sh: pass through ioremap() for non-mmu processors

PATCH In review Magnus Damm 2 participants 4 messages Apr 28, 2009 – Apr 30, 2009 SuperH

lore

Share & tools

lore

Magnus Damm

author

2009-04-28 07:59 UTC #0 lore raw

PATCH 1/1 sh: pass through ioremap() for non-mmu processors

From: Magnus Damm <[email protected]>

This patch modifies the ioremap() handling for CONFIG_MMU=n.

All 32-bit SuperH processors currently go through __ioremap_mode() and check for IO_TRAPPED and directly mapped segments. With this patch we simplify the MMU less case with a pass through version of
__ioremap_mode() which just returns the physical address.

The effects of this is change are:
- fix non-MMU ioremap() of high address hardware blocks (sh7203 CMT)
- IO_TRAPPED is never checked for

Signed-off-by Magnus Damm <[email protected]>

---

 arch/sh/Kconfig          |    1 +
 arch/sh/include/asm/io.h |   13 +++++++------
 2 files changed, 8 insertions(+), 6 deletions(-)

arch/sh/Kconfig +1 -0

@@ -136,6 +136,7 @@ config ARCH_HAS_DEFAULT_IDLE
 
 config IO_TRAPPED
 	bool
+	depends on MMU
 
 source "init/Kconfig"

arch/sh/include/asm/io.h +7 -6

@@ -228,12 +228,6 @@ void __iounmap(void __iomem *addr);
 unsigned long onchip_remap(unsigned long addr, unsigned long size,
 			   const char *name);
 extern void onchip_unmap(unsigned long vaddr);
-#else
-#define __ioremap(offset, size, flags)	((void __iomem *)(offset))
-#define __iounmap(addr)			do { } while (0)
-#define onchip_remap(addr, size, name)	(addr)
-#define onchip_unmap(addr)		do { } while (0)
-#endif /* CONFIG_MMU */
 
 static inline void __iomem *
 __ioremap_mode(unsigned long offset, unsigned long size, unsigned long flags)

@@ -269,6 +263,13 @@ __ioremap_mode(unsigned long offset, uns
 	return __ioremap(offset, size, flags);
 }
 
+#else
+#define onchip_remap(addr, size, name)	(addr)
+#define onchip_unmap(addr)		do { } while (0)
+#define __ioremap_mode(offset, size, flags) (void __iomem *)(offset)
+#define __iounmap(addr)			do { } while (0)
+#endif /* CONFIG_MMU */
+
 #define ioremap(offset, size)				\
 	__ioremap_mode((offset), (size), 0)
 #define ioremap_nocache(offset, size)			\

Paul Mundt

2009-04-28 08:49 UTC #1 lore raw

Quoted 23 lines · Magnus Damm

From: Magnus Damm <[email protected]>
This patch modifies the ioremap() handling for CONFIG_MMU=n.
All 32-bit SuperH processors currently go through __ioremap_mode() and check for IO_TRAPPED and directly mapped segments. With this patch we simplify the MMU less case with a pass through version of
__ioremap_mode() which just returns the physical address.
The effects of this is change are:
- fix non-MMU ioremap() of high address hardware blocks (sh7203 CMT)
- IO_TRAPPED is never checked for
Signed-off-by Magnus Damm <[email protected]>
---
 arch/sh/Kconfig          |    1 +
 arch/sh/include/asm/io.h |   13 +++++++------
 2 files changed, 8 insertions(+), 6 deletions(-)

arch/sh/Kconfig config ARCH_HAS_DEFAULT_IDLE · line 140

Quoted 5 lines

arch/sh/Kconfig +1 -0

@@ -136,6 +136,7 @@ config ARCH_HAS_DEFAULT_IDLE
 
 config IO_TRAPPED
 	bool
+	depends on MMU

You have failed at Kconfig. select trumps depends, so this has no impact for platforms that select IO_TRAPPED. You will have to switch the platforms that select this over to handling the dependency properly on their own.

