Files
oldlinux-files/ftp-archives/tsx-11.mit.edu/1996-10-07/mail-archive/linux-devel/Volume2/digest164
2024-02-19 00:24:15 -05:00

634 lines
23 KiB
Plaintext
Raw Blame History

From: Digestifier <Linux-Development-Request@senator-bedfellow.mit.edu>
To: Linux-Development@senator-bedfellow.mit.edu
Reply-To: Linux-Development@senator-bedfellow.mit.edu
Date: Tue, 13 Sep 94 02:13:05 EDT
Subject: Linux-Development Digest #164
Linux-Development Digest #164, Volume #2 Tue, 13 Sep 94 02:13:05 EDT
Contents:
Re: News Spool File System - new filesystem type?? (Alan Barrow)
Re: Linux Kernel's and ATDISK-Patches (Po-An Hsiung)
Re: Why I cannot mount a PhotoCD on Mitsumi ? (Fockert)
Re: -fPIC flag in gcc (Mitchum DSouza)
Re: Multiprocessing Pentium Systems (Colin Frank)
Re: DTC 3280 SCSI Driver info? (Trump66613)
Re: Has ARP been fixed ? (Howard "the Duck" Taylor)
Re: Slow curses - is there a better/faster curses? (Michael Engel)
Partitioning Question (David J Topper)
How to do file-io in a device driver (Matthias Sattler)
Re: nvi has a seriouis bug (Re: Help with development using vi.) (Spencer PriceNash)
Re: DOOM for Linux problem - help. (Dane Jasper)
Re: nvi has a seriouis bug (Re: Help with development using vi.) (Drew Eckhardt)
Re: PRIORITY make an undelete command (Matthias Urlichs)
----------------------------------------------------------------------------
From: jab@narcesc.atl.hp.com (Alan Barrow)
Crossposted-To: news.software.b
Subject: Re: News Spool File System - new filesystem type??
Date: 13 Sep 1994 00:37:48 GMT
In <3522qu$l22@rodan.UU.NET> rsalz@uunet.uu.net (Rich Salz) writes:
>I think I'm flattered that INN is becoming a generic term. :-)
More so to be flattered that INN was written well to allow such
"evolution".
>and might even get better performance
I am not convinced we would get better performance based on current news
volume and system capabilities.
I am more concerned about 5-10X volumes. CPU's are getting faster.
Disk interfaces and overall FS performance is not improving as fast as
news volume is growing.
CNEWS first, and then INN have made great strides in news system performance.
I am grateful! :-)
I could not run this site with CNEWS now, and I worry about where the
next /r$ will come from for the next perf leap. IE: we need to do what
you and Henry/Geoff did, start the "next generation" before we
absolutely need it. :-)
>and (I'm not holding my breath on these last two)
>be portable and free, >>it won't be INN.<<
Agreed. When I hear my pet DB labbie talking in 30k transactions per second
or minute, (it does not matter). I realize it will be hard to "home grow"
something that performs at that level.
I am no SQL fan, but it should allow portability at least at the DB
level. I see new (serious) apps that do not specify the DB, but
allow you to use the one your company has selected. (for presumably the
right reasons.) I like this idea.
So how do we provide this fuctionality to people who do not have a serious
DB? I guess we just don't!
If you are at the 1-5GB news spool level, you are probably not an
individual, and would be willing to spend some money (on a commercial DB)
to achieve performance and add new features. (I like my CDROM style total
keyword idea.) :-)
We can "specify/design" what we need to include in the DB as data or
index's for free. Then and only then do we code. Then someone with INN
and an ORACLE/INGRESS/INFORMIX/SYBASE/ALLBASE license tests it.
(probably at least the code person, as a start) :-) No big deal, JASMOP. :-)
Oh well, in everyone's copious spare time, I guess. :-)
Alan Barrow km4ba | If a little knowledge.....
Work: jab@atl.hp.com | is a dangerous thing.....
