From: Digestifier To: Linux-Development@senator-bedfellow.mit.edu Reply-To: Linux-Development@senator-bedfellow.mit.edu Date: Tue, 28 Sep 93 14:18:49 EDT Subject: Linux-Development Digest #133 Linux-Development Digest #133, Volume #1 Tue, 28 Sep 93 14:18:49 EDT Contents: Re: Pascal compiler. (Byron A Jeff) Re: MFM + IDE (Delman Lee) Re: Lpmud on Linux. Alarm Clock. (Matthias Urlichs) Re: xcalentool, xview, and /usr/include/sys/ucontext.h (Kenneth Osterberg) Re: Please no /config. (Harald T. Alvestrand) Re: MFM + IDE (Drew Eckhardt) Re: linux scheduler alternatives??? (Linus Torvalds) Opening a deleted file though /proc (revisited). (Brian McCauley) Re: Opening a deleted file though /proc (revisited). (Linus Torvalds) Bug in Xfree with tvga9000? (Liefting W) Driver for 3c507 anyone? (Steve Starck) Re: Pascal compiler. (Masami Ogoshi) MIT Scheme? (Theodore A. Serreyn) Re: Whence 1.0? (K J MacDonald) ASM I/O (Hugo Eide Gunnarsen) Re: Net-2 Patches are GREAT Re: ASM I/O (Mahesh Neelakanta) ---------------------------------------------------------------------------- From: byron@cc.gatech.edu (Byron A Jeff) Subject: Re: Pascal compiler. Date: Tue, 28 Sep 1993 04:11:47 GMT In article <287tc9$pjq@mtu.edu>, STEVEN J. KANGAS wrote: > Has anyone ported a pascal compiler to linux yet? It seems like >all the compilers (p4, etc.) I can find are all written in pascal. If I >had a pascal compiler to compiler the compiler I wouldn't need a stupid >compiler now would I? Any info would be appreciated. How about using p2c to break the chain? It's not a true pascal compiler. It translates Pascal to C code which can be compiled and linked with the p2c library producing a working program. So maybe you can bootstrap a compiler using p2c. BAJ --- Another random extraction from the mental bit stream of... Byron A. Jeff - PhD student operating in parallel! Georgia Tech, Atlanta GA 30332 Internet: byron@cc.gatech.edu ------------------------------ From: delman@mipg.upenn.edu (Delman Lee) Subject: Re: MFM + IDE Date: 28 Sep 93 03:30:47 GMT In article <1993Sep27.191815.24389@mnemosyne.cs.du.edu> dsharp@nyx.cs.du.edu (Dave Sharp) writes: angelos@scus1.ctstateu.edu (Angelos Karageorgiou Greek and Macedon the only combination) writes: > Is it possible to have both IDE and MFM controllers on the same >machine ? > I know that the bios try to boot from both controllers so it >locks up. Is there a work around ? like disabling some roms ? > Just curious MFM and IDE can not coexist on the same machine. I understand that you can get IDE and SCSI -or- MFM and SCSI to work together, tho. Yes, the clean distributed Linux kernel does not support the co-existence of MFM and IDE drives. However, there is a patch to the kernel that allows Linux to run with two AT-type controllers, one such example would be an IDE and MFM controller. In fact, that's what I am running --- I have two MFM drives and 1 IDE drive. The patch is in, for example, sunsite.unc.edu:pub/Linux/kernel/misc-patches/atdisk* Hope this helps. Delman. -- ______________________________________________________________________ Delman Lee Tel.: +1-215-662-6780 Medical Image Processing Group, Fax.: +1-215-898-9145 University of Pennsylvania, 4/F Blockley Hall, 418 Service Drive, Philadelphia, PA 19104-6021, U.S.A.. Internet: delman@mipg.upenn.edu ______________________________________________________________________ ------------------------------ From: urlichs@smurf.sub.org (Matthias Urlichs) Subject: Re: Lpmud on Linux. Alarm Clock. Date: 27 Sep 1993 22:48:02 +0100 In comp.os.linux.development, article <286nsvINN3uq@mailhost.uni-koblenz.de>, hodgen@ozzy.uni-koblenz.de (Wayne Hodgen) writes: > > I'm having problems getting the driver for our Lpmud working under Linux. On > a Sun, no problem. Under Linux (Slackware 1.0.2) it keeps crashing with an > Alarm clock. > Use BSD signals or re-signal() your alarm handler at its end. -- The price of seeking to force our beliefs on others is that someday they might force their beliefs on us. -- Mario Cuomo -- Matthias Urlichs \ XLink-POP Nürnberg | EMail: urlichs@smurf.sub.org Schleiermacherstraße 12 \ Unix+Linux+Mac | Phone: ...please use email. 