[PATCH 0/3] MIPS: CI20: Enable SMP with two CPUs

[PATCH 0/3] MIPS: CI20: Enable SMP with two CPUs

[Florian Fainelli]

This patch series resurrects SMP support for the CI20 board. This was
submitted back in 2020 by Zhou Yanjie.

Florian Fainelli (1):
  clk: ingenic: Poll JZ4780 CPU1 power status atomically

周琰杰 (Zhou Yanjie) (2):
  MIPS: Ingenic: Add JZ4780 SMP support to the generic platform
  MIPS: CI20: Enable SMP with two CPUs

 MAINTAINERS                           |   1 +
 arch/mips/configs/ci20_defconfig      |   2 +
 arch/mips/generic/Makefile            |   1 +
 arch/mips/generic/ingenic-smp-entry.S |  54 ++++++
 arch/mips/generic/ingenic-smp.c       | 240 ++++++++++++++++++++++++++
 arch/mips/generic/ingenic-smp.h       |  20 +++
 arch/mips/generic/init.c              |   4 +
 arch/mips/ingenic/Kconfig             |   6 +
 arch/mips/kernel/idle.c               |  25 ++-
 drivers/clk/ingenic/jz4780-cgu.c      |   2 +-
 10 files changed, 353 insertions(+), 2 deletions(-)
 create mode 100644 arch/mips/generic/ingenic-smp-entry.S
 create mode 100644 arch/mips/generic/ingenic-smp.c
 create mode 100644 arch/mips/generic/ingenic-smp.h

-- 
2.53.0

[PATCH 1/3] clk: ingenic: Poll JZ4780 CPU1 power status atomically

[Florian Fainelli]

The CPU1 clock enable callback runs under the clock framework's spinlock
with interrupts disabled. Its readl_poll_timeout() can sleep between
reads, which is invalid in this context when SMP powers up CPU1.

Use readl_poll_timeout_atomic() with the existing interval and timeout.
This is a prerequisite for the JZ4780 SMP bring-up path. The original
CPU1 clock implementation by Zhou Yanjie is already upstream.

Cross-built for CI20; the subsequent SMP-enabled image brought up both
CPU cores. Development and validation used Codex assistance.

Fixes: 6673db4f3f26 ("clk: JZ4780: Add function for enable the second core.")
Link: https://lkml.kernel.org/r/1582215889-113034-3-git-send-email-zhouyanjie@wanyeetech.com
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
---
 drivers/clk/ingenic/jz4780-cgu.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/clk/ingenic/jz4780-cgu.c b/drivers/clk/ingenic/jz4780-cgu.c
index 1e88aef7ac0f..f5983033583a 100644
--- a/drivers/clk/ingenic/jz4780-cgu.c
+++ b/drivers/clk/ingenic/jz4780-cgu.c
@@ -242,7 +242,7 @@ static int jz4780_core1_enable(struct clk_hw *hw)
 	spin_unlock_irqrestore(&cgu->lock, flags);
 
 	/* wait for the CPU to be powered up */
-	retval = readl_poll_timeout(cgu->base + CGU_REG_LCR, lcr,
+	retval = readl_poll_timeout_atomic(cgu->base + CGU_REG_LCR, lcr,
 				 !(lcr & LCR_SCPUS), 10, timeout);
 	if (retval == -ETIMEDOUT) {
 		pr_err("%s: Wait for power up core1 timeout\n", __func__);
-- 
2.53.0

[PATCH 2/3] MIPS: Ingenic: Add JZ4780 SMP support to the generic platform

[Florian Fainelli]

From: 周琰杰 (Zhou Yanjie) <zhouyanjie@wanyeetech.com>

Port the JZ4780 reset-entry and CP0 mailbox implementation from Zhou
Yanjie's 2020 v8 series, based on Paul Burton's original CI20 SMP code.
The old arch/mips/jz4740 platform has moved into the generic MIPS code,
so register the Ingenic SMP operations after unflattening the device
tree and before selecting the generic fallback implementations.

Reuse the upstream CPU clock gates, CPU device-tree nodes and per-CPU
TCU timers. Enumerate only available CPU nodes within the two-core
hardware limit. Prepare the CPU1 clock before boot, propagate clock
failures, serialize mailbox accesses with a raw spinlock and publish
bootstrap state before releasing the secondary core from reset. Align
the secondary reset entry to the REIM register's 64 KiB granularity.
Retain the UP fallback for other SoCs in a shared Ingenic kernel.

Write back the private data cache before WAIT, since XBurst gates its
clock during idle and another core must not access dirty data in that
sleeping cache. Return with interrupts disabled as required by the
current idle API. Existing per-CPU TCU events make the old forced
broadcast-timer workaround unnecessary.

This adaptation incorporates the CPU-enumeration review feedback and
removes the old unused CPU mask. Preserve the original source credits;
historical Tested-by and Reviewed-by tags are not carried to this
substantially changed port. Adaptation and validation used Codex
assistance.

Link: https://lkml.iu.edu/hypermail/linux/kernel/2005.2/03488.html
Link: https://lists.openwall.net/linux-kernel/2020/05/13/1482
Link: https://lists.openwall.net/linux-kernel/2020/05/19/1099
Signed-off-by: 周琰杰 (Zhou Yanjie) <zhouyanjie@wanyeetech.com>
[florian: forward ported to latest version]
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
---
 MAINTAINERS                           |   1 +
 arch/mips/generic/Makefile            |   1 +
 arch/mips/generic/ingenic-smp-entry.S |  54 ++++++
 arch/mips/generic/ingenic-smp.c       | 240 ++++++++++++++++++++++++++
 arch/mips/generic/ingenic-smp.h       |  20 +++
 arch/mips/generic/init.c              |   4 +
 arch/mips/ingenic/Kconfig             |   6 +
 arch/mips/kernel/idle.c               |  25 ++-
 8 files changed, 350 insertions(+), 1 deletion(-)
 create mode 100644 arch/mips/generic/ingenic-smp-entry.S
 create mode 100644 arch/mips/generic/ingenic-smp.c
 create mode 100644 arch/mips/generic/ingenic-smp.h

