{"id":"AZL-102617","summary":"CVE-2026-92525 affecting package kernel 6.6.150.1-1","details":"In the Linux kernel, the following vulnerability has been resolved:\n\nRDMA/rxe: Validate num_sge/cur_sge before indexing wqe-\u003edma.sge[]\n\nFor a user QP, qp-\u003esq.queue is a ring the application writes directly,\nso rxe_post_send() takes the is_user branch and only schedules send_task\nwithout validating the WQE. rxe_requester() consumes it in place via\nreq_next_wqe() and calls copy_data(), which indexes\n&wqe-\u003edma.sge[cur_sge] with the attacker-controlled num_sge/cur_sge.\nOnly the kernel path bounds num_sge (validate_send_wr()); the user WQE\nis never checked, so a local unprivileged user can post a WQE with an\nout-of-range cur_sge or oversized num_sge and force an out-of-bounds\nread of the per-WQE sge array in copy_data() (vmalloc OOB read, local\nDoS).\n\nBound num_sge to qp-\u003esq.max_sge in rxe_requester() before use, the way\nget_srq_wqe() already guards SRQ entries, and bound cur_sge only when\nthe WQE carries payload (dma.resid): copy_data() returns early on a\nzero-length copy before touching dma-\u003esge[], so a zero-payload WQE --\nthe only kind a max_sge == 0 QP can post -- stays valid.\n\nReproduced under KASAN; the vmalloc-out-of-bounds in copy_data() is gone.","modified":"2026-09-19T05:33:49Z","published":"2026-09-17T17:17:55Z","upstream":["CVE-2026-92525"],"references":[{"type":"WEB","url":"https://nvd.nist.gov/vuln/detail/CVE-2026-92525"}],"affected":[{"package":{"name":"kernel","ecosystem":"Azure Linux:3","purl":"pkg:rpm/azure-linux/kernel"},"ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"last_affected":"6.6.150.1-1"}]}],"database_specific":{"source":"https://github.com/microsoft/AzureLinuxVulnerabilityData/blob/main/osv/AZL-102617.json"}}],"schema_version":"1.9.0"}