On Nov 12, 2011, at 2:07 PM, Simon Glass wrote:
Hi Kumar,
On Sat, Nov 12, 2011 at 10:02 AM, Kumar Gala galak@kernel.crashing.org wrote:
This reverts commit 355a835747c6f7c5071ead295a7dfc489c73cb03.
The original commit broke long standing assumption that md commands work on effective addresses. This normally isn't an issue for most systems that map 1:1, however on systems with a 36-bit address map it breaks.
What is the system that shows this problem?
Any PPC system that have _36BIT and some that imply it in boards.cfg
With sandbox we need some sort of mapping here - what do you suggest? map_virtmem?
Not sure, what the need is w/sandbox so hard to suggest a solution.
It would be good to get a solution to this rather than just reverting.
Sure, but I'd rather revert in short term so the mainline tree works for all existing boards, rather than for the new feature until we come up with a solution.
Regards, Simon
Signed-off-by: Kumar Gala galak@kernel.crashing.org
common/cmd_mem.c | 9 ++------- 1 files changed, 2 insertions(+), 7 deletions(-)
diff --git a/common/cmd_mem.c b/common/cmd_mem.c index 461ee19..28476d7 100644 --- a/common/cmd_mem.c +++ b/common/cmd_mem.c @@ -33,7 +33,6 @@ #include <dataflash.h> #endif #include <watchdog.h> -#include <asm/io.h>
#ifdef CMD_MEM_DEBUG #define PRINTF(fmt,args...) printf (fmt ,##args) @@ -142,13 +141,9 @@ int do_mem_md ( cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[]) # endif
{
ulong bytes = size * length;void *buf = map_physmem(addr, bytes, MAP_WRBACK);/* Print the lines. */print_buffer(addr, buf, size, length, DISP_LINE_LEN / size);addr += bytes;unmap_physmem(buf, bytes);
print_buffer(addr, (void*)addr, size, length, DISP_LINE_LEN/size);addr += size*length; }#endif
-- 1.7.3.4