Re: [PATCH v3 03/10] dmaengine: dw-edma: Move ll_region from struct dw_edma_chunk to struct dw_edma_chan

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

Pre-existing issues:
- [High] Missing lock acquisition in dw_edma_device_resume() when accessing vchan descriptors.
--

commit 80df5f151f3cb77451f6a2e4ec10cbaad5b50565
Author: Frank Li <Frank.Li@nxp.com>

dmaengine: dw-edma: Move ll_region from struct dw_edma_chunk to struct dw_edma_chan

This commit moves the ll_region field from struct dw_edma_chunk to struct
dw_edma_chan to avoid redundant copies. Since the ll_region is identical
for all chunks belonging to the same DMA channel, it can be shared.

> diff --git a/drivers/dma/dw-edma/dw-edma-core.c b/drivers/dma/dw-edma/dw-edma-core.c
> index 53469c8c8b82e..2652ad8e7a8f6 100644
> --- a/drivers/dma/dw-edma/dw-edma-core.c
> +++ b/drivers/dma/dw-edma/dw-edma-core.c

[ ... ]

> static int dw_edma_device_resume(struct dma_chan *dchan)
> {
>         ...
>         dw_edma_start_transfer(chan);
>         ...
> }

[Severity: High]
This is a pre-existing issue, but can this sequence cause a data race on
the descriptor lists?

Looking at dw_edma_device_resume(), dw_edma_start_transfer() is called
without holding the chan->vc.lock. This invokes vchan_next_desc(), which
reads from the desc_issued list.

If another CPU concurrently calls dw_edma_device_issue_pending(), it uses
vchan_issue_pending() which runs list_splice_tail_init() on desc_issued
while holding the lock.

Could this concurrent access corrupt the list pointers when a client driver
dynamically pauses and resumes channels?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260702-edma_ll-v3-0-877aa463740c@nxp.com?part=3