Linux-Development-Sys Digest #762, Volume #6 Tue, 1 Jun 99 13:13:47 EDT
Contents:
Re: Reliable (!) nic for 2.2 kernel? (bryan)
*Very* large physical buffers. (Marius Vollmer)
Re: Still need help with Dell PowerEdge SP 5166-2 ("Laurent Jacquot")
Re: Terabite Plus Filesystems (Bob Hoekstra)
Re: device driver in Kernel 2.2.9 (Gerd Rausch)
Re: uid and gid assignments as distributed (H. Peter Anvin)
how to compile and make ("Chris Fanning")
Re: Terratec Troubles (Gerd Rausch)
Anyone know if Linux supports the IBM 6157 tape on PS/2-MCA? (Bob Keys)
Re: You can earn $50,000 40686 ("Kent Dahl")
Reliable MGETTY Conections NOT ? (Charles Connolly)
Re: Terabite Plus Filesystems (Bob Hoekstra)
----------------------------------------------------------------------------
From: bryan <[EMAIL PROTECTED]>
Subject: Re: Reliable (!) nic for 2.2 kernel?
Crossposted-To: comp.os.linux.networking
Date: Tue, 01 Jun 1999 10:36:33 GMT
In comp.os.linux.development.system H. Peter Anvin <[EMAIL PROTECTED]> wrote:
: Followup to: <[EMAIL PROTECTED]>
: By author: Johan Kullstam <[EMAIL PROTECTED]>
: In newsgroup: comp.os.linux.development.system
: >
: > > : >gear at a computer show, i was advised to get a hub for the fast
: > > : >ethernet.
: > >
: > > computer SHOW. meaning SALES? I wonder if that had anything to do
: > > with it? ;-)
: >
: > it didn't. the computer show had many vendors.
: >
: > one of my friends from work was working there and sold me a system
: > (mobo case cpu mem). he then pointed me at the card guy from whom i
: > bought a scsi adapter and a couple of tulips. the friend advised me
: > to get a hub too. he didn't sell hubs so had no interest in selling
: > me more stuff than i would need.
: >
: There *is* a problem with using crossover with 100Base-TX; not that it
: can't be overcome: virtually all 100Base-TX cards on the market are
: autodetecting 10/100 and half/full duplex. For many cards, the
: autodetect algorithms will fail if used over a crossover cable,
: causing it to fall down to 10/half when it should be able to do
: 100/full over that arrangement.
obviously that's not the cable's fault. its the [by definition]
broken auto-duplex alg. when one end is fixed, auto-neg seems to
work, mostly. its when both are sitting there saying 'well I can do
both, what do you want to do?' and not ever deciding right ;-)
: The solution is to force the media type on at least one side of the
: cable.
right. that's the pragmatic thing to do. have to do that with some
switches, too!
: How to do that is driver-dependent;
that's what I hate. it should be standard in IFCONFIG. irix has an
ifconfig switch to set speed and duplex. why doesn't linux? when I
proposed this to some linux kernel and networking folks, I got totally
ignored. not sure why, really - its overdue that linux should have a
standard way of setting this and ifconfig is the proper place for it.
: most Linux drivers will
: let you do that by passing module arguments. Probably ought to be
: done via /proc/sys or ifconfig or something similar, but currently no
: generic or runtime interface exists.
maybe I'll write it. even if its not acceptable to some kernel folks,
at least 2 people think this is the right place for it (ifconfig).. ;-)
--
Bryan
------------------------------
From: Marius Vollmer <[EMAIL PROTECTED]>
Subject: *Very* large physical buffers.
Date: 01 Jun 1999 12:38:12 +0200
Hi,
I have a generic question about virtual memory management and how it
relates to large physical buffers for DMA. It's not really Linux
specific, but I figure that you guys here know this stuff.
For convenience, we would like to have a 8MB large DMA buffer in
physical memory (that is, it should have continues physical
addresses). Our WindowsNT box wont give it to us, of course.
Now, I guess this is due to the fragmentation of the physical address
space. But, as long as there is no fragmentation or memory shortage
due to competing requests for *physical* addresses, the system should
be able to simply reshuffle its mapping of physical pages to virtual
addresses to make room for the large physical buffer.
