nb/intel/sandybridge/raminit: Fix normalize_training
Remove cross rank/cross channel dependency.
I guess this is a mistake that could lead to instabilities.
Tested on Lenovo T430 (Intel IvyBridge).
Change-Id: I899db907cd2d2197fd81eda4c4656fb1e570c18f
Signed-off-by: Patrick Rudolph <siro@das-labor.org>
Reviewed-on: https://review.coreboot.org/17610
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-by: Philippe Mathieu-Daudé <philippe.mathieu.daude@gmail.com>
Tested-by: build bot (Jenkins)
diff --git a/src/northbridge/intel/sandybridge/raminit_common.c b/src/northbridge/intel/sandybridge/raminit_common.c
index 8baa0f2..0c77004 100644
--- a/src/northbridge/intel/sandybridge/raminit_common.c
+++ b/src/northbridge/intel/sandybridge/raminit_common.c
@@ -3023,10 +3023,11 @@
void normalize_training(ramctr_timing * ctrl)
{
int channel, slotrank, lane;
- int mat = 0;
+ int mat;
FOR_ALL_CHANNELS FOR_ALL_POPULATED_RANKS {
int delta;
+ mat = 0;
FOR_ALL_LANES mat =
max(ctrl->timings[channel][slotrank].lanes[lane].timA, mat);
printram("normalize %d, %d, %d: mat %d\n",