You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
For AXIS, we must assert the PLD_LAST signal on the
last word for a transaction before issuing a new command
on the CMD interface. The condition to detect the last
word has 2 possibilities:
1) Specified number of words written to the DDR3 controller,
typical case in "immediate"" acquisition, and this worked fine.
2) DDR3 AXIS address going over the limit and wrapping back
to the start address. This was always asserting the last word
detection (ddr_eop signal) on the first word of the next
transaction, causing the last word to be written past
the DDR3 end address and the next word to be written
correctly at the DDR start address.
The fix for the case #2 was a simple "=", as we need to consider
the case for the DDR end address to be aligned with the
"ddr_addr_cnt_axis + c_addr_ddr_inc_axis", which is the
condition to the counter to be the last one on the next word.
This was reported by Scott Cogan, from FRIB. Thanks Scott.
Problem: There are use cases in which we may want to set the synchronizing RX edge of a pulse via an external pin.
Solution: Change the input sync module to one that allows setting the sync_edge by an external driving pin
The text was updated successfully, but these errors were encountered: