ppc32: Update/cleanup low level POWER4 & G5 CPU support

arch/ppc/kernel/Makefile

@@ -22,6 +22,7 @@ obj-y				:= entry.o traps.o irq.o idle.o time.o misc.o \
 					semaphore.o syscalls.o setup.o \
 					cputable.o ppc_htab.o
 obj-$(CONFIG_6xx)		+= l2cr.o cpu_setup_6xx.o
+obj-$(CONFIG_POWER4)		+= cpu_setup_power4.o
 obj-$(CONFIG_MODULES)		+= module.o ppc_ksyms.o
 obj-$(CONFIG_PCI)		+= pci.o
 obj-$(CONFIG_PCI)		+= pci-dma.o

arch/ppc/kernel/cpu_setup_power4.S

+/*
+ * This file contains low level CPU setup functions.
+ *    Copyright (C) 2003 Benjamin Herrenschmidt (benh@kernel.crashing.org)
+ *
+ * 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 <linux/config.h>
+#include <asm/processor.h>
+#include <asm/page.h>
+#include <asm/ppc_asm.h>
+#include <asm/cputable.h>
+#include <asm/ppc_asm.h>
+#include <asm/offsets.h>
+#include <asm/cache.h>
+
+_GLOBAL(__power4_cpu_preinit)
+	/*
+	 * On the PPC970, we have to turn off real-mode cache inhibit
+	 * early, before we first turn the MMU off.
+	 */
+	mfspr	r0,SPRN_PVR
+	srwi	r0,r0,16
+	cmpwi	r0,0x39
+	bnelr
+
+	li	r0,0
+	sync
+	mtspr	SPRN_HID4,r0
+	isync
+	sync
+	mtspr	SPRN_HID5,r0
+	isync
+
+	mfspr	r0,SPRN_HID1
+	li	r11,0x1200		/* enable i-fetch cacheability */
+	sldi	r11,r11,44		/* and prefetch */
+	or	r0,r0,r11
+	mtspr	SPRN_HID1,r0
+	mtspr	SPRN_HID1,r0
+	isync
+	li	r0,0
+	sync
+	mtspr	SPRN_HIOR,0		/* Clear interrupt prefix */
+	isync
+	blr
+
+_GLOBAL(__setup_cpu_power4)
+	blr
+_GLOBAL(__setup_cpu_ppc970)
+	mfspr	r0,SPRN_HID0
+	li	r11,5			/* clear DOZE and SLEEP */
+	rldimi	r0,r11,52,8		/* set NAP and DPM */
+	mtspr	SPRN_HID0,r0
+	mfspr	r0,SPRN_HID0
+	mfspr	r0,SPRN_HID0
+	mfspr	r0,SPRN_HID0
+	mfspr	r0,SPRN_HID0
+	mfspr	r0,SPRN_HID0
+	mfspr	r0,SPRN_HID0
+	sync
+	isync
+	blr
+
+/* Definitions for the table use to save CPU states */
+#define CS_HID0		0
+#define CS_HID1		8
+#define	CS_HID4		16
+#define CS_HID5		24
+#define CS_SIZE		32
+
+	.data
+	.balign	L1_CACHE_LINE_SIZE
+cpu_state_storage:	
+	.space	CS_SIZE
+	.balign	L1_CACHE_LINE_SIZE,0
+	.text
+	
+/* Called in normal context to backup CPU 0 state. This
+ * does not include cache settings. This function is also
+ * called for machine sleep. This does not include the MMU
+ * setup, BATs, etc... but rather the "special" registers
+ * like HID0, HID1, HID4, etc...
+ */
+_GLOBAL(__save_cpu_setup)
+	/* Some CR fields are volatile, we back it up all */
+	mfcr	r7
+
+	/* Get storage ptr */
+	lis	r5,cpu_state_storage@h
+	ori	r5,r5,cpu_state_storage@l
+
+	/* We only deal with 970 for now */
+	mfspr	r0,SPRN_PVR
+	srwi	r0,r0,16
+	cmpwi	r0,0x39
+	bne	1f
+
+	/* Save HID0,1,4 and 5 */
+	mfspr	r3,SPRN_HID0
+	std	r3,CS_HID0(r5)
+	mfspr	r3,SPRN_HID1
+	std	r3,CS_HID1(r5)
+	mfspr	r3,SPRN_HID4
+	std	r3,CS_HID4(r5)
+	mfspr	r3,SPRN_HID5
+	std	r3,CS_HID5(r5)
+	
+1:
+	mtcr	r7
+	blr
+
+/* Called with no MMU context (typically MSR:IR/DR off) to
+ * restore CPU state as backed up by the previous
+ * function. This does not include cache setting
+ */
+_GLOBAL(__restore_cpu_setup)
+	/* Some CR fields are volatile, we back it up all */
+	mfcr	r7
+
+	/* Get storage ptr */
+	lis	r5,(cpu_state_storage-KERNELBASE)@h
+	ori	r5,r5,cpu_state_storage@l
+
+	/* We only deal with 970 for now */
+	mfspr	r0,SPRN_PVR
+	srwi	r0,r0,16
+	cmpwi	r0,0x39
+	bne	1f
+
+	/* Clear interrupt prefix */
+	li	r0,0
+	sync
+	mtspr	SPRN_HIOR,0
+	isync
+
+	/* Restore HID0 */
+	ld	r3,CS_HID0(r5)
+	sync
+	isync
+	mtspr	SPRN_HID0,r3
+	mfspr	r3,SPRN_HID0
+	mfspr	r3,SPRN_HID0
+	mfspr	r3,SPRN_HID0
+	mfspr	r3,SPRN_HID0
+	mfspr	r3,SPRN_HID0
+	mfspr	r3,SPRN_HID0
+	sync
+	isync
+
+	/* Restore HID1 */
+	ld	r3,CS_HID1(r5)
+	sync
+	isync
+	mtspr	SPRN_HID1,r3
+	mtspr	SPRN_HID1,r3
+	sync
+	isync
+	
+	/* Restore HID4 */
+	ld	r3,CS_HID4(r5)
+	sync
+	isync
+	mtspr	SPRN_HID4,r3
+	sync
+	isync
+
+	/* Restore HID5 */
+	ld	r3,CS_HID5(r5)
+	sync
+	isync
+	mtspr	SPRN_HID5,r3
+	sync
+	isync
+1:
+	mtcr	r7
+	blr
+

