12 files changed, 102 insertions, 115 deletions

diff --git a/include/asm-sparc64/cacheflush.h b/include/asm-sparc64/cacheflush.h
index 51b26e81d828..b3f61659ba81 100644
--- a/include/asm-sparc64/cacheflush.h
+++ b/include/asm-sparc64/cacheflush.h
@@ -4,13 +4,6 @@
 #include <linux/config.h>
 #include <asm/page.h>
 
-/* Flushing for D-cache alias handling is only needed if
- * the page size is smaller than 16K.
- */
-#if PAGE_SHIFT < 14
-#define DCACHE_ALIASING_POSSIBLE
-#endif
-
 #ifndef __ASSEMBLY__
 
 #include <linux/mm.h>
@@ -73,6 +66,11 @@ extern void flush_ptrace_access(struct vm_area_struct *, struct page *,
 #define flush_cache_vmap(start, end)            do { } while (0)
 #define flush_cache_vunmap(start, end)          do { } while (0)
 
+#ifdef CONFIG_DEBUG_PAGEALLOC
+/* internal debugging function */
+void kernel_map_pages(struct page *page, int numpages, int enable);
+#endif
+
 #endif /* !__ASSEMBLY__ */
 
 #endif /* _SPARC64_CACHEFLUSH_H */
diff --git a/include/asm-sparc64/cpudata.h b/include/asm-sparc64/cpudata.h
index 9a3a81f1cc58..74de79dca915 100644
--- a/include/asm-sparc64/cpudata.h
+++ b/include/asm-sparc64/cpudata.h
@@ -22,6 +22,16 @@ typedef struct {
         unsigned int    __pad1;
         unsigned long   *pte_cache[2];
         unsigned long   *pgd_cache;
+
+        /* Dcache line 3, rarely used */
+        unsigned int    dcache_size;
+        unsigned int    dcache_line_size;
+        unsigned int    icache_size;
+        unsigned int    icache_line_size;
+        unsigned int    ecache_size;
+        unsigned int    ecache_line_size;
+        unsigned int    __pad2;
+        unsigned int    __pad3;
 } cpuinfo_sparc;
 
 DECLARE_PER_CPU(cpuinfo_sparc, __cpu_data);
diff --git a/include/asm-sparc64/futex.h b/include/asm-sparc64/futex.h
index 2cac5ecd9d00..9feff4ce1424 100644
--- a/include/asm-sparc64/futex.h
+++ b/include/asm-sparc64/futex.h
@@ -14,7 +14,7 @@ futex_atomic_op_inuser (int encoded_op, int __user *uaddr)
         int cmp = (encoded_op >> 24) & 15;
         int oparg = (encoded_op << 8) >> 20;
         int cmparg = (encoded_op << 20) >> 20;
-        int oldval = 0, ret, tem;
+        int oldval = 0, ret;
         if (encoded_op & (FUTEX_OP_OPARG_SHIFT << 28))
                 oparg = 1 << oparg;
 
diff --git a/include/asm-sparc64/head.h b/include/asm-sparc64/head.h
index b63a33cf4971..0abd3a674e8f 100644
--- a/include/asm-sparc64/head.h
+++ b/include/asm-sparc64/head.h
@@ -12,9 +12,12 @@
 #define __JALAPENO_ID   0x003e0016
 
 #define CHEETAH_MANUF           0x003e
-#define CHEETAH_IMPL            0x0014
-#define CHEETAH_PLUS_IMPL       0x0015
-#define JALAPENO_IMPL           0x0016
+#define CHEETAH_IMPL            0x0014 /* Ultra-III   */
+#define CHEETAH_PLUS_IMPL       0x0015 /* Ultra-III+  */
+#define JALAPENO_IMPL           0x0016 /* Ultra-IIIi  */
+#define JAGUAR_IMPL             0x0018 /* Ultra-IV    */
+#define PANTHER_IMPL            0x0019 /* Ultra-IV+   */
+#define SERRANO_IMPL            0x0022 /* Ultra-IIIi+ */
 
