Re: [PATCH 4/5] [POWERPC] Fix kernel build errors for mpc8272ads and mpc8560ads

On Wed, 10 Jan 2007 14:58:31 -0600
Kumar Gala wrote:

On Jan 10, 2007, at 2:50 PM, Vitaly Bordug wrote:

quoted

On Wed, 10 Jan 2007 11:11:05 -0600
Kumar Gala wrote:

[snip]

quoted

quoted

quoted

b/include/asm-powerpc/fs_pd.h index 3d0e819..2ba51b9 100644

--- a/include/asm-powerpc/fs_pd.h
+++ b/include/asm-powerpc/fs_pd.h

@@ -25,6 +25,12 @@ static inline int uart_clock(void)
         return ppc_proc_freq;
 }

+#if defined(CONFIG_8260)
+#include <asm/mpc8260.h>
+#elif defined(CONFIG_85xx)
+#include <asm/mpc85xx.h>
+#endif
+

what do we need these includes for?

This is to keep ppc stuff still working mostly. Basically, whole
this header is for such aim.
To be specific, both fs_enet and cpm_uart will bail the
compilation out
failing to find CPM_MAP_ADDR:

drivers/net/fs_enet/fs_enet-main.c:1144: error: 'CPM_MAP_ADDR'
undeclared (first use in this function)

and so forth.

How about wrapping it a ifndef CONFIG_PPC_MERGE so we know its just  
for that and can get ride of it in the future.

I afraid it will make arch/ppc unhappy - these got included to both.
And the headers required because there used to be same asm/io.h where this stuff used to reside, and from 
the recent time, arch/powerpc has separate io.h that made all the CPM2-related kinda hosed.

As a good thing, this header is only for the time we'll have to have same target working both in ppc and powerpc.
I think this can be revisited after the patch about cleaning-up CPM_MAP_ADDR from the drivers and core code.

quoted

quoted

quoted

 #define
cpm2_map(member)						\
({
\ u32 offset = offsetof(cpm2_map_t, member);
\

- k