[PATCH 1/7] powerpc/85xx: re-enable timebase sync disabled by KEXEC patch

[PATCH 1/7] powerpc/85xx: re-enable timebase sync disabled by KEXEC patch

[Li Yang]

The timebase sync is not only necessary when using KEXEC.  It should also
be used by normal boot up and cpu hotplug.  Remove the ifdef added by
the KEXEC patch.  Fix a problem that cpu hotplugging freezes the whole system.

Signed-off-by: Jin Qing <b24347@freescale.com>
Singed-off-by: Li Yang <leoli@freescale.com>
---
 arch/powerpc/platforms/85xx/smp.c |    4 +---
 1 files changed, 1 insertions(+), 3 deletions(-)

diff --git a/arch/powerpc/platforms/85xx/smp.c b/arch/powerpc/platforms/85xx/smp.c
index 5c91a99..1e8aec8 100644
--- a/arch/powerpc/platforms/85xx/smp.c
+++ b/arch/powerpc/platforms/85xx/smp.c
@@ -2,7 +2,7 @@
  * Author: Andy Fleming <afleming@freescale.com>
  * 	   Kumar Gala <galak@kernel.crashing.org>
  *
- * Copyright 2006-2008 Freescale Semiconductor Inc.
+ * Copyright 2006-2010 Freescale Semiconductor Inc.
  *
  * This program is free software; you can redistribute  it and/or modify it
  * under  the terms of  the GNU General  Public License as published by the
@@ -115,10 +115,8 @@ smp_85xx_setup_cpu(int cpu_nr)
 
 struct smp_ops_t smp_85xx_ops = {
 	.kick_cpu = smp_85xx_kick_cpu,
-#ifdef CONFIG_KEXEC
 	.give_timebase	= smp_generic_give_timebase,
 	.take_timebase	= smp_generic_take_timebase,
-#endif
 };
 
 #ifdef CONFIG_KEXEC
-- 
1.6.6-rc1.GIT

[PATCH 2/7] powerpc/85xx: add HOTPLUG_CPU support

[Li Yang]

Add support to disable and re-enable individual cores at runtime
on MPC85xx/QorIQ SMP machines.

This makes suspend/resume possible for SMP systems, as the power management
code on SMP always disable non-boot cpus on suspend.

MPC85xx machines use ePAPR spin-table in boot page for CPU kick-off.
This patch brings the bootpage and spin-table from bootloader into
kernel because the bootpage in bootloader might have been lost at
runtime.  Also add support to boot from physical address larger than
32-bit.

Signed-off-by: Yutaka Ando <y.ando@freescale.com>
Signed-off-by: Li Yang <leoli@freescale.com>
---
 arch/powerpc/Kconfig                   |    2 +-
 arch/powerpc/kernel/Makefile           |    2 +-
 arch/powerpc/kernel/head_fsl_booke.S   |   32 +++++
 arch/powerpc/kernel/smp.c              |    4 +-
 arch/powerpc/platforms/85xx/Makefile   |    4 +-
 arch/powerpc/platforms/85xx/bootpage.S |  206 +++++++++++++++++++++++++++++
 arch/powerpc/platforms/85xx/smp.c      |  222 ++++++++++++++++++++++++++------
 7 files changed, 428 insertions(+), 44 deletions(-)
 create mode 100644 arch/powerpc/platforms/85xx/bootpage.S

diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig
index e625e9e..b1982dd 100644
--- a/arch/powerpc/Kconfig
+++ b/arch/powerpc/Kconfig
@@ -320,7 +320,7 @@ config SWIOTLB
 
 config HOTPLUG_CPU
 	bool "Support for enabling/disabling CPUs"
-	depends on SMP && HOTPLUG && EXPERIMENTAL && (PPC_PSERIES || PPC_PMAC)
+	depends on SMP && HOTPLUG && EXPERIMENTAL && (PPC_PSERIES || PPC_PMAC || E500)
 	---help---
 	  Say Y here to be able to disable and re-enable individual
 	  CPUs at runtime on SMP machines.
diff --git a/arch/powerpc/kernel/Makefile b/arch/powerpc/kernel/Makefile
index 36c30f3..bb20496 100644
--- a/arch/powerpc/kernel/Makefile
+++ b/arch/powerpc/kernel/Makefile
@@ -56,7 +56,7 @@ obj-$(CONFIG_IBMEBUS)           += ibmebus.o
 obj-$(CONFIG_GENERIC_TBSYNC)	+= smp-tbsync.o
 obj-$(CONFIG_CRASH_DUMP)	+= crash_dump.o
 ifeq ($(CONFIG_PPC32),y)
-obj-$(CONFIG_E500)		+= idle_e500.o
+obj-$(CONFIG_E500)		+= idle_e500.o l2cr_85xx.o
 endif
 obj-$(CONFIG_6xx)		+= idle_6xx.o l2cr_6xx.o cpu_setup_6xx.o
 obj-$(CONFIG_TAU)		+= tau_6xx.o
diff --git a/arch/powerpc/kernel/head_fsl_booke.S b/arch/powerpc/kernel/head_fsl_booke.S
index 529b817..61d9c46 100644
--- a/arch/powerpc/kernel/head_fsl_booke.S
+++ b/arch/powerpc/kernel/head_fsl_booke.S
@@ -23,6 +23,7 @@
  *	PowerPC 44x support, Matt Porter <mporter@kernel.crashing.org>
  *    Copyright 2004 Freescale Semiconductor, Inc
  *	PowerPC e500 modifications, Kumar Gala <galak@kernel.crashing.org>
+ *    Copyright 2008, 2010 Freescale Semiconductor, Inc.
  *
  * This program is free software; you can redistribute  it and/or modify it
  * under  the terms of  the GNU General  Public License as published by the
@@ -254,6 +255,37 @@ _ENTRY(__early_start)
 	lwz	r11, 0(r12);		/* Get Linux PTE */
 #endif
 
+_GLOBAL(flush_disable_L1)
+/*
+ * Flush L1 d-cache, invalidate and disable d-cache,
+ * invalidate and disable i-cache
+ */
+	mflr	r10
+	bl	flush_dcache_L1	/* Flush L1 d-cache */
+	mtlr	r10
+
+	mfspr	r4, SPRN_L1CSR0	/* Invalidate and disable d-cache */
+	li	r5, 2
+	rlwimi	r4, r5, 0, 3
+
+	msync
+	isync
+	mtspr	SPRN_L1CSR0, r4
+	isync
+
+1:	mfspr	r4, SPRN_L1CSR0	/* Wait for the invalidate to finish */
+	andi.	r4, r4, 2
+	bne	1b
+
+	mfspr	r4, SPRN_L1CSR1	/* Invalidate and disable i-cache */
+	li	r5, 2
+	rlwimi	r4, r5, 0, 3
+
+	mtspr	SPRN_L1CSR1, r4
+	isync
+
+	blr
+
 /*
  * Interrupt vector entry code
  *
diff --git a/arch/powerpc/kernel/smp.c b/arch/powerpc/kernel/smp.c
index 68034bb..321cf2e 100644
--- a/arch/powerpc/kernel/smp.c
+++ b/arch/powerpc/kernel/smp.c
@@ -317,6 +317,8 @@ int generic_cpu_disable(void)
 	set_cpu_online(cpu, false);
 #ifdef CONFIG_PPC64
 	vdso_data->processorCount--;
+#endif
+#if defined(CONFIG_PPC64) || defined(CONFIG_E500)
 	fixup_irqs(cpu_online_mask);
 #endif
 	return 0;
@@ -336,7 +338,7 @@ int generic_cpu_enable(unsigned int cpu)
 	while (!cpu_online(cpu))
 		cpu_relax();
 
-#ifdef CONFIG_PPC64
+#if defined(CONFIG_PPC64) || defined(CONFIG_E500)
 	fixup_irqs(cpu_online_mask);
 	/* counter the irq disable in fixup_irqs */
 	local_irq_enable();
diff --git a/arch/powerpc/platforms/85xx/Makefile b/arch/powerpc/platforms/85xx/Makefile
index dd70db7..6bbcf22 100644
--- a/arch/powerpc/platforms/85xx/Makefile
+++ b/arch/powerpc/platforms/85xx/Makefile
@@ -1,7 +1,9 @@
 #
 # Makefile for the PowerPC 85xx linux kernel.
 #
-obj-$(CONFIG_SMP) += smp.o
+obj-$(CONFIG_SMP)         += smp.o
+obj-$(CONFIG_HOTPLUG_CPU) += bootpage.o
+obj-$(CONFIG_SUSPEND)     += suspend-asm.o
 
 obj-$(CONFIG_MPC8540_ADS) += mpc85xx_ads.o
 obj-$(CONFIG_MPC8560_ADS) += mpc85xx_ads.o
diff --git a/arch/powerpc/platforms/85xx/bootpage.S b/arch/powerpc/platforms/85xx/bootpage.S
new file mode 100644
index 0000000..ff0ca10
--- /dev/null
+++ b/arch/powerpc/platforms/85xx/bootpage.S
@@ -0,0 +1,206 @@
+/*
+ * Copyright 2008-2010 Freescale Semiconductor, Inc.
+ * Kumar Gala <kumar.gala@freescale.com>
+ * This file is taken from u-boot
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation;
+ */
+#include <linux/init.h>
+#include <linux/threads.h>
+#include <asm/processor.h>
+#include <asm/page.h>
+#include <asm/mmu.h>
+#include <asm/pgtable.h>
+#include <asm/cputable.h>
+#include <asm/thread_info.h>
+#include <asm/ppc_asm.h>
+#include <asm/asm-offsets.h>
+#include <asm/cache.h>
+
+/* To boot secondary cpus, we need a place for them to start up.
+ * Normally, they start at 0xfffffffc, but that's usually the
+ * firmware, and we don't want to have to run the firmware again.
+ * Instead, the primary cpu will set the BPTR to point here to
+ * this page.  We then set up the core, and head to
+ * start_secondary.  Note that this means that the code below
+ * must never exceed 1023 instructions (the branch at the end
+ * would then be the 1024th).
+ */
+	.globl	__secondary_start_page
+	.align	12
+__secondary_start_page:
+	lis	r3, 0x8000		/* enable machine check */
+#ifndef CONFIG_PPC_E500MC
+	ori	r3,r3,0x4000		/* enable Timebase */
+#endif
+#ifdef CONFIG_PHYS_64BIT
+	/* for 36-bit addressing */
+	ori	r3,r3,0x0080		/* enable MAS7 updates */
+#endif
+	mtspr	SPRN_HID0,r3
+
+#ifndef CONFIG_PPC_E500MC
+	li	r3,0x3000		/* Addr streaming & broadcast */
+	mtspr	SPRN_HID1,r3
+#endif
+
+	/* Enable branch prediction */
+	li	r3,0x201
+	mtspr	SPRN_BUCSR,r3
+
+	/* Ensure TB is 0 */
+	li	r3,0
+	mttbl	r3
+	mttbu	r3
+
+	mfspr	r0,SPRN_L1CSR1
+	ori	r0,r0,0x0003		/* Enable/invalidate the I-Cache */
+	mtspr	SPRN_L1CSR1,r0
+	isync
+
+
+	mfspr	r0,SPRN_L1CSR0
+	ori	r0,r0,0x0003		/* Enable/invalidate the D-Cache */
+	msync
+	isync
+	mtspr	SPRN_L1CSR0,r0
+	isync
+
+#define toreset(x) (x - __secondary_start_page + 0xfffff000)
+
+	/* get our PIR to figure out our table entry */
+	lis	r3,toreset(__spin_table)@h
+	ori	r3,r3,toreset(__spin_table)@l
+
+	/* r10 has the base address for the entry */
+	mfspr	r0,SPRN_PIR
+#ifdef CONFIG_PPC_E500MC
+	rlwinm	r4,r0,27,27,31
+#else
+	mr	r4,r0
+#endif
+	slwi	r8,r4,5
+	add	r10,r3,r8
+
+#define EPAPR_MAGIC		(0x45504150)
+#define ENTRY_ADDR_UPPER	0
+#define ENTRY_ADDR_LOWER	4
+#define ENTRY_R3_UPPER		8
+#define ENTRY_R3_LOWER		12
+#define ENTRY_RESV		16
+#define ENTRY_PIR		20
+#define ENTRY_R6_UPPER		24
+#define ENTRY_R6_LOWER		28
+#define ENTRY_SIZE		32
+
+	/* setup the entry */
+	li	r3,0
+	li	r8,1
+	stw	r0,ENTRY_PIR(r10)
+	stw	r3,ENTRY_ADDR_UPPER(r10)
+	stw	r8,ENTRY_ADDR_LOWER(r10)
+	stw	r3,ENTRY_R3_UPPER(r10)
+	stw	r4,ENTRY_R3_LOWER(r10)
+	stw	r3,ENTRY_R6_UPPER(r10)
+	stw	r3,ENTRY_R6_LOWER(r10)
+
+	/* setup mapping for AS = 1, and jump there */
+	lis	r11,(MAS0_TLBSEL(1)|MAS0_ESEL(1))@h
+	mtspr	SPRN_MAS0,r11
+	lis	r11,(MAS1_VALID|MAS1_IPROT)@h
+	ori	r11,r11,(MAS1_TS|MAS1_TSIZE(BOOK3E_PAGESZ_4K))@l
+	mtspr	SPRN_MAS1,r11
+	lis	r11,(0xfffff000|MAS2_I|MAS2_G)@h
+	ori	r11,r11,(0xfffff000|MAS2_I|MAS2_G)@l
+	mtspr	SPRN_MAS2,r11
+	lis	r11,(0xfffff000|MAS3_SX|MAS3_SW|MAS3_SR)@h
+	ori	r11,r11,(0xfffff000|MAS3_SX|MAS3_SW|MAS3_SR)@l
+	mtspr	SPRN_MAS3,r11
+	tlbwe
+
+	bl	1f
+1:	mflr	r11
+	addi	r11,r11,28
+	mfmsr	r13
+	ori	r12,r13,MSR_IS|MSR_DS@l
+
+	mtspr	SPRN_SRR0,r11
+	mtspr	SPRN_SRR1,r12
+	rfi
+
+	/* spin waiting for addr */
+2:
+	lwz	r4,ENTRY_ADDR_LOWER(r10)
+	andi.	r11,r4,1
+	bne	2b
+	isync
+
+	/* get the upper bits of the addr */
+	lwz	r11,ENTRY_ADDR_UPPER(r10)
+
+	/* setup branch addr */
+	mtspr	SPRN_SRR0,r4
+
+	/* mark the entry as released */
+	li	r8,3
+	stw	r8,ENTRY_ADDR_LOWER(r10)
+
+	/* mask by ~64M to setup our tlb we will jump to */
+	rlwinm	r12,r4,0,0,5
+
+	/* setup r3, r4, r5, r6, r7, r8, r9 */
+	lwz	r3,ENTRY_R3_LOWER(r10)
+	li	r4,0
+	li	r5,0
+	lwz	r6,ENTRY_R6_LOWER(r10)
+	lis	r7,(64*1024*1024)@h
+	li	r8,0
+	li	r9,0
+
+	/* load up the pir */
+	lwz	r0,ENTRY_PIR(r10)
+	mtspr	SPRN_PIR,r0
+	mfspr	r0,SPRN_PIR
+	stw	r0,ENTRY_PIR(r10)
+
+	mtspr	SPRN_IVPR,r12
+/*
+ * Coming here, we know the cpu has one TLB mapping in TLB1[0]
+ * which maps 0xfffff000-0xffffffff one-to-one.  We set up a
+ * second mapping that maps addr 1:1 for 64M, and then we jump to
+ * addr
+ */
+	lis	r10,(MAS0_TLBSEL(1)|MAS0_ESEL(0))@h
+	mtspr	SPRN_MAS0,r10
+	lis	r10,(MAS1_VALID|MAS1_IPROT)@h
+	ori	r10,r10,(MAS1_TSIZE(BOOK3E_PAGESZ_64M))@l
+	mtspr	SPRN_MAS1,r10
+	/* WIMGE = 0b00000 for now */
+	mtspr	SPRN_MAS2,r12
+	ori	r12,r12,(MAS3_SX|MAS3_SW|MAS3_SR)
+	mtspr	SPRN_MAS3,r12
+#ifdef CONFIG_PHYS_64BIT
+	mtspr	SPRN_MAS7,r11
+#endif
+	tlbwe
+
+/* Now we have another mapping for this page, so we jump to that
+ * mapping
+ */
+	mtspr	SPRN_SRR1,r13
+	rfi
+
+	.align L1_CACHE_SHIFT
+	.globl __spin_table
+__spin_table:
+	.space NR_CPUS*ENTRY_SIZE
+
+	/* Fill in the empty space.  The actual reset vector is
+	 * the last word of the page */
+__secondary_start_code_end:
+	.space 4092 - (__secondary_start_code_end - __secondary_start_page)
+
+__secondary_reset_vector:
+	b	__secondary_start_page
diff --git a/arch/powerpc/platforms/85xx/smp.c b/arch/powerpc/platforms/85xx/smp.c
index 1e8aec8..2ef3e8e 100644
--- a/arch/powerpc/platforms/85xx/smp.c
+++ b/arch/powerpc/platforms/85xx/smp.c
@@ -17,6 +17,7 @@
 #include <linux/of.h>
 #include <linux/kexec.h>
 #include <linux/highmem.h>
+#include <linux/cpu.h>
 
