A method and system for adjusting data line transmission rate based on communication protocol

By monitoring adaptive equalizer parameters, analyzing compensation spectrum characteristics, identifying link physical degradation or external interference, and generating transmission rate adjustment suggestions, the problem of inflexible rate adjustment in data cable transmission is solved, improving data transmission stability and user experience.

CN123074640APending Publication Date: 2026-11-10DONGGUAN KAIKE ELECTRONICS TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611330861.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-31
Publication Date
2026-11-10

AI Technical Summary

Technical Problem

In current data transmission, data lines are physically degraded or subject to external interference, making it impossible to finely and gradually adjust the traditional fixed-rate degradation strategy, resulting in a poor user experience.

Method used

By monitoring the parameters of the adaptive equalizer, during the non-link reconstruction phase of the state machine specified by the communication protocol, the compensation spectrum characteristics are analyzed to identify the source of physical degradation of the link or external interference, generate suggestions for transmission rate adjustment or anti-interference, and achieve fine-grained and gradual rate adjustment through communication protocol negotiation.

Benefits of technology

It enables dynamic adjustment of transmission rate based on real-time link quality, improving data transmission stability and user experience, and avoiding the problem of blindly and drastically reducing speed in traditional methods.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123074640A_ABST
    Figure CN123074640A_ABST
Patent Text Reader

Abstract

This application relates to the field of data transmission technology, and discloses a data line transmission rate adjustment method and system based on a communication protocol. By introducing early monitoring of equalizer parameters, extraction and matching of compensation spectral features, and a refined suggestion generation and protocol negotiation mechanism based on the source of the problem, an intelligent and adaptive data line transmission rate adjustment method is constructed. This not only solves the problems of inflexible rate adjustment and inability to adapt to cable physical defects and external interference in traditional solutions, but also significantly improves the stability and efficiency of data transmission, providing users with a better experience.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the field of data transmission technology, and more specifically, to a method and system for adjusting the data line transmission rate based on a communication protocol. Background Technology

[0002] In modern data transmission, devices need to be compatible with multiple communication protocols (such as USB-C and HDMI) when connected via data cables. Protocol negotiation often fails due to link quality issues, resulting in a decrease in transmission speed or even connection interruption.

[0003] In typical scenarios, repeated bending of USB-C data cables can cause physical fatigue in the differential pairs near the connectors. This fatigue results in uneven cross-sectional area and unbalanced stress distribution due to conductor lattice slip and dislocation accumulation, causing the differential impedance to deviate from the design value (e.g., from 90Ω to 85Ω or 95Ω). For high-frequency signals above gigabits, this impedance shift can lead to reflections and attenuation, compromising signal integrity.

[0004] During the link training phase, the receiver equalizer is unable to converge to stable parameters for persistent impedance offset, resulting in intermittent increases in the bit error rate of high-frequency signals at the highest speed level (such as the error bits exceeding the correctable range at USB 3.2 Gen 2x2 20Gbps), and the link cannot be stabilized.

[0005] When the bit error rate exceeds the tolerance threshold, the device can only trigger a preset fixed rate degradation strategy (such as 20Gbps→10Gbps→5Gbps), and cannot make fine-grained and gradual adjustments based on real-time link quality. This results in either a significant speed reduction or frequent reconnection, leading to a poor user experience.

[0006] Therefore, there is an urgent need for a method that can dynamically and gradually adjust the transmission rate based on real-time link quality. Summary of the Invention

[0007] This application discloses a data line transmission rate adjustment method and system based on a communication protocol, aiming to solve the problem in the existing data transmission field that traditional fixed rate degradation strategies cannot finely and gradually adjust the transmission rate under physical degradation or external interference of data lines, resulting in poor user experience.

[0008] The technical solution of this application is as follows: In a first aspect, this application discloses a data line transmission rate adjustment method based on a communication protocol, comprising: During the operation of the data line link, when the data bit error rate has not yet reached the preset error threshold and the tap coefficients of the adaptive equalizer at the data line receiver are working in the unsaturated range, during the non-link reconstruction phase of the state machine specified by the communication protocol, the equalizer parameters adjusted by the adaptive equalizer for the training signal specified by the communication protocol are monitored. Analyze the timing changes of equalizer parameters over one or more complete transmission cycles of the training signal, and extract compensation spectral features corresponding to the protocol structure of the training signal. The compensation spectral features include at least two of the following: the relative changes between tap coefficients, the differences in adjustment rates of tap coefficients, and the compensation modes for the frequency response of the signal. The compensation spectrum features are matched with a preset set of link physical damage spectrum features and a set of external interference spectrum features to identify the sources of equalizer parameter fluctuations; the sources include at least one of link physical degradation and external interference. When the source is physical degradation of the link, a channel warning message containing channel warning information and suggested rate step size is generated; based on the channel warning message and the position of the current transmission rate in the preset transmission rate level set, a transmission rate adjustment suggestion is generated; the transmission rate adjustment suggestion includes: reducing the current transmission rate to a level in the level set that is lower than the current transmission rate, and / or maintaining the current transmission rate at the original level and evaluating the stability of the equalizer parameters at the target level to determine whether to upgrade to the target level; When the source is external interference, anti-interference suggestions are generated; the anti-interference suggestions include at least one of maintaining the current transmission rate and increasing forward error correction redundancy, and enabling data retransmission protocol. According to the protocol data unit format specified in the communication protocol, the transmission rate adjustment suggestion or anti-interference suggestion is encapsulated into a negotiation request; the negotiation request is sent to the other end of the data line through the negotiation channel specified in the communication protocol, and the working status is determined by negotiating with the other end of the data line; the negotiation request carries at least one of the following: target transmission rate, suggested rate step size, channel warning level, and warning confidence level; According to the communication protocol, the link is retrained based on the negotiated working status to confirm that the working status can work stably.

[0009] Secondly, this application also discloses a data line transmission rate adjustment system based on a communication protocol, comprising: The parameter monitoring module is used to monitor the equalizer parameters adjusted by the adaptive equalizer to the training signal specified by the communication protocol during the non-link reconstruction phase of the state machine specified by the communication protocol, when the data error rate has not yet reached the preset error threshold and the tap coefficients of the adaptive equalizer at the data line receiver are working in the unsaturated range. The feature extraction module is used to analyze the temporal changes of equalizer parameters within one or more complete transmission cycles of the training signal and extract compensation spectral features corresponding to the protocol structure of the training signal. The compensation spectral features include at least two of the following: the relative changes between tap coefficients, the differences in adjustment rates of tap coefficients, and the compensation modes for the frequency response of the signal. The source identification module is used to match the compensation spectrum features with a preset set of link physical damage spectrum features and a set of external interference spectrum features to identify the source of equalizer parameter fluctuations; the source includes at least one of link physical degradation and external interference. The negotiation request generation module is used to generate a channel warning message containing channel warning information and a suggested rate step size, and to generate a transmission rate adjustment suggestion when the source is link physical degradation; and to generate anti-interference suggestions when the source is external interference. The transmission rate adjustment suggestions include reducing the current transmission rate to a level lower than the current transmission rate, and / or maintaining the current transmission rate at the original level and evaluating the stability of the equalizer parameters at the target level to determine whether to upgrade to the target level; the anti-interference suggestions include maintaining the current transmission rate and increasing forward error correction redundancy, and enabling at least one of the data retransmission protocol; the negotiation request carries at least one of the following: target transmission rate, suggested rate step size, channel warning level, and warning confidence level. The negotiation training module is used to encapsulate transmission rate adjustment suggestions or anti-interference suggestions into negotiation requests according to the protocol data unit format specified by the communication protocol, and send them to the other end of the data line through the negotiation channel specified by the communication protocol to negotiate and determine the working status with the other end of the data line; and trigger link retraining according to the working status determined by the negotiation according to the communication protocol to confirm that the working status can work stably.