diff --git a/MAINTAINERS b/MAINTAINERS
index 09713dc4c6a7..101027579ccd 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -13117,6 +13117,7 @@ L:	linux-mips@vger.kernel.org
 S:	Maintained
 F:	arch/mips/boot/dts/ingenic/
 F:	arch/mips/generic/board-ingenic.c
+F:	arch/mips/generic/ingenic-smp*
 F:	arch/mips/include/asm/mach-ingenic/
 F:	arch/mips/ingenic/Kconfig
 F:	drivers/clk/ingenic/
diff --git a/arch/mips/generic/Makefile b/arch/mips/generic/Makefile
index ea0e4ad5e600..df92650c0f15 100644
--- a/arch/mips/generic/Makefile
+++ b/arch/mips/generic/Makefile
@@ -12,5 +12,6 @@ obj-$(CONFIG_YAMON_DT_SHIM)		+= yamon-dt.o
 obj-$(CONFIG_LEGACY_BOARD_SEAD3)	+= board-sead3.o
 obj-$(CONFIG_LEGACY_BOARD_OCELOT)	+= board-ocelot.o
 obj-$(CONFIG_MACH_INGENIC)			+= board-ingenic.o
+obj-$(CONFIG_MIPS_INGENIC_SMP)	+= ingenic-smp.o ingenic-smp-entry.o
 obj-$(CONFIG_VIRT_BOARD_RANCHU)		+= board-ranchu.o
 obj-$(CONFIG_MACH_REALTEK_RTL)		+= board-realtek.o
diff --git a/arch/mips/generic/ingenic-smp-entry.S b/arch/mips/generic/ingenic-smp-entry.S
new file mode 100644
index 000000000000..828df4687e84
--- /dev/null
+++ b/arch/mips/generic/ingenic-smp-entry.S
@@ -0,0 +1,54 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copyright (C) 2013 Paul Burton <paul.burton@imgtec.com>
+ * JZ4780 secondary CPU reset entry
+ */
+
+#include <asm/addrspace.h>
+#include <asm/asm.h>
+#include <asm/asmmacro.h>
+#include <asm/cacheops.h>
+#include <asm/mipsregs.h>
+
+#define CACHE_SIZE	(32 * 1024)
+#define CACHE_LINESIZE	32
+
+	/* The CP0 REIM register only stores the upper 16 address bits. */
+	.section .text.smp-entry, "ax"
+	.balign 0x10000
+	.set noreorder
+LEAF(jz4780_secondary_cpu_entry)
+	mtc0	zero, CP0_CAUSE
+	li	t0, ST0_CU0
+	mtc0	t0, CP0_STATUS
+	ehb
+
+	/* Invalidate both private caches before entering cached code. */
+	li	t0, KSEG0
+	ori	t1, t0, CACHE_SIZE
+	mtc0	zero, CP0_TAGLO
+	ehb
+1:	cache	Index_Store_Tag_I, 0(t0)
+	cache	Index_Store_Tag_D, 0(t0)
+	addiu	t0, t0, CACHE_LINESIZE
+	bne	t0, t1, 1b
+	 nop
+
+	mfc0	t0, CP0_CONFIG
+	li	t1, ~CONF_CM_CMASK
+	and	t0, t0, t1
+	ori	t0, t0, CONF_CM_CACHABLE_NONCOHERENT
+	mtc0	t0, CP0_CONFIG
+	mtc0	zero, CP0_PAGEMASK
+	ehb
+
+	/* CPU0 publishes these in RAM before releasing CPU1 from reset. */
+	la	t0, jz4780_cpu_entry_sp
+	lw	sp, 0(t0)
+	la	t0, jz4780_cpu_entry_gp
+	lw	gp, 0(t0)
+
+	la	t0, smp_bootstrap
+	jr	t0
+	 nop
+END(jz4780_secondary_cpu_entry)
diff --git a/arch/mips/generic/ingenic-smp.c b/arch/mips/generic/ingenic-smp.c
new file mode 100644
index 000000000000..98c1fd7483db
--- /dev/null
+++ b/arch/mips/generic/ingenic-smp.c
@@ -0,0 +1,240 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * JZ4780 SMP support, based on Paul Burton's CI20 implementation and
+ * Zhou Yanjie's 2020 JZ4780 SMP series.
+ * Copyright (C) 2013 Paul Burton <paul.burton@imgtec.com>
+ */
+
+#include <linux/clk.h>
+#include <linux/cpu.h>
+#include <linux/interrupt.h>
+#include <linux/of.h>
+#include <linux/of_clk.h>
+#include <linux/sched/task_stack.h>
+#include <linux/smp.h>
+
+#include <asm/addrspace.h>
+#include <asm/irq_cpu.h>
+#include <asm/mipsregs.h>
+#include <asm/r4kcache.h>
+#include <asm/smp-ops.h>
+
+#include "ingenic-smp.h"
+
+#define read_c0_corectrl()	__read_32bit_c0_register($12, 2)
+#define write_c0_corectrl(val)	__write_32bit_c0_register($12, 2, val)
+#define read_c0_corestatus()	__read_32bit_c0_register($12, 3)
+#define write_c0_corestatus(val)	__write_32bit_c0_register($12, 3, val)
+#define read_c0_reim()		__read_32bit_c0_register($12, 4)
+#define write_c0_reim(val)	__write_32bit_c0_register($12, 4, val)
+#define read_c0_mailbox0()	__read_32bit_c0_register($20, 0)
+#define write_c0_mailbox0(val)	__write_32bit_c0_register($20, 0, val)
+#define read_c0_mailbox1()	__read_32bit_c0_register($20, 1)
+#define write_c0_mailbox1(val)	__write_32bit_c0_register($20, 1, val)
+
+#define JZ4780_NR_CPUS		2
+#define JZ4780_MBOX_IRQ		(MIPS_CPU_IRQ_BASE + 3)
+#define CORECTRL_SWRST(cpu)	BIT(cpu)
+#define CORECTRL_RPC(cpu)	BIT(8 + (cpu))
+#define CORESTATUS_MIRQ(cpu)	BIT(cpu)
+#define REIM_ENTRY		GENMASK(31, 16)
+#define REIM_MBOXIRQ(cpu)	BIT(cpu)
+
+static struct clk *cpu1_clock;
+static DEFINE_RAW_SPINLOCK(smp_lock);
+
+/* Referenced by the uncached secondary CPU reset entry. */
+unsigned long jz4780_cpu_entry_sp;
+unsigned long jz4780_cpu_entry_gp;
+
+static irqreturn_t jz4780_mbox_interrupt(int irq, void *dev_id)
+{
+	unsigned int cpu = cpu_logical_map(smp_processor_id());
+	u32 action, status;
+
+	raw_spin_lock(&smp_lock);
+	if (cpu == 0) {
+		action = read_c0_mailbox0();
+		write_c0_mailbox0(0);
+	} else {
+		action = read_c0_mailbox1();
+		write_c0_mailbox1(0);
+	}
+
+	status = read_c0_corestatus();
+	write_c0_corestatus(status & ~CORESTATUS_MIRQ(cpu));
+	irq_disable_hazard();
+	raw_spin_unlock(&smp_lock);
+
+	if (action & SMP_RESCHEDULE_YOURSELF)
+		scheduler_ipi();
+	if (action & SMP_CALL_FUNCTION)
+		generic_smp_call_function_interrupt();
+
+	return IRQ_HANDLED;
+}
+
+static void __init jz4780_smp_setup(void)
+{
+	struct device_node *np;
+	u32 entry, reim;
+	u64 hwid;
+
+	for_each_of_cpu_node(np) {
+		hwid = of_get_cpu_hwid(np, 0);
+		if (hwid >= nr_cpu_ids || hwid >= JZ4780_NR_CPUS ||
+		    !of_device_is_available(np))
+			continue;
+		__cpu_number_map[hwid] = hwid;
+		__cpu_logical_map[hwid] = hwid;
+		set_cpu_possible(hwid, true);
+	}
+
+	/* Program the shared reset vector with mailbox delivery masked. */
+	reim = read_c0_reim();
+	reim &= ~(REIM_MBOXIRQ(0) | REIM_MBOXIRQ(1));
+	write_c0_reim(reim);
+	write_c0_mailbox0(0);
+	write_c0_mailbox1(0);
+	write_c0_corestatus(0);
+
+	entry = KSEG1ADDR((unsigned long)jz4780_secondary_cpu_entry);
+	write_c0_reim((reim & ~REIM_ENTRY) | (entry & REIM_ENTRY));
+}
+
+static void __init jz4780_prepare_cpus(unsigned int max_cpus)
+{
+	struct device_node *np;
+	u32 reim;
+	int ret;
+
+	ret = request_irq(JZ4780_MBOX_IRQ, jz4780_mbox_interrupt,
+			  IRQF_PERCPU | IRQF_NO_THREAD, "core mailbox", NULL);
+	if (ret)
+		panic("JZ4780: cannot request mailbox IRQ: %d", ret);
+
+	reim = read_c0_reim();
+	write_c0_reim(reim | REIM_MBOXIRQ(0));
+	set_c0_status(STATUSF_IP3);
+	irq_enable_hazard();
+
+	if (!cpu_possible(1))
+		return;
+
+	np = of_get_cpu_node(1, NULL);
+	cpu1_clock = of_clk_get(np, 0);
+	of_node_put(np);
+	if (IS_ERR(cpu1_clock)) {
+		pr_err("JZ4780: cannot get CPU1 clock: %ld\n",
+		       PTR_ERR(cpu1_clock));
+		return;
+	}
+
+	ret = clk_prepare(cpu1_clock);
+	if (ret) {
+		clk_put(cpu1_clock);
+		cpu1_clock = ERR_PTR(ret);
+		pr_err("JZ4780: cannot prepare CPU1 clock: %d\n", ret);
+	}
+}
+
+static int jz4780_boot_secondary(int cpu, struct task_struct *idle)
+{
+	unsigned long flags;
+	u32 ctrl;
+	int ret;
+
+	if (cpu_logical_map(cpu) != 1 || IS_ERR_OR_NULL(cpu1_clock))
+		return -ENODEV;
+
+	raw_spin_lock_irqsave(&smp_lock, flags);
+	ctrl = read_c0_corectrl() | CORECTRL_SWRST(1) | CORECTRL_RPC(1);
+	write_c0_corectrl(ctrl);
+	back_to_back_c0_hazard();
+
+	ret = clk_enable(cpu1_clock);
+	if (ret)
+		goto unlock;
+
+	jz4780_cpu_entry_sp = __KSTK_TOS(idle);
+	jz4780_cpu_entry_gp = (unsigned long)task_thread_info(idle);
+	/* Publish bootstrap state before CPU1 starts using its private cache. */
+	r4k_blast_dcache();
+	/* Complete cache writeback before releasing reset. */
+	mb();
+	write_c0_corectrl(ctrl & ~CORECTRL_SWRST(1));
+	back_to_back_c0_hazard();
+unlock:
+	raw_spin_unlock_irqrestore(&smp_lock, flags);
+	return ret;
+}
+
+static void jz4780_init_secondary(void)
+{
+	clear_c0_status(ST0_IM);
+	irq_disable_hazard();
+}
+
+static void jz4780_smp_finish(void)
+{
+	unsigned int cpu = cpu_logical_map(smp_processor_id());
+	u32 reim;
+
+	raw_spin_lock(&smp_lock);
+	reim = read_c0_reim();
+	write_c0_reim(reim | REIM_MBOXIRQ(cpu));
+	raw_spin_unlock(&smp_lock);
+
+	change_c0_status(ST0_IM, STATUSF_IP3 | STATUSF_IP2 |
+			 STATUSF_IP1 | STATUSF_IP0);
+	irq_enable_hazard();
+	local_irq_enable();
+}
+
+static void jz4780_send_ipi_locked(int cpu, unsigned int action)
+{
+	if (cpu_logical_map(cpu) == 0)
+		write_c0_mailbox0(read_c0_mailbox0() | action);
+	else
+		write_c0_mailbox1(read_c0_mailbox1() | action);
+	back_to_back_c0_hazard();
+}
+
+static void jz4780_send_ipi_single(int cpu, unsigned int action)
+{
+	unsigned long flags;
+
+	raw_spin_lock_irqsave(&smp_lock, flags);
+	jz4780_send_ipi_locked(cpu, action);
+	raw_spin_unlock_irqrestore(&smp_lock, flags);
+}
+
+static void jz4780_send_ipi_mask(const struct cpumask *mask, unsigned int action)
+{
+	unsigned long flags;
+	int cpu;
+
+	raw_spin_lock_irqsave(&smp_lock, flags);
+	for_each_cpu(cpu, mask)
+		jz4780_send_ipi_locked(cpu, action);
+	raw_spin_unlock_irqrestore(&smp_lock, flags);
+}
+
+static const struct plat_smp_ops jz4780_smp_ops = {
+	.send_ipi_single = jz4780_send_ipi_single,
+	.send_ipi_mask = jz4780_send_ipi_mask,
+	.init_secondary = jz4780_init_secondary,
+	.smp_finish = jz4780_smp_finish,
+	.boot_secondary = jz4780_boot_secondary,
+	.smp_setup = jz4780_smp_setup,
+	.prepare_cpus = jz4780_prepare_cpus,
+};
+
+int __init register_ingenic_smp_ops(void)
+{
+	if (!of_machine_is_compatible("ingenic,jz4780"))
+		return -ENODEV;
+
+	register_smp_ops(&jz4780_smp_ops);
+	return 0;
+}
diff --git a/arch/mips/generic/ingenic-smp.h b/arch/mips/generic/ingenic-smp.h
new file mode 100644
index 000000000000..54191be8731a
--- /dev/null
+++ b/arch/mips/generic/ingenic-smp.h
@@ -0,0 +1,20 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/* Copyright (C) 2013 Paul Burton <paul.burton@imgtec.com> */
+#ifndef __MIPS_GENERIC_INGENIC_SMP_H__
+#define __MIPS_GENERIC_INGENIC_SMP_H__
+
+#include <linux/errno.h>
+#include <linux/init.h>
+#include <linux/kconfig.h>
+
+#ifdef CONFIG_MIPS_INGENIC_SMP
+int __init register_ingenic_smp_ops(void);
+void jz4780_secondary_cpu_entry(void);
+#else
+static inline int register_ingenic_smp_ops(void)
+{
+	return -ENODEV;
+}
+#endif
+
+#endif /* __MIPS_GENERIC_INGENIC_SMP_H__ */
diff --git a/arch/mips/generic/init.c b/arch/mips/generic/init.c
index 1d712eac1617..00149bd1e9a9 100644
--- a/arch/mips/generic/init.c
+++ b/arch/mips/generic/init.c
@@ -20,6 +20,8 @@
 #include <asm/smp-ops.h>
 #include <asm/time.h>
 
