Tegra: organize memory/mmio apertures to decrease memmap latency
This patch organizes the memory and mmio maps linearly, to make the
mmap_add_region process faster. The microsecond timer has been moved
to individual platforms instead of making it a common step, as it
further speeds up the memory map creation process.
Change-Id: I6fdaee392f7ac5d99daa182380ca9116a001f5d6
Signed-off-by: Varun Wadekar <vwadekar@nvidia.com>
diff --git a/plat/nvidia/tegra/soc/t186/plat_setup.c b/plat/nvidia/tegra/soc/t186/plat_setup.c
index fd109e5..c56f71c 100644
--- a/plat/nvidia/tegra/soc/t186/plat_setup.c
+++ b/plat/nvidia/tegra/soc/t186/plat_setup.c
@@ -91,6 +91,8 @@
MT_DEVICE | MT_RW | MT_SECURE),
MAP_REGION_FLAT(TEGRA_PMC_BASE, 0x40000U, /* 256KB */
MT_DEVICE | MT_RW | MT_SECURE),
+ MAP_REGION_FLAT(TEGRA_TMRUS_BASE, 0x1000U, /* 4KB */
+ MT_DEVICE | MT_RO | MT_SECURE),
MAP_REGION_FLAT(TEGRA_SCRATCH_BASE, 0x10000U, /* 64KB */
MT_DEVICE | MT_RW | MT_SECURE),
MAP_REGION_FLAT(TEGRA_MMCRAB_BASE, 0x60000U, /* 384KB */