[0010] This application discloses a data line transmission rate adjustment method and system based on a communication protocol. By compensating for spectral characteristics, it distinguishes between physical degradation and external interference, and provides targeted speed reduction or anti-interference strategies to replace the traditional fixed step-by-step degradation. This achieves refined and gradual rate adjustment, improving link stability while avoiding blind and drastic speed reduction and optimizing user experience. Attached Figure Description

[0011] Figure 1 This is a flowchart illustrating a data line transmission rate adjustment method based on a communication protocol provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of a data line transmission rate adjustment system based on a communication protocol provided in an embodiment of the present invention. Detailed Implementation

[0012] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0013] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0014] A method for adjusting the data cable transmission rate based on a communication protocol is proposed. In traditional data transmission, when devices are connected via data cables, transmission rate reductions or even connection interruptions often occur due to protocol negotiation failures. Especially when cables have physical defects, traditional fixed transmission rate modes struggle to adapt flexibly to dynamic demands. For example, after repeated bending, the internal wire pairs of interfaces such as USB-C and HDMI may experience physical fatigue, leading to impedance shifts, increased high-frequency signal bit error rates, and the link failing to stabilize at its maximum speed. Existing devices typically employ preset fixed-rate degradation strategies, failing to provide fine-grained adjustments based on real-time link quality, resulting in a poor user experience.

[0015] In this regard, refer to Figure 1 , Figure 1 This is a flowchart illustrating a data line transmission rate adjustment method based on a communication protocol provided in an embodiment of the present invention, comprising: S11, During the operation of the data line link, when the data bit error rate has not yet reached the preset error threshold and the tap coefficients of the adaptive equalizer at the data line receiver are working in the unsaturated range, during the non-link reconstruction phase of the state machine specified by the communication protocol, monitor the equalizer parameters adjusted by the adaptive equalizer for the training signal specified by the communication protocol. S12, Analyze the timing changes of equalizer parameters within one or more complete transmission cycles of the training signal, and extract the compensation spectral features corresponding to the protocol structure of the training signal; the compensation spectral features include at least two of the following: the relative changes between tap coefficients, the differences in adjustment rates of tap coefficients, and the compensation modes for the frequency response of the signal. S13, Match the compensation spectrum features with the preset set of link physical damage spectrum and external interference spectrum to identify the source of equalizer parameter fluctuations; the source includes at least one of link physical degradation and external interference; S14, when the source is link physical degradation, generate a channel warning message containing channel warning information and suggested rate step size; based on the channel warning message and the position of the current transmission rate in the preset transmission rate level set, generate a transmission rate adjustment suggestion; the transmission rate adjustment suggestion includes: reducing the current transmission rate to a level in the level set that is lower than the current transmission rate, and / or maintaining the current transmission rate at the original level and evaluating the stability of the equalizer parameters at the target level to determine whether to upgrade to the target level; S15, when the source is external interference, generate anti-interference suggestions; the anti-interference suggestions include at least one of maintaining the current transmission rate and increasing forward error correction redundancy, and enabling data retransmission protocol; S16, according to the protocol data unit format specified in the communication protocol, the transmission rate adjustment suggestion or anti-interference suggestion is encapsulated into a negotiation request; the negotiation request is sent to the other end of the data line through the negotiation channel specified in the communication protocol, and the working status is determined by negotiating with the other end of the data line; the negotiation request carries at least one of the following: target transmission rate, suggested rate step size, channel warning level, and warning confidence level; S17, according to the communication protocol, trigger link retraining based on the negotiated working state to confirm that the working state can work stably.

[0016] This application dynamically monitors adaptive equalizer parameters during link operation and analyzes their temporal changes to extract compensation spectrum features, thereby identifying whether the source of equalizer parameter fluctuations is link physical degradation or external interference. Based on the different sources, it generates refined transmission rate adjustment suggestions or anti-interference suggestions, and negotiates and retrains the link through a communication protocol, thereby achieving intelligent adjustment of the data line transmission rate. This effectively solves the problems of inflexible rate adjustment and inability to adapt to cable physical defects and external interference in traditional solutions, significantly improving data transmission stability and user experience.

[0017] To better understand this application, some key terms involved need to be explained.

[0018] A "data cable" is a physical medium used to transfer data between two or more devices, such as a USB-C cable or an HDMI cable.

[0019] "Communication protocol" refers to a set of standards that regulate the data transmission format, timing, error handling, etc., such as PCIe protocol and USB protocol.

[0020] An "adaptive equalizer" is a circuit module at the receiver used to compensate for channel loss and restore signal integrity. It adapts to different channel characteristics by adjusting the "tap coefficients".

[0021] "Tap coefficients" are adjustable parameters in adaptive equalizers used to adjust the frequency response of the equalizer to counteract the distortion caused by the channel to the signal.

[0022] The "unsaturated range" refers to the state where the tap coefficients of the adaptive equalizer have not yet reached their maximum or minimum adjustment range, indicating that the equalizer still has adjustment margin.

[0023] The "state machine non-link reconstruction phase" refers to the data transmission phase defined in the communication protocol where the link has been established and is working normally, such as the L0 / L0s state of PCIe or the U0 state of USB. At this time, the link does not need to be reconfigured on a large scale.

[0024] "Training signals" are specific signal sequences specified by communication protocols for link training and equalizer parameter adjustment, such as the TS1 / TS2 sequence for PCIe or the Polling / LFPS signals for USB.

[0025] "Compensation spectral characteristics" refers to the patterns extracted by analyzing changes in equalizer parameters that reflect channel impairment or interference characteristics, including the relative changes between tap coefficients, the differences in adjustment rates of tap coefficients, and the compensation patterns for signal frequency response.

[0026] The “Link Physical Damage Spectrum Set” and the “External Interference Spectrum Set” are pre-stored feature pattern libraries used to identify the sources of equalizer parameter fluctuations.

[0027] "Channel warning message" is a message generated by the system when physical degradation of the link is identified. It contains channel status warning information and suggested rate adjustment step size.

[0028] A “transmission rate level set” is a series of selectable transmission rates preset by a communication protocol or system, such as 5Gbps, 10Gbps, 20Gbps, etc.

[0029] A "negotiation request" is a message that the system encapsulates transmission rate adjustment suggestions or anti-interference suggestions and sends them to the peer device through a communication protocol to negotiate a new working state.

[0030] A "negotiation channel" is a logical or physical channel specified in a communication protocol for exchanging control information between devices.

[0031] "Link retraining" refers to retraining the link according to the communication protocol after a new working state is determined through negotiation, so that the link can work stably in the new state.

[0032] During the operation of the data line link, when the data error rate has not yet reached the preset error threshold and the tap coefficients of the adaptive equalizer at the data line receiver are operating in the unsaturated range, it is necessary to monitor the equalizer parameters adjusted by the adaptive equalizer for the training signal specified by the communication protocol during the non-link reconstruction phase of the state machine as defined by the communication protocol. For example, a parameter monitoring unit can be integrated at the receiver, which reads the tap coefficient register values ​​inside the adaptive equalizer in real time. When the link is in a normal data transmission state (e.g., PCIe L0 state), this unit will periodically capture the parameter adjustments made by the equalizer when processing the training signal (e.g., PCIe TS1 sequence). Another implementation method is to reserve a debugging interface in the adaptive equalizer design, through which the tap coefficient change curve of the equalizer during the training signal period can be obtained in real time.

[0033] Next, it is necessary to analyze the temporal variations of the equalizer parameters over one or more complete transmission cycles of the training signal to extract the compensation spectral features corresponding to the protocol structure of the training signal. These compensation spectral features can include at least two of the following: the relative changes between tap coefficients, the differences in adjustment rates of the tap coefficients, and the compensation modes for the signal frequency response. For example, a feature extraction module can receive the monitored equalizer parameter sequence and calculate the differences between the tap coefficients in adjacent training signal cycles to obtain the relative changes. Simultaneously, by performing time series analysis on these differences, the differences in adjustment rates of the tap coefficients can be obtained. Furthermore, by mapping the equalizer parameters to the frequency domain, the compensation intensity and mode for different frequency components can be analyzed, thereby extracting the compensation mode for the signal frequency response.

[0034] Subsequently, the compensation spectral features are matched with a preset set of link physical damage spectral features and a set of external interference spectral features to identify the sources of equalizer parameter fluctuations. These sources include at least one of link physical degradation and external interference. For example, a pattern matching algorithm can compare the extracted compensation spectral features with a preset set of spectral features stored in memory. If the compensation spectral features exhibit a pattern of continuously increasing high-frequency attenuation compensation demand and little change in low-frequency compensation, it may match the "cable aging" pattern in the link physical damage spectral feature set. If the compensation spectral features exhibit a periodic, narrow-band frequency response compensation pattern, it may match the "electromagnetic interference" pattern in the external interference spectral feature set.