 #include <asm/machdep.h>
 #include <asm/pgtable.h>
@@ -28,26 +29,116 @@
 #include <sysdev/fsl_soc.h>
 #include <sysdev/mpic.h>
 
+#define MPC85xx_BPTR_OFF		0x00020
+#define MPC85xx_BPTR_EN			0x80000000
+#define MPC85xx_ECM_EEBPCR_OFF		0x01010
+#define MPC85xx_PIC_PIR_OFF		0x41090
+
+extern void mpc85xx_cpu_down(void) __attribute__((noreturn));
 extern void __early_start(void);
+extern void __secondary_start_page(void);
+extern volatile unsigned long __spin_table;
+
+struct epapr_entry {
+	u32	addr_h;
+	u32	addr_l;
+	u32	r3_h;
+	u32	r3_l;
+	u32	reserved;
+	u32	pir;
+	u32	r6_h;
+	u32	r6_l;
+};
 
-#define BOOT_ENTRY_ADDR_UPPER	0
-#define BOOT_ENTRY_ADDR_LOWER	1
-#define BOOT_ENTRY_R3_UPPER	2
-#define BOOT_ENTRY_R3_LOWER	3
-#define BOOT_ENTRY_RESV		4
-#define BOOT_ENTRY_PIR		5
-#define BOOT_ENTRY_R6_UPPER	6
-#define BOOT_ENTRY_R6_LOWER	7
-#define NUM_BOOT_ENTRY		8
-#define SIZE_BOOT_ENTRY		(NUM_BOOT_ENTRY * sizeof(u32))
+/* access per cpu vars from generic smp.c */
+DECLARE_PER_CPU(int, cpu_state);
 
-static void __init
+#ifdef CONFIG_HOTPLUG_CPU
+static void __cpuinit
+smp_85xx_mach_cpu_die(void)
+{
+	__get_cpu_var(cpu_state) = CPU_DEAD;
+	smp_wmb();
+
+	local_irq_disable();
+	idle_task_exit();
+	mtspr(SPRN_TSR, TSR_ENW | TSR_WIS | TSR_DIS | TSR_FIS);
+	mtspr(SPRN_TCR, 0);
+	mpc85xx_cpu_down();
+}
+
+static void __cpuinit
+smp_85xx_reset_core(int nr)
+{
+	__iomem u32 *ecm_vaddr;
+	__iomem u32 *pic_vaddr;
+	u32 pcr, pir, cpu;
+
+	cpu = (1 << 24) << nr;
+	ecm_vaddr = ioremap(get_immrbase() + MPC85xx_ECM_EEBPCR_OFF, 4);
+	pcr = in_be32(ecm_vaddr);
+	if (pcr & cpu) {
+		pic_vaddr = ioremap(get_immrbase() + MPC85xx_PIC_PIR_OFF, 4);
+		pir = in_be32(pic_vaddr);
+		/* reset assert */
+		pir |= (1 << nr);
+		out_be32(pic_vaddr, pir);
+		pir = in_be32(pic_vaddr);
+		pir &= ~(1 << nr);
+		/* reset negate */
+		out_be32(pic_vaddr, pir);
+		(void)in_be32(pic_vaddr);
+		iounmap(pic_vaddr);
+	} else {
+		out_be32(ecm_vaddr, pcr | cpu);
+		(void)in_be32(ecm_vaddr);
+	}
+	iounmap(ecm_vaddr);
+}
+
+static int __cpuinit
+smp_85xx_map_bootpg(unsigned long pa)
+{
+	__iomem u32 *bootpg_ptr;
+	u32 bptr;
+
+	/* Get the BPTR */
+	bootpg_ptr = ioremap(get_immrbase() + MPC85xx_BPTR_OFF, 4);
+
+	/* Set the BPTR to the secondary boot page */
+	(void)in_be32(bootpg_ptr);
+
+	bptr = (MPC85xx_BPTR_EN | (pa >> 12));
+	out_be32(bootpg_ptr, bptr);
+	(void)in_be32(bootpg_ptr);
+	iounmap(bootpg_ptr);
+	return 0;
+}
+
+static int __cpuinit
+smp_85xx_unmap_bootpg(void)
+{
+	__iomem u32 *bootpg_ptr;
+
+	/* Get the BPTR */
+	bootpg_ptr = ioremap(get_immrbase() + MPC85xx_BPTR_OFF, 4);
+
+	/* Restore the BPTR */
+	if (in_be32(bootpg_ptr) & MPC85xx_BPTR_EN) {
+		out_be32(bootpg_ptr, 0);
+		(void)in_be32(bootpg_ptr);
+	}
+	iounmap(bootpg_ptr);
+	return 0;
+}
+#endif
+
+static void __cpuinit
 smp_85xx_kick_cpu(int nr)
 {
 	unsigned long flags;
-	const u64 *cpu_rel_addr;
-	__iomem u32 *bptr_vaddr;
-	struct device_node *np;
+	phys_addr_t cpu_rel_addr;
+	__iomem struct epapr_entry *epapr;
 	int n = 0;
 	int ioremappable;
 
@@ -55,41 +146,83 @@ smp_85xx_kick_cpu(int nr)
 
 	pr_debug("smp_85xx_kick_cpu: kick CPU #%d\n", nr);
 
-	np = of_get_cpu_node(nr, NULL);
-	cpu_rel_addr = of_get_property(np, "cpu-release-addr", NULL);
+	if (system_state < SYSTEM_RUNNING) {
+		/* booting, using __spin_table from u-boot */
+		struct device_node *np;
+		const u64 *prop;
 
-	if (cpu_rel_addr == NULL) {
-		printk(KERN_ERR "No cpu-release-addr for cpu %d\n", nr);
-		return;
-	}
+		np = of_get_cpu_node(nr, NULL);
+		if (np == NULL)
+			return;
 
-	/*
-	 * A secondary core could be in a spinloop in the bootpage
-	 * (0xfffff000), somewhere in highmem, or somewhere in lowmem.
-	 * The bootpage and highmem can be accessed via ioremap(), but
-	 * we need to directly access the spinloop if its in lowmem.
-	 */
-	ioremappable = *cpu_rel_addr > virt_to_phys(high_memory);
+		prop = of_get_property(np, "cpu-release-addr", NULL);
+		if (prop == NULL) {
+			of_node_put(np);
+			printk(KERN_ERR "No cpu-release-addr for cpu %d\n", nr);
+			return;
+		}
+		cpu_rel_addr = (phys_addr_t)*prop;
+		of_node_put(np);
+
+		/*
+		 * A secondary core could be in a spinloop in the bootpage
+		 * (0xfffff000), somewhere in highmem, or somewhere in lowmem.
+		 * The bootpage and highmem can be accessed via ioremap(), but
+		 * we need to directly access the spinloop if its in lowmem.
+		 */
+		ioremappable = cpu_rel_addr > virt_to_phys(high_memory);
+
+		if (ioremappable)
+			epapr = ioremap(cpu_rel_addr,
+						sizeof(struct epapr_entry));
+		else
+			epapr = phys_to_virt(cpu_rel_addr);
+
+		local_irq_save(flags);
+	} else {
+#ifdef CONFIG_HOTPLUG_CPU
+		/* spin table in kernel, no need to remap */
+		ioremappable = 0;
+		epapr = (void *)&__spin_table + nr * sizeof(struct epapr_entry);
 
-	/* Map the spin table */
-	if (ioremappable)
-		bptr_vaddr = ioremap(*cpu_rel_addr, SIZE_BOOT_ENTRY);
-	else
-		bptr_vaddr = phys_to_virt(*cpu_rel_addr);
+		/* prevent bootpage from being accessed by others */
+		local_irq_save(flags);
+
+		smp_85xx_map_bootpg(__pa(__secondary_start_page));
 
-	local_irq_save(flags);
+		smp_85xx_reset_core(nr);
 
-	out_be32(bptr_vaddr + BOOT_ENTRY_PIR, nr);
+		/* wait until core(nr) is ready... */
+		while ((in_be32(&epapr->addr_l) != 1) && (++n < 1000))
+			udelay(100);
+
+		if (n == 1000) {
+			pr_err("timeout waiting for core%d to reset\n",
+					nr);
+			goto out;
+		}
+#else
+		pr_err("runtime kick cpu not supported\n");
+		return;
+#endif
+	}
+
+	out_be32(&epapr->pir, nr);
 #ifdef CONFIG_PPC32
-	out_be32(bptr_vaddr + BOOT_ENTRY_ADDR_LOWER, __pa(__early_start));
+	/* clear the acknowledge status */
+	__secondary_hold_acknowledge = -1;
+	out_be32(&epapr->addr_l, __pa(__early_start));
 
 	if (!ioremappable)
-		flush_dcache_range((ulong)bptr_vaddr,
-				(ulong)(bptr_vaddr + SIZE_BOOT_ENTRY));
+		flush_dcache_range((ulong)epapr,
+				(ulong)epapr + sizeof(struct epapr_entry));
 
 	/* Wait a bit for the CPU to ack. */
+	n = 0;
 	while ((__secondary_hold_acknowledge != nr) && (++n < 1000))
 		mdelay(1);
+	if (n == 1000)
+		pr_err("timeout waiting for core%d to ack\n", nr);
 #else
 	out_be64((u64 *)(bptr_vaddr + BOOT_ENTRY_ADDR_UPPER),
 		__pa((u64)*((unsigned long long *) generic_secondary_smp_init)));
@@ -97,12 +230,15 @@ smp_85xx_kick_cpu(int nr)
 	smp_generic_kick_cpu(nr);
 #endif
 
+out:
+#ifdef CONFIG_HOTPLUG_CPU
+	if (system_state >= SYSTEM_RUNNING)
+		smp_85xx_unmap_bootpg();
+#endif
 	local_irq_restore(flags);
 
 	if (ioremappable)
-		iounmap(bptr_vaddr);
-
-	pr_debug("waited %d msecs for CPU #%d.\n", n, nr);
+		iounmap(epapr);
 }
 
 static void __init
@@ -232,6 +368,12 @@ void __init mpc85xx_smp_init(void)
 
 	BUG_ON(!smp_85xx_ops.message_pass);
 
+#ifdef CONFIG_HOTPLUG_CPU
+	smp_85xx_ops.cpu_disable   = generic_cpu_disable;
+	smp_85xx_ops.cpu_die	= generic_cpu_die;
+	ppc_md.cpu_die		= smp_85xx_mach_cpu_die;
+#endif
+
 	smp_ops = &smp_85xx_ops;
 
 #ifdef CONFIG_KEXEC
-- 
1.6.6-rc1.GIT

[PATCH 3/7] powerpc/85xx: add the deep sleep support

[Li Yang]

Some Freescale chips like MPC8536 and P1022 has deep sleep PM mode
in addtion to the sleep PM mode.

In sleep power management mode, the clock of e500 core
is turned off, the clocks of most IP blocks are shut off,
only the modules clocks which allows to wake up the
processor are still running.

While in deep sleep PM mode, additionally, the power supply is
removed for e500 core and most devices. Only the blocks needed
to detect wakeup and sequence the chip out of deep sleep are ON.

Also add APIs for setting wakeup source and lossless Ethernet in
low power modes.

The deep sleep mode is equal to the Suspend-to-RAM state of Linux
Power Management.

echo mem > /sys/power/state

to enter deep sleep mode.

Signed-off-by: Dave Liu <daveliu@freescale.com>
Signed-off-by: Li Yang <leoli@freescale.com>
Signed-off-by: Jin Qing <b24347@freescale.com>
Cc: Scott Wood <scottwood@freescale.com>
---
 arch/powerpc/kernel/l2cr_85xx.S           |   53 +++
 arch/powerpc/platforms/85xx/suspend-asm.S |  625 +++++++++++++++++++++++++++++
 arch/powerpc/sysdev/fsl_pmc.c             |  153 ++++++-
 arch/powerpc/sysdev/fsl_soc.h             |   11 +
 4 files changed, 823 insertions(+), 19 deletions(-)
 create mode 100644 arch/powerpc/kernel/l2cr_85xx.S
 create mode 100644 arch/powerpc/platforms/85xx/suspend-asm.S

