From: Digestifier To: Linux-Development@senator-bedfellow.mit.edu Reply-To: Linux-Development@senator-bedfellow.mit.edu Date: Tue, 18 Oct 94 03:13:14 EDT Subject: Linux-Development Digest #328 Linux-Development Digest #328, Volume #2 Tue, 18 Oct 94 03:13:14 EDT Contents: Re: ext2fs vs. Berkeley FFS (Bryan Sparks) Slackware boot disk with kernel 1.1.52(or up) and ncr 53c8xx driver (Jin Yang) programming each pixel of a vga/ega screen in Linux (Pritish Shah) Re: A badly missed feature in gcc (Orc) Re: 8-bit ASCII for comm. (A.E. Brouwer) Re: A badly missed feature in gcc (Grant Edwards) Re: We a FAQ: Linux vs. *BSD!!! (Grant Edwards) Lies, Damned Lies, and Benchmarks (take two) (Jeff Kuehn) Re: ext2fs vs. Berkeley FFS (Joerg Mertin) Re: Question about ext2fs (Joerg Mertin) wxWindows 1.50k shared lib wanted (Tall Sword) ---------------------------------------------------------------------------- From: bryan@infonaut.com (Bryan Sparks) Subject: Re: ext2fs vs. Berkeley FFS Date: 17 Oct 1994 15:14:33 -0600 Matthew Dillon (dillon@apollo.west.oic.com) wrote: : :In article <199410141322.OAA05432@gblinux.demon.co.uk> pdcawley@ftech.co.uk (Piers Cawley) writes: : :> : :>In article : : : : Don't be influenced by the over-featurism that NT offers. There's no need : : for this crud. : : : :Actually, my model for this sort of thing is the Mac filesystem. I wouldn't : :call it overfeatured, but it does a lot of stuff very well, and very easily, : :that is a royal PITA to accomplish with Unix. : Frankly, I see nothing here that can't be EASILY implemented with : directories and file extensions. IMHO, Mac resource forks were the : second biggest mistake Apple made after the operating system. : It sounds to me that you want a database, not a filesystem. : -Matt : -- : Matthew Dillon dillon@apollo.west.oic.com : 1005 Apollo Way ham: KC6LVW (no mail drop) : Incline Village, NV. 89451 Obvious Implementations Corporation : USA Sandel-Avery Engineering : [always include a portion of the original email in any response!] Why not Bento (part of OpenDoc)? It seems that is offers many advantages and it is available and it is a standard (debatable, I guess). Another thing, it can all be implemented in user space. Bryan --- Bryan Sparks bryan@infonaut.com, bryan@kachina.orem.ut.us -- --- Bryan Sparks Phone: 801-226-6056 (home) E-Mail: bryan@infonaut.com, bryan@kachina.orem.ut.us ------------------------------ From: jinyang@cs.utexas.edu (Jin Yang) Crossposted-To: comp.os.linux.help Subject: Slackware boot disk with kernel 1.1.52(or up) and ncr 53c8xx driver Date: 17 Oct 1994 16:46:00 -0500 I would like to install linux on my Intel Premier/PCI pentinum computer with a PCI SCSI adapter and a SCSI hard drive. I have downloaded the slackware distribution from tsx-11.mit.edu. However, the bootdisk disk (ncr.gz, kernel 1.1.19) does not recognize the PCI SCSI adapter (ncr 53c8xx). Some very kind people on the net told me that kernel 1.1.52 and up will be able to recognize the adapter. Could some one on the net tell me where I can get a slackware bootdisk with kernel 1.1.52 (or up) and ncr 53c8xx driver? In fact, I already got a 1.1.56 kernel image file (zImage) from a friend but I don't know how to make a bootdisk for slackware. I have another DX33 computer running an old version of linux (kernel 0.98). I tried to follow the instruction in SLAKWARE.FAQ to make a boot disk myself but failed. The instruction in the faq is as follows: ... well you grab an existing one like the "bare.gz", put it on a floppy, and do this: mount /dev/fd0 /mnt (this