That is, the very mechanisms of the virtual memory machine should make
it possible to fabricate large continues areas of physical addresses.
Does that make sense? How does Linux go about allocating physical
memory for DMA transfers?
thanks,
Marius
------------------------------
From: "Laurent Jacquot" <[EMAIL PROTECTED]>
Subject: Re: Still need help with Dell PowerEdge SP 5166-2
Date: Tue, 1 Jun 1999 12:33:40 +0100
It happens to too,
i used this work arround to install RH5.2
Unfortunately, you must have dos on your box.
boot on dos then put your RH cd in your cdrom
d:
cd dosutils
autoboot
hope this helps
laurent
Leslie Smith wrote in message <7ioitr$[EMAIL PROTECTED]>...
>Can help me out here, I have got a Dell PowerEdge server
>SP 5166-2 Dual CPU.
>
>I'm using Red Hat 5.2, when I am using the boot disk to boot
>the machine. It seems to load the boot.img OK, but when it
>tries to load initrd.img the system seems to hang at ... .
>This also happens with Red Hat 5.1 too. I am able to load Red
>Hat 5.0 OK.
>
>Machine Spec:
>
>CPU = 2 x 166
>SCSI Controller = NCR 53c810
>SCSI HD = 4 x 2GB
>SCSI CD = x32
>MEM = 128MB
>Display = 1 x Vodoo 4MB
>
>REgards
>
>Leslie...UK:-)
>
------------------------------
From: Bob Hoekstra <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.ms-windows.nt.admin.misc,comp.sys.sun.admin,comp.sys.hp.misc,comp.os.linux.hardware
Subject: Re: Terabite Plus Filesystems
Date: Tue, 01 Jun 1999 14:53:13 +0200
OK, "Al in Seattle" (isn't that near Redmond?), as you quote mainly from my
posting, I think I'd better answer at least some of the comments:
Al in Seattle wrote:
> I don't see where money is an issue in his original mail.
Read it again.
> Other than the fact that you folks all use Unix based systems that are
> recommending Unix based system, what technical reason are you siting for not
> using an NT based system?
I use many systems, from mainframes to PCs, as do several other regular
readers/contributors. As a professional contractor (I alternate between APL
programming on any platform and Unix system administration) I will work on the
platform of my client's choice and keep quiet about it unless I'm asked. While I
admit that I have a preference for Unix systems, that preference is borne out of
experience, not ignorance.
> Some of the quotes:
> "I feel that if your data is important and you want a file server that comes
> up and stays up, you should discount NT immediately. I have heard some
> horror stories about NT with very large directories " no basis in fact
> here.
I have a dual boot PC, Linux and NT 4 server. I will admit that I do things to
this machine that I would never do to a production machine, but I have never had
problems from Linux. NT has provided regular cause for concern, including 2
total reinstallations, many, many blue screens of death, lost and corrupted
files from NTFS and FAT file systems, and the total inability to keep running
for more than a week without a reboot.
OSses which force users to reboot when they make small changes (e.g. changing
screen resolution) have no place as servers of any sort. Servers are meant to
serve, ideally 24 hours/day, 365.25 days/year.There is no place for instability.
NT is improving, but it is not there yet. At the very least it must protect its
kernel at all times against all eventualities. Any software crash or hiccup must
never, ever bring the entire machine down.
As for the "large directories": all OSses dislike very large (i.e. many files)
directories (excluding those mainframe OSses where the concept of a directory
doesn't exist in the same sense) but NT reacts worse to this than most. Try the
suggestion I made. This is made worse by the fact that MS OSses encourage the
use of GUI tools, in this case Windows Explorer. WE will re-evaluate the
contents of the directory at the drop of a hat, and with many files in a
directory this can take a long time. Why should a server run a GUI front-end
anyway?
> "PCs are just not
> built to the same standard as most of the "real" Unix boxes from Sun, HP,
> IBM, SGI, etc. The one exception that comes to mind would be the Sequent
> range." pure bs. It simply depends on what you are willing to spend.
> Compaq and others have totally capable boxes if you want to spend the same
> kind of money that the Unix crowd delivers.