[0035] When the source is physical link degradation, a channel warning message is generated, containing channel warning information and a suggested rate step size. Based on the channel warning message and the position of the current transmission rate in a preset set of transmission rate levels, a transmission rate adjustment suggestion is generated. The transmission rate adjustment suggestion may include: reducing the current transmission rate to a level lower than the current transmission rate in the level set, or maintaining the current transmission rate at the original level and evaluating the stability of the equalizer parameters at the target level to determine whether to increase it to the target level. For example, when physical link degradation is identified, the system can select a lower rate level from the preset set of transmission rate levels as a suggestion based on the degree of degradation and the current rate level. If the degradation is minor, it may also suggest maintaining the current rate, but simultaneously performing a more rigorous equalizer parameter stability evaluation at the next target rate level to determine whether to attempt to increase the rate.

[0036] When the source of interference is external, anti-interference recommendations are generated. These recommendations may include at least one of maintaining the current transmission rate and increasing forward error correction redundancy, or enabling a data retransmission protocol. For example, when external interference is identified, the system may recommend not reducing the transmission rate, but instead enhancing the robustness of data transmission by increasing the redundancy of the forward error correction codes, or enabling a data retransmission protocol to compensate for data loss caused by interference through a retransmission mechanism.

[0037] Based on the protocol data unit format specified in the communication protocol, transmission rate adjustment suggestions or anti-interference suggestions are encapsulated into negotiation requests. These requests are sent to the other end of the data line through the negotiation channel defined by the communication protocol to negotiate and determine the operating state. The negotiation request may carry at least one of the following: target transmission rate, suggested rate step size, channel warning level, and warning confidence level. For example, a negotiation request generation module can encapsulate the generated suggestions into a specific control message according to the TLP (Transaction Layer Packet) or DLLP (Data Link Layer Packet) format of a communication protocol (such as PCIe). This message is sent to the other end device through the PCIe internal loopback negotiation channel. Upon receiving and parsing the message, the other end device responds according to its own state and strategy, thereby jointly determining the final operating state.

[0038] Finally, according to the communication protocol, link retraining is triggered based on the negotiated operating state to confirm stable operation. For example, after a new transmission rate level is determined through negotiation, the system triggers the state machine reconfiguration process specified in the communication protocol. For instance, the PCIe LTSSM (Link Training and Status State Machine) enters Recovery state and retrains the link. During retraining, the device sends specific training sequences and adjusts its physical layer parameters to ensure stable and reliable data transmission under the new negotiated operating state.

[0039] The overall working principle of this application is to continuously and meticulously monitor changes in adaptive equalizer parameters during the operation of the data line link, and to deeply analyze the channel characteristics reflected by these changes, thereby achieving early warning and accurate diagnosis of potential link problems. Traditional methods often rely on the bit error rate reaching a threshold before performing coarse rate degradation, while this application captures signs of channel changes through minute fluctuations in equalizer parameters before the bit error rate deteriorates to the point of affecting data transmission.

[0040] Scenario (PCIe Gen5 link): The CPU and GPU are connected via a PCIe Gen5x4 link, initially operating at Gen5 32GT / s. The PHY contains 5 segments of CTLE + 3 taps of DFE (h1 / h2 / h3, 7-bit signed, saturation threshold ±127).

[0041] S11, Monitoring: After the link enters the L0 state of LTSSM, the local controller periodically reads the CTLE level and DFE tap coefficients in the PHY management register. When the bit error rate is below 1e... Under condition 6, both situations of "any tap saturated but the rest unsaturated" and "all unsaturated but with continuous drift" are considered "unsaturated intervals" and are monitored online. The monitoring object is the parameter adjustment amount of the equalizer to the EQTS2 sequence (once every 256 UI).

[0042] S12, Feature Extraction: Record the CTLE tap sequence and h1 / h2 / h3 coefficient sequences for N consecutive EQTS2 cycles, and perform time-series differencing to obtain Δh1, Δh2, Δh3, and ΔCTLE. Extract three types of compensation spectral features: relative changes between taps (e.g., |Δh1| / |Δh2|), adjustment rate differences (second-order difference of Δhi / Δt), and frequency response compensation mode (coupling direction between the high-frequency CTLE tap and h1). These features correspond to the EQTS2 protocol structure—the low-frequency band is dominated by h2 / h3, while the high-frequency band is jointly dominated by h1 and CTLE.

[0043] S13, Source Identification: Perform cosine similarity matching between the current compensation spectrum and two spectrum sets. The physical damage set includes templates such as "connector aging" (h1 monotonically drifts positively, CTLE increases at high frequencies step by step, h2 / h3 remain almost unchanged); the external interference set includes templates such as "nearby crosstalk bursts" (three-tap synchronous transient jitter, CTLE remains unchanged). In this embodiment, the current feature has a similarity of 0.93 with the "connector aging" template and a similarity of 0.41 with the "nearby crosstalk bursts" template, indicating that the source is physical degradation of the link.

[0044] S14, Rate Adjustment Recommendation: Generate a channel early warning message, including warning information ("Connector aging trend, confidence level 0.93"), suggested adjustment step size ("Target Gen416GT / s"), and warning level ("Medium"). The transmission rate level set is five levels: Gen532→Gen516→Gen416→Gen48→Gen38GT / s. Generate two parallel sub-recommendations: ① Reduce to Gen416GT / s; ② Maintain Gen532GT / s and evaluate the variance of h1 / h2 / h3 at the Gen416GT / s target level to determine whether to increase. Controller selection ①.

[0045] S15, Anti-interference branch: If the source of interference in this example is not external interference, skip it; if it is external interference, generate at least one of the following: maintain the current rate + increase FEC redundancy / enable CPLRetry retransmission.

[0046] S16, negotiation request encapsulation and transmission: The rate adjustment suggestion is encapsulated as a PCIeVDMType0 negotiation request, the VendorID is filled with a custom value, and the payload is written with the target rate, suggested step size, warning level, and confidence level in a custom format. It is sent to the peer through the PCIe transaction layer VDM channel. The peer sends a completion message to confirm, and the negotiation determines the working status as "switch to Gen416GT / s and enter Recovery for retraining".

[0047] S17, Retraining and Stability Confirmation: The local LTSSM switches to Recovery and sends TS1 / TS2 / EQTS1 / EQTS2 at Gen4 rate to complete PHY rebalancing. After returning to L0, continuously observe the variance of h1 / h2 / h3 for 1024 EQTS2 cycles. If it is lower than the stability threshold, it is confirmed that "the working state can work stably". If unstable, enter the bypass state of reduced speed to maintain transmission or data buffering according to weight 5, triggering a new round of negotiation or continuing to reduce to Gen4 8GT / s to re-execute.

[0048] In summary, this application constructs an intelligent and adaptive data line transmission rate adjustment method by introducing early monitoring of equalizer parameters, extraction and matching of compensation spectral features, and a refined suggestion generation and protocol negotiation mechanism based on the source of the problem. This not only solves the problems of inflexible rate adjustment and inability to adapt to cable physical defects and external interference in traditional solutions, but also significantly improves the stability and efficiency of data transmission, providing users with a better experience.

[0049] Specifically, in some implementations of the above methods, in order to more clearly define and implement the specific forms of the communication protocol, the non-link reconstruction phase of the state machine, the training signal, the negotiation channel, and the link retraining, this application further defines these technical features.

[0050] In the above method, the communication protocol includes the PCIe communication protocol and / or the USB communication protocol; the non-link reconstruction phase of the state machine specified by the communication protocol includes the L0 / L0s state of PCIe LTSSM and / or the U0 state of USB; the training signal specified by the communication protocol includes the PCIe TS1 / TS2 sequence and / or the USB Polling.LFPS signal; the negotiation channel specified by the communication protocol includes the PCIe internal loopback negotiation channel and / or the USB protocol layer control channel; the link retraining includes the state machine reconfiguration process specified by the communication protocol.

[0051] Specifically, the communication protocols mentioned can be understood as industry standards widely used in data line transmission, such as the PCIe (Peripheral Component Interconnect Express) communication protocol or the USB (Universal Serial Bus) communication protocol. These protocols define the physical layer, data link layer, and transaction layer specifications for data transmission, providing a basic framework for high-speed data transmission.

[0052] The non-link reconstruction phase of the state machine refers to the state where the link is functioning normally and no full link initialization or reconnection is required. For the PCIe communication protocol, this typically corresponds to the L0 (Active State) or L0s (Low Power Standby State) states in its Link Training and State Machine (LTSSM), indicating that the link is in normal data transmission or low-power data transmission mode. For the USB communication protocol, this can correspond to the U0 state, i.e., the normal operating state. Monitoring equalizer parameters in these states can avoid introducing additional complexity or interrupting normal data flow during link reconstruction.

