[PATCH 00/11] Fix up more asm/system.h fallout

[PATCH 00/11] Fix up more asm/system.h fallout

[Paul Gortmaker]

Like others, I was surprised it didn't get a soak in linux-next
before being merged, but I guess what is done is done, so lets
try and limit the fallout from having any wider of a commit range
than it already does.

It is all system.h stuff except for three things.

 1) an avr32 patch I'd created but the avr32 guys hadn't got
    signed keys yet, so I might as well get it in here along
    with other similar build fail fixes (the atmel one).

 2) A cosmetic fix to the device.h changeset from me earlier.

 3) A trivial warning fix to an arch/frv specific file.

I'm going to queue this up in my feed branch to linux-next
so we can start reducing the number of build failures from what
we had on Friday.

Paul.
---

The following changes since commit dd775ae2549217d3ae09363e3edb305d0fa19928:

  Linux 3.4-rc1 (2012-03-31 16:24:09 -0700)

are available in the git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/paulg/linux.git buildfixes

for you to fetch changes up to af901116760f93297ad53cc3a0cc5e370e9ccb5a:

  asm-generic: add linux/types.h to cmpxchg.h (2012-04-01 15:59:54 -0400)

----------------------------------------------------------------

Paul Gortmaker (11):
  irq_work: fix compile failure on MIPS from system.h split
  ARM: mach-msm: fix compile fail from system.h fallout
  avr32: fix nop compile fails from system.h split up
  avr32: fix build failures from mis-naming of atmel_nand.h
  blackfin: fix cmpxchg build fails from system.h fallout
  parisc: fix missing cmpxchg file error from system.h split
  sh: fix several build failures from system.h fallout
  frv: fix warnings in mb93090-mb00/pci-dma.c about implicit EXPORT_SYMBOL
  firewire: restore the device.h include in linux/firewire.h
  tile: fix multiple build failures from system.h dismantle
  asm-generic: add linux/types.h to cmpxchg.h

 arch/arm/mach-msm/include/mach/uncompress.h |    1 +
 arch/avr32/include/asm/mmu_context.h        |    1 +
 arch/avr32/mach-at32ap/include/mach/board.h |    2 +-
 arch/blackfin/include/asm/cmpxchg.h         |    3 +-
 arch/frv/mb93090-mb00/pci-dma.c             |    1 +
 arch/parisc/include/asm/atomic.h            |  107 +------------------------
 arch/parisc/include/asm/cmpxchg.h           |  116 +++++++++++++++++++++++++++
 arch/sh/kernel/cpu/fpu.c                    |    1 +
 arch/sh/kernel/cpu/sh4/fpu.c                |    1 +
 arch/sh/mm/cache-sh4.c                      |    1 +
 arch/sh/mm/flush-sh4.c                      |    1 +
 arch/tile/include/asm/atomic.h              |   50 +-----------
 arch/tile/include/asm/cmpxchg.h             |   73 +++++++++++++++++
 arch/tile/kernel/process.c                  |    1 +
 arch/tile/kernel/stack.c                    |    1 +
 include/asm-generic/cmpxchg.h               |    1 +
 include/linux/firewire.h                    |    3 +-
 include/linux/platform_data/atmel.h         |    6 +-
 kernel/irq_work.c                           |    1 +
 19 files changed, 210 insertions(+), 161 deletions(-)
 create mode 100644 arch/parisc/include/asm/cmpxchg.h
 create mode 100644 arch/tile/include/asm/cmpxchg.h

-- 
1.7.9.1

[PATCH 01/11] irq_work: fix compile failure on MIPS from system.h split

[Paul Gortmaker]

Builds of the MIPS platform ip32_defconfig fails as of
commit 0195c00244dc2e9f522475868fa278c473ba7339

    "Merge tag 'split-asm_system_h-for-linus-20120328' ..."

because MIPS xchg() macro uses BUILD_BUG_ON and it was moved
in commit b81947c646bfefdf98e2fde5d7d39cbbda8525d4

    "Disintegrate asm/system.h for MIPS"

The root cause is that the system.h split wasn't tested on
a baseline with commit 6c03438edeb5c359af35f060ea016ca65671c269

    "kernel.h: doesn't explicitly use bug.h, so don't include it."

Since this file uses BUG code in several other places besides
the xchg call, simply make the inclusion explicit.

Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
---
 kernel/irq_work.c |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/kernel/irq_work.c b/kernel/irq_work.c
index c3c46c7..0c56d44 100644
--- a/kernel/irq_work.c
+++ b/kernel/irq_work.c
@@ -5,6 +5,7 @@
  * context. The enqueueing is NMI-safe.
  */
 
+#include <linux/bug.h>
 #include <linux/kernel.h>
 #include <linux/export.h>
 #include <linux/irq_work.h>
-- 
1.7.9.1

[PATCH 02/11] ARM: mach-msm: fix compile fail from system.h fallout

[Paul Gortmaker]

To fix:

  In file included from arm/boot/compressed/misc.c:28:0:
  arm/mach-msm/include/mach/uncompress.h: In function 'putc':
  arch/arm/mach-msm/include/mach/uncompress.h:48:3: error: implicit
  declaration of function 'smp_mb' [-Werror=implicit-function-declaration]

The putc does a cpu_relax which for this platform is smp_mb.

Bisect indicates the 1st failing commit as:

    commit 0195c00244dc2e9f522475868fa278c473ba7339
    "Merge tag 'split-asm_system_h-for-linus-20120328' ..."

Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
---
 arch/arm/mach-msm/include/mach/uncompress.h |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/arch/arm/mach-msm/include/mach/uncompress.h b/arch/arm/mach-msm/include/mach/uncompress.h
index 169a840..c14011f 100644
--- a/arch/arm/mach-msm/include/mach/uncompress.h
+++ b/arch/arm/mach-msm/include/mach/uncompress.h
@@ -16,6 +16,7 @@
 #ifndef __ASM_ARCH_MSM_UNCOMPRESS_H
 #define __ASM_ARCH_MSM_UNCOMPRESS_H
 
+#include <asm/barrier.h>
 #include <asm/processor.h>
 #include <mach/msm_iomap.h>
 
-- 
1.7.9.1

[PATCH 03/11] avr32: fix nop compile fails from system.h split up

[Paul Gortmaker]

To fix:

  In file included from kernel/exit.c:61:
  arch/avr32/include/asm/mmu_context.h: In function 'enable_mmu':
  arch/avr32/include/asm/mmu_context.h:135: error: implicit
  declaration of function 'nop'

It needs an include of the new file created in:

commit ae473946586680b01c13975a3b674de23ad7c93e

    "Disintegrate asm/system.h for AVR32"

Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
---
 arch/avr32/include/asm/mmu_context.h |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/arch/avr32/include/asm/mmu_context.h b/arch/avr32/include/asm/mmu_context.h
index 27ff234..f1b2fec 100644
--- a/arch/avr32/include/asm/mmu_context.h
+++ b/arch/avr32/include/asm/mmu_context.h
@@ -14,6 +14,7 @@
 
 #include <asm/tlbflush.h>
 #include <asm/sysreg.h>
+#include <asm/special_insns.h>
 #include <asm-generic/mm_hooks.h>
 
 /*
-- 
1.7.9.1

[PATCH 04/11] avr32: fix build failures from mis-naming of atmel_nand.h

[Paul Gortmaker]

Commit bf4289cba02b8cf770ecd7959ca70839f0dd9d3c

    "ATMEL: fix nand ecc support"

indicated that it wanted to "Move platform data to a common
header include/linux/platform_data/atmel_nand.h" and the new
header even had re-include protectors with:

    #ifndef __ATMEL_NAND_H__

However, the file that was added was simply called atmel.h
and this caused avr32 defconfig to fail with:

  In file included from arch/avr32/boards/atstk1000/setup.c:22:
  arch/avr32/mach-at32ap/include/mach/board.h:10:44: error: linux/platform_data/atmel_nand.h: No such file or directory
  In file included from arch/avr32/boards/atstk1000/setup.c:22:
  arch/avr32/mach-at32ap/include/mach/board.h:121: warning: 'struct atmel_nand_data' declared inside parameter list
  arch/avr32/mach-at32ap/include/mach/board.h:121: warning: its scope is only this definition or declaration, which is probably not what you want
  make[2]: *** [arch/avr32/boards/atstk1000/setup.o] Error 1

It seems the scope of the file contents will expand beyond
just nand, so ignore the original intention, and fix up the
users who reference the bad name with the _nand suffix.