diff --git a/arch/powerpc/kernel/l2cr_85xx.S b/arch/powerpc/kernel/l2cr_85xx.S
new file mode 100644
index 0000000..5283d31
--- /dev/null
+++ b/arch/powerpc/kernel/l2cr_85xx.S
@@ -0,0 +1,53 @@
+/*
+ * Copyright (C) 2009-2010 Freescale Semiconductor, Inc. All rights reserved.
+ *	Scott Wood <scottwood@freescale.com>
+ *	Dave Liu <daveliu@freescale.com>
+ * implement the L2 cache operations of e500 based L2 controller
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version
+ * 2 of the License, or (at your option) any later version.
+ */
+
+#include <asm/reg.h>
+#include <asm/cputable.h>
+#include <asm/ppc_asm.h>
+#include <asm/asm-offsets.h>
+
+	.section .text
+
+	/* r3 = virtual address of L2 controller, WIMG = 01xx */
+_GLOBAL(flush_disable_L2)
+	/* It's a write-through cache, so only invalidation is needed. */
+	mbar
+	isync
+	lwz	r4, 0(r3)
+	li	r5, 1
+	rlwimi	r4, r5, 30, 0xc0000000
+	stw	r4, 0(r3)
+
+	/* Wait for the invalidate to finish */
+1:	lwz	r4, 0(r3)
+	andis.	r4, r4, 0x4000
+	bne	1b
+	mbar
+
+	blr
+
+	/* r3 = virtual address of L2 controller, WIMG = 01xx */
+_GLOBAL(invalidate_enable_L2)
+	mbar
+	isync
+	lwz	r4, 0(r3)
+	li	r5, 3
+	rlwimi	r4, r5, 30, 0xc0000000
+	stw	r4, 0(r3)
+
+	/* Wait for the invalidate to finish */
+1:	lwz	r4, 0(r3)
+	andis.	r4, r4, 0x4000
+	bne	1b
+	mbar
+
+	blr
diff --git a/arch/powerpc/platforms/85xx/suspend-asm.S b/arch/powerpc/platforms/85xx/suspend-asm.S
new file mode 100644
index 0000000..99de481
--- /dev/null
+++ b/arch/powerpc/platforms/85xx/suspend-asm.S
@@ -0,0 +1,625 @@
+/*
+ * Enter and leave deep sleep/sleep state on MPC85xx
+ *
+ * Author: Scott Wood <scottwood@freescale.com>
+ *
+ * Copyright (C) 2006-2010 Freescale Semiconductor, Inc. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation.
+ */
+
+#include <asm/page.h>
+#include <asm/ppc_asm.h>
+#include <asm/reg.h>
+#include <asm/asm-offsets.h>
+
+#define SS_TB		0x00
+#define SS_HID		0x08 /* 2 HIDs */
+#define SS_IAC		0x10 /* 2 IACs */
+#define SS_DAC		0x18 /* 2 DACs */
+#define SS_DBCR		0x20 /* 3 DBCRs */
+#define SS_PID		0x2c /* 3 PIDs */
+#define SS_SPRG		0x38 /* 8 SPRGs */
+#define SS_IVOR		0x58 /* 20 interrupt vectors */
+#define SS_TCR		0xa8
+#define SS_BUCSR	0xac
+#define SS_L1CSR	0xb0 /* 2 L1CSRs */
+#define SS_MSR		0xb8
+#define SS_USPRG	0xbc
+#define SS_GPREG	0xc0 /* r12-r31 */
+#define SS_LR		0x110
+#define SS_CR		0x114
+#define SS_SP		0x118
+#define SS_CURRENT	0x11c
+#define SS_IVPR		0x120
+#define SS_BPTR		0x124
+
+#define STATE_SAVE_SIZE 0x128
+
+	.section .data
+	.align	5
+mpc85xx_sleep_save_area:
+	.space	STATE_SAVE_SIZE
+ccsrbase:
+	.long	0
+powmgtreq:
+	.long	0
+
+	.section .text
+	.align	12
+
+	/*
+	 * r3 = physical address of CCSR
+	 * r4 = JOG or deep sleep request
+	 *      JOG-0x00200000, deep sleep-0x00100000
+	 */
+_GLOBAL(mpc85xx_enter_deep_sleep)
+	lis	r5, ccsrbase@ha
+	stw	r3, ccsrbase@l(r5)
+
+	lis	r5, powmgtreq@ha
+	stw	r4, powmgtreq@l(r5)
+
+	lis	r10, mpc85xx_sleep_save_area@h
+	ori	r10, r10, mpc85xx_sleep_save_area@l
+
+	mfspr	r5, SPRN_HID0
+	mfspr	r6, SPRN_HID1
+
+	stw	r5, SS_HID+0(r10)
+	stw	r6, SS_HID+4(r10)
+
+	mfspr	r4, SPRN_IAC1
+	mfspr	r5, SPRN_IAC2
+	mfspr	r6, SPRN_DAC1
+	mfspr	r7, SPRN_DAC2
+
+	stw	r4, SS_IAC+0(r10)
+	stw	r5, SS_IAC+4(r10)
+	stw	r6, SS_DAC+0(r10)
+	stw	r7, SS_DAC+4(r10)
+
+	mfspr	r4, SPRN_DBCR0
+	mfspr	r5, SPRN_DBCR1
+	mfspr	r6, SPRN_DBCR2
+
+	stw	r4, SS_DBCR+0(r10)
+	stw	r5, SS_DBCR+4(r10)
+	stw	r6, SS_DBCR+8(r10)
+
+	mfspr	r4, SPRN_PID0
+	mfspr	r5, SPRN_PID1
+	mfspr	r6, SPRN_PID2
+
+	stw	r4, SS_PID+0(r10)
+	stw	r5, SS_PID+4(r10)
+	stw	r6, SS_PID+8(r10)
+
+	mfspr	r4, SPRN_SPRG0
+	mfspr	r5, SPRN_SPRG1
+	mfspr	r6, SPRN_SPRG2
+	mfspr	r7, SPRN_SPRG3
+
+	stw	r4, SS_SPRG+0x00(r10)
+	stw	r5, SS_SPRG+0x04(r10)
+	stw	r6, SS_SPRG+0x08(r10)
+	stw	r7, SS_SPRG+0x0c(r10)
+
+	mfspr	r4, SPRN_SPRG4
+	mfspr	r5, SPRN_SPRG5
+	mfspr	r6, SPRN_SPRG6
+	mfspr	r7, SPRN_SPRG7
+
+	stw	r4, SS_SPRG+0x10(r10)
+	stw	r5, SS_SPRG+0x14(r10)
+	stw	r6, SS_SPRG+0x18(r10)
+	stw	r7, SS_SPRG+0x1c(r10)
+
+	mfspr	r4, SPRN_IVPR
+	stw	r4, SS_IVPR(r10)
+
+	mfspr	r4, SPRN_IVOR0
+	mfspr	r5, SPRN_IVOR1
+	mfspr	r6, SPRN_IVOR2
+	mfspr	r7, SPRN_IVOR3
+
+	stw	r4, SS_IVOR+0x00(r10)
+	stw	r5, SS_IVOR+0x04(r10)
+	stw	r6, SS_IVOR+0x08(r10)
+	stw	r7, SS_IVOR+0x0c(r10)
+
+	mfspr	r4, SPRN_IVOR4
+	mfspr	r5, SPRN_IVOR5
+	mfspr	r6, SPRN_IVOR6
+	mfspr	r7, SPRN_IVOR7
+
+	stw	r4, SS_IVOR+0x10(r10)
+	stw	r5, SS_IVOR+0x14(r10)
+	stw	r6, SS_IVOR+0x18(r10)
+	stw	r7, SS_IVOR+0x1c(r10)
+
+	mfspr	r4, SPRN_IVOR8
+	mfspr	r5, SPRN_IVOR9
+	mfspr	r6, SPRN_IVOR10
+	mfspr	r7, SPRN_IVOR11
+
+	stw	r4, SS_IVOR+0x20(r10)
+	stw	r5, SS_IVOR+0x24(r10)
+	stw	r6, SS_IVOR+0x28(r10)
+	stw	r7, SS_IVOR+0x2c(r10)
+
+	mfspr	r4, SPRN_IVOR12
+	mfspr	r5, SPRN_IVOR13
+	mfspr	r6, SPRN_IVOR14
+	mfspr	r7, SPRN_IVOR15
+
+	stw	r4, SS_IVOR+0x30(r10)
+	stw	r5, SS_IVOR+0x34(r10)
+	stw	r6, SS_IVOR+0x38(r10)
+	stw	r7, SS_IVOR+0x3c(r10)
+
+	mfspr	r4, SPRN_IVOR32
+	mfspr	r5, SPRN_IVOR33
+	mfspr	r6, SPRN_IVOR34
+	mfspr	r7, SPRN_IVOR35
+
+	stw	r4, SS_IVOR+0x40(r10)
+	stw	r5, SS_IVOR+0x44(r10)
+	stw	r6, SS_IVOR+0x48(r10)
+	stw	r7, SS_IVOR+0x4c(r10)
+
+	mfspr	r4, SPRN_TCR
+	mfspr	r5, SPRN_BUCSR
+	mfspr	r6, SPRN_L1CSR0
+	mfspr	r7, SPRN_L1CSR1
+	mfspr	r8, SPRN_USPRG0
+
+	stw	r4, SS_TCR(r10)
+	stw	r5, SS_BUCSR(r10)
+	stw	r6, SS_L1CSR+0(r10)
+	stw	r7, SS_L1CSR+4(r10)
+	stw	r8, SS_USPRG+0(r10)
+
+	stmw	r12, SS_GPREG(r10)
+
+	mfmsr	r4
+	mflr	r5
+	mfcr	r6
+
+	stw	r4, SS_MSR(r10)
+	stw	r5, SS_LR(r10)
+	stw	r6, SS_CR(r10)
+	stw	r1, SS_SP(r10)
+	stw	r2, SS_CURRENT(r10)
+
+1:	mftbu	r4
+	mftb	r5
+	mftbu	r6
+	cmpw	r4, r6
+	bne	1b
+
+	stw	r4, SS_TB+0(r10)
+	stw	r5, SS_TB+4(r10)
+
+	lis	r4, ccsrbase@ha
+	lwz	r3, ccsrbase@l(r4)
+
+	/* Disable machine checks and critical exceptions */
+	mfmsr	r4
+	rlwinm	r4, r4, 0, ~MSR_CE
+	rlwinm	r4, r4, 0, ~MSR_ME
+	mtmsr	r4
+	isync
+
+	/* Use TLB1[15] to map the CCSR at 0xf0000000 */
+	lis	r4, 0x100f
+	mtspr	SPRN_MAS0, r4
+	lis	r4, 0xc000
+	ori	r4, r4, 0x0500
+	mtspr	SPRN_MAS1, r4
+	lis	r4, 0xf000
+	ori	r4, r4, 0x000a
+	mtspr	SPRN_MAS2, r4
+	rlwinm	r4, r3, 0, 0xfffff000
+	ori	r4, r4, 0x0005
+	mtspr	SPRN_MAS3, r4
+	li	r4, 0
+	mtspr	SPRN_MAS7, r4
+	isync
+	tlbwe
+	isync
+
+	lis	r3, 0xf000
+	lwz	r4, 0x20(r3)
+	stw	r4, SS_BPTR(r10)
+
+	lis	r3, 0xf002		/* L2 cache controller at CCSR+0x20000 */
+	bl	flush_disable_L2
+	bl	flush_disable_L1
+
+	/* Enable I-cache, so as not to upset the bus
+	 * with our loop.
+	 */
+
+	mfspr	r4, SPRN_L1CSR1
+	ori	r4, r4, 1
+	mtspr	SPRN_L1CSR1, r4
+	isync
+
+	/* Set boot page translation */
+	lis	r3, 0xf000
+	lis	r4, (mpc85xx_deep_resume - PAGE_OFFSET)@h
+	ori	r4, r4, (mpc85xx_deep_resume - PAGE_OFFSET)@l
+	rlwinm	r4, r4, 20, 0x000fffff
+	oris	r4, r4, 0x8000
+	stw	r4, 0x20(r3)
+	lwz	r4, 0x20(r3)		/* read-back to flush write */
+	twi	0, r4, 0
+	isync
+
+	/* Disable the decrementer */
+	mfspr	r4, SPRN_TCR
+	rlwinm	r4, r4, 0, ~TCR_DIE
+	mtspr	SPRN_TCR, r4
+
+	mfspr	r4, SPRN_TSR
+	oris	r4, r4, TSR_DIS@h
+	mtspr	SPRN_TSR, r4
+
+	/* set PMRCCR[VRCNT] to wait power stable for 40ms */
+	lis	r3, 0xf00e
+	lwz	r4, 0x84(r3)
+	clrlwi	r4, r4, 16
+	oris	r4, r4, 0x12a3
+	stw	r4, 0x84(r3)
+	lwz	r4, 0x84(r3)
+
+	/* set deep sleep bit in POWMGTSCR */
+	lis	r3, powmgtreq@ha
+	lwz	r8, powmgtreq@l(r3)
+
+	lis	r3, 0xf00e
+	mr	r4, r8
+	stw	r4, 0x80(r3)
+	lwz	r4, 0x80(r3)		/* read-back to flush write */
+	twi	0, r4, 0
+	isync
+
+	mftb	r5
+1:	/* spin until either we enter deep sleep, or the sleep process is
+	 * aborted due to a pending wakeup event.  Wait some time between
+	 * accesses, so we don't flood the bus and prevent the pmc from
+	 * detecting an idle system.
+	 */
+
+	mftb	r4
+	subf	r7, r5, r4
+	cmpwi	r7, 1000
+	blt	1b
+	mr	r5, r4
+
+	lwz	r6, 0x80(r3)
+	andis.	r6, r6, 0x0010
+	bne	1b
+	b	2f
+
+2:	mfspr	r4, SPRN_PIR
+	andi.	r4, r4, 1
+99:	bne	99b
+	/* Establish a temporary 64MB 0->0 mapping in TLB1[1]. */
+	lis	r4, 0x1001
+	mtspr	SPRN_MAS0, r4
+	lis	r4, 0xc000
+	ori	r4, r4, 0x0800
+	mtspr	SPRN_MAS1, r4
+	li	r4, 0
+	mtspr	SPRN_MAS2, r4
+	li	r4, 0x0015
+	mtspr	SPRN_MAS3, r4
+	li	r4, 0
+	mtspr	SPRN_MAS7, r4
+	isync
+	tlbwe
+	isync
+
+	lis	r3, (3f - PAGE_OFFSET)@h
+	ori	r3, r3, (3f - PAGE_OFFSET)@l
+	mtctr	r3
+	bctr
+
+	/* Locate the resume vector in the last word of the current page. */
+	. = mpc85xx_enter_deep_sleep + 0xffc
+mpc85xx_deep_resume:
+	b	2b
+
+3:
+	/* Restore the contents of TLB1[0].  It is assumed that it covers
+	 * the currently executing code and the sleep save area, and that
+	 * it does not alias our temporary mapping (which is at virtual zero).
+	 */
+	lis	r3, (TLBCAM - PAGE_OFFSET)@h
+	ori	r3, r3, (TLBCAM - PAGE_OFFSET)@l
+
+	lwz	r4, 0(r3)
+	lwz	r5, 4(r3)
+	lwz	r6, 8(r3)
+	lwz	r7, 12(r3)
+	lwz	r8, 16(r3)
+
+	mtspr	SPRN_MAS0, r4
+	mtspr	SPRN_MAS1, r5
+	mtspr	SPRN_MAS2, r6
+	mtspr	SPRN_MAS3, r7
+	mtspr	SPRN_MAS7, r8
+
+	isync
+	tlbwe
+	isync
+
+	/* Access the ccsrbase address with TLB1[0] */
+	lis	r4, ccsrbase@ha
+	lwz	r3, ccsrbase@l(r4)
+
+	/* Use TLB1[15] to map the CCSR at 0xf0000000 */
+	lis	r4, 0x100f
+	mtspr	SPRN_MAS0, r4
+	lis	r4, 0xc000
+	ori	r4, r4, 0x0500
+	mtspr	SPRN_MAS1, r4
+	lis	r4, 0xf000
+	ori	r4, r4, 0x000a
+	mtspr	SPRN_MAS2, r4
+	rlwinm	r4, r3, 0, 0xfffff000
+	ori	r4, r4, 0x0005
+	mtspr	SPRN_MAS3, r4
+	li	r4, 0
+	mtspr	SPRN_MAS7, r4
+	isync
+	tlbwe
+	isync
+
+	lis	r3, 0xf002		/* L2 cache controller at CCSR+0x20000 */
+	bl	invalidate_enable_L2
+
+	/* Access the MEM(r10) with TLB1[0] */
+	lis	r10, mpc85xx_sleep_save_area@h
+	ori	r10, r10, mpc85xx_sleep_save_area@l
+
+	lis	r3, 0xf000
+	lwz	r4, SS_BPTR(r10)
+	stw	r4, 0x20(r3)		/* restore BPTR */
+
+	/* Program shift running space to PAGE_OFFSET */
+	mfmsr	r3
+	lis	r4, 1f@h
+	ori	r4, r4, 1f@l
+
+	mtsrr1	r3
+	mtsrr0	r4
+	rfi
+
+1:	/* Restore the rest of TLB1, in ascending order so that
+	 * the TLB1[1] gets invalidated first.
+	 *
+	 * XXX: It's better to invalidate the temporary mapping
+	 * TLB1[15] for CCSR before restore any TLB1 entry include 0.
+	 */
+	lis	r4, 0x100f
+	mtspr	SPRN_MAS0, r4
+	lis	r4, 0
+	mtspr	SPRN_MAS1, r4
+	isync
+	tlbwe
+	isync
+
+	lis	r3, (TLBCAM + 5*4 - 4)@h
+	ori	r3, r3, (TLBCAM + 5*4 - 4)@l
+	li	r4, 15
+	mtctr	r4
+
+2:
+	lwz	r5, 4(r3)
+	lwz	r6, 8(r3)
+	lwz	r7, 12(r3)
+	lwz	r8, 16(r3)
+	lwzu	r9, 20(r3)
+
+	mtspr	SPRN_MAS0, r5
+	mtspr	SPRN_MAS1, r6
+	mtspr	SPRN_MAS2, r7
+	mtspr	SPRN_MAS3, r8
+	mtspr	SPRN_MAS7, r9
+
+	isync
+	tlbwe
+	isync
+	bdnz	2b
+
+	lis	r10, mpc85xx_sleep_save_area@h
+	ori	r10, r10, mpc85xx_sleep_save_area@l
+
+	lwz	r5, SS_HID+0(r10)
+	lwz	r6, SS_HID+4(r10)
+
+	isync
+	mtspr	SPRN_HID0, r5
+	isync
+
+	msync
+	mtspr	SPRN_HID1, r6
+	isync
+
+	lwz	r4, SS_IAC+0(r10)
+	lwz	r5, SS_IAC+4(r10)
+	lwz	r6, SS_DAC+0(r10)
+	lwz	r7, SS_DAC+4(r10)
+
+	mtspr	SPRN_IAC1, r4
+	mtspr	SPRN_IAC2, r5
+	mtspr	SPRN_DAC1, r6
+	mtspr	SPRN_DAC2, r7
+
+	lwz	r4, SS_DBCR+0(r10)
+	lwz	r5, SS_DBCR+4(r10)
+	lwz	r6, SS_DBCR+8(r10)
+
+	mtspr	SPRN_DBCR0, r4
+	mtspr	SPRN_DBCR1, r5
+	mtspr	SPRN_DBCR2, r6
+
+	lwz	r4, SS_PID+0(r10)
+	lwz	r5, SS_PID+4(r10)
+	lwz	r6, SS_PID+8(r10)
+
+	mtspr	SPRN_PID0, r4
+	mtspr	SPRN_PID1, r5
+	mtspr	SPRN_PID2, r6
+
+	lwz	r4, SS_SPRG+0x00(r10)
+	lwz	r5, SS_SPRG+0x04(r10)
+	lwz	r6, SS_SPRG+0x08(r10)
+	lwz	r7, SS_SPRG+0x0c(r10)
+
+	mtspr	SPRN_SPRG0, r4
+	mtspr	SPRN_SPRG1, r5
+	mtspr	SPRN_SPRG2, r6
+	mtspr	SPRN_SPRG3, r7
+
+	lwz	r4, SS_SPRG+0x10(r10)
+	lwz	r5, SS_SPRG+0x14(r10)
+	lwz	r6, SS_SPRG+0x18(r10)
+	lwz	r7, SS_SPRG+0x1c(r10)
+
+	mtspr	SPRN_SPRG4, r4
+	mtspr	SPRN_SPRG5, r5
+	mtspr	SPRN_SPRG6, r6
+	mtspr	SPRN_SPRG7, r7
+
+	lwz	r4, SS_IVPR(r10)
+	mtspr	SPRN_IVPR, r4
+
+	lwz	r4, SS_IVOR+0x00(r10)
+	lwz	r5, SS_IVOR+0x04(r10)
+	lwz	r6, SS_IVOR+0x08(r10)
+	lwz	r7, SS_IVOR+0x0c(r10)
+
+	mtspr	SPRN_IVOR0, r4
+	mtspr	SPRN_IVOR1, r5
+	mtspr	SPRN_IVOR2, r6
+	mtspr	SPRN_IVOR3, r7
+
+	lwz	r4, SS_IVOR+0x10(r10)
+	lwz	r5, SS_IVOR+0x14(r10)
+	lwz	r6, SS_IVOR+0x18(r10)
+	lwz	r7, SS_IVOR+0x1c(r10)
+
+	mtspr	SPRN_IVOR4, r4
+	mtspr	SPRN_IVOR5, r5
+	mtspr	SPRN_IVOR6, r6
+	mtspr	SPRN_IVOR7, r7
+
+	lwz	r4, SS_IVOR+0x20(r10)
+	lwz	r5, SS_IVOR+0x24(r10)
+	lwz	r6, SS_IVOR+0x28(r10)
+	lwz	r7, SS_IVOR+0x2c(r10)
+
+	mtspr	SPRN_IVOR8, r4
+	mtspr	SPRN_IVOR9, r5
+	mtspr	SPRN_IVOR10, r6
+	mtspr	SPRN_IVOR11, r7
+
+	lwz	r4, SS_IVOR+0x30(r10)
+	lwz	r5, SS_IVOR+0x34(r10)
+	lwz	r6, SS_IVOR+0x38(r10)
+	lwz	r7, SS_IVOR+0x3c(r10)
+
+	mtspr	SPRN_IVOR12, r4
+	mtspr	SPRN_IVOR13, r5
+	mtspr	SPRN_IVOR14, r6
+	mtspr	SPRN_IVOR15, r7
+
+	lwz	r4, SS_IVOR+0x40(r10)
+	lwz	r5, SS_IVOR+0x44(r10)
+	lwz	r6, SS_IVOR+0x48(r10)
+	lwz	r7, SS_IVOR+0x4c(r10)
+
+	mtspr	SPRN_IVOR32, r4
+	mtspr	SPRN_IVOR33, r5
+	mtspr	SPRN_IVOR34, r6
+	mtspr	SPRN_IVOR35, r7
+
+	lwz	r4, SS_TCR(r10)
+	lwz	r5, SS_BUCSR(r10)
+	lwz	r6, SS_L1CSR+0(r10)
+	lwz	r7, SS_L1CSR+4(r10)
+	lwz	r8, SS_USPRG+0(r10)
+
+	mtspr	SPRN_TCR, r4
+	mtspr	SPRN_BUCSR, r5
+
+	msync
+	isync
+	mtspr	SPRN_L1CSR0, r6
+	isync
+
+	mtspr	SPRN_L1CSR1, r7
+	isync
+
+	mtspr	SPRN_USPRG0, r8
+
+	lmw	r12, SS_GPREG(r10)
+
+	lwz	r1, SS_SP(r10)
+	lwz	r2, SS_CURRENT(r10)
+	lwz	r4, SS_MSR(r10)
+	lwz	r5, SS_LR(r10)
+	lwz	r6, SS_CR(r10)
+
+	msync
+	mtmsr	r4
+	isync
+
+	mtlr	r5
+	mtcr	r6
+
+	li	r4, 0
+	mtspr	SPRN_TBWL, r4
+
+	lwz	r4, SS_TB+0(r10)
+	lwz	r5, SS_TB+4(r10)
+
+	mtspr	SPRN_TBWU, r4
+	mtspr	SPRN_TBWL, r5
+
+	lis	r3, 1
+	mtdec	r3
+
+	blr
+
+
+_GLOBAL(mpc85xx_cpu_down)
+       stwu    r1,-16(r1)
+       mflr    r0
+       stw     r0,20(r1)
+       bl      flush_dcache_L1
+       lwz     r0,20(r1)
+       addi    r1,r1,16
+       mtlr    r0
+       lis     r3,HID0_NAP@h
+
+       /* Go to NAP or DOZE now */
+       mfspr   r4,SPRN_HID0
+       rlwinm  r4,r4,0,~(HID0_DOZE|HID0_NAP|HID0_SLEEP)
+       or      r4,r4,r3
+       isync
+       mtspr   SPRN_HID0,r4
+       isync
+
+       mfmsr   r7
+       oris    r7,r7,MSR_WE@h
+       msync
+       mtmsr   r7
+       isync
+99:    b       99b
diff --git a/arch/powerpc/sysdev/fsl_pmc.c b/arch/powerpc/sysdev/fsl_pmc.c
index 44de855..00014b5 100644
--- a/arch/powerpc/sysdev/fsl_pmc.c
+++ b/arch/powerpc/sysdev/fsl_pmc.c
@@ -2,6 +2,7 @@
  * Suspend/resume support
  *
  * Copyright 2009  MontaVista Software, Inc.
+ * Copyright 2007-2010 Freescale Semiconductor Inc.
  *
  * Author: Anton Vorontsov <avorontsov@ru.mvista.com>
  *
@@ -18,39 +19,137 @@
 #include <linux/delay.h>
 #include <linux/device.h>
 #include <linux/of_platform.h>
+#include <linux/pm.h>
+#include <linux/interrupt.h>
+
+#include <sysdev/fsl_soc.h>
 
