Merge git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc-merge

Documentation/filesystems/spufs.txt

@@ -0,0 +1,521 @@
+SPUFS(2)                   Linux Programmer's Manual                  SPUFS(2)
+
+
+
+NAME
+       spufs - the SPU file system
+
+
+DESCRIPTION
+       The SPU file system is used on PowerPC machines that implement the Cell
+       Broadband Engine Architecture in order to access Synergistic  Processor
+       Units (SPUs).
+
+       The file system provides a name space similar to posix shared memory or
+       message queues. Users that have write permissions on  the  file  system
+       can use spu_create(2) to establish SPU contexts in the spufs root.
+
+       Every SPU context is represented by a directory containing a predefined
+       set of files. These files can be used for manipulating the state of the
+       logical SPU. Users can change permissions on those files, but not actu-
+       ally add or remove files.
+
+
+MOUNT OPTIONS
+       uid=<uid>
+              set the user owning the mount point, the default is 0 (root).
+
+       gid=<gid>
+              set the group owning the mount point, the default is 0 (root).
+
+
+FILES
+       The files in spufs mostly follow the standard behavior for regular sys-
+       tem  calls like read(2) or write(2), but often support only a subset of
+       the operations supported on regular file systems. This list details the
+       supported  operations  and  the  deviations  from  the behaviour in the
+       respective man pages.
+
+       All files that support the read(2) operation also support readv(2)  and
+       all  files  that support the write(2) operation also support writev(2).
+       All files support the access(2) and stat(2) family of  operations,  but
+       only  the  st_mode,  st_nlink,  st_uid and st_gid fields of struct stat
+       contain reliable information.
+
+       All files support the chmod(2)/fchmod(2) and chown(2)/fchown(2)  opera-
+       tions,  but  will  not be able to grant permissions that contradict the
+       possible operations, e.g. read access on the wbox file.
+
+       The current set of files is:
+
+
+   /mem
+       the contents of the local storage memory  of  the  SPU.   This  can  be
+       accessed  like  a regular shared memory file and contains both code and
+       data in the address space of the SPU.  The possible  operations  on  an
+       open mem file are:
+
+       read(2), pread(2), write(2), pwrite(2), lseek(2)
+              These  operate  as  documented, with the exception that seek(2),
+              write(2) and pwrite(2) are not supported beyond the end  of  the
+              file. The file size is the size of the local storage of the SPU,
+              which normally is 256 kilobytes.
+
+       mmap(2)
+              Mapping mem into the process address space gives access  to  the
+              SPU  local  storage  within  the  process  address  space.  Only
+              MAP_SHARED mappings are allowed.
+
+
+   /mbox
+       The first SPU to CPU communication mailbox. This file is read-only  and
+       can  be  read  in  units of 32 bits.  The file can only be used in non-
+       blocking mode and it even poll() will not block on  it.   The  possible
+       operations on an open mbox file are:
+
+       read(2)
+              If  a  count smaller than four is requested, read returns -1 and
+              sets errno to EINVAL.  If there is no data available in the mail
+              box,  the  return  value  is set to -1 and errno becomes EAGAIN.
+              When data has been read successfully, four bytes are  placed  in
+              the data buffer and the value four is returned.
+
+
+   /ibox
+       The  second  SPU  to CPU communication mailbox. This file is similar to
+       the first mailbox file, but can be read in blocking I/O mode,  and  the
+       poll  familiy of system calls can be used to wait for it.  The possible
+       operations on an open ibox file are:
+
+       read(2)
+              If a count smaller than four is requested, read returns  -1  and
+              sets errno to EINVAL.  If there is no data available in the mail
+              box and the file descriptor has been opened with O_NONBLOCK, the
+              return value is set to -1 and errno becomes EAGAIN.
+
+              If  there  is  no  data  available  in the mail box and the file
+              descriptor has been opened without  O_NONBLOCK,  the  call  will
+              block  until  the  SPU  writes to its interrupt mailbox channel.
+              When data has been read successfully, four bytes are  placed  in
+              the data buffer and the value four is returned.
+
+       poll(2)
+              Poll  on  the  ibox  file returns (POLLIN | POLLRDNORM) whenever
+              data is available for reading.
+
+
+   /wbox
+       The CPU to SPU communation mailbox. It is write-only can can be written
+       in  units  of  32  bits. If the mailbox is full, write() will block and
+       poll can be used to wait for it becoming  empty  again.   The  possible
+       operations  on  an open wbox file are: write(2) If a count smaller than
+       four is requested, write returns -1 and sets errno to EINVAL.  If there
+       is  no space available in the mail box and the file descriptor has been
+       opened with O_NONBLOCK, the return value is set to -1 and errno becomes
+       EAGAIN.
+
+       If  there is no space available in the mail box and the file descriptor
+       has been opened without O_NONBLOCK, the call will block until  the  SPU
+       reads  from  its PPE mailbox channel.  When data has been read success-
+       fully, four bytes are placed in the data buffer and the value  four  is
+       returned.
+
+       poll(2)
+              Poll  on  the  ibox file returns (POLLOUT | POLLWRNORM) whenever
+              space is available for writing.
+
+
+   /mbox_stat
+   /ibox_stat
+   /wbox_stat
+       Read-only files that contain the length of the current queue, i.e.  how
+       many  words  can  be  read  from  mbox or ibox or how many words can be
+       written to wbox without blocking.  The files can be read only in 4-byte
+       units  and  return  a  big-endian  binary integer number.  The possible
+       operations on an open *box_stat file are:
+
+       read(2)
+              If a count smaller than four is requested, read returns  -1  and
+              sets errno to EINVAL.  Otherwise, a four byte value is placed in
+              the data buffer, containing the number of elements that  can  be
+              read  from  (for  mbox_stat  and  ibox_stat)  or written to (for
+              wbox_stat) the respective mail box without blocking or resulting
+              in EAGAIN.
+
+
+   /npc
+   /decr
+   /decr_status
+   /spu_tag_mask
+   /event_mask
+   /srr0
+       Internal  registers  of  the SPU. The representation is an ASCII string
+       with the numeric value of the next instruction to  be  executed.  These
+       can  be  used in read/write mode for debugging, but normal operation of
+       programs should not rely on them because access to any of  them  except
+       npc requires an SPU context save and is therefore very inefficient.
+
+       The contents of these files are:
+
+       npc                 Next Program Counter
+
+       decr                SPU Decrementer
+
+       decr_status         Decrementer Status
+
+       spu_tag_mask        MFC tag mask for SPU DMA
+
+       event_mask          Event mask for SPU interrupts
+
+       srr0                Interrupt Return address register
+
+
+       The   possible   operations   on   an   open  npc,  decr,  decr_status,
+       spu_tag_mask, event_mask or srr0 file are:
+
+       read(2)
+              When the count supplied to the read call  is  shorter  than  the
+              required  length for the pointer value plus a newline character,
+              subsequent reads from the same file descriptor  will  result  in
+              completing  the string, regardless of changes to the register by
+              a running SPU task.  When a complete string has been  read,  all
+              subsequent read operations will return zero bytes and a new file
+              descriptor needs to be opened to read the value again.
+
+       write(2)
+              A write operation on the file results in setting the register to
+              the  value  given  in  the string. The string is parsed from the
+              beginning to the first non-numeric character or the end  of  the
+              buffer.  Subsequent writes to the same file descriptor overwrite
+              the previous setting.
+
+
+   /fpcr
+       This file gives access to the Floating Point Status and Control  Regis-
+       ter as a four byte long file. The operations on the fpcr file are:
+
+       read(2)
+              If  a  count smaller than four is requested, read returns -1 and
+              sets errno to EINVAL.  Otherwise, a four byte value is placed in
+              the data buffer, containing the current value of the fpcr regis-
+              ter.
+
+       write(2)
+              If a count smaller than four is requested, write returns -1  and
+              sets  errno  to  EINVAL.  Otherwise, a four byte value is copied
+              from the data buffer, updating the value of the fpcr register.
+
+
+   /signal1
+   /signal2
+       The two signal notification channels of an SPU.  These  are  read-write
+       files  that  operate  on  a 32 bit word.  Writing to one of these files
+       triggers an interrupt on the SPU. The  value  writting  to  the  signal
+       files can be read from the SPU through a channel read or from host user
+       space through the file.  After the value has been read by the  SPU,  it
+       is  reset  to zero.  The possible operations on an open signal1 or sig-
+       nal2 file are:
+
+       read(2)
+              If a count smaller than four is requested, read returns  -1  and
+              sets errno to EINVAL.  Otherwise, a four byte value is placed in
+              the data buffer, containing the current value of  the  specified
+              signal notification register.
+
+       write(2)
+              If  a count smaller than four is requested, write returns -1 and
+              sets errno to EINVAL.  Otherwise, a four byte  value  is  copied
+              from the data buffer, updating the value of the specified signal
+              notification register.  The signal  notification  register  will
+              either be replaced with the input data or will be updated to the
+              bitwise OR or the old value and the input data, depending on the
+              contents  of  the  signal1_type,  or  signal2_type respectively,
+              file.
+
+
+   /signal1_type
+   /signal2_type
+       These two files change the behavior of the signal1 and signal2  notifi-
+       cation  files.  The  contain  a numerical ASCII string which is read as
+       either "1" or "0".  In mode 0 (overwrite), the  hardware  replaces  the
+       contents of the signal channel with the data that is written to it.  in
+       mode 1 (logical OR), the hardware accumulates the bits that are  subse-
+       quently written to it.  The possible operations on an open signal1_type
+       or signal2_type file are:
+
+       read(2)
+              When the count supplied to the read call  is  shorter  than  the
+              required  length  for the digit plus a newline character, subse-
+              quent reads from the same file descriptor will  result  in  com-
+              pleting  the  string.  When a complete string has been read, all
+              subsequent read operations will return zero bytes and a new file
+              descriptor needs to be opened to read the value again.
+
+       write(2)
+              A write operation on the file results in setting the register to
+              the value given in the string. The string  is  parsed  from  the
+              beginning  to  the first non-numeric character or the end of the
+              buffer.  Subsequent writes to the same file descriptor overwrite
+              the previous setting.
+
+
+EXAMPLES
+       /etc/fstab entry
+              none      /spu      spufs     gid=spu   0    0
+
+
+AUTHORS
+       Arnd  Bergmann  <arndb@de.ibm.com>,  Mark  Nutter <mnutter@us.ibm.com>,
+       Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
+
+SEE ALSO
+       capabilities(7), close(2), spu_create(2), spu_run(2), spufs(7)
+
+
+
+Linux                             2005-09-28                          SPUFS(2)
+
+------------------------------------------------------------------------------
+
+SPU_RUN(2)                 Linux Programmer's Manual                SPU_RUN(2)
+
+
+
+NAME
+       spu_run - execute an spu context
+
+
+SYNOPSIS
+       #include <sys/spu.h>
+
+       int spu_run(int fd, unsigned int *npc, unsigned int *event);
+
+DESCRIPTION
+       The  spu_run system call is used on PowerPC machines that implement the
+       Cell Broadband Engine Architecture in order to access Synergistic  Pro-
+       cessor  Units  (SPUs).  It  uses the fd that was returned from spu_cre-
+       ate(2) to address a specific SPU context. When the context gets  sched-
+       uled  to a physical SPU, it starts execution at the instruction pointer
+       passed in npc.
+
+       Execution of SPU code happens synchronously, meaning that spu_run  does
+       not  return  while the SPU is still running. If there is a need to exe-
+       cute SPU code in parallel with other code on either  the  main  CPU  or
+       other  SPUs,  you  need to create a new thread of execution first, e.g.
+       using the pthread_create(3) call.
+
+       When spu_run returns, the current value of the SPU instruction  pointer
+       is  written back to npc, so you can call spu_run again without updating
+       the pointers.
+
+       event can be a NULL pointer or point to an extended  status  code  that
+       gets  filled  when spu_run returns. It can be one of the following con-
+       stants:
+
+       SPE_EVENT_DMA_ALIGNMENT
+              A DMA alignment error
+
+       SPE_EVENT_SPE_DATA_SEGMENT
+              A DMA segmentation error
+
+       SPE_EVENT_SPE_DATA_STORAGE
+              A DMA storage error
+
+       If NULL is passed as the event argument, these errors will result in  a
+       signal delivered to the calling process.
+
+RETURN VALUE
+       spu_run  returns the value of the spu_status register or -1 to indicate
+       an error and set errno to one of the error  codes  listed  below.   The
+       spu_status  register  value  contains  a  bit  mask of status codes and
+       optionally a 14 bit code returned from the stop-and-signal  instruction
+       on the SPU. The bit masks for the status codes are:
+
+       0x02   SPU was stopped by stop-and-signal.
+
+       0x04   SPU was stopped by halt.
+
+       0x08   SPU is waiting for a channel.
+
+       0x10   SPU is in single-step mode.
+
+       0x20   SPU has tried to execute an invalid instruction.
+
+       0x40   SPU has tried to access an invalid channel.
+
+       0x3fff0000
+              The  bits  masked with this value contain the code returned from
+              stop-and-signal.
+
+       There are always one or more of the lower eight bits set  or  an  error
+       code is returned from spu_run.
+
+ERRORS
+       EAGAIN or EWOULDBLOCK
+              fd is in non-blocking mode and spu_run would block.
+
+       EBADF  fd is not a valid file descriptor.
+
+       EFAULT npc is not a valid pointer or status is neither NULL nor a valid
+              pointer.
+
+       EINTR  A signal occured while spu_run was in progress.  The  npc  value
+              has  been updated to the new program counter value if necessary.
+
+       EINVAL fd is not a file descriptor returned from spu_create(2).
+
+       ENOMEM Insufficient memory was available to handle a page fault result-
+              ing from an MFC direct memory access.
+
+       ENOSYS the functionality is not provided by the current system, because
+              either the hardware does not provide SPUs or the spufs module is
+              not loaded.
+
+
+NOTES
+       spu_run  is  meant  to  be  used  from  libraries that implement a more
+       abstract interface to SPUs, not to be used from  regular  applications.
+       See  http://www.bsc.es/projects/deepcomputing/linuxoncell/ for the rec-
+       ommended libraries.
+
+
+CONFORMING TO
+       This call is Linux specific and only implemented by the ppc64 architec-
+       ture. Programs using this system call are not portable.
+
+
+BUGS
+       The code does not yet fully implement all features lined out here.
+
+
+AUTHOR
+       Arnd Bergmann <arndb@de.ibm.com>
+
+SEE ALSO
+       capabilities(7), close(2), spu_create(2), spufs(7)
+
+
+
+Linux                             2005-09-28                        SPU_RUN(2)
+
+------------------------------------------------------------------------------
+
+SPU_CREATE(2)              Linux Programmer's Manual             SPU_CREATE(2)
+
+
+
+NAME
+       spu_create - create a new spu context
+
+
+SYNOPSIS
+       #include <sys/types.h>
+       #include <sys/spu.h>
+
+       int spu_create(const char *pathname, int flags, mode_t mode);
+
+DESCRIPTION
+       The  spu_create  system call is used on PowerPC machines that implement
+       the Cell Broadband Engine Architecture in order to  access  Synergistic
+       Processor  Units (SPUs). It creates a new logical context for an SPU in
+       pathname and returns a handle to associated  with  it.   pathname  must
+       point  to  a  non-existing directory in the mount point of the SPU file
+       system (spufs).  When spu_create is successful, a directory  gets  cre-
+       ated on pathname and it is populated with files.
+
+       The  returned  file  handle can only be passed to spu_run(2) or closed,
+       other operations are not defined on it. When it is closed, all  associ-
+       ated  directory entries in spufs are removed. When the last file handle
+       pointing either inside  of  the  context  directory  or  to  this  file
+       descriptor is closed, the logical SPU context is destroyed.
+
+       The  parameter flags can be zero or any bitwise or'd combination of the
+       following constants:
+
+       SPU_RAWIO
+              Allow mapping of some of the hardware registers of the SPU  into
+              user space. This flag requires the CAP_SYS_RAWIO capability, see
+              capabilities(7).
+
+       The mode parameter specifies the permissions used for creating the  new
+       directory  in  spufs.   mode is modified with the user's umask(2) value
+       and then used for both the directory and the files contained in it. The
+       file permissions mask out some more bits of mode because they typically
+       support only read or write access. See stat(2) for a full list  of  the
+       possible mode values.
+
+
+RETURN VALUE
+       spu_create  returns a new file descriptor. It may return -1 to indicate
+       an error condition and set errno to  one  of  the  error  codes  listed
+       below.
+
+
+ERRORS
+       EACCESS
+              The  current  user does not have write access on the spufs mount
+              point.
+
+       EEXIST An SPU context already exists at the given path name.
+
+       EFAULT pathname is not a valid string pointer in  the  current  address
+              space.
+
+       EINVAL pathname is not a directory in the spufs mount point.
+
+       ELOOP  Too many symlinks were found while resolving pathname.
+
+       EMFILE The process has reached its maximum open file limit.
+
+       ENAMETOOLONG
+              pathname was too long.
+
+       ENFILE The system has reached the global open file limit.
+
+       ENOENT Part of pathname could not be resolved.
+
+       ENOMEM The kernel could not allocate all resources required.
+
+       ENOSPC There  are  not  enough  SPU resources available to create a new
+              context or the user specific limit for the number  of  SPU  con-
+              texts has been reached.
+
+       ENOSYS the functionality is not provided by the current system, because
+              either the hardware does not provide SPUs or the spufs module is
+              not loaded.
+
+       ENOTDIR
+              A part of pathname is not a directory.
+
+
+
+NOTES
+       spu_create  is  meant  to  be used from libraries that implement a more
+       abstract interface to SPUs, not to be used from  regular  applications.
+       See  http://www.bsc.es/projects/deepcomputing/linuxoncell/ for the rec-
+       ommended libraries.
+
+
+FILES
+       pathname must point to a location beneath the mount point of spufs.  By
+       convention, it gets mounted in /spu.
+
+
+CONFORMING TO
+       This call is Linux specific and only implemented by the ppc64 architec-
+       ture. Programs using this system call are not portable.
+
+
+BUGS
+       The code does not yet fully implement all features lined out here.
+
+
+AUTHOR
+       Arnd Bergmann <arndb@de.ibm.com>
+
+SEE ALSO
+       capabilities(7), close(2), spu_run(2), spufs(7)
+
+
+
+Linux                             2005-09-28                     SPU_CREATE(2)

Documentation/powerpc/00-INDEX

@@ -8,12 +8,18 @@ please mail me.
 cpu_features.txt
 	- info on how we support a variety of CPUs with minimal compile-time
 	options.
+eeh-pci-error-recovery.txt
+	- info on PCI Bus EEH Error Recovery
+hvcs.txt
+	- IBM "Hypervisor Virtual Console Server" Installation Guide
+mpc52xx.txt
+	- Linux 2.6.x on MPC52xx family
 ppc_htab.txt
 	- info about the Linux/PPC /proc/ppc_htab entry
-smp.txt
-	- use and state info about Linux/PPC on MP machines
 SBC8260_memory_mapping.txt
 	- EST SBC8260 board info
+smp.txt
+	- use and state info about Linux/PPC on MP machines
 sound.txt
 	- info on sound support under Linux/PPC
 zImage_layout.txt

arch/powerpc/Kconfig

@@ -47,7 +47,7 @@ config PPC
 
 config EARLY_PRINTK
 	bool
-	default y if PPC64
+	default y
 
 config COMPAT
 	bool
@@ -297,6 +297,7 @@ config PPC_PMAC64
 	bool
 	depends on PPC_PMAC && POWER4
 	select U3_DART
+	select MPIC_BROKEN_U3
 	select GENERIC_TBSYNC
 	default y
 
@@ -325,9 +326,7 @@ config PPC_CELL
 	select MMIO_NVRAM
 
 config PPC_OF
-	bool
-	depends on PPC_MULTIPLATFORM	# for now
-	default y
+	def_bool y
 
 config XICS
 	depends on PPC_PSERIES
@@ -376,11 +375,28 @@ config CELL_IIC
 	bool
 	default y
 
+config CRASH_DUMP
+	bool "kernel crash dumps (EXPERIMENTAL)"
+	depends on PPC_MULTIPLATFORM
+	depends on EXPERIMENTAL
+	help
+	  Build a kernel suitable for use as a kdump capture kernel.
+	  The kernel will be linked at a different address than normal, and
+	  so can only be used for Kdump.
+
+	  Don't change this unless you know what you are doing.
+
 config IBMVIO
 	depends on PPC_PSERIES || PPC_ISERIES
 	bool
 	default y
 
+config IBMEBUS
+	depends on PPC_PSERIES
+	bool "Support for GX bus based adapters"
+	help
+	  Bus device driver for GX bus based adapters.
+
 config PPC_MPC106
 	bool
 	default n
@@ -472,6 +488,7 @@ source arch/powerpc/platforms/embedded6xx/Kconfig
 source arch/powerpc/platforms/4xx/Kconfig
 source arch/powerpc/platforms/85xx/Kconfig
 source arch/powerpc/platforms/8xx/Kconfig
+source arch/powerpc/platforms/cell/Kconfig
 
 menu "Kernel options"
 
@@ -575,11 +592,12 @@ config ARCH_SELECT_MEMORY_MODEL
 	depends on PPC64
 
 config ARCH_FLATMEM_ENABLE
-       def_bool y
-       depends on PPC64 && !NUMA
+	def_bool y
+	depends on (PPC64 && !NUMA) || PPC32
 
 config ARCH_SPARSEMEM_ENABLE
 	def_bool y
+	depends on PPC64
 
 config ARCH_SPARSEMEM_DEFAULT
 	def_bool y

arch/powerpc/Makefile

@@ -151,7 +151,7 @@ CPPFLAGS_vmlinux.lds	:= -Upowerpc
 # All the instructions talk about "make bzImage".
 bzImage: zImage
 
-BOOT_TARGETS = zImage zImage.initrd znetboot znetboot.initrd vmlinux.sm
+BOOT_TARGETS = zImage zImage.initrd znetboot znetboot.initrd vmlinux.sm uImage
 
 .PHONY: $(BOOT_TARGETS)
 

arch/powerpc/boot/Makefile

@@ -143,6 +143,36 @@ $(obj)/zImage.initrd: $(obj)/zImage.initrd.vmode $(obj)/addnote
 	@cp -f $< $@
 	$(call if_changed,addnote)
 
+#-----------------------------------------------------------
+# build u-boot images
+#-----------------------------------------------------------
+quiet_cmd_mygzip = GZIP $@
+cmd_mygzip = gzip -f -9 < $< > $@.$$$$ && mv $@.$$$$ $@
+
+quiet_cmd_objbin = OBJCOPY $@
+      cmd_objbin = $(OBJCOPY) -O binary $< $@
+
+quiet_cmd_uimage = UIMAGE $@
+      cmd_uimage = $(CONFIG_SHELL) $(MKIMAGE) -A ppc -O linux -T kernel \
+               -C gzip -a 00000000 -e 00000000 -n 'Linux-$(KERNELRELEASE)' \
+               -d $< $@
+
+MKIMAGE		:= $(srctree)/scripts/mkuboot.sh
+targets		+= uImage
+extra-y		+= vmlinux.bin vmlinux.gz
+
+$(obj)/vmlinux.bin: vmlinux FORCE
+	$(call if_changed,objbin)
+
+$(obj)/vmlinux.gz: $(obj)/vmlinux.bin FORCE
+	$(call if_changed,mygzip)
+
+$(obj)/uImage: $(obj)/vmlinux.gz
+	$(Q)rm -f $@
+	$(call cmd,uimage)
+	@echo -n '  Image: $@ '
+	@if [ -f $@ ]; then echo 'is ready' ; else echo 'not made'; fi
+
 install: $(CONFIGURE) $(BOOTIMAGE)
 	sh -x $(srctree)/$(src)/install.sh "$(KERNELRELEASE)" vmlinux System.map "$(INSTALL_PATH)" "$(BOOTIMAGE)"
 

arch/powerpc/kernel/Makefile

@@ -17,11 +17,11 @@ obj-y				+= vdso32/
 obj-$(CONFIG_PPC64)		+= setup_64.o binfmt_elf32.o sys_ppc32.o \
 				   signal_64.o ptrace32.o systbl.o \
 				   paca.o ioctl32.o cpu_setup_power4.o \
-				   firmware.o sysfs.o udbg.o idle_64.o
+				   firmware.o sysfs.o idle_64.o
 obj-$(CONFIG_PPC64)		+= vdso64/
 obj-$(CONFIG_ALTIVEC)		+= vecemu.o vector.o
 obj-$(CONFIG_POWER4)		+= idle_power4.o
-obj-$(CONFIG_PPC_OF)		+= of_device.o
+obj-$(CONFIG_PPC_OF)		+= of_device.o prom_parse.o
 procfs-$(CONFIG_PPC64)		:= proc_ppc64.o
 obj-$(CONFIG_PROC_FS)		+= $(procfs-y)
 rtaspci-$(CONFIG_PPC64)		:= rtas_pci.o
@@ -30,12 +30,10 @@ obj-$(CONFIG_RTAS_FLASH)	+= rtas_flash.o
 obj-$(CONFIG_RTAS_PROC)		+= rtas-proc.o
 obj-$(CONFIG_LPARCFG)		+= lparcfg.o
 obj-$(CONFIG_IBMVIO)		+= vio.o
+obj-$(CONFIG_IBMEBUS)           += ibmebus.o
 obj-$(CONFIG_GENERIC_TBSYNC)	+= smp-tbsync.o
-obj-$(CONFIG_PPC_PSERIES)	+= udbg_16550.o
-obj-$(CONFIG_PPC_MAPLE)		+= udbg_16550.o
-udbgscc-$(CONFIG_PPC64)		:= udbg_scc.o
-obj-$(CONFIG_PPC_PMAC)		+= $(udbgscc-y)
 obj64-$(CONFIG_PPC_MULTIPLATFORM) += nvram_64.o
+obj-$(CONFIG_CRASH_DUMP)	+= crash_dump.o
 
 ifeq ($(CONFIG_PPC_MERGE),y)
 
@@ -48,25 +46,25 @@ extra-$(CONFIG_8xx)		:= head_8xx.o
 extra-y				+= vmlinux.lds
 
 obj-y				+= process.o init_task.o time.o \
-				   prom.o traps.o setup-common.o
+				   prom.o traps.o setup-common.o udbg.o
 obj-$(CONFIG_PPC32)		+= entry_32.o setup_32.o misc_32.o systbl.o
 obj-$(CONFIG_PPC64)		+= misc_64.o dma_64.o iommu.o
-obj-$(CONFIG_PPC_OF)		+= prom_init.o
+obj-$(CONFIG_PPC_MULTIPLATFORM)	+= prom_init.o
 obj-$(CONFIG_MODULES)		+= ppc_ksyms.o
 obj-$(CONFIG_BOOTX_TEXT)	+= btext.o
 obj-$(CONFIG_6xx)		+= idle_6xx.o
 obj-$(CONFIG_SMP)		+= smp.o
 obj-$(CONFIG_KPROBES)		+= kprobes.o
-
+obj-$(CONFIG_SERIAL_8250)	+= legacy_serial.o udbg_16550.o
 module-$(CONFIG_PPC64)		+= module_64.o
 obj-$(CONFIG_MODULES)		+= $(module-y)
 
 pci64-$(CONFIG_PPC64)		+= pci_64.o pci_dn.o pci_iommu.o \
 				   pci_direct_iommu.o iomap.o
 obj-$(CONFIG_PCI)		+= $(pci64-y)
-
-kexec64-$(CONFIG_PPC64)		+= machine_kexec_64.o
-obj-$(CONFIG_KEXEC)		+= $(kexec64-y)
+kexec-$(CONFIG_PPC64)		:= machine_kexec_64.o
+kexec-$(CONFIG_PPC32)		:= machine_kexec_32.o
+obj-$(CONFIG_KEXEC)		+= machine_kexec.o crash.o $(kexec-y)
 
 ifeq ($(CONFIG_PPC_ISERIES),y)
 $(obj)/head_64.o: $(obj)/lparmap.s

arch/powerpc/kernel/asm-offsets.c

@@ -92,9 +92,9 @@ int main(void)
 
 	DEFINE(TI_FLAGS, offsetof(struct thread_info, flags));
 	DEFINE(TI_PREEMPT, offsetof(struct thread_info, preempt_count));
-	DEFINE(TI_SC_NOERR, offsetof(struct thread_info, syscall_noerror));
-#ifdef CONFIG_PPC32
+	DEFINE(TI_SIGFRAME, offsetof(struct thread_info, nvgprs_frame));
 	DEFINE(TI_TASK, offsetof(struct thread_info, task));
+#ifdef CONFIG_PPC32
 	DEFINE(TI_EXECDOMAIN, offsetof(struct thread_info, exec_domain));
 	DEFINE(TI_CPU, offsetof(struct thread_info, cpu));
 #endif /* CONFIG_PPC32 */
@@ -131,11 +131,9 @@ int main(void)
 	DEFINE(PACALOWHTLBAREAS, offsetof(struct paca_struct, context.low_htlb_areas));
 	DEFINE(PACAHIGHHTLBAREAS, offsetof(struct paca_struct, context.high_htlb_areas));
 #endif /* CONFIG_HUGETLB_PAGE */
-	DEFINE(PACADEFAULTDECR, offsetof(struct paca_struct, default_decr));
 	DEFINE(PACA_EXGEN, offsetof(struct paca_struct, exgen));
 	DEFINE(PACA_EXMC, offsetof(struct paca_struct, exmc));
 	DEFINE(PACA_EXSLB, offsetof(struct paca_struct, exslb));
-	DEFINE(PACA_EXDSI, offsetof(struct paca_struct, exdsi));
 	DEFINE(PACAEMERGSP, offsetof(struct paca_struct, emergency_sp));
 	DEFINE(PACALPPACA, offsetof(struct paca_struct, lppaca));
 	DEFINE(PACAHWCPUID, offsetof(struct paca_struct, hw_cpu_id));

arch/powerpc/kernel/btext.c

@@ -31,15 +31,18 @@ static void draw_byte_32(unsigned char *bits, unsigned int *base, int rb);
 static void draw_byte_16(unsigned char *bits, unsigned int *base, int rb);
 static void draw_byte_8(unsigned char *bits, unsigned int *base, int rb);
 
-static int g_loc_X;
-static int g_loc_Y;
-static int g_max_loc_X;
-static int g_max_loc_Y;
+#define __force_data __attribute__((__section__(".data")))
 
-static int dispDeviceRowBytes;
-static int dispDeviceDepth;
-static int dispDeviceRect[4];
-static unsigned char *dispDeviceBase, *logicalDisplayBase;
+static int g_loc_X __force_data;
+static int g_loc_Y __force_data;
+static int g_max_loc_X __force_data;
+static int g_max_loc_Y __force_data;
+
+static int dispDeviceRowBytes __force_data;
+static int dispDeviceDepth  __force_data;
+static int dispDeviceRect[4] __force_data;
+static unsigned char *dispDeviceBase __force_data;
+static unsigned char *logicalDisplayBase __force_data;
 
 unsigned long disp_BAT[2] __initdata = {0, 0};
 
@@ -47,7 +50,7 @@ unsigned long disp_BAT[2] __initdata = {0, 0};
 
 static unsigned char vga_font[cmapsz];
 
-int boot_text_mapped;
+int boot_text_mapped __force_data = 0;
 int force_printk_to_btext = 0;
 
 #ifdef CONFIG_PPC32
@@ -57,7 +60,7 @@ int force_printk_to_btext = 0;
  *
  * The display is mapped to virtual address 0xD0000000, rather
  * than 1:1, because some some CHRP machines put the frame buffer
- * in the region starting at 0xC0000000 (KERNELBASE).
+ * in the region starting at 0xC0000000 (PAGE_OFFSET).
  * This mapping is temporary and will disappear as soon as the
  * setup done by MMU_Init() is applied.
  *
@@ -66,10 +69,9 @@ int force_printk_to_btext = 0;
  * is really badly aligned, but I didn't encounter this case
  * yet.
  */
-void __init
-btext_prepare_BAT(void)
+void __init btext_prepare_BAT(void)
 {
-	unsigned long vaddr = KERNELBASE + 0x10000000;
+	unsigned long vaddr = PAGE_OFFSET + 0x10000000;
 	unsigned long addr;
 	unsigned long lowbits;
 
@@ -95,12 +97,13 @@ btext_prepare_BAT(void)
 }
 #endif
 
-/* This function will enable the early boot text when doing OF booting. This
- * way, xmon output should work too
+
+/* This function can be used to enable the early boot text when doing
+ * OF booting or within bootx init. It must be followed by a btext_unmap()
+ * call before the logical address becomes unuseable
  */
-void __init
-btext_setup_display(int width, int height, int depth, int pitch,
-		    unsigned long address)
+void __init btext_setup_display(int width, int height, int depth, int pitch,
+				unsigned long address)
 {
 	g_loc_X = 0;
 	g_loc_Y = 0;
@@ -116,6 +119,11 @@ btext_setup_display(int width, int height, int depth, int pitch,
 	boot_text_mapped = 1;
 }
 
+void __init btext_unmap(void)
+{
+	boot_text_mapped = 0;
+}
+
 /* Here's a small text engine to use during early boot
  * or for debugging purposes
  *
@@ -127,7 +135,7 @@ btext_setup_display(int width, int height, int depth, int pitch,
  *    changes.
  */
 
-void map_boot_text(void)
+static void map_boot_text(void)
 {
 	unsigned long base, offset, size;
 	unsigned char *vbase;
@@ -175,8 +183,9 @@ int btext_initialize(struct device_node *np)
 	if (prop)
 		address = *prop;
 
-	/* FIXME: Add support for PCI reg properties */
-
+	/* FIXME: Add support for PCI reg properties. Right now, only
+	 * reliable on macs
+	 */
 	if (address == 0)
 		return -EINVAL;
 
@@ -184,7 +193,6 @@ int btext_initialize(struct device_node *np)
 	g_loc_Y = 0;
 	g_max_loc_X = width / 8;
 	g_max_loc_Y = height / 16;
-	logicalDisplayBase = (unsigned char *)address;
 	dispDeviceBase = (unsigned char *)address;
 	dispDeviceRowBytes = pitch;
 	dispDeviceDepth = depth;
@@ -197,14 +205,12 @@ int btext_initialize(struct device_node *np)
 	return 0;
 }
 
-void __init init_boot_display(void)
+int __init btext_find_display(int allow_nonstdout)
 {
 	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);
@@ -218,8 +224,8 @@ void __init init_boot_display(void)
 	}
 	if (np)
 		rc = btext_initialize(np);
-	if (rc == 0)
-		return;
+	if (rc == 0 || !allow_nonstdout)
+		return rc;
 
 	for (np = NULL; (np = of_find_node_by_type(np, "display"));) {
 		if (get_property(np, "linux,opened", NULL)) {
@@ -228,8 +234,9 @@ void __init init_boot_display(void)
 			printk("result: %d\n", rc);
 		}
 		if (rc == 0)
-			return;
+			break;
 	}
+	return rc;
 }
 
 /* Calc the base address of a given point (x,y) */
@@ -277,44 +284,83 @@ EXPORT_SYMBOL(btext_update_display);
 
 void btext_clearscreen(void)
 {
-	unsigned long *base	= (unsigned long *)calc_base(0, 0);
+	unsigned int *base	= (unsigned int *)calc_base(0, 0);
 	unsigned long width 	= ((dispDeviceRect[2] - dispDeviceRect[0]) *
-					(dispDeviceDepth >> 3)) >> 3;
+					(dispDeviceDepth >> 3)) >> 2;
 	int i,j;
 
 	for (i=0; i<(dispDeviceRect[3] - dispDeviceRect[1]); i++)
 	{
-		unsigned long *ptr = base;
+		unsigned int *ptr = base;
 		for(j=width; j; --j)
 			*(ptr++) = 0;
-		base += (dispDeviceRowBytes >> 3);
+		base += (dispDeviceRowBytes >> 2);
 	}
 }
 
+void btext_flushscreen(void)
+{
+	unsigned int *base	= (unsigned int *)calc_base(0, 0);
+	unsigned long width 	= ((dispDeviceRect[2] - dispDeviceRect[0]) *
+					(dispDeviceDepth >> 3)) >> 2;
+	int i,j;
+
+	for (i=0; i < (dispDeviceRect[3] - dispDeviceRect[1]); i++)
+	{
+		unsigned int *ptr = base;
+		for(j = width; j > 0; j -= 8) {
+			__asm__ __volatile__ ("dcbst 0,%0" :: "r" (ptr));
+			ptr += 8;
+		}
+		base += (dispDeviceRowBytes >> 2);
+	}
+	__asm__ __volatile__ ("sync" ::: "memory");
+}
+
+void btext_flushline(void)
+{
+	unsigned int *base	= (unsigned int *)calc_base(0, g_loc_Y << 4);
+	unsigned long width 	= ((dispDeviceRect[2] - dispDeviceRect[0]) *
+					(dispDeviceDepth >> 3)) >> 2;
+	int i,j;
+
+	for (i=0; i < 16; i++)
+	{
+		unsigned int *ptr = base;
+		for(j = width; j > 0; j -= 8) {
+			__asm__ __volatile__ ("dcbst 0,%0" :: "r" (ptr));
+			ptr += 8;
+		}
+		base += (dispDeviceRowBytes >> 2);
+	}
+	__asm__ __volatile__ ("sync" ::: "memory");
+}
+
+
 #ifndef NO_SCROLL
 static void scrollscreen(void)
 {
-	unsigned long *src     	= (unsigned long *)calc_base(0,16);
-	unsigned long *dst     	= (unsigned long *)calc_base(0,0);
+	unsigned int *src     	= (unsigned int *)calc_base(0,16);
+	unsigned int *dst     	= (unsigned int *)calc_base(0,0);
 	unsigned long width    	= ((dispDeviceRect[2] - dispDeviceRect[0]) *
-				   (dispDeviceDepth >> 3)) >> 3;
+				   (dispDeviceDepth >> 3)) >> 2;
 	int i,j;
 
 	for (i=0; i<(dispDeviceRect[3] - dispDeviceRect[1] - 16); i++)
 	{
-		unsigned long *src_ptr = src;
-		unsigned long *dst_ptr = dst;
+		unsigned int *src_ptr = src;
+		unsigned int *dst_ptr = dst;
 		for(j=width; j; --j)
 			*(dst_ptr++) = *(src_ptr++);
-		src += (dispDeviceRowBytes >> 3);
-		dst += (dispDeviceRowBytes >> 3);
+		src += (dispDeviceRowBytes >> 2);
+		dst += (dispDeviceRowBytes >> 2);
 	}
 	for (i=0; i<16; i++)
 	{
-		unsigned long *dst_ptr = dst;
+		unsigned int *dst_ptr = dst;
 		for(j=width; j; --j)
 			*(dst_ptr++) = 0;
-		dst += (dispDeviceRowBytes >> 3);
+		dst += (dispDeviceRowBytes >> 2);
 	}
 }
 #endif /* ndef NO_SCROLL */
@@ -377,6 +423,14 @@ void btext_drawstring(const char *c)
 		btext_drawchar(*c++);
 }
 
+void btext_drawtext(const char *c, unsigned int len)
+{
+	if (!boot_text_mapped)
+		return;
+	while (len--)
+		btext_drawchar(*c++);
+}
+
 void btext_drawhex(unsigned long v)
 {
 	char *hex_table = "0123456789abcdef";

arch/powerpc/kernel/cputable.c

@@ -78,10 +78,8 @@ struct cpu_spec	cpu_specs[] = {
 		.dcache_bsize		= 128,
 		.num_pmcs		= 8,
 		.cpu_setup		= __setup_cpu_power3,
-#ifdef CONFIG_OPROFILE
 		.oprofile_cpu_type	= "ppc64/power3",
-		.oprofile_model		= &op_model_rs64,
-#endif
+		.oprofile_type		= RS64,
 	},
 	{	/* Power3+ */
 		.pvr_mask		= 0xffff0000,
@@ -93,10 +91,8 @@ struct cpu_spec	cpu_specs[] = {
 		.dcache_bsize		= 128,
 		.num_pmcs		= 8,
 		.cpu_setup		= __setup_cpu_power3,
-#ifdef CONFIG_OPROFILE
 		.oprofile_cpu_type	= "ppc64/power3",
-		.oprofile_model		= &op_model_rs64,
-#endif
+		.oprofile_type		= RS64,
 	},
 	{	/* Northstar */
 		.pvr_mask		= 0xffff0000,
@@ -108,10 +104,8 @@ struct cpu_spec	cpu_specs[] = {
 		.dcache_bsize		= 128,
 		.num_pmcs		= 8,
 		.cpu_setup		= __setup_cpu_power3,
-#ifdef CONFIG_OPROFILE
 		.oprofile_cpu_type	= "ppc64/rs64",
-		.oprofile_model		= &op_model_rs64,
-#endif
+		.oprofile_type		= RS64,
 	},
 	{	/* Pulsar */
 		.pvr_mask		= 0xffff0000,
@@ -123,10 +117,8 @@ struct cpu_spec	cpu_specs[] = {
 		.dcache_bsize		= 128,
 		.num_pmcs		= 8,
 		.cpu_setup		= __setup_cpu_power3,
-#ifdef CONFIG_OPROFILE
 		.oprofile_cpu_type	= "ppc64/rs64",
-		.oprofile_model		= &op_model_rs64,
-#endif
+		.oprofile_type		= RS64,
 	},
 	{	/* I-star */
 		.pvr_mask		= 0xffff0000,
@@ -138,10 +130,8 @@ struct cpu_spec	cpu_specs[] = {
 		.dcache_bsize		= 128,
 		.num_pmcs		= 8,
 		.cpu_setup		= __setup_cpu_power3,
-#ifdef CONFIG_OPROFILE
 		.oprofile_cpu_type	= "ppc64/rs64",
-		.oprofile_model		= &op_model_rs64,
-#endif
+		.oprofile_type		= RS64,
 	},
 	{	/* S-star */
 		.pvr_mask		= 0xffff0000,
@@ -153,10 +143,8 @@ struct cpu_spec	cpu_specs[] = {
 		.dcache_bsize		= 128,
 		.num_pmcs		= 8,
 		.cpu_setup		= __setup_cpu_power3,
-#ifdef CONFIG_OPROFILE
 		.oprofile_cpu_type	= "ppc64/rs64",
-		.oprofile_model		= &op_model_rs64,
-#endif
+		.oprofile_type		= RS64,
 	},
 	{	/* Power4 */
 		.pvr_mask		= 0xffff0000,
@@ -168,10 +156,8 @@ struct cpu_spec	cpu_specs[] = {
 		.dcache_bsize		= 128,
 		.num_pmcs		= 8,
 		.cpu_setup		= __setup_cpu_power4,
-#ifdef CONFIG_OPROFILE
 		.oprofile_cpu_type	= "ppc64/power4",
-		.oprofile_model		= &op_model_rs64,
-#endif
+		.oprofile_type		= POWER4,
 	},
 	{	/* Power4+ */
 		.pvr_mask		= 0xffff0000,
@@ -183,10 +169,8 @@ struct cpu_spec	cpu_specs[] = {
 		.dcache_bsize		= 128,
 		.num_pmcs		= 8,
 		.cpu_setup		= __setup_cpu_power4,
-#ifdef CONFIG_OPROFILE
 		.oprofile_cpu_type	= "ppc64/power4",
-		.oprofile_model		= &op_model_power4,
-#endif
+		.oprofile_type		= POWER4,
 	},
 	{	/* PPC970 */
 		.pvr_mask		= 0xffff0000,
@@ -199,10 +183,8 @@ struct cpu_spec	cpu_specs[] = {
 		.dcache_bsize		= 128,
 		.num_pmcs		= 8,
 		.cpu_setup		= __setup_cpu_ppc970,
-#ifdef CONFIG_OPROFILE
 		.oprofile_cpu_type	= "ppc64/970",
-		.oprofile_model		= &op_model_power4,
-#endif
+		.oprofile_type		= POWER4,
 	},
 #endif /* CONFIG_PPC64 */
 #if defined(CONFIG_PPC64) || defined(CONFIG_POWER4)
@@ -221,10 +203,8 @@ struct cpu_spec	cpu_specs[] = {
 		.dcache_bsize		= 128,
 		.num_pmcs		= 8,
 		.cpu_setup		= __setup_cpu_ppc970,
-#ifdef CONFIG_OPROFILE
 		.oprofile_cpu_type	= "ppc64/970",
-		.oprofile_model		= &op_model_power4,
-#endif
+		.oprofile_type		= POWER4,
 	},
 #endif /* defined(CONFIG_PPC64) || defined(CONFIG_POWER4) */
 #ifdef CONFIG_PPC64
@@ -238,10 +218,8 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 128,
 		.dcache_bsize		= 128,
 		.cpu_setup		= __setup_cpu_ppc970,
-#ifdef CONFIG_OPROFILE
 		.oprofile_cpu_type	= "ppc64/970",
-		.oprofile_model		= &op_model_power4,
-#endif
+		.oprofile_type		= POWER4,
 	},
 	{	/* Power5 GR */
 		.pvr_mask		= 0xffff0000,
@@ -253,27 +231,23 @@ struct cpu_spec	cpu_specs[] = {
 		.dcache_bsize		= 128,
 		.num_pmcs		= 6,
 		.cpu_setup		= __setup_cpu_power4,
-#ifdef CONFIG_OPROFILE
 		.oprofile_cpu_type	= "ppc64/power5",
-		.oprofile_model		= &op_model_power4,
-#endif
+		.oprofile_type		= POWER4,
 	},
 	{	/* Power5 GS */
 		.pvr_mask		= 0xffff0000,
 		.pvr_value		= 0x003b0000,
-		.cpu_name		= "POWER5 (gs)",
+		.cpu_name		= "POWER5+ (gs)",
 		.cpu_features		= CPU_FTRS_POWER5,
 		.cpu_user_features	= COMMON_USER_POWER5_PLUS,
 		.icache_bsize		= 128,
 		.dcache_bsize		= 128,
 		.num_pmcs		= 6,
 		.cpu_setup		= __setup_cpu_power4,
-#ifdef CONFIG_OPROFILE
-		.oprofile_cpu_type	= "ppc64/power5",
-		.oprofile_model		= &op_model_power4,
-#endif
+		.oprofile_cpu_type	= "ppc64/power5+",
+		.oprofile_type		= POWER4,
 	},
-	{	/* BE DD1.x */
+	{	/* Cell Broadband Engine */
 		.pvr_mask		= 0xffff0000,
 		.pvr_value		= 0x00700000,
 		.cpu_name		= "Cell Broadband Engine",
@@ -545,7 +519,9 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 32,
 		.dcache_bsize		= 32,
 		.num_pmcs		= 6,
-		.cpu_setup		= __setup_cpu_745x
+		.cpu_setup		= __setup_cpu_745x,
+		.oprofile_cpu_type      = "ppc/7450",
+		.oprofile_type		= G4,
 	},
 	{	/* 7450 2.1 */
 		.pvr_mask		= 0xffffffff,
@@ -556,7 +532,9 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 32,
 		.dcache_bsize		= 32,
 		.num_pmcs		= 6,
-		.cpu_setup		= __setup_cpu_745x
+		.cpu_setup		= __setup_cpu_745x,
+		.oprofile_cpu_type      = "ppc/7450",
+		.oprofile_type		= G4,
 	},
 	{	/* 7450 2.3 and newer */
 		.pvr_mask		= 0xffff0000,
@@ -567,7 +545,9 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 32,
 		.dcache_bsize		= 32,
 		.num_pmcs		= 6,
-		.cpu_setup		= __setup_cpu_745x
+		.cpu_setup		= __setup_cpu_745x,
+		.oprofile_cpu_type      = "ppc/7450",
+		.oprofile_type		= G4,
 	},
 	{	/* 7455 rev 1.x */
 		.pvr_mask		= 0xffffff00,
@@ -578,7 +558,9 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 32,
 		.dcache_bsize		= 32,
 		.num_pmcs		= 6,
-		.cpu_setup		= __setup_cpu_745x
+		.cpu_setup		= __setup_cpu_745x,
+		.oprofile_cpu_type      = "ppc/7450",
+		.oprofile_type		= G4,
 	},
 	{	/* 7455 rev 2.0 */
 		.pvr_mask		= 0xffffffff,
@@ -589,7 +571,9 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 32,
 		.dcache_bsize		= 32,
 		.num_pmcs		= 6,
-		.cpu_setup		= __setup_cpu_745x
+		.cpu_setup		= __setup_cpu_745x,
+		.oprofile_cpu_type      = "ppc/7450",
+		.oprofile_type		= G4,
 	},
 	{	/* 7455 others */
 		.pvr_mask		= 0xffff0000,
@@ -600,7 +584,9 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 32,
 		.dcache_bsize		= 32,
 		.num_pmcs		= 6,
-		.cpu_setup		= __setup_cpu_745x
+		.cpu_setup		= __setup_cpu_745x,
+		.oprofile_cpu_type      = "ppc/7450",
+		.oprofile_type		= G4,
 	},
 	{	/* 7447/7457 Rev 1.0 */
 		.pvr_mask		= 0xffffffff,
@@ -611,7 +597,9 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 32,
 		.dcache_bsize		= 32,
 		.num_pmcs		= 6,
-		.cpu_setup		= __setup_cpu_745x
+		.cpu_setup		= __setup_cpu_745x,
+		.oprofile_cpu_type      = "ppc/7450",
+		.oprofile_type		= G4,
 	},
 	{	/* 7447/7457 Rev 1.1 */
 		.pvr_mask		= 0xffffffff,
@@ -622,7 +610,9 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 32,
 		.dcache_bsize		= 32,
 		.num_pmcs		= 6,
-		.cpu_setup		= __setup_cpu_745x
+		.cpu_setup		= __setup_cpu_745x,
+		.oprofile_cpu_type      = "ppc/7450",
+		.oprofile_type		= G4,
 	},
 	{	/* 7447/7457 Rev 1.2 and later */
 		.pvr_mask		= 0xffff0000,
@@ -633,7 +623,9 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 32,
 		.dcache_bsize		= 32,
 		.num_pmcs		= 6,
-		.cpu_setup		= __setup_cpu_745x
+		.cpu_setup		= __setup_cpu_745x,
+		.oprofile_cpu_type      = "ppc/7450",
+		.oprofile_type		= G4,
 	},
 	{	/* 7447A */
 		.pvr_mask		= 0xffff0000,
@@ -644,7 +636,9 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 32,
 		.dcache_bsize		= 32,
 		.num_pmcs		= 6,
-		.cpu_setup		= __setup_cpu_745x
+		.cpu_setup		= __setup_cpu_745x,
+		.oprofile_cpu_type      = "ppc/7450",
+		.oprofile_type		= G4,
 	},
 	{	/* 7448 */
 		.pvr_mask		= 0xffff0000,
@@ -655,7 +649,9 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 32,
 		.dcache_bsize		= 32,
 		.num_pmcs		= 6,
-		.cpu_setup		= __setup_cpu_745x
+		.cpu_setup		= __setup_cpu_745x,
+		.oprofile_cpu_type      = "ppc/7450",
+		.oprofile_type		= G4,
 	},
 	{	/* 82xx (8240, 8245, 8260 are all 603e cores) */
 		.pvr_mask		= 0x7fff0000,
@@ -979,6 +975,8 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 32,
 		.dcache_bsize		= 32,
 		.num_pmcs		= 4,
+		.oprofile_cpu_type	= "ppc/e500",
+		.oprofile_type		= BOOKE,
 	},
 	{	/* e500v2 */
 		.pvr_mask		= 0xffff0000,
@@ -992,6 +990,8 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 32,
 		.dcache_bsize		= 32,
 		.num_pmcs		= 4,
+		.oprofile_cpu_type	= "ppc/e500",
+		.oprofile_type		= BOOKE,
 	},
 #endif
 #if !CLASSIC_PPC

arch/powerpc/kernel/crash.c

@@ -0,0 +1,264 @@
+/*
+ * Architecture specific (PPC64) functions for kexec based crash dumps.
+ *
+ * Copyright (C) 2005, IBM Corp.
+ *
+ * Created by: Haren Myneni
+ *
+ * This source code is licensed under the GNU General Public License,
+ * Version 2.  See the file COPYING for more details.
+ *
+ */
+
+#undef DEBUG
+
+#include <linux/kernel.h>
+#include <linux/smp.h>
+#include <linux/reboot.h>
+#include <linux/kexec.h>
+#include <linux/bootmem.h>
+#include <linux/crash_dump.h>
+#include <linux/irq.h>
+#include <linux/delay.h>
+#include <linux/elf.h>
+#include <linux/elfcore.h>
+#include <linux/init.h>
+#include <linux/types.h>
+
+#include <asm/processor.h>
+#include <asm/machdep.h>
+#include <asm/kdump.h>
+#include <asm/lmb.h>
+#include <asm/firmware.h>
+
+#ifdef DEBUG
+#include <asm/udbg.h>
+#define DBG(fmt...) udbg_printf(fmt)
+#else
+#define DBG(fmt...)
+#endif
+
+/* This keeps a track of which one is crashing cpu. */
+int crashing_cpu = -1;
+
+static u32 *append_elf_note(u32 *buf, char *name, unsigned type, void *data,
+							       size_t data_len)
+{
+	struct elf_note note;
+
+	note.n_namesz = strlen(name) + 1;
+	note.n_descsz = data_len;
+	note.n_type   = type;
+	memcpy(buf, &note, sizeof(note));
+	buf += (sizeof(note) +3)/4;
+	memcpy(buf, name, note.n_namesz);
+	buf += (note.n_namesz + 3)/4;
+	memcpy(buf, data, note.n_descsz);
+	buf += (note.n_descsz + 3)/4;
+
+	return buf;
+}
+
+static void final_note(u32 *buf)
+{
+	struct elf_note note;
+
+	note.n_namesz = 0;
+	note.n_descsz = 0;
+	note.n_type   = 0;
+	memcpy(buf, &note, sizeof(note));
+}
+
+static void crash_save_this_cpu(struct pt_regs *regs, int cpu)
+{
+	struct elf_prstatus prstatus;
+	u32 *buf;
+
+	if ((cpu < 0) || (cpu >= NR_CPUS))
+		return;
+
+	/* Using ELF notes here is opportunistic.
+	 * I need a well defined structure format
+	 * for the data I pass, and I need tags
+	 * on the data to indicate what information I have
+	 * squirrelled away.  ELF notes happen to provide
+	 * all of that that no need to invent something new.
+	 */
+	buf = &crash_notes[cpu][0];
+	memset(&prstatus, 0, sizeof(prstatus));
+	prstatus.pr_pid = current->pid;
+	elf_core_copy_regs(&prstatus.pr_reg, regs);
+	buf = append_elf_note(buf, "CORE", NT_PRSTATUS, &prstatus,
+			sizeof(prstatus));
+	final_note(buf);
+}
+
+/* FIXME Merge this with xmon_save_regs ?? */
+static inline void crash_get_current_regs(struct pt_regs *regs)
+{
+	unsigned long tmp1, tmp2;
+
+	__asm__ __volatile__ (
+		"std    0,0(%2)\n"
+		"std    1,8(%2)\n"
+		"std    2,16(%2)\n"
+		"std    3,24(%2)\n"
+		"std    4,32(%2)\n"
+		"std    5,40(%2)\n"
+		"std    6,48(%2)\n"
+		"std    7,56(%2)\n"
+		"std    8,64(%2)\n"
+		"std    9,72(%2)\n"
+		"std    10,80(%2)\n"
+		"std    11,88(%2)\n"
+		"std    12,96(%2)\n"
+		"std    13,104(%2)\n"
+		"std    14,112(%2)\n"
+		"std    15,120(%2)\n"
+		"std    16,128(%2)\n"
+		"std    17,136(%2)\n"
+		"std    18,144(%2)\n"
+		"std    19,152(%2)\n"
+		"std    20,160(%2)\n"
+		"std    21,168(%2)\n"
+		"std    22,176(%2)\n"
+		"std    23,184(%2)\n"
+		"std    24,192(%2)\n"
+		"std    25,200(%2)\n"
+		"std    26,208(%2)\n"
+		"std    27,216(%2)\n"
+		"std    28,224(%2)\n"
+		"std    29,232(%2)\n"
+		"std    30,240(%2)\n"
+		"std    31,248(%2)\n"
+		"mfmsr  %0\n"
+		"std    %0, 264(%2)\n"
+		"mfctr  %0\n"
+		"std    %0, 280(%2)\n"
+		"mflr   %0\n"
+		"std    %0, 288(%2)\n"
+		"bl     1f\n"
+	"1:      mflr   %1\n"
+		"std    %1, 256(%2)\n"
+		"mtlr   %0\n"
+		"mfxer  %0\n"
+		"std    %0, 296(%2)\n"
+		: "=&r" (tmp1), "=&r" (tmp2)
+		: "b" (regs));
+}
+
+/* We may have saved_regs from where the error came from
+ * or it is NULL if via a direct panic().
+ */
+static void crash_save_self(struct pt_regs *saved_regs)
+{
+	struct pt_regs regs;
+	int cpu;
+
+	cpu = smp_processor_id();
+	if (saved_regs)
+		memcpy(&regs, saved_regs, sizeof(regs));
+	else
+		crash_get_current_regs(&regs);
+	crash_save_this_cpu(&regs, cpu);
+}
+
+#ifdef CONFIG_SMP
+static atomic_t waiting_for_crash_ipi;
+
+void crash_ipi_callback(struct pt_regs *regs)
+{
+	int cpu = smp_processor_id();
+
+	if (cpu == crashing_cpu)
+		return;
+
+	if (!cpu_online(cpu))
+		return;
+
+	if (ppc_md.kexec_cpu_down)
+		ppc_md.kexec_cpu_down(1, 1);
+
+	local_irq_disable();
+
+	crash_save_this_cpu(regs, cpu);
+	atomic_dec(&waiting_for_crash_ipi);
+	kexec_smp_wait();
+	/* NOTREACHED */
+}
+
+static void crash_kexec_prepare_cpus(void)
+{
+	unsigned int msecs;
+
+	atomic_set(&waiting_for_crash_ipi, num_online_cpus() - 1);
+
+	crash_send_ipi(crash_ipi_callback);
+	smp_wmb();
+
+	/*
+	 * FIXME: Until we will have the way to stop other CPUSs reliabally,
+	 * the crash CPU will send an IPI and wait for other CPUs to
+	 * respond. If not, proceed the kexec boot even though we failed to
+	 * capture other CPU states.
+	 */
+	msecs = 1000000;
+	while ((atomic_read(&waiting_for_crash_ipi) > 0) && (--msecs > 0)) {
+		barrier();
+		mdelay(1);
+	}
+
+	/* Would it be better to replace the trap vector here? */
+
+	/*
+	 * FIXME: In case if we do not get all CPUs, one possibility: ask the
+	 * user to do soft reset such that we get all.
+	 * IPI handler is already set by the panic cpu initially. Therefore,
+	 * all cpus could invoke this handler from die() and the panic CPU
+	 * will call machine_kexec() directly from this handler to do
+	 * kexec boot.
+	 */
+	if (atomic_read(&waiting_for_crash_ipi))
+		printk(KERN_ALERT "done waiting: %d cpus not responding\n",
+			atomic_read(&waiting_for_crash_ipi));
+	/* Leave the IPI callback set */
+}
+#else
+static void crash_kexec_prepare_cpus(void)
+{
+	/*
+	 * move the secondarys to us so that we can copy
+	 * the new kernel 0-0x100 safely
+	 *
+	 * do this if kexec in setup.c ?
+	 */
+	smp_release_cpus();
+}
+
+#endif
+
+void default_machine_crash_shutdown(struct pt_regs *regs)
+{
+	/*
+	 * This function is only called after the system
+	 * has paniced or is otherwise in a critical state.
+	 * The minimum amount of code to allow a kexec'd kernel
+	 * to run successfully needs to happen here.
+	 *
+	 * In practice this means stopping other cpus in
+	 * an SMP system.
+	 * The kernel is broken so disable interrupts.
+	 */
+	local_irq_disable();
+
+	if (ppc_md.kexec_cpu_down)
+		ppc_md.kexec_cpu_down(1, 0);
+
+	/*
+	 * Make a note of crashing cpu. Will be used in machine_kexec
+	 * such that another IPI will not be sent.
+	 */
+	crashing_cpu = smp_processor_id();
+	crash_kexec_prepare_cpus();
+	crash_save_self(regs);
+}

arch/powerpc/kernel/crash_dump.c

@@ -0,0 +1,109 @@
+/*
+ * Routines for doing kexec-based kdump.
+ *
+ * Copyright (C) 2005, IBM Corp.
+ *
+ * Created by: Michael Ellerman
+ *
+ * This source code is licensed under the GNU General Public License,
+ * Version 2.  See the file COPYING for more details.
+ */
+
+#undef DEBUG
+
+#include <linux/crash_dump.h>
+#include <linux/bootmem.h>
+#include <asm/kdump.h>
+#include <asm/lmb.h>
+#include <asm/firmware.h>
+#include <asm/uaccess.h>
+
+#ifdef DEBUG
+#include <asm/udbg.h>
+#define DBG(fmt...) udbg_printf(fmt)
+#else
+#define DBG(fmt...)
+#endif
+
+static void __init create_trampoline(unsigned long addr)
+{
+	/* The maximum range of a single instruction branch, is the current
+	 * instruction's address + (32 MB - 4) bytes. For the trampoline we
+	 * need to branch to current address + 32 MB. So we insert a nop at
+	 * the trampoline address, then the next instruction (+ 4 bytes)
+	 * does a branch to (32 MB - 4). The net effect is that when we
+	 * branch to "addr" we jump to ("addr" + 32 MB). Although it requires
+	 * two instructions it doesn't require any registers.
+	 */
+	create_instruction(addr, 0x60000000); /* nop */
+	create_branch(addr + 4, addr + PHYSICAL_START, 0);
+}
+
+void __init kdump_setup(void)
+{
+	unsigned long i;
+
+	DBG(" -> kdump_setup()\n");
+
+	for (i = KDUMP_TRAMPOLINE_START; i < KDUMP_TRAMPOLINE_END; i += 8) {
+		create_trampoline(i);
+	}
+
+	create_trampoline(__pa(system_reset_fwnmi) - PHYSICAL_START);
+	create_trampoline(__pa(machine_check_fwnmi) - PHYSICAL_START);
+
+	DBG(" <- kdump_setup()\n");
+}
+
+static int __init parse_elfcorehdr(char *p)
+{
+	if (p)
+		elfcorehdr_addr = memparse(p, &p);
+
+	return 0;
+}
+__setup("elfcorehdr=", parse_elfcorehdr);
+
+static int __init parse_savemaxmem(char *p)
+{
+	if (p)
+		saved_max_pfn = (memparse(p, &p) >> PAGE_SHIFT) - 1;
+
+	return 0;
+}
+__setup("savemaxmem=", parse_savemaxmem);
+
+/*
+ * copy_oldmem_page - copy one page from "oldmem"
+ * @pfn: page frame number to be copied
+ * @buf: target memory address for the copy; this can be in kernel address
+ *      space or user address space (see @userbuf)
+ * @csize: number of bytes to copy
+ * @offset: offset in bytes into the page (based on pfn) to begin the copy
+ * @userbuf: if set, @buf is in user address space, use copy_to_user(),
+ *      otherwise @buf is in kernel address space, use memcpy().
+ *
+ * Copy a page from "oldmem". For this page, there is no pte mapped
+ * in the current kernel. We stitch up a pte, similar to kmap_atomic.
+ */
+ssize_t copy_oldmem_page(unsigned long pfn, char *buf,
+			size_t csize, unsigned long offset, int userbuf)
+{
+	void  *vaddr;
+
+	if (!csize)
+		return 0;
+
+	vaddr = __ioremap(pfn << PAGE_SHIFT, PAGE_SIZE, 0);
+
+	if (userbuf) {
+		if (copy_to_user((char __user *)buf, (vaddr + offset), csize)) {
+			iounmap(vaddr);
+			return -EFAULT;
+		}
+	} else
+		memcpy(buf, (vaddr + offset), csize);
+
+	iounmap(vaddr);
+	return csize;
+}

arch/powerpc/kernel/dma_64.c

@@ -10,6 +10,7 @@
 /* Include the busses we support */
 #include <linux/pci.h>
 #include <asm/vio.h>
+#include <asm/ibmebus.h>
 #include <asm/scatterlist.h>
 #include <asm/bug.h>
 
@@ -22,6 +23,10 @@ static struct dma_mapping_ops *get_dma_ops(struct device *dev)
 #ifdef CONFIG_IBMVIO
 	if (dev->bus == &vio_bus_type)
 		return &vio_dma_ops;
+#endif
+#ifdef CONFIG_IBMEBUS
+	if (dev->bus == &ibmebus_bus_type)
+		return &ibmebus_dma_ops;
 #endif
 	return NULL;
 }
@@ -47,6 +52,10 @@ int dma_set_mask(struct device *dev, u64 dma_mask)
 	if (dev->bus == &vio_bus_type)
 		return -EIO;
 #endif /* CONFIG_IBMVIO */
+#ifdef CONFIG_IBMEBUS
+	if (dev->bus == &ibmebus_bus_type)
+		return -EIO;
+#endif
 	BUG();
 	return 0;
 }

arch/powerpc/kernel/entry_32.S

@@ -200,8 +200,6 @@ _GLOBAL(DoSyscall)
 	bl	do_show_syscall
 #endif /* SHOW_SYSCALLS */
 	rlwinm	r10,r1,0,0,(31-THREAD_SHIFT)	/* current_thread_info() */
-	li	r11,0
-	stb	r11,TI_SC_NOERR(r10)
 	lwz	r11,TI_FLAGS(r10)
 	andi.	r11,r11,_TIF_SYSCALL_T_OR_A
 	bne-	syscall_dotrace
@@ -222,25 +220,21 @@ ret_from_syscall:
 	bl	do_show_syscall_exit
 #endif
 	mr	r6,r3
-	li	r11,-_LAST_ERRNO
-	cmplw	0,r3,r11
 	rlwinm	r12,r1,0,0,(31-THREAD_SHIFT)	/* current_thread_info() */
-	blt+	30f
-	lbz	r11,TI_SC_NOERR(r12)
-	cmpwi	r11,0
-	bne	30f
-	neg	r3,r3
-	lwz	r10,_CCR(r1)	/* Set SO bit in CR */
-	oris	r10,r10,0x1000
-	stw	r10,_CCR(r1)
-
 	/* disable interrupts so current_thread_info()->flags can't change */
-30:	LOAD_MSR_KERNEL(r10,MSR_KERNEL)	/* doesn't include MSR_EE */
+	LOAD_MSR_KERNEL(r10,MSR_KERNEL)	/* doesn't include MSR_EE */
 	SYNC
 	MTMSRD(r10)
 	lwz	r9,TI_FLAGS(r12)
-	andi.	r0,r9,(_TIF_SYSCALL_T_OR_A|_TIF_SIGPENDING|_TIF_NEED_RESCHED)
+	li	r8,-_LAST_ERRNO
+	andi.	r0,r9,(_TIF_SYSCALL_T_OR_A|_TIF_SIGPENDING|_TIF_NEED_RESCHED|_TIF_RESTOREALL)
 	bne-	syscall_exit_work
+	cmplw	0,r3,r8
+	blt+	syscall_exit_cont
+	lwz	r11,_CCR(r1)			/* Load CR */
+	neg	r3,r3
+	oris	r11,r11,0x1000	/* Set SO bit in CR */
+	stw	r11,_CCR(r1)
 syscall_exit_cont:
 #if defined(CONFIG_4xx) || defined(CONFIG_BOOKE)
 	/* If the process has its own DBCR0 value, load it up.  The single
@@ -292,46 +286,113 @@ syscall_dotrace:
 	b	syscall_dotrace_cont
 
 syscall_exit_work:
-	stw	r6,RESULT(r1)	/* Save result */
+	andi.	r0,r9,_TIF_RESTOREALL
+	bne-	2f
+	cmplw	0,r3,r8
+	blt+	1f
+	andi.	r0,r9,_TIF_NOERROR
+	bne-	1f
+	lwz	r11,_CCR(r1)			/* Load CR */
+	neg	r3,r3
+	oris	r11,r11,0x1000	/* Set SO bit in CR */
+	stw	r11,_CCR(r1)
+
+1:	stw	r6,RESULT(r1)	/* Save result */
 	stw	r3,GPR3(r1)	/* Update return value */
-	andi.	r0,r9,_TIF_SYSCALL_T_OR_A
-	beq	5f
-	ori	r10,r10,MSR_EE
-	SYNC
-	MTMSRD(r10)		/* re-enable interrupts */
+2:	andi.	r0,r9,(_TIF_PERSYSCALL_MASK)
+	beq	4f
+
+	/* Clear per-syscall TIF flags if any are set, but _leave_
+	_TIF_SAVE_NVGPRS set in r9 since we haven't dealt with that
+	yet.  */
+
+	li	r11,_TIF_PERSYSCALL_MASK
+	addi	r12,r12,TI_FLAGS
+3:	lwarx	r8,0,r12
+	andc	r8,r8,r11
+#ifdef CONFIG_IBM405_ERR77
+	dcbt	0,r12
+#endif
+	stwcx.	r8,0,r12
+	bne-	3b
+	subi	r12,r12,TI_FLAGS
+	
+4:	/* Anything which requires enabling interrupts? */
+	andi.	r0,r9,(_TIF_SYSCALL_T_OR_A|_TIF_SINGLESTEP|_TIF_SAVE_NVGPRS)
+	beq	7f
+
+	/* Save NVGPRS if they're not saved already */
 	lwz	r4,_TRAP(r1)
 	andi.	r4,r4,1
-	beq	4f
+	beq	5f
 	SAVE_NVGPRS(r1)
 	li	r4,0xc00
 	stw	r4,_TRAP(r1)
-4:
+
+	/* Re-enable interrupts */
+5:	ori	r10,r10,MSR_EE
+	SYNC
+	MTMSRD(r10)
+
+	andi.	r0,r9,_TIF_SAVE_NVGPRS
+	bne	save_user_nvgprs
+
+save_user_nvgprs_cont:
+	andi.	r0,r9,(_TIF_SYSCALL_T_OR_A|_TIF_SINGLESTEP)
+	beq	7f
+
 	addi	r3,r1,STACK_FRAME_OVERHEAD
 	bl	do_syscall_trace_leave
 	REST_NVGPRS(r1)
-2:
-	lwz	r3,GPR3(r1)
+
+6:	lwz	r3,GPR3(r1)
 	LOAD_MSR_KERNEL(r10,MSR_KERNEL)	/* doesn't include MSR_EE */
 	SYNC
 	MTMSRD(r10)		/* disable interrupts again */
 	rlwinm	r12,r1,0,0,(31-THREAD_SHIFT)	/* current_thread_info() */
 	lwz	r9,TI_FLAGS(r12)
-5:
+7:
 	andi.	r0,r9,_TIF_NEED_RESCHED
-	bne	1f
+	bne	8f
 	lwz	r5,_MSR(r1)
 	andi.	r5,r5,MSR_PR
-	beq	syscall_exit_cont
+	beq	ret_from_except
 	andi.	r0,r9,_TIF_SIGPENDING
-	beq	syscall_exit_cont
+	beq	ret_from_except
 	b	do_user_signal
-1:
+8:
 	ori	r10,r10,MSR_EE
 	SYNC
 	MTMSRD(r10)		/* re-enable interrupts */
 	bl	schedule
-	b	2b
+	b	6b
 
