Linux-Misc Digest #937, Volume #20 Tue, 6 Jul 99 00:13:08 EDT
Contents:
Is there a variant of telnetd which logs in users? (Bernie)
Re: Linux vs Solaris (Dave)
Re: Help with X: can a user detect mouse activity on a remote X console? (Christos
Siopis)
Re: ApplixWare (Carl Fink)
Re: Linux vs Solaris (DeAnn Iwan)
Re: Need Iomega Ditto 3200 (3.2G) Tape Drive info (Scott Alfter)
Re: Linux vs Solaris (Cameron L. Spitzer)
Setting IRQ and making it stick ("Joseph S. White")
Any bookmark manager applications for Linux?? (Basel Shishani)
Re: Resize linux partition w/ Partition Magic!! ("Charles Sullivan")
Re: Linux and Viruses - Not the same old question (Robert Heller)
Re: Help with X: can a user detect mouse activity on a remote X console? (Christos
Siopis)
Re: Linux and Viruses - Not the same old question ([EMAIL PROTECTED])
Re: Running another OS under Linux (Deryk Barker)
Re: Linux and Viruses - Not the same old question (William Burrow)
Re: Solving the 1024 cylinder LILO problem ("Charles Sullivan")
Re: Linux vs Solaris (Dave)
----------------------------------------------------------------------------
From: Bernie <[EMAIL PROTECTED]>
Subject: Is there a variant of telnetd which logs in users?
Date: Mon, 05 Jul 1999 22:15:10 -0400
Is there a variant of telnetd, or a separate program which
will listen on a non-telnet port, login the user, then run
a script or binary? I've looked in the contrib directories,
and cannot find anything like this.
-Thanks
------------------------------
From: [EMAIL PROTECTED] (Dave)
Crossposted-To: comp.unix.solaris
Subject: Re: Linux vs Solaris
Date: Tue, 06 Jul 1999 02:33:08 GMT
>
> Although the artical probably didn't mention the fact that the
> "test" was sponsored by M$, and the compamy doing the testing
> heavily tuned the NT machines and ran the Linux one out of the
> box, with no tuning whatsoever...
>
Actually...read it for yourself.
I'm not talking about the original article ("Heavily tuned NT beats
barely tuned Linux" - or something to that effect)...THIS WAS THE
REMATCH.
"PC Week Labs' tests show what path Linux
must take"
By Henry Baltazar and Pankaj Chowdhry,
PC Week Labs
June 25, 1999 3:45 PM ET
In a first-of-its-kind open benchmark
comparing the performance of Linux
and Windows NT, PC Week Labs not only
found that NT remains substantially faster but
also isolated Linux's shortcomings and gained
insight into where future development of the
open-source operating system should be
headed.
These tests follow the controversial NT/Linux
comparison conducted by independent
benchmarking operation Mindcraft Inc. and
sponsored by Microsoft Corp. The Linux
community had cried foul after this April match,
which found NT to be 400 percent faster than
Linux. All the principals agreed to let PC Week
Labs arbitrate a rematch.
After a tortuous five days of tests, audited by
the best and the brightest from Mindcraft,
Microsoft and Red Hat Software Inc., and
despite significant tuning improvements made
on the Linux side, Windows NT 4.0 still beat
Linux using the Apache Web server and
Samba in every performance category,
although the margin of victory was smaller than
in Mindcraft's tests.
But far more interesting is that, in all the areas
in which the Linux community cried foul, its
assumptions were wrong. Where kernel
problems were found, fixes are already under
way.
For instance, the open-source community
objected to Mindcraft's use of the Apache Web
server in its benchmarks, claiming that using
the fastest open-source Web server, Zeus,
would improve results. We tested Zeus on
Linux and found its performance peaked
almost exactly where Apache's did.
Working with Red Hat programmer Zach
Brown, we traced the problem back to the lack
of a multithreaded IP stack in the Linux
networking subsystem, which caused a
performance plateau in the operating system,
not in the Web server.
