[PATCH 1/5] OpenRISC: Select GENERIC_ATOMIC64

[PATCH 1/5] OpenRISC: Select GENERIC_ATOMIC64

[Richard Weinberger]

To have atomit64_t on OpenRISC we need GENERIC_ATOMIC64.

Signed-off-by: Richard Weinberger <richard@nod.at>
---
 arch/openrisc/Kconfig |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/arch/openrisc/Kconfig b/arch/openrisc/Kconfig
index bc428b5..a478719 100644
--- a/arch/openrisc/Kconfig
+++ b/arch/openrisc/Kconfig
@@ -16,6 +16,7 @@ config OPENRISC
 	select GENERIC_IRQ_SHOW
 	select GENERIC_IOMAP
 	select GENERIC_CPU_DEVICES
+	select GENERIC_ATOMIC64
 
 config MMU
 	def_bool y
-- 
1.7.7.3

[PATCH 2/5] OpenRISC: Use generic bitops.h

[Richard Weinberger]

There is no need to reinvent the wheel.

Signed-off-by: Richard Weinberger <richard@nod.at>
---
 arch/openrisc/include/asm/Kbuild |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/arch/openrisc/include/asm/Kbuild b/arch/openrisc/include/asm/Kbuild
index 11162e6..a283165 100644
--- a/arch/openrisc/include/asm/Kbuild
+++ b/arch/openrisc/include/asm/Kbuild
@@ -4,6 +4,7 @@ header-y += spr_defs.h
 
 generic-y += atomic.h
 generic-y += auxvec.h
+generic-y += bitops.h
 generic-y += bitsperlong.h
 generic-y += bug.h
 generic-y += bugs.h
-- 
1.7.7.3

[PATCH 3/5] OpenRISC: Export memory_start and memory_end

[Richard Weinberger]

At least virt_addr_valid() makes use of memory_end.

Signed-off-by: Richard Weinberger <richard@nod.at>
---
 arch/openrisc/include/asm/page.h |    4 ++--
 arch/openrisc/kernel/setup.c     |    2 +-
 2 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/arch/openrisc/include/asm/page.h b/arch/openrisc/include/asm/page.h
index b041b34..4dfe477 100644
--- a/arch/openrisc/include/asm/page.h
+++ b/arch/openrisc/include/asm/page.h
@@ -71,8 +71,8 @@ typedef struct page *pgtable_t;
 #define __pgd(x)	((pgd_t) { (x) })
 #define __pgprot(x)	((pgprot_t) { (x) })
 
-extern unsigned long memory_start;
-extern unsigned long memory_end;
+extern phys_addr_t memory_start;
+extern phys_addr_t memory_end;
 
 #endif /* !__ASSEMBLY__ */
 
diff --git a/arch/openrisc/kernel/setup.c b/arch/openrisc/kernel/setup.c
index bf5eba2..c993c92 100644
--- a/arch/openrisc/kernel/setup.c
+++ b/arch/openrisc/kernel/setup.c
@@ -52,6 +52,7 @@
 #include "vmlinux.h"
 
 char __initdata cmd_line[COMMAND_LINE_SIZE] = CONFIG_CMDLINE;
+phys_addr_t memory_start, memory_end;
 
 static unsigned long __init setup_memory(void)
 {
@@ -59,7 +60,6 @@ static unsigned long __init setup_memory(void)
 	unsigned long ram_start_pfn;
 	unsigned long free_ram_start_pfn;
 	unsigned long ram_end_pfn;
-	phys_addr_t memory_start, memory_end;
 	struct memblock_region *region;
 
 	memory_end = memory_start = 0;
-- 
1.7.7.3

[PATCH 4/5] OpenRISC: Include export.h

[Richard Weinberger]

EXPORT_SYMBOL* needs export.h.

