Re: The new SSB subsystem for bcm43xx (and others)

[Matthew Garrett <mjg59@srcf.ucam.org>]

On Sat, Feb 10, 2007 at 10:03:57PM +0100, Michael Buesch wrote:
> On Saturday 10 February 2007 21:46, Matthew Garrett wrote:
> > I'm testing with your bcm43xx git tree, which I'm guessing is the 
> > current ssb code. The only problem I've found is that there doesn't seem 
> > to be any sysfs relationship between the ssb bus and (in this case) the 
> > PCI device that it's associated with. Is this fixable? Right now it 
> > appears as an entirely separate branch of the device tree, which doesn't 
> > seem quite right.
> 
> I guess that's fixable, but I didn't care too much, yet.

Ok, here's a patch. The stack of PCMCIA headers are needed to get the 
pcmcia_device structure. Seems to work fine for PCI - I don't have any 
PCMCIA devices. The SSB devices now appear underneath the PCI device 
rather than in the top level of /sys/devices.

Signed-off-by: Matthew Garrett <mjg59@srcf.ucam.org>

---

diff --git a/drivers/ssb/scan.c b/drivers/ssb/scan.c
index 64f94b8..b16cee9 100644
--- a/drivers/ssb/scan.c
+++ b/drivers/ssb/scan.c
@@ -15,6 +15,11 @@
 #include <linux/ssb/ssb.h>
 #include <linux/ssb/ssb_regs.h>
 #include <linux/pci.h>
+#include <pcmcia/cs_types.h>
+#include <pcmcia/cs.h>
+#include <pcmcia/cistpl.h>
+#include <pcmcia/ciscode.h>
+#include <pcmcia/ds.h>
 #include <asm/io.h>
 
 #include "ssb_private.h"
@@ -306,8 +311,12 @@ int ssb_bus_scan(struct ssb_bus *bus,
 		dev->id.vendor = (idhi & SSB_IDHIGH_VC) >> SSB_IDHIGH_VC_SHIFT;
 		dev->core_index = i;
 		dev->bus = bus;
-		if ((dev->bus->bustype == SSB_BUSTYPE_PCI) && (bus->host_pci))
+		if ((dev->bus->bustype == SSB_BUSTYPE_PCI) && (bus->host_pci)) {
 			dev->irq = bus->host_pci->irq;
+			dev->dev.parent = &bus->host_pci->dev;
+		} else if (dev->bus->bustype == SSB_BUSTYPE_PCMCIA) {
+		        dev->dev.parent = &bus->host_pcmcia->dev;
+		}
 
 		ssb_dprintk(KERN_INFO PFX
 			    "Core %d found: %s "

-- 
Matthew Garrett | mjg59@srcf.ucam.org