 #define BRANCH_IF_CHEETAH_BASE(tmp1,tmp2,label) \
         rdpr    %ver, %tmp1;                    \
diff --git a/include/asm-sparc64/ide.h b/include/asm-sparc64/ide.h
index 4c1098474c73..c393f815b0be 100644
--- a/include/asm-sparc64/ide.h
+++ b/include/asm-sparc64/ide.h
@@ -15,6 +15,7 @@
 #include <asm/io.h>
 #include <asm/spitfire.h>
 #include <asm/cacheflush.h>
+#include <asm/page.h>
 
 #ifndef MAX_HWIFS
 # ifdef CONFIG_BLK_DEV_IDEPCI
diff --git a/include/asm-sparc64/openprom.h b/include/asm-sparc64/openprom.h
index 0a336901d585..b4959d2b0d99 100644
--- a/include/asm-sparc64/openprom.h
+++ b/include/asm-sparc64/openprom.h
@@ -186,8 +186,8 @@ struct linux_prom_registers {
 };
 
 struct linux_prom64_registers {
-        long phys_addr;
-        long reg_size;
+        unsigned long phys_addr;
+        unsigned long reg_size;
 };
 
 struct linux_prom_irqs {
diff --git a/include/asm-sparc64/oplib.h b/include/asm-sparc64/oplib.h
index a432d9e7daaa..d02f1e8ae1a6 100644
--- a/include/asm-sparc64/oplib.h
+++ b/include/asm-sparc64/oplib.h
@@ -38,6 +38,20 @@ extern int prom_stdin, prom_stdout;
  */
 extern int prom_chosen_node;
 
+/* Helper values and strings in arch/sparc64/kernel/head.S */
+extern const char prom_finddev_name[];
+extern const char prom_chosen_path[];
+extern const char prom_getprop_name[];
+extern const char prom_mmu_name[];
+extern const char prom_callmethod_name[];
+extern const char prom_translate_name[];
+extern const char prom_map_name[];
+extern const char prom_unmap_name[];
+extern int prom_mmu_ihandle_cache;
+extern unsigned int prom_boot_mapped_pc;
+extern unsigned int prom_boot_mapping_mode;
+extern unsigned long prom_boot_mapping_phys_high, prom_boot_mapping_phys_low;
+
 struct linux_mlist_p1275 {
         struct linux_mlist_p1275 *theres_more;
         unsigned long start_adr;
@@ -68,7 +82,7 @@ extern char *prom_getbootargs(void);
  * of the string is different on V0 vs. V2->higher proms.  The caller must
  * know what he/she is doing!  Returns the device descriptor, an int.
  */
-extern int prom_devopen(char *device_string);
+extern int prom_devopen(const char *device_string);
 
 /* Close a previously opened device described by the passed integer
  * descriptor.
@@ -81,27 +95,13 @@ extern int prom_devclose(int device_handle);
 extern void prom_seek(int device_handle, unsigned int seek_hival,
                       unsigned int seek_lowval);
 
-/* Machine memory configuration routine. */
-
-/* This function returns a V0 format memory descriptor table, it has three
- * entries.  One for the total amount of physical ram on the machine, one
- * for the amount of physical ram available, and one describing the virtual
- * areas which are allocated by the prom.  So, in a sense the physical
- * available is a calculation of the total physical minus the physical mapped
- * by the prom with virtual mappings.
- *
- * These lists are returned pre-sorted, this should make your life easier
- * since the prom itself is way too lazy to do such nice things.
- */
-extern struct linux_mem_p1275 *prom_meminfo(void);
-
 /* Miscellaneous routines, don't really fit in any category per se. */
 
 /* Reboot the machine with the command line passed. */
-extern void prom_reboot(char *boot_command);
+extern void prom_reboot(const char *boot_command);
 
 /* Evaluate the forth string passed. */
-extern void prom_feval(char *forth_string);
+extern void prom_feval(const char *forth_string);
 
 /* Enter the prom, with possibility of continuation with the 'go'
  * command in newer proms.
@@ -154,7 +154,7 @@ extern char prom_getchar(void);
 extern void prom_putchar(char character);
 
 /* Prom's internal routines, don't use in kernel/boot code. */
-extern void prom_printf(char *fmt, ...);
+extern void prom_printf(const char *fmt, ...);
 extern void prom_write(const char *buf, unsigned int len);
 