[0053] Furthermore, the training signals are specific signal sequences used in communication protocols for link training and equalizer adjustment. In the PCIe communication protocol, this could be a TS1 (Training Sequence 1) or TS2 (Training Sequence 2) sequence, which contains specific code patterns that allow the receiver to adjust its equalizer to optimize signal quality. In the USB communication protocol, Polling.LFPS (Low Frequency Periodic Signaling) signals can be used for similar training purposes, helping the receiver evaluate and adjust link characteristics. By monitoring the equalizer's response to these standard training signals, the real-time status of the link can be accurately obtained.

[0054] Furthermore, the negotiation channel refers to a dedicated logical or physical channel in a communication protocol used to exchange control information and negotiate operating parameters between devices at both ends of a data line. In the PCIe communication protocol, the internal loopback negotiation channel can be used to negotiate parameters without affecting the main data path. In the USB protocol, the protocol layer control channel provides a mechanism for transmitting control commands and status information between the device and the host. Through these negotiation channels, rate adjustment suggestions or anti-interference suggestions can be reliably transmitted and agreed upon with the peer device.

[0055] Therefore, the link retraining can be understood as the process of re-initializing the link and reconfiguring its parameters based on the new working state determined through negotiation. In communication protocols such as PCIe and USB, this typically involves a state machine reconfiguration process, such as transitioning from the L0 state to other training states and then returning to the L0 state, to ensure that the link can work stably and efficiently under new transmission rates or anti-interference strategies.

[0056] This application's solution explicitly specifies the specific type of communication protocol (such as PCIe or USB) and its corresponding state machine non-link reconstruction phase, training signals, and negotiation channels, enabling the aforementioned data line transmission rate adjustment method to be directly applied to existing mainstream high-speed data transmission interfaces. For example, in a PCIe link, by monitoring the equalizer's adjustment of the TS1 / TS2 sequence in the L0 / L0s state, the link health status can be assessed using the mechanism provided by the protocol itself, without introducing additional, non-standardized detection methods. Similarly, the USB protocol's U0 state and Polling.LFPS signal also provide similar standardized interfaces. This specificity allows the method to be seamlessly integrated into existing hardware and software stacks, utilizing existing protocol specifications for parameter monitoring, negotiation, and link retraining, thereby ensuring the feasibility and compatibility of the solution. By utilizing the protocol's internal negotiation channels, interference with the main data path can be avoided, ensuring the efficiency and reliability of the negotiation process.

[0057] This application further proposes the above method, characterized in that the unsaturated interval includes a mixed state in which at least one tap coefficient of the adaptive equalizer has reached a saturation threshold and other tap coefficients have not reached a saturation threshold; the method further includes: when any tap coefficient reaches a saturation threshold, triggering the monitoring step, and taking the direction of the saturated tap coefficient as the priority extraction direction of the compensation spectral feature.

[0058] Specifically, the aforementioned unsaturated region is defined as a mixed state, where at least one tap coefficient of the adaptive equalizer has reached its preset saturation threshold, while the other tap coefficients have not yet reached their saturation thresholds. This definition expands the understanding of the "unsaturated region" to encompass channel states that, while not fully saturated overall, show signs of local saturation. Reaching the saturation threshold for a tap coefficient typically means that the tap coefficient can no longer compensate the signal significantly, indicating that the channel impairment in its corresponding frequency or delay components has reached the equalizer's compensation limit.

[0059] Furthermore, when any tap coefficient in the adaptive equalizer reaches its saturation threshold, a monitoring step of the equalizer parameters is immediately triggered. This triggering mechanism ensures that detailed parameter monitoring can be initiated promptly when a local critical state occurs in the channel, without waiting for the overall equalizer to reach saturation.

[0060] Furthermore, once a tap coefficient reaches the saturation threshold, the direction in which the saturated tap coefficient is located will be prioritized for extracting the compensation spectral features. This means that when analyzing the temporal variations of equalizer parameters and extracting compensation spectral features, more emphasis will be placed on the frequency response or delay compensation mode related to the saturated tap coefficient, thereby more accurately locating and analyzing the specific channel problem that caused the tap coefficient to saturate.

[0061] This application's solution effectively addresses the issue of insufficient sensitivity to local channel degradation inherent in basic schemes by redefining the unsaturated region and introducing a specific monitoring triggering mechanism. Specifically, the definition of the unsaturated region includes a "mixed state where at least one tap coefficient has reached the saturation threshold while other tap coefficients have not," enabling the system to identify areas where compensation has reached their limits before channel damage causes overall equalizer saturation. This more refined state identification allows for immediate triggering of monitoring steps when any tap coefficient reaches the saturation threshold, avoiding potential early warning opportunities missed while waiting for overall saturation. Furthermore, prioritizing the direction of saturated tap coefficients for extracting compensation spectral features allows subsequent feature analysis to focus on the most affected frequency bands or delay components in the channel, improving the targeting and accuracy of compensation spectral feature extraction and providing a more reliable basis for subsequent identification of link physical damage or external interference.

[0062] In some preferred embodiments, it is assumed that a PCIe data link is operating stably at Gen4 rate. The adaptive equalizer at the data line receiver contains multiple tap coefficients to compensate for channel loss. At some point, due to increased ambient temperature or a slight loosening of a connector, the response at a specific frequency of the link deteriorates slightly. At this time, the data error rate has not yet reached a preset error threshold, and most of the equalizer tap coefficients are still unsaturated. However, a pre-emphasis tap coefficient responsible for compensating for this specific frequency component gradually increases due to the need for greater compensation, eventually reaching its preset saturation threshold.

[0063] According to the scheme of this application, even if other tap coefficients are not yet saturated, the system will immediately identify the "mixed state" currently in the "unsaturated interval" because this tap coefficient has reached the saturation threshold. At this time, the monitoring step is triggered, and detailed monitoring of the equalizer parameters corresponding to the PCIe TS1 / TS2 training signals adjusted by the adaptive equalizer begins. At the same time, since the pre-taper coefficient is saturated, its direction (e.g., the direction corresponding to high-frequency compensation) is designated as the priority extraction direction for compensation spectral features. In the subsequent feature extraction process, the system will pay more attention to the relative changes and adjustment rate differences between the tap coefficients related to high-frequency compensation, so as to identify physical degradation in the link (e.g., increased high-frequency loss) earlier and more accurately, and generate corresponding channel warning information and transmission rate adjustment suggestions, such as suggesting that the transmission rate be reduced to Gen3 to avoid potential data transmission errors.

[0064] This application further proposes that the steps for generating the above-mentioned transmission rate adjustment suggestions also include: The channel warning information is subjected to de-jitter processing; within the preset confirmation window corresponding to the non-link reconstruction phase of the state machine specified in the communication protocol, the channel warning information continuously meets the warning conditions, and only then is the transmission rate adjustment suggestion encapsulated as a negotiation request. And / or, Record the timestamp of the most recent negotiation request; suppress the generation of new negotiation requests within the preset suppression window corresponding to the non-link reconstruction phase of the state machine specified in the communication protocol.

[0065] Specifically, de-jittering of channel warning information means that after identifying link physical degradation as the source of equalizer parameter fluctuations and generating channel warning information, the system does not immediately use it to generate a negotiation request. Instead, the system waits for a period of time, i.e., a preset confirmation window. Within this confirmation window, if the channel warning information continuously meets the warning conditions, it is considered stable and reliable, and the transmission rate adjustment suggestion is then encapsulated as a negotiation request. This de-jittering process aims to filter out transient or unstable channel conditions, avoiding unnecessary rate adjustments triggered by brief fluctuations.

[0066] The timestamp of the most recent negotiation request refers to the specific time the request was sent after it was dispatched. This timestamp is used to determine whether to suppress the generation of new negotiation requests. The preset suppression window is a predetermined time period during which the system will suppress the generation of new negotiation requests even if the conditions for generating a negotiation request are met. Its purpose is to prevent frequent negotiation initiation within a short period, thereby reducing protocol overhead, avoiding frequent link state switching, and improving system stability.

[0067] This application's solution effectively addresses the system instability and efficiency degradation caused by instantaneous fluctuations in channel conditions or excessively frequent negotiation requests by introducing jitter reduction processing and negotiation request suppression mechanisms. Through the above technical solutions, this application can significantly improve the robustness and efficiency of data line transmission rate adjustment.