The problem is being fixed in the next version
of the Linux kernel, and a beta is available in
the 2.3-kernel series. However, the upcoming
improvements don't stop there. The next
version of Apache will have a new static page
engine, similar to the fast path in NT's IIS
(Internet Information Server).
Zeroing in on the Web server
In the open benchmark using Ziff-Davis
Benchmark Operation's WebBench 2.0, which
Mindcraft used as well, the IIS 4.0 Web server
pumped out an impressive 4,166 requests per
second, compared with 1,842 requests per
second on Red Hat's best run (see top
benchmark chart).
In comparison with Mindcraft's initial report, IIS
4.0's performance was 226 percent better than
Apache, as opposed to the 400 percent
difference Mindcraft found. And where Linux's
performance collapsed after the client load
exceeded 16 computers in Mindcraft's original
Apache benchmark, Linux fared better under
heavier loads in our tests. The discrepancy in
results may stem from the differences between
the Mindcraft and PC Week Labs testbeds.
We also tested both operating systems on a
single-processor box with 256MB of RAM to
evaluate performance on lower-end hardware.
NT still had a performance edge over Linux
(1,863 requests per second compared with
1,314 requests per second in WebBench, for
example). This amounted to a 41 percent
performance difference but showed that, even
on cheaper systems, NT came out ahead.
File server: Also not on top
Running ZDBop's NetBench 5.1 to test file
service performance, we found that Linux
running the Samba file gateway did not top NT
in any test (see bottom benchmark chart).
Linux and Samba's best numbers (which were
achieved while running on NT workstation
clients) were considerably slower than NT's
best numbers (155.9M bps vs. 338.3M bps).
Previously, we had found that Samba could
outperform NT 4.0 using NT workstation clients
(see "NOS crossroads," PC Week's
Shoot-Out of network operating systems). As a
direct result of that Shoot-Out, Microsoft's
performance engineers were prompted to find
ways to enhance file throughput from NT
servers to NT workstation clients.
Upon closer examination of the NTFS (NT File
System), Microsoft found that when the data
volume was configured as a single partition,
the contention for NTFS' transaction log
slowed file transactions significantly.
By breaking up the data volume into four NTFS
partitions (each with its own transaction log),
we were able to distribute transaction entries,
eliminating the bottleneck.
Linux and Samba had their most favorable
performance comparison with NT when we ran
both on a single-processor server with 256MB
of RAM. In this configuration, NT performed 52
percent faster than Linux running Samba
(165.2M bps vs. 108.7M bps).
======
Yup. Linux ain't bad. It's Free. It's Unix. It's flexible...
But it's not (yet) ready for prime time.
(Neither is NT, imho)
Cheers,
Dave
+---------------------+---------------+
|[EMAIL PROTECTED]| mindspring.com|
|[EMAIL PROTECTED]| davegrantier@ |
+---------------------+---------------+
------------------------------
From: Christos Siopis <[EMAIL PROTECTED]>
Crossposted-To: comp.unix.questions,comp.os.linux.x,comp.unix.programmer
Subject: Re: Help with X: can a user detect mouse activity on a remote X console?
Date: Mon, 5 Jul 1999 22:33:14 -0400
On 4 Jul 1999, Helmut Kreft wrote:
> Christos Siopis <[EMAIL PROTECTED]> wrote:
> >What I really want to do: I want to have a process owned by user
> >B that runs continuously in the background and checks whether
> >someone (e.g., user A) is using the console. If yes, the process
> >would suspend (kill -STOP) execution of a CPU-intensive program
> >(also owned by user B) until user A activity ceases, at which
> >point it will reinstate the CPU-intensive job (kill -CONT). The
> >reason I want to do this is to run a process without slowing
> >down or otherwise perturbing user A (no, nice +19 is not an
> >option!)
> >
> Why? Just curious...
You mean why nice +19 is not an option? Three reasons:
1. If a job swaps, it slows down the entire system even at nice +19. It
may also cause other jobs to swap because of memory space it occupies.
2. A job at nice +19 still takes 7-8% of CPU time when competing against a
non-niced job.
