core/arm: Improve timing accuracy before service calls in JIT
We also correct the CPU JIT's implementation of Step.
This commit is contained in:
parent
dc030c78c3
commit
7cd8b437aa
9 changed files with 59 additions and 53 deletions
|
@ -50,10 +50,10 @@ public:
|
|||
* is not required to do a full dispatch with each instruction. NOTE: the number of instructions
|
||||
* requested is not guaranteed to run, as this will be interrupted preemptively if a hardware
|
||||
* update is requested (e.g. on a thread switch).
|
||||
* @param tight_loop Number of instructions to execute.
|
||||
* @param tight_loop If false, the CPU single-steps.
|
||||
* @return Result status, indicating whethor or not the operation succeeded.
|
||||
*/
|
||||
ResultStatus RunLoop(int tight_loop = 1000);
|
||||
ResultStatus RunLoop(bool tight_loop = true);
|
||||
|
||||
/**
|
||||
* Step the CPU one instruction
|
||||
|
|
Loading…
Add table
Add a link
Reference in a new issue