Linux-Development-Sys Digest #810, Volume #6 Thu, 10 Jun 99 13:14:26 EDT
Contents:
Re: Kernel and Modules ("David Bell")
Re: SMP Problem (Karsten Hartmann)
Re: gcc: Internal compiler error: program cc1 got fatal signal 11 (Villy Kruse)
Re: Linux & Cybercafe (David Knight)
porting from NT to LINUX (filippo sartori)
Controlling a external LCD Display over parallel port (Richard Rohan)
Re: porting from NT to LINUX (J.H.M. Dassen (Ray))
Re: Strange MO drive behavior (Dr H. T. Leung)
Re: /etc/inittab suggestion for all distributions (Abdullah Ramazanoglu)
Re: the ultimate OS (Donal K. Fellows)
Re: IDE problem (UDMA) (Robert Wolter)
Re: Red Hat 6.0 not fscking the root partition! (Nelson Minar)
Re: Controlling a external LCD Display over parallel port (Medical Electronics Lab)
Filesystems: Get full pathname from I-node ?? ("ELSID Software Systems LTD.")
Testing a kernel, boot from floppy? (Medical Electronics Lab)
Re: Configuration Manager for Linux (Leslie Mikesell)
Re: How to understand /proc/net/dev (Paul Kimoto)
Re: Configuration Manager for Linux (Jonathan Abbey)
----------------------------------------------------------------------------
From: "David Bell" <[EMAIL PROTECTED]>
Crossposted-To: hp.os.linux
Subject: Re: Kernel and Modules
Date: Wed, 9 Jun 1999 14:33:26 -0700
>On Wed, 9 Jun 1999 10:57:16 -0700, David Bell <[EMAIL PROTECTED]> wrote:
>>Does the kernel (and probably depmod) use the lib/modules/2.2.5-15smp
>>directory when I boot the vmlinuz-2.2.5-15smp kernel? If so, how does it
>>know to do that?
Sylvan Butler wrote:
>Check the output of 'uname -r' when running under each kernel.
Alas... I would if I could. I can't seem to get booted SMP! This is what
has given rise to my questions.
David
------------------------------
From: Karsten Hartmann <[EMAIL PROTECTED]>
Subject: Re: SMP Problem
Date: 9 Jun 1999 17:31:19 GMT
some days ago i read a similar article in a mailing list
Someone gave a tip to use the tmscsim (Tekram DC390/Generic AM53C974)
driver.
Hope this helps
David Bell wrote:
>
> I am trying to run SMP Linux on an older PC that has an integrated AMD
> 53C974 SCSI controller. The system boots fine on a uni-processor (UP)
> kernel, but has trouble booting the SMP kernel. The trouble seems to be
> related to the SCSI driver. The basic question is: what could be
different
> about the way which the driver is being called by the UP kernel vs. the
> SMP kernel. I know that the driver is old and not maintained, but it is
my
> understatnding that the drivers are called in a single threaded manner so
> the driver should not need to be aware of the UP or SMP situation. Here
is
> some info from the boot:
>
> <detect the SCSI disks -- all looks fine>
> <other stuff>
> Trying to unmount old root ... okay
> Freeing unused kernel memory: 68k freed
> scsi: aborting command due to timeout : pid 67, scsi0, channel 0, id 1,
lun
> 0 Read (6), 07ae ed 80 00
> SCSI host 0 abort (pid 67) timed out - resetting
> SCSI bus is being reset for host 0 channel 0
> AM53C974_reset called
> AM53C974 register dump:
> IO base: 0xff00; CTCREG: 0x0000; CMDREG: 0x90; STATREG: 0x11; ISREG: 0xc4
> CFIREG: 0x80; CNTLREG1-4: 0x57; 0x40; 0x18; 0x44
> DMACMD: 0xc3; DMASTC: 0x0400; CMASPA: 0x22a000
> DMAWBC: 0x0000; DMAWAC: 0x22a400; DMASTATUS: 0x08
>
>
================== Posted via SearchLinux ==================
http://www.searchlinux.com
------------------------------
From: [EMAIL PROTECTED] (Villy Kruse)
Subject: Re: gcc: Internal compiler error: program cc1 got fatal signal 11
Date: 9 Jun 1999 19:32:26 +0200
In article <7jm5b1$ki6$[EMAIL PROTECTED]>,
Jonathan Stott <[EMAIL PROTECTED]> wrote:
>If it *always* happens on the exact same file, my suggestion is to
>ignore everyone screaming "memory error" and try compiling the problem
>file without optimization. If the file compiles without optimization
>and segfaults every time with it, I would assume that the compiler is
>at fault.