 /* Query for input device type */
@@ -215,7 +215,7 @@ extern int prom_getunumber(int syndrome_code,
                            char *buf, int buflen);
 
 /* Retain physical memory to the caller across soft resets. */
-extern unsigned long prom_retain(char *name,
+extern unsigned long prom_retain(const char *name,
                                  unsigned long pa_low, unsigned long pa_high,
                                  long size, long align);
 
@@ -269,28 +269,28 @@ extern int prom_getsibling(int node);
 /* Get the length, at the passed node, of the given property type.
  * Returns -1 on error (ie. no such property at this node).
  */
-extern int prom_getproplen(int thisnode, char *property);
+extern int prom_getproplen(int thisnode, const char *property);
 
 /* Fetch the requested property using the given buffer.  Returns
  * the number of bytes the prom put into your buffer or -1 on error.
  */
-extern int prom_getproperty(int thisnode, char *property,
+extern int prom_getproperty(int thisnode, const char *property,
                             char *prop_buffer, int propbuf_size);
 
 /* Acquire an integer property. */
-extern int prom_getint(int node, char *property);
+extern int prom_getint(int node, const char *property);
 
 /* Acquire an integer property, with a default value. */
-extern int prom_getintdefault(int node, char *property, int defval);
+extern int prom_getintdefault(int node, const char *property, int defval);
 
 /* Acquire a boolean property, 0=FALSE 1=TRUE. */
-extern int prom_getbool(int node, char *prop);
+extern int prom_getbool(int node, const char *prop);
 
 /* Acquire a string property, null string on error. */
-extern void prom_getstring(int node, char *prop, char *buf, int bufsize);
+extern void prom_getstring(int node, const char *prop, char *buf, int bufsize);
 
 /* Does the passed node have the given "name"? YES=1 NO=0 */
-extern int prom_nodematch(int thisnode, char *name);
+extern int prom_nodematch(int thisnode, const char *name);
 
 /* Puts in buffer a prom name in the form name@x,y or name (x for which_io 
  * and y for first regs phys address
@@ -300,7 +300,7 @@ extern int prom_getname(int node, char *buf, int buflen);
 /* Search all siblings starting at the passed node for "name" matching
  * the given string.  Returns the node on success, zero on failure.
  */
-extern int prom_searchsiblings(int node_start, char *name);
+extern int prom_searchsiblings(int node_start, const char *name);
 
 /* Return the first property type, as a string, for the given node.
  * Returns a null string on error. Buffer should be at least 32B long.
@@ -310,21 +310,21 @@ extern char *prom_firstprop(int node, char *buffer);
 /* Returns the next property after the passed property for the given
  * node.  Returns null string on failure. Buffer should be at least 32B long.
  */
-extern char *prom_nextprop(int node, char *prev_property, char *buffer);
+extern char *prom_nextprop(int node, const char *prev_property, char *buffer);
 
 /* Returns 1 if the specified node has given property. */
-extern int prom_node_has_property(int node, char *property);
+extern int prom_node_has_property(int node, const char *property);
 
 /* Returns phandle of the path specified */
-extern int prom_finddevice(char *name);
+extern int prom_finddevice(const char *name);
 
 /* Set the indicated property at the given node with the passed value.
  * Returns the number of bytes of your value that the prom took.
  */
-extern int prom_setprop(int node, char *prop_name, char *prop_value,
+extern int prom_setprop(int node, const char *prop_name, char *prop_value,
                         int value_size);
                         
-extern int prom_pathtoinode(char *path);
+extern int prom_pathtoinode(const char *path);
 extern int prom_inst2pkg(int);
 
 /* CPU probing helpers.  */
@@ -334,7 +334,7 @@ int cpu_find_by_mid(int mid, int *prom_node);
 /* Client interface level routines. */
 extern void prom_set_trap_table(unsigned long tba);
 
-extern long p1275_cmd (char *, long, ...);
+extern long p1275_cmd(const char *, long, ...);
                                    
