powerpc/powernv: Basic support for OPAL

Add definition of OPAL interfaces along with  the wrappers to call
into OPAL runtime and the early device-tree parsing hook to locate
the OPAL runtime firmware.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

arch/powerpc/include/asm/firmware.h

@@ -48,6 +48,8 @@
 #define FW_FEATURE_CMO		ASM_CONST(0x0000000002000000)
 #define FW_FEATURE_VPHN		ASM_CONST(0x0000000004000000)
 #define FW_FEATURE_XCMO		ASM_CONST(0x0000000008000000)
+#define FW_FEATURE_OPAL		ASM_CONST(0x0000000010000000)
+#define FW_FEATURE_OPALv2	ASM_CONST(0x0000000020000000)
 
 #ifndef __ASSEMBLY__
 
@@ -65,6 +67,8 @@ enum {
 	FW_FEATURE_PSERIES_ALWAYS = 0,
 	FW_FEATURE_ISERIES_POSSIBLE = FW_FEATURE_ISERIES | FW_FEATURE_LPAR,
 	FW_FEATURE_ISERIES_ALWAYS = FW_FEATURE_ISERIES | FW_FEATURE_LPAR,
+	FW_FEATURE_POWERNV_POSSIBLE = FW_FEATURE_OPAL | FW_FEATURE_OPALv2,
+	FW_FEATURE_POWERNV_ALWAYS = 0,
 	FW_FEATURE_PS3_POSSIBLE = FW_FEATURE_LPAR | FW_FEATURE_PS3_LV1,
 	FW_FEATURE_PS3_ALWAYS = FW_FEATURE_LPAR | FW_FEATURE_PS3_LV1,
 	FW_FEATURE_CELLEB_POSSIBLE = FW_FEATURE_LPAR | FW_FEATURE_BEAT,
@@ -78,6 +82,9 @@ enum {
 #ifdef CONFIG_PPC_ISERIES
 		FW_FEATURE_ISERIES_POSSIBLE |
 #endif
+#ifdef CONFIG_PPC_POWERNV
+		FW_FEATURE_POWERNV_POSSIBLE |
+#endif
 #ifdef CONFIG_PPC_PS3
 		FW_FEATURE_PS3_POSSIBLE |
 #endif
@@ -95,6 +102,9 @@ enum {
 #ifdef CONFIG_PPC_ISERIES
 		FW_FEATURE_ISERIES_ALWAYS &
 #endif
+#ifdef CONFIG_PPC_POWERNV
+		FW_FEATURE_POWERNV_ALWAYS &
+#endif
 #ifdef CONFIG_PPC_PS3
 		FW_FEATURE_PS3_ALWAYS &
 #endif

arch/powerpc/kernel/prom.c

@@ -54,6 +54,8 @@
 #include <asm/pci-bridge.h>
 #include <asm/phyp_dump.h>
 #include <asm/kexec.h>
+#include <asm/opal.h>
+
 #include <mm/mmu_decl.h>
 
 #ifdef DEBUG
@@ -707,6 +709,11 @@ void __init early_init_devtree(void *params)
 	of_scan_flat_dt(early_init_dt_scan_rtas, NULL);
 #endif
 
+#ifdef CONFIG_PPC_POWERNV
+	/* Some machines might need OPAL info for debugging, grab it now. */
+	of_scan_flat_dt(early_init_dt_scan_opal, NULL);
+#endif
+
 #ifdef CONFIG_PHYP_DUMP
 	/* scan tree to see if dump occurred during last boot */
 	of_scan_flat_dt(early_init_dt_scan_phyp_dump, NULL);

arch/powerpc/platforms/powernv/Kconfig

 config PPC_POWERNV
 	depends on PPC64 && PPC_BOOK3S
 	bool "IBM PowerNV (Non-Virtualized) platform support"
-	select PPC_RTAS
 	select PPC_NATIVE
 	select PPC_XICS
 	select PPC_ICP_NATIVE
-	select PPC_ICS_RTAS
 	select PPC_P7_NAP
 	select PPC_PCI_CHOICE if EMBEDDED
 	default y
+
+config PPC_POWERNV_RTAS
+	depends on PPC_POWERNV
+	bool "Support for RTAS based PowerNV platforms such as BML"
+	default y
+	select PPC_ICS_RTAS
+	select PPC_RTAS

arch/powerpc/platforms/powernv/Makefile

-obj-y			+= setup.o opal-takeover.o
+obj-y			+= setup.o opal-takeover.o opal-wrappers.o opal.o
 obj-$(CONFIG_SMP)	+= smp.o

arch/powerpc/platforms/powernv/opal-wrappers.S

+/*
+ * PowerNV OPAL API wrappers
+ *
+ * Copyright 2011 IBM Corp.
+ *
+ * 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/ppc_asm.h>
+#include <asm/hvcall.h>
+#include <asm/asm-offsets.h>
+#include <asm/opal.h>
+
+/* TODO:
+ *
+ * - Trace irqs in/off (needs saving/restoring all args, argh...)
+ * - Get r11 feed up by Dave so I can have better register usage
+ */
+#define OPAL_CALL(name, token)		\
+ _GLOBAL(name);				\
+	mflr	r0;			\
+	mfcr	r12;			\
+	std	r0,16(r1);		\
+	std	r12,8(r1);		\
+	std	r1,PACAR1(r13);		\
+	li	r0,0;			\
+	mfmsr	r12;			\
+	ori	r0,r0,MSR_EE;		\
+	std	r12,PACASAVEDMSR(r13);	\
+	andc	r12,r12,r0;		\
+	mtmsrd	r12,1;			\
+	LOAD_REG_ADDR(r0,.opal_return);	\
+	mtlr	r0;			\
+	li	r0,MSR_DR|MSR_IR;	\
+	andc	r12,r12,r0;		\
+	li	r0,token;		\
+	mtspr	SPRN_HSRR1,r12;		\
+	LOAD_REG_ADDR(r11,opal);	\
+	ld	r12,8(r11);		\
+	ld	r2,0(r11);		\
+	mtspr	SPRN_HSRR0,r12;		\
+	hrfid
+
+_STATIC(opal_return)
+	ld	r2,PACATOC(r13);
+	ld	r4,8(r1);
+	ld	r5,16(r1);
+	ld	r6,PACASAVEDMSR(r13);
+	mtspr	SPRN_SRR0,r5;
+	mtspr	SPRN_SRR1,r6;
+	mtcr	r4;
+	rfid
+
+OPAL_CALL(opal_console_write,			OPAL_CONSOLE_WRITE);
+OPAL_CALL(opal_console_read,			OPAL_CONSOLE_READ);
+OPAL_CALL(opal_console_write_buffer_space,	OPAL_CONSOLE_WRITE_BUFFER_SPACE);
+OPAL_CALL(opal_rtc_read,			OPAL_RTC_READ);
+OPAL_CALL(opal_rtc_write,			OPAL_RTC_WRITE);
+OPAL_CALL(opal_cec_power_down,			OPAL_CEC_POWER_DOWN);
+OPAL_CALL(opal_cec_reboot,			OPAL_CEC_REBOOT);
+OPAL_CALL(opal_read_nvram,			OPAL_READ_NVRAM);
+OPAL_CALL(opal_write_nvram,			OPAL_WRITE_NVRAM);
