[PATCH] ppc64: monster cleanup

This is the third & hopefully final version of the monster cleanup
patch. It does significant cleanups of the early boot code of the
ppc64 kernel, and begins the long process of cleaning up & splitting
properly the platform support.

It completely reworks the interface between the early code that is
run in the firmware context (prom_init) and the rest of the kernel,
in such a way that will make kexec or static device-tree for embedded
people possible. The early init code can eventually be moved to a
separate link entity, it no longer touches any of the kernel globals,
everything is passed via a single blob of data in memory containing
a flattened version of the device-tree and a memory reserve map.

While doing it, I also cut the ties between pSeries and Powermac. Now,
the kernel config provides a choice between legacy iSeries and
"multiplatform". The later is a set of various supported platform,
each of them beeing a boolean switch, currently defined beeing pSeries
and PowerMac. You can enable both or just one of them. CONFIG_PPC_PSERIES
is now specifically set for IBM pSeries support, you can build a PowerMac
kernel without pSeries support if you which.
The main goal here is to simplify addition of new platform types.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

arch/ppc64/Kconfig

@@ -58,16 +58,27 @@ menu "Platform support"
 
 choice
 	prompt "Platform Type"
-	default PPC_PSERIES
+	default PPC_MULTIPLATFORM
 
 config PPC_ISERIES
-	bool "iSeries"
+	bool "IBM Legacy iSeries"
 
-config PPC_PSERIES
-	bool "pSeries / PowerMac G5"
+config PPC_MULTIPLATFORM
+	bool "Generic"
 
 endchoice
 
+config PPC_PSERIES
+	depends on PPC_MULTIPLATFORM
+	bool "  IBM pSeries & new iSeries"
+	default y
+
+config PPC_PMAC
+	depends on PPC_MULTIPLATFORM
+	bool "  Apple G5 based machines"
+	default y
+	select ADB_PMU
+
 config PPC
 	bool
 	default y
@@ -77,7 +88,7 @@ config PPC64
 	default y
 
 config PPC_OF
-	depends on PPC_PSERIES
+	depends on PPC_MULTIPLATFORM
 	bool
 	default y
 
@@ -85,14 +96,9 @@ config PPC_OF
 # exception vectors for it
 config ALTIVEC
 	bool "Support for VMX (Altivec) vector unit"
-	depends on PPC_PSERIES
+	depends on PPC_MULTIPLATFORM
 	default y
 
-config PPC_PMAC
-	depends on PPC_PSERIES
-	bool "Apple PowerMac G5 support"
-	select ADB_PMU
-
 config PPC_SPLPAR
 	depends on PPC_PSERIES
 	bool "Support for shared-processor logical partitions"
@@ -163,7 +169,7 @@ config SMP
 
 config IRQ_ALL_CPUS
 	bool "Distribute interrupts on all CPUs by default"
-	depends on SMP && PPC_PSERIES
+	depends on SMP && PPC_MULTIPLATFORM
 	help
 	  This option gives the kernel permission to distribute IRQs across
 	  multiple CPUs.  Saying N here will route all IRQs to the first

arch/ppc64/boot/Makefile

@@ -31,6 +31,7 @@ BOOTAFLAGS	:= -D__ASSEMBLY__ $(BOOTCFLAGS) -traditional
 BOOTLD		:= $(CROSS32_COMPILE)ld
 BOOTLFLAGS	:= -Ttext 0x00400000 -e _start -T $(srctree)/$(src)/zImage.lds
 BOOTOBJCOPY	:= $(CROSS32_COMPILE)objcopy
+BOOTSTRIP	:= $(CROSS32_COMPILE)strip
 OBJCOPYFLAGS    := contents,alloc,load,readonly,data
 
 src-boot := crt0.S string.S prom.c main.c zlib.c imagesize.c div64.S
@@ -51,36 +52,31 @@ $(patsubst %.S,%.o, $(filter %.S, $(src-boot))): %.o: %.S
 #-----------------------------------------------------------
 # ELF sections within the zImage bootloader/wrapper
 #-----------------------------------------------------------
-required := vmlinux .config System.map
+required := vmlinux.strip
 initrd   := initrd
 
 obj-sec = $(foreach section, $(1), $(patsubst %,$(obj)/kernel-%.o, $(section)))
 src-sec = $(foreach section, $(1), $(patsubst %,$(obj)/kernel-%.c, $(section)))
 gz-sec  = $(foreach section, $(1), $(patsubst %,$(obj)/kernel-%.gz, $(section)))
 
-hostprogs-y		:= piggy addnote addSystemMap addRamDisk
+hostprogs-y		:= piggy addnote addRamDisk
 targets 		+= zImage zImage.initrd imagesize.c \
 			   $(patsubst $(obj)/%,%, $(call obj-sec, $(required) $(initrd))) \
 			   $(patsubst $(obj)/%,%, $(call src-sec, $(required) $(initrd))) \
 			   $(patsubst $(obj)/%,%, $(call gz-sec, $(required) $(initrd))) \
-			   vmlinux.sm vmlinux.initrd vmlinux.sminitrd
-extra-y			:= sysmap.o initrd.o
-
-quiet_cmd_sysmap = SYSMAP   $@
-      cmd_sysmap = $(obj)/addSystemMap System.map $< $@
-$(obj)/vmlinux.sm: vmlinux $(obj)/addSystemMap System.map FORCE
-	$(call if_changed,sysmap)
+			   vmlinux.initrd
+extra-y			:= initrd.o
 
 quiet_cmd_ramdisk = RAMDISK $@
-      cmd_ramdisk = $(obj)/addRamDisk $(obj)/ramdisk.image.gz System.map $< $@
-$(obj)/vmlinux.initrd: vmlinux $(obj)/addRamDisk $(obj)/ramdisk.image.gz System.map FORCE
-	$(call if_changed,ramdisk)
+      cmd_ramdisk = $(obj)/addRamDisk $(obj)/ramdisk.image.gz $< $@
 
-$(obj)/vmlinux.sminitrd: $(obj)/vmlinux.sm $(obj)/addRamDisk $(obj)/ramdisk.image.gz FORCE
-	$(call if_changed,ramdisk)
+quiet_cmd_stripvm = STRIP $@
+      cmd_stripvm = $(BOOTSTRIP) -s $< -o $@
 
-$(obj)/sysmap.o: System.map $(obj)/piggyback FORCE
-	$(call if_changed,piggy)
+vmlinux.strip: vmlinux FORCE
+	$(call if_changed,stripvm)
+$(obj)/vmlinux.initrd: vmlinux.strip $(obj)/addRamDisk $(obj)/ramdisk.image.gz FORCE
+	$(call if_changed,ramdisk)
 
 addsection = $(BOOTOBJCOPY) $(1) \
 		--add-section=.kernel:$(strip $(patsubst $(obj)/kernel-%.o,%, $(1)))=$(patsubst %.o,%.gz, $(1)) \
@@ -113,9 +109,9 @@ $(obj)/zImage.initrd: obj-boot += $(call obj-sec, $(required) $(initrd))
 $(obj)/zImage.initrd: $(call obj-sec, $(required) $(initrd)) $(obj-boot) $(obj)/addnote FORCE
 	$(call if_changed,addnote)
 
-$(obj)/imagesize.c: vmlinux
+$(obj)/imagesize.c: vmlinux.strip
 	@echo Generating $@
-	ls -l vmlinux | \
+	ls -l vmlinux.strip | \
 	awk '{printf "/* generated -- do not edit! */\n" \
 		"unsigned long vmlinux_filesize = %d;\n", $$5}' > $(obj)/imagesize.c
 	$(CROSS_COMPILE)nm -n vmlinux | tail -n 1 | \
@@ -123,6 +119,6 @@ $(obj)/imagesize.c: vmlinux
 		>> $(obj)/imagesize.c
 
 install: $(CONFIGURE) $(obj)/$(BOOTIMAGE)
-	sh -x $(src)/install.sh "$(KERNELRELEASE)" "$(obj)/$(BOOTIMAGE)" "$(TOPDIR)/System.map" "$(INSTALL_PATH)"
+	sh -x $(src)/install.sh "$(KERNELRELEASE)" "$(obj)/$(BOOTIMAGE)" "$(INSTALL_PATH)"
 
 clean-files := $(patsubst $(obj)/%,%, $(obj-boot))

arch/ppc64/boot/addSystemMap.c

-#include <stdio.h>
-#include <stdlib.h>
-#include <byteswap.h>
-#include <sys/types.h>
-#include <sys/stat.h>
-#include <string.h>
-
-void xlate( char * inb, char * trb, unsigned len )
-{
-	unsigned i;
-	for (  i=0; i<len; ++i )
-	{
-		char c = *inb++;
-		char c1 = c >> 4;
-		char c2 = c & 0xf;
-		if ( c1 > 9 )
-			c1 = c1 + 'A' - 10;
-		else
-			c1 = c1 + '0';
-		if ( c2 > 9 )
-			c2 = c2 + 'A' - 10;
-		else
-			c2 = c2 + '0';
-		*trb++ = c1;
-		*trb++ = c2;
-	}
-	*trb = 0;
-}
-
-#define ElfHeaderSize  (64 * 1024)
-#define ElfPages  (ElfHeaderSize / 4096)
-
-void get4k( /*istream *inf*/FILE *file, char *buf )
-{
-	unsigned j;
-	unsigned num = fread(buf, 1, 4096, file);
-	for ( j=num; j<4096; ++j )
-		buf[j] = 0;
-}
-
-void put4k( /*ostream *outf*/FILE *file, char *buf )
-{
-	fwrite(buf, 1, 4096, file);
-}
-
-int main(int argc, char **argv)
-{
-	char inbuf[4096];
-	FILE *sysmap = NULL;
-	char* ptr_end = NULL; 
-	FILE *inputVmlinux = NULL;
-	FILE *outputVmlinux = NULL;
-	long i = 0;
-	unsigned long sysmapFileLen = 0;
-	unsigned long sysmapLen = 0;
-	unsigned long roundR = 0;
-	unsigned long kernelLen = 0;
-	unsigned long actualKernelLen = 0;
-	unsigned long round = 0;
-	unsigned long roundedKernelLen = 0;
-	unsigned long sysmapStartOffs = 0;
-	unsigned long sysmapPages = 0;
-	unsigned long roundedKernelPages = 0;
-	long padPages = 0;
-	if ( argc < 2 )
-	{
-		fprintf(stderr, "Name of System Map file missing.\n");
-		exit(1);
-	}
-
-	if ( argc < 3 )
-	{
-		fprintf(stderr, "Name of vmlinux file missing.\n");
-		exit(1);
-	}
-
-	if ( argc < 4 )
-	{
-		fprintf(stderr, "Name of vmlinux output file missing.\n");
-		exit(1);
-	}
-
-	sysmap = fopen(argv[1], "r");
-	if ( ! sysmap )
-	{
-		fprintf(stderr, "System Map file \"%s\" failed to open.\n", argv[1]);
-		exit(1);
-	}
-	inputVmlinux = fopen(argv[2], "r");
-	if ( ! inputVmlinux )
-	{
-		fprintf(stderr, "vmlinux file \"%s\" failed to open.\n", argv[2]);
-		exit(1);
-	}
-	outputVmlinux = fopen(argv[3], "w");
-	if ( ! outputVmlinux )
-	{
-		fprintf(stderr, "output vmlinux file \"%s\" failed to open.\n", argv[3]);
-		exit(1);
-	}
-
-
-  
-	fseek(inputVmlinux, 0, SEEK_END);
-	kernelLen = ftell(inputVmlinux);
-	fseek(inputVmlinux, 0, SEEK_SET);
-	printf("kernel file size = %ld\n", kernelLen);
-	if ( kernelLen == 0 )
-	{
-		fprintf(stderr, "You must have a linux kernel specified as argv[2]\n");
-		exit(1);
-	}
-
-
-	actualKernelLen = kernelLen - ElfHeaderSize;
-
-	printf("actual kernel length (minus ELF header) = %ld/%lxx \n", actualKernelLen, actualKernelLen);
-
-	round = actualKernelLen % 4096;
-	roundedKernelLen = actualKernelLen;
-	if ( round )
-		roundedKernelLen += (4096 - round);
-
-	printf("Kernel length rounded up to a 4k multiple = %ld/%lxx \n", roundedKernelLen, roundedKernelLen);
-	roundedKernelPages = roundedKernelLen / 4096;
-	printf("Kernel pages to copy = %ld/%lxx\n", roundedKernelPages, roundedKernelPages);
-
-
-
-	/* Sysmap file */
-	fseek(sysmap, 0, SEEK_END);
-	sysmapFileLen = ftell(sysmap);
-	fseek(sysmap, 0, SEEK_SET);
-	printf("%s file size = %ld\n", argv[1], sysmapFileLen);
-
-	sysmapLen = sysmapFileLen;
-
-	roundR = 4096 - (sysmapLen % 4096);
-	if (roundR)
-	{
-		printf("Rounding System Map file up to a multiple of 4096, adding %ld\n", roundR);
-		sysmapLen += roundR;
-	}
-	printf("Rounded System Map size is %ld\n", sysmapLen);
-  
-  /* Process the Sysmap file to determine the true end of the kernel */
-	sysmapPages = sysmapLen / 4096;
-	printf("System map pages to copy = %ld\n", sysmapPages);
-	/* read the whole file line by line, expect that it doesn't fail */
-	while ( fgets(inbuf, 4096, sysmap) )  ;
-	/* search for _end in the last page of the system map */
-	ptr_end = strstr(inbuf, " _end");
-	if (!ptr_end)
-	{
-		fprintf(stderr, "Unable to find _end in the sysmap file \n");
-		fprintf(stderr, "inbuf: \n");
-		fprintf(stderr, "%s \n", inbuf);
-		exit(1);
-	}
-	printf("Found _end in the last page of the sysmap - backing up 10 characters it looks like %s", ptr_end-10);
-	sysmapStartOffs = (unsigned int)strtol(ptr_end-10, NULL, 16);
-	/* calc how many pages we need to insert between the vmlinux and the start of the sysmap */
-	padPages = sysmapStartOffs/4096 - roundedKernelPages;
-
-	/* Check and see if the vmlinux is larger than _end in System.map */
-	if (padPages < 0)
-	{ /* vmlinux is larger than _end - adjust the offset to start the embedded system map */ 
-		sysmapStartOffs = roundedKernelLen;
-		printf("vmlinux is larger than _end indicates it needs to be - sysmapStartOffs = %lx \n", sysmapStartOffs);
-		padPages = 0;
-		printf("will insert %lx pages between the vmlinux and the start of the sysmap \n", padPages);
-	}
-	else
-	{ /* _end is larger than vmlinux - use the sysmapStartOffs we calculated from the system map */
-		printf("vmlinux is smaller than _end indicates is needed - sysmapStartOffs = %lx \n", sysmapStartOffs);
-		printf("will insert %lx pages between the vmlinux and the start of the sysmap \n", padPages);
-	}
-
-
-
-
-	/* Copy 64K ELF header */
-	for (i=0; i<(ElfPages); ++i)
-	{
-		get4k( inputVmlinux, inbuf );
-		put4k( outputVmlinux, inbuf );
-	}
-
-  
-	/* Copy the vmlinux (as full pages). */
-	fseek(inputVmlinux, ElfHeaderSize, SEEK_SET);
-	for ( i=0; i<roundedKernelPages; ++i )
-	{
-		get4k( inputVmlinux, inbuf );
-    
-		/* Set the offsets (of the start and end) of the embedded sysmap so it is set in the vmlinux.sm */
-		if ( i == 0 )
-		{
-			unsigned long * p;
-			printf("Storing embedded_sysmap_start at 0x3c\n");
-			p = (unsigned long *)(inbuf + 0x3c);
-
-#if (BYTE_ORDER == __BIG_ENDIAN)
-			*p = sysmapStartOffs;
-#else
-			*p = bswap_32(sysmapStartOffs);
-#endif
-
-			printf("Storing embedded_sysmap_end at 0x44\n");
-			p = (unsigned long *)(inbuf + 0x44);
-
-#if (BYTE_ORDER == __BIG_ENDIAN)
-			*p = sysmapStartOffs + sysmapFileLen;
-#else
-			*p = bswap_32(sysmapStartOffs + sysmapFileLen);
-#endif
-		}
-    
-		put4k( outputVmlinux, inbuf );
-	}
-  
-  
-	/* Insert any pad pages between the end of the vmlinux and where the system map needs to be. */
-	for (i=0; i<padPages; ++i)
-	{
-		memset(inbuf, 0, 4096);
-		put4k(outputVmlinux, inbuf);
-	}
-
-
-	/* Copy the system map (as full pages). */
-	fseek(sysmap, 0, SEEK_SET);  /* start reading from begining of the system map */
-	for ( i=0; i<sysmapPages; ++i )
-	{
-		get4k( sysmap, inbuf );
-		put4k( outputVmlinux, inbuf );
-	}
-
-
-	fclose(sysmap);
-	fclose(inputVmlinux);
-	fclose(outputVmlinux);
-	/* Set permission to executable */
-	chmod(argv[3], S_IRUSR|S_IWUSR|S_IXUSR|S_IRGRP|S_IXGRP|S_IROTH|S_IXOTH);
-
-	return 0;
-}
-

