Linux-Development-Sys Digest #510, Volume #6 Fri, 19 Mar 99 21:16:05 EST
Contents:
Re: Verifying memory of running process (Bengt Richter)
Wall Street Developer Opportunities ([EMAIL PROTECTED])
Problem with sunrpc when building GLIBC-2.1 (John Stapleton)
Re: Linux e2fs file size limit! (Erik de Castro Lopo)
Re: *BIG* Problem with /etc/rc.d/rc.local (Mike McDonald)
Re: Riva TNT AGP ??? (Peter Samuelson)
pcmcia freezes me to death ("Mario Schrijver")
Re: DMA question ("Ron Reaugh")
Re: Building Linux Shared Libraries ("John Burton")
Re: After Week 1 With Linux -- licking wounds. ([EMAIL PROTECTED])
select & poll ("John Burton")
Re: How do you get drivers for ATI Pro Rage AGP? (Kevin White)
Re: device/inode to filename ([EMAIL PROTECTED])
Re: Device Driver...what for? (ellis)
Re: Help with a C program in Linux (James Youngman)
Re: Device Driver...what for? (ellis)
Re: *BIG* Problem with /etc/rc.d/rc.local (ellis)
Re: Verifying memory of running process (Daniel R. Grayson)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (Bengt Richter)
Subject: Re: Verifying memory of running process
Date: Fri, 19 Mar 1999 22:05:46 GMT
On Fri, 19 Mar 1999 17:50:12 GMT, [EMAIL PROTECTED] wrote:
>On Fri, 19 Mar 1999 04:59:51 GMT, [EMAIL PROTECTED] (Phil Howard)
>wrote:
>
>>On Thu, 18 Mar 1999 22:32:05 GMT [EMAIL PROTECTED] wrote:
>>
>>| For a secure application running under Linux (version 2.X or higher),
>>| I need to be able to verify that the memory of the running process
>>| (executable memory) does not change. We would like to be able to
>>| either checksum or do a binary comparison of the process's memory with
>>| a known value. Has anyone ever done this? Any suggestions on how to
>>| proceed?
>>
>>Unless the process is stopped, it's memory is going to change. If it
>>does anything at all, each "snapshot" of memory could see it in a
>>different state, as reflected in the stack frame, etc.
>>
>>Are you trying to check it from outside the process, or would it be OK
>>for the program to check itself?
>>
>>--
>>Phil Howard KA9WGN
>>[EMAIL PROTECTED] [EMAIL PROTECTED]
>
>We are trying to check it from an external process (preferably one
>that we could run from ROM (good luck on that one). We want to check
>only the CODE portion of the program, not the stack frame, data, or
>anything else.
>
>We suspect that we can somehow query the kernel (or /proc) to get the
>address information of the process and then verify it through /dev/mem
>or /dev/kmem. We are willing to modify the kernel if necessary, but
>we would prefer not to.
>
>FYI: We are writing code that must go through a certification process
>and part of the certification process specifies that the code must be
>checked periodically while running to make sure that no "patches" have
>been made to it. If we cannot manage to run the checking process from
>ROM, then the checking process must check itself and the information
>it verifies the running process against must come from ROM (CD is
>okay). We have considered running the main code itself out of ROM,
>but we suspect that it would be infeasible due to size restrictions.
>
What do you trust, and what don't you? Starting with the folks who
have keys to the computer room, do maintenance/installation of h/w
and s/w, have remote access to your system (??!), etc.? Then what do
you trust, continuing with the s/w and data involved in the boot
process and so forth until your secure app is loaded and running?
Do you trust any of the access control mechanisms of the OS?
Just wondering, sounds interesting...
Regards,
Bengt Richter
------------------------------
From: [EMAIL PROTECTED]
Subject: Wall Street Developer Opportunities
Date: Fri, 19 Mar 1999 21:23:21 GMT
Wall Street Developer Opportunities
A small but prestigious financial software company in the center of Manhattan
is seeking developers who are not afraid a challenge. As a leading provider of
solutions to financial traders in the US and abroad the company is lives with
the highly competitive and fast paced Wall Street environment.
Applicants must be self-motivated and prepared to learn in an entrepreneurial
and casual setting.
