sep 1994 12:11:21 -0700 Received: from gateway.mwc.com by relay1.UU.NET with SMTP id QQxibk18488; Mon, 19 Sep 1994 15:10:42 -0400 Received: from mwc by gateway.mwc.com with uucp (Smail3.1.28.1 #23) id m0qmnus-000216C; Mon, 19 Sep 94 13:58 CDT Received: by mwc.com (smail2.5.3-coh) id AA15358; 19 Sep 94 18:53:15 GMT (Mon) Received: by invid.mwc.com (smail2.5.3-coh) id AA00383; 19 Sep 94 14:58:21 CDT (Mon) To: vsta@cisco.com Date: Mon, 19 Sep 1994 14:58:21 +0200 (CDT) From: "ed" X-Mailer: ELM [version 2.4 PL23] Content-Type: text Content-Length: 100 Message-Id: <9409191458.AA00381@invid.mwc.com> subscribe vista (Ed Bravo) Requesting to be placed on the vsta mailing list. Ed Bravo ed@mwc.com From vandys@glare.cisco.com Tue Sep 20 07:37:49 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id HAA01277; Tue, 20 Sep 1994 07:37:48 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id IAA05720 for ; Tue, 20 Sep 1994 08:49:13 -0700 Message-Id: <199409201549.IAA05720@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: vsta@amri.cisco.com Subject: VSTa status Date: Tue, 20 Sep 1994 08:49:13 -0700 From: Andrew Valencia Hello, folks. Well, I was on the road with an i486 laptop with VSTa, and did some real damage. I have symlinks and environment links working, and I have shared libraries pretty close to working. I had to do some surgery to enable MAP_FILE under mmap(), but mostly I could move code from exec() support, so it didn't really grow the kernel much. Being on the laptop, networking couldn't move forward much. I have part of a /tcp server written. Regards, Andy From vandys@glare.cisco.com Tue Sep 20 07:48:35 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id HAA01280; Tue, 20 Sep 1994 07:48:33 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id IAA06151; Tue, 20 Sep 1994 08:59:58 -0700 Message-Id: <199409201559.IAA06151@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: Tim Newsham Cc: vsta@cisco.com Subject: Re: ATOMIC_*() -- endian In-Reply-To: Your message of "Sat, 17 Sep 1994 18:16:05 -1000." <94Sep17.181615hst.184460@uhunix.uhcc.Hawaii.Edu> Date: Tue, 20 Sep 1994 08:59:57 -0700 From: Andrew Valencia [Tim Newsham writes:] >the ATOMIC_INC() and ATOMIC_DEC() functions are currently >dependant on the machine being little endian. There are both >short (2 byte) and long (4 byte) values passed in. Not on purpose! This is a bug. There should be ATOMIC_INCW() or something to keep it honest. >The prototypes specified void * previously >which I changed to ushort *. There were a few values that >were being passed in as long values: cpu ticks, number of threads >and the statistics (if DEBUG is on) for malloc. Yuck. Maybe Dave Hudson remembers more about the why's and wherefore's of this? I will hunt down each reference and convert the calls accessing a short value to use ATOMIC_INCW/_DECW() unless I hear something else. Andy From vandys@glare.cisco.com Tue Sep 20 09:07:01 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id JAA01292; Tue, 20 Sep 1994 09:07:00 -0700 Received: from ftp.std.com (ftp.std.com [192.74.137.7]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with ESMTP id KAA02406 for ; Tue, 20 Sep 1994 10:18:26 -0700 Received: from world.std.com by ftp.std.com (8.6.8.1/Spike-8-1.0) id NAA03349; Tue, 20 Sep 1994 13:18:25 -0400 Received: by world.std.com (5.65c/Spike-2.0) id AA05663; Tue, 20 Sep 1994 13:18:24 -0400 From: larz@world.std.com (Mike A Larson) Message-Id: <199409201718.AA05663@world.std.com> Subject: Re: VSTa status To: vsta@cisco.com Date: Tue, 20 Sep 1994 13:18:23 -0400 (EDT) X-Mailer: ELM [version 2.4 PL23] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 542 Hi, Andrew Valencia writes: > Well, I was on the road with an i486 laptop with VSTa, and did some real > damage. I have symlinks and environment links working, and I have shared > libraries pretty close to working. What changes need to be made to the filesystem servers to support symbolic links? The reason I ask is because I'm about to add support for the Rock Ridge Extensions to the CDROM server and I'd like to coordinate the symbolic link part of this work with what you've done. Thanks. Mike Larson From vandys@glare.cisco.com Tue Sep 20 09:32:00 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id JAA01300; Tue, 20 Sep 1994 09:31:58 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id KAA13805; Tue, 20 Sep 1994 10:43:24 -0700 Message-Id: <199409201743.KAA13805@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: larz@world.std.com (Mike A Larson) Cc: vsta@cisco.com Subject: Re: VSTa status In-Reply-To: Your message of "Tue, 20 Sep 1994 13:18:23 EDT." <199409201718.AA05663@world.std.com> Date: Tue, 20 Sep 1994 10:43:23 -0700 From: Andrew Valencia [larz@world.std.com (Mike A Larson) writes:] >What changes need to be made to the filesystem servers to support >symbolic links? The reason I ask is because I'm about to add >support for the Rock Ridge Extensions to the CDROM server and I'd >like to coordinate the symbolic link part of this work with what you've >done. Thanks. Not too much. I stole the "hidden" attribute bit to represent the file being a symlink. So you need to map that bit into filetype "symlink", add some code so the type can be switched from "f" to "symlink" and back, and add code to the dos_open() routine so that ESYMLINK is returned if an open is attempted on a symlink and ACC_SYMLINK isn't set in the modes bitmask. When the path lookup loop in libc/open.c gets ESYMLINK, it opens the file with ACC_SYMLINK, gets the contents, replaces the current path element with this contents, closes the symlink file, and continues. Thus my work on shared libraries. An important change like this requires each and every a.out to be rebuilt to get the new C library code. I've known shared libraries are needed for a LONG time, but this was the event which kicked me into action. Now, administrivia. It looks like the list is picking up steam. Remember that we have a digest format available; send mail to vsta-request@cisco.com. Also, we have a shared library list (vsta-shlib@cisco.com); also send a request to vsta-request@cisco.com to join it. Let's keep shlib discussion on that list; if I write up a decent summary of how VSTa shared libraries work, I'll still send it to the main list. If somebody comes up with an eloquent summary of why my design proves I'm crazed out of my mind, I guess that's OK to post to the main list as well. :-) Andy From vandys@glare.cisco.com Tue Sep 20 12:37:40 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id MAA01377; Tue, 20 Sep 1994 12:37:39 -0700 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with ESMTP id NAA26403 for ; Tue, 20 Sep 1994 13:48:59 -0700 Received: from max.tiac.net (max.tiac.net [199.0.65.4]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with ESMTP id NAA13294 for ; Tue, 20 Sep 1994 13:48:58 -0700 Received: (from chrisp@localhost) by max.tiac.net (8.6.9/8.6.6.Beta9) id QAA10306 for vsta@glare.cisco.com; Tue, 20 Sep 1994 16:46:06 -0400 Date: Tue, 20 Sep 1994 16:46:06 -0400 From: Chris Patti { Feoh } Message-Id: <199409202046.QAA10306@max.tiac.net> To: vsta@cisco.com Subject: Will The New VSTa ever be released? I've been following VSTa's progress for quite a while now and it sounds like it's evolved to a much higher plane than when I last installed it. Will the new VSTa be released anytime soon? I heard talk about this a few months ago and haven't heard anything since. -Chris From vandys@glare.cisco.com Tue Sep 20 16:55:41 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id QAA01421; Tue, 20 Sep 1994 16:55:39 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id SAA15224; Tue, 20 Sep 1994 18:07:06 -0700 Message-Id: <199409210107.SAA15224@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: Chris Patti { Feoh } Cc: vsta@cisco.com Subject: Re: Will The New VSTa ever be released? In-Reply-To: Your message of "Tue, 20 Sep 1994 16:46:06 EDT." <199409202046.QAA10306@max.tiac.net> Date: Tue, 20 Sep 1994 18:07:06 -0700 From: Andrew Valencia [Chris Patti { Feoh } writes:] >I've been following VSTa's progress for quite a while now and it sounds like >it's evolved to a much higher plane than when I last installed it. Will the >new VSTa be released anytime soon? I heard talk about this a few months ago >and haven't heard anything since. Well, having children will do that to you.... Plans are progressing for a Internet-accessible machine to host the shared source. At this point I think v1.4 will come out about the time this machine becomes available. From v1.4 on we will start working in a shared source environment. This will eliminate me as a bottleneck in VSTa development, and will also allow anybody to grab a snapshot any time they desire of the "in progress" source. Andy From vandys@glare.cisco.com Wed Sep 21 02:10:35 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id CAA01449; Wed, 21 Sep 1994 02:10:34 -0700 Received: from gwar.mit.edu (root@GWAR.MIT.EDU [18.244.0.47]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id DAA16773 for ; Wed, 21 Sep 1994 03:22:04 -0700 Received: by gwar.mit.edu id AA00109 (5.67b/IDA-1.5 for vsta@cisco.com); Wed, 21 Sep 1994 06:22:03 -0400 Date: Wed, 21 Sep 1994 06:22:03 -0400 From: Cotton Seed Message-Id: <199409211022.AA00109@gwar.mit.edu> To: vsta@cisco.com Subject: VSTa installation failure i installed VSTa as per the instructions in `doc/roadmap.txt'. i left `boot.lst' as unchanged, the default values seemed reasonably sane. when i typed `go' at the dos prompt in `C:\VSTA\BOOT' i got a vsta banner, a list of the drivers started, and `Launch at 0x1020' then my machine promptly rebooted. this is a ZEOS PANTERA PENTIUM 90. i have a single western digital IDE drive for dos (and several large scsi drives with a ncr 53c810 controller-- currently for NeXTstep/Linux use). i have been meaning to try out VSTa for some time, it is disappointing that it failed outright. any ideas about how to get VSTa running? - Cotton --- Cotton Seed (cottons@mit.edu) From vandys@glare.cisco.com Wed Sep 21 07:04:49 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id HAA01506; Wed, 21 Sep 1994 07:04:47 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id IAA13132; Wed, 21 Sep 1994 08:16:19 -0700 Message-Id: <199409211516.IAA13132@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: Cotton Seed Cc: vsta@cisco.com Subject: Re: VSTa installation failure In-Reply-To: Your message of "Wed, 21 Sep 1994 06:22:03 EDT." <199409211022.AA00109@gwar.mit.edu> Date: Wed, 21 Sep 1994 08:16:19 -0700 From: Andrew Valencia [Cotton Seed writes:] >a list of the drivers started, >and `Launch at 0x1020' >then my machine promptly rebooted. Yuck. >this is a ZEOS PANTERA PENTIUM 90. i have a single western digital >IDE drive for dos (and several large scsi drives with a ncr 53c810 >controller-- currently for NeXTstep/Linux use). Well, I've never had my hands on a Pentium before, so that jumps out as a possible culprit. On the other hand, I had heard they were pretty compatible. You might try unloading all extra DOS drivers to see if there's a memory conflict before VSTa is even done loading. You might also disable any extra caches, turbo modes, and so forth. Otherwise, you have to debug VSTa startup to see where we're dying. If you're up for it, send me E-mail directly and we can walk through it. Andy From vandys@glare.cisco.com Wed Sep 21 07:39:07 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id HAA01510; Wed, 21 Sep 1994 07:39:04 -0700 Received: from ebt.com (spock.ebt.com [192.111.115.3]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id IAA26675; Wed, 21 Sep 1994 08:50:30 -0700 Received: from ebt-inc.ebt.com by ebt.com (4.1/SMI-4.1) id AA16317; Wed, 21 Sep 94 11:50:44 EDT Received: by ebt-inc.ebt.com (5.0/SMI-SVR4) id AA29026; Wed, 21 Sep 1994 11:50:45 +0500 Date: Wed, 21 Sep 1994 11:50:45 +0500 From: gtn@ebt.com (Gavin Nicol) Message-Id: <9409211550.AA29026@ebt-inc.ebt.com> To: vandys@cisco.com Cc: cottons@gwar.mit.edu, vsta@cisco.com In-Reply-To: <199409211516.IAA13132@glare.cisco.com> (message from Andrew Valencia on Wed, 21 Sep 1994 08:16:19 -0700) Subject: Re: VSTa installation failure Content-Length: 578 No problems with Pentiums. My Gateway P5-66 flies with VSTa on it. I had similar problems, though my machine just died. My problem was due to the flaky video card in my machine. Despite displaying graphics in color it thought it was in monochrome mode, and cons killed the machine. Yours sounds like it occurs just at the jump to kernel space, so my guess is cache, or DOS; same as Andy. Last time I ran VSTa it was on a 386SX 16mhz machine with 3MB of main memory. I almost dies at the difference between that and a 16MB Pentium 66... The 386 wasn't even IBM compatible... From vandys@glare.cisco.com Wed Sep 21 09:54:54 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id JAA01542; Wed, 21 Sep 1994 09:54:52 -0700 Received: from grunt.ksu.ksu.edu (root@grunt.ksu.ksu.edu [129.130.12.17]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with ESMTP id LAA04174; Wed, 21 Sep 1994 11:06:24 -0700 From: brtmac@ksu.ksu.edu Received: from mort by grunt.ksu.ksu.edu (8.6.8/1.34) id NAA16988; Wed, 21 Sep 1994 13:05:45 -0500 Received: by mort (8.6.8/1.34) id SAA26947; Wed, 21 Sep 1994 18:05:37 GMT Date: Wed, 21 Sep 1994 18:05:37 GMT Message-Id: <199409211805.SAA26947@mort> To: Andrew Valencia Cc: Cotton Seed , vsta@cisco.com Subject: Re: VSTa installation failure In-Reply-To: <199409211516.IAA13132@glare.cisco.com> References: <199409211022.AA00109@gwar.mit.edu> <199409211516.IAA13132@glare.cisco.com> Verily did Andrew Valencia say on September 21, 1994: >>this is a ZEOS PANTERA PENTIUM 90. i have a single western digital >>IDE drive for dos (and several large scsi drives with a ncr 53c810 >>controller-- currently for NeXTstep/Linux use). > >Well, I've never had my hands on a Pentium before, so that jumps out >as a possible culprit. On the other hand, I had heard they were >pretty compatible. You might try unloading all extra DOS drivers to >see if there's a memory conflict before VSTa is even done loading. >You might also disable any extra caches, turbo modes, and so forth. On my machine if I've got any device drivers loaded or TSR there is almost 0 chance that vsta will start. I've created a DOS startup menu that has a VSTA selection on it which skips everything in the config.sys file and does a 'cd vsta_boot_dir; go.bat' in the autoexec.bat. The other thing to try, if you are running DOS 6 or later, is to hit F5 when you see the Starting MS-DOS... message. That will cause it to skip the config.sys and autoexec.bat files completely, then you can cd to the boot directory and do a 'go.bat'. Then again, none of this may have any effect for anyone else. Brett McCoy, UNIX Systems Administrator Computing and Network Services Kansas State University, Manhattan KS 66506 vox: (913) 532-4908 / fax: (913) 532-5914 / e-mail: brtmac@ksu.ksu.edu From vandys@glare.cisco.com Wed Sep 21 19:46:02 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id TAA01581; Wed, 21 Sep 1994 19:46:01 -0700 Received: from wotan.compaq.com (wotan.compaq.com [131.168.249.254]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id UAA00687 for ; Wed, 21 Sep 1994 20:54:40 -0700 Received: from twisto by wotan.compaq.com with uucp (Smail3.1.28.1 #11) id m0qnfDb-000vIgC; Wed, 21 Sep 94 22:53 CDT Received: from gocart.eng.hou.compaq.com by twisto.eng.hou.compaq.com with smtp (Smail3.1.28.1 #10) id m0qnf96-000uGqC; Wed, 21 Sep 94 22:48 CDT Received: by gocart.eng.hou.compaq.com (Smail3.1.26.7/COMPAQ-HESIOD) id m0qnf95-00036pC; Wed, 21 Sep 94 22:48 CDT Message-Id: From: edm@gocart.eng.hou.compaq.com (Edward Mccreary) Subject: Looking for a project... To: vsta@cisco.com Date: Wed, 21 Sep 94 22:48:31 CDT X-Mailer: ELM [version 2.4dev PL11] Hi guys, I've been trying to think of project to work on and wanted to get a little feedback from people on the list. What I'd like to do is write a filesystem server for either Linux or one of the BSD's. My two questions are... 1) Is anybody already working on a Linux or BSD server? If so, I won't bother. 2) What's the preference? My first inclination is towards BSD since that's what I'm more familiar with. But, it seems that more people on the list are using Linux so it may be more helpful to write one for it. (Of course, another factor is which is easier to setup dual-boot with dos. :) thanks for any feedback, ed -- Eddie McCreary edm@twisto.compaq.com Graphics Development "Do or do not, there is no 'try'." In the event of my capture, Compaq will disavow any and all knowledge of my operations. Of course I don't speak for them. From vandys@glare.cisco.com Wed Sep 21 20:25:28 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id UAA01588; Wed, 21 Sep 1994 20:25:27 -0700 Received: from relay1.Hawaii.Edu (relay1.Hawaii.Edu [128.171.41.53]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id VAA01611 for ; Wed, 21 Sep 1994 21:37:01 -0700 Received: from uhunix.uhcc.Hawaii.Edu ([128.171.44.6]) by relay1.Hawaii.Edu with SMTP id <11340>; Wed, 21 Sep 1994 18:36:53 -1000 Received: by uhunix.uhcc.Hawaii.Edu id <184427>; Wed, 21 Sep 1994 18:36:44 -1000 Subject: Re: Looking for a project... From: Tim Newsham To: edm@gocart.eng.hou.compaq.com (Edward Mccreary) Date: Wed, 21 Sep 1994 18:36:32 -1000 Cc: vsta@cisco.com In-Reply-To: ; from "Edward Mccreary" at Sep 21, 94 5:48 pm X-Mailer: ELM [version 2.3 PL11] Message-Id: <94Sep21.183644hst.184427@uhunix.uhcc.Hawaii.Edu> > > Hi guys, > > 2) What's the preference? My first inclination is towards BSD since > that's what I'm more familiar with. But, it seems that more people > on the list are using Linux so it may be more helpful to write one for > it. My vote is for BSD. I already have several BSD partitions on my machine so the reason is obvious :). I also heard some rumours about linux's ext. fs not being so great. > Eddie McCreary edm@twisto.compaq.com From vandys@glare.cisco.com Wed Sep 21 21:27:24 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id VAA01599; Wed, 21 Sep 1994 21:27:23 -0700 Received: from alterdial.UU.NET (0@alterdial.UU.NET [192.48.96.22]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with ESMTP id WAA02837 for ; Wed, 21 Sep 1994 22:38:56 -0700 Received: from pyro.thought.com by alterdial.UU.NET with SMTP id QQxikk21185; Thu, 22 Sep 1994 01:38:48 -0400 Received: by pyro.thought.com (NX5.67d/NX3.0M) id AA01100; Thu, 22 Sep 94 00:43:46 -0500 Date: Thu, 22 Sep 94 00:43:46 -0500 From: Benjamin Black (Mr. Jinx) Message-Id: <9409220543.AA01100@pyro.thought.com> Received: by NeXT.Mailer (1.100) Received: by NeXT Mailer (1.100) To: vsta@cisco.com Subject: Re: Looking for a project... Reply-To: ben@pyro.thought.com I vote for BSD4.4 for those lovely 64-bit file descriptors (among other cool things). How reliant is VSTa on 32-bit processors (not that this relates to 64-bit file descriptors)? Just wondering since an Alpha port would be cool, and 64-bit Intel, SPARC, and PPC silicon is just around the corner... Ben From vandys@glare.cisco.com Thu Sep 22 00:20:43 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id AAA01615; Thu, 22 Sep 1994 00:20:42 -0700 Received: from swann.irisa.fr (swann.irisa.fr [131.254.43.10]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with ESMTP id BAA07475 for ; Thu, 22 Sep 1994 01:32:11 -0700 Received: from oblomov.irisa.fr (oblomov.irisa.fr [131.254.43.11]) by swann.irisa.fr (8.6.9/8.6.9) with ESMTP id KAA12605 for ; Thu, 22 Sep 1994 10:31:38 +0200 From: Tommy Thorn Date: Thu, 22 Sep 1994 10:31:37 +0200 Message-Id: <199409220831.KAA16644@oblomov.irisa.fr> To: vsta@cisco.com Subject: Re: Looking for a project... > > 2) What's the preference? My first inclination is towards BSD since > > that's what I'm more familiar with. But, it seems that more people > > on the list are using Linux so it may be more helpful to write one for > > it. If this is going to be a vote-thing, my vote is for Linux, but what exactly do you mean by a Linux server? A filesystem server? Which? Linux currently support Minix+ext, ext, xia, ext2, sysv, proc, .. etc. There is _no_ native Linux fs, though nearly everybody uses ext2. > My vote is for BSD. I already have several BSD partitions on > my machine so the reason is obvious :). I also heard some rumours > about linux's ext. fs not being so great. While this is very true, who cares? Everybody uses ext2 which is one of the best filesystem implementation I know. Easily outperforms BSD's Fast Filesystem. Ext2 is _very_ different from ext, but both are really bad names. (ext2 is also known by e2fs). The real reason you might prefer ext2 to any other is the fact that all the filesystem code has been carefully isolated an put in a library. So far there exist an ext2 server for hurd. If anybody should be interested in more, I've the source and some slides at ftp://ftp.daimi.aau.dk/pub/linux/ext2. Now, back to VSTa issues. /Tommy Thorn From vandys@glare.cisco.com Thu Sep 22 00:20:14 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id AAA01612; Thu, 22 Sep 1994 00:20:13 -0700 Received: from post.demon.co.uk (post.demon.co.uk [158.152.1.72]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id BAA07471 for ; Thu, 22 Sep 1994 01:31:47 -0700 Received: from humbug.demon.co.uk by post.demon.co.uk id aa18279; 22 Sep 94 9:26 GMT-60:00 Received: by humbug.demon.co.uk (Linux Smail3.1.28.1 #1) id m0qnjTa-00036wC; Thu, 22 Sep 94 09:25 BST Message-Id: From: Dave Hudson Subject: Re: Looking for a project... To: VSTa mailing list Date: Thu, 22 Sep 1994 09:25:57 +0100 (BST) X-Mailer: ELM [version 2.4 PL23] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 2005 Tim Newsham wrote: > > > > > Hi guys, > > > > 2) What's the preference? My first inclination is towards BSD since > > that's what I'm more familiar with. But, it seems that more people > > on the list are using Linux so it may be more helpful to write one for > > it. > > My vote is for BSD. I already have several BSD partitions on > my machine so the reason is obvious :). I also heard some rumours > about linux's ext. fs not being so great. The ext fs wasn't so hot and has been pretty much abandoned. ext2 is much better. The benchmarks floating around in the Linux development groups seem to suggest that ext2 is on a par with BSD. I know that there's been mention of both ext2 and minix servers on this list but I don't know what's happened yet. If you plan on doing ext2 you'll find that as well as the kernel source implementation there's now a HURD server. FWIW the thing I'd like to see (and if I ever get a spare few hours I'll look into doing) would be to write something like the Linux umsdos fs. For the non-Linux types on the list, this basically takes a standard dos fs but uses a special file (if present) in each subdirectory to manage additional attributes such as ownership and to provide additional filename space. The beauty of such a system is that it's overlayed on top of the dos fs and doesn't stop it from being used normally under dos (it's just some of the dos versions of long filenames can look a bit strange). In VSTa such a server could be implemented as something that layers on top of an existing fs to provide VSTa permissions and enhanced timestamping - for now this only really applies to the dos fs but the same technique could be applied to any "restrictive" fs. What would be an interesting project would be to extract all of the commonality from the current fs's and create a multithreaded fs library that would manage block caching etc and that could be used for creating any new fs. I think HURD has something like this. Regards, Dave From vandys@glare.cisco.com Thu Sep 22 07:08:23 1994 Received: from dustbin.cisco.com (dustbin.cisco.com [131.108.1.27]) by amri.cisco.com (8.3/8.3) with ESMTP id HAA01767; Thu, 22 Sep 1994 07:08:22 -0700 Received: from ftp.std.com (ftp.std.com [192.74.137.7]) by dustbin.cisco.com (8.6.8+c/CISCO.SUB_GATE.1.1) with ESMTP id IAA29152 for ; Thu, 22 Sep 1994 08:19:59 -0700 Received: from world.std.com by ftp.std.com (8.6.8.1/Spike-8-1.0) id LAA09465; Thu, 22 Sep 1994 11:18:33 -0400 Received: by world.std.com (5.65c/Spike-2.0) id AA25754; Thu, 22 Sep 1994 11:18:26 -0400 From: larz@world.std.com (Mike A Larson) Message-Id: <199409221518.AA25754@world.std.com> Subject: Re: Looking for a project... To: vsta@cisco.com Date: Thu, 22 Sep 1994 11:18:26 -0400 (EDT) X-Mailer: ELM [version 2.4 PL23] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 1085 Hi, Dave Hudson writes: > FWIW the thing I'd like to see (and if I ever get a spare few hours I'll > look into doing) would be to write something like the Linux umsdos fs. For > the non-Linux types on the list, this basically takes a standard dos fs but > uses a special file (if present) in each subdirectory to manage additional > attributes such as ownership and to provide additional filename space. The > beauty of such a system is that it's overlayed on top of the dos fs and > doesn't stop it from being used normally under dos (it's just some of the > dos versions of long filenames can look a bit strange). Note that the ISO-9660 CDROM filesystem does a similar thing. There is a variable length System Use Area within each directory record to support extensions to the specification. There is also a protocol, called the System Use Sharing Protocol or SUSP, for managing this area. Layered on top of the SUSP, the Rock Ridge extensions define records for long filenames, symbolic links, and Posix file attributes, etc. Mike Larson From vandys@glare.cisco.com Thu Sep 22 08:57:17 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id IAA01779; Thu, 22 Sep 1994 08:57:16 -0700 Received: from roxanne.nuclecu.unam.mx (roxanne.nuclecu.unam.mx [132.248.29.2]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id KAA26754 for ; Thu, 22 Sep 1994 10:08:52 -0700 Received: by roxanne.nuclecu.unam.mx (5.65/Ultrix3.0-C) id AA03297; Thu, 22 Sep 1994 11:07:57 -0600 Date: Thu, 22 Sep 1994 11:07:57 -0600 From: miguel@roxanne.nuclecu.unam.mx (Miguel de Icaza) Message-Id: <9409221707.AA03297@roxanne.nuclecu.unam.mx> To: dave@humbug.demon.co.uk Cc: vsta@cisco.com In-Reply-To: (message from Dave Hudson on Thu, 22 Sep 1994 09:25:57 +0100 (BST)) Subject: Re: Looking for a project... X-Tc: get it at dairy mart > FWIW the thing I'd like to see (and if I ever get a spare few hours I'll > look into doing) would be to write something like the Linux umsdos fs. For > the non-Linux types on the list, this basically takes a standard dos fs but > uses a special file (if present) in each subdirectory to manage additional > attributes such as ownership and to provide additional filename space. The > beauty of such a system is that it's overlayed on top of the dos fs and > doesn't stop it from being used normally under dos (it's just some of the > dos versions of long filenames can look a bit strange). Maybe it would be better to port CMU's DOS-FS, it's a file system based on a normal DOS FAT layout but with some enhacements that make it much faster than BSD's FFS. When the paper was presented on Winter Usenix '94 the people at CMU didn't have 4.4 booting in their machines so they didn't know if it was actually faster than the new 4.4 FFS with the clusering enhacements. The paper should be available in mach.cs.cmu.edu. You can use a DOS-FS fie system as your root partition, just like umsdos, and you can still use the file system under DOS. The designer of the DOS-FS, Alessandro Forin is currently working at Microsoft, so I won't be surprised if the VFAT system of Chicago is actually using the same tricks for speeding up the fs. Miguel. From vandys@glare.cisco.com Thu Sep 22 10:58:18 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id KAA01798; Thu, 22 Sep 1994 10:58:16 -0700 Received: from wally.cs.washington.edu (wally.cs.washington.edu [128.95.2.122]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with ESMTP id MAA03289 for ; Thu, 22 Sep 1994 12:09:55 -0700 Received: from localhost (localhost [127.0.0.1]) by wally.cs.washington.edu (8.6.8/7.2ws+) with SMTP id MAA00975; Thu, 22 Sep 1994 12:09:41 -0700 To: Tommy Thorn Cc: vsta@cisco.com Subject: Re: Looking for a project... In-Reply-To: Your message of "Thu, 22 Sep 1994 10:31:37 +0200." <199409220831.KAA16644@oblomov.irisa.fr> Date: Thu, 22 Sep 1994 12:09:41 PDT Message-Id: <973.780260981@wally.cs.washington.edu> From: Chris Maeda Date: Thu, 22 Sep 1994 10:31:37 +0200 To: vsta@cisco.com From: Tommy Thorn Subject: Re: Looking for a project... The real reason you might prefer ext2 to any other is the fact that all the filesystem code has been carefully isolated an put in a library. So far there exist an ext2 server for hurd. You could say the same thing for any file system implementation that supports the vnode interface. Building in support for vnodes will allow you to easily drop in any vnode file system implementation (cd-rom, nfs, dos, afs, ufs, etc). From vandys@glare.cisco.com Thu Sep 22 14:09:12 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id OAA01820; Thu, 22 Sep 1994 14:09:10 -0700 Received: from virginia.edu (uvaarpa.Virginia.EDU [128.143.2.7]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id PAA12014 for ; Thu, 22 Sep 1994 15:20:49 -0700 Received: from hopper.itc.virginia.edu by uvaarpa.virginia.edu id aa14099; 22 Sep 94 18:20 EDT Received: (from mike@localhost) by Hopper.itc.Virginia.EDU (8.6.8/8.6.6) id SAA16733 for vsta@cisco.com; Thu, 22 Sep 1994 18:20:42 -0400 From: "Michael M. Chapman" Message-Id: <199409222220.SAA16733@Hopper.itc.Virginia.EDU> Subject: weird permissions problems To: vsta@cisco.com Date: Thu, 22 Sep 1994 18:20:42 -0400 (EDT) X-Mailer: ELM [version 2.4 PL23] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 313 I cannot compile with gcc because when it goes to write the object file it prints "perm" before exiting and then perm after exiting. What does this mean? Also, I cannot use the arrow keys with the microemacs port. I'd rather use vi anyway, but seeing as I can't compile yet... Any help *greatly* appreciated. From vandys@glare.cisco.com Thu Sep 22 14:48:58 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id OAA01836; Thu, 22 Sep 1994 14:48:57 -0700 Received: from wyn.fandom-house.org (root@wyn.fandom-house.org [128.127.21.25]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with ESMTP id QAA14035 for ; Thu, 22 Sep 1994 16:00:33 -0700 Received: (from feoh@localhost) by wyn.fandom-house.org (8.6.9/8.6.9) id GAA00539; Thu, 22 Sep 1994 06:59:37 GMT Date: Thu, 22 Sep 1994 06:59:37 GMT From: Chris Patti Message-Id: <199409220659.GAA00539@wyn.fandom-house.org> To: mike@hopper.itc.virginia.edu, vsta@cisco.com Subject: Re: weird permissions problems This bug was fixed *awhile* ago. But the problem is that so were about *9,000* other bugs. VSTa, from what I can tell, is vastly improved from what you and I have. I guess we'll be able to get usable systems going when the 'shared internet source code' site gets going. I have no idea when this will actually exist, since no dates have been mentioned. From vandys@glare.cisco.com Thu Sep 22 14:42:39 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id OAA01833; Thu, 22 Sep 1994 14:42:38 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id PAA05980; Thu, 22 Sep 1994 15:54:15 -0700 Message-Id: <199409222254.PAA05980@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: "Michael M. Chapman" Cc: vsta@cisco.com Subject: Re: weird permissions problems In-Reply-To: Your message of "Thu, 22 Sep 1994 18:20:42 EDT." <199409222220.SAA16733@Hopper.itc.Virginia.EDU> Date: Thu, 22 Sep 1994 15:54:14 -0700 From: Andrew Valencia ["Michael M. Chapman" writes:] >I cannot compile with gcc because when it goes to write the object >file it prints "perm" before exiting and then perm after >exiting. What does this mean? What's your user ID? If you're using the DOS filesystem, you need to have group sys (or sys.sys) as one of your ID's. Otherwise, you don't get write access to the filesystem. >Also, I cannot use the arrow keys with the microemacs port. I'd >rather use vi anyway, but seeing as I can't compile yet... Any >help *greatly* appreciated. Let me take a look at when escape sequences got added for arrow keys. Also, you may need to use the customization language of microemacs to map the keys to their desired function. On the other hand, I have a port of vim done, which I use all the time. I'll break out a v1.3.1 system and see if it runs. Yes, I know, I should hurry up and get v1.4 out. Andy From vandys@glare.cisco.com Thu Sep 22 16:55:45 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id QAA01864; Thu, 22 Sep 1994 16:55:44 -0700 Received: from virginia.edu (uvaarpa.Virginia.EDU [128.143.2.7]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id SAA19411 for ; Thu, 22 Sep 1994 18:07:23 -0700 Received: from hopper.itc.virginia.edu by uvaarpa.virginia.edu id aa10623; 22 Sep 94 21:07 EDT Received: (from mike@localhost) by Hopper.itc.Virginia.EDU (8.6.8/8.6.6) id VAA56202 for vsta@cisco.com; Thu, 22 Sep 1994 21:07:14 -0400 From: "Michael M. Chapman" Message-Id: <199409230107.VAA56202@Hopper.itc.Virginia.EDU> Subject: perm problem... To: vsta@cisco.com Date: Thu, 22 Sep 1994 21:07:14 -0400 (EDT) X-Mailer: ELM [version 2.4 PL23] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 342 I have group sys.sys... I'll hack around some more with it, but I guess I'd do best to use a flat filesystem. What's the timing on a 1.4 release? Thanks. Also I am receiving two copies of each mailing list post and they're coming from the user who sent them so I cannot reply straight to the list... Having fun even w/ the problems. :-) From vandys@glare.cisco.com Fri Sep 23 00:02:27 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id AAA01906; Fri, 23 Sep 1994 00:02:26 -0700 Received: from swann.irisa.fr (swann.irisa.fr [131.254.43.10]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with ESMTP id BAA01014 for ; Fri, 23 Sep 1994 01:14:06 -0700 Received: from oblomov.irisa.fr (oblomov.irisa.fr [131.254.43.11]) by swann.irisa.fr (8.6.9/8.6.9) with ESMTP id KAA24620; Fri, 23 Sep 1994 10:13:31 +0200 From: Tommy Thorn Date: Fri, 23 Sep 1994 10:13:30 +0200 Message-Id: <199409230813.KAA22208@oblomov.irisa.fr> To: cmaeda@cs.washington.edu Subject: Re: Looking for a project... Cc: vsta@cisco.com > From: Chris Maeda > > You could say the same thing for any file system implementation that > supports the vnode interface. Building in support for vnodes will > allow you to easily drop in any vnode file system implementation > (cd-rom, nfs, dos, afs, ufs, etc). Right. BSD uses some kind of "vfs", but actually dropping in FFS proved so difficult, that all the people I know try it, has given up on it. Linux also uses vfs, so you have both choises. Back to VSTa. Period. /Tommy From vandys@glare.cisco.com Fri Sep 23 00:00:29 1994 Received: from dustbin.cisco.com (dustbin.cisco.com [131.108.1.27]) by amri.cisco.com (8.3/8.3) with ESMTP id AAA01903; Fri, 23 Sep 1994 00:00:27 -0700 Received: from post.demon.co.uk (post.demon.co.uk [158.152.1.72]) by dustbin.cisco.com (8.6.8+c/CISCO.SUB_GATE.1.1) with SMTP id BAA10031 for ; Fri, 23 Sep 1994 01:12:07 -0700 Received: from humbug.demon.co.uk by post.demon.co.uk id aa00872; 23 Sep 94 8:57 GMT-60:00 Received: by humbug.demon.co.uk (Linux Smail3.1.28.1 #1) id m0qo5Sh-0003BSC; Fri, 23 Sep 94 08:54 BST Message-Id: From: Dave Hudson Subject: Re: weird permissions problems To: "Michael M. Chapman" Date: Fri, 23 Sep 1994 08:54:31 +0100 (BST) Cc: VSTa mailing list In-Reply-To: <199409222220.SAA16733@Hopper.itc.Virginia.EDU> from "Michael M. Chapman" at Sep 22, 94 06:20:42 pm X-Mailer: ELM [version 2.4 PL23] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 642 > > I cannot compile with gcc because when it goes to write the object > file it prints "perm" before exiting and then perm after > exiting. What does this mean? I think this sounds like the stack growth problem (cc1 dies if memory serves). I don't have the patch handy (my mailer's not that clever) but Andy posted it on 3rd June 1994 so it should be in the archive. > Also, I cannot use the arrow keys with the microemacs port. I'd > rather use vi anyway, but seeing as I can't compile yet... Any > help *greatly* appreciated. If someone has the keybindings I'd be grateful - I'm fed up with Ctrl-P and Ctrl-N! Regards, Dave From vandys@glare.cisco.com Fri Sep 23 05:50:02 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id FAA01961; Fri, 23 Sep 1994 05:50:00 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id HAA29900; Fri, 23 Sep 1994 07:01:42 -0700 Message-Id: <199409231401.HAA29900@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: Dave Hudson Cc: "Michael M. Chapman" , VSTa mailing list Subject: Re: weird permissions problems In-Reply-To: Your message of "Fri, 23 Sep 1994 08:54:31 BST." Date: Fri, 23 Sep 1994 07:01:41 -0700 From: Andrew Valencia [Dave Hudson writes:] >> I cannot compile with gcc because when it goes to write the object >> file it prints "perm" before exiting and then perm after >> exiting. What does this mean? >I think this sounds like the stack growth problem (cc1 dies if memory >serves). I don't have the patch handy (my mailer's not that clever) but >Andy posted it on 3rd June 1994 so it should be in the archive. I had another idea. Did you clone the "root" user entry, by chance? This causes problems, because the first ID a user possesses is used to tag the default permission of objects, and a zero-length ID, while granting all privileges, is not useful as a default permission. Andy From vandys@glare.cisco.com Fri Sep 23 06:26:37 1994 Received: from dustbin.cisco.com (dustbin.cisco.com [131.108.1.27]) by amri.cisco.com (8.3/8.3) with ESMTP id GAA01971; Fri, 23 Sep 1994 06:26:35 -0700 Received: from wotan.compaq.com (wotan.compaq.com [131.168.249.254]) by dustbin.cisco.com (8.6.8+c/CISCO.SUB_GATE.1.1) with SMTP id HAA15737 for ; Fri, 23 Sep 1994 07:38:18 -0700 Received: from twisto by wotan.compaq.com with uucp (Smail3.1.28.1 #11) id m0qoBgw-000vIpC; Fri, 23 Sep 94 09:33 CDT Received: from gocart.eng.hou.compaq.com by twisto.eng.hou.compaq.com with smtp (Smail3.1.28.1 #10) id m0qoBen-000uHSC; Fri, 23 Sep 94 09:31 CDT Received: by gocart.eng.hou.compaq.com (Smail3.1.26.7/COMPAQ-HESIOD) id m0qoBem-00036pC; Fri, 23 Sep 94 09:31 CDT Message-Id: From: edm@gocart.eng.hou.compaq.com (Edward Mccreary) Subject: Re: Looking for a