riscv: Use separate IRQ shadow call stacks

When both CONFIG_IRQ_STACKS and SCS are enabled, also use a separate
per-CPU shadow call stack.

Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
Tested-by: Nathan Chancellor <nathan@kernel.org>
Link: https://lore.kernel.org/r/20230927224757.1154247-13-samitolvanen@google.com
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>

Author: samitolvanen

arch/riscv/include/asm/scs.h

@@ -13,6 +13,11 @@
 	XIP_FIXUP_OFFSET gp
 .endm
 
+/* Load the per-CPU IRQ shadow call stack to gp. */
+.macro scs_load_irq_stack tmp
+	load_per_cpu gp, irq_shadow_call_stack_ptr, \tmp
+.endm
+
 /* Load task_scs_sp(current) to gp. */
 .macro scs_load_current
 	REG_L	gp, TASK_TI_SCS_SP(tp)
@@ -34,6 +39,8 @@
 
 .macro scs_load_init_stack
 .endm
+.macro scs_load_irq_stack tmp
+.endm
 .macro scs_load_current
 .endm
 .macro scs_load_current_if_task_changed prev

arch/riscv/kernel/entry.S

@@ -237,12 +237,19 @@ SYM_FUNC_START(call_on_irq_stack)
 	REG_S	s0, STACKFRAME_FP(sp)
 	addi	s0, sp, STACKFRAME_SIZE_ON_STACK
 
+	/* Switch to the per-CPU shadow call stack */
+	scs_save_current
+	scs_load_irq_stack t0
+
 	/* Switch to the per-CPU IRQ stack and call the handler */
 	load_per_cpu t0, irq_stack_ptr, t1
 	li	t1, IRQ_STACK_SIZE
 	add	sp, t0, t1
 	jalr	a1
 
+	/* Switch back to the thread shadow call stack */
+	scs_load_current
+
 	/* Switch back to the thread stack and restore ra and s0 */
 	addi	sp, s0, -STACKFRAME_SIZE_ON_STACK
 	REG_L	ra, STACKFRAME_RA(sp)

arch/riscv/kernel/irq.c

@@ -9,6 +9,7 @@
 #include <linux/irqchip.h>
 #include <linux/irqdomain.h>
 #include <linux/module.h>
+#include <linux/scs.h>
 #include <linux/seq_file.h>
 #include <asm/sbi.h>
 #include <asm/smp.h>
@@ -34,6 +35,24 @@ EXPORT_SYMBOL_GPL(riscv_get_intc_hwnode);
 #ifdef CONFIG_IRQ_STACKS
 #include <asm/irq_stack.h>
 
+DECLARE_PER_CPU(ulong *, irq_shadow_call_stack_ptr);
+
+#ifdef CONFIG_SHADOW_CALL_STACK
+DEFINE_PER_CPU(ulong *, irq_shadow_call_stack_ptr);
+#endif
+
+static void init_irq_scs(void)
+{
+	int cpu;
+
+	if (!scs_is_enabled())
+		return;
+
+	for_each_possible_cpu(cpu)
+		per_cpu(irq_shadow_call_stack_ptr, cpu) =
+			scs_alloc(cpu_to_node(cpu));
+}
+
 DEFINE_PER_CPU(ulong *, irq_stack_ptr);
 
 #ifdef CONFIG_VMAP_STACK
@@ -76,6 +95,7 @@ void do_softirq_own_stack(void)
 #endif /* CONFIG_HAVE_SOFTIRQ_ON_OWN_STACK */
 
 #else
+static void init_irq_scs(void) {}
 static void init_irq_stacks(void) {}
 #endif /* CONFIG_IRQ_STACKS */
 
@@ -87,6 +107,7 @@ int arch_show_interrupts(struct seq_file *p, int prec)
 
 void __init init_IRQ(void)
 {
+	init_irq_scs();
 	init_irq_stacks();
 	irqchip_init();
 	if (!handle_arch_irq)