 
 #if 0
diff --git a/include/asm-sparc64/page.h b/include/asm-sparc64/page.h
index c9f8ef208ea5..5426bb28a993 100644
--- a/include/asm-sparc64/page.h
+++ b/include/asm-sparc64/page.h
@@ -21,6 +21,13 @@
 #define PAGE_SIZE    (_AC(1,UL) << PAGE_SHIFT)
 #define PAGE_MASK    (~(PAGE_SIZE-1))
 
+/* Flushing for D-cache alias handling is only needed if
+ * the page size is smaller than 16K.
+ */
+#if PAGE_SHIFT < 14
+#define DCACHE_ALIASING_POSSIBLE
+#endif
+
 #ifdef __KERNEL__
 
 #ifndef __ASSEMBLY__
@@ -133,23 +140,6 @@ extern unsigned long page_to_pfn(struct page *);
 #define virt_to_phys __pa
 #define phys_to_virt __va
 
-/* The following structure is used to hold the physical
- * memory configuration of the machine.  This is filled in
- * probe_memory() and is later used by mem_init() to set up
- * mem_map[].  We statically allocate SPARC_PHYS_BANKS of
- * these structs, this is arbitrary.  The entry after the
- * last valid one has num_bytes==0.
- */
-
-struct sparc_phys_banks {
-        unsigned long base_addr;
-        unsigned long num_bytes;
-};
-
-#define SPARC_PHYS_BANKS 32
-
-extern struct sparc_phys_banks sp_banks[SPARC_PHYS_BANKS];
-
 #endif /* !(__ASSEMBLY__) */
 
 #define VM_DATA_DEFAULT_FLAGS   (VM_READ | VM_WRITE | VM_EXEC | \
diff --git a/include/asm-sparc64/pbm.h b/include/asm-sparc64/pbm.h
index 38bbbccb4068..dd35a2c7798a 100644
--- a/include/asm-sparc64/pbm.h
+++ b/include/asm-sparc64/pbm.h
@@ -27,23 +27,27 @@
  * PCI bus.
  */
 
-#define PBM_LOGCLUSTERS 3
-#define PBM_NCLUSTERS (1 << PBM_LOGCLUSTERS)
-
 struct pci_controller_info;
 
 /* This contains the software state necessary to drive a PCI
  * controller's IOMMU.
  */
+struct pci_iommu_arena {
+        unsigned long   *map;
+        unsigned int    hint;
+        unsigned int    limit;
+};
+
 struct pci_iommu {
         /* This protects the controller's IOMMU and all
          * streaming buffers underneath.
          */
         spinlock_t      lock;
 
+        struct pci_iommu_arena arena;
+
         /* IOMMU page table, a linear array of ioptes. */
         iopte_t         *page_table;            /* The page table itself. */
-        int             page_table_sz_bits;     /* log2 of ow many pages does it map? */
 
         /* Base PCI memory space address where IOMMU mappings
          * begin.
@@ -62,12 +66,6 @@ struct pci_iommu {
          */
         unsigned long   write_complete_reg;
 
-        /* The lowest used consistent mapping entry.  Since
-         * we allocate consistent maps out of cluster 0 this
-         * is relative to the beginning of closter 0.
-         */
-        u32             lowest_consistent_map;
-
         /* In order to deal with some buggy third-party PCI bridges that
          * do wrong prefetching, we never mark valid mappings as invalid.
          * Instead we point them at this dummy page.
@@ -75,16 +73,6 @@ struct pci_iommu {
         unsigned long   dummy_page;
         unsigned long   dummy_page_pa;
 
-        /* If PBM_NCLUSTERS is ever decreased to 4 or lower,
-         * or if largest supported page_table_sz * 8K goes above
-         * 2GB, you must increase the size of the type of
-         * these counters.  You have been duly warned. -DaveM
-         */
-        struct {
-                u16     next;
-                u16     flush;
-        } alloc_info[PBM_NCLUSTERS];
-
         /* CTX allocation. */
         unsigned long ctx_lowest_free;
         unsigned long ctx_bitmap[IOMMU_NUM_CTXS / (sizeof(unsigned long) * 8)];
@@ -102,7 +90,7 @@ struct pci_iommu {
         u32 dma_addr_mask;
 };
 
