3 comments:
Patch Set #7, Line 26: * to make *one* call per stage, otherwise some timestamps will be lost.
So is this an intentional change in behavior now? Previously this wasn't one call *per stage*, this was one call *in total*, in the earliest stage where you want to use timestamps. I don't think we should change that, requiring chipset code to do extra work in every stage is cumbersome and seems unnecessary in this case. You should be able to just keep setting up the pointer to the cache or CBMEM the first time it is needed in every stage.
Patch Set #7, Line 124: return car_get_ptr(glob_ts_table);
I think this should keep doing a cbmem_find() and/or a timestamp_cache_get() if the pointer is currently NULL, so we don't have to init it explicitly in every stage.
Patch Set #7, Line 172: timestamp_add_table_entry(ts_table, id, ts_time);
Why are we removing the error message? I think that was useful (since this points to a programming problem).
To view, visit change 35032. To unsubscribe, or for help writing mail filters, visit settings.