 struct pmc_regs {
 	__be32 devdisr;
-	__be32 devdisr2;
+	__be32 :32;
 	__be32 :32;
 	__be32 :32;
 	__be32 pmcsr;
-#define PMCSR_SLP	(1 << 17)
+	__be32 :32;
+	__be32 :32;
+	__be32 pmcdr;
 };
-
 static struct device *pmc_dev;
 static struct pmc_regs __iomem *pmc_regs;
 
+#define PMCSR_SLP	0x00020000
+#define PMCSR_LOSSLESS	0x00400000
+static int has_deep_sleep, has_lossless;
+
+void mpc85xx_enter_deep_sleep(phys_addr_t ccsrbar, u32 powmgtreq);
+
+/**
+ * pmc_enable_wake - enable OF device as wakeup event source
+ * @pdev: platform device affected
+ * @state: PM state from which device will issue wakeup events
+ * @enable: True to enable event generation; false to disable
+ *
+ * This enables the device as a wakeup event source, or disables it.
+ *
+ * RETURN VALUE:
+ * 0 is returned on success
+ * -EINVAL is returned if device is not supposed to wake up the system
+ * Error code depending on the platform is returned if both the platform and
+ * the native mechanism fail to enable the generation of wake-up events
+ */
+int pmc_enable_wake(struct platform_device *pdev, suspend_state_t state, bool enable)
+{
+	int ret = 0;
+	struct device_node *clk_np;
+	u32 *pmcdr_mask;
+
+	if (enable && !device_may_wakeup(&pdev->dev))
+		return -EINVAL;
+
+	clk_np = of_parse_phandle(pdev->dev.of_node, "clk-handle", 0);
+	if (!clk_np)
+		return -EINVAL;
+
+	pmcdr_mask = (u32 *)of_get_property(clk_np, "fsl,pmcdr-mask", NULL);
+	if (!pmcdr_mask) {
+		ret = -EINVAL;
+		goto out;
+	}
+
+	/* clear to enable clock in low power mode */
+	if (enable)
+		clrbits32(&pmc_regs->pmcdr, *pmcdr_mask);
+	else
+		setbits32(&pmc_regs->pmcdr, *pmcdr_mask);
+
+out:
+	of_node_put(clk_np);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(pmc_enable_wake);
+
+/**
+ * pmc_enable_lossless - enable lossless ethernet in low power mode
+ * @enable: True to enable event generation; false to disable
+ */
+void pmc_enable_lossless(int enable)
+{
+	if (enable && has_lossless)
+		setbits32(&pmc_regs->pmcsr, PMCSR_LOSSLESS);
+	else
+		clrbits32(&pmc_regs->pmcsr, PMCSR_LOSSLESS);
+}
+EXPORT_SYMBOL_GPL(pmc_enable_lossless);
+
 static int pmc_suspend_enter(suspend_state_t state)
 {
 	int ret;
+	u32 powmgtreq = 0x00500000;
 
-	setbits32(&pmc_regs->pmcsr, PMCSR_SLP);
-	/* At this point, the CPU is asleep. */
+	switch (state) {
+	case PM_SUSPEND_MEM:
+#ifdef CONFIG_SPE
+		enable_kernel_spe();
+#endif
+		pr_debug("Entering deep sleep\n");
+
+		local_irq_disable();
+		mpc85xx_enter_deep_sleep(get_immrbase(),
+				powmgtreq);
+		pr_debug("Resumed from deep sleep\n");
+
+		return 0;
+
+	/* else fall-through */
+	case PM_SUSPEND_STANDBY:
+		local_irq_disable();
+
+		setbits32(&pmc_regs->pmcsr, PMCSR_SLP);
+
+		/* At this point, the CPU is asleep. */
+		/* Upon resume, wait for SLP bit to be clear. */
+		ret = spin_event_timeout((in_be32(&pmc_regs->pmcsr) & PMCSR_SLP)
+				== 0, 10000, 10) ? 0 : -ETIMEDOUT;
+		if (ret)
+			dev_err(pmc_dev,
+				"timeout waiting for SLP bit to be cleared\n");
+
+		return 0;
+
+	default:
+		return -EINVAL;
+
+	}
 
-	/* Upon resume, wait for SLP bit to be clear. */
-	ret = spin_event_timeout((in_be32(&pmc_regs->pmcsr) & PMCSR_SLP) == 0,
-				 10000, 10) ? 0 : -ETIMEDOUT;
-	if (ret)
-		dev_err(pmc_dev, "tired waiting for SLP bit to clear\n");
-	return ret;
 }
 
 static int pmc_suspend_valid(suspend_state_t state)
 {
-	if (state != PM_SUSPEND_STANDBY)
-		return 0;
-	return 1;
+	if (state == PM_SUSPEND_STANDBY)
+		return 1;
+	if (has_deep_sleep && (state == PM_SUSPEND_MEM))
+		return 1;
+	return 0;
 }
 
 static struct platform_suspend_ops pmc_suspend_ops = {
@@ -58,14 +157,30 @@ static struct platform_suspend_ops pmc_suspend_ops = {
 	.enter = pmc_suspend_enter,
 };
 
-static int pmc_probe(struct platform_device *ofdev,
+static int pmc_probe(struct platform_device *pdev,
 		     const struct of_device_id *id)
 {
-	pmc_regs = of_iomap(ofdev->dev.of_node, 0);
-	if (!pmc_regs)
-		return -ENOMEM;
+	struct device_node *np = pdev->dev.of_node;
+	struct device_node *node;
+
+	node = of_find_compatible_node(NULL, NULL, "fsl,mpc8548-pmc");
+
+	if (node) {
+		pmc_regs = of_iomap(pdev->dev.of_node, 0);
+		if (!pmc_regs)
+			return -ENOMEM;
+
+		if (of_device_is_compatible(np, "fsl,mpc8536-pmc")) {
+			has_deep_sleep = 1;
+		}
+		if (of_device_is_compatible(np, "fsl,p1022-pmc")) {
+			has_lossless = 1;
+		}
+
+		of_node_put(node);
+	}
 
-	pmc_dev = &ofdev->dev;
+	pmc_dev = &pdev->dev;
 	suspend_set_ops(&pmc_suspend_ops);
 	return 0;
 }
diff --git a/arch/powerpc/sysdev/fsl_soc.h b/arch/powerpc/sysdev/fsl_soc.h
index 5360948..5250539 100644
--- a/arch/powerpc/sysdev/fsl_soc.h
+++ b/arch/powerpc/sysdev/fsl_soc.h
@@ -3,6 +3,8 @@
 #ifdef __KERNEL__
 
 #include <asm/mmu.h>
+#include <linux/platform_device.h>
+#include <linux/suspend.h>
 
 struct spi_device;
 
@@ -21,6 +23,15 @@ struct device_node;
 
 extern void fsl_rstcr_restart(char *cmd);
 
+#ifdef CONFIG_FSL_PMC
+int pmc_enable_wake(struct platform_device *pdev, suspend_state_t state,
+		bool enable);
+void pmc_enable_lossless(int enable);
+#else
+#define pmc_enable_wake(pdev, state, enable)	(-EINVAL)
+#define pmc_enable_lossless(enable) do {} while (0);
+#endif
+
 #if defined(CONFIG_FB_FSL_DIU) || defined(CONFIG_FB_FSL_DIU_MODULE)
 struct platform_diu_data_ops {
 	unsigned int (*get_pixel_format) (unsigned int bits_per_pixel,
-- 
1.6.6-rc1.GIT

[PATCH 4/7] powerpc/85xx: add support to JOG feature using cpufreq interface

[Li Yang]

Some 85xx silicons like MPC8536 and P1022 has the JOG PM feature.

The patch adds the support to change CPU frequency using the standard
cpufreq interface.

