[PATCH] bcm1480 serial build fix

[PATCH] bcm1480 serial build fix

[Thiemo Seufer]

The appended patch restores serial functionality for the bcm1480.

I glued this together without reading documentation, so I'm not sure if
it is fully correct. It is good enough to build a kernel and have a
working serial console.


Signed-Off-By: Thiemo Seufer <ths@networkno.de>

diff --git a/drivers/serial/sb1250-duart.c b/drivers/serial/sb1250-duart.c
index 1d9d728..e7f5c0e 100644
--- a/drivers/serial/sb1250-duart.c
+++ b/drivers/serial/sb1250-duart.c
@@ -57,6 +57,12 @@
 #define SBD_CTRLREGS(line)	A_BCM1480_DUART_CTRLREG((line), 0)
 #define SBD_INT(line)		(K_BCM1480_INT_UART_0 + (line))
 
+#define DUART_CHANREG_SPACING	BCM1480_DUART_CHANREG_SPACING
+
+#define R_DUART_IMRREG(line)	R_BCM1480_DUART_IMRREG(line)
+#define R_DUART_INCHREG(line)	R_BCM1480_DUART_INCHREG(line)
+#define R_DUART_ISRREG(line)	R_BCM1480_DUART_ISRREG(line)
+
 #elif defined(CONFIG_SIBYTE_SB1250) || defined(CONFIG_SIBYTE_BCM112X)
 #include <asm/sibyte/sb1250_regs.h>
 #include <asm/sibyte/sb1250_int.h>
diff --git a/include/asm-mips/sibyte/bcm1480_regs.h b/include/asm-mips/sibyte/bcm1480_regs.h
index 2738c13..c34d36b 100644
--- a/include/asm-mips/sibyte/bcm1480_regs.h
+++ b/include/asm-mips/sibyte/bcm1480_regs.h
@@ -227,10 +227,15 @@
 	(A_BCM1480_DUART(chan) +					\
 	 BCM1480_DUART_CHANREG_SPACING * 3 + (reg))
 
+#define DUART_IMRISR_SPACING	    0x20
+#define DUART_INCHNG_SPACING	    0x10
+
 #define R_BCM1480_DUART_IMRREG(chan)					\
 	(R_DUART_IMR_A + ((chan) & 1) * DUART_IMRISR_SPACING)
 #define R_BCM1480_DUART_ISRREG(chan)					\
 	(R_DUART_ISR_A + ((chan) & 1) * DUART_IMRISR_SPACING)
+#define R_BCM1480_DUART_INCHREG(chan)					\
+	(R_DUART_IN_CHNG_A + ((chan) & 1) * DUART_INCHNG_SPACING)
 
 #define A_BCM1480_DUART_IMRREG(chan)					\
 	(A_BCM1480_DUART_CTRLREG((chan), R_BCM1480_DUART_IMRREG(chan)))

Re: [PATCH] bcm1480 serial build fix

[Maciej W. Rozycki]

On Sun, 22 Jul 2007, Thiemo Seufer wrote:

> diff --git a/include/asm-mips/sibyte/bcm1480_regs.h b/include/asm-mips/sibyte/bcm1480_regs.h
> index 2738c13..c34d36b 100644
> --- a/include/asm-mips/sibyte/bcm1480_regs.h
> +++ b/include/asm-mips/sibyte/bcm1480_regs.h
> @@ -227,10 +227,15 @@
>  	(A_BCM1480_DUART(chan) +					\
>  	 BCM1480_DUART_CHANREG_SPACING * 3 + (reg))
>  
> +#define DUART_IMRISR_SPACING	    0x20
> +#define DUART_INCHNG_SPACING	    0x10
> +

 Aren't all the bits in "bcm1480_regs.h" meant to be prefixed with 
BCM1480_DUART?  If these are to be the same as for the BCM1250, then they 
can probably be defined "in sb1250_regs.h" unconditionally.

 These headers are a horrible mess anyway -- a single definition should be 
enough to access the two DUARTs the BCM1480 seems to have...

  Maciej

Re: [PATCH] bcm1480 serial build fix

[Ralf Baechle]

On Mon, Jul 23, 2007 at 02:00:16PM +0100, Maciej W. Rozycki wrote:

>  Aren't all the bits in "bcm1480_regs.h" meant to be prefixed with 
> BCM1480_DUART?  If these are to be the same as for the BCM1250, then they 
> can probably be defined "in sb1250_regs.h" unconditionally.
> 
>  These headers are a horrible mess anyway -- a single definition should be 
> enough to access the two DUARTs the BCM1480 seems to have...

The headers are autogenerated by Broadcom; they got some header-o-matic
which from a common source base will generate headers for Linux and
several other operating systems by adding (C) headers etc.  Which makes
sorting them tricky ...

  Ralf

Re: [PATCH] bcm1480 serial build fix

[Maciej W. Rozycki]

On Mon, 23 Jul 2007, Ralf Baechle wrote:

> The headers are autogenerated by Broadcom; they got some header-o-matic
> which from a common source base will generate headers for Linux and
> several other operating systems by adding (C) headers etc.  Which makes
> sorting them tricky ...

 Well, I find it hardly an excuse...  Garbage In, Garbage Out.