Home: alan@km4ba.ampr.org | then what is the Anti-Dote???
------------------------------
From: eric@laplace (Po-An Hsiung)
Subject: Re: Linux Kernel's and ATDISK-Patches
Date: 12 Sep 1994 09:50:13 GMT
Even after applying the atdisk patches to the latest kernel 1.1.50, I
cannot get it to recognize my second HD controller.
When I reboot, the following error occurs:
HD1 controller still busy: status 0xd0
HD1 controller times out ...
and the error message keeps repeating ...
I don't know why but I will try the latest 1.0 patch and see if it
works for my card.
BTW, I use IRQ 10 both on the card and in hd1.c
Anyone knows whats the problem>?
Eric.
Delman Lee (delman@mipg.upenn.edu) wrote:
: In article <Cvr24L.49@cove.han.de> knick@cove.han.de (Michael Knigge) writes:
: Will the atdisk-Patch (up to four IDE-Drives) be "standard" in the next
: Linux-Releases? I hope so because I need them and the Patch works without
: any Problems.....
: Probably not. The old atdisk2 patch is not a very clean solution. The
: new atdisk2 patch, which is a cleaner solution, is under testing. If
: you feel adventurous, you could try the alpha test code at
: mipgsun.mipg.upenn.edu:pub/delman/atdisk2-1.0alpha?.tgz
: Delman.
: --
: ______________________________________________________________________
: Delman Lee Tel.: +1-215-662-6780
: Medical Image Processing Group, Fax.: +1-215-898-9145
: University of Pennsylvania,
: 4/F Blockley Hall, 423 Guardian Drive,
: Philadelphia, PA 19104-6021,
: U.S.A.. Internet: delman@mipg.upenn.edu
: ______________________________________________________________________
--
~\\|//~
-(o o)-
+---------------------------oOOOo--(_)--oOOOo----------------------------+
|Eric, E-mail addr: eric@math.ntu.edu.tw |
|System Manager, Tel. ext. : 2903 (886-02-363-0231) |
|Department of Mathematics, Office : New Math. Bldg., Rm. # 402.|
|National Taiwan University. Time avail.: 12:30 to 17:30. |
+------------------------------------------------------------------------+
U U
------------------------------
Crossposted-To: comp.os.linux.help
From: fockert@prl.philips.nl (Fockert)
Subject: Re: Why I cannot mount a PhotoCD on Mitsumi ?
Date: Mon, 12 Sep 1994 13:25:40 GMT
badics@rutcor.rutgers.edu (Tamas Badics) writes:
>jeffpk@netcom.com (Jeff Kesselman) writes:
>>The poitn of all this is that ALL CD-ROM types are fully ISO9660
^^^^^ ^^^^^^^
But linux is not fully ISO9660 compatible !!!!
>>jeffk@crystald.com
>OK, it is all nice to know, but how can I read a PhotoCD on a Mitsumi drive?
>(My MS-DOS driver can read them without problems.)
The linux versions I have used (yggdrasil fall 93, summer 94) have
a hard coded restriction on the block size.
(look in the /usr/src/linux/fs/isofs/*.[ch] files for ISO_BLOCKSIZE)
When inserting a iso9660 cdrom with a blocksize != 2048
fall 93 : wrong fs type, already mounted, busy or other error.
summer 94 : blocksize != 1<< blockbits
I don't know the reason for this restriction, but I have a CDROM
with 512 byte blocks, and will try to modify the isofs.
I think you must do the same.