+#include "ingenic-smp.h"
+
 static __initconst const void *fdt;
 static __initconst const struct mips_machine *mach;
 static __initconst const void *mach_match_data;
@@ -113,6 +115,8 @@ void __init device_tree_init(void)
 	unflatten_and_copy_device_tree();
 	mips_cpc_probe();
 
+	if (!register_ingenic_smp_ops())
+		return;
 	if (!register_cps_smp_ops())
 		return;
 	if (!register_vsmp_smp_ops())
diff --git a/arch/mips/ingenic/Kconfig b/arch/mips/ingenic/Kconfig
index f595b339a4b8..1dfd0bfd0193 100644
--- a/arch/mips/ingenic/Kconfig
+++ b/arch/mips/ingenic/Kconfig
@@ -64,6 +64,12 @@ config MACH_JZ4780
 	select MIPS_CPU_SCACHE
 	select SYS_HAS_CPU_MIPS32_R2
 	select SYS_SUPPORTS_HIGHMEM
+	select SYS_SUPPORTS_SMP
+
+config MIPS_INGENIC_SMP
+	def_bool y
+	depends on MACH_JZ4780 && SMP
+	select SMP_UP
 
 config MACH_X1000
 	bool
diff --git a/arch/mips/kernel/idle.c b/arch/mips/kernel/idle.c
index 80e8a04a642e..6cabdd4825a7 100644
--- a/arch/mips/kernel/idle.c
+++ b/arch/mips/kernel/idle.c
@@ -18,6 +18,7 @@
 #include <asm/cpu-type.h>
 #include <asm/idle.h>
 #include <asm/mipsregs.h>
+#include <asm/r4kcache.h>
 
 /*
  * Not all of the MIPS CPUs have the "wait" instruction available. Moreover,
@@ -71,6 +72,23 @@ static void __cpuidle rm7k_wait_irqoff(void)
 		"	.set	pop					\n");
 }
 
+/* XBurst gates its private cache clock in WAIT; write back dirty data first. */
+static void __cpuidle ingenic_smp_wait_irqoff(void)
+{
+	unsigned long pending = read_c0_cause() & read_c0_status() & CAUSEF_IP;
+
+	if (!need_resched() && !pending) {
+		r4k_blast_dcache();
+		__asm__ __volatile__(
+			".set push\n"
+			".set mips3\n"
+			"sync\n"
+			"wait\n"
+			".set pop\n"
+			: : : "memory");
+	}
+}
+
 /*
  * Au1 'wait' is only useful when the 32kHz counter is used as timer,
  * since coreclock (and the cp0 counter) stops upon executing it. Only an
@@ -156,10 +174,15 @@ void __init check_wait(void)
 	case CPU_CAVIUM_OCTEON_PLUS:
 	case CPU_CAVIUM_OCTEON2:
 	case CPU_CAVIUM_OCTEON3:
-	case CPU_XBURST:
 	case CPU_LOONGSON32:
 		cpu_wait = r4k_wait;
 		break;
+	case CPU_XBURST:
+		if (IS_ENABLED(CONFIG_MIPS_INGENIC_SMP))
+			cpu_wait = ingenic_smp_wait_irqoff;
+		else
+			cpu_wait = r4k_wait;
+		break;
 	case CPU_LOONGSON64:
 		if ((c->processor_id & (PRID_IMP_MASK | PRID_REV_MASK)) >=
 				(PRID_IMP_LOONGSON_64C | PRID_REV_LOONGSON3A_R2_0) ||
-- 
2.53.0

[PATCH 3/3] MIPS: CI20: Enable SMP with two CPUs

[Florian Fainelli]

From: 周琰杰 (Zhou Yanjie) <zhouyanjie@wanyeetech.com>

Enable both JZ4780 cores now that the generic Ingenic platform provides
secondary CPU reset and mailbox support. The existing device tree
already reserves TCU channels for both per-CPU clock events.

Rebase Zhou Yanjie's original defconfig update onto the current CI20
configuration. The resulting image brought up both cores on the board.
The rebase and validation used Codex assistance.

Link: https://lists.openwall.net/linux-kernel/2020/05/19/739
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
---
 arch/mips/configs/ci20_defconfig | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/arch/mips/configs/ci20_defconfig b/arch/mips/configs/ci20_defconfig
index 90536cab417e..09d65094b1bd 100644
--- a/arch/mips/configs/ci20_defconfig
+++ b/arch/mips/configs/ci20_defconfig
@@ -22,6 +22,8 @@ CONFIG_EXPERT=y
 CONFIG_MACH_INGENIC_SOC=y
 CONFIG_JZ4780_CI20=y
 CONFIG_HIGHMEM=y
+CONFIG_SMP=y
+CONFIG_NR_CPUS=2
 CONFIG_HZ_100=y
 # CONFIG_SUSPEND is not set
 # CONFIG_SECCOMP is not set
-- 
2.53.0

Re: [PATCH 1/3] clk: ingenic: Poll JZ4780 CPU1 power status atomically

[Paul Cercueil]

Hi Florian,

Le jeudi 05 novembre 2026 à 12:32 -0800, Florian Fainelli a écrit :
> The CPU1 clock enable callback runs under the clock framework's spinlock
> with interrupts disabled. Its readl_poll_timeout() can sleep between
> reads, which is invalid in this context when SMP powers up CPU1.
> 
> Use readl_poll_timeout_atomic() with the existing interval and timeout.
> This is a prerequisite for the JZ4780 SMP bring-up path. The original
> CPU1 clock implementation by Zhou Yanjie is already upstream.
> 
> Cross-built for CI20; the subsequent SMP-enabled image brought up both
> CPU cores. Development and validation used Codex assistance.
> 
> Fixes: 6673db4f3f26 ("clk: JZ4780: Add function for enable the second core.")
> Link:
> https://lkml.kernel.org/r/1582215889-113034-3-git-send-email-zhouyanjie@wanyeetech.com
> Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>

Reviewed-by: Paul Cercueil <paul@crapouillou.net>

Cheers,
-Paul

> ---
>  drivers/clk/ingenic/jz4780-cgu.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/clk/ingenic/jz4780-cgu.c b/drivers/clk/ingenic/jz4780-
> cgu.c
> index 1e88aef7ac0f..f5983033583a 100644
> --- a/drivers/clk/ingenic/jz4780-cgu.c
> +++ b/drivers/clk/ingenic/jz4780-cgu.c
> @@ -242,7 +242,7 @@ static int jz4780_core1_enable(struct clk_hw *hw)
>  	spin_unlock_irqrestore(&cgu->lock, flags);
>  
>  	/* wait for the CPU to be powered up */
> -	retval = readl_poll_timeout(cgu->base + CGU_REG_LCR, lcr,
> +	retval = readl_poll_timeout_atomic(cgu->base + CGU_REG_LCR, lcr,
>  				 !(lcr & LCR_SCPUS), 10, timeout);
>  	if (retval == -ETIMEDOUT) {
>  		pr_err("%s: Wait for power up core1 timeout\n", __func__);

Re: [PATCH 0/3] MIPS: CI20: Enable SMP with two CPUs

[H. Nikolaus Schaller]

Hi,

> Am 05.11.2026 um 21:32 schrieb Florian Fainelli <f.fainelli@gmail.com>:
> 
> This patch series resurrects SMP support for the CI20 board. This was
> submitted back in 2020 by Zhou Yanjie.

Great!

We are carrying some of these patches along locally for a long time so it
is good to integrate them.

What should be considered in the future is to extend this SMP stuff to
x2000 and x2600. We have work in progress but nothing to submit yet.