+OPAL_CALL(opal_handle_interrupt,		OPAL_HANDLE_INTERRUPT);
+OPAL_CALL(opal_poll_events,			OPAL_POLL_EVENTS);
+OPAL_CALL(opal_pci_set_hub_tce_memory,		OPAL_PCI_SET_HUB_TCE_MEMORY);
+OPAL_CALL(opal_pci_set_phb_tce_memory,		OPAL_PCI_SET_PHB_TCE_MEMORY);
+OPAL_CALL(opal_pci_config_read_byte,		OPAL_PCI_CONFIG_READ_BYTE);
+OPAL_CALL(opal_pci_config_read_half_word,	OPAL_PCI_CONFIG_READ_HALF_WORD);
+OPAL_CALL(opal_pci_config_read_word,		OPAL_PCI_CONFIG_READ_WORD);
+OPAL_CALL(opal_pci_config_write_byte,		OPAL_PCI_CONFIG_WRITE_BYTE);
+OPAL_CALL(opal_pci_config_write_half_word,	OPAL_PCI_CONFIG_WRITE_HALF_WORD);
+OPAL_CALL(opal_pci_config_write_word,		OPAL_PCI_CONFIG_WRITE_WORD);
+OPAL_CALL(opal_set_xive,			OPAL_SET_XIVE);
+OPAL_CALL(opal_get_xive,			OPAL_GET_XIVE);
+OPAL_CALL(opal_register_exception_handler,	OPAL_REGISTER_OPAL_EXCEPTION_HANDLER);
+OPAL_CALL(opal_pci_eeh_freeze_status,		OPAL_PCI_EEH_FREEZE_STATUS);
+OPAL_CALL(opal_pci_eeh_freeze_clear,		OPAL_PCI_EEH_FREEZE_CLEAR);
+OPAL_CALL(opal_pci_shpc,			OPAL_PCI_SHPC);
+OPAL_CALL(opal_pci_phb_mmio_enable,		OPAL_PCI_PHB_MMIO_ENABLE);
+OPAL_CALL(opal_pci_set_phb_mem_window,		OPAL_PCI_SET_PHB_MEM_WINDOW);
+OPAL_CALL(opal_pci_map_pe_mmio_window,		OPAL_PCI_MAP_PE_MMIO_WINDOW);
+OPAL_CALL(opal_pci_set_phb_table_memory,	OPAL_PCI_SET_PHB_TABLE_MEMORY);
+OPAL_CALL(opal_pci_set_pe,			OPAL_PCI_SET_PE);
+OPAL_CALL(opal_pci_set_peltv,			OPAL_PCI_SET_PELTV);
+OPAL_CALL(opal_pci_set_mve,			OPAL_PCI_SET_MVE);
+OPAL_CALL(opal_pci_set_mve_enable,		OPAL_PCI_SET_MVE_ENABLE);
+OPAL_CALL(opal_pci_get_xive_reissue,		OPAL_PCI_GET_XIVE_REISSUE);
+OPAL_CALL(opal_pci_set_xive_reissue,		OPAL_PCI_SET_XIVE_REISSUE);
+OPAL_CALL(opal_pci_set_xive_pe,			OPAL_PCI_SET_XIVE_PE);
+OPAL_CALL(opal_get_xive_source,			OPAL_GET_XIVE_SOURCE);
+OPAL_CALL(opal_get_msi_32,			OPAL_GET_MSI_32);
+OPAL_CALL(opal_get_msi_64,			OPAL_GET_MSI_64);
+OPAL_CALL(opal_start_cpu,			OPAL_START_CPU);
+OPAL_CALL(opal_query_cpu_status,		OPAL_QUERY_CPU_STATUS);
+OPAL_CALL(opal_write_oppanel,			OPAL_WRITE_OPPANEL);
+OPAL_CALL(opal_pci_map_pe_dma_window,		OPAL_PCI_MAP_PE_DMA_WINDOW);
+OPAL_CALL(opal_pci_map_pe_dma_window_real,	OPAL_PCI_MAP_PE_DMA_WINDOW_REAL);
+OPAL_CALL(opal_pci_reset,			OPAL_PCI_RESET);