[0068] In some preferred embodiments, a specific example is given below. Assume the data line is operating in PCIe L0 mode. The adaptive equalizer detects fluctuations in its tap coefficients and identifies the source of these fluctuations as physical link degradation, generating a channel warning message. Without jitter reduction, even if the warning message is caused by a brief external electromagnetic interference, the system might immediately generate and send a negotiation request to reduce the transmission rate. However, by introducing jitter reduction, the system sets an acknowledgment window, for example, 500 milliseconds. If the channel warning message persists and meets the warning conditions within these 500 milliseconds, the system will confirm that this is a genuine degradation requiring a response, and then encapsulate the transmission rate adjustment suggestion as a negotiation request.

[0069] Furthermore, suppose the system sends a negotiation request at a certain moment, requesting a reduction in the transmission rate from Gen4 to Gen3. The system records the timestamp of this request. To avoid sending another negotiation request within a short period, the system sets a suppression window, for example, of 5 seconds. If, within the next 5 seconds, link degradation is detected again and a channel warning message is generated (and jitter reduction is implemented), the system will suppress the generation of new negotiation requests. Only after the 5-second suppression window expires will the system consider generating and sending the next negotiation request. This mechanism ensures that the system does not experience frequent link training due to frequent negotiations, thus maintaining the stability and efficiency of data transmission.

[0070] This application further proposes that, in accordance with the aforementioned communication protocol, the step of triggering link retraining based on the negotiated working state includes: During the negotiation and the link retraining, the data stream is switched to a preset bypass state according to the communication protocol; the bypass state includes at least one of data buffering, slowing down transmission, and skipping training frames. If the negotiation fails to reach an agreement or the link retraining fails to confirm that the working state is stable, the system switches to the bypass state for reduced transmission or data caching according to the fallback procedure specified in the communication protocol, and continues to trigger negotiation or reduces to the next lower level before re-executing the negotiation and the link retraining.

[0071] Specifically, during negotiation and link retraining, the data stream is switched to a preset bypass state. The bypass state is a temporary data processing mode designed to prevent data loss or corruption during link parameter adjustments. This bypass state can include at least one of data buffering, rate-reduction transmission, or skipping training frames. Data buffering can be understood as temporarily storing data to be transmitted until the link stabilizes, ensuring data integrity. Rate-reduction transmission can be understood as continuing to transmit some critical data at a lower rate during negotiation and retraining to ensure uninterrupted basic services, maintaining basic business continuity. Skipping training frames can be understood as temporarily stopping the transmission of non-critical data or data with low real-time requirements during link retraining to avoid conflicts with training signals or data errors caused by link instability, optimizing training efficiency and avoiding data conflicts.

[0072] Furthermore, if the negotiation fails to reach an agreement—that is, the data line peer fails to accept or acknowledge the negotiation request, or the link retraining fails to confirm that the negotiated working state can operate stably—the system will handle the situation according to the fallback procedure specified in the communication protocol. This fallback procedure includes switching to a bypass state with reduced transmission speed or data buffering to ensure minimum link availability. Based on this, the system will continue to trigger the negotiation process, attempting to re-reach an agreement with the peer, or reducing the transmission rate to a lower level, and then re-execute the negotiation and link retraining until a stable working state is found.

[0073] This application's solution effectively isolates the data flow from the link configuration adjustment process by introducing a bypass state during the critical transition period of negotiation and link retraining. When link parameters change, the data flow is no longer directly exposed to the unstable link environment, but is protected through buffering, reduced transmission speed, or temporary suspension of transmission. It is precisely because of this isolation mechanism that even if the link experiences temporary instability during adjustment, data loss or corruption can be avoided. Furthermore, when negotiation or retraining fails, through a pre-defined fallback procedure, the system can automatically switch to a bypass state such as reduced transmission speed or data buffering, and attempt to renegotiate or reduce the rate level, thereby ensuring the resilience and reliability of the link and preventing service interruption due to a single failure.

[0074] In some preferred embodiments, assuming the data line is currently operating at PCIe Gen4 x4, and link physical degradation is identified based on equalizer parameter fluctuations, requiring a reduction in transmission rate to PCIe Gen3 x4, the data stream is switched to a data buffer bypass state to avoid data loss during the negotiation request sent to the peer and the triggering of link retraining. At this time, all packets to be transmitted are temporarily stored in the sender's buffer until link retraining is complete and PCIe Gen3 x4 is confirmed to be stable. If the peer rejects the negotiation request, or link retraining fails to establish a PCIe Gen3 x4 connection, the system will follow a fallback procedure, switching the data stream to a reduced-speed, transmission-preserving bypass state, for example, continuing to transmit critical data at a PCIe Gen2 x4 rate. Simultaneously, the system will trigger negotiation again, attempting to negotiate and retrain at a PCIe Gen3 x4 or lower PCIe Gen2 x4 rate until the link stabilizes. As a specific implementation method, when link retraining requires a long time, a bypass state that skips training frames can be adopted. That is, during the transmission of training sequence, the transmission of non-critical data streams is suspended and resumed after training is completed, so as to ensure the purity and efficiency of the training process.

[0075] The steps for generating transmission rate adjustment recommendations also include: Based on the channel warning information and the critical point information corresponding to each level in the transmission rate level set, determine the rate reduction step size or evaluate the strictness of equalizer parameter stability; wherein, the critical point information includes at least one of the following: communication protocol handshake overhead corresponding to each level, communication protocol state transition cost between adjacent levels, and link retraining time required when each level is the negotiation target.

[0076] Specifically, channel warning information refers to a message generated when physical degradation of the link is detected. It contains warning information about the channel condition and a suggested rate adjustment step size. The transmission rate class set can be understood as an ordered list of all possible transmission rates supported by the data line, such as the different rate classes Gen1, Gen2, Gen3, and Gen4 in the PCIe communication protocol.

[0077] Critical point information is a key parameter used to quantify the cost and complexity required to switch between different transmission rate levels or maintain a certain level. Communication protocol handshake overhead refers to the time or resources consumed by the protocol layer during the handshake process when transitioning between different transmission rate levels. For example, switching from one PCIe Gen level to another may require a series of link training and state machine transitions, which introduce latency. The communication protocol state transition cost between adjacent levels refers to the computational resources, time overhead, and potential data interruption risk required for protocol state machine transitions when switching from the current transmission rate level to an adjacent higher or lower level. The link retraining time required when each level is used as a negotiation target refers to the time consumed by the link retraining process after the data line ends have negotiated and determined a target transmission rate level, in order to ensure stable operation of the link at that level. This critical point information can be pre-stored in the system or dynamically updated and learned through statistical data from actual operation.

[0078] This application's solution introduces critical point information, making the generation process of transmission rate adjustment suggestions more intelligent and refined. When the system needs to reduce the transmission rate, it no longer simply reduces the rate in fixed steps. Instead, it dynamically calculates the optimal rate reduction step size based on the degradation level indicated by the current channel warning information, combined with critical point information such as the communication protocol handshake overhead, state transition cost, and link retraining time for different rate levels. For example, if the handshake overhead and retraining time for reducing the rate by one level are small, while the overhead for reducing it by two levels is significantly increased, the system may prioritize reducing it by one level. Similarly, when the system decides to maintain the current transmission rate and evaluate the stability of the equalizer parameters at the target level, critical point information can be used to determine the rigor of the evaluation. If the link retraining time at the target level is long, or the state transition cost from the current level to the target level is high, the system will adopt stricter evaluation criteria to ensure that the link has higher stability at the target level before making any actual upgrades, thereby avoiding link instability or frequent retraining caused by blind upgrades.

[0079] In some preferred embodiments, a specific example is given below. Assume the data line is currently operating at PCIe Gen4 rate level, and the system detects fluctuations in equalizer parameters, identifies physical degradation in the link, and generates a channel warning. At this point, the system needs to generate a transmission rate adjustment suggestion.

[0080] First, the system queries a preset set of transmission rate levels, which may include levels such as Gen4, Gen3, and Gen2, and then queries the critical point information corresponding to each level. For example, the system may learn that: 1. The communication protocol handshake overhead and state transition cost are relatively low when downgrading from Gen4 to Gen3, and the link retraining time is also shorter.

[0081] 2. Downgrading directly from Gen4 to Gen2 incurs higher handshake overhead and state transition costs, resulting in a significant increase in link retraining time.