Keywords are: C++, Java, Tcl/TK, HTML, UNIX, NT, Real-time, Equities, Fixed
Income, and Risk Management, and dedication. If by reading the previous
enumeration you think "me" four or more times, we'd like to hear from you.
Previous professional experience is not required although much appreciated.
Typical candidates will hold a Bachelor of Science. Students may contact us
for an internship - which may lead to a job offer.
Salaries start at $45,000 and include health benefits and other incentives.
Email your resume to:
mailto:[EMAIL PROTECTED]
or fax to: (212) 679-4613
============= Posted via Deja News, The Discussion Network ============
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
------------------------------
From: [EMAIL PROTECTED] (John Stapleton)
Subject: Problem with sunrpc when building GLIBC-2.1
Date: Fri, 19 Mar 1999 19:00:03 GMT
I ran into a problem late in the process of building GLIBC-2.1, during
the ``make others'' section. In the sunrpc sub-directory, the
``rpcgen'' program is run using the new ``ld-linux.so.2'' program.
Here is the command broken into separate line to make it easier to
read:
/soft/local/glibc-2.1/elf/ld-linux.so.2 \
--library-path /soft/local/glibc-2.1:/soft/local/glibc-2.1/math: \
/soft/local/glibc-2.1/elf:/soft/local/glibc-2.1/nss: \
/soft/local/glibc-2.1/nis:/soft/local/glibc-2.1/db2: \
/soft/local/glibc-2.1/rt:/soft/local/glibc-2.1/resolv: \
/soft/local/glibc-2.1/linuxthreads \
/soft/local/glibc-2.1/sunrpc/rpcgen \
-Y `gcc -print-file-name=cpp | sed "s|/cpp$||"` \
-c rpcsvc/bootparam.x \
-o /soft/local/glibc-2.1/sunrpc/xbootparam.T
This is the error message that I received:
waitpid: Operation not permitted
This is on a Redhat Linux Alpha System, distribution version 5.1
running kernel version 2.0.34. I did configure GLIBC to use the
headers from version 2.2.3 of the kernel. Here is the configure
command that I used.
../../src/glibc-2.1/configure --prefix=/usr/glibc-2.1 \
--host=alpha-linux --with-headers=/usr/src/linux-2.2.3/include \
--enable-add-ons
If I run the ``rpcgen'' which came with the Redhat distribution, it
works. I spent some time looking in various documentation and FAQ
files, but I could not find anything related to this problem.
The version of the compiler, and the tools that I built along with
the compiler are:
compiler: egcs-2.91.60 19981201 (egcs-1.1.1 release)
binutils: binutils-2.9.1.0.22b
bison: bison-1.27 [see note]
flex: flex-2.5.4a
Note: The bison distribution was built separately before building the
EGCS 1.1.1 release because I had a problem when I made it part of the
EGCS directory tree. If it is part of the EGCS directory tree, then
the EGCS main Makefile defines YACC as
$OBJ/egcs-1.1.1/bison/bison -y -L $SRC/egcs-1.1.1/bison/
and then bison complains when trying to build flex.
Can anyone help?
John
============================================================================
In an attempt to avoid unwanted junk mail, I haven't put my email address
in the header of this article. My username is "jstapes" and my domain is
"gis" + "." + "net". A human reading this should be able to figure it out,
but a program culling addresses will have a difficult time.
--
John T. Stapleton
[EMAIL PROTECTED]
------------------------------
From: Erik de Castro Lopo <[EMAIL PROTECTED]>
Subject: Re: Linux e2fs file size limit!
Date: Sat, 20 Mar 1999 08:49:26 +1100
Phil Howard wrote:
>
> On 19 Mar 1999 14:39:49 +0100 Miquel van Smoorenburg ([EMAIL PROTECTED]) wrote:
>
> | In article <[EMAIL PROTECTED]>,
> | Jeremy Main <[EMAIL PROTECTED]> wrote:
> | >
> | >We are trying to write some files to our filesystem but cannot
> | >create/write anything greater than 1.0 Gb (2^30 bytes) and the
> | >theoretical filesize limit for e2fs is 2.0 Gb (2^31 bytes)
> |
> | Check your ulimit settings. Max filesize defaults to 1GB on many systems
>
> Actually, my ulimit settings are currently "unlimited" but I hit the 1GB limit.
