OMAP hwmod: convert hwmod to use hardware clock names rather than clkdev dev+con

The OMAP hwmod core code is intended to use SoC IP block description
structures that are autogenerated from TI's OMAP hardware database.
Currently the hwmod code uses clkdev device + connection addressing to
identify clocks.  This causes problems in the hwmod autogeneration
process, since the TI hardware database doesn't use platform_device or
clkdev addressing; it uses a single clock signal name string, which
tends to bear some resemblance to what is used in the OMAP TRMs.  This
patch converts the hwmod code and existing data to use omap_clk_get_by_name(),
introduced in the previous patch.

Signed-off-by: Paul Walmsley <paul@pwsan.com>
Cc: BenoƮt Cousson <b-cousson@ti.com>
Cc: Kevin Hilman <khilman@deeprootsystems.com>
diff --git a/arch/arm/mach-omap2/omap_hwmod_2420.h b/arch/arm/mach-omap2/omap_hwmod_2420.h
index a9ca1b9..5932c1d 100644
--- a/arch/arm/mach-omap2/omap_hwmod_2420.h
+++ b/arch/arm/mach-omap2/omap_hwmod_2420.h
@@ -117,8 +117,7 @@
 /* MPU */
 static struct omap_hwmod omap2420_mpu_hwmod = {
 	.name		= "mpu_hwmod",
-	.clkdev_dev_id	= NULL,
-	.clkdev_con_id	= "mpu_ck",
+	.main_clk	= "mpu_ck",
 	.masters	= omap2420_mpu_masters,
 	.masters_cnt	= ARRAY_SIZE(omap2420_mpu_masters),
 	.omap_chip	= OMAP_CHIP_INIT(CHIP_IS_OMAP2420),