From: Digestifier To: Linux-Development@senator-bedfellow.mit.edu Reply-To: Linux-Development@senator-bedfellow.mit.edu Date: Sat, 10 Sep 94 06:13:07 EDT Subject: Linux-Development Digest #148 Linux-Development Digest #148, Volume #2 Sat, 10 Sep 94 06:13:07 EDT Contents: Re: Don't use Linux?! (Ahmed Naas) Re: Alpha Linux (H. Peter Anvin) 320x200 X resolution? (Sam Oscar Lantinga) Re: Survey: who wants f77,cc,c++,hpf for linux? (stockman harlan w) Re: Survey: who wants f77,cc,c++,hpf for linux? (stockman harlan w) Re: ATI Mach64... Does it work...? (Thomas Pfarr) Re: Mach64 XServer 90MHz limitation (Baba Buehler) Re: Developing Distributed Filesystems for Linux? (John F Carr) AFS Client Kernel Patch. . . (Grungie The Wise) Re: A thought to improve security (Craig Woodward) Re: Future of linux -- the sequel (Darin Johnson) Re: DOSEMU 0.53p17 & mouse (Alberto Alonso; shisho or albund) Re: Future of linux -- the sequel (Philippe Steindl) IP multicast with linux? (Yufan Hu) ---------------------------------------------------------------------------- From: ahmed@oea.xs4all.nl (Ahmed Naas) Subject: Re: Don't use Linux?! Date: Fri, 9 Sep 1994 17:51:09 GMT Michael Schumacher (hightec@sbusol.rz.uni-sb.de) wrote: : PS: See 4. ;-) Ok, so all this racket is about introducing your shareware program :-> Were can we get it from? -- The above is a result of random neuron activity in the writer's brain. Ahmed M. Naas ahmed@oea.xs4all.nl ====================================================================== ------------------------------ Crossposted-To: comp.lang.c From: hpa@ahab.eecs.nwu.edu (H. Peter Anvin) Subject: Re: Alpha Linux Reply-To: hpa@nwu.edu (H. Peter Anvin) Date: Sat, 10 Sep 1994 04:33:24 GMT [This is a general C issue so I am crossposting to comp.lang.c and setting followups there.] Followup to: By author: coleman@math36.gatech.edu (Richard Coleman) In newsgroup: comp.os.linux.development > > > : Why drop one? > > : 16 bits = short int > > : 32 bits = int > > : 64 bits = long > > > > 128 bits = long long :) > > > I've always thought that C should have some way > of letting you decided how many bytes to use > for your computation. It could be as simple as > having the types > > int8 ( = char ) > int16 ( = short ) > int24 > int32 ( = int ) > ... > > up to some reasonable value like > int128. Of course, other syntatic sugar > would probably be necessary, but the idea > *appears* simple enough. > Here is a program I wrote that does just that: /* * typesize.c -- Generate typesize.h header file * * Written in 1994 by H. Peter Anvin * * This program is public domain, and makes the assumptions that: * * sizeof(char) == 1 (believe this one is true by definition) * char == 8 bits unless CHAR_BIT is defined (in limits.h) * * The following external defines needs to be found out by an autoconf- * style configure script or something like it: * * HAS_LIMITS_H -- The include file exists * HAS_LONG_LONG -- The integral datatype "long long" exists * HAS_QUAD -- The integral datatype "quad" exists * HAS_LONG_QUAD -- The integral datatype "long quad" exists * HAS_LONG_DOUBLE -- The floating-point datatype "long double" exists * * (The above defines do *not* require the datatypes to be distinct, * just that they will compile without error.) * */ #ifdef HAS_LIMITS_H #include #endif #include char taken[256]; /* A primitive way of marking territory */ #ifdef CHAR_BIT int bytesize = CHAR_BIT; /* The size of a byte */ #else int bytesize = 0; /* Find out experimentally */ #endif markint(f,length,name) FILE *f; int length; char *name; { if ( length < 1 || length > 256 ) return 1; /* Don't know what this is */ if ( !taken[length-1] ) { taken[length-1] = 1; fprintf(f,"\n/* %d-bit integers */\n", length*bytesize); fprintf(f,"#define HAS_INT%d 1\n", length*bytesize); fprintf(f,"typedef %s int%d;\n", name, length*bytesize); fprintf(f,"typedef unsigned %s uint%d;\n", name, length*bytesize); } return 0; } markfloat(f,length,name) FILE *f; int length; char *name; { if ( length < 1 || length > 256 ) return 1; /* Don't know what this is */ if ( !taken[length-1] ) { taken[length-1] = 1; fprintf(f,"\n/* %d-bit floats */\n", length*bytesize); fprintf(f,"#define HAS_FLOAT%d 1\n", length*bytesize); fprintf(f,"typedef %s float%d;\n", name, length*bytesize); } return 0; } main() { FILE *f; int i; unsigned char ch; if ( sizeof(char) != 1 ) { fprintf(stderr,"Weird compiler: sizeof(char) != 1\n"); exit(1); } f = fopen("typesize.h","wt"); if ( f == NULL ) { fprintf(stderr,"Could not create typesize.h\n"); exit(1); } fprintf(f,"\ /*\n\ * typesize.h\n\ *\n\ * Defines size-specific aliases for integer and floating-point data\n\ * types, and defines conditionals to find out about them. This file\n\ * is compiler- and architecture specific and was automatically generated\n\ * by typesize.c. Do not use except on the system and compiler that\n\ * compiled typesize.c.\n\ *\n\ */\n\n"); if ( !bytesize ) for ( ch = 1 ; ch ; ch = ch << 1 ) /* Not all C compilers like <<= */ bytesize++; fprintf(f,"/* The size of a byte (might be important) */\n"); #ifdef HAS_LIMITS_H fprintf(f,"#include \n"); #ifdef CHAR_BIT fprintf(f,"#define BYTESIZE CHAR_BIT\n"); #else fprintf(f,"#define BYTESIZE %d\n", bytesize); #endif #else fprintf(f,"#define BYTESIZE %d\n", bytesize); #endif for ( i = 0 ; i < 256 ; i++ ) taken[i] = 0; markint(f,sizeof(int),"int"); markint(f,sizeof(char),"char"); markint(f,sizeof(short),"short"); markint(f,sizeof(long),"long"); #ifdef HAS_LONG_LONG markint(f,sizeof(long long),"long long"); #endif #ifdef HAS_QUAD markint(f,sizeof(quad),"quad"); #endif #ifdef HAS_LONG_QUAD markint(f,sizeof(long quad),"long quad"); #endif for ( i = 0 ; i < 256 ; i++ ) taken[i] = 0; markfloat(f,sizeof(double),"double"); markfloat(f,sizeof(float),"float"); #ifdef HAS_LONG_DOUBLE markfloat(f,sizeof(long double),"long double"); #endif fclose(f); exit(0); } -- INTERNET: hpa@nwu.edu --- Allah'u'abha --- IBM MAIL: I0050052 at IBMMAIL HAM RADIO: N9ITP or SM4TKN FIDONET: 1:115/511 or 1:115/512 STORMNET: 181:294/1 or 181:294/101 "NT is not a bad thing if I don't have to use it..." -- xmsb@borland.com ------------------------------ Crossposted-To: comp.os.linux.misc From: slouken@cs.ucdavis.edu (Sam Oscar Lantinga) Subject: 320x200 X resolution? Date: Thu, 8 Sep 1994 22:54:22 GMT Does such a thing exist? How would I go about finding the dot-clocks, etc for this resolution? Does XFree86 3.1 have it? Where can I get it? Thanks for any information.... -Sam Lantinga (slouken@cs.ucdavis.edu) P.S. DOOM for X exists, and will hopefully be released soon. ------------------------------ From: hwstock@snll-arpagw.ca.sandia.gov (stockman harlan w) Crossposted-To: comp.lang.fortran Subject: Re: Survey: who wants f77,cc,c++,hpf for linux? Date: 8 Sep 94 15:02:23 GMT In article , Michel Anders wrote: >In <34m769$bju@indy.pgroup.com> lfm@pgroup.com (Larry Meadows) writes: >>1. Are people interested in a commercial compiler suite for Linux on >> Intel Architecture platforms? The suite would include true compilers >> for extended Fortran 77, ANSI C, Draft-ANSI C++ with extensions, and >> High Performance Fortran. C, f77, and C++ could support shared memory [...] >GCC/G++ (and f2c) are excellent compilers and free..., and since i'm not [...] >For non-commercial programmers like me, i fail to see how a commercial >compiler could surpass the available GNU software. It would