Merge branch 'fixes' of git://git.linaro.org/people/rmk/linux-arm

Pull ARM fixes from Russell King:
 "The biggest two fixes are fixing a compilation error with the
  decompressor, and a problem with our __my_cpu_offset implementation.

  Other changes are very trivial and small, which seems to be the way
  for most -rc stuff."

* 'fixes' of git://git.linaro.org/people/rmk/linux-arm:
  ARM: 7747/1: pcpu: ensure __my_cpu_offset cannot be re-ordered across barrier()
  ARM: 7750/1: update legacy CPU ID in decompressor cache support jump table
  ARM: 7743/1: compressed/head.S: work around new binutils warning
  ARM: 7742/1: topology: export cpu_topology
  ARM: 7737/1: fix kernel decompressor compilation error with CONFIG_DEBUG_SEMIHOSTING

7 files changed, 45 insertions, 5 deletions

diff --git a/arch/arm/boot/compressed/Makefile b/arch/arm/boot/compressed/Makefile
index 3580d57ea218..79e9bdbfc491 100644
--- a/arch/arm/boot/compressed/Makefile
+++ b/arch/arm/boot/compressed/Makefile
@@ -124,7 +124,7 @@ KBUILD_CFLAGS = $(subst -pg, , $(ORIG_CFLAGS))
 endif
 
 ccflags-y := -fpic -mno-single-pic-base -fno-builtin -I$(obj)
-asflags-y := -Wa,-march=all -DZIMAGE
+asflags-y := -DZIMAGE
 
 # Supply kernel BSS size to the decompressor via a linker symbol.
 KBSS_SZ = $(shell $(CROSS_COMPILE)size $(obj)/../../../../vmlinux | \
diff --git a/arch/arm/boot/compressed/debug.S b/arch/arm/boot/compressed/debug.S
index 6e8382d5b7a4..5392ee63338f 100644
--- a/arch/arm/boot/compressed/debug.S
+++ b/arch/arm/boot/compressed/debug.S
@@ -1,6 +1,8 @@
 #include <linux/linkage.h>
 #include <asm/assembler.h>
 
+#ifndef CONFIG_DEBUG_SEMIHOSTING
+
 #include CONFIG_DEBUG_LL_INCLUDE
 
 ENTRY(putc)
@@ -10,3 +12,29 @@ ENTRY(putc)
         busyuart r3, r1
         mov      pc, lr
 ENDPROC(putc)
+
+#else
+
+ENTRY(putc)
+        adr     r1, 1f
+        ldmia   r1, {r2, r3}
+        add     r2, r2, r1
+        ldr     r1, [r2, r3]
+        strb    r0, [r1]
+        mov     r0, #0x03               @ SYS_WRITEC
+   ARM( svc     #0x123456       )
+ THUMB( svc     #0xab           )
+        mov     pc, lr
+        .align  2
+1:      .word   _GLOBAL_OFFSET_TABLE_ - .
+        .word   semi_writec_buf(GOT)
+ENDPROC(putc)
+
+        .bss
+        .global semi_writec_buf
+        .type   semi_writec_buf, %object
+semi_writec_buf:
+        .space  4
+        .size   semi_writec_buf, 4
+
+#endif
diff --git a/arch/arm/boot/compressed/head-sa1100.S b/arch/arm/boot/compressed/head-sa1100.S
index 6179d94dd5c6..3115e313d9f6 100644
--- a/arch/arm/boot/compressed/head-sa1100.S
+++ b/arch/arm/boot/compressed/head-sa1100.S
@@ -11,6 +11,7 @@
 #include <asm/mach-types.h>
 
                 .section        ".start", "ax"
+                .arch   armv4
 
 __SA1100_start:
 
diff --git a/arch/arm/boot/compressed/head-shark.S b/arch/arm/boot/compressed/head-shark.S
index 089c560e07f1..92b56897ed64 100644
--- a/arch/arm/boot/compressed/head-shark.S
+++ b/arch/arm/boot/compressed/head-shark.S
@@ -18,6 +18,7 @@
         
                 .section        ".start", "ax"
 
+                .arch armv4
                 b       __beginning
         
 __ofw_data:     .long   0                               @ the number of memory blocks
diff --git a/arch/arm/boot/compressed/head.S b/arch/arm/boot/compressed/head.S
index fe4d9c3ad761..032a8d987148 100644
--- a/arch/arm/boot/compressed/head.S
+++ b/arch/arm/boot/compressed/head.S
@@ -11,6 +11,7 @@
 #include <linux/linkage.h>
 #include <asm/assembler.h>
 
+        .arch   armv7-a
 /*
  * Debugging stuff
  *
@@ -805,8 +806,8 @@ call_cache_fn:	adr	r12, proc_types
                 .align  2
                 .type   proc_types,#object
 proc_types:
-                .word   0x00000000              @ old ARM ID
-                .word   0x0000f000
+                .word   0x41000000              @ old ARM ID
+                .word   0xff00f000
                 mov     pc, lr
  THUMB(         nop                             )
                 mov     pc, lr
diff --git a/arch/arm/include/asm/percpu.h b/arch/arm/include/asm/percpu.h
index 968c0a14e0a3..209e6504922e 100644
--- a/arch/arm/include/asm/percpu.h
+++ b/arch/arm/include/asm/percpu.h
@@ -30,8 +30,15 @@ static inline void set_my_cpu_offset(unsigned long off)
 static inline unsigned long __my_cpu_offset(void)
 {
         unsigned long off;
-        /* Read TPIDRPRW */
-        asm("mrc p15, 0, %0, c13, c0, 4" : "=r" (off) : : "memory");
+        register unsigned long *sp asm ("sp");
+
+        /*
+         * Read TPIDRPRW.
+         * We want to allow caching the value, so avoid using volatile and
+         * instead use a fake stack read to hazard against barrier().
+         */
+        asm("mrc p15, 0, %0, c13, c0, 4" : "=r" (off) : "Q" (*sp));
+
         return off;
 }
 #define __my_cpu_offset __my_cpu_offset()
diff --git a/arch/arm/kernel/topology.c b/arch/arm/kernel/topology.c
index f10316b4ecdc..c5a59546a256 100644
--- a/arch/arm/kernel/topology.c
+++ b/arch/arm/kernel/topology.c
@@ -13,6 +13,7 @@
 
 #include <linux/cpu.h>
 #include <linux/cpumask.h>
+#include <linux/export.h>
 #include <linux/init.h>
 #include <linux/percpu.h>
 #include <linux/node.h>
@@ -200,6 +201,7 @@ static inline void update_cpu_power(unsigned int cpuid, unsigned int mpidr) {}
  * cpu topology table
  */
 struct cputopo_arm cpu_topology[NR_CPUS];
+EXPORT_SYMBOL_GPL(cpu_topology);
 
 const struct cpumask *cpu_coregroup_mask(int cpu)
 {