Received: from PACIFIC-CARRIER-ANNEX.MIT.EDU by po7.MIT.EDU (5.61/4.7) id AA12683; Mon, 4 Dec 95 22:05:29 EST Received: from SENATOR-BEDFELLOW.MIT.EDU by MIT.EDU with SMTP id AB03268; Mon, 4 Dec 95 16:23:37 EST Received: (from root@localhost) by senator-bedfellow.MIT.EDU (8.6.12/2.3JIK) id QAA28023 for Linux-Development-System-Dist@senator-bedfellow.mit.edu; Mon, 4 Dec 1995 16:13:50 -0500 Message-Id: <199512042113.QAA28023@senator-bedfellow.MIT.EDU> From: Digestifier To: Linux-Development-System@senator-bedfellow.MIT.EDU Reply-To: Linux-Development-System@senator-bedfellow.MIT.EDU Date: Mon, 4 Dec 95 16:13:46 EST Subject: Linux-Development-System Digest #86 Linux-Development-System Digest #86, Volume #2 Mon, 4 Dec 95 16:13:46 EST Contents: Re: Standardized Printing for Linux (Gregory H. Margo) compiling problems (Hong Niu) ELF linking bug? (Jan Wielemaker) Re: Taligent is disbanded - could Linux pick up the pieces? (Nicolas Grilly) Re: How do you compile an old module? (Thomas Koenig) Re: Q: joystick-0.7.3 driver, thrustmaster programming specs (Eyal Lebedinsky) Re: ELF linking bug? (Wolfram Gloger) libg++.so.27? (Stefan Froehlich) Re: Where the offical place for libc? (Thomas Quinot) Re: Where the offical place for libc? (Louis J. LaBash Jr.) Re: can't compile with inb(), outb(); linking errors (Riku Saikkonen) Re: How do I call my own system calls.. (Michael Elizabeth Chastain) pb with 1.2.13 (Jerome Etienne) Re: Standardized Printing for Linux (David Fox) Re: Problems with 1.2.13? (Conor John Cunningham) Re: Latest Stable Kernel ??? (Sean O'Halloran) ---------------------------------------------------------------------------- From: newton@shellx.best.com (Gregory H. Margo) Subject: Re: Standardized Printing for Linux Date: 1 Dec 1995 09:37:03 -0800 Reply-To: gmargo@newton.vip.best.com In article , H. Peter Anvin wrote: >Followup to: <49iegd$g2h@kirin.wwa.com> >By author: kermit@wwa.com (Matthew D. Isleb) >In newsgroup: comp.os.linux.development.system >> >> Has anyone heard anything about stadarizing printing for linux? >> I have heard that SCO has some kind of standarization of printing. >> > >Check out magicfilter: > > sunsite.unc.edu:/pub/Linux/system/Printing/magicfilter-1.1b.tar.gz > > /hpa Ditto this. I use magicfilter for my non-postscript HP laserjet, and merrily print ascii, postscript, and .dvi formats transparently. Completely configurable, so it's easy to support different formats or different rendering tools. gm -- +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Gregory H. Margo No free speech restrictions on the Internet! gmargo@newton.vip.best.com ms windows makes me ill newton@best.com ------------------------------ From: s923298@minyos.xx.rmit.EDU.AU (Hong Niu) Subject: compiling problems Date: 4 Dec 1995 12:23:20 GMT ------------------------------ From: jan@swisun31 (Jan Wielemaker) Subject: ELF linking bug? Date: 4 Dec 1995 10:11:08 GMT Hi, I stubled on the following bug. I tried to redefine __morecore() from GNU malloc (living in libc on Linux) to get a malloc returning addresses where I want them. If you compile and link the code below using gcc -DTEST morecore.c, and run the result, you get a nasty crash in the dynamic linker. Using -static, it compiles and runs fine. Regards --- Jan --morecore.c-------------------------------------------------------------- /* - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Let GNU malloc() work from an arbitrary base! - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - */ #include #include #include #include #include #include extern void * (*__morecore)(long size); static void *base; /* base */ static void *top; /* external top */ static void *atop; /* allocated top */ static int pagsize; #ifndef roundup #define roundup(p, n) (((p)+(n)-1 / (n)) * (n)) #endif #if !defined(MAP_ANON) && defined(MAP_ANONYMOUS) #define