[0082] 3. Evaluate the stability of the equalizer parameters at Gen3 level. If the conditions are met, you can try to upgrade to Gen4, but the increased handshake overhead and retraining time should also be considered.

[0083] Based on this critical information, the system makes decisions. If the channel warning information indicates a manageable level of degradation, the system may prioritize recommending a reduction of the current transmission rate to Gen3 level, as its switching cost is low and link stability can be quickly restored. If the channel warning information indicates a severe level of degradation, but not enough to directly reduce to the lowest level, the system may recommend maintaining the current Gen4 level, while simultaneously conducting a longer-term or more stringent equalizer parameter stability assessment at the target Gen3 level to determine whether an actual rate reduction or an attempt to increase the rate is necessary. In this way, the system can make more targeted and efficient rate adjustment decisions based on the actual situation and protocol characteristics.

[0084] In some of the embodiments described above in this application, although a scheme for dynamically adjusting the data line transmission rate based on channel conditions has been proposed, and the step size and evaluation strictness of the rate adjustment have been considered, in practical applications, negotiating and adjusting the transmission rate only at the link layer may not fully meet the overall system requirements for Quality of Service (QoS). If the negotiation results at the link layer are not effectively linked with the upper-layer QoS protocol, the upper-layer application may not be able to perceive changes in the link status in a timely manner, thus failing to optimize traffic allocation or service priority, affecting overall system performance and user experience. Therefore, this application further proposes a scheme to report the negotiation results to the upper-layer QoS protocol to achieve collaborative optimization between the link layer and the upper-layer application.

[0085] The above methods also include: The negotiation results are reported to the upper-layer Quality of Service (QoS) protocol; the upper-layer QoS protocol adjusts traffic allocation or service priority according to the negotiated transmission rate; the upper-layer QoS protocol includes PCIe Virtual Channel Protocol and / or IETF Differentiated Services Protocol.

[0086] Specifically, after the data line ends negotiate and determine the working state and trigger link retraining, the obtained negotiation results, such as the final determined transmission rate and channel warning level, are reported to the upper-layer Quality of Service (QoS) protocol. The upper-layer QoS protocol refers to the protocol or mechanism responsible for managing and optimizing the QoS of different types of data flows in the network. Its main purpose is to ensure that critical services or high-priority data can obtain the required bandwidth, latency, and jitter performance. For example, the upper-layer QoS protocol can dynamically adjust the traffic allocation of different service flows or reset the priority of each service based on the negotiated transmission rate. Traffic allocation refers to allocating different bandwidth resources to different data flows to avoid network congestion and ensure QoS; service priority refers to setting different processing priorities for different data flows to ensure that high-priority services can obtain network resources first. As a preferred implementation, the upper-layer QoS protocol may include the PCIe Virtual Channel Protocol and / or the IETF Differentiated Services Protocol. The PCIe Virtual Channel Protocol is a mechanism in the PCIe standard used to support different QoS levels, achieving traffic isolation and priority management by allocating independent virtual channels to different types of traffic. The IETF Differentiated Services (DiffServ) is an IP layer quality of service architecture that enables network devices to differentiate traffic based on the differentiating service code points (DSCPs) marked in the IP packet header.

[0087] This application's solution addresses the disconnect between link-layer rate adjustment and upper-layer application service quality management by promptly reporting the negotiation results to the upper-layer service quality protocol after the link-layer transmission rate negotiation is completed. Because the upper-layer service quality protocol can obtain real-time information on link-layer transmission rate changes, it is able to intelligently and dynamically adjust traffic allocation and service priorities based on the latest link conditions.

[0088] In some preferred embodiments, a specific example is given below. Suppose that in a system employing the PCIe communication protocol, the adaptive equalizer at the data line receiver detects physical link degradation and, after negotiation, determines that the transmission rate needs to be reduced from Gen4 to Gen3. After the negotiation is complete, this result—the transmission rate has been adjusted to Gen3—is reported to the PCIe Virtual Channel Protocol. As a higher-layer Quality of Service (QoS) protocol, the PCIe Virtual Channel Protocol will reassess and adjust the bandwidth allocation of different virtual channels (e.g., one virtual channel for real-time video streaming and another for file transfer) based on this new transmission rate information. Specifically, the Virtual Channel Protocol may reduce the available bandwidth of the file transfer virtual channel to ensure that the real-time video stream still receives sufficient bandwidth and priority at a lower link rate, thereby avoiding video stuttering or latency.

[0089] For another example, in a system using the USB communication protocol, when the data line transmission rate needs to maintain the current rate but increase forward error correction redundancy due to external interference, this negotiation result will be reported to the IETF Differentiated Services Protocol. Based on the information of "increasing forward error correction redundancy" from the link layer, the Differentiated Services Protocol can interpret this as a decrease in link quality but still a need to maintain the current rate. Therefore, it may more strictly mark certain packet loss-sensitive service traffic, such as adjusting its DSCP value to a higher priority or allocating it to a smaller queue in network devices, to ensure that these critical data packets can still pass through with priority even when link quality is poor, thereby maintaining service quality.

[0090] In some embodiments described above, a data line transmission rate adjustment method based on a communication protocol is proposed, which can dynamically adjust the transmission rate according to link conditions. However, during the generation and transmission of negotiation requests and subsequent link retraining, the system may undergo temporary state transitions and resource scheduling. If power consumption during these processes is not effectively managed, unnecessary energy consumption may occur, especially in scenarios involving frequent rate adjustments or link state transitions. Therefore, this application further proposes a scheme to synchronously coordinate the power management protocol during negotiation request generation to achieve more efficient energy utilization.

[0091] Specifically, the method further includes: When the negotiation request is generated, a negotiation preparation notification is simultaneously sent to the power management protocol; the power management protocol enters the power saving preparation state according to the negotiation preparation notification; the power management protocol includes the PCIeASPM protocol and / or the USB U2 / U3 protocol.

[0092] When the system identifies the source of equalizer parameter fluctuations based on link conditions and generates transmission rate adjustment suggestions or anti-interference suggestions, these suggestions are encapsulated as negotiation requests. Simultaneously with the generation of these negotiation requests, a negotiation preparation notification is sent to the system's internal power management protocol. The power management protocol can be understood as a software or hardware module responsible for managing the power consumption of the system or device. Its purpose is to dynamically adjust the power supply based on system load and operating status to achieve energy saving. For example, the power management protocol may include the PCIeASPM protocol or the USB U2 / U3 protocol. The PCIeASPM protocol is a mechanism used by the PCIe bus to manage the power state of the link, allowing the link to enter a low-power state when idle. The USB U2 / U3 protocol defines two low-power states in the USB protocol, used to reduce power consumption when the device is inactive. Upon receiving the negotiation preparation notification, the power management protocol enters an energy-saving preparation state based on the notification. Power-saving readiness refers to the power management protocol pre-adjusting its internal parameters or states to prepare for upcoming link state changes (such as link retraining) and manage power consumption. For example, it can reduce the power supply of certain non-critical modules in advance or prepare to quickly enter a low-power mode during link retraining.

[0093] The solution proposed in this application synchronously notifies the power management protocol when the negotiation request is generated, enabling the power management module to anticipate upcoming link state changes, thereby avoiding potential power management lag issues during link negotiation and retraining.

[0094] In some preferred embodiments, a specific example is given below. Suppose a PCIe data line is transmitting data, and the adaptive equalizer parameters at its receiving end show signs of physical link degradation. The system therefore generates a transmission rate adjustment suggestion, recommending a reduction of the current transmission rate from Gen4 to Gen3. Before encapsulating this suggestion into a negotiation request and sending it to the peer, the system synchronously sends a negotiation preparation notification to the PCIeASPM protocol. Upon receiving this notification, the PCIeASPM protocol pre-enters a power-saving preparation state. For example, it may assess which non-core components can temporarily reduce power consumption during link retraining, or prepare to quickly switch the link to a lower L0s or L1 power state after successful speed reduction and stabilization. Once the negotiation request is sent and a new operating state (Gen3) is successfully negotiated with the peer, the link triggers retraining. During this retraining, because the PCIeASPM protocol has been prepared in advance, it can manage power consumption more effectively. For example, during training sequence transmission, it can ensure that non-essential power domains are in a minimum power state, thereby reducing energy consumption throughout the rate adjustment process.

