PPC860 cachelines corrected for glibc, need to correct also in kernel ?

[Kari Hameenaho]

There was previous correction in this mailing list for cachelines in glibc
to be used in PPC860. 

>Ok, here's how to build glibc for embedded PPC:

>First you must remove the assumption that cachelines are 32 bytes:
>Apply this diff, and simply move sysdeps/powerpc/memset.S out of the
>way for now:

>diff -ur orig/glibc-2.1.1/sysdeps/powerpc/dl-machine.c glibc-2.1.1/sysdeps/powerpc/dl-machine.c
>--- orig/glibc-2.1.1/sysdeps/powerpc/dl-machine.c       Fri Mar  5 23:41:23 1999
>+++ glibc-2.1.1/sysdeps/powerpc/dl-machine.c    Mon May 17 20:59:06 1999
>@@ -250,7 +250,11 @@
>         PowerPC processors have line sizes of exactly 32 bytes.  */
> 
>       size_modified = lazy ? rel_offset_words : PLT_INITIAL_ENTRY_WORDS;
>+#ifdef PPC_CACHELINESIZE_32
>       for (i = 0; i < size_modified; i+= 8)
>+#else
>+      for (i = 0; i < size_modified; i+= 4)
>+#endif
>        PPC_DCBST (plt + i);
>       PPC_DCBST (plt + size_modified - 1);
>       PPC_SYNC;
>

We have trouble with PCI bus master boards in Motorola MBX board (PPC860): 
seems like a cache problem, all data not correct as read from PCI bus. 
We have used kernels 2.2.5, 2.3.10 and 2.3.18. 
The memory area is reserved by __get_free_pages(GFP_KERNEL | GFP_DMA,5);

The situation is improved a lot (it may even be correctred totally) by
changing 
cache setting in file ppc-asm/cache.h:

/* bytes per L1 cache line */
#define        L1_CACHE_BYTES  32      

=>

/* bytes per L1 cache line */
#define        L1_CACHE_BYTES  16

There is also L1_CACHE_PAGES in this file, but it seems that that is not used
anywhere.

Is this correct change or should we change something else to make access to
DRAM more reliable
from PCI ?
Is anybody else having problems with PCI bus master access in PPC860 ?

---
Kari Hämeenaho

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/