project To: vsta@cisco.com Date: Fri, 23 Sep 94 9:31:24 CDT X-Mailer: ELM [version 2.4dev PL11] Hi guys, I've decide to work on a ext2 filesystem server instead of a BSD one for a couple of reasons. First, Linux was much easier to install on my second hard drive than the BSD's. Also, a great deal of the filesystem code is very nicely isolated in it's own library making it easy to browse. Finally, there's a Mach server in development which gives me another code base to browse through to steal^H^H^H^H^H borrow ideas from. This is beyond anything I've attepmpted before so please excuse any stupid questions I post over the next few weeks. Then again, if it was easy it wouldn't be any fun. :) One thing I was thinking about was the buffer cache. Should this be written as a seperate server? I was thinking that if it was isolated from the ext2 code it could be used again to write a BSD FFS or Minix fs server. I think Dave Hudson mentioned that Mach does something similar. One other thing. Does the wd server work with a second IDE driver right now? I could have sworn that I read a comment saying not to try it but when I went throught the src to vsta last night I couldn't find it anywhere. It won't be an issue right away, I'm planning on writing server to work with a file in the same was vstafs and vfs does, but eventually I'd like to be able to read from my Linux partition. (btw, I'm posting this to the list instead of replying individually. I hope y'all don't mind but I'm already behind in my mail. ) regards, eddie -- Eddie McCreary edm@twisto.compaq.com Graphics Development "Do or do not, there is no 'try'." In the event of my capture, Compaq will disavow any and all knowledge of my operations. Of course I don't speak for them. From vandys@glare.cisco.com Fri Sep 23 06:35:29 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id GAA01974; Fri, 23 Sep 1994 06:35:28 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id HAA00642; Fri, 23 Sep 1994 07:47:11 -0700 Message-Id: <199409231447.HAA00642@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: edm@gocart.eng.hou.compaq.com (Edward Mccreary) Cc: vsta@cisco.com Subject: Re: Looking for a project In-Reply-To: Your message of "Fri, 23 Sep 1994 09:31:24 CDT." Date: Fri, 23 Sep 1994 07:47:10 -0700 From: Andrew Valencia [edm@gocart.eng.hou.compaq.com (Edward Mccreary) writes:] >One thing I was thinking about was the buffer cache. Should this be >written as a seperate server? I was thinking that if it was isolated >from the ext2 code it could be used again to write a BSD FFS or Minix >fs server. I think Dave Hudson mentioned that Mach does something >similar. I'd make it a library. Make sure it supports variable-length blocks (or expect me to come in hacking on it as soon as you're done!). We should try to make the interface allow multi-threading, even if you don't do that for the first iteration. >One other thing. Does the wd server work with a second IDE driver >right now? I could have sworn that I read a comment saying not to >try it but when I went throught the src to vsta last night I couldn't >find it anywhere. It won't be an issue right away, I'm planning on >writing server to work with a file in the same was vstafs and vfs does, >but eventually I'd like to be able to read from my Linux partition. Yup, I believe Dave got multiple IDE drives working. I'm thinking about dumping the current state of the world as v1.3.2, just to let you all get access to the work in progress. In most ways it's an improvement over 1.3.1, but I'm sure there are places where we've moved backwards. Send private E-mail to me if you have strong feelings one way or the other. Andy From vandys@glare.cisco.com Fri Sep 23 08:44:23 1994 Received: from dustbin.cisco.com (dustbin.cisco.com [131.108.1.27]) by amri.cisco.com (8.3/8.3) with ESMTP id IAA02002; Fri, 23 Sep 1994 08:44:22 -0700 Received: from zork.tiac.net (zork.tiac.net [199.0.65.2]) by dustbin.cisco.com (8.6.8+c/CISCO.SUB_GATE.1.1) with ESMTP id JAA18074; Fri, 23 Sep 1994 09:56:03 -0700 Received: from max.tiac.net (max.tiac.net [199.0.65.4]) by zork.tiac.net (8.6.9/8.6.6.Beta9) with ESMTP id MAA20729; Fri, 23 Sep 1994 12:45:31 -0400 Received: (from chrisp@localhost) by max.tiac.net (8.6.9/8.6.6.Beta9) id MAA04539; Fri, 23 Sep 1994 12:45:31 -0400 Date: Fri, 23 Sep 1994 12:45:31 -0400 From: Chris Patti { Feoh } Message-Id: <199409231645.MAA04539@max.tiac.net> To: edm@gocart.eng.hou.compaq.com, vandys@cisco.com Subject: Re: Looking for a project Cc: vsta@cisco.com Yes! Please make the 1.3.2 dump. I and about 5 others who've been waiting with bated breath for the new release will be thrilled! :) From vandys@glare.cisco.com Fri Sep 23 11:40:05 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id LAA02052; Fri, 23 Sep 1994 11:40:04 -0700 Received: from wotan.compaq.com (wotan.compaq.com [131.168.249.254]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id MAA02240 for ; Fri, 23 Sep 1994 12:51:41 -0700 Received: from twisto by wotan.compaq.com with uucp (Smail3.1.28.1 #11) id m0qoGaL-000vJOC; Fri, 23 Sep 94 14:47 CDT Received: from gocart.eng.hou.compaq.com by twisto.eng.hou.compaq.com with smtp (Smail3.1.28.1 #10) id m0qoGY9-000uHrC; Fri, 23 Sep 94 14:44 CDT Received: by gocart.eng.hou.compaq.com (Smail3.1.26.7/COMPAQ-HESIOD) id m0qoGY8-00036pC; Fri, 23 Sep 94 14:44 CDT Message-Id: From: edm@gocart.eng.hou.compaq.com (Edward Mccreary) Subject: re: Looking for a project To: vsta@cisco.com Date: Fri, 23 Sep 94 14:44:52 CDT X-Mailer: ELM [version 2.4dev PL11] Bryan Ford wrote... > Since VSTa and the Hurd have many similarities in overall file system architec ture, > you might want to try this buffer cache library (and possibly other libraries) > to match the interfaces of the corresponding Hurd libraries. > That way you could simply "drop in" Louis-D. Dubeau's ext2fs-on-Hurd code, > _and_ Michael Bushnell's BSD FFS Hurd code, and any other Hurd file systems > that may come into existence in the future. I think the Hurd's file system > helper libraries encapsulate almost all of the major Mach dependencies, so > this may not be all that difficult. > Good idea. I think I'll download the latest Hurd snapshot this weekend and see how they implemented their filesystem libraries. I've already downloaded the ext2fs Hurd server but I haven't had time to look at it yet. regards, eddie -- Eddie McCreary edm@twisto.compaq.com Graphics Development "Do or do not, there is no 'try'." In the event of my capture, Compaq will disavow any and all knowledge of my operations. Of course I don't speak for them. From vandys@glare.cisco.com Sun Sep 25 09:51:40 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id JAA02382; Sun, 25 Sep 1994 09:51:39 -0700 Received: from virginia.edu (uvaarpa.Virginia.EDU [128.143.2.7]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id LAA12782 for ; Sun, 25 Sep 1994 11:03:34 -0700 Received: from hopper.itc.virginia.edu by uvaarpa.virginia.edu id aa13600; 25 Sep 94 14:03 EDT Received: (from mike@localhost) by Hopper.itc.Virginia.EDU (8.6.8/8.6.6) id NAA30875 for vsta@cisco.com; Sun, 25 Sep 1994 13:54:09 -0400 From: "Michael M. Chapman" Message-Id: <199409251754.NAA30875@Hopper.itc.Virginia.EDU> Subject: sites, newsgroup, etc. To: vsta@cisco.com Date: Sun, 25 Sep 1994 13:54:09 -0400 (EDT) X-Mailer: ELM [version 2.4 PL23] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 440 To encourage large scale porting of software and development, perhaps we should have a standard site with an incoming dir for people to put stuff they've already ported in. Also, a newsgroup would be *great* - I've been telling everyone I can to try out VSTa, but that would help spread the word even better. Some coordination and easier access to ported software and different distributions would be GREAT. What does everyone think? From vandys@glare.cisco.com Sun Sep 25 13:18:09 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id NAA02402; Sun, 25 Sep 1994 13:18:08 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id OAA18041; Sun, 25 Sep 1994 14:30:04 -0700 Message-Id: <199409252130.OAA18041@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: "Michael M. Chapman" Cc: vsta@cisco.com Subject: Re: sites, newsgroup, etc. In-Reply-To: Your message of "Sun, 25 Sep 1994 13:54:09 EDT." <199409251754.NAA30875@Hopper.itc.Virginia.EDU> Date: Sun, 25 Sep 1994 14:30:04 -0700 From: Andrew Valencia ["Michael M. Chapman" writes:] >To encourage large scale porting of software and development, >perhaps we should have a standard site with an incoming dir >for people to put stuff they've already ported in. Sounds good! >Also, a newsgroup would be *great* - I've been telling everyone >I can to try out VSTa, but that would help spread the word >even better. Well, I asked around a bit. The "rule of thumb" is that a mailing list is ripe for conversion to a newsgroup when it reaches 500 people. We're at 150, which is why I haven't really been pushing this idea. >Some coordination and easier access to ported software and >different distributions would be GREAT. What does everyone >think? We have an offer of some space on a Linux machine which will soon be installed. I also am pursuing the possibility of placing a VSTa server on the Internet via a site in Palo Alto, California. The former would clearly have fewer initial glitches; the latter shakes out the software for us. I'm well into coding of the equivalent of "ld.so" right now; the utility to create the mappable libraries is done and working. When it works I'm going to recompile everything (except boot servers), package it up, and call it v1.3.1. Andy From vandys@glare.cisco.com Sun Sep 25 14:25:25 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id OAA02427; Sun, 25 Sep 1994 14:25:24 -0700 Received: from ebt.com (spock.ebt.com [192.111.115.3]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id PAA18167 for ; Sun, 25 Sep 1994 15:37:21 -0700 Received: from ebt-inc.ebt.com by ebt.com (4.1/SMI-4.1) id AA14143; Sun, 25 Sep 94 18:38:26 EDT Received: by ebt-inc.ebt.com (5.0/SMI-SVR4) id AA00492; Sun, 25 Sep 1994 18:38:26 +0500 Date: Sun, 25 Sep 1994 18:38:26 +0500 From: gtn@ebt.com (Gavin Nicol) Message-Id: <9409252238.AA00492@ebt-inc.ebt.com> To: vsta@cisco.com Subject: ld.so, threads Content-Length: 1029 Andy, what is the general technique used fir the shared libraries? We don't need specifics, but I for one would like to know whether it's jump tables, PIC or something else. Also, has anyone thought of writing some wrappers for VSTa threads to make them POSIX threads compatible? Is this worth it? I think that as everyone gets used to multithreaded programming, more and more of the servers will become multithreaded; MADO certainly will be. We also need more threadsafe library code; things like lists, queues, trees, semaphores... Finally, now that shared libraries look like they'll be coming along, I think we should probably start moving over toward unicode support. I have the beginnings of a wide string library done, but I stopped work on it partly because of the size of the table needed for things like ispunct() et al. Once I get a few hours, I'll dust off the code, and finish it off. I'll also hack on doprnt. I want MADO and bitblt to be able to handle Unicode... Gavin "Threads help structure programs" Nicol From vandys@glare.cisco.com Sun Sep 25 15:39:24 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id PAA02441; Sun, 25 Sep 1994 15:39:23 -0700 Received: from post.demon.co.uk (post.demon.co.uk [158.152.1.72]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id QAA19456 for ; Sun, 25 Sep 1994 16:51:19 -0700 Received: from humbug.demon.co.uk by post.demon.co.uk id aa23793; 26 Sep 94 0:47 GMT-60:00 Received: by humbug.demon.co.uk (Linux Smail3.1.28.1 #1) id m0qp3B1-0003VqC; Mon, 26 Sep 94 00:40 BST Message-Id: From: Dave Hudson Subject: Re: ld.so, threads To: VSTa mailing list Date: Mon, 26 Sep 1994 00:40:14 +0100 (BST) X-Mailer: ELM [version 2.4 PL23] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 2221 Gavin Nicol wrote: > > Andy, what is the general technique used fir the shared libraries? We > don't need specifics, but I for one would like to know whether it's > jump tables, PIC or something else. I'd be interested in this one as judging from the debates raging in the Linux world this could have an interesting effect on performance. > Also, has anyone thought of writing some wrappers for VSTa threads to > make them POSIX threads compatible? Is this worth it? I think that as > everyone gets used to multithreaded programming, more and more of the > servers will become multithreaded; MADO certainly will be. We also > need more threadsafe library code; things like lists, queues, trees, > semaphores... I didn't think the POSIX spec had stabilised yet (or am I just misinformed?). The biggest problem I've seen with implementing thread safe libraries (eg libc) is that we're using kernel threads. We need a mechanism that would allow any per-thread data used in the libraries to be changed during a thread context-switch. This must have been solved before, but I don't know how (the only solution I've thought of so far wasn't exactly elegant). I think that for any sort of async I/O multithreaded must be the way to go - the biggest argument against this seems to be debug support (there seem to be some good flame wars raging on this in some of the O/S newsgroups). Having written MS-Windows apps in the past though I'm no fan of purely event-driven apps. > Finally, now that shared libraries look like they'll be coming along, > I think we should probably start moving over toward unicode support. I > have the beginnings of a wide string library done, but I stopped work > on it partly because of the size of the table needed for things like > ispunct() et al. Once I get a few hours, I'll dust off the code, and > finish it off. I'll also hack on doprnt. I want MADO and bitblt to be > able to handle Unicode... Unicode would be really good generally - would we need to see about supporting multibyte characters in gcc? (I seem to remember reading that the Plan 9 folks had done that). As for bitblt, I don't doubt for a second that we we'll have multibyte support :-) :-) Regards, Dave From vandys@glare.cisco.com Sun Sep 25 17:14:16 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id RAA02462; Sun, 25 Sep 1994 17:14:15 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id SAA20392 for ; Sun, 25 Sep 1994 18:26:12 -0700 Message-Id: <199409260126.SAA20392@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol Date: Sun, 25 Sep 1994 18:26:10 -0700 From: Andrew Valencia Subject: Re: ld.so, threads Apparently-To: ------- Blind-Carbon-Copy To: gtn@ebt.com (Gavin Nicol) Subject: Re: ld.so, threads In-reply-to: Your message of "Sun, 25 Sep 1994 18:38:26 +0500." <9409252238.AA00492@ebt-inc.ebt.com> Date: Sun, 25 Sep 1994 18:26:10 -0700 From: Andrew Valencia [gtn@ebt.com (Gavin Nicol) writes:] >Andy, what is the general technique used fir the shared libraries? We >don't need specifics, but I for one would like to know whether it's >jump tables, PIC or something else. Ok, here's a brief summary of what I'm coding up. Please send comments and discussion to vsta-shlib@cisco.com, not the main list! Each library to be shared has a .db file which tabulates which globals should be visible to library clients, and which should be hidden. By splitting this into its own file, it's easy to spot if an existing entry point is "lost", or spot something new which you forgot to place in the .db. This file is a simple text file which you edit, and is maintained under RCS. All the .o's of a library are put into _s.a. This is the static version, and would be used by boot servers and others who would need to run before filesystems and name servers are available. All the .o's are also built, using -r, into .tmp. This is a single object file, but is still relocatable (because of -r). This is the file which is processed to create the sharable version. The utility "mkshlib" is invoked with a list of .db's corresponding to the libraries which are to be shared. For each library, it reads the .db and then reads the .tmp, checking for extra or missing symbols. For each entry, it generates a .s file in /tmp which essentially says: 0: load index of this entry point load base of jump table if non-zero call via jump table indexed by entry index else call loader goto 0, and try again These .s's are all assembled and placed into .a, along with an extra .o at the end which provides the "loader" function. The "loader" is done in two stages. The code in each shared library knows just enough to load the main code for dealing which mapping in a library. This is so new functionality can be added to the lookup/load phase in the future (load paths, different library formats, whatever). The real "loader" initially will just look up the library name in the root filesystem, read in its header to get the address it was relocated for, and then mmap() the file at that address. To be more exact, it mmap()'s the text read only, the data copy-on-write, and mmap()'s zero-fill-on-demand for the BSS. An extra .o is added to .tmp which provides a jump table out to each entry point in the code. Shortcomings... no data is made available from the C library, nor can the C library make use of any data from user code. I fiddled things like stdio's __iob[] array (C startup calls __get_iob() to get the address now) and __ctab[] for (same trick). getopt() had similar variables, which were replaced with macros much like errno's (check the source). Strengths... except for the initial pause while the library is mapped in, calling speed is slowed down only by the cost of one extra table lookup and one extra jump. C library code is fully optimized, so it'll run as fast as ever. The C library memory is shared among all clients. The library setup should be pretty quick, since the code is already relocated. The .db file should prevent unexpected missing entry points from crashing your application at an inconvenient time. Like all shared libraries, you can update all clients by just updating the library. Unlike some, a client can't override a function in the C library. Well, he can for himself, but not for the library code. Support for multiple versions of libraries is possible, since the loader is also picked up dynamically from the filesystem during shared library map-in. I won't support this on first release; usually it ends up being a big can of worms, rather than much help. Since each entry is burst into its own .o in the .a, it doesn't matter that you placed a bunch of functions into a single .o in your C library source. The granularity on the client side is always per-function. There you have it! Andy ------- End of Blind-Carbon-Copy From vandys@glare.cisco.com Mon Sep 26 04:01:00 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id EAA02628; Mon, 26 Sep 1994 04:00:59 -0700 Received: from ftp.std.com (ftp.std.com [192.74.137.7]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with ESMTP id FAA09032 for ; Mon, 26 Sep 1994 05:12:58 -0700 Received: from world.std.com by ftp.std.com (8.6.8.1/Spike-8-1.0) id IAA28005; Mon, 26 Sep 1994 08:12:27 -0400 Received: by world.std.com (5.65c/Spike-2.0) id AA13726; Mon, 26 Sep 1994 08:12:26 -0400 From: larz@world.std.com (Mike A Larson) Message-Id: <199409261212.AA13726@world.std.com> Subject: deref_port: messages To: vsta@cisco.com Date: Mon, 26 Sep 1994 08:12:26 -0400 (EDT) X-Mailer: ELM [version 2.4 PL23] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 345 Hi, > Assertion failed line 75 file ../kern/msgcon.c > deref_port: messages Has this been fixed? The person that is testing SCSI/CAM MO support, Alistair G. Crooks (agc@uts.amdahl.com), is seeing this problem on his system. I can supply additional info. (system type, stack trace, etc.) if it would be helpful. Thanks. Mike Larson From vandys@glare.cisco.com Mon Sep 26 05:46:00 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id FAA02637; Mon, 26 Sep 1994 05:45:59 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id GAA06322; Mon, 26 Sep 1994 06:57:59 -0700 Message-Id: <199409261357.GAA06322@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: larz@world.std.com (Mike A Larson) Cc: vsta@cisco.com Subject: Re: deref_port: messages In-Reply-To: Your message of "Mon, 26 Sep 1994 08:12:26 EDT." <199409261212.AA13726@world.std.com> Date: Mon, 26 Sep 1994 06:57:59 -0700 From: Andrew Valencia [larz@world.std.com (Mike A Larson) writes:] >> Assertion failed line 75 file ../kern/msgcon.c >> deref_port: messages Yes, I finally hunted this down when one of my laptops started doing it sporadically. It's an M_ISR message which pops up between the time the first client tries to connect and when it is disconnected because the driver isn't ready for clients (he's reading the partition table). When the kernel throws out the last client, he notices that there are no more clients, and asserts that the message queue should be empty. Of course, interrupt messages can/should still arrive, so I changed the assert to be a little more sophisticated. As a workaround, you can just comment out that assertion. Andy From vandys@glare.cisco.com Mon Sep 26 08:11:57 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id IAA02705; Mon, 26 Sep 1994 08:11:56 -0700 Received: from ns.Novell.COM (ns.Novell.COM [137.65.1.1]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id JAA15442 for ; Mon, 26 Sep 1994 09:23:57 -0700 Received: from glare.cisco.com by ns.Novell.COM (4.1/SMI-4.1) id AA20979; Mon, 26 Sep 94 10:23:31 MDT Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id JAA10104; Mon, 26 Sep 1994 09:23:15 -0700 Message-Id: <199409261623.JAA10104@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: gary_shea@novell.com Cc: vsta <@novell.com:vsta@cisco.com> Subject: Re: VSTa won't boot past the 'Launch at 0x1020' line In-Reply-To: Your message of "Mon, 26 Sep 1994 10:16:00 MDT." <2e86f3eb0.31c@orb.usg.sandy.novell.com> Date: Mon, 26 Sep 1994 09:23:15 -0700 From: Andrew Valencia [gary_shea@novell.com writes:] >65504+36864+4224 >Launch at 0x1020 >the disk whirs a bit.... then nothing. Have to press the reset button >to restart. Can you verify whether your display puts text at the CGA or MGA locations? If you're seeing nothing, but disk accesses are happening, this usually means that your display is set for MGA, whereas VSTa uses the CGA locations by default. >I'm using a 33MHz 386SX board with genuine MS-DOS, a vanilla >monochrome monitor (i.e., no graphics), 4 Meg of memory all of which >tests out ok, and a 60 Meg IDE drive on a separate conroller card >(I don't think there are controllers on this mini-386 motherboard). Mono monitor, but what kind of display card? A Hercules-style would indeed use MGA locations, which means you have to tell the console server about it. Andy From vandys@glare.cisco.com Mon Sep 26 08:07:12 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id IAA02702; Mon, 26 Sep 1994 08:07:11 -0700 Received: from ns.Novell.COM (ns.Novell.COM [137.65.1.1]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id JAA15135 for ; Mon, 26 Sep 1994 09:19:11 -0700 From: gary_shea@novell.com Received: from orb.usg.sandy.novell.com by ns.Novell.COM (4.1/SMI-4.1) id AA20851; Mon, 26 Sep 94 10:18:53 MDT To: @novell.com:vsta@cisco.com Date: Mon, 26 Sep 1994 10:16 MDT Subject: VSTa won't boot past the 'Launch at 0x1020' line Content-Length: 1039 Content-Type: text/plain Message-Id: <2e86f3eb0.31c@orb.usg.sandy.novell.com> Finally got tired of just reading the code and decided to try and boot the beast! Having some problems, though. I get (roughly): 65504+36864+4224 Launch at 0x1020 the disk whirs a bit.... then nothing. Have to press the reset button to restart. I seem to remember seeing this before on the list, the fix being to remove all drivers from the config files. I renamed my config files, and it still won't boot. I'm using the vsta_bin binaries... Hardly needed to touch the boot.lst (?) file, except for the disk params. Tried all three disk-config styles, and they all seem to work, i.e., the disk goes crazy while loading the boot image and drivers (I think!), the light goes out for a sec, then it does a bit more accessing.... then silence. I'm using a 33MHz 386SX board with genuine MS-DOS, a vanilla monochrome monitor (i.e., no graphics), 4 Meg of memory all of which tests out ok, and a 60 Meg IDE drive on a separate conroller card (I don't think there are controllers on this mini-386 motherboard). Any suggestions? Gary From vandys@glare.cisco.com Mon Sep 26 08:57:43 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id IAA02727; Mon, 26 Sep 1994 08:57:42 -0700 Received: from netcom.netcom.com (root@netcom.netcom.com [192.100.81.100]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with ESMTP id KAA17705 for ; Mon, 26 Sep 1994 10:09:43 -0700 Received: by netcom.netcom.com (8.6.9/Netcom) id KAA04084; Mon, 26 Sep 1994 10:03:29 -0700 Date: Mon, 26 Sep 1994 10:03:29 -0700 From: kan@netcom.com (Kan Yu) Message-Id: <199409261703.KAA04084@netcom.netcom.com> To: vsta@cisco.com Subject: fdisk utility and others Hi! I finally got the time and space to boot up the VSTA on my 486 clone over the weekend. Although thing was running pretty well I did have some questions to ask here. 1) How can I mount the second partition on the drive? Is "dos" the utility to use? What is the command convention? 2) Is there utiltiy that allows you to reboot the system? How'bout shutdown? 3) I have changed the login program so that it reads machine name from a file under "/vsta/etc/machid" and use it as the login prompt. Would this be alright? In the same program, I also skip printing out the "Password" prompt if user has no password associated with his/her login. 4) Is there vi-clone available? How'bout the windows system? Will the next release include these and possibly KAQ9? I plan to run SLIP client on my VSTA machine. Any suggestion? Thanks for bringing such a wonderful environment to the public domain? Jacob From vandys@glare.cisco.com Mon Sep 26 09:10:05 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id JAA02742; Mon, 26 Sep 1994 09:10:04 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id KAA13287; Mon, 26 Sep 1994 10:22:01 -0700 Message-Id: <199409261722.KAA13287@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: kan@netcom.com (Kan Yu) Cc: vsta@cisco.com Subject: Re: fdisk utility and others In-Reply-To: Your message of "Mon, 26 Sep 1994 10:03:29 PDT." <199409261703.KAA04084@netcom.netcom.com> Date: Mon, 26 Sep 1994 10:22:00 -0700 From: Andrew Valencia [kan@netcom.com (Kan Yu) writes:] >1) How can I mount the second partition on the drive? Is "dos" the utility > to use? What is the command convention? Well, first you'd best take a look at what partitions the driver found. Assuming this is an IDE disk, you'd: vsta$ mount disk/wd /wd vsta$ ls /wd and you might get a list including: wd0 wd0_dos0 wd0_dos1 Then you could unmount /wd from your private namespace: vsta$ umount /wd And start up a DOS filesystem server for the second partition (in background): vsta$ /vsta/boot/dos -p disk/wd:wd0_dos1 fs/dos2 & Hopefully it won't complain. Then mount the new service in your namespace: vsta$ mount fs/dos2 /dos vsta$ ls /dos and you should see your DOS files on the second partition. >2) Is there utiltiy that allows you to reboot the system? How'bout shutdown? Well, ^Z puts you in the kernel debugger, and then "reboot" will do it. >3) I have changed the login program so that it reads machine name from a > file under "/vsta/etc/machid" and use it as the login prompt. Would > this be alright? In the same program, I also skip printing out the > "Password" prompt if user has no password associated with his/her > login. Well, the former is more usually handled with the "banner" file in /vsta/etc. The latter has already been added to the 1.3.2 source. Host names are tricky, as we want it to fit with networking, which may have multiple interfaces with distinct IP addresses for each (and by implication, distinct hostnames). >4) Is there vi-clone available? How'bout the windows system? Will the next > release include these and possibly KAQ9? I plan to run SLIP client on > my VSTA machine. SLIP will take more work; ne2000's work pretty well. This will be in 1.3.2. vi will, too. >Any suggestion? Thanks for bringing such a wonderful environment to the >public domain? Thanks, but it's not public domain. I continue to hold copyright to the system, and allow (plural) you to use it under Copyleft. See /vsta/license to see what the terms of the system's use are. Hopefully this is acceptable to you; it's the same terms as GNU C uses, for instance. Andy From vandys@glare.cisco.com Mon Sep 26 10:26:20 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id KAA02777; Mon, 26 Sep 1994 10:26:19 -0700 Received: from post.demon.co.uk (post.demon.co.uk [158.152.1.72]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id LAA22691 for ; Mon, 26 Sep 1994 11:38:18 -0700 Received: from humbug.demon.co.uk by post.demon.co.uk id aa24015; 26 Sep 94 19:35 GMT-60:00 Received: by humbug.demon.co.uk (Linux Smail3.1.28.1 #1) id m0qpKXo-0003VpC; Mon, 26 Sep 94 19:12 BST Message-Id: From: Dave Hudson Subject: Re: VSTa won't boot past the 'Launch at 0x1020' line To: VSTa mailing list Date: Mon, 26 Sep 1994 19:12:56 +0100 (BST) In-Reply-To: <199409261623.JAA10104@glare.cisco.com> from "Andrew Valencia" at Sep 26, 94 09:23:15 am X-Mailer: ELM [version 2.4 PL23] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 694 Hi, Andrew Valencia wrote: > > >I'm using a 33MHz 386SX board with genuine MS-DOS, a vanilla > >monochrome monitor (i.e., no graphics), 4 Meg of memory all of which > >tests out ok, and a 60 Meg IDE drive on a separate conroller card > >(I don't think there are controllers on this mini-386 motherboard). > > Mono monitor, but what kind of display card? A Hercules-style would indeed > use MGA locations, which means you have to tell the console server about it. It won't hurt to rebuild the kernel debug code to use mono support as well, otherwise you won't be able to see any traps, etc. I have to do this on a SVGA system that's attached to an old mono monitor :-( Regards, Dave From vandys@glare.cisco.com Mon Sep 26 10:47:43 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id KAA02780; Mon, 26 Sep 1994 10:47:36 -0700 Received: from ns.Novell.COM (ns.Novell.COM [137.65.1.1]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id LAA23790 for ; Mon, 26 Sep 1994 11:59:34 -0700 From: gary_shea@novell.com Received: from by ns.Novell.COM (4.1/SMI-4.1) id AB24613; Mon, 26 Sep 94 12:59:15 MDT To: Andrew Valencia Cc: @novell.com:vsta@cisco.com Date: Mon, 26 Sep 1994 11:35 MDT Subject: Re: VSTa won't boot past the 'Launch at 0x1020' line Content-Length: 1299 Content-Type: text/plain Message-Id: <2e8706540.386@orb.usg.sandy.novell.com> > From: Andrew Valencia > Status: R > > [gary_shea@novell.com writes:] > > >65504+36864+4224 > >Launch at 0x1020 > >the disk whirs a bit.... then nothing. Have to press the reset button > >to restart. > > Can you verify whether your display puts text at the CGA or MGA locations? > If you're seeing nothing, but disk accesses are happening, this usually > means that your display is set for MGA, whereas VSTa uses the CGA locations > by default. I believe that my video card is as dumb as one can be -- too dumb to run Windows, for instance. I don't think it's Hercules. I didn't try using the CGA or MGA settings. Ahhhhh.... VSTa assumes I have a slightly smarter display, eh? Well, I can either buy a new card or try to write a server that works with dodo-mono. I wouldn't mind the latter as a learning experience... would anyone care to comment on whether it's worth doing? Is there a freely available tool I can use to check memory on the DOS beast to see where it's putting the screen stuff? Also, can anyone suggest a book for finding out where DOS keeps such goodies as the screen map (assuming it's mapped, that is!). I don't currently have any kind of compiler or assembler that will run on DOS, although I wouldn't mind getting Borland or something... Gary From vandys@glare.cisco.com Mon Sep 26 10:55:34 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id KAA02784; Mon, 26 Sep 1994 10:55:32 -0700 Received: from ns.Novell.COM (ns.Novell.COM [137.65.1.1]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id MAA24297 for ; Mon, 26 Sep 1994 12:07:30 -0700 Received: from glare.cisco.com by ns.Novell.COM (4.1/SMI-4.1) id AA24814; Mon, 26 Sep 94 13:07:09 MDT Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id MAA19488; Mon, 26 Sep 1994 12:06:55 -0700 Message-Id: <199409261906.MAA19488@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: gary_shea@novell.com Cc: vsta <@novell.com:vsta@cisco.com> Subject: Re: VSTa won't boot past the 'Launch at 0x1020' line In-Reply-To: Your message of "Mon, 26 Sep 1994 11:35:00 MDT." <2e8706540.386@orb.usg.sandy.novell.com> Date: Mon, 26 Sep 1994 12:06:54 -0700 From: Andrew Valencia [gary_shea@novell.com writes:] > Ahhhhh.... VSTa assumes I have a slightly smarter display, eh? >Well, I can either buy a new card or try to write a server that works with dod >o-mono. No, we support mono. You just have to provide the switch to the console server, as I recollect. I think it's the -mono argument to cons in your boot/boot.lst file. >Is there a freely available tool I can use to check memory on >the DOS beast to see where it's putting the screen stuff? Also, can >anyone suggest a book for finding out where DOS keeps such goodies >as the screen map (assuming it's mapped, that is!). I have my IBM-PC/XT hardware reference manual. Some newfangled weirdos have an AT hardware reference manual, but it's multiple volumes, so clearly is padded with "this page intentionally left blank" stuff. :-) >I don't currently have any kind of compiler or assembler that will run on DOS, > although >I wouldn't mind getting Borland or something... Unneeded. Hopefully you can add -mono and get up and running. Then you can run the native compiler. Andy From vandys@glare.cisco.com Mon Sep 26 14:02:56 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id OAA02857; Mon, 26 Sep 1994 14:02:55 -0700 Received: from relay1.Hawaii.Edu (relay1.Hawaii.Edu [128.171.41.53]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id PAA03395 for ; Mon, 26 Sep 1994 15:14:57 -0700 Received: from uhunix.uhcc.Hawaii.Edu ([128.171.44.6]) by relay1.Hawaii.Edu with SMTP id <11444>; Mon, 26 Sep 1994 12:14:43 -1000 Received: by uhunix.uhcc.Hawaii.Edu id <184418>; Mon, 26 Sep 1994 12:14:04 -1000 Subject: Re: VSTa won't boot past the 'Launch at 0x1020' line From: Tim Newsham To: dave@humbug.demon.co.uk (Dave Hudson) Date: Mon, 26 Sep 1994 12:14:00 -1000 Cc: vsta@cisco.com In-Reply-To: ; from "Dave Hudson" at Sep 26, 94 8:12 am X-Mailer: ELM [version 2.3 PL11] Message-Id: <94Sep26.121404hst.184418@uhunix.uhcc.Hawaii.Edu> > > Hi, > > Andrew Valencia wrote: > > Mono monitor, but what kind of display card? A Hercules-style would indeed > > use MGA locations, which means you have to tell the console server about it. > > It won't hurt to rebuild the kernel debug code to use mono support as well, > otherwise you won't be able to see any traps, etc. I have to do this on a > SVGA system that's attached to an old mono monitor :-( hmm.. Maybe this should be part of boot? boot -mono vsta ? > Regards, > Dave From vandys@glare.cisco.com Mon Sep 26 14:21:19 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id OAA02866; Mon, 26 Sep 1994 14:21:18 -0700 Received: from post.demon.co.uk (post.demon.co.uk [158.152.1.72]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id PAA04162 for ; Mon, 26 Sep 1994 15:33:19 -0700 Received: from humbug.demon.co.uk by post.demon.co.uk id aa17275; 26 Sep 94 23:25 GMT-60:00 Received: by humbug.demon.co.uk (Linux Smail3.1.28.1 #1) id m0qpONb-0003VpC; Mon, 26 Sep 94 23:18 BST Message-Id: From: Dave Hudson Subject: Re: VSTa won't boot past the 'Launch at 0x1020' line To: VSTa mailing list Date: Mon, 26 Sep 1994 23:18:39 +0100 (BST) In-Reply-To: <199409261906.MAA19488@glare.cisco.com> from "Andrew Valencia" at Sep 26, 94 12:06:54 pm X-Mailer: ELM [version 2.4 PL23] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 1615 Hi, Andrew Valencia wrote: > > [gary_shea@novell.com writes:] > > > Ahhhhh.... VSTa assumes I have a slightly smarter display, eh? > >Well, I can either buy a new card or try to write a server that works with dod > >o-mono. > > No, we support mono. You just have to provide the switch to the console > server, as I recollect. I think it's the -mono argument to cons in your > boot/boot.lst file. Should do the trick - to force colour usage -colour or -color work :-) FWIW I'd be surprised if there's been a mono card manufactured since the days of the original XTs that doesn't support Hercules 720x348 mono graphics. When bitblt's released it will support this resolution. > >Is there a freely available tool I can use to check memory on > >the DOS beast to see where it's putting the screen stuff? Also, can > >anyone suggest a book for finding out where DOS keeps such goodies > >as the screen map (assuming it's mapped, that is!). > > I have my IBM-PC/XT hardware reference manual. Some newfangled weirdos have > an AT hardware reference manual, but it's multiple volumes, so clearly is > padded with "this page intentionally left blank" stuff. :-) The AT tech reference is good, but showing its age these days. The one I got the other day which really impressed me was one called "The Undocumented PC - A Programmer's Guide To I/O, CPUs And Fixed Memory Areas" by Frank Van Gilluwe. It costs $45 (US) (or 36 quid in the UK - someone's exchange conversion isn't too clever :-(). Even after 4 years designing and spec'ing PCs I found things I'd never come across before :-) Regards, Dave From vandys@glare.cisco.com Tue Sep 27 14:28:48 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id OAA03125; Tue, 27 Sep 1994 14:28:47 -0700 Received: from ns.Novell.COM (ns.Novell.COM [137.65.1.1]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id PAA26026 for ; Tue, 27 Sep 1994 15:40:54 -0700 From: gary_shea@novell.com Received: from orb.usg.sandy.novell.com by ns.Novell.COM (4.1/SMI-4.1) id AA27317; Tue, 27 Sep 94 16:40:36 MDT To: @novell.com:vsta@cisco.com Date: Tue, 27 Sep 1994 16:38 MDT Subject: More boot problems: Can't open /vsta/etc/inittab