[64] in linux-net channel archive
Modules 1.1.89
daemon@ATHENA.MIT.EDU (Lars Marowsky-Bree)
Mon Feb 6 20:05:46 1995
Date: Mon, 6 Feb 1995 21:55:11 +0000 (GMT)
From: Lars Marowsky-Bree <lmb@pointer.in-minden.de>
To: Rutgers Majordomo <linux-kernel@vger.rutgers.edu>
cc: Rutgers Majordomo <linux-net@vger.rutgers.edu>
I just wanted to say that I really like the idea of modules, but not the way
they are now. Well, I do even like the way - but not the way how they are
implemented.
While compiling the kernel with module support, it first choked on the
genksyms, but the second attempt suddenly succeeded, although I did a make
clean ; make dep between the two tries.
Then, during make modules, it suddenly choked on slhc.c, saying that it
couldn't find some include files (to be exact, every file included in the
""-way, I changed it to <../net/inet/..> and it compiled cleanly...).
--- slhc.c.orig Mon Feb 6 19:57:58 1995
+++ slhc.c Mon Feb 6 20:03:24 1995
@@ -67,12 +67,12 @@
#include <linux/fcntl.h>
#include <linux/inet.h>
#include <linux/netdevice.h>
-#include "ip.h"
-#include "protocol.h"
-#include "icmp.h"
-#include "tcp.h"
+#include <../net/inet/ip.h>
+#include <../net/inet/protocol.h>
+#include <../net/inet/icmp.h>
+#include <../net/inet/tcp.h>
#include <linux/skbuff.h>
-#include "sock.h"
+#include <../net/inet/sock.h>
#include <linux/errno.h>
#include <linux/timer.h>
#include <asm/system.h>
----------
After the compile, I thought, cool, now lets install the modules and reboot.
So I did. depmod -a in rc.S first died because it couldn't find nm. I found
out that this one was in /usr/bin/, which could only be mounted after the
modules were installed. No big deal, perhaps a setup error on my side, I
just moved it to /bin/nm. But I would wish this to be document if a program
which is expected to be run before mount requires external files...
Then, everything seemed to work. Cool. But suddenly, it stopped working. The
reason was that the dummy device - I NEED it - is suddenly called dummy0 if
loaded as a module, instead of just dummy. I corrected this in the file, ok,
worked.
--- dummy.c~ Mon Feb 6 21:12:00 1995
+++ dummy.c Mon Feb 6 21:12:00 1995
@@ -142,7 +142,7 @@
}
static struct device dev_dummy = {
- "dummy0\0 ",
+ "dummy\0 ",
0, 0, 0, 0,
0x0, 0,
0, 0, 0, NULL, dummy_probe };
----------
I expect that this is just a typie, or is there any special reason for this
that I might have missed? Perhaps it would be cool if the thing would be
called dummy0, and if I install another module, that device would be called
dummy1 etc? I can't imagine a use for it, but perhaps it would be cool ;-)
So, I then started a slip connection. (modules slip and slhc installed, of
course) Didn't work. No ping, nothing, the route was correct, dip -v looked
exactly the same, except that obviously NOTHING was send to the modem after
mode CSLIP. Since the module slhc printed the CSLIP at startup, I assumed
that this would work. It didn't. I don't know why, I would be more than
happy to know.
I don't know if this just was a connection of unlucky events, but I never
had so much trouble updating to a new kernel. Right now I am building the
static version, without modules...
Perhaps at least my fixes(?) will get into the next kernel patch.
--
Lars Marowsky-Bree Voice: +49-571-63663 PGP-key via return receipt
VirNet: 9:492/7158 Fido: 2:2449/620.16 Mail: lmb@pointer.in-minden.de
PGP fingerprint: CF FC 3A F0 86 F1 D3 EB 79 8A CF 75 4F 4C 81 DF