CC: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
CC: David Woodhouse <dwmw2@infradead.org>
Acked-by: Hans-Christian Egtvedt <egtvedt@samfundet.no>
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
---
 arch/avr32/mach-at32ap/include/mach/board.h |    2 +-
 include/linux/platform_data/atmel.h         |    6 +++---
 2 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/arch/avr32/mach-at32ap/include/mach/board.h b/arch/avr32/mach-at32ap/include/mach/board.h
index 7173386..70742ec 100644
--- a/arch/avr32/mach-at32ap/include/mach/board.h
+++ b/arch/avr32/mach-at32ap/include/mach/board.h
@@ -7,7 +7,7 @@
 #include <linux/types.h>
 #include <linux/serial.h>
 #include <linux/platform_data/macb.h>
-#include <linux/platform_data/atmel_nand.h>
+#include <linux/platform_data/atmel.h>
 
 #define GPIO_PIN_NONE	(-1)
 
diff --git a/include/linux/platform_data/atmel.h b/include/linux/platform_data/atmel.h
index d056263..b0f2c56 100644
--- a/include/linux/platform_data/atmel.h
+++ b/include/linux/platform_data/atmel.h
@@ -4,8 +4,8 @@
  * GPL v2 Only
  */
 
-#ifndef __ATMEL_NAND_H__
-#define __ATMEL_NAND_H__
+#ifndef __ATMEL_H__
+#define __ATMEL_H__
 
 #include <linux/mtd/nand.h>
 
@@ -24,4 +24,4 @@ struct atmel_nand_data {
 	unsigned int	num_parts;
 };
 
-#endif /* __ATMEL_NAND_H__ */
+#endif /* __ATMEL_H__ */
-- 
1.7.9.1

[PATCH 05/11] blackfin: fix cmpxchg build fails from system.h fallout

[Paul Gortmaker]

Commit 3bed8d67469cc7129b0babc0211c32fa68408ce0

    "Disintegrate asm/system.h for Blackfin [ver #2]"

introduced arch/blackfin/include/asm/cmpxchg.h but has it also
including the asm-generic one which causes this:

  CC      arch/blackfin/kernel/asm-offsets.s
  In file included from arch/blackfin/include/asm/cmpxchg.h:125:0,
                 from arch/blackfin/include/asm/atomic.h:10,
                 from include/linux/atomic.h:4,
                 from include/linux/spinlock.h:384,
                 from include/linux/seqlock.h:29,
                 from include/linux/time.h:8,
                 from include/linux/timex.h:56,
                 from include/linux/sched.h:57,
                 from arch/blackfin/kernel/asm-offsets.c:10:
  include/asm-generic/cmpxchg.h:24:15: error: redefinition of '__xchg'
  arch/blackfin/include/asm/cmpxchg.h:82:29: note: previous definition of '__xchg' was here
  In file included from arch/blackfin/include/asm/atomic.h:10:0,
                 from include/linux/atomic.h:4,
                 from include/linux/spinlock.h:384,
                 from include/linux/seqlock.h:29,
                 from include/linux/time.h:8,
                 from include/linux/timex.h:56,
                 from include/linux/sched.h:57,
                 from arch/blackfin/kernel/asm-offsets.c:10:
  arch/blackfin/include/asm/cmpxchg.h:129:0: warning: "xchg" redefined [enabled by default]
  include/asm-generic/cmpxchg.h:81:0: note: this is the location of the previous definition
  make[2]: *** [arch/blackfin/kernel/asm-offsets.s] Error 1

It really only needs two simple defines from asm-generic, so just
use those instead.

Cc: Bob Liu <lliubbo@gmail.com>
Cc: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
---
 arch/blackfin/include/asm/cmpxchg.h |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)

diff --git a/arch/blackfin/include/asm/cmpxchg.h b/arch/blackfin/include/asm/cmpxchg.h
index ba2484f..c05868c 100644
--- a/arch/blackfin/include/asm/cmpxchg.h
+++ b/arch/blackfin/include/asm/cmpxchg.h
@@ -122,7 +122,8 @@ static inline unsigned long __xchg(unsigned long x, volatile void *ptr,
 			(unsigned long)(n), sizeof(*(ptr))))
 #define cmpxchg64_local(ptr, o, n) __cmpxchg64_local_generic((ptr), (o), (n))
 
-#include <asm-generic/cmpxchg.h>
+#define cmpxchg(ptr, o, n)	cmpxchg_local((ptr), (o), (n))
+#define cmpxchg64(ptr, o, n)	cmpxchg64_local((ptr), (o), (n))
 
 #endif /* !CONFIG_SMP */
 
-- 
1.7.9.1

[PATCH 06/11] parisc: fix missing cmpxchg file error from system.h split

[Paul Gortmaker]

Commit b4816afa3986704d1404fc48e931da5135820472

    "Move the asm-generic/system.h xchg() implementation to asm-generic/cmpxchg.h"

introduced the concept of asm/cmpxchg.h but the parisc arch
never got one.  Fork the cmpxchg content out of the asm/atomic.h
file to create one.

Some minor whitespace fixups were done on the block of code that
created the new file.

Cc: "James E.J. Bottomley" <jejb@parisc-linux.org>
Cc: Helge Deller <deller@gmx.de>
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
---
 arch/parisc/include/asm/atomic.h  |  107 +---------------------------------
 arch/parisc/include/asm/cmpxchg.h |  116 +++++++++++++++++++++++++++++++++++++
 2 files changed, 117 insertions(+), 106 deletions(-)
 create mode 100644 arch/parisc/include/asm/cmpxchg.h

diff --git a/arch/parisc/include/asm/atomic.h b/arch/parisc/include/asm/atomic.h
index 3ae5607..6c6defc 100644
--- a/arch/parisc/include/asm/atomic.h
+++ b/arch/parisc/include/asm/atomic.h
@@ -6,6 +6,7 @@
 #define _ASM_PARISC_ATOMIC_H_
 
 #include <linux/types.h>
+#include <asm/cmpxchg.h>
 
 /*
  * Atomic operations that C can't guarantee us.  Useful for
@@ -48,112 +49,6 @@ extern arch_spinlock_t __atomic_hash[ATOMIC_HASH_SIZE] __lock_aligned;
 #  define _atomic_spin_unlock_irqrestore(l,f) do { local_irq_restore(f); } while (0)
 #endif
 