  Maciej

Re: [PATCH] bcm1480 serial build fix

[Thiemo Seufer]

Maciej W. Rozycki wrote:
> On Sun, 22 Jul 2007, Thiemo Seufer wrote:
> 
> > diff --git a/include/asm-mips/sibyte/bcm1480_regs.h b/include/asm-mips/sibyte/bcm1480_regs.h
> > index 2738c13..c34d36b 100644
> > --- a/include/asm-mips/sibyte/bcm1480_regs.h
> > +++ b/include/asm-mips/sibyte/bcm1480_regs.h
> > @@ -227,10 +227,15 @@
> >  	(A_BCM1480_DUART(chan) +					\
> >  	 BCM1480_DUART_CHANREG_SPACING * 3 + (reg))
> >  
> > +#define DUART_IMRISR_SPACING	    0x20
> > +#define DUART_INCHNG_SPACING	    0x10
> > +
> 
>  Aren't all the bits in "bcm1480_regs.h" meant to be prefixed with 
> BCM1480_DUART?

Appatenly not, guessing from the header's contents.

> If these are to be the same as for the BCM1250, then they 
> can probably be defined "in sb1250_regs.h" unconditionally.
> 
>  These headers are a horrible mess anyway -- a single definition should be 
> enough to access the two DUARTs the BCM1480 seems to have...

Indeed. I just took the path of least resistance to make it work again.


Thiemo

Re: [PATCH] bcm1480 serial build fix

[Maciej W. Rozycki]

On Mon, 23 Jul 2007, Thiemo Seufer wrote:

> >  Aren't all the bits in "bcm1480_regs.h" meant to be prefixed with 
> > BCM1480_DUART?
> 
> Appatenly not, guessing from the header's contents.

 There are two or three exceptions buried indeed.

> >  These headers are a horrible mess anyway -- a single definition should be 
> > enough to access the two DUARTs the BCM1480 seems to have...
> 
> Indeed. I just took the path of least resistance to make it work again.

 Sure. :)  But it looks like you merely follow the tradition here...

  Maciej

Re: [PATCH] bcm1480 serial build fix

[Martin Michlmayr]

* Thiemo Seufer <ths@networkno.de> [2007-07-23 14:44]:
> >  These headers are a horrible mess anyway -- a single definition should be 
> > enough to access the two DUARTs the BCM1480 seems to have...
> Indeed. I just took the path of least resistance to make it work again.

Maybe Ralf can apply it (at least for now).  BCM1480 in git is
currently broken because of this.

From: Thiemo Seufer <ths@networkno.de>

Restore serial functionality for the bcm1480.

I glued this together without reading documentation, so I'm not sure if
it is fully correct. It is good enough to build a kernel and have a
working serial console.


Signed-Off-By: Thiemo Seufer <ths@networkno.de>

diff --git a/drivers/serial/sb1250-duart.c b/drivers/serial/sb1250-duart.c
index 1d9d728..e7f5c0e 100644
--- a/drivers/serial/sb1250-duart.c
+++ b/drivers/serial/sb1250-duart.c
@@ -57,6 +57,12 @@
 #define SBD_CTRLREGS(line)	A_BCM1480_DUART_CTRLREG((line), 0)
 #define SBD_INT(line)		(K_BCM1480_INT_UART_0 + (line))
 
+#define DUART_CHANREG_SPACING	BCM1480_DUART_CHANREG_SPACING
+
+#define R_DUART_IMRREG(line)	R_BCM1480_DUART_IMRREG(line)
+#define R_DUART_INCHREG(line)	R_BCM1480_DUART_INCHREG(line)
+#define R_DUART_ISRREG(line)	R_BCM1480_DUART_ISRREG(line)
+
 #elif defined(CONFIG_SIBYTE_SB1250) || defined(CONFIG_SIBYTE_BCM112X)
 #include <asm/sibyte/sb1250_regs.h>
 #include <asm/sibyte/sb1250_int.h>
diff --git a/include/asm-mips/sibyte/bcm1480_regs.h b/include/asm-mips/sibyte/bcm1480_regs.h
index 2738c13..c34d36b 100644
--- a/include/asm-mips/sibyte/bcm1480_regs.h
+++ b/include/asm-mips/sibyte/bcm1480_regs.h
@@ -227,10 +227,15 @@
 	(A_BCM1480_DUART(chan) +					\
 	 BCM1480_DUART_CHANREG_SPACING * 3 + (reg))
 
+#define DUART_IMRISR_SPACING	    0x20
+#define DUART_INCHNG_SPACING	    0x10
+
 #define R_BCM1480_DUART_IMRREG(chan)					\
 	(R_DUART_IMR_A + ((chan) & 1) * DUART_IMRISR_SPACING)
 #define R_BCM1480_DUART_ISRREG(chan)					\
 	(R_DUART_ISR_A + ((chan) & 1) * DUART_IMRISR_SPACING)
+#define R_BCM1480_DUART_INCHREG(chan)					\
+	(R_DUART_IN_CHNG_A + ((chan) & 1) * DUART_INCHNG_SPACING)
 
 #define A_BCM1480_DUART_IMRREG(chan)					\
 	(A_BCM1480_DUART_CTRLREG((chan), R_BCM1480_DUART_IMRREG(chan)))

-- 
Martin Michlmayr
http://www.cyrius.com/