[U-Boot] [PATCH] net: Fix faulty definition of uec_initialize()

[U-Boot] [PATCH] net: Fix faulty definition of uec_initialize()

[Joakim Tjernlund]

The correct definition is in drivers/qe/uec.h so just
include that one instead.

Signed-off-by: Joakim Tjernlund <Joakim.Tjernlund@transmode.se>
---

 This is only compile tested on my board but hopefully
 it will do. If not, I urge Freescale to clean up this mess.

 include/netdev.h |    4 +---
 1 files changed, 1 insertions(+), 3 deletions(-)

diff --git a/include/netdev.h b/include/netdev.h
index 882642a..0873baf 100644
--- a/include/netdev.h
+++ b/include/netdev.h
@@ -83,12 +83,10 @@ int skge_initialize(bd_t *bis);
 int smc911x_initialize(u8 dev_num, int base_addr);
 int smc91111_initialize(u8 dev_num, int base_addr);
 int tsi108_eth_initialize(bd_t *bis);
-int uec_initialize(int index);
-int uec_standard_init(bd_t *bis);
 int uli526x_initialize(bd_t *bis);
 int sh_eth_initialize(bd_t *bis);
 int dm9000_initialize(bd_t *bis);
-
+#include "../drivers/qe/uec.h"
 /* Boards with PCI network controllers can call this from their board_eth_init()
  * function to initialize whatever's on board.
  * Return value is total # of devices found */
-- 
1.7.1

[U-Boot] [PATCH] net: Fix faulty definition of uec_initialize()

[Mike Frysinger]

On Mon, Aug 9, 2010 at 3:15 PM, Joakim Tjernlund wrote:
> The correct definition is in drivers/qe/uec.h so just
> include that one instead.

... except that bleeds a ton of crap from uec.h into the global
namespace for everyone.  better would be to move the prototype out of
uec.h and into netdev.h and make sure uec.h includes netdev.h.
-mike

[U-Boot] [PATCH] net: Fix faulty definition of uec_initialize()

[Ben Warren]

  On 8/9/2010 1:18 PM, Mike Frysinger wrote:
> On Mon, Aug 9, 2010 at 3:15 PM, Joakim Tjernlund wrote:
>> The correct definition is in drivers/qe/uec.h so just
>> include that one instead.
> ... except that bleeds a ton of crap from uec.h into the global
> namespace for everyone.  better would be to move the prototype out of
> uec.h and into netdev.h and make sure uec.h includes netdev.h.
> -mike
> _______
Agreed.  Please find a way to have UEC-related stuff only on boards that 
care about it.

regards,
Ben

[U-Boot] [PATCH] net: Fix faulty definition of uec_initialize()

[Joakim Tjernlund]

Ben Warren <biggerbadderben@gmail.com> wrote on 2010/08/09 22:54:10:
>
>   On 8/9/2010 1:18 PM, Mike Frysinger wrote:
> > On Mon, Aug 9, 2010 at 3:15 PM, Joakim Tjernlund wrote:
> >> The correct definition is in drivers/qe/uec.h so just
> >> include that one instead.
> > ... except that bleeds a ton of crap from uec.h into the global
> > namespace for everyone.  better would be to move the prototype out of
> > uec.h and into netdev.h and make sure uec.h includes netdev.h.

hmm, then the public include file for UEC users become "../drivers/uec.h"
while every one else uses netdev.h, doesn't seem the right way go

BTW, UEC users needs the macro STD_UEC_INFO() too.

> > -mike
> > _______
> Agreed.  Please find a way to have UEC-related stuff only on boards that
> care about it.

OK, but I will let Freescale sort this out. I just wanted to fix
the compile breakage I noticed while upgrading my board to latest
u-boot(still in progress)

[U-Boot] [PATCH] net: Fix faulty definition of uec_initialize()

[Mike Frysinger]

On Tue, Aug 10, 2010 at 1:52 AM, Joakim Tjernlund wrote:
> Ben Warren wrote on 2010/08/09 22:54:10:
>> ?On 8/9/2010 1:18 PM, Mike Frysinger wrote:
>> > On Mon, Aug 9, 2010 at 3:15 PM, Joakim Tjernlund wrote:
>> >> The correct definition is in drivers/qe/uec.h so just
>> >> include that one instead.
>> > ... except that bleeds a ton of crap from uec.h into the global
>> > namespace for everyone. ?better would be to move the prototype out of
>> > uec.h and into netdev.h and make sure uec.h includes netdev.h.
>
> hmm, then the public include file for UEC users become "../drivers/uec.h"
> while every one else uses netdev.h, doesn't seem the right way go

which isnt uncommon today.  simply grep the tree for
"include.*drivers/net/".  while i agree it isnt exactly a great idea,
there is no clean standard today in u-boot for drivers to expose their
internals to random consumers.  some drivers dump their header files
into include/, so personally, i prefer the include ../driver/net/..
style over that.