> 
> Florian Fainelli (1):
>  clk: ingenic: Poll JZ4780 CPU1 power status atomically
> 
> 周琰杰 (Zhou Yanjie) (2):
>  MIPS: Ingenic: Add JZ4780 SMP support to the generic platform
>  MIPS: CI20: Enable SMP with two CPUs
> 
> MAINTAINERS                           |   1 +
> arch/mips/configs/ci20_defconfig      |   2 +
> arch/mips/generic/Makefile            |   1 +
> arch/mips/generic/ingenic-smp-entry.S |  54 ++++++
> arch/mips/generic/ingenic-smp.c       | 240 ++++++++++++++++++++++++++
> arch/mips/generic/ingenic-smp.h       |  20 +++
> arch/mips/generic/init.c              |   4 +
> arch/mips/ingenic/Kconfig             |   6 +
> arch/mips/kernel/idle.c               |  25 ++-
> drivers/clk/ingenic/jz4780-cgu.c      |   2 +-
> 10 files changed, 353 insertions(+), 2 deletions(-)
> create mode 100644 arch/mips/generic/ingenic-smp-entry.S
> create mode 100644 arch/mips/generic/ingenic-smp.c
> create mode 100644 arch/mips/generic/ingenic-smp.h
> 
> -- 
> 2.53.0
> 
> 

BR and thanks,
Nikolaus

Re: [PATCH 2/3] MIPS: Ingenic: Add JZ4780 SMP support to the generic platform

[Paul Cercueil]

Hi Florian,

Le jeudi 05 novembre 2026 à 12:32 -0800, Florian Fainelli a écrit :
> From: 周琰杰 (Zhou Yanjie) <zhouyanjie@wanyeetech.com>
> 
> Port the JZ4780 reset-entry and CP0 mailbox implementation from Zhou
> Yanjie's 2020 v8 series, based on Paul Burton's original CI20 SMP code.
> The old arch/mips/jz4740 platform has moved into the generic MIPS code,
> so register the Ingenic SMP operations after unflattening the device
> tree and before selecting the generic fallback implementations.
> 
> Reuse the upstream CPU clock gates, CPU device-tree nodes and per-CPU
> TCU timers. Enumerate only available CPU nodes within the two-core
> hardware limit. Prepare the CPU1 clock before boot, propagate clock
> failures, serialize mailbox accesses with a raw spinlock and publish
> bootstrap state before releasing the secondary core from reset. Align
> the secondary reset entry to the REIM register's 64 KiB granularity.
> Retain the UP fallback for other SoCs in a shared Ingenic kernel.
> 
> Write back the private data cache before WAIT, since XBurst gates its
> clock during idle and another core must not access dirty data in that
> sleeping cache. Return with interrupts disabled as required by the
> current idle API. Existing per-CPU TCU events make the old forced
> broadcast-timer workaround unnecessary.

This reads like a report from AI. Please use your own words, mention what's
worthwhile to mention (e.g. the Xburst-specific wait mechanism) and drop what's
implementation details.

> This adaptation incorporates the CPU-enumeration review feedback and
> removes the old unused CPU mask. Preserve the original source credits;
> historical Tested-by and Reviewed-by tags are not carried to this
> substantially changed port. Adaptation and validation used Codex
> assistance.
> 
> Link: https://lkml.iu.edu/hypermail/linux/kernel/2005.2/03488.html
> Link: https://lists.openwall.net/linux-kernel/2020/05/13/1482
> Link: https://lists.openwall.net/linux-kernel/2020/05/19/1099
> Signed-off-by: 周琰杰 (Zhou Yanjie) <zhouyanjie@wanyeetech.com>
> [florian: forward ported to latest version]
> Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
> ---
>  MAINTAINERS                           |   1 +
>  arch/mips/generic/Makefile            |   1 +
>  arch/mips/generic/ingenic-smp-entry.S |  54 ++++++
>  arch/mips/generic/ingenic-smp.c       | 240 ++++++++++++++++++++++++++
>  arch/mips/generic/ingenic-smp.h       |  20 +++
>  arch/mips/generic/init.c              |   4 +
>  arch/mips/ingenic/Kconfig             |   6 +
>  arch/mips/kernel/idle.c               |  25 ++-
>  8 files changed, 350 insertions(+), 1 deletion(-)
>  create mode 100644 arch/mips/generic/ingenic-smp-entry.S
>  create mode 100644 arch/mips/generic/ingenic-smp.c
>  create mode 100644 arch/mips/generic/ingenic-smp.h

These files really don't belong in arch/mips/generic/. We have a
arch/mips/ingenic/, put them there. You'll have to add a small Makefile there
and list the folder in arch/mips/Kbuild.platforms.

> 
> diff --git a/MAINTAINERS b/MAINTAINERS
> index 09713dc4c6a7..101027579ccd 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -13117,6 +13117,7 @@ L:	linux-mips@vger.kernel.org
>  S:	Maintained
>  F:	arch/mips/boot/dts/ingenic/
>  F:	arch/mips/generic/board-ingenic.c
> +F:	arch/mips/generic/ingenic-smp*

That's sneaky. Giving me the maintainership like that.
(I'm OK with it though)

>  F:	arch/mips/include/asm/mach-ingenic/
>  F:	arch/mips/ingenic/Kconfig
>  F:	drivers/clk/ingenic/
> diff --git a/arch/mips/generic/Makefile b/arch/mips/generic/Makefile
> index ea0e4ad5e600..df92650c0f15 100644
> --- a/arch/mips/generic/Makefile
> +++ b/arch/mips/generic/Makefile
> @@ -12,5 +12,6 @@ obj-$(CONFIG_YAMON_DT_SHIM)		+= yamon-dt.o
>  obj-$(CONFIG_LEGACY_BOARD_SEAD3)	+= board-sead3.o
>  obj-$(CONFIG_LEGACY_BOARD_OCELOT)	+= board-ocelot.o
>  obj-$(CONFIG_MACH_INGENIC)			+= board-ingenic.o
> +obj-$(CONFIG_MIPS_INGENIC_SMP)	+= ingenic-smp.o ingenic-smp-entry.o
>  obj-$(CONFIG_VIRT_BOARD_RANCHU)		+= board-ranchu.o
>  obj-$(CONFIG_MACH_REALTEK_RTL)		+= board-realtek.o
> diff --git a/arch/mips/generic/ingenic-smp-entry.S
> b/arch/mips/generic/ingenic-smp-entry.S
> new file mode 100644
> index 000000000000..828df4687e84
> --- /dev/null
> +++ b/arch/mips/generic/ingenic-smp-entry.S
> @@ -0,0 +1,54 @@
> +/* SPDX-License-Identifier: GPL-2.0-or-later */
> +/*
> + * Copyright (C) 2013 Paul Burton <paul.burton@imgtec.com>

Use his kernel.org address.

> + * JZ4780 secondary CPU reset entry
> + */
> +
> +#include <asm/addrspace.h>
> +#include <asm/asm.h>
> +#include <asm/asmmacro.h>
> +#include <asm/cacheops.h>
> +#include <asm/mipsregs.h>
> +
> +#define CACHE_SIZE	(32 * 1024)
> +#define CACHE_LINESIZE	32
> +
> +	/* The CP0 REIM register only stores the upper 16 address bits. */
> +	.section .text.smp-entry, "ax"
> +	.balign 0x10000
> +	.set noreorder
> +LEAF(jz4780_secondary_cpu_entry)
> +	mtc0	zero, CP0_CAUSE
> +	li	t0, ST0_CU0
> +	mtc0	t0, CP0_STATUS
> +	ehb
> +
> +	/* Invalidate both private caches before entering cached code. */
> +	li	t0, KSEG0
> +	ori	t1, t0, CACHE_SIZE
> +	mtc0	zero, CP0_TAGLO
> +	ehb
> +1:	cache	Index_Store_Tag_I, 0(t0)
> +	cache	Index_Store_Tag_D, 0(t0)
> +	addiu	t0, t0, CACHE_LINESIZE
> +	bne	t0, t1, 1b
> +	 nop
> +
> +	mfc0	t0, CP0_CONFIG
> +	li	t1, ~CONF_CM_CMASK
> +	and	t0, t0, t1
> +	ori	t0, t0, CONF_CM_CACHABLE_NONCOHERENT
> +	mtc0	t0, CP0_CONFIG
> +	mtc0	zero, CP0_PAGEMASK
> +	ehb
> +
> +	/* CPU0 publishes these in RAM before releasing CPU1 from reset. */
> +	la	t0, jz4780_cpu_entry_sp
> +	lw	sp, 0(t0)
> +	la	t0, jz4780_cpu_entry_gp
> +	lw	gp, 0(t0)
> +
> +	la	t0, smp_bootstrap
> +	jr	t0
> +	 nop
> +END(jz4780_secondary_cpu_entry)
> diff --git a/arch/mips/generic/ingenic-smp.c b/arch/mips/generic/ingenic-smp.c
> new file mode 100644
> index 000000000000..98c1fd7483db
> --- /dev/null
> +++ b/arch/mips/generic/ingenic-smp.c
> @@ -0,0 +1,240 @@
> +// SPDX-License-Identifier: GPL-2.0-or-later
> +/*
> + * JZ4780 SMP support, based on Paul Burton's CI20 implementation and
> + * Zhou Yanjie's 2020 JZ4780 SMP series.
> + * Copyright (C) 2013 Paul Burton <paul.burton@imgtec.com>

Use his kernel.org address.

> + */
> +
> +#include <linux/clk.h>
> +#include <linux/cpu.h>
> +#include <linux/interrupt.h>
> +#include <linux/of.h>
> +#include <linux/of_clk.h>
> +#include <linux/sched/task_stack.h>
> +#include <linux/smp.h>
> +
> +#include <asm/addrspace.h>
> +#include <asm/irq_cpu.h>
> +#include <asm/mipsregs.h>
> +#include <asm/r4kcache.h>
> +#include <asm/smp-ops.h>
> +
> +#include "ingenic-smp.h"
> +
> +#define read_c0_corectrl()	__read_32bit_c0_register($12, 2)
> +#define write_c0_corectrl(val)	__write_32bit_c0_register($12, 2, val)
> +#define read_c0_corestatus()	__read_32bit_c0_register($12, 3)
> +#define write_c0_corestatus(val)	__write_32bit_c0_register($12, 3,
> val)
> +#define read_c0_reim()		__read_32bit_c0_register($12, 4)
> +#define write_c0_reim(val)	__write_32bit_c0_register($12, 4, val)
> +#define read_c0_mailbox0()	__read_32bit_c0_register($20, 0)
> +#define write_c0_mailbox0(val)	__write_32bit_c0_register($20, 0, val)
> +#define read_c0_mailbox1()	__read_32bit_c0_register($20, 1)
> +#define write_c0_mailbox1(val)	__write_32bit_c0_register($20, 1, val)
> +
> +#define JZ4780_NR_CPUS		2
> +#define JZ4780_MBOX_IRQ		(MIPS_CPU_IRQ_BASE + 3)
> +#define CORECTRL_SWRST(cpu)	BIT(cpu)
> +#define CORECTRL_RPC(cpu)	BIT(8 + (cpu))
> +#define CORESTATUS_MIRQ(cpu)	BIT(cpu)
> +#define REIM_ENTRY		GENMASK(31, 16)
> +#define REIM_MBOXIRQ(cpu)	BIT(cpu)
> +
> +static struct clk *cpu1_clock;
> +static DEFINE_RAW_SPINLOCK(smp_lock);
> +
> +/* Referenced by the uncached secondary CPU reset entry. */
> +unsigned long jz4780_cpu_entry_sp;
> +unsigned long jz4780_cpu_entry_gp;
> +
> +static irqreturn_t jz4780_mbox_interrupt(int irq, void *dev_id)
> +{
> +	unsigned int cpu = cpu_logical_map(smp_processor_id());
> +	u32 action, status;
> +
> +	raw_spin_lock(&smp_lock);
> +	if (cpu == 0) {
> +		action = read_c0_mailbox0();
> +		write_c0_mailbox0(0);
> +	} else {
> +		action = read_c0_mailbox1();
> +		write_c0_mailbox1(0);
> +	}

Zhou's v8 supported the 4 mailboxes, this feels like a regression.

