]> git.saurik.com Git - apple/xnu.git/blobdiff - osfmk/ppc/commpage/commpage.c
xnu-1228.3.13.tar.gz
[apple/xnu.git] / osfmk / ppc / commpage / commpage.c
index 907ae701d00dd40b5e2453eaf84b77d3767e1c3b..6b02273225b7abb9bd491b24811dc0c2ee3fd20d 100644 (file)
@@ -1,16 +1,19 @@
 /*
- * Copyright (c) 2003 Apple Computer, Inc. All rights reserved.
+ * Copyright (c) 2003-2007 Apple Inc. All rights reserved.
  *
- * @APPLE_LICENSE_HEADER_START@
- * 
- * Copyright (c) 1999-2003 Apple Computer, Inc.  All Rights Reserved.
+ * @APPLE_OSREFERENCE_LICENSE_HEADER_START@
  * 
  * This file contains Original Code and/or Modifications of Original Code
  * as defined in and that are subject to the Apple Public Source License
  * Version 2.0 (the 'License'). You may not use this file except in
- * compliance with the License. Please obtain a copy of the License at
- * http://www.opensource.apple.com/apsl/ and read it before using this
- * file.
+ * compliance with the License. The rights granted to you under the License
+ * may not be used to create, or enable the creation or redistribution of,
+ * unlawful or unlicensed copies of an Apple operating system, or to
+ * circumvent, violate, or enable the circumvention or violation of, any
+ * terms of an Apple operating system software license agreement.
+ * 
+ * Please obtain a copy of the License at
+ * http://www.opensource.apple.com/apsl/ and read it before using this file.
  * 
  * The Original Code and all software distributed under the License are
  * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
  * Please see the License for the specific language governing rights and
  * limitations under the License.
  * 
- * @APPLE_LICENSE_HEADER_END@
+ * @APPLE_OSREFERENCE_LICENSE_HEADER_END@
  */
 
 /*
- *  This is a simplifed version of the commpage support from 10.3.
- *  The supported feature is the tuning of _cpu_capabilities.
- *  There is no shared page for user processes.
+ *     Here's what to do if you want to add a new routine to the comm page:
+ *
+ *             1. Add a definition for it's address in osfmk/ppc/cpu_capabilities.h,
+ *                being careful to reserve room for future expansion.
+ *
+ *             2. Write one or more versions of the routine, each with it's own
+ *                commpage_descriptor.  The tricky part is getting the "special",
+ *                "musthave", and "canthave" fields right, so that exactly one
+ *                version of the routine is selected for every machine.
+ *                The source files should be in osfmk/ppc/commpage/.
+ *
+ *             3. Add a ptr to your new commpage_descriptor(s) in the "routines"
+ *                static array below.  Of course, you'll also have to declare them 
+ *                "extern".
+ *
+ *             4. Write the code in Libc to use the new routine.
  */
 
 #include <mach/mach_types.h>
 #include <mach/machine.h>
+#include <mach/vm_map.h>
 #include <ppc/exception.h>
 #include <ppc/machine_routines.h>
 #include <machine/cpu_capabilities.h>
 #include <machine/commpage.h>
+#include <machine/pmap.h>
+#include <vm/vm_kern.h>
+#include <vm/vm_map.h>
+#include <ipc/ipc_port.h>
+
+extern vm_map_t        commpage32_map;   // the 32-bit shared submap, set up in vm init
+extern  vm_map_t       commpage64_map;   // the 64-bit shared submap
 
+char   *commPagePtr32 = NULL;                  // virtual address of 32-bit comm page in kernel map
+char   *commPagePtr64 = NULL;                  // and 64-bit commpage
 int            _cpu_capabilities = 0;                  // define the capability vector
 
+static char    *next;                                  // next available byte in comm page
+static int             cur_routine;                    // comm page address of "current" routine
+static int             matched;                                // true if we've found a match for "current" routine
+static  char   *commPagePtr;                   // virtual address in kernel of commpage we are working on
+
+extern commpage_descriptor     compare_and_swap32_on32;
+extern commpage_descriptor     compare_and_swap32_on64;
+extern commpage_descriptor     compare_and_swap64;
+extern commpage_descriptor     atomic_enqueue32;
