Linux kernel and device drivers for NXP i.MX platforms
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: phucduc.bui@gmail.com
Cc: imx@lists.linux.dev, Frank.Li@kernel.org
Subject: Re: [PATCH v2 01/11] ASoC: fsl_asrc: Use guard() for spin locks
Date: Mon, 15 Jun 2026 09:51:45 +0000	[thread overview]
Message-ID: <20260615095146.CD1781F000E9@smtp.kernel.org> (raw)
In-Reply-To: <20260615093824.115751-2-phucduc.bui@gmail.com>

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

Pre-existing issues:
- [High] Lockless access to the `asrc->pair` array in `fsl_asrc_isr` can lead to Use-After-Free and reading uninitialized memory.
--

commit 4b6342441bc61476e92408bc85fb134b9ad8e9f7
Author: bui duc phuc <phucduc.bui@gmail.com>

ASoC: fsl_asrc: Use guard() for spin locks

Clean up the code using guard() for spin locks.
Merely code refactoring, and no behavior change.

> diff --git a/sound/soc/fsl/fsl_asrc.c b/sound/soc/fsl/fsl_asrc.c
> index 5fda9b647c704..0b28bcfa47fed 100644
> --- a/sound/soc/fsl/fsl_asrc.c
> +++ b/sound/soc/fsl/fsl_asrc.c

[ ... ]

> @@ -265,19 +262,16 @@ static void fsl_asrc_release_pair(struct fsl_asrc_pair *pair)
>  {
>  	struct fsl_asrc *asrc = pair->asrc;
>  	enum asrc_pair_index index = pair->index;
> -	unsigned long lock_flags;
>  
>  	/* Make sure the pair is disabled */
>  	regmap_update_bits(asrc->regmap, REG_ASRCTR,
>  			   ASRCTR_ASRCEi_MASK(index), 0);
>  
> -	spin_lock_irqsave(&asrc->lock, lock_flags);
> +	guard(spinlock_irqsave)(&asrc->lock);
>  
>  	asrc->channel_avail += pair->channels;
>  	asrc->pair[index] = NULL;

[Severity: High]
This isn't a bug introduced by this patch, but can the lockless access to the
asrc->pair array in fsl_asrc_isr() lead to a use-after-free or reading
uninitialized memory?

In fsl_asrc_release_pair(), the pair is disabled and removed via
asrc->pair[index] = NULL under the lock. However, fsl_asrc_isr() iterates
over the asrc->pair array and dereferences its elements locklessly:

sound/soc/fsl/fsl_asrc.c:fsl_asrc_isr() {
    ...
		if (!asrc->pair[index])
			continue;

		if (status & ASRSTR_ATQOL) {
			asrc->pair[index]->error |= ASRC_TASK_Q_OVERLOAD;
    ...
}

Because the ISR does not take asrc->lock, an ISR executing on another CPU can
read the pointer just before it is nullified. ALSA subsequently calls
fsl_asrc_dma_shutdown(), which frees the memory:

sound/soc/fsl/fsl_asrc_dma.c:fsl_asrc_dma_shutdown() {
    ...
	kfree(pair);
    ...
}

Since there is no synchronize_irq() or lock-based synchronization, the
concurrent ISR can proceed to write to pair->error after the memory is freed,
causing a use-after-free in hardirq context.

Additionally, fsl_asrc_request_pair() publishes the newly allocated pair via a
plain store:

		asrc->pair[index] = pair;

Without a memory barrier (like smp_store_release()), the lockless ISR may see
the pointer before the initialization of the pair structure is visible to its
CPU.

>  	pair->error = 0;
> -
> -	spin_unlock_irqrestore(&asrc->lock, lock_flags);
>  }
>  
>  /**

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260615093824.115751-1-phucduc.bui@gmail.com?part=1

  reply	other threads:[~2026-06-15  9:51 UTC|newest]

Thread overview: 31+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-06-15  9:38 [PATCH v2 00/11] ASoC: fsl: Use guard() for mutex & spin locks phucduc.bui
2026-06-15  9:38 ` [PATCH v2 01/11] ASoC: fsl_asrc: Use guard() for " phucduc.bui
2026-06-15  9:51   ` sashiko-bot [this message]
2026-06-15 14:18   ` Frank Li
2026-06-15  9:38 ` [PATCH v2 02/11] ASoC: fsl_audmix: " phucduc.bui
2026-06-15 14:30   ` Frank Li
2026-06-15  9:38 ` [PATCH v2 03/11] ASoC: fsl_easrc: " phucduc.bui
2026-06-15  9:48   ` sashiko-bot
2026-06-15 14:24   ` Frank Li
2026-06-15  9:38 ` [PATCH v2 04/11] ASoC: fsl_esai: " phucduc.bui
2026-06-15  9:52   ` sashiko-bot
2026-06-15 14:15   ` Frank Li
2026-06-15  9:38 ` [PATCH v2 05/11] ASoC: fsl_spdif: " phucduc.bui
2026-06-15  9:52   ` sashiko-bot
2026-06-15 14:17   ` Frank Li
2026-06-15  9:38 ` [PATCH v2 06/11] ASoC: fsl_ssi: Use guard() for mutex locks phucduc.bui
2026-06-15 14:25   ` Frank Li
2026-06-15  9:38 ` [PATCH v2 07/11] ASoC: fsl_xcvr: Use guard() for spin locks phucduc.bui
2026-06-15  9:49   ` sashiko-bot
2026-06-15 14:19   ` Frank Li
2026-06-15  9:38 ` [PATCH v2 08/11] ASoC: imx-audio-rpmsg: " phucduc.bui
2026-06-15  9:51   ` sashiko-bot
2026-06-15 14:17   ` Frank Li
2026-06-15  9:38 ` [PATCH v2 09/11] ASoC: fsl_rpmsg: Use guard() for mutex & " phucduc.bui
2026-06-15  9:59   ` sashiko-bot
2026-06-15 14:12   ` Frank Li
2026-06-15  9:38 ` [PATCH v2 10/11] ASoC: fsl: mpc5200_dma: Use guard() for " phucduc.bui
2026-06-15  9:57   ` sashiko-bot
2026-06-15 14:14   ` Frank Li
2026-06-15  9:38 ` [PATCH v2 11/11] ASoC: fsl: mpc5200_psc_ac97: Use guard() for mutex locks phucduc.bui
2026-06-15 14:26   ` Frank Li

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=20260615095146.CD1781F000E9@smtp.kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=Frank.Li@kernel.org \
    --cc=imx@lists.linux.dev \
    --cc=phucduc.bui@gmail.com \
    --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