#2 Backport of pci cardbus number enumeration from 2.6 to 2.4.29

From: Paul Marrons
Date: Wed Jan 19 2005 - 22:05:05 EST


Martin,

I forgot to attach the actually patch file in my last mail. It's attached this time with a copy of my previous email below:

Hi Martin,

To overcome a problem with my laptop cardbus not being assigned the correct bus number in 2.4.29 (I originally did this change for 2.4.27) I backported a portion of the code in the 2.6 kernel drivers/pci/pci.c file. I did this because I noticed that only 2.6 assigned the correct bus number and I specifically need to run 2.4.X because of a driver I need that is not 2.6 compatible. Basically without this change on my laptop (Thinkpad 240) both the main PCI bus and cardbus bridge both get assigned bus#0 and as a result any cardbus devices present are not correctly detected and allocated any resources, in addition the /proc/bus/pci contains two '0' entries and tools such as lspci fail to work.

I am aware of people overcoming this problem (with my model of laptop) by setting defining pcibios_assign_all_busses() as 1. But this backport is a superior solution to the problem.

The few small changes are isolated to pci_add_new_bus and pci_scan_bridge. I hope you will be able to incorporate them into the next 2.4 kernel release.

If there is anything else I need to do please let me know.

Regards,

Paul Marrons. --- pci.c.orig 2004-11-17 11:54:21.000000000 +0000
+++ pci.c 2005-01-20 02:01:26.000000000 +0000
@@ -1,3 +1,4 @@
+
/*
* $Id: pci.c,v 1.91 1999/01/21 13:34:01 davem Exp $
*
@@ -1238,8 +1239,6 @@
* Allocate a new bus, and inherit stuff from the parent..
*/
child = pci_alloc_bus();
- if (!child)
- return NULL;

list_add_tail(&child->node, &parent->children);
child->self = dev;
@@ -1286,29 +1285,25 @@

pci_read_config_dword(dev, PCI_PRIMARY_BUS, &buses);
DBG("Scanning behind PCI bridge %s, config %06x, pass %d\n", dev->slot_name, buses & 0xffffff, pass);
- if ((buses & 0xffff00) && !pcibios_assign_all_busses()) {
+ if ((buses & 0xffff00) && !pcibios_assign_all_busses() && !is_cardbus) {
/*
* Bus already configured by firmware, process it in the first
* pass and just note the configuration.
*/
+ unsigned busnr,cmax;
+ busnr = (buses >> 8)&0xFF;
if (pass)
return max;
-
- child = pci_add_new_bus(bus, dev, 0);
- if (!child)
- return max;
-
+ child = pci_add_new_bus(bus, dev, busnr);
child->primary = buses & 0xFF;
- child->secondary = (buses >> 8) & 0xFF;
child->subordinate = (buses >> 16) & 0xFF;
child->number = child->secondary;
- if (!is_cardbus) {
- unsigned int cmax = pci_do_scan_bus(child);
- if (cmax > max) max = cmax;
- } else {
- unsigned int cmax = child->subordinate;
- if (cmax > max) max = cmax;
- }
+
+ cmax = pci_do_scan_bus(child);
+ if(cmax > max)
+ max=cmax;
+ if(child->subordinate > max)
+ max = child->subordinate;
} else {
/*
* We need to assign a number to this bus which we always
@@ -1322,13 +1317,19 @@
pci_write_config_word(dev, PCI_STATUS, 0xffff);

child = pci_add_new_bus(bus, dev, ++max);
- if (!child)
- return max;
-
buses = (buses & 0xff000000)
| ((unsigned int)(child->primary) << 0)
| ((unsigned int)(child->secondary) << 8)
| ((unsigned int)(child->subordinate) << 16);
+ /*
+ * yenta.c forces a secondary latency timer of 176.
+ * Copy that behaviour here.
+ */
+ if (is_cardbus) {
+ buses &= ~0xff000000;
+ buses |= 176 << 24;
+ }
+
/*
* We need to blast all three values with a single write.
*/