+extern commpage_descriptor     atomic_enqueue64;
+extern commpage_descriptor     atomic_dequeue32_on32;
+extern commpage_descriptor     atomic_dequeue32_on64;
+extern commpage_descriptor     atomic_dequeue64;
+extern commpage_descriptor     memory_barrier_up;
+extern commpage_descriptor     memory_barrier_mp32;
+extern commpage_descriptor     memory_barrier_mp64;
+extern commpage_descriptor     atomic_add32;
+extern commpage_descriptor     atomic_add64;
+extern commpage_descriptor     mach_absolute_time_32;
+extern commpage_descriptor     mach_absolute_time_64;
+extern commpage_descriptor     mach_absolute_time_lp64;
+extern commpage_descriptor     spinlock_32_try_mp;
+extern commpage_descriptor     spinlock_32_try_up;
+extern commpage_descriptor     spinlock_64_try_mp;
+extern commpage_descriptor     spinlock_64_try_up;
+extern commpage_descriptor     spinlock_32_lock_mp;
+extern commpage_descriptor     spinlock_32_lock_up;
+extern commpage_descriptor     spinlock_64_lock_mp;
+extern commpage_descriptor     spinlock_64_lock_up;
+extern commpage_descriptor     spinlock_32_unlock_mp;
+extern commpage_descriptor     spinlock_32_unlock_up;
+extern commpage_descriptor     spinlock_64_unlock_mp;
+extern commpage_descriptor     spinlock_64_unlock_up;
+extern commpage_descriptor     pthread_getspecific_sprg3_32;
+extern commpage_descriptor     pthread_getspecific_sprg3_64;
+extern commpage_descriptor     pthread_getspecific_uftrap;
+extern commpage_descriptor     gettimeofday_32;
+extern commpage_descriptor     gettimeofday_g5_32;
+extern commpage_descriptor     gettimeofday_g5_64;
+extern commpage_descriptor     commpage_flush_dcache;
+extern commpage_descriptor     commpage_flush_icache;
+extern commpage_descriptor     pthread_self_sprg3;
+extern commpage_descriptor     pthread_self_uftrap;
+extern commpage_descriptor     spinlock_relinquish;
+extern commpage_descriptor     bzero_32;
+extern commpage_descriptor     bzero_128;
+extern commpage_descriptor     bcopy_g3;
+extern commpage_descriptor     bcopy_g4;
+extern commpage_descriptor     bcopy_970;
+extern commpage_descriptor     bcopy_64;
+extern commpage_descriptor     compare_and_swap32_on32b;
+extern commpage_descriptor     compare_and_swap32_on64b;
+extern commpage_descriptor     compare_and_swap64b;
+extern  commpage_descriptor memset_64;
+extern  commpage_descriptor memset_g3;
+extern  commpage_descriptor memset_g4;
+extern  commpage_descriptor memset_g5;
+extern commpage_descriptor     bigcopy_970;
+
+/* The list of all possible commpage routines.  WARNING: the check for overlap
+ * assumes that these routines are in strictly ascending order, sorted by address
+ * in the commpage.  We panic if not.
+ */
+static commpage_descriptor     *routines[] = {
+    &compare_and_swap32_on32,
+    &compare_and_swap32_on64,
+    &compare_and_swap64,
+    &atomic_enqueue32,
+    &atomic_enqueue64,
+    &atomic_dequeue32_on32,
+    &atomic_dequeue32_on64,
+    &atomic_dequeue64,
+    &memory_barrier_up,
+    &memory_barrier_mp32,
+    &memory_barrier_mp64,
+    &atomic_add32,
+    &atomic_add64,
+    &mach_absolute_time_32,
+    &mach_absolute_time_64,
+    &mach_absolute_time_lp64,
+    &spinlock_32_try_mp,
+    &spinlock_32_try_up,
+    &spinlock_64_try_mp,
+    &spinlock_64_try_up,
+    &spinlock_32_lock_mp,
+    &spinlock_32_lock_up,
+    &spinlock_64_lock_mp,
+    &spinlock_64_lock_up,
+    &spinlock_32_unlock_mp,
+    &spinlock_32_unlock_up,
+    &spinlock_64_unlock_mp,
+    &spinlock_64_unlock_up,
+    &pthread_getspecific_sprg3_32,
+    &pthread_getspecific_sprg3_64,
+    &pthread_getspecific_uftrap,
+    &gettimeofday_32,
+    &gettimeofday_g5_32,
+    &gettimeofday_g5_64,
+    &commpage_flush_dcache,
+    &commpage_flush_icache,
+    &pthread_self_sprg3,
+    &pthread_self_uftrap,
+    &spinlock_relinquish,
+    &bzero_32,
+    &bzero_128,
+    &bcopy_g3,
+    &bcopy_g4,
+    &bcopy_970,
+    &bcopy_64,
+    &compare_and_swap32_on32b,
+    &compare_and_swap32_on64b,
+    &compare_and_swap64b,
+    &memset_64,
+    &memset_g3,
+    &memset_g4,
+    &memset_g5,
+    &bigcopy_970,
+    NULL };
+
+
+/* Allocate the commpages and add to one of the shared submaps created by vm.
