cve/2024/CVE-2024-26584.md

18 lines
1.3 KiB
Markdown
Raw Permalink Normal View History

2024-05-25 21:48:12 +02:00
### [CVE-2024-26584](https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2024-26584)
![](https://img.shields.io/static/v1?label=Product&message=Linux&color=blue)
2024-05-25 23:14:53 +00:00
![](https://img.shields.io/static/v1?label=Version&message=a54667f6728c%3C%203ade391adc58%20&color=brighgreen)
2024-05-25 21:48:12 +02:00
![](https://img.shields.io/static/v1?label=Vulnerability&message=n%2Fa&color=brighgreen)
### Description
In the Linux kernel, the following vulnerability has been resolved:net: tls: handle backlogging of crypto requestsSince we're setting the CRYPTO_TFM_REQ_MAY_BACKLOG flag on ourrequests to the crypto API, crypto_aead_{encrypt,decrypt} can return -EBUSY instead of -EINPROGRESS in valid situations. For example, whenthe cryptd queue for AESNI is full (easy to trigger with anartificially low cryptd.cryptd_max_cpu_qlen), requests will be enqueuedto the backlog but still processed. In that case, the async callbackwill also be called twice: first with err == -EINPROGRESS, which itseems we can just ignore, then with err == 0.Compared to Sabrina's original patch this version uses the newtls_*crypt_async_wait() helpers and converts the EBUSY toEINPROGRESS to avoid having to modify all the error handlingpaths. The handling is identical.
### POC
#### Reference
No PoCs from references.
#### Github
- https://github.com/fkie-cad/nvd-json-data-feeds