Merge tag 'powerpc-4.9-1' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux

Pull powerpc updates from Michael Ellerman:
 "Highlights:
   - Major rework of Book3S 64-bit exception vectors (Nicholas Piggin)
   - Use gas sections for arranging exception vectors et. al.
   - Large set of TM cleanups and selftests (Cyril Bur)
   - Enable transactional memory (TM) lazily for userspace (Cyril Bur)
   - Support for XZ compression in the zImage wrapper (Oliver
     O'Halloran)
   - Add support for bpf constant blinding (Naveen N. Rao)
   - Beginnings of upstream support for PA Semi Nemo motherboards
     (Darren Stevens)

  Fixes:
   - Ensure .mem(init|exit).text are within _stext/_etext (Michael
     Ellerman)
   - xmon: Don't use ld on 32-bit (Michael Ellerman)
   - vdso64: Use double word compare on pointers (Anton Blanchard)
   - powerpc/nvram: Fix an incorrect partition merge (Pan Xinhui)
   - powerpc: Fix usage of _PAGE_RO in hugepage (Christophe Leroy)
   - powerpc/mm: Update FORCE_MAX_ZONEORDER range to allow hugetlb w/4K
     (Aneesh Kumar K.V)
   - Fix memory leak in queue_hotplug_event() error path (Andrew
     Donnellan)
   - Replay hypervisor maintenance interrupt first (Nicholas Piggin)

  Various performance optimisations (Anton Blanchard):
   - Align hot loops of memset() and backwards_memcpy()
   - During context switch, check before setting mm_cpumask
   - Remove static branch prediction in atomic{, 64}_add_unless
   - Only disable HAVE_EFFICIENT_UNALIGNED_ACCESS on POWER7 little
     endian
   - Set default CPU type to POWER8 for little endian builds

  Cleanups & features:
   - Sparse fixes/cleanups (Daniel Axtens)
   - Preserve CFAR value on SLB miss caused by access to bogus address
     (Paul Mackerras)
   - Radix MMU fixups for POWER9 (Aneesh Kumar K.V)
   - Support for setting used_(vsr|vr|spe) in sigreturn path (for CRIU)
     (Simon Guo)
   - Optimise syscall entry for virtual, relocatable case (Nicholas
     Piggin)
   - Optimise MSR handling in exception handling (Nicholas Piggin)
   - Support for kexec with Radix MMU (Benjamin Herrenschmidt)
   - powernv EEH fixes (Russell Currey)
   - Suprise PCI hotplug support for powernv (Gavin Shan)
   - Endian/sparse fixes for powernv PCI (Gavin Shan)
   - Defconfig updates (Anton Blanchard)
   - KVM: PPC: Book3S HV: Migrate pinned pages out of CMA (Balbir Singh)
   - cxl: Flush PSL cache before resetting the adapter (Frederic Barrat)
   - cxl: replace loop with for_each_child_of_node(), remove unneeded
     of_node_put() (Andrew Donnellan)
   - Fix HV facility unavailable to use correct handler (Nicholas
     Piggin)
   - Remove unnecessary syscall trampoline (Nicholas Piggin)
   - fadump: Fix build break when CONFIG_PROC_VMCORE=n (Michael
     Ellerman)
   - Quieten EEH message when no adapters are found (Anton Blanchard)
   - powernv: Add PHB register dump debugfs handle (Russell Currey)
   - Use kprobe blacklist for exception handlers & asm functions
     (Nicholas Piggin)
   - Document the syscall ABI (Nicholas Piggin)
   - MAINTAINERS: Update cxl maintainers (Michael Neuling)
   - powerpc: Remove all usages of NO_IRQ (Michael Ellerman)

  Minor cleanups:
   - Andrew Donnellan, Christophe Leroy, Colin Ian King, Cyril Bur,
     Frederic Barrat, Pan Xinhui, PrasannaKumar Muralidharan, Rui Teng,
     Simon Guo"

* tag 'powerpc-4.9-1' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux: (156 commits)
  powerpc/bpf: Add support for bpf constant blinding
  powerpc/bpf: Implement support for tail calls
  powerpc/bpf: Introduce accessors for using the tmp local stack space
  powerpc/fadump: Fix build break when CONFIG_PROC_VMCORE=n
  powerpc: tm: Enable transactional memory (TM) lazily for userspace
  powerpc/tm: Add TM Unavailable Exception
  powerpc: Remove do_load_up_transact_{fpu,altivec}
  powerpc: tm: Rename transct_(*) to ck(\1)_state
  powerpc: tm: Always use fp_state and vr_state to store live registers
  selftests/powerpc: Add checks for transactional VSXs in signal contexts
  selftests/powerpc: Add checks for transactional VMXs in signal contexts
  selftests/powerpc: Add checks for transactional FPUs in signal contexts
  selftests/powerpc: Add checks for transactional GPRs in signal contexts
  selftests/powerpc: Check that signals always get delivered
  selftests/powerpc: Add TM tcheck helpers in C
  selftests/powerpc: Allow tests to extend their kill timeout
  selftests/powerpc: Introduce GPR asm helper header file
  selftests/powerpc: Move VMX stack frame macros to header file
  selftests/powerpc: Rework FPU stack placement macros and move to header file
  selftests/powerpc: Check for VSX preservation across userspace preemption
  ...