Would you get sig 11 if you run out of both memory and swap space?
Villy
------------------------------
From: David Knight <[EMAIL PROTECTED]>
Crossposted-To: alt.uu.comp.os.linux.questions,comp.os.linux.networking
Subject: Re: Linux & Cybercafe
Date: 09 Jun 1999 15:19:42 +0100
Maurice Kemmann <[EMAIL PROTECTED]> writes:
> MicroNg schrieb:
> > however, how to limit the access that so the user can only access the
> > browser but NOT any other
> > program ? ( to access other program, for eg for the webmaster to shutdown
> > the computer, an passcode
> > is required).
> You can configure the windowmanager without any xterm. So the users are
> not able to start any other program. On the desktop you offer only the
> browser. Login should be via xdm !
> That's it i think !
That it itself probably wont work, just put in netscape
telnet://localhost
and it will open an xterm running the telnet session, from which they
would then be able to start other programs.
David
--
The superior man understands what is right;
the inferior man understands what will sell.
-- Confucius
------------------------------
From: filippo sartori <[EMAIL PROTECTED]>
Subject: porting from NT to LINUX
Date: Thu, 10 Jun 1999 12:21:33 +0100
HI!
I am new to linux.
I have to port several NT and OS/2 applications.
Fortunately they are all based on a common portability library.
So I have to port that library and all will be well.
The problem is that these applications are semi real time text mode.
They use threads heavily, and make use of shared mutex sempahores and
shared event semaphores.
Is there any way to get the same functionality under linux?
Do I have to extend the kernel?
Honestly having many of the kernel32 api available on LINUX would really
be
a bonus for the SO.
If I manage to do this, then it is likely that we will drop NT were it
belong to...
"the toilet"
Thanks for any help.
Please mail me, since I do not read the news every day.
Filippo
------------------------------
From: Richard Rohan <[EMAIL PROTECTED]>
Subject: Controlling a external LCD Display over parallel port
Date: Thu, 10 Jun 1999 13:32:02 +0200
hi there !!!
hope to find some answers or -better- allready existing code for
controlling an lcd display on the parallel port via a device driver ...
some requierments:
1. 2 or more applications must have access to the display (but it should
not replace the vga monitor)
2. the display size should but variable
any ideas ???
Best Regards,
Richard Rohan
------------------------------
From: [EMAIL PROTECTED] (J.H.M. Dassen (Ray))
Subject: Re: porting from NT to LINUX
Date: 10 Jun 1999 12:12:19 GMT
[F'up set]
filippo sartori <[EMAIL PROTECTED]> wrote:
>The problem is that these applications are semi real time text mode. They
>use threads heavily, and make use of shared mutex sempahores and shared
>event semaphores.
>Is there any way to get the same functionality under linux?
Linux supports the System V IPC (inter process communication); see
<URL:http://dwww.jimpick.com/cgi-bin/info2www?(ipc)>.
Current Linux distributions ship with the GNU C Library and its LinuxThreads
addon (see <URL:http://www.numeric-quest.com/lang/multithreading.html>),
which allows you to use kernel-level threads via the POSIX 1003.1c thread
standard.
HTH,
Ray
--
PATRIOTISM A great British writer once said that if he had to choose
between betraying his country and betraying a friend he hoped he would
have the decency to betray his country.
- The Hipcrime Vocab by Chad C. Mulligan
------------------------------
From: [EMAIL PROTECTED] (Dr H. T. Leung)
Subject: Re: Strange MO drive behavior
Date: 10 Jun 1999 14:02:20 GMT
If you looked at recent or not so recent posts about MO drive/MO discs, you would
learn that in kernel 2.2.x (as compared to 2.0.x or late 2.1.x)
for removable media, they changed the partition table
size to 2048 for 2048byte/sec media rather than the usual 512 for the usual
512byte/sec medias (besically every writable media else); I think the
recommendation was to use a new filesystem for whole disk without a partition
table. I don't have the details, (as I generally only do a whole new
filesystem on the whole device) , but you might choose to search older posts on
this topic.