mounts it) cat zImage > /mnt/vmlinuz (put the new kernel in place) rdev -R /mnt/vmlinuz 0 (mount read-write) rdev /mnt/vmlinuz /dev/fd0H1440 (use that drive, or /dev/fd0h1200) rdev -r /mnt/vmlinuz 1440 (set the ramdisk size, or use 1200 with 1.2 meg) lilo -r /mnt (reinstall lilo) umount /mnt (that's it! you're done! :^) First, the rdev that I have does not support -R option. Second, when I typed lilo -r /mnt, I got the following error message: /etc/lilo/config: no such file or directory Could someone tell me what is the correct way to make a boot disk from an image file? Thanks in advance, - Jin Yang jinyang@cs.utexas.edu ------------------------------ From: pritish@nic.cic.net (Pritish Shah) Crossposted-To: comp.os.linux.help,comp.os.linux.misc Subject: programming each pixel of a vga/ega screen in Linux Date: 17 Oct 1994 22:22:08 GMT Hello all, I am trying to write a display system on Linux that will display different messages on the vga/ega screen using different fonts and different attributes. I was informed that there might be something called getpixel/putpixel function that could work for the screen. Any help would be appreciated. This is for my project at school. Pritish -- ********************************************************************** Pritish M Shah * Contact: Phone 1-313-998-6700 #include * Fax 1-313-998-6105 http://www.cic.net/~pritish/ * email pritish@cic.net ------------------------------ From: orc@pell.com (Orc) Subject: Re: A badly missed feature in gcc Date: Thu, 13 Oct 1994 03:19:44 GMT In article , H. Peter Anvin wrote: > /* ... */ is supported as backward >compatibility, but you will be hard-pressed to find a C++ programmer >who uses them more than once in a blue moon. I prefer to use them when writing C++ code. // makes my code look like assembly language. ____ david parsons \bi/ followups set out of c.o.l.d \/ ------------------------------ From: aeb@wsdw01.win.tue.nl (A.E. Brouwer) Subject: Re: 8-bit ASCII for comm. Date: 17 Oct 1994 17:21:04 +0100 gavrie@pesach.jct.ac.il (Prof. Gavrie Philipson) writes: : Is there any way to display 8-bit ASCII on the Linux console? : When I call PC-based BBSes from Linux, I just get rubbish : for >127 characters. Well, first of all, there is no such thing as 8-bit ASCII, and if there were, then no doubt it would not be what you need. You want IBM code page 437 or something similar, I think. Try to give the command echo -e '\033(U' before calling that BBS. ------------------------------ From: grante@reddwarf.rosemount.com (Grant Edwards) Subject: Re: A badly missed feature in gcc Date: Sun, 16 Oct 1994 01:12:48 GMT H. Peter Anvin (hpa@ahab.eecs.nwu.edu) wrote: : In newsgroup: comp.os.linux.development : : > This is doubtful. The problem is that making this a 'feature' of ANSI c : > will all of a sudden make previously syntacticly correct code now fail to : > compile or, worse, compile with a different symantic meaning. This woudl : > be DISASTEROUS to the attempt to standardize C. : It is *already* put into quite a number of compilers (making it a : portability issue already), and anyone writing code that uses that : type of constructs should be shot anyway. True, though I suspect that anybody writing code like that has enough problems that a bullet wound would go unnoticed. That aside, it is fine and proper for you and I to have opinions on what is good style and what isn't. I actually enjoy discussing the pro's and con's of various "style" questions with other people (like whether it's a good idea to have code/data and a comment on the