On Sun, Oct 31, 1999 at 11:59:58PM -0500, [EMAIL PROTECTED] wrote:
> On Sun, 31 Oct 1999, Shawn T. Rutledge wrote:
> > Serial driver version 4.27 with HUB-6 MANY_PORTS SHARE_IRQ enabled
>
> I don't think this proves the kernel recognizes the board. This just
> says support for such a board is enabled, if one should crop up.
Good point.
> > Failing actual knowledge of this old beastie on the list do we know who
> > wrote the driver support for it so I can ask him?
> >
> Use the source, Shawn. If you are lucky the author and/or maintainer
> will have left their names. I can't even figure out where is the source
> for this one, though.
I did that. If you rgrep -i for hub6 you find a few references to it;
mostly in include/asm-i386/serial.h. That's how I figured out the major/
minor dev numbers (as shown on the web page I mentioned in the last email).
But there is no authorship credit, either there or in serial.c where credit
is given to some other contributors.
Well now I have changed it to irq7 so it won't conflict, and -I think- to
I/O port 362. It has DIP switches labeled A1-A8 which I assume are
part of the address part of the ISA bus - e.g. it's zero-based so A1 is
really the 2's bit. The board labeling appears to indicate that the
default settings are A1 and A8 on, the rest off. So if A9 (for which
there isn't a switch) is also held high, then 0x200 + 0x100 + 0x002 does
indeed add up to 0x302 which is what is claimed as the default port in
serial.h. So following that logic I also set A5 and A6 high to get 0x362
(because 0x302 overlaps eth0's port range).
So then I messed around with setserial. I added this line to 0setserial
${SETSERIAL} -b /dev/ttyS32 uart 8250 port 0x362 irq 7 ${STD_FLAGS}
which made it somewhat happier - I could actually connect to the port
in kermit without it complaining ttyS32's not a valid port. But no lights
flickered on the modem when I typed and "atz" didn't return anything.
If I put in the lines for the other 5 ports it complains that the IO
port is already in use - but for the HUB6 that's normal as I understand
it; selecting which UART to talk to is part of the data being sent to the
port, so you use the same two IO ports for all 6 UARTS. So says serial.h
anyway. setserial evidently doesn't know this.
I also tried the autoconfig option to setserial, and in that case after
a reboot it doesn't detect the port at all.
So help me out guys, I'm all out of clues.
--
_______ 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]
