core: mm: increase xlat tables for core ASan builds

[alexhoppus]

Building OP-TEE with core AddressSanitizer enabled
(CFG_CORE_SANITIZE_KADDRESS) increases MMU translation table usage.

This is caused by ASan shadow regions and by the
overall growth of the core memory layout: code, data and stack
regions become larger, which often forces additional page table splits
and results in a higher number of xlat tables being allocated during
early boot.

With the current MAX_XLAT_TABLES calculation, some ASan-enabled builds
can exhaust the xlat table pool and panic in core_mmu_map_pages() with
xlat tables exhausted.

I/TC: Normal World virtualization support is disabled
I/TC: Asynchronous notifications are enabled
D/TC:1   atomic_console_notif:158 Asynchronous notifications started, event 0 (vm 0)
E/TC:1 0 core_mmu_xlat_table_alloc:702 5 xlat tables exhausted
E/TC:1 0 Panic 'Failed to spread pgdir on small tables' at core/mm/core_mmu.c:2087 <core_mmu_map_pages>
E/TC:1 0 TEE load address @ 0xe100000
E/TC:1 0 Call stack:
E/TC:1 0  0x0e10d130
E/TC:1 0  0x0e13384c
E/TC:1 0  0x0e153bd0
E/TC:1 0  0x0e159c84
E/TC:1 0  0x0e159168
E/TC:1 0  0x0e15a764
E/TC:1 0  0x0e1089d0
E/TC:1 0  0x0e108da4
E/TC:1 0  0x0e108e30
I/TC: Halting CPU 0

[jenswi-linaro]

This can be simplified as IS_ENABLED(CFG_CORE_SANITIZE_KADDRESS).

[alexhoppus]

Done

[jenswi-linaro]

Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>