Extended HWM commands refactor - #12682
Conversation
|
|
||
| auto ans = hw_monitor::send(cmd, p_response, locked_transfer); | ||
| recv_msg.insert(recv_msg.end(), ans.begin(), ans.end()); | ||
| // currently we assume HKR send only the table size when sending 0/0 chanks command above. |
| auto ans = hw_monitor::send(cmd, p_response, locked_transfer); | ||
| recv_msg.insert(recv_msg.end(), ans.begin(), ans.end()); | ||
| // currently we assume HKR send only the table size when sending 0/0 chanks command above. | ||
| // in the future the FW should send chunk 1 in 0/0 command and than we can use it here. |
| // actual size will be returned as part for the response header and will be used | ||
| // to calculate the extended loop range | ||
| auto ans = hw_monitor::send(cmd, p_response, locked_transfer); | ||
| recv_msg.insert(recv_msg.end(), ans.begin(), ans.end()); |
There was a problem hiding this comment.
if I understand correctly, this is an API change:
instead of getting the first chunk, we get the message, in which only the header is needed, and then, we get the chunks in other messages.
Please confirm this change has been implemented also in HKR side.
There was a problem hiding this comment.
As discusses,
Current FW : we get the last chunk size of the first messages which is wrong, but since we get a valid header we extract the size from there and restart the querying process.
FW should supply a fix where 0/0 we send the first chunk with full size, then we can optimize and take this chunk and continue the loop from index 1.
This will be done on another PR once we have their fix.
remibettan
left a comment
There was a problem hiding this comment.
few comments - thanks!
When we get config table we get the size in the header.
This PR remove the hard coded tables sizes, and leverage the header size value.
Also did some renaming
Tracked on [RSDEV-1735]