In article <[EMAIL PROTECTED]>, Safuat Hamdy
<[EMAIL PROTECTED]> writes:
|> Hi,
|>
|>
|> with recent Linux versions (see below) my MO-drive shows up a strange behavior:
|> First I created a new filesystem (ext2) on a whole MO-disk (/dev/dsk/sdc), no
|problems.
|> Then I created some partitions on the disk and tried to build a new filesystem on a
|partition
|> (/dev/dsk/sdc1). I got the following messages
|>
|> Jun 9 21:48:00 euler kernel: SCSI disk error : host 0 channel 0 id 2 lun 0
|return code = 28000002
|> Jun 9 21:48:00 euler kernel: extra data not valid Current error sd08:21: sense
|key Illegal Request
|>
|> Jun 9 21:48:00 euler kernel: Additional sense indicates Logical block address
|out of range
|> Jun 9 21:48:00 euler kernel: scsidisk I/O error: dev 08:21, sector 1245448
|> [more than 10000 times more]
|>
|> Moreover, when I mounted the partition, I found in the syslog
|>
|> Jun 9 22:25:17 euler kernel: 3)): ext2_check_blocks_bitmap: Block #373 of the
|inode table in group 0 is marked free
|> Jun 9 22:25:17 euler kernel: EXT2-fs error (device sd(8,33)):
|ext2_check_blocks_bitmap: Block #374 of the inode table in group 0 is marked free
|> Jun 9 22:25:17 euler kernel: EXT2-fs error (device sd(8,33)):
|ext2_check_blocks_bitmap:
|> Block #375 of the inode table in group 0 is marked free
|> [some 1000 times more]
|>
|> intermixed with
|>
|> Jun 9 22:25:19 euler kernel: SCSI disk error : host 0 channel 0 id 2 lun 0
|return code = 28000002
|> Jun 9 22:25:19 euler kernel: extra data not valid Current error sd08:21: sense
|key Illegal Request
|> Jun 9 22:25:19 euler kernel: Additional sense indicates Logical block address
|out of range
|> Jun 9 22:25:19 euler kernel: scsidisk I/O error: dev 08:21, sector 3407884
|> [also some 1000 times more]
|>
|> Even more strange:
|>
|> # cd /mo
|> /mo
|> # df .
|> Filesystem 1k-blocks Used Available Use% Mounted on
|> /dev/dsk/sdc1 2403680 26 2403654 0% /mo
|> # touch foo
|> touch: foo: No space left on device
|>
|> I've not even the foggiest idea, whats going on.
|>
|> Technical data:
|> Linux version: 2.2.9
|> GCC: egcs 1.1.2
|> libc: glibc-2.1.1
|>
|> --
|>
|> S. Hamdy | All primes are odd except 2,
|> [EMAIL PROTECTED] | which is the oddest of all.
|> |
|> unsolicited commercial e-mail | D.E. Knuth
|> is strictly not welcome |
--
--------------------------------------------------
"What you don't care cannot hurt you." Chap. 7a, AMS-NS
--
--------------------------------------------------
"What you don't care cannot hurt you." Chap. 7a, AMS-NS
------------------------------
From: Abdullah Ramazanoglu <[EMAIL PROTECTED]>
Subject: Re: /etc/inittab suggestion for all distributions
Date: Thu, 10 Jun 1999 20:27:16 +0300
Miquel van Smoorenburg wrote:
> You can ofcourse also boot the machine, interrupt LILO (keep shift pressed),
> and boot the machine with the 'single' option. That will boot into single
> user mode for repairs.
>
> You can even boot the machine with '2' as option, it will boot into
> runlevel 2 directly.
>
Oops, it's news to me, so I accept my ignorance. But, anyway, a
kbrequest entry does no harm and it's a backup way for bringing system
to single user. Just in case LILO is not used and boot interruption is
disabled...
--
Abdullah Ramazanoglu ( aramazanoglu AT demirbank DOT com DOT tr )
------------------------------
From: [EMAIL PROTECTED] (Donal K. Fellows)
Crossposted-To: alt.os.linux,comp.os.linux.advocacy,comp.os.misc,comp.unix.advocacy
Subject: Re: the ultimate OS
Date: 10 Jun 1999 14:24:38 GMT
In article <[EMAIL PROTECTED]>,
Vladimir Z. Nuri <[EMAIL PROTECTED]> wrote:
> good item-per-item analysis of the difficulty of accomplishing the
> vision... with nice focus on what I actually wrote.
I thought about giving you a darned good grilling (which would have
made me popular with the peanut gallery) but decided that seeing as
you had typed in a fair amount and had actually appeared to have put
some thought in, a more reasoned reply was worth it.
> I freely acknowledge
> feasibility is steep. even in a best case scenario, I can't imagine
> the OS happening in sooner than a few years. but you don't really
> object to the inherent feasibility of each of the design goals.
