linux-kernel - Re: [PATCH v5] media: pci: intel: ivsc: improve device reference counting in mei

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] [day] [month] [year] [list]

Message-ID: <aN0XszBNLZPjoRfg@kekkonen.localdomain>
Date: Wed, 1 Oct 2025 14:59:47 +0300
From: Sakari Ailus <sakari.ailus@...ux.intel.com>
To: Ma Ke <make24@...as.ac.cn>
Cc: bingbu.cao@...el.com, lixu.zhang@...el.com,
	stanislaw.gruszka@...ux.intel.com, mchehab@...nel.org,
	wentong.wu@...el.com, linux-media@...r.kernel.org,
	linux-kernel@...r.kernel.org, akpm@...ux-foundation.org
Subject: Re: [PATCH v5] media: pci: intel: ivsc: improve device reference
 counting in mei_ace driver

Hi Ma,

Thanks for the update.

On Mon, Sep 29, 2025 at 09:30:07AM +0800, Ma Ke wrote:
> The device reference counting in mei_ace_setup_dev_link() was
> incomplete, as the reference acquired by device_find_child_by_name()
> was not released immediately on the success path. Add put_device() to
> properly balance the reference count. Additionally, the redundant
> put_device() in mei_ace_remove() is removed.
> 
> Found by code review.

The patch seems fine but the commit message is a bit more dramatic than
what reality justifies. I'll use this:

Put the reference to csi_dev acquired during driver probe in
mei_ace_setup_dev_link() inside the same function, instead of during driver
unbind in mei_ace_remove(). This can be done as device_link_add() already
takes a reference to csi_dev.

> 
> Signed-off-by: Ma Ke <make24@...as.ac.cn>
> ---
> Changes in v5:
> - jumped to err instead of err_put to avoid calling put_device() again in err_put as reviewer's instructions;
> Changes in v4:
> - updated the subject as suggestions;
> Changes in v3:
> - deleted the tag of Fixes and CC, and moved put_device() to immediately after device_link_add() as suggestions;
> Changes in v2:
> - modified the put_device() operations and the patch title as suggestions.
> ---
>  drivers/media/pci/intel/ivsc/mei_ace.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/media/pci/intel/ivsc/mei_ace.c b/drivers/media/pci/intel/ivsc/mei_ace.c
> index 98310b8511b1..b306a320b70f 100644
> --- a/drivers/media/pci/intel/ivsc/mei_ace.c
> +++ b/drivers/media/pci/intel/ivsc/mei_ace.c
> @@ -414,10 +414,11 @@ static int mei_ace_setup_dev_link(struct mei_ace *ace)
>  	/* setup link between mei_ace and mei_csi */
>  	ace->csi_link = device_link_add(csi_dev, dev, DL_FLAG_PM_RUNTIME |
>  					DL_FLAG_RPM_ACTIVE | DL_FLAG_STATELESS);
> +	put_device(csi_dev);
>  	if (!ace->csi_link) {
>  		ret = -EINVAL;
>  		dev_err(dev, "failed to link to %s\n", dev_name(csi_dev));
> -		goto err_put;
> +		goto err;
>  	}
>  
>  	ace->csi_dev = csi_dev;
> @@ -522,7 +523,6 @@ static void mei_ace_remove(struct mei_cl_device *cldev)
>  	cancel_work_sync(&ace->work);
>  
>  	device_link_del(ace->csi_link);
> -	put_device(ace->csi_dev);
>  
>  	pm_runtime_disable(&cldev->dev);
>  	pm_runtime_set_suspended(&cldev->dev);

-- 
Regards,

Sakari Ailus