From: Digestifier To: Linux-Development@senator-bedfellow.mit.edu Reply-To: Linux-Development@senator-bedfellow.mit.edu Date: Sun, 21 Nov 93 00:13:13 EST Subject: Linux-Development Digest #243 Linux-Development Digest #243, Volume #1 Sun, 21 Nov 93 00:13:13 EST Contents: Re: Thought... (Grant Edwards) Re: BusLogic SCSI under Linux...compatible with AHA 1740 or not? (Orest Zborowski) Re: rawrite.exe is too slow (Orest Zborowski) Re: ObjectBuilder/OI discussion group? (Warner Losh) Re: Don't use Motif for free sw: it now requires runtime royalties! (Ian McCloghrie) Re: How many BogoMips on a washing machine? (Charles Hedrick) Re: BusLogic SCSI under Linux...compatible with AHA 1740 or not? (Eric Youngdale) Re: Don't use Motif for free sw: it now requires runtime royalties! (Marty Leisner x71348) Re: Don't use Motif for free sw: it now requires runtime royalties! (Anthony A. Datri) Re: Don't use Motif for free sw: it now requires runtime royalties! (Anthony A. Datri) Re: Don't use Motif for free sw: it now requires runtime royalties! (Anthony A. Datri) Re: Berkeley Fast Filesystem (J. Dean Brock) Re: TAMU X INSTALL (Phil Perucci) Re: How many BogoMips on a Pentium? (Mike Horwath) ---------------------------------------------------------------------------- From: grante@hydro.rosemount.com (Grant Edwards) Subject: Re: Thought... Date: Sat, 20 Nov 1993 16:45:47 GMT Colin Owen Rafferty (craffert@nostril.lehman.com) wrote: : aos@rainbow.sosi.com (Michael Aos) writes: : > Can I set up something so when I speak into my microphone it comes out : > the other guys speakers? Some sort of socket thing? : all that you need to do is: : : beavis% record | rsh butthead play : : and it will work. : : In fact, if /dev/audio (or the equivalent) can be `cat'ed to/from it, : you can just do this (without even the record/play programs): : : beavis% cat /dev/audio | rsh butthead cat \> /dev/audio I tried this for grins and it does work under SunOS 4.1, but there is a delay of between 1/2 and 1 seconds between the microphone on machine A and the speaker on machine B. I assume that this is due to an I/O buffer somewhere that is getting filled before the data goes out. At 8000 bytes/second there must be a 4K or 8K buffer somewhere. If I can scrounge up another microphone I'm going to see if it works full duplex. Can one process have /dev/audio open for reading and another process have it open for writing? Yea, I know, I should get a life. -- Grant Edwards |Yow! Is something VIOLENT Rosemount Inc. |going to happen to a GARBAGE |CAN? grante@rosemount.com | ------------------------------ Crossposted-To: comp.os.linux.misc From: orestz@microsoft.com (Orest Zborowski) Subject: Re: BusLogic SCSI under Linux...compatible with AHA 1740 or not? Date: Fri, 19 Nov 1993 18:30:00 GMT In article <1993Nov16.051918.16449@news.clarkson.edu> chengb@logic.camp.clarkson.edu wrote: > Hi, I just received my BusLogic BT-747S. So far, I have only > gotten it to work under the AHA-1542 driver in the kernel, and it > doesn't seem to work when I have only AHA-1740 support compiled in. As > far as I know, the driver for 1542 probably isn't taking full advantage > of the EISA speed (10-33 Mbits/sec!!!), so is there a possibility to > hack the existing driver for one that takes full advantage of BusLogic > card? I have the same card, and after speaking with the people who sold me my system, they said that even with the 1542 driver you're still getting the full speed performance. The reason is that the BT747 loosely emulates the 1542, but does full busmastering transfers. I did some iozone timings and discovered that was more or less true. Of course, none of this is very scientific. I'd love to know more about this board. It seems to have trouble talking to a CD-ROM player I bought (SCSI-1, IBM PS/2 box, I forgot the actual drive...) Linux boots and the BT747 detects the drive, but accessing it generates tons of timeout messages (in the SCSI sense values)... -orest ------------------------------ From: orestz@microsoft.com (Orest Zborowski) Subject: Re: rawrite.exe is too slow Date: Fri, 19 Nov 1993 18:33:35 GMT One of the reasons rawrite is slow is that it only writes 3 sectors at a time, probably because that's a safe limit for all drives. I changed my copy to write a whole track at a time without errors. Makes for a great improvement and I use it as a DOS dd. -orest ------------------------------ From: imp@boulder.parcplace.com (Warner Losh) Subject: Re: ObjectBuilder/OI discussion group? Date: Sat, 20 Nov 1993 20:40:52 GMT In article <2cj7du$3ok@nova.umd.edu> jsmith@nova.umd.edu (James H. Smith) writes: >Does anyone >know of a mailing list or discussion group anywhere for Objectbuilder/OI >developers? Or is this the place? > >The release notes for Objectbuilder/OI reference a mailing list that can >be joined by sending email to "oi-users-request@bbn.com". However my >email gets returned from bbn.com with an unknown user error message! :-( The list have moved from bbn.com to cdrom.com. So, to get added to the list, please send mail to oi-users-request@freefall.cdrom.com. The developers do read that list. Warner -- Warner Losh imp@boulder.parcplace.COM ParcPlace Boulder I've almost finished my brute force solution to subtlety. ------------------------------ From: imcclogh@cs.ucsd.edu (Ian McCloghrie) Crossposted-To: comp.windows.x,comp.windows.x.i386unix,comp.windows.x.motif,gnu.misc.discuss,comp.sources.d Subject: Re: Don't use Motif for free sw: it now requires runtime royalties! Date: 20 Nov 93 20:48:49 GMT doug@aer.com. (Douglas W. Johnson) writes: >which tells me that no Motif runtime libraries are needed, nor is Mosaic >statically linked. I expect that your binary is statically linked *to the Motif libraries*, but not to the other shared libs that you have on your system. -- /~> Ian McCloghrie | Commandant of Secret Police - Cal Animage Beta. < < /~\ |~\ |~> | | <~ | email: ian@ucsd.edu Net/2, USL 0! \_> \_/ |_/ |~\ |__| _> | Card Carrying Member, UCSD Secret Islandia Club ------------------------------ From: hedrick@geneva.rutgers.edu (Charles Hedrick) Subject: Re: How many BogoMips on a washing machine? Date: 20 Nov 93 22:43:14 GMT wirzeniu@klaava.Helsinki.FI (Lars Wirzenius) writes: >I really wish I had a way to record and digitize the evil laughter that >Linus emits every time he starts reading comp.os.linux.development (and >.misc) and sees that the BogoMips threads still continue. I would be the last to deprive Linux of any source of amusement. However I think it's useful to tabulate the expected bogomips for each processor type and speed. I think one of reasons to continue printing the bogomips figure is for people to verify that their machine is properly configured. We've caught a system that was inadvertently running with reduced clock speed. For this to work, we had to know roughly what the expected bogomips figure was for that processor type. ------------------------------ Crossposted-To: comp.os.linux.misc From: eric@tantalus.nrl.navy.mil (Eric Youngdale) Subject: Re: BusLogic SCSI under Linux...compatible with AHA 1740 or not? Date: Sat, 20 Nov 1993 23:45:27 GMT In article orestz@microsoft.com (Orest Zborowski) writes: >In article <1993Nov16.051918.16449@news.clarkson.edu> chengb@logic.camp.clarkson.edu wrote: >> Hi, I just received my BusLogic BT-747S. So far, I have only >> gotten it to work under the AHA-1542 driver in the kernel, and it >> doesn't seem to work when I have only AHA-1740 support compiled in. As >> far as I know, the driver for 1542 probably isn't taking full advantage >> of the EISA speed (10-33 Mbits/sec!!!), so is there a possibility to >> hack the existing driver for one that takes full advantage of BusLogic >> card? > >I have the same card, and after speaking with the people who sold me my >system, they said that even with the 1542 driver you're still getting >the full speed performance. The reason is that the BT747 loosely emulates >the 1542, but does full busmastering transfers. I did some iozone timings >and discovered that was more or less true. Of course, none of this is >very scientific. As I have said before, I do not find it surprising that the Bustek EISA boards in the 1542 emulation give much better performance than the 1542. Anyway, there is a native driver for BusTek boards under development, and very soon people will be able to try both and see what the actual difference is. >I'd love to know more about this board. It seems to have trouble talking >to a CD-ROM player I bought (SCSI-1, IBM PS/2 box, I forgot the actual drive...) >Linux boots and the BT747 detects the drive, but accessing it generates >tons of timeout messages (in the SCSI sense values)... The sense buffers can be decoded and they provide a lot of information about what might be going wrong. Lots of timeout messages from a cdrom can be a symptom of a number of things. You might want to check: * Bus termination. * Flakey scsi cable. * You are not overloading the power supply on your system. * Thumbprints on the cdrom. * The laser head is clean (you can get special cleaning discs). * The SCSI mode page for error control is not set up to retry read errors. -Eric With very inexpensive power supplies that come with clones like what I have, running too many things at the same time can result in a lot of data errors on a cdrom. Unplugging some unused things seems to help quite a bit (actually at the moment I have an