Signed-off-by: Richard Weinberger <richard@nod.at>
---
 arch/openrisc/kernel/init_task.c |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/arch/openrisc/kernel/init_task.c b/arch/openrisc/kernel/init_task.c
index 45744a38..70381d4 100644
--- a/arch/openrisc/kernel/init_task.c
+++ b/arch/openrisc/kernel/init_task.c
@@ -15,6 +15,7 @@
  *      2 of the License, or (at your option) any later version.
  */
 
+#include <linux/export.h>
 #include <linux/init_task.h>
 #include <linux/mqueue.h>
 
-- 
1.7.7.3

[PATCH 5/5] OpenRISC: Export dump_stack()

[Richard Weinberger]

dump_stack() is used by modules and needs to be exported.

Signed-off-by: Richard Weinberger <richard@nod.at>
---
 arch/openrisc/kernel/traps.c |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/arch/openrisc/kernel/traps.c b/arch/openrisc/kernel/traps.c
index a4ec44a..ab68ffc 100644
--- a/arch/openrisc/kernel/traps.c
+++ b/arch/openrisc/kernel/traps.c
@@ -115,6 +115,7 @@ void dump_stack(void)
 
 	show_stack(current, &stack);
 }
+EXPORT_SYMBOL(dump_stack);
 
 void show_registers(struct pt_regs *regs)
 {
-- 
1.7.7.3

Re: [PATCH 2/5] OpenRISC: Use generic bitops.h

[Jonas Bonn]

[-- Attachment #1: Type: text/plain, Size: 916 bytes --]

On Fri, 2012-03-02 at 01:55 +0100, Richard Weinberger wrote:
> There is no need to reinvent the wheel.

NAK... we have optimized versions for some of these operations so the
generic header isn't right for us.

Have a look at arch/openrisc/asm/include/bitops.h to see which functions
we override with accelerated versions.

/Jonas

> 
> Signed-off-by: Richard Weinberger <richard@nod.at>
> ---
>  arch/openrisc/include/asm/Kbuild |    1 +
>  1 files changed, 1 insertions(+), 0 deletions(-)
> 
> diff --git a/arch/openrisc/include/asm/Kbuild b/arch/openrisc/include/asm/Kbuild
> index 11162e6..a283165 100644
> --- a/arch/openrisc/include/asm/Kbuild
> +++ b/arch/openrisc/include/asm/Kbuild
> @@ -4,6 +4,7 @@ header-y += spr_defs.h
>  
>  generic-y += atomic.h
>  generic-y += auxvec.h
> +generic-y += bitops.h
>  generic-y += bitsperlong.h
>  generic-y += bug.h
>  generic-y += bugs.h


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

Re: [PATCH 3/5] OpenRISC: Export memory_start and memory_end

[Jonas Bonn]

[-- Attachment #1: Type: text/plain, Size: 1702 bytes --]

On Fri, 2012-03-02 at 01:55 +0100, Richard Weinberger wrote:
> At least virt_addr_valid() makes use of memory_end.

virt_addr_valid looks broken... it shouldn't be comparing a virtual
address against memory_end which is phys_addr_t.

/Jonas

> 
> Signed-off-by: Richard Weinberger <richard@nod.at>
> ---
>  arch/openrisc/include/asm/page.h |    4 ++--
>  arch/openrisc/kernel/setup.c     |    2 +-
>  2 files changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/arch/openrisc/include/asm/page.h b/arch/openrisc/include/asm/page.h
> index b041b34..4dfe477 100644
> --- a/arch/openrisc/include/asm/page.h
> +++ b/arch/openrisc/include/asm/page.h
> @@ -71,8 +71,8 @@ typedef struct page *pgtable_t;
>  #define __pgd(x)	((pgd_t) { (x) })
>  #define __pgprot(x)	((pgprot_t) { (x) })
>  
> -extern unsigned long memory_start;
> -extern unsigned long memory_end;
> +extern phys_addr_t memory_start;
> +extern phys_addr_t memory_end;
>  
>  #endif /* !__ASSEMBLY__ */
>  
> diff --git a/arch/openrisc/kernel/setup.c b/arch/openrisc/kernel/setup.c
> index bf5eba2..c993c92 100644
> --- a/arch/openrisc/kernel/setup.c
> +++ b/arch/openrisc/kernel/setup.c
> @@ -52,6 +52,7 @@
>  #include "vmlinux.h"
>  
>  char __initdata cmd_line[COMMAND_LINE_SIZE] = CONFIG_CMDLINE;
> +phys_addr_t memory_start, memory_end;
>  
>  static unsigned long __init setup_memory(void)
>  {
> @@ -59,7 +60,6 @@ static unsigned long __init setup_memory(void)
>  	unsigned long ram_start_pfn;
>  	unsigned long free_ram_start_pfn;
>  	unsigned long ram_end_pfn;
> -	phys_addr_t memory_start, memory_end;
>  	struct memblock_region *region;
>  
>  	memory_end = memory_start = 0;