+save_user_nvgprs:
+	lwz	r8,TI_SIGFRAME(r12)
+
+.macro savewords start, end
+  1:	stw \start,4*(\start)(r8)
+	.section __ex_table,"a"
+	.align	2
+	.long	1b,save_user_nvgprs_fault
+	.previous
+	.if \end - \start
+	savewords "(\start+1)",\end
+	.endif
+.endm	
+	savewords 14,31
+	b	save_user_nvgprs_cont
+
+	
+save_user_nvgprs_fault:
+	li	r3,11		/* SIGSEGV */
+	lwz	r4,TI_TASK(r12)
+	bl	force_sigsegv
+
+	rlwinm	r12,r1,0,0,(31-THREAD_SHIFT)	/* current_thread_info() */
+	lwz	r9,TI_FLAGS(r12)
+	b	save_user_nvgprs_cont
+	
 #ifdef SHOW_SYSCALLS
 do_show_syscall:
 #ifdef SHOW_SYSCALLS_TASK
@@ -401,28 +462,10 @@ show_syscalls_task:
 #endif /* SHOW_SYSCALLS */
 
 /*
- * The sigsuspend and rt_sigsuspend system calls can call do_signal
- * and thus put the process into the stopped state where we might
- * want to examine its user state with ptrace.  Therefore we need
- * to save all the nonvolatile registers (r13 - r31) before calling
- * the C code.
+ * The fork/clone functions need to copy the full register set into
+ * the child process. Therefore we need to save all the nonvolatile
+ * registers (r13 - r31) before calling the C code.
  */
-	.globl	ppc_sigsuspend
-ppc_sigsuspend:
-	SAVE_NVGPRS(r1)
-	lwz	r0,_TRAP(r1)
-	rlwinm	r0,r0,0,0,30		/* clear LSB to indicate full */
-	stw	r0,_TRAP(r1)		/* register set saved */
-	b	sys_sigsuspend
-
-	.globl	ppc_rt_sigsuspend
-ppc_rt_sigsuspend:
-	SAVE_NVGPRS(r1)
-	lwz	r0,_TRAP(r1)
-	rlwinm	r0,r0,0,0,30
-	stw	r0,_TRAP(r1)
-	b	sys_rt_sigsuspend
-
 	.globl	ppc_fork
 ppc_fork:
 	SAVE_NVGPRS(r1)
@@ -447,14 +490,6 @@ ppc_clone:
 	stw	r0,_TRAP(r1)		/* register set saved */
 	b	sys_clone
 
-	.globl	ppc_swapcontext
-ppc_swapcontext:
-	SAVE_NVGPRS(r1)
-	lwz	r0,_TRAP(r1)
-	rlwinm	r0,r0,0,0,30		/* clear LSB to indicate full */
-	stw	r0,_TRAP(r1)		/* register set saved */
-	b	sys_swapcontext
-
 /*
  * Top-level page fault handling.
  * This is in assembler because if do_page_fault tells us that
@@ -626,16 +661,6 @@ END_FTR_SECTION_IFSET(CPU_FTR_601)
 	.long	ret_from_except
 #endif
 
-	.globl	sigreturn_exit
-sigreturn_exit:
-	subi	r1,r3,STACK_FRAME_OVERHEAD
-	rlwinm	r12,r1,0,0,(31-THREAD_SHIFT)	/* current_thread_info() */
-	lwz	r9,TI_FLAGS(r12)
-	andi.	r0,r9,_TIF_SYSCALL_T_OR_A
-	beq+	ret_from_except_full
-	bl	do_syscall_trace_leave
-	/* fall through */
-
 	.globl	ret_from_except_full
 ret_from_except_full:
 	REST_NVGPRS(r1)
@@ -658,7 +683,7 @@ user_exc_return:		/* r10 contains MSR_KERNEL here */
 	/* Check current_thread_info()->flags */
 	rlwinm	r9,r1,0,0,(31-THREAD_SHIFT)
 	lwz	r9,TI_FLAGS(r9)
-	andi.	r0,r9,(_TIF_SIGPENDING|_TIF_NEED_RESCHED)
+	andi.	r0,r9,(_TIF_SIGPENDING|_TIF_NEED_RESCHED|_TIF_RESTOREALL)
 	bne	do_work
 
 restore_user:

arch/powerpc/kernel/entry_64.S

@@ -113,9 +113,7 @@ system_call_common:
 	addi	r9,r1,STACK_FRAME_OVERHEAD
 #endif
 	clrrdi	r11,r1,THREAD_SHIFT
-	li	r12,0
 	ld	r10,TI_FLAGS(r11)
-	stb	r12,TI_SC_NOERR(r11)
 	andi.	r11,r10,_TIF_SYSCALL_T_OR_A
 	bne-	syscall_dotrace
 syscall_dotrace_cont:
@@ -144,24 +142,12 @@ system_call:			/* label this so stack traces look sane */
 	bctrl			/* Call handler */
 
 syscall_exit:
-#ifdef SHOW_SYSCALLS
-	std	r3,GPR3(r1)
-	bl	.do_show_syscall_exit
-	ld	r3,GPR3(r1)
-#endif
 	std	r3,RESULT(r1)
-	ld	r5,_CCR(r1)
-	li	r10,-_LAST_ERRNO
-	cmpld	r3,r10
+#ifdef SHOW_SYSCALLS
+	bl	.do_show_syscall_exit
+	ld	r3,RESULT(r1)
+#endif
 	clrrdi	r12,r1,THREAD_SHIFT
-	bge-	syscall_error
-syscall_error_cont:
-
-	/* check for syscall tracing or audit */
-	ld	r9,TI_FLAGS(r12)
-	andi.	r0,r9,(_TIF_SYSCALL_T_OR_A|_TIF_SINGLESTEP)
-	bne-	syscall_exit_trace
-syscall_exit_trace_cont:
 
 	/* disable interrupts so current_thread_info()->flags can't change,
 	   and so that we don't get interrupted after loading SRR0/1. */
@@ -173,8 +159,13 @@ syscall_exit_trace_cont:
 	rotldi	r10,r10,16
 	mtmsrd	r10,1
 	ld	r9,TI_FLAGS(r12)
-	andi.	r0,r9,(_TIF_SYSCALL_T_OR_A|_TIF_SIGPENDING|_TIF_NEED_RESCHED)
+	li	r11,-_LAST_ERRNO
+	andi.	r0,r9,(_TIF_SYSCALL_T_OR_A|_TIF_SINGLESTEP|_TIF_SIGPENDING|_TIF_NEED_RESCHED|_TIF_RESTOREALL|_TIF_SAVE_NVGPRS|_TIF_NOERROR)
 	bne-	syscall_exit_work
+	cmpld	r3,r11
+	ld	r5,_CCR(r1)
+	bge-	syscall_error
+syscall_error_cont:
 	ld	r7,_NIP(r1)
 	stdcx.	r0,0,r1			/* to clear the reservation */
 	andi.	r6,r8,MSR_PR
@@ -193,21 +184,12 @@ syscall_exit_trace_cont:
 	rfid
 	b	.	/* prevent speculative execution */
 
-syscall_enosys:
-	li	r3,-ENOSYS
-	std	r3,RESULT(r1)
-	clrrdi	r12,r1,THREAD_SHIFT
-	ld	r5,_CCR(r1)
-
-syscall_error:
-	lbz	r11,TI_SC_NOERR(r12)
-	cmpwi	0,r11,0
-	bne-	syscall_error_cont
-	neg	r3,r3
+syscall_error:	
 	oris	r5,r5,0x1000	/* Set SO bit in CR */
+	neg	r3,r3
 	std	r5,_CCR(r1)
 	b	syscall_error_cont
-        
+	
 /* Traced system call support */
 syscall_dotrace:
 	bl	.save_nvgprs
@@ -225,21 +207,69 @@ syscall_dotrace:
 	ld	r10,TI_FLAGS(r10)
 	b	syscall_dotrace_cont
 
-syscall_exit_trace:
-	std	r3,GPR3(r1)
-	bl	.save_nvgprs
+syscall_enosys:
+	li	r3,-ENOSYS
+	b	syscall_exit
+	
+syscall_exit_work:
+	/* If TIF_RESTOREALL is set, don't scribble on either r3 or ccr.
+	 If TIF_NOERROR is set, just save r3 as it is. */
+
+	andi.	r0,r9,_TIF_RESTOREALL
+	bne-	2f
+	cmpld	r3,r11		/* r10 is -LAST_ERRNO */
+	blt+	1f
+	andi.	r0,r9,_TIF_NOERROR
+	bne-	1f
+	ld	r5,_CCR(r1)
+	neg	r3,r3
+	oris	r5,r5,0x1000	/* Set SO bit in CR */
+	std	r5,_CCR(r1)
+1:	std	r3,GPR3(r1)
+2:	andi.	r0,r9,(_TIF_PERSYSCALL_MASK)
+	beq	4f
+
+	/* Clear per-syscall TIF flags if any are set, but _leave_
+	_TIF_SAVE_NVGPRS set in r9 since we haven't dealt with that
+	yet.  */
+
+	li	r11,_TIF_PERSYSCALL_MASK
+	addi	r12,r12,TI_FLAGS
+3:	ldarx	r10,0,r12
+	andc	r10,r10,r11
+	stdcx.	r10,0,r12
+	bne-	3b
+	subi	r12,r12,TI_FLAGS
+	
+4:	bl	.save_nvgprs
+	/* Anything else left to do? */
+	andi.	r0,r9,(_TIF_SYSCALL_T_OR_A|_TIF_SINGLESTEP|_TIF_SAVE_NVGPRS)
+	beq	.ret_from_except_lite
+
+	/* Re-enable interrupts */
+	mfmsr	r10
+	ori	r10,r10,MSR_EE
+	mtmsrd	r10,1
+
+	andi.	r0,r9,_TIF_SAVE_NVGPRS
+	bne	save_user_nvgprs
+
+	/* If tracing, re-enable interrupts and do it */
+save_user_nvgprs_cont:	
+	andi.	r0,r9,(_TIF_SYSCALL_T_OR_A|_TIF_SINGLESTEP)
+	beq	5f
+	
 	addi	r3,r1,STACK_FRAME_OVERHEAD
 	bl	.do_syscall_trace_leave
 	REST_NVGPRS(r1)
-	ld	r3,GPR3(r1)
-	ld	r5,_CCR(r1)
 	clrrdi	r12,r1,THREAD_SHIFT
-	b	syscall_exit_trace_cont
 
-/* Stuff to do on exit from a system call. */
-syscall_exit_work:
-	std	r3,GPR3(r1)
-	std	r5,_CCR(r1)
+	/* Disable interrupts again and handle other work if any */
+5:	mfmsr	r10
+	rldicl	r10,r10,48,1
+	rotldi	r10,r10,16
+	mtmsrd	r10,1
+
 	b	.ret_from_except_lite
 
 /* Save non-volatile GPRs, if not already saved. */
@@ -252,6 +282,52 @@ _GLOBAL(save_nvgprs)
 	std	r0,_TRAP(r1)
 	blr
 
+
+save_user_nvgprs:
+	ld	r10,TI_SIGFRAME(r12)
+	andi.	r0,r9,_TIF_32BIT
+	beq-	save_user_nvgprs_64
+
+	/* 32-bit save to userspace */
+
+.macro savewords start, end
+  1:	stw \start,4*(\start)(r10)
+	.section __ex_table,"a"
+	.align	3
+	.llong	1b,save_user_nvgprs_fault
+	.previous
+	.if \end - \start
+	savewords "(\start+1)",\end
+	.endif
+.endm	
+	savewords 14,31
+	b	save_user_nvgprs_cont
+
+save_user_nvgprs_64:
+	/* 64-bit save to userspace */
+
+.macro savelongs start, end
+  1:	std \start,8*(\start)(r10)
+	.section __ex_table,"a"
+	.align	3
+	.llong	1b,save_user_nvgprs_fault
+	.previous
+	.if \end - \start
+	savelongs "(\start+1)",\end
+	.endif
+.endm	
+	savelongs 14,31
+	b	save_user_nvgprs_cont
+
+save_user_nvgprs_fault:
+	li	r3,11		/* SIGSEGV */
+	ld	r4,TI_TASK(r12)
+	bl	.force_sigsegv
+
+	clrrdi	r12,r1,THREAD_SHIFT
+	ld	r9,TI_FLAGS(r12)
+	b	save_user_nvgprs_cont
+	
 /*
  * The sigsuspend and rt_sigsuspend system calls can call do_signal
  * and thus put the process into the stopped state where we might
@@ -260,35 +336,6 @@ _GLOBAL(save_nvgprs)
  * the C code.  Similarly, fork, vfork and clone need the full
  * register state on the stack so that it can be copied to the child.
  */
-_GLOBAL(ppc32_sigsuspend)
-	bl	.save_nvgprs
-	bl	.compat_sys_sigsuspend
-	b	70f
-
-_GLOBAL(ppc64_rt_sigsuspend)
-	bl	.save_nvgprs
-	bl	.sys_rt_sigsuspend
-	b	70f
-
-_GLOBAL(ppc32_rt_sigsuspend)
-	bl	.save_nvgprs
-	bl	.compat_sys_rt_sigsuspend
-70:	cmpdi	0,r3,0
-	/* If it returned an error, we need to return via syscall_exit to set
-	   the SO bit in cr0 and potentially stop for ptrace. */
-	bne	syscall_exit
-	/* If sigsuspend() returns zero, we are going into a signal handler. We
-	   may need to call audit_syscall_exit() to mark the exit from sigsuspend() */
-#ifdef CONFIG_AUDITSYSCALL
-	ld	r3,PACACURRENT(r13)
-	ld	r4,AUDITCONTEXT(r3)
-	cmpdi	0,r4,0
-	beq	.ret_from_except	/* No audit_context: Leave immediately. */
-	li	r4, 2			/* AUDITSC_FAILURE */
-	li	r5,-4			/* It's always -EINTR */
-	bl	.audit_syscall_exit
-#endif
-	b	.ret_from_except
 
 _GLOBAL(ppc_fork)
 	bl	.save_nvgprs
@@ -305,37 +352,6 @@ _GLOBAL(ppc_clone)
 	bl	.sys_clone
 	b	syscall_exit
 
-_GLOBAL(ppc32_swapcontext)
-	bl	.save_nvgprs
-	bl	.compat_sys_swapcontext
-	b	80f
-	
-_GLOBAL(ppc64_swapcontext)
-	bl	.save_nvgprs
-	bl	.sys_swapcontext
-	b	80f
-
-_GLOBAL(ppc32_sigreturn)
-	bl	.compat_sys_sigreturn
-	b	80f
-
-_GLOBAL(ppc32_rt_sigreturn)
-	bl	.compat_sys_rt_sigreturn
-	b	80f
-
-_GLOBAL(ppc64_rt_sigreturn)
-	bl	.sys_rt_sigreturn
-
-80:	cmpdi	0,r3,0
-	blt	syscall_exit
-	clrrdi	r4,r1,THREAD_SHIFT
-	ld	r4,TI_FLAGS(r4)
-	andi.	r4,r4,(_TIF_SYSCALL_T_OR_A|_TIF_SINGLESTEP)
-	beq+	81f
-	addi	r3,r1,STACK_FRAME_OVERHEAD
-	bl	.do_syscall_trace_leave
-81:	b	.ret_from_except
-
 _GLOBAL(ret_from_fork)
 	bl	.schedule_tail
 	REST_NVGPRS(r1)
@@ -674,7 +690,7 @@ _GLOBAL(enter_rtas)
 
 	/* Setup our real return addr */	
 	SET_REG_TO_LABEL(r4,.rtas_return_loc)
-	SET_REG_TO_CONST(r9,KERNELBASE)
+	SET_REG_TO_CONST(r9,PAGE_OFFSET)
 	sub	r4,r4,r9
        	mtlr	r4
 
@@ -702,7 +718,7 @@ _GLOBAL(enter_rtas)
 _STATIC(rtas_return_loc)
 	/* relocation is off at this point */
 	mfspr	r4,SPRN_SPRG3	        /* Get PACA */
-	SET_REG_TO_CONST(r5, KERNELBASE)
+	SET_REG_TO_CONST(r5, PAGE_OFFSET)
         sub     r4,r4,r5                /* RELOC the PACA base pointer */
 
 	mfmsr   r6

arch/powerpc/kernel/head_32.S

@@ -120,10 +120,25 @@ __start:
  * because OF may have I/O devices mapped into that area
  * (particularly on CHRP).
  */
+#ifdef CONFIG_PPC_MULTIPLATFORM
 	cmpwi	0,r5,0
 	beq	1f
 	bl	prom_init
 	trap
+#endif
+
+/*
+ * Check for BootX signature when supporting PowerMac and branch to
+ * appropriate trampoline if it's present
+ */
+#ifdef CONFIG_PPC_PMAC
+1:	lis	r31,0x426f
+	ori	r31,r31,0x6f58
+	cmpw	0,r3,r31
+	bne	1f
+	bl	bootx_init
+	trap
+#endif /* CONFIG_PPC_PMAC */
 
 1:	mr	r31,r3			/* save parameters */
 	mr	r30,r4
@@ -153,6 +168,9 @@ __after_mmu_off:
 	bl	flush_tlbs
 
 	bl	initial_bats
+#if !defined(CONFIG_APUS) && defined(CONFIG_BOOTX_TEXT)
+	bl	setup_disp_bat
+#endif
 
 /*
  * Call setup_cpu for CPU 0 and initialize 6xx Idle
@@ -450,16 +468,11 @@ SystemCall:
  * by executing an altivec instruction.
  */
 	. = 0xf00
-	b	Trap_0f
+	b	PerformanceMonitor
 
 	. = 0xf20
 	b	AltiVecUnavailable
 
-Trap_0f:
-	EXCEPTION_PROLOG
-	addi	r3,r1,STACK_FRAME_OVERHEAD
-	EXC_XFER_EE(0xf00, unknown_exception)
-
 /*
  * Handle TLB miss for instruction on 603/603e.
  * Note: we get an alternate set of r0 - r3 to use automatically.
@@ -703,6 +716,11 @@ AltiVecUnavailable:
 #endif /* CONFIG_ALTIVEC */
 	EXC_XFER_EE_LITE(0xf20, altivec_unavailable_exception)
 
+PerformanceMonitor:
+	EXCEPTION_PROLOG
+	addi	r3,r1,STACK_FRAME_OVERHEAD
+	EXC_XFER_STD(0xf00, performance_monitor_exception)
+
 #ifdef CONFIG_ALTIVEC
 /* Note that the AltiVec support is closely modeled after the FP
  * support.  Changes to one are likely to be applicable to the
@@ -1306,6 +1324,32 @@ initial_bats:
 	blr
 
 
+#if !defined(CONFIG_APUS) && defined(CONFIG_BOOTX_TEXT)
+setup_disp_bat:
+	/*
+	 * setup the display bat prepared for us in prom.c
+	 */
+	mflr	r8
+	bl	reloc_offset
+	mtlr	r8
+	addis	r8,r3,disp_BAT@ha
+	addi	r8,r8,disp_BAT@l
+	cmpwi	cr0,r8,0
+	beqlr
+	lwz	r11,0(r8)
+	lwz	r8,4(r8)
+	mfspr	r9,SPRN_PVR
+	rlwinm	r9,r9,16,16,31		/* r9 = 1 for 601, 4 for 604 */
+	cmpwi	0,r9,1
+	beq	1f
+	mtspr	SPRN_DBAT3L,r8
+	mtspr	SPRN_DBAT3U,r11
+	blr
+1:	mtspr	SPRN_IBAT3L,r8
+	mtspr	SPRN_IBAT3U,r11
+	blr
+#endif /* !defined(CONFIG_APUS) && defined(CONFIG_BOOTX_TEXT) */
+
 #ifdef CONFIG_8260
 /* Jump into the system reset for the rom.
  * We first disable the MMU, and then jump to the ROM reset address.

arch/powerpc/kernel/head_64.S

@@ -154,11 +154,15 @@ _GLOBAL(__secondary_hold)
 	bne	100b
 
 #ifdef CONFIG_HMT
-	b	.hmt_init
+	LOADADDR(r4, .hmt_init)
+	mtctr	r4
+	bctr
 #else
 #ifdef CONFIG_SMP
+	LOADADDR(r4, .pSeries_secondary_smp_init)
+	mtctr	r4
 	mr	r3,r24
-	b	.pSeries_secondary_smp_init
+	bctr
 #else
 	BUG_OPCODE
 #endif
@@ -200,6 +204,20 @@ exception_marker:
 #define EX_R3		64
 #define EX_LR		72
 
+/*
+ * We're short on space and time in the exception prolog, so we can't use
+ * the normal LOADADDR macro. Normally we just need the low halfword of the
+ * address, but for Kdump we need the whole low word.
+ */
+#ifdef CONFIG_CRASH_DUMP
+#define LOAD_HANDLER(reg, label)					\
+	oris	reg,reg,(label)@h;	/* virt addr of handler ... */	\
+	ori	reg,reg,(label)@l;	/* .. and the rest */
+#else
+#define LOAD_HANDLER(reg, label)					\
+	ori	reg,reg,(label)@l;	/* virt addr of handler ... */
+#endif
+
 #define EXCEPTION_PROLOG_PSERIES(area, label)				\
 	mfspr	r13,SPRN_SPRG3;		/* get paca address into r13 */	\
 	std	r9,area+EX_R9(r13);	/* save r9 - r12 */		\
@@ -212,7 +230,7 @@ exception_marker:
 	clrrdi	r12,r13,32;		/* get high part of &label */	\
 	mfmsr	r10;							\
 	mfspr	r11,SPRN_SRR0;		/* save SRR0 */			\
-	ori	r12,r12,(label)@l;	/* virt addr of handler */	\
+	LOAD_HANDLER(r12,label)						\
 	ori	r10,r10,MSR_IR|MSR_DR|MSR_RI;				\
 	mtspr	SPRN_SRR0,r12;						\
 	mfspr	r12,SPRN_SRR1;		/* and SRR1 */			\
@@ -553,6 +571,7 @@ slb_miss_user_pseries:
  * Vectors for the FWNMI option.  Share common code.
  */
 	.globl system_reset_fwnmi
+      .align 7
 system_reset_fwnmi:
 	HMT_MEDIUM
 	mtspr	SPRN_SPRG1,r13		/* save r13 */
@@ -560,6 +579,7 @@ system_reset_fwnmi:
 	EXCEPTION_PROLOG_PSERIES(PACA_EXGEN, system_reset_common)
 
 	.globl machine_check_fwnmi
+      .align 7
 machine_check_fwnmi:
 	HMT_MEDIUM
 	mtspr	SPRN_SPRG1,r13		/* save r13 */
@@ -726,7 +746,8 @@ iSeries_secondary_smp_loop:
 decrementer_iSeries_masked:
 	li	r11,1
 	stb	r11,PACALPPACA+LPPACADECRINT(r13)
-	lwz	r12,PACADEFAULTDECR(r13)
+	LOADBASE(r12,tb_ticks_per_jiffy)
+	lwz	r12,OFF(tb_ticks_per_jiffy)(r12)
 	mtspr	SPRN_DEC,r12
 	/* fall through */
 
@@ -1345,7 +1366,7 @@ _GLOBAL(do_stab_bolted)
  * fixed address (the linker can't compute (u64)&initial_stab >>
  * PAGE_SHIFT).
  */
-	. = STAB0_PHYS_ADDR	/* 0x6000 */
+	. = STAB0_OFFSET	/* 0x6000 */
 	.globl initial_stab
 initial_stab:
 	.space	4096
@@ -1485,11 +1506,13 @@ _STATIC(__mmu_off)
  *
  */
 _GLOBAL(__start_initialization_multiplatform)
+#ifdef CONFIG_PPC_MULTIPLATFORM
 	/*
 	 * Are we booted from a PROM Of-type client-interface ?
 	 */
 	cmpldi	cr0,r5,0
 	bne	.__boot_from_prom		/* yes -> prom */
+#endif
 
 	/* Save parameters */
 	mr	r31,r3
@@ -1510,6 +1533,7 @@ _GLOBAL(__start_initialization_multiplatform)
 	bl	.__mmu_off
 	b	.__after_prom_start
 
+#ifdef CONFIG_PPC_MULTIPLATFORM
 _STATIC(__boot_from_prom)
 	/* Save parameters */
 	mr	r31,r3
@@ -1542,6 +1566,7 @@ _STATIC(__boot_from_prom)
 	bl	.prom_init
 	/* We never return */
 	trap
+#endif
 
 /*
  * At this point, r3 contains the physical address we are running at,
@@ -1550,7 +1575,7 @@ _STATIC(__boot_from_prom)
 _STATIC(__after_prom_start)
 
 /*
- * We need to run with __start at physical address 0.
+ * We need to run with __start at physical address PHYSICAL_START.
  * This will leave some code in the first 256B of
  * real memory, which are reserved for software use.
  * The remainder of the first page is loaded with the fixed
@@ -1565,7 +1590,7 @@ _STATIC(__after_prom_start)
 	mr	r26,r3
 	SET_REG_TO_CONST(r27,KERNELBASE)
 
-	li	r3,0			/* target addr */
+	LOADADDR(r3, PHYSICAL_START)	/* target addr */
 
 	// XXX FIXME: Use phys returned by OF (r30)
 	add	r4,r27,r26 		/* source addr			 */
@@ -1846,7 +1871,7 @@ _STATIC(start_here_multiplatform)
 	mulli	r13,r27,PACA_SIZE	/* Calculate vaddr of right paca */
 	add	r13,r13,r24		/* for this processor.		 */
 	add	r13,r13,r26		/* convert to physical addr	 */
-	mtspr	SPRN_SPRG3,r13		/* PPPBBB: Temp... -Peter */
+	mtspr	SPRN_SPRG3,r13
 	
 	/* Do very early kernel initializations, including initial hash table,
 	 * stab and slb setup before we turn on relocation.	*/

arch/powerpc/kernel/ibmebus.c

@@ -0,0 +1,396 @@
+/*
+ * IBM PowerPC IBM eBus Infrastructure Support.
+ *
+ * Copyright (c) 2005 IBM Corporation
+ *  Heiko J Schick <schickhj@de.ibm.com>
+ *    
+ * All rights reserved.
+ *
+ * This source code is distributed under a dual license of GPL v2.0 and OpenIB 
+ * BSD. 
+ *
+ * OpenIB BSD License
+ *
+ * Redistribution and use in source and binary forms, with or without 
+ * modification, are permitted provided that the following conditions are met: 
+ *
+ * Redistributions of source code must retain the above copyright notice, this 
+ * list of conditions and the following disclaimer. 
+ *
+ * Redistributions in binary form must reproduce the above copyright notice, 
+ * this list of conditions and the following disclaimer in the documentation 
+ * and/or other materials
+ * provided with the distribution. 
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" 
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE 
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
+ * BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER
+ * IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include <linux/init.h>
+#include <linux/console.h>
+#include <linux/kobject.h>
+#include <linux/dma-mapping.h>
+#include <linux/interrupt.h>
+#include <asm/ibmebus.h>
+#include <asm/abs_addr.h>
+
+static struct ibmebus_dev ibmebus_bus_device = { /* fake "parent" device */
+	.name = ibmebus_bus_device.ofdev.dev.bus_id,
+	.ofdev.dev.bus_id = "ibmebus",
+	.ofdev.dev.bus    = &ibmebus_bus_type,
+};
+
+static void *ibmebus_alloc_coherent(struct device *dev,
+				    size_t size,
+				    dma_addr_t *dma_handle,
+				    gfp_t flag)
+{
+	void *mem;
+	
+	mem = kmalloc(size, flag);
+	*dma_handle = (dma_addr_t)mem;
+
+	return mem;
+}
+
+static void ibmebus_free_coherent(struct device *dev,
+				  size_t size, void *vaddr, 
+				  dma_addr_t dma_handle)
+{
+	kfree(vaddr);
+}
+
+static dma_addr_t ibmebus_map_single(struct device *dev,
+				     void *ptr,
+				     size_t size,
+				     enum dma_data_direction direction)
+{
+	return (dma_addr_t)(ptr);
+}
+
+static void ibmebus_unmap_single(struct device *dev,
+				 dma_addr_t dma_addr,
+				 size_t size, 
+				 enum dma_data_direction direction)
+{
+	return;
+}
+
+static int ibmebus_map_sg(struct device *dev,
+			  struct scatterlist *sg,
+			  int nents, enum dma_data_direction direction)
+{
+	int i;
+	
+	for (i = 0; i < nents; i++) {
+		sg[i].dma_address = (dma_addr_t)page_address(sg[i].page) 
+			+ sg[i].offset;
+		sg[i].dma_length = sg[i].length;
+	}
+	
+	return nents;
+}
+
+static void ibmebus_unmap_sg(struct device *dev,
+			     struct scatterlist *sg,
+			     int nents, enum dma_data_direction direction)
+{
+	return;
+}
+
+static int ibmebus_dma_supported(struct device *dev, u64 mask)
+{
+	return 1;
+}
+
+struct dma_mapping_ops ibmebus_dma_ops = {
+	.alloc_coherent = ibmebus_alloc_coherent,
+	.free_coherent  = ibmebus_free_coherent,
+	.map_single     = ibmebus_map_single,
+	.unmap_single   = ibmebus_unmap_single,
+	.map_sg         = ibmebus_map_sg,
+	.unmap_sg       = ibmebus_unmap_sg,
+	.dma_supported  = ibmebus_dma_supported,
+};
+
+static int ibmebus_bus_probe(struct device *dev)
+{
+	struct ibmebus_dev *ibmebusdev    = to_ibmebus_dev(dev);
+	struct ibmebus_driver *ibmebusdrv = to_ibmebus_driver(dev->driver);
+	const struct of_device_id *id;
+	int error = -ENODEV;
+	
+	if (!ibmebusdrv->probe)
+		return error;
+	
+	id = of_match_device(ibmebusdrv->id_table, &ibmebusdev->ofdev);
+	if (id) {
+		error = ibmebusdrv->probe(ibmebusdev, id);
+	}
+	
+	return error;
+}
+
+static int ibmebus_bus_remove(struct device *dev)
+{
+	struct ibmebus_dev *ibmebusdev    = to_ibmebus_dev(dev);
+	struct ibmebus_driver *ibmebusdrv = to_ibmebus_driver(dev->driver);
+	
+	if (ibmebusdrv->remove) {
+		return ibmebusdrv->remove(ibmebusdev);
+	}
+	
+	return 0;
+}
+
+static void __devinit ibmebus_dev_release(struct device *dev)
+{
+	of_node_put(to_ibmebus_dev(dev)->ofdev.node);
+	kfree(to_ibmebus_dev(dev));
+}
+
+static ssize_t ibmebusdev_show_name(struct device *dev, 
+				    struct device_attribute *attr, char *buf)
+{
+	return sprintf(buf, "%s\n", to_ibmebus_dev(dev)->name);
+}
+static DEVICE_ATTR(name, S_IRUSR | S_IRGRP | S_IROTH, ibmebusdev_show_name, 
+		   NULL);
+
+static struct ibmebus_dev* __devinit ibmebus_register_device_common(
+	struct ibmebus_dev *dev, char *name)
+{
+	int err = 0;
+
+	dev->name = name;
+	dev->ofdev.dev.parent  = &ibmebus_bus_device.ofdev.dev;
+	dev->ofdev.dev.bus     = &ibmebus_bus_type;
+	dev->ofdev.dev.release = ibmebus_dev_release;
+
+	/* An ibmebusdev is based on a of_device. We have to change the
+	 * bus type to use our own DMA mapping operations. 
+	 */       
+	if ((err = of_device_register(&dev->ofdev)) != 0) {
+		printk(KERN_ERR "%s: failed to register device (%d).\n",
+		       __FUNCTION__, err);
+		return NULL;
+	}
+	
+	device_create_file(&dev->ofdev.dev, &dev_attr_name);
+	
+	return dev;
+}
+
+static struct ibmebus_dev* __devinit ibmebus_register_device_node(
+	struct device_node *dn)
+{
+	struct ibmebus_dev *dev;
+	char *loc_code;
+	int length;
+
+	loc_code = (char *)get_property(dn, "ibm,loc-code", NULL);
+	if (!loc_code) {
+                printk(KERN_WARNING "%s: node %s missing 'ibm,loc-code'\n",
+		       __FUNCTION__, dn->name ? dn->name : "<unknown>");
+		return NULL;
+        }
+	
+	if (strlen(loc_code) == 0) {
+	        printk(KERN_WARNING "%s: 'ibm,loc-code' is invalid\n",
+		       __FUNCTION__);
+		return NULL;
+	}
+
+	dev = kmalloc(sizeof(struct ibmebus_dev), GFP_KERNEL);
+	if (!dev) {
+		return NULL;
+	}
+	memset(dev, 0, sizeof(struct ibmebus_dev));
+
+	dev->ofdev.node = of_node_get(dn);
+       
+	length = strlen(loc_code);
+	memcpy(dev->ofdev.dev.bus_id, loc_code 
+		+ (length - min(length, BUS_ID_SIZE - 1)), 
+		min(length, BUS_ID_SIZE - 1));
+
+	/* Register with generic device framework. */
+	if (ibmebus_register_device_common(dev, dn->name) == NULL) {
+		kfree(dev);
+		return NULL;
+	}
+
+	return dev;
+}
+
+static void ibmebus_probe_of_nodes(char* name)
+{
+	struct device_node *dn = NULL;
+	
+	while ((dn = of_find_node_by_name(dn, name))) {
+		if (ibmebus_register_device_node(dn) == NULL) {
+			of_node_put(dn);
+			
+			return;
+		}
+	}
+	
+	of_node_put(dn);
+	
+	return;
+}
+
+static void ibmebus_add_devices_by_id(struct of_device_id *idt)
+{
+	while (strlen(idt->name) > 0) {
+		ibmebus_probe_of_nodes(idt->name);
+		idt++;
+	}
+
+	return;
+}
+
+static int ibmebus_match_helper(struct device *dev, void *data)
+{
+	if (strcmp((char*)data, to_ibmebus_dev(dev)->name) == 0)
+		return 1;
+	
+	return 0;
+}
+
+static int ibmebus_unregister_device(struct device *dev)
+{
+	device_remove_file(dev, &dev_attr_name);
+	of_device_unregister(to_of_device(dev));
+
+	return 0;
+}
+
+static void ibmebus_remove_devices_by_id(struct of_device_id *idt)
+{
+	struct device *dev;
+	
+	while (strlen(idt->name) > 0) {
+		while ((dev = bus_find_device(&ibmebus_bus_type, NULL, 
+					      (void*)idt->name,
+					      ibmebus_match_helper))) {
+			ibmebus_unregister_device(dev);
+		}
+		idt++;
+		
+	}
+	
+	return;
+}
+
+int ibmebus_register_driver(struct ibmebus_driver *drv)
+{
+	int err = 0;
+
+	drv->driver.name   = drv->name;
+	drv->driver.bus    = &ibmebus_bus_type;
+	drv->driver.probe  = ibmebus_bus_probe;
+	drv->driver.remove = ibmebus_bus_remove;
+
+	if ((err = driver_register(&drv->driver) != 0))
+		return err;
+
+	ibmebus_add_devices_by_id(drv->id_table);
+	
+	return 0;
+}
+EXPORT_SYMBOL(ibmebus_register_driver);
+
+void ibmebus_unregister_driver(struct ibmebus_driver *drv)
+{	
+	driver_unregister(&drv->driver);
+	ibmebus_remove_devices_by_id(drv->id_table);
+}
+EXPORT_SYMBOL(ibmebus_unregister_driver);
+
+int ibmebus_request_irq(struct ibmebus_dev *dev,
+			u32 ist, 
+			irqreturn_t (*handler)(int, void*, struct pt_regs *),
+			unsigned long irq_flags, const char * devname,
+			void *dev_id)
+{
+	unsigned int irq = virt_irq_create_mapping(ist);
+	
+	if (irq == NO_IRQ)
+		return -EINVAL;
+	
+	irq = irq_offset_up(irq);
+	
+	return request_irq(irq, handler,
+			   irq_flags, devname, dev_id);
+}
+EXPORT_SYMBOL(ibmebus_request_irq);
+
+void ibmebus_free_irq(struct ibmebus_dev *dev, u32 ist, void *dev_id)
+{
+	unsigned int irq = virt_irq_create_mapping(ist);
+	
+	irq = irq_offset_up(irq);
+	free_irq(irq, dev_id);
+	
+	return;
+}
+EXPORT_SYMBOL(ibmebus_free_irq);
+
+static int ibmebus_bus_match(struct device *dev, struct device_driver *drv)
+{	
+	const struct ibmebus_dev *ebus_dev = to_ibmebus_dev(dev);
+	struct ibmebus_driver *ebus_drv    = to_ibmebus_driver(drv);
+	const struct of_device_id *ids     = ebus_drv->id_table;
+	const struct of_device_id *found_id;
+	
+	if (!ids)
+		return 0;
+	
+	found_id = of_match_device(ids, &ebus_dev->ofdev);
+	if (found_id)
+		return 1;
+	
+	return 0;
+}
+
+struct bus_type ibmebus_bus_type = {
+	.name = "ibmebus",
+	.match = ibmebus_bus_match,
+};
+EXPORT_SYMBOL(ibmebus_bus_type);
+
+static int __init ibmebus_bus_init(void)
+{
+	int err;
+	
+	printk(KERN_INFO "IBM eBus Device Driver\n");
+	
+	err = bus_register(&ibmebus_bus_type);
+	if (err) {
+		printk(KERN_ERR ":%s: failed to register IBM eBus.\n",
+		       __FUNCTION__);
+		return err;
+	}
+	
+	err = device_register(&ibmebus_bus_device.ofdev.dev);
+	if (err) {
+		printk(KERN_WARNING "%s: device_register returned %i\n", 
+		       __FUNCTION__, err);
+		bus_unregister(&ibmebus_bus_type);
+
+		return err;
+	}
+	
+	return 0;
+}
+__initcall(ibmebus_bus_init);

arch/powerpc/kernel/irq.c

@@ -31,7 +31,6 @@
  * to reduce code space and undefined function references.
  */
 
-#include <linux/errno.h>
 #include <linux/module.h>
 #include <linux/threads.h>
 #include <linux/kernel_stat.h>
@@ -44,18 +43,12 @@
 #include <linux/config.h>
 #include <linux/init.h>
 #include <linux/slab.h>
-#include <linux/pci.h>
 #include <linux/delay.h>
 #include <linux/irq.h>
-#include <linux/proc_fs.h>
-#include <linux/random.h>
 #include <linux/seq_file.h>
 #include <linux/cpumask.h>
 #include <linux/profile.h>
 #include <linux/bitops.h>
-#ifdef CONFIG_PPC64
-#include <linux/kallsyms.h>
-#endif
 
 #include <asm/uaccess.h>
 #include <asm/system.h>
@@ -66,8 +59,7 @@
 #include <asm/prom.h>
 #include <asm/ptrace.h>
 #include <asm/machdep.h>
-#ifdef CONFIG_PPC64
-#include <asm/iseries/it_lp_queue.h>
+#ifdef CONFIG_PPC_ISERIES
 #include <asm/paca.h>
 #endif
 
@@ -78,10 +70,6 @@ EXPORT_SYMBOL(__irq_offset_value);
 
 static int ppc_spurious_interrupts;
 
-#if defined(CONFIG_PPC_ISERIES) && defined(CONFIG_SMP)
-extern void iSeries_smp_message_recv(struct pt_regs *);
-#endif
-
 #ifdef CONFIG_PPC32
 #define NR_MASK_WORDS	((NR_IRQS + 31) / 32)
 
@@ -195,49 +183,6 @@ void fixup_irqs(cpumask_t map)
 }
 #endif
 
-#ifdef CONFIG_PPC_ISERIES
-void do_IRQ(struct pt_regs *regs)
-{
-	struct paca_struct *lpaca;
-
-	irq_enter();
-
-#ifdef CONFIG_DEBUG_STACKOVERFLOW
-	/* Debugging check for stack overflow: is there less than 2KB free? */
-	{
-		long sp;
-
-		sp = __get_SP() & (THREAD_SIZE-1);
-
-		if (unlikely(sp < (sizeof(struct thread_info) + 2048))) {
-			printk("do_IRQ: stack overflow: %ld\n",
-				sp - sizeof(struct thread_info));
-			dump_stack();
-		}
-	}
-#endif
-
-	lpaca = get_paca();
-#ifdef CONFIG_SMP
-	if (lpaca->lppaca.int_dword.fields.ipi_cnt) {
-		lpaca->lppaca.int_dword.fields.ipi_cnt = 0;
-		iSeries_smp_message_recv(regs);
-	}
-#endif /* CONFIG_SMP */
-	if (hvlpevent_is_pending())
-		process_hvlpevents(regs);
-
-	irq_exit();
-
-	if (lpaca->lppaca.int_dword.fields.decr_int) {
-		lpaca->lppaca.int_dword.fields.decr_int = 0;
-		/* Signal a fake decrementer interrupt */
-		timer_interrupt(regs);
-	}
-}
-
-#else	/* CONFIG_PPC_ISERIES */
-
 void do_IRQ(struct pt_regs *regs)
 {
 	int irq;
@@ -286,16 +231,24 @@ void do_IRQ(struct pt_regs *regs)
 		} else
 #endif
 			__do_IRQ(irq, regs);
-	} else
-#ifdef CONFIG_PPC32
-		if (irq != -2)
-#endif
-			/* That's not SMP safe ... but who cares ? */
-			ppc_spurious_interrupts++;
-        irq_exit();
-}
+	} else if (irq != -2)
+		/* That's not SMP safe ... but who cares ? */
+		ppc_spurious_interrupts++;
 
-#endif	/* CONFIG_PPC_ISERIES */
+        irq_exit();
+
+#ifdef CONFIG_PPC_ISERIES
+	{
+		struct paca_struct *lpaca = get_paca();
+
+		if (lpaca->lppaca.int_dword.fields.decr_int) {
+			lpaca->lppaca.int_dword.fields.decr_int = 0;
+			/* Signal a fake decrementer interrupt */
+			timer_interrupt(regs);
+		}
+	}
+#endif
+}
 
 void __init init_IRQ(void)
 {

arch/powerpc/kernel/legacy_serial.c

@@ -0,0 +1,557 @@
+#include <linux/config.h>
+#include <linux/kernel.h>
+#include <linux/serial.h>
+#include <linux/serial_8250.h>
+#include <linux/serial_core.h>
+#include <linux/console.h>
+#include <linux/pci.h>
+#include <asm/io.h>
+#include <asm/mmu.h>
+#include <asm/prom.h>
+#include <asm/serial.h>
+#include <asm/udbg.h>
+#include <asm/pci-bridge.h>
+#include <asm/ppc-pci.h>
+
+#undef DEBUG
+
+#ifdef DEBUG
+#define DBG(fmt...) do { printk(fmt); } while(0)
+#else
+#define DBG(fmt...) do { } while(0)
+#endif
+
+#define MAX_LEGACY_SERIAL_PORTS	8
+
+static struct plat_serial8250_port
+legacy_serial_ports[MAX_LEGACY_SERIAL_PORTS+1];
+static struct legacy_serial_info {
+	struct device_node		*np;
+	unsigned int			speed;
+	unsigned int			clock;
+	phys_addr_t			taddr;
+} legacy_serial_infos[MAX_LEGACY_SERIAL_PORTS];
+static unsigned int legacy_serial_count;
+static int legacy_serial_console = -1;
+
+static int __init add_legacy_port(struct device_node *np, int want_index,
+				  int iotype, phys_addr_t base,
+				  phys_addr_t taddr, unsigned long irq,
+				  unsigned int flags)
+{
+	u32 *clk, *spd, clock = BASE_BAUD * 16;
+	int index;
+
+	/* get clock freq. if present */
+	clk = (u32 *)get_property(np, "clock-frequency", NULL);
+	if (clk && *clk)
+		clock = *clk;
+
+	/* get default speed if present */
+	spd = (u32 *)get_property(np, "current-speed", NULL);
+
+	/* If we have a location index, then try to use it */
+	if (want_index >= 0 && want_index < MAX_LEGACY_SERIAL_PORTS)
+		index = want_index;
+	else
+		index = legacy_serial_count;
+
+	/* if our index is still out of range, that mean that
+	 * array is full, we could scan for a free slot but that
+	 * make little sense to bother, just skip the port
+	 */
+	if (index >= MAX_LEGACY_SERIAL_PORTS)
+		return -1;
+	if (index >= legacy_serial_count)
+		legacy_serial_count = index + 1;
+
+	/* Check if there is a port who already claimed our slot */
+	if (legacy_serial_infos[index].np != 0) {
+		/* if we still have some room, move it, else override */
+		if (legacy_serial_count < MAX_LEGACY_SERIAL_PORTS) {
+			printk(KERN_INFO "Moved legacy port %d -> %d\n",
+			       index, legacy_serial_count);
+			legacy_serial_ports[legacy_serial_count] =
+				legacy_serial_ports[index];
+			legacy_serial_infos[legacy_serial_count] =
+				legacy_serial_infos[index];
+			legacy_serial_count++;
+		} else {
+			printk(KERN_INFO "Replacing legacy port %d\n", index);
+		}
+	}
+
+	/* Now fill the entry */
+	memset(&legacy_serial_ports[index], 0,
+	       sizeof(struct plat_serial8250_port));
+	if (iotype == UPIO_PORT)
+		legacy_serial_ports[index].iobase = base;
+	else
+		legacy_serial_ports[index].mapbase = base;
+	legacy_serial_ports[index].iotype = iotype;
+	legacy_serial_ports[index].uartclk = clock;
+	legacy_serial_ports[index].irq = irq;
+	legacy_serial_ports[index].flags = flags;
+	legacy_serial_infos[index].taddr = taddr;
+	legacy_serial_infos[index].np = of_node_get(np);
+	legacy_serial_infos[index].clock = clock;
+	legacy_serial_infos[index].speed = spd ? *spd : 0;
+
+	printk(KERN_INFO "Found legacy serial port %d for %s\n",
+	       index, np->full_name);
+	printk(KERN_INFO "  %s=%llx, taddr=%llx, irq=%lx, clk=%d, speed=%d\n",
+	       (iotype == UPIO_PORT) ? "port" : "mem",
+	       (unsigned long long)base, (unsigned long long)taddr, irq,
+	       legacy_serial_ports[index].uartclk,
+	       legacy_serial_infos[index].speed);
+
+	return index;
+}
+
+static int __init add_legacy_soc_port(struct device_node *np,
+				      struct device_node *soc_dev)
+{
+	phys_addr_t addr;
+	u32 *addrp;
+	unsigned int flags = UPF_BOOT_AUTOCONF | UPF_SKIP_TEST | UPF_SHARE_IRQ;
+
+	/* We only support ports that have a clock frequency properly
+	 * encoded in the device-tree.
+	 */
+	if (get_property(np, "clock-frequency", NULL) == NULL)
+		return -1;
+
+	/* Get the address */
+	addrp = of_get_address(soc_dev, 0, NULL, NULL);
+	if (addrp == NULL)
+		return -1;
+
+	addr = of_translate_address(soc_dev, addrp);
+
+	/* Add port, irq will be dealt with later. We passed a translated
+	 * IO port value. It will be fixed up later along with the irq
+	 */
+	return add_legacy_port(np, -1, UPIO_MEM, addr, addr, NO_IRQ, flags);
+}
+
+#ifdef CONFIG_ISA
+static int __init add_legacy_isa_port(struct device_node *np,
+				      struct device_node *isa_brg)
+{
+	u32 *reg;
+	char *typep;
+	int index = -1;
+	phys_addr_t taddr;
+
+	/* Get the ISA port number */
+	reg = (u32 *)get_property(np, "reg", NULL);
+	if (reg == NULL)
+		return -1;
+
+	/* Verify it's an IO port, we don't support anything else */
+	if (!(reg[0] & 0x00000001))
+		return -1;
+
+	/* Now look for an "ibm,aix-loc" property that gives us ordering
+	 * if any...
+	 */
+	typep = (char *)get_property(np, "ibm,aix-loc", NULL);
+
+	/* If we have a location index, then use it */
+	if (typep && *typep == 'S')
+		index = simple_strtol(typep+1, NULL, 0) - 1;
+
+	/* Translate ISA address */
+	taddr = of_translate_address(np, reg);
+
+	/* Add port, irq will be dealt with later */
+	return add_legacy_port(np, index, UPIO_PORT, reg[1], taddr, NO_IRQ, UPF_BOOT_AUTOCONF);
+
+}
+#endif
+
+#ifdef CONFIG_PCI
+static int __init add_legacy_pci_port(struct device_node *np,
+				      struct device_node *pci_dev)
+{
+	phys_addr_t addr, base;
+	u32 *addrp;
+	unsigned int flags;
+	int iotype, index = -1, lindex = 0;
+
+	/* We only support ports that have a clock frequency properly
+	 * encoded in the device-tree (that is have an fcode). Anything
+	 * else can't be used that early and will be normally probed by
+	 * the generic 8250_pci driver later on. The reason is that 8250
+	 * compatible UARTs on PCI need all sort of quirks (port offsets
+	 * etc...) that this code doesn't know about
+	 */
+	if (get_property(np, "clock-frequency", NULL) == NULL)
+		return -1;
+
+	/* Get the PCI address. Assume BAR 0 */
+	addrp = of_get_pci_address(pci_dev, 0, NULL, &flags);
+	if (addrp == NULL)
+		return -1;
+
+	/* We only support BAR 0 for now */
+	iotype = (flags & IORESOURCE_MEM) ? UPIO_MEM : UPIO_PORT;
+	addr = of_translate_address(pci_dev, addrp);
+
+	/* Set the IO base to the same as the translated address for MMIO,
+	 * or to the domain local IO base for PIO (it will be fixed up later)
+	 */
+	if (iotype == UPIO_MEM)
+		base = addr;
+	else
+		base = addrp[2];
+
+	/* Try to guess an index... If we have subdevices of the pci dev,
+	 * we get to their "reg" property
+	 */
+	if (np != pci_dev) {
+		u32 *reg = (u32 *)get_property(np, "reg", NULL);
+		if (reg && (*reg < 4))
+			index = lindex = *reg;
+	}
+
+	/* Local index means it's the Nth port in the PCI chip. Unfortunately
+	 * the offset to add here is device specific. We know about those
+	 * EXAR ports and we default to the most common case. If your UART
+	 * doesn't work for these settings, you'll have to add your own special
+	 * cases here
+	 */
+	if (device_is_compatible(pci_dev, "pci13a8,152") ||
+	    device_is_compatible(pci_dev, "pci13a8,154") ||
+	    device_is_compatible(pci_dev, "pci13a8,158")) {
+		addr += 0x200 * lindex;
+		base += 0x200 * lindex;
+	} else {
+		addr += 8 * lindex;
+		base += 8 * lindex;
+	}
+
+	/* Add port, irq will be dealt with later. We passed a translated
+	 * IO port value. It will be fixed up later along with the irq
+	 */
+	return add_legacy_port(np, index, iotype, base, addr, NO_IRQ, UPF_BOOT_AUTOCONF);
+}
+#endif
+
+/*
+ * This is called very early, as part of setup_system() or eventually
+ * setup_arch(), basically before anything else in this file. This function
+ * will try to build a list of all the available 8250-compatible serial ports
+ * in the machine using the Open Firmware device-tree. It currently only deals
+ * with ISA and PCI busses but could be extended. It allows a very early boot
+ * console to be initialized, that list is also used later to provide 8250 with
+ * the machine non-PCI ports and to properly pick the default console port
+ */
+void __init find_legacy_serial_ports(void)
+{
+	struct device_node *np, *stdout = NULL;
+	char *path;
+	int index;
+
+	DBG(" -> find_legacy_serial_port()\n");
+
+	/* Now find out if one of these is out firmware console */
+	path = (char *)get_property(of_chosen, "linux,stdout-path", NULL);
+	if (path != NULL) {
+		stdout = of_find_node_by_path(path);
+		if (stdout)
+			DBG("stdout is %s\n", stdout->full_name);
+	} else {
+		DBG(" no linux,stdout-path !\n");
+	}
+
+	/* First fill our array with SOC ports */
+	for (np = NULL; (np = of_find_compatible_node(np, "serial", "ns16550")) != NULL;) {
+		struct device_node *soc = of_get_parent(np);
+		if (soc && !strcmp(soc->type, "soc")) {
+			index = add_legacy_soc_port(np, np);
+			if (index >= 0 && np == stdout)
+				legacy_serial_console = index;
+		}
+		of_node_put(soc);
+	}
+
+#ifdef CONFIG_ISA
+	/* First fill our array with ISA ports */
+	for (np = NULL; (np = of_find_node_by_type(np, "serial"));) {
+		struct device_node *isa = of_get_parent(np);
+		if (isa && !strcmp(isa->name, "isa")) {
+			index = add_legacy_isa_port(np, isa);
+			if (index >= 0 && np == stdout)
+				legacy_serial_console = index;
+		}
+		of_node_put(isa);
+	}
+#endif
+
+#ifdef CONFIG_PCI
+	/* Next, try to locate PCI ports */
+	for (np = NULL; (np = of_find_all_nodes(np));) {
+		struct device_node *pci, *parent = of_get_parent(np);
+		if (parent && !strcmp(parent->name, "isa")) {
+			of_node_put(parent);
+			continue;
+		}
+		if (strcmp(np->name, "serial") && strcmp(np->type, "serial")) {
+			of_node_put(parent);
+			continue;
+		}
+		/* Check for known pciclass, and also check wether we have
+		 * a device with child nodes for ports or not
+		 */
+		if (device_is_compatible(np, "pciclass,0700") ||
+		    device_is_compatible(np, "pciclass,070002"))
+			pci = np;
+		else if (device_is_compatible(parent, "pciclass,0700") ||
+			 device_is_compatible(parent, "pciclass,070002"))
+			pci = parent;
+		else {
+			of_node_put(parent);
+			continue;
+		}
+		index = add_legacy_pci_port(np, pci);
+		if (index >= 0 && np == stdout)
+			legacy_serial_console = index;
+		of_node_put(parent);
+	}
+#endif
+
+	DBG("legacy_serial_console = %d\n", legacy_serial_console);
+
+	/* udbg is 64 bits only for now, that will change soon though ... */
+	while (legacy_serial_console >= 0) {
+		struct legacy_serial_info *info =
+			&legacy_serial_infos[legacy_serial_console];
+		void __iomem *addr;
+
+		if (info->taddr == 0)
+			break;
+		addr = ioremap(info->taddr, 0x1000);
+		if (addr == NULL)
+			break;
+		if (info->speed == 0)
+			info->speed = udbg_probe_uart_speed(addr, info->clock);
+		DBG("default console speed = %d\n", info->speed);
+		udbg_init_uart(addr, info->speed, info->clock);
+		break;
+	}
+
+	DBG(" <- find_legacy_serial_port()\n");
+}
+
+static struct platform_device serial_device = {
+	.name	= "serial8250",
+	.id	= PLAT8250_DEV_PLATFORM,
+	.dev	= {
+		.platform_data = legacy_serial_ports,
+	},
+};
+
+static void __init fixup_port_irq(int index,
+				  struct device_node *np,
+				  struct plat_serial8250_port *port)
+{
+	DBG("fixup_port_irq(%d)\n", index);
+
+	/* Check for interrupts in that node */
+	if (np->n_intrs > 0) {
+		port->irq = np->intrs[0].line;
+		DBG(" port %d (%s), irq=%d\n",
+		    index, np->full_name, port->irq);
+		return;
+	}
+
+	/* Check for interrupts in the parent */
+	np = of_get_parent(np);
+	if (np == NULL)
+		return;
+
+	if (np->n_intrs > 0) {
+		port->irq = np->intrs[0].line;
+		DBG(" port %d (%s), irq=%d\n",
+		    index, np->full_name, port->irq);
+	}
+	of_node_put(np);
+}
+
+static void __init fixup_port_pio(int index,
+				  struct device_node *np,
+				  struct plat_serial8250_port *port)
+{
+#ifdef CONFIG_PCI
+	struct pci_controller *hose;
+
+	DBG("fixup_port_pio(%d)\n", index);
+
+	hose = pci_find_hose_for_OF_device(np);
+	if (hose) {
+		unsigned long offset = (unsigned long)hose->io_base_virt -
+#ifdef CONFIG_PPC64
+			pci_io_base;
+#else
+			isa_io_base;
+#endif
+		DBG("port %d, IO %lx -> %lx\n",
+		    index, port->iobase, port->iobase + offset);
+		port->iobase += offset;
+	}
+#endif
+}
+
+static void __init fixup_port_mmio(int index,
+				   struct device_node *np,
+				   struct plat_serial8250_port *port)
+{
+	DBG("fixup_port_mmio(%d)\n", index);
+
+	port->membase = ioremap(port->mapbase, 0x100);
+}
+
+/*
+ * This is called as an arch initcall, hopefully before the PCI bus is
+ * probed and/or the 8250 driver loaded since we need to register our
+ * platform devices before 8250 PCI ones are detected as some of them
+ * must properly "override" the platform ones.
+ *
+ * This function fixes up the interrupt value for platform ports as it
+ * couldn't be done earlier before interrupt maps have been parsed. It
+ * also "corrects" the IO address for PIO ports for the same reason,
+ * since earlier, the PHBs virtual IO space wasn't assigned yet. It then
+ * registers all those platform ports for use by the 8250 driver when it
+ * finally loads.
+ */
+static int __init serial_dev_init(void)
+{
+	int i;
+
+	if (legacy_serial_count == 0)
+		return -ENODEV;
+
+	/*
+	 * Before we register the platfrom serial devices, we need
+	 * to fixup their interrutps and their IO ports.
+	 */
+	DBG("Fixing serial ports interrupts and IO ports ...\n");
+
+	for (i = 0; i < legacy_serial_count; i++) {
+		struct plat_serial8250_port *port = &legacy_serial_ports[i];
+		struct device_node *np = legacy_serial_infos[i].np;
+
+		if (port->irq == NO_IRQ)
+			fixup_port_irq(i, np, port);
+		if (port->iotype == UPIO_PORT)
+			fixup_port_pio(i, np, port);
+		if (port->iotype == UPIO_MEM)
+			fixup_port_mmio(i, np, port);
+	}
+
+	DBG("Registering platform serial ports\n");
+
+	return platform_device_register(&serial_device);
+}
+arch_initcall(serial_dev_init);
+
+
+/*
+ * This is called very early, as part of console_init() (typically just after
+ * time_init()). This function is respondible for trying to find a good
+ * default console on serial ports. It tries to match the open firmware
+ * default output with one of the available serial console drivers, either
+ * one of the platform serial ports that have been probed earlier by
+ * find_legacy_serial_ports() or some more platform specific ones.
+ */
+static int __init check_legacy_serial_console(void)
+{
+	struct device_node *prom_stdout = NULL;
+	int speed = 0, offset = 0;
+	char *name;
+	u32 *spd;
+
+	DBG(" -> check_legacy_serial_console()\n");
+
+	/* The user has requested a console so this is already set up. */
+	if (strstr(saved_command_line, "console=")) {
+		DBG(" console was specified !\n");
+		return -EBUSY;
+	}
+
+	if (!of_chosen) {
+		DBG(" of_chosen is NULL !\n");
+		return -ENODEV;
+	}
+
+	if (legacy_serial_console < 0) {
+		DBG(" legacy_serial_console not found !\n");
+		return -ENODEV;
+	}
+	/* We are getting a weird phandle from OF ... */
+	/* ... So use the full path instead */
+	name = (char *)get_property(of_chosen, "linux,stdout-path", NULL);
+	if (name == NULL) {
+		DBG(" no linux,stdout-path !\n");
+		return -ENODEV;
+	}
+	prom_stdout = of_find_node_by_path(name);
+	if (!prom_stdout) {
+		DBG(" can't find stdout package %s !\n", name);
+		return -ENODEV;
+	}
+	DBG("stdout is %s\n", prom_stdout->full_name);
+
+	name = (char *)get_property(prom_stdout, "name", NULL);
+	if (!name) {
+		DBG(" stdout package has no name !\n");
+		goto not_found;
+	}
+	spd = (u32 *)get_property(prom_stdout, "current-speed", NULL);
+	if (spd)
+		speed = *spd;
+
+	if (0)
+		;
+#ifdef CONFIG_SERIAL_8250_CONSOLE
+	else if (strcmp(name, "serial") == 0) {
+		int i;
+		/* Look for it in probed array */
+		for (i = 0; i < legacy_serial_count; i++) {
+			if (prom_stdout != legacy_serial_infos[i].np)
+				continue;
+			offset = i;
+			speed = legacy_serial_infos[i].speed;
+			break;
+		}
+		if (i >= legacy_serial_count)
+			goto not_found;
+	}
+#endif /* CONFIG_SERIAL_8250_CONSOLE */
+#ifdef CONFIG_SERIAL_PMACZILOG_CONSOLE
+	else if (strcmp(name, "ch-a") == 0)
+		offset = 0;
+	else if (strcmp(name, "ch-b") == 0)
+		offset = 1;
+#endif /* CONFIG_SERIAL_PMACZILOG_CONSOLE */
+	else
+		goto not_found;
+	of_node_put(prom_stdout);
+
+	DBG("Found serial console at ttyS%d\n", offset);
+
+	if (speed) {
+		static char __initdata opt[16];
+		sprintf(opt, "%d", speed);
+		return add_preferred_console("ttyS", offset, opt);
+	} else
+		return add_preferred_console("ttyS", offset, NULL);
+
+ not_found:
+	DBG("No preferred console found !\n");
+	of_node_put(prom_stdout);
+	return -ENODEV;
+}
+console_initcall(check_legacy_serial_console);
+

arch/powerpc/kernel/lparmap.c

@@ -7,7 +7,7 @@
  * 2 of the License, or (at your option) any later version.
  */
 #include <asm/mmu.h>
-#include <asm/page.h>
+#include <asm/pgtable.h>
 #include <asm/iseries/lpar_map.h>
 
 const struct LparMap __attribute__((__section__(".text"))) xLparMap = {
@@ -16,16 +16,16 @@ const struct LparMap __attribute__((__section__(".text"))) xLparMap = {
 	.xSegmentTableOffs = STAB0_PAGE,
 
 	.xEsids = {
-		{ .xKernelEsid = GET_ESID(KERNELBASE),
-		  .xKernelVsid = KERNEL_VSID(KERNELBASE), },
-		{ .xKernelEsid = GET_ESID(VMALLOCBASE),
-		  .xKernelVsid = KERNEL_VSID(VMALLOCBASE), },
+		{ .xKernelEsid = GET_ESID(PAGE_OFFSET),
+		  .xKernelVsid = KERNEL_VSID(PAGE_OFFSET), },
+		{ .xKernelEsid = GET_ESID(VMALLOC_START),
+		  .xKernelVsid = KERNEL_VSID(VMALLOC_START), },
 	},
 
 	.xRanges = {
 		{ .xPages = HvPagesToMap,
 		  .xOffset = 0,
-		  .xVPN = KERNEL_VSID(KERNELBASE) << (SID_SHIFT - HW_PAGE_SHIFT),
+		  .xVPN = KERNEL_VSID(PAGE_OFFSET) << (SID_SHIFT - HW_PAGE_SHIFT),
 		},
 	},
 };

arch/powerpc/kernel/machine_kexec.c

@@ -0,0 +1,67 @@
+/*
+ * Code to handle transition of Linux booting another kernel.
+ *
+ * Copyright (C) 2002-2003 Eric Biederman  <ebiederm@xmission.com>
+ * GameCube/ppc32 port Copyright (C) 2004 Albert Herranz
+ * Copyright (C) 2005 IBM Corporation.
+ *
+ * This source code is licensed under the GNU General Public License,
+ * Version 2.  See the file COPYING for more details.
+ */
+
+#include <linux/kexec.h>
+#include <linux/reboot.h>
+#include <linux/threads.h>
+#include <asm/machdep.h>
+
+/*
+ * Provide a dummy crash_notes definition until crash dump is implemented.
+ * This prevents breakage of crash_notes attribute in kernel/ksysfs.c.
+ */
+note_buf_t crash_notes[NR_CPUS];
+
+void machine_crash_shutdown(struct pt_regs *regs)
+{
+	if (ppc_md.machine_crash_shutdown)
+		ppc_md.machine_crash_shutdown(regs);
+}
+
+/*
+ * Do what every setup is needed on image and the
+ * reboot code buffer to allow us to avoid allocations
+ * later.
+ */
+int machine_kexec_prepare(struct kimage *image)
+{
+	if (ppc_md.machine_kexec_prepare)
+		return ppc_md.machine_kexec_prepare(image);
+	/*
+	 * Fail if platform doesn't provide its own machine_kexec_prepare
+	 * implementation.
+	 */
+	return -ENOSYS;
+}
+
+void machine_kexec_cleanup(struct kimage *image)
+{
+	if (ppc_md.machine_kexec_cleanup)
+		ppc_md.machine_kexec_cleanup(image);
+}
+
+/*
+ * Do not allocate memory (or fail in any way) in machine_kexec().
+ * We are past the point of no return, committed to rebooting now.
+ */
+NORET_TYPE void machine_kexec(struct kimage *image)
+{
+	if (ppc_md.machine_kexec)
+		ppc_md.machine_kexec(image);
+	else {
+		/*
+		 * Fall back to normal restart if platform doesn't provide
+		 * its own kexec function, and user insist to kexec...
+		 */
+		machine_restart(NULL);
+	}
+	for(;;);
+}

arch/powerpc/kernel/machine_kexec_32.c

@@ -0,0 +1,65 @@
+/*
+ * PPC32 code to handle Linux booting another kernel.
+ *
+ * Copyright (C) 2002-2003 Eric Biederman  <ebiederm@xmission.com>
+ * GameCube/ppc32 port Copyright (C) 2004 Albert Herranz
+ * Copyright (C) 2005 IBM Corporation.
+ *
+ * This source code is licensed under the GNU General Public License,
+ * Version 2.  See the file COPYING for more details.
+ */
+
+#include <linux/kexec.h>
+#include <linux/mm.h>
+#include <linux/string.h>
+#include <asm/cacheflush.h>
+#include <asm/hw_irq.h>
+#include <asm/io.h>
+
+typedef NORET_TYPE void (*relocate_new_kernel_t)(
+				unsigned long indirection_page,
+				unsigned long reboot_code_buffer,
+				unsigned long start_address) ATTRIB_NORET;
+
+/*
+ * This is a generic machine_kexec function suitable at least for
+ * non-OpenFirmware embedded platforms.
+ * It merely copies the image relocation code to the control page and
+ * jumps to it.
+ * A platform specific function may just call this one.
+ */
+void default_machine_kexec(struct kimage *image)
+{
+	const extern unsigned char relocate_new_kernel[];
+	const extern unsigned int relocate_new_kernel_size;
+	unsigned long page_list;
+	unsigned long reboot_code_buffer, reboot_code_buffer_phys;
+	relocate_new_kernel_t rnk;
+
+	/* Interrupts aren't acceptable while we reboot */
+	local_irq_disable();
+
+	page_list = image->head;
+
+	/* we need both effective and real address here */
+	reboot_code_buffer =
+			(unsigned long)page_address(image->control_code_page);
+	reboot_code_buffer_phys = virt_to_phys((void *)reboot_code_buffer);
+
+	/* copy our kernel relocation code to the control code page */
+	memcpy((void *)reboot_code_buffer, relocate_new_kernel,
+						relocate_new_kernel_size);
+
+	flush_icache_range(reboot_code_buffer,
+				reboot_code_buffer + KEXEC_CONTROL_CODE_SIZE);
+	printk(KERN_INFO "Bye!\n");
+
+	/* now call it */
+	rnk = (relocate_new_kernel_t) reboot_code_buffer;
+	(*rnk)(page_list, reboot_code_buffer_phys, image->start);
+}
+
+int default_machine_kexec_prepare(struct kimage *image)
+{
+	return 0;
+}

arch/powerpc/kernel/machine_kexec_64.c

@@ -1,5 +1,5 @@
 /*
- * machine_kexec.c - handle transition of Linux booting another kernel
+ * PPC64 code to handle Linux booting another kernel.
  *
  * Copyright (C) 2004-2005, IBM Corp.
  *
@@ -28,21 +28,7 @@
 