+ * Called once each for the 32 and 64-bit submaps.
+ *     1. allocate pages in the kernel map (RW)
+ *     2. wire them down
+ *     3. make a memory entry out of them
+ *     4. map that entry into the shared comm region map (R-only)
+ */
+static void*
+commpage_allocate( 
+       vm_map_t                        submap )                                        // commpage32_map or commpage64_map
+{
+    vm_offset_t                        kernel_addr = 0;                // address of commpage in kernel map
+    vm_offset_t                        zero = 0;
+    vm_size_t                  size = _COMM_PAGE_AREA_USED;    // size actually populated
+    vm_map_entry_t             entry;
+    ipc_port_t                 handle;
+    
+    if (submap == NULL)
+        panic("commpage submap is null");
+    
+    if (vm_map(kernel_map,&kernel_addr,_COMM_PAGE_AREA_USED,0,VM_FLAGS_ANYWHERE,NULL,0,FALSE,VM_PROT_ALL,VM_PROT_ALL,VM_INHERIT_NONE))
+        panic("cannot allocate commpage");
+        
+    if (vm_map_wire(kernel_map,kernel_addr,kernel_addr+_COMM_PAGE_AREA_USED,VM_PROT_DEFAULT,FALSE))
+        panic("cannot wire commpage");
+
+    /* 
+     * Now that the object is created and wired into the kernel map, mark it so that no delay
+     * copy-on-write will ever be performed on it as a result of mapping it into user-space.
+     * If such a delayed copy ever occurred, we could remove the kernel's wired mapping - and
+     * that would be a real disaster.
+     *
+     * JMM - What we really need is a way to create it like this in the first place.
+     */
+    if (!vm_map_lookup_entry( kernel_map, vm_map_trunc_page(kernel_addr), &entry) || entry->is_sub_map)
+        panic("cannot find commpage entry");
+    entry->object.vm_object->copy_strategy = MEMORY_OBJECT_COPY_NONE;
+       
+    if (mach_make_memory_entry(        kernel_map,                     // target map
+                                &size,                         // size
+                                kernel_addr,           // offset (address in kernel map)
+                                VM_PROT_ALL,           // map it RWX
+                                &handle,                       // this is the object handle we get
+                                NULL ))                                // parent_entry
+        panic("cannot make entry for commpage");
+    
+    if (vm_map_64(     submap,                                                 // target map (shared submap)
+                    &zero,                                                     // address (map into 1st page in submap)
+                    _COMM_PAGE_AREA_USED,                      // size
+                    0,                                                         // mask
+                    VM_FLAGS_FIXED,                                    // flags (it must be 1st page in submap)
+                    handle,                                                    // port is the memory entry we just made
+                    0,                                                         // offset (map 1st page in memory entry)
+                    FALSE,                                                     // copy
+                    VM_PROT_READ|VM_PROT_EXECUTE,                              // cur_protection (R-only in user map)
+                    VM_PROT_READ|VM_PROT_EXECUTE,                              // max_protection
+                    VM_INHERIT_SHARE ))                                // inheritance
+        panic("cannot map commpage");
+        
+    ipc_port_release(handle);
+        
+    return (void*) kernel_addr;                                                // return address in kernel map
+}
+
+
+/* Get address (in kernel map) of a commpage field. */
+
+static void*
+commpage_addr_of(
+    int        addr_at_runtime )
+{
+    return     (void*) (commPagePtr + addr_at_runtime - _COMM_PAGE_BASE_ADDRESS);
+}
+
+
 /* Determine number of CPUs on this system.  We cannot rely on
  * machine_info.max_cpus this early in the boot.