13 files changed, 361 insertions, 314 deletions

diff --git a/arch/powerpc/boot/Makefile b/arch/powerpc/boot/Makefile
index 1a2a6e8dc40d..eae2dc8bc218 100644
--- a/arch/powerpc/boot/Makefile
+++ b/arch/powerpc/boot/Makefile
@@ -19,10 +19,15 @@
 
 all: $(obj)/zImage
 
+compress-$(CONFIG_KERNEL_GZIP) := CONFIG_KERNEL_GZIP
+compress-$(CONFIG_KERNEL_XZ)   := CONFIG_KERNEL_XZ
+
 BOOTCFLAGS    := -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs \
 		 -fno-strict-aliasing -Os -msoft-float -pipe \
 		 -fomit-frame-pointer -fno-builtin -fPIC -nostdinc \
-		 -isystem $(shell $(CROSS32CC) -print-file-name=include)
+		 -isystem $(shell $(CROSS32CC) -print-file-name=include) \
+		 -D$(compress-y)
+
 ifdef CONFIG_PPC64_BOOT_WRAPPER
 BOOTCFLAGS	+= -m64
 endif
@@ -59,13 +64,30 @@ $(obj)/treeboot-currituck.o: BOOTCFLAGS += -mcpu=405
 $(obj)/treeboot-akebono.o: BOOTCFLAGS += -mcpu=405
 $(obj)/virtex405-head.o: BOOTAFLAGS += -mcpu=405
 
+# The pre-boot decompressors pull in a lot of kernel headers and other source
+# files. This creates a bit of a dependency headache since we need to copy
+# these files into the build dir, fix up any includes and ensure that dependent
+# files are copied in the right order.
+
+# these need to be seperate variables because they are copied out of different
+# directories in the kernel tree. Sure you COULd merge them, but it's a
+# cure-is-worse-than-disease situation.
+zlib-decomp-$(CONFIG_KERNEL_GZIP) := decompress_inflate.c
+zlib-$(CONFIG_KERNEL_GZIP) := inffast.c inflate.c inftrees.c
+zlibheader-$(CONFIG_KERNEL_GZIP) := inffast.h inffixed.h inflate.h inftrees.h infutil.h
+zliblinuxheader-$(CONFIG_KERNEL_GZIP) := zlib.h zconf.h zutil.h
 
-zlib       := inffast.c inflate.c inftrees.c
-zlibheader := inffast.h inffixed.h inflate.h inftrees.h infutil.h
-zliblinuxheader := zlib.h zconf.h zutil.h
+$(addprefix $(obj)/, decompress.o): \
+	$(addprefix $(obj)/,$(zlib-decomp-y))
 
-$(addprefix $(obj)/,$(zlib) cuboot-c2k.o gunzip_util.o main.o): \
-	$(addprefix $(obj)/,$(zliblinuxheader)) $(addprefix $(obj)/,$(zlibheader))
+$(addprefix $(obj)/, $(zlib-decomp-y)): \
+	$(addprefix $(obj)/,$(zliblinuxheader-y)) \
+	$(addprefix $(obj)/,$(zlibheader-y)) \
+	$(addprefix $(obj)/,$(zlib-y))
+
+$(addprefix $(obj)/,$(zlib-y)): \
+	$(addprefix $(obj)/,$(zliblinuxheader-y)) \
+	$(addprefix $(obj)/,$(zlibheader-y))
 
 libfdt       := fdt.c fdt_ro.c fdt_wip.c fdt_sw.c fdt_rw.c fdt_strerror.c
 libfdtheader := fdt.h libfdt.h libfdt_internal.h
@@ -73,10 +95,10 @@ libfdtheader := fdt.h libfdt.h libfdt_internal.h
 $(addprefix $(obj)/,$(libfdt) libfdt-wrapper.o simpleboot.o epapr.o opal.o): \
 	$(addprefix $(obj)/,$(libfdtheader))
 
-src-wlib-y := string.S crt0.S crtsavres.S stdio.c main.c \
+src-wlib-y := string.S crt0.S crtsavres.S stdio.c decompress.c main.c \
 		$(libfdt) libfdt-wrapper.c \
 		ns16550.c serial.c simple_alloc.c div64.S util.S \
-		gunzip_util.c elf_util.c $(zlib) devtree.c stdlib.c \
+		elf_util.c $(zlib-y) devtree.c stdlib.c \
 		oflib.c ofconsole.c cuboot.c mpsc.c cpm-serial.c \
 		uartlite.c mpc52xx-psc.c opal.c opal-calls.S
 src-wlib-$(CONFIG_40x) += 4xx.c planetcore.c
@@ -125,23 +147,20 @@ obj-wlib := $(addsuffix .o, $(basename $(addprefix $(obj)/, $(src-wlib))))
 obj-plat := $(addsuffix .o, $(basename $(addprefix $(obj)/, $(src-plat))))
 obj-plat: $(libfdt)
 
-quiet_cmd_copy_zlib = COPY    $@
-      cmd_copy_zlib = sed "s@__used@@;s@<linux/\([^>]*\).*@\"\1\"@" $< > $@
+quiet_cmd_copy_kern_src = COPY    $@
+      cmd_copy_kern_src = sed -f $(srctree)/arch/powerpc/boot/fixup-headers.sed $< > $@
 
-quiet_cmd_copy_zlibheader = COPY    $@
-      cmd_copy_zlibheader = sed "s@<linux/\([^>]*\).*@\"\1\"@" $< > $@
-# stddef.h for NULL
-quiet_cmd_copy_zliblinuxheader = COPY    $@
-      cmd_copy_zliblinuxheader = sed "s@<linux/string.h>@\"string.h\"@;s@<linux/kernel.h>@<stddef.h>@;s@<linux/\([^>]*\).*@\"\1\"@" $< > $@
+$(addprefix $(obj)/,$(zlib-y)): $(obj)/%: $(srctree)/lib/zlib_inflate/%
+	$(call cmd,copy_kern_src)
 