> However, actually setting it to 2GB-1K (full 2GB does not work) works if you
> also set the hard limit. Something is a little goofy and misleading.
>
> With 32 bits to represent size, I'd think the limit could be 4GB-1. Well, at
> least *I* know how to use _unsigned_ integers. I suppose the POSIX API has
> its faults for using signed where unsigned will work, but I don't see why ext2
> had to have the same internal limit. And there's no reason you can't have a
> new FS type that goes well beyond 4GB although you won't necessarily be able
> to map it and seek it all with 32-bit POSIX.
The reason its 2GB instead of 4GB is becasue it needs to be a signed int
so you can do an lseek () by a negative amount (ie backwards).
Hope this helps,
Erik
--
+-------------------------------------------------+
Erik de Castro Lopo [EMAIL PROTECTED]
+-------------------------------------------------+
Orcad Express 9: "Its a nice demo but I wouldn't want to use it
on a day-to-day basis" -- me
------------------------------
From: [EMAIL PROTECTED] (Mike McDonald)
Subject: Re: *BIG* Problem with /etc/rc.d/rc.local
Date: 19 Mar 1999 19:49:32 GMT
Reply-To: [EMAIL PROTECTED]
In article <[EMAIL PROTECTED]>,
"April Escamilla" <[EMAIL PROTECTED]> writes:
> Please help me!!
>
> So I tried to delete the offending lines of code, but was unable to do
> so, as the file is write protected. I thought that this was due to the
> file permissions, so I tried to change them with chmod. Sadly, nothing
> helped, and all chmod calls failed to change its permissions. Despite the
> file being owned by root and its ls -l configuration set as
> -rwxr-xr-x
> I was unable to change it in vi.
> Any thoughts out there? Should I try a different editor? Are the file
> permissions set somewhere else that may be affecting this file? Is chmod
> even the right way to go? One thing I do have is a copy of the original
> "rc.local" on another system. Is there anyway to override this write
> protection and copy the original back in?
> I will be grateful for any info you can pass my way!!
>
> Thanks in Advance,
> April Escamilla
I'll bet the file system was still mounted read only. Try remounting it
read/write.
Mike McDonald
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Peter Samuelson)
Subject: Re: Riva TNT AGP ???
Date: 19 Mar 1999 18:05:57 -0600
Reply-To: Peter Samuelson <[EMAIL PROTECTED]>
[Frank Post <[EMAIL PROTECTED]>]
> Is the above card supported ?
> Is AGP Supported by Linux ?
XFree86 supports AGP: cards that have PCI and AGP versions are
generally either both supported or neither. According to XFree86 docs,
AGP is pretty similar to PCI anyway, as far as interfacing to them.
RIVA TNT is supported, but only in recent versions of XF86_SVGA.
The TNT driver first appeared in XFree86 3.3.2.3 and caused an uproar
because some pencil pusher at NVidia balked at letting XFree86
distribute the source code they'd contributed until they removed a
couple header files (and preprocessed the corresponding source files so
they would still work). XFree86 was within a day or two of the release
so they couldn't just pull the driver, so they shipped with the loud
warning that one driver was not open source. There was much yelling
and screaming on Slashdot over this one. But since then NVidia gave
XFree86 a truly open-source driver, included in 3.3.3, so I guess
they're cool again. Speculation ran rampant about what this was all
about: one theory was that the original header files in question had
people's names in them that could be used in a pending lawsuit about
hiring away someone else's intellectual property (don't remember
whose).
--
Peter Samuelson
<sampo.creighton.edu!psamuels>
------------------------------
From: "Mario Schrijver" <[EMAIL PROTECTED]>
Crossposted-To:
apana.lists.os.linux.kernel,apana.lists.os.linux.ppp,apana.lists.os.linux.redhat,at.fido.linux,comp.os.linux.hardware,comp.os.linux.networking,comp.os.linux.portable,comp.os.linux.setup,de.comp.os.unix.linux.hardw
Subject: pcmcia freezes me to death
Date: 19 Mar 1999 20:11:22 GMT
Hi,
I,ve got a Micropoint Iliission with the pcmcia-cs-3.0.8 module combined
with 2.2.3 Linux Kernel. Starting PCMCIA freezes the machine.
probe -m shows an i82365 like pcmcia chipset in my machine. Still, loading
the modules by hand shows that all other related modules don't cause a
crash, but the i82365 does
When i relay the output to logfiles it only shows the module going in for a
restart... and then it stops.
