powerpc/book3s: Add basic infrastructure to handle HMI in Linux.

Handle Hypervisor Maintenance Interrupt (HMI) in Linux. This patch implements
basic infrastructure to handle HMI in Linux host. The design is to invoke
opal handle hmi in real mode for recovery and set irq_pending when we hit HMI.
During check_irq_replay pull opal hmi event and print hmi info on console.
Signed-off-by: Mahesh Salgaonkar <mahesh@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

arch/powerpc/include/asm/exception-64s.h

@@ -425,6 +425,8 @@ label##_relon_hv:						\
 #define SOFTEN_VALUE_0xa00	PACA_IRQ_DBELL
 #define SOFTEN_VALUE_0xe80	PACA_IRQ_DBELL
 #define SOFTEN_VALUE_0xe82	PACA_IRQ_DBELL
+#define SOFTEN_VALUE_0xe60	PACA_IRQ_HMI
+#define SOFTEN_VALUE_0xe62	PACA_IRQ_HMI
 
 #define __SOFTEN_TEST(h, vec)						\
 	lbz	r10,PACASOFTIRQEN(r13);					\

arch/powerpc/include/asm/hardirq.h

@@ -11,6 +11,7 @@ typedef struct {
 	unsigned int pmu_irqs;
 	unsigned int mce_exceptions;
 	unsigned int spurious_irqs;
+	unsigned int hmi_exceptions;
 #ifdef CONFIG_PPC_DOORBELL
 	unsigned int doorbell_irqs;
 #endif

arch/powerpc/include/asm/hw_irq.h

@@ -25,6 +25,7 @@
 #define PACA_IRQ_EE		0x04
 #define PACA_IRQ_DEC		0x08 /* Or FIT */
 #define PACA_IRQ_EE_EDGE	0x10 /* BookE only */
+#define PACA_IRQ_HMI		0x20
 
 #endif /* CONFIG_PPC64 */
 

arch/powerpc/include/asm/kvm_asm.h

@@ -98,6 +98,7 @@
 #define BOOK3S_INTERRUPT_H_DATA_STORAGE	0xe00
 #define BOOK3S_INTERRUPT_H_INST_STORAGE	0xe20
 #define BOOK3S_INTERRUPT_H_EMUL_ASSIST	0xe40
+#define BOOK3S_INTERRUPT_HMI		0xe60
 #define BOOK3S_INTERRUPT_H_DOORBELL	0xe80
 #define BOOK3S_INTERRUPT_PERFMON	0xf00
 #define BOOK3S_INTERRUPT_ALTIVEC	0xf20

arch/powerpc/include/asm/machdep.h

@@ -174,6 +174,10 @@ struct machdep_calls {
 	/* Exception handlers */
 	int		(*system_reset_exception)(struct pt_regs *regs);
 	int 		(*machine_check_exception)(struct pt_regs *regs);
+	int		(*handle_hmi_exception)(struct pt_regs *regs);
+
+	/* Early exception handlers called in realmode */
+	int		(*hmi_exception_early)(struct pt_regs *regs);
 
 	/* Called during machine check exception to retrive fixup address. */
 	bool		(*mce_check_early_recovery)(struct pt_regs *regs);

arch/powerpc/include/asm/opal.h

@@ -915,6 +915,8 @@ extern void opal_msglog_init(void);
 
 extern int opal_machine_check(struct pt_regs *regs);
 extern bool opal_mce_check_early_recovery(struct pt_regs *regs);
+extern int opal_hmi_exception_early(struct pt_regs *regs);
+extern int opal_handle_hmi_exception(struct pt_regs *regs);
 
 extern void opal_shutdown(void);
 extern int opal_resync_timebase(void);

arch/powerpc/kernel/entry_64.S

@@ -915,6 +915,11 @@ restore_check_irq_replay:
 	addi	r3,r1,STACK_FRAME_OVERHEAD;
  	bl	do_IRQ
 	b	ret_from_except
+1:	cmpwi	cr0,r3,0xe60
+	bne	1f
+	addi	r3,r1,STACK_FRAME_OVERHEAD;
+	bl	handle_hmi_exception
+	b	ret_from_except
 1:	cmpwi	cr0,r3,0x900
 	bne	1f
 	addi	r3,r1,STACK_FRAME_OVERHEAD;

arch/powerpc/kernel/exceptions-64s.S

@@ -335,7 +335,7 @@ emulation_assist_trampoline:
 hv_exception_trampoline:
 	SET_SCRATCH0(r13)
 	EXCEPTION_PROLOG_0(PACA_EXGEN)
-	b	hmi_exception_hv
+	b	hmi_exception_early
 