arch/ppc64/boot/main.c

@@ -31,13 +31,9 @@ unsigned long strlen(const char *s);
 void *memmove(void *dest, const void *src, unsigned long n);
 void *memcpy(void *dest, const void *src, unsigned long n);
 
-static struct bi_record *make_bi_recs(unsigned long);
-
-#define RAM_START	0x00000000
-#define RAM_END		(64<<20)
-
 /* Value picked to match that used by yaboot */
 #define PROG_START	0x01400000
+#define RAM_END		(256<<20) // Fixme: use OF */
 
 char *avail_ram;
 char *begin_avail, *end_avail;
@@ -48,8 +44,6 @@ unsigned int heap_max;
 extern char _start[];
 extern char _vmlinux_start[];
 extern char _vmlinux_end[];
-extern char _sysmap_start[];
-extern char _sysmap_end[];
 extern char _initrd_start[];
 extern char _initrd_end[];
 extern unsigned long vmlinux_filesize;
@@ -62,7 +56,6 @@ struct addr_range {
 };
 struct addr_range vmlinux = {0, 0, 0};
 struct addr_range vmlinuz = {0, 0, 0};
-struct addr_range sysmap  = {0, 0, 0};
 struct addr_range initrd  = {0, 0, 0};
 
 static char scratch[128<<10];	/* 128kB of scratch space for gunzip */
@@ -70,7 +63,7 @@ static char scratch[128<<10];	/* 128kB of scratch space for gunzip */
 typedef void (*kernel_entry_t)( unsigned long,
                                 unsigned long,
                                 void *,
-				struct bi_record *);
+				void *);
 
 
 int (*prom)(void *);
@@ -80,12 +73,31 @@ void *stdin;
 void *stdout;
 void *stderr;
 
+#define DEBUG
+
+static unsigned long claim_base = PROG_START;
+
+static unsigned long try_claim(unsigned long size)
+{
+	unsigned long addr = 0;
+
+	for(; claim_base < RAM_END; claim_base += 0x100000) {
+#ifdef DEBUG
+		printf("    trying: 0x%08lx\n\r", claim_base);
+#endif
+		addr = (unsigned long)claim(claim_base, size, 0);
+		if ((void *)addr != (void *)-1)
+			break;
+	}
+	if (addr == 0)
+		return 0;
+	claim_base = PAGE_ALIGN(claim_base + size);
+	return addr;
+}
 
-void
-start(unsigned long a1, unsigned long a2, void *promptr)
+void start(unsigned long a1, unsigned long a2, void *promptr)
 {
-	unsigned long i, claim_addr, claim_size;
-	struct bi_record *bi_recs;
+	unsigned long i;
 	kernel_entry_t kernel_entry;
 	Elf64_Ehdr *elf64;
 	Elf64_Phdr *elf64ph;
@@ -102,61 +114,59 @@ start(unsigned long a1, unsigned long a2, void *promptr)
 
 	printf("zImage starting: loaded at 0x%x\n\r", (unsigned)_start);
 
-#if 0
-	sysmap.size = (unsigned long)(_sysmap_end - _sysmap_start);
-	sysmap.memsize = sysmap.size;
-	if ( sysmap.size > 0 ) {
-		sysmap.addr = (RAM_END - sysmap.size) & ~0xFFF;
-		claim(sysmap.addr, RAM_END - sysmap.addr, 0);
-		printf("initial ramdisk moving 0x%lx <- 0x%lx (%lx bytes)\n\r",
-		       sysmap.addr, (unsigned long)_sysmap_start, sysmap.size);
-		memcpy((void *)sysmap.addr, (void *)_sysmap_start, sysmap.size);
+	/*
+	 * Now we try to claim some memory for the kernel itself
+	 * our "vmlinux_memsize" is the memory footprint in RAM, _HOWEVER_, what
+	 * our Makefile stuffs in is an image containing all sort of junk including
+	 * an ELF header. We need to do some calculations here to find the right
+	 * size... In practice we add 1Mb, that is enough, but we should really
+	 * consider fixing the Makefile to put a _raw_ kernel in there !
+	 */
+	vmlinux_memsize += 0x100000;
+	printf("Allocating 0x%lx bytes for kernel ...\n\r", vmlinux_memsize);
+	vmlinux.addr = try_claim(vmlinux_memsize);
+	if (vmlinux.addr == 0) {
+		printf("Can't allocate memory for kernel image !\n\r");
+		exit();
 	}
-#endif
+	vmlinuz.addr = (unsigned long)_vmlinux_start;
+	vmlinuz.size = (unsigned long)(_vmlinux_end - _vmlinux_start);
+	vmlinux.size = PAGE_ALIGN(vmlinux_filesize);
+	vmlinux.memsize = vmlinux_memsize;
 
+	/*
+	 * Now we try to claim memory for the initrd (and copy it there)
+	 */
 	initrd.size = (unsigned long)(_initrd_end - _initrd_start);
 	initrd.memsize = initrd.size;
 	if ( initrd.size > 0 ) {
-		initrd.addr = (RAM_END - initrd.size) & ~0xFFF;
-		a1 = a2 = 0;
-		claim(initrd.addr, RAM_END - initrd.addr, 0);
+		printf("Allocating 0x%lx bytes for initrd ...\n\r", initrd.size);
+		initrd.addr = try_claim(initrd.size);
+		if (initrd.addr == 0) {
+			printf("Can't allocate memory for initial ramdisk !\n\r");
+			exit();
+		}
+		a1 = initrd.addr;
+		a2 = initrd.size;
 		printf("initial ramdisk moving 0x%lx <- 0x%lx (%lx bytes)\n\r",
 		       initrd.addr, (unsigned long)_initrd_start, initrd.size);
-		memcpy((void *)initrd.addr, (void *)_initrd_start, initrd.size);
-	}
-
-	vmlinuz.addr = (unsigned long)_vmlinux_start;
-	vmlinuz.size = (unsigned long)(_vmlinux_end - _vmlinux_start);
-	vmlinux.addr = (unsigned long)(void *)-1;
-	vmlinux.size = PAGE_ALIGN(vmlinux_filesize);
-	vmlinux.memsize = vmlinux_memsize;
-
-	claim_size = vmlinux.memsize /* PPPBBB: + fudge for bi_recs */;
-	for(claim_addr = PROG_START; 
-	    claim_addr <= PROG_START * 8; 
-	    claim_addr += 0x100000) {
-#ifdef DEBUG
-		printf("    trying: 0x%08lx\n\r", claim_addr);
-#endif
-		vmlinux.addr = (unsigned long)claim(claim_addr, claim_size, 0);
-		if ((void *)vmlinux.addr != (void *)-1) break;
-	}
-	if ((void *)vmlinux.addr == (void *)-1) {
-		printf("claim error, can't allocate kernel memory\n\r");
-		exit();
+		memmove((void *)initrd.addr, (void *)_initrd_start, initrd.size);
+		printf("initrd head: 0x%lx\n", *((u32 *)initrd.addr));
 	}
 
-	/* PPPBBB: should kernel always be gziped? */
+	/* Eventually gunzip the kernel */
 	if (*(unsigned short *)vmlinuz.addr == 0x1f8b) {
+		int len;
 		avail_ram = scratch;
 		begin_avail = avail_high = avail_ram;
 		end_avail = scratch + sizeof(scratch);
 		printf("gunzipping (0x%lx <- 0x%lx:0x%0lx)...",
 		       vmlinux.addr, vmlinuz.addr, vmlinuz.addr+vmlinuz.size);
+		len = vmlinuz.size;
 		gunzip((void *)vmlinux.addr, vmlinux.size,
-			(unsigned char *)vmlinuz.addr, (int *)&vmlinuz.size);
-		printf("done %lu bytes\n\r", vmlinuz.size);
-		printf("%u bytes of heap consumed, max in use %u\n\r",
+			(unsigned char *)vmlinuz.addr, &len);
+		printf("done 0x%lx bytes\n\r", len);
+		printf("0x%x bytes of heap consumed, max in use 0x%\n\r",
 		       (unsigned)(avail_high - begin_avail), heap_max);
 	} else {
 		memmove((void *)vmlinux.addr,(void *)vmlinuz.addr,vmlinuz.size);
@@ -192,7 +202,8 @@ start(unsigned long a1, unsigned long a2, void *promptr)
 
 	flush_cache((void *)vmlinux.addr, vmlinux.memsize);
 