same line). However, it is _not_ acceptable for a compiler to have an "opinion" about style. Given sytactically valid source code, it shall parse it and generate correct machine code -- no matter how ugly, stupid, counter-productive, or obfuscated that source code might be. : It is true there are such cases, but they are throuroughly and : completely artificial. If construct X is legal then rest assured that somebody, somewhere, has written a real program that relies on that fact. If you've ever tried to remove a "useless" feature from a product, you've found this out. Me: "But Mr. Customer, the fact that status bit XYZ did that was a _bug_ as you should be able to tell by the documentation, and now that bug has been fixed." Customer: "I don't care. It used to do that, we depended on it, now it doesn't. Fix it back." -- Grant Edwards |Yow! You can't hurt me!! I Rosemount Inc. |have an ASSUMABLE MORTGAGE!! | grante@rosemount.com | ------------------------------ Crossposted-To: comp.os.386bsd.development,comp.sys.unix From: grante@reddwarf.rosemount.com (Grant Edwards) Subject: Re: We a FAQ: Linux vs. *BSD!!! Date: Sun, 16 Oct 1994 01:21:47 GMT Jesus Monroy Jr (jmonroy@netcom.com) wrote: : Can we get together and write a single FAQ on this? Probably not... : I propose a single FAQ to answer the question: : : Which is better Linux or *BSD? You'd be better off with the question: "What are the similarities/differences between Linux and *BSD?" : Do I get a Yeah on this? I won't say Nay, but I suspect that it would be impossible to keep it current enough to be useful. -- Grant Edwards |Yow! I hope I bought the Rosemount Inc. |right relish... |zzzzzzzzz... grante@rosemount.com | ------------------------------ From: kuehn@citadel.scd.ucar.edu (Jeff Kuehn) Subject: Lies, Damned Lies, and Benchmarks (take two) Date: 17 Oct 1994 21:34:40 GMT Well, as promised, here's the corrected results for the Byte Magazine UNIX benchmark comparing Linux versions 1.0.4 through 1.1.54. As before, I haven't run every patchlevel (only 21 of them :-) but I'm willing to run individual patchlevels on request, and can provide detailed information on each benchmark result for any patchlevel which has been run. (see the data for the 1.1.54 run after the table) Please don't send me code which hasn't been thoroughly tested... each benchmark requires a kernel build, a reboot and an hour to run. I you have code which has been carefully tested, I'm willing to benchmark it. Please, send diffs against PL52 (the most recent version for which Linus has posted a full source tar). xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx The headers in the table below refer to the various tests in the Byte Magazine UNIX Benchmark as follows. Versn linux kernel version Ar Arithmetic Test (type = double) Dr Dhrystone 2 w/o register vars Exc Execl Throughput Test FC1 File Copy (10 seconds) FC3 File Copy (30 seconds) FR1 File Read (10 seconds) FR3 File Read (30 seconds) FW1 File Write (10 seconds) FW3 File Write (30 seconds) Pip Pipe-based Context Switch Test PCT Process Creation Test Sh1 Shell scripts (1 concurrent) Sh2 Shell scripts (2 concurrent) Sh4 Shell scripts (4 concurrent) Sh8 Shell scripts (8 concurrent) Sys System Call Overhead Test PTT Pipe Throughput Test. SUM sum of values in row The numbers in the table, represent the ratio of test(pl)/test(1.1.0), Bigger numbers indicate better performance. Also note that a pipe throughput test has been added. xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx Hardware: iNTEL 66MHz 486DX/2 AMI Enterprise III EISA/VLB Motherboard (AMI BIOS/Chipset) 256K 4-way set associative writeback cache 16MB 70ns RAM AHA-1742 EISA SCSI host adaptor Toshiba MK538 1.2GB SCSI drive Versn |Ar|Dr|Exc|FC1|FC3|FR1|FR3|FW1|FW3|PTT|Pip|PCT|Sh1|Sh2|Sh4|Sh8|Sys|SUM 1.0.04|1.