Linux MIPS Architecture development
 help / color / mirror / Atom feed
* [PATCH] Always include PHYS_OFFSET in PAGE_OFFSET
@ 2007-03-19 15:59 Franck Bui-Huu
  2007-03-20 16:21 ` Franck Bui-Huu
  0 siblings, 1 reply; 2+ messages in thread
From: Franck Bui-Huu @ 2007-03-19 15:59 UTC (permalink / raw)
  To: Ralf Baechle; +Cc: mbizon, post, linux-mips

From: Franck Bui-Huu <fbuihuu@gmail.com>

For platforms that use PHYS_OFFSET and do not use a mapped kernel,
this patch automatically adds PHYS_OFFSET into PAGE_OFFSET.
Therefore for these platforms there are no more needs to change
PAGE_OFFSET.

For mapped kernel, they need to redefine PAGE_OFFSET anyways.

Signed-off-by: Franck Bui-Huu <fbuihuu@gmail.com>
---

 Maxime,

 Could you give a try to this patch ? It removes the need to
 change your PAGE_OFFSET. If I remember correctly it's now
 0x90000000, and you should be able to restore back to
 0x80000000.

		Franck

 include/asm-mips/mach-generic/spaces.h |    6 +++---
 1 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/include/asm-mips/mach-generic/spaces.h b/include/asm-mips/mach-generic/spaces.h
index 0ae9997..600561f 100644
--- a/include/asm-mips/mach-generic/spaces.h
+++ b/include/asm-mips/mach-generic/spaces.h
@@ -22,7 +22,7 @@
  * This handles the memory map.
  * We handle pages at KSEG0 for kernels with 32 bit address space.
  */
-#define PAGE_OFFSET		0x80000000UL
+#define PAGE_OFFSET		(0x80000000UL + PHYS_OFFSET)
 
 /*
  * Memory above this physical address will be considered highmem.
@@ -39,9 +39,9 @@
  * This handles the memory map.
  */
 #ifdef CONFIG_DMA_NONCOHERENT
-#define PAGE_OFFSET	0x9800000000000000UL
+#define PAGE_OFFSET	(0x9800000000000000UL + PHYS_OFFSET)
 #else
-#define PAGE_OFFSET	0xa800000000000000UL
+#define PAGE_OFFSET	(0xa800000000000000UL + PHYS_OFFSET)
 #endif
 
 /*
-- 
1.4.4.3.ge6d4

^ permalink raw reply related	[flat|nested] 2+ messages in thread

* Re: [PATCH] Always include PHYS_OFFSET in PAGE_OFFSET
  2007-03-19 15:59 [PATCH] Always include PHYS_OFFSET in PAGE_OFFSET Franck Bui-Huu
@ 2007-03-20 16:21 ` Franck Bui-Huu
  0 siblings, 0 replies; 2+ messages in thread
From: Franck Bui-Huu @ 2007-03-20 16:21 UTC (permalink / raw)
  To: Ralf Baechle; +Cc: mbizon, post, linux-mips

Ralf,

On 3/19/07, Franck Bui-Huu <vagabon.xyz@gmail.com> wrote:
> From: Franck Bui-Huu <fbuihuu@gmail.com>
>
> For platforms that use PHYS_OFFSET and do not use a mapped kernel,
> this patch automatically adds PHYS_OFFSET into PAGE_OFFSET.
> Therefore for these platforms there are no more needs to change
> PAGE_OFFSET.
>

This patch is actually useless. I didn't notice that spaces.h must be
totaly rewritten by platforms when they need to customize one value.
Why, I dunno ?

Why not doing something like:

-- >8 --
diff --git a/include/asm-mips/mach-generic/spaces.h
b/include/asm-mips/mach-generic/spaces.h
index 0ae9997..beec80e 100644
--- a/include/asm-mips/mach-generic/spaces.h
+++ b/include/asm-mips/mach-generic/spaces.h
@@ -16,13 +16,18 @@
 #define CAC_BASE               0x80000000
 #define IO_BASE                        0xa0000000
 #define UNCAC_BASE             0xa0000000
+
+#ifndef MAP_BASE
 #define MAP_BASE               0xc0000000
+#endif

 /*
  * This handles the memory map.
  * We handle pages at KSEG0 for kernels with 32 bit address space.
  */
+#ifndef PAGE_OFFSET
 #define PAGE_OFFSET            0x80000000UL
+#endif

 /*
  * Memory above this physical address will be considered highmem.
@@ -38,11 +43,13 @@
 /*
  * This handles the memory map.
  */
+#ifndef PAGE_OFFSET
 #ifdef CONFIG_DMA_NONCOHERENT
 #define PAGE_OFFSET    0x9800000000000000UL
 #else
 #define PAGE_OFFSET    0xa800000000000000UL
 #endif
+#endif

 /*
  * Memory above this physical address will be considered highmem.
-- >8 --

and doing in platform's spaces.h:

        [ redefine a couple of constants]

        #include <asm/mach-generic/spaces.h>

AFAIK, {CAC,UNCAC,IO}_BASE are likely to be the same for all platforms, no ?

Does that need some cleanup ? if so I'm your man ;)
-- 
               Franck

^ permalink raw reply related	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2007-03-20 16:22 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2007-03-19 15:59 [PATCH] Always include PHYS_OFFSET in PAGE_OFFSET Franck Bui-Huu
2007-03-20 16:21 ` Franck Bui-Huu

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox