Linux-Development-Sys Digest #511, Volume #6 Sat, 20 Mar 99 05:13:58 EST
Contents:
Re: After Week 1 With Linux -- licking wounds. (Triple-T)
Re: Verifying memory of running process (Peter Samuelson)
Qt-1.42 + Glibc-2.1 problem (Nobody)
Re: After Week 1 With Linux -- licking wounds. ("Misterfixit")
DMA question (David Scott Peterson)
Re: Linux e2fs file size limit! (Peter Samuelson)
Re: device/inode to filename (Peter Samuelson)
Re: Verifying memory of running process (Daniel R. Grayson)
Pentium Protection Rings (Joydeep Sen Sarma)
Re: gcc problems (Jeffrey L Straszheim)
MediaGXm and sound ([EMAIL PROTECTED])
Re: select & poll ("G. Sumner Hayes")
linux newbie needs help installing 2.2 kernel (Michael)
Re: Building Linux Shared Libraries ("David Z. Maze")
Re: After Week 1 With Linux -- licking wounds. (David Kurtz)
Affordable and reliable IrDA infrared communications for 8/16/32/64 bit CPU's
([EMAIL PROTECTED])
Re: Qt-1.42 + Glibc-2.1 problem (Andreas Jaeger)
Re: Barclays Internet banking under Linux? (how to change uname()) (Robert Billing)
----------------------------------------------------------------------------
From: Triple-T <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.advocacy,comp.os.linux.questions
Subject: Re: After Week 1 With Linux -- licking wounds.
Date: Fri, 19 Mar 1999 20:05:59 -0600
personally I think it's because people are cheap and it's cheaper to
print text than images
Paul Flinders wrote:
>
> [EMAIL PROTECTED] (Peter Samuelson) writes:
>
> >
> > Or again, why do ballots not usually just print pictures of the
> > candidates instead of their names? Because it would result in mass
> > confusion. Because people can't be relied on to be able to recognize
> > their favorite candidates by face. The visual medium is simply not
> > reliable enough to entrust something like an election to.
>
> I expect that's because most people vote for parties, not individuals
--
Great spirits often encounter violent opposition from mediocre minds.
-- Albert Einstein
------------------------------
From: [EMAIL PROTECTED] (Peter Samuelson)
Subject: Re: Verifying memory of running process
Date: 19 Mar 1999 19:37:29 -0600
Reply-To: Peter Samuelson <[EMAIL PROTECTED]>
[<[EMAIL PROTECTED]>]
> My /proc/<process id>/maps file does not include the translation of
> device/inode to filename. What is the easiest (best?) way to do this
> translation? Also, is there any good documentation on proc besides
> the man page?
Hmmm, dunno, I've been using Linux 2.1.x so long I don't remember the
differences. 2.2.x gives filenames. Also, the 2.2.x source tree
includes the file Documentation/proc.txt which gives quite a bit of
information, though it doesn't seem to be quite up-to-date. (Which is
odd, since it was patched into the official source tree pretty
recently.)
> To answer some of my own questions: /proc/devices will list the
> devices. Thus, I know that (on my system) 08:01 refers to the first
> SCSI drive.
On kernel 2.2.x, use /proc/partitions which has both major and minor.
Unfortunately I guess that doesn't help you on 2.0.
--
Peter Samuelson
<sampo.creighton.edu!psamuels>
------------------------------
From: [EMAIL PROTECTED] (Nobody)
Subject: Qt-1.42 + Glibc-2.1 problem
Date: 19 Mar 1999 12:02:00 GMT
Reply-To: nobody@nowhere
I installed (successfully) glibc-2.1 on my system, and have compiled
(with egcs-1.1.1) several programs that work fine with it (I first
recompiled things like ncurses, though). I am now trying to compile
qt-1.42, and have a problem with the 'moc' program. When I run it,
it seems to do its job, but it finishes with a segmentation fault.
Here's what gdb reports:
Program received signal SIGSEGV, Segmentation fault.
0x400cc0c1 in _IO_unbuffer_write ()
I have built the program with and without optimizations, and
the problem persists. Does anybody have a clue?
Thanks,
Ricard
------------------------------
From: "Misterfixit" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.advocacy,comp.os.linux.questions
Subject: Re: After Week 1 With Linux -- licking wounds.
Date: 19 Mar 1999 18:54:33 PST
There are many replies to this original post which are quite reminiscent to
the "Team OS2" ranting back when the OS2?Warp zealots were circling like
vultures over anyone failing to immediately be absorbed by the Body of Warp.
I sure hope we are not developing a sub-genus of "Team Linux" ravers like
what happened in the embryonic Warp days.
Cut the guy some slack already ... we've all been pissed about something in
our lives ... Different strokes for different folks.