George
------------------------------
From: Mitchum.DSouza@mrc-apu.cam.ac.uk (Mitchum DSouza)
Subject: Re: -fPIC flag in gcc
Date: 12 Sep 1994 10:30:20 GMT
In article <34tleb$ih7@senator-bedfellow.MIT.EDU>, Ted Harding
<Ted.Harding@nessie.mcc.ac.uk> writes:
|> Not sure where best to post this, but anywhere that compiler buffs abound
|> looks good ....
|>
|> I'm using gcc version 2.5.8 on SlackWare Linux 1.2, and having a problem
|> compiling a big software package (sipp) which uses flag -fPIC (and probably
|> absolutely necessary for this package). This flag causes trouble. The
|> trouble is reproduced in miniature by the following little program.
|> Basically, it seems the compiler phase produces -fPIC code, but the
|> assembler phase won't accept it.
|>
|> Try it with 'gcc junk.c' -> OK; then 'gcc -fPIC junk.c' -> lots of
|> invalid "[" and "@" characters (use 'gcc -fPIC -S junk.c' to view
|> assembler).
|> Same effect with -fpic.
You can't use PIC with linux currently. This will change soon. Basically our
assember cannot handle PIC syntax generated from gcc. I am working on it...
Mitch
------------------------------
From: cfrank@sybase.com (Colin Frank)
Subject: Re: Multiprocessing Pentium Systems
Date: Mon, 12 Sep 1994 22:18:44 GMT
iialan@iifeak.swan.ac.uk (Alan Cox) uttered:
>In article <CvszIu.BJ@ndl.co.uk> rad@ndl.co.uk (Rich Deighton) writes:
>>It has to be said that Linux is a very good OS, but there is little need to
>>flame something just because it comes from Microsoft. Be open minded about
>>such things :-)
>
>I think the phrase is 'Bits of NT are pretty neat'. I certainly reckon you
>could run Linux as an NT subsystem instead of the psuedo posix bit.
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Yuckh! Isn't that like using an 1100 horsepower racing boat as the lifeboat
for an infaltable paddle boat.
Yeh, lets build a big bathroom with a house and a bedroom in it.
I think it would make more sense to creat an NT like widget set under X
since U*IX is what NT wants to be when it grows up. Or have WINE execute
windows 3.1 binaries, and put NT where the sun doesN'T shine.
C...
------------------------------
From: trump66613@aol.com (Trump66613)
Subject: Re: DTC 3280 SCSI Driver info?
Date: 12 Sep 1994 13:45:03 -0400
i have a DTC scsi card and have wanted to put together a linux drver for
it but i have had a heck of a time getting to DTC for any support. There
tech support just does nt work... i am uncertain as to the model # of the
DTC scsi card that i have. It works fine with my HD through linux using an
adaptec driver. Has anyone had any luck with DTC's bbs?
------------------------------
From: taylor@stimpy.eecis.udel.edu (Howard "the Duck" Taylor)
Subject: Re: Has ARP been fixed ?
Date: 13 Sep 1994 01:54:20 GMT
In article <CvuwCJ.1wB@info.swan.ac.uk> iialan@iifeak.swan.ac.uk (Alan Cox) writes:
>In article <1994Sep6.184123.52@elmrd6.ineab.ikea.se> anos@elmrd6.ineab.ikea.se (Anders Ostling) writes:
>>It seems like the ARP module is broken. It lists all my entries with
>>completely invalid IP addresses, but correct MAC address. Is somebody
>>working on this ? Is it solved ? When did it break ?
>
>It's not broken. The /proc/net/arp file format got changed to match the
>other /proc/net files so you need newer tools - changed about 1.1.12 I
>think.
>
>Alan
>--
> ..-----------,,----------------------------,,----------------------------,,
> // Alan Cox // iialan@www.linux.org.uk // GW4PTS@GB7SWN.#45.GBR.EU //
> ``----------'`----------------------------'`----------------------------''
can you bee a little more specific? I have this very same problem, and
I did get the new tools (I think) - net-0.32d .. I got that version of
rarp and also arp. this is the output that I see now:
========================================================================
[76] distraction ~ > tail /etc/rc.d/rc.local
# /sbin/dip /etc/CMF > /dev/null 2>&1 &
#
# This is for the addition of the sun3/50 as an xterminal
#
/sbin/rarp -s weenie 08:00:20:06:38:E8
# starting XDM
echo "Starting XDM..."
/usr/bin/X11/xdm &
[77] distraction ~ > rarp -a
IP address HW type HW address
134.207.6.42 10Mbps Ethernet 08:00:20:06:38:e8
[78] distraction ~ > arp -a weenie
SIOCGARP: Invalid argument
Address HW type HW address Flags
0.0.134.207 10Mbps Ethernet 00:00:AD:74:01:60 CMP
[79] distraction ~ >
========================================================================
anybody else see this kind of problem with the shifting of the ip addrs,
and the _way_ incorrect hw addr?
a few details:
1.1.49
slackware 1.2.0
net-0.32d
anyone else?
thanks,
--> howie
--
=============================================================================
Howard Taylor <taylor@ee.udel.edu>
Dept of Electrical Engineering
University of Delaware
------------------------------
From: engel@numerik.fb6.uni-siegen.de (Michael Engel)
Crossposted-To: comp.os.linux.help,comp.os.linux.misc,comp.os.linux.admin
Subject: Re: Slow curses - is there a better/faster curses?
Date: 12 Sep 1994 17:44:22 GMT
James Deibele (jamesd@teleport.com) wrote:
[ lots of text deleted ]
: elvis is a pretty faithful implementation of vi although it's noticeably
: slower in screen I/O than vi on a SPARC. And that's when I'm only the
: user on the linux box and there's nothing else happening on it. And
: with relatively small - 400-line - programs.
: Console output under Linux was very quick and I'm sure X performance is
: pretty good. But curses performance is a little sluggish and adding
: lines near the bottom of the screen is a real killer - curses seems to
: clear the screen with blank lines <then> adds the new text.
This seems to be an error in ncurses-1.8.5 :-(
I did experience exactly the same problem with my curses-based menu system.
Things are still worse: Under some circumstances, ncurses fills the screen
with another color than the background color...
In most other cases, I experienced that ncurses is about twice as fast as
SCO System V curses, at least on the console and on newer serial terminals
like a DEC VT-420/510. I didn't try it on a VT100, though ...
Btw, did you check that your terminfo entries are OK ? This has caused me
lots of trouble on some other versions of Unix.
: I know that curses and terminals are going the way of the dinosaur but
: like the dinosaur they're going to take a long time to die. We were
: looking at what it would cost to outfit our very small office with a
: PC/Mac/Unix solution that ran under Windows, MacOS, and UNIX and decided
: we could live with a curses interface.
Well, I have a Mac, a Sun-3 (with X) and a Linux box standing around here.
I am getting most work done when using the Linux console or the QVT-101
text terminal...
: But it would be nice if it were a little quicker. Is there a
: replacement curses out there somewhere? I checked the FAQs, How-Tos,
: and Meta-FAQ and didn't see a word about curses.
There is another version of curses out there, I think it's called PCcurses.
I remember having compiled it under SunOS, there was a Unix makefile included.
Sorry, I don't have it anymore. Ask archie.
: Thanks!
: --
: jamesd@teleport.com "Slowly cursing he deleted the word"
Michael Engel (engel@numerik.fb6.uni-siegen.de)
------------------------------
From: djt1@aloha.cc.columbia.edu (David J Topper)
Crossposted-To: comp.os.linux.admin,comp.os.liinux.help,comp.os.linux.misc
Subject: Partitioning Question
Date: 12 Sep 1994 18:09:47 GMT
So once I set my partitions (say 250 for Linux and 270 for DOS) using FIPS
or FDISK - is there a way to resize the Linux partition? I mean, is there a
way other than backing up all the data on the Linux partition, deleting that
partition, then resizing the DOS partition, then remounting the Linux partition,
then restoring the data on it.
I've downloaded quite a bit of stuff for installing Linux. I don't remember
exactly, but I think it was all the a,ap,x,xd,xv and a few other disks (all the
gnu stuff too). All tolled, I think I've got about 40+. The HOWTO says a
ROUGH estimate is about 2.5 megs per disk. I'm just not sure how much to
allocate for this thing. 250 seems fine but I would love to get away with 200
so DOS would have more room to breathe.