I understand that we can't really test a quad-core system (unless Nikolaus can?)
so I won't object, but that might need to change in the future.

> +
> +	status = read_c0_corestatus();
> +	write_c0_corestatus(status & ~CORESTATUS_MIRQ(cpu));
> +	irq_disable_hazard();
> +	raw_spin_unlock(&smp_lock);
> +
> +	if (action & SMP_RESCHEDULE_YOURSELF)
> +		scheduler_ipi();
> +	if (action & SMP_CALL_FUNCTION)
> +		generic_smp_call_function_interrupt();
> +
> +	return IRQ_HANDLED;
> +}
> +
> +static void __init jz4780_smp_setup(void)
> +{
> +	struct device_node *np;
> +	u32 entry, reim;
> +	u64 hwid;
> +
> +	for_each_of_cpu_node(np) {
> +		hwid = of_get_cpu_hwid(np, 0);
> +		if (hwid >= nr_cpu_ids || hwid >= JZ4780_NR_CPUS ||

I think these checks on hwid are pointless;
- nr_cpu_ids will never be below hwid unless you read the CPU nodes in reverse
(in which case that's a DT issue),
- hwid >= JZ4780_NR_CPUS will only be true if the DT is broken, too.

> +		    !of_device_is_available(np))
> +			continue;
> +		__cpu_number_map[hwid] = hwid;
> +		__cpu_logical_map[hwid] = hwid;

That looks wrong. The map index should be incremented for each entry you add,
here you'll have holes if one CPU is offline.

> +		set_cpu_possible(hwid, true);
> +	}
> +
> +	/* Program the shared reset vector with mailbox delivery masked. */
> +	reim = read_c0_reim();
> +	reim &= ~(REIM_MBOXIRQ(0) | REIM_MBOXIRQ(1));
> +	write_c0_reim(reim);
> +	write_c0_mailbox0(0);
> +	write_c0_mailbox1(0);
> +	write_c0_corestatus(0);
> +
> +	entry = KSEG1ADDR((unsigned long)jz4780_secondary_cpu_entry);
> +	write_c0_reim((reim & ~REIM_ENTRY) | (entry & REIM_ENTRY));
> +}
> +
> +static void __init jz4780_prepare_cpus(unsigned int max_cpus)
> +{
> +	struct device_node *np;
> +	u32 reim;
> +	int ret;
> +
> +	ret = request_irq(JZ4780_MBOX_IRQ, jz4780_mbox_interrupt,
> +			  IRQF_PERCPU | IRQF_NO_THREAD, "core mailbox",
> NULL);
> +	if (ret)
> +		panic("JZ4780: cannot request mailbox IRQ: %d", ret);

I think this panic is superfluous - if you return here instead I believe that
the system will still be working, although in single-CPU mode.

> +
> +	reim = read_c0_reim();
> +	write_c0_reim(reim | REIM_MBOXIRQ(0));
> +	set_c0_status(STATUSF_IP3);
> +	irq_enable_hazard();
> +
> +	if (!cpu_possible(1))
> +		return;
> +
> +	np = of_get_cpu_node(1, NULL);
> +	cpu1_clock = of_clk_get(np, 0);
> +	of_node_put(np);
> +	if (IS_ERR(cpu1_clock)) {
> +		pr_err("JZ4780: cannot get CPU1 clock: %ld\n",
> +		       PTR_ERR(cpu1_clock));
> +		return;
> +	}
> +
> +	ret = clk_prepare(cpu1_clock);
> +	if (ret) {
> +		clk_put(cpu1_clock);
> +		cpu1_clock = ERR_PTR(ret);
> +		pr_err("JZ4780: cannot prepare CPU1 clock: %d\n", ret);
> +	}
> +}
> +
> +static int jz4780_boot_secondary(int cpu, struct task_struct *idle)
> +{
> +	unsigned long flags;
> +	u32 ctrl;
> +	int ret;
> +
> +	if (cpu_logical_map(cpu) != 1 || IS_ERR_OR_NULL(cpu1_clock))
> +		return -ENODEV;
> +
> +	raw_spin_lock_irqsave(&smp_lock, flags);
> +	ctrl = read_c0_corectrl() | CORECTRL_SWRST(1) | CORECTRL_RPC(1);
> +	write_c0_corectrl(ctrl);
> +	back_to_back_c0_hazard();
> +
> +	ret = clk_enable(cpu1_clock);
> +	if (ret)
> +		goto unlock;
> +
> +	jz4780_cpu_entry_sp = __KSTK_TOS(idle);
> +	jz4780_cpu_entry_gp = (unsigned long)task_thread_info(idle);
> +	/* Publish bootstrap state before CPU1 starts using its private
> cache. */
> +	r4k_blast_dcache();
> +	/* Complete cache writeback before releasing reset. */
> +	mb();
> +	write_c0_corectrl(ctrl & ~CORECTRL_SWRST(1));
> +	back_to_back_c0_hazard();
> +unlock:
> +	raw_spin_unlock_irqrestore(&smp_lock, flags);
> +	return ret;
> +}
> +
> +static void jz4780_init_secondary(void)
> +{
> +	clear_c0_status(ST0_IM);
> +	irq_disable_hazard();
> +}
> +
> +static void jz4780_smp_finish(void)
> +{
> +	unsigned int cpu = cpu_logical_map(smp_processor_id());
> +	u32 reim;
> +
> +	raw_spin_lock(&smp_lock);
> +	reim = read_c0_reim();
> +	write_c0_reim(reim | REIM_MBOXIRQ(cpu));
> +	raw_spin_unlock(&smp_lock);
> +
> +	change_c0_status(ST0_IM, STATUSF_IP3 | STATUSF_IP2 |
> +			 STATUSF_IP1 | STATUSF_IP0);
> +	irq_enable_hazard();
> +	local_irq_enable();
> +}
> +
> +static void jz4780_send_ipi_locked(int cpu, unsigned int action)
> +{
> +	if (cpu_logical_map(cpu) == 0)
> +		write_c0_mailbox0(read_c0_mailbox0() | action);
> +	else
> +		write_c0_mailbox1(read_c0_mailbox1() | action);
> +	back_to_back_c0_hazard();
> +}
> +
> +static void jz4780_send_ipi_single(int cpu, unsigned int action)
> +{
> +	unsigned long flags;
> +
> +	raw_spin_lock_irqsave(&smp_lock, flags);
> +	jz4780_send_ipi_locked(cpu, action);
> +	raw_spin_unlock_irqrestore(&smp_lock, flags);
> +}
> +
> +static void jz4780_send_ipi_mask(const struct cpumask *mask, unsigned int
> action)
> +{
> +	unsigned long flags;
> +	int cpu;
> +
> +	raw_spin_lock_irqsave(&smp_lock, flags);
> +	for_each_cpu(cpu, mask)
> +		jz4780_send_ipi_locked(cpu, action);
> +	raw_spin_unlock_irqrestore(&smp_lock, flags);
> +}
> +
> +static const struct plat_smp_ops jz4780_smp_ops = {
> +	.send_ipi_single = jz4780_send_ipi_single,
> +	.send_ipi_mask = jz4780_send_ipi_mask,
> +	.init_secondary = jz4780_init_secondary,
> +	.smp_finish = jz4780_smp_finish,
> +	.boot_secondary = jz4780_boot_secondary,
> +	.smp_setup = jz4780_smp_setup,
> +	.prepare_cpus = jz4780_prepare_cpus,
> +};
> +
> +int __init register_ingenic_smp_ops(void)
> +{
> +	if (!of_machine_is_compatible("ingenic,jz4780"))
> +		return -ENODEV;
> +
> +	register_smp_ops(&jz4780_smp_ops);
> +	return 0;
> +}

Overall, the code itself is pretty much a Xburst (CPU) driver that gives the
idea (through function names, etc) that it is SoC-specific (JZ4780), while the
source file has a very generic "ingenic-smp.c" name.

So I'd suggest to replace all 'jz4780' by 'xburst'.

> diff --git a/arch/mips/generic/ingenic-smp.h b/arch/mips/generic/ingenic-smp.h
> new file mode 100644
> index 000000000000..54191be8731a
> --- /dev/null
> +++ b/arch/mips/generic/ingenic-smp.h
> @@ -0,0 +1,20 @@
> +/* SPDX-License-Identifier: GPL-2.0-or-later */
> +/* Copyright (C) 2013 Paul Burton <paul.burton@imgtec.com> */

Use his kernel.org address.

