Merge samba.org:/stuff/paulus/kernel/linux-2.5

into samba.org:/stuff/paulus/kernel/for-linus-ppc

arch/ppc/boot/of1275/map.c

@@ -23,11 +23,10 @@ map(unsigned int phys, unsigned int virt, unsigned int size)
 	char *method;
 	ihandle mmu_ihandle;
 	int misc;
-	unsigned int phys;
-	unsigned int virt;
 	unsigned int size;
+	unsigned int virt;
+	unsigned int phys;
 	int ret0;
-	int ret1;
     } args;
 
     if (of_prom_mmu == 0) {
@@ -36,10 +35,10 @@ map(unsigned int phys, unsigned int virt, unsigned int size)
     }
     args.service = "call-method";
     args.nargs = 6;
-    args.nret = 2;
+    args.nret = 1;
     args.method = "map";
     args.mmu_ihandle = of_prom_mmu;
-    args.misc = -1;
+    args.misc = 0;
     args.phys = phys;
     args.virt = virt;
     args.size = size;

arch/ppc/boot/openfirmware/coffmain.c

@@ -38,9 +38,9 @@ static char heap[SCRATCH_SIZE];
 
 static unsigned long ram_start = 0;
 static unsigned long ram_end = 0x1000000;
-static unsigned long prog_start = 0x800000;
-static unsigned long prog_size = 0x800000;
 
+static unsigned long prog_start = 0x900000;
+static unsigned long prog_size = 0x700000;
 
 typedef void (*kernel_start_t)(int, int, void *);
 

arch/ppc/kernel/head.S

@@ -181,26 +181,8 @@ __after_mmu_off:
 	bl	setup_disp_bat
 #endif
 #else /* CONFIG_POWER4 */
-/*
- * Load up the SDR1 and segment register values now
- * since we don't have the BATs.
- * Also make sure we are running in 32-bit mode.
- */
 	bl	reloc_offset
-	addis	r14,r3,_SDR1@ha		/* get the value from _SDR1 */
-	lwz	r14,_SDR1@l(r14)	/* assume hash table below 4GB */
-	mtspr	SDR1,r14
-	slbia
-	lis	r4,0x2000		/* set pseudo-segment reg 12 */
-	ori	r5,r4,0x0ccc
-	mtsr	12,r5
-	ori	r4,r4,0x0888		/* set pseudo-segment reg 8 */
-	mtsr	8,r4			/* (for access to serial port) */
-	mfmsr	r0
-	clrldi	r0,r0,1
-	sync
-	mtmsr	r0
-	isync
+	bl	initial_mm_power4
 #endif /* CONFIG_POWER4 */
 
 /*
@@ -1637,6 +1619,34 @@ setup_disp_bat:
 #endif /* !defined(CONFIG_APUS) && defined(CONFIG_BOOTX_TEXT) */
 
 #else /* CONFIG_POWER4 */
+/*
+ * Load up the SDR1 and segment register values now
+ * since we don't have the BATs.
+ * Also make sure we are running in 32-bit mode.
+ */
+
+initial_mm_power4:
+	addis	r14,r3,_SDR1@ha		/* get the value from _SDR1 */
+	lwz	r14,_SDR1@l(r14)	/* assume hash table below 4GB */
+	mtspr	SDR1,r14
+	slbia
+	lis	r4,0x2000		/* set pseudo-segment reg 12 */
+	ori	r5,r4,0x0ccc
+	mtsr	12,r5
+	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 */
+	mtsr	9,r5			/* (for access to screen) */
+	mfmsr	r0
+	clrldi	r0,r0,1
+	sync
+	mtmsr	r0
+	isync
+	blr
+
+/*
+ * On 970 (G5), we pre-set a few bits in HID0 & HID1
+ */
 ppc970_setup_hid:
 	li	r0,0
 	sync

arch/ppc/kernel/ppc_ksyms.c

@@ -367,11 +367,17 @@ EXPORT_SYMBOL(next_mmu_context);
 EXPORT_SYMBOL(set_context);
 EXPORT_SYMBOL(handle_mm_fault); /* For MOL */
 EXPORT_SYMBOL_NOVERS(disarm_decr);
+extern long mol_trampoline;
+EXPORT_SYMBOL(mol_trampoline); /* For MOL */
 #ifdef CONFIG_PPC_STD_MMU
 EXPORT_SYMBOL(flush_hash_pages); /* For MOL */
+#ifdef CONFIG_SMP
+extern int mmu_hash_lock;
+EXPORT_SYMBOL(mmu_hash_lock); /* For MOL */
+#endif /* CONFIG_SMP */
 extern long *intercept_table;
 EXPORT_SYMBOL(intercept_table);
-#endif
+#endif /* CONFIG_PPC_STD_MMU */
 EXPORT_SYMBOL(cur_cpu_spec);
 #ifdef CONFIG_PPC_PMAC
 extern unsigned long agp_special_page;