90491 Nürnberg (Germany) \ Consulting+Networking+Programming+etc'ing 42 ------------------------------ From: lmfken@lmf.ericsson.se (Kenneth Osterberg) Subject: Re: xcalentool, xview, and /usr/include/sys/ucontext.h Date: Tue, 28 Sep 1993 07:56:12 GMT "Eric Jeschke" writes: >I am trying to compile a version of calentool 2.2X using Ken >Osterberg's Linux XView port. I grabbed the stock sources off >of sunsite (/pub/X11/contrib), configured a makefile and tried >to build. It bombs trying to include >which is nowhere to be found on my box. I'm running .99p12 Yes it is, you just haven't looked closely enough. >upgraded periodically from back to .99p6 or so. I looked at >the openwin/include/xview files and it looks like it assumes >sys/ucontext.h to be in "standard" Linux. This is a kludge I did in the port. Take a look in xview3L5/patches, and you'll see ucontext.h. Copy this to /usr/include/sys as described in the README. Others have pointed out that a better solution is to #undef SYSV_UCONTEXT in base.h in the xview header files. By doing this, you won't need the dummy ucontext.h while compiling the xview library. Either way should work. >Q: has this file been added to the default distributions at some >point and I have missed it? Anybody else compiled calentool for >Linux? If you have could you drop me a line about how you did it? Kenneth -- Kenneth Osterberg lmfken@lmf.ericsson.se or lmfken@bluese1.ericsson.fi ------------------------------ From: hta@uninett.no (Harald T. Alvestrand) Subject: Re: Please no /config. Date: 28 Sep 1993 10:41:14 GMT A not uncommon practice for both keeping the configuration files with the packages they belong to and keeping them in /etc is to make (for instance) /usr/lib/X11/xdm-config be a symlink to /etc/xdm-config. This puts the name in the "standard" place, and the content in some place that changes by CPU. -- Harald Tveit Alvestrand Harald.T.Alvestrand@uninett.no G=Harald;I=T;S=Alvestrand;O=uninett;P=uninett;C=no +47 7 59 70 94 My son's name is Torbjørn. The letter between "j" and "r" is o with a slash. ------------------------------ From: drew@pipkin.cs.colorado.edu (Drew Eckhardt) Subject: Re: MFM + IDE Date: Tue, 28 Sep 1993 09:35:33 GMT In article <1993Sep27.191815.24389@mnemosyne.cs.du.edu> dsharp@nyx.cs.du.edu (Dave Sharp) writes: >angelos@scus1.ctstateu.edu (Angelos Karageorgiou Greek and Macedon the only combination) writes: > > >> Is it possible to have both IDE and MFM controllers on the same >>machine ? >> I know that the bios try to boot from both controllers so it >>locks up. Is there a work around ? like disabling some roms ? >> Just curious > >MFM and IDE can not coexist on the same machine. I understand that >you can get IDE and SCSI -or- MFM and SCSI to work together, tho. I have IDE, MFM, and SCSI disks in my system (two SCSI boards), and it does work. With the IDE/MFM boards, you need to jumper one at a secondary address, or use an XT MFM board. You also need to jumper one for a different IRQ - if this is not possible, you'll have to perform some corrective surgery on the board, I used an exacto knife and conductive pen to change the IRQ to 15. -- Boycott USL/Novell for their absurd anti-BSDI lawsuit. | Drew Eckhardt, Condemn Colorado for Amendment Two. | Professional Linux Use Linux, the fast, flexible, and free 386 unix | Consultant Will administer Unix for food | drew@cs.Colorado.EDU ------------------------------ From: torvalds@klaava.Helsinki.FI (Linus Torvalds) Subject: Re: linux scheduler alternatives??? Date: 28 Sep 1993 13:18:01 +0200 In article <286i88INN9rh@fstgds15.tu-graz.ac.at>, Klaus ZLOEBL wrote: >I found some time last weekend and read the khg >after looking around in the kernel sources i had to notice >that i don't fine the place where the timeout timer is set. >So could somebody be so nice and tell me where to find it. It's not there. Linux just uses the normal timer to keep the heartbeat: after all, we don't allow just anybody to mess with interrupts, so the timeout timer (I assume you meant the one that sends an NMI) isn't needed. Actually, I've been thinking of using it for a better kernel profiler, as the current profiler doesn't work too well in interrupt handlers (understatement of the week), and one based on NMI would work. I've never bothered to find out how it works, though. >Also i'd