be interesting In my experience, the executables produced by f2c + C compilers can be 3 times slower than executables produced by a good FORTRAN compiler. You won't see this big a difference with a 486, where the average floating point instruction takes about 14 clock ticks, but it will show up on a Pentium or RISC chip. The problem has to do with the way f2c translates FORTRAN multi-dimensional arrays into 1-D arrays with lots of integer multiplications. Most C compilers end up assuming you must really want all those integer multiplications for some reason. ------------------------------ From: hwstock@snll-arpagw.ca.sandia.gov (stockman harlan w) Crossposted-To: comp.lang.fortran Subject: Re: Survey: who wants f77,cc,c++,hpf for linux? Date: 8 Sep 94 15:08:19 GMT In article <34mton$289@charm.magnus.acs.ohio-state.edu>, >reasons: 1.) no native FORTRAN compiler, and 2.) i[345]86 floating point >performance sucks big fuzzy Volkswagens through a straw. Anybody tried That's a rather large range. A 90 MHz Pentium runs linpack 60 times faster than a 33 MHz 386/387. ------------------------------ From: tpfarr@cstgmail.gsfc.nasa.gov (Thomas Pfarr) Crossposted-To: comp.os.linux.help,comp.windows.x.i386unix Subject: Re: ATI Mach64... Does it work...? Date: Fri, 9 Sep 1994 12:44:10 >In article , deuelpm@craft.camp.clarkson.edu (Pete Deuel) says: >> >>In article pwalker@pinocchio.encore.com (Pete Walker) writes: >>>From: pwalker@pinocchio.encore.com (Pete Walker) >>>Subject: ATI Mach64... Does it work...? Lets try this again...... I have successfully used the alpha SVGA driver for the mach64 PCI card. never the less I have the following observation/questions... I have a 21" monitor (Nanao f760i) and the current Alpha driver limits the dot clock to a maximum of 90MHz if you specify the clocks in Xconfig. If you don't specify clocks and the driver probes the card then the code will set it to the proper 135MHz maximum. Does anyone know if you can override the setting of the max clock in the code at runtime? If I don't specify clocks in Xconfig then it will allow the 110Mhz clock needed for 1280x1024 @60HzV. But I have not found a mode setup which will work at this dot clock. (works fine with the 1152x8xx 90Mhz setting in the Xconfig supplied by the Alpha driver example) Does anyone have 110, 1280x1024 settings for the Nanao f760i monitor? Thanks in advance... Tom Pfarr with the 1152x8xx 90Mhz setting in the Xconfig supplied by the Alpha driver example) Does anyone have 110, 1280x1024 settings for the Nanao f760i monitor? Thanks in advance... Tom Pfarr Computer Sciences Corporation tpfarr@cstgmail.gsfc.nasa.gov ------------------------------ From: baba@ph-meter.beckman.uiuc.edu (Baba Buehler) Subject: Re: Mach64 XServer 90MHz limitation Date: 9 Sep 94 13:29:15 GMT Reply-To: baba@beckman.uiuc.edu wloh@panix.com (Weng Loh) writes: >I am currently playing with "alpha" driver for the ATI >Mach64 chipset XServer posted on sunsite.unc.edu. >The configured clocks for that chipset goes up to 135Mhz >(or thereabout) but the server does not like seeing any >clock set above 90Mhz. the "alpha" SVGA server for the Mach64 cannot do anything over a 90Mhz dot-clock, because it is using the card as a simple svga card, not as an accelerator. -- %>- Baba Z Buehler %>- Beckman Institute Systems Services, Urbana Illinois %>- WWW: http://www.beckman.uiuc.edu/groups/biss/people/baba/ %>- PGP Public Key available via WWW & public key servers ------------------------------ From: jfc@ATHENA.MIT.EDU (John F Carr) Crossposted-To: alt.filesystems.afs Subject: Re: Developing Distributed Filesystems for Linux? Date: 10 Sep 1994 08:00:44 GMT There is nearly enough freely available