Dear All,
Would you please help me to understand why every slaves always hreadyout declared as '1' in Cortex M Design Kit and can't modify it in slave status?
I found a declaration assign hreadyouts = 1'b1;
Actually I'm trying to monitor only HWDATA and HRDATA when they are executed as the below
As you can see the above waveform, there are 4 HWDATA and 4HRDATA . But there is no way what I want to capture the only executed point. because Hready is always '1' and else signals also doesn't help much to me.
Could you guide me how I can check the only read and write event for HRDATA and HWDATA ?
If you are a CMSDK licensee you should submit a support request to Arm on the licensee only support portal that your company should be aware of.
Please also say where in the CMSDK deliverables you see the hreadyouts assignment.
Initial guess without knowing where in the CMSDK system you are looking at is that this module never needs to signal wait states, hence the wait signal being tied high.
Hi Colin,
As I know, the HRDATA comes out from slave to Master after address phase when the Read Operation. I wonder that if there is something big skew between Master and Salve so the HRDATA doesn't go to the Master's Positive sensitive clock edge slightly. I think this could be happened in everywhere at the real field.
So to get the HRDATA safely from Slave in the MASTER, I'd like to take the F/F synchronize with input HRDATA in the MASTER. but The master use one cycle for delay with BUSY. Is this make sense?
Unfortunately you didn't answer my question about where in CMSDK you are looking, so I'm still not sure what you are asking about.
From your latest question, if I understand it correctly you are either worried about a delay from the data phase starting to when HRDATA can be sampled on an HCLK rising edge by the master, or else you are worried about meeting single cycle timing for a long path on HRDATA from the slave back to the master.
If it is the HRDATA not reaching the master in time to meet the next HCLK rising edge timing, the slave will have to add a wait state on HREADY.
If you are concerned about a long timing path on HRDATA, you can't just add a register to this path as that will break the relationship between address and data phases of a transfer, so you would need to add an AHB to AHB bridge between the master and slave, and this would then break the long timing path (at a cost of latency).
Sorry, I'm not sure what the BUSY cycle is that you are referring to.
Thank you Colin,
Yes, actually I was concerning that long timing path on HRDATA, But I am clear Now.
Thank you Colin again and sorry for late,