MAP_ANON MAP_ANONYMOUS #endif #define STACK_MAP_TYPE MAP_ANON|MAP_PRIVATE|MAP_FIXED void * mycore(long size) /* similar to sbrk() */ { void *mem = top; if ( top + size >= atop ) { void *ntop = (void *)roundup((ulong) top + size, pagsize); if ( mmap(atop, ntop - atop, PROT_READ|PROT_WRITE, STACK_MAP_TYPE, -1, 0L) != atop ) { fprintf(stderr, "mycore(): %s\n", strerror(errno)); return NULL; } atop = ntop; } top = top + size; return mem; } void start_memory(void * address) { pagsize = getpagesize(); base = (void *)roundup((ulong) address, pagsize); top = base; atop = base; __morecore = mycore; } #ifdef TEST #ifndef HAVE_STRTOUL ulong strtoul(const char *s, char **eptr, int base) { static unsigned char xmap[256]; static int xmap_initialised = 0; unsigned long val; if ( !xmap_initialised ) { int n; xmap_initialised++; for(n=0; n<256; n++) xmap[n] = 0xff; for(n='0'; n<='9'; n++) xmap[n] = n - '0'; for(n='a'; n<='z'; n++) xmap[n] = n - 'a' + 10; for(n='A'; n<='Z'; n++) xmap[n] = n - 'A' + 10; } if ( s[0] == '0' && (s[1] == 'x' || s[1] == 'X') ) s += 2; for(val = 0; *s; s++) { unsigned long dval = xmap[*s]; if ( dval < (unsigned) base ) val = val * base + dval; else { if ( eptr ) *eptr = (char *)s; } } return val; } #endif main(int argc, char **argv) { static int testvals[] = { 1, 1000, 10000, 100000, -1 }; int *tv = testvals; ulong base = 0x40000000L; if ( argc == 2 ) base = strtoul(argv[1], NULL, 16); start_memory((void *)base); for(; *tv != -1; tv++) { char *rval = malloc(*tv); printf("malloc(%-6d) --> 0x%08x\n", *tv, rval); memset(rval, 0xbf, *tv); } return 0; } #endif -- Jan Wielemaker ------------------------------ From: grilly@edwood.emi.u-bordeaux.fr (Nicolas Grilly) Subject: Re: Taligent is disbanded - could Linux pick up the pieces? Date: 02 Dec 1995 23:25:56 GMT > <> supposedly has a story, but I am not > registered ;-) You can register in a page of this web site. It's free ! > If this is truly the case, any adovates out there want to lobby code > donations to Linux? Under the terms of the arrangement, Apple and HP have rights to use Taligent's technology, but IBM beholds control. So, I think it will be difficult to convinve IBM to release Taligent's code in public domain... But we can try ! For now, it's a great victory for Microsoft, but it's not a great victory for free software... May the force be with us ! -- // Nicolas GRILLY // email: grilly@emi.u-bordeaux.fr // // 24 rue Bossuet // // // 33000 Bordeaux // // // France // // ------------------------------ From: ig25@fg70.rz.uni-karlsruhe.de (Thomas Koenig) Subject: Re: How do you compile an old module? Date: 4 Dec 1995 13:26:57 +0100 Reply-To: Thomas.Koenig@ciw.uni-karlsruhe.de In comp.os.linux.development.system, Thomas.Koenig@ciw.uni-karlsruhe.de wrote: >What's changed in module handling? I'd like to run the iBCS module under >1.3.45, but I'm stuck with an older version until I can figure out >what to change... If you got the rather largish patch I posted before I'd canceled it, please don't use it :-) A simpler solution is to - change all current->sigaction to current->sig->action (which is required since 1.3.something anyway) - compile your kernel with -DCONFIG_MODVERSIONS - add the flags -DCONFIG_MODVERSIONS -D__NO_VERSION__ to the Makefile Works for me, at least. -- Thomas Kvnig, Thomas.Koenig@ciw.uni-karlsruhe.de, ig25@dkauni2.bitnet. The joy of engineering is to find a straight line on a double logarithmic diagram. ------------------------------ From: eyal@ise.canberra.edu.au (Eyal Lebedinsky) Subject: Re: Q: joystick-0.7.3 driver, thrustmaster programming specs Date: 4 Dec 1995 12:44:46 GMT In article <49h5o7$rem@josie.abo.fi>, mandtbac@news.abo.fi (Mats Andtbacka) writes: |> Eyal Lebedinsky, in <49dgbd$8s5@supreme.pcug.org.au>: |> [...] |> >The CHpro hat is very different, it returns the hat position in |> >the four-button