> +#ifndef __MIPS_GENERIC_INGENIC_SMP_H__
> +#define __MIPS_GENERIC_INGENIC_SMP_H__
> +
> +#include <linux/errno.h>
> +#include <linux/init.h>
> +#include <linux/kconfig.h>
> +
> +#ifdef CONFIG_MIPS_INGENIC_SMP
> +int __init register_ingenic_smp_ops(void);
> +void jz4780_secondary_cpu_entry(void);
> +#else
> +static inline int register_ingenic_smp_ops(void)
> +{
> +	return -ENODEV;
> +}
> +#endif
> +
> +#endif /* __MIPS_GENERIC_INGENIC_SMP_H__ */
> diff --git a/arch/mips/generic/init.c b/arch/mips/generic/init.c
> index 1d712eac1617..00149bd1e9a9 100644
> --- a/arch/mips/generic/init.c
> +++ b/arch/mips/generic/init.c
> @@ -20,6 +20,8 @@
>  #include <asm/smp-ops.h>
>  #include <asm/time.h>
>  
> +#include "ingenic-smp.h"
> +
>  static __initconst const void *fdt;
>  static __initconst const struct mips_machine *mach;
>  static __initconst const void *mach_match_data;
> @@ -113,6 +115,8 @@ void __init device_tree_init(void)
>  	unflatten_and_copy_device_tree();
>  	mips_cpc_probe();
>  
> +	if (!register_ingenic_smp_ops())
> +		return;
>  	if (!register_cps_smp_ops())
>  		return;
>  	if (!register_vsmp_smp_ops())
> diff --git a/arch/mips/ingenic/Kconfig b/arch/mips/ingenic/Kconfig
> index f595b339a4b8..1dfd0bfd0193 100644
> --- a/arch/mips/ingenic/Kconfig
> +++ b/arch/mips/ingenic/Kconfig
> @@ -64,6 +64,12 @@ config MACH_JZ4780
>  	select MIPS_CPU_SCACHE
>  	select SYS_HAS_CPU_MIPS32_R2
>  	select SYS_SUPPORTS_HIGHMEM
> +	select SYS_SUPPORTS_SMP
> +
> +config MIPS_INGENIC_SMP
> +	def_bool y
> +	depends on MACH_JZ4780 && SMP
> +	select SMP_UP

I don't think you need an extra symbol here, you could just add 'select SMP_UP
if SMP' to the MACH_JZ4780 symbol.

>  
>  config MACH_X1000
>  	bool
> diff --git a/arch/mips/kernel/idle.c b/arch/mips/kernel/idle.c
> index 80e8a04a642e..6cabdd4825a7 100644
> --- a/arch/mips/kernel/idle.c
> +++ b/arch/mips/kernel/idle.c
> @@ -18,6 +18,7 @@
>  #include <asm/cpu-type.h>
>  #include <asm/idle.h>
>  #include <asm/mipsregs.h>
> +#include <asm/r4kcache.h>
>  
>  /*
>   * Not all of the MIPS CPUs have the "wait" instruction available. Moreover,
> @@ -71,6 +72,23 @@ static void __cpuidle rm7k_wait_irqoff(void)
>  		"	.set	pop					\n");
>  }
>  
> +/* XBurst gates its private cache clock in WAIT; write back dirty data first.
> */
> +static void __cpuidle ingenic_smp_wait_irqoff(void)
> +{
> +	unsigned long pending = read_c0_cause() & read_c0_status() &
> CAUSEF_IP;
> +
> +	if (!need_resched() && !pending) {
> +		r4k_blast_dcache();
> +		__asm__ __volatile__(
> +			".set push\n"
> +			".set mips3\n"
> +			"sync\n"
> +			"wait\n"
> +			".set pop\n"
> +			: : : "memory");
> +	}
> +}
> +
>  /*
>   * Au1 'wait' is only useful when the 32kHz counter is used as timer,
>   * since coreclock (and the cp0 counter) stops upon executing it. Only an
> @@ -156,10 +174,15 @@ void __init check_wait(void)
>  	case CPU_CAVIUM_OCTEON_PLUS:
>  	case CPU_CAVIUM_OCTEON2:
>  	case CPU_CAVIUM_OCTEON3:
> -	case CPU_XBURST:
>  	case CPU_LOONGSON32:
>  		cpu_wait = r4k_wait;
>  		break;
> +	case CPU_XBURST:
> +		if (IS_ENABLED(CONFIG_MIPS_INGENIC_SMP))
> +			cpu_wait = ingenic_smp_wait_irqoff;
> +		else
> +			cpu_wait = r4k_wait;
> +		break;
>  	case CPU_LOONGSON64:
>  		if ((c->processor_id & (PRID_IMP_MASK | PRID_REV_MASK)) >=
>  				(PRID_IMP_LOONGSON_64C |
> PRID_REV_LOONGSON3A_R2_0) ||

Cheers,
-Paul

Re: [PATCH 2/3] MIPS: Ingenic: Add JZ4780 SMP support to the generic platform

[H. Nikolaus Schaller]

> Am 06.11.2026 um 12:07 schrieb Paul Cercueil <paul@crapouillou.net>:
> 
> Zhou's v8 supported the 4 mailboxes, this feels like a regression.
> 
> I understand that we can't really test a quad-core system (unless Nikolaus can?)
> so I won't object, but that might need to change in the future.

Well, I am not aware that a quad core Xburst processor exists... I can
try to ask Lumissil staff next week on the Electronica Trade Fair.

BR,
Nikolaus

Re: [PATCH 2/3] MIPS: Ingenic: Add JZ4780 SMP support to the generic platform

[H. Nikolaus Schaller]