 	. = 0xe80
 hv_doorbell_trampoline:
@@ -589,8 +589,64 @@ END_FTR_SECTION_IFSET(CPU_FTR_CFAR)
 	KVM_HANDLER(PACA_EXGEN, EXC_HV, 0xe22)
 	STD_EXCEPTION_HV_OOL(0xe42, emulation_assist)
 	KVM_HANDLER(PACA_EXGEN, EXC_HV, 0xe42)
-	STD_EXCEPTION_HV_OOL(0xe62, hmi_exception) /* need to flush cache ? */
+	MASKABLE_EXCEPTION_HV_OOL(0xe62, hmi_exception)
 	KVM_HANDLER(PACA_EXGEN, EXC_HV, 0xe62)
+
+	.globl hmi_exception_early
+hmi_exception_early:
+	EXCEPTION_PROLOG_1(PACA_EXGEN, NOTEST, 0xe60)
+	mr	r10,r1			/* Save r1			*/
+	ld	r1,PACAEMERGSP(r13)	/* Use emergency stack		*/
+	subi	r1,r1,INT_FRAME_SIZE	/* alloc stack frame		*/
+	std	r9,_CCR(r1)		/* save CR in stackframe	*/
+	mfspr	r11,SPRN_HSRR0		/* Save HSRR0 */
+	std	r11,_NIP(r1)		/* save HSRR0 in stackframe	*/
+	mfspr	r12,SPRN_HSRR1		/* Save SRR1 */
+	std	r12,_MSR(r1)		/* save SRR1 in stackframe	*/
+	std	r10,0(r1)		/* make stack chain pointer	*/
+	std	r0,GPR0(r1)		/* save r0 in stackframe	*/
+	std	r10,GPR1(r1)		/* save r1 in stackframe	*/
+	EXCEPTION_PROLOG_COMMON_2(PACA_EXGEN)
+	EXCEPTION_PROLOG_COMMON_3(0xe60)
+	addi	r3,r1,STACK_FRAME_OVERHEAD
+	bl	hmi_exception_realmode
+	/* Windup the stack. */
+	/* Clear MSR_RI before setting SRR0 and SRR1. */
+	li	r0,MSR_RI
+	mfmsr	r9			/* get MSR value */
+	andc	r9,r9,r0
+	mtmsrd	r9,1			/* Clear MSR_RI */
+	/* Move original HSRR0 and HSRR1 into the respective regs */
+	ld	r9,_MSR(r1)
+	mtspr	SPRN_HSRR1,r9
+	ld	r3,_NIP(r1)
+	mtspr	SPRN_HSRR0,r3
+	ld	r9,_CTR(r1)
+	mtctr	r9
+	ld	r9,_XER(r1)
+	mtxer	r9
+	ld	r9,_LINK(r1)
+	mtlr	r9
+	REST_GPR(0, r1)
+	REST_8GPRS(2, r1)
+	REST_GPR(10, r1)
+	ld	r11,_CCR(r1)
+	mtcr	r11
+	REST_GPR(11, r1)
+	REST_2GPRS(12, r1)
+	/* restore original r1. */
+	ld	r1,GPR1(r1)
+
+	/*
+	 * Go to virtual mode and pull the HMI event information from
+	 * firmware.
+	 */
+	.globl hmi_exception_after_realmode
+hmi_exception_after_realmode:
+	SET_SCRATCH0(r13)
+	EXCEPTION_PROLOG_0(PACA_EXGEN)
+	b	hmi_exception_hv
+
 	MASKABLE_EXCEPTION_HV_OOL(0xe82, h_doorbell)
 	KVM_HANDLER(PACA_EXGEN, EXC_HV, 0xe82)
 
@@ -611,6 +667,8 @@ END_FTR_SECTION_IFSET(CPU_FTR_CFAR)
  * - If it was a decrementer interrupt, we bump the dec to max and and return.
  * - If it was a doorbell we return immediately since doorbells are edge
  *   triggered and won't automatically refire.
+ * - If it was a HMI we return immediately since we handled it in realmode
+ *   and it won't refire.
  * - else we hard disable and return.
  * This is called with r10 containing the value to OR to the paca field.
  */
@@ -627,6 +685,8 @@ masked_##_H##interrupt:					\
 	mtspr	SPRN_DEC,r10;				\
 	b	2f;					\
 1:	cmpwi	r10,PACA_IRQ_DBELL;			\
+	beq	2f;					\
+	cmpwi	r10,PACA_IRQ_HMI;			\
 	beq	2f;					\
 	mfspr	r10,SPRN_##_H##SRR1;			\
 	rldicl	r10,r10,48,1; /* clear MSR_EE */	\
@@ -767,7 +827,7 @@ kvmppc_skip_Hinterrupt:
 	STD_EXCEPTION_COMMON(0xd00, single_step, single_step_exception)
 	STD_EXCEPTION_COMMON(0xe00, trap_0e, unknown_exception)
 	STD_EXCEPTION_COMMON(0xe40, emulation_assist, emulation_assist_interrupt)
-	STD_EXCEPTION_COMMON(0xe60, hmi_exception, unknown_exception)
+	STD_EXCEPTION_COMMON_ASYNC(0xe60, hmi_exception, handle_hmi_exception)
 #ifdef CONFIG_PPC_DOORBELL
 	STD_EXCEPTION_COMMON_ASYNC(0xe80, h_doorbell, doorbell_exception)
 #else

arch/powerpc/kernel/irq.c

@@ -189,6 +189,11 @@ notrace unsigned int __check_irq_replay(void)
 	}
 #endif /* CONFIG_PPC_BOOK3E */
 