external power supply connected to my cdrom). I do not have a good sense of where the actual problem lies - it could be noise coming through because of poor filtering, or it could be cheap power connectors that do not make good contact. I used to have an NEC drive that would intermittently give me all sorts of trouble, but when I put in a Dell box with SVr4 it has started working like a champ. The Texel drive that I replaced the NEC with is also giving intermittent errors. Anyone have a 300W Dell power supply for sale? (-: -- "The woods are lovely, dark and deep. But I have promises to keep, And lines to code before I sleep, And lines to code before I sleep." ------------------------------ Crossposted-To: comp.infosystems.www,comp.windows.x,comp.windows.x.i386unix,comp.windows.x.motif,gnu.misc.discuss,comp.sources.d From: leisner@hydrus (Marty Leisner x71348 ) Subject: Re: Don't use Motif for free sw: it now requires runtime royalties! Reply-To: leisner@eso.mc.xerox.com Date: Sat, 20 Nov 1993 20:19:45 GMT I recall when Motif was coming out there was a request not to use it since its not free... I'm not an X programmer (yet) but an X user/builder/debugger of C code... From what I've seen its not clear what you accomplish with motif you can't accomplish with the Athena widgets... maybe its a little harder (?? -- how much?) but at least its free to give away... -- Marty Leisner leisner@eso.mc.xerox.com leisner.henr801c@xerox.com "I just know I'm a better manager when I have Joe DiMaggio in center field" -- Casey Stengel ------------------------------ Crossposted-To: comp.windows.x,comp.windows.x.i386unix,comp.windows.x.motif,gnu.misc.discuss,comp.sources.d From: aad@dvorak.amd.com (Anthony A. Datri) Subject: Re: Don't use Motif for free sw: it now requires runtime royalties! Date: Sun, 21 Nov 1993 01:36:24 GMT >Just to make this subtle point absolutely clear: OSF does not require >the repeated payment of the runtime library feee: if it is _proven_ that >the end user has already paid the royalty for a copy of the OSF Motif >runtime library, either standalone or not, the supplier of the software >need not pay the royalty again. In reality, though, what will happen is the the developer will pay the royalty, and the end-user will also pay it. Developers will not exert the effort to have different pricing, or to verify anything. -- ======================================================================8--< ------------------------------ Crossposted-To: comp.infosystems.www,comp.windows.x,comp.windows.x.i386unix,comp.windows.x.motif,gnu.misc.discuss,comp.sources.d From: aad@dvorak.amd.com (Anthony A. Datri) Subject: Re: Don't use Motif for free sw: it now requires runtime royalties! Date: Sun, 21 Nov 1993 01:55:48 GMT >From what I've seen its not clear what you accomplish with motif you >can't accomplish with the Athena widgets... >maybe its a little harder (?? -- how much?) but at >least its free to give away... You fill up your disk with huge executables and spend all of your memory and cycles drawing pseudo-3D art-deco borders around every linear displayed feature. How a button drawn this way is supposed to be easier to click on than one drawn with normal lines (a la Xaw) is beyond me. I also don't see how something like an xbiff is easier to use when its screen area is doubled by the abovementioned borders. -- ======================================================================8--< ------------------------------ Crossposted-To: comp.windows.x,comp.windows.x.i386unix,comp.windows.x.motif,gnu.misc.discuss,comp.sources.d From: aad@dvorak.amd.com (Anthony A. Datri) Subject: Re: Don't use Motif for free sw: it now requires runtime royalties! Date: Sun, 21 Nov 1993 01:51:55 GMT >The principle effect of specifying OSF/Motif as part of the CDE is >that libXm is guaranteed to be available on any machine running the >CDE. I predict that the same thing will happen as has happened with "DCE". It will be so expensive that nobody will buy it, and it will be similarly ignored. -- ======================================================================8--< ------------------------------ From: brock@cs.unca.edu (J. Dean Brock) Crossposted-To: comp.os.linux.admin Subject: Re: Berkeley Fast Filesystem Date: 21 Nov 1993 02:11:59 GMT In article pcg@aber.ac.uk (Piercarlo Grandi) writes: > ... a lot of stuff about extending files under FFS > ... in response to an earlier message of mine In my message, I was talking about the allocation of fragments on the disk. Piercarlo Grandi seems to be talking about "the really terrifying details of the consequences of the FFS fragmentation scheme on incore buffers". I'll agree that I was terrified. Again, look at The Design and