[LinuxBIOS] get some output on serial console of commel lv671 board, but not readable one

Corey Osgood corey.osgood at gmail.com
Sat Jun 16 05:31:41 CEST 2007

Stefan Reinauer wrote:
> * Corey Osgood <corey.osgood at gmail.com> [070615 21:51]:
>> In the future, "svn diff" is much easier to use and more readable.
>>> +
>>> +	device_t dev;
>>> +	/* enable the decoding of superio ranges to the LPC interface */
>>> +	dev = pci_locate_device(PCI_ID(0x8086, 0x24c0), 0);
>>> +	pci_write_config16(dev, 0xE6, 0x3cff);
>> Please add this line:
>> pci_write_config16(dev, 0xE0, 0x10);
>> and see if it helps. If it doesn't, remove that block entirely, and try
>> again (this shouldn't be necessary for basic early serial anyways)
> Why do you think so? The ICH4 has LPC decodes and Serial port decodes
> disabled on poweron. Or the datasheets lie. (Both is possible, in any
> case)
Interesting...the ICH also says the same thing, but I didn't have to do
any such thing to get serial output. The intel xe7501devkit, which is
ICH3, doesn't do it either, nor does the jarrell, which is ICH5R. Given
that, I'd say the datasheets lie, and in any case it couldn't hurt to
try it ;)

Also, the tyan s2875 has the same superio as yours, and doesn't do any
special setup for it at all, just w83627hf_enable_serial. You might want
to take a quick peak at the auto.c there. From a running system, you
should be able to determine if it's on 0x2e or 0x4e, just start doing
inb from the command line, using 0x2e/4e as a starting point and
counting up, when you start finding non-0xff values you've found
_something_, although there's no guarantee its the serial port. Some
distros (*cough* ubuntu *cough*) don't allow you to do anything under
0x100 though, but iirc arch and debian do it fine.


More information about the coreboot mailing list