-extern void pci_iommu_table_init(struct pci_iommu *, int);
+extern void pci_iommu_table_init(struct pci_iommu *iommu, int tsbsize, u32 dma_offset, u32 dma_addr_mask);
 
 /* This describes a PCI bus module's streaming buffer. */
 struct pci_strbuf {
diff --git a/include/asm-sparc64/pgalloc.h b/include/asm-sparc64/pgalloc.h
index b9b1914aae63..a96067cca963 100644
--- a/include/asm-sparc64/pgalloc.h
+++ b/include/asm-sparc64/pgalloc.h
@@ -10,6 +10,7 @@
 #include <asm/spitfire.h>
 #include <asm/cpudata.h>
 #include <asm/cacheflush.h>
+#include <asm/page.h>
 
 /* Page table allocation/freeing. */
 #ifdef CONFIG_SMP
diff --git a/include/asm-sparc64/pgtable.h b/include/asm-sparc64/pgtable.h
index a2b4f5ed4625..8c6dfc6c7af6 100644
--- a/include/asm-sparc64/pgtable.h
+++ b/include/asm-sparc64/pgtable.h
@@ -24,21 +24,23 @@
 #include <asm/processor.h>
 #include <asm/const.h>
 
-/* The kernel image occupies 0x4000000 to 0x1000000 (4MB --> 16MB).
- * The page copy blockops use 0x1000000 to 0x18000000 (16MB --> 24MB).
+/* The kernel image occupies 0x4000000 to 0x1000000 (4MB --> 32MB).
+ * The page copy blockops can use 0x2000000 to 0x10000000.
  * The PROM resides in an area spanning 0xf0000000 to 0x100000000.
- * The vmalloc area spans 0x140000000 to 0x200000000.
+ * The vmalloc area spans 0x100000000 to 0x200000000.
+ * Since modules need to be in the lowest 32-bits of the address space,
+ * we place them right before the OBP area from 0x10000000 to 0xf0000000.
  * There is a single static kernel PMD which maps from 0x0 to address
  * 0x400000000.
  */
-#define TLBTEMP_BASE            _AC(0x0000000001000000,UL)
-#define MODULES_VADDR           _AC(0x0000000002000000,UL)
-#define MODULES_LEN             _AC(0x000000007e000000,UL)
-#define MODULES_END             _AC(0x0000000080000000,UL)
-#define VMALLOC_START           _AC(0x0000000140000000,UL)
-#define VMALLOC_END             _AC(0x0000000200000000,UL)
+#define TLBTEMP_BASE            _AC(0x0000000002000000,UL)
+#define MODULES_VADDR           _AC(0x0000000010000000,UL)
+#define MODULES_LEN             _AC(0x00000000e0000000,UL)
+#define MODULES_END             _AC(0x00000000f0000000,UL)
 #define LOW_OBP_ADDRESS         _AC(0x00000000f0000000,UL)
 #define HI_OBP_ADDRESS          _AC(0x0000000100000000,UL)
+#define VMALLOC_START           _AC(0x0000000100000000,UL)
+#define VMALLOC_END             _AC(0x0000000200000000,UL)
 
 /* XXX All of this needs to be rethought so we can take advantage
  * XXX cheetah's full 64-bit virtual address space, ie. no more hole
@@ -58,13 +60,13 @@
  * table can map
  */
 #define PMD_SHIFT       (PAGE_SHIFT + (PAGE_SHIFT-3))
-#define PMD_SIZE        (1UL << PMD_SHIFT)
+#define PMD_SIZE        (_AC(1,UL) << PMD_SHIFT)
 #define PMD_MASK        (~(PMD_SIZE-1))
 #define PMD_BITS        (PAGE_SHIFT - 2)
 
 /* PGDIR_SHIFT determines what a third-level page table entry can map */
 #define PGDIR_SHIFT     (PAGE_SHIFT + (PAGE_SHIFT-3) + PMD_BITS)
-#define PGDIR_SIZE      (1UL << PGDIR_SHIFT)
+#define PGDIR_SIZE      (_AC(1,UL) << PGDIR_SHIFT)
 #define PGDIR_MASK      (~(PGDIR_SIZE-1))
 #define PGDIR_BITS      (PAGE_SHIFT - 2)
 