|1.|1. |1. | .9|1. |1. | .9| .9|1. | .9|1. |1. |1. |1. |1. |1. |16.7 1.0.09|1.|1.|1. |1. | .9|1. |1. |1. |1. |1. | .9|1. |1. |1. |1. |1. |1. |16.8 1.1.00|1.|1.|1. |1. |1. |1. |1. |1. |1. |1. |1. |1. |1. |1. |1. |1. |1. |17.0 1.1.01|1.|1.|1. |1. |1. |1. |1. | .9|1. |1. |1. |1. |1. |1. |1. |1. |1. |16.9 1.1.02|1.|1.|1. |1. | .9|1. |1. |1. |1. |1. |1. |1. |1. |1. |1. |1. |1. |16.9 1.1.03|1.|1.| .9| .9|1. |1. |1. | .7| .9|1. | .9|1. |1. |1. |1. |1. |1. |16.1 1.1.04|1.|1.| .9| .9|1. |1. |1. | .7| .9|1. | .9|1. |1. |1. |1. |1. |1. |16.1 1.1.05|1.|1.| .9| .9|1. |1. |1. | .7| .9|1. |1. |1. |1. |1. |1. |1. |1. |16.2 1.1.10|1.|1.| .9|1.3|1.4|1. |1. | .7| .9|1. |1. |1. |1. |1. |1. |1. |1. |17.0 1.1.15|1.|1.| .9|1.3|1.5|1. |1. | .7| .9|1. |1. |1. | .9| .9| .9|1. |1. |17.0 1.1.20|1.|1.| .9|1.2|1.4| .9| .8| .6| .8| .9| .9|1. | .9| .9| .9|1. |1. |16.3 1.1.25|1.|1.| .9|1.2|1.4| .9| .9| .7| .9| .9| .9|1. |1. |1. | .9|1. |1. |16.6 1.1.30|1.|1.| .9|1.2|1.5| .9| .9| .7| .9| .9| .9|1. |1. |1. | .9|1. |1. |16.6 1.1.35|1.|1.| .9|1.2|1.4| .9| .9| .6| .9| .9|1. |1. |1. |1. |1. |1. |1. |16.5 1.1.40|1.|1.|1. |1.2|1.3| .8| .8| .6| .8| .9| .6|1. |1. |1. |1. |1. | .9|16.0 1.1.45|1.|1.|1. |1.2|1.3| .9| .9| .6| .9| .9| .9|1. |1. |1. |1. |1. |1. |16.7 1.1.50|1.|1.|1. |1.2|1.3| .8| .8| .6| .9| .9| .8| .9|1. |1. |1. |1. |1. |16.4 1.1.51|1.|1.|1. |1.2|1.3| .8| .8| .6| .9| .9| .9|1. |1. |1. |1. |1. | .9|16.5 1.1.52|1.|1.|1. |1.3|1.4| .8| .8| .9| .9| .9| .9|1. |1. |1. |1. |1. |1. |16.8 1.1.53|1.|1.|1. |1.1|1.3| .8| .9| .6| .9| .8| .8|1. |1. |1. |1. |1. |1. |16.3 1.1.54|1.|1.|1. |1.1|1.3| .8| .8| .6| .9| .9| .9|1. |1. |1. |1. |1. | .9|16.2 SunOS 4.1.3 |.5|.4| .1| .9|1.0| .1| .1| .1| .1| .1| .2| .4| .2| .2| .2| .2| .2| 5.1 (SPARCstation IPC, 24MB RAM, 200MB SCSI drive) xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx Here's the data for Linux 1.1.54 vs. the baseline 1.1.0. (note: lps==loops/second and lpm==loops/minute). BYTE UNIX Benchmarks (Version 3.11) System -- Linux cyberdeck 1.1.54 #5 Sun Oct 16 17:10:43 MDT 1994 i486 Start Benchmark Run: Sun Oct 16 17:31:47 MDT 1994 1 interactive users. Arithmetic Test (type = double) 5069.1 lps (10 secs, 6 samples) Dhrystone 2 without register variables 51635.1 lps (10 secs, 6 samples) File Read (10 seconds) 94368.0 KBps (10 secs, 6 samples) File Write (10 seconds) 7871.0 KBps (10 secs, 6 samples) File Copy (10 seconds) 1312.0 KBps (10 secs, 6 samples) File Read (30 seconds) 91781.0 KBps (30 secs, 6 samples) File Write (30 seconds) 11640.0 KBps (30 secs, 6 samples) File Copy (30 seconds) 1224.0 KBps (30 secs, 6 samples) Pipe-based Context Switching Test 6796.6 lps (10 secs, 6 samples) Pipe Throughput Test 18999.9 lps (10 secs, 6 samples) Execl Throughput Test 63.2 lps (9 secs, 6 samples) Process Creation Test 109.0 lps (10 secs, 6 samples) System Call Overhead Test 29827.4 lps (10 secs, 6 samples) Shell scripts (1 concurrent) 98.1 lpm (60 secs, 3 samples) Shell scripts (2 concurrent) 51.3 lpm (60 secs, 3 samples) Shell scripts (4 concurrent) 26.0 lpm (60 secs, 3 samples) Shell scripts (8 concurrent) 13.0 lpm (60 secs, 3 samples) INDEX VALUES TEST BASELINE RESULT INDEX Arithmetic Test (type = double) 5062.4 5069.1 1.0 Dhrystone 2 without register variables 51685.8 51635.1 1.0 Execl Throughput Test 64.1 63.2 1.0 File Copy (10 seconds) 1172.0 