Regards
Dave
Linux 24/7/365
------------------------------
From: [EMAIL PROTECTED] (David Scott Peterson)
Crossposted-To: comp.periphs.scsi,comp.sys.ibm.pc.hardware.chips,comp.unix.internals
Subject: DMA question
Date: 19 Mar 1999 23:48:27 GMT
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.) ?
------------------------------
From: [EMAIL PROTECTED] (Peter Samuelson)
Subject: Re: Linux e2fs file size limit!
Date: 19 Mar 1999 17:43:27 -0600
Reply-To: Peter Samuelson <[EMAIL PROTECTED]>
[Erik de Castro Lopo <[EMAIL PROTECTED]>]
> 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).
Shouldn't matter. Since when did anyone guarantee that you could get
to any file offset in a single system call? That shouldn't affect the
design of a filesystem.
Now if SEEK_CUR didn't exist, there would be a problem.
--
Peter Samuelson
<sampo.creighton.edu!psamuels>
------------------------------
From: [EMAIL PROTECTED] (Peter Samuelson)
Subject: Re: device/inode to filename
Date: 19 Mar 1999 20:06:06 -0600
Reply-To: Peter Samuelson <[EMAIL PROTECTED]>
[<[EMAIL PROTECTED]>]
> 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?
Nope. It would bypass the permission checks on the directory (and its
parents) the file is in. If you need to do this, write a new system
call and restrict its use to root or whatever. I don't think it would
be all that hairy.
--
Peter Samuelson
<sampo.creighton.edu!psamuels>
------------------------------
From: [EMAIL PROTECTED] (Daniel R. Grayson)
Subject: Re: Verifying memory of running process
Date: 19 Mar 1999 22:10:04 -0600
[EMAIL PROTECTED] writes:
> My /proc/<process id>/maps file does not include the translation of
> device/inode to filename. What is the easiest (best?) way to do this
> translation?
Why bother? You know the executable name, since you know what program is
running, so use ldd to figure out what libraries will be used. Ldd will also
tell you where the libraries will likely load. Now use ls to find out the
inode numbers.
geometry% sleep 1000 &
[1] 312
geometry% cat /proc/312/maps
08048000-0804a000 r-xp 00000000 03:03 188217 /usr/bin/sleep
0804a000-0804b000 rw-p 00001000 03:03 188217 /usr/bin/sleep
40000000-40013000 r-xp 00000000 03:03 122602 /lib/ld-2.1.so
40013000-40014000 rw-p 00012000 03:03 122602 /lib/ld-2.1.so
40014000-40015000 rwxp 00000000 00:00 0
40018000-400f7000 r-xp 00000000 03:03 122603 /usr2.1/lib/libc-2.1.so
400f7000-400fc000 rw-p 000de000 03:03 122603 /usr2.1/lib/libc-2.1.so
400fc000-400ff000 rw-p 00000000 00:00 0
bfffe000-c0000000 rwxp fffff000 00:00 0
geometry% ldd /usr/bin/sleep
libc.so.6 => /usr2.1/lib/libc.so.6 (0x40018000)
/lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x40000000)
geometry% ls -li /usr/bin/sleep /usr2.1/lib/`readlink /usr2.1/lib/libc.so.6`
/lib/ld-linux.so.2
51088 lrwxrwxrwx 1 root root 9 Mar 18 09:18 /lib/ld-linux.so.2
-> ld-2.1.so
188217 -rwxr-xr-x 1 root root 8276 Jan 10 14:27 /usr/bin/sleep
122603 -rwxr-xr-x 1 root root 3985785 Mar 4 20:51
/usr2.1/lib/libc-2.1.so
------------------------------
From: Joydeep Sen Sarma <[EMAIL PROTECTED]>
Subject: Pentium Protection Rings
Date: 20 Mar 1999 01:47:31 GMT
Hi,
I want to restrict access to a particular
threads shared memory segments from other
threads in the same process. I was planning
to do it by changing the privilege levels
of these threads and the shared memory segments
to 2 (or 1). Can anyone give me some idea
on how this can be done ? - or if there are any
other better ways of enforcing thread level
protection on memory ?
Thanks,
Joydeep
--
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
@ @
@ Residence : Office: @
@ 411 N. Craig St., #5 303A Mineral Industries Bldg.@
@ Pittsburgh, PA15213 Dept. of Computer Science @
@ Phone : (412) 802-0217 University of Pittsburgh @
@ EMail : [EMAIL PROTECTED] Pittsburgh, PA15260 @
@ URL : www.cs.pitt.edu/~sens Phone: (412) 624-9955 @
@ FAX : (412) 624-8854 @
@ @
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
------------------------------
From: Jeffrey L Straszheim
<[EMAIL PROTECTED]>
Subject: Re: gcc problems
Date: Sat, 20 Mar 1999 00:14:15 -0500
David B Anderson wrote:
[snip]
> I'd suggest the book
>
> Linux Application Development
> authors: Johnson and Troad
> publisher: Addison Wesley.