-	bi_recs = make_bi_recs(vmlinux.addr + vmlinux.memsize);
+	if (a1)
+		printf("initrd head: 0x%lx\n\r", *((u32 *)initrd.addr));
 
 	kernel_entry = (kernel_entry_t)vmlinux.addr;
 #ifdef DEBUG
@@ -203,65 +214,16 @@ start(unsigned long a1, unsigned long a2, void *promptr)
 		"        prom       = 0x%lx,\n\r"
 		"        bi_recs    = 0x%lx,\n\r",
 		(unsigned long)kernel_entry, a1, a2,
-		(unsigned long)prom, (unsigned long)bi_recs);
+		(unsigned long)prom, NULL);
 #endif
 
-	kernel_entry( a1, a2, prom, bi_recs );
+	kernel_entry( a1, a2, prom, NULL );
 
 	printf("Error: Linux kernel returned to zImage bootloader!\n\r");
 
 	exit();
 }
 
-static struct bi_record *
-make_bi_recs(unsigned long addr)
-{
-	struct bi_record *bi_recs;
-	struct bi_record *rec;
-
-	bi_recs = rec = bi_rec_init(addr);
-
-	rec = bi_rec_alloc(rec, 2);
-	rec->tag = BI_FIRST;
-	/* rec->data[0] = ...;	# Written below before return */
-	/* rec->data[1] = ...;	# Written below before return */
-
-	rec = bi_rec_alloc_bytes(rec, strlen("chrpboot")+1);
-	rec->tag = BI_BOOTLOADER_ID;
-	sprintf( (char *)rec->data, "chrpboot");
-
-	rec = bi_rec_alloc(rec, 2);
-	rec->tag = BI_MACHTYPE;
-	rec->data[0] = PLATFORM_PSERIES;
-	rec->data[1] = 1;
-
-	if ( initrd.size > 0 ) {
-		rec = bi_rec_alloc(rec, 2);
-		rec->tag = BI_INITRD;
-		rec->data[0] = initrd.addr;
-		rec->data[1] = initrd.size;
-	}
-
-	if ( sysmap.size > 0 ) {
-		rec = bi_rec_alloc(rec, 2);
-		rec->tag = BI_SYSMAP;
-		rec->data[0] = (unsigned long)sysmap.addr;
-		rec->data[1] = (unsigned long)sysmap.size;
-	}
-
-	rec = bi_rec_alloc(rec, 1);
-	rec->tag = BI_LAST;
-	rec->data[0] = (bi_rec_field)bi_recs;
-
-	/* Save the _end_ address of the bi_rec's in the first bi_rec
-	 * data field for easy access by the kernel.
-	 */
-	bi_recs->data[0] = (bi_rec_field)rec;
-	bi_recs->data[1] = (bi_rec_field)rec + rec->size - (bi_rec_field)bi_recs;
-
-	return bi_recs;
-}
-
 struct memchunk {
 	unsigned int size;
 	unsigned int pad;

arch/ppc64/boot/zImage.lds

@@ -61,19 +61,9 @@ SECTIONS
 
   . = ALIGN(4096);
   _vmlinux_start =  .;
-  .kernel:vmlinux : { *(.kernel:vmlinux) }
+  .kernel:vmlinux.strip : { *(.kernel:vmlinux.strip) }
   _vmlinux_end =  .;
 
-  . = ALIGN(4096);
-  _dotconfig_start =  .;
-  .kernel:.config : { *(.kernel:.config) }
-  _dotconfig_end =  .;
-
-  . = ALIGN(4096);
-  _sysmap_start =  .;
-  .kernel:System.map : { *(.kernel:System.map) }
-  _sysmap_end =  .;
-
   . = ALIGN(4096);
   _initrd_start =  .;
   .kernel:initrd : { *(.kernel:initrd) }

arch/ppc64/configs/pSeries_defconfig

 #
 # Automatically generated make config: don't edit
+# Linux kernel version: 2.6.9-rc2
+# Thu Sep 23 16:45:05 2004
 #
 CONFIG_64BIT=y
 CONFIG_MMU=y
@@ -20,6 +22,7 @@ CONFIG_CLEAN_COMPILE=y
 #
 # General setup
 #
+CONFIG_LOCALVERSION=""
 CONFIG_SWAP=y
 CONFIG_SYSVIPC=y
 CONFIG_POSIX_MQUEUE=y
@@ -61,12 +64,13 @@ CONFIG_SYSVIPC_COMPAT=y
 # Platform support
 #
 # CONFIG_PPC_ISERIES is not set
+CONFIG_PPC_MULTIPLATFORM=y
 CONFIG_PPC_PSERIES=y
+# CONFIG_PPC_PMAC is not set
 CONFIG_PPC=y
 CONFIG_PPC64=y
 CONFIG_PPC_OF=y
 CONFIG_ALTIVEC=y
-# CONFIG_PPC_PMAC is not set
 CONFIG_PPC_SPLPAR=y
 # CONFIG_BOOTX_TEXT is not set
 # CONFIG_POWER4_ONLY is not set
@@ -189,7 +193,6 @@ CONFIG_BLK_DEV_IDEDMA_PCI=y
 # CONFIG_BLK_DEV_IDEDMA_FORCED is not set
 CONFIG_IDEDMA_PCI_AUTO=y
 # CONFIG_IDEDMA_ONLYDISK is not set
-CONFIG_BLK_DEV_ADMA=y
 # CONFIG_BLK_DEV_AEC62XX is not set
 # CONFIG_BLK_DEV_ALI15X3 is not set
 CONFIG_BLK_DEV_AMD74XX=y
@@ -365,6 +368,8 @@ CONFIG_NETFILTER=y
 # IP: Netfilter Configuration
 #
 CONFIG_IP_NF_CONNTRACK=m
+# CONFIG_IP_NF_CT_ACCT is not set
+CONFIG_IP_NF_CT_PROTO_SCTP=m
 CONFIG_IP_NF_FTP=m
 CONFIG_IP_NF_IRC=m
 CONFIG_IP_NF_TFTP=m
@@ -389,8 +394,14 @@ CONFIG_IP_NF_MATCH_HELPER=m
 CONFIG_IP_NF_MATCH_STATE=m
 CONFIG_IP_NF_MATCH_CONNTRACK=m
 CONFIG_IP_NF_MATCH_OWNER=m
+CONFIG_IP_NF_MATCH_ADDRTYPE=m
+CONFIG_IP_NF_MATCH_REALM=m
+CONFIG_IP_NF_MATCH_SCTP=m
 CONFIG_IP_NF_FILTER=m
 CONFIG_IP_NF_TARGET_REJECT=m
+CONFIG_IP_NF_TARGET_LOG=m
+CONFIG_IP_NF_TARGET_ULOG=m
+CONFIG_IP_NF_TARGET_TCPMSS=m
 CONFIG_IP_NF_NAT=m
 CONFIG_IP_NF_NAT_NEEDED=y
 CONFIG_IP_NF_TARGET_MASQUERADE=m
@@ -409,21 +420,13 @@ CONFIG_IP_NF_TARGET_ECN=m
 CONFIG_IP_NF_TARGET_DSCP=m
 CONFIG_IP_NF_TARGET_MARK=m
 CONFIG_IP_NF_TARGET_CLASSIFY=m
-CONFIG_IP_NF_TARGET_LOG=m
-CONFIG_IP_NF_TARGET_ULOG=m
-CONFIG_IP_NF_TARGET_TCPMSS=m
+CONFIG_IP_NF_RAW=m
+CONFIG_IP_NF_TARGET_NOTRACK=m
 CONFIG_IP_NF_ARPTABLES=m
 CONFIG_IP_NF_ARPFILTER=m
 CONFIG_IP_NF_ARP_MANGLE=m
 CONFIG_IP_NF_COMPAT_IPCHAINS=m
 CONFIG_IP_NF_COMPAT_IPFWADM=m
-CONFIG_IP_NF_TARGET_NOTRACK=m
-CONFIG_IP_NF_RAW=m
-CONFIG_IP_NF_MATCH_ADDRTYPE=m
-CONFIG_IP_NF_MATCH_REALM=m
-# CONFIG_IP_NF_CT_ACCT is not set
-CONFIG_IP_NF_MATCH_SCTP=m
-CONFIG_IP_NF_CT_PROTO_SCTP=m
 CONFIG_XFRM=y
 CONFIG_XFRM_USER=m
 
@@ -604,6 +607,7 @@ CONFIG_SERIO_I8042=y
 # CONFIG_SERIO_SERPORT is not set
 # CONFIG_SERIO_CT82C710 is not set
 # CONFIG_SERIO_PCIPS2 is not set
+# CONFIG_SERIO_RAW is not set
 
 #
 # Input Device Drivers
@@ -621,6 +625,7 @@ CONFIG_MOUSE_PS2=y
 # CONFIG_INPUT_JOYSTICK is not set
 # CONFIG_INPUT_TOUCHSCREEN is not set
 CONFIG_INPUT_MISC=y
+# CONFIG_INPUT_PCSPKR is not set
 # CONFIG_INPUT_UINPUT is not set
 
 #

arch/ppc64/kernel/Makefile

@@ -17,7 +17,7 @@ obj-$(CONFIG_PPC_OF) +=	of_device.o
 
 pci-obj-$(CONFIG_PPC_ISERIES)	+= iSeries_pci.o iSeries_pci_reset.o \
 				     iSeries_IoMmTable.o
-pci-obj-$(CONFIG_PPC_PSERIES)	+= pci_dn.o pci_dma_direct.o
+pci-obj-$(CONFIG_PPC_MULTIPLATFORM)	+= pci_dn.o pci_dma_direct.o
 
 obj-$(CONFIG_PCI)	+= pci.o pci_iommu.o $(pci-obj-y)
 
@@ -28,10 +28,11 @@ obj-$(CONFIG_PPC_ISERIES) += iSeries_irq.o \
 			     mf.o HvLpEvent.o iSeries_proc.o iSeries_htab.o \
 			     iSeries_iommu.o
 
+obj-$(CONFIG_PPC_MULTIPLATFORM) += nvram.o open_pic.o i8259.o prom_init.o prom.o
+
 obj-$(CONFIG_PPC_PSERIES) += pSeries_pci.o pSeries_lpar.o pSeries_hvCall.o \
-			     eeh.o nvram.o pSeries_nvram.o rtasd.o ras.o \
-			     open_pic.o xics.o pSeries_htab.o rtas.o \
-			     chrp_setup.o i8259.o prom.o pSeries_iommu.o
+			     eeh.o pSeries_nvram.o rtasd.o ras.o \
+			     xics.o rtas.o pSeries_setup.o pSeries_iommu.o
 
 obj-$(CONFIG_PROC_FS)		+= proc_ppc64.o
 obj-$(CONFIG_RTAS_FLASH)	+= rtas_flash.o

arch/ppc64/kernel/asm-offsets.c

@@ -33,7 +33,6 @@
 #include <asm/iSeries/ItLpPaca.h>
 #include <asm/iSeries/ItLpQueue.h>
 #include <asm/iSeries/HvLpEvent.h>
-#include <asm/prom.h>
 #include <asm/rtas.h>
 #include <asm/cputable.h>
 
@@ -108,7 +107,6 @@ int main(void)
         DEFINE(LPPACASRR1, offsetof(struct ItLpPaca, xSavedSrr1));
 	DEFINE(LPPACAANYINT, offsetof(struct ItLpPaca, xIntDword.xAnyInt));
 	DEFINE(LPPACADECRINT, offsetof(struct ItLpPaca, xIntDword.xFields.xDecrInt));
-	DEFINE(PROMENTRY, offsetof(struct prom_t, entry));
 
 	/* RTAS */
 	DEFINE(RTASBASE, offsetof(struct rtas_t, base));

arch/ppc64/kernel/entry.S

@@ -746,6 +746,10 @@ _STATIC(rtas_restore_regs)
 	mtlr    r0
         blr				/* return to caller */
 
+#endif /* CONFIG_PPC_PSERIES */
+
+#ifdef CONFIG_PPC_MULTIPLATFORM
+
 _GLOBAL(enter_prom)
 	mflr	r0
 	std	r0,16(r1)
@@ -777,11 +781,8 @@ _GLOBAL(enter_prom)
 	std	r11,_MSR(r1)
 
 	/* Get the PROM entrypoint */
-        bl      .reloc_offset
-	LOADADDR(r12,prom)
-	sub	r12,r12,r3
-	ld	r12,PROMENTRY(r12)
-	mtlr	r12
+	ld	r0,GPR4(r1)
+	mtlr	r0
 
 	/* Switch MSR to 32 bits mode
 	 */
@@ -834,4 +835,4 @@ _GLOBAL(enter_prom)
 	mtlr    r0
         blr
 	