arch/ppc/kernel/head.S

@@ -141,17 +141,6 @@ __start:
 	mr	r27,r7
 	li	r24,0			/* cpu # */
 
-#ifdef CONFIG_POWER4
-/*
- * On the PPC970, we have to turn off real-mode cache inhibit
- * early, before we first turn the MMU off.
- */
-	mfspr	r0,SPRN_PVR
-	srwi	r0,r0,16
-	cmpwi	r0,0x39
-	beql	ppc970_setup_hid
-#endif /* CONFIG_POWER4 */
-
 /*
  * early_init() does the early machine identification and does
  * the necessary low-level setup and clears the BSS
@@ -159,6 +148,14 @@ __start:
  */
 	bl	early_init
 
+/*
+ * On POWER4, we first need to tweak some CPU configuration registers
+ * like real mode cache inhibit or exception base
+ */
+#ifdef CONFIG_POWER4
+	bl	__power4_cpu_preinit
+#endif /* CONFIG_POWER4 */
+
 #ifdef CONFIG_APUS
 /* On APUS the __va/__pa constants need to be set to the correct
  * values before continuing.
@@ -1278,26 +1275,15 @@ __secondary_start:
  */
 _GLOBAL(__setup_cpu_power3)
 	blr
-_GLOBAL(__setup_cpu_power4)
-	blr
-_GLOBAL(__setup_cpu_ppc970)
-	blr
 _GLOBAL(__setup_cpu_generic)
 	blr
 
-#ifndef CONFIG_6xx
+#if !defined(CONFIG_6xx) && !defined(CONFIG_POWER4)
 _GLOBAL(__save_cpu_setup)
 	blr
 _GLOBAL(__restore_cpu_setup)
-#ifdef CONFIG_POWER4
-	/* turn off real-mode cache inhibit on the PPC970 */
-	mfspr	r0,SPRN_PVR
-	srwi	r0,r0,16
-	cmpwi	r0,0x39
-	beq	ppc970_setup_hid
-#endif
 	blr
-#endif /* CONFIG_6xx */
+#endif /* !defined(CONFIG_6xx) && !defined(CONFIG_POWER4) */
 
 
 /*
@@ -1633,10 +1619,14 @@ initial_mm_power4:
 	lis	r4,0x2000		/* set pseudo-segment reg 12 */
 	ori	r5,r4,0x0ccc
 	mtsr	12,r5
+#if 0
 	ori	r5,r4,0x0888		/* set pseudo-segment reg 8 */
 	mtsr	8,r5			/* (for access to serial port) */
-	ori	r5,r4,0x0999		/* set pseudo-segment reg 8 */
+#endif
+#ifdef CONFIG_BOOTX_TEXT
+	ori	r5,r4,0x0999		/* set pseudo-segment reg 9 */
 	mtsr	9,r5			/* (for access to screen) */
+#endif
 	mfmsr	r0
 	clrldi	r0,r0,1
 	sync
@@ -1644,41 +1634,6 @@ initial_mm_power4:
 	isync
 	blr
 
