Config: Exit with fatal if a port is already connected

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

Description

Config: Exit with fatal if a port is already connected

This patch turns the existing warning into a fatal, as there should
never be any cases where a (non-vector) port is assigned to and then
later connected to something else. If this behaviour is allowed, as it
used to be, there are cases where the wrong number of C++ ports are
created when instantiating objects with VectorPorts (obviously that
could be fixed, but the better approach is to simply not allow it).

Testing Done

util/regress all passing (disregarding t1000 and eio)

hmm.. do we want to do it this way, or fix the issue? It seems like there may be value in allowing the configuration to change before it's instantiated (e.g create a perfectly workable system with only l1s and the later on decide you want to wedge in an l2).

Andreas Hansson May 20, 2012, 5:17 p.m. (May 20, 2012, 5:17 p.m.)

From my point of view what I propose is definitely the cleaner option, and no doubt the common case. I would suggest starting out with this patch to fix the "brokenness". 

If someone would like to construct a system the way you suggest, i.e. partially undoing previous connections, we can extend the port functionality (mostly the vector ports) to match the assumptions.

Ali Saidi May 21, 2012, 12:30 p.m. (May 21, 2012, 12:30 p.m.)

It's just an idea that I've had. If you could do this then you could imagine building a configuration by using decorators on a basic system. Something like

@o3_cpu
@l2caches
@l1caches
createBasicSystem()

Andreas Hansson May 21, 2012, 5:21 p.m. (May 21, 2012, 5:21 p.m.)

It's a nice idea...but also requires quite some work beyond the port connectivity. Shall we save the more extensive changes for later and just firm up the checks for now?

If overwrites of ports are currently causing problems, then I think the fatal error is appropriate.  If we decide to go with a scheme like Ali is proposing, we'll have to fix the overwrite behavior anyway, and at that point we can get rid of the error message.

Ship It!

You have a pending review.

Review Board 2.0.15

This change has been marked as submitted.

Screenshots

Files

Status: Closed (submitted)