-#endif	/* defined(CONFIG_PPC_PSERIES) */
+#endif	/* CONFIG_PPC_MULTIPLATFORM */

arch/ppc64/kernel/head.S

@@ -83,13 +83,14 @@
 	.text
 	.globl  _stext
 _stext:
-#ifdef CONFIG_PPC_PSERIES
-_STATIC(__start)
+#ifdef CONFIG_PPC_MULTIPLATFORM
+_GLOBAL(__start)
 	/* NOP this out unconditionally */
 BEGIN_FTR_SECTION
-	b .__start_initialization_pSeries
+	b .__start_initialization_multiplatform
 END_FTR_SECTION(0, 1)
-#endif
+#endif /* CONFIG_PPC_MULTIPLATFORM */
+
 	/* Catch branch to 0 in real mode */
 	trap
 #ifdef CONFIG_PPC_ISERIES
@@ -117,7 +118,8 @@ embedded_sysmap_start:
 	.globl	embedded_sysmap_end
 embedded_sysmap_end:
 	.llong	0
-#else
+
+#else /* CONFIG_PPC_ISERIES */
 
 	/* Secondary processors spin on this value until it goes to 1. */
 	.globl  __secondary_hold_spinloop
@@ -1223,12 +1225,10 @@ _GLOBAL(pseries_secondary_smp_init)
 #endif
 	b 	1b			/* Loop until told to go	 */
 #ifdef CONFIG_PPC_ISERIES
-_GLOBAL(__start_initialization_iSeries)
+_STATIC(__start_initialization_iSeries)
 	/* Clear out the BSS */
 	LOADADDR(r11,__bss_stop)
-
 	LOADADDR(r8,__bss_start)
-
 	sub	r11,r11,r8		/* bss size			*/
 	addi	r11,r11,7		/* round up to an even double word */
 	rldicl. r11,r11,61,3		/* shift right by 3		*/
@@ -1265,32 +1265,73 @@ _GLOBAL(__start_initialization_iSeries)
 	ld	r6,PACA(r4)		/* Get the base paca pointer	*/
 	ld	r4,PACASTABVIRT(r6)
 
-	bl	.iSeries_fixup_klimit
+	bl	.iSeries_early_setup
 
 	/* relocation is on at this point */
 
 	b	.start_here_common
-#endif
+#endif /* CONFIG_PPC_ISERIES */
 
-#ifdef CONFIG_PPC_PSERIES
+#ifdef CONFIG_PPC_MULTIPLATFORM
 
-_STATIC(mmu_off)
+_STATIC(__mmu_off)
 	mfmsr	r3
 	andi.	r0,r3,MSR_IR|MSR_DR
 	beqlr
 	andc	r3,r3,r0
-	mtspr	SRR0,r4
-	mtspr	SRR1,r3
+	mtspr	SPRN_SRR0,r4
+	mtspr	SPRN_SRR1,r3
 	sync
 	rfid
 	b	.	/* prevent speculative execution */
-_GLOBAL(__start_initialization_pSeries)
-	mr	r31,r3			/* save parameters */
+
+
+/*
+ * Here is our main kernel entry point. We support currently 2 kind of entries
+ * depending on the value of r5.
+ *
+ *   r5 != NULL -> OF entry, we go to prom_init, "legacy" parameter content
+ *                 in r3...r7
+ *   
+ *   r5 == NULL -> kexec style entry. r3 is a physical pointer to the
+ *                 DT block, r4 is a physical pointer to the kernel itself
+ *
+ */
+_GLOBAL(__start_initialization_multiplatform)
+	/*
+	 * Are we booted from a PROM Of-type client-interface ?
+	 */
+	cmpldi	cr0,r5,0
+	bne	.__boot_from_prom		/* yes -> prom */
+
+	/* Save parameters */
+	mr	r31,r3
+	mr	r30,r4
+
+	/* Make sure we are running in 64 bits mode */
+	bl	.enable_64b_mode
+
+	/* Setup some critical 970 SPRs before switching MMU off */
+	bl	.__970_cpu_preinit
+
+	/* cpu # */
+	li	r24,0
+
+	/* Switch off MMU if not already */
+	LOADADDR(r4, .__after_prom_start - KERNELBASE)
+	add	r4,r4,r30
+	bl	.__mmu_off
+	b	.__after_prom_start
+
+_STATIC(__boot_from_prom)
+	/* Save parameters */
+	mr	r31,r3
 	mr	r30,r4
 	mr	r29,r5
 	mr	r28,r6
 	mr	r27,r7
 
+	/* Make sure we are running in 64 bits mode */
 	bl	.enable_64b_mode
 
 	/* put a relocation offset into r3 */
@@ -1303,7 +1344,7 @@ _GLOBAL(__start_initialization_pSeries)
 	/* Relocate the TOC from a virt addr to a real addr */
 	sub	r2,r2,r3
 
-	/* Save parameters */
+	/* Restore parameters */
 	mr	r3,r31
 	mr	r4,r30
 	mr	r5,r29
@@ -1312,17 +1353,8 @@ _GLOBAL(__start_initialization_pSeries)
 
 	/* Do all of the interaction with OF client interface */
 	bl	.prom_init
-	mr	r23,r3			/* Save phys address we are running at */
-
-	/* Setup some critical 970 SPRs before switching MMU off */
-	bl	.__970_cpu_preinit
-
-	li	r24,0			/* cpu # */
-
-	/* Switch off MMU if not already */
-	LOADADDR(r4, .__after_prom_start - KERNELBASE)
-	add	r4,r4,r23
-	bl	.mmu_off	
+	/* We never return */
+	trap
 
 /*
  * At this point, r3 contains the physical address we are running at,
@@ -1348,7 +1380,7 @@ _STATIC(__after_prom_start)
 
 	li	r3,0			/* target addr */
 
-	// XXX FIXME: Use phys returned by OF (r23)
+	// XXX FIXME: Use phys returned by OF (r30)
 	sub	r4,r27,r26 		/* source addr			 */
 					/* current address of _start	 */
 					/*   i.e. where we are running	 */
@@ -1373,8 +1405,9 @@ _STATIC(__after_prom_start)
 	ld	r5,0(r5)		/* get the value of klimit */
 	sub	r5,r5,r27
 	bl	.copy_and_flush		/* copy the rest */
-	b	.start_here_pSeries
-#endif
+	b	.start_here_multiplatform
+
+#endif /* CONFIG_PPC_MULTIPLATFORM */
 
 /*
  * Copy routine used to copy the kernel to start at physical address 0
@@ -1794,15 +1827,33 @@ _GLOBAL(enable_64b_mode)
 	isync
 	blr
 
-#ifdef CONFIG_PPC_PSERIES
+#ifdef CONFIG_PPC_MULTIPLATFORM
 /*
  * This is where the main kernel code starts.
  */
-_STATIC(start_here_pSeries)
+_STATIC(start_here_multiplatform)
 	/* get a new offset, now that the kernel has moved. */
 	bl	.reloc_offset
 	mr	r26,r3
 
+	/* Clear out the BSS. It may have been done in prom_init,
+	 * already but that's irrelevant since prom_init will soon
+	 * be detached from the kernel completely. Besides, we need
+	 * to clear it now for kexec-style entry.
+	 */
+	LOADADDR(r11,__bss_stop)
+	LOADADDR(r8,__bss_start)
+	sub	r11,r11,r8		/* bss size			*/
+	addi	r11,r11,7		/* round up to an even double word */
+	rldicl. r11,r11,61,3		/* shift right by 3		*/
+	beq	4f
+	addi	r8,r8,-8
+	li	r0,0
+	mtctr	r11			/* zero this many doublewords	*/
+3:	stdu	r0,8(r8)
+	bdnz	3b
+4:
+
 	mfmsr	r6
 	ori	r6,r6,MSR_RI
 	mtmsrd	r6			/* RI on */
@@ -1875,8 +1926,11 @@ _STATIC(start_here_pSeries)
 	mr	r5,r26
 	bl	.identify_cpu
 
-	/* Get the pointer to the segment table which is used by		*/
-	/* stab_initialize						 */
+	/* Setup a valid physical PACA pointer in SPRG3 for early_setup
+	 * note that boot_cpuid can always be 0 nowadays since there is
+	 * nowhere it can be initialized differently before we reach this
+	 * code
+	 */
 	LOADADDR(r27, boot_cpuid)
 	sub	r27,r27,r26
 	lwz	r27,0(r27)
@@ -1885,12 +1939,18 @@ _STATIC(start_here_pSeries)
 	mulli	r13,r27,PACA_SIZE	/* Calculate vaddr of right paca */
 	add	r13,r13,r24		/* for this processor.		 */
 	sub	r13,r13,r26		/* convert to physical addr	 */
-
 	mtspr	SPRG3,r13		/* PPPBBB: Temp... -Peter */
-	ld	r3,PACASTABREAL(r13)
-	ori	r4,r3,1			/* turn on valid bit		 */
 	
+	/* Do very early kernel initializations, including initial hash table,
+	 * stab and slb setup before we turn on relocation.	*/
+
+	/* Restore parameters passed from prom_init/kexec */
+	mr	r3,r31
+ 	bl	.early_setup
+
 	/* set the ASR */
+	ld	r3,PACASTABREAL(r13)
+	ori	r4,r3,1			/* turn on valid bit		 */
 	li	r3,SYSTEMCFG_PHYS_ADDR	/* r3 = ptr to systemcfg */
 	lwz	r3,PLATFORM(r3)		/* r3 = platform flags */
 	cmpldi 	r3,PLATFORM_PSERIES_LPAR
@@ -1909,13 +1969,7 @@ _STATIC(start_here_pSeries)
 98:					/* !(rpa hypervisor) || !(star) */
 	mtasr	r4			/* set the stab location	*/
 99:
-	mfspr	r6,SPRG3
-	ld	r3,PACASTABREAL(r6)	/* restore r3 for stab_initialize */
-
-	/* Initialize an initial memory mapping and turn on relocation.	*/
-	bl	.stab_initialize
-	bl	.htab_initialize
-
+	/* Set SDR1 (hash table pointer) */
 	li	r3,SYSTEMCFG_PHYS_ADDR	/* r3 = ptr to systemcfg */
 	lwz	r3,PLATFORM(r3)		/* r3 = platform flags */
 	/* Test if bit 0 is set (LPAR bit) */
@@ -1932,7 +1986,7 @@ _STATIC(start_here_pSeries)
 	mtspr	SRR1,r4
 	rfid
 	b	.	/* prevent speculative execution */
-#endif /* CONFIG_PPC_PSERIES */
+#endif /* CONFIG_PPC_MULTIPLATFORM */
 	
 	/* This is where all platforms converge execution */
 _STATIC(start_here_common)
@@ -1980,13 +2034,6 @@ _STATIC(start_here_common)
 	ld	r2,PACATOC(r13)
 	std	r1,PACAKSAVE(r13)
 
-	/* Restore the parms passed in from the bootloader. */
-	mr	r3,r31
-	mr	r4,r30
-	mr	r5,r29
-	mr	r6,r28
-	mr	r7,r27
-
 	bl	.setup_system
 
 	/* Load up the kernel context */

arch/ppc64/kernel/iSeries_htab.c

@@ -233,4 +233,6 @@ void hpte_init_iSeries(void)
 	ppc_md.hpte_updateboltedpp = iSeries_hpte_updateboltedpp;
 	ppc_md.hpte_insert	= iSeries_hpte_insert;
 	ppc_md.hpte_remove     	= iSeries_hpte_remove;
+
+	htab_finish_init();
 }

arch/ppc64/kernel/iSeries_setup.c

@@ -16,6 +16,8 @@
  *      2 of the License, or (at your option) any later version.
  */
  
+#undef DEBUG
+
 #include <linux/config.h>
 #include <linux/init.h>
 #include <linux/threads.h>
@@ -54,6 +56,14 @@
 #include <asm/iSeries/iSeries_proc.h>
 #include <asm/iSeries/mf.h>
 
+extern void hvlog(char *fmt, ...);
+
+#ifdef DEBUG
+#define DBG(fmt...) hvlog(fmt)
+#else
+#define DBG(fmt...)
+#endif
+
 /* Function Prototypes */
 extern void ppcdbg_initialize(void);
 extern void tce_init_iSeries(void);
@@ -75,8 +85,6 @@ static unsigned long tbFreqMhzHundreths;
 
 int piranha_simulator;
 
-int boot_cpuid;
-
 extern int rd_size;		/* Defined in drivers/block/rd.c */
 extern unsigned long klimit;
 extern unsigned long embedded_sysmap_start;
@@ -275,8 +283,28 @@ unsigned long iSeries_process_mainstore_vpd(struct MemoryBlock *mb_array,
 	return mem_blocks;
 }
 