-/*
- * On 970 (G5), we pre-set a few bits in HID0 & HID1
- */
-ppc970_setup_hid:
-	li	r0,0
-	sync
-	mtspr	0x3f4,r0
-	isync
-	sync
-	mtspr	0x3f6,r0
-	isync
-	mfspr	r0,SPRN_HID0
-	li	r11,1			/* clear DOZE, NAP and SLEEP */
-	rldimi	r0,r11,52,8		/* set DPM */
-	mtspr	SPRN_HID0,r0
-	mfspr	r0,SPRN_HID0
-	mfspr	r0,SPRN_HID0
-	mfspr	r0,SPRN_HID0
-	mfspr	r0,SPRN_HID0
-	mfspr	r0,SPRN_HID0
-	mfspr	r0,SPRN_HID0
-	sync
-	isync
-	mfspr	r0,SPRN_HID1
-	li	r11,0x1200		/* enable i-fetch cacheability */
-	sldi	r11,r11,44		/* and prefetch */
-	or	r0,r0,r11
-	mtspr	SPRN_HID1,r0
-	mtspr	SPRN_HID1,r0
-	isync
-	li	r0,0
-	sync
-	mtspr	0x137,0
-	isync
-	blr
 #endif /* CONFIG_POWER4 */
 	
 #ifdef CONFIG_8260

arch/ppc/kernel/idle_power4.S

@@ -28,17 +28,11 @@
 
 /*
  * Init idle, called at early CPU setup time from head.S for each CPU
- * Make sure no rest of NAP mode remains in HID0, save default
- * values for some CPU specific registers. Called with r24
- * containing CPU number and r3 reloc offset
+ * So nothing for now. Called with r24 containing CPU number and r3
+ * reloc offset
  */
  	.globl	init_idle_power4
 init_idle_power4:
-BEGIN_FTR_SECTION
-	mfspr	r4,SPRN_HID0
-	rlwinm	r4,r4,0,10,8	/* Clear NAP */
-	mtspr	SPRN_HID0, r4
-END_FTR_SECTION_IFSET(CPU_FTR_CAN_NAP)
 	blr
 
 /*
@@ -48,10 +42,9 @@ END_FTR_SECTION_IFSET(CPU_FTR_CAN_NAP)
  */
 	.globl	power4_idle
 power4_idle:
-	/* Check if we can nap or doze, put HID0 mask in r3
-	 */
-	lis	r3, 0
 BEGIN_FTR_SECTION
+	blr
+END_FTR_SECTION_IFCLR(CPU_FTR_CAN_NAP)
 	/* We must dynamically check for the NAP feature as it
 	 * can be cleared by CPU init after the fixups are done
 	 */
@@ -59,16 +52,11 @@ BEGIN_FTR_SECTION
 	lwz	r4,cur_cpu_spec@l(r4)
 	lwz	r4,CPU_SPEC_FEATURES(r4)
 	andi.	r0,r4,CPU_FTR_CAN_NAP
-	beq	1f
+	beqlr
 	/* Now check if user or arch enabled NAP mode */
 	lis	r4,powersave_nap@ha
 	lwz	r4,powersave_nap@l(r4)
 	cmpi	0,r4,0
-	beq	1f
-	lis	r3,HID0_NAP@h
-1:	
-END_FTR_SECTION_IFSET(CPU_FTR_CAN_NAP)
-	cmpi	0,r3,0
 	beqlr
 
 	/* Clear MSR:EE */
@@ -85,18 +73,6 @@ END_FTR_SECTION_IFSET(CPU_FTR_CAN_NAP)
 	blr
 1:	
 	/* Go to NAP now */	
-	mfspr	r4,SPRN_HID0
-	lis	r5,(HID0_NAP|HID0_SLEEP)@h
-	andc	r4,r4,r5
-	or	r4,r4,r3
-	oris	r4,r4,HID0_DPM@h	/* that should be done once for all  */
-	mtspr	SPRN_HID0,r4
-	mfspr	r0,SPRN_HID0
-	mfspr	r0,SPRN_HID0
-	mfspr	r0,SPRN_HID0
-	mfspr	r0,SPRN_HID0
-	mfspr	r0,SPRN_HID0
-	mfspr	r0,SPRN_HID0
 BEGIN_FTR_SECTION
 	DSSALL
 	sync

arch/ppc/kernel/misc.S

@@ -201,7 +201,7 @@ _GLOBAL(call_setup_cpu)
 	mr	r4,r24
 	bctr
 
-#ifdef CONFIG_CPU_FREQ_PMAC
+#if defined(CONFIG_CPU_FREQ_PMAC) && defined(CONFIG_6xx)
 