like to write a new scheduler (multilevel feedback queue) There has been one that used something like this, and reportedly worked better under heavy load. I have to admit to liking the current one, though: simple and reasonably well-behaved. >What algorithm is use for paging? It's essentially a modified "clock" algorithm, which seems to work reasonably well. I once tried a NRU-like thing, but it was so good at finding pages that belonged to programs that were waiting for user input that I gave up on it. I could probably have tuned it to give better performance, but the clock algorithm is reasonably fair, needs no tuning, and doesn't give *too* bad results. Linus ------------------------------ From: bam@wcl-l.bham.ac.uk (Brian McCauley) Subject: Opening a deleted file though /proc (revisited). Date: 28 Sep 1993 12:50:44 GMT Reply-To: B.A.McCauley@bham.ac.uk A while pack I posted to c.o.l.help to find out if anyone knew why: $cat /dev/stdin < X cat: /dev/stdin: Permission denied (Note /dev/stdin -> /proc/self/fd/0) Linus said the reason was in linux/fs/namei.c: int permission(struct inode * inode,int mask) { int mode = inode->i_mode; /* special case: not even root can read/write a deleted file */ if (inode->i_dev && !inode->i_nlink) return 0; But he also said he couldn't remember why this was there. I've commented it out in my kernel now - anyone got any ideas why this could be a bad idea. The most obvious reason (the possiblity of someone else deleting your pwd) seems not to be a problem. What others can people think of? -- \\ ( ) No Bullshit! | Email: B.A.McCauley@bham.ac.uk . _\\__[oo from | Voice: +44 21 471 3789 (home) .__/ \\ /\@ /~) /~[ /\/[ | +44 21 627 2171 (work) . l___\\ /~~) /~~[ / [ | Fax: +44 21 627 2175 (work) # ll l\\ ~~~~ ~ ~ ~ ~ | Snail: 197 Harborne Lane, B29 6SS, UK ###LL LL\\ (Brian McCauley) | ICBM: 52.5N 1.9W ------------------------------ From: torvalds@klaava.Helsinki.FI (Linus Torvalds) Subject: Re: Opening a deleted file though /proc (revisited). Date: 28 Sep 1993 16:22:46 +0200 In article , Brian McCauley wrote: >A while pack I posted to c.o.l.help to find out if anyone knew why: > >$cat /dev/stdin <> X >cat: /dev/stdin: Permission denied > >(Note /dev/stdin -> /proc/self/fd/0) > >Linus said the reason was in linux/fs/namei.c: > >int permission(struct inode * inode,int mask) >{ > int mode = inode->i_mode; > >/* special case: not even root can read/write a deleted file */ > if (inode->i_dev && !inode->i_nlink) > return 0; > >But he also said he couldn't remember why this was there. I've >commented it out in my kernel now - anyone got any ideas why this >could be a bad idea. The most obvious reason (the possiblity of >someone else deleting your pwd) seems not to be a problem. What others >can people think of? Actually, I thought about it, and as far as I can remember, the check is actually there due to a *very* early kernel that wanted this to handle deleted working directories gracefully. We are talking version 0.10 or something like that: deleting a directory from under a user resulted in weird errors without this, I seem to remember. Note that my memory isn't what it used to be (and it has never really been that good), but the above is kind of a hazy recollection of what might have been the reason. The test should be safe to remove now (deleting a pwd should be checked elsewhere in the current kernels), and I just did so in my kernel. Linus ------------------------------ From: wlieftin@cs.vu.nl (Liefting W) Subject: Bug in Xfree with tvga9000? Date: Tue, 28 Sep 1993 14:46:44 GMT Hello. I have just installed X on my system, and I have a funny problem. 