Sync SiByte system code to the new DUART driver

Sync SiByte system code to the new DUART driver

[Imre Kaloz]

The new upstream SiByte DUART driver uses a different config option then
the old one, so the SiByte target doesn't compile currently.
This patch fixes the problem.

Signed-off-by: Imre Kaloz <kaloz@openwrt.org>

---

diff --git a/arch/mips/sibyte/cfe/console.c b/arch/mips/sibyte/cfe/console.c
index c6ec748..4cec9d7 100644
--- a/arch/mips/sibyte/cfe/console.c
+++ b/arch/mips/sibyte/cfe/console.c
@@ -46,7 +46,7 @@ static int cfe_console_setup(struct console *cons, char *str)
  	/* XXXKW think about interaction with 'console=' cmdline arg */
  	/* If none of the console options are configured, the build will break. */
  	if (cfe_getenv("BOOT_CONSOLE", consdev, 32) >= 0) {
-#ifdef CONFIG_SIBYTE_SB1250_DUART
+#ifdef CONFIG_SERIAL_SB1250_DUART
  		if (!strcmp(consdev, "uart0")) {
  			setleds("u0cn");
  		} else if (!strcmp(consdev, "uart1")) {

Re: Sync SiByte system code to the new DUART driver

[Ralf Baechle]

On Thu, Aug 02, 2007 at 12:58:50PM +0200, Imre Kaloz wrote:

> The new upstream SiByte DUART driver uses a different config option then
> the old one, so the SiByte target doesn't compile currently.
> This patch fixes the problem.
> 
> Signed-off-by: Imre Kaloz <kaloz@openwrt.org>

Hmm...  You just found the tip of the iceberg ...

$ git grep -w CONFIG_SIBYTE_SB1250_DUART | cat -
arch/mips/configs/bigsur_defconfig:CONFIG_SIBYTE_SB1250_DUART=y
arch/mips/configs/sb1250-swarm_defconfig:CONFIG_SIBYTE_SB1250_DUART=y
arch/mips/sibyte/bcm1480/irq.c:#ifdef CONFIG_SIBYTE_SB1250_DUART
arch/mips/sibyte/bcm1480/irq.c:#ifdef CONFIG_SIBYTE_SB1250_DUART
arch/mips/sibyte/cfe/console.c:#ifdef CONFIG_SIBYTE_SB1250_DUART
arch/mips/sibyte/sb1250/irq.c:#ifdef CONFIG_SIBYTE_SB1250_DUART
arch/mips/sibyte/sb1250/irq.c:#ifdef CONFIG_SIBYTE_SB1250_DUART

  Ralf

Re: Sync SiByte system code to the new DUART driver

[Maciej W. Rozycki]

On Thu, 2 Aug 2007, Ralf Baechle wrote:

> Hmm...  You just found the tip of the iceberg ...
> 
> $ git grep -w CONFIG_SIBYTE_SB1250_DUART | cat -
> arch/mips/configs/bigsur_defconfig:CONFIG_SIBYTE_SB1250_DUART=y
> arch/mips/configs/sb1250-swarm_defconfig:CONFIG_SIBYTE_SB1250_DUART=y
> arch/mips/sibyte/bcm1480/irq.c:#ifdef CONFIG_SIBYTE_SB1250_DUART
> arch/mips/sibyte/bcm1480/irq.c:#ifdef CONFIG_SIBYTE_SB1250_DUART
> arch/mips/sibyte/cfe/console.c:#ifdef CONFIG_SIBYTE_SB1250_DUART
> arch/mips/sibyte/sb1250/irq.c:#ifdef CONFIG_SIBYTE_SB1250_DUART
> arch/mips/sibyte/sb1250/irq.c:#ifdef CONFIG_SIBYTE_SB1250_DUART

 Most of that stuff does not work anymore anyway and now you risk link 
errors. ;-)  I do not think there is place for driver-related #ifdefs 
under arch/ anyway, the answer being platform devices.  Though chances are 
nobody might be bothered to ever implement them here.

 Also the #ifdefs in arch/mips/sibyte/cfe/console.c do not make sense to 
me.

  Maciej

Re: Sync SiByte system code to the new DUART driver

[Ralf Baechle]

On Thu, Aug 02, 2007 at 12:26:22PM +0100, Maciej W. Rozycki wrote:

>  Most of that stuff does not work anymore anyway and now you risk link 
> errors. ;-)  I do not think there is place for driver-related #ifdefs 
> under arch/ anyway, the answer being platform devices.  Though chances are 
> nobody might be bothered to ever implement them here.
> 
>  Also the #ifdefs in arch/mips/sibyte/cfe/console.c do not make sense to 
> me.

Unfortunately you're right ...

  Ralf

Re: Sync SiByte system code to the new DUART driver

[Imre Kaloz]

On Thu, 02 Aug 2007 13:26:22 +0200, Maciej W. Rozycki  
<macro@linux-mips.org> wrote:

>  Most of that stuff does not work anymore anyway and now you risk link
> errors. ;-)  I do not think there is place for driver-related #ifdefs
> under arch/ anyway, the answer being platform devices.  Though chances  
> are
> nobody might be bothered to ever implement them here.
>
>  Also the #ifdefs in arch/mips/sibyte/cfe/console.c do not make sense to
> me.
>


True, but as I've said Ralf on IRC, I was fixing a bug rather then doing  
an overhaul. My TODO list for today is too long to save the world, but  
maybe I'll have time for that tomorrow ;)
I think the whole sibyte target should be cleaned up and changed to  
support runtime board detection like we do it on bcm47xx/53xx and  
ar231x/5312 in OpenWrt.


Imre