Newsgroups: comp.os.linux
From: Eric Youngdale <youngdale@v6550c.nrl.navy.mil>
Subject: Problem with tar reader in rootimage 0.95a
Reply-To: youngdale@v6550c.nrl.navy.mil
Organization: The Internet
Date: Sat, 9 May 1992 18:42:29 GMT
Linux Activists:
I just installed linux on my 486-33 system over the past
couple of days, and first of all I wanted to congradulate you all.
I had been planning to wait for hurd to reach a state where it was useful,
but linux fills the bill quite nicely. I must admit that my experience
in managing a SVr4 unix machine came in *very* handy during installation.
There are a couple of rough edges that could be polished off. To
make it easier for people who work on these things, I will make one point here,
and save the rest for other mail messages.
First of all the tar reader in the rootimage-0.95a apparently has a bug
in it. I am sorry that I do not have enough utilities going to be able
to come up with a patch, but it appears that a patch is not needed (I will
explain later). I noticed the problem when trying to add the 2.1-lib.tar.Z
and 2.1-misc.tar.Z distributions to my system. Having used the command
tar xZf /dev/PS0
to unpack, and then having done the installation, I found that the gcc
image (i.e. the compiler driver itself) was adding a unparsable option
to the list of options for ld. On the screen it looked almost line
four dashes, except that the second one was a little box shaped
creature. The cc1 image was also corrupt, because it would report a
syntax error for the simplest programs, complaining about the '(' in
a function definiition or prototype. I tried making a second compiler disk and
installing it, and no dice. It was not until I tried using the tar on the
utils.tar.Z distribution that I was able to unpack a compiler that would work
for me.
-Eric
Newsgroups: comp.os.linux
From: Eric Youngdale <youngdale@v6550c.nrl.navy.mil>
Subject: Problem building kernel 0.95c+ (and pre-0.96?)
Reply-To: youngdale@v6550c.nrl.navy.mil
Organization: The Internet
Date: Sat, 9 May 1992 18:47:42 GMT
Linux Activists:
The file tools/build in linux-0.95c.tar.Z distribution will not compile
because of the line:
#include < linux/fs.h>
Basically the compiler cannot find linux/fs.h because there was no -Iinclude
specified for this file in the Makefile. I tried adding a -Iinclude, which
helped a little, but this created other problems because some of the data types
in linux/fs.h were declared somewhere else. To fix this, I ended up adding
the lines:
#define MAJOR(a) (((unsigned)(a))>>8)
#define MINOR(a) ((a)&0xff)
to build.c and removing the above include in order to get the thing to fly.
I looked at the sources to pre-0.96, and it looked like situation
had not changed.
-Eric
Newsgroups: comp.os.linux
From: Eric Youngdale <youngdale@v6550c.nrl.navy.mil>
Subject: Lack of send_break for serial I/O
Reply-To: youngdale@v6550c.nrl.navy.mil
Organization: The Internet
Date: Sat, 9 May 1992 18:53:31 GMT
Linux Activists:
I was having trouble with Kermit, because the hangup command was not
working, and I am guessing that the underlying problem is that the function
send_break in linux/kerel/chr_dev/tty_ioctl.c is essentially a null function.
Is there a good reason that this has not yet been implemented? Are
there kernel patches to implement this? Is someone working on this? Finally,
if the answer to the above questions is "no", would you be interested in
having me do the job?
-Eric
(I will be out of town all next week, so I would not be able to get to it until
after that).
From: youngdale@v6550c.nrl.navy.mil (Eric Youngdale)
Newsgroups: comp.os.linux
Subject: Re: send_break
Date: 9 May 92 19:52:54 GMT
Reply-To: youngdale@v6550c.nrl.navy.mil
Organization: The Internet
Linux Activists:
| I was having trouble with Kermit, because the hangup command was not
|working, and I am guessing that the underlying problem is that the function
|send_break in linux/kerel/chr_dev/tty_ioctl.c is essentially a null function.
I realized that the above statement was far too vague. The problem
that I am having is that the DTR line is not dropped, and as a result the modem
does not hang up the phone. The only way that I can get the modem to hang up
the phone is to power-cycle the modem, which is a real pain.
I saw that Linus had commented in comp.os.linux that 0.96 will drop DTR
when the serial line is closed. If I am reading this correctly, then all I
need to do is to wait until 0.96 comes out.
-Eric