Magnus Damm

author

2009-04-29 14:44 UTC #2 lore raw

PATCH 1/1 sh: pass through ioremap() for non-mmu processors V2

In reply to Magnus Damm

From: Magnus Damm <[email protected]>

This is V2 of the CONFIG_MMU=n pass through ioremap() patch.

The effects of this is change are:
- fix non-MMU ioremap() of high address hardware blocks (sh7203 CMT)
- make sure IO_TRAPPED is not selected

Signed-off-by Magnus Damm <[email protected]>

---

Changes since V1:
- rework Kconfig bits

 arch/sh/boards/Kconfig   |    6 +++---
 arch/sh/include/asm/io.h |   13 +++++++------
 2 files changed, 10 insertions(+), 9 deletions(-)

arch/sh/boards/Kconfig +3 -3

@@ -121,7 +121,7 @@ config SH_RTS7751R2D
 	bool "RTS7751R2D"
 	depends on CPU_SUBTYPE_SH7751R
 	select SYS_SUPPORTS_PCI
-	select IO_TRAPPED
+	select IO_TRAPPED if MMU
 	help
 	  Select RTS7751R2D if configuring for a Renesas Technology
 	  Sales SH-Graphics board.

@@ -145,13 +145,13 @@ config SH_HIGHLANDER
 	bool "Highlander"
 	depends on CPU_SUBTYPE_SH7780 || CPU_SUBTYPE_SH7785
 	select SYS_SUPPORTS_PCI
-	select IO_TRAPPED
+	select IO_TRAPPED if MMU
 
 config SH_SH7785LCR
 	bool "SH7785LCR"
 	depends on CPU_SUBTYPE_SH7785
 	select SYS_SUPPORTS_PCI
-	select IO_TRAPPED
+	select IO_TRAPPED if MMU
 
 config SH_SH7785LCR_29BIT_PHYSMAPS
 	bool "SH7785LCR 29bit physmaps"

arch/sh/include/asm/io.h +7 -6

@@ -228,12 +228,6 @@ void __iounmap(void __iomem *addr);
 unsigned long onchip_remap(unsigned long addr, unsigned long size,
 			   const char *name);
 extern void onchip_unmap(unsigned long vaddr);
-#else
-#define __ioremap(offset, size, flags)	((void __iomem *)(offset))
-#define __iounmap(addr)			do { } while (0)
-#define onchip_remap(addr, size, name)	(addr)
-#define onchip_unmap(addr)		do { } while (0)
-#endif /* CONFIG_MMU */
 
 static inline void __iomem *
 __ioremap_mode(unsigned long offset, unsigned long size, unsigned long flags)

@@ -269,6 +263,13 @@ __ioremap_mode(unsigned long offset, uns
 	return __ioremap(offset, size, flags);
 }
 
+#else
+#define onchip_remap(addr, size, name)	(addr)
+#define onchip_unmap(addr)		do { } while (0)
+#define __ioremap_mode(offset, size, flags) (void __iomem *)(offset)
+#define __iounmap(addr)			do { } while (0)
+#endif /* CONFIG_MMU */
+
 #define ioremap(offset, size)				\
 	__ioremap_mode((offset), (size), 0)
 #define ioremap_nocache(offset, size)			\

Paul Mundt

2009-04-30 03:53 UTC #3 lore raw

Re: [PATCH] sh: pass through ioremap() for non-mmu processors V2

In reply to Magnus Damm

Quoted 10 lines · Magnus Damm

This is V2 of the CONFIG_MMU=n pass through ioremap() patch.
All 32-bit SuperH processors currently go through __ioremap_mode() and check for IO_TRAPPED and directly mapped segments. With this patch we simplify the MMU less case with a pass through version of
__ioremap_mode() which just returns the physical address.
The effects of this is change are:
- fix non-MMU ioremap() of high address hardware blocks (sh7203 CMT)
- make sure IO_TRAPPED is not selected

Quoted 2 lines · Magnus Damm

This patch contains various fixes for 16-bit cmt hardware. With this applied periodic clockevents work fine on sh7203.

Both applied, thanks.