arch/ppc/kernel/process.c

@@ -56,6 +56,7 @@ static struct files_struct init_files = INIT_FILES;
 static struct signal_struct init_signals = INIT_SIGNALS(init_signals);
 static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand);
 struct mm_struct init_mm = INIT_MM(init_mm);
+EXPORT_SYMBOL(init_mm);
 
 /* this is 8kB-aligned so we can get to the thread_info struct
    at the base of it from the stack pointer with 1 integer instruction. */

arch/ppc/mm/hashtable.S

@@ -417,6 +417,21 @@ _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

@@ -211,6 +211,17 @@ void __init MMU_init_hw(void)
 #define MIN_N_HPTEG	1024		/* min 64kB hash table */
 #endif
 
+#ifdef CONFIG_POWER4
+	/* The hash table has already been allocated and initialized
+	   in prom.c */
+	n_hpteg = Hash_size >> LG_HPTEG_SIZE;
+	lg_n_hpteg = __ilog2(n_hpteg);
+
+	/* Remove the hash table from the available memory */
+	if (Hash)
+		reserve_phys_mem(__pa(Hash), Hash_size);
+
+#else /* CONFIG_POWER4 */
 	/*
 	 * Allow 1 HPTE (1/8 HPTEG) for each page of memory.
 	 * This is less than the recommended amount, but then
@@ -224,13 +235,7 @@ void __init MMU_init_hw(void)
 		++lg_n_hpteg;		/* round up if not power of 2 */
 		n_hpteg = 1 << lg_n_hpteg;
 	}
-
 	Hash_size = n_hpteg << LG_HPTEG_SIZE;
-	Hash_mask = n_hpteg - 1;
-	hmask = Hash_mask >> (16 - LG_HPTEG_SIZE);
-	mb2 = mb = 32 - LG_HPTEG_SIZE - lg_n_hpteg;
-	if (lg_n_hpteg > 16)
-		mb2 = 16 - LG_HPTEG_SIZE;
 
 	/*
 	 * Find some memory for the hash table.
@@ -240,6 +245,7 @@ void __init MMU_init_hw(void)
 	cacheable_memzero(Hash, Hash_size);
 	_SDR1 = __pa(Hash) | SDR1_LOW_BITS;
 	Hash_end = (PTE *) ((unsigned long)Hash + Hash_size);
+#endif /* CONFIG_POWER4 */
 
 	printk("Total memory = %ldMB; using %ldkB for hash table (at %p)\n",
 	       total_memory >> 20, Hash_size >> 10, Hash);
@@ -249,6 +255,12 @@ void __init MMU_init_hw(void)
 	 * Patch up the instructions in hashtable.S:create_hpte
 	 */
 	if ( ppc_md.progress ) ppc_md.progress("hash:patch", 0x345);
+	Hash_mask = n_hpteg - 1;
+	hmask = Hash_mask >> (16 - LG_HPTEG_SIZE);
+	mb2 = mb = 32 - LG_HPTEG_SIZE - lg_n_hpteg;
+	if (lg_n_hpteg > 16)
+		mb2 = 16 - LG_HPTEG_SIZE;
+
 	hash_page_patch_A[0] = (hash_page_patch_A[0] & ~0xffff)
 		| ((unsigned int)(Hash) >> 16);
 	hash_page_patch_A[1] = (hash_page_patch_A[1] & ~0x7c0) | (mb << 6);

arch/ppc/platforms/pmac_smp.c

@@ -51,6 +51,7 @@
 #include <asm/time.h>
 #include <asm/open_pic.h>
 #include <asm/cacheflush.h>
+#include <asm/keylargo.h>
 
 /*
  * Powersurge (old powermac SMP) support.
@@ -108,9 +109,16 @@ static volatile u32 *psurge_start;
 /* what sort of powersurge board we have */
 static int psurge_type = PSURGE_NONE;
 
+/* L2 and L3 cache settings to pass from CPU0 to CPU1 */
 volatile static long int core99_l2_cache;
 volatile static long int core99_l3_cache;
 