[0095] In some embodiments of this application, the aforementioned preset set of link physical impairment patterns and the aforementioned preset set of external interference patterns are adaptively updated according to the link operating state; the aforementioned adaptive update is performed during the non-link reconstruction phase of the state machine specified in the aforementioned communication protocol; the aforementioned adaptive update includes: When the compensation spectral feature corresponding to the above-mentioned channel warning information does not match any of the spectral features in the above-mentioned link physical damage spectral feature set and the above-mentioned external interference spectral feature set, the compensation spectral feature corresponding to the above-mentioned channel warning information is added as a new spectral feature to the corresponding spectral feature set. When the link meets the stability condition within the preset observation window after the above anti-interference suggestion is triggered, the corresponding compensation spectral feature will be added to the above external interference spectral set as the above new spectral feature.

[0096] Specifically, the preset set of physical link damage patterns and the preset set of external interference patterns are reference pattern libraries used to identify the sources of equalizer parameter fluctuations. In practical applications, these pattern sets are not static but need to be dynamically adjusted and expanded according to the actual operating state of the data line. The adaptive update mechanism ensures that these pattern sets can continuously learn and adapt to new link conditions. The state machine specified in the communication protocol refers to the logical framework for switching and managing the state of the data line between different operating modes, such as the LTSSM of PCIe or the state machine of USB. Performing update operations during the non-link reconstruction phase of these state machines ensures that the update process remains synchronized with the normal operation and state transitions of the link, avoiding unnecessary interference to data transmission.

[0097] Furthermore, when the system detects channel warning information and extracts the corresponding compensation spectral features, it attempts to match them with existing sets of link physical damage spectral features and external interference spectral features. If the compensation spectral feature does not match any known spectral feature in the existing sets, it indicates that this may be a new, unrecorded link problem pattern. In this case, to enhance the system's identification capability, the new compensation spectral feature is added as a new spectral feature to the corresponding spectral set. For example, if the problem is judged to be more likely to be physical damage based on other contexts, it is added to the link physical damage spectral set; otherwise, it is added to the external interference spectral set.

[0098] Furthermore, once the system generates and triggers anti-interference suggestions in response to identified external interference, it enters a preset observation window. Within this window, if the link meets stability conditions (e.g., bit error rate returns to normal, equalizer parameter fluctuations stabilize), it indicates that the previously identified compensation spectral characteristics were indeed caused by external interference, and the anti-interference measures are effective. At this point, to enhance the system's ability to identify this type of external interference, the corresponding compensation spectral characteristics are added as a new spectral type to the external interference spectral type set. This helps the system to identify and take appropriate anti-interference measures more quickly and accurately when encountering similar interference in the future.

[0099] This application's solution effectively addresses the limitations of traditional methods where the preset spectrum set is static and unable to cope with unknown or dynamically changing link conditions by introducing an adaptive update mechanism for the spectrum set. Through the above technical solution, this application can significantly improve the accuracy and adaptability of the data line transmission rate adjustment system in identifying physical damage to the link and external interference sources.

[0100] For example, on the CPU-GPU PCIe Gen5 link, the controller maintains two spectrum sets: a physical damage set (containing templates such as "connector aging", "PCB trace aging", and "via poor contact") and an external interference set (containing templates such as "nearby crosstalk burst" and "periodic EMI"). Each template is a compensated spectrum feature vector, label, and weight.

[0101] Scenario 1 (Mismatch → Add): During a large GPU training task in the third month, transient synchronization jitter strongly correlated with business traffic was detected in h1 / h2 / h3, with the CTLE level remaining unchanged. This feature has a similarity of 0.62 with "nearby crosstalk burst" and a similarity of less than 0.5 with all templates in the physical damage set, indicating a mismatch with both sets. According to the first strategy (weight 9), this feature is added to the "to be observed" area of ​​the external interference set within L0 / L0s and labeled "unconfirmed".

[0102] Scenario 2 (Observation Passed → Migration): A new feature emerges in the ninth month: h1 experiences a step jump synchronized with a fixed business cycle, and CTLE level stabilizes after a one-time jump. The similarity to all templates in both sets is below 0.9, so it is added to the "Under Observation" area as in Scenario 1. Anti-interference recommendations are triggered (maintain current rate, improve FEC, enable retransmission). Within the 30-day observation window, there are no new alarms on the link, and the tap variance remains below the stable threshold. Using the second strategy (weight 9), this feature is migrated from the "Under Observation" area to the "Confirmed" area of ​​the external interference set and assigned a formal template number.

[0103] Results: The spectral set has been expanded to include two new types of spectral patterns without relying on external configuration: "periodic business-related interference" and "step-like periodic crosstalk". Subsequent fluctuations of the same type can directly hit the template, and the recognition accuracy continues to improve over time.

[0104] In some preferred embodiments, assuming the data line operates under a specific environment and a novel type of electromagnetic interference causes the tap coefficients of the adaptive equalizer to exhibit a unique fluctuation pattern never recorded in the preset external interference spectrum set. In this case, the system monitors the abnormal fluctuations in the equalizer parameters and extracts the corresponding compensation spectrum features. When these compensation spectrum features do not match any spectrum in the existing set of link physical damage spectrum features or the external interference spectrum set, the system identifies it as a new, unknown interference pattern. Based on its characteristics, the system initially determines it to be external interference and adds it as a new spectrum to the external interference spectrum set. Subsequently, the system generates and triggers corresponding anti-interference suggestions, such as increasing forward error correction redundancy. Within a preset observation window, if the link's bit error rate significantly decreases and stabilizes, the system confirms that the new spectrum indeed represents external interference and permanently retains it in the external interference spectrum set. In this way, when the system encounters the same type of electromagnetic interference in the future, it can directly match the learned spectrum, thereby quickly and accurately identifying the source of interference and immediately taking effective anti-interference measures without having to go through the "unknown pattern recognition" process again, which greatly improves the response efficiency and the level of system intelligence.

[0105] refer to Figure 2 , Figure 2 This is a schematic diagram of a data line transmission rate adjustment system based on a communication protocol provided in an embodiment of the present invention, comprising: The parameter monitoring module is used to monitor the equalizer parameters adjusted by the adaptive equalizer for the training signal specified by the communication protocol during the non-link reconstruction phase of the state machine specified by the communication protocol, when the data error rate has not reached the preset error threshold and the tap coefficients of the adaptive equalizer at the data line receiving end are working in the unsaturated range during the data line link operation process. The feature extraction module is used to analyze the temporal changes of the equalizer parameters within one or more complete transmission cycles of the training signal, and extract the compensation spectral features corresponding to the protocol structure of the training signal; the compensation spectral features include at least two of the following: the relative change between each tap coefficient, the difference in the adjustment rate of each tap coefficient, and the compensation mode for the signal frequency response. The source identification module is used to match the compensation spectrum features with a preset set of link physical damage spectrum features and a set of external interference spectrum features to identify the source of the equalizer parameter fluctuations; the source includes at least one of link physical degradation and external interference; The negotiation request generation module is configured to generate a channel warning message containing channel warning information and a suggested rate step size, and generate a transmission rate adjustment suggestion when the source is link physical degradation; and to generate an anti-interference suggestion when the source is external interference. The transmission rate adjustment suggestion includes reducing the current transmission rate to a level lower than the current transmission rate, and / or maintaining the current transmission rate at the original level and evaluating the stability of the equalizer parameters at the target level to determine whether to increase it to the target level; the anti-interference suggestion includes maintaining the current transmission rate and increasing forward error correction redundancy, and enabling a data retransmission protocol, at least one of the following: the negotiation request carries at least one of the following: target transmission rate, suggested rate step size, channel warning level, and warning confidence level. The negotiation training module is used to encapsulate the transmission rate adjustment suggestion or the anti-interference suggestion into a negotiation request according to the protocol data unit format specified by the communication protocol, send it to the data line peer through the negotiation channel specified by the communication protocol, negotiate with the data line peer to determine the working state; and trigger link retraining according to the working state determined by the negotiation according to the communication protocol to confirm that the working state can work stably.

[0106] This application's system, through a modular design, achieves intelligent and dynamic adjustment of the data line transmission rate. The parameter monitoring module is responsible for acquiring subtle changes in the link status in real time, while the feature extraction module transforms these changes into identifiable compensation spectral features. Subsequently, the source identification module can accurately determine whether the problem originates from physical link degradation or external interference. Based on this, the negotiation request generation module can generate targeted transmission rate adjustment suggestions or anti-interference suggestions, and through the negotiation training module, conduct efficient negotiation and link retraining with the peer device. This allows for intervention before the bit error rate deteriorates, effectively solving the problems of inflexible rate adjustment and inability to adapt to cable physical defects and external interference in traditional solutions, significantly improving data transmission stability and user experience.

