CVE-2026-43445
Memory Leak in Linux Kernel e1000/e1000e Network Drivers
Publication date: 2026-05-08
Last updated on: 2026-05-08
Assigner: kernel.org
Description
Description
CVSS Scores
EPSS Scores
| Probability: | |
| Percentile: |
Meta Information
Affected Vendors & Products
| Vendor | Product | Version / Range |
|---|---|---|
| intel | e1000 | * |
| intel | e1000e | * |
Helpful Resources
Exploitability
| CWE ID | Description |
|---|---|
| CWE-UNKNOWN |
Attack-Flow Graph
AI Powered Q&A
Can you explain this vulnerability to me?
This vulnerability exists in the Linux kernel's e1000 and e1000e network drivers. It involves a flaw in the handling of DMA (Direct Memory Access) error cleanup when mapping transmit (TX) buffers. Specifically, if an error occurs while mapping TX buffers, the driver is supposed to unmap any buffers that were already successfully mapped for the socket buffer (skb).
The problem arises because the count of successfully mapped buffers is incremented after each successful mapping, but the code decrements this count before a cleanup loop, causing an off-by-one error. This means that if at least one buffer was successfully mapped before an error, exactly one DMA mapping would leak (not be properly unmapped).
Additionally, earlier commits introduced a faulty while loop condition that caused an infinite loop during DMA error handling, which was later fixed but introduced this off-by-one error. The issue may also exist in the igbvf driver but was not addressed in the patch.
How can this vulnerability impact me? :
This vulnerability can lead to a resource leak in the form of DMA mappings that are not properly cleaned up when an error occurs during buffer mapping. Such leaks can cause increased memory usage and potentially degrade system performance over time.
In worst cases, resource leaks in kernel drivers can contribute to system instability or crashes if resources become exhausted. However, the description does not specify any direct security impact such as privilege escalation or data corruption.