How to debug this?
Anyone has a clue how to handle this.?
Regards,
Mario Schrijver
------------------------------
From: "Ron Reaugh" <[EMAIL PROTECTED]>
Crossposted-To: comp.periphs.scsi,comp.sys.ibm.pc.hardware.chips,comp.unix.internals
Subject: Re: DMA question
Date: 20 Mar 1999 00:42:55 GMT
David Scott Peterson wrote in message <7cunob$7a1$[EMAIL PROTECTED]>...
>It is my understanding that the typical x86 PC has 7 useable DMA
>channels and a DMA channel can be used to transfer up to 128k of
>data at a time. How does this compare to what is typically found
>on other architectures (SPARC, Alpha, PowerPC, etc.) ?
That's classic ISA DMA and has nothing to do with modern PCI busmastering
DMA.
------------------------------
From: "John Burton" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development.apps
Subject: Re: Building Linux Shared Libraries
Date: Fri, 19 Mar 1999 20:46:17 -0000
I just use :-
'gcc -shared -o lib.so lib.o'
Then I link my apps :-
'gcc -o testapp testapp.c ./lib.so'
Seems to work for me.
Does anyone know what happens if you DONT build your shared library using
the position
independent code flag for gcc? I asked on comp.unix.programmer but so far no
replies.
Scott Ratliff wrote in message <7cu102$ht5$[EMAIL PROTECTED]>...
>
>I am trying to create a share library that is linked with other
>shared libraries on Linux (RH 5.0 on Intel). I have tried
>basically two different methods of creating these libraries
>and neither one seems to work for me.
>
>The first method is to use 'ld -shared'. When I do this I can
>create all of the shared libraries with no problem. However,
>when I go to link and executable with this shared library I
>get a bunch of undefined reference errors from the shared library
>of functions that are in the initial shared libraries. I have
>tried setting 'LD_LIBRARY_PATH' to '`pwd`:/lib:/usr/lib' and
>adding -rpath parameters but nothing seems to work.
>
>The second method I've tried is to use 'ld -Ur'. When I do
>this, creating the shared library that links with the other
>shared libraries chokes because of multiple definition problems.
>These errors are from global variables that are in include
>files that are shared amongst the libraries.
>
>The kicker is that building these shared libraries works for
>me on other platforms using the vendors compilers/linkers.
>So the question is, is there a GNU'ism I'm missing
>(gcc/ld parameters/versions) or do I need to do a boat load
>of ifdef'ing to get this shared library created on Linux.
>
>Thanks in advance for any assistance.
>
>Scott Ratliff
>
>scottr (NOSPAM)
>@ (NOSPAM)
>carsinfo.com (NOSPAM)
>
>
>P.S. I can provide small examples if needed.
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To: comp.os.linux.advocacy,comp.os.linux.questions
Subject: Re: After Week 1 With Linux -- licking wounds.
Reply-To: [EMAIL PROTECTED]
Date: Fri, 19 Mar 1999 20:47:27 GMT
[EMAIL PROTECTED] (Christopher Browne) writes:
>On Thu, 18 Mar 1999 15:05:55 GMT, [EMAIL PROTECTED]
><[EMAIL PROTECTED]> wrote:
>>I have no problem believing that more than half of the information we _take
>>in_ is in visual form. After all, most of the information we take in is
>>about things that just happen to sit around somewhere. And we are extremely
>>good at pattern recognition.
>Counterpoint...
>How many ballot cards have you been asked to fill in at election time
>that presented pictures of candidates rather than their names?
I don't see the counterpoint in that. As I said, we are extremely good at
pattern recognition --- which, among other things, means recognizing written
text.
Now, how many ballots have you been asked to fill out where the candidates
were presented by sound, smell, taste or touch?
Bernie
--
============================================================================
"It's a magical world, Hobbes ol' buddy...