-/* This should get optimized out since it's never called.
-** Or get a link error if xchg is used "wrong".
-*/
-extern void __xchg_called_with_bad_pointer(void);
-
-
-/* __xchg32/64 defined in arch/parisc/lib/bitops.c */
-extern unsigned long __xchg8(char, char *);
-extern unsigned long __xchg32(int, int *);
-#ifdef CONFIG_64BIT
-extern unsigned long __xchg64(unsigned long, unsigned long *);
-#endif
-
-/* optimizer better get rid of switch since size is a constant */
-static __inline__ unsigned long
-__xchg(unsigned long x, __volatile__ void * ptr, int size)
-{
-	switch(size) {
-#ifdef CONFIG_64BIT
-	case 8: return __xchg64(x,(unsigned long *) ptr);
-#endif
-	case 4: return __xchg32((int) x, (int *) ptr);
-	case 1: return __xchg8((char) x, (char *) ptr);
-	}
-	__xchg_called_with_bad_pointer();
-	return x;
-}
-
-
-/*
-** REVISIT - Abandoned use of LDCW in xchg() for now:
-** o need to test sizeof(*ptr) to avoid clearing adjacent bytes
-** o and while we are at it, could CONFIG_64BIT code use LDCD too?
-**
-**	if (__builtin_constant_p(x) && (x == NULL))
-**		if (((unsigned long)p & 0xf) == 0)
-**			return __ldcw(p);
-*/
-#define xchg(ptr,x) \
-	((__typeof__(*(ptr)))__xchg((unsigned long)(x),(ptr),sizeof(*(ptr))))
-
-
-#define __HAVE_ARCH_CMPXCHG	1
-
-/* bug catcher for when unsupported size is used - won't link */
-extern void __cmpxchg_called_with_bad_pointer(void);
-
-/* __cmpxchg_u32/u64 defined in arch/parisc/lib/bitops.c */
-extern unsigned long __cmpxchg_u32(volatile unsigned int *m, unsigned int old, unsigned int new_);
-extern unsigned long __cmpxchg_u64(volatile unsigned long *ptr, unsigned long old, unsigned long new_);
-
-/* don't worry...optimizer will get rid of most of this */
-static __inline__ unsigned long
-__cmpxchg(volatile void *ptr, unsigned long old, unsigned long new_, int size)
-{
-	switch(size) {
-#ifdef CONFIG_64BIT
-	case 8: return __cmpxchg_u64((unsigned long *)ptr, old, new_);
-#endif
-	case 4: return __cmpxchg_u32((unsigned int *)ptr, (unsigned int) old, (unsigned int) new_);
-	}
-	__cmpxchg_called_with_bad_pointer();
-	return old;
-}
-
-#define cmpxchg(ptr,o,n)						 \
-  ({									 \
-     __typeof__(*(ptr)) _o_ = (o);					 \
-     __typeof__(*(ptr)) _n_ = (n);					 \
-     (__typeof__(*(ptr))) __cmpxchg((ptr), (unsigned long)_o_,		 \
-				    (unsigned long)_n_, sizeof(*(ptr))); \
-  })
-
-#include <asm-generic/cmpxchg-local.h>
-
-static inline unsigned long __cmpxchg_local(volatile void *ptr,
-				      unsigned long old,
-				      unsigned long new_, int size)
-{
-	switch (size) {
-#ifdef CONFIG_64BIT
-	case 8:	return __cmpxchg_u64((unsigned long *)ptr, old, new_);
-#endif
-	case 4:	return __cmpxchg_u32(ptr, old, new_);
-	default:
-		return __cmpxchg_local_generic(ptr, old, new_, size);
-	}
-}
-
-/*
- * cmpxchg_local and cmpxchg64_local are atomic wrt current CPU. Always make
- * them available.
- */
-#define cmpxchg_local(ptr, o, n)				  	\
-	((__typeof__(*(ptr)))__cmpxchg_local((ptr), (unsigned long)(o),	\
-			(unsigned long)(n), sizeof(*(ptr))))
-#ifdef CONFIG_64BIT
-#define cmpxchg64_local(ptr, o, n)					\
-  ({									\
-	BUILD_BUG_ON(sizeof(*(ptr)) != 8);				\
-	cmpxchg_local((ptr), (o), (n));					\
-  })
-#else
-#define cmpxchg64_local(ptr, o, n) __cmpxchg64_local_generic((ptr), (o), (n))
-#endif
-
 /*
  * Note that we need not lock read accesses - aligned word writes/reads
  * are atomic, so a reader never sees inconsistent values.
diff --git a/arch/parisc/include/asm/cmpxchg.h b/arch/parisc/include/asm/cmpxchg.h
new file mode 100644
index 0000000..dbd1335
--- /dev/null
+++ b/arch/parisc/include/asm/cmpxchg.h
@@ -0,0 +1,116 @@
+/*
+ * forked from parisc asm/atomic.h which was:
+ *	Copyright (C) 2000 Philipp Rumpf <prumpf@tux.org>
+ *	Copyright (C) 2006 Kyle McMartin <kyle@parisc-linux.org>
+ */
+
+#ifndef _ASM_PARISC_CMPXCHG_H_
+#define _ASM_PARISC_CMPXCHG_H_
+
+/* This should get optimized out since it's never called.
+** Or get a link error if xchg is used "wrong".
+*/
+extern void __xchg_called_with_bad_pointer(void);
+
+/* __xchg32/64 defined in arch/parisc/lib/bitops.c */
+extern unsigned long __xchg8(char, char *);
+extern unsigned long __xchg32(int, int *);
+#ifdef CONFIG_64BIT
+extern unsigned long __xchg64(unsigned long, unsigned long *);
+#endif
+
+/* optimizer better get rid of switch since size is a constant */
+static inline unsigned long
+__xchg(unsigned long x, __volatile__ void *ptr, int size)
+{
+	switch (size) {
+#ifdef CONFIG_64BIT
+	case 8: return __xchg64(x, (unsigned long *) ptr);
+#endif
+	case 4: return __xchg32((int) x, (int *) ptr);
+	case 1: return __xchg8((char) x, (char *) ptr);
+	}
+	__xchg_called_with_bad_pointer();
+	return x;
+}
+
+/*
+** REVISIT - Abandoned use of LDCW in xchg() for now:
+** o need to test sizeof(*ptr) to avoid clearing adjacent bytes
+** o and while we are at it, could CONFIG_64BIT code use LDCD too?
+**
+**	if (__builtin_constant_p(x) && (x == NULL))
+**		if (((unsigned long)p & 0xf) == 0)
+**			return __ldcw(p);
+*/
+#define xchg(ptr, x) \
+	((__typeof__(*(ptr)))__xchg((unsigned long)(x), (ptr), sizeof(*(ptr))))
+
+#define __HAVE_ARCH_CMPXCHG	1
+
+/* bug catcher for when unsupported size is used - won't link */
+extern void __cmpxchg_called_with_bad_pointer(void);
+
+/* __cmpxchg_u32/u64 defined in arch/parisc/lib/bitops.c */
+extern unsigned long __cmpxchg_u32(volatile unsigned int *m, unsigned int old,
+				   unsigned int new_);
+extern unsigned long __cmpxchg_u64(volatile unsigned long *ptr,
+				   unsigned long old, unsigned long new_);
+
+/* don't worry...optimizer will get rid of most of this */
+static inline unsigned long
+__cmpxchg(volatile void *ptr, unsigned long old, unsigned long new_, int size)
+{
+	switch (size) {
+#ifdef CONFIG_64BIT
+	case 8: return __cmpxchg_u64((unsigned long *)ptr, old, new_);
+#endif
+	case 4: return __cmpxchg_u32((unsigned int *)ptr,
+				     (unsigned int)old, (unsigned int)new_);
+	}
+	__cmpxchg_called_with_bad_pointer();
+	return old;
+}
+
+#define cmpxchg(ptr, o, n)						 \
+({									 \
+	__typeof__(*(ptr)) _o_ = (o);					 \
+	__typeof__(*(ptr)) _n_ = (n);					 \
+	(__typeof__(*(ptr))) __cmpxchg((ptr), (unsigned long)_o_,	 \
+				    (unsigned long)_n_, sizeof(*(ptr))); \
+})
+
+#include <asm-generic/cmpxchg-local.h>
+
+static inline unsigned long __cmpxchg_local(volatile void *ptr,
+				      unsigned long old,
+				      unsigned long new_, int size)
+{
+	switch (size) {
+#ifdef CONFIG_64BIT
+	case 8:	return __cmpxchg_u64((unsigned long *)ptr, old, new_);
+#endif
+	case 4:	return __cmpxchg_u32(ptr, old, new_);
+	default:
+		return __cmpxchg_local_generic(ptr, old, new_, size);
+	}
+}
+
+/*
+ * cmpxchg_local and cmpxchg64_local are atomic wrt current CPU. Always make
+ * them available.
+ */
+#define cmpxchg_local(ptr, o, n)					\
+	((__typeof__(*(ptr)))__cmpxchg_local((ptr), (unsigned long)(o),	\
+			(unsigned long)(n), sizeof(*(ptr))))
+#ifdef CONFIG_64BIT
+#define cmpxchg64_local(ptr, o, n)					\
+({									\
+	BUILD_BUG_ON(sizeof(*(ptr)) != 8);				\
+	cmpxchg_local((ptr), (o), (n));					\
+})
+#else
+#define cmpxchg64_local(ptr, o, n) __cmpxchg64_local_generic((ptr), (o), (n))
+#endif
+
+#endif /* _ASM_PARISC_CMPXCHG_H_ */
-- 
1.7.9.1

[PATCH 07/11] sh: fix several build failures from system.h fallout

[Paul Gortmaker]

commit e839ca528718e68cad32a307dc9aabf01ef3eb05

    "Disintegrate asm/system.h for SH"

created the new file asm/cache_insns.h but it didn't add
an include for all the users of it.

