dmaengine: mmp_pdma: Fix race condition in mmp_pdma_residue()
Linux
Affected versions
| Product | Affected | Fixed in |
|---|---|---|
| Linux Linux |
|
Linux
| Product | Affected | Fixed in |
|---|---|---|
| Linux Linux |
|
≥ 1b38da264674d6a0fe26a63996b8f88b88c3da48 · < a143545855bc2c6e1330f6f57ae375ac44af00a7| No fix yet |
| Linux Linux |
|
|
In the Linux kernel, the following vulnerability has been resolved:
dmaengine: mmp_pdma: Fix race condition in mmp_pdma_residue()
Add proper locking in mmp_pdma_residue() to prevent use-after-free when accessing descriptor list and descriptor contents.
The race occurs when multiple threads call tx_status() while the tasklet on another CPU is freeing completed descriptors:
CPU 0 CPU 1 ----- ----- mmp_pdma_tx_status() mmp_pdma_residue() -> NO LOCK held list_for_each_entry(sw, ..) DMA interrupt dma_do_tasklet() -> spin_lock(&desc_lock) list_move(sw->node, ...) spin_unlock(&desc_lock) | dma_pool_free(sw) <- FREED! -> access sw->desc <- UAF!
This issue can be reproduced when running dmatest on the same channel with multiple threads (threads_per_chan > 1).
Fix by protecting the chain_running list iteration and descriptor access with the chan->desc_lock spinlock.
From the CVE record · Linux
The Linux kernel’s mmp_pdma_residue() can access a freed DMA descriptor because it does not lock the descriptor list during iteration. The race can occur when multiple threads call tx_status() while a tasklet on another CPU frees completed descriptors; it was reproduced with multithreaded dmatest on one channel. The input does not specify authentication requirements or a remote attack path.
Engineers running Linux kernels with the mmp_pdma driver should check their installations; affected versions are not specified.