patch to fix the IDE configuration on EPIA boards. At some point this

broke and stopped FILO
from being able to boot.

The fix is a simple one line change  plus a comment to
src/mainboard/via/epia/auto.c to write to the IDE
configuration register 0x42 .  This has always been done here, however
at some point
something broke it.

The same register was also being set correctly  in ide_init(), however
for some reason
this does not work. Possibly the register needs to be set before the IDE
peripheral is enabled
or maybe it is a timing issue.

The section of code in ide_init() (
src/southbridge/via/vt8231/vt8231_ide.c ) that does
write to register 0x42 has been commented out as it is superfluous
and I have added a comment to indicate the reason, should someone at a
future date wonder
why.

I have also changed the default COM speed from 19200 to 115200 in
src/mainboard/via/epia/Options.lb
There has been mention before about the EPIA board not being able to use
115200 but I have seen
no such problems with my board.

Signed-off-by: Ben Hewson <ben@hewson-venieri.com>
This patch worked for me and allowed me to boot Debian kernel
2.5.16-4-486  on an epia 800 mhz system.  It is able to consistently get
through the initialization and start init now.

However, after that it crashes at various points in the boot process.

Acked-by: Alex Mauer <hawke@hawkesnest.net>

Note from comitter: I am commiting this, although:
1. it's not the exact right way to fix it up, the chip.h for the sb
should change
2. Alex reports problems, which are almost certainly memory issues. 

But it is as close as we've gotten. I can't test it.

Ron Minnich



git-svn-id: svn://svn.coreboot.org/coreboot/trunk@2633 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1

src/mainboard/via/epia/Options.lb

@@ -51,7 +51,7 @@ uses MAXIMUM_CONSOLE_LOGLEVEL
 
 default CONFIG_CONSOLE_SERIAL8250=1
 ## Select the serial console baud rate
-default TTYS0_BAUD=19200
+default TTYS0_BAUD=115200
 
 # Select the serial console base port
 default TTYS0_BASE=0x3f8

src/mainboard/via/epia/auto.c

@@ -66,8 +66,15 @@ static void enable_mainboard_devices(void)
 	/* we do this here as in V2, we can not yet do raw operations 
 	 * to pci!
 	 */
-        dev += 0x100; /* ICKY */
+	/* changed this to work correctly on later revisions of LB.
+	* The original dev += 0x100; stopped working. It also appears
+	* that if this is not set here, but in ide_init() only, the IDE
+	* does not work at all. I assume it needs to be set before something else,
+	* possibly before enabling the IDE peripheral, or it is a timing issue.
+	* Ben Hewson 29 Apr 2007.
+	*/
 
+	dev = pci_locate_device(PCI_ID(0x1106,0x0571), 0);
 	pci_write_config8(dev, 0x42, 0);
 }
 

src/southbridge/via/vt8231/vt8231_ide.c

@@ -14,7 +14,12 @@ static void ide_init(struct device *dev)
 	if (!conf->enable_native_ide) {
 		// Run the IDE controller in 'compatiblity mode - i.e. don't use PCI
 		// interrupts.  Using PCI ints confuses linux for some reason.
-		
+		/* Setting reg 0x42 here does not work. It is set in mainboard/auto.c
+		* It probably can only be changed while the IDE is disabled
+		* or it is possibly a timing issue. Ben Hewson 29 Apr 2007.
+		*/
+
+		/*		
 		printk_info("%s: enabling compatibility IDE addresses\n", __FUNCTION__);
 		enables = pci_read_config8(dev, 0x42);
 		printk_debug("enables in reg 0x42 0x%x\n", enables);
@@ -22,6 +27,7 @@ static void ide_init(struct device *dev)
 		pci_write_config8(dev, 0x42, enables);
 		enables = pci_read_config8(dev, 0x42);
 		printk_debug("enables in reg 0x42 read back as 0x%x\n", enables);
+		*/
 	}
 	
 	enables = pci_read_config8(dev, 0x40);