diff mbox

[U-Boot] Allow the "reset" command to be omitted with CONFIG_CMD_RESET

Message ID 1324006361-25456-1-git-send-email-Kyle.D.Moffett@boeing.com
State Changes Requested
Headers show

Commit Message

Kyle Moffett Dec. 16, 2011, 3:32 a.m. UTC
This new #define is set in config_cmd_defaults.h (which is automatically
included on every board by "mkconfig"), but this allows boards to elect
to omit the "reset" command if necessary with "#undef CONFIG_CMD_RESET".

Signed-off-by: Kyle Moffett <Kyle.D.Moffett@boeing.com>
Cc: Wolfgang Denk <wd@denx.de>
Cc: Mike Frysinger <vapier@gentoo.org>
---
 README                        |    1 +
 common/cmd_boot.c             |    2 ++
 include/config_cmd_defaults.h |    1 +
 3 files changed, 4 insertions(+), 0 deletions(-)

Comments

Mike Frysinger Dec. 16, 2011, 5:05 a.m. UTC | #1
On Thursday 15 December 2011 22:32:41 Kyle Moffett wrote:
> This new #define is set in config_cmd_defaults.h (which is automatically
> included on every board by "mkconfig"), but this allows boards to elect
> to omit the "reset" command if necessary with "#undef CONFIG_CMD_RESET".

NAK: doesn't seem to address the feedback i posted last time ...
-mike
Kyle Moffett Dec. 16, 2011, 6:49 p.m. UTC | #2
On Dec 16, 2011, at 00:05, Mike Frysinger wrote:
> On Thursday 15 December 2011 22:32:41 Kyle Moffett wrote:
>> This new #define is set in config_cmd_defaults.h (which is automatically
>> included on every board by "mkconfig"), but this allows boards to elect
>> to omit the "reset" command if necessary with "#undef CONFIG_CMD_RESET".
> 
> NAK: doesn't seem to address the feedback i posted last time ...

Hmm, I thought I addressed these:

>> --- a/include/config_cmd_defaults.h
>> +++ b/include/config_cmd_defaults.h
> 
> updating these files is not sufficient and will break boards.  drop the hunks to 
> these files and update include/config_defaults.h instead.

The "mkconfig" program automatically includes both config_defaults.h and
config_cmd_defaults.h, so this is sufficient to not break boards.  This
is exactly how CONFIG_CMD_GO works.  Since it's always on, there's no need
to add it to config_cmd_all.h either (which is why I removed that hunk).

> would be a good time to split this into a dedicated common/cmd_reset.c,
> we can also then fix cmd_boot.c to only build when CONFIG_CMD_GO is enabled

The only references to CONFIG_CMD_GO are in config_cmd_defaults.h, README,
and common/cmd_boot.c, there is nothing that ever turns it off, so perhaps
the config symbol should just be removed?

Cheers,
Kyle Moffett
Mike Frysinger Dec. 16, 2011, 7:30 p.m. UTC | #3
On Friday 16 December 2011 13:49:15 Moffett, Kyle D wrote:
> On Dec 16, 2011, at 00:05, Mike Frysinger wrote:
> > On Thursday 15 December 2011 22:32:41 Kyle Moffett wrote:
> >> This new #define is set in config_cmd_defaults.h (which is automatically
> >> included on every board by "mkconfig"), but this allows boards to elect
> >> to omit the "reset" command if necessary with "#undef CONFIG_CMD_RESET".
> > 
> > NAK: doesn't seem to address the feedback i posted last time ...
> 
> Hmm, I thought I addressed these:
> >> --- a/include/config_cmd_defaults.h
> >> +++ b/include/config_cmd_defaults.h
> > 
> > updating these files is not sufficient and will break boards.  drop the
> > hunks to these files and update include/config_defaults.h instead.
> 
> The "mkconfig" program automatically includes both config_defaults.h and
> config_cmd_defaults.h, so this is sufficient to not break boards.

ah, you're right.  sorry about that.  i was thinking of "config_cmd_default.h" 
which is not the same as "config_cmd_defaults.h" ;).