documentation to figure out and implement the AFS protocol _given a working RPC layer_. AFS uses an RPC protocol called "RX". RX has no value except for compatibility with other software using it -- you wouldn't choose it for a new product. There is enough freely available documentation to get a general idea how RX works; a lot of information has been spread by word of mouth. I spent some time a while back reverse engineering the AFS and RX protocols based on publicly available documentation and a network monitor (because even if you make the mostly valid assumption that C data structures map directly to network data structures, the documentation about the meaning of the fields is incomplete and at times incorrect). I ended up with a proof of concept: it is possible to write a program independent of Transarc copyrighted source which will read or write an AFS file most of the time. I don't have anything I would call an AFS implementation, though I may turn it into one eventually. I don't have any code I'm willing to release. Since this is crossposted to a Linux group I should also point out that I no longer use Linux (I saw the article in an AFS group) and I do not use the GPL for my software. If a group does form to work on a new implementation of AFS I may be able to answer questions from developers about the protocol. In the near term there is a port of Transarc AFS to Linux in testing at MIT. If it is released you would still need an AFS license to use it. If the RMS "contributory GPL infringement" theory is correct then releasing this version would be illegal. Note that AFS uses encryption so a full-featured version may have export restrictions. -- John Carr (jfc@mit.edu) ------------------------------ From: slash@cyclone.Stanford.EDU (Grungie The Wise) Subject: AFS Client Kernel Patch. . . Date: 9 Sep 1994 16:48:36 -0700 Ive heard rumors that AFS support is being developed for linux. Please tell me they were true? If so, what is the state of the project? Is there a beta site for it? Any info you have on it would be a great help! Thanks, Jeff -- __________------== Jeff Townsend ==------___________ ____----DCC Consultant - Guitarist - CS Major - Simpsons Fan----____ -==slash@cyclone.stanford.edu jefft@xenon god@cs grungie@leland ==- ----_______"Yes sir. Very much so sir. Obviously insane."_______---- ------------------------------ From: cdw@cci.com (Craig Woodward) Subject: Re: A thought to improve security Date: Thu, 8 Sep 1994 21:26:57 GMT In article <34jvth$ovl@nic.wi.leidenuniv.nl>, J.A.vanderMost wrote: >Just a thought : > >Some programs in a Unix system have to be SUID root to do the things they do. >Like /bin/passwd /bin/login /usr/bin/lp /user/bin/at etc.. are all SUID root. > >But being root is MUCH to powerful for these programs, they don't need all Agreed, thats why most run SGID on my system. I give then GROUP access to things. lp runs as group lpdev, mtools runs as user/group disk. The big thing with Unix (and Linux) is that you CAN give things partial permissions, or ALL of them if you want. >Let's say lpr has a security bug in it, it allows a normal user that knows >witch options, etc. to use, to modify a file that this user couldn't normally >modify. This is a very harmful bug, because this user can easily become root! Same with finger, if someone's stupid enough to run it as root. It should run as nobody, with no privs what so ever (thus requiring ***4 permissions on .plan files). Simply put, your contrived system eats user ids, and does nothing that you can't do with groups. (Thats why groups exist after all...) >Now my suggestion : >Let's modify the kernel a bit,.... If I had a penny for every time someone wanted