640x480x256color no problem, but when I Ctrl-Alt-+ to 800x600x256color, the screen gets messed up. But... not like it would when I had the wrong sync numbers. The screen _IS_ stable, but black spickles get all over it. And when I switch back to 640x480, they stay so. I have to refresh the screen to get them away. Also, after stopping X, some letters (in normal screen mode) seem to have some parts missing. I have a trident 9000 with 512K mem. Any solution (besides buying something like an ET4000)? Wouter. ------------------------------ From: sstarck@uswnvg.com (Steve Starck) Crossposted-To: comp.os.linux.misc Subject: Driver for 3c507 anyone? Date: 28 Sep 93 14:39:43 GMT Realizing that the Ethernet HOWTO contains the line: 3c507 -- Not available at present. I was wondering if anyone has been working on a driver for this Ethernet card, since my PC at work has a 3c507-TP and I'd like to put Linux on it and have it connected to our net. I might end up rolling my own, but with my current project load, I don't really have the time right now and was wondering if anyone was working on something. I might be able to provide some time to "alpha" or "beta" test whatever is available. Thanks In Advance -- Steve Starck sstarck@uswnvg.com Sr. Systems Administrator U S WEST NewVector Group, Inc. (206)562-5024 ------------------------------ From: ogochan@jh4tjwgw.jh4tjw.prug.or.jp (Masami Ogoshi) Subject: Re: Pascal compiler. Date: 28 Sep 93 11:04:04 GMT Reply-To: ogochan@jh4tjwgw.jh4tjw.prug.or.jp (Masami Ogoshi) In article <287tc9$pjq@mtu.edu> sjkangas@major.cs.mtu.edu (STEVEN J. KANGAS) writes: > Has anyone ported a pascal compiler to linux yet? It seems like >all the compilers (p4, etc.) I wrote a Pascal compiler on MS-DOS, which likes Pascal-P4. It was written in C and it generates byte code, And It have little optimizer. It is easy to port for Linux (I think). -- ogochan@jh4tjw.prug.or.jp Masami Ogoshi 109 Okudani-cho Matsue city Shimane pref 690 JAPAN ------------------------------ Crossposted-To: comp.os.linux.help From: tserreyn@nyx.cs.du.edu (Theodore A. Serreyn) Subject: MIT Scheme? Date: Tue, 28 Sep 93 15:47:04 GMT Has anyone ported the 7.1 or 7.2 versions of MIT Scheme to LINUX yet? I'm looking for a decent version that doesn't respind #undefined when I define a procedure. Any help or pointer to good scheme compilers would be much appreciated. Ted Serreyn tserreyn@nyx.cs.du.edu serreyta@cnsvax.uwec.edu serreyta@adam.cs.uwec.edu ------------------------------ From: kenny@festival.ed.ac.uk (K J MacDonald) Subject: Re: Whence 1.0? Date: Tue, 28 Sep 1993 16:44:05 GMT Kevin Brown (kevin@frobozz.sccsi.com) wrote: : In article <285l1d$4u6@senator-bedfellow.MIT.EDU> dthumim@athena.mit.edu (Daniel J Thumim) writes: : > It seems that the last few patchlevel releases have not been any more : >stable than previous ones. This is not good because people usually just : >grab the latest kernel release, and if it's buggy that looks bad for : >linux. Maybe we can split the kernel into two "releases"... one which : >would have no new non-essential features added, but would only be debugged : > Just my two cents... [stuff deleted] : I basically agree, though I think it would work best overall if done this : way: : There are two versions in each release. One is a bug fix version. One : is a "new features" version. : For each release, the "bug fix" version is the prior release's "new : features" version with bug fixes, and the "new features" version is Perhaps everybody should remember that Linux is BETA software and if you don't want a new (buggier) kernel, then just wait for a week or so before upgrading. It's all in the USERS hands as to which version of the kernel he/she uses. Don't load Linus with extra stuff, he's doing a marvellous job at the moment. This isn't a flame, just a reminder. Happy Linuxing, Kenny. -- ============================================================================== Kenneth MacDonald E-mail: kenny@ed Dept. of Geology & Geophysics "Allow me to introduce myself, Major Dennis University of Edinburgh Bloodnok, International Christmas Pudding ------------------------------ From: hugogu@slagg.met.sintef.no (Hugo Eide Gunnarsen) Subject: ASM I/O Date: Tue, 28 Sep 93 16:19:11 GMT Hi. I want to be able to write/read a word from a I/O port. Now I'm using inb/outb, but what I need is inw/outw. I tried to change the code for inb/outb, but it failed to work. SO it anybody know how to read/write a word (16b) from/to a I/O port, pleas tell me. Hugo E. Gunnarsen hugogu@slagg.met.sintef.no ------------------------------ Crossposted-To: comp.os.linux.misc From: dan@archimedes.lance.colostate.edu () Subject: Re: Net-2 Patches are GREAT Date: Tue, 28 Sep 1993 17:30:12 GMT Reply-To: dan@lance.colostate.edu In article , jan@swi.psy.uva.nl (Jan Wielemaker) writes: |> jmorriso@rflab.ee.ubc.ca (John Paul