bits. This is why you cannot use multiple |> >button presses on this stick (a terrible disadvantage in my opinion). |> |> Just out of curiosity, which buttons? It would need three, of course, |> to represent the five possible positions of the thing; did they at |> least have the sense to leave button #1 out of the chording, to let |> you combine it with any one of the other ones? Not so lucky... here is a file I have put together a long while back: CH Pro buttons ============== The button presses generate the following codes: (these are the four bits at io port hex 201, high nibble): Hat left 0011 Hat down 0111 Hat right 1011 Hat top 1111 Trigger 0001 left 0010 Right 0100 Low 1000 Multiple presses are not recognized, the button which is earlier in the above list will override a button that follows it (you cannot have the hat in more than one position at a time). -- Eyal Lebedinsky (eyal@ise.canberra.edu.au, eyal@pcug.org.au) ------------------------------ From: Gloger@lrz.uni-muenchen.de (Wolfram Gloger) Subject: Re: ELF linking bug? Date: 4 Dec 1995 13:56:02 GMT Reply-To: Gloger@lrz.uni-muenchen.de jan@swisun31 (Jan Wielemaker) writes: >If you compile and link the code below using gcc -DTEST morecore.c, and >run the result, you get a nasty crash in the dynamic linker. Using >-static, it compiles and runs fine. I can't reproduce the problem, I get % ldd a.out libc.so.5 => /lib/libc.so.5.2.13 % ./a.out malloc(1 ) --> 0x40003000 malloc(1000 ) --> 0x40004c00 malloc(10000 ) --> 0x40005000 malloc(100000) --> 0x40008000 This is with binutils-2.5.2l.17 and ld.so-1.7.10. Regards, Wolfram. ------------------------------ From: e9025800@stud1.tuwien.ac.at (Stefan Froehlich) Subject: libg++.so.27? Date: 4 Dec 1995 08:54:12 GMT Hello all, does anyone know where I can find libg++.so.27 and libstdc++.so.27? I have got a binary distribution of octave-1.1.1 and it complains about not finding those libraries. On sunsite and some other servers I have only found *.so.26. Any hints are appreciated. Bye, Stefan ------------------------------ From: thomas@cuivre.fdn.fr (Thomas Quinot) Subject: Re: Where the offical place for libc? Date: 3 Dec 1995 17:05:34 GMT Joseph Vigneau (joev@pyramid.res.wpi.edu) icrit : > Now, what exactly is this? The GNU glibc is up to version 1.0.9, and libg++ > is up to 2.7.1... Are these just specialized versions for Linux? GNU libc != Linux libc. Some code is shared, perhaps both libraries will merge in the future, but not yet. Latest official Linux libc is 5.2.16, now available at tsx-11. -- Grand.Bwana@cuivre.fdn.fr | Linux : the choice of a GNU generation ------------------------------ From: louis@lcjones.aclib.siue.edu (Louis J. LaBash Jr.) Subject: Re: Where the offical place for libc? Date: 4 Dec 1995 17:20:22 GMT On 3 Dec 1995 01:08:15 GMT, Joseph Vigneau wrote: |In article <49prlm$4vm@news.ox.ac.uk>, |Daniel Barlow wrote: |> |>5.0.9 _is_ the most recent update. 5.2.x is beta code for testers to |>test. | |Now, what exactly is this? The GNU glibc is up to version 1.0.9, and libg++ |is up to 2.7.1... Are these just specialized versions for Linux? The short answer is yes, there are *differences*. -- Louis-ljl-{LLaBash@eniac.ac.siue.edu | lou@minuet.siue.edu} ------------------------------ From: rjs@spider.compart.fi (Riku Saikkonen) Crossposted-To: comp.os.linux.development.apps Subject: Re: can't compile with inb(), outb(); linking errors Date: Fri, 1 Dec 1995 17:35:47 GMT Randall Jones (rjones@newshost.aoc.nrao.edu) wrote: > I'm trying to write a *very* basic hardware interface to a card I have. >I'm trying to use the outb() and inb() functions under C. See ftp://sunsite.unc.edu/pub/Linux/docs/HOWTO/mini/IO-Port-Programming or http://isil.lloke.dna.fi/rjs/tech-writing/ioportprog.html for my I/O port programming mini-HOWTO. -- -=- Rjs -=- rjs@spider.compart.fi - http://isil.lloke.dna.fi/rjs/ "And with great power comes great responsibility. Use it wisely." - AD&D 2nd edition Dungeon Master's Guide ------------------------------ From: mec@treflan.shout.net (Michael Elizabeth Chastain) Subject: Re: How do I call my own system calls.. Date: 30 Nov 1995 23:26:24 GMT Reply-To: mec@duracef.shout.net In article <49l662$kun@gopher.cs.uofs.edu>, Thomas W. Janofsky wrote: > 4. My question now is how do I call sys_whatever from a user program -- > khg was a little unclear. Here's an example program. #include #define __NR_sys_200 200 #define __NR_sys_300 300 #define __NR_m1 -1 _syscall1( int, afs_syscall, int, i ); _syscall1( int, sys_200, int, i ); _syscall1( int, sys_300, int, i ); _syscall1( int, m1, int, i ); int main( ) { afs_syscall( 0 ); sys_200( 200 ); sys_300( 300 ); m1( -1 ); return 0; } Michael Chastain mec@duracef.shout.net Check out my pioneer trace-and-replay debugger: ftp://tsx-11.mit.edu/pub/linux/sources/usr.bin/mec-0.3.tar.gz ------------------------------ From: poipoi@hwi.resi.insa-lyon.fr (Jerome Etienne) Subject: pb with 1.2.13 Date: 4 Dec 1995 18:37:06 GMT i got a application compile with -g option which produces this output. Oops: 0002 EIP: 0010:00142570 EFLAGS: 00010046 eax: 00000000 ebx: 00275018 ecx: ffffffff edx: 00000000 esi: 009f0000 edi: 0000000a ebp: 00000300 esp: 00288f68 ds: 0018 es: 0018 fs: 002b gs: 002b ss: 0018 Process a.out (pid: 193, process nr: 25, stackpage=00288000) Stack: 00275018 009fb3f4 00142f96 00000300 00275018 0019cfe4 009fb3f4 00000300 001316dd 009fb3f4 00000300 bffff378 50000000 50001df0 bffff3b4 0013250c 00000002 0000000a 00000300 00a79000 001102f9 00000001 bffff378 00000001 Call Trace: 00142f96 001316dd 0013250c 001102f9 0011002b Code: 89 5a 48 fb 5b 5e c3 90 90 90 90 90 90 90 90 90 56 53 8b 4c My question is really simple, how can i use this to fix the bug ? :))) the bug comes from my application or from the kernel? i assume hardly the bug comes from the kernel... so which kernel has a fix for this problem? information about my configuration: - i use kernel1.2.13+elf. and a slackware 3.0 distribution. - uname -a -> Linux hwi 1.2.13 #3 Fri Nov 24 15:38:52 MET 1995 i486. - i can have a list of symbols for my kernel and my application. my application uses socket. 001422b0 T ip_dev_check 00142320 t gcc2_compiled. 00142320 t sk_inuse 00142360 T get_new_socknum 00142470 T put_sock <--- seems to be here. 00142580 t remove_sock 00142620 T destroy_sock 00142890 t inet_fcntl 00142910 t inet_setsockopt 00142980 t inet_getsockopt -- Jerome Etienne. computer science student (4th year). email: poipoi@hwi.resi.insa-lyon.fr quote: Fear is the mind-killer. Informations want to be free. ------------------------------ From: fox@graphics.cs.nyu.edu (David Fox) Subject: Re: Standardized Printing for Linux Date: 03 Dec 1995 08:44:53 -0500 In article <49qvj5$fas@muirwood.convex.com> ldaffner@convex.com (Larry Daffner) writes: ] Putting the drivers in the kernel won't help get them developed any ] faster. Indeed, it reduces your list of potential authors from the entire free software community to just the Linux kernel hackers. Please, lets not re-invent Unix badly. Someone else mentioned they would like to see a new page description language that took less compute power than Postscript. This is another heinous suggestion. Ghostscript is already faster than my printer, and likely to remain so. Lets not re-invent Postscript badly either. -- David Fox http://found.cs.nyu.edu/fox xoF divaD NYU Media Research Lab baL hcraeseR aideM UYN ------------------------------ From: conroy@cs.utexas.edu (Conor John Cunningham) Subject: Re: Problems with 1.2.13? Date: 4 Dec 1995 09:02:48 -0600 In article <48qvff$9e4@news.duke.edu>, Richard Lee Kern wrote: >is anyone aware of problems with 1.2.13 and the adaptec 1542 SCSI card. >Our kernel had been working well on it's