+/* Timebase freeze GPIO */
+static unsigned int core99_tb_gpio;
+
+/* Sync flag for HW tb sync */
+static volatile int sec_tb_reset = 0;
+
 static void __init
 core99_init_caches(int cpu)
 {
@@ -380,7 +388,6 @@ static void __init smp_psurge_kick_cpu(int nr)
  */
 static void __init psurge_dual_sync_tb(int cpu_nr)
 {
-	static volatile int sec_tb_reset = 0;
 	int t;
 
 	set_dec(tb_ticks_per_jiffy);
@@ -408,8 +415,15 @@ smp_psurge_setup_cpu(int cpu_nr)
 {
 
 	if (cpu_nr == 0) {
-		if (num_online_cpus() < 2)
+		/* If we failed to start the second CPU, we should still
+		 * send it an IPI to start the timebase & DEC or we might
+		 * have them stuck.
+		 */
+		if (num_online_cpus() < 2) {
+			if (psurge_type == PSURGE_DUAL)
+				psurge_set_ipi(1);
 			return;
+		}
 		/* reset the entry point so if we get another intr we won't
 		 * try to startup again */
 		out_be32(psurge_start, 0x100);
@@ -421,19 +435,42 @@ smp_psurge_setup_cpu(int cpu_nr)
 		psurge_dual_sync_tb(cpu_nr);
 }
 
+void __init
+smp_psurge_take_timebase(void)
+{
+	/* Dummy implementation */
+}
+
+void __init
+smp_psurge_give_timebase(void)
+{
+	/* Dummy implementation */
+}
+
 static int __init
 smp_core99_probe(void)
 {
+	extern int powersave_nap;
 	struct device_node *cpus;
 	int i, ncpus = 1;
-	extern int powersave_nap;
+	u32 *tbprop;
 
 	if (ppc_md.progress) ppc_md.progress("smp_core99_probe", 0x345);
 	cpus = find_type_devices("cpu");
-	if (cpus)
-		while ((cpus = cpus->next) != NULL)
-			++ncpus;
+	if (cpus == NULL)
+		return 0;
+
+	tbprop = (u32 *)get_property(cpus, "timebase-enable", NULL);
+	if (tbprop)
+		core99_tb_gpio = *tbprop;
+	else
+		core99_tb_gpio = KL_GPIO_TB_ENABLE;
+
+       	while ((cpus = cpus->next) != NULL)
+	       	++ncpus;
+
 	printk("smp_core99_probe: found %d cpus\n", ncpus);
+
 	if (ncpus > 1) {
 		openpic_request_IPIs();
 		for (i = 1; i < ncpus; ++i)
@@ -517,14 +554,59 @@ smp_core99_setup_cpu(int cpu_nr)
 		if (ppc_md.progress) ppc_md.progress("core99_setup_cpu 0 done", 0x349);
 }
 
+void __init
+smp_core99_take_timebase(void)
+{
+	/* Secondary processor "takes" the timebase by freezing
+	 * it, resetting its local TB and telling CPU 0 to go on
+	 */
+	pmac_call_feature(PMAC_FTR_WRITE_GPIO, NULL, core99_tb_gpio, 4);
+	pmac_call_feature(PMAC_FTR_READ_GPIO, NULL, core99_tb_gpio, 0);
+	mb();
+
+	set_dec(tb_ticks_per_jiffy);
+	set_tb(0, 0);
+	last_jiffy_stamp(smp_processor_id()) = 0;
+
+	mb();
+       	sec_tb_reset = 1;
+}
+
+void __init
+smp_core99_give_timebase(void)
+{
+	unsigned int t;
+
+	/* Primary processor waits for secondary to have frozen
+	 * the timebase, resets local TB, and kick timebase again
+	 */
+	/* wait for the secondary to have reset its TB before proceeding */
+	for (t = 1000; t > 0 && !sec_tb_reset; --t)
+		udelay(1000);
+	if (t == 0)
+		printk(KERN_WARNING "Timeout waiting sync on second CPU\n");
+
+       	set_dec(tb_ticks_per_jiffy);
+	set_tb(0, 0);
+	last_jiffy_stamp(smp_processor_id()) = 0;
+	mb();
+
+	/* Now, restart the timebase by leaving the GPIO to an open collector */
+       	pmac_call_feature(PMAC_FTR_WRITE_GPIO, NULL, core99_tb_gpio, 0);
+        pmac_call_feature(PMAC_FTR_READ_GPIO, NULL, core99_tb_gpio, 0);
+
+	smp_tb_synchronized = 1;
+}
+
+
 /* PowerSurge-style Macs */
 struct smp_ops_t psurge_smp_ops __pmacdata = {
 	.message_pass	= smp_psurge_message_pass,
 	.probe		= smp_psurge_probe,
 	.kick_cpu	= smp_psurge_kick_cpu,
 	.setup_cpu	= smp_psurge_setup_cpu,
-	.give_timebase	= smp_generic_give_timebase,
-	.take_timebase	= smp_generic_take_timebase,
+	.give_timebase	= smp_psurge_give_timebase,
+	.take_timebase	= smp_psurge_take_timebase,
 };
 
