Логотип exploitDog
Консоль
Логотип exploitDog

exploitDog

redhat логотип

CVE-2026-74353

Опубликовано: 15 авг. 2026
Источник: redhat
CVSS3: 5.5
EPSS Низкий

Описание

In the Linux kernel, the following vulnerability has been resolved: drm/amdkfd: always resume_all after suspend_all Need to restore any good queues even if the suspend_all failed for some. Always run remove_queue as that will schedule a GPU reset is removing the queue fails. v2: move resume_all after remove

A flaw was found in the drm/amdkfd component of the Linux kernel. This vulnerability occurs when the system fails to properly resume all queues after a suspend operation, even if some suspend operations fail. An attacker could potentially exploit this by triggering a scenario where queue removal fails, leading to an unscheduled Graphics Processing Unit (GPU) reset. This could result in a Denial of Service (DoS) for the affected GPU functionality.

Отчет

A flaw of low impact was found in the AMD KFD (drm/amdkfd) driver of the Linux kernel as shipped in Red Hat Enterprise Linux and Red Hat In-Vehicle OS. The driver did not reliably resume all queues after a suspend_all operation: when suspending some queues failed, otherwise healthy queues could remain suspended, and when queue removal failed, the required GPU reset was not scheduled. A local user with access to the AMD KFD/GPU device could trigger a bad GPU queue condition that leaves good queues suspended, resulting in a denial of service for GPU compute workloads or forcing GPU reset-based recovery. The flaw has no confidentiality or integrity impact and is not network-reachable. The affected streams are Red Hat Enterprise Linux 9 (rhel-9.6.z, rhel-9.8.z, rhel-9.9) and Red Hat Enterprise Linux 10 (rhel-10.0.z, rhel-10.2.z, rhel-10.3), and Red Hat In-Vehicle OS 2 streams derived from them, where the fix is deferred to a later update; the residual risk is a local denial of service limited to AMD GPU functionality. Red Hat Enterprise Linux 7 and 8, and the earlier RHEL 9.2 and 9.4 update streams, do not contain the vulnerable code and are therefore not affected. Although the Red Hat CVSS:3.1 base score is 5.5 (Moderate), Red Hat rates the impact of this flaw as LOW because exploitation is limited to a local denial of service of AMD GPU workloads.

Меры по смягчению последствий

Mitigation for this issue is either not available or the currently available options do not meet the Red Hat Product Security criteria comprising ease of use and deployment, applicability to widespread installation base, or stability.

Затронутые пакеты

ПлатформаПакетСостояниеРекомендацияРелиз
Red Hat Enterprise Linux 10kernelFix deferred
Red Hat Enterprise Linux 6kernelOut of support scope
Red Hat Enterprise Linux 7kernelNot affected
Red Hat Enterprise Linux 7kernel-rtNot affected
Red Hat Enterprise Linux 8kernelNot affected
Red Hat Enterprise Linux 8kernel-rtNot affected
Red Hat Enterprise Linux 9kernelFix deferred
Red Hat Enterprise Linux 9kernel-rtFix deferred

Показывать по

Дополнительная информация

Статус:

Low
Дефект:
CWE-390
https://bugzilla.redhat.com/show_bug.cgi?id=2516862kernel: drm/amdkfd: always resume_all after suspend_all

EPSS

Процентиль: 10%
0.00209
Низкий

5.5 Medium

CVSS3

Связанные уязвимости

ubuntu
2 месяца назад

In the Linux kernel, the following vulnerability has been resolved: drm/amdkfd: always resume_all after suspend_all Need to restore any good queues even if the suspend_all failed for some. Always run remove_queue as that will schedule a GPU reset is removing the queue fails. v2: move resume_all after remove

nvd
2 месяца назад

In the Linux kernel, the following vulnerability has been resolved: drm/amdkfd: always resume_all after suspend_all Need to restore any good queues even if the suspend_all failed for some. Always run remove_queue as that will schedule a GPU reset is removing the queue fails. v2: move resume_all after remove

debian
2 месяца назад

In the Linux kernel, the following vulnerability has been resolved: d ...

CVSS3: 5.5
redos
около 1 месяца назад

Уязвимость kernel-lt

github
2 месяца назад

In the Linux kernel, the following vulnerability has been resolved: drm/amdkfd: always resume_all after suspend_all Need to restore any good queues even if the suspend_all failed for some. Always run remove_queue as that will schedule a GPU reset is removing the queue fails. v2: move resume_all after remove

EPSS

Процентиль: 10%
0.00209
Низкий

5.5 Medium

CVSS3