The failure on fpu.c (also seen below) showed up roughly
at the same time, and since the fix is trivial (include traps.h)
I did not bother to bisect to 100% confirm it was caused by
the system.h disintegration.

arch/sh/mm/flush-sh4.c:16:2: error: implicit declaration of function 'register_align' [-Werror=implicit-function-declaration]
arch/sh/mm/flush-sh4.c:23:3: error: implicit declaration of function '__ocbwb' [-Werror=implicit-function-declaration]
arch/sh/mm/flush-sh4.c:57:3: error: implicit declaration of function '__ocbp' [-Werror=implicit-function-declaration]
arch/sh/mm/flush-sh4.c:87:3: error: implicit declaration of function '__ocbi' [-Werror=implicit-function-declaration]
cc1: all warnings being treated as errors
make[2]: *** [arch/sh/mm/flush-sh4.o] Error 1
arch/sh/mm/cache-sh4.c:70:3: error: implicit declaration of function '__ocbwb' [-Werror=implicit-function-declaration]
cc1: all warnings being treated as errors
make[2]: *** [arch/sh/mm/cache-sh4.o] Error 1
arch/sh/kernel/cpu/fpu.c:81:2: error: 'TRAP_HANDLER_DECL' undeclared (first use in this function)
arch/sh/kernel/cpu/fpu.c:83:20: error: 'regs' undeclared (first use in this function)
arch/sh/kernel/cpu/fpu.c:83:19: error: called object 'fpu_state_restore' is not a function
make[3]: *** [arch/sh/kernel/cpu/fpu.o] Error 1
arch/sh/kernel/cpu/sh4/fpu.c:407:2: error: 'TRAP_HANDLER_DECL' undeclared (first use in this function)
arch/sh/kernel/cpu/sh4/fpu.c:409:20: error: 'regs' undeclared (first use in this function)
make[4]: *** [arch/sh/kernel/cpu/sh4/fpu.o] Error 1

Cc: Paul Mundt <lethal@linux-sh.org>
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
---
 arch/sh/kernel/cpu/fpu.c     |    1 +
 arch/sh/kernel/cpu/sh4/fpu.c |    1 +
 arch/sh/mm/cache-sh4.c       |    1 +
 arch/sh/mm/flush-sh4.c       |    1 +
 4 files changed, 4 insertions(+), 0 deletions(-)

diff --git a/arch/sh/kernel/cpu/fpu.c b/arch/sh/kernel/cpu/fpu.c
index 7f1b70c..07fb02f 100644
--- a/arch/sh/kernel/cpu/fpu.c
+++ b/arch/sh/kernel/cpu/fpu.c
@@ -1,6 +1,7 @@
 #include <linux/sched.h>
 #include <linux/slab.h>
 #include <asm/processor.h>
+#include <asm/traps.h>
 #include <asm/fpu.h>
 
 int init_fpu(struct task_struct *tsk)
diff --git a/arch/sh/kernel/cpu/sh4/fpu.c b/arch/sh/kernel/cpu/sh4/fpu.c
index e74cd6c..36538ff 100644
--- a/arch/sh/kernel/cpu/sh4/fpu.c
+++ b/arch/sh/kernel/cpu/sh4/fpu.c
@@ -15,6 +15,7 @@
 #include <linux/io.h>
 #include <cpu/fpu.h>
 #include <asm/processor.h>
+#include <asm/traps.h>
 #include <asm/fpu.h>
 
 /* The PR (precision) bit in the FP Status Register must be clear when
diff --git a/arch/sh/mm/cache-sh4.c b/arch/sh/mm/cache-sh4.c
index 112fea1..c9667d9 100644
--- a/arch/sh/mm/cache-sh4.c
+++ b/arch/sh/mm/cache-sh4.c
@@ -19,6 +19,7 @@
 #include <asm/pgtable.h>
 #include <asm/mmu_context.h>
 #include <asm/cacheflush.h>
+#include <asm/cache_insns.h>
 
 /*
  * The maximum number of pages we support up to when doing ranged dcache
diff --git a/arch/sh/mm/flush-sh4.c b/arch/sh/mm/flush-sh4.c
index 75a17f5..59acd8a 100644
--- a/arch/sh/mm/flush-sh4.c
+++ b/arch/sh/mm/flush-sh4.c
@@ -1,6 +1,7 @@
 #include <linux/mm.h>
 #include <asm/mmu_context.h>
 #include <asm/cacheflush.h>
+#include <asm/cache_insns.h>
 #include <asm/traps.h>
 
 /*
-- 
1.7.9.1

[PATCH 08/11] frv: fix warnings in mb93090-mb00/pci-dma.c about implicit EXPORT_SYMBOL

[Paul Gortmaker]

To fix:

arch/frv/mb93090-mb00/pci-dma.c:31:1: warning: data definition has no type or storage class [enabled by default]
arch/frv/mb93090-mb00/pci-dma.c:31:1: warning: type defaults to 'int' in declaration of 'EXPORT_SYMBOL' [-Wimplicit-int]
arch/frv/mb93090-mb00/pci-dma.c:31:1: warning: parameter names (without types) in function declaration [enabled by default]
arch/frv/mb93090-mb00/pci-dma.c:38:1: warning: data definition has no type or storage class [enabled by default]

Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
---
 arch/frv/mb93090-mb00/pci-dma.c |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/arch/frv/mb93090-mb00/pci-dma.c b/arch/frv/mb93090-mb00/pci-dma.c
index 41098a3..4f8d8bc 100644
--- a/arch/frv/mb93090-mb00/pci-dma.c
+++ b/arch/frv/mb93090-mb00/pci-dma.c
@@ -13,6 +13,7 @@
 #include <linux/dma-mapping.h>
 #include <linux/list.h>
 #include <linux/pci.h>
+#include <linux/export.h>
 #include <linux/highmem.h>
 #include <linux/scatterlist.h>
 #include <asm/io.h>
-- 
1.7.9.1

[PATCH 09/11] firewire: restore the device.h include in linux/firewire.h

[Paul Gortmaker]

Commit 313162d0b83836e2f57e51b9b8650fb4b9c396ea

    "device.h: audit and cleanup users in main include dir"

exchanged an include <linux/device.h> for a struct *device
but in actuality I misread this file when creating 313162d
and it should have remained an include.

There were no build regressions since all consumers were
already getting device.h anyway, but make it right regardless.

Reported-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
---
 include/linux/firewire.h |    3 +--
 1 files changed, 1 insertions(+), 2 deletions(-)

diff --git a/include/linux/firewire.h b/include/linux/firewire.h
index 4db7b68..cdc9b71 100644
--- a/include/linux/firewire.h
+++ b/include/linux/firewire.h
@@ -2,6 +2,7 @@
 #define _LINUX_FIREWIRE_H
 
 #include <linux/completion.h>
+#include <linux/device.h>
 #include <linux/dma-mapping.h>
 #include <linux/kernel.h>
 #include <linux/kref.h>
@@ -64,8 +65,6 @@
 #define CSR_MODEL		0x17
 #define CSR_DIRECTORY_ID	0x20
 
-struct device;
-
 struct fw_csr_iterator {
 	const u32 *p;
 	const u32 *end;
-- 
1.7.9.1

[PATCH 10/11] tile: fix multiple build failures from system.h dismantle

[Paul Gortmaker]

Commit bd119c69239322caafdb64517a806037d0d0c70a

    "Disintegrate asm/system.h for Tile"

created the asm/switch_to.h file, but did not add an include
of it to all its users.

Also, commit b4816afa3986704d1404fc48e931da5135820472

        "Move the asm-generic/system.h xchg() implementation to asm-generic/cmpxchg.h"

introduced the concept of asm/cmpxchg.h but the tile arch
never got one.  Fork the cmpxchg content out of the asm/atomic.h
file to create one.

