Update resume from Debug Mode behavior with Smdbltrp #1051
Conversation
There was a problem hiding this comment.
I see. MRET/SRET clear MDT so that you won't get a critical error on the next trap from U mode (or whatever). Resuming should clear it for the same reason.
This is very slightly different than the MRET/SRET requirements in the priv spec because you are able to go back to M mode with MDT=1 (whereas MRET to M would clear MDT). But we don't want to clear MDT in that case because the debugger needs to be allowed to resume to the critical part of the M mode handler where another trap should be a critical error. So this all makes sense.
|
Additionally, debug mode can resume to S with |
|
@pdonahue-ventana - I had missed qualifying the |
|
It's not entirely clear that the third sentence has the same Ssdbltrp condition as the second sentence. It really depends on both H and Ssdbltrp because (for obvious reasons) Ssdbltrp only adds the vsstatus bit if H is implemented. |
|
I placed the second and third sentence in a separate list item to avoid confusion that sentence 3 is not related to Ssdbltrp. I dont say "if H is implemented" as it is implied by VU mode which is provided only when H is implemented. |
|
A separate bullet is perfect. |
Sdext.adoc
Outdated
| . If the Smdbltrp extension is implemented and the new privilege mode is not M, | ||
| then the `MDT` bit is set to 0. | ||
| . If the Ssdbltrp extension is implemented and the new privilege mode is U, VS, | ||
| or VU, then `sstatus.SDT` is also set to 0. Additionally, if it is VU, then |
There was a problem hiding this comment.
| or VU, then `sstatus.SDT` is also set to 0. Additionally, if it is VU, then | |
| or VU, then `sstatus.SDT` is set to 0. Additionally, if it is VU, then |
|
@pdonahue-ventana @rtwfroody |
|
The cetrig change looks good. |
No description provided.