1312.0 1.1 File Copy (30 seconds) 951.0 1224.0 1.3 File Read (10 seconds) 117544.0 94368.0 0.8 File Read (30 seconds) 119451.0 91781.0 0.8 File Write (10 seconds) 12389.0 7871.0 0.6 File Write (30 seconds) 13393.0 11640.0 0.9 Pipe Throughput Test 21779.3 18999.9 0.9 Pipe-based Context Switching Test 7620.8 6796.6 0.9 Process Creation Test 112.2 109.0 1.0 Shell scripts (1 concurrent) 97.1 98.1 1.0 Shell scripts (2 concurrent) 51.0 51.3 1.0 Shell scripts (4 concurrent) 26.0 26.0 1.0 Shell scripts (8 concurrent) 12.6 13.0 1.0 System Call Overhead Test 31504.0 29827.4 0.9 ========= SUM of 17 items 16.2 AVERAGE 1.0 xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx Observations: 0) My Linux (home) system soundly stomps the stuffings out of my SunOS (work) system in terms of performance... by about a factor of three. 1) The scheduler, while not perfect, is better than first estimated due to the awk(1)/time(1) problems in the first benchmark. 2) The I/O performance is pretty touch-and-go with a lot of variation in the numbers. IMHO, since the benchmark ran almost entirely in memory, the problem lies above the device drivers, probably in either the buffer cache code or filesystem code. It has already been suggested (by Eric) that the problem could lie in the memory allocation which is heavily used by the buffer cache. Were this the case, I think I would expect to see the File Read tests showing more problems than the File Write tests. The opposite was true -- File Write did more poorly than File Read. I'd like to hear more discussion on this. File Write dropped off at 1.1.3 and File Read dropped off between 1.1.15 and 1.1.20 (remember these numbers). 3) Somewhere between 1.1.15 and 1.1.35 the File Copy test did the best and since dropped off a bit. Again, most of this happened in cache on my machine, so IMHO this doesn't represent device driver performance, but rather reflects the performance of filesystem/buffer cache and associated algorithms. 4) Pipes seem to have become more expensive to use (Pipe-based Context Switching Test and Pipe Throughput Test) since somewhere between 1.1.15 and 1.1.20. Could this be another symptom of whatever happened to File Read between 1.1.15 and 1.1.20? 5) System call overhead is up a little as well since somewhere between 1.1.35 and 1.1.40. Jeff Kuehn Scientific Computing Division National Center For Atmospheric Research Boulder, Colorado USA ------------------------------ From: smurphy@stardust.bln.sub.org (Joerg Mertin) Subject: Re: ext2fs vs. Berkeley FFS Date: 16 Oct 1994 10:09:38 GMT Reply-To: smurphy@stardust.bln.sub.org In comp.os.linux.development Theodore Ts'o (tytso@MIT.EDU) wrote: : From: mike@majestix.cs.uoregon.edu (Mike Haertel) : Date: 03 Oct 1994 20:46:49 GMT [...] Stuff deleted : In any case, it's a good idea to optimize for the common case, not the : exception case. Thus, I believe ext2fs made a pretty good tradeoff. : Not doing syncronous meta-data writes gives the ext2fs a blinding speed : advantage over the FFS. From my experience with both FFS and ext2fs, : ext2fs is just as reliable --- if not more --- than FFS in surviving : unclean shutdowns. That's right. I would say, under Xwindows I have better chances of saving unclosed files than under MSWindog while editing a file and having a system-Crash (Which happens once all 2 houres). In fact, I never loosed any file with ext2fs since I use it. And my first release was 0.2 :-) Just Good. Thanks to all developers of this fs ! Just my 2 cents :^) -- Solong & Happy Hacking ************************************************************************ * Joerg Mertin : smurphy@stardust.bln.sub.org (Home) * * in Berlin Spandau at : jorgagif@w203zrz.zrz.tu-berlin.de * * Stardust's Linux System : Data, Fax & Voice 49 30 3615569 * ************************************************************************ `Fatal Error: Found [MS-Windows] System -> Terminating Virus...' ------------------------------ From: smurphy@stardust.bln.sub.org (Joerg Mertin) Crossposted-To: comp.os.linux.help Subject: Re: Question about ext2fs Date: 16 Oct 1994 18:06:37 GMT Reply-To: smurphy@stardust.bln.sub.org In article <1994Oct11.150318.28373@news.cs.indiana.edu>, "Eric Jeschke" writes: : |> Can someone with some knowledge of the ext2fs internals tell me : |> how intensive the disk activity is to certain blocks like the : |> superblock and inode bitmap blocks? extfs is working fine, but : |> I'm afraid I have a poor quality disk (Seagate) that is not able : |> to handle the intense repeated disk activity to these blocks. : |> I am slowly developing bad blocks on various inode bitmap blocks : |> (and now the superblock). The kernel complains about getting a : |> "short read" on those blocks. I am able to map them out successfully : |> using the -L option with ext2fs and the system recovers admirably, : |> but sooner or later the problem recurs with another block. The : |> latest victim is the superblock on the root partition. I am able : |> to e2fsck the partition using the -b 8193 option and averything seems : |> fine, however the system fails to mount it at bootup even though : |> I added sb=8193 to the mount options in /etc/fstab for /. : |> : |> How can I successfully boot from this partition now? Well, I've recently made such an experience on my system with an ST1480A Seagate 420 MB Harddisk. The only thing to do was to make a backup, getting a low-level Formating & Testing programm from seagate and test the entire disk so the Disk-Controller could remmap all bad-Blocks found. After it, all worked well again, but this should't happen anyway. I noticed it, since I had a bad block on one Startup File, so I had to fix this problem :^< cu. -- Solong & Happy Hacking ************************************************************************ * Joerg Mertin : smurphy@stardust.bln.sub.org (Home) * * in Berlin Spandau at : jorgagif@w203zrz.zrz.tu-berlin.de * * Stardust's Linux System : Data, Fax & Voice 49 30 3615569 * ************************************************************************ `Fatal Error: Found [MS-Windows] System -> Terminating Virus...' ------------------------------ From: cs_kokim@dmf123.ust.hk (Tall Sword) Subject: wxWindows 1.50k shared lib wanted Date: Thu, 13 Oct 1994 07:48:50 GMT I have compiled wxWindows 1.50k in my Linux box and found it quite useful for developing cross platform GUI program. But the wxWindows 1.50k only generate a static lib for me and it make the executable file quite large and the time of compiling is longer too. Anyone else where got a copy of shared lib of wxWindows 1.50k? I realized that there is copy in sunsite that support shared lib but that version is older that only support XView 3.1. Please direct me to the right site and directories to get the shared lib. -- * Origin: TallSword, Computer Science Year 2, HKUST internet: cs_kokim@dmf123.ust.hk, cs_kokim@stu.ust.hk root@dmf123.ust.hk, raymond@dmf123.ust.hk Raymond.Ko@f15.n700.z6.ftn.air.org fidonet: Raymond Ko, 6:700/15@fidonet.org ------------------------------ ** 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 ****************************** .