@@ -96,7 +98,9 @@
 #define _PAGE_NFO       _AC(0x1000000000000000,UL) /* No Fault Only          */
 #define _PAGE_IE        _AC(0x0800000000000000,UL) /* Invert Endianness      */
 #define _PAGE_SOFT2     _AC(0x07FC000000000000,UL) /* Software bits, set 2   */
-#define _PAGE_RES1      _AC(0x0003000000000000,UL) /* Reserved               */
+#define _PAGE_RES1      _AC(0x0002000000000000,UL) /* Reserved               */
+#define _PAGE_SZ32MB    _AC(0x0001000000000000,UL) /* (Panther) 32MB page    */
+#define _PAGE_SZ256MB   _AC(0x2001000000000000,UL) /* (Panther) 256MB page   */
 #define _PAGE_SN        _AC(0x0000800000000000,UL) /* (Cheetah) Snoop        */
 #define _PAGE_RES2      _AC(0x0000780000000000,UL) /* Reserved               */
 #define _PAGE_PADDR_SF  _AC(0x000001FFFFFFE000,UL) /* (Spitfire) paddr[40:13]*/
@@ -334,7 +338,11 @@ static inline void set_pte_at(struct mm_struct *mm, unsigned long addr, pte_t *p
 #define pte_clear(mm,addr,ptep)         \
         set_pte_at((mm), (addr), (ptep), __pte(0UL))
 
-extern pgd_t swapper_pg_dir[1];
+extern pgd_t swapper_pg_dir[2048];
+extern pmd_t swapper_low_pmd_dir[2048];
+
+extern void paging_init(void);
+extern unsigned long find_ecache_flush_span(unsigned long size);
 
 /* These do nothing with the way I have things setup. */
 #define mmu_lockarea(vaddr, len)                (vaddr)
diff --git a/include/asm-sparc64/uaccess.h b/include/asm-sparc64/uaccess.h
index 80a65d7e3dbf..203e8eee6351 100644
--- a/include/asm-sparc64/uaccess.h
+++ b/include/asm-sparc64/uaccess.h
@@ -70,26 +70,14 @@ static inline int access_ok(int type, const void __user * addr, unsigned long si
  * with the main instruction path.  This means when everything is well,
  * we don't even have to jump over them.  Further, they do not intrude
  * on our cache or tlb entries.
- *
- * There is a special way how to put a range of potentially faulting
- * insns (like twenty ldd/std's with now intervening other instructions)
- * You specify address of first in insn and 0 in fixup and in the next
- * exception_table_entry you specify last potentially faulting insn + 1
- * and in fixup the routine which should handle the fault.
- * That fixup code will get
- * (faulting_insn_address - first_insn_in_the_range_address)/4
- * in %g2 (ie. index of the faulting instruction in the range).
  */
 
-struct exception_table_entry
-{
-        unsigned insn, fixup;
+struct exception_table_entry {
+        unsigned int insn, fixup;
 };
 
-/* Special exable search, which handles ranges.  Returns fixup */
-unsigned long search_extables_range(unsigned long addr, unsigned long *g2);
-
 extern void __ret_efault(void);
+extern void __retl_efault(void);
 
 /* Uh, these should become the main single-value transfer routines..
  * They automatically use the right size if we just have the right
@@ -263,7 +251,7 @@ copy_from_user(void *to, const void __user *from, unsigned long size)
 {
         unsigned long ret = ___copy_from_user(to, from, size);
 
-        if (ret)
+        if (unlikely(ret))
                 ret = copy_from_user_fixup(to, from, size);
         return ret;
 }
@@ -279,7 +267,7 @@ copy_to_user(void __user *to, const void *from, unsigned long size)
 {
         unsigned long ret = ___copy_to_user(to, from, size);
 
-        if (ret)
+        if (unlikely(ret))
                 ret = copy_to_user_fixup(to, from, size);
         return ret;
 }
@@ -295,7 +283,7 @@ copy_in_user(void __user *to, void __user *from, unsigned long size)
 {
         unsigned long ret = ___copy_in_user(to, from, size);
 
-        if (ret)
+        if (unlikely(ret))
                 ret = copy_in_user_fixup(to, from, size);
         return ret;
 }