Content-Length: 528 Content-Type: text/plain Message-Id: <2e889ee80.bf7@orb.usg.sandy.novell.com> I used cons -mono with good effect... I now get two syslog messages before death :) The init program is unable to open /vsta/etc/inittab. It's there.... as c:/vsta/etc/inittab Grep & I have looked throught the archives and have found no evidence of this ever happening before. The connect just before the call to read_inittab (in init) goes fine, then the open() inittab fails... and whatever testsh is, it never shows up. The system just hangs and the reset button must be pushed. Thanks for the ongoing help... Gary From vandys@glare.cisco.com Tue Sep 27 15:15:30 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id PAA03138; Tue, 27 Sep 1994 15:15:28 -0700 Received: from ns.Novell.COM (ns.Novell.COM [137.65.1.1]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id QAA28051 for ; Tue, 27 Sep 1994 16:27:36 -0700 Received: from glare.cisco.com by ns.Novell.COM (4.1/SMI-4.1) id AA28302; Tue, 27 Sep 94 17:27:14 MDT Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id QAA14226; Tue, 27 Sep 1994 16:26:57 -0700 Message-Id: <199409272326.QAA14226@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: gary_shea@novell.com Cc: vsta <@novell.com:vsta@cisco.com> Subject: Re: More boot problems: Can't open /vsta/etc/inittab In-Reply-To: Your message of "Tue, 27 Sep 1994 16:38:00 MDT." <2e889ee80.bf7@orb.usg.sandy.novell.com> Date: Tue, 27 Sep 1994 16:26:56 -0700 From: Andrew Valencia [gary_shea@novell.com writes:] >The init program is unable to open /vsta/etc/inittab. It's there.... as c:/vs >ta/etc/inittab Hmmm... something funny here. Can you provide your inittab? >Grep & I have looked throught the archives and have found no evidence >of this ever happening before. The connect just before the call to read_initt >ab (in init) >goes fine, then the open() inittab fails... and whatever testsh is, it never s >hows up. I'm puzzled. Usually you don't have init *and* testsh launching from boot.lst. It would work after a fashion, though. Anyway, you're saying that init successfully connected to fs/root (which is the connect right before the call to read_inittab()), which means your DOS filesystem is on-line. Is there any chance that the filesystem it's opening is the right one? Do you have more than one DOS filesystem around? Andy From vandys@glare.cisco.com Tue Sep 27 15:32:57 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id PAA03145; Tue, 27 Sep 1994 15:32:56 -0700 Received: from ns.Novell.COM (ns.Novell.COM [137.65.1.1]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id QAA29068 for ; Tue, 27 Sep 1994 16:45:04 -0700 From: gary_shea@novell.com Received: from orb.usg.sandy.novell.com by ns.Novell.COM (4.1/SMI-4.1) id AA28726; Tue, 27 Sep 94 17:44:46 MDT To: @novell.com:vsta@cisco.com Date: Tue, 27 Sep 1994 17:42 MDT Subject: Re: More boot problems: Can't open /vsta/etc/inittab Content-Length: 2079 Content-Type: text/plain Message-Id: <2e88adf30.c67@orb.usg.sandy.novell.com> From vandys@cisco.com Tue Sep 27 17:25 MDT 1994 > >The init program is unable to open /vsta/etc/inittab. It's there.... as c:/vs > >ta/etc/inittab > Hmmm... something funny here. Can you provide your inittab? I'll bring it in tomorrow... > >Grep & I have looked throught the archives and have found no evidence > >of this ever happening before. The connect just before the call to read_initt > >ab (in init) > >goes fine, then the open() inittab fails... and whatever testsh is, it never s > >hows up. > > I'm puzzled. Usually you don't have init *and* testsh launching from > boot.lst. It would work after a fashion, though. Ooops sorry! You may not remember writing it, but if the open of inittab fails, there is an automatic call to testsh -- it's not in the inittab at all. As per Dave's comments, it may be that I'm going into testsh as I should, but that it's not supporting my display? > Anyway, you're saying that init successfully connected to fs/root (which is > the connect right before the call to read_inittab()), which means your DOS > filesystem is on-line. Is there any chance that the filesystem it's opening > is the right one? Do you have more than one DOS filesystem around? I have, to the best of my knowledge, only one fs. This is a 60M disk :) And yes, the connection to fs/root goes ok. Sorry for the vagueness. From: Dave Hudson >Which two servers are displaying syslog info? I suspect that there's a >server dying somewhere, but without having the kernel compiled for a mono >display you won't see the debugger messages. >If you don't have the mono support let me know and I'll uuencode one and >mail it to you - there are a couple of bug fixes in the kernels I'm running >so maybe that will help as well :-) I wish I could remember what the first syslog message was, but it looked benign, and I totally forgot it. I could really use mono support. I'm hesitant about switching kernels after I finally got all the source printed out :) but.... I would love to get this thing running. Thanks! Gary From vandys@glare.cisco.com Wed Sep 28 12:30:24 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id MAA03309; Wed, 28 Sep 1994 12:30:23 -0700 Received: from divi.com (denver.divi.com [199.97.190.1]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id NAA11381 for ; Wed, 28 Sep 1994 13:42:35 -0700 Date: Wed, 28 Sep 94 13:42:59 PDT From: jacob@divi.com (Jacob Yu) Received: from jaroso.divicom ([199.97.190.29]) by divi.com (4.1/3.1.090690-Divicom) id AA03760; Wed, 28 Sep 94 13:42:59 PDT Message-Id: <9409282042.AA03760@divi.com> To: vsta@cisco.com Subject: Re: Re: fdisk utility and others >[kan@netcom.com (Kan Yu) writes:] > >>1) How can I mount the second partition on the drive? Is "dos" the utility >> to use? What is the command convention? > >Well, first you'd best take a look at what partitions the driver found. >Assuming this is an IDE disk, you'd: > >vsta$ mount disk/wd /wd >vsta$ ls /wd > >and you might get a list including: > >wd0 wd0_dos0 wd0_dos1 > >Then you could unmount /wd from your private namespace: > >vsta$ umount /wd > >And start up a DOS filesystem server for the second partition (in background): > >vsta$ /vsta/boot/dos -p disk/wd:wd0_dos1 fs/dos2 & > >Hopefully it won't complain. Then mount the new service in your namespace: > >vsta$ mount fs/dos2 /dos >vsta$ ls /dos > >and you should see your DOS files on the second partition. Currently I have VSTA 1.3.1 installed (from vandys/vsta at ftp.cisco.com) and I couldn't find the "mount" command from the "/vsta/bin" or any of the *.tz files. Is there newer release of VSTA? >>3) I have changed the login program so that it reads machine name from a >> file under "/vsta/etc/machid" and use it as the login prompt. Would >> this be alright? In the same program, I also skip printing out the >> "Password" prompt if user has no password associated with his/her >> login. > >Well, the former is more usually handled with the "banner" file in >/vsta/etc. The latter has already been added to the 1.3.2 source. Host >names are tricky, as we want it to fit with networking, which may have >multiple interfaces with distinct IP addresses for each (and by implication, >distinct hostnames). Maybe I see the "banner" file a little different here. I though the "banner" file was like the "motd" file and will be printed out to the screen when the machine first came up. Is there other way to customize the login prompt? >>4) Is there vi-clone available? How'bout the windows system? Will the next >> release include these and possibly KAQ9? I plan to run SLIP client on >> my VSTA machine. > >SLIP will take more work; ne2000's work pretty well. This will be in 1.3.2. >vi will, too. Will 1.3.2 be available any time soon? Also, is there anything I can help with? I am currently running two 486 PC's connected by a Lantastic network. >Thanks, but it's not public domain. I continue to hold copyright to the >system, and allow (plural) you to use it under Copyleft. See /vsta/license >to see what the terms of the system's use are. > >Hopefully this is acceptable to you; it's the same terms as GNU C uses, for >instance. Sorry, this is my mistake and yes, I have read the copyright statement. Also, is there a porting going on for 68K? What is the possiblity of using VSTA in a commerical word? For instance, use it in a demo package for MPEG-2 decoder development kit? Thanks. Jacob From vandys@glare.cisco.com Wed Sep 28 16:44:37 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id QAA03359; Wed, 28 Sep 1994 16:44:36 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id RAA09071; Wed, 28 Sep 1994 17:56:49 -0700 Message-Id: <199409290056.RAA09071@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: jacob@divi.com (Jacob Yu) Cc: vsta@cisco.com Subject: Re: fdisk utility and others In-Reply-To: Your message of "Wed, 28 Sep 1994 13:42:59 PDT." <9409282042.AA03760@divi.com> Date: Wed, 28 Sep 1994 17:56:48 -0700 From: Andrew Valencia [jacob@divi.com (Jacob Yu) writes:] >Currently I have VSTA 1.3.1 installed (from vandys/vsta at ftp.cisco.com) and >I couldn't find the "mount" command from the "/vsta/bin" or any of the *.tz >files. Is there newer release of VSTA? Since mount tables are per-process and inherited, it has to be a shell built-in. I could've *sworn* we had added it to all shells by 1.3.1.... >Maybe I see the "banner" file a little different here. I though the "banner" >file was like the "motd" file and will be printed out to the screen when >the machine first came up. Is there other way to customize the login prompt? Nope, we have motd too. Banner is emitted before the login prompt. >Will 1.3.2 be available any time soon? Also, is there anything I can help >with? I am currently running two 486 PC's connected by a Lantastic network. 1.3.2 is on the home stretch. I have shared libraries working (plus or minus a bug) and am rebuilding all the binaries, things like that. Out in a week, I'd guess. >Sorry, this is my mistake and yes, I have read the copyright statement. Also, >is there a porting going on for 68K? What is the possiblity of using VSTA in >a commerical word? For instance, use it in a demo package for MPEG-2 >decoder development kit? There is a 68030 port coming alive. I understand it boots and even runs processes. I think a disk driver is in the works next. Commercial products based on VSTa... the copyleft license answers this. Rather than hash this over again, I'd recommend reading the gnu newsgroups. I believe they have a FAQ which covers this in depth. The short answer is that there are restrictions, but many companies have lived within them and offered commercial products. Andy From vandys@glare.cisco.com Thu Sep 29 00:48:09 1994 Received: from beasley.cisco.com (beasley.cisco.com [171.69.2.135]) by amri.cisco.com (8.3/8.3) with ESMTP id AAA03391; Thu, 29 Sep 1994 00:48:07 -0700 Received: from post.demon.co.uk (post.demon.co.uk [158.152.1.72]) by beasley.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id BAA01270 for ; Thu, 29 Sep 1994 01:57:57 -0700 Received: from humbug.demon.co.uk by post.demon.co.uk id aa25485; 29 Sep 94 9:54 GMT-60:00 Received: by humbug.demon.co.uk (Linux Smail3.1.28.1 #1) id m0qqGwG-0004lqC; Thu, 29 Sep 94 09:34 BST Message-Id: From: Dave Hudson Subject: Re: fdisk utility and others To: VSTa mailing list Date: Thu, 29 Sep 1994 09:34:04 +0100 (BST) In-Reply-To: <199409290056.RAA09071@glare.cisco.com> from "Andrew Valencia" at Sep 28, 94 05:56:48 pm X-Mailer: ELM [version 2.4 PL23] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 615 Hi, Andrew Valencia wrote: > > [jacob@divi.com (Jacob Yu) writes:] > > >Currently I have VSTA 1.3.1 installed (from vandys/vsta at ftp.cisco.com) and > >I couldn't find the "mount" command from the "/vsta/bin" or any of the *.tz > >files. Is there newer release of VSTA? > > Since mount tables are per-process and inherited, it has to be a shell > built-in. I could've *sworn* we had added it to all shells by 1.3.1.... They're in there - my notes suggest they got added in late April and then the patches got re-written in early May. Should find mount and umount in ash, rc and testsh. Regards, Dave From vandys@glare.cisco.com Thu Sep 29 05:33:30 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id FAA03528; Thu, 29 Sep 1994 05:33:29 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id GAA04954; Thu, 29 Sep 1994 06:45:36 -0700 Message-Id: <199409291345.GAA04954@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: Dave Hudson Cc: vsta@amri.cisco.com Subject: Re: Copyrights In-Reply-To: Your message of "Thu, 29 Sep 1994 09:52:42 BST." Date: Thu, 29 Sep 1994 06:45:35 -0700 From: Andrew Valencia [Dave Hudson writes:] >The Linux crowd thrashed this one around for a while and in the end moved >all of the libc code over to LGPL'd stuff (coding replacements for GPL'd >things) and then placed the crt0 file under a BSD style copyright since it's >the thing that must be linked with the user's software. With shared libraries we can go one better. I'm pretty sure that I can set things up so that the stub library--which is all that is linked with the application--is public domain. I *think* I can even get crt0 into the shared library. This would allow people to generate commercial products with a minimum of hassle. BTW, I have shared libraries working under VSTa. Right now I'm fiddling with a second-level shared library loader. The extra level of indirection is so that the policy of loading can be updated by replacing the second-level loader. Andy From vandys@glare.cisco.com Thu Sep 29 06:00:48 1994 Received: from beasley.cisco.com (beasley.cisco.com [171.69.2.135]) by amri.cisco.com (8.3/8.3) with ESMTP id GAA03532; Thu, 29 Sep 1994 06:00:47 -0700 Received: from ns.Novell.COM (ns.Novell.COM [137.65.1.1]) by beasley.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id HAA00836; Thu, 29 Sep 1994 07:10:52 -0700 From: gary_shea@novell.com Received: from orb.usg.sandy.novell.com by ns.Novell.COM (4.1/SMI-4.1) id AA14955; Thu, 29 Sep 94 08:12:47 MDT To: @novell.com:vsta@cisco.com Cc: @novell.com:vandys@cisco.com Date: Thu, 29 Sep 1994 08:10 MDT Subject: bootage Content-Length: 1268 Content-Type: text/plain Message-Id: <2e8acaed0.4c9c@orb.usg.sandy.novell.com> Please ignore the '>' prefix... vsta@cisco.com didn't seem to exist when I sent this the first couple times... so I'm still trying. If someone sees this, please let me know, 'cause I'm getting it kicked back undeliverable. > I fiddled with the arguments to the wd server and eventually it booted! > When I fed wd explicit args (gleaned from the cmos) it wouldn't boot, > although it did report those same args in the first (info) syslog message. > Likewise with using the "d0:cmos" option. I got the "right" parameters > from syslog but no boot. When I used the "read from the controller" > option, the syslog message gave different parameters than > the cmos showed, but it booted. > > Pretty exciting! Amazing to see a perfectly usable unix-ish system, brought > over on one diskette (the binaries, anyway!), and running quite quickly in > 4 Megs. I am running a pre-release of a Unix system that really isn't usable > with 16Meg... needs 24-32 to get work done. > > Well, now to start hacking :) I think my first changes will be to get mono > working in the kernel debugger (which I know has already been done). > Does anyone know if it's possible to auto-sense the type of video card in use? > Maybe that would be a useful contribution... > > Gary > From vandys@glare.cisco.com Thu Sep 29 06:01:19 1994 Received: from beasley.cisco.com (beasley.cisco.com [171.69.2.135]) by amri.cisco.com (8.3/8.3) with ESMTP id GAA03535; Thu, 29 Sep 1994 06:01:18 -0700 Received: from ns.Novell.COM (ns.Novell.COM [137.65.1.1]) by beasley.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id HAA00851 for ; Thu, 29 Sep 1994 07:11:25 -0700 From: gary_shea@novell.com Received: from by ns.Novell.COM (4.1/SMI-4.1) id AB14955; Thu, 29 Sep 94 08:13:19 MDT To: @novell.com:vsta@cisco.com Cc: vandys@cisco.com Date: Wed, 28 Sep 1994 11:31 MDT Subject: We have bootage Content-Length: 1140 Content-Type: text/plain Message-Id: <2e89a85b0.1f0f@orb.usg.sandy.novell.com> Please ignore all the '>' char's... for some reason vsta@cisco.com didn't exist this morning :) > I fiddled with the arguments to the wd server and eventually it booted. > When I fed wd explicit args (gleaned from the cmos) it wouldn't boot, > although it did report those same args in the first (info) syslog message. > Likewise with using the "d0:cmos" option. I got the "right" parameters > from syslog but no boot. When I used the "read from the controller" > option, the syslog message gave different parameters than > the cmos showed, but it booted. > > Pretty exciting! Amazing to see a perfectly usable unix-ish system, brought > over on one diskette (the binaries, anyway!), and running quite quickly in > 4 Megs. I am running a pre-release of a Unix system that really isn't usable > with 16Meg... needs 24-32 to get work done. > > Well, now to start hacking :) I think my first changes will be to get mono > working in the kernel debugger (which I know has already been done). > Does anyone know if it's possible to auto-sense the type of video card in use? > Maybe that would be a useful contribution... > > Gary > From vandys@glare.cisco.com Thu Sep 29 06:49:02 1994 Received: from beasley.cisco.com (beasley.cisco.com [171.69.2.135]) by amri.cisco.com (8.3/8.3) with ESMTP id GAA03545; Thu, 29 Sep 1994 06:49:01 -0700 Received: from ns.Novell.COM (ns.Novell.COM [137.65.1.1]) by beasley.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id HAA02003 for ; Thu, 29 Sep 1994 07:58:49 -0700 From: gary_shea@novell.com Received: from orb.usg.sandy.novell.com by ns.Novell.COM (4.1/SMI-4.1) id AA16319; Thu, 29 Sep 94 09:00:41 MDT To: @novell.com:vsta@cisco.com Date: Thu, 29 Sep 1994 08:58 MDT Subject: More questions! Content-Length: 1007 Content-Type: text/plain Message-Id: <2e8ad6280.4cda@orb.usg.sandy.novell.com> If you've been seeing virtually identical messages from me, sorry; I mailed it twice yesterday and they both bounced; once today and it made it. Now the ones I mailed yesterday are showing up!!! Sigh... Last night I tried to compile the system and the first gcc from mkall hung the system. I tried compiling a trivial one-line hello world program and that hung the system. I don't think it's hanging in the kernel debugger (since I have mono I can't see the kernel debugger to tell for sure) because reboot didn't work until after I typed ^Z. Right now reboot is the only kernel debugger command I know :) Any ideas on this one? Most of the other commands I've tried worked fine. The gcc I'm using is the one that's in the binary distribution. The one that's in the vsta distribution directory is huge and I haven't been able to get it to my machine -- I have no modem access and the backup format appears to be incompatible with Novell DOS 7 backup format (I'm going to try again today). Gary From vandys@glare.cisco.com Thu Sep 29 08:35:44 1994 Received: from beasley.cisco.com (beasley.cisco.com [171.69.2.135]) by amri.cisco.com (8.3/8.3) with ESMTP id IAA03564; Thu, 29 Sep 1994 08:35:43 -0700 Received: from post.demon.co.uk (post.demon.co.uk [158.152.1.72]) by beasley.