Patchwork [U-Boot] corenet: Disable video on P2020DS

login
register
mail settings
Submitter Andy Fleming
Date Jan. 25, 2013, 4:38 p.m.
Message ID <1359131881-29799-1-git-send-email-afleming@freescale.com>
Download mbox | patch
Permalink /patch/215809/
State Accepted
Delegated to: Andy Fleming
Headers show

Comments

Andy Fleming - Jan. 25, 2013, 4:38 p.m.
The P2020DS build had grown too large, and video support isn't enabled
in almost any other Freescale board. Disabling it allows us to keep
building, and provides options for reenabling it later.

Signed-off-by: Andy Fleming <afleming@freescale.com>
---
 include/configs/P2020DS.h |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
Tom Rini - Jan. 25, 2013, 6:50 p.m.
On Fri, Jan 25, 2013 at 10:38:01AM -0600, Andy Fleming wrote:

> The P2020DS build had grown too large, and video support isn't enabled
> in almost any other Freescale board. Disabling it allows us to keep
> building, and provides options for reenabling it later.
> 
> Signed-off-by: Andy Fleming <afleming@freescale.com>

Now we may start having dead code around, yes?  Can you perhaps get away
with making this be disable video or something else and add a
P2020DS_video boards.cfg entry or similar?  Thanks!
Scott Wood - Jan. 25, 2013, 7:25 p.m.
On 01/25/2013 12:50:59 PM, Tom Rini wrote:
> On Fri, Jan 25, 2013 at 10:38:01AM -0600, Andy Fleming wrote:
> 
> > The P2020DS build had grown too large, and video support isn't  
> enabled
> > in almost any other Freescale board. Disabling it allows us to keep
> > building, and provides options for reenabling it later.
> >
> > Signed-off-by: Andy Fleming <afleming@freescale.com>
> 
> Now we may start having dead code around, yes?  Can you perhaps get  
> away
> with making this be disable video or something else and add a
> P2020DS_video boards.cfg entry or similar?  Thanks!

There are already 5 P2020DS targets, and there *should* be 8 (why is  
there no 36BIT version of DDR2, SDCARD, or SPIFLASH?).  This would  
expand it to 16.  Ideally we would have something like kconfig, but  
until then I don't see a reasonable alternative to saying that certain  
config symbols are user-settable by tweaking the board config file.

-Scott
Tom Rini - Jan. 25, 2013, 7:35 p.m.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 01/25/2013 02:25 PM, Scott Wood wrote:
> On 01/25/2013 12:50:59 PM, Tom Rini wrote:
>> On Fri, Jan 25, 2013 at 10:38:01AM -0600, Andy Fleming wrote:
>> 
>>> The P2020DS build had grown too large, and video support isn't
>>>  enabled in almost any other Freescale board. Disabling it 
>>> allows us to keep building, and provides options for reenabling
>>> it later.
>>> 
>>> Signed-off-by: Andy Fleming <afleming@freescale.com>
>> 
>> Now we may start having dead code around, yes?  Can you perhaps 
>> get away with making this be disable video or something else and
>>  add a P2020DS_video boards.cfg entry or similar?  Thanks!
> 
> There are already 5 P2020DS targets, and there *should* be 8 (why 
> is there no 36BIT version of DDR2, SDCARD, or SPIFLASH?).  This 
> would expand it to 16.  Ideally we would have something like 
> kconfig, but until then I don't see a reasonable alternative to 
> saying that certain config symbols are user-settable by tweaking 
> the board config file.

That's fine, in general.  But does this patch now leave us with
non-build testing video code?  That way lies bitrot, so yes, please
add a 6th target so that when someone needs to hand tweak their
P2020DS setup for this, not that, yes this and not that, oh and video,
they can have some confidence the code still builds.  Or say that
P1020/1022 having video on still too means the code in question is
still used.  That would also be fine.  Thanks.

- -- 
Tom
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/
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=ZCH1
-----END PGP SIGNATURE-----
Andy Fleming - Jan. 25, 2013, 8:43 p.m.
On Jan 25, 2013, at 12:50 PM, Tom Rini wrote:

> On Fri, Jan 25, 2013 at 10:38:01AM -0600, Andy Fleming wrote:
> 
>> The P2020DS build had grown too large, and video support isn't enabled
>> in almost any other Freescale board. Disabling it allows us to keep
>> building, and provides options for reenabling it later.
>> 
>> Signed-off-by: Andy Fleming <afleming@freescale.com>
> 
> Now we may start having dead code around, yes?  Can you perhaps get away
> with making this be disable video or something else and add a
> P2020DS_video boards.cfg entry or similar?  Thanks!

There doesn't appear to be any 2020-specific code in drivers/video/. The truth is, I doubt the code gets more than compile-tested by the setting being there. Also, as noted, P1022 defines it, in addition to MPC8536, MPC8544, MPC8572, MPC8610, and MPC8641. I'd rather not hack up a "video" config just for P2020, if possible.

Andy
Tom Rini - Jan. 25, 2013, 8:48 p.m.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 01/25/2013 03:43 PM, Fleming Andy-AFLEMING wrote:
> 
> On Jan 25, 2013, at 12:50 PM, Tom Rini wrote:
> 
>> On Fri, Jan 25, 2013 at 10:38:01AM -0600, Andy Fleming wrote:
>> 
>>> The P2020DS build had grown too large, and video support isn't
>>>  enabled in almost any other Freescale board. Disabling it 
>>> allows us to keep building, and provides options for reenabling
>>> it later.
>>> 
>>> Signed-off-by: Andy Fleming <afleming@freescale.com>
>> 
>> Now we may start having dead code around, yes?  Can you perhaps 
>> get away with making this be disable video or something else and
>>  add a P2020DS_video boards.cfg entry or similar?  Thanks!
> 
> There doesn't appear to be any 2020-specific code in 
> drivers/video/. The truth is, I doubt the code gets more than 
> compile-tested by the setting being there. Also, as noted, P1022 
> defines it, in addition to MPC8536, MPC8544, MPC8572, MPC8610, and
>  MPC8641. I'd rather not hack up a "video" config just for P2020, 
> if possible.

That's all I was looking for, that yes, we don't now have very dead
code there.  Thanks! (and I kicked off my loop of testing on the PR
that includes this change a while ago, still spinning over everything).

- -- 
Tom
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/
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=C7Yu
-----END PGP SIGNATURE-----

Patch

diff --git a/include/configs/P2020DS.h b/include/configs/P2020DS.h
index 0cc5781..a975ee1 100644
--- a/include/configs/P2020DS.h
+++ b/include/configs/P2020DS.h
@@ -490,7 +490,7 @@ 
 #define VIDEO_IO_OFFSET		CONFIG_SYS_PCIE1_IO_VIRT
 
 /* video */
-#define CONFIG_VIDEO
+#undef CONFIG_VIDEO
 
 #if defined(CONFIG_VIDEO)
 #define CONFIG_BIOSEMU