From: sashiko-bot@kernel.org
To: Frank.Li@oss.nxp.com
Cc: Frank.Li@kernel.org, imx@lists.linux.dev, linux-serial@vger.kernel.org
Subject: Re: [PATCH 1/2] serial: use dmaengine_get_dma_device() instead of chan->device->dev
Date: Fri, 25 Sep 2026 20:19:57 +0000 [thread overview]
Message-ID: <20260925201958.2FC1A1F000FF@smtp.kernel.org> (raw)
In-Reply-To: <20260925200850.1678179-1-Frank.Li@oss.nxp.com>
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] dma-mapping: asymmetric dma_map_single/unmap pair in pl011
--
commit fb77536d75ed603638aebc851f73960ddf8efe10
Author: Frank Li <Frank.Li@nxp.com>
serial: use dmaengine_get_dma_device() instead of chan->device->dev
This commit replaces direct access to chan->device->dev with the
dmaengine_get_dma_device() API across multiple serial drivers. This correctly
handles per-channel IOMMU mappings and prepares for making DMA engine
provider structures private.
Link: https://lore.kernel.org/imx/67966b47-22cb-4b07-80c7-2044f557dbcb@app.fastmail.com/
> diff --git a/drivers/tty/serial/amba-pl011.c b/drivers/tty/serial/amba-pl011.c
> index c4824c201e1c3..86729ddbdadc8 100644
> --- a/drivers/tty/serial/amba-pl011.c
> +++ b/drivers/tty/serial/amba-pl011.c
[ ... ]
> @@ -609,7 +609,7 @@ static void pl011_dma_tx_callback(void *data)
>
> uart_port_lock_irqsave(&uap->port, &flags);
> if (uap->dmatx.queued)
> - dma_unmap_single(dmatx->chan->device->dev, dmatx->dma,
> + dma_unmap_single(dmaengine_get_dma_device(dmatx->chan), dmatx->dma,
[Severity: High]
Does this change create an asymmetric map/unmap pair in the PL011 driver?
By updating pl011_dma_tx_callback() to use dmaengine_get_dma_device() for the
unmap operation, it appears to no longer match the map operation in
pl011_dma_tx_refill() which was not updated in this patch:
drivers/tty/serial/amba-pl011.c:pl011_dma_tx_refill() {
...
dmatx->dma = dma_map_single(dma_dev->dev, dmatx->buf, count,
DMA_TO_DEVICE);
...
}
The DMA API requires the exact same device pointer to be used for both
mapping and unmapping. On platforms where the DMA engine provides per-channel
devices, dmaengine_get_dma_device() will return a different device pointer
than the raw device pointer used during mapping.
Could this break DMA tracking or cause IOMMU faults during standard TX
operations? Should the map operation in pl011_dma_tx_refill() be updated to
use the new API as well?
> dmatx->len, DMA_TO_DEVICE);
>
> dmacr = uap->dmacr;
--
Sashiko AI review · https://sashiko.dev/#/patchset/20260925200850.1678179-1-Frank.Li@oss.nxp.com?part=1
next prev parent reply other threads:[~2026-09-25 20:20 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-09-25 20:08 [PATCH 1/2] serial: use dmaengine_get_dma_device() instead of chan->device->dev Frank.Li
2026-09-25 20:08 ` [PATCH 2/2] serial: amba-pl011: use dmaengine public API instead of raw ops Frank.Li
2026-09-25 20:16 ` sashiko-bot
2026-09-25 20:19 ` sashiko-bot [this message]
2026-09-28 8:51 ` [PATCH 1/2] serial: use dmaengine_get_dma_device() instead of chan->device->dev Andy Shevchenko
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20260925201958.2FC1A1F000FF@smtp.kernel.org \
--to=sashiko-bot@kernel.org \
--cc=Frank.Li@kernel.org \
--cc=Frank.Li@oss.nxp.com \
--cc=imx@lists.linux.dev \
--cc=linux-serial@vger.kernel.org \
--cc=sashiko-reviews@lists.linux.dev \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox