In the Linux kernel, the following vulnerability has been resolved:
iio: buffer: Fix DMA fence leak in iiobufferenqueue_dmabuf()
iiobufferenqueuedmabuf() allocates a struct iiodmafence (104 bytes, kmalloc-128) via kmallocobj()+dmafenceinit(), which sets the initial kref to 1. It then calls dmaresvadd_fence() which takes a second reference (kref=2), and stores a raw pointer in block->fence.
On the success path the function returns without calling dmafenceput() to release the initial reference, so every buffer enqueue permanently leaks one kmalloc-128 allocation.
The iiobuffercleanup() work item only releases the temporary reference taken during completion signalling by iiobuffersignaldmabufdone(); the initial reference from dmafenceinit() is never released.
With four iiorwdev instances at 240kHz and 512 samples per buffer, this produces ~1875 kmalloc-128 allocations per second matching the observed slab growth exactly. A test with ftrace confirmed that the dmafence_destroy event was never triggered.
Fix by calling dmafenceput() after dmaresvadd_fence(), transferring ownership of the fence to the DMA reservation object. The DMA fence then gets properly discarded after being signalled.
{
"cna_assigner": "Linux",
"osv_generated_from": "https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/63xxx/CVE-2026-63929.json"
}