Attention is currently required from: Bora Guvendik, Furquan Shaikh, Selma Bensaid, Tim Wawrzynczak, Angel Pons, Julius Werner, Aaron Durbin. Subrata Banik has posted comments on this change. ( https://review.coreboot.org/c/coreboot/+/51445 )
Change subject: timestamp: Add new helper functions ......................................................................
Patch Set 8:
(1 comment)
Commit Message:
https://review.coreboot.org/c/coreboot/+/51445/comment/bf71f630_ccf5cf13 PS3, Line 7: Add helper fucntions
I wouldn't move the base timestamp, but instead add these timestamps with a negative delta. […]
Thanks Angel and Julius for your suggestion. I believe we all are in same page but we are using different notation to explain our understanding.
Let me put the question straight here:
if you take a look into below table, the timestamp ids are in serialized order where ``` 944: CSE sent boot stall done to PMC is the first activity for CSE RO post coming out from reset.
Followed by
945, 946 and finally 947 where CSE receives CPU reset ack from PMC and we are moving towards CPU coming out from reset ```
now if we subtract the CSE timestamp from base or `0` and the cbmem does its sorting, it breaks the pre-cpu reset order (as you can see below, 947 becoming the least negative number). This is what my concern was about. I believe subtract is the final solution for simplicity but followed by sorting makes it weird as below.
0:1st timestamp 0 947:CSE received 'CPU Reset Done Ack sent' from PMC 1,074,000 946:CSE sent 'Host BIOS Prep Done' to PMC 308,000 (18,483,711,495,936,957) 945:CSE started to handle ICC configuration 303,000 (18,483,711,496,697,957) 944:CSE sent 'Boot Stall Done' to PMC 296,000 (18,483,711,496,695,957)