Files
2024-02-19 00:25:02 -05:00

535 lines
21 KiB
Plaintext

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: Fri, 1 Apr 94 07:13:04 EST
Subject: Linux-Development Digest #594
Linux-Development Digest #594, Volume #1 Fri, 1 Apr 94 07:13:04 EST
Contents:
HELP ME --- Running X with ORCHID KELVIN 64 VLB (Etienne Provencher)
Re: IDE Performance Package (Wayne Schlitt)
Soundblaster 16 SCSI - Supported by Linux? (Frank Luthe)
Re: profiling anyone? (Ivan)
Re: Slackware as a tar.gz file? (Paul Tomblin)
Re: PC as C64 file server (Charles T Wilson -- Personal Account)
Re: Soundblaster 16 SCSI - Supported by Linux? (Jon Cardwell)
Re: IDE Performance Package (Daniel Aaron Supernaw-Issen)
Re: BusLogic BT445S driver? (Rob Janssen)
Speed problem with more then one ext2-partition (Aurel Balmosan)
Re: IDE Performance Package (David Monro)
Re: LINUX port to a transputer system (Karri Kaksonen)
Re: NFS timeouts (Frank Lofaro)
----------------------------------------------------------------------------
From: provench@cs.unc.edu (Etienne Provencher)
Crossposted-To: comp.os.linux.help,comp.windows.x.i386unix
Subject: HELP ME --- Running X with ORCHID KELVIN 64 VLB
Date: 29 Mar 1994 10:21:37 -0500
I have just purchased the new Orchid Kelvin 64-bit graphics card. It is
not supported in the current version of XFree86 as far as I know so I
don't know how to run X at this point. It uses a new Cirrus Logic chipset, I
think it is a 5434. Does anyone else have this card up and running
with the correct configurations and/or how may I go about getting this
card to work?
Also, someone mentioned that I may be able to set it up as a Cirrus
Logic 5426 (although not taking advantage of its much greater speed
capabilities). What would I need to change to try this out? I used to
have a Trident 8900C w/1 meg (now you know why I upgraded...SLOOOOOW)
On another note...what is a good way to test the "speed" of this card
to compare it to others...it seems fairly fast in Windoze (yuch)
Be Kind...I am a newbie trying to learn
Thanks in advance!
Etienne Provenher
provench@cs.unc.edu
Respone to e-mail would be preferred
------------------------------
From: wayne@backbone.uucp (Wayne Schlitt)
Subject: Re: IDE Performance Package
Date: Thu, 31 Mar 1994 02:18:47 GMT
Reply-To: wayne@cse.unl.edu
In article <2nccga$3f0@bmerha64.bnr.ca> mlord@bnr.ca (Mark Lord) writes:
>
> Hi. The patches are deliberately suspicious of any drive which supports
> fewer than 32 sectors in multiple mode. Your drives support only about 16,
> so it skips them by default.
Just out of curiosity, why are you suspicious of any drive that
supports fewer than 32 sectors? I would think that 16 or even 8
sectors would go a long way in reducing the overhead, and that you
probably aren't saving much by increasing it to 32. Cutting it down
to 16 (or 8) might also help those serial port overruns.
-wayne
--
The Fundamental Problem with USENET is that you have at least a couple
of hours, if not a day or so to think up that witty, absolutely
devastating retort... The other Fundamental Problem is people don't
even take a couple of minutes to think before they hit that send key...
------------------------------
From: fl@Germany.EU.net (Frank Luthe)
Subject: Soundblaster 16 SCSI - Supported by Linux?
Date: 30 Mar 1994 22:41:53 +0200
Hi there,
today I examined the SCSI-Howto and found the following lines:
Subsection C : Adaptec 152x, 151x, Sound Blaster 16 SCSI,
AIC 6260 chips (Standard)
Supported Configurations :
BIOS addresses : 0xd8000, 0xdc000, 0xd0000, 0xd4000, 0xc8000, 0xcc000, 0xe0000,
0xe4000.
Ports : 0x140, 0x340
IRQs : 9, 10, 11, 12
DMA is not used
IO : port mapped
Does this definitely mean that Linux can support the SB 16 SCSI,
because it is Adaptec 152x-compatible ????
That would be great... any clue would be appreciated!
--
Frank Luthe
=== ____ === fl@Germany.EU.net
=== / / / ___ ___ _/_ === EUnet Deutschland GmbH
=== /---- / / / / /___/ / === Emil-Figge-Str. 80
=== /____ /___/ / / /___ / === D-44227 Dortmund
===== ===== Tel. +49 231 972 00
===== Connecting Europe since 1982 ===== Fax +49 231 972 1111
------------------------------
From: ivan@djomolungma.Eng.Sun.COM (Ivan)
Subject: Re: profiling anyone?
Date: 30 Mar 1994 21:15:48 GMT
Reply-To: ivan@djomolungma.Eng.Sun.COM
! From bas@phys.uva.nl (Bas de Bakker)
! Date: 30 Mar 1994 06:58:13 GMT
!
!I'm not sure what it is exactly that you want. There is indeed no
!profil() system call in the Linux kernel, but there is a routine in
!the C library under that name which does the same thing.
!
!As to profiling tools: gprof is in the binutils package
Yes, except that the output is function count based ... there is no
timing information. That is probably because 'profil(2)' isn't implemented.
'profil(3)' obviously isn't doing the "same thing", otherwise gprof
would've provided me with timing data.
!and recently I wrote ...
!sunsite.unc.edu:/pub/Linux/devel/bprof-0.1.tar.gz
Just got it, will give it a try tonight.
------------------------------
From: ptomblin@gandalf.ca (Paul Tomblin)
Subject: Re: Slackware as a tar.gz file?
Date: 29 Mar 1994 20:28:53 -0500
jkaidor@synoptics.com (Jerome Kaidor) writes:
> I dreamt of a script that would activate FTP, tell it to get slackware.tar, and pipe its
>output straight up to tar on my machine, which would then spew out files and directories.
>Probably an impossible dream......
Hmmmmm - wouldn't
get slackware.tar |tar xvf -
work?
I'll have to try it.
--
Paul Tomblin, Head - Automation Design Group.
Gandalf Canada Limited
This is not an official statement of Gandalf, or of Vicki Robinson.
"Hello, this is Linus Torvalds, and I pronounce Linux as Linux"
------------------------------
From: ctwilson@rock.concert.net (Charles T Wilson -- Personal Account)
Subject: Re: PC as C64 file server
Date: 30 Mar 1994 04:47:31 GMT
In article <1994Mar30.034709.4583@taylor.wyvern.com>,
Mark A. Davis <mark@taylor.wyvern.com> wrote:
>k-garner@ux4.cso.uiuc.edu (Garner Keith Thomas) writes:
>>acbul1@lindblat.cc.monash.edu.au (Andrew Bulhak) writes:
>>>Sven Goldt (goldt@math.tu-berlin.de) wrote:
>>>: paul (paul@dino.eng.monash.edu.au) wrote:
>>>: : Ok,
>>>: : It seems quite clear that there is a need for a device that allows
>>>: : a standard ibm pc to be used as a file server for our humble ol' Commodore
>>>: : 64's. Is anyone working on such a device? What do people think about the idea?
>>>: : Is it possible ??
>
>It seems like it would be a lot easier to use a better obsolete system, like
>the Tandy COCO's running a real OS; one which is semi-multi-user, fully
>multitasking, re-entrant, kernel/driver designed, multi-windowing, etc
>...... OS-9 :) Hard to believe, isn't it! It was my start before I jumped
>into Unix. It is still impressive, even today.
It's pretty impressive all right, especially when you consider that it ran
in as little as 64K. It was my first exposure to something unix-like, too.
Too bad they had such anemic keyboards...it was fun to play with.
--
/-----------------------------------------------------------------------\
| Tom Wilson | "I can't complain, but sometimes |
| ctwilson@rock.concert.net | I still do." |
| | -Joe Walsh |
------------------------------
From: jcardwell@umi.com (Jon Cardwell)
Subject: Re: Soundblaster 16 SCSI - Supported by Linux?
Date: 31 Mar 1994 10:32:19 -0500
In article <fl.765059503@orca> fl@Germany.EU.net (Frank Luthe) writes:
>Hi there,
>
>today I examined the SCSI-Howto and found the following lines:
>
>Subsection C : Adaptec 152x, 151x, Sound Blaster 16 SCSI,
> AIC 6260 chips (Standard)
>Supported Configurations :
>BIOS addresses : 0xd8000, 0xdc000, 0xd0000, 0xd4000, 0xc8000, 0xcc000, 0xe0000,
> 0xe4000.
>Ports : 0x140, 0x340
>IRQs : 9, 10, 11, 12
>DMA is not used
>IO : port mapped
>
>Does this definitely mean that Linux can support the SB 16 SCSI,
>because it is Adaptec 152x-compatible ????
>
>That would be great... any clue would be appreciated!
I have a Soundblaster-16scsi-2 card, and use external SCSI-2
devices under linux successfully. What I did was turn on the
Adaptec 152x support when I build my kernels. Also, since there
is no boot-ROM on the SB16SCSI2 card, one must tell the kernel
to look for the chip upon boot-up at the "LILO: " prompt:
LILO: linux aha152x=0x140,11,7,1
Arguments:
0x140 = I/O port base address.
11 = IRQ that the chip is set to (jumper)
7 = SCSI ID of the chip itself.
1 = flag for SCSI reconnect feature (I Think... May be optional...)
Unfortunately though, the aha152x driver seems to have a bug in it
that crashes the system completely when I transfer LARGE files to/from
a Syquest 105M removeable HD. I can do normal I/O to a toshiba XM3301B
cdrom drive and the syquest, but the system hangs with a message to
the effect "more data than expected &^%&$&..." and hangs.. :-(
Anybody know of a patch to the aha152x driver? I haven't looked
at the source code yet to see who the author(s) is/are yet...
>--
> Frank Luthe
> === ____ === fl@Germany.EU.net
> === / / / ___ ___ _/_ === EUnet Deutschland GmbH
> === /---- / / / / /___/ / === Emil-Figge-Str. 80
> === /____ /___/ / / /___ / === D-44227 Dortmund
> ===== ===== Tel. +49 231 972 00
> ===== Connecting Europe since 1982 ===== Fax +49 231 972 1111
--Jon Cardwell
University Microfilms International
------------------------------
From: danielsi@cs.utexas.edu (Daniel Aaron Supernaw-Issen)
Subject: Re: IDE Performance Package
Date: 30 Mar 1994 22:19:43 -0600
I've installed the ide performance package upon linux 1.0 and have found
the following: Whenever I have disk activity, the mouse jumps around under X.
This has made the system unusable whenever there is any real swapping going
on. Worse yet, I can't manage to control the mouse enough to be able to
kill the offending apps gracefully. Enevitably, I end up exiting X and killing
the app by hand. Not good. btw I'm running linux 1.0 on a 386-33 with 8M
ram and a 100M Conner ide drive. I hope that this can be fixed before
inclusion into the standard kernel - it really makes machine quite unusable.
Daniel Supernaw-Issen
------------------------------
From: rob@pe1chl.ampr.org (Rob Janssen)
Subject: Re: BusLogic BT445S driver?
Date: Thu, 31 Mar 1994 11:31:40 GMT
Reply-To: pe1chl@rabo.nl
In <CnHBnB.9rJ@eskimo.com> vicki@eskimo.com (Victoria Harrington) writes:
>pyeatt@CS.ColoState.EDU (Larry Pyeatt) writes:
>>I have a BusLogic BT445S fast SCSI adapter. Is there a driver for
>>this board, or should I start writing my own. Any pointers would be
>>appreciated.
>I have such a board on order also. There are drivers available for
>in at tsx-11.mit.edu:/pub/linux/drivers/ALPHA (I think). Check the
>ls-lR for the exact location. However, ONLY the drivers are
>provided (buslogic.tar.gz = buslogic.c+buslogic.h). Installation
>instructions are most notable by their absence.
It does say ALPHA, doesn't it?
Rob
--
=========================================================================
| Rob Janssen | AMPRnet: rob@pe1chl.ampr.org |
| e-mail: pe1chl@rabo.nl | AX.25 BBS: PE1CHL@PI8UTR.#UTR.NLD.EU |
=========================================================================
------------------------------
From: aurel@perseus.uni-paderborn.de (Aurel Balmosan)
Subject: Speed problem with more then one ext2-partition
Date: 31 Mar 1994 17:55:09 +0200
I have a speed problem using more then one ext2 partition on one harddisk.
I have recognized that if I using two and more (ext2) partitions the raw
harddiskspeed drop from >1M per sec (15.15 sec for 16Mbytes) down to
0.8M per sec (20.00 sec for 16Mbytes) with two ext2 partitions and
0.64M per sec (25.00 sec for 16Mbytes) with three ext2 partitions mounted.
I used following speed test:
dd if=/dev/hda1 of=/dev/null bs=8k count=2k
I am using linux under following hardware condition:
486DX33 8Mbyte RAM,ISA, AHA1542CF SCSI:600M, IDE:420M(Western digital), 160M
(Conner)
The SCSI-Drive contains one ext2-partition. The 160M(Conner) contains one dos
-partition and the 420M(Western digital) contains 1 dos, 1 swap, and 1-3 ext2-
partitions.
The speed data above are for the IDE:420M(Western digital). I have not tested
if the SCSI-Drive behave like the IDE-Drive.
My question is: Where does the kernel lose the speed performance.
--
Alexandru-Aurel Balmosan aurel@uni-paderborn.de
University of Paderborn (Germany)
------------------------------
From: davem@extro.ucc.su.OZ.AU (David Monro)
Subject: Re: IDE Performance Package
Date: Wed, 30 Mar 1994 17:20:19 GMT
jcgreen@iastate.edu (Jon Green) writes:
>I just installed the patch to enable Multimode with my IDE drives, and
>got the following message on bootup:
>hda: WDC AC2340H (325MB IDE w/128KB Cache)
>hda: older drive, multiple mode not enabled
>hda: hda1 hda2
>hdb: st3144AT (124MB IDE w/32KB Cache)
>hdb: older drive, multiple mode not enabled
>hdb: hdb1
Firstly, try defining VERBOSE_DRIVEID as 1 at the top of hd.c - you will
get lots of info. The key is the MaxMultSect field in this extra info. As
the kernel patch stands, it only enables the multi mode stuff if this is
>= 32, as there were problems with some drives with MaxMultSect = 16.`
>I would have thought the Western Digital drive would have this feature, as
>it is practically brand new. Oh well, it matches the rest of my system. :)
This time, you are (sort of) in luck. The WDC AC2540H, which is the 540Mb
version of the same drive, supports 16 sector multimode, so I would be
extremly surprised if the 340Mb version doesn't. The trick is as follows:
change line 324 of hd.c from
if ((i = ib[47] & 0xff) >= 32)
to
if ((i = ib[47] & 0xff) >= 16)
and it should work.
However, I have (I think) the same seagate drive, and it doesn't support
multimode at all.
Before installing the patches I was getting a raw transfer rate off the WD
of around 730 kb/s, now I get about 1200k/s, and reading from the drive
consumes a bit less cpu. Your mileage may vary.
>My question is this: Since multimode is not available, can I expect to see
>any performance increase by installing this patch?
Possibly a little - interrupts are enabled for more of the time. Should
allow your machine to get a little more other work done while something is
waiting for the drive.
FYI, my other drive is a "Conner Peripherals 170MB - CP30174E", which also
has 16 sector multimode capability - with that drive I get no speedup, in
terms of elapsed time to read data, but it consumes a lot less cpu (like
22% instead of 60%). Again, your mileage may vary.
>--
>* Jon Green * Still searching for the * Friley 5646 Lorch-Russell *
>* jcgreen@iastate.edu * queen of my double-wide * Ames, Iowa 50012-0001 *
>* Jon2@irc * trailer :) * Phone (515) 296-0648 *
--
I stood at the edge and I'm looking down
Caught in the danger zone
I feel like a king that has lost his crown
And now I stand here alone
------------------------------
From: karri@cute.unda.fi (Karri Kaksonen)
Subject: Re: LINUX port to a transputer system
Date: 31 Mar 1994 15:48:13 GMT
> > I said I wanted to
> > have the 486 do all the I/O work and thus working as a server with the
> > transputer as a client.
It has been done the other way around.
I once attended an Inmos conference where they told about a successfull
experiment of writing a X-server completely in Occam. They defined the
X-protocol as a protocol for a channel and implemented the functionality
from scratch. This resulted in a fast X-terminal.
> > The port wouldn't be written in OCCAM 2 because that would give me a HUGE pain
> > in the BUM!!!!
> > Because of the way how OCCAM 2 is written. But in C and compiled with a 3L-C
> > Compiler.
>
> What's wrong with OCCAM? You have to 'at best' extend the syntax of C or depend
> on run-time library support to make use of the Transputer's implementation of
> parallelism. Okay, I guess it's up to you!
One word about Occam 2 - I _love_ it! I wrote software for a NMR-scanner
in Occam 2 about 4 years ago. Now I am struggling with Solaris and C++.
And I really miss the simplicity and power behind the most ingenious language
in the world. It would be a great thing to build a kernel that could use
the lightning-fast task switching and channel interrupts that the transputer is good at. Shared memory can also be addressed by accessing memory through
channels.
Best of luck!
karri
--
Karri Kaksonen OH2BEK .......................... Beam me up Scotty,
Unda Oy - A Scitex subsidiary .................. there's no intelligent
tel +358-0-52558522 fax +358-0-52558585 ........ life down here.
Kutojantie 7, FIN-02630 Espoo, Finland ......... [Dr. Spock to Enterprise]
------------------------------
From: ftlofaro@unlv.edu (Frank Lofaro)
Subject: Re: NFS timeouts
Date: Tue, 29 Mar 94 22:31:05 GMT
In article <1994Mar29.013504.25381@cc.gatech.edu> byron@cc.gatech.edu (Byron A Jeff) writes:
>In article <9403282352.AA26078@cs.utexas.edu>,
>Frank Lofaro <ftlofaro@mayall.CS.UNLV.EDU> wrote:
>>In article <1994Mar28.133906.8797@cc.gatech.edu> byron@cc.gatech.edu (Byron A Jeff) writes:
>>>In article <proff.764778560@suburbia>,
>>>Julian Assange <proff@suburbia.apana.org.au> wrote:
>>>>ward@crl.com (Ward Mullins) writes:
>>>>>I'm trying to use a Linux Box as an NFS server to a Sparc running Solaris
>>>>>2.3, and I keep getting thousands of server timeouts, followed by server
>>>>>OK messages.
>>>>Linux nfs is broken, large reads kill it. Small read (around 512 bytes) are ok,
>>>>with solaris.
>
>>>Linux NFS is not broken. It has different buffer sizes than the Sun OS's
>>>(SunOS and Solaris). It's the NFS clients's responsibility to set buffer
>>>sizes. So if anything is broken (and nothing is) then it's Solaris ;-)
>>>
>>>Anyway the solution. When you mount set the buffer size to max 1k. Example:
>>>
>>>mount linux:/ /solaris -o rsize=1024 wsize=1024
>>>
>>>End of problem. I've transferred upwards of 120M at a time (tar backup)
>>>over this kind of interface. No program necessary. Inventive though.
>>>
>>
>>Linux NFS _is_ broken. You don't have to use the rsize wsize kludge for
>>other OS's. It is a restriction that is unique to Linux (possibly plus a
>>small handful of other OS's). This is BAD. I think it is very good that
>>we have Linux and NFS for Linux for free and I am not flaming the net people,
>>they have done a good job so far. It is not yet finished however. This is
>>one of the things which should be given a very high priority.
>
>It works. What's the problem? Just because it doesn't have the same size
>buffers as other O.S. doesn't mean it's broken. By that reasoning then
>the Shareware DOS NFS client I'm testing now is broken because it only
>has 1K buffers (actually it is broken, long story).
...
>But back to the questions: why is Linux NFS limited to 1K buffers? How
>difficult would it be to fix?
>
First you say it is not broken, then you ask how hard it would be to _fix_.
A slight contradiction there.
It should be easy to get the buffers up to almost 4k, trivial in fact.
After that you'd need to hack kmalloc, use vmalloc, or have the net code
use 2 buffers per large packet or somesuch.
P.S. Is calling vmalloc from an interrupt bad?
------------------------------
** 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
******************************