Peter Stuge wrote:
On Fri, Nov 02, 2007 at 01:30:25PM -0400, Corey Osgood wrote:
+void smbus_fixup(const struct mem_controller *ctrl) +{
- int i, ram_slots, current_slot = 0;
- u8 result = 0;
- ram_slots = ARRAY_SIZE(ctrl->channel0);
- if (!ram_slots) {
print_err("smbus_fixup thinks there are no ram slots!\r\n");
return;
- }
- PRINT_DEBUG("Waiting for smbus to warm up");
- /* Bad SPD data should be either 0 or 0xff, but YMMV. So we look for the
* ID bytes of SDRAM, DDR, DDR2, and DDR3 (and anything in between).
* vt8237r has only been seen on DDR and DDR2 based systems, so far */
- for(i = 0; (i < SMBUS_TIMEOUT && ((result < SPD_MEMORY_TYPE_SDRAM) ||
(result > SPD_MEMORY_TYPE_SDRAM_DDR3))); i++)
- {
if (current_slot > ram_slots) j = 0;
What is j here?
Oops, should have been current_slot. I can fix it before commit.
+++ src/include/spd.h (working copy) @@ -105,6 +105,7 @@ #define SPD_MEMORY_TYPE_SGRAM_DDR 6 #define SPD_MEMORY_TYPE_SDRAM_DDR 7 #define SPD_MEMORY_TYPE_SDRAM_DDR2 8 +#define SPD_MEMORY_TYPE_SDRAM_DDR3 0xb
Looks like bad whitespace, but that's a separate patch.
Weird, it looks fine in the file.
Otherwise I'll ack. Can you commit?
Yep, just gotta figure out how. I'll hop on IRC if I have a problem.
-Corey