The companies you mention have some very nice boxes. They also come at very nice
prices. I would rather have a small box from Sun on my desk than a more
expensive one from "Compaq and others", as the cost is less and the depreciation
much less. As an example, I have a Sun SPARCstation 2 at home, which hasn't been
powered down in the last 5 years except when I moved house recently. This
machine does important work in my network reliably. It dates from around the
same time as the 80386 chip. How many '386 (or early '486) PCs are still going?
The SS2 was expensive compared to PCs of its day, but the extra expense was a
sound investment. I am still amazed when opening a modern, high priced PC at
just how badly they are put together. Loose cables dangle everywhere across a
badly fitting motherboard without a thought given to air circulation around
sometimes quite expensive components. On the SS2 the few cables are routed
neatly around the edge of the board and components (even later add-ins) are of a
consistently high quality. Modern Unix machines are even better. They are built
to last.
Getting back to the original question: by anybody's standards, 2TB is a lot of
data. The poster did not say what it was or what it cost, but that amount of
data presumably has a price attached to it. Using cheap hardware and unreliable
software to maintain it does not make sense.
The contents of this message express only the sender's opinion.
This message does not necessarily reflect the policy or views of
my employer, Merck & Co., Inc. All responsibility for the statements
made in this Usenet posting resides solely and completely with the
sender.
------------------------------
From: Gerd Rausch <[EMAIL PROTECTED]>
Subject: Re: device driver in Kernel 2.2.9
Date: 01 Jun 1999 15:08:28 +0200
>>>>> "±èÇü¼®" == ±èÇü¼® <[EMAIL PROTECTED]> writes:
±èÇü¼®> I make a device driver in kernel 2.0.36 . But after I install kernel
±èÇü¼®> 2.2.9,
±èÇü¼®> it is not loaded to kernel!! error message is followed:
±èÇü¼®> % insmod fsex.o
±èÇü¼®> So, I change memcpy_xxfs to copy_xxx_user, but same...
Uhhh, strange concept.
Really, the memcpy_fromfs and memcpy_tofs were using the FS segment
register in the old days, and have been accessing user space that way.
The semantics of copy_{from,to}_user is slightly different in
Linux-2.2. It's using the memory manager to catch the condition when
you're trying to access a bad memory area, and make the function
return an error code.
That way you can save an additional verify_area call.
So you shouldn't be replacing memcpy_{from,to}fs with
copy_{from,to}_user, as they do have a different meaning !
You can replace memcpy_{from,to}fs with a simple memcpy in Linux-2.2,
provided you're 100% sure that the memory area is accessible
(readable/writable, depending on what you're doing).
Anyway, a transition from 2.0 -> 2.2 Linux kernel usually requires
slightly more work than just replacing these functions, so you better
get in contact with the author of that "fsex" module.
If you're the author by yourself, compare some old kernel (2.0) with
newer ones (2.2) and find out what needs to be changed the hard way
;-)
Regards,
Gerd
--
voice: +49-2407-575-353
email: [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (H. Peter Anvin)
Subject: Re: uid and gid assignments as distributed
Date: 1 Jun 1999 11:09:48 GMT
Reply-To: [EMAIL PROTECTED] (H. Peter Anvin)
Followup to: <[EMAIL PROTECTED]>
By author: Glen Turner <[EMAIL PROTECTED]>
In newsgroup: comp.os.linux.development.system
>
> "H. Peter Anvin" wrote:
> >
> > Followup to: <j%i43.2964$[EMAIL PROTECTED]>
> > By author: [EMAIL PROTECTED] (Phil Howard)
> > In newsgroup: comp.os.linux.development.system
> > >
> > > Again, I upgrade, and again, uid and/or gid number conflicts.
> > >
> > > Is there a standard anywhere that indicates what range of uid/gid
> > > numbers a distribution/vendor may use, vs. which ones I may use
> > > (any administrator) such that I can expect in the future there will
> > > never be any conflicts?
> > >
> >
> > Yes, 0-99 are assigned by the distribution (with 0, of course, being root).
>
> Which was my understanding, then RH6.0 installed xfs as
> UID 100 <sigh>
>
> ASAIK there is *no* standard for system UID and GID usage.
>
> We have Linux, FreeBSD, DEC, Sun and HP gear and finding
> safe UID ranges is getting so difficult as to be silly.
>
Seems like an off-by-one error on their part. This probably is
something for the LSB/FHS to take up.
-hpa
--
"The user's computer downloads the ActiveX code and simulates a 'Blue
Screen' crash, a generally benign event most users are familiar with
and that would not necessarily arouse suspicions."
-- Security exploit description on http://www.zks.net/p3/how.asp
------------------------------
From: "Chris Fanning" <[EMAIL PROTECTED]>
Subject: how to compile and make
Date: Tue, 1 Jun 1999 13:51:54 +0200
Hi all,
I'm new at this.
Can you please tell me where I can get some info on how to
Compile and make
I'm at a loss and it seems that it's just something I can't ignore.
Thanks in advance.
Chris.
------------------------------
From: Gerd Rausch <[EMAIL PROTECTED]>
Subject: Re: Terratec Troubles
Date: 01 Jun 1999 15:12:21 +0200
>>>>> "Adam" == Adam Langley <[EMAIL PROTECTED]> writes:
Adam> Could anyone tell me how to get my EWS64 working in Linux. I know it runs
Adam> with the commercial version of oss but I don't relly want to have to pay
Adam> just to get sound on what is currentley my second OS.
Adam> Email me on : [EMAIL PROTECTED]
Depends wether you own an EWS64XL or an EWS64S.
Terratec doesn't hand out any information about their internal routing add-on's.
They require to sign NDA's and distribute the modules as binary only.
Read more about that topic at:
http://www.anime.net/~sam9407/
(Download the sam9407 device driver and browse through the README file).
Regards,
Gerd
--
voice: +49-2407-575-353
email: [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Bob Keys)
Subject: Anyone know if Linux supports the IBM 6157 tape on PS/2-MCA?
Date: 1 Jun 1999 13:47:56 GMT
Does anyone know for sure if Linux on a PS/2 supports the IBM 6157
tape drive system (Cypher 1/4 inch streamer)?
Thanks
Bob Keys
------------------------------
From: "Kent Dahl" <[EMAIL PROTECTED]>
Crossposted-To:
comp.lang.java.programmer,comp.lang.javascript,comp.lang.perl.misc,comp.lang.tcl,comp.mail.eudora.ms-windows,comp.os.linux.advocacy,comp.os.linux.hardware,comp.os.linux.misc,comp.os.linux.networking
Subject: Re: You can earn $50,000 40686
Date: Tue, 1 Jun 1999 17:09:40 +0200
Jochem Huhmann skrev i meldingen <7j0s5m$hi8$[EMAIL PROTECTED]>...
>(I don't recommend mail bombing, it does more harm to others than to
>it's target)
I wasn't suggesting bombing as such (repeated large e-mails), but a single
one from me, a single one from you... makes one big whale on their account,
and doesn't block the line I'm on too much.
Students in Norway protested against... uh, what was it again? Can't
remember... Anyway, we didn't "bomb" the government representatives, but
someone make a script that would send a generic letter, with the name you
wrote, to a dousin representatives. Small thing, but a lot of students did
it...
I do not call this bombing, but the communications company just dropped the
whole bunch of e-mails. Sort of the postal service burning postcards if the
sack gets too heavy. *grrr*
Democracy my butt-cheeks!
--
// =========================================================
/** @author Kent Dahl - stud.techn.; ind.øk. data, 1. år
* @url http://www.bigfoot.com/~MenThal/
* { Suicide in the cyberpunk era;
* run a Microsoft product
* on your cyberspace deck. } */
------------------------------
From: [EMAIL PROTECTED] (Charles Connolly)
Subject: Reliable MGETTY Conections NOT ?
Date: Tue, 1 Jun 1999 15:17:30 GMT
I am having problems with loging on to my linux system from a
terminal.
The login propmt is presented and I can type the login name fine
but after typing password the system often just does nothing.
sumtimes the shell I have initated the MGETTY from on the consol
gets closed at this time.
Sometimes I am able to get on ok.
The problem may be more with login then AGETTY I think an alternative
login can be invoked but I do not know of one. Could It Just be bash
or mc if you are not worried about security ? Or does a LOGIN
program need to satisfy certian criteria.
--
- Charles.
------------------------------
From: Bob Hoekstra <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.ms-windows.nt.admin.misc,comp.sys.sun.admin,comp.sys.hp.misc,comp.os.linux.hardware
Subject: Re: Terabite Plus Filesystems
Date: Tue, 01 Jun 1999 17:52:17 +0200
Al in Seattle wrote:
> Some comments below...
> David T. Blake wrote in message ...
>
> >We regularly find 2-3 times the sys admin manpower required
> >for NT boxes compared to UNIX boxes in the same setting. And
> >I've had more than enough experience with the blue screen of
> >death.
>
> Whatever. I haven't seen this issue from the ISPs I've chatted with, except
> for the fact that they are Unix trained and are trying to support a non-Unix
> OS. I would bet that the reverse would hold true also.
No, it's not. Unix boxes are easier to maintain in large numbers. When problems
occur, they are easier to fix, even for fledgling sysadmins.
> There are a number of core problems with NTs usability.
>
> >Such as,
> >1) Want to put in a new video card - reboot about 10 times.
>
> I have not experienced this in 5 years working with NT.
No? You've never had chains of interrupt conflicts? Magical changes due to
Plug&Pray? Video drivers that are supposed to be compatible but aren't? Another
reboot demanded because you decided to change the screen resolution? And then
the resolution didn't work (even though it was fine when you tested it) and you
get horrible lines and you can't see anything and your only choice is to switch
off? An the next reboot takes 4 hours while scandisk repairs your trashed NTFS
filesystem and the 4 FAT filesystems? And you fall asleep before you get to the
promt where you can choose the VGA mode so you go straight back to the squiggly
lines?....
> >2) Video server built in to the kernel
> >3) Want to extend functionality - send a check for $10k to
> > Redmond
>
> What is this supposed to mean?
NT sucks. M$ sucks. When you want it to do something properly, you have to pay
them money. It's what BG calls good business.
> >4) Want to program - send another check to Redmond - one
> > for each language
>
> What is this supposed to mean, compared to Unix?
You can't install free languages. Those that exist are hardly worth the prices
you pay for them. Many Unixes come with C compilers, and Gnu C is available in
open source. As are many other languages.> >5) Having problems with the OS - too
bad. Call Redmond and
> > pay out the nose while you wait on hold, and then talk to
> > someone who knows horribly less than you about the OS.
> >6) Remote administration
>
> We routinely support our NT servers in our web farm remotely. there are a
> number of tools out there to do that. You are rapidly losing all credibility
> on this one.
Yep, but the "tools out there" are weak compared to the abilities that come with
every version of Unix.
> >7) Lack of a respectable scripting language for administration
> > purposes
>
> Other than Perl or it's built in Scripting language?
Perl is a scripting languages, but it is available for NT. I rather fancy he
means the ability to write shell scripts. Fortunately, a Korn shell-like program
is available for NT. However NT's ability to call the written scripts at
predetermined times is weak compare to Unix.
> >PCs with *x86 architecture have about 1/3 the computing power
> >of an alpha at the same clock speed. That is the penalty for
> >keeping legacy chip architecture around.
> >
>
> As mentioned in other posts, you are losing all credibility here with
> statements like that. There are not benchmarks that I have ever seen
> pointing to that big a difference.
I'm not sure how big the difference is, but I do know it exists. The Alpha is a
great chip. Try http://www.sgi.com/t3e/ to see...
> >I understand the Compaq XP-2000 is a quite capable box. If you
> >don't have the $$ for that you can try the DS-10 from Compaq for
> >about $3500 + $1200 or so for Tru64Unix (or linux for free).
> >
> >--
> >Dave Blake
> >[EMAIL PROTECTED]
The contents of this message express only the sender's opinion.
This message does not necessarily reflect the policy or views of
my employer, Merck & Co., Inc. All responsibility for the statements
made in this Usenet posting resides solely and completely with the
sender.
------------------------------
** 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.development.system) 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-Development-System Digest
******************************
