[PATCH] Fix 8xx compile errors

[PATCH] Fix 8xx compile errors

[John Tyner]

[-- Attachment #1: Type: text/plain, Size: 518 bytes --]

Building for 8xx fails to compile due to errors in a couple of places. 
The first is due to the casting of an lvalue (if I remember correctly), 
and the second was due to the cpmp variable being declared static even 
though the headers previously defined it as extern. The following patch 
corrects these errors. The patch is against 2.4 since that's what I'm 
working with. (I've been unable to get 2.6 to run properly on my 
hardware so far.)

Please CC me on any responses since I'm not subscribed.

Thanks,
John



[-- Attachment #2: 8xx_compile.patch --]
[-- Type: text/plain, Size: 1125 bytes --]

diff -ruNa linux-2.4.35.4.orig/arch/ppc/8xx_io/uart.c linux-2.4.35.4/arch/ppc/8xx_io/uart.c
--- linux-2.4.35.4.orig/arch/ppc/8xx_io/uart.c	2007-11-17 09:23:15.000000000 -0800
+++ linux-2.4.35.4/arch/ppc/8xx_io/uart.c	2007-11-27 11:28:09.000000000 -0800
@@ -2292,7 +2292,8 @@
 
 		/* Get the address of the host memory buffer.*/
 		info = &consinfo;
-		info->tx_bd_base = (cbd_t *)bdbase = (cbd_t *)&cpmp->cp_dpmem[up->smc_tbase];
+		bdbase = (cbd_t *)&cpmp->cp_dpmem[up->smc_tbase];
+		info->tx_bd_base = (cbd_t *)bdbase;
 		info->tx_cur = (cbd_t *)bdbase;
 	}
 	max_tx_size = console_tx_buf_len;
diff -ruNa linux-2.4.35.4.orig/arch/ppc/boot/simple/m8xx_tty.c linux-2.4.35.4/arch/ppc/boot/simple/m8xx_tty.c
--- linux-2.4.35.4.orig/arch/ppc/boot/simple/m8xx_tty.c	2007-11-17 09:23:15.000000000 -0800
+++ linux-2.4.35.4/arch/ppc/boot/simple/m8xx_tty.c	2007-11-27 11:28:42.000000000 -0800
@@ -30,7 +30,7 @@
 #define SMC_INDEX	0
 #endif
 
-static cpm8xx_t	*cpmp = (cpm8xx_t *)&(((immap_t *)IMAP_ADDR)->im_cpm);
+cpm8xx_t	*cpmp = (cpm8xx_t *)&(((immap_t *)IMAP_ADDR)->im_cpm);
 
 unsigned long
 serial_init(int ignored, bd_t *bd)

Re: [PATCH] Fix 8xx compile errors

[Vitaly Bordug]

On Mon, 3 Dec 2007 12:58:38 -0800
John Tyner wrote:

> Building for 8xx fails to compile due to errors in a couple of
> places. The first is due to the casting of an lvalue (if I remember
> correctly), and the second was due to the cpmp variable being
> declared static even though the headers previously defined it as
> extern. The following patch corrects these errors. The patch is
> against 2.4 since that's what I'm working with. (I've been unable to
> get 2.6 to run properly on my hardware so far.)
> 
> Please CC me on any responses since I'm not subscribed.
> 
You don't need these quirks with 2.6 and arch/powerpc... It does not use 8xx_io
and arch/powerpc/boot/* having a bit different meaning.

What is your platform btw? Adding something to 8xx camp should not be huge effort with powerpc... 

-- 
Sincerely, Vitaly