linux-kernel - Re: [PATCH v3 2/3] counter: 104-quad-8: Add lock guards

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: <20200312153450.GB3250@icarus>
Date:   Thu, 12 Mar 2020 11:34:50 -0400
From:   William Breathitt Gray <vilhelm.gray@...il.com>
To:     Syed Nayyar Waris <syednwaris@...il.com>
Cc:     jic23@...nel.org, linux-iio@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3 2/3] counter: 104-quad-8: Add lock guards -
 differential encoder cable

On Thu, Mar 12, 2020 at 04:55:17PM +0530, Syed Nayyar Waris wrote:
> Add lock protection from race conditions in the 104-quad-8 counter
> driver for differential encoder cable status changes. There is no IRQ
> handling so spin_lock calls are used for protection.
> 
> Fixes: bbef69e088c3 ("counter: 104-quad-8: Support Differential Encoder
> Cable Status")
> 
> Signed-off-by: Syed Nayyar Waris <syednwaris@...il.com>
> 
> Split the patch from generic driver interface and clock prescaler
> related code changes. Also, include more code statements for protection
> using spin_lock calls and remove protection from few code statements as
> they are unnecessary.
> ---

Hello Syed,

Just like in the first patch, move these comments below the "---" line.

>  drivers/counter/104-quad-8.c | 12 +++++++++++-
>  1 file changed, 11 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/counter/104-quad-8.c b/drivers/counter/104-quad-8.c
> index 9dab190..1ce9660 100644
> --- a/drivers/counter/104-quad-8.c
> +++ b/drivers/counter/104-quad-8.c
> @@ -1153,16 +1153,22 @@ static ssize_t quad8_signal_cable_fault_read(struct counter_device *counter,
>  {
>  	const struct quad8_iio *const priv = counter->priv;
>  	const size_t channel_id = signal->id / 2;
> -	const bool disabled = !(priv->cable_fault_enable & BIT(channel_id));
> +	bool disabled;
>  	unsigned int status;
>  	unsigned int fault;
>  
> +	spin_lock(&((struct quad8_iio *)priv)->lock);

You can redeclare priv whenever you need to avoid these casts:

	struct quad8_iio *const priv = counter->priv;
	...
	spin_lock(&priv->lock);
	...
	spin_unlock(&priv->lock);

> +
> +	disabled = !(priv->cable_fault_enable & BIT(channel_id));
> +
>  	if (disabled)
>  		return -EINVAL;

This return statement can cause a deadlock. You can avoid that by
calling spin_unlock before the return:

	if (disabled) {
		spin_unlock(&priv->lock);
		return -EINVAL;
	}

Sincerely,

William Breathitt Gray

>  
>  	/* Logic 0 = cable fault */
>  	status = inb(priv->base + QUAD8_DIFF_ENCODER_CABLE_STATUS);
>  
> +	spin_unlock(&((struct quad8_iio *)priv)->lock);
> +
>  	/* Mask respective channel and invert logic */
>  	fault = !(status & BIT(channel_id));
>  
> @@ -1194,6 +1200,8 @@ static ssize_t quad8_signal_cable_fault_enable_write(
>  	if (ret)
>  		return ret;
>  
> +	spin_lock(&priv->lock);
> +
>  	if (enable)
>  		priv->cable_fault_enable |= BIT(channel_id);
>  	else
> @@ -1204,6 +1212,8 @@ static ssize_t quad8_signal_cable_fault_enable_write(
>  
>  	outb(cable_fault_enable, priv->base + QUAD8_DIFF_ENCODER_CABLE_STATUS);
>  
> +	spin_unlock(&priv->lock);
> +
>  	return len;
>  }
>  
> -- 
> 2.7.4
> 

Download attachment "signature.asc" of type "application/pgp-signature" (834 bytes)