-void __init iSeries_init_early(void)
+static void __init iSeries_parse_cmdline(void)
 {
+	char *p, *q;
+
+	/* copy the command line parameter from the primary VSP  */
+	HvCallEvent_dmaToSp(cmd_line, 2 * 64* 1024, 256,
+			HvLpDma_Direction_RemoteToLocal);
+
+	p = cmd_line;
+	q = cmd_line + 255;
+	while(p < q) {
+		if (!*p || *p == '\n')
+			break;
+		++p;
+	}
+	*p = 0;
+}
+
+/*static*/ void __init iSeries_init_early(void)
+{
+	DBG(" -> iSeries_init_early()\n");
+
 	ppcdbg_initialize();
 
 #if defined(CONFIG_BLK_DEV_INITRD)
@@ -300,25 +328,20 @@ void __init iSeries_init_early(void)
 	iSeries_recal_tb = get_tb();
 	iSeries_recal_titan = HvCallXm_loadTod();
 
-	ppc_md.setup_arch = iSeries_setup_arch;
-	ppc_md.get_cpuinfo = iSeries_get_cpuinfo;
-	ppc_md.init_IRQ = iSeries_init_IRQ;
-	ppc_md.get_irq = iSeries_get_irq;
-	ppc_md.init = NULL;
-
-	ppc_md.pcibios_fixup  = iSeries_pci_final_fixup;
-
-	ppc_md.restart = iSeries_restart;
-	ppc_md.power_off = iSeries_power_off;
-	ppc_md.halt = iSeries_halt;
-
-	ppc_md.get_boot_time = iSeries_get_boot_time;
-	ppc_md.set_rtc_time = iSeries_set_rtc_time;
-	ppc_md.get_rtc_time = iSeries_get_rtc_time;
-	ppc_md.calibrate_decr = iSeries_calibrate_decr;
-	ppc_md.progress = iSeries_progress;
+	/*
+	 * Cache sizes must be initialized before hpte_init_iSeries is called
+	 * as the later need them for flush_icache_range()
+	 */
+	setup_iSeries_cache_sizes();
 
+	/*
+	 * Initialize the hash table management pointers
+	 */
 	hpte_init_iSeries();
+
+	/*
+	 * Initialize the DMA/TCE management
+	 */
 	tce_init_iSeries();
 
 	/*
@@ -326,7 +349,7 @@ void __init iSeries_init_early(void)
 	 * AS/400 absolute addresses
 	 */
 	build_iSeries_Memory_Map();
-	setup_iSeries_cache_sizes();
+
 	/* Initialize machine-dependency vectors */
 #ifdef CONFIG_SMP
 	smp_init_iSeries();
@@ -340,24 +363,22 @@ void __init iSeries_init_early(void)
 	mf_init();
 	mf_initialized = 1;
 	mb();
-}
 
-void __init iSeries_parse_cmdline(void)
-{
-	char *p, *q;
+	/* If we were passed an initrd, set the ROOT_DEV properly if the values
+	 * look sensible. If not, clear initrd reference.
+	 */
+#ifdef CONFIG_BLK_DEV_INITRD
+	if (initrd_start >= KERNELBASE && initrd_end >= KERNELBASE &&
+	    initrd_end > initrd_start)
+		ROOT_DEV = Root_RAM0;
+	else
+		initrd_start = initrd_end = 0;
+#endif /* CONFIG_BLK_DEV_INITRD */
 
-	/* copy the command line parameter from the primary VSP  */
-	HvCallEvent_dmaToSp(cmd_line, 2 * 64* 1024, 256,
-			HvLpDma_Direction_RemoteToLocal);
 
-	p = cmd_line;
-	q = cmd_line + 255;
-	while(p < q) {
-		if (!*p || *p == '\n')
-			break;
-		++p;
-	}
-	*p = 0;
+	iSeries_parse_cmdline();
+
+	DBG(" <- iSeries_init_early()\n");
 }
 
 /*
@@ -781,7 +802,7 @@ void __init iSeries_progress(char * st, unsigned short code)
 	}
 }
 
-void iSeries_fixup_klimit(void)
+static void __init iSeries_fixup_klimit(void)
 {
 	/*
 	 * Change klimit to take into account any ram disk
@@ -810,3 +831,27 @@ int __init iSeries_src_init(void)
 }
 
 late_initcall(iSeries_src_init);
+
+void __init iSeries_early_setup(void)
+{
+	iSeries_fixup_klimit();
+
+	ppc_md.setup_arch = iSeries_setup_arch;
+	ppc_md.get_cpuinfo = iSeries_get_cpuinfo;
+	ppc_md.init_IRQ = iSeries_init_IRQ;
+	ppc_md.get_irq = iSeries_get_irq;
+	ppc_md.init_early = iSeries_init_early,
+
+	ppc_md.pcibios_fixup  = iSeries_pci_final_fixup;
+
+	ppc_md.restart = iSeries_restart;
+	ppc_md.power_off = iSeries_power_off;
+	ppc_md.halt = iSeries_halt;
+
+	ppc_md.get_boot_time = iSeries_get_boot_time;
+	ppc_md.set_rtc_time = iSeries_set_rtc_time;
+	ppc_md.get_rtc_time = iSeries_get_rtc_time;
+	ppc_md.calibrate_decr = iSeries_calibrate_decr;
+	ppc_md.progress = iSeries_progress;
+}
+

arch/ppc64/kernel/iSeries_setup.h

@@ -19,10 +19,6 @@
 #ifndef	__ISERIES_SETUP_H__
 #define	__ISERIES_SETUP_H__
 
-extern void iSeries_init_early(void);
-extern void iSeries_init(unsigned long r3, unsigned long ird_start,
-		unsigned long ird_end, unsigned long cline_start,
-		unsigned long cline_end);
 extern void iSeries_setup_arch(void);
 extern void iSeries_setup_residual(struct seq_file *m, int cpu_id);
 extern void iSeries_get_cpuinfo(struct seq_file *m);

arch/ppc64/kernel/idle.c

@@ -30,9 +30,8 @@
 #include <asm/time.h>
 #include <asm/iSeries/HvCall.h>
 #include <asm/iSeries/ItLpQueue.h>
+#include <asm/plpar_wrappers.h>
 
-extern long cede_processor(void);
-extern long poll_pending(void);
 extern void power4_idle(void);
 
 static int (*idle_loop)(void);
@@ -153,6 +152,8 @@ static int default_idle(void)
 	return 0;
 }
 
+#ifdef CONFIG_PPC_PSERIES
+
 DECLARE_PER_CPU(unsigned long, smt_snooze_delay);
 
 int dedicated_idle(void)
@@ -280,9 +281,12 @@ static int shared_idle(void)
 	return 0;
 }
 
-static int powermac_idle(void)
+#endif /* CONFIG_PPC_PSERIES */
+
+static int native_idle(void)
 {
 	while(1) {
+		/* check CPU type here */
 		if (!need_resched())
 			power4_idle();
 		if (need_resched())
@@ -290,7 +294,8 @@ static int powermac_idle(void)
 	}
 	return 0;
 }
-#endif
+
+#endif /* CONFIG_PPC_ISERIES */
 
 int cpu_idle(void)
 {
@@ -332,30 +337,38 @@ __initcall(register_powersave_nap_sysctl);
 
 int idle_setup(void)
 {
+	/*
+	 * Move that junk to each platform specific file, eventually define
+	 * a pSeries_idle for shared processor stuff
+	 */
 #ifdef CONFIG_PPC_ISERIES
 	idle_loop = iSeries_idle;
+	return 1;
 #else
+	idle_loop = default_idle;
+#endif
+#ifdef CONFIG_PPC_PSERIES
 	if (systemcfg->platform & PLATFORM_PSERIES) {
 		if (cur_cpu_spec->firmware_features & FW_FEATURE_SPLPAR) {
 			if (get_paca()->lppaca.xSharedProc) {
-				printk("idle = shared_idle\n");
+				printk(KERN_INFO "Using shared processor idle loop\n");
 				idle_loop = shared_idle;
 			} else {
-				printk("idle = dedicated_idle\n");
+				printk(KERN_INFO "Using dedicated idle loop\n");
 				idle_loop = dedicated_idle;
 			}
 		} else {
-			printk("idle = default_idle\n");
+			printk(KERN_INFO "Using default idle loop\n");
 			idle_loop = default_idle;
 		}
-	} else if (systemcfg->platform == PLATFORM_POWERMAC) {
-		printk("idle = powermac_idle\n");
-		idle_loop = powermac_idle;
-	} else {
-		printk("idle_setup: unknown platform, use default_idle\n");
-		idle_loop = default_idle;
 	}
-#endif
+#endif /* CONFIG_PPC_PSERIES */
+#ifdef CONFIG_PPC_PMAC
+	if (systemcfg->platform == PLATFORM_POWERMAC) {
+		printk(KERN_INFO "Using native/NAP idle loop\n");
+		idle_loop = native_idle;
+	}
+#endif /* CONFIG_PPC_PMAC */
 
 	return 1;
 }

arch/ppc64/kernel/lmb.c

@@ -22,6 +22,43 @@
 
 struct lmb lmb;
 
+#undef DEBUG
+
+void lmb_dump_all(void)
+{
+#ifdef DEBUG
+	unsigned long i;
+	struct lmb *_lmb  = &lmb;
+
+	udbg_printf("lmb_dump_all:\n");
+	udbg_printf("    memory.cnt		  = 0x%lx\n",
+		    _lmb->memory.cnt);
+	udbg_printf("    memory.size		  = 0x%lx\n",
+		    _lmb->memory.size);
+	for (i=0; i < _lmb->memory.cnt ;i++) {
+		udbg_printf("    memory.region[0x%x].base       = 0x%lx\n",
+			    i, _lmb->memory.region[i].base);
+		udbg_printf("		      .physbase = 0x%lx\n",
+			    _lmb->memory.region[i].physbase);
+		udbg_printf("		      .size     = 0x%lx\n",
+			    _lmb->memory.region[i].size);
+	}
+
+	udbg_printf("\n    reserved.cnt	  = 0x%lx\n",
+		    _lmb->reserved.cnt);
+	udbg_printf("    reserved.size	  = 0x%lx\n",
+		    _lmb->reserved.size);
+	for (i=0; i < _lmb->reserved.cnt ;i++) {
+		udbg_printf("    reserved.region[0x%x].base       = 0x%lx\n",
+			    i, _lmb->reserved.region[i].base);
+		udbg_printf("		      .physbase = 0x%lx\n",
+			    _lmb->reserved.region[i].physbase);
+		udbg_printf("		      .size     = 0x%lx\n",
+			    _lmb->reserved.region[i].size);
+	}
+#endif /* DEBUG */
+}
+
 static unsigned long __init
 lmb_addrs_overlap(unsigned long base1, unsigned long size1,
                   unsigned long base2, unsigned long size2)
@@ -71,8 +108,7 @@ lmb_coalesce_regions(struct lmb_region *rgn, unsigned long r1, unsigned long r2)
 void __init
 lmb_init(void)
 {
-	unsigned long offset = reloc_offset();
-	struct lmb *_lmb = PTRRELOC(&lmb);
+	struct lmb *_lmb = &lmb;
 
 	/* Create a dummy zero size LMB which will get coalesced away later.
 	 * This simplifies the lmb_add() code below...
@@ -94,8 +130,7 @@ lmb_analyze(void)
 	unsigned long i;
 	unsigned long mem_size = 0;
 	unsigned long size_mask = 0;
-	unsigned long offset = reloc_offset();
-	struct lmb *_lmb = PTRRELOC(&lmb);
+	struct lmb *_lmb = &lmb;
 #ifdef CONFIG_MSCHUNKS
 	unsigned long physbase = 0;
 #endif
@@ -178,8 +213,7 @@ lmb_add_region(struct lmb_region *rgn, unsigned long base, unsigned long size)
 long __init
 lmb_add(unsigned long base, unsigned long size)
 {
-	unsigned long offset = reloc_offset();
-	struct lmb *_lmb = PTRRELOC(&lmb);
+	struct lmb *_lmb = &lmb;
 	struct lmb_region *_rgn = &(_lmb->memory);
 
 	/* On pSeries LPAR systems, the first LMB is our RMO region. */
@@ -193,8 +227,7 @@ lmb_add(unsigned long base, unsigned long size)
 long __init
 lmb_reserve(unsigned long base, unsigned long size)
 {
-	unsigned long offset = reloc_offset();
-	struct lmb *_lmb = PTRRELOC(&lmb);
+	struct lmb *_lmb = &lmb;
 	struct lmb_region *_rgn = &(_lmb->reserved);
 
 	return lmb_add_region(_rgn, base, size);
@@ -227,8 +260,7 @@ lmb_alloc_base(unsigned long size, unsigned long align, unsigned long max_addr)
 {
 	long i, j;
 	unsigned long base = 0;
-	unsigned long offset = reloc_offset();
-	struct lmb *_lmb = PTRRELOC(&lmb);
+	struct lmb *_lmb = &lmb;
 	struct lmb_region *_mem = &(_lmb->memory);
 	struct lmb_region *_rsv = &(_lmb->reserved);
 