Cc: Chris Metcalf <cmetcalf@tilera.com>
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
---
 arch/tile/include/asm/atomic.h  |   50 +-------------------------
 arch/tile/include/asm/cmpxchg.h |   73 +++++++++++++++++++++++++++++++++++++++
 arch/tile/kernel/process.c      |    1 +
 arch/tile/kernel/stack.c        |    1 +
 4 files changed, 77 insertions(+), 48 deletions(-)
 create mode 100644 arch/tile/include/asm/cmpxchg.h

diff --git a/arch/tile/include/asm/atomic.h b/arch/tile/include/asm/atomic.h
index bb696da..f246142 100644
--- a/arch/tile/include/asm/atomic.h
+++ b/arch/tile/include/asm/atomic.h
@@ -17,6 +17,8 @@
 #ifndef _ASM_TILE_ATOMIC_H
 #define _ASM_TILE_ATOMIC_H
 
+#include <asm/cmpxchg.h>
+
 #ifndef __ASSEMBLY__
 
 #include <linux/compiler.h>
@@ -121,54 +123,6 @@ static inline int atomic_read(const atomic_t *v)
  */
 #define atomic_add_negative(i, v)	(atomic_add_return((i), (v)) < 0)
 
-/* Nonexistent functions intended to cause link errors. */
-extern unsigned long __xchg_called_with_bad_pointer(void);
-extern unsigned long __cmpxchg_called_with_bad_pointer(void);
-
-#define xchg(ptr, x)							\
-	({								\
-		typeof(*(ptr)) __x;					\
-		switch (sizeof(*(ptr))) {				\
-		case 4:							\
-			__x = (typeof(__x))(typeof(__x-__x))atomic_xchg( \
-				(atomic_t *)(ptr),			\
-				(u32)(typeof((x)-(x)))(x));		\
-			break;						\
-		case 8:							\
-			__x = (typeof(__x))(typeof(__x-__x))atomic64_xchg( \
-				(atomic64_t *)(ptr),			\
-				(u64)(typeof((x)-(x)))(x));		\
-			break;						\
-		default:						\
-			__xchg_called_with_bad_pointer();		\
-		}							\
-		__x;							\
-	})
-
-#define cmpxchg(ptr, o, n)						\
-	({								\
-		typeof(*(ptr)) __x;					\
-		switch (sizeof(*(ptr))) {				\
-		case 4:							\
-			__x = (typeof(__x))(typeof(__x-__x))atomic_cmpxchg( \
-				(atomic_t *)(ptr),			\
-				(u32)(typeof((o)-(o)))(o),		\
-				(u32)(typeof((n)-(n)))(n));		\
-			break;						\
-		case 8:							\
-			__x = (typeof(__x))(typeof(__x-__x))atomic64_cmpxchg( \
-				(atomic64_t *)(ptr),			\
-				(u64)(typeof((o)-(o)))(o),		\
-				(u64)(typeof((n)-(n)))(n));		\
-			break;						\
-		default:						\
-			__cmpxchg_called_with_bad_pointer();		\
-		}							\
-		__x;							\
-	})
-
-#define tas(ptr) (xchg((ptr), 1))
-
 #endif /* __ASSEMBLY__ */
 
 #ifndef __tilegx__
diff --git a/arch/tile/include/asm/cmpxchg.h b/arch/tile/include/asm/cmpxchg.h
new file mode 100644
index 0000000..276f067
--- /dev/null
+++ b/arch/tile/include/asm/cmpxchg.h
@@ -0,0 +1,73 @@
+/*
+ * cmpxchg.h -- forked from asm/atomic.h with this copyright:
+ *
+ * Copyright 2010 Tilera Corporation. All Rights Reserved.
+ *
+ *   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, version 2.
+ *
+ *   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, GOOD TITLE or
+ *   NON INFRINGEMENT.  See the GNU General Public License for
+ *   more details.
+ *
+ */
+
+#ifndef _ASM_TILE_CMPXCHG_H
+#define _ASM_TILE_CMPXCHG_H
+
+#ifndef __ASSEMBLY__
+
+/* Nonexistent functions intended to cause link errors. */
+extern unsigned long __xchg_called_with_bad_pointer(void);
+extern unsigned long __cmpxchg_called_with_bad_pointer(void);
+
+#define xchg(ptr, x)							\
+	({								\
+		typeof(*(ptr)) __x;					\
+		switch (sizeof(*(ptr))) {				\
+		case 4:							\
+			__x = (typeof(__x))(typeof(__x-__x))atomic_xchg( \
+				(atomic_t *)(ptr),			\
+				(u32)(typeof((x)-(x)))(x));		\
+			break;						\
+		case 8:							\
+			__x = (typeof(__x))(typeof(__x-__x))atomic64_xchg( \
+				(atomic64_t *)(ptr),			\
+				(u64)(typeof((x)-(x)))(x));		\
+			break;						\
+		default:						\
+			__xchg_called_with_bad_pointer();		\
+		}							\
+		__x;							\
+	})
+
+#define cmpxchg(ptr, o, n)						\
+	({								\
+		typeof(*(ptr)) __x;					\
+		switch (sizeof(*(ptr))) {				\
+		case 4:							\
+			__x = (typeof(__x))(typeof(__x-__x))atomic_cmpxchg( \
+				(atomic_t *)(ptr),			\
+				(u32)(typeof((o)-(o)))(o),		\
+				(u32)(typeof((n)-(n)))(n));		\
+			break;						\
+		case 8:							\
+			__x = (typeof(__x))(typeof(__x-__x))atomic64_cmpxchg( \
+				(atomic64_t *)(ptr),			\
+				(u64)(typeof((o)-(o)))(o),		\
+				(u64)(typeof((n)-(n)))(n));		\
+			break;						\
+		default:						\
+			__cmpxchg_called_with_bad_pointer();		\
+		}							\
+		__x;							\
+	})
+
+#define tas(ptr) (xchg((ptr), 1))
+
+#endif /* __ASSEMBLY__ */
+
+#endif /* _ASM_TILE_CMPXCHG_H */
diff --git a/arch/tile/kernel/process.c b/arch/tile/kernel/process.c
index 30caeca..ee01b1c 100644
--- a/arch/tile/kernel/process.c
+++ b/arch/tile/kernel/process.c
@@ -28,6 +28,7 @@
 #include <linux/tracehook.h>
 #include <linux/signal.h>
 #include <asm/stack.h>
+#include <asm/switch_to.h>
 #include <asm/homecache.h>
 #include <asm/syscalls.h>
 #include <asm/traps.h>
diff --git a/arch/tile/kernel/stack.c b/arch/tile/kernel/stack.c
index 37ee4d0..0be6b01 100644
--- a/arch/tile/kernel/stack.c
+++ b/arch/tile/kernel/stack.c
@@ -25,6 +25,7 @@
 #include <asm/page.h>
 #include <asm/tlbflush.h>
 #include <asm/ucontext.h>
+#include <asm/switch_to.h>
 #include <asm/sigframe.h>
 #include <asm/stack.h>
 #include <arch/abi.h>
-- 
1.7.9.1

[PATCH 11/11] asm-generic: add linux/types.h to cmpxchg.h

[Paul Gortmaker]

Builds of the openrisc or1ksim_defconfig show the following:

  In file included from arch/openrisc/include/generated/asm/cmpxchg.h:1:0,
                   from include/asm-generic/atomic.h:18,
                   from arch/openrisc/include/generated/asm/atomic.h:1,
                   from include/linux/atomic.h:4,
                   from include/linux/dcache.h:4,
                   from fs/notify/fsnotify.c:19:
  include/asm-generic/cmpxchg.h: In function '__xchg':
  include/asm-generic/cmpxchg.h:34:20: error: expected ')' before 'u8'
  include/asm-generic/cmpxchg.h:34:20: warning: type defaults to 'int' in type name

and many more lines of similar errors.  It seems specific to the or32
because most other platforms have an arch specific component that would
have already included types.h ahead of time, but the o32 does not.

Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Jonas Bonn <jonas@southpole.se>
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
---
 include/asm-generic/cmpxchg.h |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/include/asm-generic/cmpxchg.h b/include/asm-generic/cmpxchg.h
index 8a36183..1488302 100644
--- a/include/asm-generic/cmpxchg.h
+++ b/include/asm-generic/cmpxchg.h
@@ -10,6 +10,7 @@
 #error "Cannot use generic cmpxchg on SMP"
 #endif
 
+#include <linux/types.h>
 #include <linux/irqflags.h>
 
 #ifndef xchg
-- 
1.7.9.1

Re: [PATCH 00/11] Fix up more asm/system.h fallout

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1226 bytes --]

Hi Paul,

On Sun,  1 Apr 2012 16:38:36 -0400 Paul Gortmaker <paul.gortmaker@windriver.com> wrote:
>
> Like others, I was surprised it didn't get a soak in linux-next
> before being merged, but I guess what is done is done, so lets
> try and limit the fallout from having any wider of a commit range
> than it already does.
> 
> It is all system.h stuff except for three things.
> 
>  1) an avr32 patch I'd created but the avr32 guys hadn't got
>     signed keys yet, so I might as well get it in here along
>     with other similar build fail fixes (the atmel one).
> 
>  2) A cosmetic fix to the device.h changeset from me earlier.
> 
>  3) A trivial warning fix to an arch/frv specific file.
> 
> I'm going to queue this up in my feed branch to linux-next
> so we can start reducing the number of build failures from what
> we had on Friday.

Thanks for all this.  Please feed them to Linus as soon as sensible i.e.
after you get some feedback from the affected arch maintainers etc.
Maybe you should repost some of these with more cc's on them.

I look forward to less blood on the ground in today's linux-next
builds :-)
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au

[-- Attachment #2: Type: application/pgp-signature, Size: 836 bytes --]

Re: [PATCH 07/11] sh: fix several build failures from system.h fallout

[Paul Mundt]

On Sun, Apr 01, 2012 at 04:38:43PM -0400, Paul Gortmaker wrote:
> commit e839ca528718e68cad32a307dc9aabf01ef3eb05
> 
>     "Disintegrate asm/system.h for SH"
> 
> created the new file asm/cache_insns.h but it didn't add
> an include for all the users of it.
> 
> The failure on fpu.c (also seen below) showed up roughly
> at the same time, and since the fix is trivial (include traps.h)
> I did not bother to bisect to 100% confirm it was caused by
> the system.h disintegration.
> 
There's still more, but they've alreadly all been fixed up in my tree.

Re: [PATCH 00/11] Fix up more asm/system.h fallout

[David Howells]

Paul Gortmaker <paul.gortmaker@windriver.com> wrote:

> Like others, I was surprised it didn't get a soak in linux-next
> before being merged

I don't think that would've helped - unless Linus took all of linux-next into
his tree.  The problem is that there were a substantial number of differences
between Linus-vanilla and linux-next (obviously), even mid-window.

To be fair, Linus had to fix up my patch set to apply to his vanilla tree, but
I did ask him to to pull the patches later in the merge window and hopefully
give me a chance to fix them up again before he did so.  But I suppose arch
tree breakage was to be expected anyway in some arches.

However, I think the better way to have done this would've been to have the
arch-specific patches go through the arch trees first (and, indeed, Blackfin
did this and I was able to drop their patch from my set) - and then have the
generic patches go at the end.

Even better might've been to add a #warning into each asm/system.h in this
merge window and defer the final deletion to the next merge window.

David

Re: [PATCH 08/11] frv: fix warnings in mb93090-mb00/pci-dma.c about implicit EXPORT_SYMBOL

[David Howells]

Paul Gortmaker <paul.gortmaker@windriver.com> wrote:

> To fix:
> 
> arch/frv/mb93090-mb00/pci-dma.c:31:1: warning: data definition has no type or storage class [enabled by default]
> arch/frv/mb93090-mb00/pci-dma.c:31:1: warning: type defaults to 'int' in declaration of 'EXPORT_SYMBOL' [-Wimplicit-int]
> arch/frv/mb93090-mb00/pci-dma.c:31:1: warning: parameter names (without types) in function declaration [enabled by default]
> arch/frv/mb93090-mb00/pci-dma.c:38:1: warning: data definition has no type or storage class [enabled by default]
> 
> Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>

Acked-by: David Howells <dhowells@redhat.com>

Re: [PATCH 02/11] ARM: mach-msm: fix compile fail from system.h fallout

[David Howells]

Paul Gortmaker <paul.gortmaker@windriver.com> wrote:

> To fix:
> 
>   In file included from arm/boot/compressed/misc.c:28:0:
>   arm/mach-msm/include/mach/uncompress.h: In function 'putc':
>   arch/arm/mach-msm/include/mach/uncompress.h:48:3: error: implicit
>   declaration of function 'smp_mb' [-Werror=implicit-function-declaration]
> 
> The putc does a cpu_relax which for this platform is smp_mb.
> 
> Bisect indicates the 1st failing commit as:
> 
>     commit 0195c00244dc2e9f522475868fa278c473ba7339
>     "Merge tag 'split-asm_system_h-for-linus-20120328' ..."
> 
> Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>

Acked-by: David Howells <dhowells@redhat.com>

Re: [PATCH 01/11] irq_work: fix compile failure on MIPS from system.h split

[David Howells]

Paul Gortmaker <paul.gortmaker@windriver.com> wrote:

> Builds of the MIPS platform ip32_defconfig fails as of
> commit 0195c00244dc2e9f522475868fa278c473ba7339
> 
>     "Merge tag 'split-asm_system_h-for-linus-20120328' ..."
> 
> because MIPS xchg() macro uses BUILD_BUG_ON and it was moved
> in commit b81947c646bfefdf98e2fde5d7d39cbbda8525d4
> 
>     "Disintegrate asm/system.h for MIPS"
> 
> The root cause is that the system.h split wasn't tested on
> a baseline with commit 6c03438edeb5c359af35f060ea016ca65671c269
> 
>     "kernel.h: doesn't explicitly use bug.h, so don't include it."
> 
> Since this file uses BUG code in several other places besides
> the xchg call, simply make the inclusion explicit.
> 
> Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>

Acked-by: David Howells <dhowells@redhat.com>

Re: [PATCH 03/11] avr32: fix nop compile fails from system.h split up

[David Howells]

Paul Gortmaker <paul.gortmaker@windriver.com> wrote:

> To fix:
> 
>   In file included from kernel/exit.c:61:
>   arch/avr32/include/asm/mmu_context.h: In function 'enable_mmu':
>   arch/avr32/include/asm/mmu_context.h:135: error: implicit
>   declaration of function 'nop'

Better would be to move the definition of nop() into asm/barrier.h to be
consistent with other arches.

David

Re: [PATCH 05/11] blackfin: fix cmpxchg build fails from system.h fallout

[David Howells]

Paul Gortmaker <paul.gortmaker@windriver.com> wrote:

> Commit 3bed8d67469cc7129b0babc0211c32fa68408ce0
> 
>     "Disintegrate asm/system.h for Blackfin [ver #2]"
> 
> introduced arch/blackfin/include/asm/cmpxchg.h but has it also
> including the asm-generic one which causes this:
> 
>   CC      arch/blackfin/kernel/asm-offsets.s
>   In file included from arch/blackfin/include/asm/cmpxchg.h:125:0,
>                  from arch/blackfin/include/asm/atomic.h:10,
>                  from include/linux/atomic.h:4,
>                  from include/linux/spinlock.h:384,
>                  from include/linux/seqlock.h:29,
>                  from include/linux/time.h:8,
>                  from include/linux/timex.h:56,
>                  from include/linux/sched.h:57,
>                  from arch/blackfin/kernel/asm-offsets.c:10:
>   include/asm-generic/cmpxchg.h:24:15: error: redefinition of '__xchg'
>   arch/blackfin/include/asm/cmpxchg.h:82:29: note: previous definition of '__xchg' was here
>   In file included from arch/blackfin/include/asm/atomic.h:10:0,
>                  from include/linux/atomic.h:4,
>                  from include/linux/spinlock.h:384,
>                  from include/linux/seqlock.h:29,
>                  from include/linux/time.h:8,
>                  from include/linux/timex.h:56,
>                  from include/linux/sched.h:57,
>                  from arch/blackfin/kernel/asm-offsets.c:10:
>   arch/blackfin/include/asm/cmpxchg.h:129:0: warning: "xchg" redefined [enabled by default]
>   include/asm-generic/cmpxchg.h:81:0: note: this is the location of the previous definition
>   make[2]: *** [arch/blackfin/kernel/asm-offsets.s] Error 1
> 
> It really only needs two simple defines from asm-generic, so just
> use those instead.
> 
> Cc: Bob Liu <lliubbo@gmail.com>
> Cc: Mike Frysinger <vapier@gentoo.org>
> Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>