> "they are very hard and not being done right now". ok, I agree.<g>
> that's why they haven't been done yet. the easy stuff has already
> been done. it's not gonna be easy. who said it was gonna be
> easy?<g>
Well you asked pretty much for something right at the very conceptual
pinnacle of software engineering. There are toolsets that are working
towards this (many of the names escape me right now, but the language
Eiffel springs to mind here, as does the B Toolkit) but they require
techniques that are utterly inimicable to the manner of working of
both Microsoft-like consumer software mega-corps[*], and the Linux
coterie of hackers[**]. This is because you have to define the exact
specification of what you are building *first* and then stick to it
precisely unless you have a schema for evolving your specification in
a compatible way. And that is an active research area, since I'm not
sure yet if there is anyone in the world that understands evolving a
component's interface sufficiently well to be able to completely
certain that doing it will not introduce serious problems - I gave you
a URL to someone I know who is working in that field at the moment...
What you asked for is not merely hard, but darn close to impossible.
Trust me on this; I would be very surprised if it becomes even
remotely feasable to start the detailed design of Tao in the next ten
years.
Donal.
[* Where time-to-market matters far more than the removal of anything
other than the absolutely grossest bugs. ]
[** Proving your modifications don't break anything critical[***] is
tough at best. So most hackers don't even think about trying... ]
[*** When you go on to consider non-critical stuff, it gets even worse! ]
--
Donal K. Fellows http://www.cs.man.ac.uk/~fellowsd/ [EMAIL PROTECTED]
-- The small advantage of not having California being part of my country would
be overweighed by having California as a heavily-armed rabid weasel on our
borders. -- David Parsons <o r c @ p e l l . p o r t l a n d . o r . u s>
------------------------------
From: Robert Wolter <[EMAIL PROTECTED]>
Subject: Re: IDE problem (UDMA)
Date: Thu, 10 Jun 1999 16:17:43 +0200
> > hdc: timeout waiting for DMA
> > hdc: irq timeout: status=0x58 { DriveReady SeekComplete DataRequest }
> > hdc: DMA disabled
> > ide1: reset: success
> >
It simple means that your EIDE controller OR/AND your hard disk doesn't support
DMA.
I have the same problem with IBM hard disks and Soyo and A-Trend main boards.
Buggy chipsets I suppose.
------------------------------
From: Nelson Minar <[EMAIL PROTECTED]>
Subject: Re: Red Hat 6.0 not fscking the root partition!
Date: 10 Jun 1999 11:26:59 -0400
Make sure your root filesystem is being mounted read-only on boot (in
/etc/lilo.conf). This is the Redhat 6.0 default. It boots, mounts /
read-only, safely fscks it, then remounts it read-write.
[EMAIL PROTECTED]
. . . . . . . . http://www.media.mit.edu/~nelson/
------------------------------
From: Medical Electronics Lab <[EMAIL PROTECTED]>
Subject: Re: Controlling a external LCD Display over parallel port
Date: Thu, 10 Jun 1999 09:58:24 -0500
Richard Rohan wrote:
> hope to find some answers or -better- allready existing code for
> controlling an lcd display on the parallel port via a device driver ...
>
> some requierments:
> 1. 2 or more applications must have access to the display (but it should
> not replace the vga monitor)
> 2. the display size should but variable
>
> any ideas ???
I'm new to linux, but from an embedded systems view this
shouldn't be too hard. The question is what kind of lcd display?
If you are talking directly to the lcd driver chip, your
device driver is "low level", but if the display has a computer
built in and you can make it look like a printer then life is
much simpler.
Patience, persistence, truth,
Dr. mike
------------------------------
From: "ELSID Software Systems LTD." <[EMAIL PROTECTED]>
Subject: Filesystems: Get full pathname from I-node ??
Date: Thu, 10 Jun 1999 10:58:00 -0400
Greetings:
Does any one know of a method of finding the complete pathname of a file
given
the i-node of that file.
>From what I have read it appears that the file system structures are
only set up
to get the I-node given a pathname but not the reverse.
Regards
Robert
[EMAIL PROTECTED]
------------------------------
From: Medical Electronics Lab <[EMAIL PROTECTED]>
Subject: Testing a kernel, boot from floppy?
Date: Thu, 10 Jun 1999 10:12:14 -0500
I got a zImage from the make sequence,
and following the "howto" I finally got a
floppy that begins to boot. However,
I'm obviously screwing something up because
it gives "Error 0x10" and starts to
reboot the floppy, causing and infinite loop
which is only halted with a power off.
I think I need to edit the lilo.conf file,
but I'm not sure what the correct magic
invocations are. I'd like it to load the
kernel from the floppy, then use the file
system on the hard drive. That way, if
the kernel is screwed, I can yank out the
floppy and reboot the original system on the
hard drive and re-edit the new kernel.
The present lilo.conf contains
image=zImage
label=Bootdisk
root=/dev/fd0
Should I change it to root=/dev/hda6 ?
(hda1 is dos)
Thanks!