3. If "it's their machine" they can ask you to not do things even if the
consequences are mostly perceived rather than real :)
Christos
------------------------------
From: [EMAIL PROTECTED] (Carl Fink)
Subject: Re: ApplixWare
Date: 6 Jul 1999 01:35:34 GMT
Reply-To: [EMAIL PROTECTED]
On Mon, 5 Jul 1999 19:15:55 -0400 AccessNet <[EMAIL PROTECTED]> wrote:
>Can anyone please tell me what is the latest version of Applixware Office
>package?
It took me at least 30 seconds of looking at the Applix web site
(www.applix.com) to find out that it's 4.4.1.
--
Carl Fink [EMAIL PROTECTED]
"This fool wishes to reverse the entire science of astronomy."
-Martin Luther on Copernicus' theory that the Earth orbits the sun
------------------------------
From: [EMAIL PROTECTED] (DeAnn Iwan)
Crossposted-To: comp.unix.solaris
Subject: Re: Linux vs Solaris
Date: Tue, 06 Jul 1999 02:09:57 GMT
On Mon, 5 Jul 1999 07:45:07 +0100, "Paul" <[EMAIL PROTECTED]> wrote:
>I'm building a mission-critical high throughput OLTP application which
>required considerable scalability. I'm trying to choose between Linux and
>Solaris for the operating system.
>
>Does anyone have any views on this matter?
>
>Any help or advice appreciated.
>
>Thanks
>
>Paul
>
>
Both are fine, stable OSes. But Solaris on Sun platforms
can probably outperform Linux on miscellaneous platforms. If you
have time, consider a trial of both. If you have to choose now, talk
to Sun and have them help you set up the hardware and software
you need.
------------------------------
From: [EMAIL PROTECTED] (Scott Alfter)
Crossposted-To:
comp.os.os2.misc,comp.os.os2.setup.misc,comp.os.ms-windows.win95.misc,comp.os.linux.hardware,comp.sys.hp.hardware,comp.sys.ibm.pc.hardware.storage,comp.sys.ibm.pc.hardware.systems
Subject: Re: Need Iomega Ditto 3200 (3.2G) Tape Drive info
Date: 6 Jul 1999 01:59:07 GMT
In article <377a0309$1$naqerl$[EMAIL PROTECTED]>, <[EMAIL PROTECTED]> wrote:
>On 06/29/99 at 10:26 PM, [EMAIL PROTECTED] (Ron Gibson) said:
>>I'm considering getting an Iomega 3200 (3.2 G) floppy connected internal
>>tape backup.
>
>Wow... that's GOTTA be slow.
>
>Why don't you try a DITTOEASY 3200 external Parallel port.
Ugh...parallel is even slower than floppy, and it brings your system to a
crawl while it runs. IDE or SCSI is the only way to fly. (I replaced a
Ditto 3200 a while back with an Exabyte Eagle TR-4i. I think Exabyte has
discontinued the drive, but it's much faster than the Ditto ever was. The
extra capacity (4GB vs. 1.6GB for the Ditto) is also nice. Comparable
drives are available from Seagate and HP.)
(BTW, I had the Ditto Dash 2MB/s accelerator with mine. It helped, but it
was still no match for the new drive.)
_/_
/ v \
(IIGS( Scott Alfter (salfter at (yo no quiero spam) delphi dot com)
\_^_/ http://people.delphi.com/salfter
------------------------------
From: [EMAIL PROTECTED] (Cameron L. Spitzer)
Crossposted-To: comp.unix.solaris
Subject: Re: Linux vs Solaris
Date: 6 Jul 1999 02:31:40 GMT
In article <Pine.GSO.4.10.9907051825460.17267-100000@zen>, Rich Teer wrote:
>On Mon, 5 Jul 1999, Dave wrote:
>
>> I read an article last month where an NT
>> IIS 4.0 web server handily beat Linux running Apache (and later Zeus)
>> web servers in nearly every category. NT! (if you can believe it!)
>
>Although the artical probably didn't mention the fact that the
>"test" was sponsored by M$, and the compamy doing the testing
>heavily tuned the NT machines and ran the Linux one out of the
>box, with no tuning whatsoever...
