DMA Engine development
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: "Wolfram Sang" <wsa+renesas@sang-engineering.com>
Cc: Frank.Li@kernel.org, dmaengine@vger.kernel.org, vkoul@kernel.org
Subject: Re: [PATCH 1/8] dmaengine: ptdma: unmap descriptor before completing cookie
Date: Sat, 10 Oct 2026 12:57:19 +0000	[thread overview]
Message-ID: <sashiko-outbox-166555@kernel.org> (raw)
In-Reply-To: <20261010124348.19055-2-wsa+renesas@sang-engineering.com>

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] dmaengine: ptdma: UAF in ae4_pending_work due to unremoved list entries
--

commit 45d9a162336ad9b005d6d181ab13de82d05d7cf3
Author: Wolfram Sang <wsa+renesas@sang-engineering.com>

dmaengine: ptdma: unmap descriptor before completing cookie

To ensure that buffers are flushed, make sure that the descriptor gets
unmapped before completing the cookie.

Link: https://lore.kernel.org/r/20261009093745.539924-1-vkoul@kernel.org

> diff --git a/drivers/dma/amd/ptdma/ptdma-dmaengine.c b/drivers/dma/amd/ptdma/ptdma-dmaengine.c
> index 628c49ce5de9d..7f51d1e6c727f 100644
> --- a/drivers/dma/amd/ptdma/ptdma-dmaengine.c
> +++ b/drivers/dma/amd/ptdma/ptdma-dmaengine.c

[ ... ]

> @@ -362,8 +362,8 @@ static void pt_cmd_callback_work(void *data, int err)
>  		if (desc->status != DMA_ERROR)
>  			desc->status = DMA_COMPLETE;
>  
> -		dma_cookie_complete(tx_desc);
>  		dma_descriptor_unmap(tx_desc);
> +		dma_cookie_complete(tx_desc);

[Severity: High]
Can this result in a use-after-free when pt_terminate_all() is called?

pt_terminate_all() extracts all descriptors from the virt-dma channel and
frees them via vchan_dma_desc_free_list():

drivers/dma/amd/ptdma/ptdma-dmaengine.c:pt_terminate_all() {
    ...
	vchan_dma_desc_free_list(&chan->vc, &head);
    ...
}

However, for AE4 DMA devices, descriptors are additionally added to the
ae4cmd_q->cmd list during pt_create_desc(). pt_terminate_all() fails to
remove the freed descriptors from this hardware pending command list, and
does not acquire ae4cmd_q->cmd_lock.

The asynchronous worker thread ae4_pending_work() then iterates over the
uncleaned list, removes the freed descriptor, and executes its callback:

drivers/dma/amd/ae4dma/ae4dma-dev.c:ae4_pending_work() {
    ...
			cmd = list_first_entry(&ae4cmd_q->cmd, struct pt_cmd, entry);
			list_del(&cmd->entry);

			ae4_check_status_error(ae4cmd_q, ae4cmd_q->dridx);
			cmd->pt_cmd_callback(cmd->data, cmd->ret);
    ...
}

Since pt_cmd_callback_work() is the callback invoked here, will this execution
dereference the freed descriptor pointers and cause list memory corruption?

>  	} else {
>  		tx_desc = NULL;
>  	}

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261010124348.19055-1-wsa+renesas@sang-engineering.com?part=1

  reply	other threads:[~2026-10-10 12:57 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-10-10 12:43 [PATCH 0/8] dmaengine: unmap descriptor before completing cookie Wolfram Sang
2026-10-10 12:43 ` [PATCH 1/8] dmaengine: ptdma: " Wolfram Sang
2026-10-10 12:57   ` sashiko-bot [this message]
2026-10-12 14:51   ` Frank Li
2026-10-10 12:43 ` [PATCH 2/8] dmaengine: fsl_raid: " Wolfram Sang
2026-10-10 12:43 ` [PATCH 3/8] dmaengine: idxd: " Wolfram Sang
2026-10-10 12:43 ` [PATCH 4/8] dmaengine: ioat: " Wolfram Sang
2026-10-10 12:43 ` [PATCH 5/8] dmaengine: mv_xor_v2: " Wolfram Sang
2026-10-10 12:54   ` sashiko-bot
2026-10-10 12:43 ` [PATCH 6/8] dmaengine: plx_dma: " Wolfram Sang
2026-10-10 12:56   ` sashiko-bot
2026-10-10 12:43 ` [PATCH 7/8] dmaengine: switchtec-dma: " Wolfram Sang
2026-10-10 12:54   ` sashiko-bot
2026-10-10 12:43 ` [PATCH 8/8] dmaengine: xgene-dma: " Wolfram Sang
2026-10-12 10:42 ` [PATCH 0/8] dmaengine: " Wolfram Sang

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=sashiko-outbox-166555@kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=Frank.Li@kernel.org \
    --cc=dmaengine@vger.kernel.org \
    --cc=sashiko-reviews@lists.linux.dev \
    --cc=vkoul@kernel.org \
    --cc=wsa+renesas@sang-engineering.com \
    /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