Patience, persistence, truth,
Dr. mike
------------------------------
From: [EMAIL PROTECTED] (Leslie Mikesell)
Subject: Re: Configuration Manager for Linux
Date: 10 Jun 1999 10:36:46 -0500
In article <7jmdln$[EMAIL PROTECTED]>,
Jonathan Abbey <[EMAIL PROTECTED]> wrote:
>
>This is where Ganymede really wins.. having a flexible central
>authority for user management that can accept change requests from a
>large number of sources and pass it through a common, custom-coded
>process to fan out to LDAP and everything else makes all of that
>very manageable.
I guess one thing that made me put off setting up Ganymede to test
it was the idea of yet-another-database. I already have several
flavors of sql servers and ldap each with their own concepts of
multi-platform clients and access controls. Wouldn't it have
been possible to use one of these instead of inventing another
one? The idea is supposed to be to avoid having to understand
and manage so much different stuff - or at least that's what
I'm looking for.
Les Mikesell
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Paul Kimoto)
Subject: Re: How to understand /proc/net/dev
Date: 9 Jun 1999 22:23:11 -0500
Reply-To: [EMAIL PROTECTED]
In article <[EMAIL PROTECTED]>,
[EMAIL PROTECTED] wrote:
> can someone tell me where to obtain detailed information (maybe
> development guide) to understand /proc/net/dev ?
Have you looked at the man page for proc(5)?
--
Paul Kimoto <[EMAIL PROTECTED]>
------------------------------
From: [EMAIL PROTECTED] (Jonathan Abbey)
Subject: Re: Configuration Manager for Linux
Date: 10 Jun 1999 11:14:15 -0500
In article <7jom2e$3155$[EMAIL PROTECTED]>,
Leslie Mikesell <[EMAIL PROTECTED]> wrote:
|
| I guess one thing that made me put off setting up Ganymede to test
| it was the idea of yet-another-database. I already have several
| flavors of sql servers and ldap each with their own concepts of
| multi-platform clients and access controls. Wouldn't it have
| been possible to use one of these instead of inventing another
| one? The idea is supposed to be to avoid having to understand
| and manage so much different stuff - or at least that's what
| I'm looking for.
I had several goals in mind that were incompatible with using an
external SQL/LDAP database.
1) I wanted Ganymede to be as independent as possible from other
software, so that it would be easy to download and install, and easy
to distribute. If Ganymede had depended on a particular flavor of
commercial database, that might have rendered it impractical as an
open source tool.
2) I wanted Ganymede to have an easily evolvable schema. I spent six
months doing design work on pen and paper before starting to implement
Ganymede. One of the sketches I did was to lay out what Ganymede
would look like on an SQL database. Just replicating the data and
logic that GASH held would have required 15 tables, and it was very
unclear how I could allow the end-user to easily add new types of
objects, fields, and so on. Doing the object-relational mapping on
the fly seemed prohibitively difficult, particularly since I was doing
the design back in early '96, before things like Java Blend and other
object-relational mapping technologies had been announced.
An object database might have gotten around that problem, but again I
didn't want Ganymede to require a for-cost database. Object databases
of any kind, and especially free ones, were thin on the ground in 96,
and generally didn't support the sophisticated transactions, locking,
and journalling I wanted.
3) I wanted Ganymede to have absolute control over the database, so
that I could be completely assured of data integrity. I didn't want
it to be possible for someone to skirt all of Ganymede's
data-integrity logic, nor to have to worry about transactions getting
messed up by outside access. The whole value of Ganymede is that it
is a central point for filtering and reacting to changes.. if changes
could be made behind its back, a lot of the value would be lost.
GASH had proved very capable of maintaining logical relationships, as
long as all changes went through GASH. GASH had difficulties when
someone went in and hand-edited its text files, breaking constraints,
and I wanted to avoid that problem with Ganymede.
4) LDAP wasn't nearly as prominent back in 96, or else I would have
taken a harder look at using it. From what I could see, LDAP didn't
seem to be a particularly natural solution for doing DNS management,
which was one of the functions of GASH and something that would be
essential to properly and flexibly support in Ganymede.
Also, from what I've seen of actual production LDAP schema
definitions, things like vector fields tend to be rather messy, with a
lot of separate entities required, which again would make it more
difficult for users to be able to easily edit the schema. The issue
of changes being made outside of Ganymede was also an issue.
| Les Mikesell
| [EMAIL PROTECTED]
--
===============================================================================
Jonathan Abbey [EMAIL PROTECTED]
Applied Research Laboratories The University of Texas at Austin
===============================================================================
------------------------------
** 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
******************************