> BTW, UEC users needs the macro STD_UEC_INFO() too.

emphasis on "UEC users" and not "everyone".
-mike

[U-Boot] [PATCH] net: Fix faulty definition of uec_initialize()

[Joakim Tjernlund]

vapierfilter at gmail.com wrote on 2010/08/10 08:04:02:
>
> On Tue, Aug 10, 2010 at 1:52 AM, Joakim Tjernlund wrote:
> > Ben Warren wrote on 2010/08/09 22:54:10:
> >> ?On 8/9/2010 1:18 PM, Mike Frysinger wrote:
> >> > On Mon, Aug 9, 2010 at 3:15 PM, Joakim Tjernlund wrote:
> >> >> The correct definition is in drivers/qe/uec.h so just
> >> >> include that one instead.
> >> > ... except that bleeds a ton of crap from uec.h into the global
> >> > namespace for everyone. ?better would be to move the prototype out of
> >> > uec.h and into netdev.h and make sure uec.h includes netdev.h.
> >
> > hmm, then the public include file for UEC users become "../drivers/uec.h"
> > while every one else uses netdev.h, doesn't seem the right way go
>
> which isnt uncommon today.  simply grep the tree for
> "include.*drivers/net/".  while i agree it isnt exactly a great idea,
> there is no clean standard today in u-boot for drivers to expose their
> internals to random consumers.  some drivers dump their header files
> into include/, so personally, i prefer the include ../driver/net/..
> style over that.

I see, but that defeats the purpose of netdev.h I think. Oh well,
just gonna have to live with that then.

 Jocke

[U-Boot] [PATCHv2] net: Fix faulty definition of uec_initialize()

[Joakim Tjernlund]

The correct definition is in drivers/qe/uec.h so just
remove this one.

Signed-off-by: Joakim Tjernlund <Joakim.Tjernlund@transmode.se>
---
 include/netdev.h |    1 -
 1 files changed, 0 insertions(+), 1 deletions(-)

diff --git a/include/netdev.h b/include/netdev.h
index 882642a..65833e2 100644
--- a/include/netdev.h
+++ b/include/netdev.h
@@ -83,7 +83,6 @@ int skge_initialize(bd_t *bis);
 int smc911x_initialize(u8 dev_num, int base_addr);
 int smc91111_initialize(u8 dev_num, int base_addr);
 int tsi108_eth_initialize(bd_t *bis);
-int uec_initialize(int index);
 int uec_standard_init(bd_t *bis);
 int uli526x_initialize(bd_t *bis);
 int sh_eth_initialize(bd_t *bis);
-- 
1.7.1

[U-Boot] [PATCHv2] net: Fix faulty definition of uec_initialize()

[Joakim Tjernlund]

The correct definition is in drivers/qe/uec.h so just
remove this one.

Signed-off-by: Joakim Tjernlund <Joakim.Tjernlund@transmode.se>
---
 include/netdev.h |    1 -
 1 files changed, 0 insertions(+), 1 deletions(-)

diff --git a/include/netdev.h b/include/netdev.h
index 882642a..65833e2 100644
--- a/include/netdev.h
+++ b/include/netdev.h
@@ -83,7 +83,6 @@ int skge_initialize(bd_t *bis);
 int smc911x_initialize(u8 dev_num, int base_addr);
 int smc91111_initialize(u8 dev_num, int base_addr);
 int tsi108_eth_initialize(bd_t *bis);
-int uec_initialize(int index);
 int uec_standard_init(bd_t *bis);
 int uli526x_initialize(bd_t *bis);
 int sh_eth_initialize(bd_t *bis);
-- 
1.7.1

[U-Boot] [PATCHv2] net: Fix faulty definition of uec_initialize()

[Ben Warren]

  Hi Jocke,

On 8/19/2010 12:37 AM, Joakim Tjernlund wrote:
> The correct definition is in drivers/qe/uec.h so just
> remove this one.
>
> Signed-off-by: Joakim Tjernlund<Joakim.Tjernlund@transmode.se>
> ---
>   include/netdev.h |    1 -
>   1 files changed, 0 insertions(+), 1 deletions(-)
>
> diff --git a/include/netdev.h b/include/netdev.h
> index 882642a..65833e2 100644
> --- a/include/netdev.h
> +++ b/include/netdev.h
> @@ -83,7 +83,6 @@ int skge_initialize(bd_t *bis);
>   int smc911x_initialize(u8 dev_num, int base_addr);
>   int smc91111_initialize(u8 dev_num, int base_addr);
>   int tsi108_eth_initialize(bd_t *bis);
> -int uec_initialize(int index);
>   int uec_standard_init(bd_t *bis);
>   int uli526x_initialize(bd_t *bis);
>   int sh_eth_initialize(bd_t *bis);
Applied to net/next

regards,
Ben