 #define HASH_GROUP_SIZE 0x80	/* size of each hash group, asm/mmu.h */
 
-/* Have this around till we move it into crash specific file */
-note_buf_t crash_notes[NR_CPUS];
-
-/* Dummy for now. Not sure if we need to have a crash shutdown in here
- * and if what it will achieve. Letting it be now to compile the code
- * in generic kexec environment
- */
-void machine_crash_shutdown(struct pt_regs *regs)
-{
-	/* do nothing right now */
-	/* smp_relase_cpus() if we want smp on panic kernel */
-	/* cpu_irq_down to isolate us until we are ready */
-}
-
-int machine_kexec_prepare(struct kimage *image)
+int default_machine_kexec_prepare(struct kimage *image)
 {
 	int i;
 	unsigned long begin, end;	/* limits of segment */
@@ -111,11 +97,6 @@ int machine_kexec_prepare(struct kimage *image)
 	return 0;
 }
 
-void machine_kexec_cleanup(struct kimage *image)
-{
-	/* we do nothing in prepare that needs to be undone */
-}
-
 #define IND_FLAGS (IND_DESTINATION | IND_INDIRECTION | IND_DONE | IND_SOURCE)
 
 static void copy_segments(unsigned long ind)
@@ -172,9 +153,8 @@ void kexec_copy_flush(struct kimage *image)
 	 * including ones that were in place on the original copy
 	 */
 	for (i = 0; i < nr_segments; i++)
-		flush_icache_range(ranges[i].mem + KERNELBASE,
-				ranges[i].mem + KERNELBASE +
-				ranges[i].memsz);
+		flush_icache_range((unsigned long)__va(ranges[i].mem),
+			(unsigned long)__va(ranges[i].mem + ranges[i].memsz));
 }
 
 #ifdef CONFIG_SMP
@@ -283,13 +263,20 @@ extern NORET_TYPE void kexec_sequence(void *newstack, unsigned long start,
 					void (*clear_all)(void)) ATTRIB_NORET;
 
 /* too late to fail here */
-void machine_kexec(struct kimage *image)
+void default_machine_kexec(struct kimage *image)
 {
-
 	/* prepare control code if any */
 
-	/* shutdown other cpus into our wait loop and quiesce interrupts */
-	kexec_prepare_cpus();
+	/*
+        * If the kexec boot is the normal one, need to shutdown other cpus
+        * into our wait loop and quiesce interrupts.
+        * Otherwise, in the case of crashed mode (crashing_cpu >= 0),
+        * stopping other CPUs and collecting their pt_regs is done before
+        * using debugger IPI.
+        */
+
+       if (crashing_cpu == -1)
+               kexec_prepare_cpus();
 
 	/* switch to a staticly allocated stack.  Based on irq stack code.
 	 * XXX: the task struct will likely be invalid once we do the copy!

arch/powerpc/kernel/misc_32.S

@@ -5,6 +5,10 @@
  * Largely rewritten by Cort Dougan (cort@cs.nmt.edu)
  * and Paul Mackerras.
  *
+ * kexec bits:
+ * Copyright (C) 2002-2003 Eric Biederman  <ebiederm@xmission.com>
+ * GameCube/ppc32 port Copyright (C) 2004 Albert Herranz
+ *
  * This program is free software; you can redistribute it and/or
  * modify it under the terms of the GNU General Public License
  * as published by the Free Software Foundation; either version
@@ -24,6 +28,8 @@
 #include <asm/ppc_asm.h>
 #include <asm/thread_info.h>
 #include <asm/asm-offsets.h>
+#include <asm/processor.h>
+#include <asm/kexec.h>
 
 	.text
 
@@ -1006,3 +1012,110 @@ _GLOBAL(execve)
  */
 _GLOBAL(__main)
 	blr
+
+#ifdef CONFIG_KEXEC
+	/*
+	 * Must be relocatable PIC code callable as a C function.
+	 */
+	.globl relocate_new_kernel
+relocate_new_kernel:
+	/* r3 = page_list   */
+	/* r4 = reboot_code_buffer */
+	/* r5 = start_address      */
+
+	li	r0, 0
+
+	/*
+	 * Set Machine Status Register to a known status,
+	 * switch the MMU off and jump to 1: in a single step.
+	 */
+
+	mr	r8, r0
+	ori     r8, r8, MSR_RI|MSR_ME
+	mtspr	SPRN_SRR1, r8
+	addi	r8, r4, 1f - relocate_new_kernel
+	mtspr	SPRN_SRR0, r8
+	sync
+	rfi
+
+1:
+	/* from this point address translation is turned off */
+	/* and interrupts are disabled */
+
+	/* set a new stack at the bottom of our page... */
+	/* (not really needed now) */
+	addi	r1, r4, KEXEC_CONTROL_CODE_SIZE - 8 /* for LR Save+Back Chain */
+	stw	r0, 0(r1)
+
+	/* Do the copies */
+	li	r6, 0 /* checksum */
+	mr	r0, r3
+	b	1f
+
+0:	/* top, read another word for the indirection page */
+	lwzu	r0, 4(r3)
+
+1:
+	/* is it a destination page? (r8) */
+	rlwinm.	r7, r0, 0, 31, 31 /* IND_DESTINATION (1<<0) */
+	beq	2f
+
+	rlwinm	r8, r0, 0, 0, 19 /* clear kexec flags, page align */
+	b	0b
+
+2:	/* is it an indirection page? (r3) */
+	rlwinm.	r7, r0, 0, 30, 30 /* IND_INDIRECTION (1<<1) */
+	beq	2f
+
+	rlwinm	r3, r0, 0, 0, 19 /* clear kexec flags, page align */
+	subi	r3, r3, 4
+	b	0b
+
+2:	/* are we done? */
+	rlwinm.	r7, r0, 0, 29, 29 /* IND_DONE (1<<2) */
+	beq	2f
+	b	3f
+
+2:	/* is it a source page? (r9) */
+	rlwinm.	r7, r0, 0, 28, 28 /* IND_SOURCE (1<<3) */
+	beq	0b
+
+	rlwinm	r9, r0, 0, 0, 19 /* clear kexec flags, page align */
+
+	li	r7, PAGE_SIZE / 4
+	mtctr   r7
+	subi    r9, r9, 4
+	subi    r8, r8, 4
+9:
+	lwzu    r0, 4(r9)  /* do the copy */
+	xor	r6, r6, r0
+	stwu    r0, 4(r8)
+	dcbst	0, r8
+	sync
+	icbi	0, r8
+	bdnz    9b
+
+	addi    r9, r9, 4
+	addi    r8, r8, 4
+	b	0b
+
+3:
+
+	/* To be certain of avoiding problems with self-modifying code
+	 * execute a serializing instruction here.
+	 */
+	isync
+	sync
+
+	/* jump to the entry point, usually the setup routine */
+	mtlr	r5
+	blrl
+
+1:	b	1b
+
+relocate_new_kernel_end:
+
+	.globl relocate_new_kernel_size
+relocate_new_kernel_size:
+	.long relocate_new_kernel_end - relocate_new_kernel
+#endif

arch/powerpc/kernel/nvram_64.c

@@ -80,80 +80,74 @@ static loff_t dev_nvram_llseek(struct file *file, loff_t offset, int origin)
 static ssize_t dev_nvram_read(struct file *file, char __user *buf,
 			  size_t count, loff_t *ppos)
 {
-	ssize_t len;
-	char *tmp_buffer;
-	int size;
+	ssize_t ret;
+	char *tmp = NULL;
+	ssize_t size;
 
-	if (ppc_md.nvram_size == NULL)
-		return -ENODEV;
+	ret = -ENODEV;
+	if (!ppc_md.nvram_size)
+		goto out;
+
+	ret = 0;
 	size = ppc_md.nvram_size();
+	if (*ppos >= size || size < 0)
+		goto out;
 
-	if (!access_ok(VERIFY_WRITE, buf, count))
-		return -EFAULT;
-	if (*ppos >= size)
-		return 0;
-	if (count > size) 
-		count = size;
+	count = min_t(size_t, count, size - *ppos);
+	count = min(count, PAGE_SIZE);
 
-	tmp_buffer = (char *) kmalloc(count, GFP_KERNEL);
-	if (!tmp_buffer) {
-		printk(KERN_ERR "dev_read_nvram: kmalloc failed\n");
-		return -ENOMEM;
-	}
+	ret = -ENOMEM;
+	tmp = kmalloc(count, GFP_KERNEL);
+	if (!tmp)
+		goto out;
 
-	len = ppc_md.nvram_read(tmp_buffer, count, ppos);
-	if ((long)len <= 0) {
-		kfree(tmp_buffer);
-		return len;
-	}
+	ret = ppc_md.nvram_read(tmp, count, ppos);
+	if (ret <= 0)
+		goto out;
 
-	if (copy_to_user(buf, tmp_buffer, len)) {
-		kfree(tmp_buffer);
-		return -EFAULT;
-	}
+	if (copy_to_user(buf, tmp, ret))
+		ret = -EFAULT;
 
-	kfree(tmp_buffer);
-	return len;
+out:
+	kfree(tmp);
+	return ret;
 
 }
 
 static ssize_t dev_nvram_write(struct file *file, const char __user *buf,
-			   size_t count, loff_t *ppos)
+			  size_t count, loff_t *ppos)
 {
-	ssize_t len;
-	char * tmp_buffer;
-	int size;
+	ssize_t ret;
+	char *tmp = NULL;
+	ssize_t size;
 
-	if (ppc_md.nvram_size == NULL)
-		return -ENODEV;
+	ret = -ENODEV;
+	if (!ppc_md.nvram_size)
+		goto out;
+
+	ret = 0;
 	size = ppc_md.nvram_size();
+	if (*ppos >= size || size < 0)
+		goto out;
 
-	if (!access_ok(VERIFY_READ, buf, count))
-		return -EFAULT;
-	if (*ppos >= size)
-		return 0;
-	if (count > size)
-		count = size;
+	count = min_t(size_t, count, size - *ppos);
+	count = min(count, PAGE_SIZE);
 
-	tmp_buffer = (char *) kmalloc(count, GFP_KERNEL);
-	if (!tmp_buffer) {
-		printk(KERN_ERR "dev_nvram_write: kmalloc failed\n");
-		return -ENOMEM;
-	}
-	
-	if (copy_from_user(tmp_buffer, buf, count)) {
-		kfree(tmp_buffer);
-		return -EFAULT;
-	}
+	ret = -ENOMEM;
+	tmp = kmalloc(count, GFP_KERNEL);
+	if (!tmp)
+		goto out;
 
-	len = ppc_md.nvram_write(tmp_buffer, count, ppos);
-	if ((long)len <= 0) {
-		kfree(tmp_buffer);
-		return len;
-	}
+	ret = -EFAULT;
+	if (copy_from_user(tmp, buf, count))
+		goto out;
+
+	ret = ppc_md.nvram_write(tmp, count, ppos);
+
+out:
+	kfree(tmp);
+	return ret;
 
-	kfree(tmp_buffer);
-	return len;
 }
 
 static int dev_nvram_ioctl(struct inode *inode, struct file *file,

arch/powerpc/kernel/paca.c

@@ -17,6 +17,7 @@
 #include <asm/page.h>
 #include <asm/lppaca.h>
 #include <asm/iseries/it_lp_queue.h>
+#include <asm/iseries/it_lp_reg_save.h>
 #include <asm/paca.h>
 
 
@@ -26,8 +27,7 @@ extern unsigned long __toc_start;
 
 /* The Paca is an array with one entry per processor.  Each contains an
  * lppaca, which contains the information shared between the
- * hypervisor and Linux.  Each also contains an ItLpRegSave area which
- * is used by the hypervisor to save registers.
+ * hypervisor and Linux.
  * On systems with hardware multi-threading, there are two threads
  * per processor.  The Paca array must contain an entry for each thread.
  * The VPD Areas will give a max logical processors = 2 * max physical
@@ -37,7 +37,6 @@ extern unsigned long __toc_start;
 #define PACA_INIT_COMMON(number, start, asrr, asrv)			    \
 	.lock_token = 0x8000,						    \
 	.paca_index = (number),		/* Paca Index */		    \
-	.default_decr = 0x00ff0000,	/* Initial Decr */		    \
 	.kernel_toc = (unsigned long)(&__toc_start) + 0x8000UL,		    \
 	.stab_real = (asrr), 		/* Real pointer to segment table */ \
 	.stab_addr = (asrv),		/* Virt pointer to segment table */ \
@@ -57,11 +56,7 @@ extern unsigned long __toc_start;
 #ifdef CONFIG_PPC_ISERIES
 #define PACA_INIT_ISERIES(number)					    \
 	.lppaca_ptr = &paca[number].lppaca,				    \
-	.reg_save_ptr = &paca[number].reg_save,				    \
-	.reg_save = {							    \
-		.xDesc = 0xd397d9e2,	/* "LpRS" */			    \
-		.xSize = sizeof(struct ItLpRegSave)			    \
-	}
+	.reg_save_ptr = &iseries_reg_save[number],
 
 #define PACA_INIT(number)						    \
 {									    \

arch/powerpc/kernel/pci_64.c

@@ -34,7 +34,7 @@
 
 #ifdef DEBUG
 #include <asm/udbg.h>
-#define DBG(fmt...) udbg_printf(fmt)
+#define DBG(fmt...) printk(fmt)
 #else
 #define DBG(fmt...)
 #endif
@@ -251,7 +251,7 @@ void pcibios_free_controller(struct pci_controller *phb)
 		kfree(phb);
 }
 
-static void __init pcibios_claim_one_bus(struct pci_bus *b)
+void __devinit pcibios_claim_one_bus(struct pci_bus *b)
 {
 	struct pci_dev *dev;
 	struct pci_bus *child_bus;
@@ -323,6 +323,7 @@ static void pci_parse_of_addrs(struct device_node *node, struct pci_dev *dev)
 	addrs = (u32 *) get_property(node, "assigned-addresses", &proplen);
 	if (!addrs)
 		return;
+	DBG("    parse addresses (%d bytes) @ %p\n", proplen, addrs);
 	for (; proplen >= 20; proplen -= 20, addrs += 5) {
 		flags = pci_parse_of_flags(addrs[0]);
 		if (!flags)
@@ -332,6 +333,9 @@ static void pci_parse_of_addrs(struct device_node *node, struct pci_dev *dev)
 		if (!size)
 			continue;
 		i = addrs[0] & 0xff;
+		DBG("  base: %llx, size: %llx, i: %x\n",
+		    (unsigned long long)base, (unsigned long long)size, i);
+
 		if (PCI_BASE_ADDRESS_0 <= i && i <= PCI_BASE_ADDRESS_5) {
 			res = &dev->resource[(i - PCI_BASE_ADDRESS_0) >> 2];
 		} else if (i == dev->rom_base_reg) {
@@ -362,6 +366,8 @@ struct pci_dev *of_create_pci_dev(struct device_node *node,
 	if (type == NULL)
 		type = "";
 
+	DBG("    create device, devfn: %x, type: %s\n", devfn, type);
+
 	memset(dev, 0, sizeof(struct pci_dev));
 	dev->bus = bus;
 	dev->sysdata = node;
@@ -381,6 +387,8 @@ struct pci_dev *of_create_pci_dev(struct device_node *node,
 		dev->bus->number, PCI_SLOT(devfn), PCI_FUNC(devfn));
 	dev->class = get_int_prop(node, "class-code", 0);
 
+	DBG("    class: 0x%x\n", dev->class);
+
 	dev->current_state = 4;		/* unknown power state */
 
 	if (!strcmp(type, "pci")) {
@@ -402,6 +410,8 @@ struct pci_dev *of_create_pci_dev(struct device_node *node,
 
 	pci_parse_of_addrs(node, dev);
 
+	DBG("    adding to system ...\n");
+
 	pci_device_add(dev, bus);
 
 	/* XXX pci_scan_msi_device(dev); */
@@ -418,15 +428,21 @@ void __devinit of_scan_bus(struct device_node *node,
 	int reglen, devfn;
 	struct pci_dev *dev;
 
+	DBG("of_scan_bus(%s) bus no %d... \n", node->full_name, bus->number);
+
 	while ((child = of_get_next_child(node, child)) != NULL) {
+		DBG("  * %s\n", child->full_name);
 		reg = (u32 *) get_property(child, "reg", &reglen);
 		if (reg == NULL || reglen < 20)
 			continue;
 		devfn = (reg[0] >> 8) & 0xff;
+
 		/* create a new pci_dev for this device */
 		dev = of_create_pci_dev(child, bus, devfn);
 		if (!dev)
 			continue;
+		DBG("dev header type: %x\n", dev->hdr_type);
+
 		if (dev->hdr_type == PCI_HEADER_TYPE_BRIDGE ||
 		    dev->hdr_type == PCI_HEADER_TYPE_CARDBUS)
 			of_scan_pci_bridge(child, dev);
@@ -446,16 +462,18 @@ void __devinit of_scan_pci_bridge(struct device_node *node,
 	unsigned int flags;
 	u64 size;
 
+	DBG("of_scan_pci_bridge(%s)\n", node->full_name);
+
 	/* parse bus-range property */
 	busrange = (u32 *) get_property(node, "bus-range", &len);
 	if (busrange == NULL || len != 8) {
-		printk(KERN_ERR "Can't get bus-range for PCI-PCI bridge %s\n",
+		printk(KERN_DEBUG "Can't get bus-range for PCI-PCI bridge %s\n",
 		       node->full_name);
 		return;
 	}
 	ranges = (u32 *) get_property(node, "ranges", &len);
 	if (ranges == NULL) {
-		printk(KERN_ERR "Can't get ranges for PCI-PCI bridge %s\n",
+		printk(KERN_DEBUG "Can't get ranges for PCI-PCI bridge %s\n",
 		       node->full_name);
 		return;
 	}
@@ -509,10 +527,13 @@ void __devinit of_scan_pci_bridge(struct device_node *node,
 	}
 	sprintf(bus->name, "PCI Bus %04x:%02x", pci_domain_nr(bus),
 		bus->number);
+	DBG("    bus name: %s\n", bus->name);
 
 	mode = PCI_PROBE_NORMAL;
 	if (ppc_md.pci_probe_mode)
 		mode = ppc_md.pci_probe_mode(bus);
+	DBG("    probe mode: %d\n", mode);
+
 	if (mode == PCI_PROBE_DEVTREE)
 		of_scan_bus(node, bus);
 	else if (mode == PCI_PROBE_NORMAL)
@@ -528,6 +549,8 @@ void __devinit scan_phb(struct pci_controller *hose)
 	int i, mode;
 	struct resource *res;
 
+	DBG("Scanning PHB %s\n", node ? node->full_name : "<NO NAME>");
+
 	bus = pci_create_bus(NULL, hose->first_busno, hose->ops, node);
 	if (bus == NULL) {
 		printk(KERN_ERR "Failed to create bus for PCI domain %04x\n",
@@ -552,8 +575,9 @@ void __devinit scan_phb(struct pci_controller *hose)
 
 	mode = PCI_PROBE_NORMAL;
 #ifdef CONFIG_PPC_MULTIPLATFORM
-	if (ppc_md.pci_probe_mode)
+	if (node && ppc_md.pci_probe_mode)
 		mode = ppc_md.pci_probe_mode(bus);
+	DBG("    probe mode: %d\n", mode);
 	if (mode == PCI_PROBE_DEVTREE) {
 		bus->subordinate = hose->last_busno;
 		of_scan_bus(node, bus);
@@ -842,8 +866,7 @@ pgprot_t pci_phys_mem_access_prot(struct file *file,
  * Returns a negative error code on failure, zero on success.
  */
 int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
-			enum pci_mmap_state mmap_state,
-			int write_combine)
+			enum pci_mmap_state mmap_state, int write_combine)
 {
 	unsigned long offset = vma->vm_pgoff << PAGE_SHIFT;
 	struct resource *rp;
@@ -896,6 +919,25 @@ static void __devinit pci_process_ISA_OF_ranges(struct device_node *isa_node,
 				      unsigned long phb_io_base_phys,
 				      void __iomem * phb_io_base_virt)
 {
+	/* Remove these asap */
+
+	struct pci_address {
+		u32 a_hi;
+		u32 a_mid;
+		u32 a_lo;
+	};
+
+	struct isa_address {
+		u32 a_hi;
+		u32 a_lo;
+	};
+
+	struct isa_range {
+		struct isa_address isa_addr;
+		struct pci_address pci_addr;
+		unsigned int size;
+	};
+
 	struct isa_range *range;
 	unsigned long pci_addr;
 	unsigned int isa_addr;
@@ -1223,6 +1265,7 @@ void __devinit pcibios_fixup_device_resources(struct pci_dev *dev,
 }
 EXPORT_SYMBOL(pcibios_fixup_device_resources);
 
+
 static void __devinit do_bus_setup(struct pci_bus *bus)
 {
 	struct pci_dev *dev;
@@ -1306,8 +1349,38 @@ void pci_resource_to_user(const struct pci_dev *dev, int bar,
 	*end = rsrc->end + offset;
 }
 
+struct pci_controller* pci_find_hose_for_OF_device(struct device_node* node)
+{
+	if (!have_of)
+		return NULL;
+	while(node) {
+		struct pci_controller *hose, *tmp;
+		list_for_each_entry_safe(hose, tmp, &hose_list, list_node)
+			if (hose->arch_data == node)
+				return hose;
+		node = node->parent;
+	}
+	return NULL;
+}
+
 #endif /* CONFIG_PPC_MULTIPLATFORM */
 
+unsigned long pci_address_to_pio(phys_addr_t address)
+{
+	struct pci_controller *hose, *tmp;
+
+	list_for_each_entry_safe(hose, tmp, &hose_list, list_node) {
+		if (address >= hose->io_base_phys &&
+		    address < (hose->io_base_phys + hose->pci_io_size)) {
+			unsigned long base =
+				(unsigned long)hose->io_base_virt - pci_io_base;
+			return base + (address - hose->io_base_phys);
+		}
+	}
+	return (unsigned int)-1;
+}
+EXPORT_SYMBOL_GPL(pci_address_to_pio);
+
 
 #define IOBASE_BRIDGE_NUMBER	0
 #define IOBASE_MEMORY		1

arch/powerpc/kernel/pmc.c

@@ -43,8 +43,13 @@ static void dummy_perf(struct pt_regs *regs)
 	mtspr(SPRN_MMCR0, mmcr0);
 }
 #else
+/* Ensure exceptions are disabled */
 static void dummy_perf(struct pt_regs *regs)
 {
+	unsigned int mmcr0 = mfspr(SPRN_MMCR0);
+
+	mmcr0 &= ~(MMCR0_PMXE);
+	mtspr(SPRN_MMCR0, mmcr0);
 }
 #endif
 

arch/powerpc/kernel/ppc_ksyms.c

@@ -76,11 +76,6 @@ EXPORT_SYMBOL(single_step_exception);
 EXPORT_SYMBOL(sys_sigreturn);
 #endif
 
-#if defined(CONFIG_PPC_PREP)
-EXPORT_SYMBOL(_prep_type);
-EXPORT_SYMBOL(ucSystemType);
-#endif
-
 EXPORT_SYMBOL(strcpy);
 EXPORT_SYMBOL(strncpy);
 EXPORT_SYMBOL(strcat);

arch/powerpc/kernel/prom.c

@@ -29,6 +29,7 @@
 #include <linux/initrd.h>
 #include <linux/bitops.h>
 #include <linux/module.h>
+#include <linux/kexec.h>
 
 #include <asm/prom.h>
 #include <asm/rtas.h>
@@ -37,6 +38,7 @@
 #include <asm/processor.h>
 #include <asm/irq.h>
 #include <asm/io.h>
+#include <asm/kdump.h>
 #include <asm/smp.h>
 #include <asm/system.h>
 #include <asm/mmu.h>
@@ -55,21 +57,6 @@
 #define DBG(fmt...)
 #endif
 
-struct pci_reg_property {
-	struct pci_address addr;
-	u32 size_hi;
-	u32 size_lo;
-};
-
-struct isa_reg_property {
-	u32 space;
-	u32 address;
-	u32 size;
-};
-
-
-typedef int interpret_func(struct device_node *, unsigned long *,
-			   int, int, int);
 
 static int __initdata dt_root_addr_cells;
 static int __initdata dt_root_size_cells;
@@ -311,6 +298,16 @@ static int __devinit finish_node_interrupts(struct device_node *np,
 	int i, j, n, sense;
 	unsigned int *irq, virq;
 	struct device_node *ic;
+	int trace = 0;
+
+	//#define TRACE(fmt...) do { if (trace) { printk(fmt); mdelay(1000); } } while(0)
+#define TRACE(fmt...)
+
+	if (!strcmp(np->name, "smu-doorbell"))
+		trace = 1;
+
+	TRACE("Finishing SMU doorbell ! num_interrupt_controllers = %d\n",
+	      num_interrupt_controllers);
 
 	if (num_interrupt_controllers == 0) {
 		/*
@@ -345,11 +342,12 @@ static int __devinit finish_node_interrupts(struct device_node *np,
 	}
 
 	ints = (unsigned int *) get_property(np, "interrupts", &intlen);
+	TRACE("ints=%p, intlen=%d\n", ints, intlen);
 	if (ints == NULL)
 		return 0;
 	intrcells = prom_n_intr_cells(np);
 	intlen /= intrcells * sizeof(unsigned int);
-
+	TRACE("intrcells=%d, new intlen=%d\n", intrcells, intlen);
 	np->intrs = prom_alloc(intlen * sizeof(*(np->intrs)), mem_start);
 	if (!np->intrs)
 		return -ENOMEM;
@@ -360,6 +358,7 @@ static int __devinit finish_node_interrupts(struct device_node *np,
 	intrcount = 0;
 	for (i = 0; i < intlen; ++i, ints += intrcells) {
 		n = map_interrupt(&irq, &ic, np, ints, intrcells);
+		TRACE("map, irq=%d, ic=%p, n=%d\n", irq, ic, n);
 		if (n <= 0)
 			continue;
 
@@ -370,6 +369,7 @@ static int __devinit finish_node_interrupts(struct device_node *np,
 			np->intrs[intrcount].sense = map_isa_senses[sense];
 		} else {
 			virq = virt_irq_create_mapping(irq[0]);
+			TRACE("virq=%d\n", virq);
 #ifdef CONFIG_PPC64
 			if (virq == NO_IRQ) {
 				printk(KERN_CRIT "Could not allocate interrupt"
@@ -379,6 +379,12 @@ static int __devinit finish_node_interrupts(struct device_node *np,
 #endif
 			np->intrs[intrcount].line = irq_offset_up(virq);
 			sense = (n > 1)? (irq[1] & 3): 1;
+
+			/* Apple uses bits in there in a different way, let's
+			 * only keep the real sense bit on macs
+			 */
+			if (_machine == PLATFORM_POWERMAC)
+				sense &= 0x1;
 			np->intrs[intrcount].sense = map_mpic_senses[sense];
 		}
 
@@ -388,12 +394,13 @@ static int __devinit finish_node_interrupts(struct device_node *np,
 			char *name = get_property(ic->parent, "name", NULL);
 			if (name && !strcmp(name, "u3"))
 				np->intrs[intrcount].line += 128;
-			else if (!(name && !strcmp(name, "mac-io")))
+			else if (!(name && (!strcmp(name, "mac-io") ||
+					    !strcmp(name, "u4"))))
 				/* ignore other cascaded controllers, such as
 				   the k2-sata-root */
 				break;
 		}
-#endif
+#endif /* CONFIG_PPC64 */
 		if (n > 2) {
 			printk("hmmm, got %d intr cells for %s:", n,
 			       np->full_name);
@@ -408,234 +415,19 @@ static int __devinit finish_node_interrupts(struct device_node *np,
 	return 0;
 }
 
-static int __devinit interpret_pci_props(struct device_node *np,
-					 unsigned long *mem_start,
-					 int naddrc, int nsizec,
-					 int measure_only)
-{
-	struct address_range *adr;
-	struct pci_reg_property *pci_addrs;
-	int i, l, n_addrs;
-
-	pci_addrs = (struct pci_reg_property *)
-		get_property(np, "assigned-addresses", &l);
-	if (!pci_addrs)
-		return 0;
-
-	n_addrs = l / sizeof(*pci_addrs);
-
-	adr = prom_alloc(n_addrs * sizeof(*adr), mem_start);
-	if (!adr)
-		return -ENOMEM;
-
- 	if (measure_only)
- 		return 0;
-
- 	np->addrs = adr;
- 	np->n_addrs = n_addrs;
-
- 	for (i = 0; i < n_addrs; i++) {
- 		adr[i].space = pci_addrs[i].addr.a_hi;
- 		adr[i].address = pci_addrs[i].addr.a_lo |
-			((u64)pci_addrs[i].addr.a_mid << 32);
- 		adr[i].size = pci_addrs[i].size_lo;
-	}
-
-	return 0;
-}
-
-static int __init interpret_dbdma_props(struct device_node *np,
-					unsigned long *mem_start,
-					int naddrc, int nsizec,
-					int measure_only)
-{
-	struct reg_property32 *rp;
-	struct address_range *adr;
-	unsigned long base_address;
-	int i, l;
-	struct device_node *db;
-
-	base_address = 0;
-	if (!measure_only) {
-		for (db = np->parent; db != NULL; db = db->parent) {
-			if (!strcmp(db->type, "dbdma") && db->n_addrs != 0) {
-				base_address = db->addrs[0].address;
-				break;
-			}
-		}
-	}
-
-	rp = (struct reg_property32 *) get_property(np, "reg", &l);
-	if (rp != 0 && l >= sizeof(struct reg_property32)) {
-		i = 0;
-		adr = (struct address_range *) (*mem_start);
-		while ((l -= sizeof(struct reg_property32)) >= 0) {
-			if (!measure_only) {
-				adr[i].space = 2;
-				adr[i].address = rp[i].address + base_address;
-				adr[i].size = rp[i].size;
-			}
-			++i;
-		}
-		np->addrs = adr;
-		np->n_addrs = i;
-		(*mem_start) += i * sizeof(struct address_range);
-	}
-
-	return 0;
-}
-
-static int __init interpret_macio_props(struct device_node *np,
-					unsigned long *mem_start,
-					int naddrc, int nsizec,
-					int measure_only)
-{
-	struct reg_property32 *rp;
-	struct address_range *adr;
-	unsigned long base_address;
-	int i, l;
-	struct device_node *db;
-
-	base_address = 0;
-	if (!measure_only) {
-		for (db = np->parent; db != NULL; db = db->parent) {
-			if (!strcmp(db->type, "mac-io") && db->n_addrs != 0) {
-				base_address = db->addrs[0].address;
-				break;
-			}
-		}
-	}
-
-	rp = (struct reg_property32 *) get_property(np, "reg", &l);
-	if (rp != 0 && l >= sizeof(struct reg_property32)) {
-		i = 0;
-		adr = (struct address_range *) (*mem_start);
-		while ((l -= sizeof(struct reg_property32)) >= 0) {
-			if (!measure_only) {
-				adr[i].space = 2;
-				adr[i].address = rp[i].address + base_address;
-				adr[i].size = rp[i].size;
-			}
-			++i;
-		}
-		np->addrs = adr;
-		np->n_addrs = i;
-		(*mem_start) += i * sizeof(struct address_range);
-	}
-
-	return 0;
-}
-
-static int __init interpret_isa_props(struct device_node *np,
-				      unsigned long *mem_start,
-				      int naddrc, int nsizec,
-				      int measure_only)
-{
-	struct isa_reg_property *rp;
-	struct address_range *adr;
-	int i, l;
-
-	rp = (struct isa_reg_property *) get_property(np, "reg", &l);
-	if (rp != 0 && l >= sizeof(struct isa_reg_property)) {
-		i = 0;
-		adr = (struct address_range *) (*mem_start);
-		while ((l -= sizeof(struct isa_reg_property)) >= 0) {
-			if (!measure_only) {
-				adr[i].space = rp[i].space;
-				adr[i].address = rp[i].address;
-				adr[i].size = rp[i].size;
-			}
-			++i;
-		}
-		np->addrs = adr;
-		np->n_addrs = i;
-		(*mem_start) += i * sizeof(struct address_range);
-	}
-
-	return 0;
-}
-
-static int __init interpret_root_props(struct device_node *np,
-				       unsigned long *mem_start,
-				       int naddrc, int nsizec,
-				       int measure_only)
-{
-	struct address_range *adr;
-	int i, l;
-	unsigned int *rp;
-	int rpsize = (naddrc + nsizec) * sizeof(unsigned int);
-
-	rp = (unsigned int *) get_property(np, "reg", &l);
-	if (rp != 0 && l >= rpsize) {
-		i = 0;
-		adr = (struct address_range *) (*mem_start);
-		while ((l -= rpsize) >= 0) {
-			if (!measure_only) {
-				adr[i].space = 0;
-				adr[i].address = rp[naddrc - 1];
-				adr[i].size = rp[naddrc + nsizec - 1];
-			}
-			++i;
-			rp += naddrc + nsizec;
-		}
-		np->addrs = adr;
-		np->n_addrs = i;
-		(*mem_start) += i * sizeof(struct address_range);
-	}
-
-	return 0;
-}
-
 static int __devinit finish_node(struct device_node *np,
 				 unsigned long *mem_start,
-				 interpret_func *ifunc,
-				 int naddrc, int nsizec,
 				 int measure_only)
 {
 	struct device_node *child;
-	int *ip, rc = 0;
-
-	/* get the device addresses and interrupts */
-	if (ifunc != NULL)
-		rc = ifunc(np, mem_start, naddrc, nsizec, measure_only);
-	if (rc)
-		goto out;
+	int rc = 0;
 
 	rc = finish_node_interrupts(np, mem_start, measure_only);
 	if (rc)
 		goto out;
 
-	/* Look for #address-cells and #size-cells properties. */
-	ip = (int *) get_property(np, "#address-cells", NULL);
-	if (ip != NULL)
-		naddrc = *ip;
-	ip = (int *) get_property(np, "#size-cells", NULL);
-	if (ip != NULL)
-		nsizec = *ip;
-
-	if (!strcmp(np->name, "device-tree") || np->parent == NULL)
-		ifunc = interpret_root_props;
-	else if (np->type == 0)
-		ifunc = NULL;
-	else if (!strcmp(np->type, "pci") || !strcmp(np->type, "vci"))
-		ifunc = interpret_pci_props;
-	else if (!strcmp(np->type, "dbdma"))
-		ifunc = interpret_dbdma_props;
-	else if (!strcmp(np->type, "mac-io") || ifunc == interpret_macio_props)
-		ifunc = interpret_macio_props;
-	else if (!strcmp(np->type, "isa"))
-		ifunc = interpret_isa_props;
-	else if (!strcmp(np->name, "uni-n") || !strcmp(np->name, "u3"))
-		ifunc = interpret_root_props;
-	else if (!((ifunc == interpret_dbdma_props
-		    || ifunc == interpret_macio_props)
-		   && (!strcmp(np->type, "escc")
-		       || !strcmp(np->type, "media-bay"))))
-		ifunc = NULL;
-
 	for (child = np->child; child != NULL; child = child->sibling) {
-		rc = finish_node(child, mem_start, ifunc,
-				 naddrc, nsizec, measure_only);
+		rc = finish_node(child, mem_start, measure_only);
 		if (rc)
 			goto out;
 	}
@@ -697,10 +489,10 @@ void __init finish_device_tree(void)
 	 * reason and then remove those additional 16 bytes
 	 */
 	size = 16;
-	finish_node(allnodes, &size, NULL, 0, 0, 1);
+	finish_node(allnodes, &size, 1);
 	size -= 16;
 	end = start = (unsigned long) __va(lmb_alloc(size, 128));
-	finish_node(allnodes, &end, NULL, 0, 0, 0);
+	finish_node(allnodes, &end, 0);
 	BUG_ON(end != start + size);
 
 	DBG(" <- finish_device_tree\n");
@@ -1197,6 +989,16 @@ static int __init early_init_dt_scan_chosen(unsigned long node,
 	}
 #endif /* CONFIG_PPC_RTAS */
 
+#ifdef CONFIG_KEXEC
+       lprop = (u64*)of_get_flat_dt_prop(node, "linux,crashkernel-base", NULL);
+       if (lprop)
+               crashk_res.start = *lprop;
+
+       lprop = (u64*)of_get_flat_dt_prop(node, "linux,crashkernel-size", NULL);
+       if (lprop)
+               crashk_res.end = crashk_res.start + *lprop - 1;
+#endif
+
 	/* break now */
 	return 1;
 }
@@ -1263,7 +1065,9 @@ static int __init early_init_dt_scan_memory(unsigned long node,
 	} else if (strcmp(type, "memory") != 0)
 		return 0;
 
-	reg = (cell_t *)of_get_flat_dt_prop(node, "reg", &l);
+	reg = (cell_t *)of_get_flat_dt_prop(node, "linux,usable-memory", &l);
+	if (reg == NULL)
+		reg = (cell_t *)of_get_flat_dt_prop(node, "reg", &l);
 	if (reg == NULL)
 		return 0;
 
@@ -1335,11 +1139,14 @@ void __init early_init_devtree(void *params)
 	of_scan_flat_dt(early_init_dt_scan_memory, NULL);
 	lmb_enforce_memory_limit(memory_limit);
 	lmb_analyze();
-	lmb_reserve(0, __pa(klimit));
 
 	DBG("Phys. mem: %lx\n", lmb_phys_mem_size());
 
 	/* Reserve LMB regions used by kernel, initrd, dt, etc... */
+	lmb_reserve(PHYSICAL_START, __pa(klimit) - PHYSICAL_START);
+#ifdef CONFIG_CRASH_DUMP
+	lmb_reserve(0, KDUMP_RESERVE_LIMIT);
+#endif
 	early_reserve_mem();
 
 	DBG("Scanning CPUs ...\n");
@@ -1802,7 +1609,6 @@ static void of_node_release(struct kref *kref)
 		prop = next;
 	}
 	kfree(node->intrs);
-	kfree(node->addrs);
 	kfree(node->full_name);
 	kfree(node->data);
 	kfree(node);
@@ -1884,9 +1690,7 @@ void of_detach_node(const struct device_node *np)
  * This should probably be split up into smaller chunks.
  */
 
-static int of_finish_dynamic_node(struct device_node *node,
-				  unsigned long *unused1, int unused2,
-				  int unused3, int unused4)
+static int of_finish_dynamic_node(struct device_node *node)
 {
 	struct device_node *parent = of_get_parent(node);
 	int err = 0;
@@ -1907,7 +1711,8 @@ static int of_finish_dynamic_node(struct device_node *node,
 		return -ENODEV;
 
 	/* fix up new node's linux_phandle field */
-	if ((ibm_phandle = (unsigned int *)get_property(node, "ibm,phandle", NULL)))
+	if ((ibm_phandle = (unsigned int *)get_property(node,
+							"ibm,phandle", NULL)))
 		node->linux_phandle = *ibm_phandle;
 
 out:
@@ -1922,7 +1727,9 @@ static int prom_reconfig_notifier(struct notifier_block *nb,
 
 	switch (action) {
 	case PSERIES_RECONFIG_ADD:
-		err = finish_node(node, NULL, of_finish_dynamic_node, 0, 0, 0);
+		err = of_finish_dynamic_node(node);
+		if (!err)
+			finish_node(node, NULL, 0);
 		if (err < 0) {
 			printk(KERN_ERR "finish_node returned %d\n", err);
 			err = NOTIFY_BAD;
@@ -1996,175 +1803,4 @@ int prom_add_property(struct device_node* np, struct property* prop)
 	return 0;
 }
 
-/* I quickly hacked that one, check against spec ! */
-static inline unsigned long
-bus_space_to_resource_flags(unsigned int bus_space)
-{
-	u8 space = (bus_space >> 24) & 0xf;
-	if (space == 0)
-		space = 0x02;
-	if (space == 0x02)
-		return IORESOURCE_MEM;
-	else if (space == 0x01)
-		return IORESOURCE_IO;
-	else {
-		printk(KERN_WARNING "prom.c: bus_space_to_resource_flags(), space: %x\n",
-		    	bus_space);
-		return 0;
-	}
-}
 
-#ifdef CONFIG_PCI
-static struct resource *find_parent_pci_resource(struct pci_dev* pdev,
-						 struct address_range *range)
-{
-	unsigned long mask;
-	int i;
-
-	/* Check this one */
-	mask = bus_space_to_resource_flags(range->space);
-	for (i=0; i<DEVICE_COUNT_RESOURCE; i++) {
-		if ((pdev->resource[i].flags & mask) == mask &&
-			pdev->resource[i].start <= range->address &&
-			pdev->resource[i].end > range->address) {
-				if ((range->address + range->size - 1) > pdev->resource[i].end) {
-					/* Add better message */
-					printk(KERN_WARNING "PCI/OF resource overlap !\n");
-					return NULL;
-				}
-				break;
-			}
-	}
-	if (i == DEVICE_COUNT_RESOURCE)
-		return NULL;
-	return &pdev->resource[i];
-}
-
-/*
- * Request an OF device resource. Currently handles child of PCI devices,
- * or other nodes attached to the root node. Ultimately, put some
- * link to resources in the OF node.
- */
-struct resource *request_OF_resource(struct device_node* node, int index,
-				     const char* name_postfix)
-{
-	struct pci_dev* pcidev;
-	u8 pci_bus, pci_devfn;
-	unsigned long iomask;
-	struct device_node* nd;
-	struct resource* parent;
-	struct resource *res = NULL;
-	int nlen, plen;
-
-	if (index >= node->n_addrs)
-		goto fail;
-
-	/* Sanity check on bus space */
-	iomask = bus_space_to_resource_flags(node->addrs[index].space);
-	if (iomask & IORESOURCE_MEM)
-		parent = &iomem_resource;
-	else if (iomask & IORESOURCE_IO)
-		parent = &ioport_resource;
-	else
-		goto fail;
-
-	/* Find a PCI parent if any */
-	nd = node;
-	pcidev = NULL;
-	while (nd) {
-		if (!pci_device_from_OF_node(nd, &pci_bus, &pci_devfn))
-			pcidev = pci_find_slot(pci_bus, pci_devfn);
-		if (pcidev) break;
-		nd = nd->parent;
-	}
-	if (pcidev)
-		parent = find_parent_pci_resource(pcidev, &node->addrs[index]);
-	if (!parent) {
-		printk(KERN_WARNING "request_OF_resource(%s), parent not found\n",
-			node->name);
-		goto fail;
-	}
-
-	res = __request_region(parent, node->addrs[index].address,
-			       node->addrs[index].size, NULL);
-	if (!res)
-		goto fail;
-	nlen = strlen(node->name);
-	plen = name_postfix ? strlen(name_postfix) : 0;
-	res->name = (const char *)kmalloc(nlen+plen+1, GFP_KERNEL);
-	if (res->name) {
-		strcpy((char *)res->name, node->name);
-		if (plen)
-			strcpy((char *)res->name+nlen, name_postfix);
-	}
-	return res;
-fail:
-	return NULL;
-}
-EXPORT_SYMBOL(request_OF_resource);
-
-int release_OF_resource(struct device_node *node, int index)
-{
-	struct pci_dev* pcidev;
-	u8 pci_bus, pci_devfn;
-	unsigned long iomask, start, end;
-	struct device_node* nd;
-	struct resource* parent;
-	struct resource *res = NULL;
-
-	if (index >= node->n_addrs)
-		return -EINVAL;
-
-	/* Sanity check on bus space */
-	iomask = bus_space_to_resource_flags(node->addrs[index].space);
-	if (iomask & IORESOURCE_MEM)
-		parent = &iomem_resource;
-	else if (iomask & IORESOURCE_IO)
-		parent = &ioport_resource;
-	else
-		return -EINVAL;
-
-	/* Find a PCI parent if any */
-	nd = node;
-	pcidev = NULL;
-	while(nd) {
-		if (!pci_device_from_OF_node(nd, &pci_bus, &pci_devfn))
-			pcidev = pci_find_slot(pci_bus, pci_devfn);
-		if (pcidev) break;
-		nd = nd->parent;
-	}
-	if (pcidev)
-		parent = find_parent_pci_resource(pcidev, &node->addrs[index]);
-	if (!parent) {
-		printk(KERN_WARNING "release_OF_resource(%s), parent not found\n",
-			node->name);
-		return -ENODEV;
-	}
-
-	/* Find us in the parent and its childs */
-	res = parent->child;
-	start = node->addrs[index].address;
-	end = start + node->addrs[index].size - 1;
-	while (res) {
-		if (res->start == start && res->end == end &&
-		    (res->flags & IORESOURCE_BUSY))
-		    	break;
-		if (res->start <= start && res->end >= end)
-			res = res->child;
-		else
-			res = res->sibling;
-	}
-	if (!res)
-		return -ENODEV;
-
-	if (res->name) {
-		kfree(res->name);
-		res->name = NULL;
-	}
-	release_resource(res);
-	kfree(res);
-
-	return 0;
-}
-EXPORT_SYMBOL(release_OF_resource);
-#endif /* CONFIG_PCI */

arch/powerpc/kernel/prom_init.c

@@ -192,6 +192,11 @@ static unsigned long __initdata alloc_bottom;
 static unsigned long __initdata rmo_top;
 static unsigned long __initdata ram_top;
 
+#ifdef CONFIG_KEXEC
+static unsigned long __initdata prom_crashk_base;
+static unsigned long __initdata prom_crashk_size;
+#endif
+
 static struct mem_map_entry __initdata mem_reserve_map[MEM_RESERVE_MAP_SIZE];
 static int __initdata mem_reserve_cnt;
 
@@ -553,7 +558,8 @@ unsigned long prom_memparse(const char *ptr, const char **retptr)
 static void __init early_cmdline_parse(void)
 {
 	struct prom_t *_prom = &RELOC(prom);
-	char *opt, *p;
+	const char *opt;
+	char *p;
 	int l = 0;
 
 	RELOC(prom_cmd_line[0]) = 0;
@@ -590,6 +596,34 @@ static void __init early_cmdline_parse(void)
 		RELOC(prom_memory_limit) = ALIGN(RELOC(prom_memory_limit), 0x1000000);
 #endif
 	}
+
+#ifdef CONFIG_KEXEC
+	/*
+	 * crashkernel=size@addr specifies the location to reserve for
+	 * crash kernel.
+	 */
+	opt = strstr(RELOC(prom_cmd_line), RELOC("crashkernel="));
+	if (opt) {
+		opt += 12;
+		RELOC(prom_crashk_size) = prom_memparse(opt, &opt);
+
+		if (ALIGN(RELOC(prom_crashk_size), 0x1000000) !=
+			RELOC(prom_crashk_size)) {
+			prom_printf("Warning: crashkernel size is not "
+					"aligned to 16MB\n");
+		}
+
+		/*
+		 * At present, the crash kernel always run at 32MB.
+		 * Just ignore whatever user passed.
+		 */
+		RELOC(prom_crashk_base) = 0x2000000;
+		if (*opt == '@') {
+			prom_printf("Warning: PPC64 kdump kernel always runs "
+					"at 32 MB\n");
+		}
+	}
+#endif
 }
 
 #ifdef CONFIG_PPC_PSERIES
@@ -1011,6 +1045,12 @@ static void __init prom_init_mem(void)
 	prom_printf("  alloc_top_hi : %x\n", RELOC(alloc_top_high));
 	prom_printf("  rmo_top      : %x\n", RELOC(rmo_top));
 	prom_printf("  ram_top      : %x\n", RELOC(ram_top));
+#ifdef CONFIG_KEXEC
+	if (RELOC(prom_crashk_base)) {
+		prom_printf("  crashk_base  : %x\n",  RELOC(prom_crashk_base));
+		prom_printf("  crashk_size  : %x\n", RELOC(prom_crashk_size));
+	}
+#endif
 }
 
 
@@ -1500,6 +1540,8 @@ static int __init prom_find_machine_type(void)
 #ifdef CONFIG_PPC64
 			if (strstr(p, RELOC("Momentum,Maple")))
 				return PLATFORM_MAPLE;
+			if (strstr(p, RELOC("IBM,CPB")))
+				return PLATFORM_CELL;
 #endif
 			i += sl + 1;
 		}
@@ -1994,7 +2036,7 @@ static void __init prom_check_initrd(unsigned long r3, unsigned long r4)
 	if (r3 && r4 && r4 != 0xdeadbeef) {
 		unsigned long val;
 
-		RELOC(prom_initrd_start) = (r3 >= KERNELBASE) ? __pa(r3) : r3;
+		RELOC(prom_initrd_start) = is_kernel_addr(r3) ? __pa(r3) : r3;
 		RELOC(prom_initrd_end) = RELOC(prom_initrd_start) + r4;
 
 		val = RELOC(prom_initrd_start);
@@ -2094,6 +2136,10 @@ unsigned long __init prom_init(unsigned long r3, unsigned long r4,
 	 */
 	prom_init_mem();
 
+#ifdef CONFIG_KEXEC
+	if (RELOC(prom_crashk_base))
+		reserve_mem(RELOC(prom_crashk_base), RELOC(prom_crashk_size));
+#endif
 	/*
 	 * Determine which cpu is actually running right _now_
 	 */
@@ -2150,6 +2196,16 @@ unsigned long __init prom_init(unsigned long r3, unsigned long r4,
 	}
 #endif
 
+#ifdef CONFIG_KEXEC
+	if (RELOC(prom_crashk_base)) {
+		prom_setprop(_prom->chosen, "/chosen", "linux,crashkernel-base",
+			PTRRELOC(&prom_crashk_base),
+			sizeof(RELOC(prom_crashk_base)));
+		prom_setprop(_prom->chosen, "/chosen", "linux,crashkernel-size",
+			PTRRELOC(&prom_crashk_size),
+			sizeof(RELOC(prom_crashk_size)));
+	}
+#endif
 	/*
 	 * Fixup any known bugs in the device-tree
 	 */

arch/powerpc/kernel/prom_parse.c

@@ -0,0 +1,547 @@
+#undef DEBUG
+
+#include <linux/kernel.h>
+#include <linux/string.h>
+#include <linux/pci_regs.h>
+#include <linux/module.h>
+#include <linux/ioport.h>
+#include <asm/prom.h>
+#include <asm/pci-bridge.h>
+
+#ifdef DEBUG
+#define DBG(fmt...) do { printk(fmt); } while(0)
+#else
+#define DBG(fmt...) do { } while(0)
+#endif
+
+#ifdef CONFIG_PPC64
+#define PRu64	"%lx"
+#else
+#define PRu64	"%llx"
+#endif
+
+/* Max address size we deal with */
+#define OF_MAX_ADDR_CELLS	4
+#define OF_CHECK_COUNTS(na, ns)	((na) > 0 && (na) <= OF_MAX_ADDR_CELLS && \
+			(ns) > 0)
+
+/* Debug utility */
+#ifdef DEBUG
+static void of_dump_addr(const char *s, u32 *addr, int na)
+{
+	printk("%s", s);
+	while(na--)
+		printk(" %08x", *(addr++));
+	printk("\n");
+}
+#else
+static void of_dump_addr(const char *s, u32 *addr, int na) { }
+#endif
+
+/* Read a big address */
+static inline u64 of_read_addr(u32 *cell, int size)
+{
+	u64 r = 0;
+	while (size--)
+		r = (r << 32) | *(cell++);
+	return r;
+}
+
+/* Callbacks for bus specific translators */
+struct of_bus {
+	const char	*name;
+	const char	*addresses;
+	int		(*match)(struct device_node *parent);
+	void		(*count_cells)(struct device_node *child,
+				       int *addrc, int *sizec);
+	u64		(*map)(u32 *addr, u32 *range, int na, int ns, int pna);
+	int		(*translate)(u32 *addr, u64 offset, int na);
+	unsigned int	(*get_flags)(u32 *addr);
+};
+
+
+/*
+ * Default translator (generic bus)
+ */
+
+static void of_bus_default_count_cells(struct device_node *dev,
+				       int *addrc, int *sizec)
+{
+	if (addrc)
+		*addrc = prom_n_addr_cells(dev);
+	if (sizec)
+		*sizec = prom_n_size_cells(dev);
+}
+
+static u64 of_bus_default_map(u32 *addr, u32 *range, int na, int ns, int pna)
+{
+	u64 cp, s, da;
+
+	cp = of_read_addr(range, na);
+	s  = of_read_addr(range + na + pna, ns);
+	da = of_read_addr(addr, na);
+
+	DBG("OF: default map, cp="PRu64", s="PRu64", da="PRu64"\n",
+	    cp, s, da);
+
+	if (da < cp || da >= (cp + s))
+		return OF_BAD_ADDR;
+	return da - cp;
+}
+
+static int of_bus_default_translate(u32 *addr, u64 offset, int na)
+{
+	u64 a = of_read_addr(addr, na);
+	memset(addr, 0, na * 4);
+	a += offset;
+	if (na > 1)
+		addr[na - 2] = a >> 32;
+	addr[na - 1] = a & 0xffffffffu;
+
+	return 0;
+}
+
+static unsigned int of_bus_default_get_flags(u32 *addr)
+{
+	return IORESOURCE_MEM;
+}
+
+
+/*
+ * PCI bus specific translator
+ */
+
+static int of_bus_pci_match(struct device_node *np)
+{
+	return !strcmp(np->type, "pci");
+}
+
+static void of_bus_pci_count_cells(struct device_node *np,
+				   int *addrc, int *sizec)
+{
+	if (addrc)
+		*addrc = 3;
+	if (sizec)
+		*sizec = 2;
+}
+
+static u64 of_bus_pci_map(u32 *addr, u32 *range, int na, int ns, int pna)
+{
+	u64 cp, s, da;
+
+	/* Check address type match */
+	if ((addr[0] ^ range[0]) & 0x03000000)
+		return OF_BAD_ADDR;
+
+	/* Read address values, skipping high cell */
+	cp = of_read_addr(range + 1, na - 1);
+	s  = of_read_addr(range + na + pna, ns);
+	da = of_read_addr(addr + 1, na - 1);
+
+	DBG("OF: PCI map, cp="PRu64", s="PRu64", da="PRu64"\n", cp, s, da);
+
+	if (da < cp || da >= (cp + s))
+		return OF_BAD_ADDR;
+	return da - cp;
+}
+
+static int of_bus_pci_translate(u32 *addr, u64 offset, int na)
+{
+	return of_bus_default_translate(addr + 1, offset, na - 1);
+}
+
+static unsigned int of_bus_pci_get_flags(u32 *addr)
+{
+	unsigned int flags = 0;
+	u32 w = addr[0];
+
+	switch((w >> 24) & 0x03) {
+	case 0x01:
+		flags |= IORESOURCE_IO;
+	case 0x02: /* 32 bits */
+	case 0x03: /* 64 bits */
+		flags |= IORESOURCE_MEM;
+	}
+	if (w & 0x40000000)
+		flags |= IORESOURCE_PREFETCH;
+	return flags;
+}
+
+/*
+ * ISA bus specific translator
+ */
+
+static int of_bus_isa_match(struct device_node *np)
+{
+	return !strcmp(np->name, "isa");
+}
+
+static void of_bus_isa_count_cells(struct device_node *child,
+				   int *addrc, int *sizec)
+{
+	if (addrc)
+		*addrc = 2;
+	if (sizec)
+		*sizec = 1;
+}
+
+static u64 of_bus_isa_map(u32 *addr, u32 *range, int na, int ns, int pna)
+{
+	u64 cp, s, da;
+
+	/* Check address type match */
+	if ((addr[0] ^ range[0]) & 0x00000001)
+		return OF_BAD_ADDR;
+
+	/* Read address values, skipping high cell */
+	cp = of_read_addr(range + 1, na - 1);
+	s  = of_read_addr(range + na + pna, ns);
+	da = of_read_addr(addr + 1, na - 1);
+
+	DBG("OF: ISA map, cp="PRu64", s="PRu64", da="PRu64"\n", cp, s, da);
+
+	if (da < cp || da >= (cp + s))
+		return OF_BAD_ADDR;
+	return da - cp;
+}
+
+static int of_bus_isa_translate(u32 *addr, u64 offset, int na)
+{
+	return of_bus_default_translate(addr + 1, offset, na - 1);
+}
+
+static unsigned int of_bus_isa_get_flags(u32 *addr)
+{
+	unsigned int flags = 0;
+	u32 w = addr[0];
+
+	if (w & 1)
+		flags |= IORESOURCE_IO;
+	else
+		flags |= IORESOURCE_MEM;
+	return flags;
+}
+
+
+/*
+ * Array of bus specific translators
+ */
+
+static struct of_bus of_busses[] = {
+	/* PCI */
+	{
+		.name = "pci",
+		.addresses = "assigned-addresses",
+		.match = of_bus_pci_match,
+		.count_cells = of_bus_pci_count_cells,
+		.map = of_bus_pci_map,
+		.translate = of_bus_pci_translate,
+		.get_flags = of_bus_pci_get_flags,
+	},
+	/* ISA */
+	{
+		.name = "isa",
+		.addresses = "reg",
+		.match = of_bus_isa_match,
+		.count_cells = of_bus_isa_count_cells,
+		.map = of_bus_isa_map,
+		.translate = of_bus_isa_translate,
+		.get_flags = of_bus_isa_get_flags,
+	},
+	/* Default */
+	{
+		.name = "default",
+		.addresses = "reg",
+		.match = NULL,
+		.count_cells = of_bus_default_count_cells,
+		.map = of_bus_default_map,
+		.translate = of_bus_default_translate,
+		.get_flags = of_bus_default_get_flags,
+	},
+};
+
+static struct of_bus *of_match_bus(struct device_node *np)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(of_busses); i ++)
+		if (!of_busses[i].match || of_busses[i].match(np))
+			return &of_busses[i];
+	BUG();
+	return NULL;
+}
+
+static int of_translate_one(struct device_node *parent, struct of_bus *bus,
+			    struct of_bus *pbus, u32 *addr,
+			    int na, int ns, int pna)
+{
+	u32 *ranges;
+	unsigned int rlen;
+	int rone;
+	u64 offset = OF_BAD_ADDR;
+
+	/* Normally, an absence of a "ranges" property means we are
+	 * crossing a non-translatable boundary, and thus the addresses
+	 * below the current not cannot be converted to CPU physical ones.
+	 * Unfortunately, while this is very clear in the spec, it's not
+	 * what Apple understood, and they do have things like /uni-n or
+	 * /ht nodes with no "ranges" property and a lot of perfectly
+	 * useable mapped devices below them. Thus we treat the absence of
+	 * "ranges" as equivalent to an empty "ranges" property which means
+	 * a 1:1 translation at that level. It's up to the caller not to try
+	 * to translate addresses that aren't supposed to be translated in
+	 * the first place. --BenH.
+	 */
+	ranges = (u32 *)get_property(parent, "ranges", &rlen);
+	if (ranges == NULL || rlen == 0) {
+		offset = of_read_addr(addr, na);
+		memset(addr, 0, pna * 4);
+		DBG("OF: no ranges, 1:1 translation\n");
+		goto finish;
+	}
+
+	DBG("OF: walking ranges...\n");
+
+	/* Now walk through the ranges */
+	rlen /= 4;
+	rone = na + pna + ns;
+	for (; rlen >= rone; rlen -= rone, ranges += rone) {
+		offset = bus->map(addr, ranges, na, ns, pna);
+		if (offset != OF_BAD_ADDR)
+			break;
+	}
+	if (offset == OF_BAD_ADDR) {
+		DBG("OF: not found !\n");
+		return 1;
+	}
+	memcpy(addr, ranges + na, 4 * pna);
+
+ finish:
+	of_dump_addr("OF: parent translation for:", addr, pna);
+	DBG("OF: with offset: "PRu64"\n", offset);
+
+	/* Translate it into parent bus space */
+	return pbus->translate(addr, offset, pna);
+}
+
+
+/*
+ * Translate an address from the device-tree into a CPU physical address,
+ * this walks up the tree and applies the various bus mappings on the
+ * way.
+ *
+ * Note: We consider that crossing any level with #size-cells == 0 to mean
+ * that translation is impossible (that is we are not dealing with a value
+ * that can be mapped to a cpu physical address). This is not really specified
+ * that way, but this is traditionally the way IBM at least do things
+ */
+u64 of_translate_address(struct device_node *dev, u32 *in_addr)
+{
+	struct device_node *parent = NULL;
+	struct of_bus *bus, *pbus;
+	u32 addr[OF_MAX_ADDR_CELLS];
+	int na, ns, pna, pns;
+	u64 result = OF_BAD_ADDR;
+
+	DBG("OF: ** translation for device %s **\n", dev->full_name);
+
+	/* Increase refcount at current level */
+	of_node_get(dev);
+
+	/* Get parent & match bus type */
+	parent = of_get_parent(dev);
+	if (parent == NULL)
+		goto bail;
+	bus = of_match_bus(parent);
+
+	/* Cound address cells & copy address locally */
+	bus->count_cells(dev, &na, &ns);
+	if (!OF_CHECK_COUNTS(na, ns)) {
+		printk(KERN_ERR "prom_parse: Bad cell count for %s\n",
+		       dev->full_name);
+		goto bail;
+	}
+	memcpy(addr, in_addr, na * 4);
+
+	DBG("OF: bus is %s (na=%d, ns=%d) on %s\n",
+	    bus->name, na, ns, parent->full_name);
+	of_dump_addr("OF: translating address:", addr, na);
+
+	/* Translate */
+	for (;;) {
+		/* Switch to parent bus */
+		of_node_put(dev);
+		dev = parent;
+		parent = of_get_parent(dev);
+
+		/* If root, we have finished */
+		if (parent == NULL) {
+			DBG("OF: reached root node\n");
+			result = of_read_addr(addr, na);
+			break;
+		}
+
+		/* Get new parent bus and counts */
+		pbus = of_match_bus(parent);
+		pbus->count_cells(dev, &pna, &pns);
+		if (!OF_CHECK_COUNTS(pna, pns)) {
+			printk(KERN_ERR "prom_parse: Bad cell count for %s\n",
+			       dev->full_name);
+			break;
+		}
+
+		DBG("OF: parent bus is %s (na=%d, ns=%d) on %s\n",
+		    pbus->name, pna, pns, parent->full_name);
+
+		/* Apply bus translation */
+		if (of_translate_one(dev, bus, pbus, addr, na, ns, pna))
+			break;
+
+		/* Complete the move up one level */
+		na = pna;
+		ns = pns;
+		bus = pbus;
+
+		of_dump_addr("OF: one level translation:", addr, na);
+	}
+ bail:
+	of_node_put(parent);
+	of_node_put(dev);
+
+	return result;
+}
+EXPORT_SYMBOL(of_translate_address);
+
+u32 *of_get_address(struct device_node *dev, int index, u64 *size,
+		    unsigned int *flags)
+{
+	u32 *prop;
+	unsigned int psize;
+	struct device_node *parent;
+	struct of_bus *bus;
+	int onesize, i, na, ns;
+
+	/* Get parent & match bus type */
+	parent = of_get_parent(dev);
+	if (parent == NULL)
+		return NULL;
+	bus = of_match_bus(parent);
+	bus->count_cells(dev, &na, &ns);
+	of_node_put(parent);
+	if (!OF_CHECK_COUNTS(na, ns))
+		return NULL;
+
+	/* Get "reg" or "assigned-addresses" property */
+	prop = (u32 *)get_property(dev, bus->addresses, &psize);
+	if (prop == NULL)
+		return NULL;
+	psize /= 4;
+
+	onesize = na + ns;
+	for (i = 0; psize >= onesize; psize -= onesize, prop += onesize, i++)
+		if (i == index) {
+			if (size)
+				*size = of_read_addr(prop + na, ns);
+			if (flags)
+				*flags = bus->get_flags(prop);
+			return prop;
+		}
+	return NULL;
+}
+EXPORT_SYMBOL(of_get_address);
+
+u32 *of_get_pci_address(struct device_node *dev, int bar_no, u64 *size,
+			unsigned int *flags)
+{
+	u32 *prop;
+	unsigned int psize;
+	struct device_node *parent;
+	struct of_bus *bus;
+	int onesize, i, na, ns;
+
+	/* Get parent & match bus type */
+	parent = of_get_parent(dev);
+	if (parent == NULL)
+		return NULL;
+	bus = of_match_bus(parent);
+	if (strcmp(bus->name, "pci"))
+		return NULL;
+	bus->count_cells(dev, &na, &ns);
+	of_node_put(parent);
+	if (!OF_CHECK_COUNTS(na, ns))
+		return NULL;
+
+	/* Get "reg" or "assigned-addresses" property */
+	prop = (u32 *)get_property(dev, bus->addresses, &psize);
+	if (prop == NULL)
+		return NULL;
+	psize /= 4;
+
+	onesize = na + ns;
+	for (i = 0; psize >= onesize; psize -= onesize, prop += onesize, i++)
+		if ((prop[0] & 0xff) == ((bar_no * 4) + PCI_BASE_ADDRESS_0)) {
+			if (size)
+				*size = of_read_addr(prop + na, ns);
+			if (flags)
+				*flags = bus->get_flags(prop);
+			return prop;
+		}
+	return NULL;
+}
+EXPORT_SYMBOL(of_get_pci_address);
+
+static int __of_address_to_resource(struct device_node *dev, u32 *addrp,
+				    u64 size, unsigned int flags,
+				    struct resource *r)
+{
+	u64 taddr;
+
+	if ((flags & (IORESOURCE_IO | IORESOURCE_MEM)) == 0)
+		return -EINVAL;
+	taddr = of_translate_address(dev, addrp);
+	if (taddr == OF_BAD_ADDR)
+		return -EINVAL;
+	memset(r, 0, sizeof(struct resource));
+	if (flags & IORESOURCE_IO) {
+		unsigned long port;
+		port = pci_address_to_pio(taddr);
+		if (port == (unsigned long)-1)
+			return -EINVAL;
+		r->start = port;
+		r->end = port + size - 1;
+	} else {
+		r->start = taddr;
+		r->end = taddr + size - 1;
+	}
+	r->flags = flags;
+	r->name = dev->name;
+	return 0;
+}
+
+int of_address_to_resource(struct device_node *dev, int index,
+			   struct resource *r)
+{
+	u32		*addrp;
+	u64		size;
+	unsigned int	flags;
+
+	addrp = of_get_address(dev, index, &size, &flags);
+	if (addrp == NULL)
+		return -EINVAL;
+	return __of_address_to_resource(dev, addrp, size, flags, r);
+}
+EXPORT_SYMBOL_GPL(of_address_to_resource);
+
+int of_pci_address_to_resource(struct device_node *dev, int bar,
+			       struct resource *r)
+{
+	u32		*addrp;
+	u64		size;
+	unsigned int	flags;
+
+	addrp = of_get_pci_address(dev, bar, &size, &flags);
+	if (addrp == NULL)
+		return -EINVAL;
+	return __of_address_to_resource(dev, addrp, size, flags, r);
+}
+EXPORT_SYMBOL_GPL(of_pci_address_to_resource);

arch/powerpc/kernel/rtas_pci.c

@@ -188,39 +188,19 @@ int is_python(struct device_node *dev)
 	return 0;
 }
 
-static int get_phb_reg_prop(struct device_node *dev,
-			    unsigned int addr_size_words,
-			    struct reg_property64 *reg)
+static void python_countermeasures(struct device_node *dev)
 {
-	unsigned int *ui_ptr = NULL, len;
-
-	/* Found a PHB, now figure out where his registers are mapped. */
-	ui_ptr = (unsigned int *)get_property(dev, "reg", &len);
-	if (ui_ptr == NULL)
-		return 1;
-
-	if (addr_size_words == 1) {
-		reg->address = ((struct reg_property32 *)ui_ptr)->address;
-		reg->size    = ((struct reg_property32 *)ui_ptr)->size;
-	} else {
-		*reg = *((struct reg_property64 *)ui_ptr);
-	}
-
-	return 0;
-}
-
-static void python_countermeasures(struct device_node *dev,
-				   unsigned int addr_size_words)
-{
-	struct reg_property64 reg_struct;
+	struct resource registers;
 	void __iomem *chip_regs;
 	volatile u32 val;
 
