Sad: tried code but no go, MAC address within problem range 

Forum: PCMCIA Network Adapter Issues
Re: Question: trouble with netgear FA410TXC (Matthew Austin)
Re: Feedback: Hardware Version (Jochen Friedrich)
Date: 2000, Mar 27
From: Matthew Austin yogi

Hi Jochen, thanks for the code, but I am still stuck. I tried your code (with some help compiling provided by my UNIX SA ;-)) and I get core dumps for any of the tried options with both the old enabler code and the new enabler code. Below is /var/log/messages and my attempt with the new code. It does look like my MAC address is within the problem range. Mar 27 12:06:33 yogimind cardmgr[367]: socket 0: NetGear FA410TXC Fast Ethernet Mar 27 12:06:33 yogimind cardmgr[367]: executing: 'insmod /lib/modules/2.2.12-20/net/8390.o' Mar 27 12:06:33 yogimind cardmgr[367]: executing: 'insmod /lib/modules/2.2.12-20/pcmcia/pcnet_cs.o' Mar 27 12:06:33 yogimind kernel: eth0: NE2000 Compatible: io 0x800, irq 5, hw_addr 00:E0:98:77:A9:5D Mar 27 12:06:33 yogimind cardmgr[367]: executing: './network start eth0' Mar 27 12:06:35 yogimind kernel: eth0: found link beat Mar 27 12:06:36 yogimind kernel: eth0: lost link beat [maustin@yogimind maustin]$ gcc -Wall -O -o fa_select eth.c eth.c: In function `writea': eth.c:104: warning: unused variable `j' eth.c: At top level: eth.c:150: warning: return type of `main' is not `int' [maustin@yogimind maustin]$ ./fa_select eth0 0 Segmentation fault (core dumped) [maustin@yogimind maustin]$ ./fa_select eth0 1 Segmentation fault (core dumped) [maustin@yogimind maustin]$ ./fa_select eth0 2 Segmentation fault (core dumped) [maustin@yogimind maustin]$ ./fa_select eth0 3 Segmentation fault (core dumped) [maustin@yogimind maustin]$ -matthew

[ Next-in-Thread ]  [ Next Message ] 

Messages Inline: [ 0 ]  [ 1 ] 

More: What version Re: Sad: tried code but no go, MAC address within problem range (Matthew Austin) Date: 2000, Mar 31 From: Jochen Friedrich jochen1

Hi Matthew, what version of linux, libc and gcc are you using? Did you run the program as root (it needs root for ioperm() to enable user space I/O access)? Do you have a chance to add "-g" to your compile parameters and run gdb ./fa_select? (Then do a "run eth0 0" and record the result of "bt"). Anyways, I placed a binary (needs glibc 2.1) on http://www.gxsnmp.org/fa_select -- Jochen

Sad: still no go even with your binary . . . here's the details Re: More: What version (Jochen Friedrich) Date: 2000, Mar 31 From: Matthew Austin yogi

#info you requested kernel-2.2.12-20 libc-5.3.12-31 glibc-2.1.2-11 #run with your binary from http://www.gxsnmp.org/fa_select [root@yogimind maustin]# gdb fa_select GNU gdb 4.18 Copyright 1998 Free Software Foundation, Inc. GDB is free software, covered by the GNU General Public License, and you are welcome to change it and/or distribute copies of it under certain conditions. Type "show copying" to see the conditions. There is absolutely no warranty for GDB. Type "show warranty" for details. This GDB was configured as "i386-redhat-linux"... (gdb) run eth0 0 Starting program: /home/maustin/fa_select eth0 0 Program received signal SIGSEGV, Segmentation fault. 0x80485d2 in outb () (gdb) bt #0 0x80485d2 in outb () #1 0x804877d in reset () #2 0x8048e2e in main () #3 0x400301eb in __libc_start_main (main=0x8048d08 <main>, argc=3,     argv=0xbffffce4, init=0x80483bc <_init>, fini=0x8048eec <_fini>,     rtld_fini=0x4000a610 <_dl_fini>, stack_end=0xbffffcdc)     at ../sysdeps/generic/libc-start.c:90 (gdb) q #run with your updated source and suggested flag [root@yogimind maustin]# gcc -g -Wall -O -o fa_select eth.c eth.c: In function `writea': eth.c:104: warning: unused variable `j' eth.c: At top level: eth.c:150: warning: return type of `main' is not `int' [root@yogimind maustin]# gdb ./fa_select GNU gdb 4.18 Copyright 1998 Free Software Foundation, Inc. GDB is free software, covered by the GNU General Public License, and you are welcome to change it and/or distribute copies of it under certain conditions. Type "show copying" to see the conditions. There is absolutely no warranty for GDB. Type "show warranty" for details. This GDB was configured as "i386-redhat-linux"... (gdb) run eth0 0 Starting program: /home/maustin/./fa_select eth0 0 Program received signal SIGSEGV, Segmentation fault. 0x8048667 in reset (port=2076) at eth.c:60 60 outb(0x08, port); (gdb) bt #0 0x8048667 in reset (port=2076) at eth.c:60 #1 0x8048a68 in main (argc=3, argv=0xbffffce4) at eth.c:184 (gdb) q The program is running. Exit anyway? (y or n) y

ioperm() vs iopl() issue Re: More: What version (Jochen Friedrich) Date: 2000, Mar 31 From: David Hinds <dhinds@pcmcia.sourceforge.org>

ioperm() only works for ports below 0x400. For higher ports, you need to use iopl(3) instead. -- Dave

Agree: success!!!!!!!! Re: ioperm() vs iopl() issue (David Hinds) Date: 2000, Apr 01 From: Matthew Austin yogi

Excellent! Dave I removed the exclusion on the port range previously in place (per trouble shooting section on this site) to try and isolate the issue from a possible hardware conflict with another device. ------------------------------------------------------------- From my original post . . . ------------------------------------------------------------- "I have forced a change in io port and IRQ to see if that was an issue as recommended by changing config.opts . . . no change." -------------------------------------------------------------   After removing those changes Jochen's binary now works, but I imagine my own compile of his source would work now as well. Thanks to you both. -matthew PS I was getting ready to purchase a 3com 3c575 off ebay as cowboyneal reported that working flawlessly for him . . . ;-)

[ Add Message ]  to: "tried code but no go, MAC address within problem range"

[ Members ]  [ Subscribe ]  [ Admin Mode ] 
[ Show Frames ]  [ Help for HyperNews at pcmcia-cs.sourceforge.net 1.10 ]