-$(addprefix $(obj)/,$(zlib)): $(obj)/%: $(srctree)/lib/zlib_inflate/%
-	$(call cmd,copy_zlib)
+$(addprefix $(obj)/,$(zlibheader-y)): $(obj)/%: $(srctree)/lib/zlib_inflate/%
+	$(call cmd,copy_kern_src)
 
-$(addprefix $(obj)/,$(zlibheader)): $(obj)/%: $(srctree)/lib/zlib_inflate/%
-	$(call cmd,copy_zlibheader)
+$(addprefix $(obj)/,$(zliblinuxheader-y)): $(obj)/%: $(srctree)/include/linux/%
+	$(call cmd,copy_kern_src)
 
-$(addprefix $(obj)/,$(zliblinuxheader)): $(obj)/%: $(srctree)/include/linux/%
-	$(call cmd,copy_zliblinuxheader)
+$(addprefix $(obj)/,$(zlib-decomp-y)): $(obj)/%: $(srctree)/lib/%
+	$(call cmd,copy_kern_src)
 
 quiet_cmd_copy_libfdt = COPY    $@
       cmd_copy_libfdt = cp $< $@
@@ -150,17 +169,17 @@ $(addprefix $(obj)/,$(libfdt) $(libfdtheader)): $(obj)/%: $(srctree)/scripts/dtc
 	$(call cmd,copy_libfdt)
 
 $(obj)/empty.c:
-	@touch $@
+	$(Q)touch $@
 
 $(obj)/zImage.lds: $(obj)/%: $(srctree)/$(src)/%.S
 	$(CROSS32CC) $(cpp_flags) -E -Wp,-MD,$(depfile) -P -Upowerpc \
 		-D__ASSEMBLY__ -DLINKER_SCRIPT -o $@ $<
 
 $(obj)/zImage.coff.lds $(obj)/zImage.ps3.lds : $(obj)/%: $(srctree)/$(src)/%.S
-	@cp $< $@
+	$(Q)cp $< $@
 
-clean-files := $(zlib) $(zlibheader) $(zliblinuxheader) \
-		$(libfdt) $(libfdtheader) \
+clean-files := $(zlib-) $(zlibheader-) $(zliblinuxheader-) \
+		$(zlib-decomp-) $(libfdt) $(libfdtheader) \
 		empty.c zImage.coff.lds zImage.ps3.lds zImage.lds
 
 quiet_cmd_bootcc = BOOTCC  $@
@@ -207,10 +226,14 @@ CROSSWRAP := -C "$(CROSS_COMPILE)"
 endif
 endif
 
+compressor-$(CONFIG_KERNEL_GZIP) := gz
+compressor-$(CONFIG_KERNEL_XZ)   := xz
+
 # args (to if_changed): 1 = (this rule), 2 = platform, 3 = dts 4=dtb 5=initrd
 quiet_cmd_wrap	= WRAP    $@
-      cmd_wrap	=$(CONFIG_SHELL) $(wrapper) -c -o $@ -p $2 $(CROSSWRAP) \
-		$(if $3, -s $3)$(if $4, -d $4)$(if $5, -i $5) vmlinux
+      cmd_wrap	=$(CONFIG_SHELL) $(wrapper) -Z $(compressor-y) -c -o $@ -p $2 \
+		$(CROSSWRAP) $(if $3, -s $3)$(if $4, -d $4)$(if $5, -i $5) \
+		vmlinux
 
 image-$(CONFIG_PPC_PSERIES)		+= zImage.pseries
 image-$(CONFIG_PPC_POWERNV)		+= zImage.pseries
@@ -391,9 +414,9 @@ image-y := vmlinux.strip
 endif
 
 $(obj)/zImage:		$(addprefix $(obj)/, $(image-y))
-	@rm -f $@; ln $< $@
+	$(Q)rm -f $@; ln $< $@
 $(obj)/zImage.initrd:	$(addprefix $(obj)/, $(initrd-y))
-	@rm -f $@; ln $< $@
+	$(Q)rm -f $@; ln $< $@
 
 # Only install the vmlinux
 install: $(CONFIGURE) $(addprefix $(obj)/, $(image-y))
@@ -410,8 +433,9 @@ clean-files += $(image-) $(initrd-) cuImage.* dtbImage.* treeImage.* \
 	zImage.maple simpleImage.* otheros.bld *.dtb
 
 # clean up files cached by wrapper
-clean-kernel := vmlinux.strip vmlinux.bin
-clean-kernel += $(addsuffix .gz,$(clean-kernel))
+clean-kernel-base := vmlinux.strip vmlinux.bin
+clean-kernel := $(addsuffix .gz,$(clean-kernel-base))
+clean-kernel += $(addsuffix .xz,$(clean-kernel-base))
 # If not absolute clean-files are relative to $(obj).
 clean-files += $(addprefix $(objtree)/, $(clean-kernel))
 
