Re: [Gems-users] Sequencer


Date: Mon, 7 Aug 2006 09:39:30 -0500 (CDT)
From: Mike Marty <mikem@xxxxxxxxxxx>
Subject: Re: [Gems-users] Sequencer
>
> I am trying to debug my code with the tester.exec and I compare my outcome
> to the original MOSI_SMP_bcast_1level outcome.
>
> What I found out is that in the original code after LD just before the DATA
> arrives I see the sequencer line that print DONE/END.
>

This is normal in the protocol trace.  They sequencer and the data message
print their transitions in the same cycle, but the sequencer prints first.

> In my code I see the data arrives but with no sequencer DONE before it. This
> is the point I get the "Possible Deadlock detected" error.
>

The sequencer completes the request when it gets a hitCallback.  If you
look in the SLICC cache controller specification, you can see where these
callbacks occur.

--Mike

[← Prev in Thread] Current Thread [Next in Thread→]