partition then on the dos >partition we installed windows 95 and when we later returned to try >to boot linux the kernel booted (with floppy LILO) and went through >part of the initialization and when it got to the hard disk part it >said that it couldn't determine DMA priority on the Adaptec card >so it disabled the SCSI card-->no hard disk availability ---> >kernel panic unable to mount root! Hmm. never happened to me. I have win95, an Adaptec 1542CF, and I _think_ I have 1.2.13, (though I haven't upgraded that machine in awhile, so perhaps I put an early 1.3 kernel on there). I _have_ been getting a weird error where it prints groups of 10 repeating lines (I don't have the exact message), about how there is weirdness from some bit being x, y, and clean - I usually reboot and it goes away, but if that comes up with you, please let me know Conor -- Conor Cunningham | CS Senior | UT ACM President,Spr/Fall 1995 |FIGHT THE MAN| University of Texas at Austin ---< Hey Bill: >---------| USE PGP |------------ "You can have my private key when you pry it from my cold, dead hands." ------------------------------ From: seano@teleport.com (Sean O'Halloran) Subject: Re: Latest Stable Kernel ??? Date: 4 Dec 1995 11:10:05 -0800 In gknitz@netcom.com (Gary K Nitzberg) writes: >I want to undate my kernel. >What is the latest stable kernel you would recommend? >Thanks. > Gary. ... and life was good ... I had a similar question, and so I have spent the weekend investigating all the kernel changes from 1.2.1 onward ... Thanks too to those who have responded to my later posting see -"Quick kernel vnum recommendation" .. Anyway, after looking at all the changes, and reading all the posts in all the linux associated groups, and mail sent to me, this is what I have found .. Changes slowed as 1.3.20 approached, and the changes I saw going in appeared to be consolidations and a few "tweaking" fixes. At 1.3.22 (I think) support started being added for multi-threading in the kernel - this did not complete till late in 1.3.2X ?? The same sort of consolidations appears to occur as 1.3.45 is approached - so I would ??guess?? that 1.3.45 should be a pretty stable release as well .. I have had three mail replys; two say that 1.3.45 works fine for them; and one with a mix of versions at 1.2.X, but stating that they use a 1.3.20 release in a critical server, and it looks to be a very stable 1.3.X.. After considering everything, fixes, posting to all the groups about problems with various releases, looking at changes and Linus's comments, I decided to go with 1.3.21 - there were a couple of fixes to in 1.3.21 that looked important to me vs 1.3.20, notably shared memory stuff .. I considered 1.3.45, but I thought it might be a little early in its' release, although the posts do suggest that it is indeed "stable". None of these releases are going to be without problems, and possibly some bugs may show up in the operation of some of the binaries not associated with kernel code - .. but hey, that's life on the edge eh? So I am running 1.3.21 now, and performance is better than 1.2.1 - and HeY !!!!! My Lightning fast, 32 bit PCI ethernet card appears to work just great !!!! Thanks again - Resp, Seano p.s - if you want to look at some rough notes on what got changed between releases, point your browser at .. http://www.crynwr.com:80/kchanges most helpful. -- Sean O'Halloran Artel Systems Inc. EMAIL: seano@teleport.com Send compliments to me: seano@teleport.com Send complaints to: president@whitehouse.gov ------------------------------ ** 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-System-Request@NEWS-DIGESTS.MIT.EDU You can send mail to the entire list (and comp.os.linux.development.system) via: Internet: Linux-Development-System@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-System Digest ******************************