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

Re: [PATCH 5/5] OpenRISC: Export dump_stack()

[Jonas Bonn]

[-- Attachment #1: Type: text/plain, Size: 786 bytes --]

Thanks.  Patches 1 and 5 from your series applied.

I already had patch 4, and comments have been sent for the others.

/Jonas

On Fri, 2012-03-02 at 01:55 +0100, Richard Weinberger wrote:
> dump_stack() is used by modules and needs to be exported.
> 
> Signed-off-by: Richard Weinberger <richard@nod.at>
> ---
>  arch/openrisc/kernel/traps.c |    1 +
>  1 files changed, 1 insertions(+), 0 deletions(-)
> 
> diff --git a/arch/openrisc/kernel/traps.c b/arch/openrisc/kernel/traps.c
> index a4ec44a..ab68ffc 100644
> --- a/arch/openrisc/kernel/traps.c
> +++ b/arch/openrisc/kernel/traps.c
> @@ -115,6 +115,7 @@ void dump_stack(void)
>  
>  	show_stack(current, &stack);
>  }
> +EXPORT_SYMBOL(dump_stack);
>  
>  void show_registers(struct pt_regs *regs)
>  {


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

Re: [PATCH 2/5] OpenRISC: Use generic bitops.h

[Richard Weinberger]

[-- Attachment #1: Type: text/plain, Size: 602 bytes --]

Am 02.03.2012 07:50, schrieb Jonas Bonn:
> On Fri, 2012-03-02 at 01:55 +0100, Richard Weinberger wrote:
>> There is no need to reinvent the wheel.
> 
> NAK... we have optimized versions for some of these operations so the
> generic header isn't right for us.

Sorry, I missed that.

> Have a look at arch/openrisc/asm/include/bitops.h to see which functions
> we override with accelerated versions.

Okay, but some operations are missing in your bitopts.h.
Like test_bit_le() or find_next_zero_bit_le().

#include <asm-generic/bitops/le.h>
Should to the job...

Thanks,
//richard


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 490 bytes --]

Re: [PATCH 3/5] OpenRISC: Export memory_start and memory_end

[Richard Weinberger]

[-- Attachment #1: Type: text/plain, Size: 390 bytes --]

Am 02.03.2012 07:51, schrieb Jonas Bonn:
> On Fri, 2012-03-02 at 01:55 +0100, Richard Weinberger wrote:
>> At least virt_addr_valid() makes use of memory_end.
> 
> virt_addr_valid looks broken... it shouldn't be comparing a virtual
> address against memory_end which is phys_addr_t.

Then it's time to implement it correctly and match OpenRISC's
semantics.

Thanks,
//richard


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 490 bytes --]

Re: [PATCH 3/5] OpenRISC: Export memory_start and memory_end

[Jonas Bonn]

On Fri, 2012-03-02 at 09:48 +0100, Richard Weinberger wrote:
> Am 02.03.2012 07:51, schrieb Jonas Bonn:
> > On Fri, 2012-03-02 at 01:55 +0100, Richard Weinberger wrote:
> >> At least virt_addr_valid() makes use of memory_end.
> > 
> > virt_addr_valid looks broken... it shouldn't be comparing a virtual
> > address against memory_end which is phys_addr_t.
> 
> Then it's time to implement it correctly and match OpenRISC's
> semantics.
> 

Give this a try...

http://git.openrisc.net/cgit.cgi/jonas/linux/commit/?id=30fbeb350cd26dc87477b9cadf0bfe132dcf5119

/Jonas