[PATCH] x86, soc: Fix fallout from the previous changes

[PATCH] x86, soc: Fix fallout from the previous changes

[Alan Cox]

From: Alan Cox <alan@linux.intel.com>

The PCI fixups get executed based upon whether they are linked in. We need
to avoid executing them if we boot a dual SoC/PC type kernel on a PC class
system.

Signed-off-by: Alan Cox <alan@linux.intel.com>
---

 arch/x86/pci/mrst.c |   12 +++++++++++-
 1 files changed, 11 insertions(+), 1 deletions(-)


diff --git a/arch/x86/pci/mrst.c b/arch/x86/pci/mrst.c
index c5e81a4..140942f 100644
--- a/arch/x86/pci/mrst.c
+++ b/arch/x86/pci/mrst.c
@@ -43,6 +43,8 @@
 #define PCI_FIXED_BAR_4_SIZE	0x14
 #define PCI_FIXED_BAR_5_SIZE	0x1c
 
+static int pci_soc_mode = 0;
+
 /**
  * fixed_bar_cap - return the offset of the fixed BAR cap if found
  * @bus: PCI bus
@@ -233,10 +235,11 @@ struct pci_ops pci_mrst_ops = {
  */
 int __init pci_mrst_init(void)
 {
-	printk(KERN_INFO "Moorestown platform detected, using MRST PCI ops\n");
+	printk(KERN_INFO "Intel MID platform detected, using MID PCI ops\n");
 	pci_mmcfg_late_init();
 	pcibios_enable_irq = mrst_pci_irq_enable;
 	pci_root_ops = pci_mrst_ops;
+	pci_soc_mode = 1;
 	/* Continue with standard init */
 	return 1;
 }
@@ -246,6 +249,10 @@ int __init pci_mrst_init(void)
  */
 static void __devinit pci_d3delay_fixup(struct pci_dev *dev)
 {
+	/* PCI fixups are effectively decided compile time. If we have a dual
+	   SoC/non-SoC kernel we don't want to mangle d3 on non SoC devices */
+        if (!pci_soc_mode)
+            return;
 	/* true pci devices in lincroft should allow type 1 access, the rest
 	 * are langwell fake pci devices.
 	 */
@@ -274,6 +281,9 @@ static void __devinit pci_fixed_bar_fixup(struct pci_dev *dev)
 	u32 size;
 	int i;
 
+	if (!pci_soc_mode)
+		return;
+
 	/* Must have extended configuration space */
 	if (dev->cfg_size < PCIE_CAP_OFFSET + 4)
 		return;

Re: [PATCH] x86, soc: Fix fallout from the previous changes

[Ingo Molnar]

* Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:

> From: Alan Cox <alan@linux.intel.com>
> 
> The PCI fixups get executed based upon whether they are linked in. We need
> to avoid executing them if we boot a dual SoC/PC type kernel on a PC class
> system.
> 
> Signed-off-by: Alan Cox <alan@linux.intel.com>
> ---
> 
>  arch/x86/pci/mrst.c |   12 +++++++++++-
>  1 files changed, 11 insertions(+), 1 deletions(-)

This would be for Jesse's PCI tree I think.

Thanks,

	Ingo