Hi Greg, Stephen,

On Thu, 10 May 2012 19:23:55 +0200
Stephen Warren [off-list ref] wrote:

On 05/10/2012 01:42 AM, Hiroshi DOYU wrote:

quoted

Tegra Memory Controller(MC) driver for Tegra20
Added to support MC General interrupts, mainly for IOMMU(GART).

Greg, you appear to have been commiting all of drivers/memory. I assume
you'll take patches 1 and 3 in this series, and I will take 2 and 4
through the Tegra tree? I can take them all through the Tegra tree if
you want. Taking the .dts patches through your tree might cause
conflicts in 3.6, since I anticipate some DT-wide cleanup in 3.6.

Hiroshi, one question below (same for patch 3)

quoted

diff --git a/arch/arm/mach-tegra/Kconfig b/arch/arm/mach-tegra/Kconfig

quoted

@@ -20,6 +20,8 @@ config ARCH_TEGRA_2x_SOC
 	select PL310_ERRATA_727915 if CACHE_L2X0
 	select PL310_ERRATA_769419 if CACHE_L2X0
 	select CPU_FREQ_TABLE if CPU_FREQ
+	select MEMORY
+	select TEGRA20_MC

Does ARCH_TEGRA_2x_SOC /have/ to select these?

I'd be inclined to drop the change to mach-tegra/Kconfig. As rationale,
the code works fine without the driver right now. You'd need to make
config TEGRA*_MC user-visible/selectable though.

You could replace this change with a patch to tegra_defconfig which
enables MEMORY and TEGRA*_MC, or perhaps make config TEGRA20_MC default y.

I'm sending the fixup patches, following this mail, since Greg has
already put the original patches in queue. I think that the following
fixup patches could be squashed into the original ones. The fixup
patches are:

  [PATCH 1/3] ARM: tegra20: Make MC optional in Kconfig
  [PATCH 2/3] ARM: tegra30: Make MC optional in Kconfig
  [PATCH 3/3] ARM: tegra: Make TEGRA{20,30}_MC selectable in defconfig

Let me know if it's better to resend the complete patches again.