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id JAA05287; Thu, 29 Sep 1994 09:45:45 -0700 Received: from humbug.demon.co.uk by post.demon.co.uk id aa00372; 29 Sep 94 17:01 GMT-60:00 Received: by humbug.demon.co.uk (Linux Smail3.1.28.1 #1) id m0qqNtg-0004lqC; Thu, 29 Sep 94 16:59 BST Message-Id: From: Dave Hudson Subject: Re: Copyrights To: Andrew Valencia Date: Thu, 29 Sep 1994 16:59:52 +0100 (BST) Cc: VSTa mailing list In-Reply-To: <199409291345.GAA04954@glare.cisco.com> from "Andrew Valencia" at Sep 29, 94 06:45:35 am X-Mailer: ELM [version 2.4 PL23] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 1346 Andrew Valencia wrote: > > [Dave Hudson writes:] > > >The Linux crowd thrashed this one around for a while and in the end moved > >all of the libc code over to LGPL'd stuff (coding replacements for GPL'd > >things) and then placed the crt0 file under a BSD style copyright since it's > >the thing that must be linked with the user's software. > > With shared libraries we can go one better. I'm pretty sure that I can set > things up so that the stub library--which is all that is linked with the > application--is public domain. I *think* I can even get crt0 into the > shared library. This would allow people to generate commercial products > with a minimum of hassle. Sounds like a good solution - I like the sound of moving crt0's functionality into the library. I guess we'd just have an even simpler stub that only knows how to load the library and run the functions from crt0? > BTW, I have shared libraries working under VSTa. Right now I'm fiddling > with a second-level shared library loader. The extra level of indirection > is so that the policy of loading can be updated by replacing the > second-level loader. Would this mean that I could go away and implement a new shared library mechanism (if I was that much of a masochist) and not have to recompile all of my binaries? Regards, Dave From vandys@glare.cisco.com Thu Sep 29 08:33:01 1994 Received: from beasley.cisco.com (beasley.cisco.com [171.69.2.135]) by amri.cisco.com (8.3/8.3) with ESMTP id IAA03561; Thu, 29 Sep 1994 08:33:00 -0700 Received: from pine.cse.nau.edu (scd@pine.cse.nau.edu [134.114.64.90]) by beasley.cisco.com (8.6.8+c/CISCO.GATE.1.1) with ESMTP id JAA05206 for ; Thu, 29 Sep 1994 09:43:07 -0700 Received: (from scd@localhost) by pine.cse.nau.edu (8.6.9/2.2-nau) id JAA03936 for vsta@cisco.com; Thu, 29 Sep 1994 09:45:10 -0700 Message-Id: <199409291645.JAA03936@pine.cse.nau.edu> From: scd@pine.cse.nau.edu (Steven Dake) Date: Thu, 29 Sep 1994 09:45:10 -0700 X-Mailer: Mail User's Shell (7.2.5 10/14/92) To: vsta@cisco.com Subject: Developing Drivers etc. I remember reading somewhere in the archives that there was a way to build vsta sources without being in vsta (dos hosted compiles). I'm in the midst of developing some bare-bones SCSI device drivers for a Buslogic scsi board and can't get vsta to boot of course because there is no driver yet. If anyone knows where I might find gcc for dos-hosted compiles (for vsta) I'd appreciate it. -- ------------------------------------------------------------------------------- | Steven C. Dake People new to unix that confuse ctrl-C | | (scd@pine.cse.nau.edu) with ctrl-Z make it all worthwhile. | ------------------------------------------------------------------------------- From vandys@glare.cisco.com Thu Sep 29 10:53:58 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id KAA03590; Thu, 29 Sep 1994 10:53:57 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id MAA17962; Thu, 29 Sep 1994 12:06:16 -0700 Message-Id: <199409291906.MAA17962@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: Dave Hudson Cc: VSTa mailing list Subject: Re: Copyrights In-Reply-To: Your message of "Thu, 29 Sep 1994 16:59:52 BST." Date: Thu, 29 Sep 1994 12:06:16 -0700 From: Andrew Valencia [Dave Hudson writes:] >Would this mean that I could go away and implement a new shared library >mechanism (if I was that much of a masochist) and not have to recompile all >of my binaries? Yes. The only restriction is that the second level shlib loader can have only text, no data or BSS. Of course, nothing stops him from loading a THIRD level loader.... In fact, one can conceive of the libraries converting to a different format, but still being callable from existing a.out programs. Andy From vandys@glare.cisco.com Thu Sep 29 10:54:52 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id KAA03595; Thu, 29 Sep 1994 10:54:51 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id MAA18013; Thu, 29 Sep 1994 12:07:09 -0700 Message-Id: <199409291907.MAA18013@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: scd@pine.cse.nau.edu (Steven Dake) Cc: vsta@cisco.com Subject: Re: Developing Drivers etc. In-Reply-To: Your message of "Thu, 29 Sep 1994 09:45:10 PDT." <199409291645.JAA03936@pine.cse.nau.edu> Date: Thu, 29 Sep 1994 12:07:09 -0700 From: Andrew Valencia [scd@pine.cse.nau.edu (Steven Dake) writes:] >I remember reading somewhere in the archives that there was a way to build >vsta sources without being in vsta (dos hosted compiles). I'm in the midst >of developing some bare-bones SCSI device drivers for a Buslogic scsi board >and can't get vsta to boot of course because there is no driver yet. If >anyone knows where I might find gcc for dos-hosted compiles (for vsta) I'd >appreciate it. Yes, older versions of djgpp, a DOS port of GNU C, use an a.out format entirely compatible with VSTa. When I get home tonight I can bundle it up and put it on FTP for you. This is how I got VSTa bootstrapped, BTW. Andy From vandys@glare.cisco.com Fri Sep 30 15:27:42 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id PAA03762; Fri, 30 Sep 1994 15:27:41 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id QAA23259 for ; Fri, 30 Sep 1994 16:40:06 -0700 Message-Id: <199409302340.QAA23259@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: vsta@amri.cisco.com Subject: Shared libraries, operational! Date: Fri, 30 Sep 1994 16:40:05 -0700 From: Andrew Valencia Hello all, This is being typed from a VSTa telnet window. All commands and non-boot servers are running from a single shared library, including networking and the ethernet driver. The vsta/bin and vsta/boot directories are much smaller. A surprising side effect is that booting up goes quicker--I guess having to demand-page in ~30K of C library for each server and command adds up to a non-trivial bit of time! I will recompile all the "auxilary" source over the weekend, then bundle this up and make it available as 1.3.2 early next week. Andy From vandys@glare.cisco.com Sat Oct 1 15:50:44 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id PAA03935; Sat, 1 Oct 1994 15:50:42 -0700 Received: from eskinews.eskimo.com (uucp@ESKINEWS.ESKIMO.COM [204.122.16.44]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id AAA07914 for ; Sat, 1 Oct 1994 00:51:11 -0700 From: songdog!roman@eskinews.eskimo.com Received: from songdog.UUCP by eskinews.eskimo.com (5.65c/1.35) id AA25855; Sat, 1 Oct 1994 00:49:27 -0700 Received: by songdog.uucp (Smail3.1.28.1 #6) id m0qqwPB-0003C7C; Fri, 30 Sep 94 21:50 PDT Message-Id: Subject: alternative to Unicode To: vsta@cisco.com Date: Fri, 30 Sep 1994 21:50:40 -0800 (PDT) X-Mailer: ELM [version 2.4 PL21] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 2287 I noticed the mention of Unicode and wide character support here a while back, and I thought people might be interested in an alternative I came across recently. P.J. Plauger writes a column for "Embedded Systems Programming." In the May 1994 issue he discusses an encoding for multibyte characters which has been developed at AT&T Bell Labs. It is called FSS-UTF; he says the "FSS" part stands for "file-system safe" and the "UTF" might mean something like "universal transfer format." The "file-system safe" part has to do with the fact that certain characters are treated specially when they appear in pathnames; conventional multi-byte encodings could, for example, produce a '/' or a '\0' as one byte of the two-byte representation of a Kanji character, with undesireable consequences. FSS-UTF gets around this by representing the 7-bit ASCII characters as themselves, and never using these values as part of a multi-byte sequence. All bytes of a multi-byte sequence have their high bit set. Here is the full description of the encoding quoted from the article: A byte with a leading 0 bit (hex 00 through hex 7F) stands for a wide character with the same value, in the range 00-7F. A byte with a leading 10 (80-BF) is never the first byte of a multibyte sequence. Such an NFB contributes its six low-order bits to a wide character, as specified below. A byte with a leading 110 (C0-DF) is the first of a 2-byte sequence. It contributes its five low-order bits to a wide character in the range 40-7FF, and a following NFB contributes the low-order six bits. A byte with a leading 1110 (E0-EF) is the first of a 3-byte sequence. It contributes its four low-order bits to a wide character in the range 1000-FFFF, and the two following NFBs contribute the low-order 12 bits, in descending order of significance. Plauger goes on to describe obvious extensions which can encode the ISO 10646 31-bit (!) character set within six bytes. He also points out FSS-UTF could be used, for example, to write integer values into a byte stream in a system-independent fashion, while also providing some amount of compression for smaller values. -- Bill Roman (songdog!roman@eskimo.com / roman@songdog.uucp) running linux From vandys@glare.cisco.com Sat Oct 1 18:54:49 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id SAA03948; Sat, 1 Oct 1994 18:54:47 -0700 Received: from ebt.com (spock.ebt.com [192.111.115.3]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id UAA03141 for ; Sat, 1 Oct 1994 20:07:16 -0700 Received: from ebt-inc.ebt.com by ebt.com (4.1/SMI-4.1) id AA18587; Sat, 1 Oct 94 23:08:24 EDT Received: by ebt-inc.ebt.com (5.0/SMI-SVR4) id AA26483; Sat, 1 Oct 1994 23:08:26 +0500 Date: Sat, 1 Oct 1994 23:08:26 +0500 From: gtn@ebt.com (Gavin Nicol) Message-Id: <9410020308.AA26483@ebt-inc.ebt.com> To: songdog!roman@eskinews.eskimo.com Cc: vsta@cisco.com In-Reply-To: (songdog!roman@eskinews.eskimo.com) Subject: Re: alternative to Unicode Content-Length: 357 Actually, FSS-UTF is an external representation for the 16 bit Unicode characters that is safe in a 7 bit environment. I am very well aware of this, and plan to make a library for VSTa which will give us much the same "universal locale" as AIX 4.1. It's not *that* easy though because a lot of changes needs to be made in many places (regex, printing etc.) From vandys@glare.cisco.com Mon Oct 3 06:27:36 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id GAA04263; Mon, 3 Oct 1994 06:27:35 -0700 Received: from wotan.compaq.com (wotan.compaq.com [131.168.249.254]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id HAA21819 for ; Mon, 3 Oct 1994 07:40:16 -0700 Received: from twisto.eng.hou.compaq.com by wotan.compaq.com with smtp (Smail3.1.28.1 #12) id m0qroUL-000vJaC; Mon, 3 Oct 94 09:35 CDT Received: from gocart.eng.hou.compaq.com by twisto.eng.hou.compaq.com with smtp (Smail3.1.28.1 #10) id m0qroRR-000uGsC; Mon, 3 Oct 94 09:32 CDT Received: by gocart.eng.hou.compaq.com (Smail3.1.26.7/COMPAQ-HESIOD) id m0qroRR-00036pC; Mon, 3 Oct 94 09:32 CDT Message-Id: From: edm@gocart.eng.hou.compaq.com (Edward Mccreary) Subject: accessing physical memory To: vsta@cisco.com Date: Mon, 3 Oct 94 9:32:37 CDT X-Mailer: ELM [version 2.4dev PL11] Ok, this is probably a *really* stupid mistake.... I'm trying to do some stuff in graphics mode but can't seem to get access to the frame buffer. What I'm doing is... ulong addr; unsigned char *frame; /* * code to get io perms and set video mode, this works ok * also sets banks to first page of video memory */ addr = 0xA0000; frame = mmap((void *)addr, 32*1024L, PROT_READ | PROT_WRITE, MAP_PHYS, 0, 0L); mmap is returning a non-NULL pointer but when I do this... /* fill first line of display to white */ int i; unsigned char *ptr = frame; i = 1024; while(i--) *ptr++ = 0xFF; nothing happens. Any ideas? I get the feeling it's something obvious I'm just not getting. eddie -- Eddie McCreary edm@twisto.compaq.com Graphics Development "Do or do not, there is no 'try'." In the event of my capture, Compaq will disavow any and all knowledge of my operations. Of course I don't speak for them. From vandys@glare.cisco.com Mon Oct 3 07:42:13 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id HAA04284; Mon, 3 Oct 1994 07:42:11 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id IAA02362; Mon, 3 Oct 1994 08:54:44 -0700 Message-Id: <199410031554.IAA02362@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: Dave Hudson Cc: vsta@amri.cisco.com Subject: Re: Generating diffs to 1.3.2 In-Reply-To: Your message of "Sun, 02 Oct 1994 21:34:29 BST." Date: Mon, 03 Oct 1994 08:54:44 -0700 From: Andrew Valencia [Dave Hudson writes:] >I've finally had some time to trawl through all of the code and some of the >patches that haven't made it (BTW I hadn't realised you were a ham :-)). >One thing that I noticed was that apart from libc things there are quite a >few server's with patches outstanding and some of the core utils. As I'm >going to be repatching (I really need the shared libs and syslog code) 'til >early in the morning do you want me to resend patch files against the 1.3.2 >sources? :-> got my general license way back in junior high. My wife's a ham, too (though that isn't what got us together)! Actually, it would be best if you prioritized and only sent back in the patches which you think REALLY should be present for the next release. Then, between the core team, I'd dearly like to get /tcp, RCS, and a remote filesystem working. Oh, also probably touch up vstafs. Then I want to put all the source in a VSTa filesystem with protection set appropriately. So /vsta/libc would be owner vsta.libc, vsta/os would be vsta.os, vsta/srv/mach/scsi might be vsta.srv.scsi. Then you can hand out ID's for people based on what they work on; I get "vsta", and thus inherit access to the tree. A server guru would get vsta.srv, and would get access to all the servers (including vsta.srv.scsi, since he dominates that label and thus inherits its full access). Default access to the tree is read, so anybody at all can mount it and read the current state. >One other thing - does mkshlib and mkshlib.c really want to live in libc >rather than bin? Yes; they're tools only used to build the libraries. >BTW I finally figured out a way to port BSD code with the nasty file >operations - it means writing a script to translate the file ops to new >names say bsd_* and then writing bsd_* functions that store extra >side-information but it should work OK and won't impose too much extra code >size overhead now that we have shared libs. I'll see if I can get some time >to write this before the end of the year :-) Yes, that would work. I don't know if you want to support this subrosa information across exec(), though. That would be harder, although the mechanism currently used should be generalized anyway. Andy From vandys@glare.cisco.com Tue Oct 4 11:31:03 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id LAA04524; Tue, 4 Oct 1994 11:31:02 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id MAA13420; Tue, 4 Oct 1994 12:43:18 -0700 Message-Id: <199410041943.MAA13420@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: David.Eagle@ragbbs.ima.infomail.com (David Eagle) Cc: vsta@amri.cisco.com Subject: Re: Internet Mail VIA RagBBS In-Reply-To: Your message of "04 Oct 1994 11:30:00 MDT." <51a_9410041335@ima.infomail.com> Date: Tue, 04 Oct 1994 12:43:17 -0700 From: Andrew Valencia [David.Eagle@ragbbs.ima.infomail.com (David Eagle) writes:] >I hope this gets to you. I used to subscribe to the VSTa mailing list >while I was on contract at DEC. I am in between contracts right now and >don't really have a permanent email address but for the next few weeks >I can be reached at David.Eagle@ragbbs.ima.infomail.com. While at DEC >I ftp'd VSTa, and now have just gotten settled in after a little vacation >and tried to decompress it and try it out. Unfortunately, the boot program >tells me I should be in V86 mode or something. I assume this has something >to do with real vs. virtual 8086 mode for the CPU, but I don't see how >anything I can do would affect this. Can someone tell me how to boot VSTa? Most often, it's an EMS emulator which causes this problem. It runs the x86 in V86 mode, which prevents the loader from going to 32-bit protected mode. If you're running DOS 6, you can hold the shift key down during DOS boot, and it'll inhibit loading of the config.sys. Otherwise you have to fiddle your config.sys/autoexec.bat files. >Also, is VSTa ready to distribute as share/freeware yet? If so, I would be >willing to spend a few bucks to get floppies from you containing the most >recent stuff (and hopefully some recent excerpts from the mailing list). Well, it's copyleft'ed. If you send me floppies with room for ~8 meg, I can dump the whole distribution. This includes source to all commands ported, networking, core distribution, and mailing list archives. Please include a SASE mailer for me to send back the floppies in. I can do 5.25" and 3.5", the latter preferred because of their greater durability. If you just want the core distribution+mail list archives, that comes to ~3 megs instead. In either case, please be aware that if you don't provide me with everything I need to do the job (i.e., send $$$'s instead of floppies, or send $$$'s which you figure can cover me going out and buying a return mailer) then I'll put it on my queue, but it will take a LONG time. I wrote an RS-232 driver for Minix 1.1, and the floppies are *still* waiting to be sent back, if you get my drift.... We're on the brink of 1.3.2, with 1.4 following (best guess) in a couple months. Be sure to specify what you'd prefer. Andy From vandys@glare.cisco.com Wed Oct 5 19:34:27 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id TAA04675; Wed, 5 Oct 1994 19:34:26 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id UAA06380 for ; Wed, 5 Oct 1994 20:47:22 -0700 Message-Id: <199410060347.UAA06380@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: vsta@amri.cisco.com Subject: v1.3.2, now available Date: Wed, 05 Oct 1994 20:47:22 -0700 From: Andrew Valencia Version 1.3.2 of VSTa is now available at the usual place. See the file doc/features.txt for a list of what's been added. Enjoy! Andy From vandys@glare.cisco.com Thu Oct 6 03:27:58 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id DAA04817; Thu, 6 Oct 1994 03:27:57 -0700 Received: from post.demon.co.uk (post.demon.co.uk [158.152.1.72]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id EAA17174 for ; Thu, 6 Oct 1994 04:40:55 -0700 Received: from humbug.demon.co.uk by post.demon.co.uk id aa28545; 6 Oct 94 11:57 GMT-60:00 Received: by humbug.demon.co.uk (Linux Smail3.1.28.1 #1) id m0qsqUk-00037ZC; Thu, 6 Oct 94 11:56 BST Message-Id: From: Dave Hudson Subject: GNU tool ports and release 1.3.2 To: VSTa mailing list Date: Thu, 6 Oct 1994 11:56:17 +0100 (BST) X-Mailer: ELM [version 2.4 PL23] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 765 Hi All, Just in case anyone's wondered what's happened to the tool ports I did a couple of months back (apart from sed and gzip which are in 1.3.2) - they should be available in release 1.4. There's a lot of library patches that need to be re-integrated into the 1.3.2 source tree and I concentrated on getting server enhancements and core tool bug fixes done. I'll see if I can get ports out against the standard 1.3.2 library (some of the libc patches are there), but some will definitely need new functions and/or enhancements to existing functions to get them running. Until 1.4's released though I have *statically* linked binaries for the following: tar, make, fileutils, find, flex, bison and yacc. If you need them, drop me a line. Regards, Dave From vandys@glare.cisco.com Thu Oct 6 06:57:15 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id GAA04836; Thu, 6 Oct 1994 06:57:14 -0700 Received: from cygnus.com (cygnus.com [140.174.1.1]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with ESMTP id IAA24254 for ; Thu, 6 Oct 1994 08:10:11 -0700 Received: (from rob@localhost) by cygnus.com (8.6.9/8.6.9) id IAA03814; Thu, 6 Oct 1994 08:09:31 -0700 Message-Id: <199410061509.IAA03814@cygnus.com> From: rob@cygnus.com (Rob Savoye) Date: Thu, 6 Oct 1994 08:09:30 PDT In-Reply-To: Dave Hudson ' GNU tool ports and release 1.3.2 Reply-To: rob@cygnus.com Phone-Number: (303) 258-0506 MST X-Mailer: Mail User's Shell (7.1.1 5/02/90) To: Dave Hudson , VSTa mailing list Subject: Re: GNU tool ports and release 1.3.2 From: Dave Hudson Subject: GNU tool ports and release 1.3.2 > Just in case anyone's wondered what's happened to the tool ports I did a > couple of months back (apart from sed and gzip which are in 1.3.2) - they > should be available in release 1.4. There's a lot of library patches that Also there is a patch out now for GCC 2.6.0 that adds support for VSTa as a cross target. (or native) With this patch, you should be able to configure using "--target i386-vsta" and get a working tool chain. The file is on ftp.cygnus.com:pub/embedded/crossgcc-2.6.0.patch.gz. I'll admit it hasn't been tested recently much, but I'm finally getting the VSTa support into the FSF GCC sources. There is already VSTa support in gas, ld, and the binutils current net releases. - rob - From vandys@glare.cisco.com Sun Oct 9 19:16:58 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id TAA05286; Sun, 9 Oct 1994 19:16:57 -0700 Received: from Trantor.DSO.gov.SG (trantor.dso.gov.sg [192.190.204.1]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id UAA27521 for ; Sun, 9 Oct 1994 20:30:13 -0700 Received: by Trantor.DSO.gov.SG (4.1/25-TRANTOR-eef) id AA05252; Mon, 10 Oct 94 11:31:23 SST From: Tan Pong Heng Message-Id: <9410100331.AA05252@Trantor.DSO.gov.SG> Subject: FPU Management for VSTa To: vsta@cisco.com Date: Mon, 10 Oct 94 11:31:21 WST X-Mailer: ELM [version 2.3 PL11] Hi, I just downloaded the 1.3.2 release of VSTa and was reading through the kernel codes. I was actually looking for the support of FPU in the codes as it was not supported (was it?) in 1.3.1. I realized that the kernel saved the state of the FPU everytime it does a task switch! This is definitely not necessary as most of the task would not be using the FPU and so it is highly likely that a task using the FPU would not need to save it's state at all. The Linux used a smarter scheme that it would save the old state when a new task starts to use the FPU and caused a FPU exception due to the TS bit. At this point it would save the state to the last task used the FPU and restore the FPU state for the current task (just like VSTa) and record that the current task is the last task using the FPU so that the next exception can save the state properly. The scheme is simple and it would save a lot of unnecessary saving of FPU states which is not cheap (around 96 words?). I hope this scheme would be implemented in the next release of VSTa. -- Tan Pong Heng Defence Science Organization Project Leader 20 Science Park Drive, S(0511) From vandys@glare.cisco.com Sun Oct 9 20:00:59 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id UAA05291; Sun, 9 Oct 1994 20:00:58 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id VAA06216; Sun, 9 Oct 1994 21:14:17 -0700 Message-Id: <199410100414.VAA06216@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: Tan Pong Heng Cc: vsta@cisco.com Subject: Re: FPU Management for VSTa In-Reply-To: Your message of "Mon, 10 Oct 1994 11:31:21 +0700." <9410100331.AA05252@Trantor.DSO.gov.SG> Date: Sun, 09 Oct 1994 21:14:17 -0700 From: Andrew Valencia [Tan Pong Heng writes:] >... I realized that the >kernel saved the state of the FPU everytime it does a task switch! Nope, shouldn't. Are you sure you read it right? You need to take a look at when T_FPU is set in the thread flags. It's set in trap.c when an access is attempted to the FPU. A non-FPU process will never execute an FPU instruction, will never get the bit turned on, and will run at full speed. In sched.c all the code to do an FPU save is conditional on T_FPU. A process which accesses a couple FPU instructions, switches away, and doesn't execute any more will also not have T_FPU set the 2nd and successive times it runs. > The Linux used a smarter scheme that it would >save the old state when a new task starts to use the FPU and caused a >FPU exception due to the TS bit. Except this has serious problems in a multiprocessor environment. It's really, really hard to grab CPU state out from under another CPU without having a performance hit at best, and nasy bugs indeed at worst. Since FPU isn't famous for being all that fast, I wanted to move what "hit" there was to the FPU-using process. >... I hope this scheme >would be implemented in the next release of VSTa. Nope, I'd rather leave things compatible with multiprocessor hardware. There are a couple groups who are looking at such ports, so it'll be more than academic with a little luck. Regards, Andy From vandys@glare.cisco.com Mon Oct 10 06:43:33 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id GAA05435; Mon, 10 Oct 1994 06:43:31 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id HAA23870 for ; Mon, 10 Oct 1994 07:56:54 -0700 Message-Id: <199410101456.HAA23870@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: vsta@amri.cisco.com Subject: Status Date: Mon, 10 Oct 1994 07:56:54 -0700 From: Andrew Valencia Well, at least *some* folks are running v1.3.2 out there. We've found one serious problem which bites executables which have a size N such that ((n % PAGESIZE) < 32). If you rebuild the RS232 server you'll see it--that's where I found it. The fix is in ld; looks like a v1.3.3 will be out soon. As a part of chasing that I discovered that each executable has an extra four bytes tacked on the end. These turn out to be the (empty) string table; the effect is that all stripped executables burn an extra disk block for no reason! ld, strip, and nm will be fixed to avoid this. I've added a sample autoexec.net to /vsta/etc; the KA9Q networking needs such a file to set up IP address, default routes, and so forth. If I hae a spare minute I'll tweak the source to use /vsta/etc/autoexec.net by default. Regards, Andy From vandys@glare.cisco.com Mon Oct 10 08:20:19 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id IAA05475; Mon, 10 Oct 1994 08:20:18 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id JAA28683; Mon, 10 Oct 1994 09:33:40 -0700 Message-Id: <199410101633.JAA28683@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: Chris Patti Cc: vsta@cisco.com Subject: Re: 1.3.2 is *very* impressive In-Reply-To: Your message of "Mon, 10 Oct 1994 00:22:32 GMT." <199410100022.AAA00620@wyn.fandom-house.org> Date: Mon, 10 Oct 1994 09:33:39 -0700 From: Andrew Valencia [Chris Patti writes:] >This could be totally unrelated, but after booting VSTa and telling it about >my NE2000 clone (CHEAP NE2000 clone, $50 :) my ethercard started blinking its >red light regularly, as if there was an error condition. Of course the grungy >vendor supplied software didn't help, but a power down , 2 min. wait, and >restart fixed the problem. The ne2000 driver no doubt could use some work. It was handed to me in somewhat broken form, and I chased it just enough to get it talking to my card and doing packets. >Glad to hear you intend to include a sample KA9Q startup file in the next >release. I was just pulling down the Ka9Q dist to try to divine what to do. You can enter the command by hand as well. I use: /vsta/boot/ne 0x300,5 & # Start the driver mount net/ne /dev/eth # Add it to my namespace net # Start ka9q attach eth eth0 1500 # Start the networking interface ip addr 198.92.29.53 # Set my node's IP address route add default eth0 # Default route via ether host baruk # My hostname domain suffix cisco.com # Default suffix domain addserver 131.108.1.111 # My DNS host of choice After this you can telnet, FTP, and so forth. To save typing, put these all in a file, and start the "net" command with the file as an argument. Regards, Andy From vandys@glare.cisco.com Mon Oct 10 08:11:08 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id IAA05472; Mon, 10 Oct 1994 08:11:07 -0700 Received: from wyn.fandom-house.org (root@wyn.fandom-house.org [128.127.21.25]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with ESMTP id JAA25001 for ; Mon, 10 Oct 1994 09:24:26 -0700 Received: (from feoh@localhost) by wyn.fandom-house.org (8.6.9/8.6.9) id AAA00620 for vsta@amri.cisco.com; Mon, 10 Oct 1994 00:22:32 GMT Date: Mon, 10 Oct 1994 00:22:32 GMT From: Chris Patti Message-Id: <199410100022.AAA00620@wyn.fandom-house.org> To: vsta@cisco.com Subject: 1.3.2 is *very* impressive Nice job folks. VSTa is now a truly useful environment in and of itself. the native gcc works great, networking, lots of rough edges polished smooth. This could be totally unrelated, but after booting VSTa and telling it about my NE2000 clone (CHEAP NE2000 clone, $50 :) my ethercard started blinking its red light regularly, as if there was an error condition. Of course the grungy vnedor supplied software didn't help, but a power down , 2 min. wait, and restart fixed the problem. Glad to hear you intend to include a sample KA9Q startup file in the next release. I was just pulling down the Ka9Q dist to try to divine what to do. From vandys@glare.cisco.com Mon Oct 10 20:23:40 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id UAA05569; Mon, 10 Oct 1994 20:23:38 -0700 Received: from post.demon.co.uk (post.demon.co.uk [158.152.1.72]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id VAA29184 for ; Mon, 10 Oct 1994 21:37:03 -0700 Received: from humbug.demon.co.uk by post.demon.co.uk id aa22272; 11 Oct 94 0:16 GMT-60:00 Received: by humbug.demon.co.uk (Linux Smail3.1.28.1 #1) id m0quSyP-0003RmC; Mon, 10 Oct 94 23:13 BST Message-Id: From: Dave Hudson Subject: Running ka9q against Linux To: VSTa mailing list Date: Mon, 10 Oct 1994 23:13:37 +0100 (BST) X-Mailer: ELM [version 2.4 PL23] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 573 Hi All, If anyone's managed to get Linux to handle telnet sessions from VSTa properly can you let me know what I need to do. I have a strong suspicion that the Linux IP stack's a little broken at the moment (based on a note from the Linux NET mailing list). FWIW I've tried Linux kernels 1.1.11, 1.1.41 and 1.1.52 (the latter two with or without the PC/TCP config). I've also tried two different types of PC and two types of NE2000 card (one of which is a genuine Novell/Eagle NE2000), but I still find that the telnet connection absolutely crawls. Regards, Dave From vandys@glare.cisco.com Tue Oct 11 00:41:12 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id AAA05585; Tue, 11 Oct 1994 00:41:11 -0700 Received: from ebt.com (spock.ebt.com [192.111.115.3]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id BAA07107 for ; Tue, 11 Oct 1994 01:54:37 -0700 Received: from ebt-inc.ebt.com by ebt.com (4.1/SMI-4.1) id AA06252; Tue, 11 Oct 94 04:55:47 EDT Received: by ebt-inc.ebt.com (5.0/SMI-SVR4) id AA21072; Tue, 11 Oct 1994 04:55:47 +0500 Date: Tue, 11 Oct 1994 04:55:47 +0500 From: gtn@ebt.com (Gavin Nicol) Message-Id: <9410110855.AA21072@ebt-inc.ebt.com> To: vsta@cisco.com Subject: wide strings Content-Length: 328 If someone has the ANSI spec handy, could you please post a list of all string/wide string and isxxxx tests defined in it. Also, I don't think POSIX specifies anything special, but if so, also post such data. Now that we have shared libraries, I'll finish off the unicode based wide string package I was working on last year. From vandys@glare.cisco.com Tue Oct 11 01:21:22 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id BAA05589; Tue, 11 Oct 1994 01:21:20 -0700 Received: from zephyr.cs.vu.nl (mmdf@[130.37.16.3]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id CAA08024 for ; Tue, 11 Oct 1994 02:34:43 -0700 Received: from galjas.cs.vu.nl by zephyr.cs.vu.nl id aa18060; 11 Oct 94 10:34 MET Date: Tue, 11 Oct 94 10:34:30 MET From: "R.Oussoren" To: vsta@cisco.com Subject: Installing v1.3.2 Message-Id: <9410111034.aa02017@galjas.cs.vu.nl> I have a little problem when installing VSTa. When I login the session no longer responds (ie. I type a valid name and a valid password and then nothing happens). The system doesn't hang, as I can switch to another virtual terminal and the login for that terminal works fine. In the kernel debugger, the process that was the login process no longer has a name. When a change the boot.lst file to start 'testsh' instead of 'init' I can run programs (I haven't done anything useful but 'run sh' and then using 'ps' works fine). How can I fix this?? Ronald. From vandys@glare.cisco.com Tue Oct 11 08:01:32 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id IAA05667; Tue, 11 Oct 1994 08:01:31 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id JAA01730; Tue, 11 Oct 1994 09:14:58 -0700 Message-Id: <199410111614.JAA01730@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: "R.Oussoren" Cc: vsta@cisco.com Subject: Re: Installing v1.3.2 In-Reply-To: Your message of "Tue, 11 Oct 1994 10:34:30 +0700." <9410111034.aa02017@galjas.cs.vu.nl> Date: Tue, 11 Oct 1994 09:14:58 -0700 From: Andrew Valencia ["R.Oussoren" writes:] >I have a little problem when installing VSTa. When I login the session >no longer responds (ie. I type a valid name and a valid password and then >nothing happens). The system doesn't hang, as I can switch to another >virtual terminal and the login for that terminal works fine. In the >kernel debugger, the process that was the login process no longer has a name. It appears that shared libraries don't work unless you're in group 0 (and thus get sys.sys). Make sure your account has group 0 and I bet things will start working. We'll get this fixed for 1.3.3, whose release appears imminent. Sorry! Andy From vandys@glare.cisco.com Tue Oct 11 10:43:59 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id KAA05691; Tue, 11 Oct 1994 10:43:58 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id LAA13053 for ; Tue, 11 Oct 1994 11:57:27 -0700 Message-Id: <199410111857.LAA13053@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: vsta@amri.cisco.com Subject: VSTa FTP mirror at ftp.cygnus.com updated Date: Tue, 11 Oct 1994 11:57:27 -0700 From: Andrew Valencia The v1.3.2 code has been successfully mirrored to ftp.cygnus.com. You can find it in pub/embedded/vsta. Andy From vandys@glare.cisco.com Tue Oct 11 13:12:28 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id NAA05709; Tue, 11 Oct 1994 13:12:27 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id OAA22019 for ; Tue, 11 Oct 1994 14:25:56 -0700 Message-Id: <199410112125.OAA22019@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: vsta@amri.cisco.com Subject: Line editing Date: Tue, 11 Oct 1994 14:25:56 -0700 From: Andrew Valencia As an amusing test of shared libraries, I took the "getline" library off the net and ported it into the VSTa C library in place of the current canonical input processing. My old Bourne shell binary now does line editing, as does anything else which uses canonical (i.e., line input) TTY mode! There were a couple interesting issues in making this work. First, getline() wanted (and needed) the "prompt". Of course, any existing binary write()'s the prompt and then read()'s a line. I wanted existing binaries to work, so changing them to call getline() was out. I did this by recording the canonical-mode write()'s data, and then riffling through it when a read() happened to see what was left on the last line. Gross, yes. It seems to work. Second, I had problems in that per-file-descriptor state wasn't going to help me--stdin and stdout are distinct file descriptors, each with its own state. I tweaked up vsta/libc/tty.c so the prompt state can be shared between the two. As another nice side effect of all this, ^D now does the right thing as EOF. Andy From vandys@glare.cisco.com Wed Oct 12 21:52:11 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id VAA05920; Wed, 12 Oct 1994 21:52:09 -0700 Received: from staff.cs.su.OZ.AU (staff.cs.su.OZ.AU [129.78.8.1]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id XAA06934 for ; Wed, 12 Oct 1994 23:05:44 -0700 Received: from sour.sw.oz.au by staff.cs.su.OZ.AU (mail from jeremy for vsta@cisco.com) with MHSnet (insertion MHSnet site: swallow.sw.oz.au); Thu, 13 Oct 1994 16:05:41 +1000 Received: from sour.sw.oz.au by swallow.sw.oz.au with SMTP id AA25638; Thu, 13 Oct 94 16:21:07 EST (4.1/Unixware) (from jeremy@sour.sw.oz.au for vsta@cisco.com) Received: by sour.sw.oz.au id AA23019; Thu, 13 Oct 1994 16:08:01 +1000 (5.65c/1.34) (from jeremy@sour.sw.oz.au for vsta@cisco.com) From: jeremy@sour.sw.oz.au (Jeremy Fitzhardinge) Message-Id: <199410130608.AA23019@sour.sw.oz.au> Subject: Problem getting VSTa 1.3.2 up To: vsta@cisco.com (VSTa list) Date: Thu, 13 Oct 1994 16:08:01 +1000 (EST) Organization: Softway Pty Ltd X-Face: '6U=%Tv\k1l-:?\$C[D@G 7(vl~w8&y}!f\bh#wL#n,TGKh>T.c7eT5-y)Hl'i;A1z$9?*lD.k}yqshddFb l[EC}c=;uc%x'}uh3E91p&oE pid 14 status 65536 user 0 system 0 Pid 14 was the testsh created by the last run. If I do this repeatedly, eventually I type it in the original testsh, which then makes the system drop into the kernel debugger. As I said, I know nothing about VSTa, so I feel stumped. Is this related to the problem of having to be in sys.sys to use shared libs? If so, how do I add users to sys.sys (I don't really understand the passwd, group and ids files)? Thanks, J From vandys@glare.cisco.com Wed Oct 12 22:02:28 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id WAA05924; Wed, 12 Oct 1994 22:02:27 -0700 Received: from staff.cs.su.OZ.AU (staff.cs.su.OZ.AU [129.78.8.1]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id XAA07334 for ; Wed, 12 Oct 1994 23:16:03 -0700 Received: from sour.sw.oz.au by staff.cs.su.OZ.AU (mail from jeremy for vsta@cisco.com) with MHSnet (insertion MHSnet site: swallow.sw.oz.au); Thu, 13 Oct 1994 16:16:02 +1000 Received: from sour.sw.oz.au by swallow.sw.oz.au with SMTP id AA25813; Thu, 13 Oct 94 16:31:15 EST (4.1/Unixware) (from jeremy@sour.sw.oz.au for vsta@cisco.com) Received: by sour.sw.oz.au id AA23152; Thu, 13 Oct 1994 16:18:08 +1000 (5.65c/1.34) (from jeremy@sour.sw.oz.au for vsta@cisco.com) From: jeremy@sour.sw.oz.au (Jeremy Fitzhardinge) Message-Id: <199410130618.AA23152@sour.sw.oz.au> Subject: Slightly embarrased... To: vsta@cisco.com (VSTa list) Date: Thu, 13 Oct 1994 16:18:07 +1000 (EST) Organization: Softway Pty Ltd X-Face: '6U=%Tv\k1l-:?