Signed-off-by: Dave Liu <daveliu@freescale.com>
Signed-off-by: Li Yang <leoli@freescale.com>
---
 arch/powerpc/platforms/85xx/Makefile  |    1 +
 arch/powerpc/platforms/85xx/cpufreq.c |  236 +++++++++++++++++++++++++++++++++
 arch/powerpc/platforms/Kconfig        |    8 +
 3 files changed, 245 insertions(+), 0 deletions(-)
 create mode 100644 arch/powerpc/platforms/85xx/cpufreq.c

diff --git a/arch/powerpc/platforms/85xx/Makefile b/arch/powerpc/platforms/85xx/Makefile
index 6bbcf22..11cedde 100644
--- a/arch/powerpc/platforms/85xx/Makefile
+++ b/arch/powerpc/platforms/85xx/Makefile
@@ -4,6 +4,7 @@
 obj-$(CONFIG_SMP)         += smp.o
 obj-$(CONFIG_HOTPLUG_CPU) += bootpage.o
 obj-$(CONFIG_SUSPEND)     += suspend-asm.o
+obj-$(CONFIG_MPC85xx_CPUFREQ) += cpufreq.o
 
 obj-$(CONFIG_MPC8540_ADS) += mpc85xx_ads.o
 obj-$(CONFIG_MPC8560_ADS) += mpc85xx_ads.o
diff --git a/arch/powerpc/platforms/85xx/cpufreq.c b/arch/powerpc/platforms/85xx/cpufreq.c
new file mode 100644
index 0000000..f729c3d
--- /dev/null
+++ b/arch/powerpc/platforms/85xx/cpufreq.c
@@ -0,0 +1,236 @@
+/*
+ * Copyright (C) 2008-2010 Freescale Semiconductor, Inc.
+ * 	Dave Liu <daveliu@freescale.com>
+ *
+ * The cpufreq driver is for Freescale 85xx processor,
+ * based on arch/powerpc/platforms/cell/cbe_cpufreq.c
+ * (C) Copyright IBM Deutschland Entwicklung GmbH 2005-2007
+ *	Christian Krafft <krafft@de.ibm.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+#include <linux/cpufreq.h>
+#include <linux/of_platform.h>
+
+#include <asm/prom.h>
+#include <asm/time.h>
+#include <asm/reg.h>
+#include <asm/io.h>
+
+#include <sysdev/fsl_soc.h>
+
+static DEFINE_MUTEX(mpc85xx_switch_mutex);
+
+static void __iomem *guts;
+
+#define PORPLLSR	0x0
+#define PMJCR		0x7c
+#define POWMGTCSR	0x80
+
+static struct cpufreq_frequency_table mpc85xx_freqs[] = {
+	{2,	0},
+	{3,	0},
+	{4,	0},
+	{0,	CPUFREQ_TABLE_END},
+};
+
+/*
+ * hardware specific functions
+ */
+static int get_pll(int cpu)
+{
+	int ret, shift;
+	u32 pll = in_be32(guts + PORPLLSR);
+	shift = (cpu == 1) ? 24 : 16;
+	ret = (pll >> shift) & 0x3f;
+
+	return ret;
+}
+
+static void set_pll(unsigned int pll, int cpu)
+{
+	int shift;
+	u32 busfreq, corefreq, val;
+	u32 core_spd, mask, tmp;
+
+	tmp = in_be32(guts + PMJCR);
+	shift = (cpu == 1) ? 24 : 16;
+	busfreq = fsl_get_sys_freq();
+	val = (pll & 0x3f) << shift;
+
+	corefreq = ((busfreq * pll) >> 1);
+	/* must set the bit[18/19] if the requested core freq > 533 MHz */
+	core_spd = (cpu == 1) ? 0x00002000 : 0x00001000;
+	if (corefreq > 533000000)
+		val |= core_spd;
+
+	mask = (cpu == 1) ? 0x3f002000 : 0x003f1000;
+	tmp &= ~mask;
+	tmp |= val;
+	out_be32(guts + PMJCR, tmp);
+	val = in_be32(guts + PMJCR);
+	out_be32(guts + POWMGTCSR, 0x00600000);
+	printk("PMJCR request %08x at CPU %d\n", tmp, cpu);
+}
+
+static void verify_pll(int cpu)
+{
+	int shift;
+	u32 busfreq, pll, corefreq;
+
+	shift = (cpu == 1) ? 24 : 16;
+	busfreq = fsl_get_sys_freq();
+	pll = (in_be32(guts + PORPLLSR) >> shift) & 0x3f;
+
+	corefreq = (busfreq * pll) >> 1;
+	corefreq /= 1000000;
+	printk("PORPLLSR core freq %dMHz at CPU %d\n", corefreq, cpu);
+}
+
+/*
+ * cpufreq functions
+ */
+
+static int mpc85xx_cpufreq_cpu_init(struct cpufreq_policy *policy)
+{
+	u32 busfreq = fsl_get_sys_freq();
+	int i, cur_pll;
+
+	/* we need the freq unit with kHz */
+	busfreq /= 1000;
+
+	/* initialize frequency table */
+	for (i = 0; mpc85xx_freqs[i].frequency != CPUFREQ_TABLE_END; i++) {
+		mpc85xx_freqs[i].frequency = (busfreq * mpc85xx_freqs[i].index) >> 1;
+		printk("%d: %dkHz\n", i, mpc85xx_freqs[i].frequency);
+	}
+
+	/* the latency of a transition, the unit is ns */
+	policy->cpuinfo.transition_latency = 2000;
+
+	cur_pll = get_pll(policy->cpu);
+	pr_debug("current pll is at %d\n", cur_pll);
+
+	for (i = 0; mpc85xx_freqs[i].frequency != CPUFREQ_TABLE_END; i++) {
+		if (mpc85xx_freqs[i].index == cur_pll)
+			policy->cur = mpc85xx_freqs[i].frequency;
+	}
+	pr_debug("current core freq is %d\n", policy->cur);
+
+	cpufreq_frequency_table_get_attr(mpc85xx_freqs, policy->cpu);
+
+	/* this ensures that policy->cpuinfo_min
+	 * and policy->cpuinfo_max are set correctly */
+	return cpufreq_frequency_table_cpuinfo(policy, mpc85xx_freqs);
+}
+
+static int mpc85xx_cpufreq_cpu_exit(struct cpufreq_policy *policy)
+{
+	cpufreq_frequency_table_put_attr(policy->cpu);
+	return 0;
+}
+
+static int mpc85xx_cpufreq_verify(struct cpufreq_policy *policy)
+{
+	return cpufreq_frequency_table_verify(policy, mpc85xx_freqs);
+}
+
+static int mpc85xx_cpufreq_target(struct cpufreq_policy *policy,
+			      unsigned int target_freq,
+			      unsigned int relation)
+{
+	struct cpufreq_freqs freqs;
+	unsigned int new;
+
+	cpufreq_frequency_table_target(policy,
+				       mpc85xx_freqs,
+				       target_freq,
+				       relation,
+				       &new);
+
+	freqs.old = policy->cur;
+	freqs.new = mpc85xx_freqs[new].frequency;
+	freqs.cpu = policy->cpu;
+
+	mutex_lock(&mpc85xx_switch_mutex);
+	cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE);
+
+	printk("setting frequency for cpu %d to %d kHz, " \
+		 "PLL ratio is %d/2\n",
+		 policy->cpu,
+		 mpc85xx_freqs[new].frequency,
+		 mpc85xx_freqs[new].index);
+
+	set_pll(mpc85xx_freqs[new].index, policy->cpu);
+
+	cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE);
+	mutex_unlock(&mpc85xx_switch_mutex);
+
+	ppc_proc_freq = freqs.new * 1000ul;
+
+	verify_pll(policy->cpu);
+
+	return 0;
+}
+
+static struct cpufreq_driver mpc85xx_cpufreq_driver = {
+	.verify		= mpc85xx_cpufreq_verify,
+	.target		= mpc85xx_cpufreq_target,
+	.init		= mpc85xx_cpufreq_cpu_init,
+	.exit		= mpc85xx_cpufreq_cpu_exit,
+	.name		= "mpc85xx-cpufreq",
+	.owner		= THIS_MODULE,
+	.flags		= CPUFREQ_CONST_LOOPS,
+};
+
+/*
+ * module init and destoy
+ */
+
+static struct of_device_id mpc85xx_jog_ids[] __initdata = {
+	{ .compatible = "fsl,mpc8536-guts", },
+	{ .compatible = "fsl,p1022-guts", },
+	{}
+};
+
+static int __init mpc85xx_cpufreq_init(void)
+{
+	struct device_node *np;
+
+	np = of_find_matching_node(NULL, mpc85xx_jog_ids);
+	if (np == NULL)
+		return -ENODEV;
+
+	guts = of_iomap(np, 0);
+	of_node_put(np);
+	if (guts == NULL)
+		return -ENOMEM;
+
+	return cpufreq_register_driver(&mpc85xx_cpufreq_driver);
+}
+
+static void __exit mpc85xx_cpufreq_exit(void)
+{
+	iounmap(guts);
+
+	cpufreq_unregister_driver(&mpc85xx_cpufreq_driver);
+}
+
+module_init(mpc85xx_cpufreq_init);
+module_exit(mpc85xx_cpufreq_exit);
+
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Dave Liu <daveliu@freescale.com>");
diff --git a/arch/powerpc/platforms/Kconfig b/arch/powerpc/platforms/Kconfig
index 956154f..df529f9 100644
--- a/arch/powerpc/platforms/Kconfig
+++ b/arch/powerpc/platforms/Kconfig
@@ -178,6 +178,14 @@ config CPU_FREQ_PMAC64
 	  This adds support for frequency switching on Apple iMac G5,
 	  and some of the more recent desktop G5 machines as well.
 
+config MPC85xx_CPUFREQ
+	bool "Support for Freescale MPC85xx CPU freq"
+	depends on PPC_85xx && PPC32
+	select CPU_FREQ_TABLE
+	help
+	  This adds support for frequency switching on Freescale MPC85xx,
+	  this currently includes P1022 processor.
+
 config PPC_PASEMI_CPUFREQ
 	bool "Support for PA Semi PWRficient"
 	depends on PPC_PASEMI
-- 
1.6.6-rc1.GIT

[PATCH 5/7] powerpc: add the mpic timer support

[Li Yang]

There are global timers in the MPIC interrupt controller.  The patch
adds support to the timers.

The timer can generate interrupt which can be used as a wakeup event.

Signed-off-by: Dave Liu <daveliu@freescale.com>
Signed-off-by: Li Yang <leoli@freescale.com>
---
 arch/powerpc/include/asm/mpic.h  |    1 +
 arch/powerpc/sysdev/Makefile     |    2 +-
 arch/powerpc/sysdev/mpic.c       |   88 +++++++++++++-
 arch/powerpc/sysdev/mpic_timer.c |  258 ++++++++++++++++++++++++++++++++++++++
 4 files changed, 346 insertions(+), 3 deletions(-)
 create mode 100644 arch/powerpc/sysdev/mpic_timer.c

diff --git a/arch/powerpc/include/asm/mpic.h b/arch/powerpc/include/asm/mpic.h
index e000cce..4272111 100644
--- a/arch/powerpc/include/asm/mpic.h
+++ b/arch/powerpc/include/asm/mpic.h
@@ -263,6 +263,7 @@ struct mpic
 #ifdef CONFIG_SMP
 	struct irq_chip		hc_ipi;
 #endif
+	struct irq_chip         hc_tm;
 	const char		*name;
 	/* Flags */
 	unsigned int		flags;
diff --git a/arch/powerpc/sysdev/Makefile b/arch/powerpc/sysdev/Makefile
index 0bef9da..d95a417 100644
--- a/arch/powerpc/sysdev/Makefile
+++ b/arch/powerpc/sysdev/Makefile
@@ -3,7 +3,7 @@ subdir-ccflags-$(CONFIG_PPC_WERROR) := -Werror
 ccflags-$(CONFIG_PPC64)		:= -mno-minimal-toc
 
 mpic-msi-obj-$(CONFIG_PCI_MSI)	+= mpic_msi.o mpic_u3msi.o mpic_pasemi_msi.o
-obj-$(CONFIG_MPIC)		+= mpic.o $(mpic-msi-obj-y)
+obj-$(CONFIG_MPIC)		+= mpic.o mpic_timer.o $(mpic-msi-obj-y)
 fsl-msi-obj-$(CONFIG_PCI_MSI)	+= fsl_msi.o
 obj-$(CONFIG_PPC_MSI_BITMAP)	+= msi_bitmap.o
 
diff --git a/arch/powerpc/sysdev/mpic.c b/arch/powerpc/sysdev/mpic.c
index 7c13426..107549d 100644
--- a/arch/powerpc/sysdev/mpic.c
+++ b/arch/powerpc/sysdev/mpic.c
@@ -6,6 +6,7 @@
  *  with various broken implementations of this HW.
  *
  *  Copyright (C) 2004 Benjamin Herrenschmidt, IBM Corp.
+ *  Copyright (C) 2006, 2008-2010 Freescale Semiconductor Inc.
  *
  *  This file is subject to the terms and conditions of the GNU General Public
  *  License.  See the file COPYING in the main directory of this archive
@@ -36,6 +37,7 @@
 #include <asm/machdep.h>
 #include <asm/mpic.h>
 #include <asm/smp.h>
+#include <asm/prom.h>
 
 #include "mpic.h"
 
@@ -208,6 +210,22 @@ static inline void _mpic_ipi_write(struct mpic *mpic, unsigned int ipi, u32 valu
 	_mpic_write(mpic->reg_type, &mpic->gregs, offset, value);
 }
 
+static inline u32 _mpic_tm_read(struct mpic *mpic, unsigned int tm)
+{
+	unsigned int offset = MPIC_INFO(TIMER_VECTOR_PRI) +
+			      (tm * MPIC_INFO(TIMER_STRIDE));
+
+	return _mpic_read(mpic->reg_type, &mpic->tmregs, offset);
+}
+
+static inline void _mpic_tm_write(struct mpic *mpic, unsigned int tm, u32 value)
+{
+	unsigned int offset = MPIC_INFO(TIMER_VECTOR_PRI) +
+			      (tm * MPIC_INFO(TIMER_STRIDE));
+
+	_mpic_write(mpic->reg_type, &mpic->tmregs, offset, value);
+}
+
 static inline u32 _mpic_cpu_read(struct mpic *mpic, unsigned int reg)
 {
 	unsigned int cpu = 0;
@@ -263,6 +281,8 @@ static inline void _mpic_irq_write(struct mpic *mpic, unsigned int src_no,
 #define mpic_write(b,r,v)	_mpic_write(mpic->reg_type,&(b),(r),(v))
 #define mpic_ipi_read(i)	_mpic_ipi_read(mpic,(i))
 #define mpic_ipi_write(i,v)	_mpic_ipi_write(mpic,(i),(v))
+#define mpic_tm_read(i)		_mpic_tm_read(mpic,(i))
+#define mpic_tm_write(i,v)	_mpic_tm_write(mpic,(i),(v))
 #define mpic_cpu_read(i)	_mpic_cpu_read(mpic,(i))
 #define mpic_cpu_write(i,v)	_mpic_cpu_write(mpic,(i),(v))
 #define mpic_irq_read(s,r)	_mpic_irq_read(mpic,(s),(r))
@@ -622,6 +642,13 @@ static unsigned int mpic_is_ipi(struct mpic *mpic, unsigned int irq)
 	return (src >= mpic->ipi_vecs[0] && src <= mpic->ipi_vecs[3]);
 }
 
+/* Determine if the linux irq is an timer IPI */
+static unsigned int mpic_is_tm(struct mpic *mpic, unsigned int irq)
+{
+	unsigned int src = mpic_irq_to_hw(irq);
+
+	return (src >= mpic->timer_vecs[0] && src <= mpic->timer_vecs[3]);
+}
 
 /* Convert a cpu mask from logical to physical cpu numbers. */
 static inline u32 mpic_physmask(u32 cpumask)
@@ -642,6 +669,12 @@ static inline struct mpic * mpic_from_ipi(unsigned int ipi)
 }
 #endif
 
+/* Get the mpic structure from the tm number */
+static inline struct mpic * mpic_from_tm(unsigned int tm)
+{
+	return irq_to_desc(tm)->chip_data;
+}
+
 /* Get the mpic structure from the irq number */
 static inline struct mpic * mpic_from_irq(unsigned int irq)
 {
@@ -800,6 +833,32 @@ static void mpic_end_ipi(unsigned int irq)
 
