Hi All,
I have a question about assert timing of PSTRB and PPROT.
I think APB4 spec (IHI0024C_amba_apb_protocol_spec.pdf) does not describe when PSTRB and PPROT should be asserted.I guess these signals should be asserted while PSEL is high like PADDR and PWDATA.
Is my understanding correct?
Regards,Taichi Ishitani
Hi Colin,
Thank you for your answer and I understand.
Colin Campbell said:All "control" signals (other than PENABLE) in an APB transfer should be constant for the entire transfer.
I cannot find the above description from the APB spec (IHI0024C_amba_apb_protocol_spec.pdf).Could you point me where this description is?
I think the APB spec does not describe when asserting PSTRB and PPROT should start clearly.Especially, I'd like to know this point.
I thought that was the point of your initial question, the spec does not show this, so my reply attempted to explain what IS required.
Regarding my earlier comment about control signals staying constant, just look at any diagram in the protocol document, they all clearly show this, so I was putting into words what the diagrams show. This is also described in sections 3.1 and 3.2, and in the later description of the ACCESS state in section 4.1.
As PSTRB and PPROT were only added in the last release of the protocol, these new signals should have been added to the above descriptions.
Thank you for your detailed answer. I understand.
Colin Campbell said:I thought that was the point of your initial question, the spec does not show this, so my reply attempted to explain what IS required.
Colin Campbell said:As PSTRB and PPROT were only added in the last release of the protocol, these new signals should have been added to the above descriptions.
Do you have a plan to update the spec document to add the above description?
To be honest I'm not sure the spec needs updating to explicitly cover these signals. PPROT is a control signal just like PADDR or PWRITE, so it seems reasonable to assume it would have exactly the same timing requirements. Similarly PSTRB is very closely related to PWDATA, so should have the same timing requirements.
However if any update to the APB protocol is proposed, I'm sure these description updates will be included just to avoid any doubt.
Thank you again for your kindly replies.