-	if (get_phb_reg_prop(dev, addr_size_words, &reg_struct))
+	if (of_address_to_resource(dev, 0, &registers)) {
+		printk(KERN_ERR "Can't get address for Python workarounds !\n");
 		return;
+	}
 
 	/* Python's register file is 1 MB in size. */
-	chip_regs = ioremap(reg_struct.address & ~(0xfffffUL), 0x100000);
+	chip_regs = ioremap(registers.start & ~(0xfffffUL), 0x100000);
 
 	/*
 	 * Firmware doesn't always clear this bit which is critical
@@ -301,11 +281,10 @@ static int phb_set_bus_ranges(struct device_node *dev,
 }
 
 static int __devinit setup_phb(struct device_node *dev,
-			       struct pci_controller *phb,
-			       unsigned int addr_size_words)
+			       struct pci_controller *phb)
 {
 	if (is_python(dev))
-		python_countermeasures(dev, addr_size_words);
+		python_countermeasures(dev);
 
 	if (phb_set_bus_ranges(dev, phb))
 		return 1;
@@ -320,8 +299,8 @@ unsigned long __init find_and_init_phbs(void)
 {
 	struct device_node *node;
 	struct pci_controller *phb;
-	unsigned int root_size_cells = 0;
 	unsigned int index;
+	unsigned int root_size_cells = 0;
 	unsigned int *opprop = NULL;
 	struct device_node *root = of_find_node_by_path("/");
 
@@ -343,10 +322,11 @@ unsigned long __init find_and_init_phbs(void)
 		phb = pcibios_alloc_controller(node);
 		if (!phb)
 			continue;
-		setup_phb(node, phb, root_size_cells);
+		setup_phb(node, phb);
 		pci_process_bridge_OF_ranges(phb, node, 0);
 		pci_setup_phb_io(phb, index == 0);
 #ifdef CONFIG_PPC_PSERIES
+		/* XXX This code need serious fixing ... --BenH */
 		if (ppc64_interrupt_controller == IC_OPEN_PIC && pSeries_mpic) {
 			int addr = root_size_cells * (index + 2) - 1;
 			mpic_assign_isu(pSeries_mpic, index, opprop[addr]);
@@ -381,22 +361,17 @@ unsigned long __init find_and_init_phbs(void)
 
 struct pci_controller * __devinit init_phb_dynamic(struct device_node *dn)
 {
-	struct device_node *root = of_find_node_by_path("/");
-	unsigned int root_size_cells = 0;
 	struct pci_controller *phb;
 	int primary;
 
-	root_size_cells = prom_n_size_cells(root);
-
 	primary = list_empty(&hose_list);
 	phb = pcibios_alloc_controller(dn);
 	if (!phb)
 		return NULL;
-	setup_phb(dn, phb, root_size_cells);
+	setup_phb(dn, phb);
 	pci_process_bridge_OF_ranges(phb, dn, primary);
 
 	pci_setup_phb_io_dynamic(phb, primary);
-	of_node_put(root);
 
 	pci_devs_phb_init_dynamic(phb);
 	scan_phb(phb);

arch/powerpc/kernel/setup-common.c

@@ -93,8 +93,8 @@ EXPORT_SYMBOL(ppc_do_canonicalize_irqs);
 /* also used by kexec */
 void machine_shutdown(void)
 {
-	if (ppc_md.nvram_sync)
-		ppc_md.nvram_sync();
+	if (ppc_md.machine_shutdown)
+		ppc_md.machine_shutdown();
 }
 
 void machine_restart(char *cmd)
@@ -294,129 +294,6 @@ struct seq_operations cpuinfo_op = {
 	.show =	show_cpuinfo,
 };
 
-#ifdef CONFIG_PPC_MULTIPLATFORM
-static int __init set_preferred_console(void)
-{
-	struct device_node *prom_stdout = NULL;
-	char *name;
-	u32 *spd;
-	int offset = 0;
-
-	DBG(" -> set_preferred_console()\n");
-
-	/* The user has requested a console so this is already set up. */
-	if (strstr(saved_command_line, "console=")) {
-		DBG(" console was specified !\n");
-		return -EBUSY;
-	}
-
-	if (!of_chosen) {
-		DBG(" of_chosen is NULL !\n");
-		return -ENODEV;
-	}
-	/* We are getting a weird phandle from OF ... */
-	/* ... So use the full path instead */
-	name = (char *)get_property(of_chosen, "linux,stdout-path", NULL);
-	if (name == NULL) {
-		DBG(" no linux,stdout-path !\n");
-		return -ENODEV;
-	}
-	prom_stdout = of_find_node_by_path(name);
-	if (!prom_stdout) {
-		DBG(" can't find stdout package %s !\n", name);
-		return -ENODEV;
-	}	
-	DBG("stdout is %s\n", prom_stdout->full_name);
-
-	name = (char *)get_property(prom_stdout, "name", NULL);
-	if (!name) {
-		DBG(" stdout package has no name !\n");
-		goto not_found;
-	}
-	spd = (u32 *)get_property(prom_stdout, "current-speed", NULL);
-
-	if (0)
-		;
-#ifdef CONFIG_SERIAL_8250_CONSOLE
-	else if (strcmp(name, "serial") == 0) {
-		int i;
-		u32 *reg = (u32 *)get_property(prom_stdout, "reg", &i);
-		if (i > 8) {
-			switch (reg[1]) {
-				case 0x3f8:
-					offset = 0;
-					break;
-				case 0x2f8:
-					offset = 1;
-					break;
-				case 0x898:
-					offset = 2;
-					break;
-				case 0x890:
-					offset = 3;
-					break;
-				default:
-					/* We dont recognise the serial port */
-					goto not_found;
-			}
-		}
-	}
-#endif /* CONFIG_SERIAL_8250_CONSOLE */
-#ifdef CONFIG_PPC_PSERIES
-	else if (strcmp(name, "vty") == 0) {
- 		u32 *reg = (u32 *)get_property(prom_stdout, "reg", NULL);
- 		char *compat = (char *)get_property(prom_stdout, "compatible", NULL);
-
- 		if (reg && compat && (strcmp(compat, "hvterm-protocol") == 0)) {
- 			/* Host Virtual Serial Interface */
- 			switch (reg[0]) {
- 				case 0x30000000:
- 					offset = 0;
- 					break;
- 				case 0x30000001:
- 					offset = 1;
- 					break;
- 				default:
-					goto not_found;
- 			}
-			of_node_put(prom_stdout);
-			DBG("Found hvsi console at offset %d\n", offset);
- 			return add_preferred_console("hvsi", offset, NULL);
- 		} else {
- 			/* pSeries LPAR virtual console */
-			of_node_put(prom_stdout);
-			DBG("Found hvc console\n");
- 			return add_preferred_console("hvc", 0, NULL);
- 		}
-	}
-#endif /* CONFIG_PPC_PSERIES */
-#ifdef CONFIG_SERIAL_PMACZILOG_CONSOLE
-	else if (strcmp(name, "ch-a") == 0)
-		offset = 0;
-	else if (strcmp(name, "ch-b") == 0)
-		offset = 1;
-#endif /* CONFIG_SERIAL_PMACZILOG_CONSOLE */
-	else
-		goto not_found;
-	of_node_put(prom_stdout);
-
-	DBG("Found serial console at ttyS%d\n", offset);
-
-	if (spd) {
-		static char __initdata opt[16];
-		sprintf(opt, "%d", *spd);
-		return add_preferred_console("ttyS", offset, opt);
-	} else
-		return add_preferred_console("ttyS", offset, NULL);
-
- not_found:
-	DBG("No preferred console found !\n");
-	of_node_put(prom_stdout);
-	return -ENODEV;
-}
-console_initcall(set_preferred_console);
-#endif /* CONFIG_PPC_MULTIPLATFORM */
-
 void __init check_for_initrd(void)
 {
 #ifdef CONFIG_BLK_DEV_INITRD
@@ -442,7 +319,7 @@ void __init check_for_initrd(void)
 	/* If we were passed an initrd, set the ROOT_DEV properly if the values
 	 * look sensible. If not, clear initrd reference.
 	 */
-	if (initrd_start >= KERNELBASE && initrd_end >= KERNELBASE &&
+	if (is_kernel_addr(initrd_start) && is_kernel_addr(initrd_end) &&
 	    initrd_end > initrd_start)
 		ROOT_DEV = Root_RAM0;
 	else

arch/powerpc/kernel/setup_32.c

@@ -39,6 +39,8 @@
 #include <asm/nvram.h>
 #include <asm/xmon.h>
 #include <asm/time.h>
+#include <asm/serial.h>
+#include <asm/udbg.h>
 
 #include "setup.h"
 
@@ -172,12 +174,23 @@ void __init platform_init(void)
  */
 void __init machine_init(unsigned long dt_ptr, unsigned long phys)
 {
+	/* If btext is enabled, we might have a BAT setup for early display,
+	 * thus we do enable some very basic udbg output
+	 */
+#ifdef CONFIG_BOOTX_TEXT
+	udbg_putc = btext_drawchar;
+#endif
+
+	/* Do some early initialization based on the flat device tree */
 	early_init_devtree(__va(dt_ptr));
 
+	/* Check default command line */
 #ifdef CONFIG_CMDLINE
-	strlcpy(cmd_line, CONFIG_CMDLINE, sizeof(cmd_line));
+	if (cmd_line[0] == 0)
+		strlcpy(cmd_line, CONFIG_CMDLINE, sizeof(cmd_line));
 #endif /* CONFIG_CMDLINE */
 
+	/* Base init based on machine type */
 	platform_init();
 
 #ifdef CONFIG_6xx
@@ -282,25 +295,22 @@ void __init setup_arch(char **cmdline_p)
 
 	unflatten_device_tree();
 	check_for_initrd();
+
+	if (ppc_md.init_early)
+		ppc_md.init_early();
+
+#ifdef CONFIG_SERIAL_8250
+	find_legacy_serial_ports();
+#endif
 	finish_device_tree();
 
 	smp_setup_cpu_maps();
 
-#ifdef CONFIG_BOOTX_TEXT
-	init_boot_display();
-#endif
-
-#ifdef CONFIG_PPC_PMAC
-	/* This could be called "early setup arch", it must be done
-	 * now because xmon need it
-	 */
-	if (_machine == _MACH_Pmac)
-		pmac_feature_init();	/* New cool way */
-#endif
-
 #ifdef CONFIG_XMON_DEFAULT
 	xmon_init(1);
 #endif
+	/* Register early console */
+	register_early_udbg_console();
 
 #if defined(CONFIG_KGDB)
 	if (ppc_md.kgdb_map_scc)

arch/powerpc/kernel/setup_64.c

@@ -34,6 +34,7 @@
 #include <linux/serial.h>
 #include <linux/serial_8250.h>
 #include <asm/io.h>
+#include <asm/kdump.h>
 #include <asm/prom.h>
 #include <asm/processor.h>
 #include <asm/pgtable.h>
@@ -268,6 +269,10 @@ void __init early_setup(unsigned long dt_ptr)
 	}
 	ppc_md = **mach;
 
+#ifdef CONFIG_CRASH_DUMP
+	kdump_setup();
+#endif
+
 	DBG("Found, Initializing memory management...\n");
 
 	/*
@@ -317,6 +322,7 @@ void early_setup_secondary(void)
 void smp_release_cpus(void)
 {
 	extern unsigned long __secondary_hold_spinloop;
+	unsigned long *ptr;
 
 	DBG(" -> smp_release_cpus()\n");
 
@@ -327,7 +333,9 @@ void smp_release_cpus(void)
 	 * This is useless but harmless on iSeries, secondaries are already
 	 * waiting on their paca spinloops. */
 
-	__secondary_hold_spinloop = 1;
+	ptr  = (unsigned long *)((unsigned long)&__secondary_hold_spinloop
+			- PHYSICAL_START);
+	*ptr = 1;
 	mb();
 
 	DBG(" <- smp_release_cpus()\n");
@@ -459,16 +467,21 @@ void __init setup_system(void)
 	 */
 	ppc_md.init_early();
 
+ 	/*
+	 * We can discover serial ports now since the above did setup the
+	 * hash table management for us, thus ioremap works. We do that early
+	 * so that further code can be debugged
+	 */
+#ifdef CONFIG_SERIAL_8250
+	find_legacy_serial_ports();
+#endif
+
 	/*
 	 * "Finish" the device-tree, that is do the actual parsing of
 	 * some of the properties like the interrupt map
 	 */
 	finish_device_tree();
 
-#ifdef CONFIG_BOOTX_TEXT
-	init_boot_display();
-#endif
-
 	/*
 	 * Initialize xmon
 	 */
@@ -507,6 +520,9 @@ void __init setup_system(void)
 	       ppc64_caches.iline_size);
 	printk("htab_address                  = 0x%p\n", htab_address);
 	printk("htab_hash_mask                = 0x%lx\n", htab_hash_mask);
+#if PHYSICAL_START > 0
+	printk("physical_start                = 0x%x\n", PHYSICAL_START);
+#endif
 	printk("-----------------------------------------------------\n");
 
 	mm_init_ppc64();
@@ -657,187 +673,6 @@ void ppc64_terminate_msg(unsigned int src, const char *msg)
 	printk("[terminate]%04x %s\n", src, msg);
 }
 
-#ifndef CONFIG_PPC_ISERIES
-/*
- * This function can be used by platforms to "find" legacy serial ports.
- * It works for "serial" nodes under an "isa" node, and will try to
- * respect the "ibm,aix-loc" property if any. It works with up to 8
- * ports.
- */
-
-#define MAX_LEGACY_SERIAL_PORTS	8
-static struct plat_serial8250_port serial_ports[MAX_LEGACY_SERIAL_PORTS+1];
-static unsigned int old_serial_count;
-
-void __init generic_find_legacy_serial_ports(u64 *physport,
-		unsigned int *default_speed)
-{
-	struct device_node *np;
-	u32 *sizeprop;
-
-	struct isa_reg_property {
-		u32 space;
-		u32 address;
-		u32 size;
-	};
-	struct pci_reg_property {
-		struct pci_address addr;
-		u32 size_hi;
-		u32 size_lo;
-	};                                                                        
-
-	DBG(" -> generic_find_legacy_serial_port()\n");
-
-	*physport = 0;
-	if (default_speed)
-		*default_speed = 0;
-
-	np = of_find_node_by_path("/");
-	if (!np)
-		return;
-
-	/* First fill our array */
-	for (np = NULL; (np = of_find_node_by_type(np, "serial"));) {
-		struct device_node *isa, *pci;
-		struct isa_reg_property *reg;
-		unsigned long phys_size, addr_size, io_base;
-		u32 *rangesp;
-		u32 *interrupts, *clk, *spd;
-		char *typep;
-		int index, rlen, rentsize;
-
-		/* Ok, first check if it's under an "isa" parent */
-		isa = of_get_parent(np);
-		if (!isa || strcmp(isa->name, "isa")) {
-			DBG("%s: no isa parent found\n", np->full_name);
-			continue;
-		}
-		
-		/* Now look for an "ibm,aix-loc" property that gives us ordering
-		 * if any...
-		 */
-	 	typep = (char *)get_property(np, "ibm,aix-loc", NULL);
-
-		/* Get the ISA port number */
-		reg = (struct isa_reg_property *)get_property(np, "reg", NULL);	
-		if (reg == NULL)
-			goto next_port;
-		/* We assume the interrupt number isn't translated ... */
-		interrupts = (u32 *)get_property(np, "interrupts", NULL);
-		/* get clock freq. if present */
-		clk = (u32 *)get_property(np, "clock-frequency", NULL);
-		/* get default speed if present */
-		spd = (u32 *)get_property(np, "current-speed", NULL);
-		/* Default to locate at end of array */
-		index = old_serial_count; /* end of the array by default */
-
-		/* If we have a location index, then use it */
-		if (typep && *typep == 'S') {
-			index = simple_strtol(typep+1, NULL, 0) - 1;
-			/* if index is out of range, use end of array instead */
-			if (index >= MAX_LEGACY_SERIAL_PORTS)
-				index = old_serial_count;
-			/* if our index is still out of range, that mean that
-			 * array is full, we could scan for a free slot but that
-			 * make little sense to bother, just skip the port
-			 */
-			if (index >= MAX_LEGACY_SERIAL_PORTS)
-				goto next_port;
-			if (index >= old_serial_count)
-				old_serial_count = index + 1;
-			/* Check if there is a port who already claimed our slot */
-			if (serial_ports[index].iobase != 0) {
-				/* if we still have some room, move it, else override */
-				if (old_serial_count < MAX_LEGACY_SERIAL_PORTS) {
-					DBG("Moved legacy port %d -> %d\n", index,
-					    old_serial_count);
-					serial_ports[old_serial_count++] =
-						serial_ports[index];
-				} else {
-					DBG("Replacing legacy port %d\n", index);
-				}
-			}
-		}
-		if (index >= MAX_LEGACY_SERIAL_PORTS)
-			goto next_port;
-		if (index >= old_serial_count)
-			old_serial_count = index + 1;
-
-		/* Now fill the entry */
-		memset(&serial_ports[index], 0, sizeof(struct plat_serial8250_port));
-		serial_ports[index].uartclk = clk ? *clk : BASE_BAUD * 16;
-		serial_ports[index].iobase = reg->address;
-		serial_ports[index].irq = interrupts ? interrupts[0] : 0;
-		serial_ports[index].flags = ASYNC_BOOT_AUTOCONF;
-
-		DBG("Added legacy port, index: %d, port: %x, irq: %d, clk: %d\n",
-		    index,
-		    serial_ports[index].iobase,
-		    serial_ports[index].irq,
-		    serial_ports[index].uartclk);
-
-		/* Get phys address of IO reg for port 1 */
-		if (index != 0)
-			goto next_port;
-
-		pci = of_get_parent(isa);
-		if (!pci) {
-			DBG("%s: no pci parent found\n", np->full_name);
-			goto next_port;
-		}
-
-		rangesp = (u32 *)get_property(pci, "ranges", &rlen);
-		if (rangesp == NULL) {
-			of_node_put(pci);
-			goto next_port;
-		}
-		rlen /= 4;
-
-		/* we need the #size-cells of the PCI bridge node itself */
-		phys_size = 1;
-		sizeprop = (u32 *)get_property(pci, "#size-cells", NULL);
-		if (sizeprop != NULL)
-			phys_size = *sizeprop;
-		/* we need the parent #addr-cells */
-		addr_size = prom_n_addr_cells(pci);
-		rentsize = 3 + addr_size + phys_size;
-		io_base = 0;
-		for (;rlen >= rentsize; rlen -= rentsize,rangesp += rentsize) {
-			if (((rangesp[0] >> 24) & 0x3) != 1)
-				continue; /* not IO space */
-			io_base = rangesp[3];
-			if (addr_size == 2)
-				io_base = (io_base << 32) | rangesp[4];
-		}
-		if (io_base != 0) {
-			*physport = io_base + reg->address;
-			if (default_speed && spd)
-				*default_speed = *spd;
-		}
-		of_node_put(pci);
-	next_port:
-		of_node_put(isa);
-	}
-
-	DBG(" <- generic_find_legacy_serial_port()\n");
-}
-
-static struct platform_device serial_device = {
-	.name	= "serial8250",
-	.id	= PLAT8250_DEV_PLATFORM,
-	.dev	= {
-		.platform_data = serial_ports,
-	},
-};
-
-static int __init serial_dev_init(void)
-{
-	return platform_device_register(&serial_device);
-}
-arch_initcall(serial_dev_init);
-
-#endif /* CONFIG_PPC_ISERIES */
-
 int check_legacy_ioport(unsigned long base_port)
 {
 	if (ppc_md.check_legacy_ioport == NULL)

arch/powerpc/kernel/signal_32.c

@@ -76,7 +76,6 @@
  * registers from *regs.  This is what we need
  * to do when a signal has been delivered.
  */
-#define sigreturn_exit(regs)	return 0
 
 #define GP_REGS_SIZE	min(sizeof(elf_gregset_t32), sizeof(struct pt_regs32))
 #undef __SIGNAL_FRAMESIZE
@@ -156,9 +155,17 @@ static inline int save_general_regs(struct pt_regs *regs,
 	elf_greg_t64 *gregs = (elf_greg_t64 *)regs;
 	int i;
 
-	for (i = 0; i <= PT_RESULT; i ++)
+	if (!FULL_REGS(regs)) {
+		set_thread_flag(TIF_SAVE_NVGPRS);
+		current_thread_info()->nvgprs_frame = frame->mc_gregs;
+	}
+
+	for (i = 0; i <= PT_RESULT; i ++) {
+		if (i == 14 && !FULL_REGS(regs))
+			i = 32;
 		if (__put_user((unsigned int)gregs[i], &frame->mc_gregs[i]))
 			return -EFAULT;
+	}
 	return 0;
 }
 
@@ -179,8 +186,6 @@ static inline int restore_general_regs(struct pt_regs *regs,
 
 #else /* CONFIG_PPC64 */
 
-extern void sigreturn_exit(struct pt_regs *);
-
 #define GP_REGS_SIZE	min(sizeof(elf_gregset_t), sizeof(struct pt_regs))
 
 static inline int put_sigset_t(sigset_t __user *uset, sigset_t *set)
@@ -214,6 +219,15 @@ static inline int get_old_sigaction(struct k_sigaction *new_ka,
 static inline int save_general_regs(struct pt_regs *regs,
 		struct mcontext __user *frame)
 {
+	if (!FULL_REGS(regs)) {
+		/* Zero out the unsaved GPRs to avoid information
+		   leak, and set TIF_SAVE_NVGPRS to ensure that the
+		   registers do actually get saved later. */
+		memset(&regs->gpr[14], 0, 18 * sizeof(unsigned long));
+		current_thread_info()->nvgprs_frame = &frame->mc_gregs;
+		set_thread_flag(TIF_SAVE_NVGPRS);
+	}
+
 	return __copy_to_user(&frame->mc_gregs, regs, GP_REGS_SIZE);
 }
 
@@ -256,8 +270,10 @@ long sys_sigsuspend(old_sigset_t mask, int p2, int p3, int p4, int p6, int p7,
 	while (1) {
 		current->state = TASK_INTERRUPTIBLE;
 		schedule();
-		if (do_signal(&saveset, regs))
-			sigreturn_exit(regs);
+		if (do_signal(&saveset, regs)) {
+			set_thread_flag(TIF_RESTOREALL);
+			return 0;
+		}
 	}
 }
 
@@ -292,8 +308,10 @@ long sys_rt_sigsuspend(
 	while (1) {
 		current->state = TASK_INTERRUPTIBLE;
 		schedule();
-		if (do_signal(&saveset, regs))
-			sigreturn_exit(regs);
+		if (do_signal(&saveset, regs)) {
+			set_thread_flag(TIF_RESTOREALL);
+			return 0;
+		}
 	}
 }
 
@@ -391,9 +409,6 @@ struct rt_sigframe {
 static int save_user_regs(struct pt_regs *regs, struct mcontext __user *frame,
 		int sigret)
 {
-#ifdef CONFIG_PPC32
-	CHECK_FULL_REGS(regs);
-#endif
 	/* Make sure floating point registers are stored in regs */
 	flush_fp_to_thread(current);
 
@@ -828,12 +843,6 @@ static int handle_rt_signal(unsigned long sig, struct k_sigaction *ka,
 	regs->gpr[6] = (unsigned long) rt_sf;
 	regs->nip = (unsigned long) ka->sa.sa_handler;
 	regs->trap = 0;
-#ifdef CONFIG_PPC64
-	regs->result = 0;
-
-	if (test_thread_flag(TIF_SINGLESTEP))
-		ptrace_notify(SIGTRAP);
-#endif
 	return 1;
 
 badframe:
@@ -911,8 +920,8 @@ long sys_swapcontext(struct ucontext __user *old_ctx,
 	 */
 	if (do_setcontext(new_ctx, regs, 0))
 		do_exit(SIGSEGV);
-	sigreturn_exit(regs);
-	/* doesn't actually return back to here */
+
+	set_thread_flag(TIF_RESTOREALL);
 	return 0;
 }
 
@@ -945,12 +954,11 @@ long sys_rt_sigreturn(int r3, int r4, int r5, int r6, int r7, int r8,
 	 * nobody does any...
 	 */
 	compat_sys_sigaltstack((u32)(u64)&rt_sf->uc.uc_stack, 0, 0, 0, 0, 0, regs);
-	return (int)regs->result;
 #else
 	do_sigaltstack(&rt_sf->uc.uc_stack, NULL, regs->gpr[1]);
-	sigreturn_exit(regs);		/* doesn't return here */
-	return 0;
 #endif
+	set_thread_flag(TIF_RESTOREALL);
+	return 0;
 
  bad:
 	force_sig(SIGSEGV, current);
@@ -1041,9 +1049,7 @@ int sys_debug_setcontext(struct ucontext __user *ctx,
 	 */
 	do_sigaltstack(&ctx->uc_stack, NULL, regs->gpr[1]);
 
-	sigreturn_exit(regs);
-	/* doesn't actually return back to here */
-
+	set_thread_flag(TIF_RESTOREALL);
  out:
 	return 0;
 }
@@ -1107,12 +1113,6 @@ static int handle_signal(unsigned long sig, struct k_sigaction *ka,
 	regs->gpr[4] = (unsigned long) sc;
 	regs->nip = (unsigned long) ka->sa.sa_handler;
 	regs->trap = 0;
-#ifdef CONFIG_PPC64
-	regs->result = 0;
-
-	if (test_thread_flag(TIF_SINGLESTEP))
-		ptrace_notify(SIGTRAP);
-#endif
 
 	return 1;
 
@@ -1160,12 +1160,8 @@ long sys_sigreturn(int r3, int r4, int r5, int r6, int r7, int r8,
 	    || restore_user_regs(regs, sr, 1))
 		goto badframe;
 
-#ifdef CONFIG_PPC64
-	return (int)regs->result;
-#else
-	sigreturn_exit(regs);		/* doesn't return */
+	set_thread_flag(TIF_RESTOREALL);
 	return 0;
-#endif
 
 badframe:
 	force_sig(SIGSEGV, current);

arch/powerpc/kernel/signal_64.c

@@ -96,8 +96,10 @@ long sys_rt_sigsuspend(sigset_t __user *unewset, size_t sigsetsize, int p3, int
 	while (1) {
 		current->state = TASK_INTERRUPTIBLE;
 		schedule();
-		if (do_signal(&saveset, regs))
+		if (do_signal(&saveset, regs)) {
+			set_thread_flag(TIF_RESTOREALL);
 			return 0;
+		}
 	}
 }
 
@@ -152,6 +154,14 @@ static long setup_sigcontext(struct sigcontext __user *sc, struct pt_regs *regs,
 	err |= __put_user(0, &sc->v_regs);
 #endif /* CONFIG_ALTIVEC */
 	err |= __put_user(&sc->gp_regs, &sc->regs);
+	if (!FULL_REGS(regs)) {
+		/* Zero out the unsaved GPRs to avoid information
+		   leak, and set TIF_SAVE_NVGPRS to ensure that the
+		   registers do actually get saved later. */
+		memset(&regs->gpr[14], 0, 18 * sizeof(unsigned long));
+		set_thread_flag(TIF_SAVE_NVGPRS);
+		current_thread_info()->nvgprs_frame = &sc->gp_regs;
+	}
 	err |= __copy_to_user(&sc->gp_regs, regs, GP_REGS_SIZE);
 	err |= __copy_to_user(&sc->fp_regs, &current->thread.fpr, FP_REGS_SIZE);
 	err |= __put_user(signr, &sc->signal);
@@ -340,6 +350,7 @@ int sys_swapcontext(struct ucontext __user *old_ctx,
 		do_exit(SIGSEGV);
 
 	/* This returns like rt_sigreturn */
+	set_thread_flag(TIF_RESTOREALL);
 	return 0;
 }
 
@@ -372,7 +383,8 @@ int sys_rt_sigreturn(unsigned long r3, unsigned long r4, unsigned long r5,
 	 */
 	do_sigaltstack(&uc->uc_stack, NULL, regs->gpr[1]);
 
-	return regs->result;
+	set_thread_flag(TIF_RESTOREALL);
+	return 0;
 
 badframe:
 #if DEBUG_SIG
@@ -454,9 +466,6 @@ static int setup_rt_frame(int signr, struct k_sigaction *ka, siginfo_t *info,
 	if (err)
 		goto badframe;
 
-	if (test_thread_flag(TIF_SINGLESTEP))
-		ptrace_notify(SIGTRAP);
-
 	return 1;
 
 badframe:
@@ -502,6 +511,8 @@ static inline void syscall_restart(struct pt_regs *regs, struct k_sigaction *ka)
 		 * we only get here if there is a handler, we dont restart.
 		 */
 		regs->result = -EINTR;
+		regs->gpr[3] = EINTR;
+		regs->ccr |= 0x10000000;
 		break;
 	case -ERESTARTSYS:
 		/* ERESTARTSYS means to restart the syscall if there is no
@@ -509,6 +520,8 @@ static inline void syscall_restart(struct pt_regs *regs, struct k_sigaction *ka)
 		 */
 		if (!(ka->sa.sa_flags & SA_RESTART)) {
 			regs->result = -EINTR;
+			regs->gpr[3] = EINTR;
+			regs->ccr |= 0x10000000;
 			break;
 		}
 		/* fallthrough */

arch/powerpc/kernel/smp.c

@@ -31,6 +31,7 @@
 #include <linux/sysdev.h>
 #include <linux/cpu.h>
 #include <linux/notifier.h>
+#include <linux/topology.h>
 
 #include <asm/ptrace.h>
 #include <asm/atomic.h>
@@ -75,6 +76,8 @@ void smp_call_function_interrupt(void);
 
 int smt_enabled_at_boot = 1;
 
+static void (*crash_ipi_function_ptr)(struct pt_regs *) = NULL;
+
 #ifdef CONFIG_MPIC
 int __init smp_mpic_probe(void)
 {
@@ -123,11 +126,16 @@ void smp_message_recv(int msg, struct pt_regs *regs)
 		/* XXX Do we have to do this? */
 		set_need_resched();
 		break;
-#ifdef CONFIG_DEBUGGER
 	case PPC_MSG_DEBUGGER_BREAK:
+		if (crash_ipi_function_ptr) {
+			crash_ipi_function_ptr(regs);
+			break;
+		}
+#ifdef CONFIG_DEBUGGER
 		debugger_ipi(regs);
 		break;
-#endif
+#endif /* CONFIG_DEBUGGER */
+		/* FALLTHROUGH */
 	default:
 		printk("SMP %d: smp_message_recv(): unknown msg %d\n",
 		       smp_processor_id(), msg);
@@ -147,6 +155,17 @@ void smp_send_debugger_break(int cpu)
 }
 #endif
 
+#ifdef CONFIG_KEXEC
+void crash_send_ipi(void (*crash_ipi_callback)(struct pt_regs *))
+{
+	crash_ipi_function_ptr = crash_ipi_callback;
+	if (crash_ipi_callback) {
+		mb();
+		smp_ops->message_pass(MSG_ALL_BUT_SELF, PPC_MSG_DEBUGGER_BREAK);
+	}
+}
+#endif
+
 static void stop_this_cpu(void *dummy)
 {
 	local_irq_disable();
@@ -452,10 +471,6 @@ int __devinit __cpu_up(unsigned int cpu)
 	if (smp_ops->cpu_bootable && !smp_ops->cpu_bootable(cpu))
 		return -EINVAL;
 
-#ifdef CONFIG_PPC64
-	paca[cpu].default_decr = tb_ticks_per_jiffy;
-#endif
-
 	/* Make sure callin-map entry is 0 (can be leftover a CPU
 	 * hotplug
 	 */
@@ -554,6 +569,8 @@ void __init smp_cpus_done(unsigned int max_cpus)
 	smp_ops->setup_cpu(boot_cpuid);
 
 	set_cpus_allowed(current, old_mask);
+
+	dump_numa_cpu_topology();
 }
 
 #ifdef CONFIG_HOTPLUG_CPU

arch/powerpc/kernel/syscalls.c

@@ -43,9 +43,6 @@
 #include <asm/time.h>
 #include <asm/unistd.h>
 
-extern unsigned long wall_jiffies;
-
-
 /*
  * sys_ipc() is the de-multiplexer for the SysV IPC calls..
  *
@@ -311,31 +308,6 @@ int sys_olduname(struct oldold_utsname __user *name)
 	return error? -EFAULT: 0;
 }
 
-#ifdef CONFIG_PPC64
-time_t sys64_time(time_t __user * tloc)
-{
-	time_t secs;
-	time_t usecs;
-
-	long tb_delta = tb_ticks_since(tb_last_stamp);
-	tb_delta += (jiffies - wall_jiffies) * tb_ticks_per_jiffy;
-
-	secs  = xtime.tv_sec;  
-	usecs = (xtime.tv_nsec/1000) + tb_delta / tb_ticks_per_usec;
-	while (usecs >= USEC_PER_SEC) {
-		++secs;
-		usecs -= USEC_PER_SEC;
-	}
-
-	if (tloc) {
-		if (put_user(secs,tloc))
-			secs = -EFAULT;
-	}
-
-	return secs;
-}
-#endif
-
 long ppc_fadvise64_64(int fd, int advice, u32 offset_high, u32 offset_low,
 		      u32 len_high, u32 len_low)
 {

arch/powerpc/kernel/systbl.S

@@ -54,7 +54,7 @@ SYSCALL(link)
 SYSCALL(unlink)
 COMPAT_SYS(execve)
 SYSCALL(chdir)
-SYSX(sys64_time,compat_sys_time,sys_time)
+COMPAT_SYS(time)
 SYSCALL(mknod)
 SYSCALL(chmod)
 SYSCALL(lchown)
@@ -113,7 +113,7 @@ SYSCALL(sgetmask)
 COMPAT_SYS(ssetmask)
 SYSCALL(setreuid)
 SYSCALL(setregid)
-SYSX(sys_ni_syscall,ppc32_sigsuspend,ppc_sigsuspend)
+SYS32ONLY(sigsuspend)
 COMPAT_SYS(sigpending)
 COMPAT_SYS(sethostname)
 COMPAT_SYS(setrlimit)
@@ -160,7 +160,7 @@ SYSCALL(swapoff)
 COMPAT_SYS(sysinfo)
 COMPAT_SYS(ipc)
 SYSCALL(fsync)
-SYSX(sys_ni_syscall,ppc32_sigreturn,sys_sigreturn)
+SYS32ONLY(sigreturn)
 PPC_SYS(clone)
 COMPAT_SYS(setdomainname)
 PPC_SYS(newuname)
@@ -213,13 +213,13 @@ COMPAT_SYS(nfsservctl)
 SYSCALL(setresgid)
 SYSCALL(getresgid)
 COMPAT_SYS(prctl)
-SYSX(ppc64_rt_sigreturn,ppc32_rt_sigreturn,sys_rt_sigreturn)
+COMPAT_SYS(rt_sigreturn)
 COMPAT_SYS(rt_sigaction)
 COMPAT_SYS(rt_sigprocmask)
 COMPAT_SYS(rt_sigpending)
 COMPAT_SYS(rt_sigtimedwait)
 COMPAT_SYS(rt_sigqueueinfo)
-SYSX(ppc64_rt_sigsuspend,ppc32_rt_sigsuspend,ppc_rt_sigsuspend)
+COMPAT_SYS(rt_sigsuspend)
 COMPAT_SYS(pread64)
 COMPAT_SYS(pwrite64)
 SYSCALL(chown)
@@ -290,7 +290,7 @@ COMPAT_SYS(clock_settime)
 COMPAT_SYS(clock_gettime)
 COMPAT_SYS(clock_getres)
 COMPAT_SYS(clock_nanosleep)
-SYSX(ppc64_swapcontext,ppc32_swapcontext,ppc_swapcontext)
+COMPAT_SYS(swapcontext)
 COMPAT_SYS(tgkill)
 COMPAT_SYS(utimes)
 COMPAT_SYS(statfs64)
@@ -319,3 +319,5 @@ COMPAT_SYS(ioprio_get)
 SYSCALL(inotify_init)
 SYSCALL(inotify_add_watch)
 SYSCALL(inotify_rm_watch)
+SYSCALL(spu_run)
+SYSCALL(spu_create)

arch/powerpc/kernel/time.c

@@ -699,10 +699,6 @@ void __init time_init(void)
 	div128_by_32(1024*1024, 0, tb_ticks_per_sec, &res);
 	tb_to_xs = res.result_low;
 
-#ifdef CONFIG_PPC64
-	get_paca()->default_decr = tb_ticks_per_jiffy;
-#endif
-
 	/*
 	 * Compute scale factor for sched_clock.
 	 * The calibrate_decr() function has set tb_ticks_per_sec,

arch/powerpc/kernel/traps.c

@@ -31,6 +31,7 @@
 #include <linux/prctl.h>
 #include <linux/delay.h>
 #include <linux/kprobes.h>
+#include <linux/kexec.h>
 
 #include <asm/kdebug.h>
 #include <asm/pgtable.h>
@@ -95,7 +96,7 @@ static DEFINE_SPINLOCK(die_lock);
 
 int die(const char *str, struct pt_regs *regs, long err)
 {
-	static int die_counter;
+	static int die_counter, crash_dump_start = 0;
 	int nl = 0;
 
 	if (debugger(regs))
@@ -156,7 +157,21 @@ int die(const char *str, struct pt_regs *regs, long err)
 	print_modules();
 	show_regs(regs);
 	bust_spinlocks(0);
+
+	if (!crash_dump_start && kexec_should_crash(current)) {
+		crash_dump_start = 1;
+		spin_unlock_irq(&die_lock);
+		crash_kexec(regs);
+		/* NOTREACHED */
+	}
 	spin_unlock_irq(&die_lock);
+	if (crash_dump_start)
+		/*
+		 * Only for soft-reset: Other CPUs will be responded to an IPI
+		 * sent by first kexec CPU.
+		 */
+		for(;;)
+			;
 
 	if (in_interrupt())
 		panic("Fatal exception in interrupt");
@@ -215,8 +230,10 @@ void _exception(int signr, struct pt_regs *regs, int code, unsigned long addr)
 void system_reset_exception(struct pt_regs *regs)
 {
 	/* See if any machine dependent calls */
-	if (ppc_md.system_reset_exception)
-		ppc_md.system_reset_exception(regs);
+	if (ppc_md.system_reset_exception) {
+		if (ppc_md.system_reset_exception(regs))
+			return;
+	}
 
 	die("System Reset", regs, SIGABRT);
 
@@ -886,12 +903,10 @@ void altivec_unavailable_exception(struct pt_regs *regs)
 	die("Unrecoverable VMX/Altivec Unavailable Exception", regs, SIGABRT);
 }
 
-#if defined(CONFIG_PPC64) || defined(CONFIG_E500)
 void performance_monitor_exception(struct pt_regs *regs)
 {
 	perf_irq(regs);
 }
-#endif
 
 #ifdef CONFIG_8xx
 void SoftwareEmulation(struct pt_regs *regs)

arch/powerpc/kernel/udbg.c

@@ -16,8 +16,8 @@
 #include <linux/console.h>
 #include <asm/processor.h>
 
-void (*udbg_putc)(unsigned char c);
-unsigned char (*udbg_getc)(void);
+void (*udbg_putc)(char c);
+int (*udbg_getc)(void);
 int (*udbg_getc_poll)(void);
 
 /* udbg library, used by xmon et al */
@@ -57,8 +57,8 @@ int udbg_write(const char *s, int n)
 
 int udbg_read(char *buf, int buflen)
 {
-	char c, *p = buf;
-	int i;
+	char *p = buf;
+	int i, c;
 
 	if (!udbg_getc)
 		return 0;
@@ -66,8 +66,11 @@ int udbg_read(char *buf, int buflen)
 	for (i = 0; i < buflen; ++i) {
 		do {
 			c = udbg_getc();
+			if (c == -1 && i == 0)
+				return -1;
+
 		} while (c == 0x11 || c == 0x13);
-		if (c == 0)
+		if (c == 0 || c == -1)
 			break;
 		*p++ = c;
 	}
@@ -78,7 +81,7 @@ int udbg_read(char *buf, int buflen)
 #define UDBG_BUFSIZE 256
 void udbg_printf(const char *fmt, ...)
 {
-	unsigned char buf[UDBG_BUFSIZE];
+	char buf[UDBG_BUFSIZE];
 	va_list args;
 
 	va_start(args, fmt);
@@ -87,6 +90,12 @@ void udbg_printf(const char *fmt, ...)
 	va_end(args);
 }
 
+void __init udbg_progress(char *s, unsigned short hex)
+{
+	udbg_puts(s);
+	udbg_puts("\n");
+}
+
 /*
  * Early boot console based on udbg
  */
@@ -99,7 +108,7 @@ static void udbg_console_write(struct console *con, const char *s,
 static struct console udbg_console = {
 	.name	= "udbg",
 	.write	= udbg_console_write,
-	.flags	= CON_PRINTBUFFER,
+	.flags	= CON_PRINTBUFFER | CON_ENABLED,
 	.index	= -1,
 };
 
@@ -107,15 +116,19 @@ static int early_console_initialized;
 
 void __init disable_early_printk(void)
 {
+#if 1
 	if (!early_console_initialized)
 		return;
 	unregister_console(&udbg_console);
 	early_console_initialized = 0;
+#endif
 }
 
 /* called by setup_system */
 void register_early_udbg_console(void)
 {
+	if (early_console_initialized)
+		return;
 	early_console_initialized = 1;
 	register_console(&udbg_console);
 }

arch/powerpc/kernel/udbg_16550.c

@@ -43,9 +43,11 @@ struct NS16550 {
 #define LSR_TEMT 0x40  /* Xmitter empty */
 #define LSR_ERR  0x80  /* Error */
 
+#define LCR_DLAB 0x80
+
 static volatile struct NS16550 __iomem *udbg_comport;
 
-static void udbg_550_putc(unsigned char c)
+static void udbg_550_putc(char c)
 {
 	if (udbg_comport) {
 		while ((in_8(&udbg_comport->lsr) & LSR_THRE) == 0)
@@ -67,39 +69,80 @@ static int udbg_550_getc_poll(void)
 	return -1;
 }
 
-static unsigned char udbg_550_getc(void)
+static int udbg_550_getc(void)
 {
 	if (udbg_comport) {
 		while ((in_8(&udbg_comport->lsr) & LSR_DR) == 0)
 			/* wait for char */;
 		return in_8(&udbg_comport->rbr);
 	}
-	return 0;
+	return -1;
 }
 
-void udbg_init_uart(void __iomem *comport, unsigned int speed)
+void udbg_init_uart(void __iomem *comport, unsigned int speed,
+		    unsigned int clock)
 {
-	u16 dll = speed ? (115200 / speed) : 12;
+	unsigned int dll, base_bauds = clock / 16;
+
+	if (speed == 0)
+		speed = 9600;
+	dll = base_bauds / speed;
 
 	if (comport) {
 		udbg_comport = (struct NS16550 __iomem *)comport;
 		out_8(&udbg_comport->lcr, 0x00);
 		out_8(&udbg_comport->ier, 0xff);
 		out_8(&udbg_comport->ier, 0x00);
-		out_8(&udbg_comport->lcr, 0x80);	/* Access baud rate */
-		out_8(&udbg_comport->dll, dll & 0xff);	/* 1 = 115200,  2 = 57600,
-							   3 = 38400, 12 = 9600 baud */
-		out_8(&udbg_comport->dlm, dll >> 8);	/* dll >> 8 which should be zero
-							   for fast rates; */
-		out_8(&udbg_comport->lcr, 0x03);	/* 8 data, 1 stop, no parity */
-		out_8(&udbg_comport->mcr, 0x03);	/* RTS/DTR */
-		out_8(&udbg_comport->fcr ,0x07);	/* Clear & enable FIFOs */
+		out_8(&udbg_comport->lcr, LCR_DLAB);
+		out_8(&udbg_comport->dll, dll & 0xff);
+		out_8(&udbg_comport->dlm, dll >> 8);
+		/* 8 data, 1 stop, no parity */
+		out_8(&udbg_comport->lcr, 0x03);
+		/* RTS/DTR */
+		out_8(&udbg_comport->mcr, 0x03);
+		/* Clear & enable FIFOs */
+		out_8(&udbg_comport->fcr ,0x07);
 		udbg_putc = udbg_550_putc;
 		udbg_getc = udbg_550_getc;
 		udbg_getc_poll = udbg_550_getc_poll;
 	}
 }
 
+unsigned int udbg_probe_uart_speed(void __iomem *comport, unsigned int clock)
+{
+	unsigned int dll, dlm, divisor, prescaler, speed;
+	u8 old_lcr;
+	volatile struct NS16550 __iomem *port = comport;
+
+	old_lcr = in_8(&port->lcr);
+
+	/* select divisor latch registers.  */
+	out_8(&port->lcr, LCR_DLAB);
+
+	/* now, read the divisor */
+	dll = in_8(&port->dll);
+	dlm = in_8(&port->dlm);
+	divisor = dlm << 8 | dll;
+
+	/* check prescaling */
+	if (in_8(&port->mcr) & 0x80)
+		prescaler = 4;
+	else
+		prescaler = 1;
+
+	/* restore the LCR */
+	out_8(&port->lcr, old_lcr);
+
+	/* calculate speed */
+	speed = (clock / prescaler) / (divisor * 16);
+
+	/* sanity check */
+	if (speed < 0 || speed > (clock / 16))
+		speed = 9600;
+
+	return speed;
+}
+
 #ifdef CONFIG_PPC_MAPLE
 void udbg_maple_real_putc(unsigned char c)
 {

arch/powerpc/mm/fault.c

@@ -81,7 +81,8 @@ static int store_updates_sp(struct pt_regs *regs)
 }
 
 #if !(defined(CONFIG_4xx) || defined(CONFIG_BOOKE))
-static void do_dabr(struct pt_regs *regs, unsigned long error_code)
+static void do_dabr(struct pt_regs *regs, unsigned long address,
+		    unsigned long error_code)
 {
 	siginfo_t info;
 
@@ -99,7 +100,7 @@ static void do_dabr(struct pt_regs *regs, unsigned long error_code)
 	info.si_signo = SIGTRAP;
 	info.si_errno = 0;
 	info.si_code = TRAP_HWBKPT;
-	info.si_addr = (void __user *)regs->nip;
+	info.si_addr = (void __user *)address;
 	force_sig_info(SIGTRAP, &info, current);
 }
 #endif /* !(CONFIG_4xx || CONFIG_BOOKE)*/
@@ -159,7 +160,7 @@ int __kprobes do_page_fault(struct pt_regs *regs, unsigned long address,
 #if !(defined(CONFIG_4xx) || defined(CONFIG_BOOKE))
   	if (error_code & DSISR_DABRMATCH) {
 		/* DABR match */
-		do_dabr(regs, error_code);
+		do_dabr(regs, address, error_code);
 		return 0;
 	}
 #endif /* !(CONFIG_4xx || CONFIG_BOOKE)*/

arch/powerpc/mm/hash_utils_64.c

@@ -456,7 +456,7 @@ void __init htab_initialize(void)
 
 	/* create bolted the linear mapping in the hash table */
 	for (i=0; i < lmb.memory.cnt; i++) {
-		base = lmb.memory.region[i].base + KERNELBASE;
+		base = (unsigned long)__va(lmb.memory.region[i].base);
 		size = lmb.memory.region[i].size;
 
 		DBG("creating mapping for region: %lx : %lx\n", base, size);
@@ -498,8 +498,8 @@ void __init htab_initialize(void)
 	 * for either 4K or 16MB pages.
 	 */
 	if (tce_alloc_start) {
-		tce_alloc_start += KERNELBASE;
-		tce_alloc_end += KERNELBASE;
+		tce_alloc_start = (unsigned long)__va(tce_alloc_start);
+		tce_alloc_end = (unsigned long)__va(tce_alloc_end);
 
 		if (base + size >= tce_alloc_start)
 			tce_alloc_start = base + size + 1;
@@ -644,6 +644,7 @@ int hash_page(unsigned long ea, unsigned long access, unsigned long trap)
 	DBG_LOW(" -> rc=%d\n", rc);
 	return rc;
 }
+EXPORT_SYMBOL_GPL(hash_page);
 
 void hash_preload(struct mm_struct *mm, unsigned long ea,
 		  unsigned long access, unsigned long trap)

arch/powerpc/mm/hugetlbpage.c

@@ -549,6 +549,17 @@ fail:
 	return addr;
 }
 
+static int htlb_check_hinted_area(unsigned long addr, unsigned long len)
+{
+	struct vm_area_struct *vma;
+
+	vma = find_vma(current->mm, addr);
+	if (!vma || ((addr + len) <= vma->vm_start))
+		return 0;
+
+	return -ENOMEM;
+}
+
 static unsigned long htlb_get_low_area(unsigned long len, u16 segmask)
 {
 	unsigned long addr = 0;
@@ -618,15 +629,28 @@ unsigned long hugetlb_get_unmapped_area(struct file *file, unsigned long addr,
 	if (!cpu_has_feature(CPU_FTR_16M_PAGE))
 		return -EINVAL;
 
+	/* Paranoia, caller should have dealt with this */
+	BUG_ON((addr + len)  < addr);
+
 	if (test_thread_flag(TIF_32BIT)) {
+		/* Paranoia, caller should have dealt with this */
+		BUG_ON((addr + len) > 0x100000000UL);
+
 		curareas = current->mm->context.low_htlb_areas;
 
-		/* First see if we can do the mapping in the existing
-		 * low areas */
+		/* First see if we can use the hint address */
+		if (addr && (htlb_check_hinted_area(addr, len) == 0)) {
+			areamask = LOW_ESID_MASK(addr, len);
+			if (open_low_hpage_areas(current->mm, areamask) == 0)
+				return addr;
+		}
+
+		/* Next see if we can map in the existing low areas */
 		addr = htlb_get_low_area(len, curareas);
 		if (addr != -ENOMEM)
 			return addr;
 
+		/* Finally go looking for areas to open */
 		lastshift = 0;
 		for (areamask = LOW_ESID_MASK(0x100000000UL-len, len);
 		     ! lastshift; areamask >>=1) {
@@ -641,12 +665,22 @@ unsigned long hugetlb_get_unmapped_area(struct file *file, unsigned long addr,
 	} else {
 		curareas = current->mm->context.high_htlb_areas;
 
-		/* First see if we can do the mapping in the existing
-		 * high areas */
+		/* First see if we can use the hint address */
+		/* We discourage 64-bit processes from doing hugepage
+		 * mappings below 4GB (must use MAP_FIXED) */
+		if ((addr >= 0x100000000UL)
+		    && (htlb_check_hinted_area(addr, len) == 0)) {
+			areamask = HTLB_AREA_MASK(addr, len);
+			if (open_high_hpage_areas(current->mm, areamask) == 0)
+				return addr;
+		}
+
+		/* Next see if we can map in the existing high areas */
 		addr = htlb_get_high_area(len, curareas);
 		if (addr != -ENOMEM)
 			return addr;
 
+		/* Finally go looking for areas to open */
 		lastshift = 0;
 		for (areamask = HTLB_AREA_MASK(TASK_SIZE_USER64-len, len);
 		     ! lastshift; areamask >>=1) {

arch/powerpc/mm/imalloc.c

@@ -107,6 +107,7 @@ static int im_region_status(unsigned long v_addr, unsigned long size,
 		if (v_addr < (unsigned long) tmp->addr + tmp->size)
 			break;
 
+	*vm = NULL;
 	if (tmp) {
 		if (im_region_overlaps(v_addr, size, tmp))
 			return IM_REGION_OVERLAP;
@@ -127,7 +128,6 @@ static int im_region_status(unsigned long v_addr, unsigned long size,
 		}
 	}
 
-	*vm = NULL;
 	return IM_REGION_UNUSED;
 }
 

arch/powerpc/mm/init_32.c

@@ -188,6 +188,11 @@ void __init MMU_init(void)
 
 	if (ppc_md.progress)
 		ppc_md.progress("MMU:exit", 0x211);
+
+	/* From now on, btext is no longer BAT mapped if it was at all */
+#ifdef CONFIG_BOOTX_TEXT
+	btext_unmap();
+#endif
 }
 
 /* This is only called until mem_init is done. */

arch/powerpc/mm/mem.c

@@ -114,19 +114,18 @@ void online_page(struct page *page)
 	num_physpages++;
 }
 
-/*
- * This works only for the non-NUMA case.  Later, we'll need a lookup
- * to convert from real physical addresses to nid, that doesn't use
- * pfn_to_nid().
- */
 int __devinit add_memory(u64 start, u64 size)
 {
-	struct pglist_data *pgdata = NODE_DATA(0);
+	struct pglist_data *pgdata;
 	struct zone *zone;
+	int nid;
 	unsigned long start_pfn = start >> PAGE_SHIFT;
 	unsigned long nr_pages = size >> PAGE_SHIFT;
 
-	start += KERNELBASE;
+	nid = hot_add_scn_to_nid(start);
+	pgdata = NODE_DATA(nid);
+
+	start = __va(start);
 	create_section_mapping(start, start + size);
 
 	/* this should work for most non-highmem platforms */

arch/powerpc/mm/numa.c

@@ -37,6 +37,7 @@ EXPORT_SYMBOL(node_data);
 
 static bootmem_data_t __initdata plat_node_bdata[MAX_NUMNODES];
 static int min_common_depth;
+static int n_mem_addr_cells, n_mem_size_cells;
 
 /*
  * We need somewhere to store start/end/node for each region until we have
@@ -254,32 +255,20 @@ static int __init find_min_common_depth(void)
 	return depth;
 }
 
-static int __init get_mem_addr_cells(void)
+static void __init get_n_mem_cells(int *n_addr_cells, int *n_size_cells)
 {
 	struct device_node *memory = NULL;
-	int rc;
 
 	memory = of_find_node_by_type(memory, "memory");
 	if (!memory)
-		return 0; /* it won't matter */
+		panic("numa.c: No memory nodes found!");
 
-	rc = prom_n_addr_cells(memory);
-	return rc;
+	*n_addr_cells = prom_n_addr_cells(memory);
+	*n_size_cells = prom_n_size_cells(memory);
+	of_node_put(memory);
 }
 
-static int __init get_mem_size_cells(void)
-{
-	struct device_node *memory = NULL;
-	int rc;
-
-	memory = of_find_node_by_type(memory, "memory");
-	if (!memory)
-		return 0; /* it won't matter */
-	rc = prom_n_size_cells(memory);
-	return rc;
-}
-
-static unsigned long __init read_n_cells(int n, unsigned int **buf)
+static unsigned long __devinit read_n_cells(int n, unsigned int **buf)
 {
 	unsigned long result = 0;
 
@@ -386,7 +375,6 @@ static int __init parse_numa_properties(void)
 {
 	struct device_node *cpu = NULL;
 	struct device_node *memory = NULL;
-	int addr_cells, size_cells;
 	int max_domain;
 	unsigned long i;
 
@@ -425,8 +413,7 @@ static int __init parse_numa_properties(void)
 		}
 	}
 
-	addr_cells = get_mem_addr_cells();
-	size_cells = get_mem_size_cells();
+	get_n_mem_cells(&n_mem_addr_cells, &n_mem_size_cells);
 	memory = NULL;
 	while ((memory = of_find_node_by_type(memory, "memory")) != NULL) {
 		unsigned long start;
@@ -436,15 +423,21 @@ static int __init parse_numa_properties(void)
 		unsigned int *memcell_buf;
 		unsigned int len;
 
-		memcell_buf = (unsigned int *)get_property(memory, "reg", &len);
+		memcell_buf = (unsigned int *)get_property(memory,
+			"linux,usable-memory", &len);
+		if (!memcell_buf || len <= 0)
+			memcell_buf =
+				(unsigned int *)get_property(memory, "reg",
+					&len);
 		if (!memcell_buf || len <= 0)
 			continue;
 
-		ranges = memory->n_addrs;
+		/* ranges in cell */
+		ranges = (len >> 2) / (n_mem_addr_cells + n_mem_size_cells);
 new_range:
 		/* these are order-sensitive, and modify the buffer pointer */
-		start = read_n_cells(addr_cells, &memcell_buf);
-		size = read_n_cells(size_cells, &memcell_buf);
+		start = read_n_cells(n_mem_addr_cells, &memcell_buf);
+		size = read_n_cells(n_mem_size_cells, &memcell_buf);
 
 		numa_domain = of_node_numa_domain(memory);
 
@@ -497,7 +490,41 @@ static void __init setup_nonnuma(void)
 	node_set_online(0);
 }
 
-static void __init dump_numa_topology(void)
+void __init dump_numa_cpu_topology(void)
+{
+	unsigned int node;
+	unsigned int cpu, count;
+
+	if (min_common_depth == -1 || !numa_enabled)
+		return;
+
+	for_each_online_node(node) {
+		printk(KERN_INFO "Node %d CPUs:", node);
+
+		count = 0;
+		/*
+		 * If we used a CPU iterator here we would miss printing
+		 * the holes in the cpumap.
+		 */
+		for (cpu = 0; cpu < NR_CPUS; cpu++) {
+			if (cpu_isset(cpu, numa_cpumask_lookup_table[node])) {
+				if (count == 0)
+					printk(" %u", cpu);
+				++count;
+			} else {
+				if (count > 1)
+					printk("-%u", cpu - 1);
+				count = 0;
+			}
+		}
+
+		if (count > 1)
+			printk("-%u", NR_CPUS - 1);
+		printk("\n");
+	}
+}
+
+static void __init dump_numa_memory_topology(void)
 {
 	unsigned int node;
 	unsigned int count;
@@ -529,7 +556,6 @@ static void __init dump_numa_topology(void)
 			printk("-0x%lx", i);
 		printk("\n");
 	}
-	return;
 }
 
 /*
@@ -591,7 +617,7 @@ void __init do_init_bootmem(void)
 	if (parse_numa_properties())
 		setup_nonnuma();
 	else
-		dump_numa_topology();
+		dump_numa_memory_topology();
 
 	register_cpu_notifier(&ppc64_numa_nb);
 
@@ -730,3 +756,60 @@ static int __init early_numa(char *p)
 	return 0;
 }
 early_param("numa", early_numa);
+
+#ifdef CONFIG_MEMORY_HOTPLUG
+/*
+ * Find the node associated with a hot added memory section.  Section
+ * corresponds to a SPARSEMEM section, not an LMB.  It is assumed that
+ * sections are fully contained within a single LMB.
+ */
+int hot_add_scn_to_nid(unsigned long scn_addr)
+{
+	struct device_node *memory = NULL;
+	nodemask_t nodes;
+	int numa_domain = 0;
+
+	if (!numa_enabled || (min_common_depth < 0))
+		return numa_domain;
+
+	while ((memory = of_find_node_by_type(memory, "memory")) != NULL) {
+		unsigned long start, size;
+		int ranges;
+		unsigned int *memcell_buf;
+		unsigned int len;
+
+		memcell_buf = (unsigned int *)get_property(memory, "reg", &len);
+		if (!memcell_buf || len <= 0)
+			continue;
+
+		/* ranges in cell */
+		ranges = (len >> 2) / (n_mem_addr_cells + n_mem_size_cells);
+ha_new_range:
+		start = read_n_cells(n_mem_addr_cells, &memcell_buf);
+		size = read_n_cells(n_mem_size_cells, &memcell_buf);
+		numa_domain = of_node_numa_domain(memory);
+
+		/* Domains not present at boot default to 0 */
+		if (!node_online(numa_domain))
+			numa_domain = any_online_node(NODE_MASK_ALL);
+
+		if ((scn_addr >= start) && (scn_addr < (start + size))) {
+			of_node_put(memory);
+			goto got_numa_domain;
+		}
+
+		if (--ranges)		/* process all ranges in cell */
+			goto ha_new_range;
+	}
+	BUG();	/* section address should be found above */
+
+	/* Temporary code to ensure that returned node is not empty */
+got_numa_domain:
+	nodes_setall(nodes);
+	while (NODE_DATA(numa_domain)->node_spanned_pages == 0) {
+		node_clear(numa_domain, nodes);
+		numa_domain = any_online_node(nodes);
+	}
+	return numa_domain;
+}
+#endif /* CONFIG_MEMORY_HOTPLUG */

arch/powerpc/mm/slb.c

@@ -75,7 +75,7 @@ static void slb_flush_and_rebolt(void)
 	vflags = SLB_VSID_KERNEL | virtual_llp;
 
 	ksp_esid_data = mk_esid_data(get_paca()->kstack, 2);
-	if ((ksp_esid_data & ESID_MASK) == KERNELBASE)
+	if ((ksp_esid_data & ESID_MASK) == PAGE_OFFSET)
 		ksp_esid_data &= ~SLB_ESID_V;
 
 	/* We need to do this all in asm, so we're sure we don't touch
@@ -87,8 +87,8 @@ static void slb_flush_and_rebolt(void)
 		     /* Slot 2 - kernel stack */
 		     "slbmte	%2,%3\n"
 		     "isync"
-		     :: "r"(mk_vsid_data(VMALLOCBASE, vflags)),
-		        "r"(mk_esid_data(VMALLOCBASE, 1)),
+		     :: "r"(mk_vsid_data(VMALLOC_START, vflags)),
+		        "r"(mk_esid_data(VMALLOC_START, 1)),
 		        "r"(mk_vsid_data(ksp_esid_data, lflags)),
 		        "r"(ksp_esid_data)
 		     : "memory");
@@ -134,14 +134,14 @@ void switch_slb(struct task_struct *tsk, struct mm_struct *mm)
 	else
 		unmapped_base = TASK_UNMAPPED_BASE_USER64;
 
-	if (pc >= KERNELBASE)
+	if (is_kernel_addr(pc))
 		return;
 	slb_allocate(pc);
 
 	if (GET_ESID(pc) == GET_ESID(stack))
 		return;
 
-	if (stack >= KERNELBASE)
+	if (is_kernel_addr(stack))
 		return;
 	slb_allocate(stack);
 
@@ -149,7 +149,7 @@ void switch_slb(struct task_struct *tsk, struct mm_struct *mm)
 	    || (GET_ESID(stack) == GET_ESID(unmapped_base)))
 		return;
 
-	if (unmapped_base >= KERNELBASE)
+	if (is_kernel_addr(unmapped_base))
 		return;
 	slb_allocate(unmapped_base);
 }
@@ -213,10 +213,10 @@ void slb_initialize(void)
 	asm volatile("isync":::"memory");
 	asm volatile("slbmte  %0,%0"::"r" (0) : "memory");
 	asm volatile("isync; slbia; isync":::"memory");
-	create_slbe(KERNELBASE, lflags, 0);
+	create_slbe(PAGE_OFFSET, lflags, 0);
 
 	/* VMALLOC space has 4K pages always for now */
-	create_slbe(VMALLOCBASE, vflags, 1);
+	create_slbe(VMALLOC_START, vflags, 1);
 
 	/* We don't bolt the stack for the time being - we're in boot,
 	 * so the stack is in the bolted segment.  By the time it goes

arch/powerpc/mm/slb_low.S

@@ -37,9 +37,9 @@ _GLOBAL(slb_allocate_realmode)
 
 	srdi	r9,r3,60		/* get region */
 	srdi	r10,r3,28		/* get esid */
-	cmpldi	cr7,r9,0xc		/* cmp KERNELBASE for later use */
+	cmpldi	cr7,r9,0xc		/* cmp PAGE_OFFSET for later use */
 
-	/* r3 = address, r10 = esid, cr7 = <>KERNELBASE */
+	/* r3 = address, r10 = esid, cr7 = <> PAGE_OFFSET */
 	blt	cr7,0f			/* user or kernel? */
 
 	/* kernel address: proto-VSID = ESID */
@@ -166,7 +166,7 @@ _GLOBAL(slb_allocate_user)
 /*
  * Finish loading of an SLB entry and return
  *
- * r3 = EA, r10 = proto-VSID, r11 = flags, clobbers r9, cr7 = <>KERNELBASE
+ * r3 = EA, r10 = proto-VSID, r11 = flags, clobbers r9, cr7 = <> PAGE_OFFSET
  */
 slb_finish_load:
 	ASM_VSID_SCRAMBLE(r10,r9)

arch/powerpc/mm/stab.c

@@ -40,7 +40,7 @@ static int make_ste(unsigned long stab, unsigned long esid, unsigned long vsid)
 	unsigned long entry, group, old_esid, castout_entry, i;
 	unsigned int global_entry;
 	struct stab_entry *ste, *castout_ste;
-	unsigned long kernel_segment = (esid << SID_SHIFT) >= KERNELBASE;
+	unsigned long kernel_segment = (esid << SID_SHIFT) >= PAGE_OFFSET;
 
 	vsid_data = vsid << STE_VSID_SHIFT;
 	esid_data = esid << SID_SHIFT | STE_ESID_KP | STE_ESID_V;
@@ -83,7 +83,7 @@ static int make_ste(unsigned long stab, unsigned long esid, unsigned long vsid)
 		}
 
 		/* Dont cast out the first kernel segment */
-		if ((castout_ste->esid_data & ESID_MASK) != KERNELBASE)
+		if ((castout_ste->esid_data & ESID_MASK) != PAGE_OFFSET)
 			break;
 
 		castout_entry = (castout_entry + 1) & 0xf;
@@ -122,7 +122,7 @@ static int __ste_allocate(unsigned long ea, struct mm_struct *mm)
 	unsigned long offset;
 
 	/* Kernel or user address? */
-	if (ea >= KERNELBASE) {
+	if (is_kernel_addr(ea)) {
 		vsid = get_kernel_vsid(ea);
 	} else {
 		if ((ea >= TASK_SIZE_USER64) || (! mm))
@@ -133,7 +133,7 @@ static int __ste_allocate(unsigned long ea, struct mm_struct *mm)
 
 	stab_entry = make_ste(get_paca()->stab_addr, GET_ESID(ea), vsid);
 
-	if (ea < KERNELBASE) {
+	if (!is_kernel_addr(ea)) {
 		offset = __get_cpu_var(stab_cache_ptr);
 		if (offset < NR_STAB_CACHE_ENTRIES)
 			__get_cpu_var(stab_cache[offset++]) = stab_entry;
@@ -190,7 +190,7 @@ void switch_stab(struct task_struct *tsk, struct mm_struct *mm)
 		     entry++, ste++) {
 			unsigned long ea;
 			ea = ste->esid_data & ESID_MASK;
-			if (ea < KERNELBASE) {
+			if (!is_kernel_addr(ea)) {
 				ste->esid_data = 0;
 			}
 		}
@@ -251,7 +251,7 @@ void stabs_alloc(void)
 			panic("Unable to allocate segment table for CPU %d.\n",
 			      cpu);
 
-		newstab += KERNELBASE;
+		newstab = (unsigned long)__va(newstab);
 
 		memset((void *)newstab, 0, HW_PAGE_SIZE);
 
@@ -270,11 +270,11 @@ void stabs_alloc(void)
  */
 void stab_initialize(unsigned long stab)
 {
-	unsigned long vsid = get_kernel_vsid(KERNELBASE);
+	unsigned long vsid = get_kernel_vsid(PAGE_OFFSET);
 	unsigned long stabreal;
 
 	asm volatile("isync; slbia; isync":::"memory");
-	make_ste(stab, GET_ESID(KERNELBASE), vsid);
+	make_ste(stab, GET_ESID(PAGE_OFFSET), vsid);
 
 	/* Order update */
 	asm volatile("sync":::"memory");

arch/powerpc/mm/tlb_64.c

@@ -168,7 +168,7 @@ void hpte_update(struct mm_struct *mm, unsigned long addr,
 		batch->mm = mm;
 		batch->psize = psize;
 	}
-	if (addr < KERNELBASE) {
+	if (!is_kernel_addr(addr)) {
 		vsid = get_vsid(mm->context.id, addr);
 		WARN_ON(vsid == 0);
 	} else

arch/powerpc/oprofile/Makefile

@@ -9,3 +9,4 @@ DRIVER_OBJS := $(addprefix ../../../drivers/oprofile/, \
 oprofile-y := $(DRIVER_OBJS) common.o
 oprofile-$(CONFIG_PPC64) += op_model_rs64.o op_model_power4.o
 oprofile-$(CONFIG_FSL_BOOKE) += op_model_fsl_booke.o
+oprofile-$(CONFIG_PPC32) += op_model_7450.o

arch/powerpc/oprofile/common.c

@@ -14,9 +14,6 @@
  */
 
 #include <linux/oprofile.h>
-#ifndef __powerpc64__
-#include <linux/slab.h>
-#endif /* ! __powerpc64__ */
 #include <linux/init.h>
 #include <linux/smp.h>
 #include <linux/errno.h>
@@ -31,10 +28,6 @@ static struct op_powerpc_model *model;
 static struct op_counter_config ctr[OP_MAX_COUNTER];
 static struct op_system_config sys;
 
-#ifndef __powerpc64__
-static char *cpu_type;
-#endif /* ! __powerpc64__ */
-
 static void op_handle_interrupt(struct pt_regs *regs)
 {
 	model->handle_interrupt(regs, ctr);
@@ -53,14 +46,7 @@ static int op_powerpc_setup(void)
 	model->reg_setup(ctr, &sys, model->num_counters);
 
