Mem: Use range operations in bus in preparation for striping

Information
Submitter:	Andreas Hansson
Repository:	gem5
Branch:	default
Bugs:
Depends On:
Reviewers
Groups:	Default
People:

Description

This patch transitions the bus to use the AddrRange operations instead
of directly accessing the start and end. The change facilitates the
move to a more elaborate AddrRange class that also supports address
striping in the bus by specifying interleaving bits in the ranges.

Two new functions are added to the AddrRange to determine if two
ranges intersect, and if one is a subset of another. The bus
propagation of address ranges is also tweaked such that an update is
only propagated if the bus received information from all the
downstream slave modules. This avoids the iteration and need for the
cycle-breaking scheme that was previously used.

Testing Done

util/regress all passing (disregarding t1000 and eio)

Issue Summary

Description	From	Last Updated	Status
Should this not be start <= r.start?	Nilay Vaish	Sept. 21, 2012, 2:35 a.m.	Resolved
I think the check being carried out is the other way around.	Nilay Vaish	Sept. 21, 2012, 2:35 a.m.	Resolved
Why do we need to do this? How many characters are we saving any way?	Nilay Vaish	Oct. 7, 2012, 7:33 p.m.	Dropped
Is this intentional?	Nilay Vaish	Sept. 21, 2012, 2:37 a.m.	Resolved

Andreas Hansson Oct. 10, 2012, 10:29 p.m. (Oct. 10, 2012, 10:29 p.m.)
```
Happy with the changes Nilay?
```

src/base/addr_range.hh (Diff revision 1)

Should this not be start <= r.start?

Show all issues

src/base/addr_range.hh (Diff revision 1)

I think the check being carried out is the other way around.

Show all issues

src/base/addr_range.hh (Diff revision 1)

Why do we need to do this? How many characters are we saving any way?

Show all issues

Andreas Hansson Sept. 21, 2012, 2:37 a.m. (Sept. 21, 2012, 2:37 a.m.)

The operators are important with the addition of striping. The whole idea is to hide all the checking for interleaving and matching inside the AddrRange and never use the start and end outside of this class. I hope that explains it.

Nilay Vaish Oct. 11, 2012, 11:26 a.m. (Oct. 11, 2012, 11:26 a.m.)

Andreas, I wrote to you almost three weeks back that I do not like overloading the
equality operator for checking whether the address lies in a range. Can you not just
create a new function?

Andreas Hansson Oct. 11, 2012, 5:38 p.m. (Oct. 11, 2012, 5:38 p.m.)

Sure, I'm happy to do either. The equality operator is only there for historical reasons, that is how it has been done up till this very point. Rather than changing all the places where it is used (and it is quite a few) I was happy to know if anyone else sees a need to change it before going and doing so.

Andreas Hansson Oct. 11, 2012, 5:46 p.m. (Oct. 11, 2012, 5:46 p.m.)

In addition, if we decide to change it, I want to keep it out of this patch. I hope that's ok.

Nilay Vaish Oct. 11, 2012, 9:10 p.m. (Oct. 11, 2012, 9:10 p.m.)
```
I am fine with having a separate patch for the change.
```

Ali Saidi Oct. 11, 2012, 11:58 p.m. (Oct. 11, 2012, 11:58 p.m.)

Personally, I'm fine with using the operator overloading here.  What problem do you have with it specifically?

src/mem/bus.hh (Diff revision 1)

Is this intentional?

Show all issues

Description:

		This patch transitions the bus to use the AddrRange operations instead
		of directly accessing the start and end. The change facilitates the
		move to a more elaborate AddrRange class that also supports address
		striping in the bus by specifying interleaving bits in the ranges.

		Two new functions are added to the AddrRange to determine if two
		ranges intersect, and if one is a subset of another. The bus
		propagation of address ranges is also tweaked such that an update is
		only propagated if the bus received information from all the
		downstream slave modules. This avoids the iteration and need for the
		cycle-breaking scheme that was previously used.

Diff:

Revision 2 (+164 -101)

Show changes

	src/base/addr_range.hh
	src/mem/bus.hh
	src/mem/bus.cc

Any feedback? Good to go?

Ship It!

You have a pending review.

Review Board 2.0.15

This change has been marked as submitted.

Screenshots

Files

Issue Summary

Description:

Diff:

Status: Closed (submitted)