Acked-by: David Howells <dhowells@redhat.com>

Re: [PATCH 06/11] parisc: fix missing cmpxchg file error from system.h split

[David Howells]

Paul Gortmaker <paul.gortmaker@windriver.com> wrote:

> Commit b4816afa3986704d1404fc48e931da5135820472
> 
>     "Move the asm-generic/system.h xchg() implementation to asm-generic/cmpxchg.h"
> 
> introduced the concept of asm/cmpxchg.h but the parisc arch
> never got one.  Fork the cmpxchg content out of the asm/atomic.h
> file to create one.
> 
> Some minor whitespace fixups were done on the block of code that
> created the new file.
> 
> Cc: "James E.J. Bottomley" <jejb@parisc-linux.org>
> Cc: Helge Deller <deller@gmx.de>
> Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>

Acked-by: David Howells <dhowells@redhat.com>

Re: [PATCH 10/11] tile: fix multiple build failures from system.h dismantle

[David Howells]

Paul Gortmaker <paul.gortmaker@windriver.com> wrote:

> Commit bd119c69239322caafdb64517a806037d0d0c70a
> 
>     "Disintegrate asm/system.h for Tile"
> 
> created the asm/switch_to.h file, but did not add an include
> of it to all its users.
> 
> Also, commit b4816afa3986704d1404fc48e931da5135820472
> 
>         "Move the asm-generic/system.h xchg() implementation to asm-generic/cmpxchg.h"
> 
> introduced the concept of asm/cmpxchg.h but the tile arch
> never got one.  Fork the cmpxchg content out of the asm/atomic.h
> file to create one.
> 
> Cc: Chris Metcalf <cmetcalf@tilera.com>
> Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>

Acked-by: David Howells <dhowells@redhat.com>

Re: [PATCH 11/11] asm-generic: add linux/types.h to cmpxchg.h

[David Howells]

Paul Gortmaker <paul.gortmaker@windriver.com> wrote:

> Builds of the openrisc or1ksim_defconfig show the following:
> 
>   In file included from arch/openrisc/include/generated/asm/cmpxchg.h:1:0,
>                    from include/asm-generic/atomic.h:18,
>                    from arch/openrisc/include/generated/asm/atomic.h:1,
>                    from include/linux/atomic.h:4,
>                    from include/linux/dcache.h:4,
>                    from fs/notify/fsnotify.c:19:
>   include/asm-generic/cmpxchg.h: In function '__xchg':
>   include/asm-generic/cmpxchg.h:34:20: error: expected ')' before 'u8'
>   include/asm-generic/cmpxchg.h:34:20: warning: type defaults to 'int' in type name
> 
> and many more lines of similar errors.  It seems specific to the or32
> because most other platforms have an arch specific component that would
> have already included types.h ahead of time, but the o32 does not.
> 
> Cc: Arnd Bergmann <arnd@arndb.de>
> Cc: Jonas Bonn <jonas@southpole.se>
> Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>

Acked-by: David Howells <dhowells@redhat.com

Re: [PATCH 10/11] tile: fix multiple build failures from system.h dismantle

[Chris Metcalf]

On 4/2/2012 1:52 AM, David Howells wrote:
> Paul Gortmaker <paul.gortmaker@windriver.com> wrote:
>
>> Commit bd119c69239322caafdb64517a806037d0d0c70a
>>
>>     "Disintegrate asm/system.h for Tile"
>>
>> created the asm/switch_to.h file, but did not add an include
>> of it to all its users.
>>
>> Also, commit b4816afa3986704d1404fc48e931da5135820472
>>
>>         "Move the asm-generic/system.h xchg() implementation to asm-generic/cmpxchg.h"
>>
>> introduced the concept of asm/cmpxchg.h but the tile arch
>> never got one.  Fork the cmpxchg content out of the asm/atomic.h
>> file to create one.
>>
>> Cc: Chris Metcalf <cmetcalf@tilera.com>
>> Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
> Acked-by: David Howells <dhowells@redhat.com>

Thanks!  I can push this through the tile tree along with some other
bug-fix changes for -rc2.

-- 
Chris Metcalf, Tilera Corp.
http://www.tilera.com

Re: [PATCH 10/11] tile: fix multiple build failures from system.h dismantle

[Paul Gortmaker]

On 12-04-02 11:59 AM, Chris Metcalf wrote:
> On 4/2/2012 1:52 AM, David Howells wrote:
>> Paul Gortmaker <paul.gortmaker@windriver.com> wrote:
>>
>>> Commit bd119c69239322caafdb64517a806037d0d0c70a
>>>
>>>     "Disintegrate asm/system.h for Tile"
>>>
>>> created the asm/switch_to.h file, but did not add an include
>>> of it to all its users.
>>>
>>> Also, commit b4816afa3986704d1404fc48e931da5135820472
>>>
>>>         "Move the asm-generic/system.h xchg() implementation to asm-generic/cmpxchg.h"
>>>
>>> introduced the concept of asm/cmpxchg.h but the tile arch
>>> never got one.  Fork the cmpxchg content out of the asm/atomic.h
>>> file to create one.
>>>
>>> Cc: Chris Metcalf <cmetcalf@tilera.com>
>>> Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
>> Acked-by: David Howells <dhowells@redhat.com>
> 
> Thanks!  I can push this through the tile tree along with some other
> bug-fix changes for -rc2.

Perfect.  I see you have a tree in linux-next, so I'll drop it
from my series here without having to worry about reintroducing
the build failure.

Thanks.
Paul.

> 

Re: [PATCH 07/11] sh: fix several build failures from system.h fallout

[Paul Gortmaker]

On 12-04-01 09:00 PM, Paul Mundt wrote:
> On Sun, Apr 01, 2012 at 04:38:43PM -0400, Paul Gortmaker wrote:
>> commit e839ca528718e68cad32a307dc9aabf01ef3eb05
>>
>>     "Disintegrate asm/system.h for SH"
>>
>> created the new file asm/cache_insns.h but it didn't add
>> an include for all the users of it.
>>
>> The failure on fpu.c (also seen below) showed up roughly
>> at the same time, and since the fix is trivial (include traps.h)
>> I did not bother to bisect to 100% confirm it was caused by
>> the system.h disintegration.
>>
> There's still more, but they've alreadly all been fixed up in my tree.

OK, I can drop this one.  Is that tree part of linux-next?

Thanks,
Paul.

Re: [PATCH 03/11] avr32: fix nop compile fails from system.h split up

[Paul Gortmaker]

On 12-04-02 01:47 AM, David Howells wrote:
> Paul Gortmaker <paul.gortmaker@windriver.com> wrote:
> 
>> To fix:
>>
>>   In file included from kernel/exit.c:61:
>>   arch/avr32/include/asm/mmu_context.h: In function 'enable_mmu':
>>   arch/avr32/include/asm/mmu_context.h:135: error: implicit
>>   declaration of function 'nop'
> 
> Better would be to move the definition of nop() into asm/barrier.h to be
> consistent with other arches.

That does seem to make sense.  But just to confirm, then I'd end up
deleting the new file, arch/avr32/include/asm/special_insns.h

------------
commit ae473946586680b01c13975a3b674de23ad7c93e
Author: David Howells <dhowells@redhat.com>
Date:   Wed Mar 28 18:30:01 2012 +0100

    Disintegrate asm/system.h for AVR32
    
    Disintegrate asm/system.h for AVR32.  This has not been compiled at all.
    