arch/powerpc/platforms/powernv/opal.c

+/*
+ * PowerNV OPAL high level interfaces
+ *
+ * Copyright 2011 IBM Corp.
+ *
+ * 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.
+ */
+
+#undef DEBUG
+
+#include <linux/types.h>
+#include <linux/of.h>
+#include <linux/of_platform.h>
+#include <asm/opal.h>
+#include <asm/firmware.h>
+
+#include "powernv.h"
+
+struct opal {
+	u64 base;
+	u64 entry;
+} opal;
+
+static struct device_node *opal_node;
+static DEFINE_SPINLOCK(opal_write_lock);
+
+int __init early_init_dt_scan_opal(unsigned long node,
+				   const char *uname, int depth, void *data)
+{
+	const void *basep, *entryp;
+	unsigned long basesz, entrysz;
+
+	if (depth != 1 || strcmp(uname, "ibm,opal") != 0)
+		return 0;
+
+	basep  = of_get_flat_dt_prop(node, "opal-base-address", &basesz);
+	entryp = of_get_flat_dt_prop(node, "opal-entry-address", &entrysz);
+
+	if (!basep || !entryp)
+		return 1;
+
+	opal.base = of_read_number(basep, basesz/4);
+	opal.entry = of_read_number(entryp, entrysz/4);
+
+	pr_debug("OPAL Base  = 0x%llx (basep=%p basesz=%ld)\n",
+		 opal.base, basep, basesz);
+	pr_debug("OPAL Entry = 0x%llx (entryp=%p basesz=%ld)\n",
+		 opal.entry, entryp, entrysz);
+
+	powerpc_firmware_features |= FW_FEATURE_OPAL;
+	if (of_flat_dt_is_compatible(node, "ibm,opal-v2")) {
+		powerpc_firmware_features |= FW_FEATURE_OPALv2;
+		printk("OPAL V2 detected !\n");
+	} else {
+		printk("OPAL V1 detected !\n");
+	}
+
+	return 1;
+}
+
+int opal_get_chars(uint32_t vtermno, char *buf, int count)
+{
+	s64 len, rc;
+	u64 evt;
+
+	if (!opal.entry)
+		return 0;
+	opal_poll_events(&evt);
+	if ((evt & OPAL_EVENT_CONSOLE_INPUT) == 0)
+		return 0;
+	len = count;
+	rc = opal_console_read(vtermno, &len, buf);
+	if (rc == OPAL_SUCCESS)
+		return len;
+	return 0;
+}
+
+int opal_put_chars(uint32_t vtermno, const char *data, int total_len)
+{
+	int written = 0;
+	s64 len, rc = OPAL_BUSY;
+	unsigned long flags;
+	u64 evt;
+
+	if (!opal.entry)
+		return 0;
+
+	/* We want put_chars to be atomic to avoid mangling of hvsi
+	 * packets. To do that, we first test for room and return
+	 * -EAGAIN if there isn't enough
+	 */
+	spin_lock_irqsave(&opal_write_lock, flags);
+	rc = opal_console_write_buffer_space(vtermno, &len);
+	if (rc || len < total_len) {
+		spin_unlock_irqrestore(&opal_write_lock, flags);
+		/* Closed -> drop characters */
+		if (rc)
+			return total_len;
+		opal_poll_events(&evt);
+		return -EAGAIN;
+	}
+
+	/* We still try to handle partial completions, though they
+	 * should no longer happen.
+	 */
+	while(total_len > 0 && (rc == OPAL_BUSY ||
+				rc == OPAL_BUSY_EVENT || rc == OPAL_SUCCESS)) {
+		len = total_len;
+		rc = opal_console_write(vtermno, &len, data);
+		if (rc == OPAL_SUCCESS) {
+			total_len -= len;
+			data += len;
+			written += len;
+		}
+		/* This is a bit nasty but we need that for the console to
+		 * flush when there aren't any interrupts. We will clean
+		 * things a bit later to limit that to synchronous path
+		 * such as the kernel console and xmon/udbg
+		 */
+		do
+			opal_poll_events(&evt);
+		while(rc == OPAL_SUCCESS && (evt & OPAL_EVENT_CONSOLE_OUTPUT));
+	}
+	spin_unlock_irqrestore(&opal_write_lock, flags);
+	return written;
+}
+
+static int __init opal_init(void)
+{
+	struct device_node *np, *consoles;
+
+	opal_node = of_find_node_by_path("/ibm,opal");
+	if (!opal_node) {
+		pr_warn("opal: Node not found\n");
+		return -ENODEV;
+	}
+	if (firmware_has_feature(FW_FEATURE_OPALv2))
+		consoles = of_find_node_by_path("/ibm,opal/consoles");
+	else
+		consoles = of_node_get(opal_node);
+
+	/* Register serial ports */
+	for_each_child_of_node(consoles, np) {
+		if (strcmp(np->name, "serial"))
+			continue;
+		of_platform_device_create(np, NULL, NULL);
+	}
+	of_node_put(consoles);
+	return 0;
+}
+subsys_initcall(opal_init);