Thanks,
DT
------------------------------
From: sattler@unix-ag.uni-kl.de (Matthias Sattler)
Subject: How to do file-io in a device driver
Date: Mon, 12 Sep 1994 17:41:34 GMT
Hiho
Before you start flaming at me that device drivers don't do file-io, I will
tell you what I'm trying to do. I'm trying to write a block device driver
that combines several block devices to one device (especially useful
for people with a few smaller hard discs). Much of the management stuff
is already done, but I need to access the devices I want to combine to one.
I've tried it the simple way by using the normal open(), close(), read(),
write() system calls, but already open() fails (with "Bad address" error).
Can someone give me a hint please?
Matthias
PS: The device driver is written as a loadable module.
PPS: How do I officially obtain a device (major) number?
--
******************************************************************************
******************************************************************************
I am pleased to see that we have differences. May we together become
greater than the sum of both of us.
-- Surak of Vulcan, TOS: "The Savage Curtain", stardate 5906.4
******************************************************************************
******************************************************************************
Matthias Sattler
email: sattler@unix-ag.uni-kl.de
Phone: GERMANY 06333/65079 (always)
" 06333/63175 (at reasonable times only)
And always remember:
When all else fails, read the instructions.
------------------------------
From: spencer@montego.umcc.umich.edu (Spencer PriceNash)
Crossposted-To: comp.os.linux.help
Subject: Re: nvi has a seriouis bug (Re: Help with development using vi.)
Date: 13 Sep 1994 00:01:44 -0400
In article <3535pi$bj2@spool.cs.wisc.edu>,
Bryan S. So <so@brownie.cs.wisc.edu> wrote:
>:
>: >Get a better vi. nvi from ftp.cs.berkeley.edu (if I remember the
>: >address correctly) is a much better vi than elvis. (And let's you
>: >cut&paste under X, which is the exact reason why I dumped elvis)
>
>The problem with nvi (mine is ver 1.03) is, you can delete a line and put it
>into a buffer. Try this:
> "add
>to delete a line and put it in register a. And use
> "ap
>to put it back. It says "buffer a is empty" ... very scary!
>
>Use another better vi -- "vim".
Hmm. Let me try this on the above line.
Hey, it worked for me. Let's see if I can drag it out of the buffer
again.
>Use another better vi -- "vim".
Yep, worked again.
Of course, I'm using nvi 1.32. vim is ok, but nvi 1.32 has nicer
extensions. I dumped vim because of that backup file nonsense,
and later learned it could be configured not to make backup files,
but by that time I was hooked on nvi. I'm even disappointed when
I log into a Unix system and it's running regular vi.
--
Spencer PriceNash spencer@spencer.ann-arbor.mi.us spencer@umcc.umich.edu
Dan Quayle via anon ftp: Quotes at umcc.umich.edu in pub/users/quayle, GIFs
and sound files at vaxa.crc.mssm.edu in quayle/gif and quayle/sound.
------------------------------
From: dane@nermal.santarosa.edu (Dane Jasper)
Crossposted-To: alt.games.doom,comp.os.linux.help
Subject: Re: DOOM for Linux problem - help.
Date: 12 Sep 1994 21:08:28 GMT
Patrick Reijnen (patrickr@cs.kun.nl) wrote:
: Looks like you need to upgrade you X environment to 2.2.1. Your libs are to old for linuxxdoom.
Problem solved - it turns out that when I upgraded to 2.2.1, I forgot to run
ldconfig. Two people pointed this out to me and all is well now. Thanks
for all of you responses.
Dane
------------------------------
From: drew@frisbee.cs.Colorado.EDU (Drew Eckhardt)
Crossposted-To: comp.os.linux.help
Subject: Re: nvi has a seriouis bug (Re: Help with development using vi.)