+	/* Check if an hypervisor Maintenance interrupt happened */
+	local_paca->irq_happened &= ~PACA_IRQ_HMI;
+	if (happened & PACA_IRQ_HMI)
+		return 0xe60;
+
 	/* There should be nothing left ! */
 	BUG_ON(local_paca->irq_happened != 0);
 
@@ -377,6 +382,14 @@ int arch_show_interrupts(struct seq_file *p, int prec)
 		seq_printf(p, "%10u ", per_cpu(irq_stat, j).mce_exceptions);
 	seq_printf(p, "  Machine check exceptions\n");
 
+	if (cpu_has_feature(CPU_FTR_HVMODE)) {
+		seq_printf(p, "%*s: ", prec, "HMI");
+		for_each_online_cpu(j)
+			seq_printf(p, "%10u ",
+					per_cpu(irq_stat, j).hmi_exceptions);
+		seq_printf(p, "  Hypervisor Maintenance Interrupts\n");
+	}
+
 #ifdef CONFIG_PPC_DOORBELL
 	if (cpu_has_feature(CPU_FTR_DBELL)) {
 		seq_printf(p, "%*s: ", prec, "DBL");
@@ -400,6 +413,7 @@ u64 arch_irq_stat_cpu(unsigned int cpu)
 	sum += per_cpu(irq_stat, cpu).mce_exceptions;
 	sum += per_cpu(irq_stat, cpu).spurious_irqs;
 	sum += per_cpu(irq_stat, cpu).timer_irqs_others;
+	sum += per_cpu(irq_stat, cpu).hmi_exceptions;
 #ifdef CONFIG_PPC_DOORBELL
 	sum += per_cpu(irq_stat, cpu).doorbell_irqs;
 #endif

arch/powerpc/kernel/traps.c

@@ -302,6 +302,16 @@ long machine_check_early(struct pt_regs *regs)
 	return handled;
 }
 
+long hmi_exception_realmode(struct pt_regs *regs)
+{
+	__get_cpu_var(irq_stat).hmi_exceptions++;
+
+	if (ppc_md.hmi_exception_early)
+		ppc_md.hmi_exception_early(regs);
+
+	return 0;
+}
+
 #endif
 
 /*
@@ -738,6 +748,20 @@ void SMIException(struct pt_regs *regs)
 	die("System Management Interrupt", regs, SIGABRT);
 }
 
+void handle_hmi_exception(struct pt_regs *regs)
+{
+	struct pt_regs *old_regs;
+
+	old_regs = set_irq_regs(regs);
+	irq_enter();
+
+	if (ppc_md.handle_hmi_exception)
+		ppc_md.handle_hmi_exception(regs);
+
+	irq_exit();
+	set_irq_regs(old_regs);
+}
+
 void unknown_exception(struct pt_regs *regs)
 {
 	enum ctx_state prev_state = exception_enter();

arch/powerpc/kvm/book3s_hv_rmhandlers.S

@@ -159,6 +159,8 @@ END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S)
 	cmpwi	r12, BOOK3S_INTERRUPT_EXTERNAL
 BEGIN_FTR_SECTION
 	beq	11f
+	cmpwi	cr2, r12, BOOK3S_INTERRUPT_HMI
+	beq	cr2, 14f			/* HMI check */
 END_FTR_SECTION_IFSET(CPU_FTR_ARCH_206)
 
 	/* RFI into the highmem handler, or branch to interrupt handler */
@@ -179,6 +181,10 @@ END_FTR_SECTION_IFSET(CPU_FTR_ARCH_206)
 
 13:	b	machine_check_fwnmi
 
+14:	mtspr	SPRN_HSRR0, r8
+	mtspr	SPRN_HSRR1, r7
+	b	hmi_exception_after_realmode
+
 kvmppc_primary_no_guest:
 	/* We handle this much like a ceded vcpu */
 	/* set our bit in napping_threads */

arch/powerpc/platforms/powernv/opal.c

@@ -514,6 +514,20 @@ int opal_machine_check(struct pt_regs *regs)
 	return 0;
 }
 
+/* Early hmi handler called in real mode. */
+int opal_hmi_exception_early(struct pt_regs *regs)
+{
+	/* TODO: Call opal hmi handler. */
+	return 0;
+}
+
+/* HMI exception handler called in virtual mode during check_irq_replay. */
+int opal_handle_hmi_exception(struct pt_regs *regs)
+{
+	/* TODO: Retrive and print HMI event from OPAL. */
+	return 0;
+}
+
 static uint64_t find_recovery_address(uint64_t nip)
 {
 	int i;

arch/powerpc/platforms/powernv/setup.c

@@ -264,6 +264,8 @@ static void __init pnv_setup_machdep_opal(void)
 	ppc_md.halt = pnv_halt;
 	ppc_md.machine_check_exception = opal_machine_check;
 	ppc_md.mce_check_early_recovery = opal_mce_check_early_recovery;
+	ppc_md.hmi_exception_early = opal_hmi_exception_early;
+	ppc_md.handle_hmi_exception = opal_handle_hmi_exception;
 }
 
 #ifdef CONFIG_PPC_POWERNV_RTAS