...let's go exploring"
Calvin's final words, on December 31st, 1995
------------------------------
From: "John Burton" <[EMAIL PROTECTED]>
Subject: select & poll
Date: Fri, 19 Mar 1999 20:51:43 -0000
In the man page for "poll" on my system (slackware 2.0.34 kernel based) it
says that
the poll system call is implemented in terms of "select".
In the book "linux device drivers" it claims that the kernel interface for
select has
been changed for poll because it is not possible to implement poll in terms
of
select.
Which one is correct?
------------------------------
From: Kevin White <[EMAIL PROTECTED]>
Reply-To: [EMAIL PROTECTED]
Subject: Re: How do you get drivers for ATI Pro Rage AGP?
Date: Fri, 19 Mar 1999 13:59:44 -0700
Ampem Darko wrote:
>
> Ive got an aTI Pro Rage AGP Graphics card,
>
> How do i set up linux to use it with X11?
I think the rage pro uses the mach 64 chipset. You can run xf86config
and select to use the mach64 accelerated x server.
I have an ATI Rage Pro 3d agp2x chipset, and the mach 64 x server runs
great. You will want to read the readme files, of course, and make sure
you monitor, etc., are setup right, you know, all the general "setting
up xf86" stuff.
--
Kevin White, Software Engineer
Envision Telephony
[EMAIL PROTECTED]
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: device/inode to filename
Date: Sat, 20 Mar 1999 01:10:56 GMT
On Sat, 20 Mar 1999 00:53:25 GMT, [EMAIL PROTECTED] (Phil Howard)
wrote:
>On Fri, 19 Mar 1999 22:51:29 GMT [EMAIL PROTECTED] wrote:
>| I need to convert a device (e.g. 08:01, which on my system is sda1)
>| and an inode to a filename. I could find where sda1 is mounted and
>| then use find -mount -inum 1234, but that is rather ridiculous...
>
>Ridiculous as it seems, that's the way it's done. The mapping between
>a filename and an inode is ONE WAY. Further, there may be many such
>filenames that map to the inode, and the only evidence of this is that
>the inode does contain a link count. You simply have to go look for
>the names that map to the inode of interest.
>
>If you're going to do this frequently, it would be better to collect
>all the name and inode mappings into a file and search that. Here is
>a quick way to do it:
>
> find /mountpoint -printf '%i %P\n' > inodefiles
>
>Note that %P does not include the command line component, whereas %p
>does include it. Then to search the file:
>
> egrep "^1234 " < inodefiles
>
>Be sure to include the ^ before and the space after the inode number.
>If you're doing it big time, you might even go so far as to store the
>mappings in DB files.
>
>--
>Phil Howard KA9WGN
>[EMAIL PROTECTED] [EMAIL PROTECTED]
You've been a lot of help. Thank you very much.
However... (of course)
Even if I can't get the file name from the device/inode, is there any
way to open the file with just the device/inode?
------------------------------
From: [EMAIL PROTECTED] (ellis)
Subject: Re: Device Driver...what for?
Date: 20 Mar 1999 01:23:35 GMT
In article <[EMAIL PROTECTED]>,
Caolan (McNamara) <[EMAIL PROTECTED]> wrote:
>Computer Based Telephony cards, theres a large lack of supported cards
>in this area. On one project that im involved in we had to reccomend
>using solaris rather that linux because there was no support for any
>cbt cards.
Aculab has release some drivers with more to come.
--
http://www.fnet.net/~ellis/photo/
------------------------------
From: James Youngman <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux,comp.os.linux.development.apps,comp.lang.c
Subject: Re: Help with a C program in Linux
Date: 18 Mar 1999 20:13:38 +0000
[EMAIL PROTECTED] (Nick) writes:
> In DOS there is a function, int86(interrupt,register_in, register_out).
> It allows you to modify the system registers.
Actually, that isn't what it does.
> Is there a way to do this in Linux?
What, change registers, or do what int86() actually does?
> Also if there is, where can I find a listing of all the interupts
> and what they do?
for m in /usr/man/man2/*; do man $m || break; done
> In DOS there is a function getch(), Linux has getc() and getchar()
> which work similiary but they require that you press enter. Is
> there anyway of getting them to work like getch() or how would I got
> about writting my own getch?