>
> For details you will have to go to the various info docs, but
> the book is a good place to start.
> They write a complete 'shell' CLI program as an example.
> Lots of http references to places for more info.
I second this suggestion, this is a good book to peruse
before starting. There are some IDE's available on the
various Linux d/l sites, but I'd recommend that you take
some time to learn a bit about makefiles before you shift
to the IDE route. Try the commands "info gcc" and
"info make" next. Also look at a few of the simpler
makefiles that come with the smaller apps. This may
seem like more work than the IDE, but it will serve you
well when you want to distribute your application in source
form (there are Linux users on Alphas and such who might
want to run your app., it is likely that you won't be able
to provide a binary for ever configuration). If then
you decide that the IDE convenience is not something that
you can do without, at least you will be comfortable
with what the IDE is doing behind the surface. This
knowledge is invaluable when something goes wrong.
--
--Jeffrey Straszheim
---Systems Engineer, Programmer
----stimuli AT shadow DOT net
------------------------------
From: [EMAIL PROTECTED]
Subject: MediaGXm and sound
Date: Sat, 20 Mar 1999 02:16:02 GMT
I have a set of LPX chassis with Cyrix MediaGXm based motherboards. The
motherboards seem to be made by a company called SuperTek Computer
Electronics Inc. The goal is to use them for public terminals, so they
need to be working nicely. Linux boots and runs well and with a good
speed expected of a 233MHz CPU. X runs nicely once acceleration is turned
off as there is at least one bug in acceleration, but we can live with that
as the speed is still acceptible.
Where we are having problems is with the sound. The MediaGXm appears to
emulate a Soundblaster 16, and is detected by the sounds drivers as such.
However the sound output comes out with a high level of distortion. It
sounds like something is interrupting the output at 100Hz. I'd like to
know if anyone has used the sound output on a MediGX-based system successfully
and if there is anything I can try to improve it.
Regards,
RH.
============= Posted via Deja News, The Discussion Network ============
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
------------------------------
From: "G. Sumner Hayes" <[EMAIL PROTECTED]>
Subject: Re: select & poll
Date: Fri, 19 Mar 1999 21:30:40 -0500
John Burton wrote:
>
> 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?
In current versions of Linux poll() is a system call. select() is a
library wrapper around poll(). poll() as an interface may be slightly
more efficient.
In older versions of Linux, poll() didn't exist -- the kernel offered
select() and some versions of libc had a poll() wrapper around it. This
was less efficient than the current setup, and also broke poll()
semantics on some not-too-uncommon corner cases IIRC.
Of course, this may vary based on what version of libc you're using
and what kernel headers it was compiled against. The best way to see
what system calls are being issued by your program is to use strace --
man strace FMI.
I expect that your slackware man page is outdated, but use strace to
check for yourself.
Hope this helps,
Sumner
------------------------------
From: Michael <[EMAIL PROTECTED]>
Subject: linux newbie needs help installing 2.2 kernel
Date: Sat, 20 Mar 1999 00:31:38 -0500
I've downloaded the 2.2.3 kernel (linux-2.2.3.tar.gz). I extracted it to
/usr/src/linux. I follow the instuctions in the readme but when I type
'make config' (without the quotes), it says no rule to make target linux
or something like that. Am i doing something wrong? I'm running RH Linux
5.0 with 2.0.32 kernel.
------------------------------
From: "David Z. Maze" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development.apps
Subject: Re: Building Linux Shared Libraries
Date: 20 Mar 1999 01:14:21 -0500
John Burton <[EMAIL PROTECTED]> writes:
JB> Does anyone know what happens if you DONT build your shared
JB> library using the position independent code flag for gcc? I asked
JB> on comp.unix.programmer but so far no replies.
I think this is a Linux-specific thing (every system does shared
libraries differently). If you don't use -fPIC, you still get a valid
.so file, but the code pages aren't shareable. So you get the
advantage that disk images of binaries linked against the library are
smaller, but you don't get the advantage of smaller memory images.
(Both of these advantages are vs. static-link [.a] libraries.)
--
David Maze [EMAIL PROTECTED] http://donut.mit.edu/dmaze/
"Hey, Doug, do you mind if I push the Emergency Booth Self-Destruct Button?"
"Oh, sure, Dave, whatever...you _do_ know what that does, right?"
------------------------------
From: [EMAIL PROTECTED] (David Kurtz)
Crossposted-To: comp.os.linux.advocacy,comp.os.linux.questions,comp.sys.mac.advocacy
Subject: Re: After Week 1 With Linux -- licking wounds.