Hi,

> Am 05.11.2026 um 21:32 schrieb Florian Fainelli <f.fainelli@gmail.com>:
> 
> From: 周琰杰 (Zhou Yanjie) <zhouyanjie@wanyeetech.com>
> 
> Port the JZ4780 reset-entry and CP0 mailbox implementation from Zhou
> Yanjie's 2020 v8 series, based on Paul Burton's original CI20 SMP code.
> The old arch/mips/jz4740 platform has moved into the generic MIPS code,
> so register the Ingenic SMP operations after unflattening the device
> tree and before selecting the generic fallback implementations.
> 
> Reuse the upstream CPU clock gates, CPU device-tree nodes and per-CPU
> TCU timers. Enumerate only available CPU nodes within the two-core
> hardware limit. Prepare the CPU1 clock before boot, propagate clock
> failures, serialize mailbox accesses with a raw spinlock and publish
> bootstrap state before releasing the secondary core from reset. Align
> the secondary reset entry to the REIM register's 64 KiB granularity.
> Retain the UP fallback for other SoCs in a shared Ingenic kernel.
> 
> Write back the private data cache before WAIT, since XBurst gates its
> clock during idle and another core must not access dirty data in that
> sleeping cache. Return with interrupts disabled as required by the
> current idle API. Existing per-CPU TCU events make the old forced
> broadcast-timer workaround unnecessary.
> 
> This adaptation incorporates the CPU-enumeration review feedback and
> removes the old unused CPU mask. Preserve the original source credits;
> historical Tested-by and Reviewed-by tags are not carried to this
> substantially changed port. Adaptation and validation used Codex
> assistance.
> 
> Link: https://lkml.iu.edu/hypermail/linux/kernel/2005.2/03488.html
> Link: https://lists.openwall.net/linux-kernel/2020/05/13/1482
> Link: https://lists.openwall.net/linux-kernel/2020/05/19/1099
> Signed-off-by: 周琰杰 (Zhou Yanjie) <zhouyanjie@wanyeetech.com>
> [florian: forward ported to latest version]
> Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
> ---
> MAINTAINERS                           |   1 +
> arch/mips/generic/Makefile            |   1 +
> arch/mips/generic/ingenic-smp-entry.S |  54 ++++++
> arch/mips/generic/ingenic-smp.c       | 240 ++++++++++++++++++++++++++
> arch/mips/generic/ingenic-smp.h       |  20 +++
> arch/mips/generic/init.c              |   4 +
> arch/mips/ingenic/Kconfig             |   6 +
> arch/mips/kernel/idle.c               |  25 ++-
> 8 files changed, 350 insertions(+), 1 deletion(-)
> create mode 100644 arch/mips/generic/ingenic-smp-entry.S
> create mode 100644 arch/mips/generic/ingenic-smp.c
> create mode 100644 arch/mips/generic/ingenic-smp.h
> 
> diff --git a/MAINTAINERS b/MAINTAINERS
> index 09713dc4c6a7..101027579ccd 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -13117,6 +13117,7 @@ L: linux-mips@vger.kernel.org
> S: Maintained
> F: arch/mips/boot/dts/ingenic/
> F: arch/mips/generic/board-ingenic.c
> +F: arch/mips/generic/ingenic-smp*
> F: arch/mips/include/asm/mach-ingenic/
> F: arch/mips/ingenic/Kconfig
> F: drivers/clk/ingenic/
> diff --git a/arch/mips/generic/Makefile b/arch/mips/generic/Makefile
> index ea0e4ad5e600..df92650c0f15 100644
> --- a/arch/mips/generic/Makefile
> +++ b/arch/mips/generic/Makefile
> @@ -12,5 +12,6 @@ obj-$(CONFIG_YAMON_DT_SHIM) += yamon-dt.o
> obj-$(CONFIG_LEGACY_BOARD_SEAD3) += board-sead3.o
> obj-$(CONFIG_LEGACY_BOARD_OCELOT) += board-ocelot.o
> obj-$(CONFIG_MACH_INGENIC) += board-ingenic.o
> +obj-$(CONFIG_MIPS_INGENIC_SMP) += ingenic-smp.o ingenic-smp-entry.o
> obj-$(CONFIG_VIRT_BOARD_RANCHU) += board-ranchu.o
> obj-$(CONFIG_MACH_REALTEK_RTL) += board-realtek.o
> diff --git a/arch/mips/generic/ingenic-smp-entry.S b/arch/mips/generic/ingenic-smp-entry.S
> new file mode 100644
> index 000000000000..828df4687e84
> --- /dev/null
> +++ b/arch/mips/generic/ingenic-smp-entry.S
> @@ -0,0 +1,54 @@
> +/* SPDX-License-Identifier: GPL-2.0-or-later */
> +/*
> + * Copyright (C) 2013 Paul Burton <paul.burton@imgtec.com>
> + * JZ4780 secondary CPU reset entry
> + */
> +
> +#include <asm/addrspace.h>
> +#include <asm/asm.h>
> +#include <asm/asmmacro.h>
> +#include <asm/cacheops.h>
> +#include <asm/mipsregs.h>
> +
> +#define CACHE_SIZE (32 * 1024)
> +#define CACHE_LINESIZE 32
> +
> + /* The CP0 REIM register only stores the upper 16 address bits. */
> + .section .text.smp-entry, "ax"
> + .balign 0x10000
> + .set noreorder
> +LEAF(jz4780_secondary_cpu_entry)
> + mtc0 zero, CP0_CAUSE
> + li t0, ST0_CU0
> + mtc0 t0, CP0_STATUS
> + ehb
> +
> + /* Invalidate both private caches before entering cached code. */
> + li t0, KSEG0
> + ori t1, t0, CACHE_SIZE
> + mtc0 zero, CP0_TAGLO
> + ehb
> +1: cache Index_Store_Tag_I, 0(t0)
> + cache Index_Store_Tag_D, 0(t0)
> + addiu t0, t0, CACHE_LINESIZE
> + bne t0, t1, 1b
> + nop
> +
> + mfc0 t0, CP0_CONFIG
> + li t1, ~CONF_CM_CMASK
> + and t0, t0, t1
> + ori t0, t0, CONF_CM_CACHABLE_NONCOHERENT
> + mtc0 t0, CP0_CONFIG
> + mtc0 zero, CP0_PAGEMASK
> + ehb
> +
> + /* CPU0 publishes these in RAM before releasing CPU1 from reset. */
> + la t0, jz4780_cpu_entry_sp
> + lw sp, 0(t0)
> + la t0, jz4780_cpu_entry_gp
> + lw gp, 0(t0)
> +
> + la t0, smp_bootstrap
> + jr t0
> + nop
> +END(jz4780_secondary_cpu_entry)
> diff --git a/arch/mips/generic/ingenic-smp.c b/arch/mips/generic/ingenic-smp.c
> new file mode 100644
> index 000000000000..98c1fd7483db
> --- /dev/null
> +++ b/arch/mips/generic/ingenic-smp.c
> @@ -0,0 +1,240 @@
> +// SPDX-License-Identifier: GPL-2.0-or-later
> +/*
> + * JZ4780 SMP support, based on Paul Burton's CI20 implementation and
> + * Zhou Yanjie's 2020 JZ4780 SMP series.
> + * Copyright (C) 2013 Paul Burton <paul.burton@imgtec.com>
> + */
> +
> +#include <linux/clk.h>
> +#include <linux/cpu.h>
> +#include <linux/interrupt.h>
> +#include <linux/of.h>
> +#include <linux/of_clk.h>
> +#include <linux/sched/task_stack.h>
> +#include <linux/smp.h>
> +
> +#include <asm/addrspace.h>
> +#include <asm/irq_cpu.h>
> +#include <asm/mipsregs.h>
> +#include <asm/r4kcache.h>
> +#include <asm/smp-ops.h>
> +
> +#include "ingenic-smp.h"
> +
> +#define read_c0_corectrl() __read_32bit_c0_register($12, 2)
> +#define write_c0_corectrl(val) __write_32bit_c0_register($12, 2, val)
> +#define read_c0_corestatus() __read_32bit_c0_register($12, 3)
> +#define write_c0_corestatus(val) __write_32bit_c0_register($12, 3, val)
> +#define read_c0_reim() __read_32bit_c0_register($12, 4)
> +#define write_c0_reim(val) __write_32bit_c0_register($12, 4, val)
> +#define read_c0_mailbox0() __read_32bit_c0_register($20, 0)
> +#define write_c0_mailbox0(val) __write_32bit_c0_register($20, 0, val)
> +#define read_c0_mailbox1() __read_32bit_c0_register($20, 1)
> +#define write_c0_mailbox1(val) __write_32bit_c0_register($20, 1, val)
> +
> +#define JZ4780_NR_CPUS 2
> +#define JZ4780_MBOX_IRQ (MIPS_CPU_IRQ_BASE + 3)
> +#define CORECTRL_SWRST(cpu) BIT(cpu)
> +#define CORECTRL_RPC(cpu) BIT(8 + (cpu))
> +#define CORESTATUS_MIRQ(cpu) BIT(cpu)
> +#define REIM_ENTRY GENMASK(31, 16)
> +#define REIM_MBOXIRQ(cpu) BIT(cpu)
> +
> +static struct clk *cpu1_clock;
> +static DEFINE_RAW_SPINLOCK(smp_lock);
> +
> +/* Referenced by the uncached secondary CPU reset entry. */
> +unsigned long jz4780_cpu_entry_sp;
> +unsigned long jz4780_cpu_entry_gp;
> +
> +static irqreturn_t jz4780_mbox_interrupt(int irq, void *dev_id)
> +{
> + unsigned int cpu = cpu_logical_map(smp_processor_id());
> + u32 action, status;
> +
> + raw_spin_lock(&smp_lock);
> + if (cpu == 0) {
> + action = read_c0_mailbox0();
> + write_c0_mailbox0(0);
> + } else {
> + action = read_c0_mailbox1();
> + write_c0_mailbox1(0);
> + }
> +
> + status = read_c0_corestatus();
> + write_c0_corestatus(status & ~CORESTATUS_MIRQ(cpu));
> + irq_disable_hazard();
> + raw_spin_unlock(&smp_lock);
> +
> + if (action & SMP_RESCHEDULE_YOURSELF)
> + scheduler_ipi();
> + if (action & SMP_CALL_FUNCTION)
> + generic_smp_call_function_interrupt();
> +
> + return IRQ_HANDLED;
> +}
> +
> +static void __init jz4780_smp_setup(void)
> +{
> + struct device_node *np;
> + u32 entry, reim;
> + u64 hwid;
> +
> + for_each_of_cpu_node(np) {
> + hwid = of_get_cpu_hwid(np, 0);
> + if (hwid >= nr_cpu_ids || hwid >= JZ4780_NR_CPUS ||
> +    !of_device_is_available(np))
> + continue;
> + __cpu_number_map[hwid] = hwid;
> + __cpu_logical_map[hwid] = hwid;
> + set_cpu_possible(hwid, true);
> + }
> +
> + /* Program the shared reset vector with mailbox delivery masked. */
> + reim = read_c0_reim();
> + reim &= ~(REIM_MBOXIRQ(0) | REIM_MBOXIRQ(1));
> + write_c0_reim(reim);
> + write_c0_mailbox0(0);
> + write_c0_mailbox1(0);
> + write_c0_corestatus(0);
> +
> + entry = KSEG1ADDR((unsigned long)jz4780_secondary_cpu_entry);
> + write_c0_reim((reim & ~REIM_ENTRY) | (entry & REIM_ENTRY));
> +}
> +
> +static void __init jz4780_prepare_cpus(unsigned int max_cpus)
> +{
> + struct device_node *np;
> + u32 reim;
> + int ret;
> +
> + ret = request_irq(JZ4780_MBOX_IRQ, jz4780_mbox_interrupt,
> +  IRQF_PERCPU | IRQF_NO_THREAD, "core mailbox", NULL);
> + if (ret)
> + panic("JZ4780: cannot request mailbox IRQ: %d", ret);
> +
> + reim = read_c0_reim();
> + write_c0_reim(reim | REIM_MBOXIRQ(0));
> + set_c0_status(STATUSF_IP3);
> + irq_enable_hazard();
> +
> + if (!cpu_possible(1))
> + return;
> +
> + np = of_get_cpu_node(1, NULL);
> + cpu1_clock = of_clk_get(np, 0);
> + of_node_put(np);
> + if (IS_ERR(cpu1_clock)) {
> + pr_err("JZ4780: cannot get CPU1 clock: %ld\n",
> +       PTR_ERR(cpu1_clock));
> + return;
> + }
> +
> + ret = clk_prepare(cpu1_clock);
> + if (ret) {
> + clk_put(cpu1_clock);
> + cpu1_clock = ERR_PTR(ret);
> + pr_err("JZ4780: cannot prepare CPU1 clock: %d\n", ret);
> + }
> +}
> +
> +static int jz4780_boot_secondary(int cpu, struct task_struct *idle)
> +{
> + unsigned long flags;
> + u32 ctrl;
> + int ret;
> +
> + if (cpu_logical_map(cpu) != 1 || IS_ERR_OR_NULL(cpu1_clock))
> + return -ENODEV;
> +
> + raw_spin_lock_irqsave(&smp_lock, flags);
> + ctrl = read_c0_corectrl() | CORECTRL_SWRST(1) | CORECTRL_RPC(1);
> + write_c0_corectrl(ctrl);
> + back_to_back_c0_hazard();
> +
> + ret = clk_enable(cpu1_clock);
> + if (ret)
> + goto unlock;
> +
> + jz4780_cpu_entry_sp = __KSTK_TOS(idle);
> + jz4780_cpu_entry_gp = (unsigned long)task_thread_info(idle);
> + /* Publish bootstrap state before CPU1 starts using its private cache. */
> + r4k_blast_dcache();
> + /* Complete cache writeback before releasing reset. */
> + mb();
> + write_c0_corectrl(ctrl & ~CORECTRL_SWRST(1));
> + back_to_back_c0_hazard();
> +unlock:
> + raw_spin_unlock_irqrestore(&smp_lock, flags);
> + return ret;
> +}
> +
> +static void jz4780_init_secondary(void)
> +{
> + clear_c0_status(ST0_IM);
> + irq_disable_hazard();
> +}
> +
> +static void jz4780_smp_finish(void)
> +{
> + unsigned int cpu = cpu_logical_map(smp_processor_id());
> + u32 reim;
> +
> + raw_spin_lock(&smp_lock);
> + reim = read_c0_reim();
> + write_c0_reim(reim | REIM_MBOXIRQ(cpu));
> + raw_spin_unlock(&smp_lock);
> +
> + change_c0_status(ST0_IM, STATUSF_IP3 | STATUSF_IP2 |
> + STATUSF_IP1 | STATUSF_IP0);
> + irq_enable_hazard();
> + local_irq_enable();
> +}
> +
> +static void jz4780_send_ipi_locked(int cpu, unsigned int action)
> +{
> + if (cpu_logical_map(cpu) == 0)
> + write_c0_mailbox0(read_c0_mailbox0() | action);
> + else
> + write_c0_mailbox1(read_c0_mailbox1() | action);
> + back_to_back_c0_hazard();
> +}
> +
> +static void jz4780_send_ipi_single(int cpu, unsigned int action)
> +{
> + unsigned long flags;
> +
> + raw_spin_lock_irqsave(&smp_lock, flags);
> + jz4780_send_ipi_locked(cpu, action);
> + raw_spin_unlock_irqrestore(&smp_lock, flags);
> +}
> +
> +static void jz4780_send_ipi_mask(const struct cpumask *mask, unsigned int action)
> +{
> + unsigned long flags;
> + int cpu;
> +
> + raw_spin_lock_irqsave(&smp_lock, flags);
> + for_each_cpu(cpu, mask)
> + jz4780_send_ipi_locked(cpu, action);
> + raw_spin_unlock_irqrestore(&smp_lock, flags);
> +}
> +
> +static const struct plat_smp_ops jz4780_smp_ops = {
> + .send_ipi_single = jz4780_send_ipi_single,
> + .send_ipi_mask = jz4780_send_ipi_mask,
> + .init_secondary = jz4780_init_secondary,
> + .smp_finish = jz4780_smp_finish,
> + .boot_secondary = jz4780_boot_secondary,
> + .smp_setup = jz4780_smp_setup,
> + .prepare_cpus = jz4780_prepare_cpus,
> +};
> +
> +int __init register_ingenic_smp_ops(void)
> +{
> + if (!of_machine_is_compatible("ingenic,jz4780"))