arch/powerpc/platforms/powernv/setup.c

@@ -74,6 +74,12 @@ static void pnv_show_cpuinfo(struct seq_file *m)
 	if (root)
 		model = of_get_property(root, "model", NULL);
 	seq_printf(m, "machine\t\t: PowerNV %s\n", model);
+	if (firmware_has_feature(FW_FEATURE_OPALv2))
+		seq_printf(m, "firmware\t: OPAL v2\n");
+	else if (firmware_has_feature(FW_FEATURE_OPAL))
+		seq_printf(m, "firmware\t: OPAL v1\n");
+	else
+		seq_printf(m, "firmware\t: BML\n");
 	of_node_put(root);
 }
 

arch/powerpc/platforms/powernv/smp.c

@@ -30,6 +30,7 @@
 #include <asm/vdso_datapage.h>
 #include <asm/cputhreads.h>
 #include <asm/xics.h>
+#include <asm/opal.h>
 
 #include "powernv.h"
 
@@ -62,6 +63,28 @@ static int pnv_smp_cpu_bootable(unsigned int nr)
 	return 1;
 }
 
+int __devinit pnv_smp_kick_cpu(int nr)
+{
+	unsigned int pcpu = get_hard_smp_processor_id(nr);
+	unsigned long start_here = __pa(*((unsigned long *)
+					  generic_secondary_smp_init));
+	long rc;
+
+	BUG_ON(nr < 0 || nr >= NR_CPUS);
+
+	/* On OPAL v2 the CPU are still spinning inside OPAL itself,
+	 * get them back now
+	 */
+	if (firmware_has_feature(FW_FEATURE_OPALv2)) {
+		pr_devel("OPAL: Starting CPU %d (HW 0x%x)...\n", nr, pcpu);
+		rc = opal_start_cpu(pcpu, start_here);
+		if (rc != OPAL_SUCCESS)
+			pr_warn("OPAL Error %ld starting CPU %d\n",
+				rc, nr);
+	}
+	return smp_generic_kick_cpu(nr);
+}
+
 #ifdef CONFIG_HOTPLUG_CPU
 
 static int pnv_smp_cpu_disable(void)
@@ -127,7 +150,7 @@ static struct smp_ops_t pnv_smp_ops = {
 	.message_pass	= smp_muxed_ipi_message_pass,
 	.cause_ipi	= NULL,	/* Filled at runtime by xics_smp_probe() */
 	.probe		= xics_smp_probe,
-	.kick_cpu	= smp_generic_kick_cpu,
+	.kick_cpu	= pnv_smp_kick_cpu,
 	.setup_cpu	= pnv_smp_setup_cpu,
 	.cpu_bootable	= pnv_smp_cpu_bootable,
 #ifdef CONFIG_HOTPLUG_CPU