@@ -263,8 +295,7 @@ lmb_alloc_base(unsigned long size, unsigned long align, unsigned long max_addr)
 unsigned long __init
 lmb_phys_mem_size(void)
 {
-	unsigned long offset = reloc_offset();
-	struct lmb *_lmb = PTRRELOC(&lmb);
+	struct lmb *_lmb = &lmb;
 #ifdef CONFIG_MSCHUNKS
 	return _lmb->memory.size;
 #else
@@ -282,8 +313,7 @@ lmb_phys_mem_size(void)
 unsigned long __init
 lmb_end_of_DRAM(void)
 {
-	unsigned long offset = reloc_offset();
-	struct lmb *_lmb = PTRRELOC(&lmb);
+	struct lmb *_lmb = &lmb;
 	struct lmb_region *_mem = &(_lmb->memory);
 	int idx = _mem->cnt - 1;
 
@@ -300,8 +330,7 @@ unsigned long __init
 lmb_abs_to_phys(unsigned long aa)
 {
 	unsigned long i, pa = aa;
-	unsigned long offset = reloc_offset();
-	struct lmb *_lmb = PTRRELOC(&lmb);
+	struct lmb *_lmb = &lmb;
 	struct lmb_region *_mem = &(_lmb->memory);
 
 	for (i=0; i < _mem->cnt; i++) {

arch/ppc64/kernel/misc.S

@@ -127,6 +127,47 @@ _GLOBAL(call_handle_irq_event)
 	blr
 #endif /* CONFIG_IRQSTACKS */
 
+	/*
+ * To be called by C code which needs to do some operations with MMU
+ * disabled. Note that interrupts have to be disabled by the caller
+ * prior to calling us. The code called _MUST_ be in the RMO of course
+ * and part of the linear mapping as we don't attempt to translate the
+ * stack pointer at all. The function is called with the stack switched
+ * to this CPU emergency stack
+ *
+ * prototype is void *call_with_mmu_off(void *func, void *data);
+ *
+ * the called function is expected to be of the form
+ *
+ * void *called(void *data); 
+ */
+_GLOBAL(call_with_mmu_off)
+	mflr	r0			/* get link, save it on stackframe */
+	std	r0,16(r1)
+	mr	r1,r5			/* save old stack ptr */
+	ld	r1,PACAEMERGSP(r13)	/* get emerg. stack */
+	subi	r1,r1,STACK_FRAME_OVERHEAD
+	std	r0,16(r1)		/* save link on emerg. stack */
+	std	r5,0(r1)		/* save old stack ptr in backchain */
+	ld	r3,0(r3)		/* get to real function ptr (assume same TOC) */
+	bl	2f			/* we need LR to return, continue at label 2 */
+
+	ld	r0,16(r1)		/* we return here from the call, get LR and */
+	ld	r1,0(r1)		/* .. old stack ptr */
+	mtspr	SPRN_SRR0,r0		/* and get back to virtual mode with these */
+	mfmsr	r4
+	ori	r4,r4,MSR_IR|MSR_DR
+	mtspr	SPRN_SRR1,r4
+	rfid
+
+2:	mtspr	SPRN_SRR0,r3		/* coming from above, enter real mode */
+	mr	r3,r4			/* get parameter */
+	mfmsr	r0
+	ori	r0,r0,MSR_IR|MSR_DR
+	xori	r0,r0,MSR_IR|MSR_DR
+	mtspr	SPRN_SRR1,r0
+	rfid
+	
 /*
  * Flush instruction cache.
  */
@@ -454,17 +495,6 @@ _GLOBAL(_outsl_ns)
 	sync
 	blr	
 
-_GLOBAL(_get_PVR)
-	mfspr	r3,PVR
-	blr
-
-_GLOBAL(_get_PIR)
-	mfspr	r3,PIR
-	blr
-
-_GLOBAL(_get_HID0)
-	mfspr	r3,HID0
-	blr
 
 _GLOBAL(cvt_fd)
 	lfd	0,0(r5)		/* load up fpscr value */
@@ -561,6 +591,69 @@ _GLOBAL(do_cpu_ftr_fixups)
 	isync
 	b	1b
 
+#ifdef CONFIG_PPC_PMAC
+/*
+ * Do an IO access in real mode
+ */
+_GLOBAL(real_readb)
+	mfmsr	r7
+	ori	r0,r7,MSR_DR
+	xori	r0,r0,MSR_DR
+	sync
+	mtmsrd	r0
+	sync
+	isync
+	mfspr	r6,SPRN_HID4
+	rldicl	r5,r6,32,0
+	ori	r5,r5,0x100
+	rldicl	r5,r5,32,0
+	sync
+	mtspr	SPRN_HID4,r5
+	isync
+	slbia
+	isync
+	lbz	r3,0(r3)
+	sync
+	mtspr	SPRN_HID4,r6
+	isync
+	slbia
+	isync
+	mtmsrd	r7
+	sync
+	isync
+	blr
+
+	/*
+ * Do an IO access in real mode
+ */
+_GLOBAL(real_writeb)
+	mfmsr	r7
+	ori	r0,r7,MSR_DR
+	xori	r0,r0,MSR_DR
+	sync
+	mtmsrd	r0
+	sync
+	isync
+	mfspr	r6,SPRN_HID4
+	rldicl	r5,r6,32,0
+	ori	r5,r5,0x100
+	rldicl	r5,r5,32,0
+	sync
+	mtspr	SPRN_HID4,r5
+	isync
+	slbia
+	isync
+	stb	r3,0(r4)
+	sync
+	mtspr	SPRN_HID4,r6
+	isync
+	slbia
+	isync
+	mtmsrd	r7
+	sync
+	isync
+	blr
+#endif /* CONFIG_PPC_PMAC */
 
 /*
  * Create a kernel thread
@@ -594,7 +687,7 @@ _GLOBAL(kernel_thread)
 	ld	r30,-16(r1)
 	blr
 
-#ifdef CONFIG_PPC_ISERIES	/* hack hack hack */
+#ifndef CONFIG_PPC_PSERIE	/* hack hack hack */
 #define ppc_rtas	sys_ni_syscall
 #endif
 

arch/ppc64/kernel/nvram.c

@@ -603,6 +603,7 @@ void __exit nvram_cleanup(void)
 }
 
 
+#ifdef CONFIG_PPC_PSERIES
 
 /* nvram_write_error_log
  *
@@ -727,6 +728,7 @@ int nvram_clear_error_log()
 	return 0;
 }
 
+#endif /* CONFIG_PPC_PSERIES */
 
 module_init(nvram_init);
 module_exit(nvram_cleanup);

arch/ppc64/kernel/open_pic.c

@@ -362,7 +362,6 @@ void __init openpic_init(int main_pic, int offset, unsigned char* chrp_ack,
 	find_ISUs();
 
 	/* Initialize timer interrupts */
-	ppc64_boot_msg(0x21, "OpenPic Timer");
 	for (i = 0; i < OPENPIC_NUM_TIMERS; i++) {
 		/* Disabled, Priority 0 */
 		openpic_inittimer(i, 0, openpic_vec_timer+i);
@@ -372,7 +371,6 @@ void __init openpic_init(int main_pic, int offset, unsigned char* chrp_ack,
 
 #ifdef CONFIG_SMP
 	/* Initialize IPI interrupts */
-	ppc64_boot_msg(0x22, "OpenPic IPI");
 	openpic_test_broken_IPI();
 	for (i = 0; i < OPENPIC_NUM_IPI; i++) {
 		/* Disabled, Priority 10..13 */
@@ -384,8 +382,6 @@ void __init openpic_init(int main_pic, int offset, unsigned char* chrp_ack,
 #endif
 
 	/* Initialize external interrupts */
-	ppc64_boot_msg(0x23, "OpenPic Ext");
-
 	openpic_set_priority(0xf);
 
 	/* SIOint (8259 cascade) is special */
@@ -420,7 +416,6 @@ void __init openpic_init(int main_pic, int offset, unsigned char* chrp_ack,
 		irq_desc[i].handler = &open_pic;
 
 	/* Initialize the spurious interrupt */
-	ppc64_boot_msg(0x24, "OpenPic Spurious");
 	openpic_set_spurious(openpic_vec_spurious);
 
 	openpic_set_priority(0);

arch/ppc64/kernel/open_pic_u3.c

@@ -344,5 +344,5 @@ void __init openpic2_init(int offset)
 	openpic2_set_priority(0);
 	openpic2_disable_8259_pass_through();
 
-	ppc64_boot_msg(0x25, "OpenPic2 Done");
+	ppc64_boot_msg(0x25, "OpenPic U3 Done");
 }

arch/ppc64/kernel/pSeries_iommu.c

@@ -42,6 +42,7 @@
 #include <asm/pci-bridge.h>
 #include <asm/machdep.h>
 #include <asm/abs_addr.h>
+#include <asm/plpar_wrappers.h>
 #include "pci.h"
 
 
@@ -88,6 +89,150 @@ static void tce_free_pSeries(struct iommu_table *tbl, long index, long npages)
 }
 
 
+static void tce_build_pSeriesLP(struct iommu_table *tbl, long tcenum,
+				long npages, unsigned long uaddr,
+				enum dma_data_direction direction)
+{
+	u64 rc;
+	union tce_entry tce;
+
+	tce.te_word = 0;
+	tce.te_rpn = (virt_to_abs(uaddr)) >> PAGE_SHIFT;
+	tce.te_rdwr = 1;
+	if (direction != DMA_TO_DEVICE)
+		tce.te_pciwr = 1;
+
+	while (npages--) {
+		rc = plpar_tce_put((u64)tbl->it_index, 
+				   (u64)tcenum << 12, 
+				   tce.te_word );
+		
+		if (rc && printk_ratelimit()) {
+			printk("tce_build_pSeriesLP: plpar_tce_put failed. rc=%ld\n", rc);
+			printk("\tindex   = 0x%lx\n", (u64)tbl->it_index);
+			printk("\ttcenum  = 0x%lx\n", (u64)tcenum);
+			printk("\ttce val = 0x%lx\n", tce.te_word );
+			show_stack(current, (unsigned long *)__get_SP());
+		}
+			
+		tcenum++;
+		tce.te_rpn++;
+	}
+}
+
+DEFINE_PER_CPU(void *, tce_page) = NULL;
+
+static void tce_buildmulti_pSeriesLP(struct iommu_table *tbl, long tcenum,
+				     long npages, unsigned long uaddr,
+				     enum dma_data_direction direction)
+{
+	u64 rc;
+	union tce_entry tce, *tcep;
+	long l, limit;
+
+	if (npages == 1)
+		return tce_build_pSeriesLP(tbl, tcenum, npages, uaddr,
+					   direction);
+
+	tcep = __get_cpu_var(tce_page);
+
+	/* This is safe to do since interrupts are off when we're called
+	 * from iommu_alloc{,_sg}()
+	 */
+	if (!tcep) {
+		tcep = (void *)__get_free_page(GFP_ATOMIC);
+		/* If allocation fails, fall back to the loop implementation */
+		if (!tcep)
+			return tce_build_pSeriesLP(tbl, tcenum, npages,
+						   uaddr, direction);
+		__get_cpu_var(tce_page) = tcep;
+	}
+
+	tce.te_word = 0;
+	tce.te_rpn = (virt_to_abs(uaddr)) >> PAGE_SHIFT;
+	tce.te_rdwr = 1;
+	if (direction != DMA_TO_DEVICE)
+		tce.te_pciwr = 1;
+
+	/* We can map max one pageful of TCEs at a time */
+	do {
+		/*
+		 * Set up the page with TCE data, looping through and setting
+		 * the values.
+		 */
+		limit = min_t(long, npages, PAGE_SIZE/sizeof(union tce_entry));
+
+		for (l = 0; l < limit; l++) {
+			tcep[l] = tce;
+			tce.te_rpn++;
+		}
+
+		rc = plpar_tce_put_indirect((u64)tbl->it_index,
+					    (u64)tcenum << 12,
+					    (u64)virt_to_abs(tcep),
+					    limit);
+
+		npages -= limit;
+		tcenum += limit;
+	} while (npages > 0 && !rc);
+
+	if (rc && printk_ratelimit()) {
+		printk("tce_buildmulti_pSeriesLP: plpar_tce_put failed. rc=%ld\n", rc);
+		printk("\tindex   = 0x%lx\n", (u64)tbl->it_index);
+		printk("\tnpages  = 0x%lx\n", (u64)npages);
+		printk("\ttce[0] val = 0x%lx\n", tcep[0].te_word);
+		show_stack(current, (unsigned long *)__get_SP());
+	}
+}
+
+static void tce_free_pSeriesLP(struct iommu_table *tbl, long tcenum, long npages)
+{
+	u64 rc;
+	union tce_entry tce;
+
+	tce.te_word = 0;
+
+	while (npages--) {
+		rc = plpar_tce_put((u64)tbl->it_index,
+				   (u64)tcenum << 12,
+				   tce.te_word);
+
+		if (rc && printk_ratelimit()) {
+			printk("tce_free_pSeriesLP: plpar_tce_put failed. rc=%ld\n", rc);
+			printk("\tindex   = 0x%lx\n", (u64)tbl->it_index);
+			printk("\ttcenum  = 0x%lx\n", (u64)tcenum);
+			printk("\ttce val = 0x%lx\n", tce.te_word );
+			show_stack(current, (unsigned long *)__get_SP());
+		}
+
+		tcenum++;
+	}
+}
+
+
+static void tce_freemulti_pSeriesLP(struct iommu_table *tbl, long tcenum, long npages)
+{
+	u64 rc;
+	union tce_entry tce;
+
+	tce.te_word = 0;
+
+	rc = plpar_tce_stuff((u64)tbl->it_index,
+			   (u64)tcenum << 12,
+			   tce.te_word,
+			   npages);
+
+	if (rc && printk_ratelimit()) {
+		printk("tce_freemulti_pSeriesLP: plpar_tce_stuff failed\n");
+		printk("\trc      = %ld\n", rc);
+		printk("\tindex   = 0x%lx\n", (u64)tbl->it_index);
+		printk("\tnpages  = 0x%lx\n", (u64)npages);
+		printk("\ttce val = 0x%lx\n", tce.te_word );
+		show_stack(current, (unsigned long *)__get_SP());
+	}
+}
+
+
 static void iommu_buses_init(void)
 {
 	struct pci_controller *phb, *tmp;
@@ -166,24 +311,25 @@ static void iommu_table_setparms(struct pci_controller *phb,
 				 struct device_node *dn,
 				 struct iommu_table *tbl) 
 {
-	phandle node;
-	unsigned long i;
-	struct of_tce_table *oft;
-
-	node = ((struct device_node *)(phb->arch_data))->node;
+	struct device_node *node;
+	unsigned long *basep;
+	unsigned int *sizep;
 