 	/* Configure the registers on all cpus.  */
-#ifdef __powerpc64__
 	on_each_cpu(model->cpu_setup, NULL, 0, 1);
-#else /* __powerpc64__ */
-#if 0
-	/* FIXME: Make multi-cpu work */
-	on_each_cpu(model->reg_setup, NULL, 0, 1);
-#endif
-#endif /* __powerpc64__ */
 
 	return 0;
 }
@@ -95,7 +81,7 @@ static int op_powerpc_create_files(struct super_block *sb, struct dentry *root)
 {
 	int i;
 
-#ifdef __powerpc64__
+#ifdef CONFIG_PPC64
 	/*
 	 * There is one mmcr0, mmcr1 and mmcra for setting the events for
 	 * all of the counters.
@@ -103,7 +89,7 @@ static int op_powerpc_create_files(struct super_block *sb, struct dentry *root)
 	oprofilefs_create_ulong(sb, root, "mmcr0", &sys.mmcr0);
 	oprofilefs_create_ulong(sb, root, "mmcr1", &sys.mmcr1);
 	oprofilefs_create_ulong(sb, root, "mmcra", &sys.mmcra);
-#endif /* __powerpc64__ */
+#endif
 
 	for (i = 0; i < model->num_counters; ++i) {
 		struct dentry *dir;
@@ -115,65 +101,68 @@ static int op_powerpc_create_files(struct super_block *sb, struct dentry *root)
 		oprofilefs_create_ulong(sb, dir, "enabled", &ctr[i].enabled);
 		oprofilefs_create_ulong(sb, dir, "event", &ctr[i].event);
 		oprofilefs_create_ulong(sb, dir, "count", &ctr[i].count);
-#ifdef __powerpc64__
+
 		/*
-		 * We dont support per counter user/kernel selection, but
-		 * we leave the entries because userspace expects them
+		 * Classic PowerPC doesn't support per-counter
+		 * control like this, but the options are
+		 * expected, so they remain.  For Freescale
+		 * Book-E style performance monitors, we do
+		 * support them.
 		 */
-#endif /* __powerpc64__ */
 		oprofilefs_create_ulong(sb, dir, "kernel", &ctr[i].kernel);
 		oprofilefs_create_ulong(sb, dir, "user", &ctr[i].user);
 
-#ifndef __powerpc64__
-		/* FIXME: Not sure if this is used */
-#endif /* ! __powerpc64__ */
 		oprofilefs_create_ulong(sb, dir, "unit_mask", &ctr[i].unit_mask);
 	}
 
 	oprofilefs_create_ulong(sb, root, "enable_kernel", &sys.enable_kernel);
 	oprofilefs_create_ulong(sb, root, "enable_user", &sys.enable_user);
-#ifdef __powerpc64__
+#ifdef CONFIG_PPC64
 	oprofilefs_create_ulong(sb, root, "backtrace_spinlocks",
 				&sys.backtrace_spinlocks);
-#endif /* __powerpc64__ */
+#endif
 
 	/* Default to tracing both kernel and user */
 	sys.enable_kernel = 1;
 	sys.enable_user = 1;
-#ifdef __powerpc64__
+#ifdef CONFIG_PPC64
 	/* Turn on backtracing through spinlocks by default */
 	sys.backtrace_spinlocks = 1;
-#endif /* __powerpc64__ */
+#endif
 
 	return 0;
 }
 
 int __init oprofile_arch_init(struct oprofile_operations *ops)
 {
-#ifndef __powerpc64__
-#ifdef CONFIG_FSL_BOOKE
-	model = &op_model_fsl_booke;
-#else
-	return -ENODEV;
-#endif
-
-	cpu_type = kmalloc(32, GFP_KERNEL);
-	if (NULL == cpu_type)
-		return -ENOMEM;
-
-	sprintf(cpu_type, "ppc/%s", cur_cpu_spec->cpu_name);
-
-	model->num_counters = cur_cpu_spec->num_pmcs;
-
-	ops->cpu_type = cpu_type;
-#else /* __powerpc64__ */
-	if (!cur_cpu_spec->oprofile_model || !cur_cpu_spec->oprofile_cpu_type)
+	if (!cur_cpu_spec->oprofile_cpu_type)
 		return -ENODEV;
-	model = cur_cpu_spec->oprofile_model;
+
+	switch (cur_cpu_spec->oprofile_type) {
+#ifdef CONFIG_PPC64
+		case RS64:
+			model = &op_model_rs64;
+			break;
+		case POWER4:
+			model = &op_model_power4;
+			break;
+#else
+		case G4:
+			model = &op_model_7450;
+			break;
+#endif
+#ifdef CONFIG_FSL_BOOKE
+		case BOOKE:
+			model = &op_model_fsl_booke;
+			break;
+#endif
+		default:
+			return -ENODEV;
+	}
+
 	model->num_counters = cur_cpu_spec->num_pmcs;
 
 	ops->cpu_type = cur_cpu_spec->oprofile_cpu_type;
-#endif /* __powerpc64__ */
 	ops->create_files = op_powerpc_create_files;
 	ops->setup = op_powerpc_setup;
 	ops->shutdown = op_powerpc_shutdown;
@@ -188,8 +177,4 @@ int __init oprofile_arch_init(struct oprofile_operations *ops)
 
 void oprofile_arch_exit(void)
 {
-#ifndef __powerpc64__
-	kfree(cpu_type);
-	cpu_type = NULL;
-#endif /* ! __powerpc64__ */
 }

arch/powerpc/oprofile/op_model_7450.c

@@ -0,0 +1,206 @@
+/*
+ * oprofile/op_model_7450.c
+ *
+ * Freescale 745x/744x oprofile support, based on fsl_booke support
+ * Copyright (C) 2004 Anton Blanchard <anton@au.ibm.com>, IBM
+ *
+ * Copyright (c) 2004 Freescale Semiconductor, Inc
+ *
+ * Author: Andy Fleming
+ * Maintainer: Kumar Gala <galak@kernel.crashing.org>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version
+ * 2 of the License, or (at your option) any later version.
+ */
+
+#include <linux/oprofile.h>
+#include <linux/init.h>
+#include <linux/smp.h>
+#include <asm/ptrace.h>
+#include <asm/system.h>
+#include <asm/processor.h>
+#include <asm/cputable.h>
+#include <asm/page.h>
+#include <asm/pmc.h>
+#include <asm/oprofile_impl.h>
+
+static unsigned long reset_value[OP_MAX_COUNTER];
+
+static int oprofile_running;
+static u32 mmcr0_val, mmcr1_val, mmcr2_val;
+
+#define MMCR0_PMC1_SHIFT	6
+#define MMCR0_PMC2_SHIFT	0
+#define MMCR1_PMC3_SHIFT	27
+#define MMCR1_PMC4_SHIFT	22
+#define MMCR1_PMC5_SHIFT	17
+#define MMCR1_PMC6_SHIFT	11
+
+#define mmcr0_event1(event) \
+	((event << MMCR0_PMC1_SHIFT) & MMCR0_PMC1SEL)
+#define mmcr0_event2(event) \
+	((event << MMCR0_PMC2_SHIFT) & MMCR0_PMC2SEL)
+
+#define mmcr1_event3(event) \
+	((event << MMCR1_PMC3_SHIFT) & MMCR1_PMC3SEL)
+#define mmcr1_event4(event) \
+	((event << MMCR1_PMC4_SHIFT) & MMCR1_PMC4SEL)
+#define mmcr1_event5(event) \
+	((event << MMCR1_PMC5_SHIFT) & MMCR1_PMC5SEL)
+#define mmcr1_event6(event) \
+	((event << MMCR1_PMC6_SHIFT) & MMCR1_PMC6SEL)
+
+#define MMCR0_INIT (MMCR0_FC | MMCR0_FCS | MMCR0_FCP | MMCR0_FCM1 | MMCR0_FCM0)
+
+/* Unfreezes the counters on this CPU, enables the interrupt,
+ * enables the counters to trigger the interrupt, and sets the
+ * counters to only count when the mark bit is not set.
+ */
+static void pmc_start_ctrs(void)
+{
+	u32 mmcr0 = mfspr(SPRN_MMCR0);
+
+	mmcr0 &= ~(MMCR0_FC | MMCR0_FCM0);
+	mmcr0 |= (MMCR0_FCECE | MMCR0_PMC1CE | MMCR0_PMCnCE | MMCR0_PMXE);
+
+	mtspr(SPRN_MMCR0, mmcr0);
+}
+
+/* Disables the counters on this CPU, and freezes them */
+static void pmc_stop_ctrs(void)
+{
+	u32 mmcr0 = mfspr(SPRN_MMCR0);
+
+	mmcr0 |= MMCR0_FC;
+	mmcr0 &= ~(MMCR0_FCECE | MMCR0_PMC1CE | MMCR0_PMCnCE | MMCR0_PMXE);
+
+	mtspr(SPRN_MMCR0, mmcr0);
+}
+
+/* Configures the counters on this CPU based on the global
+ * settings */
+static void fsl7450_cpu_setup(void *unused)
+{
+	/* freeze all counters */
+	pmc_stop_ctrs();
+
+	mtspr(SPRN_MMCR0, mmcr0_val);
+	mtspr(SPRN_MMCR1, mmcr1_val);
+	mtspr(SPRN_MMCR2, mmcr2_val);
+}
+
+#define NUM_CTRS 6
+
+/* Configures the global settings for the countes on all CPUs. */
+static void fsl7450_reg_setup(struct op_counter_config *ctr,
+			     struct op_system_config *sys,
+			     int num_ctrs)
+{
+	int i;
+
+	/* Our counters count up, and "count" refers to
+	 * how much before the next interrupt, and we interrupt
+	 * on overflow.  So we calculate the starting value
+	 * which will give us "count" until overflow.
+	 * Then we set the events on the enabled counters */
+	for (i = 0; i < NUM_CTRS; ++i)
+		reset_value[i] = 0x80000000UL - ctr[i].count;
+
+	/* Set events for Counters 1 & 2 */
+	mmcr0_val = MMCR0_INIT | mmcr0_event1(ctr[0].event)
+		| mmcr0_event2(ctr[1].event);
+
+	/* Setup user/kernel bits */
+	if (sys->enable_kernel)
+		mmcr0_val &= ~(MMCR0_FCS);
+
+	if (sys->enable_user)
+		mmcr0_val &= ~(MMCR0_FCP);
+
+	/* Set events for Counters 3-6 */
+	mmcr1_val = mmcr1_event3(ctr[2].event)
+		| mmcr1_event4(ctr[3].event)
+		| mmcr1_event5(ctr[4].event)
+		| mmcr1_event6(ctr[5].event);
+
+	mmcr2_val = 0;
+}
+
+/* Sets the counters on this CPU to the chosen values, and starts them */
+static void fsl7450_start(struct op_counter_config *ctr)
+{
+	int i;
+
+	mtmsr(mfmsr() | MSR_PMM);
+
+	for (i = 0; i < NUM_CTRS; ++i) {
+		if (ctr[i].enabled)
+			ctr_write(i, reset_value[i]);
+		else
+			ctr_write(i, 0);
+	}
+
+	/* Clear the freeze bit, and enable the interrupt.
+	 * The counters won't actually start until the rfi clears
+	 * the PMM bit */
+	pmc_start_ctrs();
+
+	oprofile_running = 1;
+}
+
+/* Stop the counters on this CPU */
+static void fsl7450_stop(void)
+{
+	/* freeze counters */
+	pmc_stop_ctrs();
+
+	oprofile_running = 0;
+
+	mb();
+}
+
+
+/* Handle the interrupt on this CPU, and log a sample for each
+ * event that triggered the interrupt */
+static void fsl7450_handle_interrupt(struct pt_regs *regs,
+				    struct op_counter_config *ctr)
+{
+	unsigned long pc;
+	int is_kernel;
+	int val;
+	int i;
+
+	/* set the PMM bit (see comment below) */
+	mtmsr(mfmsr() | MSR_PMM);
+
+	pc = mfspr(SPRN_SIAR);
+	is_kernel = (pc >= KERNELBASE);
+
+	for (i = 0; i < NUM_CTRS; ++i) {
+		val = ctr_read(i);
+		if (val < 0) {
+			if (oprofile_running && ctr[i].enabled) {
+				oprofile_add_pc(pc, is_kernel, i);
+				ctr_write(i, reset_value[i]);
+			} else {
+				ctr_write(i, 0);
+			}
+		}
+	}
+
+	/* The freeze bit was set by the interrupt. */
+	/* Clear the freeze bit, and reenable the interrupt.
+	 * The counters won't actually start until the rfi clears
+	 * the PMM bit */
+	pmc_start_ctrs();
+}
+
+struct op_powerpc_model op_model_7450= {
+	.reg_setup		= fsl7450_reg_setup,
+	.cpu_setup		= fsl7450_cpu_setup,
+	.start			= fsl7450_start,
+	.stop			= fsl7450_stop,
+	.handle_interrupt	= fsl7450_handle_interrupt,
+};

arch/powerpc/oprofile/op_model_power4.c

@@ -252,7 +252,7 @@ static unsigned long get_pc(struct pt_regs *regs)
 		return (unsigned long)__va(pc);
 
 	/* Not sure where we were */
-	if (pc < KERNELBASE)
+	if (!is_kernel_addr(pc))
 		/* function descriptor madness */
 		return *((unsigned long *)kernel_unknown_bucket);
 
@@ -264,7 +264,7 @@ static int get_kernel(unsigned long pc)
 	int is_kernel;
 
 	if (!mmcra_has_sihv) {
-		is_kernel = (pc >= KERNELBASE);
+		is_kernel = is_kernel_addr(pc);
 	} else {
 		unsigned long mmcra = mfspr(SPRN_MMCRA);
 		is_kernel = ((mmcra & MMCRA_SIPR) == 0);

arch/powerpc/oprofile/op_model_rs64.c

@@ -178,7 +178,6 @@ static void rs64_handle_interrupt(struct pt_regs *regs,
 	int val;
 	int i;
 	unsigned long pc = mfspr(SPRN_SIAR);
-	int is_kernel = (pc >= KERNELBASE);
 
 	/* set the PMM bit (see comment below) */
 	mtmsrd(mfmsr() | MSR_PMM);
@@ -187,7 +186,7 @@ static void rs64_handle_interrupt(struct pt_regs *regs,
 		val = ctr_read(i);
 		if (val < 0) {
 			if (ctr[i].enabled) {
-				oprofile_add_pc(pc, is_kernel, i);
+				oprofile_add_pc(pc, is_kernel_addr(pc), i);
 				ctr_write(i, reset_value[i]);
 			} else {
 				ctr_write(i, 0);

arch/powerpc/platforms/cell/Kconfig

@@ -0,0 +1,13 @@
+menu "Cell Broadband Engine options"
+	depends on PPC_CELL
+
+config SPU_FS
+	tristate "SPU file system"
+	default m
+	depends on PPC_CELL
+	help
+	  The SPU file system is used to access Synergistic Processing
+	  Units on machines implementing the Broadband Processor
+	  Architecture.
+
+endmenu

arch/powerpc/platforms/cell/Makefile

@@ -1,2 +1,10 @@
 obj-y			+= interrupt.o iommu.o setup.o spider-pic.o
+obj-y			+= pervasive.o
+
 obj-$(CONFIG_SMP)	+= smp.o
+obj-$(CONFIG_SPU_FS)	+= spufs/ spu-base.o
+
+spu-base-y		+= spu_base.o spu_priv1.o
+
+builtin-spufs-$(CONFIG_SPU_FS)	+= spu_syscalls.o
+obj-y			+= $(builtin-spufs-m)

arch/powerpc/platforms/cell/interrupt.c

@@ -23,6 +23,7 @@
 #include <linux/config.h>
 #include <linux/interrupt.h>
 #include <linux/irq.h>
+#include <linux/module.h>
 #include <linux/percpu.h>
 #include <linux/types.h>
 
@@ -55,6 +56,7 @@ struct iic_regs {
 
 struct iic {
 	struct iic_regs __iomem *regs;
+	u8 target_id;
 };
 
 static DEFINE_PER_CPU(struct iic, iic);
@@ -172,12 +174,11 @@ int iic_get_irq(struct pt_regs *regs)
 	return irq;
 }
 
-static struct iic_regs __iomem *find_iic(int cpu)
+static int setup_iic(int cpu, struct iic *iic)
 {
 	struct device_node *np;
 	int nodeid = cpu / 2;
 	unsigned long regs;
-	struct iic_regs __iomem *iic_regs;
 
 	for (np = of_find_node_by_type(NULL, "cpu");
 	     np;
@@ -188,20 +189,23 @@ static struct iic_regs __iomem *find_iic(int cpu)
 
 	if (!np) {
 		printk(KERN_WARNING "IIC: CPU %d not found\n", cpu);
-		iic_regs = NULL;
-	} else {
-		regs = *(long *)get_property(np, "iic", NULL);
-
-		/* hack until we have decided on the devtree info */
-		regs += 0x400;
-		if (cpu & 1)
-			regs += 0x20;
-
-		printk(KERN_DEBUG "IIC for CPU %d at %lx\n", cpu, regs);
-		iic_regs = __ioremap(regs, sizeof(struct iic_regs),
-						 _PAGE_NO_CACHE);
+		iic->regs = NULL;
+		iic->target_id = 0xff;
+		return -ENODEV;
 	}
-	return iic_regs;
+
+	regs = *(long *)get_property(np, "iic", NULL);
+
+	/* hack until we have decided on the devtree info */
+	regs += 0x400;
+	if (cpu & 1)
+		regs += 0x20;
+
+	printk(KERN_DEBUG "IIC for CPU %d at %lx\n", cpu, regs);
+	iic->regs = __ioremap(regs, sizeof(struct iic_regs),
+					 _PAGE_NO_CACHE);
+	iic->target_id = (nodeid << 4) + ((cpu & 1) ? 0xf : 0xe);
+	return 0;
 }
 
 #ifdef CONFIG_SMP
@@ -227,6 +231,12 @@ void iic_cause_IPI(int cpu, int mesg)
 	out_be64(&per_cpu(iic, cpu).regs->generate, (IIC_NUM_IPIS - 1 - mesg) << 4);
 }
 
+u8 iic_get_target_id(int cpu)
+{
+	return per_cpu(iic, cpu).target_id;
+}
+EXPORT_SYMBOL_GPL(iic_get_target_id);
+
 static irqreturn_t iic_ipi_action(int irq, void *dev_id, struct pt_regs *regs)
 {
 	smp_message_recv(iic_irq_to_ipi(irq), regs);
@@ -276,7 +286,7 @@ void iic_init_IRQ(void)
 	irq_offset = 0;
 	for_each_cpu(cpu) {
 		iic = &per_cpu(iic, cpu);
-		iic->regs = find_iic(cpu);
+		setup_iic(cpu, iic);
 		if (iic->regs)
 			out_be64(&iic->regs->prio, 0xff);
 	}

arch/powerpc/platforms/cell/interrupt.h

@@ -54,6 +54,7 @@ extern void iic_setup_cpu(void);
 extern void iic_local_enable(void);
 extern void iic_local_disable(void);
 
+extern u8 iic_get_target_id(int cpu);
 
 extern void spider_init_IRQ(void);
 extern int spider_get_irq(unsigned long int_pending);

arch/powerpc/platforms/cell/iommu.c

@@ -29,6 +29,8 @@
 #include <linux/bootmem.h>
 #include <linux/mm.h>
 #include <linux/dma-mapping.h>
+#include <linux/kernel.h>
+#include <linux/compiler.h>
 
 #include <asm/sections.h>
 #include <asm/iommu.h>
@@ -40,6 +42,7 @@
 #include <asm/abs_addr.h>
 #include <asm/system.h>
 #include <asm/ppc-pci.h>
+#include <asm/udbg.h>
 
 #include "iommu.h"
 
@@ -220,8 +223,6 @@ set_iopt_cache(void __iomem *base, unsigned long index,
 {
 	unsigned long __iomem *tags = base + IOC_PT_CACHE_DIR;
 	unsigned long __iomem *p = base + IOC_PT_CACHE_REG;
-	pr_debug("iopt %02lx was v%016lx/t%016lx, store v%016lx/t%016lx\n",
-		index, get_iopt_cache(base, index, &oldtag), oldtag, val, tag);
 
 	out_be64(p, val);
 	out_be64(&tags[index], tag);
@@ -248,66 +249,175 @@ set_iocmd_config(void __iomem *base)
 	out_be64(p, conf | IOCMD_CONF_TE);
 }
 
-/* FIXME: get these from the device tree */
-#define ioc_base	0x20000511000ull
-#define ioc_mmio_base	0x20000510000ull
-#define ioid		0x48a
-#define iopt_phys_offset (- 0x20000000) /* We have a 512MB offset from the SB */
-#define io_page_size	0x1000000
-
-static unsigned long map_iopt_entry(unsigned long address)
+static void enable_mapping(void __iomem *base, void __iomem *mmio_base)
 {
-	switch (address >> 20) {
-	case 0x600:
-		address = 0x24020000000ull; /* spider i/o */
-		break;
-	default:
-		address += iopt_phys_offset;
-		break;
-	}
-
-	return get_iopt_entry(address, ioid, IOPT_PROT_RW);
+	set_iocmd_config(base);
+	set_iost_origin(mmio_base);
 }
 
-static void iommu_bus_setup_null(struct pci_bus *b) { }
 static void iommu_dev_setup_null(struct pci_dev *d) { }
+static void iommu_bus_setup_null(struct pci_bus *b) { }
+
+struct cell_iommu {
+	unsigned long base;
+	unsigned long mmio_base;
+	void __iomem *mapped_base;
+	void __iomem *mapped_mmio_base;
+};
+
+static struct cell_iommu cell_iommus[NR_CPUS];
 
 /* initialize the iommu to support a simple linear mapping
  * for each DMA window used by any device. For now, we
  * happen to know that there is only one DMA window in use,
  * starting at iopt_phys_offset. */
-static void cell_map_iommu(void)
+static void cell_do_map_iommu(struct cell_iommu *iommu,
+			      unsigned int ioid,
+			      unsigned long map_start,
+			      unsigned long map_size)
 {
-	unsigned long address;
-	void __iomem *base;
+	unsigned long io_address, real_address;
+	void __iomem *ioc_base, *ioc_mmio_base;
 	ioste ioste;
 	unsigned long index;
 
-	base = __ioremap(ioc_base, 0x1000, _PAGE_NO_CACHE);
-	pr_debug("%lx mapped to %p\n", ioc_base, base);
-	set_iocmd_config(base);
-	iounmap(base);
+	/* we pretend the io page table was at a very high address */
+	const unsigned long fake_iopt = 0x10000000000ul;
+	const unsigned long io_page_size = 0x1000000; /* use 16M pages */
+	const unsigned long io_segment_size = 0x10000000; /* 256M */
 
-	base = __ioremap(ioc_mmio_base, 0x1000, _PAGE_NO_CACHE);
-	pr_debug("%lx mapped to %p\n", ioc_mmio_base, base);
+	ioc_base = iommu->mapped_base;
+	ioc_mmio_base = iommu->mapped_mmio_base;
 
-	set_iost_origin(base);
-
-	for (address = 0; address < 0x100000000ul; address += io_page_size) {
-		ioste = get_iost_entry(0x10000000000ul, address, io_page_size);
-		if ((address & 0xfffffff) == 0) /* segment start */
-			set_iost_cache(base, address >> 28, ioste);
-		index = get_ioc_hash_1way(ioste, address);
+	for (real_address = 0, io_address = 0;
+	     io_address <= map_start + map_size;
+	     real_address += io_page_size, io_address += io_page_size) {
+		ioste = get_iost_entry(fake_iopt, io_address, io_page_size);
+		if ((real_address % io_segment_size) == 0) /* segment start */
+			set_iost_cache(ioc_mmio_base,
+				       io_address >> 28, ioste);
+		index = get_ioc_hash_1way(ioste, io_address);
 		pr_debug("addr %08lx, index %02lx, ioste %016lx\n",
-					 address, index, ioste.val);
-		set_iopt_cache(base,
-			get_ioc_hash_1way(ioste, address),
-			get_ioc_tag(ioste, address),
-			map_iopt_entry(address));
+					 io_address, index, ioste.val);
+		set_iopt_cache(ioc_mmio_base,
+			get_ioc_hash_1way(ioste, io_address),
+			get_ioc_tag(ioste, io_address),
+			get_iopt_entry(real_address-map_start, ioid, IOPT_PROT_RW));
 	}
-	iounmap(base);
 }
 
+static void iommu_devnode_setup(struct device_node *d)
+{
+	unsigned int *ioid;
+	unsigned long *dma_window, map_start, map_size, token;
+	struct cell_iommu *iommu;
+
+	ioid = (unsigned int *)get_property(d, "ioid", NULL);
+	if (!ioid)
+		pr_debug("No ioid entry found !\n");
+
+	dma_window = (unsigned long *)get_property(d, "ibm,dma-window", NULL);
+	if (!dma_window)
+		pr_debug("No ibm,dma-window entry found !\n");
+
+	map_start = dma_window[1];
+	map_size = dma_window[2];
+	token = dma_window[0] >> 32;
+
+	iommu = &cell_iommus[token];
+
+	cell_do_map_iommu(iommu, *ioid, map_start, map_size);
+}
+
+static void iommu_bus_setup(struct pci_bus *b)
+{
+	struct device_node *d = (struct device_node *)b->sysdata;
+	iommu_devnode_setup(d);
+}
+
+
+static int cell_map_iommu_hardcoded(int num_nodes)
+{
+	struct cell_iommu *iommu = NULL;
+
+	pr_debug("%s(%d): Using hardcoded defaults\n", __FUNCTION__, __LINE__);
+
+	/* node 0 */
+	iommu = &cell_iommus[0];
+	iommu->mapped_base = __ioremap(0x20000511000, 0x1000, _PAGE_NO_CACHE);
+	iommu->mapped_mmio_base = __ioremap(0x20000510000, 0x1000, _PAGE_NO_CACHE);
+
+	enable_mapping(iommu->mapped_base, iommu->mapped_mmio_base);
+
+	cell_do_map_iommu(iommu, 0x048a,
+			  0x20000000ul,0x20000000ul);
+
+	if (num_nodes < 2)
+		return 0;
+
+	/* node 1 */
+	iommu = &cell_iommus[1];
+	iommu->mapped_base = __ioremap(0x30000511000, 0x1000, _PAGE_NO_CACHE);
+	iommu->mapped_mmio_base = __ioremap(0x30000510000, 0x1000, _PAGE_NO_CACHE);
+
+	enable_mapping(iommu->mapped_base, iommu->mapped_mmio_base);
+
+	cell_do_map_iommu(iommu, 0x048a,
+			  0x20000000,0x20000000ul);
+
+	return 0;
+}
+
+
+static int cell_map_iommu(void)
+{
+	unsigned int num_nodes = 0, *node_id;
+	unsigned long *base, *mmio_base;
+	struct device_node *dn;
+	struct cell_iommu *iommu = NULL;
+
+	/* determine number of nodes (=iommus) */
+	pr_debug("%s(%d): determining number of nodes...", __FUNCTION__, __LINE__);
+	for(dn = of_find_node_by_type(NULL, "cpu");
+	    dn;
+	    dn = of_find_node_by_type(dn, "cpu")) {
+		node_id = (unsigned int *)get_property(dn, "node-id", NULL);
+
+		if (num_nodes < *node_id)
+			num_nodes = *node_id;
+		}
+
+	num_nodes++;
+	pr_debug("%i found.\n", num_nodes);
+
+	/* map the iommu registers for each node */
+	pr_debug("%s(%d): Looping through nodes\n", __FUNCTION__, __LINE__);
+	for(dn = of_find_node_by_type(NULL, "cpu");
+	    dn;
+	    dn = of_find_node_by_type(dn, "cpu")) {
+
+		node_id = (unsigned int *)get_property(dn, "node-id", NULL);
+		base = (unsigned long *)get_property(dn, "ioc-cache", NULL);
+		mmio_base = (unsigned long *)get_property(dn, "ioc-translation", NULL);
+
+		if (!base || !mmio_base || !node_id)
+			return cell_map_iommu_hardcoded(num_nodes);
+
+		iommu = &cell_iommus[*node_id];
+		iommu->base = *base;
+		iommu->mmio_base = *mmio_base;
+
+		iommu->mapped_base = __ioremap(*base, 0x1000, _PAGE_NO_CACHE);
+		iommu->mapped_mmio_base = __ioremap(*mmio_base, 0x1000, _PAGE_NO_CACHE);
+
+		enable_mapping(iommu->mapped_base,
+			       iommu->mapped_mmio_base);
+
+		/* everything else will be done in iommu_bus_setup */
+	}
+
+	return 1;
+}
 
 static void *cell_alloc_coherent(struct device *hwdev, size_t size,
 			   dma_addr_t *dma_handle, gfp_t flag)
@@ -365,11 +475,28 @@ static int cell_dma_supported(struct device *dev, u64 mask)
 
 void cell_init_iommu(void)
 {
-	cell_map_iommu();
+	int setup_bus = 0;
 
-	/* Direct I/O, IOMMU off */
-	ppc_md.iommu_dev_setup = iommu_dev_setup_null;
-	ppc_md.iommu_bus_setup = iommu_bus_setup_null;
+	if (of_find_node_by_path("/mambo")) {
+		pr_info("Not using iommu on systemsim\n");
+	} else {
+
+		if (!(of_chosen &&
+		      get_property(of_chosen, "linux,iommu-off", NULL)))
+			setup_bus = cell_map_iommu();
+
+		if (setup_bus) {
+			pr_debug("%s: IOMMU mapping activated\n", __FUNCTION__);
+			ppc_md.iommu_dev_setup = iommu_dev_setup_null;
+			ppc_md.iommu_bus_setup = iommu_bus_setup;
+		} else {
+			pr_debug("%s: IOMMU mapping activated, "
+				 "no device action necessary\n", __FUNCTION__);
+			/* Direct I/O, IOMMU off */
+			ppc_md.iommu_dev_setup = iommu_dev_setup_null;
+			ppc_md.iommu_bus_setup = iommu_bus_setup_null;
+		}
+	}
 
 	pci_dma_ops.alloc_coherent = cell_alloc_coherent;
 	pci_dma_ops.free_coherent = cell_free_coherent;

arch/powerpc/platforms/cell/pervasive.c

@@ -0,0 +1,229 @@
+/*
+ * CBE Pervasive Monitor and Debug
+ *
+ * (C) Copyright IBM Corporation 2005
+ *
+ * Authors: Maximino Aguilar (maguilar@us.ibm.com)
+ *          Michael N. Day (mnday@us.ibm.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+#undef DEBUG
+
+#include <linux/config.h>
+#include <linux/interrupt.h>
+#include <linux/irq.h>
+#include <linux/percpu.h>
+#include <linux/types.h>
+#include <linux/kallsyms.h>
+
+#include <asm/io.h>
+#include <asm/machdep.h>
+#include <asm/prom.h>
+#include <asm/pgtable.h>
+#include <asm/reg.h>
+
+#include "pervasive.h"
+
+static DEFINE_SPINLOCK(cbe_pervasive_lock);
+struct cbe_pervasive {
+	struct pmd_regs __iomem *regs;
+	unsigned int thread;
+};
+
+/* can't use per_cpu from setup_arch */
+static struct cbe_pervasive cbe_pervasive[NR_CPUS];
+
+static void __init cbe_enable_pause_zero(void)
+{
+	unsigned long thread_switch_control;
+	unsigned long temp_register;
+	struct cbe_pervasive *p;
+	int thread;
+
+	spin_lock_irq(&cbe_pervasive_lock);
+	p = &cbe_pervasive[smp_processor_id()];
+
+	if (!cbe_pervasive->regs)
+		goto out;
+
+	pr_debug("Power Management: CPU %d\n", smp_processor_id());
+
+	 /* Enable Pause(0) control bit */
+	temp_register = in_be64(&p->regs->pm_control);
+
+	out_be64(&p->regs->pm_control,
+		 temp_register|PMD_PAUSE_ZERO_CONTROL);
+
+	/* Enable DEC and EE interrupt request */
+	thread_switch_control  = mfspr(SPRN_TSC_CELL);
+	thread_switch_control |= TSC_CELL_EE_ENABLE | TSC_CELL_EE_BOOST;
+
+	switch ((mfspr(SPRN_CTRLF) & CTRL_CT)) {
+	case CTRL_CT0:
+		thread_switch_control |= TSC_CELL_DEC_ENABLE_0;
+		thread = 0;
+		break;
+	case CTRL_CT1:
+		thread_switch_control |= TSC_CELL_DEC_ENABLE_1;
+		thread = 1;
+		break;
+	default:
+		printk(KERN_WARNING "%s: unknown configuration\n",
+			__FUNCTION__);
+		thread = -1;
+		break;
+	}
+
+	if (p->thread != thread)
+		printk(KERN_WARNING "%s: device tree inconsistant, "
+				     "cpu %i: %d/%d\n", __FUNCTION__,
+				     smp_processor_id(),
+				     p->thread, thread);
+
+	mtspr(SPRN_TSC_CELL, thread_switch_control);
+
+out:
+	spin_unlock_irq(&cbe_pervasive_lock);
+}
+
+static void cbe_idle(void)
+{
+	unsigned long ctrl;
+
+	cbe_enable_pause_zero();
+
+	while (1) {
+		if (!need_resched()) {
+			local_irq_disable();
+			while (!need_resched()) {
+				/* go into low thread priority */
+				HMT_low();
+
+				/*
+				 * atomically disable thread execution
+				 * and runlatch.
+				 * External and Decrementer exceptions
+				 * are still handled when the thread
+				 * is disabled but now enter in
+				 * cbe_system_reset_exception()
+				 */
+				ctrl = mfspr(SPRN_CTRLF);
+				ctrl &= ~(CTRL_RUNLATCH | CTRL_TE);
+				mtspr(SPRN_CTRLT, ctrl);
+			}
+			/* restore thread prio */
+			HMT_medium();
+			local_irq_enable();
+		}
+
+		/*
+		 * turn runlatch on again before scheduling the
+		 * process we just woke up
+		 */
+		ppc64_runlatch_on();
+
+		preempt_enable_no_resched();
+		schedule();
+		preempt_disable();
+	}
+}
+
+int cbe_system_reset_exception(struct pt_regs *regs)
+{
+	switch (regs->msr & SRR1_WAKEMASK) {
+	case SRR1_WAKEEE:
+		do_IRQ(regs);
+		break;
+	case SRR1_WAKEDEC:
+		timer_interrupt(regs);
+		break;
+	case SRR1_WAKEMT:
+		/* no action required */
+		break;
+	default:
+		/* do system reset */
+		return 0;
+	}
+	/* everything handled */
+	return 1;
+}
+
+static int __init cbe_find_pmd_mmio(int cpu, struct cbe_pervasive *p)
+{
+	struct device_node *node;
+	unsigned int *int_servers;
+	char *addr;
+	unsigned long real_address;
+	unsigned int size;
+
+	struct pmd_regs __iomem *pmd_mmio_area;
+	int hardid, thread;
+	int proplen;
+
+	pmd_mmio_area = NULL;
+	hardid = get_hard_smp_processor_id(cpu);
+	for (node = NULL; (node = of_find_node_by_type(node, "cpu"));) {
+		int_servers = (void *) get_property(node,
+				"ibm,ppc-interrupt-server#s", &proplen);
+		if (!int_servers) {
+			printk(KERN_WARNING "%s misses "
+				"ibm,ppc-interrupt-server#s property",
+				node->full_name);
+			continue;
+		}
+		for (thread = 0; thread < proplen / sizeof (int); thread++) {
+			if (hardid == int_servers[thread]) {
+				addr = get_property(node, "pervasive", NULL);
+				goto found;
+			}
+		}
+	}
+
+	printk(KERN_WARNING "%s: CPU %d not found\n", __FUNCTION__, cpu);
+	return -EINVAL;
+
+found:
+	real_address = *(unsigned long*) addr;
+	addr += sizeof (unsigned long);
+	size = *(unsigned int*) addr;
+
+	pr_debug("pervasive area for CPU %d at %lx, size %x\n",
+			cpu, real_address, size);
+	p->regs = __ioremap(real_address, size, _PAGE_NO_CACHE);
+	p->thread = thread;
+	return 0;
+}
+
+void __init cell_pervasive_init(void)
+{
+	struct cbe_pervasive *p;
+	int cpu;
+	int ret;
+
+	if (!cpu_has_feature(CPU_FTR_PAUSE_ZERO))
+		return;
+
+	for_each_cpu(cpu) {
+		p = &cbe_pervasive[cpu];
+		ret = cbe_find_pmd_mmio(cpu, p);
+		if (ret)
+			return;
+	}
+
+	ppc_md.idle_loop = cbe_idle;
+	ppc_md.system_reset_exception = cbe_system_reset_exception;
+}

arch/powerpc/platforms/cell/pervasive.h

@@ -0,0 +1,62 @@
+/*
+ * Cell Pervasive Monitor and Debug interface and HW structures
+ *
+ * (C) Copyright IBM Corporation 2005
+ *
+ * Authors: Maximino Aguilar (maguilar@us.ibm.com)
+ *          David J. Erb (djerb@us.ibm.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+
+#ifndef PERVASIVE_H
+#define PERVASIVE_H
+
+struct pmd_regs {
+	u8 pad_0x0000_0x0800[0x0800 - 0x0000];			/* 0x0000 */
+
+	/* Thermal Sensor Registers */
+	u64  ts_ctsr1;						/* 0x0800 */
+	u64  ts_ctsr2;						/* 0x0808 */
+	u64  ts_mtsr1;						/* 0x0810 */
+	u64  ts_mtsr2;						/* 0x0818 */
+	u64  ts_itr1;						/* 0x0820 */
+	u64  ts_itr2;						/* 0x0828 */
+	u64  ts_gitr;						/* 0x0830 */
+	u64  ts_isr;						/* 0x0838 */
+	u64  ts_imr;						/* 0x0840 */
+	u64  tm_cr1;						/* 0x0848 */
+	u64  tm_cr2;						/* 0x0850 */
+	u64  tm_simr;						/* 0x0858 */
+	u64  tm_tpr;						/* 0x0860 */
+	u64  tm_str1;						/* 0x0868 */
+	u64  tm_str2;						/* 0x0870 */
+	u64  tm_tsr;						/* 0x0878 */
+
+	/* Power Management */
+	u64  pm_control;					/* 0x0880 */
+#define PMD_PAUSE_ZERO_CONTROL		0x10000
+	u64  pm_status;						/* 0x0888 */
+
+	/* Time Base Register */
+	u64  tbr;						/* 0x0890 */
+
+	u8   pad_0x0898_0x1000 [0x1000 - 0x0898];		/* 0x0898 */
+};
+
+void __init cell_pervasive_init(void);
+
+#endif

arch/powerpc/platforms/cell/setup.c

@@ -33,6 +33,7 @@
 #include <asm/mmu.h>
 #include <asm/processor.h>
 #include <asm/io.h>
+#include <asm/kexec.h>
 #include <asm/pgtable.h>
 #include <asm/prom.h>
 #include <asm/rtas.h>
@@ -48,6 +49,7 @@
 
 #include "interrupt.h"
 #include "iommu.h"
+#include "pervasive.h"
 
 #ifdef DEBUG
 #define DBG(fmt...) udbg_printf(fmt)
@@ -67,6 +69,77 @@ void cell_show_cpuinfo(struct seq_file *m)
 	of_node_put(root);
 }
 
+#ifdef CONFIG_SPARSEMEM
+static int __init find_spu_node_id(struct device_node *spe)
+{
+	unsigned int *id;
+#ifdef CONFIG_NUMA
+	struct device_node *cpu;
+	cpu = spe->parent->parent;
+	id = (unsigned int *)get_property(cpu, "node-id", NULL);
+#else
+	id = NULL;
+#endif
+	return id ? *id : 0;
+}
+
+static void __init cell_spuprop_present(struct device_node *spe,
+				       const char *prop, int early)
+{
+	struct address_prop {
+		unsigned long address;
+		unsigned int len;
+	} __attribute__((packed)) *p;
+	int proplen;
+
+	unsigned long start_pfn, end_pfn, pfn;
+	int node_id;
+
+	p = (void*)get_property(spe, prop, &proplen);
+	WARN_ON(proplen != sizeof (*p));
+
+	node_id = find_spu_node_id(spe);
+
+	start_pfn = p->address >> PAGE_SHIFT;
+	end_pfn = (p->address + p->len + PAGE_SIZE - 1) >> PAGE_SHIFT;
+
+	/* We need to call memory_present *before* the call to sparse_init,
+	   but we can initialize the page structs only *after* that call.
+	   Thus, we're being called twice. */
+	if (early)
+		memory_present(node_id, start_pfn, end_pfn);
+	else {
+		/* As the pages backing SPU LS and I/O are outside the range
+		   of regular memory, their page structs were not initialized
+		   by free_area_init. Do it here instead. */
+		for (pfn = start_pfn; pfn < end_pfn; pfn++) {
+			struct page *page = pfn_to_page(pfn);
+			set_page_links(page, ZONE_DMA, node_id, pfn);
+			set_page_count(page, 1);
+			reset_page_mapcount(page);
+			SetPageReserved(page);
+			INIT_LIST_HEAD(&page->lru);
+		}
+	}
+}
+
+static void __init cell_spumem_init(int early)
+{
+	struct device_node *node;
+	for (node = of_find_node_by_type(NULL, "spe");
+			node; node = of_find_node_by_type(node, "spe")) {
+		cell_spuprop_present(node, "local-store", early);
+		cell_spuprop_present(node, "problem", early);
+		cell_spuprop_present(node, "priv1", early);
+		cell_spuprop_present(node, "priv2", early);
+	}
+}
+#else
+static void __init cell_spumem_init(int early)
+{
+}
+#endif
+
 static void cell_progress(char *s, unsigned short hex)
 {
 	printk("*** %04x : %s\n", hex, s ? s : "");
@@ -93,11 +166,14 @@ static void __init cell_setup_arch(void)
 	init_pci_config_tokens();
 	find_and_init_phbs();
 	spider_init_IRQ();
+	cell_pervasive_init();
 #ifdef CONFIG_DUMMY_CONSOLE
 	conswitchp = &dummy_con;
 #endif
 
 	mmio_nvram_init();
+
+	cell_spumem_init(0);
 }
 
 /*
@@ -113,6 +189,8 @@ static void __init cell_init_early(void)
 
 	ppc64_interrupt_controller = IC_CELL_PIC;
 
+	cell_spumem_init(1);
+
 	DBG(" <- cell_init_early()\n");
 }
 
@@ -125,6 +203,15 @@ static int __init cell_probe(int platform)
 	return 1;
 }
 
+/*
+ * Cell has no legacy IO; anything calling this function has to
+ * fail or bad things will happen
+ */
+static int cell_check_legacy_ioport(unsigned int baseport)
+{
+	return -ENODEV;
+}
+
 struct machdep_calls __initdata cell_md = {
 	.probe			= cell_probe,
 	.setup_arch		= cell_setup_arch,
@@ -137,5 +224,11 @@ struct machdep_calls __initdata cell_md = {
 	.get_rtc_time		= rtas_get_rtc_time,
 	.set_rtc_time		= rtas_set_rtc_time,
 	.calibrate_decr		= generic_calibrate_decr,
+	.check_legacy_ioport	= cell_check_legacy_ioport,
 	.progress		= cell_progress,
+#ifdef CONFIG_KEXEC
+	.machine_kexec		= default_machine_kexec,
+	.machine_kexec_prepare	= default_machine_kexec_prepare,
+	.machine_crash_shutdown	= default_machine_crash_shutdown,
+#endif
 };

arch/powerpc/platforms/cell/spu_base.c

@@ -0,0 +1,711 @@
+/*
+ * Low-level SPU handling
+ *
+ * (C) Copyright IBM Deutschland Entwicklung GmbH 2005
+ *
+ * Author: Arnd Bergmann <arndb@de.ibm.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+#undef DEBUG
+
+#include <linux/interrupt.h>
+#include <linux/list.h>
+#include <linux/module.h>
+#include <linux/poll.h>
+#include <linux/ptrace.h>
+#include <linux/slab.h>
+#include <linux/wait.h>
+
+#include <asm/io.h>
+#include <asm/prom.h>
+#include <asm/semaphore.h>
+#include <asm/spu.h>
+#include <asm/mmu_context.h>
+
+#include "interrupt.h"
+
+static int __spu_trap_invalid_dma(struct spu *spu)
+{
+	pr_debug("%s\n", __FUNCTION__);
+	force_sig(SIGBUS, /* info, */ current);
+	return 0;
+}
+
+static int __spu_trap_dma_align(struct spu *spu)
+{
+	pr_debug("%s\n", __FUNCTION__);
+	force_sig(SIGBUS, /* info, */ current);
+	return 0;
+}
+
+static int __spu_trap_error(struct spu *spu)
+{
+	pr_debug("%s\n", __FUNCTION__);
+	force_sig(SIGILL, /* info, */ current);
+	return 0;
+}
+
+static void spu_restart_dma(struct spu *spu)
+{
+	struct spu_priv2 __iomem *priv2 = spu->priv2;
+
+	if (!test_bit(SPU_CONTEXT_SWITCH_PENDING, &spu->flags))
+		out_be64(&priv2->mfc_control_RW, MFC_CNTL_RESTART_DMA_COMMAND);
+}
+
+static int __spu_trap_data_seg(struct spu *spu, unsigned long ea)
+{
+	struct spu_priv2 __iomem *priv2 = spu->priv2;
+	struct mm_struct *mm = spu->mm;
+	u64 esid, vsid;
+
+	pr_debug("%s\n", __FUNCTION__);
+
+	if (test_bit(SPU_CONTEXT_SWITCH_ACTIVE, &spu->flags)) {
+		/* SLBs are pre-loaded for context switch, so
+		 * we should never get here!
+		 */
+		printk("%s: invalid access during switch!\n", __func__);
+		return 1;
+	}
+	if (!mm || (REGION_ID(ea) != USER_REGION_ID)) {
+		/* Future: support kernel segments so that drivers
+		 * can use SPUs.
+		 */
+		pr_debug("invalid region access at %016lx\n", ea);
+		return 1;
+	}
+
+	esid = (ea & ESID_MASK) | SLB_ESID_V;
+	vsid = (get_vsid(mm->context.id, ea) << SLB_VSID_SHIFT) | SLB_VSID_USER;
+	if (in_hugepage_area(mm->context, ea))
+		vsid |= SLB_VSID_L;
+
+	out_be64(&priv2->slb_index_W, spu->slb_replace);
+	out_be64(&priv2->slb_vsid_RW, vsid);
+	out_be64(&priv2->slb_esid_RW, esid);
+
+	spu->slb_replace++;
+	if (spu->slb_replace >= 8)
+		spu->slb_replace = 0;
+
+	spu_restart_dma(spu);
+
+	return 0;
+}
+
+extern int hash_page(unsigned long ea, unsigned long access, unsigned long trap); //XXX
+static int __spu_trap_data_map(struct spu *spu, unsigned long ea, u64 dsisr)
+{
+	pr_debug("%s\n", __FUNCTION__);
+
+	/* Handle kernel space hash faults immediately.
+	   User hash faults need to be deferred to process context. */
+	if ((dsisr & MFC_DSISR_PTE_NOT_FOUND)
+	    && REGION_ID(ea) != USER_REGION_ID
+	    && hash_page(ea, _PAGE_PRESENT, 0x300) == 0) {
+		spu_restart_dma(spu);
+		return 0;
+	}
+
+	if (test_bit(SPU_CONTEXT_SWITCH_ACTIVE, &spu->flags)) {
+		printk("%s: invalid access during switch!\n", __func__);
+		return 1;
+	}
+
+	spu->dar = ea;
+	spu->dsisr = dsisr;
+	mb();
+	if (spu->stop_callback)
+		spu->stop_callback(spu);
+	return 0;
+}
+
+static int __spu_trap_mailbox(struct spu *spu)
+{
+	if (spu->ibox_callback)
+		spu->ibox_callback(spu);
+
+	/* atomically disable SPU mailbox interrupts */
+	spin_lock(&spu->register_lock);
+	spu_int_mask_and(spu, 2, ~0x1);
+	spin_unlock(&spu->register_lock);
+	return 0;
+}
+
+static int __spu_trap_stop(struct spu *spu)
+{
+	pr_debug("%s\n", __FUNCTION__);
+	spu->stop_code = in_be32(&spu->problem->spu_status_R);
+	if (spu->stop_callback)
+		spu->stop_callback(spu);
+	return 0;
+}
+
+static int __spu_trap_halt(struct spu *spu)
+{
+	pr_debug("%s\n", __FUNCTION__);
+	spu->stop_code = in_be32(&spu->problem->spu_status_R);
+	if (spu->stop_callback)
+		spu->stop_callback(spu);
+	return 0;
+}
+
+static int __spu_trap_tag_group(struct spu *spu)
+{
+	pr_debug("%s\n", __FUNCTION__);
+	/* wake_up(&spu->dma_wq); */
+	return 0;
+}
+
+static int __spu_trap_spubox(struct spu *spu)
+{
+	if (spu->wbox_callback)
+		spu->wbox_callback(spu);
+
+	/* atomically disable SPU mailbox interrupts */
+	spin_lock(&spu->register_lock);
+	spu_int_mask_and(spu, 2, ~0x10);
+	spin_unlock(&spu->register_lock);
+	return 0;
+}
+
+static irqreturn_t
+spu_irq_class_0(int irq, void *data, struct pt_regs *regs)
+{
+	struct spu *spu;
+
+	spu = data;
+	spu->class_0_pending = 1;
+	if (spu->stop_callback)
+		spu->stop_callback(spu);
+
+	return IRQ_HANDLED;
+}
+
+int
+spu_irq_class_0_bottom(struct spu *spu)
+{
+	unsigned long stat, mask;
+
+	spu->class_0_pending = 0;
+
+	mask = spu_int_mask_get(spu, 0);
+	stat = spu_int_stat_get(spu, 0);
+
+	stat &= mask;
+
+	if (stat & 1) /* invalid MFC DMA */
+		__spu_trap_invalid_dma(spu);
+
+	if (stat & 2) /* invalid DMA alignment */
+		__spu_trap_dma_align(spu);
+
+	if (stat & 4) /* error on SPU */
+		__spu_trap_error(spu);
+
+	spu_int_stat_clear(spu, 0, stat);
+
+	return (stat & 0x7) ? -EIO : 0;
+}
+EXPORT_SYMBOL_GPL(spu_irq_class_0_bottom);
+
+static irqreturn_t
+spu_irq_class_1(int irq, void *data, struct pt_regs *regs)
+{
+	struct spu *spu;
+	unsigned long stat, mask, dar, dsisr;
+
+	spu = data;
+
+	/* atomically read & clear class1 status. */
+	spin_lock(&spu->register_lock);
+	mask  = spu_int_mask_get(spu, 1);
+	stat  = spu_int_stat_get(spu, 1) & mask;
+	dar   = spu_mfc_dar_get(spu);
+	dsisr = spu_mfc_dsisr_get(spu);
+	if (stat & 2) /* mapping fault */
+		spu_mfc_dsisr_set(spu, 0ul);
+	spu_int_stat_clear(spu, 1, stat);
+	spin_unlock(&spu->register_lock);
+
+	if (stat & 1) /* segment fault */
+		__spu_trap_data_seg(spu, dar);
+
+	if (stat & 2) { /* mapping fault */
+		__spu_trap_data_map(spu, dar, dsisr);
+	}
+
+	if (stat & 4) /* ls compare & suspend on get */
+		;
+
+	if (stat & 8) /* ls compare & suspend on put */
+		;
+
+	return stat ? IRQ_HANDLED : IRQ_NONE;
+}
+EXPORT_SYMBOL_GPL(spu_irq_class_1_bottom);
+
+static irqreturn_t
+spu_irq_class_2(int irq, void *data, struct pt_regs *regs)
+{
+	struct spu *spu;
+	unsigned long stat;
+	unsigned long mask;
+
+	spu = data;
+	stat = spu_int_stat_get(spu, 2);
+	mask = spu_int_mask_get(spu, 2);
+
+	pr_debug("class 2 interrupt %d, %lx, %lx\n", irq, stat, mask);
+
+	stat &= mask;
+
+	if (stat & 1)  /* PPC core mailbox */
+		__spu_trap_mailbox(spu);
+
+	if (stat & 2) /* SPU stop-and-signal */
+		__spu_trap_stop(spu);
+
+	if (stat & 4) /* SPU halted */
+		__spu_trap_halt(spu);
+
+	if (stat & 8) /* DMA tag group complete */
+		__spu_trap_tag_group(spu);
+
+	if (stat & 0x10) /* SPU mailbox threshold */
+		__spu_trap_spubox(spu);
+
+	spu_int_stat_clear(spu, 2, stat);
+	return stat ? IRQ_HANDLED : IRQ_NONE;
+}
+
+static int
+spu_request_irqs(struct spu *spu)
+{
+	int ret;
+	int irq_base;
+
+	irq_base = IIC_NODE_STRIDE * spu->node + IIC_SPE_OFFSET;
+
+	snprintf(spu->irq_c0, sizeof (spu->irq_c0), "spe%02d.0", spu->number);
+	ret = request_irq(irq_base + spu->isrc,
+		 spu_irq_class_0, 0, spu->irq_c0, spu);
+	if (ret)
+		goto out;
+
+	snprintf(spu->irq_c1, sizeof (spu->irq_c1), "spe%02d.1", spu->number);
+	ret = request_irq(irq_base + IIC_CLASS_STRIDE + spu->isrc,
+		 spu_irq_class_1, 0, spu->irq_c1, spu);
+	if (ret)
+		goto out1;
+
+	snprintf(spu->irq_c2, sizeof (spu->irq_c2), "spe%02d.2", spu->number);
+	ret = request_irq(irq_base + 2*IIC_CLASS_STRIDE + spu->isrc,
+		 spu_irq_class_2, 0, spu->irq_c2, spu);
+	if (ret)
+		goto out2;
+	goto out;
+
+out2:
+	free_irq(irq_base + IIC_CLASS_STRIDE + spu->isrc, spu);
+out1:
+	free_irq(irq_base + spu->isrc, spu);
+out:
+	return ret;
+}
+
+static void
+spu_free_irqs(struct spu *spu)
+{
+	int irq_base;
+
+	irq_base = IIC_NODE_STRIDE * spu->node + IIC_SPE_OFFSET;
+
+	free_irq(irq_base + spu->isrc, spu);
+	free_irq(irq_base + IIC_CLASS_STRIDE + spu->isrc, spu);
+	free_irq(irq_base + 2*IIC_CLASS_STRIDE + spu->isrc, spu);
+}
+
+static LIST_HEAD(spu_list);
+static DECLARE_MUTEX(spu_mutex);
+
+static void spu_init_channels(struct spu *spu)
+{
+	static const struct {
+		 unsigned channel;
+		 unsigned count;
+	} zero_list[] = {
+		{ 0x00, 1, }, { 0x01, 1, }, { 0x03, 1, }, { 0x04, 1, },
+		{ 0x18, 1, }, { 0x19, 1, }, { 0x1b, 1, }, { 0x1d, 1, },
+	}, count_list[] = {
+		{ 0x00, 0, }, { 0x03, 0, }, { 0x04, 0, }, { 0x15, 16, },
+		{ 0x17, 1, }, { 0x18, 0, }, { 0x19, 0, }, { 0x1b, 0, },
+		{ 0x1c, 1, }, { 0x1d, 0, }, { 0x1e, 1, },
+	};
+	struct spu_priv2 __iomem *priv2;
+	int i;
+
+	priv2 = spu->priv2;
+
+	/* initialize all channel data to zero */
+	for (i = 0; i < ARRAY_SIZE(zero_list); i++) {
+		int count;
+
+		out_be64(&priv2->spu_chnlcntptr_RW, zero_list[i].channel);
+		for (count = 0; count < zero_list[i].count; count++)
+			out_be64(&priv2->spu_chnldata_RW, 0);
+	}
+
+	/* initialize channel counts to meaningful values */
+	for (i = 0; i < ARRAY_SIZE(count_list); i++) {
+		out_be64(&priv2->spu_chnlcntptr_RW, count_list[i].channel);
+		out_be64(&priv2->spu_chnlcnt_RW, count_list[i].count);
+	}
+}
+
+struct spu *spu_alloc(void)
+{
+	struct spu *spu;
+
+	down(&spu_mutex);
+	if (!list_empty(&spu_list)) {
+		spu = list_entry(spu_list.next, struct spu, list);
+		list_del_init(&spu->list);
+		pr_debug("Got SPU %x %d\n", spu->isrc, spu->number);
+	} else {
+		pr_debug("No SPU left\n");
+		spu = NULL;
+	}
+	up(&spu_mutex);
+
+	if (spu)
+		spu_init_channels(spu);
+
+	return spu;
+}
+EXPORT_SYMBOL_GPL(spu_alloc);
+
+void spu_free(struct spu *spu)
+{
+	down(&spu_mutex);
+	list_add_tail(&spu->list, &spu_list);
+	up(&spu_mutex);
+}
+EXPORT_SYMBOL_GPL(spu_free);
+
+static int spu_handle_mm_fault(struct spu *spu)
+{
+	struct mm_struct *mm = spu->mm;
+	struct vm_area_struct *vma;
+	u64 ea, dsisr, is_write;
+	int ret;
+
+	ea = spu->dar;
+	dsisr = spu->dsisr;
+#if 0
+	if (!IS_VALID_EA(ea)) {
+		return -EFAULT;
+	}
+#endif /* XXX */
+	if (mm == NULL) {
+		return -EFAULT;
+	}
+	if (mm->pgd == NULL) {
+		return -EFAULT;
+	}
+
+	down_read(&mm->mmap_sem);
+	vma = find_vma(mm, ea);
+	if (!vma)
+		goto bad_area;
+	if (vma->vm_start <= ea)
+		goto good_area;
+	if (!(vma->vm_flags & VM_GROWSDOWN))
+		goto bad_area;
+#if 0
+	if (expand_stack(vma, ea))
+		goto bad_area;
+#endif /* XXX */
+good_area:
+	is_write = dsisr & MFC_DSISR_ACCESS_PUT;
+	if (is_write) {
+		if (!(vma->vm_flags & VM_WRITE))
+			goto bad_area;
+	} else {
+		if (dsisr & MFC_DSISR_ACCESS_DENIED)
+			goto bad_area;
+		if (!(vma->vm_flags & (VM_READ | VM_EXEC)))
+			goto bad_area;
+	}
+	ret = 0;
+	switch (handle_mm_fault(mm, vma, ea, is_write)) {
+	case VM_FAULT_MINOR:
+		current->min_flt++;
+		break;
+	case VM_FAULT_MAJOR:
+		current->maj_flt++;
+		break;
+	case VM_FAULT_SIGBUS:
+		ret = -EFAULT;
+		goto bad_area;
+	case VM_FAULT_OOM:
+		ret = -ENOMEM;
+		goto bad_area;
+	default:
+		BUG();
+	}
+	up_read(&mm->mmap_sem);
+	return ret;
+
+bad_area:
+	up_read(&mm->mmap_sem);
+	return -EFAULT;
+}
+
+int spu_irq_class_1_bottom(struct spu *spu)
+{
+	u64 ea, dsisr, access, error = 0UL;
+	int ret = 0;
+
+	ea = spu->dar;
+	dsisr = spu->dsisr;
+	if (dsisr & MFC_DSISR_PTE_NOT_FOUND) {
+		access = (_PAGE_PRESENT | _PAGE_USER);
+		access |= (dsisr & MFC_DSISR_ACCESS_PUT) ? _PAGE_RW : 0UL;
+		if (hash_page(ea, access, 0x300) != 0)
+			error |= CLASS1_ENABLE_STORAGE_FAULT_INTR;
+	}
+	if ((error & CLASS1_ENABLE_STORAGE_FAULT_INTR) ||
+	    (dsisr & MFC_DSISR_ACCESS_DENIED)) {
+		if ((ret = spu_handle_mm_fault(spu)) != 0)
+			error |= CLASS1_ENABLE_STORAGE_FAULT_INTR;
+		else
+			error &= ~CLASS1_ENABLE_STORAGE_FAULT_INTR;
+	}
+	spu->dar = 0UL;
+	spu->dsisr = 0UL;
+	if (!error) {
+		spu_restart_dma(spu);
+	} else {
+		__spu_trap_invalid_dma(spu);
+	}
+	return ret;
+}
+
+void spu_irq_setaffinity(struct spu *spu, int cpu)
+{
+	u64 target = iic_get_target_id(cpu);
+	u64 route = target << 48 | target << 32 | target << 16;
+	spu_int_route_set(spu, route);
+}
+EXPORT_SYMBOL_GPL(spu_irq_setaffinity);
+
+static void __iomem * __init map_spe_prop(struct device_node *n,
+						 const char *name)
+{
+	struct address_prop {
+		unsigned long address;
+		unsigned int len;
+	} __attribute__((packed)) *prop;
+
+	void *p;
+	int proplen;
+
+	p = get_property(n, name, &proplen);
+	if (proplen != sizeof (struct address_prop))
+		return NULL;
+
+	prop = p;
+
+	return ioremap(prop->address, prop->len);
+}
+
+static void spu_unmap(struct spu *spu)
+{
+	iounmap(spu->priv2);
+	iounmap(spu->priv1);
+	iounmap(spu->problem);
+	iounmap((u8 __iomem *)spu->local_store);
+}
+
+static int __init spu_map_device(struct spu *spu, struct device_node *spe)
+{
+	char *prop;
+	int ret;
+
+	ret = -ENODEV;
+	prop = get_property(spe, "isrc", NULL);
+	if (!prop)
+		goto out;
+	spu->isrc = *(unsigned int *)prop;
+
+	spu->name = get_property(spe, "name", NULL);
+	if (!spu->name)
+		goto out;
+
+	prop = get_property(spe, "local-store", NULL);
+	if (!prop)
+		goto out;
+	spu->local_store_phys = *(unsigned long *)prop;
+
+	/* we use local store as ram, not io memory */
+	spu->local_store = (void __force *)map_spe_prop(spe, "local-store");
+	if (!spu->local_store)
+		goto out;
+
+	spu->problem= map_spe_prop(spe, "problem");
+	if (!spu->problem)
+		goto out_unmap;
+
+	spu->priv1= map_spe_prop(spe, "priv1");
+	/* priv1 is not available on a hypervisor */
+
+	spu->priv2= map_spe_prop(spe, "priv2");
+	if (!spu->priv2)
+		goto out_unmap;
+	ret = 0;
+	goto out;
+
+out_unmap:
+	spu_unmap(spu);
+out:
+	return ret;
+}
+
+static int __init find_spu_node_id(struct device_node *spe)
+{
+	unsigned int *id;
+	struct device_node *cpu;
+
+	cpu = spe->parent->parent;
+	id = (unsigned int *)get_property(cpu, "node-id", NULL);
+
+	return id ? *id : 0;
+}
+
+static int __init create_spu(struct device_node *spe)
+{
+	struct spu *spu;
+	int ret;
+	static int number;
+
+	ret = -ENOMEM;
+	spu = kmalloc(sizeof (*spu), GFP_KERNEL);
+	if (!spu)
+		goto out;
+
+	ret = spu_map_device(spu, spe);
+	if (ret)
+		goto out_free;
+
+	spu->node = find_spu_node_id(spe);
+	spu->stop_code = 0;
+	spu->slb_replace = 0;
+	spu->mm = NULL;
+	spu->ctx = NULL;
+	spu->rq = NULL;
+	spu->pid = 0;
+	spu->class_0_pending = 0;
+	spu->flags = 0UL;
+	spu->dar = 0UL;
+	spu->dsisr = 0UL;
+	spin_lock_init(&spu->register_lock);
+
+	spu_mfc_sdr_set(spu, mfspr(SPRN_SDR1));
+	spu_mfc_sr1_set(spu, 0x33);
+
+	spu->ibox_callback = NULL;
+	spu->wbox_callback = NULL;
+	spu->stop_callback = NULL;
+
+	down(&spu_mutex);
+	spu->number = number++;
+	ret = spu_request_irqs(spu);
+	if (ret)
+		goto out_unmap;
+
+	list_add(&spu->list, &spu_list);
+	up(&spu_mutex);
+
+	pr_debug(KERN_DEBUG "Using SPE %s %02x %p %p %p %p %d\n",
+		spu->name, spu->isrc, spu->local_store,
+		spu->problem, spu->priv1, spu->priv2, spu->number);
+	goto out;
+
+out_unmap:
+	up(&spu_mutex);
+	spu_unmap(spu);
+out_free:
+	kfree(spu);
+out:
+	return ret;
+}
+
+static void destroy_spu(struct spu *spu)
+{
+	list_del_init(&spu->list);
+
+	spu_free_irqs(spu);
+	spu_unmap(spu);
+	kfree(spu);
+}
+
+static void cleanup_spu_base(void)
+{
+	struct spu *spu, *tmp;
+	down(&spu_mutex);
+	list_for_each_entry_safe(spu, tmp, &spu_list, list)
+		destroy_spu(spu);
+	up(&spu_mutex);
+}
+module_exit(cleanup_spu_base);
+
+static int __init init_spu_base(void)
+{
+	struct device_node *node;
+	int ret;
+
+	ret = -ENODEV;
+	for (node = of_find_node_by_type(NULL, "spe");
+			node; node = of_find_node_by_type(node, "spe")) {
+		ret = create_spu(node);
+		if (ret) {
+			printk(KERN_WARNING "%s: Error initializing %s\n",
+				__FUNCTION__, node->name);
+			cleanup_spu_base();
+			break;
+		}
+	}
+	/* in some old firmware versions, the spe is called 'spc', so we
+	   look for that as well */
+	for (node = of_find_node_by_type(NULL, "spc");
+			node; node = of_find_node_by_type(node, "spc")) {
+		ret = create_spu(node);
+		if (ret) {
+			printk(KERN_WARNING "%s: Error initializing %s\n",
+				__FUNCTION__, node->name);
+			cleanup_spu_base();
+			break;
+		}
+	}
+	return ret;
+}
+module_init(init_spu_base);
+
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Arnd Bergmann <arndb@de.ibm.com>");

arch/powerpc/platforms/cell/spu_priv1.c

@@ -0,0 +1,133 @@
+/*
+ * access to SPU privileged registers
+ */
+#include <linux/module.h>
+
+#include <asm/io.h>
+#include <asm/spu.h>
+
+void spu_int_mask_and(struct spu *spu, int class, u64 mask)
+{
+	u64 old_mask;
+
+	old_mask = in_be64(&spu->priv1->int_mask_RW[class]);
+	out_be64(&spu->priv1->int_mask_RW[class], old_mask & mask);
+}
+EXPORT_SYMBOL_GPL(spu_int_mask_and);
+
+void spu_int_mask_or(struct spu *spu, int class, u64 mask)
+{
+	u64 old_mask;
+
+	old_mask = in_be64(&spu->priv1->int_mask_RW[class]);
+	out_be64(&spu->priv1->int_mask_RW[class], old_mask | mask);
+}
+EXPORT_SYMBOL_GPL(spu_int_mask_or);
+
+void spu_int_mask_set(struct spu *spu, int class, u64 mask)
+{
+	out_be64(&spu->priv1->int_mask_RW[class], mask);
+}
+EXPORT_SYMBOL_GPL(spu_int_mask_set);
+
+u64 spu_int_mask_get(struct spu *spu, int class)
+{
+	return in_be64(&spu->priv1->int_mask_RW[class]);
+}
+EXPORT_SYMBOL_GPL(spu_int_mask_get);
+
+void spu_int_stat_clear(struct spu *spu, int class, u64 stat)
+{
+	out_be64(&spu->priv1->int_stat_RW[class], stat);
+}
+EXPORT_SYMBOL_GPL(spu_int_stat_clear);
+
+u64 spu_int_stat_get(struct spu *spu, int class)
+{
+	return in_be64(&spu->priv1->int_stat_RW[class]);
+}
+EXPORT_SYMBOL_GPL(spu_int_stat_get);
+
+void spu_int_route_set(struct spu *spu, u64 route)
+{
+	out_be64(&spu->priv1->int_route_RW, route);
+}
+EXPORT_SYMBOL_GPL(spu_int_route_set);
+
+u64 spu_mfc_dar_get(struct spu *spu)
+{
+	return in_be64(&spu->priv1->mfc_dar_RW);
+}
+EXPORT_SYMBOL_GPL(spu_mfc_dar_get);
+
+u64 spu_mfc_dsisr_get(struct spu *spu)
+{
+	return in_be64(&spu->priv1->mfc_dsisr_RW);
+}
+EXPORT_SYMBOL_GPL(spu_mfc_dsisr_get);
+
+void spu_mfc_dsisr_set(struct spu *spu, u64 dsisr)
+{
+	out_be64(&spu->priv1->mfc_dsisr_RW, dsisr);
+}
+EXPORT_SYMBOL_GPL(spu_mfc_dsisr_set);
+
+void spu_mfc_sdr_set(struct spu *spu, u64 sdr)
+{
+	out_be64(&spu->priv1->mfc_sdr_RW, sdr);
+}
+EXPORT_SYMBOL_GPL(spu_mfc_sdr_set);
+
+void spu_mfc_sr1_set(struct spu *spu, u64 sr1)
+{
+	out_be64(&spu->priv1->mfc_sr1_RW, sr1);
+}
+EXPORT_SYMBOL_GPL(spu_mfc_sr1_set);
+
+u64 spu_mfc_sr1_get(struct spu *spu)
+{
+	return in_be64(&spu->priv1->mfc_sr1_RW);
+}
+EXPORT_SYMBOL_GPL(spu_mfc_sr1_get);
+
+void spu_mfc_tclass_id_set(struct spu *spu, u64 tclass_id)
+{
+	out_be64(&spu->priv1->mfc_tclass_id_RW, tclass_id);
+}
+EXPORT_SYMBOL_GPL(spu_mfc_tclass_id_set);
+
+u64 spu_mfc_tclass_id_get(struct spu *spu)
+{
+	return in_be64(&spu->priv1->mfc_tclass_id_RW);
+}
+EXPORT_SYMBOL_GPL(spu_mfc_tclass_id_get);
+
+void spu_tlb_invalidate(struct spu *spu)
+{
+	out_be64(&spu->priv1->tlb_invalidate_entry_W, 0ul);
+}
+EXPORT_SYMBOL_GPL(spu_tlb_invalidate);
+
+void spu_resource_allocation_groupID_set(struct spu *spu, u64 id)
+{
+	out_be64(&spu->priv1->resource_allocation_groupID_RW, id);
+}
+EXPORT_SYMBOL_GPL(spu_resource_allocation_groupID_set);
+
+u64 spu_resource_allocation_groupID_get(struct spu *spu)
+{
+	return in_be64(&spu->priv1->resource_allocation_groupID_RW);
+}
+EXPORT_SYMBOL_GPL(spu_resource_allocation_groupID_get);
+
+void spu_resource_allocation_enable_set(struct spu *spu, u64 enable)
+{
+	out_be64(&spu->priv1->resource_allocation_enable_RW, enable);
+}
+EXPORT_SYMBOL_GPL(spu_resource_allocation_enable_set);
+
+u64 spu_resource_allocation_enable_get(struct spu *spu)
+{
+	return in_be64(&spu->priv1->resource_allocation_enable_RW);
+}
+EXPORT_SYMBOL_GPL(spu_resource_allocation_enable_get);