 #endif /* CONFIG_SMP */
 
+static void mpic_unmask_tm(unsigned int irq)
+{
+	struct mpic *mpic = mpic_from_tm(irq);
+	unsigned int src = mpic_irq_to_hw(irq) - mpic->timer_vecs[0];
+
+	DBG("%s: enable_tm: %d (tm %d)\n", mpic->name, irq, src);
+	mpic_tm_write(src, mpic_tm_read(src) & ~MPIC_VECPRI_MASK);
+	mpic_tm_read(src);
+}
+
+static void mpic_mask_tm(unsigned int irq)
+{
+	struct mpic *mpic = mpic_from_tm(irq);
+	unsigned int src = mpic_irq_to_hw(irq) - mpic->timer_vecs[0];
+
+	mpic_tm_write(src, mpic_tm_read(src) | MPIC_VECPRI_MASK);
+	mpic_tm_read(src);
+}
+
+static void mpic_end_tm(unsigned int irq)
+{
+	struct mpic *mpic = mpic_from_tm(irq);
+
+	mpic_eoi(mpic);
+}
+
 int mpic_set_affinity(unsigned int irq, const struct cpumask *cpumask)
 {
 	struct mpic *mpic = mpic_from_irq(irq);
@@ -919,6 +978,12 @@ static struct irq_chip mpic_ipi_chip = {
 };
 #endif /* CONFIG_SMP */
 
+static struct irq_chip mpic_tm_chip = {
+	.mask		= mpic_mask_tm,
+	.unmask		= mpic_unmask_tm,
+	.eoi		= mpic_end_tm,
+};
+
 #ifdef CONFIG_MPIC_U3_HT_IRQS
 static struct irq_chip mpic_irq_ht_chip = {
 	.startup	= mpic_startup_ht_irq,
@@ -950,6 +1015,15 @@ static int mpic_host_map(struct irq_host *h, unsigned int virq,
 	if (mpic->protected && test_bit(hw, mpic->protected))
 		return -EINVAL;
 
+	else if (hw >= mpic->timer_vecs[0] && hw <= mpic->timer_vecs[3]) {
+		WARN_ON(!(mpic->flags & MPIC_PRIMARY));
+
+		DBG("mpic: mapping as timer\n");
+		set_irq_chip_data(virq, mpic);
+		set_irq_chip_and_handler(virq, &mpic->hc_tm,
+					 handle_fasteoi_irq);
+		return 0;
+	}
 #ifdef CONFIG_SMP
 	else if (hw >= mpic->ipi_vecs[0]) {
 		WARN_ON(!(mpic->flags & MPIC_PRIMARY));
@@ -1071,6 +1145,9 @@ struct mpic * __init mpic_alloc(struct device_node *node,
 	mpic->hc_ipi.name = name;
 #endif /* CONFIG_SMP */
 
+	mpic->hc_tm = mpic_tm_chip;
+	mpic->hc_tm.name = name;
+
 	mpic->flags = flags;
 	mpic->isu_size = isu_size;
 	mpic->irq_count = irq_count;
@@ -1277,15 +1354,17 @@ void __init mpic_init(struct mpic *mpic)
 	/* Set current processor priority to max */
 	mpic_cpu_write(MPIC_INFO(CPU_CURRENT_TASK_PRI), 0xf);
 
-	/* Initialize timers: just disable them all */
+	/* Initialize timers to our reserved vectors and mask them for now */
 	for (i = 0; i < 4; i++) {
 		mpic_write(mpic->tmregs,
 			   i * MPIC_INFO(TIMER_STRIDE) +
-			   MPIC_INFO(TIMER_DESTINATION), 0);
+			   MPIC_INFO(TIMER_DESTINATION),
+			   1 << hard_smp_processor_id());
 		mpic_write(mpic->tmregs,
 			   i * MPIC_INFO(TIMER_STRIDE) +
 			   MPIC_INFO(TIMER_VECTOR_PRI),
 			   MPIC_VECPRI_MASK |
+			   (9 << MPIC_VECPRI_PRIORITY_SHIFT) |
 			   (mpic->timer_vecs[0] + i));
 	}
 
@@ -1395,6 +1474,11 @@ void mpic_irq_set_priority(unsigned int irq, unsigned int pri)
 			~MPIC_VECPRI_PRIORITY_MASK;
 		mpic_ipi_write(src - mpic->ipi_vecs[0],
 			       reg | (pri << MPIC_VECPRI_PRIORITY_SHIFT));
+	} else if (mpic_is_tm(mpic, irq)) {
+		reg = mpic_tm_read(src - mpic->timer_vecs[0]) &
+			~MPIC_VECPRI_PRIORITY_MASK;
+		mpic_tm_write(src - mpic->timer_vecs[0],
+			      reg | (pri << MPIC_VECPRI_PRIORITY_SHIFT));
 	} else {
 		reg = mpic_irq_read(src, MPIC_INFO(IRQ_VECTOR_PRI))
 			& ~MPIC_VECPRI_PRIORITY_MASK;
diff --git a/arch/powerpc/sysdev/mpic_timer.c b/arch/powerpc/sysdev/mpic_timer.c
new file mode 100644
index 0000000..cdc2438
--- /dev/null
+++ b/arch/powerpc/sysdev/mpic_timer.c
@@ -0,0 +1,258 @@
+/*
+ * Copyright (c) 2008-2010 Freescale Semiconductor, Inc. All rights reserved.
+ * Dave Liu <daveliu@freescale.com>
+ * copy from the 83xx GTM driver and modify for MPIC global timer,
+ * implement the global timer 0 function.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation.
+ */
+
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/sched.h>
+#include <linux/errno.h>
+#include <linux/module.h>
+#include <linux/mm.h>
+#include <linux/delay.h>
+#include <linux/fs.h>
+#include <linux/string.h>
+#include <linux/interrupt.h>
+#include <linux/sysfs.h>
+#include <linux/slab.h>
+#include <linux/of_platform.h>
+
+#include <linux/io.h>
+#include <linux/irq.h>
+
+#include <sysdev/fsl_soc.h>
+
+#define MPIC_TIMER_TCR_OFFSET		0x200
+#define MPIC_TIMER_TCR_CLKDIV_64	0x00000300
+#define MPIC_TIMER_STOP			0x80000000
+
+struct mpic_tm_regs {
+	u32	gtccr;
+	u32	res0[3];
+	u32	gtbcr;
+	u32	res1[3];
+	u32	gtvpr;
+	u32	res2[3];
+	u32	gtdr;
+	u32	res3[3];
+};
+
+struct mpic_tm_priv {
+	struct mpic_tm_regs __iomem *regs;
+	int irq;
+	int ticks_per_sec;
+	spinlock_t lock;
+};
+
+struct mpic_type {
+	int has_tcr;
+};
+
+static irqreturn_t mpic_tm_isr(int irq, void *dev_id)
+{
+	struct mpic_tm_priv *priv = dev_id;
+	unsigned long flags;
+	unsigned long temp;
+
+	spin_lock_irqsave(&priv->lock, flags);
+	temp = in_be32(&priv->regs->gtbcr);
+	temp |= MPIC_TIMER_STOP; /* counting inhibited */
+	out_be32(&priv->regs->gtbcr, temp);
+	spin_unlock_irqrestore(&priv->lock, flags);
+
+	return IRQ_HANDLED;
+}
+
+static ssize_t mpic_tm_timeout_store(struct device *dev,
+				struct device_attribute *attr,
+				const char *buf, size_t count)
+{
+	struct mpic_tm_priv *priv = dev_get_drvdata(dev);
+	unsigned long interval = simple_strtoul(buf, NULL, 0);
+	unsigned long temp;
+
+	if (interval > 0x7fffffff) {
+		dev_dbg(dev, "mpic_tm: interval %lu (in s) too long\n", interval);
+		return -EINVAL;
+	}
+
+	temp = interval;
+	interval *= priv->ticks_per_sec;
+
+	if (interval > 0x7fffffff || (interval / priv->ticks_per_sec) != temp) {
+		dev_dbg(dev, "mpic_tm: interval %lu (in ticks) too long\n",
+			interval);
+		return -EINVAL;
+	}
+
+	spin_lock_irq(&priv->lock);
+
+	/* stop timer 0 */
+	temp = in_be32(&priv->regs->gtbcr);
+	temp |= MPIC_TIMER_STOP; /* counting inhibited */
+	out_be32(&priv->regs->gtbcr, temp);
+
+	if (interval != 0) {
+		/* start timer */
+		out_be32(&priv->regs->gtbcr, interval | MPIC_TIMER_STOP);
+		out_be32(&priv->regs->gtbcr, interval);
+	}
+
+	spin_unlock_irq(&priv->lock);
+	return count;
+}
+
+static ssize_t mpic_tm_timeout_show(struct device *dev,
+				struct device_attribute *attr,
+				char *buf)
+{
+	struct mpic_tm_priv *priv = dev_get_drvdata(dev);
+	int timeout = 0;
+
+	spin_lock_irq(&priv->lock);
+
+	if (!(in_be32(&priv->regs->gtbcr) & MPIC_TIMER_STOP)) {
+		timeout = in_be32(&priv->regs->gtccr);
+		timeout += priv->ticks_per_sec - 1;
+		timeout /= priv->ticks_per_sec;
+	}
+
+	spin_unlock_irq(&priv->lock);
+	return sprintf(buf, "%u\n", timeout);
+}
+
+static DEVICE_ATTR(timeout, 0660, mpic_tm_timeout_show, mpic_tm_timeout_store);
+
+static int __devinit mpic_tm_probe(struct platform_device *dev,
+				const struct of_device_id *match)
+{
+	struct device_node *np = dev->dev.of_node;
+	struct resource res;
+	struct mpic_tm_priv *priv;
+	struct mpic_type *type = match->data;
+	int has_tcr = type->has_tcr;
+	u32 busfreq = fsl_get_sys_freq();
+	int ret = 0;
+
+	if (busfreq == 0) {
+		dev_err(&dev->dev, "mpic_tm: No bus frequency in device tree.\n");
+		return -ENODEV;
+	}
+
+	priv = kmalloc(sizeof(struct mpic_tm_priv), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	spin_lock_init(&priv->lock);
+	dev_set_drvdata(&dev->dev, priv);
+
+	ret = of_address_to_resource(np, 0, &res);
+	if (ret)
+		goto out;
+
+	priv->irq = irq_of_parse_and_map(np, 0);
+	if (priv->irq == NO_IRQ) {
+		dev_err(&dev->dev, "MPIC global timer0 exists in device tree "
+				"without an IRQ.\n");
+		ret = -ENODEV;
+		goto out;
+	}
+
+	ret = request_irq(priv->irq, mpic_tm_isr, 0, "mpic timer 0", priv);
+	if (ret)
+		goto out;
+
+	priv->regs = ioremap(res.start, res.end - res.start + 1);
+	if (!priv->regs) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	/*
+	 * MPIC implementation from Freescale has the TCR register,
+	 * the MPIC_TIMER_TCR_OFFSET is 0x200 from global timer base
+	 * the default clock source to the MPIC timer 0 is CCB freq / 8.
+	 * to extend the timer period, we divide the timer clock source
+	 * as CCB freq / 64, so the max timer period is 336 seconds
+	 * when the CCB frequence is 400MHz.
+	 */
+	if (!has_tcr) {
+		priv->ticks_per_sec = busfreq / 8;
+	} else {
+		u32 __iomem *tcr;
+		tcr = (u32 __iomem *)((u32)priv->regs + MPIC_TIMER_TCR_OFFSET);
+		out_be32(tcr, in_be32(tcr) | MPIC_TIMER_TCR_CLKDIV_64);
+		priv->ticks_per_sec = busfreq / 64;
+	}
+
+	ret = device_create_file(&dev->dev, &dev_attr_timeout);
+	if (ret)
+		goto out;
+
+	printk("MPIC global timer init done.\n");
+
+	return 0;
+
+out:
+	kfree(priv);
+	return ret;
+}
+
+static int __devexit mpic_tm_remove(struct platform_device *dev)
+{
+	struct mpic_tm_priv *priv = dev_get_drvdata(&dev->dev);
+
+	device_remove_file(&dev->dev, &dev_attr_timeout);
+	free_irq(priv->irq, priv);
+	iounmap(priv->regs);
+
+	dev_set_drvdata(&dev->dev, NULL);
+	kfree(priv);
+	return 0;
+}
+
+static struct mpic_type mpic_types[] = {
+	{
+		.has_tcr = 0,
+	},
+	{
+		.has_tcr = 1,
+	}
+};
+
+static struct of_device_id mpic_tm_match[] = {
+	{
+		.compatible = "fsl,mpic-global-timer",
+		.data = &mpic_types[1],
+	},
+	{},
+};
+
+static struct of_platform_driver mpic_tm_driver = {
+	.driver = {
+		.name = "mpic-global-timer",
+		.owner = THIS_MODULE,
+		.of_match_table = mpic_tm_match,
+	},
+	.probe = mpic_tm_probe,
+	.remove = __devexit_p(mpic_tm_remove)
+};
+
+static int __init mpic_tm_init(void)
+{
+	return of_register_platform_driver(&mpic_tm_driver);
+}
+
+static void __exit mpic_tm_exit(void)
+{
+	of_unregister_platform_driver(&mpic_tm_driver);
+}
+
+module_init(mpic_tm_init);
+module_exit(mpic_tm_exit);
-- 
1.6.6-rc1.GIT

[PATCH 6/7] fsl_pmc: update device bindings

[Li Yang]

Signed-off-by: Li Yang <leoli@freescale.com>
---
 Documentation/powerpc/dts-bindings/fsl/pmc.txt |   63 +++++++++++++----------
 1 files changed, 36 insertions(+), 27 deletions(-)

diff --git a/Documentation/powerpc/dts-bindings/fsl/pmc.txt b/Documentation/powerpc/dts-bindings/fsl/pmc.txt
index 07256b7..d84b4f8 100644
--- a/Documentation/powerpc/dts-bindings/fsl/pmc.txt
+++ b/Documentation/powerpc/dts-bindings/fsl/pmc.txt
@@ -9,22 +9,27 @@ Properties:
 
   "fsl,mpc8548-pmc" should be listed for any chip whose PMC is
   compatible.  "fsl,mpc8536-pmc" should also be listed for any chip
-  whose PMC is compatible, and implies deep-sleep capability.
+  whose PMC is compatible, and implies deep-sleep capability and
+  wake on user defined packet(wakeup on ARP).
+
+  "fsl,p1022-pmc" should be listed for any chip whose PMC is
+  compatible, and implies lossless Ethernet capability during sleep.
 
   "fsl,mpc8641d-pmc" should be listed for any chip whose PMC is
   compatible; all statements below that apply to "fsl,mpc8548-pmc" also
   apply to "fsl,mpc8641d-pmc".
 
   Compatibility does not include bit assignments in SCCR/PMCDR/DEVDISR; these
-  bit assignments are indicated via the sleep specifier in each device's
-  sleep property.
+  bit assignments are indicated via the clock nodes.  Device which has a
+  controllable clock source should have a "clk-handle" property pointing
+  to the clock node.
 
 - reg: For devices compatible with "fsl,mpc8349-pmc", the first resource
   is the PMC block, and the second resource is the Clock Configuration
   block.
 
-  For devices compatible with "fsl,mpc8548-pmc", the first resource
-  is a 32-byte block beginning with DEVDISR.
+  For devices compatible with "fsl,mpc8548-pmc", the second resource
+  is a 32-byte block beginning with DEVDISR if supported.
 
 - interrupts: For "fsl,mpc8349-pmc"-compatible devices, the first
   resource is the PMC block interrupt.
@@ -33,31 +38,35 @@ Properties:
   this is a phandle to an "fsl,gtm" node on which timer 4 can be used as
   a wakeup source from deep sleep.
 
-Sleep specifiers:
+Clock nodes:
+The clock nodes are to describe the masks in PM controller registers for each
+soc clock.
+- fsl,pmcdr-mask: For "fsl,mpc8548-pmc"-compatible devices, the mask will be
+  ORed into PMCDR before suspend if the device using this clock is the wake-up
+  source and need to be running during low power mode; clear the mask if
+  otherwise.
 