  */
@@ -62,13 +283,11 @@ commpage_cpus( void )
 static void
 commpage_init_cpu_capabilities( void )
 {
-    struct per_proc_info *pp;
     procFeatures       *pfp;
     int        cpus;
     int        available;
 
-    pp = per_proc_info;                                        // use CPU 0's per-proc
-    pfp = &pp->pf;                                             // point to features in per-proc
+    pfp = &(PerProcTable[0].ppe_vaddr->pf);                    // point to features in per-proc
     available = pfp->Available;
 
     // If AltiVec is disabled make sure it is not reported as available.
@@ -80,19 +299,381 @@ commpage_init_cpu_capabilities( void )
         _cpu_capabilities |= commpage_time_dcba();     // ...and set kDcbaRecomended if it helps.
     }
 
-    cpus = commpage_cpus();                            // how many CPUs do we have
+    cpus = commpage_cpus();                         // how many CPUs do we have
     if (cpus == 1) _cpu_capabilities |= kUP;
     _cpu_capabilities |= (cpus << kNumCPUsShift);
+
+    if (_cpu_capabilities & k64Bit)                 // 64-bit processors use SPRG3 for TLS
+        _cpu_capabilities |= kFastThreadLocalStorage;
+}
+
+
+/* Copy data into commpage. */
+
+static void
+commpage_stuff(
+    int         address,
+    const void         *source,
+    int         length )
+{    
+    char       *dest = commpage_addr_of(address);
+    
+    if (dest < next)
+        panic("commpage overlap: %p - %p", dest, next);
+    
+    bcopy((const char*)source,dest,length);
+    
+    next = (dest + length);
+}
+
+
+/* Modify commpage code in-place for this specific platform. */
+
+static void
+commpage_change(
+    uint32_t   *ptr,
+    int                bytes,
+    uint32_t   search_mask, 
+    uint32_t   search_pattern,
+    uint32_t   new_mask,
+    uint32_t   new_pattern,
+    int                        (*check)(uint32_t instruction)  )
+{
+    int                        words = bytes >> 2;
+    uint32_t   word;
+
+    while( (--words) >= 0 ) {
+        word = *ptr;
+        if ((word & search_mask)==search_pattern) {
+            if ((check==NULL) || (check(word))) {      // check instruction if necessary
+                word &= ~new_mask;
+                word |= new_pattern;
+                *ptr = word;
+            }
+        }
+        ptr++;
+    }
+}
+
+
+/* Check to see if exactly one bit is set in a MTCRF instruction's FXM field.
+ */
+static int
+commpage_onebit(
+    uint32_t   mtcrf )
+{
+    int x = (mtcrf >> 12) & 0xFF;              // isolate the FXM field of the MTCRF
+    
+    if (x==0)
+        panic("commpage bad mtcrf");
+        
+    return     (x & (x-1))==0 ? 1 : 0;         // return 1 iff exactly 1 bit set in FXM field
+}
+
+
+/* Check to see if a RLWINM (whose ME is 31) is a SRWI.  Since to shift right n bits
+ * you must "RLWINM ra,rs,32-n,n,31", if (SH+MB)==32 then we have a SRWI.
+ */
+static int
+commpage_srwi(
+       uint32_t        rlwinm )
+{
+       int                     sh = (rlwinm >> 11) & 0x1F;             // extract SH field of RLWINM, ie bits 16-20
+       int                     mb = (rlwinm >> 6 ) & 0x1F;             // extract MB field of RLWINM, ie bits 21-25
+       
+       return  (sh + mb) == 32;                                        // it is a SRWI if (SH+MB)==32
+}
+
+
+/* Handle kCommPageDCBA bit: the commpage routine uses DCBA.  If the machine we're
+ * running on doesn't benefit from use of that instruction, map them to NOPs
+ * in the commpage.