Date: Fri, 19 Mar 1999 22:37:20 -0800
In article <36f31070$1$xnlawnl$[EMAIL PROTECTED]>,
[EMAIL PROTECTED] wrote:
> In <7csprv$5t1$[EMAIL PROTECTED]>, on 03/19/99
> at 06:12 AM, Cari D. Burstein <[EMAIL PROTECTED]> said:
>
> >|Ye have enemies? Good, good- that means ye've stood up for |
> >| something, sometime in thy life.... -Elminster of Shadowdale |
>
> I really like this line. What is the source? A book ... ? (Shadowdale?)
>
> Kenward
Elminster of Shadowdale is a character out of Ed Greenwood's (and
subsequently other authors who wrote for the series) Forgotten Realms
stories. Forgotten Realms was a fantasy world he created which was later
picked up by TSR, Inc. for use as the "default world" for all Advanced
Dungeons & Dragons 2nd Edition rules, scenarios, and stories. A number of
relatively successful novels, RPG scenarios and computer games (I'm
reliving one of them now on my Mac II: Pool of Radiance) have been based
on the world.
Elminster is one of those characters who has been everywhere, seen nearly
everything, knows everyone, and is hundreds of years older than you think.
Cranky old wizard in a tower type of guy.
--
David Kurtz -- remove the underscore from my email address to reply
PGP key and more... http://www.lightside.net/~david/
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To:
comp.arch,comp.arch.embedded,comp.arch.fpga,comp.arch.fpga,comp.lang.c,comp.os.linus.m68k,comp.os.msdos.apps
Subject: Affordable and reliable IrDA infrared communications for 8/16/32/64 bit CPU's
Date: Sat, 20 Mar 1999 08:20:11 GMT
IrPro(Tm) is An IrDA compliant software protocol stack that supports
reliable infrared communications for portable devices.
For more information check out:
http://www.embednet.com
Our customers have successfully used our stack on popular
8 bit CPU's like: 8051 derivatives, PIC.
We also support Atmel's AVR CPU's .
This software stack is:
. modular and written in "clean" C.
. small (<11K bytes)
. efficient ( Zero copy buffer management )
. able to support speeds from 9.6Kbps to 4Mbps
. able interface to any commercial RTOS or a custom OS
. has direct support for popular CPU's or is simple to port to your
custom CPU.
. robust and mature.
. reliable, and provides inexpensive data transfers
. supported on all popular platforms: MIPS, ARM, Motorola, Intel,
Hitachi SH, etc ...
. affordable (<$6k, no royalties, all source code)
Advantages of using IrDA infrared:
. requires no connectors
. is low power and low cost.
. infrared is not regulated like RF and requires no certification
. data transfers are reliable and robust
. IrDA is supported by MS Windows operating systems.
------------------------------
From: Andreas Jaeger <[EMAIL PROTECTED]>
Subject: Re: Qt-1.42 + Glibc-2.1 problem
Date: 20 Mar 1999 09:09:40 +0100
>>>>> Nobody writes:
> I installed (successfully) glibc-2.1 on my system, and have compiled
> (with egcs-1.1.1) several programs that work fine with it (I first
> recompiled things like ncurses, though). I am now trying to compile
> qt-1.42, and have a problem with the 'moc' program. When I run it,
> it seems to do its job, but it finishes with a segmentation fault.
> Here's what gdb reports:
> Program received signal SIGSEGV, Segmentation fault.
> 0x400cc0c1 in _IO_unbuffer_write ()
> I have built the program with and without optimizations, and
> the problem persists. Does anybody have a clue?
That looks like calling write without an opened file. Have a look
with gdb if the FILE parameter is ok - and if not this is a bug in qt.
Andreas
--
Andreas Jaeger [EMAIL PROTECTED] [EMAIL PROTECTED]
for pgp-key finger [EMAIL PROTECTED]
------------------------------
From: Robert Billing <[EMAIL PROTECTED]>
Crossposted-To: uk.finance,uk.comp.os.linux
Subject: Re: Barclays Internet banking under Linux? (how to change uname())
Date: Sat, 20 Mar 1999 08:43:14 +0000
Paul Ashton wrote:
> Personally, I use lloyds online with Linux. They don't use any java
> or javascript but force me to accept one cookie which I find acceptable.
Can you tell me how? It just gives me "incorrect browser".
--
I am Robert Billing, Christian, inventor, traveller, cook and animal
lover, I live near 0:46W 51:22N. http://www.tnglwood.demon.co.uk/
"Bother," said Pooh, "Eeyore, ready two photon torpedoes and lock
phasers on the Heffalump, Piglet, meet me in transporter room three"
------------------------------
** 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
******************************
