refactor(qemu): make L0GPT size configurable

Add a new parameter to make the size of the L0GPT configurable based on
the amount of memory available on a platform. That way platform with a
wider physical address range can be supported.

No change in functionality.

Change-Id: I5b7b4968636d61929ad6ebdc05c389291cf510b1
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
diff --git a/plat/qemu/common/qemu_bl31_setup.c b/plat/qemu/common/qemu_bl31_setup.c
index 961f87c..33354dc 100644
--- a/plat/qemu/common/qemu_bl31_setup.c
+++ b/plat/qemu/common/qemu_bl31_setup.c
@@ -135,7 +135,7 @@
 	 * 256TB of RAM (48-bit PA) would require a 2MB L0 region. At the
 	 * moment we use a 8KB table, which covers 1TB of RAM (40-bit PA).
 	 */
-	if (gpt_init_l0_tables(GPCCR_PPS_1TB, PLAT_QEMU_L0_GPT_BASE,
+	if (gpt_init_l0_tables(PLATFORM_GPCCR_PPS, PLAT_QEMU_L0_GPT_BASE,
 			       PLAT_QEMU_L0_GPT_SIZE +
 			       PLAT_QEMU_GPT_BITLOCK_SIZE) < 0) {
 		ERROR("gpt_init_l0_tables() failed!\n");