Review Board 2.0.15


O3: Support timing translations for O3 CPU fetch.

Review Request #340 - Created Dec. 6, 2010 and submitted

Information
Ali Saidi
gem5
Reviewers
Default
ali, gblack, nate, stever
O3: Support timing translations for O3 CPU fetch.

   
Posted (Dec. 8, 2010, 3:18 p.m.)



  
src/cpu/o3/fetch.hh (Diff revision 1)
 
 
Neither the tid or _mem_req seems to be used. These can get nuked, right?
src/cpu/o3/fetch.hh (Diff revision 1)
 
 
It's not necessary, but maybe an assert here that the "mode" parameter indicates a fetch?
src/cpu/o3/fetch_impl.hh (Diff revision 1)
 
 
Why should this be removed?
src/cpu/o3/fetch_impl.hh (Diff revision 1)
 
 
It's not part of your change, but while you're dicing up this function lets get rid of this commented out code.
src/cpu/o3/fetch_impl.hh (Diff revision 1)
 
 
This looks more involved than the original version of the code below. Why does it need to do more work?
src/cpu/o3/fetch_impl.hh (Diff revision 1)
 
 
star->start
src/cpu/o3/fetch_impl.hh (Diff revision 1)
 
 
Ah... This is why that block of code above can be removed?
src/cpu/o3/fetch_impl.hh (Diff revision 1)
 
 
This if and the one above it really are begging to be merged. If the condition turns into a big mess then never mind, but otherwise I think we definitely want to do that.
Posted (Jan. 9, 2011, 11:40 a.m.)



  
src/cpu/o3/fetch.hh (Diff revision 1)
 
 
done
src/cpu/o3/fetch.hh (Diff revision 1)
 
 
done
src/cpu/o3/fetch_impl.hh (Diff revision 1)
 
 
It's not gone, it's moved... Need to check that the physical address still is what we think it is after translation. 
src/cpu/o3/fetch_impl.hh (Diff revision 1)
 
 
it's really dated.. done
src/cpu/o3/fetch_impl.hh (Diff revision 1)
 
 
you're right buildInst should have been used and then it will be ~ the same size
src/cpu/o3/fetch_impl.hh (Diff revision 1)
 
 
done
src/cpu/o3/fetch_impl.hh (Diff revision 1)
 
 
yes