arch/powerpc/platforms/cell/spu_syscalls.c

@@ -0,0 +1,88 @@
+/*
+ * SPU file system -- system call stubs
+ *
+ * (C) Copyright IBM Deutschland Entwicklung GmbH 2005
+ *
+ * Author: Arnd Bergmann <arndb@de.ibm.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+#include <linux/file.h>
+#include <linux/module.h>
+#include <linux/syscalls.h>
+
+#include <asm/spu.h>
+
+struct spufs_calls spufs_calls = {
+	.owner = NULL,
+};
+
+/* These stub syscalls are needed to have the actual implementation
+ * within a loadable module. When spufs is built into the kernel,
+ * this file is not used and the syscalls directly enter the fs code */
+
+asmlinkage long sys_spu_create(const char __user *name,
+		unsigned int flags, mode_t mode)
+{
+	long ret;
+	struct module *owner = spufs_calls.owner;
+
+	ret = -ENOSYS;
+	if (owner && try_module_get(owner)) {
+		ret = spufs_calls.create_thread(name, flags, mode);
+		module_put(owner);
+	}
+	return ret;
+}
+
+asmlinkage long sys_spu_run(int fd, __u32 __user *unpc, __u32 __user *ustatus)
+{
+	long ret;
+	struct file *filp;
+	int fput_needed;
+	struct module *owner = spufs_calls.owner;
+
+	ret = -ENOSYS;
+	if (owner && try_module_get(owner)) {
+		ret = -EBADF;
+		filp = fget_light(fd, &fput_needed);
+		if (filp) {
+			ret = spufs_calls.spu_run(filp, unpc, ustatus);
+			fput_light(filp, fput_needed);
+		}
+		module_put(owner);
+	}
+	return ret;
+}
+
+int register_spu_syscalls(struct spufs_calls *calls)
+{
+	if (spufs_calls.owner)
+		return -EBUSY;
+
+	spufs_calls.create_thread = calls->create_thread;
+	spufs_calls.spu_run = calls->spu_run;
+	smp_mb();
+	spufs_calls.owner = calls->owner;
+	return 0;
+}
+EXPORT_SYMBOL_GPL(register_spu_syscalls);
+
+void unregister_spu_syscalls(struct spufs_calls *calls)
+{
+	BUG_ON(spufs_calls.owner != calls->owner);
+	spufs_calls.owner = NULL;
+}
+EXPORT_SYMBOL_GPL(unregister_spu_syscalls);

arch/powerpc/platforms/cell/spufs/Makefile

@@ -0,0 +1,54 @@
+obj-$(CONFIG_SPU_FS) += spufs.o
+spufs-y += inode.o file.o context.o switch.o syscalls.o
+spufs-y += sched.o backing_ops.o hw_ops.o run.o
+
+# Rules to build switch.o with the help of SPU tool chain
+SPU_CROSS	:= spu-
+SPU_CC		:= $(SPU_CROSS)gcc
+SPU_AS		:= $(SPU_CROSS)gcc
+SPU_LD		:= $(SPU_CROSS)ld
+SPU_OBJCOPY	:= $(SPU_CROSS)objcopy
+SPU_CFLAGS	:= -O2 -Wall -I$(srctree)/include -I$(objtree)/include2
+SPU_AFLAGS	:= -c -D__ASSEMBLY__ -I$(srctree)/include -I$(objtree)/include2
+SPU_LDFLAGS	:= -N -Ttext=0x0
+
+$(obj)/switch.o: $(obj)/spu_save_dump.h $(obj)/spu_restore_dump.h
+
+# Compile SPU files
+      cmd_spu_cc = $(SPU_CC) $(SPU_CFLAGS) -c -o $@ $<
+quiet_cmd_spu_cc = SPU_CC  $@
+$(obj)/spu_%.o: $(src)/spu_%.c
+	$(call if_changed,spu_cc)
+
+# Assemble SPU files
+      cmd_spu_as = $(SPU_AS) $(SPU_AFLAGS) -o $@ $<
+quiet_cmd_spu_as = SPU_AS  $@
+$(obj)/spu_%.o: $(src)/spu_%.S
+	$(call if_changed,spu_as)
+
+# Link SPU Executables
+      cmd_spu_ld = $(SPU_LD) $(SPU_LDFLAGS) -o $@ $^
+quiet_cmd_spu_ld = SPU_LD  $@
+$(obj)/spu_%: $(obj)/spu_%_crt0.o $(obj)/spu_%.o
+	$(call if_changed,spu_ld)
+
+# Copy into binary format
+      cmd_spu_objcopy = $(SPU_OBJCOPY) -O binary $< $@
+quiet_cmd_spu_objcopy = OBJCOPY $@
+$(obj)/spu_%.bin: $(src)/spu_%
+	$(call if_changed,spu_objcopy)
+
+# create C code from ELF executable
+cmd_hexdump   = ( \
+		echo "/*" ; \
+		echo " * $*_dump.h: Copyright (C) 2005 IBM." ; \
+		echo " * Hex-dump auto generated from $*.c." ; \
+		echo " * Do not edit!" ; \
+		echo " */" ; \
+		echo "static unsigned int $*_code[] __page_aligned = {" ; \
+		hexdump -v -e '"0x" 4/1 "%02x" "," "\n"' $< ; \
+		echo "};" ; \
+		) > $@
+quiet_cmd_hexdump = HEXDUMP $@
+$(obj)/%_dump.h: $(obj)/%.bin
+	$(call if_changed,hexdump)

arch/powerpc/platforms/cell/spufs/backing_ops.c

@@ -0,0 +1,308 @@
+/* backing_ops.c - query/set operations on saved SPU context.
+ *
+ * Copyright (C) IBM 2005
+ * Author: Mark Nutter <mnutter@us.ibm.com>
+ *
+ * These register operations allow SPUFS to operate on saved
+ * SPU contexts rather than hardware.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+#include <linux/config.h>
+#include <linux/module.h>
+#include <linux/errno.h>
+#include <linux/sched.h>
+#include <linux/kernel.h>
+#include <linux/mm.h>
+#include <linux/vmalloc.h>
+#include <linux/smp.h>
+#include <linux/smp_lock.h>
+#include <linux/stddef.h>
+#include <linux/unistd.h>
+#include <linux/poll.h>
+
+#include <asm/io.h>
+#include <asm/spu.h>
+#include <asm/spu_csa.h>
+#include <asm/mmu_context.h>
+#include "spufs.h"
+
+/*
+ * Reads/writes to various problem and priv2 registers require
+ * state changes, i.e.  generate SPU events, modify channel
+ * counts, etc.
+ */
+
+static void gen_spu_event(struct spu_context *ctx, u32 event)
+{
+	u64 ch0_cnt;
+	u64 ch0_data;
+	u64 ch1_data;
+
+	ch0_cnt = ctx->csa.spu_chnlcnt_RW[0];
+	ch0_data = ctx->csa.spu_chnldata_RW[0];
+	ch1_data = ctx->csa.spu_chnldata_RW[1];
+	ctx->csa.spu_chnldata_RW[0] |= event;
+	if ((ch0_cnt == 0) && !(ch0_data & event) && (ch1_data & event)) {
+		ctx->csa.spu_chnlcnt_RW[0] = 1;
+	}
+}
+
+static int spu_backing_mbox_read(struct spu_context *ctx, u32 * data)
+{
+	u32 mbox_stat;
+	int ret = 0;
+
+	spin_lock(&ctx->csa.register_lock);
+	mbox_stat = ctx->csa.prob.mb_stat_R;
+	if (mbox_stat & 0x0000ff) {
+		/* Read the first available word.
+		 * Implementation note: the depth
+		 * of pu_mb_R is currently 1.
+		 */
+		*data = ctx->csa.prob.pu_mb_R;
+		ctx->csa.prob.mb_stat_R &= ~(0x0000ff);
+		ctx->csa.spu_chnlcnt_RW[28] = 1;
+		gen_spu_event(ctx, MFC_PU_MAILBOX_AVAILABLE_EVENT);
+		ret = 4;
+	}
+	spin_unlock(&ctx->csa.register_lock);
+	return ret;
+}
+
+static u32 spu_backing_mbox_stat_read(struct spu_context *ctx)
+{
+	return ctx->csa.prob.mb_stat_R;
+}
+
+static unsigned int spu_backing_mbox_stat_poll(struct spu_context *ctx,
+					  unsigned int events)
+{
+	int ret;
+	u32 stat;
+
+	ret = 0;
+	spin_lock_irq(&ctx->csa.register_lock);
+	stat = ctx->csa.prob.mb_stat_R;
+
+	/* if the requested event is there, return the poll
+	   mask, otherwise enable the interrupt to get notified,
+	   but first mark any pending interrupts as done so
+	   we don't get woken up unnecessarily */
+
+	if (events & (POLLIN | POLLRDNORM)) {
+		if (stat & 0xff0000)
+			ret |= POLLIN | POLLRDNORM;
+		else {
+			ctx->csa.priv1.int_stat_class0_RW &= ~0x1;
+			ctx->csa.priv1.int_mask_class2_RW |= 0x1;
+		}
+	}
+	if (events & (POLLOUT | POLLWRNORM)) {
+		if (stat & 0x00ff00)
+			ret = POLLOUT | POLLWRNORM;
+		else {
+			ctx->csa.priv1.int_stat_class0_RW &= ~0x10;
+			ctx->csa.priv1.int_mask_class2_RW |= 0x10;
+		}
+	}
+	spin_unlock_irq(&ctx->csa.register_lock);
+	return ret;
+}
+
+static int spu_backing_ibox_read(struct spu_context *ctx, u32 * data)
+{
+	int ret;
+
+	spin_lock(&ctx->csa.register_lock);
+	if (ctx->csa.prob.mb_stat_R & 0xff0000) {
+		/* Read the first available word.
+		 * Implementation note: the depth
+		 * of puint_mb_R is currently 1.
+		 */
+		*data = ctx->csa.priv2.puint_mb_R;
+		ctx->csa.prob.mb_stat_R &= ~(0xff0000);
+		ctx->csa.spu_chnlcnt_RW[30] = 1;
+		gen_spu_event(ctx, MFC_PU_INT_MAILBOX_AVAILABLE_EVENT);
+		ret = 4;
+	} else {
+		/* make sure we get woken up by the interrupt */
+		ctx->csa.priv1.int_mask_class2_RW |= 0x1UL;
+		ret = 0;
+	}
+	spin_unlock(&ctx->csa.register_lock);
+	return ret;
+}
+
+static int spu_backing_wbox_write(struct spu_context *ctx, u32 data)
+{
+	int ret;
+
+	spin_lock(&ctx->csa.register_lock);
+	if ((ctx->csa.prob.mb_stat_R) & 0x00ff00) {
+		int slot = ctx->csa.spu_chnlcnt_RW[29];
+		int avail = (ctx->csa.prob.mb_stat_R & 0x00ff00) >> 8;
+
+		/* We have space to write wbox_data.
+		 * Implementation note: the depth
+		 * of spu_mb_W is currently 4.
+		 */
+		BUG_ON(avail != (4 - slot));
+		ctx->csa.spu_mailbox_data[slot] = data;
+		ctx->csa.spu_chnlcnt_RW[29] = ++slot;
+		ctx->csa.prob.mb_stat_R = (((4 - slot) & 0xff) << 8);
+		gen_spu_event(ctx, MFC_SPU_MAILBOX_WRITTEN_EVENT);
+		ret = 4;
+	} else {
+		/* make sure we get woken up by the interrupt when space
+		   becomes available */
+		ctx->csa.priv1.int_mask_class2_RW |= 0x10;
+		ret = 0;
+	}
+	spin_unlock(&ctx->csa.register_lock);
+	return ret;
+}
+
+static u32 spu_backing_signal1_read(struct spu_context *ctx)
+{
+	return ctx->csa.spu_chnldata_RW[3];
+}
+
+static void spu_backing_signal1_write(struct spu_context *ctx, u32 data)
+{
+	spin_lock(&ctx->csa.register_lock);
+	if (ctx->csa.priv2.spu_cfg_RW & 0x1)
+		ctx->csa.spu_chnldata_RW[3] |= data;
+	else
+		ctx->csa.spu_chnldata_RW[3] = data;
+	ctx->csa.spu_chnlcnt_RW[3] = 1;
+	gen_spu_event(ctx, MFC_SIGNAL_1_EVENT);
+	spin_unlock(&ctx->csa.register_lock);
+}
+
+static u32 spu_backing_signal2_read(struct spu_context *ctx)
+{
+	return ctx->csa.spu_chnldata_RW[4];
+}
+
+static void spu_backing_signal2_write(struct spu_context *ctx, u32 data)
+{
+	spin_lock(&ctx->csa.register_lock);
+	if (ctx->csa.priv2.spu_cfg_RW & 0x2)
+		ctx->csa.spu_chnldata_RW[4] |= data;
+	else
+		ctx->csa.spu_chnldata_RW[4] = data;
+	ctx->csa.spu_chnlcnt_RW[4] = 1;
+	gen_spu_event(ctx, MFC_SIGNAL_2_EVENT);
+	spin_unlock(&ctx->csa.register_lock);
+}
+
+static void spu_backing_signal1_type_set(struct spu_context *ctx, u64 val)
+{
+	u64 tmp;
+
+	spin_lock(&ctx->csa.register_lock);
+	tmp = ctx->csa.priv2.spu_cfg_RW;
+	if (val)
+		tmp |= 1;
+	else
+		tmp &= ~1;
+	ctx->csa.priv2.spu_cfg_RW = tmp;
+	spin_unlock(&ctx->csa.register_lock);
+}
+
+static u64 spu_backing_signal1_type_get(struct spu_context *ctx)
+{
+	return ((ctx->csa.priv2.spu_cfg_RW & 1) != 0);
+}
+
+static void spu_backing_signal2_type_set(struct spu_context *ctx, u64 val)
+{
+	u64 tmp;
+
+	spin_lock(&ctx->csa.register_lock);
+	tmp = ctx->csa.priv2.spu_cfg_RW;
+	if (val)
+		tmp |= 2;
+	else
+		tmp &= ~2;
+	ctx->csa.priv2.spu_cfg_RW = tmp;
+	spin_unlock(&ctx->csa.register_lock);
+}
+
+static u64 spu_backing_signal2_type_get(struct spu_context *ctx)
+{
+	return ((ctx->csa.priv2.spu_cfg_RW & 2) != 0);
+}
+
+static u32 spu_backing_npc_read(struct spu_context *ctx)
+{
+	return ctx->csa.prob.spu_npc_RW;
+}
+
+static void spu_backing_npc_write(struct spu_context *ctx, u32 val)
+{
+	ctx->csa.prob.spu_npc_RW = val;
+}
+
+static u32 spu_backing_status_read(struct spu_context *ctx)
+{
+	return ctx->csa.prob.spu_status_R;
+}
+
+static char *spu_backing_get_ls(struct spu_context *ctx)
+{
+	return ctx->csa.lscsa->ls;
+}
+
+static void spu_backing_runcntl_write(struct spu_context *ctx, u32 val)
+{
+	spin_lock(&ctx->csa.register_lock);
+	ctx->csa.prob.spu_runcntl_RW = val;
+	if (val & SPU_RUNCNTL_RUNNABLE) {
+		ctx->csa.prob.spu_status_R |= SPU_STATUS_RUNNING;
+	} else {
+		ctx->csa.prob.spu_status_R &= ~SPU_STATUS_RUNNING;
+	}
+	spin_unlock(&ctx->csa.register_lock);
+}
+
+static void spu_backing_runcntl_stop(struct spu_context *ctx)
+{
+	spu_backing_runcntl_write(ctx, SPU_RUNCNTL_STOP);
+}
+
+struct spu_context_ops spu_backing_ops = {
+	.mbox_read = spu_backing_mbox_read,
+	.mbox_stat_read = spu_backing_mbox_stat_read,
+	.mbox_stat_poll = spu_backing_mbox_stat_poll,
+	.ibox_read = spu_backing_ibox_read,
+	.wbox_write = spu_backing_wbox_write,
+	.signal1_read = spu_backing_signal1_read,
+	.signal1_write = spu_backing_signal1_write,
+	.signal2_read = spu_backing_signal2_read,
+	.signal2_write = spu_backing_signal2_write,
+	.signal1_type_set = spu_backing_signal1_type_set,
+	.signal1_type_get = spu_backing_signal1_type_get,
+	.signal2_type_set = spu_backing_signal2_type_set,
+	.signal2_type_get = spu_backing_signal2_type_get,
+	.npc_read = spu_backing_npc_read,
+	.npc_write = spu_backing_npc_write,
+	.status_read = spu_backing_status_read,
+	.get_ls = spu_backing_get_ls,
+	.runcntl_write = spu_backing_runcntl_write,
+	.runcntl_stop = spu_backing_runcntl_stop,
+};

arch/powerpc/platforms/cell/spufs/context.c

@@ -0,0 +1,167 @@
+/*
+ * SPU file system -- SPU context management
+ *
+ * (C) Copyright IBM Deutschland Entwicklung GmbH 2005
+ *
+ * Author: Arnd Bergmann <arndb@de.ibm.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+#include <linux/fs.h>
+#include <linux/mm.h>
+#include <linux/slab.h>
+#include <asm/spu.h>
+#include <asm/spu_csa.h>
+#include "spufs.h"
+
+struct spu_context *alloc_spu_context(struct address_space *local_store)
+{
+	struct spu_context *ctx;
+	ctx = kmalloc(sizeof *ctx, GFP_KERNEL);
+	if (!ctx)
+		goto out;
+	/* Binding to physical processor deferred
+	 * until spu_activate().
+	 */
+	spu_init_csa(&ctx->csa);
+	if (!ctx->csa.lscsa) {
+		goto out_free;
+	}
+	spin_lock_init(&ctx->mmio_lock);
+	kref_init(&ctx->kref);
+	init_rwsem(&ctx->state_sema);
+	init_MUTEX(&ctx->run_sema);
+	init_waitqueue_head(&ctx->ibox_wq);
+	init_waitqueue_head(&ctx->wbox_wq);
+	init_waitqueue_head(&ctx->stop_wq);
+	ctx->ibox_fasync = NULL;
+	ctx->wbox_fasync = NULL;
+	ctx->state = SPU_STATE_SAVED;
+	ctx->local_store = local_store;
+	ctx->spu = NULL;
+	ctx->ops = &spu_backing_ops;
+	ctx->owner = get_task_mm(current);
+	goto out;
+out_free:
+	kfree(ctx);
+	ctx = NULL;
+out:
+	return ctx;
+}
+
+void destroy_spu_context(struct kref *kref)
+{
+	struct spu_context *ctx;
+	ctx = container_of(kref, struct spu_context, kref);
+	down_write(&ctx->state_sema);
+	spu_deactivate(ctx);
+	ctx->ibox_fasync = NULL;
+	ctx->wbox_fasync = NULL;
+	up_write(&ctx->state_sema);
+	spu_fini_csa(&ctx->csa);
+	kfree(ctx);
+}
+
+struct spu_context * get_spu_context(struct spu_context *ctx)
+{
+	kref_get(&ctx->kref);
+	return ctx;
+}
+
+int put_spu_context(struct spu_context *ctx)
+{
+	return kref_put(&ctx->kref, &destroy_spu_context);
+}
+
+/* give up the mm reference when the context is about to be destroyed */
+void spu_forget(struct spu_context *ctx)
+{
+	struct mm_struct *mm;
+	spu_acquire_saved(ctx);
+	mm = ctx->owner;
+	ctx->owner = NULL;
+	mmput(mm);
+	spu_release(ctx);
+}
+
+void spu_acquire(struct spu_context *ctx)
+{
+	down_read(&ctx->state_sema);
+}
+
+void spu_release(struct spu_context *ctx)
+{
+	up_read(&ctx->state_sema);
+}
+
+void spu_unmap_mappings(struct spu_context *ctx)
+{
+	unmap_mapping_range(ctx->local_store, 0, LS_SIZE, 1);
+}
+
+int spu_acquire_runnable(struct spu_context *ctx)
+{
+	int ret = 0;
+
+	down_read(&ctx->state_sema);
+	if (ctx->state == SPU_STATE_RUNNABLE) {
+		ctx->spu->prio = current->prio;
+		return 0;
+	}
+	up_read(&ctx->state_sema);
+
+	down_write(&ctx->state_sema);
+	/* ctx is about to be freed, can't acquire any more */
+	if (!ctx->owner) {
+		ret = -EINVAL;
+		goto out;
+	}
+
+	if (ctx->state == SPU_STATE_SAVED) {
+		ret = spu_activate(ctx, 0);
+		if (ret)
+			goto out;
+		ctx->state = SPU_STATE_RUNNABLE;
+	}
+
+	downgrade_write(&ctx->state_sema);
+	/* On success, we return holding the lock */
+
+	return ret;
+out:
+	/* Release here, to simplify calling code. */
+	up_write(&ctx->state_sema);
+
+	return ret;
+}
+
+void spu_acquire_saved(struct spu_context *ctx)
+{
+	down_read(&ctx->state_sema);
+
+	if (ctx->state == SPU_STATE_SAVED)
+		return;
+
+	up_read(&ctx->state_sema);
+	down_write(&ctx->state_sema);
+
+	if (ctx->state == SPU_STATE_RUNNABLE) {
+		spu_deactivate(ctx);
+		ctx->state = SPU_STATE_SAVED;
+	}
+
+	downgrade_write(&ctx->state_sema);
+}

arch/powerpc/platforms/cell/spufs/file.c

@@ -0,0 +1,794 @@
+/*
+ * SPU file system -- file contents
+ *
+ * (C) Copyright IBM Deutschland Entwicklung GmbH 2005
+ *
+ * Author: Arnd Bergmann <arndb@de.ibm.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+#include <linux/fs.h>
+#include <linux/ioctl.h>
+#include <linux/module.h>
+#include <linux/pagemap.h>
+#include <linux/poll.h>
+#include <linux/ptrace.h>
+
+#include <asm/io.h>
+#include <asm/semaphore.h>
+#include <asm/spu.h>
+#include <asm/uaccess.h>
+
+#include "spufs.h"
+
+
+static int
+spufs_mem_open(struct inode *inode, struct file *file)
+{
+	struct spufs_inode_info *i = SPUFS_I(inode);
+	file->private_data = i->i_ctx;
+	file->f_mapping = i->i_ctx->local_store;
+	return 0;
+}
+
+static ssize_t
+spufs_mem_read(struct file *file, char __user *buffer,
+				size_t size, loff_t *pos)
+{
+	struct spu_context *ctx = file->private_data;
+	char *local_store;
+	int ret;
+
+	spu_acquire(ctx);
+
+	local_store = ctx->ops->get_ls(ctx);
+	ret = simple_read_from_buffer(buffer, size, pos, local_store, LS_SIZE);
+
+	spu_release(ctx);
+	return ret;
+}
+
+static ssize_t
+spufs_mem_write(struct file *file, const char __user *buffer,
+					size_t size, loff_t *pos)
+{
+	struct spu_context *ctx = file->private_data;
+	char *local_store;
+	int ret;
+
+	size = min_t(ssize_t, LS_SIZE - *pos, size);
+	if (size <= 0)
+		return -EFBIG;
+	*pos += size;
+
+	spu_acquire(ctx);
+
+	local_store = ctx->ops->get_ls(ctx);
+	ret = copy_from_user(local_store + *pos - size,
+			     buffer, size) ? -EFAULT : size;
+
+	spu_release(ctx);
+	return ret;
+}
+
+#ifdef CONFIG_SPARSEMEM
+static struct page *
+spufs_mem_mmap_nopage(struct vm_area_struct *vma,
+		      unsigned long address, int *type)
+{
+	struct page *page = NOPAGE_SIGBUS;
+
+	struct spu_context *ctx = vma->vm_file->private_data;
+	unsigned long offset = address - vma->vm_start;
+	offset += vma->vm_pgoff << PAGE_SHIFT;
+
+	spu_acquire(ctx);
+
+	if (ctx->state == SPU_STATE_SAVED)
+		page = vmalloc_to_page(ctx->csa.lscsa->ls + offset);
+	else
+		page = pfn_to_page((ctx->spu->local_store_phys + offset)
+				   >> PAGE_SHIFT);
+
+	spu_release(ctx);
+
+	if (type)
+		*type = VM_FAULT_MINOR;
+
+	page_cache_get(page);
+	return page;
+}
+
+static struct vm_operations_struct spufs_mem_mmap_vmops = {
+	.nopage = spufs_mem_mmap_nopage,
+};
+
+static int
+spufs_mem_mmap(struct file *file, struct vm_area_struct *vma)
+{
+	if (!(vma->vm_flags & VM_SHARED))
+		return -EINVAL;
+
+	/* FIXME: */
+	vma->vm_page_prot = __pgprot(pgprot_val(vma->vm_page_prot)
+				     | _PAGE_NO_CACHE);
+
+	vma->vm_ops = &spufs_mem_mmap_vmops;
+	return 0;
+}
+#endif
+
+static struct file_operations spufs_mem_fops = {
+	.open	 = spufs_mem_open,
+	.read    = spufs_mem_read,
+	.write   = spufs_mem_write,
+	.llseek  = generic_file_llseek,
+#ifdef CONFIG_SPARSEMEM
+	.mmap    = spufs_mem_mmap,
+#endif
+};
+
+static int
+spufs_regs_open(struct inode *inode, struct file *file)
+{
+	struct spufs_inode_info *i = SPUFS_I(inode);
+	file->private_data = i->i_ctx;
+	return 0;
+}
+
+static ssize_t
+spufs_regs_read(struct file *file, char __user *buffer,
+		size_t size, loff_t *pos)
+{
+	struct spu_context *ctx = file->private_data;
+	struct spu_lscsa *lscsa = ctx->csa.lscsa;
+	int ret;
+
+	spu_acquire_saved(ctx);
+
+	ret = simple_read_from_buffer(buffer, size, pos,
+				      lscsa->gprs, sizeof lscsa->gprs);
+
+	spu_release(ctx);
+	return ret;
+}
+
+static ssize_t
+spufs_regs_write(struct file *file, const char __user *buffer,
+		 size_t size, loff_t *pos)
+{
+	struct spu_context *ctx = file->private_data;
+	struct spu_lscsa *lscsa = ctx->csa.lscsa;
+	int ret;
+
+	size = min_t(ssize_t, sizeof lscsa->gprs - *pos, size);
+	if (size <= 0)
+		return -EFBIG;
+	*pos += size;
+
+	spu_acquire_saved(ctx);
+
+	ret = copy_from_user(lscsa->gprs + *pos - size,
+			     buffer, size) ? -EFAULT : size;
+
+	spu_release(ctx);
+	return ret;
+}
+
+static struct file_operations spufs_regs_fops = {
+	.open	 = spufs_regs_open,
+	.read    = spufs_regs_read,
+	.write   = spufs_regs_write,
+	.llseek  = generic_file_llseek,
+};
+
+static ssize_t
+spufs_fpcr_read(struct file *file, char __user * buffer,
+		size_t size, loff_t * pos)
+{
+	struct spu_context *ctx = file->private_data;
+	struct spu_lscsa *lscsa = ctx->csa.lscsa;
+	int ret;
+
+	spu_acquire_saved(ctx);
+
+	ret = simple_read_from_buffer(buffer, size, pos,
+				      &lscsa->fpcr, sizeof(lscsa->fpcr));
+
+	spu_release(ctx);
+	return ret;
+}
+
+static ssize_t
+spufs_fpcr_write(struct file *file, const char __user * buffer,
+		 size_t size, loff_t * pos)
+{
+	struct spu_context *ctx = file->private_data;
+	struct spu_lscsa *lscsa = ctx->csa.lscsa;
+	int ret;
+
+	size = min_t(ssize_t, sizeof(lscsa->fpcr) - *pos, size);
+	if (size <= 0)
+		return -EFBIG;
+	*pos += size;
+
+	spu_acquire_saved(ctx);
+
+	ret = copy_from_user((char *)&lscsa->fpcr + *pos - size,
+			     buffer, size) ? -EFAULT : size;
+
+	spu_release(ctx);
+	return ret;
+}
+
+static struct file_operations spufs_fpcr_fops = {
+	.open = spufs_regs_open,
+	.read = spufs_fpcr_read,
+	.write = spufs_fpcr_write,
+	.llseek = generic_file_llseek,
+};
+
+/* generic open function for all pipe-like files */
+static int spufs_pipe_open(struct inode *inode, struct file *file)
+{
+	struct spufs_inode_info *i = SPUFS_I(inode);
+	file->private_data = i->i_ctx;
+
+	return nonseekable_open(inode, file);
+}
+
+static ssize_t spufs_mbox_read(struct file *file, char __user *buf,
+			size_t len, loff_t *pos)
+{
+	struct spu_context *ctx = file->private_data;
+	u32 mbox_data;
+	int ret;
+
+	if (len < 4)
+		return -EINVAL;
+
+	spu_acquire(ctx);
+	ret = ctx->ops->mbox_read(ctx, &mbox_data);
+	spu_release(ctx);
+
+	if (!ret)
+		return -EAGAIN;
+
+	if (copy_to_user(buf, &mbox_data, sizeof mbox_data))
+		return -EFAULT;
+
+	return 4;
+}
+
+static struct file_operations spufs_mbox_fops = {
+	.open	= spufs_pipe_open,
+	.read	= spufs_mbox_read,
+};
+
+static ssize_t spufs_mbox_stat_read(struct file *file, char __user *buf,
+			size_t len, loff_t *pos)
+{
+	struct spu_context *ctx = file->private_data;
+	u32 mbox_stat;
+
+	if (len < 4)
+		return -EINVAL;
+
+	spu_acquire(ctx);
+
+	mbox_stat = ctx->ops->mbox_stat_read(ctx) & 0xff;
+
+	spu_release(ctx);
+
+	if (copy_to_user(buf, &mbox_stat, sizeof mbox_stat))
+		return -EFAULT;
+
+	return 4;
+}
+
+static struct file_operations spufs_mbox_stat_fops = {
+	.open	= spufs_pipe_open,
+	.read	= spufs_mbox_stat_read,
+};
+
+/* low-level ibox access function */
+size_t spu_ibox_read(struct spu_context *ctx, u32 *data)
+{
+	return ctx->ops->ibox_read(ctx, data);
+}
+
+static int spufs_ibox_fasync(int fd, struct file *file, int on)
+{
+	struct spu_context *ctx = file->private_data;
+
+	return fasync_helper(fd, file, on, &ctx->ibox_fasync);
+}
+
+/* interrupt-level ibox callback function. */
+void spufs_ibox_callback(struct spu *spu)
+{
+	struct spu_context *ctx = spu->ctx;
+
+	wake_up_all(&ctx->ibox_wq);
+	kill_fasync(&ctx->ibox_fasync, SIGIO, POLLIN);
+}
+
+static ssize_t spufs_ibox_read(struct file *file, char __user *buf,
+			size_t len, loff_t *pos)
+{
+	struct spu_context *ctx = file->private_data;
+	u32 ibox_data;
+	ssize_t ret;
+
+	if (len < 4)
+		return -EINVAL;
+
+	spu_acquire(ctx);
+
+	ret = 0;
+	if (file->f_flags & O_NONBLOCK) {
+		if (!spu_ibox_read(ctx, &ibox_data))
+			ret = -EAGAIN;
+	} else {
+		ret = spufs_wait(ctx->ibox_wq, spu_ibox_read(ctx, &ibox_data));
+	}
+
+	spu_release(ctx);
+
+	if (ret)
+		return ret;
+
+	ret = 4;
+	if (copy_to_user(buf, &ibox_data, sizeof ibox_data))
+		ret = -EFAULT;
+
+	return ret;
+}
+
+static unsigned int spufs_ibox_poll(struct file *file, poll_table *wait)
+{
+	struct spu_context *ctx = file->private_data;
+	unsigned int mask;
+
+	poll_wait(file, &ctx->ibox_wq, wait);
+
+	spu_acquire(ctx);
+	mask = ctx->ops->mbox_stat_poll(ctx, POLLIN | POLLRDNORM);
+	spu_release(ctx);
+
+	return mask;
+}
+
+static struct file_operations spufs_ibox_fops = {
+	.open	= spufs_pipe_open,
+	.read	= spufs_ibox_read,
+	.poll	= spufs_ibox_poll,
+	.fasync	= spufs_ibox_fasync,
+};
+
+static ssize_t spufs_ibox_stat_read(struct file *file, char __user *buf,
+			size_t len, loff_t *pos)
+{
+	struct spu_context *ctx = file->private_data;
+	u32 ibox_stat;
+
+	if (len < 4)
+		return -EINVAL;
+
+	spu_acquire(ctx);
+	ibox_stat = (ctx->ops->mbox_stat_read(ctx) >> 16) & 0xff;
+	spu_release(ctx);
+
+	if (copy_to_user(buf, &ibox_stat, sizeof ibox_stat))
+		return -EFAULT;
+
+	return 4;
+}
+
+static struct file_operations spufs_ibox_stat_fops = {
+	.open	= spufs_pipe_open,
+	.read	= spufs_ibox_stat_read,
+};
+
+/* low-level mailbox write */
+size_t spu_wbox_write(struct spu_context *ctx, u32 data)
+{
+	return ctx->ops->wbox_write(ctx, data);
+}
+
+static int spufs_wbox_fasync(int fd, struct file *file, int on)
+{
+	struct spu_context *ctx = file->private_data;
+	int ret;
+
+	ret = fasync_helper(fd, file, on, &ctx->wbox_fasync);
+
+	return ret;
+}
+
+/* interrupt-level wbox callback function. */
+void spufs_wbox_callback(struct spu *spu)
+{
+	struct spu_context *ctx = spu->ctx;
+
+	wake_up_all(&ctx->wbox_wq);
+	kill_fasync(&ctx->wbox_fasync, SIGIO, POLLOUT);
+}
+
+static ssize_t spufs_wbox_write(struct file *file, const char __user *buf,
+			size_t len, loff_t *pos)
+{
+	struct spu_context *ctx = file->private_data;
+	u32 wbox_data;
+	int ret;
+
+	if (len < 4)
+		return -EINVAL;
+
+	if (copy_from_user(&wbox_data, buf, sizeof wbox_data))
+		return -EFAULT;
+
+	spu_acquire(ctx);
+
+	ret = 0;
+	if (file->f_flags & O_NONBLOCK) {
+		if (!spu_wbox_write(ctx, wbox_data))
+			ret = -EAGAIN;
+	} else {
+		ret = spufs_wait(ctx->wbox_wq, spu_wbox_write(ctx, wbox_data));
+	}
+
+	spu_release(ctx);
+
+	return ret ? ret : sizeof wbox_data;
+}
+
+static unsigned int spufs_wbox_poll(struct file *file, poll_table *wait)
+{
+	struct spu_context *ctx = file->private_data;
+	unsigned int mask;
+
+	poll_wait(file, &ctx->wbox_wq, wait);
+
+	spu_acquire(ctx);
+	mask = ctx->ops->mbox_stat_poll(ctx, POLLOUT | POLLWRNORM);
+	spu_release(ctx);
+
+	return mask;
+}
+
+static struct file_operations spufs_wbox_fops = {
+	.open	= spufs_pipe_open,
+	.write	= spufs_wbox_write,
+	.poll	= spufs_wbox_poll,
+	.fasync	= spufs_wbox_fasync,
+};
+
+static ssize_t spufs_wbox_stat_read(struct file *file, char __user *buf,
+			size_t len, loff_t *pos)
+{
+	struct spu_context *ctx = file->private_data;
+	u32 wbox_stat;
+
+	if (len < 4)
+		return -EINVAL;
+
+	spu_acquire(ctx);
+	wbox_stat = (ctx->ops->mbox_stat_read(ctx) >> 8) & 0xff;
+	spu_release(ctx);
+
+	if (copy_to_user(buf, &wbox_stat, sizeof wbox_stat))
+		return -EFAULT;
+
+	return 4;
+}
+
+static struct file_operations spufs_wbox_stat_fops = {
+	.open	= spufs_pipe_open,
+	.read	= spufs_wbox_stat_read,
+};
+
+static ssize_t spufs_signal1_read(struct file *file, char __user *buf,
+			size_t len, loff_t *pos)
+{
+	struct spu_context *ctx = file->private_data;
+	u32 data;
+
+	if (len < 4)
+		return -EINVAL;
+
+	spu_acquire(ctx);
+	data = ctx->ops->signal1_read(ctx);
+	spu_release(ctx);
+
+	if (copy_to_user(buf, &data, 4))
+		return -EFAULT;
+
+	return 4;
+}
+
+static ssize_t spufs_signal1_write(struct file *file, const char __user *buf,
+			size_t len, loff_t *pos)
+{
+	struct spu_context *ctx;
+	u32 data;
+
+	ctx = file->private_data;
+
+	if (len < 4)
+		return -EINVAL;
+
+	if (copy_from_user(&data, buf, 4))
+		return -EFAULT;
+
+	spu_acquire(ctx);
+	ctx->ops->signal1_write(ctx, data);
+	spu_release(ctx);
+
+	return 4;
+}
+
+static struct file_operations spufs_signal1_fops = {
+	.open = spufs_pipe_open,
+	.read = spufs_signal1_read,
+	.write = spufs_signal1_write,
+};
+
+static ssize_t spufs_signal2_read(struct file *file, char __user *buf,
+			size_t len, loff_t *pos)
+{
+	struct spu_context *ctx;
+	u32 data;
+
+	ctx = file->private_data;
+
+	if (len < 4)
+		return -EINVAL;
+
+	spu_acquire(ctx);
+	data = ctx->ops->signal2_read(ctx);
+	spu_release(ctx);
+
+	if (copy_to_user(buf, &data, 4))
+		return -EFAULT;
+
+	return 4;
+}
+
+static ssize_t spufs_signal2_write(struct file *file, const char __user *buf,
+			size_t len, loff_t *pos)
+{
+	struct spu_context *ctx;
+	u32 data;
+
+	ctx = file->private_data;
+
+	if (len < 4)
+		return -EINVAL;
+
+	if (copy_from_user(&data, buf, 4))
+		return -EFAULT;
+
+	spu_acquire(ctx);
+	ctx->ops->signal2_write(ctx, data);
+	spu_release(ctx);
+
+	return 4;
+}
+
+static struct file_operations spufs_signal2_fops = {
+	.open = spufs_pipe_open,
+	.read = spufs_signal2_read,
+	.write = spufs_signal2_write,
+};
+
+static void spufs_signal1_type_set(void *data, u64 val)
+{
+	struct spu_context *ctx = data;
+
+	spu_acquire(ctx);
+	ctx->ops->signal1_type_set(ctx, val);
+	spu_release(ctx);
+}
+
+static u64 spufs_signal1_type_get(void *data)
+{
+	struct spu_context *ctx = data;
+	u64 ret;
+
+	spu_acquire(ctx);
+	ret = ctx->ops->signal1_type_get(ctx);
+	spu_release(ctx);
+
+	return ret;
+}
+DEFINE_SIMPLE_ATTRIBUTE(spufs_signal1_type, spufs_signal1_type_get,
+					spufs_signal1_type_set, "%llu");
+
+static void spufs_signal2_type_set(void *data, u64 val)
+{
+	struct spu_context *ctx = data;
+
+	spu_acquire(ctx);
+	ctx->ops->signal2_type_set(ctx, val);
+	spu_release(ctx);
+}
+
+static u64 spufs_signal2_type_get(void *data)
+{
+	struct spu_context *ctx = data;
+	u64 ret;
+
+	spu_acquire(ctx);
+	ret = ctx->ops->signal2_type_get(ctx);
+	spu_release(ctx);
+
+	return ret;
+}
+DEFINE_SIMPLE_ATTRIBUTE(spufs_signal2_type, spufs_signal2_type_get,
+					spufs_signal2_type_set, "%llu");
+
+static void spufs_npc_set(void *data, u64 val)
+{
+	struct spu_context *ctx = data;
+	spu_acquire(ctx);
+	ctx->ops->npc_write(ctx, val);
+	spu_release(ctx);
+}
+
+static u64 spufs_npc_get(void *data)
+{
+	struct spu_context *ctx = data;
+	u64 ret;
+	spu_acquire(ctx);
+	ret = ctx->ops->npc_read(ctx);
+	spu_release(ctx);
+	return ret;
+}
+DEFINE_SIMPLE_ATTRIBUTE(spufs_npc_ops, spufs_npc_get, spufs_npc_set, "%llx\n")
+
+static void spufs_decr_set(void *data, u64 val)
+{
+	struct spu_context *ctx = data;
+	struct spu_lscsa *lscsa = ctx->csa.lscsa;
+	spu_acquire_saved(ctx);
+	lscsa->decr.slot[0] = (u32) val;
+	spu_release(ctx);
+}
+
+static u64 spufs_decr_get(void *data)
+{
+	struct spu_context *ctx = data;
+	struct spu_lscsa *lscsa = ctx->csa.lscsa;
+	u64 ret;
+	spu_acquire_saved(ctx);
+	ret = lscsa->decr.slot[0];
+	spu_release(ctx);
+	return ret;
+}
+DEFINE_SIMPLE_ATTRIBUTE(spufs_decr_ops, spufs_decr_get, spufs_decr_set,
+			"%llx\n")
+
+static void spufs_decr_status_set(void *data, u64 val)
+{
+	struct spu_context *ctx = data;
+	struct spu_lscsa *lscsa = ctx->csa.lscsa;
+	spu_acquire_saved(ctx);
+	lscsa->decr_status.slot[0] = (u32) val;
+	spu_release(ctx);
+}
+
+static u64 spufs_decr_status_get(void *data)
+{
+	struct spu_context *ctx = data;
+	struct spu_lscsa *lscsa = ctx->csa.lscsa;
+	u64 ret;
+	spu_acquire_saved(ctx);
+	ret = lscsa->decr_status.slot[0];
+	spu_release(ctx);
+	return ret;
+}
+DEFINE_SIMPLE_ATTRIBUTE(spufs_decr_status_ops, spufs_decr_status_get,
+			spufs_decr_status_set, "%llx\n")
+
+static void spufs_spu_tag_mask_set(void *data, u64 val)
+{
+	struct spu_context *ctx = data;
+	struct spu_lscsa *lscsa = ctx->csa.lscsa;
+	spu_acquire_saved(ctx);
+	lscsa->tag_mask.slot[0] = (u32) val;
+	spu_release(ctx);
+}
+
+static u64 spufs_spu_tag_mask_get(void *data)
+{
+	struct spu_context *ctx = data;
+	struct spu_lscsa *lscsa = ctx->csa.lscsa;
+	u64 ret;
+	spu_acquire_saved(ctx);
+	ret = lscsa->tag_mask.slot[0];
+	spu_release(ctx);
+	return ret;
+}
+DEFINE_SIMPLE_ATTRIBUTE(spufs_spu_tag_mask_ops, spufs_spu_tag_mask_get,
+			spufs_spu_tag_mask_set, "%llx\n")
+
+static void spufs_event_mask_set(void *data, u64 val)
+{
+	struct spu_context *ctx = data;
+	struct spu_lscsa *lscsa = ctx->csa.lscsa;
+	spu_acquire_saved(ctx);
+	lscsa->event_mask.slot[0] = (u32) val;
+	spu_release(ctx);
+}
+
+static u64 spufs_event_mask_get(void *data)
+{
+	struct spu_context *ctx = data;
+	struct spu_lscsa *lscsa = ctx->csa.lscsa;
+	u64 ret;
+	spu_acquire_saved(ctx);
+	ret = lscsa->event_mask.slot[0];
+	spu_release(ctx);
+	return ret;
+}
+DEFINE_SIMPLE_ATTRIBUTE(spufs_event_mask_ops, spufs_event_mask_get,
+			spufs_event_mask_set, "%llx\n")
+
+static void spufs_srr0_set(void *data, u64 val)
+{
+	struct spu_context *ctx = data;
+	struct spu_lscsa *lscsa = ctx->csa.lscsa;
+	spu_acquire_saved(ctx);
+	lscsa->srr0.slot[0] = (u32) val;
+	spu_release(ctx);
+}
+
+static u64 spufs_srr0_get(void *data)
+{
+	struct spu_context *ctx = data;
+	struct spu_lscsa *lscsa = ctx->csa.lscsa;
+	u64 ret;
+	spu_acquire_saved(ctx);
+	ret = lscsa->srr0.slot[0];
+	spu_release(ctx);
+	return ret;
+}
+DEFINE_SIMPLE_ATTRIBUTE(spufs_srr0_ops, spufs_srr0_get, spufs_srr0_set,
+			"%llx\n")
+
+struct tree_descr spufs_dir_contents[] = {
+	{ "mem",  &spufs_mem_fops,  0666, },
+	{ "regs", &spufs_regs_fops,  0666, },
+	{ "mbox", &spufs_mbox_fops, 0444, },
+	{ "ibox", &spufs_ibox_fops, 0444, },
+	{ "wbox", &spufs_wbox_fops, 0222, },
+	{ "mbox_stat", &spufs_mbox_stat_fops, 0444, },
+	{ "ibox_stat", &spufs_ibox_stat_fops, 0444, },
+	{ "wbox_stat", &spufs_wbox_stat_fops, 0444, },
+	{ "signal1", &spufs_signal1_fops, 0666, },
+	{ "signal2", &spufs_signal2_fops, 0666, },
+	{ "signal1_type", &spufs_signal1_type, 0666, },
+	{ "signal2_type", &spufs_signal2_type, 0666, },
+	{ "npc", &spufs_npc_ops, 0666, },
+	{ "fpcr", &spufs_fpcr_fops, 0666, },
+	{ "decr", &spufs_decr_ops, 0666, },
+	{ "decr_status", &spufs_decr_status_ops, 0666, },
+	{ "spu_tag_mask", &spufs_spu_tag_mask_ops, 0666, },
+	{ "event_mask", &spufs_event_mask_ops, 0666, },
+	{ "srr0", &spufs_srr0_ops, 0666, },
+	{},
+};

arch/powerpc/platforms/cell/spufs/hw_ops.c

@@ -0,0 +1,255 @@
+/* hw_ops.c - query/set operations on active SPU context.
+ *
+ * Copyright (C) IBM 2005
+ * Author: Mark Nutter <mnutter@us.ibm.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+#include <linux/config.h>
+#include <linux/module.h>
+#include <linux/errno.h>
+#include <linux/sched.h>
+#include <linux/kernel.h>
+#include <linux/mm.h>
+#include <linux/poll.h>
+#include <linux/smp.h>
+#include <linux/smp_lock.h>
+#include <linux/stddef.h>
+#include <linux/unistd.h>
+
+#include <asm/io.h>
+#include <asm/spu.h>
+#include <asm/spu_csa.h>
+#include <asm/mmu_context.h>
+#include "spufs.h"
+
+static int spu_hw_mbox_read(struct spu_context *ctx, u32 * data)
+{
+	struct spu *spu = ctx->spu;
+	struct spu_problem __iomem *prob = spu->problem;
+	u32 mbox_stat;
+	int ret = 0;
+
+	spin_lock_irq(&spu->register_lock);
+	mbox_stat = in_be32(&prob->mb_stat_R);
+	if (mbox_stat & 0x0000ff) {
+		*data = in_be32(&prob->pu_mb_R);
+		ret = 4;
+	}
+	spin_unlock_irq(&spu->register_lock);
+	return ret;
+}
+
+static u32 spu_hw_mbox_stat_read(struct spu_context *ctx)
+{
+	return in_be32(&ctx->spu->problem->mb_stat_R);
+}
+
+static unsigned int spu_hw_mbox_stat_poll(struct spu_context *ctx,
+					  unsigned int events)
+{
+	struct spu *spu = ctx->spu;
+	int ret = 0;
+	u32 stat;
+
+	spin_lock_irq(&spu->register_lock);
+	stat = in_be32(&spu->problem->mb_stat_R);
+
+	/* if the requested event is there, return the poll
+	   mask, otherwise enable the interrupt to get notified,
+	   but first mark any pending interrupts as done so
+	   we don't get woken up unnecessarily */
+
+	if (events & (POLLIN | POLLRDNORM)) {
+		if (stat & 0xff0000)
+			ret |= POLLIN | POLLRDNORM;
+		else {
+			spu_int_stat_clear(spu, 2, 0x1);
+			spu_int_mask_or(spu, 2, 0x1);
+		}
+	}
+	if (events & (POLLOUT | POLLWRNORM)) {
+		if (stat & 0x00ff00)
+			ret = POLLOUT | POLLWRNORM;
+		else {
+			spu_int_stat_clear(spu, 2, 0x10);
+			spu_int_mask_or(spu, 2, 0x10);
+		}
+	}
+	spin_unlock_irq(&spu->register_lock);
+	return ret;
+}
+
+static int spu_hw_ibox_read(struct spu_context *ctx, u32 * data)
+{
+	struct spu *spu = ctx->spu;
+	struct spu_problem __iomem *prob = spu->problem;
+	struct spu_priv2 __iomem *priv2 = spu->priv2;
+	int ret;
+
+	spin_lock_irq(&spu->register_lock);
+	if (in_be32(&prob->mb_stat_R) & 0xff0000) {
+		/* read the first available word */
+		*data = in_be64(&priv2->puint_mb_R);
+		ret = 4;
+	} else {
+		/* make sure we get woken up by the interrupt */
+		spu_int_mask_or(spu, 2, 0x1);
+		ret = 0;
+	}
+	spin_unlock_irq(&spu->register_lock);
+	return ret;
+}
+
+static int spu_hw_wbox_write(struct spu_context *ctx, u32 data)
+{
+	struct spu *spu = ctx->spu;
+	struct spu_problem __iomem *prob = spu->problem;
+	int ret;
+
+	spin_lock_irq(&spu->register_lock);
+	if (in_be32(&prob->mb_stat_R) & 0x00ff00) {
+		/* we have space to write wbox_data to */
+		out_be32(&prob->spu_mb_W, data);
+		ret = 4;
+	} else {
+		/* make sure we get woken up by the interrupt when space
+		   becomes available */
+		spu_int_mask_or(spu, 2, 0x10);
+		ret = 0;
+	}
+	spin_unlock_irq(&spu->register_lock);
+	return ret;
+}
+
+static u32 spu_hw_signal1_read(struct spu_context *ctx)
+{
+	return in_be32(&ctx->spu->problem->signal_notify1);
+}
+
+static void spu_hw_signal1_write(struct spu_context *ctx, u32 data)
+{
+	out_be32(&ctx->spu->problem->signal_notify1, data);
+}
+
+static u32 spu_hw_signal2_read(struct spu_context *ctx)
+{
+	return in_be32(&ctx->spu->problem->signal_notify1);
+}
+
+static void spu_hw_signal2_write(struct spu_context *ctx, u32 data)
+{
+	out_be32(&ctx->spu->problem->signal_notify2, data);
+}
+
+static void spu_hw_signal1_type_set(struct spu_context *ctx, u64 val)
+{
+	struct spu *spu = ctx->spu;
+	struct spu_priv2 __iomem *priv2 = spu->priv2;
+	u64 tmp;
+
+	spin_lock_irq(&spu->register_lock);
+	tmp = in_be64(&priv2->spu_cfg_RW);
+	if (val)
+		tmp |= 1;
+	else
+		tmp &= ~1;
+	out_be64(&priv2->spu_cfg_RW, tmp);
+	spin_unlock_irq(&spu->register_lock);
+}
+
+static u64 spu_hw_signal1_type_get(struct spu_context *ctx)
+{
+	return ((in_be64(&ctx->spu->priv2->spu_cfg_RW) & 1) != 0);
+}
+
+static void spu_hw_signal2_type_set(struct spu_context *ctx, u64 val)
+{
+	struct spu *spu = ctx->spu;
+	struct spu_priv2 __iomem *priv2 = spu->priv2;
+	u64 tmp;
+
+	spin_lock_irq(&spu->register_lock);
+	tmp = in_be64(&priv2->spu_cfg_RW);
+	if (val)
+		tmp |= 2;
+	else
+		tmp &= ~2;
+	out_be64(&priv2->spu_cfg_RW, tmp);
+	spin_unlock_irq(&spu->register_lock);
+}
+
+static u64 spu_hw_signal2_type_get(struct spu_context *ctx)
+{
+	return ((in_be64(&ctx->spu->priv2->spu_cfg_RW) & 2) != 0);
+}
+
+static u32 spu_hw_npc_read(struct spu_context *ctx)
+{
+	return in_be32(&ctx->spu->problem->spu_npc_RW);
+}
+
+static void spu_hw_npc_write(struct spu_context *ctx, u32 val)
+{
+	out_be32(&ctx->spu->problem->spu_npc_RW, val);
+}
+
+static u32 spu_hw_status_read(struct spu_context *ctx)
+{
+	return in_be32(&ctx->spu->problem->spu_status_R);
+}
+
+static char *spu_hw_get_ls(struct spu_context *ctx)
+{
+	return ctx->spu->local_store;
+}
+
+static void spu_hw_runcntl_write(struct spu_context *ctx, u32 val)
+{
+	eieio();
+	out_be32(&ctx->spu->problem->spu_runcntl_RW, val);
+}
+
+static void spu_hw_runcntl_stop(struct spu_context *ctx)
+{
+	spin_lock_irq(&ctx->spu->register_lock);
+	out_be32(&ctx->spu->problem->spu_runcntl_RW, SPU_RUNCNTL_STOP);
+	while (in_be32(&ctx->spu->problem->spu_status_R) & SPU_STATUS_RUNNING)
+		cpu_relax();
+	spin_unlock_irq(&ctx->spu->register_lock);
+}
+
+struct spu_context_ops spu_hw_ops = {
+	.mbox_read = spu_hw_mbox_read,
+	.mbox_stat_read = spu_hw_mbox_stat_read,
+	.mbox_stat_poll = spu_hw_mbox_stat_poll,
+	.ibox_read = spu_hw_ibox_read,
+	.wbox_write = spu_hw_wbox_write,
+	.signal1_read = spu_hw_signal1_read,
+	.signal1_write = spu_hw_signal1_write,
+	.signal2_read = spu_hw_signal2_read,
+	.signal2_write = spu_hw_signal2_write,
+	.signal1_type_set = spu_hw_signal1_type_set,
+	.signal1_type_get = spu_hw_signal1_type_get,
+	.signal2_type_set = spu_hw_signal2_type_set,
+	.signal2_type_get = spu_hw_signal2_type_get,
+	.npc_read = spu_hw_npc_read,
+	.npc_write = spu_hw_npc_write,
+	.status_read = spu_hw_status_read,
+	.get_ls = spu_hw_get_ls,
+	.runcntl_write = spu_hw_runcntl_write,
+	.runcntl_stop = spu_hw_runcntl_stop,
+};

arch/powerpc/platforms/cell/spufs/inode.c

@@ -0,0 +1,486 @@
+/*
+ * SPU file system
+ *
+ * (C) Copyright IBM Deutschland Entwicklung GmbH 2005
+ *
+ * Author: Arnd Bergmann <arndb@de.ibm.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+#include <linux/file.h>
+#include <linux/fs.h>
+#include <linux/backing-dev.h>
+#include <linux/init.h>
+#include <linux/ioctl.h>
+#include <linux/module.h>
+#include <linux/mount.h>
+#include <linux/namei.h>
+#include <linux/pagemap.h>
+#include <linux/poll.h>
+#include <linux/slab.h>
+#include <linux/parser.h>
+
+#include <asm/io.h>
+#include <asm/semaphore.h>
+#include <asm/spu.h>
+#include <asm/uaccess.h>
+
+#include "spufs.h"
+
+static kmem_cache_t *spufs_inode_cache;
+
+static struct inode *
+spufs_alloc_inode(struct super_block *sb)
+{
+	struct spufs_inode_info *ei;
+
+	ei = kmem_cache_alloc(spufs_inode_cache, SLAB_KERNEL);
+	if (!ei)
+		return NULL;
+	return &ei->vfs_inode;
+}
+
+static void
+spufs_destroy_inode(struct inode *inode)
+{
+	kmem_cache_free(spufs_inode_cache, SPUFS_I(inode));
+}
+
+static void
+spufs_init_once(void *p, kmem_cache_t * cachep, unsigned long flags)
+{
+	struct spufs_inode_info *ei = p;
+
+	if ((flags & (SLAB_CTOR_VERIFY|SLAB_CTOR_CONSTRUCTOR)) ==
+	    SLAB_CTOR_CONSTRUCTOR) {
+		inode_init_once(&ei->vfs_inode);
+	}
+}
+
+static struct inode *
+spufs_new_inode(struct super_block *sb, int mode)
+{
+	struct inode *inode;
+
+	inode = new_inode(sb);
+	if (!inode)
+		goto out;
+
+	inode->i_mode = mode;
+	inode->i_uid = current->fsuid;
+	inode->i_gid = current->fsgid;
+	inode->i_blksize = PAGE_CACHE_SIZE;
+	inode->i_blocks = 0;
+	inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
+out:
+	return inode;
+}
+
+static int
+spufs_setattr(struct dentry *dentry, struct iattr *attr)
+{
+	struct inode *inode = dentry->d_inode;
+
+	if ((attr->ia_valid & ATTR_SIZE) &&
+	    (attr->ia_size != inode->i_size))
+		return -EINVAL;
+	return inode_setattr(inode, attr);
+}
+
+
+static int
+spufs_new_file(struct super_block *sb, struct dentry *dentry,
+		struct file_operations *fops, int mode,
+		struct spu_context *ctx)
+{
+	static struct inode_operations spufs_file_iops = {
+		.setattr = spufs_setattr,
+	};
+	struct inode *inode;
+	int ret;
+
+	ret = -ENOSPC;
+	inode = spufs_new_inode(sb, S_IFREG | mode);
+	if (!inode)
+		goto out;
+
+	ret = 0;
+	inode->i_op = &spufs_file_iops;
+	inode->i_fop = fops;
+	inode->u.generic_ip = SPUFS_I(inode)->i_ctx = get_spu_context(ctx);
+	d_add(dentry, inode);
+out:
+	return ret;
+}
+
+static void
+spufs_delete_inode(struct inode *inode)
+{
+	if (SPUFS_I(inode)->i_ctx)
+		put_spu_context(SPUFS_I(inode)->i_ctx);
+	clear_inode(inode);
+}
+
+static void spufs_prune_dir(struct dentry *dir)
+{
+	struct dentry *dentry, *tmp;
+	down(&dir->d_inode->i_sem);
+	list_for_each_entry_safe(dentry, tmp, &dir->d_subdirs, d_child) {
+		spin_lock(&dcache_lock);
+		spin_lock(&dentry->d_lock);
+		if (!(d_unhashed(dentry)) && dentry->d_inode) {
+			dget_locked(dentry);
+			__d_drop(dentry);
+			spin_unlock(&dentry->d_lock);
+			simple_unlink(dir->d_inode, dentry);
+			spin_unlock(&dcache_lock);
+			dput(dentry);
+		} else {
+			spin_unlock(&dentry->d_lock);
+			spin_unlock(&dcache_lock);
+		}
+	}
+	shrink_dcache_parent(dir);
+	up(&dir->d_inode->i_sem);
+}
+
+static int spufs_rmdir(struct inode *root, struct dentry *dir_dentry)
+{
+	struct spu_context *ctx;
+
+	/* remove all entries */
+	down(&root->i_sem);
+	spufs_prune_dir(dir_dentry);
+	up(&root->i_sem);
+
+	/* We have to give up the mm_struct */
+	ctx = SPUFS_I(dir_dentry->d_inode)->i_ctx;
+	spu_forget(ctx);
+
+	/* XXX Do we need to hold i_sem here ? */
+	return simple_rmdir(root, dir_dentry);
+}
+
+static int spufs_fill_dir(struct dentry *dir, struct tree_descr *files,
+			  int mode, struct spu_context *ctx)
+{
+	struct dentry *dentry;
+	int ret;
+
+	while (files->name && files->name[0]) {
+		ret = -ENOMEM;
+		dentry = d_alloc_name(dir, files->name);
+		if (!dentry)
+			goto out;
+		ret = spufs_new_file(dir->d_sb, dentry, files->ops,
+					files->mode & mode, ctx);
+		if (ret)
+			goto out;
+		files++;
+	}
+	return 0;
+out:
+	spufs_prune_dir(dir);
+	return ret;
+}
+
+static int spufs_dir_close(struct inode *inode, struct file *file)
+{
+	struct inode *dir;
+	struct dentry *dentry;
+	int ret;
+
+	dentry = file->f_dentry;
+	dir = dentry->d_parent->d_inode;
+
+	ret = spufs_rmdir(dir, dentry);
+	WARN_ON(ret);
+
+	return dcache_dir_close(inode, file);
+}
+
+struct inode_operations spufs_dir_inode_operations = {
+	.lookup = simple_lookup,
+};
+
+struct file_operations spufs_context_fops = {
+	.open		= dcache_dir_open,
+	.release	= spufs_dir_close,
+	.llseek		= dcache_dir_lseek,
+	.read		= generic_read_dir,
+	.readdir	= dcache_readdir,
+	.fsync		= simple_sync_file,
+};
+
+static int
+spufs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
+{
+	int ret;
+	struct inode *inode;
+	struct spu_context *ctx;
+
+	ret = -ENOSPC;
+	inode = spufs_new_inode(dir->i_sb, mode | S_IFDIR);
+	if (!inode)
+		goto out;
+
+	if (dir->i_mode & S_ISGID) {
+		inode->i_gid = dir->i_gid;
+		inode->i_mode &= S_ISGID;
+	}
+	ctx = alloc_spu_context(inode->i_mapping);
+	SPUFS_I(inode)->i_ctx = ctx;
+	if (!ctx)
+		goto out_iput;
+
+	inode->i_op = &spufs_dir_inode_operations;
+	inode->i_fop = &simple_dir_operations;
+	ret = spufs_fill_dir(dentry, spufs_dir_contents, mode, ctx);
+	if (ret)
+		goto out_free_ctx;
+
+	d_instantiate(dentry, inode);
+	dget(dentry);
+	dir->i_nlink++;
+	dentry->d_inode->i_nlink++;
+	goto out;
+
+out_free_ctx:
+	put_spu_context(ctx);
+out_iput:
+	iput(inode);
+out:
+	return ret;
+}
+
+static int spufs_context_open(struct dentry *dentry, struct vfsmount *mnt)
+{
+	int ret;
+	struct file *filp;
+
+	ret = get_unused_fd();
+	if (ret < 0) {
+		dput(dentry);
+		mntput(mnt);
+		goto out;
+	}
+
+	filp = dentry_open(dentry, mnt, O_RDONLY);
+	if (IS_ERR(filp)) {
+		put_unused_fd(ret);
+		ret = PTR_ERR(filp);
+		goto out;
+	}
+
+	filp->f_op = &spufs_context_fops;
+	fd_install(ret, filp);
+out:
+	return ret;
+}
+
+static struct file_system_type spufs_type;
+
+long spufs_create_thread(struct nameidata *nd,
+			 unsigned int flags, mode_t mode)
+{
+	struct dentry *dentry;
+	int ret;
+
+	/* need to be at the root of spufs */
+	ret = -EINVAL;
+	if (nd->dentry->d_sb->s_type != &spufs_type ||
+	    nd->dentry != nd->dentry->d_sb->s_root)
+		goto out;
+
+	dentry = lookup_create(nd, 1);
+	ret = PTR_ERR(dentry);
+	if (IS_ERR(dentry))
+		goto out_dir;
+
+	ret = -EEXIST;
+	if (dentry->d_inode)
+		goto out_dput;
+
+	mode &= ~current->fs->umask;
+	ret = spufs_mkdir(nd->dentry->d_inode, dentry, mode & S_IRWXUGO);
+	if (ret)
+		goto out_dput;
+
+	/*
+	 * get references for dget and mntget, will be released
+	 * in error path of *_open().
+	 */
+	ret = spufs_context_open(dget(dentry), mntget(nd->mnt));
+	if (ret < 0)
+		spufs_rmdir(nd->dentry->d_inode, dentry);
+
+out_dput:
+	dput(dentry);
+out_dir:
+	up(&nd->dentry->d_inode->i_sem);
+out:
+	return ret;
+}
+
+/* File system initialization */
+enum {
+	Opt_uid, Opt_gid, Opt_err,
+};
+
+static match_table_t spufs_tokens = {
+	{ Opt_uid, "uid=%d" },
+	{ Opt_gid, "gid=%d" },
+	{ Opt_err, NULL  },
+};
+
+static int
+spufs_parse_options(char *options, struct inode *root)
+{
+	char *p;
+	substring_t args[MAX_OPT_ARGS];
+
+	while ((p = strsep(&options, ",")) != NULL) {
+		int token, option;
+
+		if (!*p)
+			continue;
+
+		token = match_token(p, spufs_tokens, args);
+		switch (token) {
+		case Opt_uid:
+			if (match_int(&args[0], &option))
+				return 0;
+			root->i_uid = option;
+			break;
+		case Opt_gid:
+			if (match_int(&args[0], &option))
+				return 0;
+			root->i_gid = option;
+			break;
+		default:
+			return 0;
+		}
+	}
+	return 1;
+}
+
+static int
+spufs_create_root(struct super_block *sb, void *data)
+{
+	struct inode *inode;
+	int ret;
+
+	ret = -ENOMEM;
+	inode = spufs_new_inode(sb, S_IFDIR | 0775);
+	if (!inode)
+		goto out;
+
+	inode->i_op = &spufs_dir_inode_operations;
+	inode->i_fop = &simple_dir_operations;
+	SPUFS_I(inode)->i_ctx = NULL;
+
+	ret = -EINVAL;
+	if (!spufs_parse_options(data, inode))
+		goto out_iput;
+
+	ret = -ENOMEM;
+	sb->s_root = d_alloc_root(inode);
+	if (!sb->s_root)
+		goto out_iput;
+
+	return 0;
+out_iput:
+	iput(inode);
+out:
+	return ret;
+}
+
+static int
+spufs_fill_super(struct super_block *sb, void *data, int silent)
+{
+	static struct super_operations s_ops = {
+		.alloc_inode = spufs_alloc_inode,
+		.destroy_inode = spufs_destroy_inode,
+		.statfs = simple_statfs,
+		.delete_inode = spufs_delete_inode,
+		.drop_inode = generic_delete_inode,
+	};
+
+	sb->s_maxbytes = MAX_LFS_FILESIZE;
+	sb->s_blocksize = PAGE_CACHE_SIZE;
+	sb->s_blocksize_bits = PAGE_CACHE_SHIFT;
+	sb->s_magic = SPUFS_MAGIC;
+	sb->s_op = &s_ops;
+
+	return spufs_create_root(sb, data);
+}
+
+static struct super_block *
+spufs_get_sb(struct file_system_type *fstype, int flags,
+		const char *name, void *data)
+{
+	return get_sb_single(fstype, flags, data, spufs_fill_super);
+}
+
+static struct file_system_type spufs_type = {
+	.owner = THIS_MODULE,
+	.name = "spufs",
+	.get_sb = spufs_get_sb,
+	.kill_sb = kill_litter_super,
+};
+
+static int spufs_init(void)
+{
+	int ret;
+	ret = -ENOMEM;
+	spufs_inode_cache = kmem_cache_create("spufs_inode_cache",
+			sizeof(struct spufs_inode_info), 0,
+			SLAB_HWCACHE_ALIGN, spufs_init_once, NULL);
+
+	if (!spufs_inode_cache)
+		goto out;
+	if (spu_sched_init() != 0) {
+		kmem_cache_destroy(spufs_inode_cache);
+		goto out;
+	}
+	ret = register_filesystem(&spufs_type);
+	if (ret)
+		goto out_cache;
+	ret = register_spu_syscalls(&spufs_calls);
+	if (ret)
+		goto out_fs;
+	return 0;
+out_fs:
+	unregister_filesystem(&spufs_type);
+out_cache:
+	kmem_cache_destroy(spufs_inode_cache);
+out:
+	return ret;
+}
+module_init(spufs_init);
+
+static void spufs_exit(void)
+{
+	spu_sched_exit();
+	unregister_spu_syscalls(&spufs_calls);
+	unregister_filesystem(&spufs_type);
+	kmem_cache_destroy(spufs_inode_cache);
+}
+module_exit(spufs_exit);
+
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Arnd Bergmann <arndb@de.ibm.com>");
+

