In the PCI Express (PCIe) architecture, the Data Link Layer serves as the crucial manager of the lower-level Link protocol. To successfully perform its duties, it relies on a specialized, behind-the-scenes messenger known as the Data Link Layer Packet (DLLP).
If you are exploring how PCIe devices maintain reliable connections, understanding DLLPs is essential. Here is a closer look at what these packets are, how they operate as “local traffic,” and why they remain completely hidden from the upper layers of the PCIe architecture.
What are DLLPs?
DLLPs are communication packets exchanged directly between the Data Link Layers of neighboring devices. They are the functional backbone of link management, specifically designed to support four major tasks:
- The Ack/Nak Protocol: Confirming the successful or failed reception of Transaction Layer Packets (TLPs) to ensure data integrity.
- Flow Control: Communicating available buffer space so transmitters know when it is safe to send data.
- Power Management: Negotiating power state transitions to save energy.
- Vendor-Specific Purposes: Carrying custom, vendor-defined information across the link.
Strictly “Local Traffic” (No Routing Required)
One of the most defining characteristics of DLLPs is that they are strictly local traffic.
Unlike TLPs, which act like long-haul freight trucks carrying payloads from one end of a complex system to another, DLLPs are more like local couriers communicating solely with the device immediately next door. Because they are exclusively used for this nearest-neighbor communication, DLLPs carry absolutely no target or routing information and are never routed through switches.
Invisible to the Transaction Layer
Because DLLPs handle the localized, “housekeeping” chores of the link, they are never seen by the Transaction Layer. The Transaction Layer is focused on high-level exchanges like reading and writing memory. Since DLLPs do not contain transactional data or routing paths, they are intercepted, processed, and consumed entirely by the Data Link Layer before they ever reach the Transaction Layer above it.
The Anatomy of a DLLP
To keep link management incredibly fast and efficient, DLLPs are built to be lightweight.
- Fixed Size: Every DLLP has a simple packet format and a fixed size of exactly 8 bytes, which includes the framing bytes.
- No Payload: There is never a data payload attached to a DLLP. All of the necessary information is carried within a 4-byte core (containing a 1-byte DLLP Type field and 3 bytes of specific attributes).
- 16-bit CRC: The core is followed by a 16-bit Cyclic Redundancy Check (CRC) used specifically to protect the DLLP’s contents.
How Receivers Handle Incoming DLLPs
When a neighboring device receives a DLLP, it processes the packet using a strict set of rules:
- Immediate Processing: DLLPs are processed immediately upon arrival. Because they are critical for link management, their flow cannot be restricted, meaning DLLPs are not subject to flow control.
- Error Checking: The packet’s 16-bit CRC is calculated and checked. If the CRC does not match, the DLLP is considered corrupted and is immediately discarded.
- No Acknowledgement: Unlike TLPs, there is no Ack/Nak acknowledgement protocol for DLLPs. If a DLLP fails the error check and is discarded, there is no mechanism to ask for a replay. Instead, the system simply relies on built-in timeout mechanisms or waits for the next periodic DLLP to arrive and update the missing information.
- Routing to Logic: If the packet passes the error check, the system reads the DLLP type and routes it to the correct internal logic to handle the Ack/Nak, Flow Control, or Power Management request.
Summary DLLPs are the, localized heroes of PCIe communication. By keeping link management traffic small, unrouted, and entirely separate from the Transaction Layer, DLLPs ensure that your system maintains reliable, efficient, and error-free data transport without clogging up the primary data pipelines.