In the MS-sponsored Mindcraft test, IIS on NT beat Apache on Linux
3:1. _PC WEEK_ repeated the test, and let some Linux professionals
set up the Linux box, and IIS on NT beat it 2:1. Or thereabouts.
Notice, this was a serving pages fast test, not a staying up for
months test or a resisting crackers test. And it reflected recent
improvements in IIS, according to the mag. (_PC WEEK_ is quite
sympathetic to Linux, even ran their Web site on 1.0.9 for the first
year or so they had one. As I recall, they even gave a Red Hat
release "Product of the Year" once and *didn't* give it to the
NT release that came out the same year.) The Linux folks have admitted
they've got some work to do on the TCP/IP stack, which is single threaded
in the test kernel. Look for this situation to be fixed in linux-2.4
(or -3.0).
BTW, I've been running Linux at home since 0.98pl18, and everywhere
else since about 1995. I run Debian GNU/Linux on *my* little Web
servers because it's the best damn OS I've ever seen. I think anyone who
buys a big server and an OS soley on the basis of a couple of serving
pages fast tests is being irresponsible. I'd fire anybody who tried
to put my company's e-commerce up on NT, because of NT's well known
security and stability problems. (Or transfer them to the Department
of Nothing Important.) In an application like that, the hardware cost
is lost in the noise: you double the hardware if that's what it takes,
you don't pick a lighter OS to make the hardware go faster. The big
server benchmark that NT won doesn't make any sense to me, as far as
representing any practical real-world application. It was just a
publicity stunt.
Linux really isn't *finished* yet, as far as many markets are
concerned. It's still got standardization and documentation and
installability problems. Linux advocates would do well to work on
those instead of whining about NT winning contrived benchmarks.
And leave the big-server jobs to FreeBSD (e.g. www.yahoo,com,
ftp.cdrom.com) for now.
Cameron
------------------------------
From: "Joseph S. White" <[EMAIL PROTECTED]>
Subject: Setting IRQ and making it stick
Date: Mon, 05 Jul 1999 19:39:07 -0600
Hi All,
Just loaded Mandrake 6.0 w/KDE 1.1.1, had some trouble getting the modem
working. I did a setserial /dev/cua2 irq 5 and now it
works just fine, but when I reboot it will lose that setting. How do I
make that change permenant? Are there any utilities that come with Linux
to do this kind of thing, or is it best to use the command line?
Thanks
Joe
--
[EMAIL PROTECTED]
[EMAIL PROTECTED]
http://www.nmia.com/~jwhite
------------------------------
Subject: Any bookmark manager applications for Linux??
From: Basel Shishani <[EMAIL PROTECTED]>
Date: 06 Jul 1999 11:45:57 +1000
Hi everyone,
I'm looking for a *desktop* bookmark manager application for Linux
(NOT a *web based service* and not a web server add-on).
Main features I'm looking for are: a tree view (folders and subfolders),
annotations on links and folders, search - including on annotations,
basic editing (move, delete ... etc), and some interaction with a
browser (launching a url, importing from Netscape and IE ... etc).
Is there something close to that under Linux? I did some search and
found lots of such applications for Windows (eg: Compass, Powermarks,
Amigo ...) but none for Linux.
Cheers.
==============
Basel Shishani
------------------------------
From: "Charles Sullivan" <[EMAIL PROTECTED]>
Subject: Re: Resize linux partition w/ Partition Magic!!
Date: Mon, 5 Jul 1999 22:37:03 -0400
If you want to boot with LILO on the hard drive, the kernel must be
entirely below cyl 1024. The usual way of insuring this is to
create a small /boot partition of about 10-15 Meg below this boundary
which contains the kernel and a couple of other small files. The rest
of Linux can be anywhere.
If you can't arrange to have this /boot partition, you can also boot
Linux by running LOADLIN.EXE from a DOS partition. Then the kernel
can reside anywhere, so long as it's accessible by DOS.
Regards,
Charles
Mohamad Termizi wrote in message <[EMAIL PROTECTED]>...