This will restrict to jz4780 and leave out x2000, x2600 in the future.

But for the moment it seems to be ok.

> + return -ENODEV;
> +
> + register_smp_ops(&jz4780_smp_ops);
> + return 0;
> +}
> diff --git a/arch/mips/generic/ingenic-smp.h b/arch/mips/generic/ingenic-smp.h
> new file mode 100644
> index 000000000000..54191be8731a
> --- /dev/null
> +++ b/arch/mips/generic/ingenic-smp.h
> @@ -0,0 +1,20 @@
> +/* SPDX-License-Identifier: GPL-2.0-or-later */
> +/* Copyright (C) 2013 Paul Burton <paul.burton@imgtec.com> */
> +#ifndef __MIPS_GENERIC_INGENIC_SMP_H__
> +#define __MIPS_GENERIC_INGENIC_SMP_H__
> +
> +#include <linux/errno.h>
> +#include <linux/init.h>
> +#include <linux/kconfig.h>
> +
> +#ifdef CONFIG_MIPS_INGENIC_SMP

is this needed if the only user is not compiled in case of undefined CONFIG_MIPS_INGENIC_SMP?

Makefile:
> +obj-$(CONFIG_MIPS_INGENIC_SMP) += ingenic-smp.o ingenic-smp-entry.o

> +int __init register_ingenic_smp_ops(void);
> +void jz4780_secondary_cpu_entry(void);
> +#else
> +static inline int register_ingenic_smp_ops(void)
> +{
> + return -ENODEV;
> +}
> +#endif
> +
> +#endif /* __MIPS_GENERIC_INGENIC_SMP_H__ */
> diff --git a/arch/mips/generic/init.c b/arch/mips/generic/init.c
> index 1d712eac1617..00149bd1e9a9 100644
> --- a/arch/mips/generic/init.c
> +++ b/arch/mips/generic/init.c
> @@ -20,6 +20,8 @@
> #include <asm/smp-ops.h>
> #include <asm/time.h>
> 
> +#include "ingenic-smp.h"
> +
> static __initconst const void *fdt;
> static __initconst const struct mips_machine *mach;
> static __initconst const void *mach_match_data;
> @@ -113,6 +115,8 @@ void __init device_tree_init(void)
> unflatten_and_copy_device_tree();
> mips_cpc_probe();
> 
> + if (!register_ingenic_smp_ops())
> + return;
> if (!register_cps_smp_ops())
> return;
> if (!register_vsmp_smp_ops())
> diff --git a/arch/mips/ingenic/Kconfig b/arch/mips/ingenic/Kconfig
> index f595b339a4b8..1dfd0bfd0193 100644
> --- a/arch/mips/ingenic/Kconfig
> +++ b/arch/mips/ingenic/Kconfig
> @@ -64,6 +64,12 @@ config MACH_JZ4780
> select MIPS_CPU_SCACHE
> select SYS_HAS_CPU_MIPS32_R2
> select SYS_SUPPORTS_HIGHMEM
> + select SYS_SUPPORTS_SMP
> +
> +config MIPS_INGENIC_SMP
> + def_bool y
> + depends on MACH_JZ4780 && SMP
> + select SMP_UP

do we need this additional CONFIG option?
Ah, yes the inclusion of the code depends on it.

> 
> config MACH_X1000
> bool
> diff --git a/arch/mips/kernel/idle.c b/arch/mips/kernel/idle.c
> index 80e8a04a642e..6cabdd4825a7 100644
> --- a/arch/mips/kernel/idle.c
> +++ b/arch/mips/kernel/idle.c
> @@ -18,6 +18,7 @@
> #include <asm/cpu-type.h>
> #include <asm/idle.h>
> #include <asm/mipsregs.h>
> +#include <asm/r4kcache.h>
> 
> /*
>  * Not all of the MIPS CPUs have the "wait" instruction available. Moreover,
> @@ -71,6 +72,23 @@ static void __cpuidle rm7k_wait_irqoff(void)
> " .set pop \n");
> }
> 
> +/* XBurst gates its private cache clock in WAIT; write back dirty data first. */
> +static void __cpuidle ingenic_smp_wait_irqoff(void)
> +{
> + unsigned long pending = read_c0_cause() & read_c0_status() & CAUSEF_IP;
> +
> + if (!need_resched() && !pending) {
> + r4k_blast_dcache();
> + __asm__ __volatile__(
> + ".set push\n"
> + ".set mips3\n"
> + "sync\n"
> + "wait\n"
> + ".set pop\n"
> + : : : "memory");
> + }
> +}
> +
> /*
>  * Au1 'wait' is only useful when the 32kHz counter is used as timer,
>  * since coreclock (and the cp0 counter) stops upon executing it. Only an
> @@ -156,10 +174,15 @@ void __init check_wait(void)
> case CPU_CAVIUM_OCTEON_PLUS:
> case CPU_CAVIUM_OCTEON2:
> case CPU_CAVIUM_OCTEON3:
> - case CPU_XBURST:
> case CPU_LOONGSON32:
> cpu_wait = r4k_wait;
> break;
> + case CPU_XBURST:
> + if (IS_ENABLED(CONFIG_SMP))
> + cpu_wait = ingenic_smp_wait_irqoff;
> + else
> + cpu_wait = r4k_wait;
> + break;
> case CPU_LOONGSON64:
> if ((c->processor_id & (PRID_IMP_MASK | PRID_REV_MASK)) >=
> (PRID_IMP_LOONGSON_64C | PRID_REV_LOONGSON3A_R2_0) ||
> -- 
> 2.53.0
> 

I just want to report that I wasn't able to bring up the second CPU:

[    0.108834] CPU 0 Unable to handle kernel paging request at virtual address 00000000, epc == 00000000, ra == 800239cc
[    0.119549] Oops[#1]:
[    0.121845] CPU: 0 UID: 0 PID: 1 Comm: swapper/0 Not tainted 7.4.0-rc1-letux+ #8714 PREEMPT 
[    0.130405] Hardware name: Imagination creator CI20
...
[    0.093381] [<8006fad4>] __cpu_up+0x24/0x98
[    0.097618] [<800891f4>] bringup_cpu+0x48/0x124
[    0.102209] [<800885c4>] __cpuhp_invoke_callback_range+0xb0/0x140
[    0.108387] [<800f359c>] __down_trylock_console_sem.isra.40+0x28/0x58
[    0.114925] [<80089480>] cpu_up.constprop.24+0x190/0x344
[    0.120307] [<800c1950>] try_to_wake_up+0xbc/0xae4
[    0.125166] [<80e0bc7c>] bringup_nonboot_cpus+0xa4/0x134
[    0.130553] [<80e13cec>] smp_init+0x30/0x80
[    0.134789] [<80e07028>] kernel_init_freeable+0x120/0x2e4
[    0.140266] [<80b6eb4c>] kernel_init+0x28/0x124
[    0.144853] [<80b6eb24>] kernel_init+0x0/0x124
[    0.061973] [<800616e4>] ret_from_kernel_thread+0x14/0x1c
[    0.067462] 
[    0.068947] Code: (Bad address in epc)
[    0.068947] 
[    0.075802] ---[ end trace 0000000000000000 ]---
[    0.080430] Kernel panic - not syncing: Fatal exception
[    0.085734] Rebooting in 10 seconds..
[    0.138377] Reboot failed -- System halted

If I disable CONFIG_SMP I can boot a single core.

The reason may be that we have other interfering private configs in our tree
and/or something else is missing.

But I remember that the original patches by Zhou Yanjie had once worked in 2020
on an older kernel release.

So at the moment I can't dig into all this deeper.

BR,
Nikolaus