    Signed-off-by: David Howells <dhowells@redhat.com>
    cc: Haavard Skinnemoen <hskinnemoen@gmail.com>

:000000 100644 0000000... f922218... A  arch/avr32/include/asm/special_insns.h
------------

...since it only contains the nop definition.

Is that what you had in mind, or is some other content going to
be moved there at a later time?

Thanks,
Paul.

> 
> David

Re: [PATCH 02/11] ARM: mach-msm: fix compile fail from system.h fallout

[David Brown]

On Sun, Apr 01, 2012 at 04:38:38PM -0400, Paul Gortmaker wrote:

> diff --git a/arch/arm/mach-msm/include/mach/uncompress.h b/arch/arm/mach-msm/include/mach/uncompress.h

Acked-by: David Brown <davidb@codeaurora.org>

-- 
Sent by an employee of the Qualcomm Innovation Center, Inc.
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum.

Re: [PATCH 07/11] sh: fix several build failures from system.h fallout

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 744 bytes --]

On Mon, 2 Apr 2012 14:15:53 -0400 Paul Gortmaker <paul.gortmaker@windriver.com> wrote:
>
> On 12-04-01 09:00 PM, Paul Mundt wrote:
>
> > There's still more, but they've alreadly all been fixed up in my tree.
> 
> OK, I can drop this one.  Is that tree part of linux-next?

I used to have sh-current and sh trees (and rmobile-current and rmobile)
trees in linux-next, but they have not been resurrected since the
kernel.org problems last year ...

They were all branches of
git://git.kernel.org/pub/scm/linux/kernel/git/lethal/sh-2.6.git which is
still listed in MAINTAINERS but does not exist :-(

Paul M, to you have an alternate place I can fetch from?
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au

[-- Attachment #2: Type: application/pgp-signature, Size: 836 bytes --]

Re: [PATCH 07/11] sh: fix several build failures from system.h fallout

[Paul Gortmaker]

On Mon, Apr 2, 2012 at 7:57 PM, Stephen Rothwell <sfr@canb.auug.org.au> wrote:
> On Mon, 2 Apr 2012 14:15:53 -0400 Paul Gortmaker <paul.gortmaker@windriver.com> wrote:
>>
>> On 12-04-01 09:00 PM, Paul Mundt wrote:
>>
>> > There's still more, but they've alreadly all been fixed up in my tree.
>>
>> OK, I can drop this one.  Is that tree part of linux-next?
>
> I used to have sh-current and sh trees (and rmobile-current and rmobile)
> trees in linux-next, but they have not been resurrected since the
> kernel.org problems last year ...
>
> They were all branches of
> git://git.kernel.org/pub/scm/linux/kernel/git/lethal/sh-2.6.git which is
> still listed in MAINTAINERS but does not exist :-(
>
> Paul M, to you have an alternate place I can fetch from?
> --
> Cheers,
> Stephen Rothwell                    sfr@canb.auug.org.au

Re: [PATCH 07/11] sh: fix several build failures from system.h fallout

[Paul Gortmaker]

[sorry for empty mail, pilot error]

On Mon, Apr 2, 2012 at 7:57 PM, Stephen Rothwell <sfr@canb.auug.org.au> wrote:

>
> I used to have sh-current and sh trees (and rmobile-current and rmobile)
> trees in linux-next, but they have not been resurrected since the
> kernel.org problems last year ...
>
> They were all branches of
> git://git.kernel.org/pub/scm/linux/kernel/git/lethal/sh-2.6.git which is
> still listed in MAINTAINERS but does not exist :-(
>
> Paul M, to you have an alternate place I can fetch from?

Hi Paul M.

Is this the tree Stephen should be using in linux-next?

        git://github.com/pmundt/linux-sh.git sh-latest

Thanks,
Paul G.


> --
> Cheers,
> Stephen Rothwell                    sfr@canb.auug.org.au

Re: [PATCH 07/11] sh: fix several build failures from system.h fallout

[Paul Mundt]

On Wed, Apr 04, 2012 at 10:49:30PM -0400, Paul Gortmaker wrote:
> [sorry for empty mail, pilot error]
> 
> On Mon, Apr 2, 2012 at 7:57 PM, Stephen Rothwell <sfr@canb.auug.org.au> wrote:
> 
> >
> > I used to have sh-current and sh trees (and rmobile-current and rmobile)
> > trees in linux-next, but they have not been resurrected since the
> > kernel.org problems last year ...
> >
> > They were all branches of
> > git://git.kernel.org/pub/scm/linux/kernel/git/lethal/sh-2.6.git which is
> > still listed in MAINTAINERS but does not exist :-(
> >
> > Paul M, to you have an alternate place I can fetch from?
> 
> Hi Paul M.
> 
> Is this the tree Stephen should be using in linux-next?
> 
>         git://github.com/pmundt/linux-sh.git sh-latest
> 
Yes, that's the current one for sh stuff for now. The rmobile stuff is
going through the ARM SoC tree nowadays, so that shouldn't matter.  I
haven't bothered getting the trees back on kernel.org yet, so I haven't
updated the MAINTAINERS entry, but will probably do so in the future.

Re: [PATCH 11/11] asm-generic: add linux/types.h to cmpxchg.h

[Jonas Bonn]

[-- Attachment #1: Type: text/plain, Size: 1623 bytes --]

Thanks, applied.
Will send this upstream via the openrisc tree.
/Jonas

On Sun, 2012-04-01 at 16:38 -0400, Paul Gortmaker wrote:
> Builds of the openrisc or1ksim_defconfig show the following:
> 
>   In file included from arch/openrisc/include/generated/asm/cmpxchg.h:1:0,
>                    from include/asm-generic/atomic.h:18,
>                    from arch/openrisc/include/generated/asm/atomic.h:1,
>                    from include/linux/atomic.h:4,
>                    from include/linux/dcache.h:4,
>                    from fs/notify/fsnotify.c:19:
>   include/asm-generic/cmpxchg.h: In function '__xchg':
>   include/asm-generic/cmpxchg.h:34:20: error: expected ')' before 'u8'
>   include/asm-generic/cmpxchg.h:34:20: warning: type defaults to 'int' in type name
> 
> and many more lines of similar errors.  It seems specific to the or32
> because most other platforms have an arch specific component that would
> have already included types.h ahead of time, but the o32 does not.
> 
> Cc: Arnd Bergmann <arnd@arndb.de>
> Cc: Jonas Bonn <jonas@southpole.se>
> Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
> ---
>  include/asm-generic/cmpxchg.h |    1 +
>  1 files changed, 1 insertions(+), 0 deletions(-)
> 
> diff --git a/include/asm-generic/cmpxchg.h b/include/asm-generic/cmpxchg.h
> index 8a36183..1488302 100644
> --- a/include/asm-generic/cmpxchg.h
> +++ b/include/asm-generic/cmpxchg.h
> @@ -10,6 +10,7 @@
>  #error "Cannot use generic cmpxchg on SMP"
>  #endif
>  
> +#include <linux/types.h>
>  #include <linux/irqflags.h>
>  
>  #ifndef xchg


[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

linux-next: sh tree (Was:Re: [PATCH 07/11] sh: fix several build failures from system.h fallout)

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 836 bytes --]

Hi Paul,

On Fri, 6 Apr 2012 00:15:59 +0900 Paul Mundt <lethal@linux-sh.org> wrote:
>
> On Wed, Apr 04, 2012 at 10:49:30PM -0400, Paul Gortmaker wrote:
> > 
> > Is this the tree Stephen should be using in linux-next?
> > 
> >         git://github.com/pmundt/linux-sh.git sh-latest
> 
> Yes, that's the current one for sh stuff for now. The rmobile stuff is
> going through the ARM SoC tree nowadays, so that shouldn't matter.  I
> haven't bothered getting the trees back on kernel.org yet, so I haven't
> updated the MAINTAINERS entry, but will probably do so in the future.

OK, I have added that back - using the sh-latest branch for the sh tree
and the sh-fixes-for-linus branch for sh-current (maybe the sh/urgent
branch would be better?).

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au

[-- Attachment #2: Type: application/pgp-signature, Size: 836 bytes --]