diff --git a/arch/powerpc/boot/cuboot-c2k.c b/arch/powerpc/boot/cuboot-c2k.c
index e43594950ba3..9309c51f1d65 100644
--- a/arch/powerpc/boot/cuboot-c2k.c
+++ b/arch/powerpc/boot/cuboot-c2k.c
@@ -18,7 +18,6 @@
 #include "io.h"
 #include "ops.h"
 #include "elf.h"
-#include "gunzip_util.h"
 #include "mv64x60.h"
 #include "cuboot.h"
 #include "ppcboot.h"
diff --git a/arch/powerpc/boot/decompress.c b/arch/powerpc/boot/decompress.c
new file mode 100644
index 000000000000..3aff4423ad01
--- /dev/null
+++ b/arch/powerpc/boot/decompress.c
@@ -0,0 +1,148 @@
+/*
+ * Wrapper around the kernel's pre-boot decompression library.
+ *
+ * Copyright (C) IBM Corporation 2016.
+ *
+ * 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 "elf.h"
+#include "page.h"
+#include "string.h"
+#include "stdio.h"
+#include "ops.h"
+#include "reg.h"
+#include "types.h"
+
+/*
+ * The decompressor_*.c files play #ifdef games so they can be used in both
+ * pre-boot and regular kernel code. We need these definitions to make the
+ * includes work.
+ */
+
+#define STATIC static
+#define INIT
+#define __always_inline inline
+
+/*
+ * The build process will copy the required zlib source files and headers
+ * out of lib/ and "fix" the includes so they do not pull in other kernel
+ * headers.
+ */
+
+#ifdef CONFIG_KERNEL_GZIP
+#	include "decompress_inflate.c"
+#endif
+
+#ifdef CONFIG_KERNEL_XZ
+#	include "xz_config.h"
+#	include "../../../lib/decompress_unxz.c"
+#endif
+
+/* globals for tracking the state of the decompression */
+static unsigned long decompressed_bytes;
+static unsigned long limit;
+static unsigned long skip;
+static char *output_buffer;
+
+/*
+ * flush() is called by __decompress() when the decompressor's scratch buffer is
+ * full.
+ */
+static long flush(void *v, unsigned long buffer_size)
+{
+	unsigned long end = decompressed_bytes + buffer_size;
+	unsigned long size = buffer_size;
+	unsigned long offset = 0;
+	char *in = v;
+	char *out;
+
+	/*
+	 * if we hit our decompression limit, we need to fake an error to abort
+	 * the in-progress decompression.
+	 */
+	if (decompressed_bytes >= limit)
+		return -1;
+
+	/* skip this entire block */
+	if (end <= skip) {
+		decompressed_bytes += buffer_size;
+		return buffer_size;
+	}
+
+	/* skip some data at the start, but keep the rest of the block */
+	if (decompressed_bytes < skip && end > skip) {
+		offset = skip - decompressed_bytes;
+
+		in += offset;
+		size -= offset;
+		decompressed_bytes += offset;
+	}
+
+	out = &output_buffer[decompressed_bytes - skip];
+	size = min(decompressed_bytes + size, limit) - decompressed_bytes;
+
+	memcpy(out, in, size);
+	decompressed_bytes += size;
+
+	return buffer_size;
+}
+
+static void print_err(char *s)
+{
+	/* suppress the "error" when we terminate the decompressor */
+	if (decompressed_bytes >= limit)
+		return;
+
+	printf("Decompression error: '%s'\n\r", s);
+}
+
+/**
+ * partial_decompress - decompresses part or all of a compressed buffer
+ * @inbuf:       input buffer
+ * @input_size:  length of the input buffer
+ * @outbuf:      input buffer
+ * @output_size: length of the input buffer
+ * @skip         number of output bytes to ignore
+ *
+ * This function takes compressed data from inbuf, decompresses and write it to
+ * outbuf. Once output_size bytes are written to the output buffer, or the
+ * stream is exhausted the function will return the number of bytes that were
+ * decompressed. Otherwise it will return whatever error code the decompressor
+ * reported (NB: This is specific to each decompressor type).
+ *
+ * The skip functionality is mainly there so the program and discover
+ * the size of the compressed image so that it can ask firmware (if present)
+ * for an appropriately sized buffer.
+ */
+long partial_decompress(void *inbuf, unsigned long input_size,
+	void *outbuf, unsigned long output_size, unsigned long _skip)
+{
+	int ret;
+
+	/*
+	 * The skipped bytes needs to be included in the size of data we want
+	 * to decompress.
+	 */
+	output_size += _skip;
+
+	decompressed_bytes = 0;
+	output_buffer = outbuf;
+	limit = output_size;
+	skip = _skip;
+
+	ret = __decompress(inbuf, input_size, NULL, flush, outbuf,
+		output_size, NULL, print_err);
+
+	/*
+	 * If decompression was aborted due to an actual error rather than
+	 * a fake error that we used to abort, then we should report it.
+	 */
+	if (decompressed_bytes < limit)
+		return ret;
+
+	return decompressed_bytes - skip;
+}
diff --git a/arch/powerpc/boot/fixup-headers.sed b/arch/powerpc/boot/fixup-headers.sed
new file mode 100644
index 000000000000..96362428eb37
--- /dev/null
+++ b/arch/powerpc/boot/fixup-headers.sed
@@ -0,0 +1,12 @@
+# Copyright 2016 IBM Corporation.
+#
+# This program is free software; you can redistribute it and/or modify it
+# under the terms of the GNU General Public License version 2 or later as
+# published by the Free Software Foundation.
+
+s@#include <linux/decompress/mm\.h>@@;
+s@\"zlib_inflate/\([^\"]*\).*@"\1"@;
+s@<linux/kernel.h>@<stddef.h>@;
+
+s@__used@@;
+s@<linux/\([^>]*\).*@"\1"@;
diff --git a/arch/powerpc/boot/gunzip_util.c b/arch/powerpc/boot/gunzip_util.c
deleted file mode 100644
index 9dc52501de83..000000000000
--- a/arch/powerpc/boot/gunzip_util.c
+++ /dev/null
@@ -1,204 +0,0 @@
-/*
- * Copyright 2007 David Gibson, IBM Corporation.
- * Based on earlier work, Copyright (C) Paul Mackerras 1997.
- *
- * 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 <stddef.h>
-#include "string.h"
-#include "stdio.h"
-#include "ops.h"
-#include "gunzip_util.h"
-
-#define HEAD_CRC	2
-#define EXTRA_FIELD	4
-#define ORIG_NAME	8
-#define COMMENT		0x10
-#define RESERVED	0xe0
-
-/**
- * gunzip_start - prepare to decompress gzip data
- * @state:     decompressor state structure to be initialized
- * @src:       buffer containing gzip compressed or uncompressed data
- * @srclen:    size in bytes of the buffer at src
- *
- * If the buffer at @src contains a gzip header, this function
- * initializes zlib to decompress the data, storing the decompression
- * state in @state.  The other functions in this file can then be used
- * to decompress data from the gzipped stream.
- *
- * If the buffer at @src does not contain a gzip header, it is assumed
- * to contain uncompressed data.  The buffer information is recorded
- * in @state and the other functions in this file will simply copy
- * data from the uncompressed data stream at @src.
- *
- * Any errors, such as bad compressed data, cause an error to be
- * printed an the platform's exit() function to be called.
- */
-void gunzip_start(struct gunzip_state *state, void *src, int srclen)
-{
-	char *hdr = src;
-	int hdrlen = 0;
-
-	memset(state, 0, sizeof(*state));
-
-	/* Check for gzip magic number */
-	if ((hdr[0] == 0x1f) && (hdr[1] == 0x8b)) {
-		/* gzip data, initialize zlib parameters */
-		int r, flags;
-
-		state->s.workspace = state->scratch;
-		if (zlib_inflate_workspacesize() > sizeof(state->scratch))
-			fatal("insufficient scratch space for gunzip\n\r");
-
-		/* skip header */