arch/powerpc/platforms/cell/spufs/run.c

@@ -0,0 +1,131 @@
+#include <linux/wait.h>
+#include <linux/ptrace.h>
+
+#include <asm/spu.h>
+
+#include "spufs.h"
+
+/* interrupt-level stop callback function. */
+void spufs_stop_callback(struct spu *spu)
+{
+	struct spu_context *ctx = spu->ctx;
+
+	wake_up_all(&ctx->stop_wq);
+}
+
+static inline int spu_stopped(struct spu_context *ctx, u32 * stat)
+{
+	struct spu *spu;
+	u64 pte_fault;
+
+	*stat = ctx->ops->status_read(ctx);
+	if (ctx->state != SPU_STATE_RUNNABLE)
+		return 1;
+	spu = ctx->spu;
+	pte_fault = spu->dsisr &
+	    (MFC_DSISR_PTE_NOT_FOUND | MFC_DSISR_ACCESS_DENIED);
+	return (!(*stat & 0x1) || pte_fault || spu->class_0_pending) ? 1 : 0;
+}
+
+static inline int spu_run_init(struct spu_context *ctx, u32 * npc,
+			       u32 * status)
+{
+	int ret;
+
+	if ((ret = spu_acquire_runnable(ctx)) != 0)
+		return ret;
+	ctx->ops->npc_write(ctx, *npc);
+	ctx->ops->runcntl_write(ctx, SPU_RUNCNTL_RUNNABLE);
+	return 0;
+}
+
+static inline int spu_run_fini(struct spu_context *ctx, u32 * npc,
+			       u32 * status)
+{
+	int ret = 0;
+
+	*status = ctx->ops->status_read(ctx);
+	*npc = ctx->ops->npc_read(ctx);
+	spu_release(ctx);
+
+	if (signal_pending(current))
+		ret = -ERESTARTSYS;
+	if (unlikely(current->ptrace & PT_PTRACED)) {
+		if ((*status & SPU_STATUS_STOPPED_BY_STOP)
+		    && (*status >> SPU_STOP_STATUS_SHIFT) == 0x3fff) {
+			force_sig(SIGTRAP, current);
+			ret = -ERESTARTSYS;
+		}
+	}
+	return ret;
+}
+
+static inline int spu_reacquire_runnable(struct spu_context *ctx, u32 *npc,
+				         u32 *status)
+{
+	int ret;
+
+	if ((ret = spu_run_fini(ctx, npc, status)) != 0)
+		return ret;
+	if (*status & (SPU_STATUS_STOPPED_BY_STOP |
+		       SPU_STATUS_STOPPED_BY_HALT)) {
+		return *status;
+	}
+	if ((ret = spu_run_init(ctx, npc, status)) != 0)
+		return ret;
+	return 0;
+}
+
+static inline int spu_process_events(struct spu_context *ctx)
+{
+	struct spu *spu = ctx->spu;
+	u64 pte_fault = MFC_DSISR_PTE_NOT_FOUND | MFC_DSISR_ACCESS_DENIED;
+	int ret = 0;
+
+	if (spu->dsisr & pte_fault)
+		ret = spu_irq_class_1_bottom(spu);
+	if (spu->class_0_pending)
+		ret = spu_irq_class_0_bottom(spu);
+	if (!ret && signal_pending(current))
+		ret = -ERESTARTSYS;
+	return ret;
+}
+
+long spufs_run_spu(struct file *file, struct spu_context *ctx,
+		   u32 * npc, u32 * status)
+{
+	int ret;
+
+	if (down_interruptible(&ctx->run_sema))
+		return -ERESTARTSYS;
+
+	ret = spu_run_init(ctx, npc, status);
+	if (ret)
+		goto out;
+
+	do {
+		ret = spufs_wait(ctx->stop_wq, spu_stopped(ctx, status));
+		if (unlikely(ret))
+			break;
+		if (unlikely(ctx->state != SPU_STATE_RUNNABLE)) {
+			ret = spu_reacquire_runnable(ctx, npc, status);
+			if (ret)
+				goto out;
+			continue;
+		}
+		ret = spu_process_events(ctx);
+
+	} while (!ret && !(*status & (SPU_STATUS_STOPPED_BY_STOP |
+				      SPU_STATUS_STOPPED_BY_HALT)));
+
+	ctx->ops->runcntl_stop(ctx);
+	ret = spu_run_fini(ctx, npc, status);
+	if (!ret)
+		ret = *status;
+	spu_yield(ctx);
+
+out:
+	up(&ctx->run_sema);
+	return ret;
+}
+

arch/powerpc/platforms/cell/spufs/sched.c

@@ -0,0 +1,461 @@
+/* sched.c - SPU scheduler.
+ *
+ * Copyright (C) IBM 2005
+ * Author: Mark Nutter <mnutter@us.ibm.com>
+ *
+ * SPU scheduler, based on Linux thread priority.  For now use
+ * a simple "cooperative" yield model with no preemption.  SPU
+ * scheduling will eventually be preemptive: When a thread with
+ * a higher static priority gets ready to run, then an active SPU
+ * context will be preempted and returned to the waitq.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+#undef DEBUG
+
+#include <linux/config.h>
+#include <linux/module.h>
+#include <linux/errno.h>
+#include <linux/sched.h>
+#include <linux/kernel.h>
+#include <linux/mm.h>
+#include <linux/completion.h>
+#include <linux/vmalloc.h>
+#include <linux/smp.h>
+#include <linux/smp_lock.h>
+#include <linux/stddef.h>
+#include <linux/unistd.h>
+
+#include <asm/io.h>
+#include <asm/mmu_context.h>
+#include <asm/spu.h>
+#include <asm/spu_csa.h>
+#include "spufs.h"
+
+#define SPU_MIN_TIMESLICE 	(100 * HZ / 1000)
+
+#define SPU_BITMAP_SIZE (((MAX_PRIO+BITS_PER_LONG)/BITS_PER_LONG)+1)
+struct spu_prio_array {
+	atomic_t nr_blocked;
+	unsigned long bitmap[SPU_BITMAP_SIZE];
+	wait_queue_head_t waitq[MAX_PRIO];
+};
+
+/* spu_runqueue - This is the main runqueue data structure for SPUs. */
+struct spu_runqueue {
+	struct semaphore sem;
+	unsigned long nr_active;
+	unsigned long nr_idle;
+	unsigned long nr_switches;
+	struct list_head active_list;
+	struct list_head idle_list;
+	struct spu_prio_array prio;
+};
+
+static struct spu_runqueue *spu_runqueues = NULL;
+
+static inline struct spu_runqueue *spu_rq(void)
+{
+	/* Future: make this a per-NODE array,
+	 * and use cpu_to_node(smp_processor_id())
+	 */
+	return spu_runqueues;
+}
+
+static inline struct spu *del_idle(struct spu_runqueue *rq)
+{
+	struct spu *spu;
+
+	BUG_ON(rq->nr_idle <= 0);
+	BUG_ON(list_empty(&rq->idle_list));
+	/* Future: Move SPU out of low-power SRI state. */
+	spu = list_entry(rq->idle_list.next, struct spu, sched_list);
+	list_del_init(&spu->sched_list);
+	rq->nr_idle--;
+	return spu;
+}
+
+static inline void del_active(struct spu_runqueue *rq, struct spu *spu)
+{
+	BUG_ON(rq->nr_active <= 0);
+	BUG_ON(list_empty(&rq->active_list));
+	list_del_init(&spu->sched_list);
+	rq->nr_active--;
+}
+
+static inline void add_idle(struct spu_runqueue *rq, struct spu *spu)
+{
+	/* Future: Put SPU into low-power SRI state. */
+	list_add_tail(&spu->sched_list, &rq->idle_list);
+	rq->nr_idle++;
+}
+
+static inline void add_active(struct spu_runqueue *rq, struct spu *spu)
+{
+	rq->nr_active++;
+	rq->nr_switches++;
+	list_add_tail(&spu->sched_list, &rq->active_list);
+}
+
+static void prio_wakeup(struct spu_runqueue *rq)
+{
+	if (atomic_read(&rq->prio.nr_blocked) && rq->nr_idle) {
+		int best = sched_find_first_bit(rq->prio.bitmap);
+		if (best < MAX_PRIO) {
+			wait_queue_head_t *wq = &rq->prio.waitq[best];
+			wake_up_interruptible_nr(wq, 1);
+		}
+	}
+}
+
+static void prio_wait(struct spu_runqueue *rq, struct spu_context *ctx,
+		      u64 flags)
+{
+	int prio = current->prio;
+	wait_queue_head_t *wq = &rq->prio.waitq[prio];
+	DEFINE_WAIT(wait);
+
+	__set_bit(prio, rq->prio.bitmap);
+	atomic_inc(&rq->prio.nr_blocked);
+	prepare_to_wait_exclusive(wq, &wait, TASK_INTERRUPTIBLE);
+	if (!signal_pending(current)) {
+		up(&rq->sem);
+		up_write(&ctx->state_sema);
+		pr_debug("%s: pid=%d prio=%d\n", __FUNCTION__,
+			 current->pid, current->prio);
+		schedule();
+		down_write(&ctx->state_sema);
+		down(&rq->sem);
+	}
+	finish_wait(wq, &wait);
+	atomic_dec(&rq->prio.nr_blocked);
+	if (!waitqueue_active(wq))
+		__clear_bit(prio, rq->prio.bitmap);
+}
+
+static inline int is_best_prio(struct spu_runqueue *rq)
+{
+	int best_prio;
+
+	best_prio = sched_find_first_bit(rq->prio.bitmap);
+	return (current->prio < best_prio) ? 1 : 0;
+}
+
+static inline void mm_needs_global_tlbie(struct mm_struct *mm)
+{
+	/* Global TLBIE broadcast required with SPEs. */
+#if (NR_CPUS > 1)
+	__cpus_setall(&mm->cpu_vm_mask, NR_CPUS);
+#else
+	__cpus_setall(&mm->cpu_vm_mask, NR_CPUS+1); /* is this ok? */
+#endif
+}
+
+static inline void bind_context(struct spu *spu, struct spu_context *ctx)
+{
+	pr_debug("%s: pid=%d SPU=%d\n", __FUNCTION__, current->pid,
+		 spu->number);
+	spu->ctx = ctx;
+	spu->flags = 0;
+	ctx->flags = 0;
+	ctx->spu = spu;
+	ctx->ops = &spu_hw_ops;
+	spu->pid = current->pid;
+	spu->prio = current->prio;
+	spu->mm = ctx->owner;
+	mm_needs_global_tlbie(spu->mm);
+	spu->ibox_callback = spufs_ibox_callback;
+	spu->wbox_callback = spufs_wbox_callback;
+	spu->stop_callback = spufs_stop_callback;
+	mb();
+	spu_unmap_mappings(ctx);
+	spu_restore(&ctx->csa, spu);
+	spu->timestamp = jiffies;
+}
+
+static inline void unbind_context(struct spu *spu, struct spu_context *ctx)
+{
+	pr_debug("%s: unbind pid=%d SPU=%d\n", __FUNCTION__,
+		 spu->pid, spu->number);
+	spu_unmap_mappings(ctx);
+	spu_save(&ctx->csa, spu);
+	spu->timestamp = jiffies;
+	ctx->state = SPU_STATE_SAVED;
+	spu->ibox_callback = NULL;
+	spu->wbox_callback = NULL;
+	spu->stop_callback = NULL;
+	spu->mm = NULL;
+	spu->pid = 0;
+	spu->prio = MAX_PRIO;
+	ctx->ops = &spu_backing_ops;
+	ctx->spu = NULL;
+	ctx->flags = 0;
+	spu->flags = 0;
+	spu->ctx = NULL;
+}
+
+static void spu_reaper(void *data)
+{
+	struct spu_context *ctx = data;
+	struct spu *spu;
+
+	down_write(&ctx->state_sema);
+	spu = ctx->spu;
+	if (spu && test_bit(SPU_CONTEXT_PREEMPT, &ctx->flags)) {
+		if (atomic_read(&spu->rq->prio.nr_blocked)) {
+			pr_debug("%s: spu=%d\n", __func__, spu->number);
+			ctx->ops->runcntl_stop(ctx);
+			spu_deactivate(ctx);
+			wake_up_all(&ctx->stop_wq);
+		} else {
+			clear_bit(SPU_CONTEXT_PREEMPT, &ctx->flags);
+		}
+	}
+	up_write(&ctx->state_sema);
+	put_spu_context(ctx);
+}
+
+static void schedule_spu_reaper(struct spu_runqueue *rq, struct spu *spu)
+{
+	struct spu_context *ctx = get_spu_context(spu->ctx);
+	unsigned long now = jiffies;
+	unsigned long expire = spu->timestamp + SPU_MIN_TIMESLICE;
+
+	set_bit(SPU_CONTEXT_PREEMPT, &ctx->flags);
+	INIT_WORK(&ctx->reap_work, spu_reaper, ctx);
+	if (time_after(now, expire))
+		schedule_work(&ctx->reap_work);
+	else
+		schedule_delayed_work(&ctx->reap_work, expire - now);
+}
+
+static void check_preempt_active(struct spu_runqueue *rq)
+{
+	struct list_head *p;
+	struct spu *worst = NULL;
+
+	list_for_each(p, &rq->active_list) {
+		struct spu *spu = list_entry(p, struct spu, sched_list);
+		struct spu_context *ctx = spu->ctx;
+		if (!test_bit(SPU_CONTEXT_PREEMPT, &ctx->flags)) {
+			if (!worst || (spu->prio > worst->prio)) {
+				worst = spu;
+			}
+		}
+	}
+	if (worst && (current->prio < worst->prio))
+		schedule_spu_reaper(rq, worst);
+}
+
+static struct spu *get_idle_spu(struct spu_context *ctx, u64 flags)
+{
+	struct spu_runqueue *rq;
+	struct spu *spu = NULL;
+
+	rq = spu_rq();
+	down(&rq->sem);
+	for (;;) {
+		if (rq->nr_idle > 0) {
+			if (is_best_prio(rq)) {
+				/* Fall through. */
+				spu = del_idle(rq);
+				break;
+			} else {
+				prio_wakeup(rq);
+				up(&rq->sem);
+				yield();
+				if (signal_pending(current)) {
+					return NULL;
+				}
+				rq = spu_rq();
+				down(&rq->sem);
+				continue;
+			}
+		} else {
+			check_preempt_active(rq);
+			prio_wait(rq, ctx, flags);
+			if (signal_pending(current)) {
+				prio_wakeup(rq);
+				spu = NULL;
+				break;
+			}
+			continue;
+		}
+	}
+	up(&rq->sem);
+	return spu;
+}
+
+static void put_idle_spu(struct spu *spu)
+{
+	struct spu_runqueue *rq = spu->rq;
+
+	down(&rq->sem);
+	add_idle(rq, spu);
+	prio_wakeup(rq);
+	up(&rq->sem);
+}
+
+static int get_active_spu(struct spu *spu)
+{
+	struct spu_runqueue *rq = spu->rq;
+	struct list_head *p;
+	struct spu *tmp;
+	int rc = 0;
+
+	down(&rq->sem);
+	list_for_each(p, &rq->active_list) {
+		tmp = list_entry(p, struct spu, sched_list);
+		if (tmp == spu) {
+			del_active(rq, spu);
+			rc = 1;
+			break;
+		}
+	}
+	up(&rq->sem);
+	return rc;
+}
+
+static void put_active_spu(struct spu *spu)
+{
+	struct spu_runqueue *rq = spu->rq;
+
+	down(&rq->sem);
+	add_active(rq, spu);
+	up(&rq->sem);
+}
+
+/* Lock order:
+ *	spu_activate() & spu_deactivate() require the
+ *	caller to have down_write(&ctx->state_sema).
+ *
+ *	The rq->sem is breifly held (inside or outside a
+ *	given ctx lock) for list management, but is never
+ *	held during save/restore.
+ */
+
+int spu_activate(struct spu_context *ctx, u64 flags)
+{
+	struct spu *spu;
+
+	if (ctx->spu)
+		return 0;
+	spu = get_idle_spu(ctx, flags);
+	if (!spu)
+		return (signal_pending(current)) ? -ERESTARTSYS : -EAGAIN;
+	bind_context(spu, ctx);
+	/*
+	 * We're likely to wait for interrupts on the same
+	 * CPU that we are now on, so send them here.
+	 */
+	spu_irq_setaffinity(spu, raw_smp_processor_id());
+	put_active_spu(spu);
+	return 0;
+}
+
+void spu_deactivate(struct spu_context *ctx)
+{
+	struct spu *spu;
+	int needs_idle;
+
+	spu = ctx->spu;
+	if (!spu)
+		return;
+	needs_idle = get_active_spu(spu);
+	unbind_context(spu, ctx);
+	if (needs_idle)
+		put_idle_spu(spu);
+}
+
+void spu_yield(struct spu_context *ctx)
+{
+	struct spu *spu;
+	int need_yield = 0;
+
+	down_write(&ctx->state_sema);
+	spu = ctx->spu;
+	if (spu && (sched_find_first_bit(spu->rq->prio.bitmap) < MAX_PRIO)) {
+		pr_debug("%s: yielding SPU %d\n", __FUNCTION__, spu->number);
+		spu_deactivate(ctx);
+		ctx->state = SPU_STATE_SAVED;
+		need_yield = 1;
+	} else if (spu) {
+		spu->prio = MAX_PRIO;
+	}
+	up_write(&ctx->state_sema);
+	if (unlikely(need_yield))
+		yield();
+}
+
+int __init spu_sched_init(void)
+{
+	struct spu_runqueue *rq;
+	struct spu *spu;
+	int i;
+
+	rq = spu_runqueues = kmalloc(sizeof(struct spu_runqueue), GFP_KERNEL);
+	if (!rq) {
+		printk(KERN_WARNING "%s: Unable to allocate runqueues.\n",
+		       __FUNCTION__);
+		return 1;
+	}
+	memset(rq, 0, sizeof(struct spu_runqueue));
+	init_MUTEX(&rq->sem);
+	INIT_LIST_HEAD(&rq->active_list);
+	INIT_LIST_HEAD(&rq->idle_list);
+	rq->nr_active = 0;
+	rq->nr_idle = 0;
+	rq->nr_switches = 0;
+	atomic_set(&rq->prio.nr_blocked, 0);
+	for (i = 0; i < MAX_PRIO; i++) {
+		init_waitqueue_head(&rq->prio.waitq[i]);
+		__clear_bit(i, rq->prio.bitmap);
+	}
+	__set_bit(MAX_PRIO, rq->prio.bitmap);
+	for (;;) {
+		spu = spu_alloc();
+		if (!spu)
+			break;
+		pr_debug("%s: adding SPU[%d]\n", __FUNCTION__, spu->number);
+		add_idle(rq, spu);
+		spu->rq = rq;
+		spu->timestamp = jiffies;
+	}
+	if (!rq->nr_idle) {
+		printk(KERN_WARNING "%s: No available SPUs.\n", __FUNCTION__);
+		kfree(rq);
+		return 1;
+	}
+	return 0;
+}
+
+void __exit spu_sched_exit(void)
+{
+	struct spu_runqueue *rq = spu_rq();
+	struct spu *spu;
+
+	if (!rq) {
+		printk(KERN_WARNING "%s: no runqueues!\n", __FUNCTION__);
+		return;
+	}
+	while (rq->nr_idle > 0) {
+		spu = del_idle(rq);
+		if (!spu)
+			break;
+		spu_free(spu);
+	}
+	kfree(rq);
+}

arch/powerpc/platforms/cell/spufs/spu_restore.c

@@ -0,0 +1,336 @@
+/*
+ * spu_restore.c
+ *
+ * (C) Copyright IBM Corp. 2005
+ *
+ * SPU-side context restore sequence outlined in
+ * Synergistic Processor Element Book IV
+ *
+ * Author: Mark Nutter <mnutter@us.ibm.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ *
+ */
+
+
+#ifndef LS_SIZE
+#define LS_SIZE                 0x40000	/* 256K (in bytes) */
+#endif
+
+typedef unsigned int u32;
+typedef unsigned long long u64;
+
+#include <spu_intrinsics.h>
+#include <asm/spu_csa.h>
+#include "spu_utils.h"
+
+#define BR_INSTR		0x327fff80	/* br -4         */
+#define NOP_INSTR		0x40200000	/* nop           */
+#define HEQ_INSTR		0x7b000000	/* heq $0, $0    */
+#define STOP_INSTR		0x00000000	/* stop 0x0      */
+#define ILLEGAL_INSTR		0x00800000	/* illegal instr */
+#define RESTORE_COMPLETE	0x00003ffc	/* stop 0x3ffc   */
+
+static inline void fetch_regs_from_mem(addr64 lscsa_ea)
+{
+	unsigned int ls = (unsigned int)&regs_spill[0];
+	unsigned int size = sizeof(regs_spill);
+	unsigned int tag_id = 0;
+	unsigned int cmd = 0x40;	/* GET */
+
+	spu_writech(MFC_LSA, ls);
+	spu_writech(MFC_EAH, lscsa_ea.ui[0]);
+	spu_writech(MFC_EAL, lscsa_ea.ui[1]);
+	spu_writech(MFC_Size, size);
+	spu_writech(MFC_TagID, tag_id);
+	spu_writech(MFC_Cmd, cmd);
+}
+
+static inline void restore_upper_240kb(addr64 lscsa_ea)
+{
+	unsigned int ls = 16384;
+	unsigned int list = (unsigned int)&dma_list[0];
+	unsigned int size = sizeof(dma_list);
+	unsigned int tag_id = 0;
+	unsigned int cmd = 0x44;	/* GETL */
+
+	/* Restore, Step 4:
+	 *    Enqueue the GETL command (tag 0) to the MFC SPU command
+	 *    queue to transfer the upper 240 kb of LS from CSA.
+	 */
+	spu_writech(MFC_LSA, ls);
+	spu_writech(MFC_EAH, lscsa_ea.ui[0]);
+	spu_writech(MFC_EAL, list);
+	spu_writech(MFC_Size, size);
+	spu_writech(MFC_TagID, tag_id);
+	spu_writech(MFC_Cmd, cmd);
+}
+
+static inline void restore_decr(void)
+{
+	unsigned int offset;
+	unsigned int decr_running;
+	unsigned int decr;
+
+	/* Restore, Step 6:
+	 *    If the LSCSA "decrementer running" flag is set
+	 *    then write the SPU_WrDec channel with the
+	 *    decrementer value from LSCSA.
+	 */
+	offset = LSCSA_QW_OFFSET(decr_status);
+	decr_running = regs_spill[offset].slot[0];
+	if (decr_running) {
+		offset = LSCSA_QW_OFFSET(decr);
+		decr = regs_spill[offset].slot[0];
+		spu_writech(SPU_WrDec, decr);
+	}
+}
+
+static inline void write_ppu_mb(void)
+{
+	unsigned int offset;
+	unsigned int data;
+
+	/* Restore, Step 11:
+	 *    Write the MFC_WrOut_MB channel with the PPU_MB
+	 *    data from LSCSA.
+	 */
+	offset = LSCSA_QW_OFFSET(ppu_mb);
+	data = regs_spill[offset].slot[0];
+	spu_writech(SPU_WrOutMbox, data);
+}
+
+static inline void write_ppuint_mb(void)
+{
+	unsigned int offset;
+	unsigned int data;
+
+	/* Restore, Step 12:
+	 *    Write the MFC_WrInt_MB channel with the PPUINT_MB
+	 *    data from LSCSA.
+	 */
+	offset = LSCSA_QW_OFFSET(ppuint_mb);
+	data = regs_spill[offset].slot[0];
+	spu_writech(SPU_WrOutIntrMbox, data);
+}
+
+static inline void restore_fpcr(void)
+{
+	unsigned int offset;
+	vector unsigned int fpcr;
+
+	/* Restore, Step 13:
+	 *    Restore the floating-point status and control
+	 *    register from the LSCSA.
+	 */
+	offset = LSCSA_QW_OFFSET(fpcr);
+	fpcr = regs_spill[offset].v;
+	spu_mtfpscr(fpcr);
+}
+
+static inline void restore_srr0(void)
+{
+	unsigned int offset;
+	unsigned int srr0;
+
+	/* Restore, Step 14:
+	 *    Restore the SPU SRR0 data from the LSCSA.
+	 */
+	offset = LSCSA_QW_OFFSET(srr0);
+	srr0 = regs_spill[offset].slot[0];
+	spu_writech(SPU_WrSRR0, srr0);
+}
+
+static inline void restore_event_mask(void)
+{
+	unsigned int offset;
+	unsigned int event_mask;
+
+	/* Restore, Step 15:
+	 *    Restore the SPU_RdEventMsk data from the LSCSA.
+	 */
+	offset = LSCSA_QW_OFFSET(event_mask);
+	event_mask = regs_spill[offset].slot[0];
+	spu_writech(SPU_WrEventMask, event_mask);
+}
+
+static inline void restore_tag_mask(void)
+{
+	unsigned int offset;
+	unsigned int tag_mask;
+
+	/* Restore, Step 16:
+	 *    Restore the SPU_RdTagMsk data from the LSCSA.
+	 */
+	offset = LSCSA_QW_OFFSET(tag_mask);
+	tag_mask = regs_spill[offset].slot[0];
+	spu_writech(MFC_WrTagMask, tag_mask);
+}
+
+static inline void restore_complete(void)
+{
+	extern void exit_fini(void);
+	unsigned int *exit_instrs = (unsigned int *)exit_fini;
+	unsigned int offset;
+	unsigned int stopped_status;
+	unsigned int stopped_code;
+
+	/* Restore, Step 18:
+	 *    Issue a stop-and-signal instruction with
+	 *    "good context restore" signal value.
+	 *
+	 * Restore, Step 19:
+	 *    There may be additional instructions placed
+	 *    here by the PPE Sequence for SPU Context
+	 *    Restore in order to restore the correct
+	 *    "stopped state".
+	 *
+	 *    This step is handled here by analyzing the
+	 *    LSCSA.stopped_status and then modifying the
+	 *    exit() function to behave appropriately.
+	 */
+
+	offset = LSCSA_QW_OFFSET(stopped_status);
+	stopped_status = regs_spill[offset].slot[0];
+	stopped_code = regs_spill[offset].slot[1];
+
+	switch (stopped_status) {
+	case SPU_STOPPED_STATUS_P_I:
+		/* SPU_Status[P,I]=1.  Add illegal instruction
+		 * followed by stop-and-signal instruction after
+		 * end of restore code.
+		 */
+		exit_instrs[0] = RESTORE_COMPLETE;
+		exit_instrs[1] = ILLEGAL_INSTR;
+		exit_instrs[2] = STOP_INSTR | stopped_code;
+		break;
+	case SPU_STOPPED_STATUS_P_H:
+		/* SPU_Status[P,H]=1.  Add 'heq $0, $0' followed
+		 * by stop-and-signal instruction after end of
+		 * restore code.
+		 */
+		exit_instrs[0] = RESTORE_COMPLETE;
+		exit_instrs[1] = HEQ_INSTR;
+		exit_instrs[2] = STOP_INSTR | stopped_code;
+		break;
+	case SPU_STOPPED_STATUS_S_P:
+		/* SPU_Status[S,P]=1.  Add nop instruction
+		 * followed by 'br -4' after end of restore
+		 * code.
+		 */
+		exit_instrs[0] = RESTORE_COMPLETE;
+		exit_instrs[1] = STOP_INSTR | stopped_code;
+		exit_instrs[2] = NOP_INSTR;
+		exit_instrs[3] = BR_INSTR;
+		break;
+	case SPU_STOPPED_STATUS_S_I:
+		/* SPU_Status[S,I]=1.  Add  illegal instruction
+		 * followed by 'br -4' after end of restore code.
+		 */
+		exit_instrs[0] = RESTORE_COMPLETE;
+		exit_instrs[1] = ILLEGAL_INSTR;
+		exit_instrs[2] = NOP_INSTR;
+		exit_instrs[3] = BR_INSTR;
+		break;
+	case SPU_STOPPED_STATUS_I:
+		/* SPU_Status[I]=1. Add illegal instruction followed
+		 * by infinite loop after end of restore sequence.
+		 */
+		exit_instrs[0] = RESTORE_COMPLETE;
+		exit_instrs[1] = ILLEGAL_INSTR;
+		exit_instrs[2] = NOP_INSTR;
+		exit_instrs[3] = BR_INSTR;
+		break;
+	case SPU_STOPPED_STATUS_S:
+		/* SPU_Status[S]=1. Add two 'nop' instructions. */
+		exit_instrs[0] = RESTORE_COMPLETE;
+		exit_instrs[1] = NOP_INSTR;
+		exit_instrs[2] = NOP_INSTR;
+		exit_instrs[3] = BR_INSTR;
+		break;
+	case SPU_STOPPED_STATUS_H:
+		/* SPU_Status[H]=1. Add 'heq $0, $0' instruction
+		 * after end of restore code.
+		 */
+		exit_instrs[0] = RESTORE_COMPLETE;
+		exit_instrs[1] = HEQ_INSTR;
+		exit_instrs[2] = NOP_INSTR;
+		exit_instrs[3] = BR_INSTR;
+		break;
+	case SPU_STOPPED_STATUS_P:
+		/* SPU_Status[P]=1. Add stop-and-signal instruction
+		 * after end of restore code.
+		 */
+		exit_instrs[0] = RESTORE_COMPLETE;
+		exit_instrs[1] = STOP_INSTR | stopped_code;
+		break;
+	case SPU_STOPPED_STATUS_R:
+		/* SPU_Status[I,S,H,P,R]=0. Add infinite loop. */
+		exit_instrs[0] = RESTORE_COMPLETE;
+		exit_instrs[1] = NOP_INSTR;
+		exit_instrs[2] = NOP_INSTR;
+		exit_instrs[3] = BR_INSTR;
+		break;
+	default:
+		/* SPU_Status[R]=1. No additonal instructions. */
+		break;
+	}
+	spu_sync();
+}
+
+/**
+ * main - entry point for SPU-side context restore.
+ *
+ * This code deviates from the documented sequence in the
+ * following aspects:
+ *
+ * 	1. The EA for LSCSA is passed from PPE in the
+ *	   signal notification channels.
+ *	2. The register spill area is pulled by SPU
+ *	   into LS, rather than pushed by PPE.
+ *	3. All 128 registers are restored by exit().
+ *	4. The exit() function is modified at run
+ *	   time in order to properly restore the
+ *	   SPU_Status register.
+ */
+int main()
+{
+	addr64 lscsa_ea;
+
+	lscsa_ea.ui[0] = spu_readch(SPU_RdSigNotify1);
+	lscsa_ea.ui[1] = spu_readch(SPU_RdSigNotify2);
+	fetch_regs_from_mem(lscsa_ea);
+
+	set_event_mask();		/* Step 1.  */
+	set_tag_mask();			/* Step 2.  */
+	build_dma_list(lscsa_ea);	/* Step 3.  */
+	restore_upper_240kb(lscsa_ea);	/* Step 4.  */
+					/* Step 5: done by 'exit'. */
+	restore_decr();			/* Step 6. */
+	enqueue_putllc(lscsa_ea);	/* Step 7. */
+	set_tag_update();		/* Step 8. */
+	read_tag_status();		/* Step 9. */
+	read_llar_status();		/* Step 10. */
+	write_ppu_mb();			/* Step 11. */
+	write_ppuint_mb();		/* Step 12. */
+	restore_fpcr();			/* Step 13. */
+	restore_srr0();			/* Step 14. */
+	restore_event_mask();		/* Step 15. */
+	restore_tag_mask();		/* Step 16. */
+					/* Step 17. done by 'exit'. */
+	restore_complete();		/* Step 18. */
+
+	return 0;
+}

arch/powerpc/platforms/cell/spufs/spu_restore_crt0.S

@@ -0,0 +1,116 @@
+/*
+ * crt0_r.S: Entry function for SPU-side context restore.
+ *
+ * Copyright (C) 2005 IBM
+ *
+ * Entry and exit function for SPU-side of the context restore
+ * sequence.  Sets up an initial stack frame, then branches to
+ * 'main'.  On return, restores all 128 registers from the LSCSA
+ * and exits.
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+#include <asm/spu_csa.h>
+
+.data
+.align 7
+.globl regs_spill
+regs_spill:
+.space SIZEOF_SPU_SPILL_REGS, 0x0
+
+.text
+.global _start
+_start:
+	/* Initialize the stack pointer to point to 16368
+	 * (16kb-16). The back chain pointer is initialized
+	 * to NULL.
+	 */
+	il      $0, 0
+	il      $SP, 16368
+	stqd    $0, 0($SP)
+
+	/* Allocate a minimum stack frame for the called main.
+	 * This is needed so that main has a place to save the
+	 * link register when it calls another function.
+	 */
+	stqd    $SP, -160($SP)
+	ai      $SP, $SP, -160
+
+	/* Call the program's main function. */
+	brsl    $0, main
+
+.global exit
+.global	_exit
+exit:
+_exit:
+	/* SPU Context Restore, Step 5: Restore the remaining 112 GPRs. */
+	ila     $3, regs_spill + 256
+restore_regs:
+	lqr     $4, restore_reg_insts
+restore_reg_loop:
+	ai      $4, $4, 4
+	.balignl 16, 0x40200000
+restore_reg_insts:       /* must be quad-word aligned. */
+	lqd     $16, 0($3)
+	lqd     $17, 16($3)
+	lqd     $18, 32($3)
+	lqd     $19, 48($3)
+	andi    $5, $4, 0x7F
+	stqr    $4, restore_reg_insts
+	ai      $3, $3, 64
+	brnz    $5, restore_reg_loop
+
+	/* SPU Context Restore Step 17: Restore the first 16 GPRs. */
+	lqa $0, regs_spill + 0
+	lqa $1, regs_spill + 16
+	lqa $2, regs_spill + 32
+	lqa $3, regs_spill + 48
+	lqa $4, regs_spill + 64
+	lqa $5, regs_spill + 80
+	lqa $6, regs_spill + 96
+	lqa $7, regs_spill + 112
+	lqa $8, regs_spill + 128
+	lqa $9, regs_spill + 144
+	lqa $10, regs_spill + 160
+	lqa $11, regs_spill + 176
+	lqa $12, regs_spill + 192
+	lqa $13, regs_spill + 208
+	lqa $14, regs_spill + 224
+	lqa $15, regs_spill + 240
+
+	/* Under normal circumstances, the 'exit' function
+	 * terminates with 'stop SPU_RESTORE_COMPLETE',
+	 * indicating that the SPU-side restore code has
+	 * completed.
+	 *
+	 * However it is possible that instructions immediately
+	 * following the 'stop 0x3ffc' have been modified at run
+	 * time so as to recreate the exact SPU_Status settings
+	 * from the application, e.g. illegal instruciton, halt,
+	 * etc.
+	 */
+.global exit_fini
+.global	_exit_fini
+exit_fini:
+_exit_fini:
+	stop	SPU_RESTORE_COMPLETE
+	stop	0
+	stop	0
+	stop	0
+
+	/* Pad the size of this crt0.o to be multiple of 16 bytes. */
+.balignl 16, 0x0

arch/powerpc/platforms/cell/spufs/spu_restore_dump.h_shipped

@@ -0,0 +1,231 @@
+/*
+ * spu_restore_dump.h: Copyright (C) 2005 IBM.
+ * Hex-dump auto generated from spu_restore.c.
+ * Do not edit!
+ */
+static unsigned int spu_restore_code[] __page_aligned = {
+0x40800000, 0x409ff801, 0x24000080, 0x24fd8081,
+0x1cd80081, 0x33001180, 0x42030003, 0x33800284,
+0x1c010204, 0x40200000, 0x40200000, 0x40200000,
+0x34000190, 0x34004191, 0x34008192, 0x3400c193,
+0x141fc205, 0x23fffd84, 0x1c100183, 0x217ffa85,
+0x3080a000, 0x3080a201, 0x3080a402, 0x3080a603,
+0x3080a804, 0x3080aa05, 0x3080ac06, 0x3080ae07,
+0x3080b008, 0x3080b209, 0x3080b40a, 0x3080b60b,
+0x3080b80c, 0x3080ba0d, 0x3080bc0e, 0x3080be0f,
+0x00003ffc, 0x00000000, 0x00000000, 0x00000000,
+0x01a00182, 0x3ec00083, 0xb0a14103, 0x01a00204,
+0x3ec10082, 0x4202800e, 0x04000703, 0xb0a14202,
+0x21a00803, 0x3fbf028d, 0x3f20068d, 0x3fbe0682,
+0x3fe30102, 0x21a00882, 0x3f82028f, 0x3fe3078f,
+0x3fbf0784, 0x3f200204, 0x3fbe0204, 0x3fe30204,
+0x04000203, 0x21a00903, 0x40848002, 0x21a00982,
+0x40800003, 0x21a00a03, 0x40802002, 0x21a00a82,
+0x21a00083, 0x40800082, 0x21a00b02, 0x10002818,
+0x40a80002, 0x32800007, 0x4207000c, 0x18008208,
+0x40a0000b, 0x4080020a, 0x40800709, 0x00200000,
+0x42070002, 0x3ac30384, 0x1cffc489, 0x00200000,
+0x18008383, 0x38830382, 0x4cffc486, 0x3ac28185,
+0xb0408584, 0x28830382, 0x1c020387, 0x38828182,
+0xb0408405, 0x1802c408, 0x28828182, 0x217ff886,
+0x04000583, 0x21a00803, 0x3fbe0682, 0x3fe30102,
+0x04000106, 0x21a00886, 0x04000603, 0x21a00903,
+0x40803c02, 0x21a00982, 0x40800003, 0x04000184,
+0x21a00a04, 0x40802202, 0x21a00a82, 0x42028005,
+0x34208702, 0x21002282, 0x21a00804, 0x21a00886,
+0x3fbf0782, 0x3f200102, 0x3fbe0102, 0x3fe30102,
+0x21a00902, 0x40804003, 0x21a00983, 0x21a00a04,
+0x40805a02, 0x21a00a82, 0x40800083, 0x21a00b83,
+0x01a00c02, 0x01a00d83, 0x3420c282, 0x21a00e02,
+0x34210283, 0x21a00f03, 0x34200284, 0x77400200,
+0x3421c282, 0x21a00702, 0x34218283, 0x21a00083,
+0x34214282, 0x21a00b02, 0x4200480c, 0x00200000,
+0x1c010286, 0x34220284, 0x34220302, 0x0f608203,
+0x5c024204, 0x3b81810b, 0x42013c02, 0x00200000,
+0x18008185, 0x38808183, 0x3b814182, 0x21004e84,
+0x4020007f, 0x35000100, 0x000004e0, 0x000002a0,
+0x000002e8, 0x00000428, 0x00000360, 0x000002e8,
+0x000004a0, 0x00000468, 0x000003c8, 0x00000360,
+0x409ffe02, 0x30801203, 0x40800204, 0x3ec40085,
+0x10009c09, 0x3ac10606, 0xb060c105, 0x4020007f,
+0x4020007f, 0x20801203, 0x38810602, 0xb0408586,
+0x28810602, 0x32004180, 0x34204702, 0x21a00382,
+0x4020007f, 0x327fdc80, 0x409ffe02, 0x30801203,
+0x40800204, 0x3ec40087, 0x40800405, 0x00200000,
+0x40800606, 0x3ac10608, 0x3ac14609, 0x3ac1860a,
+0xb060c107, 0x20801203, 0x41004003, 0x38810602,
+0x4020007f, 0xb0408188, 0x4020007f, 0x28810602,
+0x41201002, 0x38814603, 0x10009c09, 0xb060c109,
+0x4020007f, 0x28814603, 0x41193f83, 0x38818602,
+0x60ffc003, 0xb040818a, 0x28818602, 0x32003080,
+0x409ffe02, 0x30801203, 0x40800204, 0x3ec40087,
+0x41201008, 0x10009c14, 0x40800405, 0x3ac10609,
+0x40800606, 0x3ac1460a, 0xb060c107, 0x3ac1860b,
+0x20801203, 0x38810602, 0xb0408409, 0x28810602,
+0x38814603, 0xb060c40a, 0x4020007f, 0x28814603,
+0x41193f83, 0x38818602, 0x60ffc003, 0xb040818b,
+0x28818602, 0x32002380, 0x409ffe02, 0x30801204,
+0x40800205, 0x3ec40083, 0x40800406, 0x3ac14607,
+0x3ac18608, 0xb0810103, 0x41004002, 0x20801204,
+0x4020007f, 0x38814603, 0x10009c0b, 0xb060c107,
+0x4020007f, 0x4020007f, 0x28814603, 0x38818602,
+0x4020007f, 0x4020007f, 0xb0408588, 0x28818602,
+0x4020007f, 0x32001780, 0x409ffe02, 0x1000640e,
+0x40800204, 0x30801203, 0x40800405, 0x3ec40087,
+0x40800606, 0x3ac10608, 0x3ac14609, 0x3ac1860a,
+0xb060c107, 0x20801203, 0x413d8003, 0x38810602,
+0x4020007f, 0x327fd780, 0x409ffe02, 0x10007f0c,
+0x40800205, 0x30801204, 0x40800406, 0x3ec40083,
+0x3ac14607, 0x3ac18608, 0xb0810103, 0x413d8002,
+0x20801204, 0x38814603, 0x4020007f, 0x327feb80,
+0x409ffe02, 0x30801203, 0x40800204, 0x3ec40087,
+0x40800405, 0x1000650a, 0x40800606, 0x3ac10608,
+0x3ac14609, 0x3ac1860a, 0xb060c107, 0x20801203,
+0x38810602, 0xb0408588, 0x4020007f, 0x327fc980,
+0x00400000, 0x40800003, 0x4020007f, 0x35000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+};

arch/powerpc/platforms/cell/spufs/spu_save.c

@@ -0,0 +1,195 @@
+/*
+ * spu_save.c
+ *
+ * (C) Copyright IBM Corp. 2005
+ *
+ * SPU-side context save sequence outlined in
+ * Synergistic Processor Element Book IV
+ *
+ * Author: Mark Nutter <mnutter@us.ibm.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ *
+ */
+
+
+#ifndef LS_SIZE
+#define LS_SIZE                 0x40000	/* 256K (in bytes) */
+#endif
+
+typedef unsigned int u32;
+typedef unsigned long long u64;
+
+#include <spu_intrinsics.h>
+#include <asm/spu_csa.h>
+#include "spu_utils.h"
+
+static inline void save_event_mask(void)
+{
+	unsigned int offset;
+
+	/* Save, Step 2:
+	 *    Read the SPU_RdEventMsk channel and save to the LSCSA.
+	 */
+	offset = LSCSA_QW_OFFSET(event_mask);
+	regs_spill[offset].slot[0] = spu_readch(SPU_RdEventStatMask);
+}
+
+static inline void save_tag_mask(void)
+{
+	unsigned int offset;
+
+	/* Save, Step 3:
+	 *    Read the SPU_RdTagMsk channel and save to the LSCSA.
+	 */
+	offset = LSCSA_QW_OFFSET(tag_mask);
+	regs_spill[offset].slot[0] = spu_readch(MFC_RdTagMask);
+}
+
+static inline void save_upper_240kb(addr64 lscsa_ea)
+{
+	unsigned int ls = 16384;
+	unsigned int list = (unsigned int)&dma_list[0];
+	unsigned int size = sizeof(dma_list);
+	unsigned int tag_id = 0;
+	unsigned int cmd = 0x24;	/* PUTL */
+
+	/* Save, Step 7:
+	 *    Enqueue the PUTL command (tag 0) to the MFC SPU command
+	 *    queue to transfer the remaining 240 kb of LS to CSA.
+	 */
+	spu_writech(MFC_LSA, ls);
+	spu_writech(MFC_EAH, lscsa_ea.ui[0]);
+	spu_writech(MFC_EAL, list);
+	spu_writech(MFC_Size, size);
+	spu_writech(MFC_TagID, tag_id);
+	spu_writech(MFC_Cmd, cmd);
+}
+
+static inline void save_fpcr(void)
+{
+	// vector unsigned int fpcr;
+	unsigned int offset;
+
+	/* Save, Step 9:
+	 *    Issue the floating-point status and control register
+	 *    read instruction, and save to the LSCSA.
+	 */
+	offset = LSCSA_QW_OFFSET(fpcr);
+	regs_spill[offset].v = spu_mffpscr();
+}
+
+static inline void save_decr(void)
+{
+	unsigned int offset;
+
+	/* Save, Step 10:
+	 *    Read and save the SPU_RdDec channel data to
+	 *    the LSCSA.
+	 */
+	offset = LSCSA_QW_OFFSET(decr);
+	regs_spill[offset].slot[0] = spu_readch(SPU_RdDec);
+}
+
+static inline void save_srr0(void)
+{
+	unsigned int offset;
+
+	/* Save, Step 11:
+	 *    Read and save the SPU_WSRR0 channel data to
+	 *    the LSCSA.
+	 */
+	offset = LSCSA_QW_OFFSET(srr0);
+	regs_spill[offset].slot[0] = spu_readch(SPU_RdSRR0);
+}
+
+static inline void spill_regs_to_mem(addr64 lscsa_ea)
+{
+	unsigned int ls = (unsigned int)&regs_spill[0];
+	unsigned int size = sizeof(regs_spill);
+	unsigned int tag_id = 0;
+	unsigned int cmd = 0x20;	/* PUT */
+
+	/* Save, Step 13:
+	 *    Enqueue a PUT command (tag 0) to send the LSCSA
+	 *    to the CSA.
+	 */
+	spu_writech(MFC_LSA, ls);
+	spu_writech(MFC_EAH, lscsa_ea.ui[0]);
+	spu_writech(MFC_EAL, lscsa_ea.ui[1]);
+	spu_writech(MFC_Size, size);
+	spu_writech(MFC_TagID, tag_id);
+	spu_writech(MFC_Cmd, cmd);
+}
+
+static inline void enqueue_sync(addr64 lscsa_ea)
+{
+	unsigned int tag_id = 0;
+	unsigned int cmd = 0xCC;
+
+	/* Save, Step 14:
+	 *    Enqueue an MFC_SYNC command (tag 0).
+	 */
+	spu_writech(MFC_TagID, tag_id);
+	spu_writech(MFC_Cmd, cmd);
+}
+
+static inline void save_complete(void)
+{
+	/* Save, Step 18:
+	 *    Issue a stop-and-signal instruction indicating
+	 *    "save complete".  Note: This function will not
+	 *    return!!
+	 */
+	spu_stop(SPU_SAVE_COMPLETE);
+}
+
+/**
+ * main - entry point for SPU-side context save.
+ *
+ * This code deviates from the documented sequence as follows:
+ *
+ *      1. The EA for LSCSA is passed from PPE in the
+ *         signal notification channels.
+ *      2. All 128 registers are saved by crt0.o.
+ */
+int main()
+{
+	addr64 lscsa_ea;
+
+	lscsa_ea.ui[0] = spu_readch(SPU_RdSigNotify1);
+	lscsa_ea.ui[1] = spu_readch(SPU_RdSigNotify2);
+
+	/* Step 1: done by exit(). */
+	save_event_mask();	/* Step 2.  */
+	save_tag_mask();	/* Step 3.  */
+	set_event_mask();	/* Step 4.  */
+	set_tag_mask();		/* Step 5.  */
+	build_dma_list(lscsa_ea);	/* Step 6.  */
+	save_upper_240kb(lscsa_ea);	/* Step 7.  */
+	/* Step 8: done by exit(). */
+	save_fpcr();		/* Step 9.  */
+	save_decr();		/* Step 10. */
+	save_srr0();		/* Step 11. */
+	enqueue_putllc(lscsa_ea);	/* Step 12. */
+	spill_regs_to_mem(lscsa_ea);	/* Step 13. */
+	enqueue_sync(lscsa_ea);	/* Step 14. */
+	set_tag_update();	/* Step 15. */
+	read_tag_status();	/* Step 16. */
+	read_llar_status();	/* Step 17. */
+	save_complete();	/* Step 18. */
+
+	return 0;
+}

arch/powerpc/platforms/cell/spufs/spu_save_crt0.S

@@ -0,0 +1,102 @@
+/*
+ * crt0_s.S: Entry function for SPU-side context save.
+ *
+ * Copyright (C) 2005 IBM
+ *
+ * Entry function for SPU-side of the context save sequence.
+ * Saves all 128 GPRs, sets up an initial stack frame, then
+ * branches to 'main'.
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+#include <asm/spu_csa.h>
+
+.data
+.align 7
+.globl regs_spill
+regs_spill:
+.space SIZEOF_SPU_SPILL_REGS, 0x0
+
+.text
+.global _start
+_start:
+	/* SPU Context Save Step 1: Save the first 16 GPRs. */
+	stqa $0, regs_spill + 0
+	stqa $1, regs_spill + 16
+	stqa $2, regs_spill + 32
+	stqa $3, regs_spill + 48
+	stqa $4, regs_spill + 64
+	stqa $5, regs_spill + 80
+	stqa $6, regs_spill + 96
+	stqa $7, regs_spill + 112
+	stqa $8, regs_spill + 128
+	stqa $9, regs_spill + 144
+	stqa $10, regs_spill + 160
+	stqa $11, regs_spill + 176
+	stqa $12, regs_spill + 192
+	stqa $13, regs_spill + 208
+	stqa $14, regs_spill + 224
+	stqa $15, regs_spill + 240
+
+	/* SPU Context Save, Step 8: Save the remaining 112 GPRs. */
+	ila     $3, regs_spill + 256
+save_regs:
+	lqr     $4, save_reg_insts
+save_reg_loop:
+	ai      $4, $4, 4
+	.balignl 16, 0x40200000
+save_reg_insts:       /* must be quad-word aligned. */
+	stqd    $16, 0($3)
+	stqd    $17, 16($3)
+	stqd    $18, 32($3)
+	stqd    $19, 48($3)
+	andi    $5, $4, 0x7F
+	stqr    $4, save_reg_insts
+	ai      $3, $3, 64
+	brnz    $5, save_reg_loop
+
+	/* Initialize the stack pointer to point to 16368
+	 * (16kb-16). The back chain pointer is initialized
+	 * to NULL.
+	 */
+	il	$0, 0
+	il	$SP, 16368
+	stqd	$0, 0($SP)
+
+	/* Allocate a minimum stack frame for the called main.
+	 * This is needed so that main has a place to save the
+	 * link register when it calls another function.
+	 */
+	stqd	$SP, -160($SP)
+	ai	$SP, $SP, -160
+
+	/* Call the program's main function. */
+	brsl	$0, main
+
+	/* In this case main should not return; if it does
+	 * there has been an error in the sequence.  Execute
+	 * stop-and-signal with code=0.
+	 */
+.global exit
+.global	_exit
+exit:
+_exit:
+	stop	0x0
+
+	/* Pad the size of this crt0.o to be multiple of 16 bytes. */
+.balignl 16, 0x0
+

arch/powerpc/platforms/cell/spufs/spu_save_dump.h_shipped

@@ -0,0 +1,191 @@
+/*
+ * spu_save_dump.h: Copyright (C) 2005 IBM.
+ * Hex-dump auto generated from spu_save.c.
+ * Do not edit!
+ */
+static unsigned int spu_save_code[] __page_aligned = {
+0x20805000, 0x20805201, 0x20805402, 0x20805603,
+0x20805804, 0x20805a05, 0x20805c06, 0x20805e07,
+0x20806008, 0x20806209, 0x2080640a, 0x2080660b,
+0x2080680c, 0x20806a0d, 0x20806c0e, 0x20806e0f,
+0x4201c003, 0x33800184, 0x1c010204, 0x40200000,
+0x24000190, 0x24004191, 0x24008192, 0x2400c193,
+0x141fc205, 0x23fffd84, 0x1c100183, 0x217ffb85,
+0x40800000, 0x409ff801, 0x24000080, 0x24fd8081,
+0x1cd80081, 0x33000180, 0x00000000, 0x00000000,
+0x01a00182, 0x3ec00083, 0xb1c38103, 0x01a00204,
+0x3ec10082, 0x4201400d, 0xb1c38202, 0x01a00583,
+0x34218682, 0x3ed80684, 0xb0408184, 0x24218682,
+0x01a00603, 0x00200000, 0x34214682, 0x3ed40684,
+0xb0408184, 0x40800003, 0x24214682, 0x21a00083,
+0x40800082, 0x21a00b02, 0x4020007f, 0x1000251e,
+0x40a80002, 0x32800008, 0x4205c00c, 0x00200000,
+0x40a0000b, 0x3f82070f, 0x4080020a, 0x40800709,
+0x3fe3078f, 0x3fbf0783, 0x3f200183, 0x3fbe0183,
+0x3fe30187, 0x18008387, 0x4205c002, 0x3ac30404,
+0x1cffc489, 0x00200000, 0x18008403, 0x38830402,
+0x4cffc486, 0x3ac28185, 0xb0408584, 0x28830402,
+0x1c020408, 0x38828182, 0xb0408385, 0x1802c387,
+0x28828182, 0x217ff886, 0x04000582, 0x32800007,
+0x21a00802, 0x3fbf0705, 0x3f200285, 0x3fbe0285,
+0x3fe30285, 0x21a00885, 0x04000603, 0x21a00903,
+0x40803c02, 0x21a00982, 0x04000386, 0x21a00a06,
+0x40801202, 0x21a00a82, 0x73000003, 0x24200683,
+0x01a00404, 0x00200000, 0x34204682, 0x3ec40683,
+0xb0408203, 0x24204682, 0x01a00783, 0x00200000,
+0x3421c682, 0x3edc0684, 0xb0408184, 0x2421c682,
+0x21a00806, 0x21a00885, 0x3fbf0784, 0x3f200204,
+0x3fbe0204, 0x3fe30204, 0x21a00904, 0x40804002,
+0x21a00982, 0x21a00a06, 0x40805a02, 0x21a00a82,
+0x04000683, 0x21a00803, 0x21a00885, 0x21a00904,
+0x40848002, 0x21a00982, 0x21a00a06, 0x40801002,
+0x21a00a82, 0x21a00a06, 0x40806602, 0x00200000,
+0x35800009, 0x21a00a82, 0x40800083, 0x21a00b83,
+0x01a00c02, 0x01a00d83, 0x00003ffb, 0x40800003,
+0x4020007f, 0x35000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+0x00000000, 0x00000000, 0x00000000, 0x00000000,
+};

arch/powerpc/platforms/cell/spufs/spu_utils.h

@@ -0,0 +1,160 @@
+/*
+ * utils.h: Utilities for SPU-side of the context switch operation.
+ *
+ * (C) Copyright IBM 2005
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+#ifndef _SPU_CONTEXT_UTILS_H_
+#define _SPU_CONTEXT_UTILS_H_
+
+/*
+ * 64-bit safe EA.
+ */
+typedef union {
+	unsigned long long ull;
+	unsigned int ui[2];
+} addr64;
+
+/*
+ * 128-bit register template.
+ */
+typedef union {
+	unsigned int slot[4];
+	vector unsigned int v;
+} spu_reg128v;
+
+/*
+ * DMA list structure.
+ */
+struct dma_list_elem {
+	unsigned int size;
+	unsigned int ea_low;
+};
+
+/*
+ * Declare storage for 8-byte aligned DMA list.
+ */
+struct dma_list_elem dma_list[15] __attribute__ ((aligned(8)));
+
+/*
+ * External definition for storage
+ * declared in crt0.
+ */
+extern spu_reg128v regs_spill[NR_SPU_SPILL_REGS];
+
+/*
+ * Compute LSCSA byte offset for a given field.
+ */
+static struct spu_lscsa *dummy = (struct spu_lscsa *)0;
+#define LSCSA_BYTE_OFFSET(_field)  \
+	((char *)(&(dummy->_field)) - (char *)(&(dummy->gprs[0].slot[0])))
+#define LSCSA_QW_OFFSET(_field)  (LSCSA_BYTE_OFFSET(_field) >> 4)
+
+static inline void set_event_mask(void)
+{
+	unsigned int event_mask = 0;
+
+	/* Save, Step 4:
+	 * Restore, Step 1:
+	 *    Set the SPU_RdEventMsk channel to zero to mask
+	 *    all events.
+	 */
+	spu_writech(SPU_WrEventMask, event_mask);
+}
+
+static inline void set_tag_mask(void)
+{
+	unsigned int tag_mask = 1;
+
+	/* Save, Step 5:
+	 * Restore, Step 2:
+	 *    Set the SPU_WrTagMsk channel to '01' to unmask
+	 *    only tag group 0.
+	 */
+	spu_writech(MFC_WrTagMask, tag_mask);
+}
+
+static inline void build_dma_list(addr64 lscsa_ea)
+{
+	unsigned int ea_low;
+	int i;
+
+	/* Save, Step 6:
+	 * Restore, Step 3:
+	 *    Update the effective address for the CSA in the
+	 *    pre-canned DMA-list in local storage.
+	 */
+	ea_low = lscsa_ea.ui[1];
+	ea_low += LSCSA_BYTE_OFFSET(ls[16384]);
+
+	for (i = 0; i < 15; i++, ea_low += 16384) {
+		dma_list[i].size = 16384;
+		dma_list[i].ea_low = ea_low;
+	}
+}
+
+static inline void enqueue_putllc(addr64 lscsa_ea)
+{
+	unsigned int ls = 0;
+	unsigned int size = 128;
+	unsigned int tag_id = 0;
+	unsigned int cmd = 0xB4;	/* PUTLLC */
+
+	/* Save, Step 12:
+	 * Restore, Step 7:
+	 *    Send a PUTLLC (tag 0) command to the MFC using
+	 *    an effective address in the CSA in order to
+	 *    remove any possible lock-line reservation.
+	 */
+	spu_writech(MFC_LSA, ls);
+	spu_writech(MFC_EAH, lscsa_ea.ui[0]);
+	spu_writech(MFC_EAL, lscsa_ea.ui[1]);
+	spu_writech(MFC_Size, size);
+	spu_writech(MFC_TagID, tag_id);
+	spu_writech(MFC_Cmd, cmd);
+}
+
+static inline void set_tag_update(void)
+{
+	unsigned int update_any = 1;
+
+	/* Save, Step 15:
+	 * Restore, Step 8:
+	 *    Write the MFC_TagUpdate channel with '01'.
+	 */
+	spu_writech(MFC_WrTagUpdate, update_any);
+}
+
+static inline void read_tag_status(void)
+{
+	/* Save, Step 16:
+	 * Restore, Step 9:
+	 *    Read the MFC_TagStat channel data.
+	 */
+	spu_readch(MFC_RdTagStat);
+}
+
+static inline void read_llar_status(void)
+{
+	/* Save, Step 17:
+	 * Restore, Step 10:
+	 *    Read the MFC_AtomicStat channel data.
+	 */
+	spu_readch(MFC_RdAtomicStat);
+}
+
+#endif				/* _SPU_CONTEXT_UTILS_H_ */

arch/powerpc/platforms/cell/spufs/spufs.h

@@ -0,0 +1,163 @@
+/*
+ * SPU file system
+ *
+ * (C) Copyright IBM Deutschland Entwicklung GmbH 2005
+ *
+ * Author: Arnd Bergmann <arndb@de.ibm.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+#ifndef SPUFS_H
+#define SPUFS_H
+
+#include <linux/kref.h>
+#include <linux/rwsem.h>
+#include <linux/spinlock.h>
+#include <linux/fs.h>
+
+#include <asm/spu.h>
+#include <asm/spu_csa.h>
+
+/* The magic number for our file system */
+enum {
+	SPUFS_MAGIC = 0x23c9b64e,
+};
+
+struct spu_context_ops;
+
+#define SPU_CONTEXT_PREEMPT          0UL
+
+struct spu_context {
+	struct spu *spu;		  /* pointer to a physical SPU */
+	struct spu_state csa;		  /* SPU context save area. */
+	spinlock_t mmio_lock;		  /* protects mmio access */
+	struct address_space *local_store;/* local store backing store */
+
+	enum { SPU_STATE_RUNNABLE, SPU_STATE_SAVED } state;
+	struct rw_semaphore state_sema;
+	struct semaphore run_sema;
+
+	struct mm_struct *owner;
+
+	struct kref kref;
+	wait_queue_head_t ibox_wq;
+	wait_queue_head_t wbox_wq;
+	wait_queue_head_t stop_wq;
+	struct fasync_struct *ibox_fasync;
+	struct fasync_struct *wbox_fasync;
+	struct spu_context_ops *ops;
+	struct work_struct reap_work;
+	u64 flags;
+};
+
+/* SPU context query/set operations. */
+struct spu_context_ops {
+	int (*mbox_read) (struct spu_context * ctx, u32 * data);
+	 u32(*mbox_stat_read) (struct spu_context * ctx);
+	unsigned int (*mbox_stat_poll)(struct spu_context *ctx,
+					unsigned int events);
+	int (*ibox_read) (struct spu_context * ctx, u32 * data);
+	int (*wbox_write) (struct spu_context * ctx, u32 data);
+	 u32(*signal1_read) (struct spu_context * ctx);
+	void (*signal1_write) (struct spu_context * ctx, u32 data);
+	 u32(*signal2_read) (struct spu_context * ctx);
+	void (*signal2_write) (struct spu_context * ctx, u32 data);
+	void (*signal1_type_set) (struct spu_context * ctx, u64 val);
+	 u64(*signal1_type_get) (struct spu_context * ctx);
+	void (*signal2_type_set) (struct spu_context * ctx, u64 val);
+	 u64(*signal2_type_get) (struct spu_context * ctx);
+	 u32(*npc_read) (struct spu_context * ctx);
+	void (*npc_write) (struct spu_context * ctx, u32 data);
+	 u32(*status_read) (struct spu_context * ctx);
+	char*(*get_ls) (struct spu_context * ctx);
+	void (*runcntl_write) (struct spu_context * ctx, u32 data);
+	void (*runcntl_stop) (struct spu_context * ctx);
+};
+
+extern struct spu_context_ops spu_hw_ops;
+extern struct spu_context_ops spu_backing_ops;
+
+struct spufs_inode_info {
+	struct spu_context *i_ctx;
+	struct inode vfs_inode;
+};
+#define SPUFS_I(inode) \
+	container_of(inode, struct spufs_inode_info, vfs_inode)
+
+extern struct tree_descr spufs_dir_contents[];
+
+/* system call implementation */
+long spufs_run_spu(struct file *file,
+		   struct spu_context *ctx, u32 *npc, u32 *status);
+long spufs_create_thread(struct nameidata *nd,
+			 unsigned int flags, mode_t mode);
+extern struct file_operations spufs_context_fops;
+
+/* context management */
+struct spu_context * alloc_spu_context(struct address_space *local_store);
+void destroy_spu_context(struct kref *kref);
+struct spu_context * get_spu_context(struct spu_context *ctx);
+int put_spu_context(struct spu_context *ctx);
+void spu_unmap_mappings(struct spu_context *ctx);
+
+void spu_forget(struct spu_context *ctx);
+void spu_acquire(struct spu_context *ctx);
+void spu_release(struct spu_context *ctx);
+int spu_acquire_runnable(struct spu_context *ctx);
+void spu_acquire_saved(struct spu_context *ctx);
+
+int spu_activate(struct spu_context *ctx, u64 flags);
+void spu_deactivate(struct spu_context *ctx);
+void spu_yield(struct spu_context *ctx);
+int __init spu_sched_init(void);
+void __exit spu_sched_exit(void);
+
+/*
+ * spufs_wait
+ * 	Same as wait_event_interruptible(), except that here
+ *	we need to call spu_release(ctx) before sleeping, and
+ *	then spu_acquire(ctx) when awoken.
+ */
+
+#define spufs_wait(wq, condition)					\
+({									\
+	int __ret = 0;							\
+	DEFINE_WAIT(__wait);						\
+	for (;;) {							\
+		prepare_to_wait(&(wq), &__wait, TASK_INTERRUPTIBLE);	\
+		if (condition)						\
+			break;						\
+		if (!signal_pending(current)) {				\
+			spu_release(ctx);				\
+			schedule();					\
+			spu_acquire(ctx);				\
+			continue;					\
+		}							\
+		__ret = -ERESTARTSYS;					\
+		break;							\
+	}								\
+	finish_wait(&(wq), &__wait);					\
+	__ret;								\
+})
+
+size_t spu_wbox_write(struct spu_context *ctx, u32 data);
+size_t spu_ibox_read(struct spu_context *ctx, u32 *data);
+
+/* irq callback funcs. */
+void spufs_ibox_callback(struct spu *spu);
+void spufs_wbox_callback(struct spu *spu);
+void spufs_stop_callback(struct spu *spu);
+
+#endif

arch/powerpc/platforms/cell/spufs/syscalls.c

@@ -0,0 +1,101 @@
+#include <linux/file.h>
+#include <linux/fs.h>
+#include <linux/module.h>
+#include <linux/mount.h>
+#include <linux/namei.h>
+
+#include <asm/uaccess.h>
+
+#include "spufs.h"
+
+/**
+ * sys_spu_run - run code loaded into an SPU
+ *
+ * @unpc:    next program counter for the SPU
+ * @ustatus: status of the SPU
+ *
+ * This system call transfers the control of execution of a
+ * user space thread to an SPU. It will return when the
+ * SPU has finished executing or when it hits an error
+ * condition and it will be interrupted if a signal needs
+ * to be delivered to a handler in user space.
+ *
+ * The next program counter is set to the passed value
+ * before the SPU starts fetching code and the user space
+ * pointer gets updated with the new value when returning
+ * from kernel space.
+ *
+ * The status value returned from spu_run reflects the
+ * value of the spu_status register after the SPU has stopped.
+ *
+ */
+long do_spu_run(struct file *filp, __u32 __user *unpc, __u32 __user *ustatus)
+{
+	long ret;
+	struct spufs_inode_info *i;
+	u32 npc, status;
+
+	ret = -EFAULT;
+	if (get_user(npc, unpc) || get_user(status, ustatus))
+		goto out;
+
+	/* check if this file was created by spu_create */
+	ret = -EINVAL;
+	if (filp->f_op != &spufs_context_fops)
+		goto out;
+
+	i = SPUFS_I(filp->f_dentry->d_inode);
+	ret = spufs_run_spu(filp, i->i_ctx, &npc, &status);
+
+	if (put_user(npc, unpc) || put_user(status, ustatus))
+		ret = -EFAULT;
+out:
+	return ret;
+}
+
+#ifndef MODULE
+asmlinkage long sys_spu_run(int fd, __u32 __user *unpc, __u32 __user *ustatus)
+{
+	int fput_needed;
+	struct file *filp;
+	long ret;
+
+	ret = -EBADF;
+	filp = fget_light(fd, &fput_needed);
+	if (filp) {
+		ret = do_spu_run(filp, unpc, ustatus);
+		fput_light(filp, fput_needed);
+	}
+
+	return ret;
+}
+#endif
+
+asmlinkage long sys_spu_create(const char __user *pathname,
+					unsigned int flags, mode_t mode)
+{
+	char *tmp;
+	int ret;
+
+	tmp = getname(pathname);
+	ret = PTR_ERR(tmp);
+	if (!IS_ERR(tmp)) {
+		struct nameidata nd;
+
+		ret = path_lookup(tmp, LOOKUP_PARENT|
+				LOOKUP_OPEN|LOOKUP_CREATE, &nd);
+		if (!ret) {
+			ret = spufs_create_thread(&nd, flags, mode);
+			path_release(&nd);
+		}
+		putname(tmp);
+	}
+
+	return ret;
+}
+
+struct spufs_calls spufs_calls = {
+	.create_thread = sys_spu_create,
+	.spu_run = do_spu_run,
+	.owner = THIS_MODULE,
+};

arch/powerpc/platforms/chrp/setup.c

@@ -49,7 +49,6 @@
 #include <asm/hydra.h>
 #include <asm/sections.h>
 #include <asm/time.h>
-#include <asm/btext.h>
 #include <asm/i8259.h>
 #include <asm/mpic.h>
 #include <asm/rtas.h>
@@ -58,7 +57,6 @@
 #include "chrp.h"
 
 void rtas_indicator_progress(char *, unsigned short);
-void btext_progress(char *, unsigned short);
 
 int _chrp_type;
 EXPORT_SYMBOL(_chrp_type);
@@ -264,11 +262,6 @@ void __init chrp_setup_arch(void)
 		ppc_md.set_rtc_time	= rtas_set_rtc_time;
 	}
 
