{"id":"AZL-106005","summary":"CVE-2026-98342 affecting package kernel 6.6.157.1-1","details":"In the Linux kernel, the following vulnerability has been resolved:\n\ndmaengine: wait for RCU readers before releasing dma_device\n\ndma_issue_pending_all() walks the dma_device_list with\nlist_for_each_entry_rcu() under rcu_read_lock(). dma_device_release()\nunlinks the device with list_del_rcu() and then calls\ndevice-\u003edevice_release() (which in many drivers, such as plx_dma.c,\ndirectly calls kfree()).\n\nBecause there is no grace period between unlinking the device and\nfreeing it, concurrent RCU readers in dma_issue_pending_all() can\naccess the device after it has been freed.\n\nThe lockless walk originally relied on clients holding a dmaengine\nreference to pin the provider module, and therefore the device, for as\nlong as they might traverse the list. Commit 8ad342a86359 (\"dmaengine:\nAdd reference counting to dma_device struct\") decoupled the dma_device\nlifetime from the module reference, so the device can now be released\nwhile a reader is still walking the list.\n\nAdd synchronize_rcu() before the device is freed, so RCU readers are\nguaranteed to have finished. Keep it unconditional: providers that do\nnot implement device_release() free the device themselves once\ndma_async_device_unregister() returns. This call will delay for a grace\nperiod with dma_list_mutex held, which is safe and only teardown path is\ndelayed.","modified":"2026-10-07T14:16:54.760669176Z","published":"2026-10-06T09:18:27Z","upstream":["CVE-2026-98342"],"references":[{"type":"WEB","url":"https://nvd.nist.gov/vuln/detail/CVE-2026-98342"}],"affected":[{"package":{"name":"kernel","ecosystem":"Azure Linux:3","purl":"pkg:rpm/azure-linux/kernel"},"ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"last_affected":"6.6.157.1-1"}]}],"database_specific":{"source":"https://github.com/microsoft/AzureLinuxVulnerabilityData/blob/main/osv/AZL-106005.json"}}],"schema_version":"1.9.0"}