-		hdrlen = 10;
-		flags = hdr[3];
-		if (hdr[2] != Z_DEFLATED || (flags & RESERVED) != 0)
-			fatal("bad gzipped data\n\r");
-		if ((flags & EXTRA_FIELD) != 0)
-			hdrlen = 12 + hdr[10] + (hdr[11] << 8);
-		if ((flags & ORIG_NAME) != 0)
-			while (hdr[hdrlen++] != 0)
-				;
-		if ((flags & COMMENT) != 0)
-			while (hdr[hdrlen++] != 0)
-				;
-		if ((flags & HEAD_CRC) != 0)
-			hdrlen += 2;
-		if (hdrlen >= srclen)
-			fatal("gunzip_start: ran out of data in header\n\r");
-
-		r = zlib_inflateInit2(&state->s, -MAX_WBITS);
-		if (r != Z_OK)
-			fatal("inflateInit2 returned %d\n\r", r);
-	}
-
-	state->s.total_in = hdrlen;
-	state->s.next_in = src + hdrlen;
-	state->s.avail_in = srclen - hdrlen;
-}
-
-/**
- * gunzip_partial - extract bytes from a gzip data stream
- * @state:     gzip state structure previously initialized by gunzip_start()
- * @dst:       buffer to store extracted data
- * @dstlen:    maximum number of bytes to extract
- *
- * This function extracts at most @dstlen bytes from the data stream
- * previously associated with @state by gunzip_start(), decompressing
- * if necessary.  Exactly @dstlen bytes are extracted unless the data
- * stream doesn't contain enough bytes, in which case the entire
- * remainder of the stream is decompressed.
- *
- * Returns the actual number of bytes extracted.  If any errors occur,
- * such as a corrupted compressed stream, an error is printed an the
- * platform's exit() function is called.
- */
-int gunzip_partial(struct gunzip_state *state, void *dst, int dstlen)
-{
-	int len;
-
-	if (state->s.workspace) {
-		/* gunzipping */
-		int r;
-
-		state->s.next_out = dst;
-		state->s.avail_out = dstlen;
-		r = zlib_inflate(&state->s, Z_FULL_FLUSH);
-		if (r != Z_OK && r != Z_STREAM_END)
-			fatal("inflate returned %d msg: %s\n\r", r, state->s.msg);
-		len = state->s.next_out - (Byte *)dst;
-	} else {
-		/* uncompressed image */
-		len = min(state->s.avail_in, (uLong)dstlen);
-		memcpy(dst, state->s.next_in, len);
-		state->s.next_in += len;
-		state->s.avail_in -= len;
-	}
-	return len;
-}
-
-/**
- * gunzip_exactly - extract a fixed number of bytes from a gzip data stream
- * @state:     gzip state structure previously initialized by gunzip_start()
- * @dst:       buffer to store extracted data
- * @dstlen:    number of bytes to extract
- *
- * This function extracts exactly @dstlen bytes from the data stream
- * previously associated with @state by gunzip_start(), decompressing
- * if necessary.
- *
- * If there are less @dstlen bytes available in the data stream, or if
- * any other errors occur, such as a corrupted compressed stream, an
- * error is printed an the platform's exit() function is called.
- */
-void gunzip_exactly(struct gunzip_state *state, void *dst, int dstlen)
-{
-	int len;
-
-	len  = gunzip_partial(state, dst, dstlen);
-	if (len < dstlen)
-		fatal("\n\rgunzip_exactly: ran out of data!"
-				" Wanted %d, got %d.\n\r", dstlen, len);
-}
-
-/**
- * gunzip_discard - discard bytes from a gzip data stream
- * @state:     gzip state structure previously initialized by gunzip_start()
- * @len:       number of bytes to discard
- *
- * This function extracts, then discards exactly @len bytes from the
- * data stream previously associated with @state by gunzip_start().
- * Subsequent gunzip_partial(), gunzip_exactly() or gunzip_finish()
- * calls will extract the data following the discarded bytes in the
- * data stream.
- *
- * If there are less @len bytes available in the data stream, or if
- * any other errors occur, such as a corrupted compressed stream, an
- * error is printed an the platform's exit() function is called.
- */
-void gunzip_discard(struct gunzip_state *state, int len)
-{
-	static char discard_buf[128];
-
-	while (len > sizeof(discard_buf)) {
-		gunzip_exactly(state, discard_buf, sizeof(discard_buf));
-		len -= sizeof(discard_buf);
-	}
-
-	if (len > 0)
-		gunzip_exactly(state, discard_buf, len);
-}
-
-/**
- * gunzip_finish - extract all remaining bytes from a gzip data stream
- * @state:     gzip state structure previously initialized by gunzip_start()
- * @dst:       buffer to store extracted data
- * @dstlen:    maximum number of bytes to extract
- *
- * This function extracts all remaining data, or at most @dstlen
- * bytes, from the stream previously associated with @state by
- * gunzip_start().  zlib is then shut down, so it is an error to use
- * any of the functions in this file on @state until it is
- * re-initialized with another call to gunzip_start().
- *
- * If any errors occur, such as a corrupted compressed stream, an
- * error is printed an the platform's exit() function is called.
- */
-int gunzip_finish(struct gunzip_state *state, void *dst, int dstlen)
-{
-	int len;
-
-	len = gunzip_partial(state, dst, dstlen);
-
-	if (state->s.workspace) {
-		zlib_inflateEnd(&state->s);
-	}
-
-	return len;
-}
diff --git a/arch/powerpc/boot/gunzip_util.h b/arch/powerpc/boot/gunzip_util.h
deleted file mode 100644
index b3dfa6e87b3a..000000000000
--- a/arch/powerpc/boot/gunzip_util.h
+++ /dev/null
@@ -1,45 +0,0 @@
-/*
- * Decompression convenience functions
- *
- * Copyright 2007 David Gibson, IBM Corporation.
- *
- * This file is licensed under the terms of the GNU General Public
- * License version 2.  This program is licensed "as is" without any
- * warranty of any kind, whether express or implied.
- */
-#ifndef _PPC_BOOT_GUNZIP_UTIL_H_
-#define _PPC_BOOT_GUNZIP_UTIL_H_
-
-#include "zlib.h"
-
-/*
- * These functions are designed to make life easy for decompressing
- * kernel images, initrd images or any other gzip compressed image,
- * particularly if its useful to decompress part of the image (e.g. to
- * examine headers) before decompressing the remainder.
- *
- * To use:
- *     - declare a gunzip_state structure
- *     - use gunzip_start() to initialize the state, associating it
- *       with a stream of compressed data
- *     - use gunzip_partial(), gunzip_exactly() and gunzip_discard()
- *       in any combination to extract pieces of data from the stream
- *     - Finally use gunzip_finish() to extract the tail of the
- *       compressed stream and wind up zlib
- */
-
-/* scratch space for gunzip; 46912 is from zlib_inflate_workspacesize() */
-#define GUNZIP_SCRATCH_SIZE	46912
-
-struct gunzip_state {
-	z_stream s;
-	char scratch[46912];
-};
-
-void gunzip_start(struct gunzip_state *state, void *src, int srclen);
-int gunzip_partial(struct gunzip_state *state, void *dst, int dstlen);
-void gunzip_exactly(struct gunzip_state *state, void *dst, int len);
-void gunzip_discard(struct gunzip_state *state, int len);
-int gunzip_finish(struct gunzip_state *state, void *dst, int len);
-
-#endif /* _PPC_BOOT_GUNZIP_UTIL_H_ */
diff --git a/arch/powerpc/boot/main.c b/arch/powerpc/boot/main.c
index d80161b633f4..f7a184b6c35b 100644
--- a/arch/powerpc/boot/main.c
+++ b/arch/powerpc/boot/main.c
@@ -15,11 +15,8 @@
 #include "string.h"
 #include "stdio.h"
 #include "ops.h"
-#include "gunzip_util.h"
 #include "reg.h"
 