> > would be a good time to split this into a dedicated common/cmd_reset.c,
> > we can also then fix cmd_boot.c to only build when CONFIG_CMD_GO is
> > enabled
> 
> The only references to CONFIG_CMD_GO are in config_cmd_defaults.h, README,
> and common/cmd_boot.c, there is nothing that ever turns it off, so perhaps
> the config symbol should just be removed?

i have some boards that turn it off
-mike
Kyle Moffett Dec. 16, 2011, 8:42 p.m. UTC | #4
On Dec 16, 2011, at 14:30, Mike Frysinger wrote:
> On Friday 16 December 2011 13:49:15 Moffett, Kyle D wrote:
>> On Dec 16, 2011, at 00:05, Mike Frysinger wrote:
>>> On Thursday 15 December 2011 22:32:41 Kyle Moffett wrote:
>>>> This new #define is set in config_cmd_defaults.h (which is automatically
>>>> included on every board by "mkconfig"), but this allows boards to elect
>>>> to omit the "reset" command if necessary with "#undef CONFIG_CMD_RESET".
>>> 
>>> would be a good time to split this into a dedicated common/cmd_reset.c,
>>> we can also then fix cmd_boot.c to only build when CONFIG_CMD_GO is
>>> enabled
>> 
>> The only references to CONFIG_CMD_GO are in config_cmd_defaults.h, README,
>> and common/cmd_boot.c, there is nothing that ever turns it off, so perhaps
>> the config symbol should just be removed?
> 
> i have some boards that turn it off

Ok,

I really don't see the point in moving these 5 lines to their own file,
especially since they are in is already just 78 lines long:
  U_BOOT_CMD(
          reset, 1, 0,    do_reset,
          "Perform RESET of the CPU",
          ""
  );

If you really think that's what I should do, though, I'll go ahead and
make that change.

Cheers,
Kyle Moffett

--
Curious about my work on the Debian powerpcspe port?
I'm keeping a blog here: http://pureperl.blogspot.com/
Mike Frysinger Dec. 16, 2011, 10:07 p.m. UTC | #5
On Friday 16 December 2011 15:42:26 Moffett, Kyle D wrote:
> On Dec 16, 2011, at 14:30, Mike Frysinger wrote:
> > On Friday 16 December 2011 13:49:15 Moffett, Kyle D wrote:
> >> On Dec 16, 2011, at 00:05, Mike Frysinger wrote:
> >>> On Thursday 15 December 2011 22:32:41 Kyle Moffett wrote:
> >>>> This new #define is set in config_cmd_defaults.h (which is
> >>>> automatically included on every board by "mkconfig"), but this allows
> >>>> boards to elect to omit the "reset" command if necessary with "#undef
> >>>> CONFIG_CMD_RESET".
> >>> 
> >>> would be a good time to split this into a dedicated common/cmd_reset.c,
> >>> we can also then fix cmd_boot.c to only build when CONFIG_CMD_GO is
> >>> enabled
> >> 
> >> The only references to CONFIG_CMD_GO are in config_cmd_defaults.h,
> >> README, and common/cmd_boot.c, there is nothing that ever turns it off,
> >> so perhaps the config symbol should just be removed?
> > 
> > i have some boards that turn it off
> 
> Ok,
> 
> I really don't see the point in moving these 5 lines to their own file,
> especially since they are in is already just 78 lines long:
>   U_BOOT_CMD(
>           reset, 1, 0,    do_reset,
>           "Perform RESET of the CPU",
>           ""
>   );
> 
> If you really think that's what I should do, though, I'll go ahead and
> make that change.

i think so, that way we can change cmd_boot.o in the Makefile to depend on 
CONFIG_CMD_GO and drop the ifdef logic from cmd_boot.c
-mike
diff mbox

Patch

diff --git a/README b/README
index ff72e47..90989db 100644
--- a/README
+++ b/README
@@ -793,6 +793,7 @@  The following options need to be configured:
 					  host
 		CONFIG_CMD_PORTIO	* Port I/O
 		CONFIG_CMD_REGINFO	* Register dump
+		CONFIG_CMD_RESET	  Reset the CPU
 		CONFIG_CMD_RUN		  run command in env variable
 		CONFIG_CMD_SAVES	* save S record dump
 		CONFIG_CMD_SCSI		* SCSI Support
diff --git a/common/cmd_boot.c b/common/cmd_boot.c
index 0afd939..f5779e9 100644
--- a/common/cmd_boot.c
+++ b/common/cmd_boot.c
@@ -71,8 +71,10 @@  U_BOOT_CMD(
 
 #endif
 
+#ifdef CONFIG_CMD_RESET
 U_BOOT_CMD(
 	reset, 1, 0,	do_reset,
 	"Perform RESET of the CPU",
 	""
 );
+#endif
diff --git a/include/config_cmd_defaults.h b/include/config_cmd_defaults.h
index a55b268..b65a945 100644
--- a/include/config_cmd_defaults.h
+++ b/include/config_cmd_defaults.h
@@ -14,5 +14,6 @@ 
 #define CONFIG_CMD_EXPORTENV 1
 #define CONFIG_CMD_GO 1
 #define CONFIG_CMD_IMPORTENV 1
+#define CONFIG_CMD_RESET 1
 
 #endif