[coreboot] [v2] r4339 - trunk/coreboot-v2/src/southbridge/amd/rs690

Author: zbao
Date: 2009-06-05 12:00:07 +0200 (Fri, 05 Jun 2009)
New Revision: 4339

Modified:
   trunk/coreboot-v2/src/southbridge/amd/rs690/rs690_pcie.c
Log:
After I modify the pci_ext_read_config32 and pci_ext_read_config32, the step 6a
starts to play its role. Then the system hangs at HDA init. I dont know what the
VC1 is. The RPR says "Optional Features (only needed if CMOS option is enabled)"
in 5.10.2. Before I know what it is, I think it is better to skip it.

Tested on dbm690t.

Add comment from Rudolf,
"
VC is virtual channel. Its used for isochronous transfer of data to sound card.
The virtual channel guarantee "on time" delivery. In other words it sets up a
channel for data to sound card, which means that that arrivs in time and there will
be no interuptions in audio stream.

http://www.microsoft.com/whdc/connect/pci/wlp_interrupt.mspx
"

Signed-off-by: Zheng Bao 
Acked-by: Carl-Daniel Hailfinger

Modified: trunk/coreboot-v2/src/southbridge/amd/rs690/rs690_pcie.c
===================================================================
--- trunk/coreboot-v2/src/southbridge/amd/rs690/rs690_pcie.c	2009-06-05 00:22:25 UTC (rev 4338)
+++ trunk/coreboot-v2/src/southbridge/amd/rs690/rs690_pcie.c	2009-06-05 10:00:07 UTC (rev 4339)
@@ -276,6 +276,9 @@
    /* step 6a: VCI */
    sb_dev = dev_find_slot(0, PCI_DEVFN(8, 0));
    if (port == 8) {
+		/* The code below between #if and #endif causes a hang on HDA init.
+		 * So we skip it. */
+#if 0
    	/* Clear bits 7:1 */
    	pci_ext_write_config32(nb_dev, sb_dev, 0x114, 0x3f << 1, 0 << 1);
    	/* Maps Traffic Class 1-7 to VC1 */
@@ -284,7 +287,7 @@
    	pci_ext_write_config32(nb_dev, sb_dev, 0x120, 7 << 24, 1 << 24);
    	/* Enables VC1 */
    	pci_ext_write_config32(nb_dev, sb_dev, 0x120, 1 << 31, 1 << 31);
-#if 0
+
    	do {
    		reg16 = pci_ext_read_config32(nb_dev, sb_dev, 0x124);
    		reg16 &= 0x2;