See the comp.unix.programmer FAQ.
> If you can e-mail me at [EMAIL PROTECTED] it would be much
> appreciated.
Sorry.
--
ACTUALLY reachable as @free-lunch.demon.(whitehouse)co.uk:james+usenet
------------------------------
From: [EMAIL PROTECTED] (ellis)
Subject: Re: Device Driver...what for?
Date: 20 Mar 1999 01:25:32 GMT
In article <[EMAIL PROTECTED]>, Vincent <[EMAIL PROTECTED]> wrote:
>I'm looking for some documentation about programming Drivers device and specialy on
>PCI
>hardware.
>Have you some source about it for me ?
There's a very good book available:
http://www.amazon.com/exec/obidos/ASIN/1565922921/ricksphotograpag
------------------------------
From: [EMAIL PROTECTED] (ellis)
Subject: Re: *BIG* Problem with /etc/rc.d/rc.local
Date: 20 Mar 1999 01:32:31 GMT
In article <[EMAIL PROTECTED]>,
April Escamilla <[EMAIL PROTECTED]> wrote:
>
> My redhat Linux 5.2 box will NOT boot completely, and drops me into a
>bash shell under root to fix the problem, which I will describe as best I
>can. Prior to this situation, the machine re-booted normally. Before this
>error occured, I changed but one thing -- I added 4 lines to the "rc.local"
>file to help install IP Masquerading. I'll list them here:
>
> /sbin/depmod -a
> /sbin/modprobe ip_masq_ftp
> /sbin/modprobe ip_masq_raudio
> /sbin/modprobe ip_masq_irc
>
> I understand that "rc.local" is one of the last files to be run during
>the boot sequence. I also understand that as root, I should be able to
>manipulate this file in the shell and *hopefully* reboot normally.
> So I tried to delete the offending lines of code, but was unable to do
>so, as the file is write protected. I thought that this was due to the
>file permissions, so I tried to change them with chmod. Sadly, nothing
>helped, and all chmod calls failed to change its permissions. Despite the
>file being owned by root and its ls -l configuration set as
> -rwxr-xr-x
> I was unable to change it in vi.
You need to remount the file system read/write before you can change
anything. What was the error that made the system fail to come up?
--
http://www.fnet.net/~ellis/photo
------------------------------
From: [EMAIL PROTECTED] (Daniel R. Grayson)
Subject: Re: Verifying memory of running process
Date: 19 Mar 1999 15:44:30 -0600
[EMAIL PROTECTED] writes:
> On Fri, 19 Mar 1999 04:59:51 GMT, [EMAIL PROTECTED] (Phil Howard)
> wrote:
>
> >On Thu, 18 Mar 1999 22:32:05 GMT [EMAIL PROTECTED] wrote:
> >
> >| For a secure application running under Linux (version 2.X or higher),
> >| I need to be able to verify that the memory of the running process
> >| (executable memory) does not change. We would like to be able to
> >| either checksum or do a binary comparison of the process's memory with
> >| a known value. Has anyone ever done this? Any suggestions on how to
> >| proceed?
> >
Say the process id is 123. Get the memory maps from /proc/123/maps. It'll
look like this:
08048000-080c8000 r-xp 00000000 03:03 187828 /usr/sbin/lpd
080c8000-080cd000 rw-p 0007f000 03:03 187828 /usr/sbin/lpd
080cd000-080e8000 rwxp 00000000 00:00 0
40000000-40006000 r-xp 00000000 03:03 51457 /lib/ld-linux.so.1.9.9
40006000-40007000 rw-p 00005000 03:03 51457 /lib/ld-linux.so.1.9.9
40007000-40008000 rw-p 00000000 00:00 0
4000b000-4003b000 r-xp 00000000 03:03 51269 /usr1/lib/libncurses.so.3.0.0
4003b000-40043000 rw-p 0002f000 03:03 51269 /usr1/lib/libncurses.so.3.0.0
40043000-40046000 rw-p 00000000 00:00 0
bfffe000-c0000000 rwxp fffff000 00:00 0
The executable ones are marked. Running as root, read each one from
/proc/123/mem at the indicated addressses to compute the checksum.
------------------------------
** 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
******************************