-#ifdef CONFIG_BOOTX_TEXT
-	if (ppc_md.progress == NULL && boot_text_mapped)
-		ppc_md.progress = btext_progress;
-#endif
-
 #ifdef CONFIG_BLK_DEV_INITRD
 	/* this is fine for chrp */
 	initrd_below_start_ok = 1;
@@ -522,12 +515,3 @@ void __init chrp_init(void)
 	smp_ops = &chrp_smp_ops;
 #endif /* CONFIG_SMP */
 }
-
-#ifdef CONFIG_BOOTX_TEXT
-void
-btext_progress(char *s, unsigned short hex)
-{
-	btext_drawstring(s);
-	btext_drawstring("\n");
-}
-#endif /* CONFIG_BOOTX_TEXT */

arch/powerpc/platforms/iseries/irq.c

@@ -35,161 +35,138 @@
 #include <linux/irq.h>
 #include <linux/spinlock.h>
 
+#include <asm/paca.h>
 #include <asm/iseries/hv_types.h>
 #include <asm/iseries/hv_lp_event.h>
 #include <asm/iseries/hv_call_xm.h>
+#include <asm/iseries/it_lp_queue.h>
 
 #include "irq.h"
 #include "call_pci.h"
 
-static long Pci_Interrupt_Count;
-static long Pci_Event_Count;
+#if defined(CONFIG_SMP)
+extern void iSeries_smp_message_recv(struct pt_regs *);
+#endif
 
-enum XmPciLpEvent_Subtype {
-	XmPciLpEvent_BusCreated		= 0,	// PHB has been created
-	XmPciLpEvent_BusError		= 1,	// PHB has failed
-	XmPciLpEvent_BusFailed		= 2,	// Msg to Secondary, Primary failed bus
-	XmPciLpEvent_NodeFailed		= 4,	// Multi-adapter bridge has failed
-	XmPciLpEvent_NodeRecovered	= 5,	// Multi-adapter bridge has recovered
-	XmPciLpEvent_BusRecovered	= 12,	// PHB has been recovered
-	XmPciLpEvent_UnQuiesceBus	= 18,	// Secondary bus unqiescing
-	XmPciLpEvent_BridgeError	= 21,	// Bridge Error
-	XmPciLpEvent_SlotInterrupt	= 22	// Slot interrupt
+enum pci_event_type {
+	pe_bus_created		= 0,	/* PHB has been created */
+	pe_bus_error		= 1,	/* PHB has failed */
+	pe_bus_failed		= 2,	/* Msg to Secondary, Primary failed bus */
+	pe_node_failed		= 4,	/* Multi-adapter bridge has failed */
+	pe_node_recovered	= 5,	/* Multi-adapter bridge has recovered */
+	pe_bus_recovered	= 12,	/* PHB has been recovered */
+	pe_unquiese_bus		= 18,	/* Secondary bus unqiescing */
+	pe_bridge_error		= 21,	/* Bridge Error */
+	pe_slot_interrupt	= 22	/* Slot interrupt */
 };
 
-struct XmPciLpEvent_BusInterrupt {
-	HvBusNumber	busNumber;
-	HvSubBusNumber	subBusNumber;
-};
-
-struct XmPciLpEvent_NodeInterrupt {
-	HvBusNumber	busNumber;
-	HvSubBusNumber	subBusNumber;
-	HvAgentId	deviceId;
-};
-
-struct XmPciLpEvent {
-	struct HvLpEvent hvLpEvent;
-
+struct pci_event {
+	struct HvLpEvent event;
 	union {
-		u64 alignData;			// Align on an 8-byte boundary
-
+		u64 __align;		/* Align on an 8-byte boundary */
 		struct {
 			u32		fisr;
-			HvBusNumber	busNumber;
-			HvSubBusNumber	subBusNumber;
-			HvAgentId	deviceId;
-		} slotInterrupt;
-
-		struct XmPciLpEvent_BusInterrupt busFailed;
-		struct XmPciLpEvent_BusInterrupt busRecovered;
-		struct XmPciLpEvent_BusInterrupt busCreated;
-
-		struct XmPciLpEvent_NodeInterrupt nodeFailed;
-		struct XmPciLpEvent_NodeInterrupt nodeRecovered;
-
-	} eventData;
-
+			HvBusNumber	bus_number;
+			HvSubBusNumber	sub_bus_number;
+			HvAgentId	dev_id;
+		} slot;
+		struct {
+			HvBusNumber	bus_number;
+			HvSubBusNumber	sub_bus_number;
+		} bus;
+		struct {
+			HvBusNumber	bus_number;
+			HvSubBusNumber	sub_bus_number;
+			HvAgentId	dev_id;
+		} node;
+	} data;
 };
 
-static void intReceived(struct XmPciLpEvent *eventParm,
-		struct pt_regs *regsParm)
+static DEFINE_SPINLOCK(pending_irqs_lock);
+static int num_pending_irqs;
+static int pending_irqs[NR_IRQS];
+
+static void int_received(struct pci_event *event, struct pt_regs *regs)
 {
 	int irq;
-#ifdef CONFIG_IRQSTACKS
-	struct thread_info *curtp, *irqtp;
-#endif
 
-	++Pci_Interrupt_Count;
-
-	switch (eventParm->hvLpEvent.xSubtype) {
-	case XmPciLpEvent_SlotInterrupt:
-		irq = eventParm->hvLpEvent.xCorrelationToken;
-		/* Dispatch the interrupt handlers for this irq */
-#ifdef CONFIG_IRQSTACKS
-		/* Switch to the irq stack to handle this */
-		curtp = current_thread_info();
-		irqtp = hardirq_ctx[smp_processor_id()];
-		if (curtp != irqtp) {
-			irqtp->task = curtp->task;
-			irqtp->flags = 0;
-			call___do_IRQ(irq, regsParm, irqtp);
-			irqtp->task = NULL;
-			if (irqtp->flags)
-				set_bits(irqtp->flags, &curtp->flags);
-		} else
-#endif
-			__do_IRQ(irq, regsParm);
-		HvCallPci_eoi(eventParm->eventData.slotInterrupt.busNumber,
-			eventParm->eventData.slotInterrupt.subBusNumber,
-			eventParm->eventData.slotInterrupt.deviceId);
+	switch (event->event.xSubtype) {
+	case pe_slot_interrupt:
+		irq = event->event.xCorrelationToken;
+		if (irq < NR_IRQS) {
+			spin_lock(&pending_irqs_lock);
+			pending_irqs[irq]++;
+			num_pending_irqs++;
+			spin_unlock(&pending_irqs_lock);
+		} else {
+			printk(KERN_WARNING "int_received: bad irq number %d\n",
+					irq);
+			HvCallPci_eoi(event->data.slot.bus_number,
+					event->data.slot.sub_bus_number,
+					event->data.slot.dev_id);
+		}
 		break;
 		/* Ignore error recovery events for now */
-	case XmPciLpEvent_BusCreated:
-		printk(KERN_INFO "intReceived: system bus %d created\n",
-			eventParm->eventData.busCreated.busNumber);
+	case pe_bus_created:
+		printk(KERN_INFO "int_received: system bus %d created\n",
+			event->data.bus.bus_number);
 		break;
-	case XmPciLpEvent_BusError:
-	case XmPciLpEvent_BusFailed:
-		printk(KERN_INFO "intReceived: system bus %d failed\n",
-			eventParm->eventData.busFailed.busNumber);
+	case pe_bus_error:
+	case pe_bus_failed:
+		printk(KERN_INFO "int_received: system bus %d failed\n",
+			event->data.bus.bus_number);
 		break;
-	case XmPciLpEvent_BusRecovered:
-	case XmPciLpEvent_UnQuiesceBus:
-		printk(KERN_INFO "intReceived: system bus %d recovered\n",
-			eventParm->eventData.busRecovered.busNumber);
+	case pe_bus_recovered:
+	case pe_unquiese_bus:
+		printk(KERN_INFO "int_received: system bus %d recovered\n",
+			event->data.bus.bus_number);
 		break;
-	case XmPciLpEvent_NodeFailed:
-	case XmPciLpEvent_BridgeError:
+	case pe_node_failed:
+	case pe_bridge_error:
 		printk(KERN_INFO
-			"intReceived: multi-adapter bridge %d/%d/%d failed\n",
-			eventParm->eventData.nodeFailed.busNumber,
-			eventParm->eventData.nodeFailed.subBusNumber,
-			eventParm->eventData.nodeFailed.deviceId);
+			"int_received: multi-adapter bridge %d/%d/%d failed\n",
+			event->data.node.bus_number,
+			event->data.node.sub_bus_number,
+			event->data.node.dev_id);
 		break;
-	case XmPciLpEvent_NodeRecovered:
+	case pe_node_recovered:
 		printk(KERN_INFO
-			"intReceived: multi-adapter bridge %d/%d/%d recovered\n",
-			eventParm->eventData.nodeRecovered.busNumber,
-			eventParm->eventData.nodeRecovered.subBusNumber,
-			eventParm->eventData.nodeRecovered.deviceId);
+			"int_received: multi-adapter bridge %d/%d/%d recovered\n",
+			event->data.node.bus_number,
+			event->data.node.sub_bus_number,
+			event->data.node.dev_id);
 		break;
 	default:
 		printk(KERN_ERR
-			"intReceived: unrecognized event subtype 0x%x\n",
-			eventParm->hvLpEvent.xSubtype);
+			"int_received: unrecognized event subtype 0x%x\n",
+			event->event.xSubtype);
 		break;
 	}
 }
 
-static void XmPciLpEvent_handler(struct HvLpEvent *eventParm,
-		struct pt_regs *regsParm)
+static void pci_event_handler(struct HvLpEvent *event, struct pt_regs *regs)
 {
-#ifdef CONFIG_PCI
-	++Pci_Event_Count;
-
-	if (eventParm && (eventParm->xType == HvLpEvent_Type_PciIo)) {
-		switch (eventParm->xFlags.xFunction) {
+	if (event && (event->xType == HvLpEvent_Type_PciIo)) {
+		switch (event->xFlags.xFunction) {
 		case HvLpEvent_Function_Int:
-			intReceived((struct XmPciLpEvent *)eventParm, regsParm);
+			int_received((struct pci_event *)event, regs);
 			break;
 		case HvLpEvent_Function_Ack:
 			printk(KERN_ERR
-				"XmPciLpEvent_handler: unexpected ack received\n");
+				"pci_event_handler: unexpected ack received\n");
 			break;
 		default:
 			printk(KERN_ERR
-				"XmPciLpEvent_handler: unexpected event function %d\n",
-				(int)eventParm->xFlags.xFunction);
+				"pci_event_handler: unexpected event function %d\n",
+				(int)event->xFlags.xFunction);
 			break;
 		}
-	} else if (eventParm)
+	} else if (event)
 		printk(KERN_ERR
-			"XmPciLpEvent_handler: Unrecognized PCI event type 0x%x\n",
-			(int)eventParm->xType);
+			"pci_event_handler: Unrecognized PCI event type 0x%x\n",
+			(int)event->xType);
 	else
-		printk(KERN_ERR "XmPciLpEvent_handler: NULL event received\n");
-#endif
+		printk(KERN_ERR "pci_event_handler: NULL event received\n");
 }
 
 /*
@@ -199,20 +176,21 @@ static void XmPciLpEvent_handler(struct HvLpEvent *eventParm,
 void __init iSeries_init_IRQ(void)
 {
 	/* Register PCI event handler and open an event path */
-	int xRc;
+	int ret;
 
-	xRc = HvLpEvent_registerHandler(HvLpEvent_Type_PciIo,
-			&XmPciLpEvent_handler);
-	if (xRc == 0) {
-		xRc = HvLpEvent_openPath(HvLpEvent_Type_PciIo, 0);
-		if (xRc != 0)
-			printk(KERN_ERR "iSeries_init_IRQ: open event path "
-					"failed with rc 0x%x\n", xRc);
+	ret = HvLpEvent_registerHandler(HvLpEvent_Type_PciIo,
+			&pci_event_handler);
+	if (ret == 0) {
+		ret = HvLpEvent_openPath(HvLpEvent_Type_PciIo, 0);
+		if (ret != 0)
+			printk(KERN_ERR "iseries_init_IRQ: open event path "
+					"failed with rc 0x%x\n", ret);
 	} else
-		printk(KERN_ERR "iSeries_init_IRQ: register handler "
-				"failed with rc 0x%x\n", xRc);
+		printk(KERN_ERR "iseries_init_IRQ: register handler "
+				"failed with rc 0x%x\n", ret);
 }
 
+#define REAL_IRQ_TO_SUBBUS(irq)	(((irq) >> 14) & 0xff)
 #define REAL_IRQ_TO_BUS(irq)	((((irq) >> 6) & 0xff) + 1)
 #define REAL_IRQ_TO_IDSEL(irq)	((((irq) >> 3) & 7) + 1)
 #define REAL_IRQ_TO_FUNC(irq)	((irq) & 7)
@@ -221,40 +199,40 @@ void __init iSeries_init_IRQ(void)
  * This will be called by device drivers (via enable_IRQ)
  * to enable INTA in the bridge interrupt status register.
  */
-static void iSeries_enable_IRQ(unsigned int irq)
+static void iseries_enable_IRQ(unsigned int irq)
 {
-	u32 bus, deviceId, function, mask;
-	const u32 subBus = 0;
+	u32 bus, dev_id, function, mask;
+	const u32 sub_bus = 0;
 	unsigned int rirq = virt_irq_to_real_map[irq];
 
 	/* The IRQ has already been locked by the caller */
 	bus = REAL_IRQ_TO_BUS(rirq);
 	function = REAL_IRQ_TO_FUNC(rirq);
-	deviceId = (REAL_IRQ_TO_IDSEL(rirq) << 4) + function;
+	dev_id = (REAL_IRQ_TO_IDSEL(rirq) << 4) + function;
 
 	/* Unmask secondary INTA */
 	mask = 0x80000000;
-	HvCallPci_unmaskInterrupts(bus, subBus, deviceId, mask);
+	HvCallPci_unmaskInterrupts(bus, sub_bus, dev_id, mask);
 }
 
-/* This is called by iSeries_activate_IRQs */
-static unsigned int iSeries_startup_IRQ(unsigned int irq)
+/* This is called by iseries_activate_IRQs */
+static unsigned int iseries_startup_IRQ(unsigned int irq)
 {
-	u32 bus, deviceId, function, mask;
-	const u32 subBus = 0;
+	u32 bus, dev_id, function, mask;
+	const u32 sub_bus = 0;
 	unsigned int rirq = virt_irq_to_real_map[irq];
 
 	bus = REAL_IRQ_TO_BUS(rirq);
 	function = REAL_IRQ_TO_FUNC(rirq);
-	deviceId = (REAL_IRQ_TO_IDSEL(rirq) << 4) + function;
+	dev_id = (REAL_IRQ_TO_IDSEL(rirq) << 4) + function;
 
 	/* Link the IRQ number to the bridge */
-	HvCallXm_connectBusUnit(bus, subBus, deviceId, irq);
+	HvCallXm_connectBusUnit(bus, sub_bus, dev_id, irq);
 
 	/* Unmask bridge interrupts in the FISR */
 	mask = 0x01010000 << function;
-	HvCallPci_unmaskFisr(bus, subBus, deviceId, mask);
-	iSeries_enable_IRQ(irq);
+	HvCallPci_unmaskFisr(bus, sub_bus, dev_id, mask);
+	iseries_enable_IRQ(irq);
 	return 0;
 }
 
@@ -279,78 +257,115 @@ void __init iSeries_activate_IRQs()
 }
 
 /*  this is not called anywhere currently */
-static void iSeries_shutdown_IRQ(unsigned int irq)
+static void iseries_shutdown_IRQ(unsigned int irq)
 {
-	u32 bus, deviceId, function, mask;
-	const u32 subBus = 0;
+	u32 bus, dev_id, function, mask;
+	const u32 sub_bus = 0;
 	unsigned int rirq = virt_irq_to_real_map[irq];
 
 	/* irq should be locked by the caller */
 	bus = REAL_IRQ_TO_BUS(rirq);
 	function = REAL_IRQ_TO_FUNC(rirq);
-	deviceId = (REAL_IRQ_TO_IDSEL(rirq) << 4) + function;
+	dev_id = (REAL_IRQ_TO_IDSEL(rirq) << 4) + function;
 
 	/* Invalidate the IRQ number in the bridge */
-	HvCallXm_connectBusUnit(bus, subBus, deviceId, 0);
+	HvCallXm_connectBusUnit(bus, sub_bus, dev_id, 0);
 
 	/* Mask bridge interrupts in the FISR */
 	mask = 0x01010000 << function;
-	HvCallPci_maskFisr(bus, subBus, deviceId, mask);
+	HvCallPci_maskFisr(bus, sub_bus, dev_id, mask);
 }
 
 /*
  * This will be called by device drivers (via disable_IRQ)
  * to disable INTA in the bridge interrupt status register.
  */
-static void iSeries_disable_IRQ(unsigned int irq)
+static void iseries_disable_IRQ(unsigned int irq)
 {
-	u32 bus, deviceId, function, mask;
-	const u32 subBus = 0;
+	u32 bus, dev_id, function, mask;
+	const u32 sub_bus = 0;
 	unsigned int rirq = virt_irq_to_real_map[irq];
 
 	/* The IRQ has already been locked by the caller */
 	bus = REAL_IRQ_TO_BUS(rirq);
 	function = REAL_IRQ_TO_FUNC(rirq);
-	deviceId = (REAL_IRQ_TO_IDSEL(rirq) << 4) + function;
+	dev_id = (REAL_IRQ_TO_IDSEL(rirq) << 4) + function;
 
 	/* Mask secondary INTA   */
 	mask = 0x80000000;
-	HvCallPci_maskInterrupts(bus, subBus, deviceId, mask);
+	HvCallPci_maskInterrupts(bus, sub_bus, dev_id, mask);
 }
 
-/*
- * This does nothing because there is not enough information
- * provided to do the EOI HvCall.  This is done by XmPciLpEvent.c
- */
-static void iSeries_end_IRQ(unsigned int irq)
+static void iseries_end_IRQ(unsigned int irq)
 {
+	unsigned int rirq = virt_irq_to_real_map[irq];
+
+	HvCallPci_eoi(REAL_IRQ_TO_BUS(rirq), REAL_IRQ_TO_SUBBUS(rirq),
+		(REAL_IRQ_TO_IDSEL(rirq) << 4) + REAL_IRQ_TO_FUNC(rirq));
 }
 
 static hw_irq_controller iSeries_IRQ_handler = {
 	.typename = "iSeries irq controller",
-	.startup = iSeries_startup_IRQ,
-	.shutdown = iSeries_shutdown_IRQ,
-	.enable = iSeries_enable_IRQ,
-	.disable = iSeries_disable_IRQ,
-	.end = iSeries_end_IRQ
+	.startup = iseries_startup_IRQ,
+	.shutdown = iseries_shutdown_IRQ,
+	.enable = iseries_enable_IRQ,
+	.disable = iseries_disable_IRQ,
+	.end = iseries_end_IRQ
 };
 
 /*
  * This is called out of iSeries_scan_slot to allocate an IRQ for an EADS slot
  * It calculates the irq value for the slot.
- * Note that subBusNumber is always 0 (at the moment at least).
+ * Note that sub_bus is always 0 (at the moment at least).
  */
-int __init iSeries_allocate_IRQ(HvBusNumber busNumber,
-		HvSubBusNumber subBusNumber, HvAgentId deviceId)
+int __init iSeries_allocate_IRQ(HvBusNumber bus,
+		HvSubBusNumber sub_bus, HvAgentId dev_id)
 {
 	int virtirq;
 	unsigned int realirq;
-	u8 idsel = (deviceId >> 4);
-	u8 function = deviceId & 7;
+	u8 idsel = (dev_id >> 4);
+	u8 function = dev_id & 7;
 
-	realirq = ((busNumber - 1) << 6) + ((idsel - 1) << 3) + function;
+	realirq = (((((sub_bus << 8) + (bus - 1)) << 3) + (idsel - 1)) << 3)
+		+ function;
 	virtirq = virt_irq_create_mapping(realirq);
 
 	irq_desc[virtirq].handler = &iSeries_IRQ_handler;
 	return virtirq;
 }
+
+/*
+ * Get the next pending IRQ.
+ */
+int iSeries_get_irq(struct pt_regs *regs)
+{
+	struct paca_struct *lpaca;
+	/* -2 means ignore this interrupt */
+	int irq = -2;
+
+	lpaca = get_paca();
+#ifdef CONFIG_SMP
+	if (lpaca->lppaca.int_dword.fields.ipi_cnt) {
+		lpaca->lppaca.int_dword.fields.ipi_cnt = 0;
+		iSeries_smp_message_recv(regs);
+	}
+#endif /* CONFIG_SMP */
+	if (hvlpevent_is_pending())
+		process_hvlpevents(regs);
+
+	if (num_pending_irqs) {
+		spin_lock(&pending_irqs_lock);
+		for (irq = 0; irq < NR_IRQS; irq++) {
+			if (pending_irqs[irq]) {
+				pending_irqs[irq]--;
+				num_pending_irqs--;
+				break;
+			}
+		}
+		spin_unlock(&pending_irqs_lock);
+		if (irq >= NR_IRQS)
+			irq = -2;
+	}
+
+	return irq;
+}

arch/powerpc/platforms/iseries/irq.h

@@ -4,5 +4,6 @@
 extern void iSeries_init_IRQ(void);
 extern int  iSeries_allocate_IRQ(HvBusNumber, HvSubBusNumber, HvAgentId);
 extern void iSeries_activate_IRQs(void);
+extern int iSeries_get_irq(struct pt_regs *);
 
 #endif /* _ISERIES_IRQ_H */

arch/powerpc/platforms/iseries/lpardata.c

@@ -225,3 +225,10 @@ struct ItVpdAreas itVpdAreas = {
 		0,0
 	}
 };
+
+struct ItLpRegSave iseries_reg_save[] = {
+	[0 ... (NR_CPUS-1)] = {
+		.xDesc = 0xd397d9e2,	/* "LpRS" */
+		.xSize = sizeof(struct ItLpRegSave),
+	},
+};

arch/powerpc/platforms/iseries/setup.c

@@ -569,16 +569,6 @@ static void iSeries_show_cpuinfo(struct seq_file *m)
 	seq_printf(m, "machine\t\t: 64-bit iSeries Logical Partition\n");
 }
 
-/*
- * Document me.
- * and Implement me.
- */
-static int iSeries_get_irq(struct pt_regs *regs)
-{
-	/* -2 means ignore this interrupt */
-	return -2;
-}
-
 /*
  * Document me.
  */

arch/powerpc/platforms/maple/setup.c

@@ -51,6 +51,7 @@
 #include <asm/pgtable.h>
 #include <asm/bitops.h>
 #include <asm/io.h>
+#include <asm/kexec.h>
 #include <asm/pci-bridge.h>
 #include <asm/iommu.h>
 #include <asm/machdep.h>
@@ -191,24 +192,10 @@ static void __init maple_init_early(void)
 	 */
 	hpte_init_native();
 
-	/* Find the serial port */
-	generic_find_legacy_serial_ports(&physport, &default_speed);
-
-	DBG("phys port addr: %lx\n", (long)physport);
-
-	if (physport) {
-		void *comport;
-		/* Map the uart for udbg. */
-		comport = (void *)ioremap(physport, 16);
-		udbg_init_uart(comport, default_speed);
-
-		DBG("Hello World !\n");
-	}
-
 	/* Setup interrupt mapping options */
 	ppc64_interrupt_controller = IC_OPEN_PIC;
 
-	iommu_init_early_u3();
+	iommu_init_early_dart();
 
 	DBG(" <- maple_init_early\n");
 }
@@ -270,7 +257,7 @@ static int __init maple_probe(int platform)
 	 * occupies having to be broken up so the DART itself is not
 	 * part of the cacheable linar mapping
 	 */
-	alloc_u3_dart_table();
+	alloc_dart_table();
 
 	return 1;
 }
@@ -292,4 +279,9 @@ struct machdep_calls __initdata maple_md = {
       	.calibrate_decr		= generic_calibrate_decr,
 	.progress		= maple_progress,
 	.idle_loop		= native_idle,
+#ifdef CONFIG_KEXEC
+	.machine_kexec		= default_machine_kexec,
+	.machine_kexec_prepare	= default_machine_kexec_prepare,
+	.machine_crash_shutdown	= default_machine_crash_shutdown,
+#endif
 };

arch/powerpc/platforms/powermac/Makefile

@@ -1,9 +1,14 @@
+CFLAGS_bootx_init.o  		+= -fPIC
+
 obj-y				+= pic.o setup.o time.o feature.o pci.o \
-				   sleep.o low_i2c.o cache.o
+				   sleep.o low_i2c.o cache.o pfunc_core.o \
+				   pfunc_base.o
 obj-$(CONFIG_PMAC_BACKLIGHT)	+= backlight.o
 obj-$(CONFIG_CPU_FREQ_PMAC)	+= cpufreq_32.o
 obj-$(CONFIG_CPU_FREQ_PMAC64)	+= cpufreq_64.o
 obj-$(CONFIG_NVRAM)		+= nvram.o
 # ppc64 pmac doesn't define CONFIG_NVRAM but needs nvram stuff
 obj-$(CONFIG_PPC64)		+= nvram.o
+obj-$(CONFIG_PPC32)		+= bootx_init.o
 obj-$(CONFIG_SMP)		+= smp.o
+obj-$(CONFIG_PPC_MERGE)		+= udbg_scc.o udbg_adb.o

arch/powerpc/platforms/powermac/bootx_init.c

@@ -0,0 +1,547 @@
+/*
+ *  Early boot support code for BootX bootloader
+ *
+ *  Copyright (C) 2005 Ben. Herrenschmidt (benh@kernel.crashing.org)
+ *
+ *  This program is free software; you can redistribute it and/or
+ *  modify it under the terms of the GNU General Public License
+ *  as published by the Free Software Foundation; either version
+ *  2 of the License, or (at your option) any later version.
+ */
+
+#include <linux/config.h>
+#include <linux/kernel.h>
+#include <linux/string.h>
+#include <linux/init.h>
+#include <linux/version.h>
+#include <asm/sections.h>
+#include <asm/prom.h>
+#include <asm/page.h>
+#include <asm/bootx.h>
+#include <asm/bootinfo.h>
+#include <asm/btext.h>
+#include <asm/io.h>
+
+#undef DEBUG
+#define SET_BOOT_BAT
+
+#ifdef DEBUG
+#define DBG(fmt...) do { bootx_printf(fmt); } while(0)
+#else
+#define DBG(fmt...) do { } while(0)
+#endif
+
+extern void __start(unsigned long r3, unsigned long r4, unsigned long r5);
+
+static unsigned long __initdata bootx_dt_strbase;
+static unsigned long __initdata bootx_dt_strend;
+static unsigned long __initdata bootx_node_chosen;
+static boot_infos_t * __initdata bootx_info;
+static char __initdata bootx_disp_path[256];
+
+/* Is boot-info compatible ? */
+#define BOOT_INFO_IS_COMPATIBLE(bi) \
+	((bi)->compatible_version <= BOOT_INFO_VERSION)
+#define BOOT_INFO_IS_V2_COMPATIBLE(bi)	((bi)->version >= 2)
+#define BOOT_INFO_IS_V4_COMPATIBLE(bi)	((bi)->version >= 4)
+
+#ifdef CONFIG_BOOTX_TEXT
+static void __init bootx_printf(const char *format, ...)
+{
+	const char *p, *q, *s;
+	va_list args;
+	unsigned long v;
+
+	va_start(args, format);
+	for (p = format; *p != 0; p = q) {
+		for (q = p; *q != 0 && *q != '\n' && *q != '%'; ++q)
+			;
+		if (q > p)
+			btext_drawtext(p, q - p);
+		if (*q == 0)
+			break;
+		if (*q == '\n') {
+			++q;
+			btext_flushline();
+			btext_drawstring("\r\n");
+			btext_flushline();
+			continue;
+		}
+		++q;
+		if (*q == 0)
+			break;
+		switch (*q) {
+		case 's':
+			++q;
+			s = va_arg(args, const char *);
+			if (s == NULL)
+				s = "<NULL>";
+			btext_drawstring(s);
+			break;
+		case 'x':
+			++q;
+			v = va_arg(args, unsigned long);
+			btext_drawhex(v);
+			break;
+		}
+	}
+}
+#else /* CONFIG_BOOTX_TEXT */
+static void __init bootx_printf(const char *format, ...) {}
+#endif /* CONFIG_BOOTX_TEXT */
+
+static void * __init bootx_early_getprop(unsigned long base,
+					 unsigned long node,
+					 char *prop)
+{
+	struct bootx_dt_node *np = (struct bootx_dt_node *)(base + node);
+	u32 *ppp = &np->properties;
+
+	while(*ppp) {
+		struct bootx_dt_prop *pp =
+			(struct bootx_dt_prop *)(base + *ppp);
+
+		if (strcmp((char *)((unsigned long)pp->name + base),
+			   prop) == 0) {
+			return (void *)((unsigned long)pp->value + base);
+		}
+		ppp = &pp->next;
+	}
+	return NULL;
+}
+
+#define dt_push_token(token, mem) \
+	do { \
+		*(mem) = _ALIGN_UP(*(mem),4); \
+		*((u32 *)*(mem)) = token; \
+		*(mem) += 4; \
+	} while(0)
+
+static unsigned long __init bootx_dt_find_string(char *str)
+{
+	char *s, *os;
+
+	s = os = (char *)bootx_dt_strbase;
+	s += 4;
+	while (s <  (char *)bootx_dt_strend) {
+		if (strcmp(s, str) == 0)
+			return s - os;
+		s += strlen(s) + 1;
+	}
+	return 0;
+}
+
+static void __init bootx_dt_add_prop(char *name, void *data, int size,
+				  unsigned long *mem_end)
+{
+	unsigned long soff = bootx_dt_find_string(name);
+	if (data == NULL)
+		size = 0;
+	if (soff == 0) {
+		bootx_printf("WARNING: Can't find string index for <%s>\n",
+			     name);
+		return;
+	}
+	if (size > 0x20000) {
+		bootx_printf("WARNING: ignoring large property ");
+		bootx_printf("%s length 0x%x\n", name, size);
+		return;
+	}
+	dt_push_token(OF_DT_PROP, mem_end);
+	dt_push_token(size, mem_end);
+	dt_push_token(soff, mem_end);
+
+	/* push property content */
+	if (size && data) {
+		memcpy((void *)*mem_end, data, size);
+		*mem_end = _ALIGN_UP(*mem_end + size, 4);
+	}
+}
+
+static void __init bootx_add_chosen_props(unsigned long base,
+					  unsigned long *mem_end)
+{
+	u32 val = _MACH_Pmac;
+
+	bootx_dt_add_prop("linux,platform", &val, 4, mem_end);
+
+	if (bootx_info->kernelParamsOffset) {
+		char *args = (char *)((unsigned long)bootx_info) +
+			bootx_info->kernelParamsOffset;
+		bootx_dt_add_prop("bootargs", args, strlen(args) + 1, mem_end);
+	}
+	if (bootx_info->ramDisk) {
+		val = ((unsigned long)bootx_info) + bootx_info->ramDisk;
+		bootx_dt_add_prop("linux,initrd-start", &val, 4, mem_end);
+		val += bootx_info->ramDiskSize;
+		bootx_dt_add_prop("linux,initrd-end", &val, 4, mem_end);
+	}
+	if (strlen(bootx_disp_path))
+		bootx_dt_add_prop("linux,stdout-path", bootx_disp_path,
+				  strlen(bootx_disp_path) + 1, mem_end);
+}
+
+static void __init bootx_add_display_props(unsigned long base,
+					   unsigned long *mem_end)
+{
+	bootx_dt_add_prop("linux,boot-display", NULL, 0, mem_end);
+	bootx_dt_add_prop("linux,opened", NULL, 0, mem_end);
+}
+
+static void __init bootx_dt_add_string(char *s, unsigned long *mem_end)
+{
+	unsigned int l = strlen(s) + 1;
+	memcpy((void *)*mem_end, s, l);
+	bootx_dt_strend = *mem_end = *mem_end + l;
+}
+
+static void __init bootx_scan_dt_build_strings(unsigned long base,
+					       unsigned long node,
+					       unsigned long *mem_end)
+{
+	struct bootx_dt_node *np = (struct bootx_dt_node *)(base + node);
+	u32 *cpp, *ppp = &np->properties;
+	unsigned long soff;
+	char *namep;
+
+	/* Keep refs to known nodes */
+	namep = np->full_name ? (char *)(base + np->full_name) : NULL;
+       	if (namep == NULL) {
+		bootx_printf("Node without a full name !\n");
+		namep = "";
+	}
+	DBG("* strings: %s\n", namep);
+
+	if (!strcmp(namep, "/chosen")) {
+		DBG(" detected /chosen ! adding properties names !\n");
+		bootx_dt_add_string("linux,platform", mem_end);
+		bootx_dt_add_string("linux,stdout-path", mem_end);
+		bootx_dt_add_string("linux,initrd-start", mem_end);
+		bootx_dt_add_string("linux,initrd-end", mem_end);
+		bootx_dt_add_string("bootargs", mem_end);
+		bootx_node_chosen = node;
+	}
+	if (node == bootx_info->dispDeviceRegEntryOffset) {
+		DBG(" detected display ! adding properties names !\n");
+		bootx_dt_add_string("linux,boot-display", mem_end);
+		bootx_dt_add_string("linux,opened", mem_end);
+		strncpy(bootx_disp_path, namep, 255);
+	}
+
+	/* get and store all property names */
+	while (*ppp) {
+		struct bootx_dt_prop *pp =
+			(struct bootx_dt_prop *)(base + *ppp);
+
+		namep = pp->name ? (char *)(base + pp->name) : NULL;
+ 		if (namep == NULL || strcmp(namep, "name") == 0)
+ 			goto next;
+		/* get/create string entry */
+		soff = bootx_dt_find_string(namep);
+		if (soff == 0)
+			bootx_dt_add_string(namep, mem_end);
+	next:
+		ppp = &pp->next;
+	}
+
+	/* do all our children */
+	cpp = &np->child;
+	while(*cpp) {
+		np = (struct bootx_dt_node *)(base + *cpp);
+		bootx_scan_dt_build_strings(base, *cpp, mem_end);
+		cpp = &np->sibling;
+	}
+}
+
+static void __init bootx_scan_dt_build_struct(unsigned long base,
+					      unsigned long node,
+					      unsigned long *mem_end)
+{
+	struct bootx_dt_node *np = (struct bootx_dt_node *)(base + node);
+	u32 *cpp, *ppp = &np->properties;
+	char *namep, *p, *ep, *lp;
+	int l;
+
+	dt_push_token(OF_DT_BEGIN_NODE, mem_end);
+
+	/* get the node's full name */
+	namep = np->full_name ? (char *)(base + np->full_name) : NULL;
+	if (namep == NULL)
+		namep = "";
+	l = strlen(namep);
+
+	DBG("* struct: %s\n", namep);
+
+	/* Fixup an Apple bug where they have bogus \0 chars in the
+	 * middle of the path in some properties, and extract
+	 * the unit name (everything after the last '/').
+	 */
+	memcpy((void *)*mem_end, namep, l + 1);
+	namep = (char *)*mem_end;
+	for (lp = p = namep, ep = namep + l; p < ep; p++) {
+		if (*p == '/')
+			lp = namep;
+		else if (*p != 0)
+			*lp++ = *p;
+	}
+	*lp = 0;
+	*mem_end = _ALIGN_UP((unsigned long)lp + 1, 4);
+
+	/* get and store all properties */
+	while (*ppp) {
+		struct bootx_dt_prop *pp =
+			(struct bootx_dt_prop *)(base + *ppp);
+
+		namep = pp->name ? (char *)(base + pp->name) : NULL;
+		/* Skip "name" */
+ 		if (namep == NULL || !strcmp(namep, "name"))
+ 			goto next;
+		/* Skip "bootargs" in /chosen too as we replace it */
+		if (node == bootx_node_chosen && !strcmp(namep, "bootargs"))
+			goto next;
+
+		/* push property head */
+		bootx_dt_add_prop(namep,
+				  pp->value ? (void *)(base + pp->value): NULL,
+				  pp->length, mem_end);
+	next:
+		ppp = &pp->next;
+	}
+
+	if (node == bootx_node_chosen)
+		bootx_add_chosen_props(base, mem_end);
+	if (node == bootx_info->dispDeviceRegEntryOffset)
+		bootx_add_display_props(base, mem_end);
+
+	/* do all our children */
+	cpp = &np->child;
+	while(*cpp) {
+		np = (struct bootx_dt_node *)(base + *cpp);
+		bootx_scan_dt_build_struct(base, *cpp, mem_end);
+		cpp = &np->sibling;
+	}
+
+	dt_push_token(OF_DT_END_NODE, mem_end);
+}
+
+static unsigned long __init bootx_flatten_dt(unsigned long start)
+{
+	boot_infos_t *bi = bootx_info;
+	unsigned long mem_start, mem_end;
+	struct boot_param_header *hdr;
+	unsigned long base;
+	u64 *rsvmap;
+
+	/* Start using memory after the big blob passed by BootX, get
+	 * some space for the header
+	 */
+	mem_start = mem_end = _ALIGN_UP(((unsigned long)bi) + start, 4);
+	DBG("Boot params header at: %x\n", mem_start);
+	hdr = (struct boot_param_header *)mem_start;
+	mem_end += sizeof(struct boot_param_header);
+	rsvmap = (u64 *)(_ALIGN_UP(mem_end, 8));
+	hdr->off_mem_rsvmap = ((unsigned long)rsvmap) - mem_start;
+	mem_end = ((unsigned long)rsvmap) + 8 * sizeof(u64);
+
+	/* Get base of tree */
+	base = ((unsigned long)bi) + bi->deviceTreeOffset;
+
+	/* Build string array */
+	DBG("Building string array at: %x\n", mem_end);
+	DBG("Device Tree Base=%x\n", base);
+	bootx_dt_strbase = mem_end;
+	mem_end += 4;
+	bootx_dt_strend = mem_end;
+	bootx_scan_dt_build_strings(base, 4, &mem_end);
+	hdr->off_dt_strings = bootx_dt_strbase - mem_start;
+	hdr->dt_strings_size = bootx_dt_strend - bootx_dt_strbase;
+
+	/* Build structure */
+	mem_end = _ALIGN(mem_end, 16);
+	DBG("Building device tree structure at: %x\n", mem_end);
+	hdr->off_dt_struct = mem_end - mem_start;
+	bootx_scan_dt_build_struct(base, 4, &mem_end);
+	dt_push_token(OF_DT_END, &mem_end);
+
+	/* Finish header */
+	hdr->boot_cpuid_phys = 0;
+	hdr->magic = OF_DT_HEADER;
+	hdr->totalsize = mem_end - mem_start;
+	hdr->version = OF_DT_VERSION;
+	/* Version 16 is not backward compatible */
+	hdr->last_comp_version = 0x10;
+
+	/* Reserve the whole thing and copy the reserve map in, we
+	 * also bump mem_reserve_cnt to cause further reservations to
+	 * fail since it's too late.
+	 */
+	mem_end = _ALIGN(mem_end, PAGE_SIZE);
+	DBG("End of boot params: %x\n", mem_end);
+	rsvmap[0] = mem_start;
+	rsvmap[1] = mem_end;
+	rsvmap[2] = 0;
+	rsvmap[3] = 0;
+
+	return (unsigned long)hdr;
+}
+
+
+#ifdef CONFIG_BOOTX_TEXT
+static void __init btext_welcome(boot_infos_t *bi)
+{
+	unsigned long flags;
+	unsigned long pvr;
+
+	bootx_printf("Welcome to Linux, kernel " UTS_RELEASE "\n");
+	bootx_printf("\nlinked at        : 0x%x", KERNELBASE);
+	bootx_printf("\nframe buffer at  : 0x%x", bi->dispDeviceBase);
+	bootx_printf(" (phys), 0x%x", bi->logicalDisplayBase);
+	bootx_printf(" (log)");
+	bootx_printf("\nklimit           : 0x%x",(unsigned long)klimit);
+	bootx_printf("\nboot_info at     : 0x%x", bi);
+	__asm__ __volatile__ ("mfmsr %0" : "=r" (flags));
+	bootx_printf("\nMSR              : 0x%x", flags);
+	__asm__ __volatile__ ("mfspr %0, 287" : "=r" (pvr));
+	bootx_printf("\nPVR              : 0x%x", pvr);
+	pvr >>= 16;
+	if (pvr > 1) {
+	    __asm__ __volatile__ ("mfspr %0, 1008" : "=r" (flags));
+	    bootx_printf("\nHID0             : 0x%x", flags);
+	}
+	if (pvr == 8 || pvr == 12 || pvr == 0x800c) {
+	    __asm__ __volatile__ ("mfspr %0, 1019" : "=r" (flags));
+	    bootx_printf("\nICTC             : 0x%x", flags);
+	}
+#ifdef DEBUG
+	bootx_printf("\n\n");
+	bootx_printf("bi->deviceTreeOffset   : 0x%x\n",
+		     bi->deviceTreeOffset);
+	bootx_printf("bi->deviceTreeSize     : 0x%x\n",
+		     bi->deviceTreeSize);
+#endif
+	bootx_printf("\n\n");
+}
+#endif /* CONFIG_BOOTX_TEXT */
+
+void __init bootx_init(unsigned long r3, unsigned long r4)
+{
+	boot_infos_t *bi = (boot_infos_t *) r4;
+	unsigned long hdr;
+	unsigned long space;
+	unsigned long ptr, x;
+	char *model;
+	unsigned long offset = reloc_offset();
+
+	reloc_got2(offset);
+
+	bootx_info = bi;
+
+	/* We haven't cleared any bss at this point, make sure
+	 * what we need is initialized
+	 */
+	bootx_dt_strbase = bootx_dt_strend = 0;
+	bootx_node_chosen = 0;
+	bootx_disp_path[0] = 0;
+
+	if (!BOOT_INFO_IS_V2_COMPATIBLE(bi))
+		bi->logicalDisplayBase = bi->dispDeviceBase;
+
+#ifdef CONFIG_BOOTX_TEXT
+	btext_setup_display(bi->dispDeviceRect[2] - bi->dispDeviceRect[0],
+			    bi->dispDeviceRect[3] - bi->dispDeviceRect[1],
+			    bi->dispDeviceDepth, bi->dispDeviceRowBytes,
+			    (unsigned long)bi->logicalDisplayBase);
+	btext_clearscreen();
+	btext_flushscreen();
+#endif /* CONFIG_BOOTX_TEXT */
+
+	/*
+	 * Test if boot-info is compatible.  Done only in config
+	 * CONFIG_BOOTX_TEXT since there is nothing much we can do
+	 * with an incompatible version, except display a message
+	 * and eventually hang the processor...
+	 *
+	 * I'll try to keep enough of boot-info compatible in the
+	 * future to always allow display of this message;
+	 */
+	if (!BOOT_INFO_IS_COMPATIBLE(bi)) {
+		bootx_printf(" !!! WARNING - Incompatible version"
+			     " of BootX !!!\n\n\n");
+		for (;;)
+			;
+	}
+	if (bi->architecture != BOOT_ARCH_PCI) {
+		bootx_printf(" !!! WARNING - Usupported machine"
+			     " architecture !\n");
+		for (;;)
+			;
+	}
+
+#ifdef CONFIG_BOOTX_TEXT
+	btext_welcome(bi);
+#endif
+	/* New BootX enters kernel with MMU off, i/os are not allowed
+	 * here. This hack will have been done by the boostrap anyway.
+	 */
+	if (bi->version < 4) {
+		/*
+		 * XXX If this is an iMac, turn off the USB controller.
+		 */
+		model = (char *) bootx_early_getprop(r4 + bi->deviceTreeOffset,
+						     4, "model");
+		if (model
+		    && (strcmp(model, "iMac,1") == 0
+			|| strcmp(model, "PowerMac1,1") == 0)) {
+			bootx_printf("iMac,1 detected, shutting down USB \n");
+			out_le32((unsigned *)0x80880008, 1);	/* XXX */
+		}
+	}
+
+	/* Get a pointer that points above the device tree, args, ramdisk,
+	 * etc... to use for generating the flattened tree
+	 */
+	if (bi->version < 5) {
+		space = bi->deviceTreeOffset + bi->deviceTreeSize;
+		if (bi->ramDisk)
+			space = bi->ramDisk + bi->ramDiskSize;
+	} else
+		space = bi->totalParamsSize;
+
+	bootx_printf("Total space used by parameters & ramdisk: %x \n", space);
+
+	/* New BootX will have flushed all TLBs and enters kernel with
+	 * MMU switched OFF, so this should not be useful anymore.
+	 */
+	if (bi->version < 4) {
+		bootx_printf("Touching pages...\n");
+
+		/*
+		 * Touch each page to make sure the PTEs for them
+		 * are in the hash table - the aim is to try to avoid
+		 * getting DSI exceptions while copying the kernel image.
+		 */
+		for (ptr = ((unsigned long) &_stext) & PAGE_MASK;
+		     ptr < (unsigned long)bi + space; ptr += PAGE_SIZE)
+			x = *(volatile unsigned long *)ptr;
+	}
+
+	/* Ok, now we need to generate a flattened device-tree to pass
+	 * to the kernel
+	 */
+	bootx_printf("Preparing boot params...\n");
+
+	hdr = bootx_flatten_dt(space);
+
+#ifdef CONFIG_BOOTX_TEXT
+#ifdef SET_BOOT_BAT
+	bootx_printf("Preparing BAT...\n");
+	btext_prepare_BAT();
+#else
+	btext_unmap();
+#endif
+#endif
+
+	reloc_got2(-offset);
+
+	__start(hdr, KERNELBASE + offset, 0);
+}

arch/powerpc/platforms/powermac/cpufreq_64.c

@@ -28,6 +28,7 @@
 #include <asm/cputable.h>
 #include <asm/time.h>
 #include <asm/smu.h>
+#include <asm/pmac_pfunc.h>
 
 #undef DEBUG
 
@@ -85,6 +86,10 @@ static u32 *g5_pmode_data;
 static int g5_pmode_max;
 static int g5_pmode_cur;
 
+static void (*g5_switch_volt)(int speed_mode);
+static int (*g5_switch_freq)(int speed_mode);
+static int (*g5_query_freq)(void);
+
 static DECLARE_MUTEX(g5_switch_mutex);
 
 
@@ -92,9 +97,11 @@ static struct smu_sdbp_fvt *g5_fvt_table;	/* table of op. points */
 static int g5_fvt_count;			/* number of op. points */
 static int g5_fvt_cur;				/* current op. point */
 
-/* ----------------- real hardware interface */
+/*
+ * SMU based voltage switching for Neo2 platforms
+ */
 
-static void g5_switch_volt(int speed_mode)
+static void g5_smu_switch_volt(int speed_mode)
 {
 	struct smu_simple_cmd	cmd;
 
@@ -105,26 +112,57 @@ static void g5_switch_volt(int speed_mode)
 	wait_for_completion(&comp);
 }
 
-static int g5_switch_freq(int speed_mode)
+/*
+ * Platform function based voltage/vdnap switching for Neo2
+ */
+
+static struct pmf_function *pfunc_set_vdnap0;
+static struct pmf_function *pfunc_vdnap0_complete;
+
+static void g5_vdnap_switch_volt(int speed_mode)
 {
-	struct cpufreq_freqs freqs;
+	struct pmf_args args;
+	u32 slew, done = 0;
+	unsigned long timeout;
+
+	slew = (speed_mode == CPUFREQ_LOW) ? 1 : 0;
+	args.count = 1;
+	args.u[0].p = &slew;
+
+	pmf_call_one(pfunc_set_vdnap0, &args);
+
+	/* It's an irq GPIO so we should be able to just block here,
+	 * I'll do that later after I've properly tested the IRQ code for
+	 * platform functions
+	 */
+	timeout = jiffies + HZ/10;
+	while(!time_after(jiffies, timeout)) {
+		args.count = 1;
+		args.u[0].p = &done;
+		pmf_call_one(pfunc_vdnap0_complete, &args);
+		if (done)
+			break;
+		msleep(1);
+	}
+	if (done == 0)
+		printk(KERN_WARNING "cpufreq: Timeout in clock slewing !\n");
+}
+
+
+/*
+ * SCOM based frequency switching for 970FX rev3
+ */
+static int g5_scom_switch_freq(int speed_mode)
+{
+	unsigned long flags;
 	int to;
 
-	if (g5_pmode_cur == speed_mode)
-		return 0;
-
-	down(&g5_switch_mutex);
-
-	freqs.old = g5_cpu_freqs[g5_pmode_cur].frequency;
-	freqs.new = g5_cpu_freqs[speed_mode].frequency;
-	freqs.cpu = 0;
-
-	cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE);
-
 	/* If frequency is going up, first ramp up the voltage */
 	if (speed_mode < g5_pmode_cur)
 		g5_switch_volt(speed_mode);
 
+	local_irq_save(flags);
+
 	/* Clear PCR high */
 	scom970_write(SCOM_PCR, 0);
 	/* Clear PCR low */
@@ -147,6 +185,8 @@ static int g5_switch_freq(int speed_mode)
 		udelay(100);
 	}
 
+	local_irq_restore(flags);
+
 	/* If frequency is going down, last ramp the voltage */
 	if (speed_mode > g5_pmode_cur)
 		g5_switch_volt(speed_mode);
@@ -154,14 +194,10 @@ static int g5_switch_freq(int speed_mode)
 	g5_pmode_cur = speed_mode;
 	ppc_proc_freq = g5_cpu_freqs[speed_mode].frequency * 1000ul;
 
-	cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE);
-
-	up(&g5_switch_mutex);
-
 	return 0;
 }
 
-static int g5_query_freq(void)
+static int g5_scom_query_freq(void)
 {
 	unsigned long psr = scom970_read(SCOM_PSR);
 	int i;
@@ -173,7 +209,104 @@ static int g5_query_freq(void)
 	return i;
 }
 
-/* ----------------- cpufreq bookkeeping */
+/*
+ * Platform function based voltage switching for PowerMac7,2 & 7,3
+ */
+
+static struct pmf_function *pfunc_cpu0_volt_high;
+static struct pmf_function *pfunc_cpu0_volt_low;
+static struct pmf_function *pfunc_cpu1_volt_high;
+static struct pmf_function *pfunc_cpu1_volt_low;
+
+static void g5_pfunc_switch_volt(int speed_mode)
+{
+	if (speed_mode == CPUFREQ_HIGH) {
+		if (pfunc_cpu0_volt_high)
+			pmf_call_one(pfunc_cpu0_volt_high, NULL);
+		if (pfunc_cpu1_volt_high)
+			pmf_call_one(pfunc_cpu1_volt_high, NULL);
+	} else {
+		if (pfunc_cpu0_volt_low)
+			pmf_call_one(pfunc_cpu0_volt_low, NULL);
+		if (pfunc_cpu1_volt_low)
+			pmf_call_one(pfunc_cpu1_volt_low, NULL);
+	}
+	msleep(10); /* should be faster , to fix */
+}
+
+/*
+ * Platform function based frequency switching for PowerMac7,2 & 7,3
+ */
+
+static struct pmf_function *pfunc_cpu_setfreq_high;
+static struct pmf_function *pfunc_cpu_setfreq_low;
+static struct pmf_function *pfunc_cpu_getfreq;
+static struct pmf_function *pfunc_slewing_done;;
+
+static int g5_pfunc_switch_freq(int speed_mode)
+{
+	struct pmf_args args;
+	u32 done = 0;
+	unsigned long timeout;
+
+	/* If frequency is going up, first ramp up the voltage */
+	if (speed_mode < g5_pmode_cur)
+		g5_switch_volt(speed_mode);
+
+	/* Do it */
+	if (speed_mode == CPUFREQ_HIGH)
+		pmf_call_one(pfunc_cpu_setfreq_high, NULL);
+	else
+		pmf_call_one(pfunc_cpu_setfreq_low, NULL);
+
+	/* It's an irq GPIO so we should be able to just block here,
+	 * I'll do that later after I've properly tested the IRQ code for
+	 * platform functions
+	 */
+	timeout = jiffies + HZ/10;
+	while(!time_after(jiffies, timeout)) {
+		args.count = 1;
+		args.u[0].p = &done;
+		pmf_call_one(pfunc_slewing_done, &args);
+		if (done)
+			break;
+		msleep(1);
+	}
+	if (done == 0)
+		printk(KERN_WARNING "cpufreq: Timeout in clock slewing !\n");
+
+	/* If frequency is going down, last ramp the voltage */
+	if (speed_mode > g5_pmode_cur)
+		g5_switch_volt(speed_mode);
+
+	g5_pmode_cur = speed_mode;
+	ppc_proc_freq = g5_cpu_freqs[speed_mode].frequency * 1000ul;
+
+	return 0;
+}
+
+static int g5_pfunc_query_freq(void)
+{
+	struct pmf_args args;
+	u32 val = 0;
+
+	args.count = 1;
+	args.u[0].p = &val;
+	pmf_call_one(pfunc_cpu_getfreq, &args);
+	return val ? CPUFREQ_HIGH : CPUFREQ_LOW;
+}
+
+/*
+ * Fake voltage switching for platforms with missing support
+ */
+
+static void g5_dummy_switch_volt(int speed_mode)
+{
+}
+
+/*
+ * Common interface to the cpufreq core
+ */
 
 static int g5_cpufreq_verify(struct cpufreq_policy *policy)
 {
@@ -183,13 +316,30 @@ static int g5_cpufreq_verify(struct cpufreq_policy *policy)
 static int g5_cpufreq_target(struct cpufreq_policy *policy,
 	unsigned int target_freq, unsigned int relation)
 {
-	unsigned int    newstate = 0;
+	unsigned int newstate = 0;
+	struct cpufreq_freqs freqs;
+	int rc;
 
 	if (cpufreq_frequency_table_target(policy, g5_cpu_freqs,
 			target_freq, relation, &newstate))
 		return -EINVAL;
 
-	return g5_switch_freq(newstate);
+	if (g5_pmode_cur == newstate)
+		return 0;
+
+	down(&g5_switch_mutex);
+
+	freqs.old = g5_cpu_freqs[g5_pmode_cur].frequency;
+	freqs.new = g5_cpu_freqs[newstate].frequency;
+	freqs.cpu = 0;
+
+	cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE);
+	rc = g5_switch_freq(newstate);
+	cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE);
+
+	up(&g5_switch_mutex);
+
+	return rc;
 }
 
 static unsigned int g5_cpufreq_get_speed(unsigned int cpu)
@@ -205,6 +355,7 @@ static int g5_cpufreq_cpu_init(struct cpufreq_policy *policy)
 	policy->governor = CPUFREQ_DEFAULT_GOVERNOR;
 	policy->cpuinfo.transition_latency = CPUFREQ_ETERNAL;
 	policy->cur = g5_cpu_freqs[g5_query_freq()].frequency;
+	policy->cpus = cpu_possible_map;
 	cpufreq_frequency_table_get_attr(g5_cpu_freqs, policy->cpu);
 
 	return cpufreq_frequency_table_cpuinfo(policy,
@@ -224,19 +375,39 @@ static struct cpufreq_driver g5_cpufreq_driver = {
 };
 
 
-static int __init g5_cpufreq_init(void)
+static int __init g5_neo2_cpufreq_init(struct device_node *cpus)
 {
 	struct device_node *cpunode;
 	unsigned int psize, ssize;
-	struct smu_sdbp_header *shdr;
 	unsigned long max_freq;
-	u32 *valp;
+	char *freq_method, *volt_method;
+	u32 *valp, pvr_hi;
+	int use_volts_vdnap = 0;
+	int use_volts_smu = 0;
 	int rc = -ENODEV;
 
-	/* Look for CPU and SMU nodes */
-	cpunode = of_find_node_by_type(NULL, "cpu");
-	if (!cpunode) {
-		DBG("No CPU node !\n");
+	/* Check supported platforms */
+	if (machine_is_compatible("PowerMac8,1") ||
+	    machine_is_compatible("PowerMac8,2") ||
+	    machine_is_compatible("PowerMac9,1"))
+		use_volts_smu = 1;
+	else if (machine_is_compatible("PowerMac11,2"))
+		use_volts_vdnap = 1;
+	else
+		return -ENODEV;
+
+	/* Get first CPU node */
+	for (cpunode = NULL;
+	     (cpunode = of_get_next_child(cpus, cpunode)) != NULL;) {
+		u32 *reg =
+			(u32 *)get_property(cpunode, "reg", NULL);
+		if (reg == NULL || (*reg) != 0)
+			continue;
+		if (!strcmp(cpunode->type, "cpu"))
+			break;
+	}
+	if (cpunode == NULL) {
+		printk(KERN_ERR "cpufreq: Can't find any CPU 0 node\n");
 		return -ENODEV;
 	}
 
@@ -246,8 +417,9 @@ static int __init g5_cpufreq_init(void)
 		DBG("No cpu-version property !\n");
 		goto bail_noprops;
 	}
-	if (((*valp) >> 16) != 0x3c) {
-		DBG("Wrong CPU version: %08x\n", *valp);
+	pvr_hi = (*valp) >> 16;
+	if (pvr_hi != 0x3c && pvr_hi != 0x44) {
+		printk(KERN_ERR "cpufreq: Unsupported CPU version\n");
 		goto bail_noprops;
 	}
 
@@ -259,18 +431,50 @@ static int __init g5_cpufreq_init(void)
 	}
 	g5_pmode_max = psize / sizeof(u32) - 1;
 
-	/* Look for the FVT table */
-	shdr = smu_get_sdb_partition(SMU_SDB_FVT_ID, NULL);
-	if (!shdr)
-		goto bail_noprops;
-	g5_fvt_table = (struct smu_sdbp_fvt *)&shdr[1];
-	ssize = (shdr->len * sizeof(u32)) - sizeof(struct smu_sdbp_header);
-	g5_fvt_count = ssize / sizeof(struct smu_sdbp_fvt);
-	g5_fvt_cur = 0;
+	if (use_volts_smu) {
+		struct smu_sdbp_header *shdr;
 
-	/* Sanity checking */
-	if (g5_fvt_count < 1 || g5_pmode_max < 1)
-		goto bail_noprops;
+		/* Look for the FVT table */
+		shdr = smu_get_sdb_partition(SMU_SDB_FVT_ID, NULL);
+		if (!shdr)
+			goto bail_noprops;
+		g5_fvt_table = (struct smu_sdbp_fvt *)&shdr[1];
+		ssize = (shdr->len * sizeof(u32)) -
+			sizeof(struct smu_sdbp_header);
+		g5_fvt_count = ssize / sizeof(struct smu_sdbp_fvt);
+		g5_fvt_cur = 0;
+
+		/* Sanity checking */
+		if (g5_fvt_count < 1 || g5_pmode_max < 1)
+			goto bail_noprops;
+
+		g5_switch_volt = g5_smu_switch_volt;
+		volt_method = "SMU";
+	} else if (use_volts_vdnap) {
+		struct device_node *root;
+
+		root = of_find_node_by_path("/");
+		if (root == NULL) {
+			printk(KERN_ERR "cpufreq: Can't find root of "
+			       "device tree\n");
+			goto bail_noprops;
+		}
+		pfunc_set_vdnap0 = pmf_find_function(root, "set-vdnap0");
+		pfunc_vdnap0_complete =
+			pmf_find_function(root, "slewing-done");
+		if (pfunc_set_vdnap0 == NULL ||
+		    pfunc_vdnap0_complete == NULL) {
+			printk(KERN_ERR "cpufreq: Can't find required "
+			       "platform function\n");
+			goto bail_noprops;
+		}
+
+		g5_switch_volt = g5_vdnap_switch_volt;
+		volt_method = "GPIO";
+	} else {
+		g5_switch_volt = g5_dummy_switch_volt;
+		volt_method = "none";
+	}
 
 	/*
 	 * From what I see, clock-frequency is always the maximal frequency.
@@ -286,19 +490,23 @@ static int __init g5_cpufreq_init(void)
 	g5_cpu_freqs[0].frequency = max_freq;
 	g5_cpu_freqs[1].frequency = max_freq/2;
 
-	/* Check current frequency */
-	g5_pmode_cur = g5_query_freq();
-	if (g5_pmode_cur > 1)
-		/* We don't support anything but 1:1 and 1:2, fixup ... */
-		g5_pmode_cur = 1;
+	/* Set callbacks */
+	g5_switch_freq = g5_scom_switch_freq;
+	g5_query_freq = g5_scom_query_freq;
+	freq_method = "SCOM";
 
 	/* Force apply current frequency to make sure everything is in
 	 * sync (voltage is right for example). Firmware may leave us with
 	 * a strange setting ...
 	 */
-	g5_switch_freq(g5_pmode_cur);
+	g5_switch_volt(CPUFREQ_HIGH);
+	msleep(10);
+	g5_pmode_cur = -1;
+	g5_switch_freq(g5_query_freq());
 
 	printk(KERN_INFO "Registering G5 CPU frequency driver\n");
+	printk(KERN_INFO "Frequency method: %s, Voltage method: %s\n",
+	       freq_method, volt_method);
 	printk(KERN_INFO "Low: %d Mhz, High: %d Mhz, Cur: %d MHz\n",
 		g5_cpu_freqs[1].frequency/1000,
 		g5_cpu_freqs[0].frequency/1000,
@@ -317,6 +525,200 @@ static int __init g5_cpufreq_init(void)
 	return rc;
 }
 
+static int __init g5_pm72_cpufreq_init(struct device_node *cpus)
+{
+	struct device_node *cpuid = NULL, *hwclock = NULL, *cpunode = NULL;
+	u8 *eeprom = NULL;
+	u32 *valp;
+	u64 max_freq, min_freq, ih, il;
+	int has_volt = 1, rc = 0;
+
+	/* Get first CPU node */
+	for (cpunode = NULL;
+	     (cpunode = of_get_next_child(cpus, cpunode)) != NULL;) {
+		if (!strcmp(cpunode->type, "cpu"))
+			break;
+	}
+	if (cpunode == NULL) {
+		printk(KERN_ERR "cpufreq: Can't find any CPU node\n");
+		return -ENODEV;
+	}
+
+	/* Lookup the cpuid eeprom node */
+        cpuid = of_find_node_by_path("/u3@0,f8000000/i2c@f8001000/cpuid@a0");
+	if (cpuid != NULL)
+		eeprom = (u8 *)get_property(cpuid, "cpuid", NULL);
+	if (eeprom == NULL) {
+		printk(KERN_ERR "cpufreq: Can't find cpuid EEPROM !\n");
+		rc = -ENODEV;
+		goto bail;
+	}
+
+	/* Lookup the i2c hwclock */
+	for (hwclock = NULL;
+	     (hwclock = of_find_node_by_name(hwclock, "i2c-hwclock")) != NULL;){
+		char *loc = get_property(hwclock, "hwctrl-location", NULL);
+		if (loc == NULL)
+			continue;
+		if (strcmp(loc, "CPU CLOCK"))
+			continue;
+		if (!get_property(hwclock, "platform-get-frequency", NULL))
+			continue;
+		break;
+	}
+	if (hwclock == NULL) {
+		printk(KERN_ERR "cpufreq: Can't find i2c clock chip !\n");
+		rc = -ENODEV;
+		goto bail;
+	}
+
+	DBG("cpufreq: i2c clock chip found: %s\n", hwclock->full_name);
+
+	/* Now get all the platform functions */
+	pfunc_cpu_getfreq =
+		pmf_find_function(hwclock, "get-frequency");
+	pfunc_cpu_setfreq_high =
+		pmf_find_function(hwclock, "set-frequency-high");
+	pfunc_cpu_setfreq_low =
+		pmf_find_function(hwclock, "set-frequency-low");
+	pfunc_slewing_done =
+		pmf_find_function(hwclock, "slewing-done");
+	pfunc_cpu0_volt_high =
+		pmf_find_function(hwclock, "set-voltage-high-0");
+	pfunc_cpu0_volt_low =
+		pmf_find_function(hwclock, "set-voltage-low-0");
+	pfunc_cpu1_volt_high =
+		pmf_find_function(hwclock, "set-voltage-high-1");
+	pfunc_cpu1_volt_low =
+		pmf_find_function(hwclock, "set-voltage-low-1");
+
+	/* Check we have minimum requirements */
+	if (pfunc_cpu_getfreq == NULL || pfunc_cpu_setfreq_high == NULL ||
+	    pfunc_cpu_setfreq_low == NULL || pfunc_slewing_done == NULL) {
+		printk(KERN_ERR "cpufreq: Can't find platform functions !\n");
+		rc = -ENODEV;
+		goto bail;
+	}
+
+	/* Check that we have complete sets */
+	if (pfunc_cpu0_volt_high == NULL || pfunc_cpu0_volt_low == NULL) {
+		pmf_put_function(pfunc_cpu0_volt_high);
+		pmf_put_function(pfunc_cpu0_volt_low);
+		pfunc_cpu0_volt_high = pfunc_cpu0_volt_low = NULL;
+		has_volt = 0;
+	}
+	if (!has_volt ||
+	    pfunc_cpu1_volt_high == NULL || pfunc_cpu1_volt_low == NULL) {
+		pmf_put_function(pfunc_cpu1_volt_high);
+		pmf_put_function(pfunc_cpu1_volt_low);
+		pfunc_cpu1_volt_high = pfunc_cpu1_volt_low = NULL;
+	}
+
+	/* Note: The device tree also contains a "platform-set-values"
+	 * function for which I haven't quite figured out the usage. It
+	 * might have to be called on init and/or wakeup, I'm not too sure
+	 * but things seem to work fine without it so far ...
+	 */
+
+	/* Get max frequency from device-tree */
+	valp = (u32 *)get_property(cpunode, "clock-frequency", NULL);
+	if (!valp) {
+		printk(KERN_ERR "cpufreq: Can't find CPU frequency !\n");
+		rc = -ENODEV;
+		goto bail;
+	}
+
+	max_freq = (*valp)/1000;
+
+	/* Now calculate reduced frequency by using the cpuid input freq
+	 * ratio. This requires 64 bits math unless we are willing to lose
+	 * some precision
+	 */
+	ih = *((u32 *)(eeprom + 0x10));
+	il = *((u32 *)(eeprom + 0x20));
+	min_freq = 0;
+	if (ih != 0 && il != 0)
+		min_freq = (max_freq * il) / ih;
+
+	/* Sanity check */
+	if (min_freq >= max_freq || min_freq < 1000) {
+		printk(KERN_ERR "cpufreq: Can't calculate low frequency !\n");
+		rc = -ENODEV;
+		goto bail;
+	}
+	g5_cpu_freqs[0].frequency = max_freq;
+	g5_cpu_freqs[1].frequency = min_freq;
+
+	/* Set callbacks */
+	g5_switch_volt = g5_pfunc_switch_volt;
+	g5_switch_freq = g5_pfunc_switch_freq;
+	g5_query_freq = g5_pfunc_query_freq;
+
+	/* Force apply current frequency to make sure everything is in
+	 * sync (voltage is right for example). Firmware may leave us with
+	 * a strange setting ...
+	 */
+	g5_switch_volt(CPUFREQ_HIGH);
+	msleep(10);
+	g5_pmode_cur = -1;
+	g5_switch_freq(g5_query_freq());
+
+	printk(KERN_INFO "Registering G5 CPU frequency driver\n");
+	printk(KERN_INFO "Frequency method: i2c/pfunc, "
+	       "Voltage method: %s\n", has_volt ? "i2c/pfunc" : "none");
+	printk(KERN_INFO "Low: %d Mhz, High: %d Mhz, Cur: %d MHz\n",
+		g5_cpu_freqs[1].frequency/1000,
+		g5_cpu_freqs[0].frequency/1000,
+		g5_cpu_freqs[g5_pmode_cur].frequency/1000);
+
+	rc = cpufreq_register_driver(&g5_cpufreq_driver);
+ bail:
+	if (rc != 0) {
+		pmf_put_function(pfunc_cpu_getfreq);
+		pmf_put_function(pfunc_cpu_setfreq_high);
+		pmf_put_function(pfunc_cpu_setfreq_low);
+		pmf_put_function(pfunc_slewing_done);
+		pmf_put_function(pfunc_cpu0_volt_high);
+		pmf_put_function(pfunc_cpu0_volt_low);
+		pmf_put_function(pfunc_cpu1_volt_high);
+		pmf_put_function(pfunc_cpu1_volt_low);
+	}
+	of_node_put(hwclock);
+	of_node_put(cpuid);
+	of_node_put(cpunode);
+
+	return rc;
+}
+
+static int __init g5_rm31_cpufreq_init(struct device_node *cpus)
+{
+	/* NYI */
+	return 0;
+}
+
+static int __init g5_cpufreq_init(void)
+{
+	struct device_node *cpus;
+	int rc;
+
+	cpus = of_find_node_by_path("/cpus");
+	if (cpus == NULL) {
+		DBG("No /cpus node !\n");
+		return -ENODEV;
+	}
+
+	if (machine_is_compatible("PowerMac7,2") ||
+	    machine_is_compatible("PowerMac7,3"))
+		rc = g5_pm72_cpufreq_init(cpus);
+	else if (machine_is_compatible("RackMac3,1"))
+		rc = g5_rm31_cpufreq_init(cpus);
+	else
+		rc = g5_neo2_cpufreq_init(cpus);
+
+	of_node_put(cpus);
+	return rc;
+}
+
 module_init(g5_cpufreq_init);