-static struct gunzip_state gzstate;
-
 struct addr_range {
 	void *addr;
 	unsigned long size;
@@ -30,15 +27,14 @@ struct addr_range {
 static struct addr_range prep_kernel(void)
 {
 	char elfheader[256];
-	void *vmlinuz_addr = _vmlinux_start;
+	unsigned char *vmlinuz_addr = (unsigned char *)_vmlinux_start;
 	unsigned long vmlinuz_size = _vmlinux_end - _vmlinux_start;
 	void *addr = 0;
 	struct elf_info ei;
-	int len;
+	long len;
 
-	/* gunzip the ELF header of the kernel */
-	gunzip_start(&gzstate, vmlinuz_addr, vmlinuz_size);
-	gunzip_exactly(&gzstate, elfheader, sizeof(elfheader));
+	partial_decompress(vmlinuz_addr, vmlinuz_size,
+		elfheader, sizeof(elfheader), 0);
 
 	if (!parse_elf64(elfheader, &ei) && !parse_elf32(elfheader, &ei))
 		fatal("Error: not a valid PPC32 or PPC64 ELF file!\n\r");
@@ -51,7 +47,7 @@ static struct addr_range prep_kernel(void)
 	 * the kernel bss must be claimed (it will be zero'd by the
 	 * kernel itself)
 	 */
-	printf("Allocating 0x%lx bytes for kernel ...\n\r", ei.memsize);
+	printf("Allocating 0x%lx bytes for kernel...\n\r", ei.memsize);
 
 	if (platform_ops.vmlinux_alloc) {
 		addr = platform_ops.vmlinux_alloc(ei.memsize);
@@ -71,16 +67,21 @@ static struct addr_range prep_kernel(void)
 					"device tree\n\r");
 	}
 