 /* This gets called by via-pmu.c to switch the PLL selection
  * on 750fx CPU. This function should really be moved to some
@@ -253,7 +253,7 @@ _GLOBAL(low_choose_750fx_pll)
 	mtmsr	r7
 	blr
 
-#endif /* CONFIG_CPU_FREQ_PMAC */
+#endif /* CONFIG_CPU_FREQ_PMAC && CONFIG_6xx */
 
 /* void local_save_flags_ptr(unsigned long *flags) */
 _GLOBAL(local_save_flags_ptr)

arch/ppc/mm/hashtable.S

@@ -417,21 +417,6 @@ _GLOBAL(hash_page_patch_C)
 	lwz	r6,next_slot@l(r4)
 	addi	r6,r6,PTE_SIZE
 	andi.	r6,r6,7*PTE_SIZE
-#ifdef CONFIG_POWER4
-	/*
-	 * Since we don't have BATs on POWER4, we rely on always having
-	 * PTEs in the hash table to map the hash table and the code
-	 * that manipulates it in virtual mode, namely flush_hash_page and
-	 * flush_hash_segments.  Otherwise we can get a DSI inside those
-	 * routines which leads to a deadlock on the hash_table_lock on
-	 * SMP machines.  We avoid this by never overwriting the first
-	 * PTE of each PTEG if it is already valid.
-	 *	-- paulus.
-	 */
-	bne	102f
-	li	r6,PTE_SIZE
-102:
-#endif /* CONFIG_POWER4 */
 	stw	r6,next_slot@l(r4)
 	add	r4,r3,r6
 

arch/ppc/mm/ppc_mmu.c

@@ -83,6 +83,9 @@ unsigned long p_mapped_by_bats(unsigned long pa)
 
 unsigned long __init mmu_mapin_ram(void)
 {
+#ifdef CONFIG_POWER4
+	return 0;
+#else
 	unsigned long tot, bl, done;
 	unsigned long max_size = (256<<20);
 	unsigned long align;
@@ -119,6 +122,7 @@ unsigned long __init mmu_mapin_ram(void)
 	}
 
 	return done;
+#endif
 }
 
 /*
@@ -244,9 +248,10 @@ void __init MMU_init_hw(void)
 	Hash = mem_pieces_find(Hash_size, Hash_size);
 	cacheable_memzero(Hash, Hash_size);
 	_SDR1 = __pa(Hash) | SDR1_LOW_BITS;
-	Hash_end = (PTE *) ((unsigned long)Hash + Hash_size);
 #endif /* CONFIG_POWER4 */
 
+	Hash_end = (PTE *) ((unsigned long)Hash + Hash_size);
+
 	printk("Total memory = %ldMB; using %ldkB for hash table (at %p)\n",
 	       total_memory >> 20, Hash_size >> 10, Hash);
 

include/asm-ppc/reg.h

@@ -91,6 +91,7 @@
 #define SPRN_TBRU	0x10D	/* Time Base Read Upper Register (user, R/O) */
 #define SPRN_TBWL	0x11C	/* Time Base Lower Register (super, R/W) */
 #define SPRN_TBWU	0x11D	/* Time Base Upper Register (super, R/W) */
+#define SPRN_HIOR	0x137	/* 970 Hypervisor interrupt offset */
 #define SPRN_DBAT0L	0x219	/* Data BAT 0 Lower Register */
 #define SPRN_DBAT0U	0x218	/* Data BAT 0 Upper Register */
 #define SPRN_DBAT1L	0x21B	/* Data BAT 1 Lower Register */
@@ -179,7 +180,10 @@
 #define HID1_PC3	(1<<13)		/* 7450 PLL_CFG[3] */
 #define HID1_SYNCBE	(1<<11)		/* 7450 ABE for sync, eieio */
 #define HID1_ABE	(1<<10)		/* 7450 Address Broadcast Enable */
+#define SPRN_HID2	0x3F8		/* Hardware Implementation Register 2 */
 #define SPRN_IABR	0x3F2	/* Instruction Address Breakpoint Register */
+#define SPRN_HID4	0x3F4		/* 970 HID4 */
+#define SPRN_HID5	0x3F6		/* 970 HID5 */
 #if !defined(SPRN_IAC1) && !defined(SPRN_IAC2)
 #define SPRN_IAC1	0x3F4		/* Instruction Address Compare 1 */
 #define SPRN_IAC2	0x3F5		/* Instruction Address Compare 2 */