[0107] In summary, the system of this application, through its modular design and intelligent adjustment mechanism, achieves dynamic, precise, and intelligent adjustment of the data line transmission rate, effectively solving the problems of inflexible rate adjustment and inability to adapt to cable physical defects and external interference in traditional solutions. It significantly improves the stability and efficiency of data transmission, providing users with a better experience.

[0108] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for adjusting the data line transmission rate based on a communication protocol, characterized in that, include: During the operation of the data line link, when the data bit error rate has not yet reached the preset error threshold and the tap coefficients of the adaptive equalizer at the data line receiver are working in the unsaturated range, during the non-link reconstruction phase of the state machine specified by the communication protocol, the equalizer parameters adjusted by the adaptive equalizer for the training signal specified by the communication protocol are monitored. The timing variations of the equalizer parameters within one or more complete transmission cycles of the training signal are analyzed, and compensation spectral features corresponding to the protocol structure of the training signal are extracted. The compensation spectral features include at least two of the following: the relative changes between tap coefficients, the differences in the adjustment rates of tap coefficients, and the compensation modes for the signal frequency response. The compensation spectrum features are matched with a preset set of link physical damage spectrum features and a set of external interference spectrum features to identify the source of the equalizer parameter fluctuations; the source includes at least one of link physical degradation and external interference. When the source is link physical degradation, a channel early warning message containing channel warning information and a suggested rate step size is generated. Based on the channel warning message and the position of the current transmission rate in the preset transmission rate level set, a transmission rate adjustment suggestion is generated; The transmission rate adjustment recommendations include: reducing the current transmission rate to a level in the set of levels that is lower than the current transmission rate, or maintaining the current transmission rate at the original level and evaluating the stability of the equalizer parameters at the target level to determine whether to increase it to the target level. When the source is external interference, an anti-interference suggestion is generated; the anti-interference suggestion includes at least one of maintaining the current transmission rate and increasing forward error correction redundancy, and enabling the data retransmission protocol. According to the protocol data unit format specified in the communication protocol, the transmission rate adjustment suggestion or the anti-interference suggestion is encapsulated into a negotiation request; the negotiation request is sent to the other end of the data line through the negotiation channel specified in the communication protocol to negotiate and determine the working state with the other end of the data line; the negotiation request carries at least one of the following: target transmission rate, suggested rate step size, channel warning level, and warning confidence level; According to the communication protocol, link retraining is triggered based on the negotiated working state to confirm that the working state can work stably.

2. The method according to claim 1, characterized in that, The communication protocol includes the PCIe communication protocol and / or the USB communication protocol; the non-link reconstruction phase of the state machine specified by the communication protocol includes the L0 / L0s state of PCIe LTSSM and / or the U0 state of USB; the training signal specified by the communication protocol includes the PCIe TS1 / TS2 sequence and / or the USB Polling.LFPS signal; The negotiation channels specified in the communication protocol include the PCIe internal loopback negotiation channel and / or the USB protocol layer control channel; the link retraining includes the state machine reconfiguration process specified in the communication protocol.

3. The method according to claim 1, characterized in that, The unsaturated region includes a mixed state where at least one tap coefficient of the adaptive equalizer has reached a saturation threshold and other tap coefficients have not reached a saturation threshold; the method further includes: when any tap coefficient reaches a saturation threshold, triggering the monitoring step, and taking the direction of the saturated tap coefficient as the priority extraction direction of the compensation spectral feature.

4. The method according to claim 1, characterized in that, The step of generating transmission rate adjustment suggestions also includes: The channel warning information is subjected to de-jitter processing; within the preset confirmation window corresponding to the non-link reconstruction phase of the state machine specified in the communication protocol, the channel warning information continuously meets the warning conditions, and only then is the transmission rate adjustment suggestion encapsulated as a negotiation request. And / or, Record the timestamp of the most recent negotiation request; suppress the generation of new negotiation requests within the preset suppression window corresponding to the non-link reconstruction phase of the state machine specified in the communication protocol.

5. The method according to claim 1, characterized in that, The step of triggering link retraining based on the working state determined through negotiation according to the communication protocol further includes: During the negotiation and the link retraining, the data stream is switched to a preset bypass state according to the communication protocol; the bypass state includes at least one of data buffering, slowing down transmission, and skipping training frames. If the negotiation fails to reach an agreement or the link retraining fails to confirm that the working state is stable, the system switches to the bypass state for reduced transmission or data caching according to the fallback procedure specified in the communication protocol, and continues to trigger negotiation or reduces to the next lower level before re-executing the negotiation and the link retraining.

6. The method according to claim 1, characterized in that, The step of generating transmission rate adjustment suggestions also includes: Based on the channel warning information and the critical point information corresponding to each level in the transmission rate level set, the reduced rate step size or the strictness of evaluating the stability of the equalizer parameters is determined; wherein, the critical point information includes at least one of the following: communication protocol handshake overhead corresponding to each level, communication protocol state transition cost between adjacent levels, and link retraining time required when each level is the negotiation target.

7. The method according to claim 6, characterized in that, The method further includes: The negotiation results are reported to the upper-layer Quality of Service (QoS) protocol; the upper-layer QoS protocol adjusts traffic allocation or service priority according to the negotiated transmission rate; the upper-layer QoS protocol includes PCIe Virtual Channel Protocol and / or IETF Differentiated Services Protocol.

8. The method according to claim 1, characterized in that, The method further includes: When the negotiation request is generated, a negotiation preparation notification is simultaneously sent to the power management protocol; the power management protocol enters the power saving preparation state according to the negotiation preparation notification; the power management protocol includes the PCIeASPM protocol and / or the USB U2 / U3 protocol.

9. The method according to claim 1, characterized in that, The preset set of physical damage patterns of the link and the preset set of external interference patterns are adaptively updated according to the link's operating status. The adaptive update is performed during the off-link reconstruction phase of the state machine as specified in the communication protocol; the adaptive update includes: When the compensation spectral feature corresponding to the channel warning information does not match any spectral feature in the link physical damage spectral feature set and the external interference spectral feature set, the compensation spectral feature corresponding to the channel warning information is added as a new spectral feature to the corresponding spectral feature set. When the anti-interference suggestion is triggered and the link meets the stability condition within the preset observation window, the corresponding compensation spectral feature is added to the external interference spectral set as the new spectral feature.

10. A data line transmission rate adjustment system based on a communication protocol, characterized in that, include: The parameter monitoring module is used to monitor the equalizer parameters adjusted by the adaptive equalizer for the training signal specified by the communication protocol during the non-link reconstruction phase of the state machine specified by the communication protocol, when the data error rate has not reached the preset error threshold and the tap coefficients of the adaptive equalizer at the data line receiving end are working in the unsaturated range during the data line link operation process. The feature extraction module is used to analyze the temporal changes of the equalizer parameters within one or more complete transmission cycles of the training signal, and extract the compensation spectral features corresponding to the protocol structure of the training signal; the compensation spectral features include at least two of the following: the relative change between each tap coefficient, the difference in the adjustment rate of each tap coefficient, and the compensation mode for the signal frequency response. The source identification module is used to match the compensation spectrum features with a preset set of link physical damage spectrum features and a preset set of external interference spectrum features to identify the source of the equalizer parameter fluctuations; the source includes at least one of link physical degradation and external interference; The negotiation request generation module is used to generate a channel early warning message containing channel early warning information and a suggested rate step size, and to generate a transmission rate adjustment suggestion when the source is link physical degradation. And for generating anti-interference suggestions when the source is external interference; the transmission rate adjustment suggestions include reducing the current transmission rate to a level below the current transmission rate, and / or maintaining the current transmission rate at the original level and evaluating the stability of the equalizer parameters at the target level to determine whether to increase it to the target level; the anti-interference suggestions include maintaining the current transmission rate and increasing forward error correction redundancy, enabling data retransmission protocol or at least one of the following; the negotiation request carries at least one of the following: target transmission rate, suggested rate step size, channel warning level, warning confidence level; The negotiation training module is used to encapsulate the transmission rate adjustment suggestion or the anti-interference suggestion into a negotiation request according to the protocol data unit format specified by the communication protocol, send it to the data line peer through the negotiation channel specified by the communication protocol, negotiate with the data line peer to determine the working state; and trigger link retraining according to the working state determined by the negotiation according to the communication protocol to confirm that the working state can work stably.