[U-Boot] [PATCH 1/1] SPARC: fixes exported function stub for standalone applications.

[U-Boot] [PATCH 1/1] SPARC: fixes exported function stub for standalone applications.

[Daniel Hellstrom]

Hello Wolfgang,

Please pull the u-boot-sparc.git master branch.

This patch fixes the SPARC support for standalone u-boot applications. 
The problem was that I neve finished the implementation in the first
place.

jmp ensures we get back to the location we came from, size(void *) make
sure we get the function addresses correctly from the table (entry*4 
instead of entry*1).


-------------------- SPARC/LEON3 --------------------
=> tftpboot 0x41000000 192.168.0.47:hello_world.bin
Using GRETH 10/100 device
TFTP from server 192.168.0.47; our IP address is 192.168.0.206
Filename 'hello_world.bin'.
Load address: 0x41000000
Loading: #
done
Bytes transferred = 1232 (4d0 hex)
=> go 41000008 hej
## Starting application at 0x41000008 ...
Example expects ABI version 5
Actual U-Boot ABI version 5
Hello World
argc = 2
argv[0] = "41000008"
argv[1] = "hej"
argv[2] = "<NULL>"
Hit any key to exit ...

## Application terminated, rc = 0x0
=>
-----------------------------------------------------


Best Regards,
Daniel Hellstrom






Signed-off-by: Daniel Hellstrom <daniel@gaisler.com>
---
 examples/standalone/stubs.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/examples/standalone/stubs.c b/examples/standalone/stubs.c
index 339bbf9..ce3371d 100644
--- a/examples/standalone/stubs.c
+++ b/examples/standalone/stubs.c
@@ -181,9 +181,9 @@ gd_t *global_data;
 "	or %%g1, %%g7, %%g1\n"				\
 "	ld [%%g1], %%g1\n"				\
 "	ld [%%g1 + %1], %%g1\n"				\
-"	call %%g1\n"					\
+"	jmp %%g1\n"					\
 "	nop\n"						\
-	: : "i"(offsetof(gd_t, jt)), "i"(XF_ ## x) : "g1" );
+	: : "i"(offsetof(gd_t, jt)), "i"(XF_ ## x * sizeof(void *)) : "g1" );
 
 #else
 #error stubs definition missing for this architecture
-- 
1.5.4

[U-Boot] [PATCH 1/1] SPARC: fixes exported function stub for standalone applications.

[Wolfgang Denk]

Dear Daniel Hellstrom,

In message <1253270387-15741-1-git-send-email-daniel@gaisler.com> you wrote:
> Hello Wolfgang,
> 
> Please pull the u-boot-sparc.git master branch.
> 
> This patch fixes the SPARC support for standalone u-boot applications. 
> The problem was that I neve finished the implementation in the first
> place.
> 
> jmp ensures we get back to the location we came from, size(void *) make
> sure we get the function addresses correctly from the table (entry*4 
> instead of entry*1).

Um...
...
> 
> Signed-off-by: Daniel Hellstrom <daniel@gaisler.com>
> ---
>  examples/standalone/stubs.c |    4 ++--
>  1 files changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/examples/standalone/stubs.c b/examples/standalone/stubs.c
> index 339bbf9..ce3371d 100644
> --- a/examples/standalone/stubs.c
> +++ b/examples/standalone/stubs.c
> @@ -181,9 +181,9 @@ gd_t *global_data;
>  "	or %%g1, %%g7, %%g1\n"				\
>  "	ld [%%g1], %%g1\n"				\
>  "	ld [%%g1 + %1], %%g1\n"				\
> -"	call %%g1\n"					\
> +"	jmp %%g1\n"					\
>  "	nop\n"						\
> -	: : "i"(offsetof(gd_t, jt)), "i"(XF_ ## x) : "g1" );
> +	: : "i"(offsetof(gd_t, jt)), "i"(XF_ ## x * sizeof(void *)) : "g1" );

This looks 100% like the patch submitted by Sergey Mironov on 16 Sep,
yet I see no Signed-off-by: line from him, nor any credit or
reference.

This can't be right?

Best regards,

Wolfgang Denk

-- 
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
"He was so narrow minded he could see through  a  keyhole  with  both
eyes ..."

[U-Boot] [PATCH 1/1] SPARC: fixes exported function stub for standalone applications.

[Daniel Hellstrom]

Wolfgang Denk wrote:

>Dear Daniel Hellstrom,
>
>In message <1253270387-15741-1-git-send-email-daniel@gaisler.com> you wrote:
>  
>
>>Hello Wolfgang,
>>
>>Please pull the u-boot-sparc.git master branch.
>>
>>This patch fixes the SPARC support for standalone u-boot applications. 
>>The problem was that I neve finished the implementation in the first
>>place.
>>
>>jmp ensures we get back to the location we came from, size(void *) make
>>sure we get the function addresses correctly from the table (entry*4 
>>instead of entry*1).
>>    
>>
>
>Um...
>...
>  
>
>>Signed-off-by: Daniel Hellstrom <daniel@gaisler.com>
>>---
>> examples/standalone/stubs.c |    4 ++--
>> 1 files changed, 2 insertions(+), 2 deletions(-)
>>
>>diff --git a/examples/standalone/stubs.c b/examples/standalone/stubs.c
>>index 339bbf9..ce3371d 100644
>>--- a/examples/standalone/stubs.c
>>+++ b/examples/standalone/stubs.c
>>@@ -181,9 +181,9 @@ gd_t *global_data;
>> "	or %%g1, %%g7, %%g1\n"				\
>> "	ld [%%g1], %%g1\n"				\
>> "	ld [%%g1 + %1], %%g1\n"				\
>>-"	call %%g1\n"					\
>>+"	jmp %%g1\n"					\
>> "	nop\n"						\
>>-	: : "i"(offsetof(gd_t, jt)), "i"(XF_ ## x) : "g1" );
>>+	: : "i"(offsetof(gd_t, jt)), "i"(XF_ ## x * sizeof(void *)) : "g1" );
>>    
>>
>
>This looks 100% like the patch submitted by Sergey Mironov on 16 Sep,
>yet I see no Signed-off-by: line from him, nor any credit or
>reference.
>
>This can't be right?
>  
>
Hello Wolfgang,

It is 100% his patch, I have only reviewed and tried it.

I admit it looks disgusting.

I will ask Sergey for a new patch with a signoff line, I will let you 
know when the repository is updated. I suppose I will use the "git am 
--signoff --whitespace=strip" you have written about on the git help page.

How do you deal with cyrillic in the sign-off line?

Best Regards,
Daniel Hellstrom