linux-kernel - Re: [PATCH v3] coresight: ETR: Fix ETR buffer use-after-free issue

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20251105152316.GA281971@e132581.arm.com>
Date: Wed, 5 Nov 2025 15:23:16 +0000
From: Leo Yan <leo.yan@....com>
To: Xiaoqi Zhuang <xiaoqi.zhuang@....qualcomm.com>
Cc: Suzuki K Poulose <suzuki.poulose@....com>,
	Mike Leach <mike.leach@...aro.org>,
	James Clark <james.clark@...aro.org>,
	Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
	Linu Cherian <lcherian@...vell.com>, coresight@...ts.linaro.org,
	linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
	linux-arm-msm@...r.kernel.org
Subject: Re: [PATCH v3] coresight: ETR: Fix ETR buffer use-after-free issue

On Tue, Oct 21, 2025 at 04:45:25PM +0800, Xiaoqi Zhuang wrote:
> When ETR is enabled as CS_MODE_SYSFS, if the buffer size is changed
> and enabled again, currently sysfs_buf will point to the newly
> allocated memory(buf_new) and free the old memory(buf_old). But the
> etr_buf that is being used by the ETR remains pointed to buf_old, not
> updated to buf_new. In this case, it will result in a memory
> use-after-free issue.
> 
> Fix this by checking ETR's mode before updating and releasing buf_old,
> if the mode is CS_MODE_SYSFS, then skip updating and releasing it.
> 
> Fixes: bd2767ec3df2 ("coresight: Fix run time warnings while reusing ETR buffer")
> Signed-off-by: Xiaoqi Zhuang <xiaoqi.zhuang@....qualcomm.com>

Tested on my Juno board with below steps:

  1) Enable the first path: ETM2 -> ETR0
  echo 1 > /sys/bus/coresight/devices/tmc_etr0/enable_sink
  echo 1 > /sys/bus/coresight/devices/etm2/enable_source

  2) Enlarge buffer size from 1MiB to 4MiB
  cat /sys/bus/coresight/devices/tmc_etr0/buffer_size
  0x100000
  echo 0x400000 > /sys/bus/coresight/devices/tmc_etr0/buffer_size

  3) Enable the second path: ETM0 -> ETR0
  echo 1 > /sys/bus/coresight/devices/etm0/enable_source

  4) Disable paths
  echo 0 > /sys/bus/coresight/devices/etm0/enable_source
  echo 0 > /sys/bus/coresight/devices/etm2/enable_source

Without this patch, the oops will be triggered when disable paths.
I can confirm this patch does dismiss the issue.

Tested-by: Leo Yan <leo.yan@....com>

> ---
> Changes in v3:
> - Add a fix tag for the fix patch.
> - Link to v2: https://lore.kernel.org/r/20251021-fix_etr_issue-v2-1-80c40c9cac8c@oss.qualcomm.com
> 
> Changes in v2:
> - Exit earlier to avoid allocating memory unnecessarily.
> - Link to v1: https://lore.kernel.org/r/20251020-fix_etr_issue-v1-1-902ab51770b4@oss.qualcomm.com
> ---
>  drivers/hwtracing/coresight/coresight-tmc-etr.c | 7 +++++++
>  1 file changed, 7 insertions(+)
> 
> diff --git a/drivers/hwtracing/coresight/coresight-tmc-etr.c b/drivers/hwtracing/coresight/coresight-tmc-etr.c
> index b07fcdb3fe1a..800be06598c1 100644
> --- a/drivers/hwtracing/coresight/coresight-tmc-etr.c
> +++ b/drivers/hwtracing/coresight/coresight-tmc-etr.c
> @@ -1250,6 +1250,13 @@ static struct etr_buf *tmc_etr_get_sysfs_buffer(struct coresight_device *csdev)
>  	 * with the lock released.
>  	 */
>  	raw_spin_lock_irqsave(&drvdata->spinlock, flags);
> +
> +	/*
> +	 * If the ETR is already enabled, continue with the existing buffer.
> +	 */
> +	if (coresight_get_mode(csdev) == CS_MODE_SYSFS)
> +		goto out;
> +
>  	sysfs_buf = READ_ONCE(drvdata->sysfs_buf);
>  	if (!sysfs_buf || (sysfs_buf->size != drvdata->size)) {
>  		raw_spin_unlock_irqrestore(&drvdata->spinlock, flags);
> 
> ---
> base-commit: 98ac9cc4b4452ed7e714eddc8c90ac4ae5da1a09
> change-id: 20251020-fix_etr_issue-02c706dbc899
> 
> Best regards,
> -- 
> Xiaoqi Zhuang <xiaoqi.zhuang@....qualcomm.com>
> 
>