 /* Core99 Macs (dual G4s) */
@@ -533,6 +615,6 @@ struct smp_ops_t core99_smp_ops __pmacdata = {
 	.probe		= smp_core99_probe,
 	.kick_cpu	= smp_core99_kick_cpu,
 	.setup_cpu	= smp_core99_setup_cpu,
-	.give_timebase	= smp_generic_give_timebase,
-	.take_timebase	= smp_generic_take_timebase,
+	.give_timebase	= smp_core99_give_timebase,
+	.take_timebase	= smp_core99_take_timebase,
 };

arch/ppc/platforms/pmac_time.c

@@ -166,7 +166,7 @@ via_calibrate_decr(void)
 {
 	struct device_node *vias;
 	volatile unsigned char *via;
-	int count = VIA_TIMER_FREQ_6 / HZ;
+	int count = VIA_TIMER_FREQ_6 / 100;
 	unsigned int dstart, dend;
 
 	vias = find_devices("via-cuda");
@@ -196,7 +196,7 @@ via_calibrate_decr(void)
 		;
 	dend = get_dec();
 
-	tb_ticks_per_jiffy = (dstart - dend) / 6;
+	tb_ticks_per_jiffy = (dstart - dend) / (6 * (HZ/100));
 	tb_to_us = mulhwu_scale_factor(dstart - dend, 60000);
 
 	printk(KERN_INFO "via_calibrate_decr: ticks per jiffy = %u (%u ticks)\n",
@@ -260,7 +260,9 @@ pmac_calibrate_decr(void)
 	 * calibration. That's better since the VIA itself seems
 	 * to be slightly off. --BenH
 	 */
-	if (!machine_is_compatible("MacRISC2"))
+	if (!machine_is_compatible("MacRISC2") &&
+	    !machine_is_compatible("MacRISC3") &&
+	    !machine_is_compatible("MacRISC4"))
 		if (via_calibrate_decr())
 			return;
 

arch/ppc/syslib/of_device.c

@@ -97,9 +97,9 @@ static int of_device_probe(struct device *dev)
 static int of_device_remove(struct device *dev)
 {
 	struct of_device * of_dev = to_of_device(dev);
-	struct of_platform_driver * drv = to_of_platform_driver(of_dev->dev.driver);
+	struct of_platform_driver * drv = to_of_platform_driver(dev->driver);
 
-	if (drv && drv->remove)
+	if (dev->driver && drv->remove)
 		drv->remove(of_dev);
 	return 0;
 }
@@ -107,10 +107,10 @@ static int of_device_remove(struct device *dev)
 static int of_device_suspend(struct device *dev, u32 state)
 {
 	struct of_device * of_dev = to_of_device(dev);
-	struct of_platform_driver * drv = to_of_platform_driver(of_dev->dev.driver);
+	struct of_platform_driver * drv = to_of_platform_driver(dev->driver);
 	int error = 0;
 
-	if (drv && drv->suspend)
+	if (dev->driver && drv->suspend)
 		error = drv->suspend(of_dev, state);
 	return error;
 }
@@ -118,10 +118,10 @@ static int of_device_suspend(struct device *dev, u32 state)
 static int of_device_resume(struct device * dev)
 {
 	struct of_device * of_dev = to_of_device(dev);
-	struct of_platform_driver * drv = to_of_platform_driver(of_dev->dev.driver);
+	struct of_platform_driver * drv = to_of_platform_driver(dev->driver);
 	int error = 0;
 
-	if (drv && drv->resume)
+	if (dev->driver && drv->resume)
 		error = drv->resume(of_dev);
 	return error;
 }

drivers/macintosh/adbhid.c

@@ -611,8 +611,8 @@ adbhid_input_register(int id, int default_id, int original_handler_id,
 		/* HACK WARNING!! This should go away as soon there is an utility
 		 * to control that for event devices.
 		 */
-		adbhid[id]->input.rep[REP_DELAY] = HZ/2;   /* input layer default: HZ/4 */
-		adbhid[id]->input.rep[REP_PERIOD] = HZ/15; /* input layer default: HZ/33 */
+		adbhid[id]->input.rep[REP_DELAY] = 500;   /* input layer default: 250 */
+		adbhid[id]->input.rep[REP_PERIOD] = 66; /* input layer default: 33 */
 	}
 }
 