-	oft = NULL;
+	node = (struct device_node *)phb->arch_data;
 
-	for (i=0; of_tce_table[i].node; i++)
-		if(of_tce_table[i].node == node) {
-			oft = &of_tce_table[i];
-			break;
-		}
-
-	if (!oft)
-		panic("PCI_DMA: iommu_table_setparms: Can't find phb named '%s' in of_tce_table\n", dn->full_name);
-
-	memset((void *)oft->base, 0, oft->size);
+	if (get_property(node, "linux,has-tce-table", NULL) == NULL) {
+		printk(KERN_ERR "PCI_DMA: iommu_table_setparms: %s has no tce table !\n",
+		      dn->full_name);
+		return;
+	}
+	basep = (unsigned long *)get_property(node, "linux,tce-base", NULL);
+	sizep = (unsigned int *)get_property(node, "linux,tce-size", NULL);
+	if (basep == NULL || sizep == NULL) {
+		printk(KERN_ERR "PCI_DMA: iommu_table_setparms: %s has missing tce"
+		       " entries !\n", dn->full_name);
+		return;
+	}
+	memset((void *)(*basep), 0, *sizep);
 
 	tbl->it_busno = phb->bus->number;
 	
@@ -207,7 +353,7 @@ static void iommu_table_setparms(struct pci_controller *phb,
 	if (phb->dma_window_base_cur > (1 << 19))
 		panic("PCI_DMA: Unexpected number of IOAs under this PHB.\n"); 
 	
-	tbl->it_base = oft->base;
+	tbl->it_base = *basep;
 	tbl->it_index = 0;
 	tbl->it_entrysize = sizeof(union tce_entry);
 	tbl->it_blocksize = 16;
@@ -295,8 +441,16 @@ void iommu_setup_pSeries(void)
 /* These are called very early. */
 void tce_init_pSeries(void)
 {
-	ppc_md.tce_build = tce_build_pSeries;
-	ppc_md.tce_free  = tce_free_pSeries;
+	if (!(systemcfg->platform & PLATFORM_LPAR)) {
+		ppc_md.tce_build = tce_build_pSeries;
+		ppc_md.tce_free  = tce_free_pSeries;
+	} else if (cur_cpu_spec->firmware_features & FW_FEATURE_MULTITCE) {
+		ppc_md.tce_build = tce_buildmulti_pSeriesLP;
+		ppc_md.tce_free	 = tce_freemulti_pSeriesLP;
+	} else {
+		ppc_md.tce_build = tce_build_pSeriesLP;
+		ppc_md.tce_free	 = tce_free_pSeriesLP;
+	}
 
 	pci_iommu_init();
 }

arch/ppc64/kernel/pci_iommu.c

@@ -54,9 +54,9 @@ static inline struct iommu_table *devnode_table(struct pci_dev *dev)
 	return ISERIES_DEVNODE(dev)->iommu_table;
 #endif /* CONFIG_PPC_ISERIES */
 
-#ifdef CONFIG_PPC_PSERIES
+#ifdef CONFIG_PPC_MULTIPLATFORM
 	return PCI_GET_DN(dev)->iommu_table;
-#endif /* CONFIG_PPC_PSERIES */
+#endif /* CONFIG_PPC_MULTIPLATFORM */
 }
 
 

arch/ppc64/kernel/pmac.h

@@ -29,4 +29,6 @@ extern void pmac_ide_init_hwif_ports(hw_regs_t *hw,
 
 extern void pmac_nvram_init(void);
 
+extern void pmac_iommu_alloc(void);
+
 #endif /* __PMAC_H__ */

arch/ppc64/kernel/pmac_feature.c

@@ -44,9 +44,9 @@
 #undef DEBUG_FEATURE
 
 #ifdef DEBUG_FEATURE
-#define DBG(fmt,...) printk(KERN_DEBUG fmt)
+#define DBG(fmt...) printk(KERN_DEBUG fmt)
 #else
-#define DBG(fmt,...)
+#define DBG(fmt...)
 #endif
 
 /*

arch/ppc64/kernel/pmac_iommu.c

@@ -37,13 +37,14 @@
 #include <linux/vmalloc.h>
 #include <asm/io.h>
 #include <asm/prom.h>
-#include <asm/rtas.h>
 #include <asm/ppcdebug.h>
 #include <asm/iommu.h>
 #include <asm/pci-bridge.h>
 #include <asm/machdep.h>
 #include <asm/abs_addr.h>
 #include <asm/cacheflush.h>
+#include <asm/lmb.h>
+
 #include "pci.h"
 
 
@@ -76,8 +77,8 @@
 #define DARTMAP_RPNMASK 0x00ffffff
 
 /* Physical base address and size of the DART table */
-unsigned long dart_tablebase;
-unsigned long dart_tablesize;
+unsigned long dart_tablebase; /* exported to htab_initialize */
+static unsigned long dart_tablesize;
 
 /* Virtual base address of the DART table */
 static u32 *dart_vbase;
@@ -263,7 +264,6 @@ static int dart_init(struct device_node *dart_node)
 	return 0;
 }
 
-
 void iommu_setup_pmac(void)
 {
 	struct pci_dev *dev = NULL;
@@ -300,6 +300,22 @@ void iommu_setup_pmac(void)
 	}
 }
 
+void __init pmac_iommu_alloc(void)
+{
+	/* Only reserve DART space if machine has more than 2GB of RAM
+	 * or if requested with iommu=on on cmdline.
+	 */
+	if (lmb_end_of_DRAM() <= 0x80000000ull &&
+	    get_property(of_chosen, "linux,iommu-force-on", NULL) == NULL)
+		return;
 
+	/* 512 pages (2MB) is max DART tablesize. */
+	dart_tablesize = 1UL << 21;
+	/* 16MB (1 << 24) alignment. We allocate a full 16Mb chuck since we
+	 * will blow up an entire large page anyway in the kernel mapping
+	 */
+	dart_tablebase = (unsigned long)
+		abs_to_virt(lmb_alloc_base(1UL<<24, 1UL<<24, 0x80000000L));
 
-
+	printk(KERN_INFO "U3-DART allocated at: %lx\n", dart_tablebase);
+}

arch/ppc64/kernel/pmac_pci.c

@@ -746,6 +746,9 @@ void __init pmac_pci_init(void)
 	 * the exception of U3/AGP (hook into pci_set_mwi)
 	 */
 	pci_cache_line_size = 16; /* 64 bytes */
+
+	/* Allow all IO */
+	io_page_mask = -1;
 }
 
 /*

arch/ppc64/kernel/pmac_setup.c

@@ -23,6 +23,8 @@
  * bootup setup stuff..
  */
 
+#undef DEBUG
+
 #include <linux/config.h>
 #include <linux/init.h>
 #include <linux/errno.h>
@@ -63,7 +65,6 @@
 #include <asm/iommu.h>
 #include <asm/machdep.h>
 #include <asm/dma.h>
-#include <asm/bootx.h>
 #include <asm/btext.h>
 #include <asm/cputable.h>
 #include <asm/pmac_feature.h>
@@ -73,6 +74,13 @@
 
 #include "pmac.h"
 
+#ifdef DEBUG
+#define DBG(fmt...) udbg_printf(fmt)
+#else
+#define DBG(fmt...)
+#endif
+
+
 static int current_root_goodness = -1;
 #define DEFAULT_ROOT_DEVICE Root_SDA1	/* sda1 - slightly silly choice */
 
@@ -81,10 +89,6 @@ int sccdbg;
 
 extern void udbg_init_scc(struct device_node *np);
 
-#ifdef CONFIG_BOOTX_TEXT
-void pmac_progress(char *s, unsigned short hex);
-#endif
-
 void __pmac pmac_show_cpuinfo(struct seq_file *m)
 {
 	struct device_node *np;
@@ -134,28 +138,20 @@ void __pmac pmac_show_cpuinfo(struct seq_file *m)
 
 void __init pmac_setup_arch(void)
 {
-	struct device_node *cpu;
-	int *fp;
-	unsigned long pvr;
-
-	pvr = PVR_VER(mfspr(PVR));
-
-	/* Set loops_per_jiffy to a half-way reasonable value,
-	   for use until calibrate_delay gets called. */
-	cpu = find_type_devices("cpu");
-	if (cpu != 0) {
-		fp = (int *) get_property(cpu, "clock-frequency", NULL);
-		if (fp != 0) {
-			if (pvr == 4 || pvr >= 8)
-				/* 604, G3, G4 etc. */
-				loops_per_jiffy = *fp / HZ;
-			else
-				/* 601, 603, etc. */
-				loops_per_jiffy = *fp / (2*HZ);
-		} else
-			loops_per_jiffy = 50000000 / HZ;
+	/* init to some ~sane value until calibrate_delay() runs */
+	loops_per_jiffy = 50000000;
+
+	/* Probe motherboard chipset */
+	pmac_feature_init();
+#if 0
+	/* Lock-enable the SCC channel used for debug */
+	if (sccdbg) {
+		np = of_find_node_by_name(NULL, "escc");
+		if (np)
+			pmac_call_feature(PMAC_FTR_SCC_ENABLE, np,
+					  PMAC_SCC_ASYNC | PMAC_SCC_FLAG_XMON, 1);
 	}
-	
+#endif
 	/* We can NAP */
 	powersave_nap = 1;
 
@@ -183,67 +179,13 @@ void __init pmac_setup_arch(void)
 #endif
 }
 
-extern char *bootpath;
-extern char *bootdevice;
-void *boot_host;
-int boot_target;
-int boot_part;
-extern dev_t boot_dev;
-
 #ifdef CONFIG_SCSI
-void __init note_scsi_host(struct device_node *node, void *host)
+void note_scsi_host(struct device_node *node, void *host)
 {
-	int l;
-	char *p;
-
-	l = strlen(node->full_name);
-	if (bootpath != NULL && bootdevice != NULL
-	    && strncmp(node->full_name, bootdevice, l) == 0
-	    && (bootdevice[l] == '/' || bootdevice[l] == 0)) {
-		boot_host = host;
-		/*
-		 * There's a bug in OF 1.0.5.  (Why am I not surprised.)
-		 * If you pass a path like scsi/sd@1:0 to canon, it returns
-		 * something like /bandit@F2000000/gc@10/53c94@10000/sd@0,0
-		 * That is, the scsi target number doesn't get preserved.
-		 * So we pick the target number out of bootpath and use that.
-		 */
-		p = strstr(bootpath, "/sd@");
-		if (p != NULL) {
-			p += 4;
-			boot_target = simple_strtoul(p, NULL, 10);
-			p = strchr(p, ':');
-			if (p != NULL)
-				boot_part = simple_strtoul(p + 1, NULL, 10);
-		}
-	}
+	/* Obsolete */
 }
 #endif
 