>Hi Charles,
>
>I have run into problem that the partition is more than 1024 in LBA mode.
Then,
>I have the problem of installing the LILO.
>
>Any tricks on how to overcome this problem?
>
>TIA
>
>Charles Sullivan wrote:
>
>> If you just resized a partition and are _sure_ that the kernel still
resides
>> below the 1024 cylinder boundary, then simply rerunning /sbin/lilo (as
root)
>> ought to restore your ability to boot from the HD.
>>
>> Alleyoop Sam wrote in message <[EMAIL PROTECTED]>...
>> >Hi, I am a new linux user, using RH 6.0:
>> >
>> > Recently, I have resized my linux partition with Partition Magic 4.
>> >After that, I can no longer boot into my linux partition without the use
>> >of linux floppy boot disk. I can dual boot my machine before using the
>> >PQ4 to resize my linux partitioin. I have NT installed and use NT OS
>> >loader to daul boot my machine. If I use the NT OS loader to boot into
>> >linux, it hangs on the prompt: "LI" and nothing goes on. Well, my linux
>> >works fine if I use the floppy to boot in!! Any ideas to fix the
>> >problem? Pls
>> >
>> >Thx for you help.
>> >
>
>
>
------------------------------
From: Robert Heller <[EMAIL PROTECTED]>
Subject: Re: Linux and Viruses - Not the same old question
Date: Tue, 06 Jul 1999 02:45:30 GMT
[EMAIL PROTECTED],
In a message on 05 Jul 1999 18:57:53 -0700, wrote :
j> [EMAIL PROTECTED] (Paul Anderson) writes:
j>
j> > Jiim McIntyre <[EMAIL PROTECTED]> writes:
j> >
j> > >I know Linux isn't affected by viruses, but I'm trying to find an
j> > >explanation for this.
j> > >
j> > It's multi-user nature. A regular cannot twiddle with the processes of any
j> > other user in memory, and with proper file permissions cannot alter other
j> > user's files. Also, most binaries(that a virus would want to attack) are
j> > owned by root, so it can't get at them. Also, a program cannot write into
j> > memory that it doesn't own. So even if root where running a virus, it would
j> > not be allowed to alter the memory of other running programs(the kernel would
j> > give it a segfault). TTYL!
j>
j> Could a bad root running program damage hardware in Linux? AFAIK it can't, but
j> I am not certain....and THAT is the most dangerous form of viruses.
j>
No. Only a 'bad' device driver or other kernel module. Merely being
root does not give you any direct access to devices. Note: a device
driver that allowed possibly 'bad' device settings can do this and may
or may not need root to do the ioctl() to do it. Or a possible kernel
bug that stomped on an I/O device in a bad way. I don't believe that
most I/O devices can be *physically* damaged by software (I don't
consider something like 'format C:' physical damage -- the drive still
works -- it just would be devoid of your data). There *is* a class of
I/O software that *can* damage hardware -- X servers. If one lies
badly about ones monitor one can toast it. That is, if you claim your
cheap little junky monitor that can only do 600x800, and you have a
video board that can do better than 1280x1024 and tell Xconfigurator or
XF86Setup that you have some wonderful fancy monitor (when you don't)
with some wonderfully high H freq., (way beyond your monitor safety
limit), and then fire up the X server in some high freq. mode, you can
literally burn-out your monitor. Obviously this requires massive
operator error / stupidity... This was more of a problem in the old
days of Slackware 1.0 / kernel 0.99 or 1.2.13 and very primitive X11
setup, etc. Along with the cheap monitors of that era as well. Unless
up pick up an old VGA monitor at a yard sale, any modern monitor will
support the highest freq. that you would have any reason to what to use.
It is conceivable to have a trojan program modify (assuming being run by
root or lax permissions) /etc/X11/XF86Config to specify a higher H Freq
than is ligit for the monitor. Assumes also that the machine's monitor
is a cheap one or that the user will every use X or use it in such a
high res mode and that the video card can run at such a high freq., etc.
Possible, but rather unlikely.
In terms of normal / common (i.e. desktop) I/O devices, I don't know of
any that can be hurt with mere software -- this usually needs really bad
design. Special lab-type I/O devices are a different issue, given the
nature of what they might be connected to and the usually experimental
nature of the device drivers in question.
--
\/
Robert Heller ||InterNet: [EMAIL PROTECTED]
http://vis-www.cs.umass.edu/~heller || [EMAIL PROTECTED]
http://www.deepsoft.com /\FidoNet: 1:321/153
------------------------------
From: Christos Siopis <[EMAIL PROTECTED]>
Crossposted-To: comp.unix.questions,comp.os.linux.x,comp.unix.programmer
Subject: Re: Help with X: can a user detect mouse activity on a remote X console?
Date: Mon, 5 Jul 1999 22:47:33 -0400
On 5 Jul 1999, Ken Pizzini wrote:
> On Sun, 04 Jul 1999 15:19:03 -0400,
> Christos Siopis <[EMAIL PROTECTED]> wrote:
> >Suppose that user A works on a (Linux, if that matters) console
> >running X. I would like to know whether it is possible for (a
> >different) user B (other than root!) who is remotely logged on
> >to the same computer as user A, to test whether the mouse
> >attached to the console is being used (by user A).
>
> How about a "ls -Llu /dev/mouse" (or whatever the name
> of the mouse device is for that system)?
Wow, yes, this seems to work, thank you! For some reason, it also seems to
sense when the *keyboard* is used --not just the mouse-- which suits my
plans even better :)
Christos
------------------------------
Subject: Re: Linux and Viruses - Not the same old question
From: [EMAIL PROTECTED]
Date: 05 Jul 1999 19:59:03 -0700
Robert Heller <[EMAIL PROTECTED]> writes:
> I don't believe that
> most I/O devices can be *physically* damaged by software
I was thinking more in terms of memory and such. I was told by one guy that
he could just tell the computer to send an electronic pulse down the wrong
side of the ROM and be done with it. Course this was in the days of Apple ][e
so....
I have also heard that DOS viruses can do this kind of thing, having direct
hardware access.
>(I don't
> consider something like 'format C:' physical damage -- the drive still
> works -- it just would be devoid of your data).
Oh neither do I, I am talking hardware 'damage'....like toasting the ROM or
BIOS or maybe a drive, to the point were it needs to be replaced.
------------------------------
From: [EMAIL PROTECTED] (Deryk Barker)
Subject: Re: Running another OS under Linux
Date: 6 Jul 1999 02:29:58 GMT
Andy Backa ([EMAIL PROTECTED]) wrote:
[...]
: Here's my scenario: I want to run Linux as my primary OS, and using
: VMWare/Win98 for Office apps (Office97, Visio, etc...). I also want to be
: able to multi-boot into *pure* Win98 for games.
:
: The big question: do I need to install Win98 once or twice? If I can get away
: with installing it once, then life would be damn sweet. If twice, then I'm
: going to have to do some number crunching to figure out the best way to
: partition this sucker.
I believe you can use a previously-installed OS partition. I haven't
had the hardware to try it yet, so am relying on what I've read. Check
the website for the definitive word.
--
|Deryk Barker, Computer Science Dept. | Music does not have to be understood|
|Camosun College, Victoria, BC, Canada| It has to be listened to. |
|email: [EMAIL PROTECTED] | |
|phone: +1 250 370 4452 | Hermann Scherchen. |
------------------------------
From: [EMAIL PROTECTED] (William Burrow)
Crossposted-To:
comp.os.ms-windows.advocacy,comp.os.msdos.programmer,alt.msdos.programmer
Subject: Re: Linux and Viruses - Not the same old question
Date: 6 Jul 1999 03:06:28 GMT
Reply-To: [EMAIL PROTECTED]
On 5 Jul 1999 21:36:14 GMT,
Cameron L. Spitzer <[EMAIL PROTECTED]> wrote:
>In article <[EMAIL PROTECTED]>, Jiim McIntyre wrote:
>The best example of this automatic execution is MS-DOS' and MacOS'
>way of finding out how a floppy was formatted. The floppy disks in
>common use today may be formatted several different ways, and the
>OS has to find out which way any particular floppy was formatted
>before they can read the directory on the floppy. They do it by
>executing a tiny program found on the very first sector of the floppy!
I suspect you mean Windows9x where you say MSDOS. MSDOS has never
executed code on the floppy to find out what the floppy is.
I doubt that MacOS does this either, though it will certainly
automatically execute things from the floppy.
These things are just too easily specified to bother putting code to do
something similar.
--
William Burrow -- New Brunswick, Canada o
Copyright 1999 William Burrow ~ /\
~ ()>()
------------------------------
From: "Charles Sullivan" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development
Subject: Re: Solving the 1024 cylinder LILO problem
Date: Mon, 5 Jul 1999 22:48:45 -0400
If you've got Linux in a partition which crosses cylinder 1024,
the simplest thing to do is use Partition Magic 4.0 to adjust its lower
boundary upward by 10-15 Meg and create a /boot partition in the
space just freed up. Then you won't have to worry about the kernel
location.
Cameron L. Spitzer wrote in message ...
>
>One of the most frequent FAQ/complaints about LILO is the requirement
>to maintain an entire partition inside the first 1024 cylinders, so
>that BIOS can see all the blocks of a bootimage located there.
>Often, the first 1024 cylinders and then some are occupied by
>an alien file system called "C:" that folks are understandably reluctant
>to adjust.
>
>Microsoft seems to get around this with a special write call, used by
>its SYS and FORMAT utilities, which can write a bootimage where it is
>guaranteed visible to BIOS, even if the whole partition isn't.
>
>I played around with a program to make a BIOS-accessible copy of a file.
>I can write a block, or several blocks, and check whether they are
>BIOS-accessible using the same ioctl( FD, HDIO_GETGEO, &hdprm)
>that lilo uses to build its map.
>
>This works for small files, such as /boot/message or even /boot/lilo-map
>if you try enough times. But the chances of randomly writing a
>350 KB bzImage onto a half-full 2000-cylinder /dev/hda1 and having
>all the blocks land in the first 1023 are astronomically small.
>
>I tried closing and reopening the file to write each block.
>(fopen(filename, "a");.) The problem is, once I've appended the first
>BIOS-inaccessible block onto a file, there seems to be no way to
>remove it. rewind(3)'ing and closing doesn't do anything; the block is
>still there.
>
>All I need to do is adjust the size of a file downward, without moving it.
>That is, throw away its last block.
>Is there any way to do this from user space?
>If not, it seems the only ways around the problem are another ioctl
>or splitting the bzImage into hundreds of 1- or 2-block files that
>can be written independently, and teaching lilo how to concatenate them
>again. And an ioctl that says "this file must be written on BIOS-accesible
>blocks" would be rather messy.
>
>Cameron
>
>
------------------------------
From: [EMAIL PROTECTED] (Dave)
Crossposted-To: comp.unix.solaris
Subject: Re: Linux vs Solaris
Date: Tue, 06 Jul 1999 02:49:19 GMT
I went back and found the website.
Here is the link for the original article,
"PC Week Labs' tests show what path Linux must take"
http://www.zdnet.com/pcweek/stories/news/0,4153,1015266,00.html
Here's the link for a chart of the test results:
"Labs benchmark results: NT vs. Linux"
http://www.zdnet.com/pcweek/stories/jumps/0,4270,2283480,00.html
and, the original "test" that started the brouhaha:
"Highly tuned NT whips barely tuned Linux in Microsoft-backed test"
http://www.zdnet.com/pcweek/stories/news/0,4153,1014383,00.html
cheers,
Dave
+---------------------+---------------+
|[EMAIL PROTECTED]| mindspring.com|
|[EMAIL PROTECTED]| davegrantier@ |
+---------------------+---------------+
------------------------------
** 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: [EMAIL PROTECTED]
You can send mail to the entire list (and comp.os.linux.misc) via:
Internet: [EMAIL PROTECTED]
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Misc Digest
******************************
