I have two systems on which I can't get the motherboard's built-in serial
ports to work. One is the system that used to run my ax.25 stuff. It
used to be a dual-pentium board that didn't detect the second chip (a
Neptune chipset ideosyncracy) so I replaced it with a Tyan dual-pentium
board. SMP is working fine now but the new board also has built-in serial
ports (the old one didn't) and so I connected my TNC to one of them. At
first I didn't change anything, it was the same port with the same startup
scripts, but it wasn't working so I've been changing everything I can think
of. First I tried the other port to make sure I hadn't gotten them mixed
up (ttyS0 vs ttyS1). I replaced the brand-new serial connectors (the kind
that has two ribbon cables going to two ports on a slot insert) with another
pair of brand-new ones. I tried hooking up a modem instead of the TNC
so I could use kermit to send commands to it. At one point the modem lights
would at least blink when I sent it characters but now it won't even do
that. I have never got any kind of response out of the modem when using
these ports. But everything looks OK config-wise; when the system boots
the ascii-graphic box that the BIOS puts up which shows all the hardware
information shows two serial ports at 3f8/4 and 2f8/3. When the kernel
boots it says the same thing, and that they are 16550s, and when 0setserial
runs and configures the ports, afterwards it says the same thing again.
This works whether I use the autoconfig lines in 0setserial or the manual
config. Oh and I tried it with both 2.0.36 kernel and a 2.2 series (I
forget which but it is the next to latest or so). There is no modem in
that system or anything else I know of that could conflict. I had a HUB6
in that system but I can't figure out how to make that work either so I
took it back out to make sure it's not conflicting.
[electron:~][2:20:33pm] cat /proc/ioports
0000-001f : dma1
0020-003f : pic1
0040-005f : timer
0060-006f : keyboard
0070-007f : rtc
0080-009f : dma page reg
00a0-00bf : pic2
00c0-00df : dma2
00f0-00ff : npu
01f0-01f7 : ide0
02f8-02ff : serial(set)
0300-030f : 3c509
0350-035f : 3c509
03c0-03df : vga+
03f6-03f6 : ide0
03f8-03ff : serial(set)
[electron:~][2:20:46pm] cat /proc/interrupts
0: 122772 timer
1: 2 keyboard
2: 0 cascade
3: 50 + serial
5: 18341 3c509
8: 2 + rtc
9: 8677 3c509
13: 1 math error
14: 65156 + ide0
[electron:~][2:22:20pm] dmesg | grep tty
tty00 at 0x03f8 (irq = 4) is a 16550A
tty01 at 0x02f8 (irq = 3) is a 16550A
Both systems are running Debian, originally 2.1 but I have upgraded a lot
of the packages to the ones from Potato. There are the following libc's:
[electron:~][2:18:03pm] ll /lib/*libc*
-rwxr-xr-x 1 root root 1068322 Jul 25 05:49 /lib/libc-2.1.2.so
lrwxrwxrwx 1 root root 13 Nov 22 1998 /lib/libc.so.4 -> libc.so.4.7.2
-r-x------ 1 root root 634880 Nov 22 1998 /lib/libc.so.4.7.2
lrwxrwxrwx 1 root root 14 Mar 7 1999 /lib/libc.so.5 -> libc.so.5.4.46
-rw-r--r-- 1 root root 586720 Feb 8 1999 /lib/libc.so.5.4.46
lrwxrwxrwx 1 root root 13 Jul 26 00:26 /lib/libc.so.6 -> libc-2.1.2.so
The other system that's having trouble is a no-name celeron board. I wanted
to use my CM11a X-10 controller on that system, and again the existing
software worked fine on a previous system, but no more. Now I have had to
relocate the CM11a to a 486 with good old reliable ISA serial ports.
I haven't spent as much time trying to troubleshoot that system but I
have a hunch the problem is the same on both of them.
So my question is what is so darn different about motherboard serial ports?
In the past I have always used ISA cards for serial. Well except on my
third system, a Tyan dual PII machine, whose built-ins work just fine now
that I have the IRQ conflicts worked out.
--
_______ http://www.bigfoot.com/~ecloud
(_ | |_) [EMAIL PROTECTED] finger [EMAIL PROTECTED]
__) | | \__________________________________________________________________
-
To unsubscribe from this list: send the line "unsubscribe linux-serial" in
the body of a message to [EMAIL PROTECTED]