+ */
+static void
+commpage_handle_dcbas(
+    int        address,
+    int        length  )
+{
+    uint32_t   *ptr, search_mask, search, replace_mask, replace;
+    
+    if ( (_cpu_capabilities & kDcbaRecommended) == 0 ) {
+        ptr = commpage_addr_of(address);
+        
+        search_mask =  0xFC0007FE;             // search x-form opcode bits
+        search =               0x7C0005EC;             // for a DCBA
+        replace_mask =         0xFFFFFFFF;             // replace all bits...
+        replace =              0x60000000;             // ...with a NOP
+    
+        commpage_change(ptr,length,search_mask,search,replace_mask,replace,NULL);
+    }
+}
+
+
+/* Handle kCommPageSYNC bit: this routine uses SYNC, LWSYNC, or EIEIO.  If we're
+ * running on a UP machine, map them to NOPs.
+ */
+static void
+commpage_handle_syncs(
+    int        address, 
+    int        length  )
+{
+    uint32_t   *ptr, search_mask, search, replace_mask, replace;
+    
+    if (_NumCPUs() == 1) {
+        ptr = commpage_addr_of(address);
+        
+        search_mask =  0xFC0005FE;             // search x-form opcode bits (but ignore bit 0x00000200)
+        search =               0x7C0004AC;             // for a SYNC, LWSYNC, or EIEIO
+        replace_mask =         0xFFFFFFFF;             // replace all bits...
+        replace =              0x60000000;             // ...with a NOP
+    
+        commpage_change(ptr,length,search_mask,search,replace_mask,replace,NULL);
+    }
+}
+
+
+/* Handle kCommPageISYNC bit: this routine uses ISYNCs.  If we're running on a UP machine,
+ * map them to NOPs.
+ */
+static void
+commpage_handle_isyncs(
+    int        address, 
+    int        length  )
+{
+    uint32_t   *ptr, search_mask, search, replace_mask, replace;
+    
+    if (_NumCPUs() == 1) {
+        ptr = commpage_addr_of(address);
+        
+        search_mask =  0xFC0007FE;             // search xl-form opcode bits
+        search =               0x4C00012C;             // for an ISYNC
+        replace_mask =         0xFFFFFFFF;             // replace all bits...
+        replace =              0x60000000;             // ...with a NOP
+    
+        commpage_change(ptr,length,search_mask,search,replace_mask,replace,NULL);
+    }
+}
+
+
+/* Handle kCommPageMTCRF bit.  When this was written (3/03), the assembler did not
+ * recognize the special form of MTCRF instructions, in which exactly one bit is set
+ * in the 8-bit mask field.  Bit 11 of the instruction should be set in this case,
+ * since the 970 and probably other 64-bit processors optimize it.  Once the assembler
+ * has been updated this code can be removed, though it need not be.
+ */
+static void
+commpage_handle_mtcrfs(
+    int        address, 
+    int        length  )
+{
+    uint32_t   *ptr, search_mask, search, replace_mask, replace;
+    
+    if (_cpu_capabilities & k64Bit) {
+        ptr = commpage_addr_of(address);
+        
+        search_mask =  0xFC0007FE;             // search x-form opcode bits
+        search =               0x7C000120;             // for a MTCRF
+        replace_mask =         0x00100000;             // replace bit 11...
+        replace =              0x00100000;             // ...with a 1-bit
+    
+        commpage_change(ptr,length,search_mask,search,replace_mask,replace,commpage_onebit);
+    }
+}
+
+
+/* Port 32-bit code to 64-bit for use in the 64-bit commpage.  This sounds fancier than
+ * it is.  We do the following:
+ *             - map "cmpw*" into "cmpd*"
+ *             - map "srwi" into "srdi"
+ * Perhaps surprisingly, this is enough to permit lots of code to run in 64-bit mode, as
+ * long as it is written with this in mind.
+ */
+static void
+commpage_port_32_to_64(
+    int        address, 
+    int        length  )
+{
+    uint32_t   *ptr, search_mask, search, replace_mask, replace;
+
+       ptr = commpage_addr_of(address);
+       
+       search_mask =   0xFC2007FE;             // search x-form opcode bits (and L bit)
+       search =                0x7C000000;             // for a CMPW
+       replace_mask =  0x00200000;             // replace bit 10 (L)...