Morrison) writes: |> |> >OK, what's the secret? I can get pl13 with net-2e4 running, although |> >it isnt very stable. I'm about to give up on net-2e, so I downloaded |> >the newest pl13 source and followed the net2-debugged instructions. The |> >instructions tell you to move your inet directory, then unpack the new |> >code. No Makefile or CONFIG is included with net2-debugged, so I |> >assume you are supposed to copy the old ones (from inet.old!). |> |> >After compiled and booting, the new net2 stuff won't even detect my ethercard. |> |> Tried exactly the same. I've got no network at home, but we're trying to |> get the network stuff going at the University. Even using the loopback |> net, just `rlogin myself' followed by logout crashed the system without |> any message!? At least that works on plain pl13. Now I trust that the |> inet-D package is debugged, but how does one get it *properly* linked |> to the kernel? |> |> The first thing I want are *reliable* remote running X11 applications. |> The second thing I want is nfs. NIS would be the greatest! |> |> Thanks --- Jan Whats the deal? I've install MCC pl10+ and XFee86-1.3. I have a cheapo NE200 clone card (Allied Telesis). Within 20 minutes of installing the base package and tcp/ip I was ftp'ing the X stuff. Within an hour of installing X I was displaying remote X stuff with no problem (xrn, xmh, netrek ;) I have no problems - none! Except that the pl10+ memory management is kind of sophomoric, but it works if I dont try to display a 2000x2000x256 jpeg. I ftp, telnet and display remote X all the time - and I've only crashed once, due to the memory management. After reading messages like the above I'm in no big hurry to update. =========================================================================== Dan Doner College of Engineering Colorado State University dan@lance.colostate.edu ------------------------------ From: mahesh@cybernet.cse.fau.edu (Mahesh Neelakanta) Subject: Re: ASM I/O Date: Tue, 28 Sep 1993 17:29:42 GMT In article <1993Sep28.161911.27630@ugle.unit.no> hugogu@slagg.met.sintef.no (Hugo Eide Gunnarsen) writes: >I tried to change the code for inb/outb, but it failed to >work. > >SO it anybody know how to read/write a word (16b) from/to a >I/O port, pleas tell me. > >Hugo E. Gunnarsen >hugogu@slagg.met.sintef.no > Here is what I use. I got this from both the XFree86 (1.3 I think) sources and the [s]vgalib packages. Linux provides two ways (that I know of) for I/O privs..the iopl() command and the ioctl method (KDADDIO). The former is quick-and-dirty but is more dangerous since it will allow access to all io ports. /* inline.h */ #ifndef LNX_INLINE_H #define LNX_INLINE_H static __inline__ void outb(port, val) short port; char val; { __asm__ __volatile__("outb %0,%1" : :"a" (val), "d" (port)); } static __inline__ void outw(port, val) short port; short val; { __asm__ __volatile__("outw %0,%1" : :"a" (val), "d" (port)); } static __inline__ unsigned int inb(port) short port; { unsigned char ret; __asm__ __volatile__("inb %1,%0" : "=a" (ret) : "d" (port)); return ret; } static __inline__ unsigned int inw(port) short port; { unsigned short ret; __asm__ __volatile__("inw %1,%0" : "=a" (ret) : "d" (port)); return ret; } /* End of inline.h */ #endif LNX_INLINE_H ============================================================================== Metro Link Incorporated. 2213 W. McNab Rd. Pompano Beach, Florida 33069 X11.5 and OSF/Motif for QNX, SVR3, SVR4.[012], SCO, Linux, UnixWare, LynxOS, AT&T, Venix, ISC, Solaris, Pyramid, SunOS Voice: +1.305.970.7353 Fax: +1.305.970.7351 Email: mahesh@metrolink.com ============================================================================== ------------------------------ ** FOR YOUR REFERENCE ** The service address, to which questions about the list itself and requests to be added to or deleted from it should be directed, is: Internet: Linux-Development-Request@NEWS-DIGESTS.MIT.EDU You can send mail to the entire list (and comp.os.linux.development) via: Internet: Linux-Development@NEWS-DIGESTS.MIT.EDU Linux may be obtained via one of these FTP sites: nic.funet.fi pub/OS/Linux tsx-11.mit.edu pub/linux sunsite.unc.edu pub/Linux End of Linux-Development Digest ******************************