-	/* Finally, gunzip the kernel */
-	printf("gunzipping (0x%p <- 0x%p:0x%p)...", addr,
+	/* Finally, decompress the kernel */
+	printf("Decompressing (0x%p <- 0x%p:0x%p)...\n\r", addr,
 	       vmlinuz_addr, vmlinuz_addr+vmlinuz_size);
-	/* discard up to the actual load data */
-	gunzip_discard(&gzstate, ei.elfoffset - sizeof(elfheader));
-	len = gunzip_finish(&gzstate, addr, ei.loadsize);
+
+	len = partial_decompress(vmlinuz_addr, vmlinuz_size,
+		addr, ei.loadsize, ei.elfoffset);
+
+	if (len < 0)
+		fatal("Decompression failed with error code %ld\n\r", len);
+
 	if (len != ei.loadsize)
-		fatal("ran out of data!  only got 0x%x of 0x%lx bytes.\n\r",
-				len, ei.loadsize);
-	printf("done 0x%x bytes\n\r", len);
+		 fatal("Decompression error: got 0x%lx bytes, expected 0x%lx.\n\r",
+			 len, ei.loadsize);
+
+	printf("Done! Decompressed 0x%lx bytes\n\r", len);
 
 	flush_cache(addr, ei.loadsize);
 
diff --git a/arch/powerpc/boot/ops.h b/arch/powerpc/boot/ops.h
index e19b64ef977a..309d1b127e96 100644
--- a/arch/powerpc/boot/ops.h
+++ b/arch/powerpc/boot/ops.h
@@ -260,4 +260,7 @@ int __ilog2_u32(u32 n)
 	return 31 - bit;
 }
 
+long partial_decompress(void *inbuf, unsigned long input_size, void *outbuf,
+	unsigned long output_size, unsigned long skip);
+
 #endif /* _PPC_BOOT_OPS_H_ */
diff --git a/arch/powerpc/boot/stdbool.h b/arch/powerpc/boot/stdbool.h
new file mode 100644
index 000000000000..f818efb08891
--- /dev/null
+++ b/arch/powerpc/boot/stdbool.h
@@ -0,0 +1,14 @@
+/*
+ * Copyright (C) IBM Corporation 2016.
+ *
+ * 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.
+ *
+ * This file is only necessary because some of the pre-boot decompressors
+ * expect stdbool.h to be available.
+ *
+ */
+
+#include "types.h"
diff --git a/arch/powerpc/boot/stdint.h b/arch/powerpc/boot/stdint.h
new file mode 100644
index 000000000000..c1c853be7490
--- /dev/null
+++ b/arch/powerpc/boot/stdint.h
@@ -0,0 +1,13 @@
+/*
+ * Copyright (C) IBM Corporation 2016.
+ *
+ * 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.
+ *
+ * This file is only necessary because some of the pre-boot decompressors
+ * expect stdint.h to be available.
+ */
+
+#include "types.h"
diff --git a/arch/powerpc/boot/types.h b/arch/powerpc/boot/types.h
index 85565a89bcc2..af6b66b842c4 100644
--- a/arch/powerpc/boot/types.h
+++ b/arch/powerpc/boot/types.h
@@ -1,6 +1,8 @@
 #ifndef _TYPES_H_
 #define _TYPES_H_
 
+#include <stdbool.h>
+
 #define ARRAY_SIZE(x) (sizeof(x) / sizeof((x)[0]))
 
 typedef unsigned char		u8;
@@ -34,4 +36,16 @@ typedef s64 int64_t;
 	(void) (&_x == &_y);	\
 	_x > _y ? _x : _y; })
 
+#define min_t(type, a, b) min(((type) a), ((type) b))
+#define max_t(type, a, b) max(((type) a), ((type) b))
+
+typedef int bool;
+
+#ifndef true
+#define true 1
+#endif
+
+#ifndef false
+#define false 0
+#endif
 #endif /* _TYPES_H_ */
diff --git a/arch/powerpc/boot/wrapper b/arch/powerpc/boot/wrapper
index 6681ec3625c9..404b3aabdb4d 100755
--- a/arch/powerpc/boot/wrapper
+++ b/arch/powerpc/boot/wrapper
@@ -20,6 +20,8 @@
 # -D dir	specify directory containing data files used by script
 #		(default ./arch/powerpc/boot)
 # -W dir	specify working directory for temporary files (default .)
+# -z		use gzip (legacy)
+# -Z zsuffix    compression to use (gz, xz or none)
 
 # Stop execution if any command fails
 set -e
@@ -38,7 +40,7 @@ dtb=
 dts=
 cacheit=
 binary=
-gzip=.gz
+compression=.gz
 pie=
 format=
 
@@ -59,7 +61,8 @@ tmpdir=.
 usage() {
     echo 'Usage: wrapper [-o output] [-p platform] [-i initrd]' >&2
     echo '       [-d devtree] [-s tree.dts] [-c] [-C cross-prefix]' >&2
-    echo '       [-D datadir] [-W workingdir] [--no-gzip] [vmlinux]' >&2
+    echo '       [-D datadir] [-W workingdir] [-Z (gz|xz|none)]' >&2
+    echo '       [--no-compression] [vmlinux]' >&2
     exit 1
 }
 