+       replace =               0x00200000;             // ...with a 1-bit, converting word to doubleword compares
+       commpage_change(ptr,length,search_mask,search,replace_mask,replace,NULL);
+
+       search_mask =   0xFC2007FE;             // search x-form opcode bits (and L bit)
+       search =                0x7C000040;             // for a CMPLW
+       replace_mask =  0x00200000;             // replace bit 10 (L)...
+       replace =               0x00200000;             // ...with a 1-bit, converting word to doubleword compares
+       commpage_change(ptr,length,search_mask,search,replace_mask,replace,NULL);
+
+       search_mask =   0xFC200000;             // search d-form opcode bits (and L bit)
+       search =                0x28000000;             // for a CMPLWI
+       replace_mask =  0x00200000;             // replace bit 10 (L)...
+       replace =               0x00200000;             // ...with a 1-bit, converting word to doubleword compares
+       commpage_change(ptr,length,search_mask,search,replace_mask,replace,NULL);
+
+       search_mask =   0xFC200000;             // search d-form opcode bits (and L bit)
+       search =                0x2C000000;             // for a CMPWI
+       replace_mask =  0x00200000;             // replace bit 10 (L)...
+       replace =               0x00200000;             // ...with a 1-bit, converting word to doubleword compares
+       commpage_change(ptr,length,search_mask,search,replace_mask,replace,NULL);
+       
+       search_mask =   0xFC00003E;             // search d-form opcode bits and ME (mask end) field
+       search =                0x5400003E;             // for an RLWINM with ME=31 (which might be a "srwi")
+       replace_mask =  0xFC00003E;             // then replace RLWINM's opcode and ME field to make a RLDICL
+       replace =               0x78000002;             // opcode is 30, ME is 0, except we add 32 to SH amount
+       commpage_change(ptr,length,search_mask,search,replace_mask,replace,commpage_srwi);
+} 
+
+
+/* Copy a routine into comm page if it matches running machine.
+ */
+static void
+commpage_stuff_routine(
+    commpage_descriptor        *rd,
+       int                                     mode )                          // kCommPage32 or kCommPage64
+{
+    char       *routine_code;
+    int                must,cant;
+       
+       if ( (rd->special & mode) == 0 )                // is this routine useable in this mode?
+               return;
+    
+    if (rd->commpage_address != cur_routine) {
+        if ((cur_routine!=0) && (matched==0))
+            panic("commpage no match for last, next address %08x", rd->commpage_address);
+        cur_routine = rd->commpage_address;
+        matched = 0;
+    }
+    
+    must = _cpu_capabilities & rd->musthave;
+    cant = _cpu_capabilities & rd->canthave;
+    
+    if ((must == rd->musthave) && (cant == 0)) {
+        if (matched)
+            panic("commpage multiple matches for address %08x", rd->commpage_address);
+        matched = 1;
+        routine_code = ((char*)rd) + rd->code_offset;
+        
+        commpage_stuff(rd->commpage_address,routine_code,rd->code_length);
+        
+        if (rd->special & kCommPageDCBA)
+            commpage_handle_dcbas(rd->commpage_address,rd->code_length);
+            
+        if (rd->special & kCommPageSYNC)
+            commpage_handle_syncs(rd->commpage_address,rd->code_length);
+            
+        if (rd->special & kCommPageISYNC)
+            commpage_handle_isyncs(rd->commpage_address,rd->code_length);
+            
+        if (rd->special & kCommPageMTCRF)
+            commpage_handle_mtcrfs(rd->commpage_address,rd->code_length);
+                       
+               if ((mode == kCommPage64) && (rd->special & kPort32to64))
+                       commpage_port_32_to_64(rd->commpage_address,rd->code_length);
+    }
+}
+
+
+/* Fill in the 32- or 64-bit commpage.  Called once for each.  */
+
+static void
+commpage_populate_one( 
+       vm_map_t        submap,                 // the map to populate
+       char    **  kernAddressPtr,     // address within kernel of this commpage
+       int                     mode,           // either kCommPage32 or kCommPage64
+    const char* signature )     // "commpage 32-bit" or "commpage 64-bit"
+{
+    char       c1;
+    short      c2;
+    addr64_t c8;
+    static double      two52 = 1048576.0 * 1048576.0 * 4096.0; // 2**52
+    static double      ten6 = 1000000.0;                                               // 10**6
+    static uint64_t magicFE = 0xFEFEFEFEFEFEFEFFLL;         // used to find 0s in strings
+    static uint64_t magic80 = 0x8080808080808080LL;         // also used to find 0s
+    commpage_descriptor        **rd;
+    short      version = _COMM_PAGE_THIS_VERSION;
+    
+    next = NULL;                                                               // initialize next available byte in the commpage
+       cur_routine = 0;                                                        // initialize comm page address of "current" routine
+       
+    commPagePtr = (char*) commpage_allocate( submap );
+    *kernAddressPtr = commPagePtr;                             // save address either in commPagePtr32 or 64
+
+    /* Stuff in the constants.  We move things into the comm page in strictly
+     * ascending order, so we can check for overlap and panic if so.
+     */
+    
+    commpage_stuff(_COMM_PAGE_SIGNATURE,signature,strlen(signature));
+    
+    commpage_stuff(_COMM_PAGE_VERSION,&version,2);
+
+    commpage_stuff(_COMM_PAGE_CPU_CAPABILITIES,&_cpu_capabilities,sizeof(int));
+    
+    c1 = (_cpu_capabilities & kHasAltivec) ? -1 : 0;
+    commpage_stuff(_COMM_PAGE_ALTIVEC,&c1,1);
+    
+    c1 = (_cpu_capabilities & k64Bit) ? -1 : 0;
+    commpage_stuff(_COMM_PAGE_64_BIT,&c1,1);
+    
+    if (_cpu_capabilities & kCache32)
+        c2 = 32;
+    else if (_cpu_capabilities & kCache64)
+        c2 = 64;
+    else if (_cpu_capabilities & kCache128)
+        c2 = 128;
+    commpage_stuff(_COMM_PAGE_CACHE_LINESIZE,&c2,2);
+    
+    commpage_stuff(_COMM_PAGE_2_TO_52,&two52,8);    
+    commpage_stuff(_COMM_PAGE_10_TO_6,&ten6,8);
+    commpage_stuff(_COMM_PAGE_MAGIC_FE,&magicFE,8);
+    commpage_stuff(_COMM_PAGE_MAGIC_80,&magic80,8);
+    
+    c8 = 0;                                                                                                    // 0 timestamp means "disabled"
+    commpage_stuff(_COMM_PAGE_TIMEBASE,&c8,8);
+    commpage_stuff(_COMM_PAGE_TIMESTAMP,&c8,8);
+    commpage_stuff(_COMM_PAGE_SEC_PER_TICK,&c8,8);
+
+    /* Now the routines.  We try each potential routine in turn,
+     * and copy in any that "match" the platform we are running on.
+     * We require that exactly one routine match for each slot in the
+     * comm page, and panic if not.
+     */
+        
+    for( rd = routines; *rd != NULL ; rd++ ) 
+        commpage_stuff_routine(*rd,mode);
+        
+    if (!matched)
+        panic("commpage no match on last routine");
+    
+    if (next > (commPagePtr + _COMM_PAGE_AREA_USED))
+        panic("commpage overflow");
+       
+       
+       // make all that new code executable
+       
+    sync_cache_virtual((vm_offset_t) commPagePtr,_COMM_PAGE_AREA_USED);
 }
 
 
 /* Fill in commpage: called once, during kernel initialization, from the
  * startup thread before user-mode code is running.
+ *
  * See the top of this file for a list of what you have to do to add
  * a new routine to the commpage.
  */  
+
 void
 commpage_populate( void )
 {
     commpage_init_cpu_capabilities();
+       commpage_populate_one( commpage32_map, &commPagePtr32, kCommPage32, "commpage 32-bit");
+       if (_cpu_capabilities & k64Bit) {
+               commpage_populate_one( commpage64_map, &commPagePtr64, kCommPage64, "commpage 64-bit");
+               pmap_init_sharedpage((vm_offset_t)commPagePtr64);                       // Do the 64-bit version        
+       }
+        
 }