Implementation of 4.3 BSD for the details of allocating fragments. It's all in Section 7.7. Here's a brief rewording that summarizes the interesting facts. If a file system has a reasonable amount of space, disk allocation is being optimized for time. In this case, the first fragment of a file is allocated on a best-fit basis. That is, the system tries to find an isolated fragment for this file. When the file grows into a second fragment, the old and new fragment are placed in the first two fragments of an empty block. The file can grow six more fragments until a new block is needed. Each time, the file uses all eight fragments of a block, a new block is allocated for the next fragments of the file. If the file system is pretty full, disk allocation is being optimized for space. In this case, fragments are allocated on a best-fit basis as the file grows. Obviously this can involve a lot of allocating and de-allocating as the file grows. So, once again, when allocating fragments on a disk using the "fast" file systems that descended from BSD: Files CAN use multiple fragments. A 42k file WILL consist of 8 full blocks and 2 fragments. ------------------------------ From: philp@universe.digex.net (Phil Perucci) Subject: Re: TAMU X INSTALL Date: 20 Nov 1993 23:17:24 -0500 In article <1993Nov20.050819.6807@msus1.msus.edu>, Patrick J. Volkerding wrote: >Quite to my surprise, I found myself flamed by some of the XFree people >over including this in the Slackware distribution. I've never had any >problems with it, and I don't know of anyone who has. I imagine there are >some people out there who never did get it to work, but I've used it to >great success on a variety of accelerated and non-accelerated cards. > >So, I'd like to hear if anyone out there has experienced trouble with >this package, such as hardware problems during the initial setup, or later >while using a mode that seems to work. > >Reports of success would be good, too. :^) > I remember reading related posts some time back. It seems the danger is for people without multisync monitors. Multisync monitors are reported to be safe in that they can not be driven beyond safe operating limits. Non-multisync monitor CAN be driven outside safe limits. This can quickly result in blown fly-back transformers! Bummer! My multisync happens to be NEC, but I believe other non-NEC multisyncs should be OK (safe)... Still, next time I eagerly install the latest and greatest Linux (Slackware) I will 1) use my current Xconfig as a go-by and 2) look at /usr/X386/lib/X11/etc/modeDB.txt. -- ============================================================================== Phil Perucci | "All postings are my own opinion - all comments Systems Programmer | are intended for research/educational purposes" ============================================================================== ------------------------------ From: root@jacobs.mn.org (Mike Horwath) Subject: Re: How many BogoMips on a Pentium? Date: 20 Nov 1993 18:47:12 GMT I tested a Pentium for a week at my job. I didn't write down the bogomips cause it wasn't worthwhile to do this, since it is just a timer setup routine for some devices (QIC-80 come to mind). What I did do was benchmark it in other ways. I was looking to see if it was a worth while processor for a future machine. Things I did: dhrystone compile of the kernel, 99pl13 clean news expiration (harder to watch, since the systems were a litte different) compiles of the same packages concurrently What were my results? dhrystone Pentium 35% faster compile of the kernel Pentium 22% faster news expiration Pentium 2-7% faster (clean disks versus other system) compiles concurrent Pentium 23% faster These are from the top of my head. It was faster, but not as much as a person would think. Configurations of the machines were basically the same: 486-DX50, 8 megs RAM, 256K cache, WD IDE drives (340, 170) Pentium, 8 megs RAM, 512K cache, WD IDE drive (170) I think I am going to run another test again, using basic same configs but using the same exact drives (dd if=/dev/hda of=/dev/hdb :) so that things like fragmentation and layout of the filesystems can be reduced. My conclusion: It was fast, but not worth the extra cash to spend on it. The concurrent compile was the best I could do at multiple users using the machine, since my machine runs multiple users here at home all the time, I had to try to get the pentium to the same area of usage. -- Mike Horwath IRC: Drechsau BBS: Drechsau LIFE: lover root@jacobs.mn.org drechsau@jacobs.mn.org Jacob's Ladder 612-588-0201 UUCP, UseNet, Linux files, BBS ------------------------------ ** 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 ******************************