linux-kernel - Re: [PATCH 04/11] cxl/core: Clamp max lock

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]

Date:   Wed, 9 Mar 2022 18:26:50 +0000
From:   Jonathan Cameron <Jonathan.Cameron@...wei.com>
To:     Dan Williams <dan.j.williams@...el.com>
CC:     <gregkh@...uxfoundation.org>, <rafael.j.wysocki@...el.com>,
        "Alison Schofield" <alison.schofield@...el.com>,
        Vishal Verma <vishal.l.verma@...el.com>,
        Ira Weiny <ira.weiny@...el.com>,
        Ben Widawsky <ben.widawsky@...el.com>,
        <linux-kernel@...r.kernel.org>, <linux-cxl@...r.kernel.org>,
        <nvdimm@...ts.linux.dev>
Subject: Re: [PATCH 04/11] cxl/core: Clamp max lock_class

On Mon, 28 Feb 2022 18:49:11 -0800
Dan Williams <dan.j.williams@...el.com> wrote:

> MAX_LOCKDEP_SUBCLASSES limits the depth of the CXL topology that can be
> validated by lockdep. Given that the cxl_test topology is already at
> this limit collapse some of the levels and clamp the max depth.
> 
> Cc: Alison Schofield <alison.schofield@...el.com>
> Cc: Vishal Verma <vishal.l.verma@...el.com>
> Cc: Ira Weiny <ira.weiny@...el.com>
> Cc: Ben Widawsky <ben.widawsky@...el.com>
> Signed-off-by: Dan Williams <dan.j.williams@...el.com>
> ---
>  drivers/cxl/cxl.h |   21 +++++++++++++++++----
>  1 file changed, 17 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/cxl/cxl.h b/drivers/cxl/cxl.h
> index 97e6ca7e4940..1357a245037d 100644
> --- a/drivers/cxl/cxl.h
> +++ b/drivers/cxl/cxl.h
> @@ -501,20 +501,33 @@ enum cxl_lock_class {
>  	CXL_ANON_LOCK,
>  	CXL_NVDIMM_LOCK,
>  	CXL_NVDIMM_BRIDGE_LOCK,

I'd be tempted to give explicit value to the one above as well
so it's immediate clear there is deliberate duplication here.

> -	CXL_PORT_LOCK,
> +	CXL_PORT_LOCK = 2,
>  	/*
>  	 * Be careful to add new lock classes here, CXL_PORT_LOCK is
>  	 * extended by the port depth, so a maximum CXL port topology
> -	 * depth would need to be defined first.
> +	 * depth would need to be defined first. Also, the max
> +	 * validation depth is limited by MAX_LOCKDEP_SUBCLASSES.
>  	 */
>  };
>  
> +static inline int clamp_lock_class(struct device *dev, int lock_class)
> +{
> +	if (lock_class >= MAX_LOCKDEP_SUBCLASSES) {
> +		dev_warn_once(dev,
> +			      "depth: %d, disabling lockdep for this device\n",
> +			      lock_class);
> +		return 0;
> +	}
> +
> +	return lock_class;
> +}
> +
>  static inline int cxl_lock_class(struct device *dev)
>  {
>  	if (is_cxl_port(dev)) {
>  		struct cxl_port *port = to_cxl_port(dev);
>  
> -		return CXL_PORT_LOCK + port->depth;
> +		return clamp_lock_class(dev, CXL_PORT_LOCK + port->depth);
>  	} else if (is_cxl_decoder(dev)) {
>  		struct cxl_port *port = to_cxl_port(dev->parent);
>  
> @@ -522,7 +535,7 @@ static inline int cxl_lock_class(struct device *dev)
>  		 * A decoder is the immediate child of a port, so set
>  		 * its lock class equal to other child device siblings.
>  		 */
> -		return CXL_PORT_LOCK + port->depth + 1;
> +		return clamp_lock_class(dev, CXL_PORT_LOCK + port->depth + 1);
>  	} else if (is_cxl_nvdimm_bridge(dev))
>  		return CXL_NVDIMM_BRIDGE_LOCK;
>  	else if (is_cxl_nvdimm(dev))
>