drivers/video/imsttfb.c

@@ -1461,6 +1461,15 @@ imsttfb_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	unsigned long addr, size;
 	struct imstt_par *par;
 	struct fb_info *info;
+#ifdef CONFIG_PPC_OF
+	struct device_node *dp;
+	
+	dp = pci_device_to_OF_node(pdev);
+	if(dp)
+		printk(KERN_INFO "%s: OF name %s\n",__FUNCTION__, dp->name);
+	else
+		printk(KERN_ERR "imsttfb: no OF node for pci device\n");
+#endif /* CONFIG_PPC_OF */
 
 	size = sizeof(struct fb_info) + sizeof(struct imstt_par) +
 		sizeof(u32) * 16;
@@ -1488,6 +1497,11 @@ imsttfb_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	switch (pdev->device) {
 		case PCI_DEVICE_ID_IMS_TT128: /* IMS,tt128mbA */
 			par->ramdac = IBM;
+#ifdef CONFIG_PPC_OF
+			if (dp && ((strcmp(dp->name, "IMS,tt128mb8") == 0) ||
+				   (strcmp(dp->name, "IMS,tt128mb8A") == 0)))
+				par->ramdac = TVP;
+#endif /* CONFIG_PPC_OF */
 			break;
 		case PCI_DEVICE_ID_IMS_TT3D:  /* IMS,tt3d */
 			par->ramdac = TVP;

include/asm-ppc/reg.h

@@ -87,6 +87,10 @@
 #define SPRN_CTR	0x009	/* Count Register */
 #define SPRN_DABR	0x3F5	/* Data Address Breakpoint Register */
 #define SPRN_DAR	0x013	/* Data Address Register */
+#define SPRN_TBRL	0x10C	/* Time Base Read Lower Register (user, R/O) */
+#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_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 */
@@ -397,6 +401,10 @@
 #define THRM2	SPRN_THRM2	/* Thermal Management Register 2 */
 #define THRM3	SPRN_THRM3	/* Thermal Management Register 3 */
 #define XER	SPRN_XER
+#define TBRL	SPRN_TBRL	/* Time Base Read Lower Register */
+#define TBRU	SPRN_TBRU	/* Time Base Read Upper Register */
+#define TBWL	SPRN_TBWL	/* Time Base Write Lower Register */
+#define TBWU	SPRN_TBWU	/* Time Base Write Upper Register */
 
 /* Processor Version Register */
 

include/asm-ppc/reg_booke.h

@@ -59,14 +59,10 @@ do {						\
 #define SPRN_SPRG5R	0x105	/* Special Purpose Register General 5 Read */
 #define SPRN_SPRG6R	0x106	/* Special Purpose Register General 6 Read */
 #define SPRN_SPRG7R	0x107	/* Special Purpose Register General 7 Read */
-#define SPRN_TBRL	0x10C	/* Time Base Read Lower Register (user, R/O) */
-#define SPRN_TBRU	0x10D	/* Time Base Read Upper Register (user, R/O) */
 #define SPRN_SPRG4W	0x114	/* Special Purpose Register General 4 Write */
 #define SPRN_SPRG5W	0x115	/* Special Purpose Register General 5 Write */
 #define SPRN_SPRG6W	0x116	/* Special Purpose Register General 6 Write */
 #define SPRN_SPRG7W	0x117	/* Special Purpose Register General 7 Write */
-#define SPRN_TBWL	0x11C	/* Time Base Lower Register (super, R/W) */
-#define SPRN_TBWU	0x11D	/* Time Base Upper Register (super, R/W) */
 #define SPRN_DBCR2	0x136	/* Debug Control Register 2 */
 #define SPRN_IAC3	0x13A	/* Instruction Address Compare 3 */
 #define SPRN_IAC4	0x13B	/* Instruction Address Compare 4 */
@@ -252,10 +248,6 @@ do {						\
 #define SPRG5W	SPRN_SPRG5W
 #define SPRG6W	SPRN_SPRG6W
 #define SPRG7W	SPRN_SPRG7W
-#define TBRL	SPRN_TBRL	/* Time Base Read Lower Register */
-#define TBRU	SPRN_TBRU	/* Time Base Read Upper Register */
-#define TBWL	SPRN_TBWL	/* Time Base Write Lower Register */
-#define TBWU	SPRN_TBWU	/* Time Base Write Upper Register */
 
 /*
  * The IBM-403 is an even more odd special case, as it is much