Date: 13 Sep 1994 04:16:02 GMT
In article <3535pi$bj2@spool.cs.wisc.edu>,
Bryan S. So <so@brownie.cs.wisc.edu> wrote:
>:
>: >Get a better vi. nvi from ftp.cs.berkeley.edu (if I remember the
>: >address correctly) is a much better vi than elvis. (And let's you
>: >cut&paste under X, which is the exact reason why I dumped elvis)
>
>The problem with nvi (mine is ver 1.03) is, you can delete a line and put it
>into a buffer. Try this:
>
> "add
>
>to delete a line and put it in register a. And use
>
> "ap
>
>to put it back. It says "buffer a is empty" ... very scary!
>
>Use another better vi -- "vim".
Better yet, don't even bother with Linux, let alone the right
vi to run under it! After all, the kernel (ver .10) goes into infinite
loops if you have any non-IDE drives in the system, runs out of memory
quickly because it doesn't swap to disk...
In other words : if you want to run old, alpha versions of software,
you're free to do so. Just don't expect them to work well, and
don't expect the same bugs to carry into newer versions.
I run nvi 1.21, and historically, it's the most accurate of
all the vi clones I've used. It's also eight bit clean,
does split screen in text mode (aka Emacs, uemacs, vile,
etc), doesn't puke on complicated regexes the way Elvis
does, etc.
--
Since our leaders won't respect The Constitution, the highest law of our
country, you can't expect them to obey lesser laws of any country.
Boycott the United States until this changes.
------------------------------
From: urlichs@smurf.noris.de (Matthias Urlichs)
Subject: Re: PRIORITY make an undelete command
Date: 12 Sep 1994 20:46:39 +0200
In comp.os.linux.development, article <346jtv$sho@crl6.crl.com>,
davidh@crl.com (David Herron) writes:
>
> The only downside I can think of is the file system might end up
> more fragmented than it is now.
>
The Mac people had the problem that "deleted" files (i.e. files in the
trash) disappeared at semi-random times, eg. when shutting down the Mac
or when starting a program. This was such an user interface hassle that
the Mac OS now only deletes the trash when you tell it to.
The second problem is that unless you're very careful, you'll get
fragmentation. Lots of it. I once had a Netnews partition on an old-style
Sys5 file system (i.e., no free block bitmap). The thing was about 80%
fragmented after a month of continuous use (i.e., a 100-k file needed 80
seeks. Ugh). Avoiding the same situation for a file system with undelete is
not at all easy.
The third problem is how to associate the file with whatever directory it
was in when it got deleted. This gets _very_ messy when there are links on
the file which you as Joe User don't know anything about. For instance:
- I create File A
- My coworker hard-links it into his directory
- I decide to delete the thing
- I change my mind, type "unrm", am very astonished that the file is not
unrecoverable, and write a confused mail to the author of the unrm
program.
If the undelete system decides to keep changes, would my coworker be able
to get at an old version? How? How do I control whether he can or cannot do
this -- a separate set of permissions for deleted files? How complex _can_
you make "rm" and why would you want to? ;-)
Related to this is a fourth problem, which is that you have to remember
where links to an unremovable file have been if you need the space.
(Remember that under Unix, the traditional directory-to-inode mapping is
one-way, which
Offhand, I can't think of a solution which is neither a performance nor a
complexity problem.
--
Great men are not always idiots.
-- Karen Elizabeth Gordon
--
Matthias Urlichs \ XLink-POP N<>rnberg | EMail: urlichs@smurf.noris.de
Schleiermacherstra<EFBFBD>e 12 \ Unix+Linux+Mac | Phone: ...please use email.
90491 N<>rnberg (Germany) \ Consulting+Networking+Programming+etc'ing 42
PGP: 1B 89 E2 1C 43 EA 80 44 15 D2 29 CF C6 C7 E0 DE
Click <A HREF="http://smurf.noris.de/~urlichs/finger">here</A>.
------------------------------
** 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
******************************