-#if defined(CONFIG_BLK_DEV_IDE) && defined(CONFIG_BLK_DEV_IDE_PMAC)
-static dev_t __init find_ide_boot(void)
-{
-	char *p;
-	int n;
-	dev_t __init pmac_find_ide_boot(char *bootdevice, int n);
-
-	if (bootdevice == NULL)
-		return 0;
-	p = strrchr(bootdevice, '/');
-	if (p == NULL)
-		return 0;
-	n = p - bootdevice;
-
-	return pmac_find_ide_boot(bootdevice, n);
-}
-#endif /* CONFIG_BLK_DEV_IDE && CONFIG_BLK_DEV_IDE_PMAC */
-
-void __init find_boot_device(void)
-{
-#if defined(CONFIG_BLK_DEV_IDE) && defined(CONFIG_BLK_DEV_IDE_PMAC)
-	boot_dev = find_ide_boot();
-#endif
-}
 
 static int initializing = 1;
 
@@ -258,7 +200,7 @@ late_initcall(pmac_late_init);
 /* can't be __init - can be called whenever a disk is first accessed */
 void __pmac note_bootable_part(dev_t dev, int part, int goodness)
 {
-	static int found_boot = 0;
+	extern dev_t boot_dev;
 	char *p;
 
 	if (!initializing)
@@ -270,10 +212,6 @@ void __pmac note_bootable_part(dev_t dev, int part, int goodness)
 	if (p != NULL && (p == saved_command_line || p[-1] == ' '))
 		return;
 
-	if (!found_boot) {
-		find_boot_device();
-		found_boot = 1;
-	}
 	if (!boot_dev || dev == boot_dev) {
 		ROOT_DEV = dev + part;
 		boot_dev = 0;
@@ -313,21 +251,73 @@ static void btext_putc(unsigned char c)
 }
 #endif /* CONFIG_BOOTX_TEXT */
 
+static void __init init_boot_display(void)
+{
+	char *name;
+	struct device_node *np = NULL; 
+	int rc = -ENODEV;
+
+	printk("trying to initialize btext ...\n");
+
+	name = (char *)get_property(of_chosen, "linux,stdout-path", NULL);
+	if (name != NULL) {
+		np = of_find_node_by_path(name);
+		if (np != NULL) {
+			if (strcmp(np->type, "display") != 0) {
+				printk("boot stdout isn't a display !\n");
+				of_node_put(np);
+				np = NULL;
+			}
+		}
+	}
+	if (np)
+		rc = btext_initialize(np);
+	if (rc == 0)
+		return;
+
+	for (np = NULL; (np = of_find_node_by_type(np, "display"));) {
+		if (get_property(np, "linux,opened", NULL)) {
+			printk("trying %s ...\n", np->full_name);
+			rc = btext_initialize(np);
+			printk("result: %d\n", rc);
+		}
+		if (rc == 0)
+			return;
+	}
+}
+
 /* 
  * Early initialization.
- * Relocation is on but do not reference unbolted pages
- * Also, device-tree hasn't been "finished", so don't muck with
- * it too much
  */
 void __init pmac_init_early(void)
 {
-	hpte_init_pSeries();
+	DBG(" -> pmac_init_early\n");
+
+	/* Initialize hash table, from now on, we can take hash faults
+	 * and call ioremap
+	 */
+	hpte_init_native();
 
+	/* Init SCC */
+       	if (strstr(cmd_line, "sccdbg")) {
+		sccdbg = 1;
+       		udbg_init_scc(NULL);
+       	}
+
+	else {
 #ifdef CONFIG_BOOTX_TEXT
-	ppc_md.udbg_putc = btext_putc;
-	ppc_md.udbg_getc = dummy_getc;
-	ppc_md.udbg_getc_poll = dummy_getc_poll;
+		init_boot_display();
+
+		ppc_md.udbg_putc = btext_putc;
+		ppc_md.udbg_getc = dummy_getc;
+		ppc_md.udbg_getc_poll = dummy_getc_poll;
 #endif /* CONFIG_BOOTX_TEXT */
+	}
+
+	/* Setup interrupt mapping options */
+	naca->interrupt_controller = IC_OPEN_PIC;
+
+	DBG(" <- pmac_init_early\n");
 }
 
 extern void* OpenPIC_Addr;
@@ -417,60 +407,19 @@ static int __init pmac_irq_cascade_init(void)
 
 core_initcall(pmac_irq_cascade_init);
 
-void __init pmac_init(unsigned long r3, unsigned long r4, unsigned long r5,
-		      unsigned long r6, unsigned long r7)
-{
-	/* Probe motherboard chipset */
-	pmac_feature_init();
-
-	/* Init SCC */
-	if (strstr(cmd_line, "sccdbg")) {
-		sccdbg = 1;
-		udbg_init_scc(NULL);
-	}
-
-	/* Fill up the machine description */
-	ppc_md.setup_arch     = pmac_setup_arch;
-       	ppc_md.get_cpuinfo    = pmac_show_cpuinfo;
-
-	ppc_md.init_IRQ       = pmac_init_IRQ;
-	ppc_md.get_irq        = openpic_get_irq;
-
-	ppc_md.pcibios_fixup  = pmac_pcibios_fixup;
-
-	ppc_md.restart        = pmac_restart;
-	ppc_md.power_off      = pmac_power_off;
-	ppc_md.halt           = pmac_halt;
-
-       	ppc_md.get_boot_time  = pmac_get_boot_time;
-       	ppc_md.set_rtc_time   = pmac_set_rtc_time;
-       	ppc_md.get_rtc_time   = pmac_get_rtc_time;
-      	ppc_md.calibrate_decr = pmac_calibrate_decr;
-
-	ppc_md.feature_call   = pmac_do_feature_call;
-
-
-#ifdef CONFIG_BOOTX_TEXT
-	ppc_md.progress       = pmac_progress;
-#endif /* CONFIG_BOOTX_TEXT */
-
-	if (ppc_md.progress) ppc_md.progress("pmac_init(): exit", 0);
-
-}
-
-#ifdef CONFIG_BOOTX_TEXT
-void __init pmac_progress(char *s, unsigned short hex)
+static void __init pmac_progress(char *s, unsigned short hex)
 {
 	if (sccdbg) {
 		udbg_puts(s);
-		udbg_putc('\n');
+		udbg_puts("\n");
 	}
+#ifdef CONFIG_BOOTX_TEXT
 	else if (boot_text_mapped) {
 		btext_drawstring(s);
-		btext_drawchar('\n');
+		btext_drawstring("\n");
 	}
-}
 #endif /* CONFIG_BOOTX_TEXT */
+}
 
 static int __init pmac_declare_of_platform_devices(void)
 {
@@ -489,3 +438,43 @@ static int __init pmac_declare_of_platform_devices(void)
 }
 
 device_initcall(pmac_declare_of_platform_devices);
+
+/*
+ * Called very early, MMU is off, device-tree isn't unflattened
+ */
+static int __init pmac_probe(int platform)
+{
+	if (platform != PLATFORM_POWERMAC)
+		return 0;
+
+#ifdef CONFIG_PMAC_DART
+	/*
+	 * On U3, the DART (iommu) must be allocated now since it
+	 * has an impact on htab_initialize (due to the large page it
+	 * occupies having to be broken up so the DART itself is not
+	 * part of the cacheable linar mapping
+	 */
+	pmac_iommu_alloc();
+#endif /* CONFIG_PMAC_DART */
+
+	return 1;
+}
+
+struct machdep_calls __initdata pmac_md = {
+	.probe			= pmac_probe,
+	.setup_arch		= pmac_setup_arch,
+	.init_early		= pmac_init_early,
+       	.get_cpuinfo		= pmac_show_cpuinfo,
+	.init_IRQ		= pmac_init_IRQ,
+	.get_irq		= openpic_get_irq,
+	.pcibios_fixup		= pmac_pcibios_fixup,
+	.restart		= pmac_restart,
+	.power_off		= pmac_power_off,
+	.halt			= pmac_halt,
+       	.get_boot_time		= pmac_get_boot_time,
+       	.set_rtc_time		= pmac_set_rtc_time,
+       	.get_rtc_time		= pmac_get_rtc_time,
+      	.calibrate_decr		= pmac_calibrate_decr,
+	.feature_call		= pmac_do_feature_call,
+	.progress		= pmac_progress,
+};

arch/ppc64/kernel/ppc_ksyms.c

@@ -119,7 +119,6 @@ EXPORT_SYMBOL(start_thread);
 EXPORT_SYMBOL(kernel_thread);
 
 EXPORT_SYMBOL(flush_instruction_cache);
-EXPORT_SYMBOL(_get_PVR);
 EXPORT_SYMBOL(giveup_fpu);
 #ifdef CONFIG_ALTIVEC
 EXPORT_SYMBOL(giveup_altivec);

arch/ppc64/kernel/time.c

@@ -84,7 +84,7 @@ static unsigned long first_settimeofday = 1;
 #define XSEC_PER_SEC (1024*1024)
 
 unsigned long tb_ticks_per_jiffy;
-unsigned long tb_ticks_per_usec;
+unsigned long tb_ticks_per_usec = 100; /* sane default */
 unsigned long tb_ticks_per_sec;
 unsigned long next_xtime_sync_tb;
 unsigned long xtime_sync_interval;

arch/ppc64/mm/Makefile

@@ -8,3 +8,4 @@ obj-y := fault.o init.o imalloc.o hash_utils.o hash_low.o tlb.o \
 	slb_low.o slb.o stab.o mmap.o
 obj-$(CONFIG_DISCONTIGMEM) += numa.o
 obj-$(CONFIG_HUGETLB_PAGE) += hugetlbpage.o
+obj-$(CONFIG_PPC_MULTIPLATFORM) += hash_native.o

arch/ppc64/oprofile/common.c

@@ -133,7 +133,7 @@ int __init oprofile_arch_init(struct oprofile_operations **ops)
 {
 	unsigned int pvr;
 
-	pvr = _get_PVR();
+	pvr = mfspr(SPRN_PVR);
 
 	switch (PVR_VER(pvr)) {
 		case PV_630:

arch/ppc64/xmon/xmon.c

@@ -252,30 +252,28 @@ extern inline void sync(void)
 
 static inline void disable_surveillance(void)
 {
-#ifndef CONFIG_PPC_ISERIES
+#ifdef CONFIG_PPC_PSERIES
 	/* Since this can't be a module, args should end up below 4GB. */
 	static struct rtas_args args;
 
-	if (systemcfg->platform & PLATFORM_PSERIES) {
-		/*
-		 * At this point we have got all the cpus we can into
-		 * xmon, so there is hopefully no other cpu calling RTAS
-		 * at the moment, even though we don't take rtas.lock.
-		 * If we did try to take rtas.lock there would be a
-		 * real possibility of deadlock.
-		 */
-		args.token = rtas_token("set-indicator");
-		if (args.token == RTAS_UNKNOWN_SERVICE)
-			return;
-		args.nargs = 3;
-		args.nret = 1;
-		args.rets = &args.args[3];
-		args.args[0] = SURVEILLANCE_TOKEN;
-		args.args[1] = 0;
-		args.args[2] = 0;
-		enter_rtas(__pa(&args));
-	}
-#endif
+	/*
+	 * At this point we have got all the cpus we can into
+	 * xmon, so there is hopefully no other cpu calling RTAS
+	 * at the moment, even though we don't take rtas.lock.
+	 * If we did try to take rtas.lock there would be a
+	 * real possibility of deadlock.
+	 */
+	args.token = rtas_token("set-indicator");
+	if (args.token == RTAS_UNKNOWN_SERVICE)
+		return;
+	args.nargs = 3;
+	args.nret = 1;
+	args.rets = &args.args[3];
+	args.args[0] = SURVEILLANCE_TOKEN;
+	args.args[1] = 0;
+	args.args[2] = 0;
+	enter_rtas(__pa(&args));
+#endif /* CONFIG_PPC_PSERIES */
 }
 
 #ifdef CONFIG_SMP

include/asm-ppc64/btext.h

@@ -7,23 +7,20 @@
 #define __PPC_BTEXT_H
 #ifdef __KERNEL__
 
-#include <asm/bootx.h>
-
 extern void btext_clearscreen(void);
 extern void btext_flushscreen(void);
 
-extern boot_infos_t disp_bi;
 extern int boot_text_mapped;
 
-void btext_setup_display(int width, int height, int depth, int pitch,
-			 unsigned long address);
-void map_boot_text(void);
-void btext_update_display(unsigned long phys, int width, int height,
-			  int depth, int pitch);
+extern int btext_initialize(struct device_node *np);
+
+extern void map_boot_text(void);
+extern void btext_update_display(unsigned long phys, int width, int height,
+				 int depth, int pitch);
 
-void btext_drawchar(char c);
-void btext_drawstring(const char *str);
-void btext_drawhex(unsigned long v);
+extern void btext_drawchar(char c);
+extern void btext_drawstring(const char *str);
+extern void btext_drawhex(unsigned long v);
 
 #endif /* __KERNEL__ */
 #endif /* __PPC_BTEXT_H */