RDMA Virtual Memory Page Fault Handling With Shared Fault Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing RDMA technology fails to efficiently handle page faults on virtual memory pages, leading to prolonged access delays and reduced communication efficiency between devices.
Innovation Solution
Implement a method where devices detect and handle page faults by allocating physical memory pages associated with virtual memory pages in real-time, using shared memory pages to store fault information and version numbers, ensuring timely detection and resolution of page faults.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the network interface card stops accessing the virtual memory page and requests the first device to process the page fault, then the page fault can be handled properly, but it takes a long time for the second device to find the page fault, resulting in low access efficiency
Solution Approach 1:
The patent applies preliminary action by having the network interface card proactively send page fault information to the second device before the access operation completes. This allows the second device to be notified in advance of the page fault condition, enabling it to prepare appropriate handling actions and avoid prolonged waiting times. The network interface card performs the notification action beforehand rather than waiting for the full access delay to occur.
Solution Approach 2:
The patent implements feedback by establishing a communication loop where the network interface card sends page fault information back to the second device, which then responds with handling instructions. This feedback mechanism ensures that the second device has real-time awareness of the page fault status and can dynamically adjust its access operations accordingly, improving both reliability and timing.
2Reliability
If the second device re-accesses the virtual memory page only after finding the page fault, then the page fault is handled, but the access efficiency remains low due to the delay
Solution Approach 1:
The patent uses feedback by having the network interface card continuously monitor and report page fault status to the second device. This real-time feedback enables the second device to maintain accurate awareness of page fault conditions without needing to perform repeated access attempts, thereby improving both detection accuracy and access efficiency simultaneously.
Solution Approach 2:
The system applies self-service by enabling the network interface card to autonomously detect and report page faults to the second device without requiring the second device to continuously poll or re-access the virtual memory page. The network interface card serves itself by managing the page fault detection and notification process, freeing the second device to optimize its access operations.
Data Source
AI summary
This application discloses a communication method. The method includes: A first device receives a first access request sent by a second device. The first access request includes identification information of a first virtual memory page in the first device. The first virtual memory page is not associated with a physical memory page corresponding to the first data. The first device sends first information to the second device based on the identification information. The first device receives an exception handling request that is sent by the second device based on the first information. The first device allocates, in a memory of the first device based on the exception handling request, a first physical memory page associated with the first virtual memory page. The first device sends, to the second device, a first access result of accessing the first physical memory page.