Work around a QEMU bug by reordering the drivers in the Linux kernel.

See http://lists.gnu.org/archive/html/qemu-devel/2010-02/msg00892.html

QEMU's -hda and -hdb options initialize drives on a power mac IDE controller,
and its -hdc and -hdd options initialize drives on a cmd646 controller.  (They
do this because they expect hdc to be a CDROM and DMA for those doesn't
work right on the pmac controller, so they added a second controller.)

A non-modular Linux kernel always initialies the cmd646 driver before the
pmac driver, and thus the disks on the cmd646 become /dev/hda and /dev/hdb,
and the pmac ones become /dev/hdc and /dev/hdd.  Thus -hda initializes hdc,
and so on, working differently than every other qemu target and confusing our
generic boot script.

The real fix is in QEMU (making their command line arguments match up
with what the Linux will name the drives), but they won't do it.  The Linux
kernel's behavior hasn't changed in 10 years, but here's a patch to do
so, to work around the QEMU bug.

diff --git a/drivers/ide/Makefile b/drivers/ide/Makefile
index 81df925..d9318a1 100644
--- a/drivers/ide/Makefile
+++ b/drivers/ide/Makefile
@@ -39,6 +39,7 @@ obj-$(CONFIG_BLK_DEV_ALI15X3)		+= alim15x3.o
 obj-$(CONFIG_BLK_DEV_AMD74XX)		+= amd74xx.o
 obj-$(CONFIG_BLK_DEV_ATIIXP)		+= atiixp.o
 obj-$(CONFIG_BLK_DEV_CELLEB)		+= scc_pata.o
+obj-$(CONFIG_BLK_DEV_IDE_PMAC)		+= pmac.o
 obj-$(CONFIG_BLK_DEV_CMD64X)		+= cmd64x.o
 obj-$(CONFIG_BLK_DEV_CS5520)		+= cs5520.o
 obj-$(CONFIG_BLK_DEV_CS5530)		+= cs5530.o
@@ -76,8 +77,6 @@ obj-$(CONFIG_IDEPCI_PCIBUS_ORDER)	+= ide-scan-pci.o
 
 obj-$(CONFIG_BLK_DEV_CMD640)		+= cmd640.o
 
-obj-$(CONFIG_BLK_DEV_IDE_PMAC)		+= pmac.o
-
 obj-$(CONFIG_IDE_H8300)			+= ide-h8300.o
 
 obj-$(CONFIG_IDE_GENERIC)		+= ide-generic.o