to change the kernel... Why not compile the shell right into the kernel? Heck, MSDOG did it. >All comments are appreciated, especially from kernel-hackers. Comment: RTFM and realize that just because a program WANTS to be SUID something, doesn't mean it NEEDS to be SUID something. >* America may be unique in being a country * >* which has leapt from barbarism to * >* decadence without touching civilization * Pahhahaa aha ahahha ha ha ha... oh.. thats funny, all considered. -Woody -- --- I represent my own ideas. Yes, I DO think, all on my own, without the collective. Sure, you may think "NT's just putting him up to this", and you my be right, but thats only what I say, not why I say it. :) ------------------------------ From: djohnson@seuss.ucsd.edu (Darin Johnson) Subject: Re: Future of linux -- the sequel Date: 10 Sep 1994 00:32:05 GMT > My little ISA bus 486 has little trouble besting low-end workstations > in anything (Sun 4/110's and Sparcstation 1's). It will do some integer > ops faster than a Sun 2. The Sun 2's FPU will eat mine alive but for > $10000 it should beat it on *ALL* counts. It doesn't. A Sun *2*? The one they stopped supported in the mid-80's? It had a 68010 in it! Maybe you mean a Sparcstation 2? So then, compare the DMA speed of your ISA system to a Sparcstation (or even a lowly VME system like a Sun 3). There's more to a computer's performance than cpu speed. -- Darin Johnson djohnson@ucsd.edu - I'm not a well adjusted person, but I play one on the net. ------------------------------ From: a_alonso@iastate.edu (Alberto Alonso; shisho or albund) Subject: Re: DOSEMU 0.53p17 & mouse Date: 9 Sep 94 16:58:16 GMT In <34k0q5$eu4@mirage.unipv.it> clint@hal9000.unipv.it (Francesco Defilippo ) writes: >Hi, when i exit from dosemu selection doesn't work, >I'v linux 1.1.49 & dosemu 053p17 A temporary fix. Kill selection and startit again. Then think about getting a diferent mouse driver (I think this is the reason, but I'm not sure) like the msc0007 (I don't know the right spelling but is something like that.) It is available at OAK. Alberto -- -- Alberto Alonso e-mail: a_alonso@iastate.edu Phone: never heard of it WWW: http://www.public.iastate.edu/~a_alonso/homepage.html Electrical Engineering. (At least trying to be one) =============================================================================== =============================================================================== For PGP signature finger a_alonso@iastate.edu ------------------------------ From: psteindl@il.us.swissbank.com (Philippe Steindl) Subject: Re: Future of linux -- the sequel Date: 8 Sep 1994 08:46:55 GMT Reply-To: ilg@imp.ch [some stuff deleted] > BTW, the entry level Indy used to be the Indy 4400PC, and the 4400's > performance without a secondary cache is pretty poor. It's still at > least as fast as a DX66. Hullo, just to kill some misinformation :-) The entry Indy used to be the R4000PC Indy, now it is the R4600PC Indy. The R4400 isn't an entry CPU and I haven't seen any configuration without second level cache. Just my 0.000002$ Philippe -- ====================+=================================================== Philippe Steindl | Any opinions expressed are my own E-mail: ilg@imp.ch | and not necessarily those of the | Swiss Bank Corporation. ====================+=================================================== ------------------------------ From: yufan@iscm.ulst.ac.uk (Yufan Hu) Subject: IP multicast with linux? Date: 10 Sep 1994 10:44:02 GMT Hello, Does Linux support or plan to support IP multicasting? Tnanks. Yufan ------------------------------ ** 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 ****************************** .