\$C[D@G 7(vl~w8&y}!f\bh#wL#n,TGKh>T.c7eT5-y)Hl'i;A1z$9?*lD.k}yqshddFb l[EC}c=;uc%x'}uh3E91p&oE; Thu, 13 Oct 1994 02:47:05 -0700 Received: from studs.sci.kun.nl by wn1.sci.kun.nl via studs.sci.kun.nl [131.174.124.5] with ESMTP id KAA27083 (8.6.9/2.4) for ; Thu, 13 Oct 1994 10:47:03 +0100 From: Marc Papen Received: by studs.sci.kun.nl id KAA26605 (8.6.9/2.1 for VSTa@cisco.com); Thu, 13 Oct 1994 10:47:02 +0100 Message-Id: <199410130947.KAA26605@studs.sci.kun.nl> Subject: New Project and a few Questions To: VSTa@cisco.com (vsta) Date: Thu, 13 Oct 1994 10:47:02 +0100 (MET) Reply-To: marcp@sci.kun.nl X-Mailer: ELM [version 2.4 PL23] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 917 Hi After playing around with VSTa, I decided to add something myself. I'm in the process of writing a server for the Mitsumi CD-ROM drives. (My first thought was to write a tape driver, but i'm not this keen on doing DMA-stuff so i'll try something easier -- I hope at least) If somebody is already working on such a server please mail me! I have a 1 week holiday next week so i might get something working soon. Now my question: I can't get longer compiles to work with the new version(1.32). I have canged the the group to 0 (sys.sys) on my account but this only helped a little bit. Is there a short solution to this problem, as i can't work on the server very well if the system locks up every few compiles? Now my thanks to all who have brought VSTa to the current state. Thanks Marc PS: I will mail again if i have a working version of the mitsumi server to test. --- Marc Papen email: marcp@sci.kun.nl From vandys@glare.cisco.com Thu Oct 13 02:10:18 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id CAA06029; Thu, 13 Oct 1994 02:10:17 -0700 Received: from wn1.sci.kun.nl (wn1.sci.kun.nl [131.174.8.1]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with ESMTP id DAA20630 for ; Thu, 13 Oct 1994 03:23:54 -0700 Received: from studs.sci.kun.nl by wn1.sci.kun.nl via studs.sci.kun.nl [131.174.124.5] with ESMTP id LAA28410 (8.6.9/2.4) for ; Thu, 13 Oct 1994 11:23:46 +0100 From: Marc Papen Received: by studs.sci.kun.nl id LAA29556 (8.6.9/2.1 for VSTa@cisco.com); Thu, 13 Oct 1994 11:23:45 +0100 Message-Id: <199410131023.LAA29556@studs.sci.kun.nl> Subject: New Project and question To: VSTa@cisco.com Date: Thu, 13 Oct 1994 11:23:44 +0100 (MET) Reply-To: marcp@sci.kun.nl X-Mailer: ELM [version 2.4 PL23] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 918 Hi After playing around with VSTa, I decided to add something myself. I'm in the process of writing a server for the Mitsumi CD-ROM drives. (My first thought was to write a tape driver, but i'm not this keen on doing DMA-stuff so i'll try something easier -- I hope at least) If somebody is already working on such a server please mail me! I have a 1 week holiday next week so i might get something working soon. Now my question: I can't get longer compiles to work with the new version(1.32). I have canged the the group to 0 (sys.sys) on my account but this only helped a little bit. Is there a short solution to this problem, as i can't work on the server very well if the system locks up every few compiles? Now my thanks to all who have brought VSTa to the current state. Thanks Marc PS: I will mail again if i have a working version of the mitsumi server to test. --- Marc Papen email: marcp@sci.kun.nl From vandys@glare.cisco.com Thu Oct 13 05:32:32 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id FAA06075; Thu, 13 Oct 1994 05:32:30 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id GAA02754; Thu, 13 Oct 1994 06:46:09 -0700 Message-Id: <199410131346.GAA02754@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: marcp@sci.kun.nl Cc: VSTa@cisco.com (vsta) Subject: Re: New Project and a few Questions In-Reply-To: Your message of "Thu, 13 Oct 1994 10:47:02 BST." <199410130947.KAA26605@studs.sci.kun.nl> Date: Thu, 13 Oct 1994 06:46:08 -0700 From: Andrew Valencia [Marc Papen writes:] >After playing around with VSTa, I decided to add something myself. >I'm in the process of writing a server for the Mitsumi CD-ROM drives. Is this CDROM a non-SCSI drive? I assume you know about Mike Larson's work. >I can't get longer compiles to work with the new version(1.32). I have canged >the the group >to 0 (sys.sys) on my account but this only helped a little bit. When it dies, do a: ^Z (you'll get the kernel debugger prompt) dv freemem If this value is 0, then you're running out of memory. >... as i can't work on the server very well if the system locks u >p every few compiles? How much are you compiling at a shot, which directory, and how much RAM does your system have? The kernel debugger also has the "memleak" command, which reports how much of what sorts of memory have been used since the last time it ran. You might use this to see if you can characterize where the memory's going. BTW, we do know of one voracious RAM eater--building the shared version of libc. A file is created in /tmp for each entry in the library, but /tmp uses 8K blocks for its RAM-based filesystem. The result is that it uses an amazing amount of RAM. You can fix this by changing the blocksize of tmpfs to something more modest--1K, for instance. Or, avoid rebuilding the entire libc when possible. Or, umount /tmp, make a /tmp in your DOS filesystem, and use a DOS directory for temp storage. This slows compiles noticably, in my experience. Andy From vandys@glare.cisco.com Thu Oct 13 15:36:26 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id PAA06188; Thu, 13 Oct 1994 15:36:25 -0700 Received: from post.demon.co.uk (post.demon.co.uk [158.152.1.72]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id QAA27455 for ; Thu, 13 Oct 1994 16:50:07 -0700 Received: from humbug.demon.co.uk by post.demon.co.uk id aa10756; 14 Oct 94 0:23 GMT-60:00 Received: by humbug.demon.co.uk (Linux Smail3.1.28.1 #1) id m0qvZ4o-0003JHC; Thu, 13 Oct 94 23:56 BST Message-Id: From: Dave Hudson Subject: Bug in the swapper code To: VSTa mailing list Date: Thu, 13 Oct 1994 23:56:46 +0100 (BST) X-Mailer: ELM [version 2.4 PL23] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 306 Hi All, After the number of problems people with 4 MByte systems seem to have been reporting I've been looking at the paging code. There are a few problems with it in the 1.3.2 release. I've got fixes for some of these already so hopefully the 4 meg systems will work OK with 1.3.3. Regards, Dave From vandys@glare.cisco.com Mon Oct 17 18:47:20 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id SAA06784; Mon, 17 Oct 1994 18:47:18 -0700 Received: from staff.cs.su.OZ.AU (staff.cs.su.OZ.AU [129.78.8.1]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id UAA14517 for ; Mon, 17 Oct 1994 20:01:21 -0700 Received: from sour.sw.oz.au by staff.cs.su.OZ.AU (mail from jeremy for vsta@cisco.com) with MHSnet (insertion MHSnet site: swallow.sw.oz.au); Tue, 18 Oct 1994 13:01:15 +1000 Received: from sour.sw.oz.au by swallow.sw.oz.au with SMTP id AA25566; Tue, 18 Oct 94 13:19:02 EST (4.1/Unixware) (from jeremy@sour.sw.oz.au for vsta@cisco.com) Received: by sour.sw.oz.au id AA28365; Tue, 18 Oct 1994 13:03:37 +1000 (5.65c/1.34) (from jeremy@sour.sw.oz.au for vsta@cisco.com) From: jeremy@sour.sw.oz.au (Jeremy Fitzhardinge) Message-Id: <199410180303.AA28365@sour.sw.oz.au> Subject: Question about swap To: vsta@cisco.com (VSTa list) Date: Tue, 18 Oct 1994 13:03:36 +1000 (EST) Organization: Softway Pty Ltd X-Face: '6U=%Tv\k1l-:?\$C[D@G 7(vl~w8&y}!f\bh#wL#n,TGKh>T.c7eT5-y)Hl'i;A1z$9?*lD.k}yqshddFb l[EC}c=;uc%x'}uh3E91p&oE X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: jeremy@sour.sw.oz.au (Jeremy Fitzhardinge) Cc: vsta@cisco.com (VSTa list) Subject: Re: Question about swap In-Reply-To: Your message of "Tue, 18 Oct 1994 13:03:36 +1000." <199410180303.AA28365@sour.sw.oz.au> Date: Tue, 18 Oct 1994 08:25:21 -0700 From: Andrew Valencia [jeremy@sour.sw.oz.au (Jeremy Fitzhardinge) writes:] >After looking at it for a few days, I still have a very simple >sounding question: where does swap go? Swap is done via the swap daemon. That is, the kernel thinks of swap as a simple array 0..n of blocks, and sends requests to the swap daemon. The swap daemon can maintain multiple swap device (see bin.src/test/swapd.c for an example of how to tell him about new partitions) and send off I/O's to their ultimate destination. He could also do remapping, allowing dynamic expansion and contraction of the swap pool. >... I have a linux swap partition on the same machine >which could be shared, but I can't see how its done. Our libdpart code appears to know about Linux swap. So the _lxswapX partition from the disk server should work fine as the destination for swap. Swapping is, ahem, "lightly" tested. If you start using it seriously you'll probably have to chase some bugs. I got basic swapin/swapout working ages ago, and haven't really hammered upon it since. I've been reviewing the code of late and it still looks workable, but I'm sure there are still basic bugs to be ironed out. Andy From vandys@glare.cisco.com Tue Oct 18 14:15:48 1994 Received: from hubbub.cisco.com (hubbub.cisco.com [198.92.30.32]) by amri.cisco.com (8.3/8.3) with ESMTP id OAA07003; Tue, 18 Oct 1994 14:15:47 -0700 Received: from ns.Novell.COM (ns.Novell.COM [137.65.1.1]) by hubbub.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id PAA08807 for ; Tue, 18 Oct 1994 15:29:57 -0700 From: gary_shea@novell.com Received: from orb.usg.sandy.novell.com by ns.Novell.COM (4.1/SMI-4.1) id AA27477; Tue, 18 Oct 94 16:29:43 MDT To: @novell.com:vsta@cisco.com Date: Tue, 18 Oct 1994 16:27 MDT Subject: Re: Bug in the swapper code Content-Length: 600 Content-Type: text/plain Message-Id: <2ea44bef0.171f@orb.usg.sandy.novell.com> From: Dave Hudson > Hi All, > > After the number of problems people with 4 MByte systems seem to have been > reporting I've been looking at the paging code. There are a few problems > with it in the 1.3.2 release. I've got fixes for some of these already so > hopefully the 4 meg systems will work OK with 1.3.3. > > > Regards, > Dave Yeah, I looked at the swap code for a long time and didn't find any problems there, but I don't know much... haven't been able to experiment yet!!! Turns out I only have 2 Meg, so it's fairly amazing how much _does_ work.... Gary From vandys@glare.cisco.com Wed Oct 19 07:14:25 1994 Received: from beasley.cisco.com (beasley.cisco.com [171.69.2.135]) by amri.cisco.com (8.3/8.3) with ESMTP id HAA07126; Wed, 19 Oct 1994 07:14:23 -0700 Received: from irz301.inf.tu-dresden.de (irz301.inf.tu-dresden.de [141.76.1.11]) by beasley.cisco.com (8.6.8+c/CISCO.GATE.1.1) with SMTP id IAA07130 for ; Wed, 19 Oct 1994 08:25:41 -0700 Received: from ibch50.inf.tu-dresden.de by irz301.inf.tu-dresden.de with SMTP (5.67b+/DEC-Ultrix/4.3) id AA13015; Wed, 19 Oct 1994 16:25:56 +0100 Received: by ibch50.inf.tu-dresden.de (AIX 3.2/UCB 5.64/4.03) id AA15555; Wed, 19 Oct 1994 16:25:09 +0100 Date: Wed, 19 Oct 1994 16:25:09 +0100 From: jw@ibch50.inf.tu-dresden.de (Wittenberger) Message-Id: <9410191525.AA15555@ibch50.inf.tu-dresden.de> To: vsta@cisco.com Cc: jw@ibch50.inf.tu-dresden.de Subject: Q: messagesystem Hello, I got questions I can't answer myself: 1) Correct me if I'm wrong: As far as I understood, a request like "read" is done by turning the given buffer to be the area where the read bytes go and send out a appropriate message. Is there any case yet, where multi segment messages are generated? 2) A message is transfered to the receiver by making segments from the buffer(s) and mapping them to the address space of the receiver. There is no copying. (?) Assumed one task of a process 1 send a message to a second process and a second task of process 1 modifies the data of the message later. What happens? Is this a caught violation, is the data process 2 will read modified or anything else? 3) The rule is: all what can be done in the library goes to the library. So for every kernel service there is a good reason to make it a kernel service. But I can't find a good reason for msg_connect, msg_accept, msg_disconnect (and msg_err). In my oppinion it could be done by the library and under certain circumstances it could be left out. - msg_connect/msg_accept just handles the connection. Easy to code in the library or did I miss some point? And could be left out if a little different protocoll would be used. - msg_disconnect: same as above. For the exit() the server had to send it to all clients at once (library code). A small problem comes up for processes which dump core. But that should not be a problem. - msg_err is a complete different question, but I can't see any problem. 4) A non-blocking send/receive pair would make it easier to build a level with async services between the file-like level and the kernel. We feel to need this. Is there something to say against it? One more idea about kernel services: The kernel entry is expensive (in intel systems at least). The L3-system saved a lot of time by providing a additional service reply_and_receive_next. I think it would be a good idea to provide a library function like that and prefer the use of it over the use of the single calls. So a later optimization by a additional kernel service won't hurd. /Joerg ----------------------------------------------------------------------------- Joerg Wittenberger | email: joerg.wittenberger@inf.tu-dresden.de Rietzstr. 32b | (eventually try: jw@ibch50.inf.tu-dresden.de 01139 Dresden | or jw6@mail.inf.tu-dresden.de) Germany | WWW: jerry PGP PUBLIC KEY: available on request or by finger From vandys@glare.cisco.com Wed Oct 19 07:37:54 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id HAA07132; Wed, 19 Oct 1994 07:37:52 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id IAA25771; Wed, 19 Oct 1994 08:52:07 -0700 Message-Id: <199410191552.IAA25771@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: jw@ibch50.inf.tu-dresden.de (Wittenberger) Cc: vsta@cisco.com Subject: Re: Q: messagesystem In-Reply-To: Your message of "Wed, 19 Oct 1994 16:25:09 BST." <9410191525.AA15555@ibch50.inf.tu-dresden.de> Date: Wed, 19 Oct 1994 08:52:06 -0700 From: Andrew Valencia [jw@ibch50.inf.tu-dresden.de (Wittenberger) writes:] >1) Correct me if I'm wrong: >As far as I understood, a request like "read" is done by turning the >given buffer to be the area where the read bytes go and send out a >appropriate message. Is there any case yet, where multi segment >messages are generated? Things are more subtle than this. The M_READ modifier to an m_op (such as FS_READ) controls whether the buffer is provided to the server, or whether the segments which the servers msg_reply()'s with will be copied out to the buffer. For a typical filesystem read, M_READ is set, the server sends back some segments and they are copied out before the msg_send() completes. For DMA, M_READ is *not* set, although the operation is still an FS_READ. The buffers are thus made available to the DMA server, who will directly write into them. Note that the buffers made available to a normal (non-DMA) server are read-only. DMA server get them read-write, so they can do either physical or programmed I/O. >2) A message is transfered to the receiver by making segments from the >buffer(s) and mapping them to the address space of the receiver. There >is no copying. (?) >Assumed one task of a process 1 send a message to a second process and a >second task of process 1 modifies the data of the message later. What >happens? Is this a caught violation, is the data process 2 will read >modified or anything else? As you can see, this race condition is why you can't just leave pieces of a server's address space attached to a client. >3) The rule is: all what can be done in the library goes to the >library. So for every kernel service there is a good reason to make it >a kernel service. But I can't find a good reason for msg_connect, >msg_accept, msg_disconnect (and msg_err). In my oppinion it could be >done by the library and under certain circumstances it could be left >out. You also can't let misbehaving processes violate security, or break the system. The connection phase authenticates the connecting user to the server. If, for instance, the user got to build his own connect message then something else would have to be done to gain this authentication. A state machine is also applied to protect against the various edge conditions of servers or clients dying during each phase of the connection and ongoing I/O. Nobody (plus or minus a bug!) is ever stranded by the failure of the "other side". >4) A non-blocking send/receive pair would make it easier to build a >level with async services between the file-like level and the >kernel. We feel to need this. Is there something to say against it? Read the QNX papers. They make a good starting point. I had coded async messaging ages ago, and the implementation was a lot more trouble, and your average client benefited little. VSTa has threads, which are used for the rare cases where we want to handle multiple I/O's in parallel (see KA9Q). >One more idea about kernel services: The kernel entry is expensive (in >intel systems at least). The L3-system saved a lot of time by >providing a additional service reply_and_receive_next. I think it >would be a good idea to provide a library function like that and >prefer the use of it over the use of the single calls. So a later >optimization by a additional kernel service won't hurd. It would be interesting to characterize how often a reply is done with another request already on the queue. I'd guess not very often, on a single user system. When we get the VSTa source server machine on the net we could measure under a multi-user load. Converting to reply-and-receive-next would break the structure of many servers (the reply is deeply buried while the receive is at the top of the server loop). It would probably require some global variables, so I'd rather verify the benefit before going in this direction. Andy From vandys@glare.cisco.com Fri Oct 21 16:21:46 1994 Received: from glare.cisco.com (glare.cisco.com [131.108.13.56]) by amri.cisco.com (8.3/8.3) with ESMTP id QAA07582; Fri, 21 Oct 1994 16:21:45 -0700 Received: from localhost.cisco.com (localhost.cisco.com [127.0.0.1]) by glare.cisco.com (8.6.8+c/CISCO.SERVER.1.1) with SMTP id RAA20139 for ; Fri, 21 Oct 1994 17:36:14 -0700 Message-Id: <199410220036.RAA20139@glare.cisco.com> X-Authentication-Warning: glare.cisco.com: Host localhost.cisco.com didn't use HELO protocol To: vsta@amri.cisco.com Subject: VSTa list down this weekend Date: Fri, 21 Oct 1994 17:36:13 -0700 From: Andrew Valencia Due to equipment moves in our data center. E-mail connectivity to me will be similarly interrupted. In a pinch, you can reach me at my wife's account: jtk@netcom.com. Have a good weekend! Andy