-  fsl,mpc8349-pmc: Sleep specifiers consist of one cell.  For each bit
-  that is set in the cell, the corresponding bit in SCCR will be saved
-  and cleared on suspend, and restored on resume.  This sleep controller
-  supports disabling and resuming devices at any time.
+- fsl,sccr-mask: For "fsl,mpc8349-pmc"-compatible devices, the corresponding
+  bit specified by the mask in SCCR will be saved and cleared on suspend, and
+  restored on resume.
 
-  fsl,mpc8536-pmc: Sleep specifiers consist of three cells, the third of
-  which will be ORed into PMCDR upon suspend, and cleared from PMCDR
-  upon resume.  The first two cells are as described for fsl,mpc8578-pmc.
-  This sleep controller only supports disabling devices during system
-  sleep, or permanently.
-
-  fsl,mpc8548-pmc: Sleep specifiers consist of one or two cells, the
-  first of which will be ORed into DEVDISR (and the second into
-  DEVDISR2, if present -- this cell should be zero or absent if the
-  hardware does not have DEVDISR2) upon a request for permanent device
-  disabling.  This sleep controller does not support configuring devices
-  to disable during system sleep (unless supported by another compatible
-  match), or dynamically.
+- fsl,devdisr-mask: Contain one or two cells, depending on the availability of
+  DEVDISR2 register.  For compatible devices, the mask will be ORed into DEVDISR
+  or DEVDISR2 when the clock should be permenently disabled.
 
 Example:
 
-	power@b00 {
-		compatible = "fsl,mpc8313-pmc", "fsl,mpc8349-pmc";
-		reg = <0xb00 0x100 0xa00 0x100>;
-		interrupts = <80 8>;
+	power@e0070 {
+		compatible = "fsl,mpc8536-pmc", "fsl,mpc8548-pmc";
+		reg = <0xe0070 0x20>;
+
+		etsec1_clk: soc-clk@24 {
+			fsl,pmcdr-mask = <0x00000080>;
+		};
+		etsec2_clk: soc-clk@25 {
+			fsl,pmcdr-mask = <0x00000040>;
+		};
+		etsec3_clk: soc-clk@26 {
+			fsl,pmcdr-mask = <0x00000020>;
+		};
 	};
-- 
1.6.6-rc1.GIT

[PATCH 7/7] P2020ds: add event button handler

[Li Yang]

This can be used as a wakeup source for power management.

Signed-off-by: Li Yang <leoli@freescale.com>
---
 arch/powerpc/boot/dts/p2020ds.dts        |    9 ++++++++-
 arch/powerpc/platforms/85xx/mpc85xx_ds.c |   26 +++++++++++++++++++++++++-
 2 files changed, 33 insertions(+), 2 deletions(-)

diff --git a/arch/powerpc/boot/dts/p2020ds.dts b/arch/powerpc/boot/dts/p2020ds.dts
index 1101914..f4c6520 100644
--- a/arch/powerpc/boot/dts/p2020ds.dts
+++ b/arch/powerpc/boot/dts/p2020ds.dts
@@ -1,7 +1,7 @@
 /*
  * P2020 DS Device Tree Source
  *
- * Copyright 2009 Freescale Semiconductor Inc.
+ * Copyright 2009-2010 Freescale Semiconductor Inc.
  *
  * This program is free software; you can redistribute  it and/or modify it
  * under  the terms of  the GNU General  Public License as published by the
@@ -155,6 +155,13 @@
 			compatible = "fsl,elbc-fcm-nand";
 			reg = <0x6 0x0 0x40000>;
 		};
+
+		ngpixis@3,0 {
+			compatible = "fsl,p2020ds-fpga";
+			reg = <0x3 0 0x30>;
+			interrupt-parent = <&mpic>;
+			interrupts = <0 0>;
+		};
 	};
 
 	soc@ffe00000 {
diff --git a/arch/powerpc/platforms/85xx/mpc85xx_ds.c b/arch/powerpc/platforms/85xx/mpc85xx_ds.c
index 8190bc2..a8807fe 100644
--- a/arch/powerpc/platforms/85xx/mpc85xx_ds.c
+++ b/arch/powerpc/platforms/85xx/mpc85xx_ds.c
@@ -4,7 +4,7 @@
  * Author Xianghua Xiao (x.xiao@freescale.com)
  * Roy Zang <tie-fei.zang@freescale.com>
  * 	- Add PCI/PCI Exprees support
- * Copyright 2007 Freescale Semiconductor Inc.
+ * Copyright 2007-2010 Freescale Semiconductor Inc.
  *
  * This program is free software; you can redistribute  it and/or modify it
  * under  the terms of  the GNU General  Public License as published by the
@@ -200,6 +200,30 @@ static void __init mpc85xx_ds_setup_arch(void)
 	printk("MPC85xx DS board from Freescale Semiconductor\n");
 }
 
+static irqreturn_t event_isr(int irq, void *dev_id)
+{
+
+	printk(KERN_INFO "MPC85xxDS: Event button been pushed.\n");
+	return IRQ_HANDLED;
+}
+
+static int __init p2020ds_ngpixis_init(void)
+{
+	int event_irq, ret;
+	struct device_node *np;
+
+	np = of_find_compatible_node(NULL, NULL, "fsl,p2020ds-fpga");
+	if (np) {
+		event_irq = irq_of_parse_and_map(np, 0);
+		ret = request_irq(event_irq, event_isr, 0, "event", NULL);
+		if (ret)
+			printk(KERN_ERR "Can't request board event int\n");
+		of_node_put(np);
+	}
+	return 0;
+}
+machine_device_initcall(p2020_ds, p2020ds_ngpixis_init);
+
 /*
  * Called very early, device-tree isn't unflattened
  */
-- 
1.6.6-rc1.GIT

Re: [PATCH 7/7] P2020ds: add event button handler

[Timur Tabi]

On Fri, Dec 3, 2010 at 6:34 AM, Li Yang <leoli@freescale.com> wrote:
> This can be used as a wakeup source for power management.

This patch doesn't actually add wake-up support.

This patch should probably be split up, since you're adding generic
functionality for the IRQ that applies to all 85xx boards, but you
only update the device tree for one board.

> +static irqreturn_t event_isr(int irq, void *dev_id)
> +{
> +
> + =A0 =A0 =A0 printk(KERN_INFO "MPC85xxDS: Event button been pushed.\n");
> + =A0 =A0 =A0 return IRQ_HANDLED;
> +}

Would it make sense to have this be a weak function, so that it would
be easier to implement board-specific support?

> +
> +static int __init p2020ds_ngpixis_init(void)

You're adding a function called "p2020ds_ngpixis_init" to the file
"mpc85xx_ds.c".  mpc85xx_ds.c supports more than just the P2020DS.

> +{
> + =A0 =A0 =A0 int event_irq, ret;
> + =A0 =A0 =A0 struct device_node *np;
> +
> + =A0 =A0 =A0 np =3D of_find_compatible_node(NULL, NULL, "fsl,p2020ds-fpg=
a");
> + =A0 =A0 =A0 if (np) {
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 event_irq =3D irq_of_parse_and_map(np, 0);
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 ret =3D request_irq(event_irq, event_isr, 0=
, "event", NULL);

You should probably choose a less generic name than "event".

> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 if (ret)
> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 printk(KERN_ERR "Can't requ=
est board event int\n");

Use pr_err()

--=20
Timur Tabi
Linux kernel developer at Freescale

Re: [PATCH 7/7] P2020ds: add event button handler

[Li Yang]

On Tue, Dec 14, 2010 at 5:56 AM, Timur Tabi <timur.tabi@gmail.com> wrote:
> On Fri, Dec 3, 2010 at 6:34 AM, Li Yang <leoli@freescale.com> wrote:
>> This can be used as a wakeup source for power management.
>
> This patch doesn't actually add wake-up support.

Any enabled IRQ is a valid wake-up source for standby.  The patch
enables a board specific interrupt for the purpose of wakeup.

>
> This patch should probably be split up, since you're adding generic
> functionality for the IRQ that applies to all 85xx boards, but you
> only update the device tree for one board.

The IRQ is a board specific one from GPIO which not applicable on all
85xx boards.

>
>> +static irqreturn_t event_isr(int irq, void *dev_id)
>> +{
>> +
>> + =C2=A0 =C2=A0 =C2=A0 printk(KERN_INFO "MPC85xxDS: Event button been pu=
shed.\n");
>> + =C2=A0 =C2=A0 =C2=A0 return IRQ_HANDLED;
>> +}
>
> Would it make sense to have this be a weak function, so that it would
> be easier to implement board-specific support?

It's already a board-specific one.

>
>> +
>> +static int __init p2020ds_ngpixis_init(void)
>
> You're adding a function called "p2020ds_ngpixis_init" to the file
> "mpc85xx_ds.c". =C2=A0mpc85xx_ds.c supports more than just the P2020DS.

I'm not sure if other DS boards covered by this file has the same functiona=
lity.

>
>> +{
>> + =C2=A0 =C2=A0 =C2=A0 int event_irq, ret;
>> + =C2=A0 =C2=A0 =C2=A0 struct device_node *np;
>> +
>> + =C2=A0 =C2=A0 =C2=A0 np =3D of_find_compatible_node(NULL, NULL, "fsl,p=
2020ds-fpga");
>> + =C2=A0 =C2=A0 =C2=A0 if (np) {
>> + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 event_irq =3D irq_of_=
parse_and_map(np, 0);
>> + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 ret =3D request_irq(e=
vent_irq, event_isr, 0, "event", NULL);
>
> You should probably choose a less generic name than "event".

Well, it's the name suggested by the board manual.  We may change it
to "event_button" if not too long.

- Leo

Re: [PATCH 7/7] P2020ds: add event button handler

[Timur Tabi]

On Mon, Dec 13, 2010 at 10:24 PM, Li Yang <leoli@freescale.com> wrote:

> Any enabled IRQ is a valid wake-up source for standby. =A0The patch
> enables a board specific interrupt for the purpose of wakeup.

Oh, right.  I should have realized that.

>> This patch should probably be split up, since you're adding generic
>> functionality for the IRQ that applies to all 85xx boards, but you
>> only update the device tree for one board.
>
> The IRQ is a board specific one from GPIO which not applicable on all
> 85xx boards.

Doesn't the device tree take care of that?  If the node exists, the
IRQ number is specified.  Otherwise, the feature is not enabled.

>>> +static irqreturn_t event_isr(int irq, void *dev_id)
>>> +{
>>> +
>>> + =A0 =A0 =A0 printk(KERN_INFO "MPC85xxDS: Event button been pushed.\n"=
);
>>> + =A0 =A0 =A0 return IRQ_HANDLED;
>>> +}
>>
>> Would it make sense to have this be a weak function, so that it would
>> be easier to implement board-specific support?
>
> It's already a board-specific one.

Hmmm.... I guess technically it is, but I wonder if it should be.

>>> +
>>> +static int __init p2020ds_ngpixis_init(void)
>>
>> You're adding a function called "p2020ds_ngpixis_init" to the file
>> "mpc85xx_ds.c". =A0mpc85xx_ds.c supports more than just the P2020DS.
>
> I'm not sure if other DS boards covered by this file has the same functio=
nality.

Well, either it does or it doesn't, but you can't add a p2020-specific
function to mpc85xx_ds.c.  You should just rename the function,
because I see no reason why it can't work on other DS boards.

The problem is the compatible string.  Each pixis implementation is
different, but they share several common traits.  I wonder if we need
to have a generic compatible string for the FPGA node.

>> You should probably choose a less generic name than "event".
>
> Well, it's the name suggested by the board manual. =A0We may change it
> to "event_button" if not too long.

I was thinking something like "fsl-fpga-event".

--=20
Timur Tabi
Linux kernel developer at Freescale

Re: [PATCH 5/7] powerpc: add the mpic timer support

[Jean-Michel Hautbois]

Hi,

I am interested by this implementation.

> +static ssize_t mpic_tm_timeout_store(struct device *dev,
> + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 struct device_attribute *attr,
> + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 const char *buf, size_t count)
> +{

<snip>

> +
> + =C2=A0 =C2=A0 =C2=A0 spin_lock_irq(&priv->lock);
> +
> + =C2=A0 =C2=A0 =C2=A0 /* stop timer 0 */
> + =C2=A0 =C2=A0 =C2=A0 temp =3D in_be32(&priv->regs->gtbcr);
> + =C2=A0 =C2=A0 =C2=A0 temp |=3D MPIC_TIMER_STOP; /* counting inhibited *=
/
> + =C2=A0 =C2=A0 =C2=A0 out_be32(&priv->regs->gtbcr, temp);
> +
> + =C2=A0 =C2=A0 =C2=A0 if (interval !=3D 0) {
> + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 /* start timer */
> + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 out_be32(&priv->regs->=
gtbcr, interval | MPIC_TIMER_STOP);
> + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 out_be32(&priv->regs->=
gtbcr, interval);
> + =C2=A0 =C2=A0 =C2=A0 }
> +
> + =C2=A0 =C2=A0 =C2=A0 spin_unlock_irq(&priv->lock);
> + =C2=A0 =C2=A0 =C2=A0 return count;
> +}

What is the delay when doing that stop/start thing ?
Is this delay variable and/or big or not ?
I would imagine a timer which would see its interval regularly
corrected, and I would like to know if this is really really fast
(say, less than some ticks) or not.

Thanks in advance for any answer !
Regards,
JM

Re: [PATCH 1/7] powerpc/85xx: re-enable timebase sync disabled by KEXEC patch

[Kumar Gala]

On Dec 3, 2010, at 6:34 AM, Li Yang wrote:

> The timebase sync is not only necessary when using KEXEC.  It should =
also
> be used by normal boot up and cpu hotplug.  Remove the ifdef added by
> the KEXEC patch.  Fix a problem that cpu hotplugging freezes the whole =
system.
>=20
> Signed-off-by: Jin Qing <b24347@freescale.com>
> Singed-off-by: Li Yang <leoli@freescale.com>
> ---
> arch/powerpc/platforms/85xx/smp.c |    4 +---
> 1 files changed, 1 insertions(+), 3 deletions(-)

But we have problems with KEXEC w/o this?  What is the issue with =
hotplugging and the generic timebase code?  When do we freeze?

>=20
> diff --git a/arch/powerpc/platforms/85xx/smp.c =
b/arch/powerpc/platforms/85xx/smp.c
> index 5c91a99..1e8aec8 100644
> --- a/arch/powerpc/platforms/85xx/smp.c
> +++ b/arch/powerpc/platforms/85xx/smp.c
> @@ -2,7 +2,7 @@
>  * Author: Andy Fleming <afleming@freescale.com>
>  * 	   Kumar Gala <galak@kernel.crashing.org>
>  *
> - * Copyright 2006-2008 Freescale Semiconductor Inc.
> + * Copyright 2006-2010 Freescale Semiconductor Inc.
>  *
>  * This program is free software; you can redistribute  it and/or =
modify it
>  * under  the terms of  the GNU General  Public License as published =
by the
> @@ -115,10 +115,8 @@ smp_85xx_setup_cpu(int cpu_nr)
>=20
> struct smp_ops_t smp_85xx_ops =3D {
> 	.kick_cpu =3D smp_85xx_kick_cpu,
> -#ifdef CONFIG_KEXEC
> 	.give_timebase	=3D smp_generic_give_timebase,
> 	.take_timebase	=3D smp_generic_take_timebase,
> -#endif
> };
>=20
> #ifdef CONFIG_KEXEC
> --=20
> 1.6.6-rc1.GIT
>=20
>=20
> _______________________________________________
> Linuxppc-dev mailing list
> Linuxppc-dev@lists.ozlabs.org
> https://lists.ozlabs.org/listinfo/linuxppc-dev

Re: [PATCH 1/7] powerpc/85xx: re-enable timebase sync disabled by KEXEC patch

[Li Yang]

On Sat, Dec 4, 2010 at 12:40 AM, Kumar Gala <galak@kernel.crashing.org> wro=
te:
>
> On Dec 3, 2010, at 6:34 AM, Li Yang wrote:
>
>> The timebase sync is not only necessary when using KEXEC. =C2=A0It shoul=
d also
>> be used by normal boot up and cpu hotplug. =C2=A0Remove the ifdef added =
by
>> the KEXEC patch. =C2=A0Fix a problem that cpu hotplugging freezes the wh=
ole system.
>>
>> Signed-off-by: Jin Qing <b24347@freescale.com>
>> Singed-off-by: Li Yang <leoli@freescale.com>
>> ---
>> arch/powerpc/platforms/85xx/smp.c | =C2=A0 =C2=A04 +---
>> 1 files changed, 1 insertions(+), 3 deletions(-)
>
> But we have problems with KEXEC w/o this? =C2=A0What is the issue with ho=
tplugging and the generic timebase code? =C2=A0When do we freeze?

