Merge pull request #897 from antmicro/jboc/dq-dqs-training
software/liblitedram: use single iteration of dq-dqs training
This commit is contained in:
commit
c4195254ed
|
@ -1075,23 +1075,17 @@ static void sdram_read_leveling_best_bitslip(int module)
|
|||
|
||||
static void sdram_write_dq_dqs_training(void)
|
||||
{
|
||||
/* In the first iteration read leveling and DQ-DQS training may not be done optimally */
|
||||
const int n_iter = 2;
|
||||
int i;
|
||||
int show;
|
||||
int module;
|
||||
|
||||
for (i = 0; i < n_iter; ++i) {
|
||||
show = i == n_iter - 1;
|
||||
for(module=0; module<SDRAM_PHY_MODULES; module++) {
|
||||
/* Find best bitslip */
|
||||
sdram_read_leveling_best_bitslip(module);
|
||||
/* Center DQ-DQS window */
|
||||
sdram_leveling_center_module(module, show, show,
|
||||
sdram_leveling_center_module(module, 1, 1,
|
||||
sdram_write_dq_dqs_training_rst_delay, sdram_write_dq_dqs_training_inc_delay);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
#endif /* SDRAM_PHY_WRITE_DQ_DQS_TRAINING_CAPABLE */
|
||||
|
||||
|
|
Loading…
Reference in New Issue