@@ -126,8 +129,24 @@ while [ "$#" -gt 0 ]; do
 	[ "$#" -gt 0 ] || usage
 	tmpdir="$1"
 	;;
+    -z)
+	compression=.gz
+	;;
+    -Z)
+	shift
+	[ "$#" -gt 0 ] || usage
+        [ "$1" != "gz" -o "$1" != "xz" -o "$1" != "none" ] || usage
+
+	compression=".$1"
+
+        if [ $compression = ".none" ]; then
+                compression=
+        fi
+	;;
     --no-gzip)
-        gzip=
+        # a "feature" of the the wrapper script is that it can be used outside
+        # the kernel tree. So keeping this around for backwards compatibility.
+        compression=
         ;;
     -?)
 	usage
@@ -140,6 +159,7 @@ while [ "$#" -gt 0 ]; do
     shift
 done
 
+
 if [ -n "$dts" ]; then
     if [ ! -r "$dts" -a -r "$object/dts/$dts" ]; then
 	dts="$object/dts/$dts"
@@ -212,7 +232,7 @@ miboot|uboot*)
     ;;
 cuboot*)
     binary=y
-    gzip=
+    compression=
     case "$platform" in
     *-mpc866ads|*-mpc885ads|*-adder875*|*-ep88xc)
         platformo=$object/cuboot-8xx.o
@@ -243,7 +263,7 @@ cuboot*)
 ps3)
     platformo="$object/ps3-head.o $object/ps3-hvcall.o $object/ps3.o"
     lds=$object/zImage.ps3.lds
-    gzip=
+    compression=
     ext=bin
     objflags="-O binary --set-section-flags=.bss=contents,alloc,load,data"
     ksection=.kernel:vmlinux.bin
@@ -310,27 +330,37 @@ mvme7100)
 esac
 
 vmz="$tmpdir/`basename \"$kernel\"`.$ext"
-if [ -z "$cacheit" -o ! -f "$vmz$gzip" -o "$vmz$gzip" -ot "$kernel" ]; then
-    ${CROSS}objcopy $objflags "$kernel" "$vmz.$$"
 
-    strip_size=$(stat -c %s $vmz.$$)
+# Calculate the vmlinux.strip size
+${CROSS}objcopy $objflags "$kernel" "$vmz.$$"
+strip_size=$(stat -c %s $vmz.$$)
 
-    if [ -n "$gzip" ]; then
+if [ -z "$cacheit" -o ! -f "$vmz$compression" -o "$vmz$compression" -ot "$kernel" ]; then
+    # recompress the image if we need to
+    case $compression in
+    .xz)
+        xz --check=crc32 -f -6 "$vmz.$$"
+        ;;
+    .gz)
         gzip -n -f -9 "$vmz.$$"
-    fi
+        ;;
+    *)
+        # drop the compression suffix so the stripped vmlinux is used
+        compression=
+	;;
+    esac
 
     if [ -n "$cacheit" ]; then
-	mv -f "$vmz.$$$gzip" "$vmz$gzip"
+	mv -f "$vmz.$$$compression" "$vmz$compression"
     else
 	vmz="$vmz.$$"
     fi
 else
-    # Calculate the vmlinux.strip size
-    ${CROSS}objcopy $objflags "$kernel" "$vmz.$$"
-    strip_size=$(stat -c %s $vmz.$$)
     rm -f $vmz.$$
 fi
 
+vmz="$vmz$compression"
+
 if [ "$make_space" = "y" ]; then
 	# Round the size to next higher MB limit
 	round_size=$(((strip_size + 0xfffff) & 0xfff00000))
@@ -346,8 +376,6 @@ if [ "$make_space" = "y" ]; then
 	fi
 fi
 
-vmz="$vmz$gzip"
-
 # Extract kernel version information, some platforms want to include
 # it in the image header
 version=`${CROSS}strings "$kernel" | grep '^Linux version [-0-9.]' | \
@@ -417,6 +445,7 @@ if [ "$platform" != "miboot" ]; then
     if [ -n "$link_address" ] ; then
         text_start="-Ttext $link_address"
     fi
+#link everything
     ${CROSS}ld -m $format -T $lds $text_start $pie -o "$ofile" \
 	$platformo $tmp $object/wrapper.a
     rm $tmp
diff --git a/arch/powerpc/boot/xz_config.h b/arch/powerpc/boot/xz_config.h
new file mode 100644
index 000000000000..5c6afdbca642
--- /dev/null
+++ b/arch/powerpc/boot/xz_config.h
@@ -0,0 +1,39 @@
+#ifndef __XZ_CONFIG_H__
+#define __XZ_CONFIG_H__
+
+/*
+ * most of this is copied from lib/xz/xz_private.h, we can't use their defines
+ * since the boot wrapper is not built in the same environment as the rest of
+ * the kernel.
+ */
+
+#include "types.h"
+#include "swab.h"
+
+static inline uint32_t swab32p(void *p)
+{
+	uint32_t *q = p;
+
+	return swab32(*q);
+}
+
+#ifdef __LITTLE_ENDIAN__
+#define get_le32(p) (*((uint32_t *) (p)))
+#else
+#define get_le32(p) swab32p(p)
+#endif
+
+#define memeq(a, b, size) (memcmp(a, b, size) == 0)
+#define memzero(buf, size) memset(buf, 0, size)
+
+/* prevent the inclusion of the xz-preboot MM headers */
+#define DECOMPR_MM_H
+#define memmove memmove
+#define XZ_EXTERN static
+
+/* xz.h needs to be included directly since we need enum xz_mode */
+#include "../../../include/linux/xz.h"
+
+#undef XZ_EXTERN
+
+#endif