Actually the KEXEC patch disables timebase sync when KEXEC is not
defined.  If the timebase sync is disabled, the timebase on non-boot
cpu will become non-consistent.  And thus ruins the scheduler when
hot-plugged.

>
>>
>> diff --git a/arch/powerpc/platforms/85xx/smp.c b/arch/powerpc/platforms/=
85xx/smp.c
>> index 5c91a99..1e8aec8 100644
>> --- a/arch/powerpc/platforms/85xx/smp.c
>> +++ b/arch/powerpc/platforms/85xx/smp.c
>> @@ -2,7 +2,7 @@
>> =C2=A0* Author: Andy Fleming <afleming@freescale.com>
>> =C2=A0* =C2=A0 =C2=A0 =C2=A0 Kumar Gala <galak@kernel.crashing.org>
>> =C2=A0*
>> - * Copyright 2006-2008 Freescale Semiconductor Inc.
>> + * Copyright 2006-2010 Freescale Semiconductor Inc.
>> =C2=A0*
>> =C2=A0* This program is free software; you can redistribute =C2=A0it and=
/or modify it
>> =C2=A0* under =C2=A0the terms of =C2=A0the GNU General =C2=A0Public Lice=
nse as published by the
>> @@ -115,10 +115,8 @@ smp_85xx_setup_cpu(int cpu_nr)
>>
>> struct smp_ops_t smp_85xx_ops =3D {
>> =C2=A0 =C2=A0 =C2=A0 .kick_cpu =3D smp_85xx_kick_cpu,
>> -#ifdef CONFIG_KEXEC
>> =C2=A0 =C2=A0 =C2=A0 .give_timebase =C2=A0=3D smp_generic_give_timebase,
>> =C2=A0 =C2=A0 =C2=A0 .take_timebase =C2=A0=3D smp_generic_take_timebase,
>> -#endif
>> };
>>
>> #ifdef CONFIG_KEXEC
>> --
>> 1.6.6-rc1.GIT
>>
>>
>> _______________________________________________
>> Linuxppc-dev mailing list
>> Linuxppc-dev@lists.ozlabs.org
>> https://lists.ozlabs.org/listinfo/linuxppc-dev
>
> _______________________________________________
> Linuxppc-dev mailing list
> Linuxppc-dev@lists.ozlabs.org
> https://lists.ozlabs.org/listinfo/linuxppc-dev
>



--=20
- Leo

[PATCH 1/7] powerpc/85xx: re-enable timebase sync disabled by KEXEC patch

[Zhao Chenhui]

From: Li Yang <leoli@freescale.com>

The timebase sync is not only necessary when using KEXEC. It should also
be used by normal boot up and cpu hotplug. Remove the ifdef added by
the KEXEC patch.

Signed-off-by: Jin Qing <b24347@freescale.com>
Signed-off-by: Li Yang <leoli@freescale.com>
---
 arch/powerpc/platforms/85xx/smp.c |    2 --
 1 files changed, 0 insertions(+), 2 deletions(-)

diff --git a/arch/powerpc/platforms/85xx/smp.c b/arch/powerpc/platforms/85xx/smp.c
index 5b9b901..9b0de9c 100644
--- a/arch/powerpc/platforms/85xx/smp.c
+++ b/arch/powerpc/platforms/85xx/smp.c
@@ -113,10 +113,8 @@ smp_85xx_kick_cpu(int nr)
 
 struct smp_ops_t smp_85xx_ops = {
 	.kick_cpu = smp_85xx_kick_cpu,
-#ifdef CONFIG_KEXEC
 	.give_timebase	= smp_generic_give_timebase,
 	.take_timebase	= smp_generic_take_timebase,
-#endif
 };
 
 #ifdef CONFIG_KEXEC
-- 
1.6.4.1

Re: [PATCH 1/7] powerpc/85xx: re-enable timebase sync disabled by KEXEC patch

[Scott Wood]

On 11/04/2011 07:29 AM, Zhao Chenhui wrote:
> From: Li Yang <leoli@freescale.com>
> 
> The timebase sync is not only necessary when using KEXEC. It should also
> be used by normal boot up and cpu hotplug. Remove the ifdef added by
> the KEXEC patch.

The KEXEC patch didn't just add the ifdef, it also added the initializers:

> @@ -105,8 +107,64 @@ smp_85xx_setup_cpu(int cpu_nr)
>  
>  struct smp_ops_t smp_85xx_ops = {
>         .kick_cpu = smp_85xx_kick_cpu,
> +#ifdef CONFIG_KEXEC
> +       .give_timebase  = smp_generic_give_timebase,
> +       .take_timebase  = smp_generic_take_timebase,
> +#endif
>  };

U-Boot synchronizes the timebase on 85xx.  With what chip and U-Boot
version are you seeing this not happen?

If you are seeing only a small (around one tick) difference, make sure
you're running a U-Boot that has this commit:

> commit 7afc45ad7d9493208d89072cbb78a5bfc8034b59
> Author: Kumar Gala <galak@kernel.crashing.org>
> Date:   Sun Mar 13 10:55:53 2011 -0500
> 
>     powerpc/85xx: Fix synchronization of timebase on MP boot
>     
>     There is a small ordering issue in the master core in that we need to
>     make sure the disabling of the timebase in the SoC is visible before we
>     set the value to 0.  We can simply just read back the value to
>     synchronizatize the write, before we set TB to 0.
>     
>     Reported-by: Dan Hettena
>     Tested-by: Dan Hettena
>     Signed-off-by: Kumar Gala <galak@kernel.crashing.org>


-Scott

Re: [PATCH 1/7] powerpc/85xx: re-enable timebase sync disabled by KEXEC patch

[Kumar Gala]

On Nov 4, 2011, at 12:33 PM, Scott Wood wrote:

> On 11/04/2011 07:29 AM, Zhao Chenhui wrote:
>> From: Li Yang <leoli@freescale.com>
>> 
>> The timebase sync is not only necessary when using KEXEC. It should also
>> be used by normal boot up and cpu hotplug. Remove the ifdef added by
>> the KEXEC patch.
> 
> The KEXEC patch didn't just add the ifdef, it also added the initializers:
> 
>> @@ -105,8 +107,64 @@ smp_85xx_setup_cpu(int cpu_nr)
>> 
>> struct smp_ops_t smp_85xx_ops = {
>>        .kick_cpu = smp_85xx_kick_cpu,
>> +#ifdef CONFIG_KEXEC
>> +       .give_timebase  = smp_generic_give_timebase,
>> +       .take_timebase  = smp_generic_take_timebase,
>> +#endif
>> };
> 
> U-Boot synchronizes the timebase on 85xx.  With what chip and U-Boot
> version are you seeing this not happen?
> 
> If you are seeing only a small (around one tick) difference, make sure
> you're running a U-Boot that has this commit:
> 
>> commit 7afc45ad7d9493208d89072cbb78a5bfc8034b59
>> Author: Kumar Gala <galak@kernel.crashing.org>
>> Date:   Sun Mar 13 10:55:53 2011 -0500
>> 
>>    powerpc/85xx: Fix synchronization of timebase on MP boot
>> 
>>    There is a small ordering issue in the master core in that we need to
>>    make sure the disabling of the timebase in the SoC is visible before we
>>    set the value to 0.  We can simply just read back the value to
>>    synchronizatize the write, before we set TB to 0.
>> 
>>    Reported-by: Dan Hettena
>>    Tested-by: Dan Hettena
>>    Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

Scott,

Aren't we going to need this when a core is woken back up w/o any state?

- k

Re: [PATCH 1/7] powerpc/85xx: re-enable timebase sync disabled by KEXEC patch

[Scott Wood]

On 11/04/2011 02:33 PM, Kumar Gala wrote:
> 
> On Nov 4, 2011, at 12:33 PM, Scott Wood wrote:
> 
>> On 11/04/2011 07:29 AM, Zhao Chenhui wrote:
>>> From: Li Yang <leoli@freescale.com>
>>>
>>> The timebase sync is not only necessary when using KEXEC. It should also
>>> be used by normal boot up and cpu hotplug. Remove the ifdef added by
>>> the KEXEC patch.
>>
>> The KEXEC patch didn't just add the ifdef, it also added the initializers:
>>
>>> @@ -105,8 +107,64 @@ smp_85xx_setup_cpu(int cpu_nr)
>>>
>>> struct smp_ops_t smp_85xx_ops = {
>>>        .kick_cpu = smp_85xx_kick_cpu,
>>> +#ifdef CONFIG_KEXEC
>>> +       .give_timebase  = smp_generic_give_timebase,
>>> +       .take_timebase  = smp_generic_take_timebase,
>>> +#endif
>>> };
>>
>> U-Boot synchronizes the timebase on 85xx.  With what chip and U-Boot
>> version are you seeing this not happen?
>>
>> If you are seeing only a small (around one tick) difference, make sure
>> you're running a U-Boot that has this commit:
[snip]
> 
> Scott,
> 
> Aren't we going to need this when a core is woken back up w/o any state?

We'll need some form of timebase resync if a core is individually
hard-reset -- I was responding to the "should also be used by normal
boot up" bit.

For kexec/hotplug, if we must reset the core (for deep sleep we must),
any reason not to do the sync the same way U-Boot does?

-Scott

RE: [PATCH 1/7] powerpc/85xx: re-enable timebase sync disabled by KEXEC patch

[Li Yang-R58472]

>-----Original Message-----
>From: linuxppc-dev-bounces+leoli=3Dfreescale.com@lists.ozlabs.org
>[mailto:linuxppc-dev-bounces+leoli=3Dfreescale.com@lists.ozlabs.org] On
>Behalf Of Scott Wood
>Sent: Saturday, November 05, 2011 1:34 AM
>To: Zhao Chenhui-B35336
>Cc: linuxppc-dev@lists.ozlabs.org
>Subject: Re: [PATCH 1/7] powerpc/85xx: re-enable timebase sync disabled by
>KEXEC patch
>
>On 11/04/2011 07:29 AM, Zhao Chenhui wrote:
>> From: Li Yang <leoli@freescale.com>
>>
>> The timebase sync is not only necessary when using KEXEC. It should also
>> be used by normal boot up and cpu hotplug. Remove the ifdef added by
>> the KEXEC patch.
>
>The KEXEC patch didn't just add the ifdef, it also added the initializers:

Yes.  But the code suggests that the timebase synchronization is only neces=
sary for KEXEC, but it turns out that sleep/wakeup also need it.  Maybe the=
 description of the patch need to be changed as KEXEC is not to be blamed.

>
>> @@ -105,8 +107,64 @@ smp_85xx_setup_cpu(int cpu_nr)
>>
>>  struct smp_ops_t smp_85xx_ops =3D {
>>         .kick_cpu =3D smp_85xx_kick_cpu,
>> +#ifdef CONFIG_KEXEC
>> +       .give_timebase  =3D smp_generic_give_timebase,
>> +       .take_timebase  =3D smp_generic_take_timebase,
>> +#endif
>>  };
>
>U-Boot synchronizes the timebase on 85xx.  With what chip and U-Boot
>version are you seeing this not happen?

I'm curious why don't we make it happen in kernel as we are against adding =
dependency to the bootloader?  Other architectures don't have this dependen=
cy, it will be better if we don't add this dependency either IMO.


>
>If you are seeing only a small (around one tick) difference, make sure
>you're running a U-Boot that has this commit:
>
>> commit 7afc45ad7d9493208d89072cbb78a5bfc8034b59
>> Author: Kumar Gala <galak@kernel.crashing.org>
>> Date:   Sun Mar 13 10:55:53 2011 -0500
>>
>>     powerpc/85xx: Fix synchronization of timebase on MP boot
>>
>>     There is a small ordering issue in the master core in that we need
>to
>>     make sure the disabling of the timebase in the SoC is visible before
>we
>>     set the value to 0.  We can simply just read back the value to
>>     synchronizatize the write, before we set TB to 0.
>>
>>     Reported-by: Dan Hettena
>>     Tested-by: Dan Hettena
>>     Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
>
>
>-Scott
>
>_______________________________________________
>Linuxppc-dev mailing list
>Linuxppc-dev@lists.ozlabs.org
>https://lists.ozlabs.org/listinfo/linuxppc-dev

Re: [PATCH 1/7] powerpc/85xx: re-enable timebase sync disabled by KEXEC patch

[Scott Wood]

On 11/08/2011 03:06 AM, Li Yang-R58472 wrote:
> 
> 
>> -----Original Message-----
>> From: linuxppc-dev-bounces+leoli=freescale.com@lists.ozlabs.org
>> [mailto:linuxppc-dev-bounces+leoli=freescale.com@lists.ozlabs.org] On
>> Behalf Of Scott Wood
>> Sent: Saturday, November 05, 2011 1:34 AM
>> To: Zhao Chenhui-B35336
>> Cc: linuxppc-dev@lists.ozlabs.org
>> Subject: Re: [PATCH 1/7] powerpc/85xx: re-enable timebase sync disabled by
>> KEXEC patch
>>
>> On 11/04/2011 07:29 AM, Zhao Chenhui wrote:
>>> From: Li Yang <leoli@freescale.com>
>>>
>>> The timebase sync is not only necessary when using KEXEC. It should also
>>> be used by normal boot up and cpu hotplug. Remove the ifdef added by
>>> the KEXEC patch.
>>
>> The KEXEC patch didn't just add the ifdef, it also added the initializers:
> 
> Yes.  But the code suggests that the timebase synchronization is only necessary for KEXEC, but it turns out that sleep/wakeup also need it.  Maybe the description of the patch need to be changed as KEXEC is not to be blamed.

It is needed when you hard reset a core.  This was something we never
did on SMP before kexec.  Now you're adding a second thing that does it,
so it'll need the sync as well, but that doesn't mean we should do it on
normal boot.

>>> @@ -105,8 +107,64 @@ smp_85xx_setup_cpu(int cpu_nr)
>>>
>>>  struct smp_ops_t smp_85xx_ops = {
>>>         .kick_cpu = smp_85xx_kick_cpu,
>>> +#ifdef CONFIG_KEXEC
>>> +       .give_timebase  = smp_generic_give_timebase,
>>> +       .take_timebase  = smp_generic_take_timebase,
>>> +#endif
>>>  };
>>
>> U-Boot synchronizes the timebase on 85xx.  With what chip and U-Boot
>> version are you seeing this not happen?
> 
> I'm curious why don't we make it happen in kernel as we are against
> adding dependency to the bootloader?

We are against adding gratuitous dependencies on the bootloader, but
some things are just a lot easier to do in that context.  Nobody
complains about Linux expecting RAM to be working on entry. :-)

While it's certainly possible to do this in Linux (and should be done
the way U-Boot does instead of the software sync, in the cases where we
need to), it's easier to do in U-Boot, before the cores are running.

It would be impossible for Linux to do this (or any other tb
modifications) when running on top of a hypervisor.

In http://lists.ozlabs.org/pipermail/linuxppc-dev/2011-June/091321.html,
Ben Herrenschmidt said, "smp-tbsync.c is and has always been a
'workaround' for broken HW."

> Other architectures don't have this dependency, 

Which "other architectures" are you referring to?

On PPC server this is handled with a firmware call to freeze the
timebase.  On x86 this is handled by the BIOS by the time the OS starts.

-Scott