fix: increase hypervisor/spmc per core stack size
There is a latent issue related with per core stack size being too small
and triggering randomly depending on test configurations.
This change increases the per core stack size from one to two pages
hopefully leaving enough room for long function call chains and local
stack allocations.
MTE is able to detect a stack overflow in certain conditions. A better
improvement will be to insert unmapped guard pages inbetween core
stacks.
Signed-off-by: Olivier Deprez <olivier.deprez@arm.com>
Change-Id: Ic864b5ff998f17e99339b9f6d3297866fa00e011
diff --git a/inc/hf/cpu.h b/inc/hf/cpu.h
index 25eab3b..74d61c1 100644
--- a/inc/hf/cpu.h
+++ b/inc/hf/cpu.h
@@ -8,6 +8,10 @@
#pragma once
+#define STACK_SIZE (8192)
+
+#if !defined(__ASSEMBLER__)
+
#include "hf/arch/cpu.h"
/* TODO: Fix alignment such that `cpu` structs are in different cache lines. */
@@ -34,3 +38,5 @@
struct cpu *cpu_find(cpu_id_t id);
uint8_t *cpu_get_buffer(struct cpu *c);
uint32_t cpu_get_buffer_size(struct cpu *c);
+
+#endif