linux-kernel - Re: [PATCH] driver core: fix async device shutdown hang

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: <20240919141658.GA3737785@thelio-3990X>
Date: Thu, 19 Sep 2024 07:16:58 -0700
From: Nathan Chancellor <nathan@...nel.org>
To: Stuart Hayes <stuart.w.hayes@...il.com>
Cc: linux-kernel@...r.kernel.org,
	Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
	"Rafael J . Wysocki" <rafael@...nel.org>,
	Martin Belanger <Martin.Belanger@...l.com>,
	Oliver O'Halloran <oohall@...il.com>,
	Daniel Wagner <dwagner@...e.de>, Keith Busch <kbusch@...nel.org>,
	Lukas Wunner <lukas@...ner.de>, David Jeffery <djeffery@...hat.com>,
	Jeremy Allison <jallison@....com>, Jens Axboe <axboe@...com>,
	Christoph Hellwig <hch@....de>, Sagi Grimberg <sagi@...mberg.me>,
	linux-nvme@...ts.infradead.org, Jan Kiszka <jan.kiszka@...mens.com>
Subject: Re: [PATCH] driver core: fix async device shutdown hang

On Wed, Sep 18, 2024 at 11:31:43PM -0500, Stuart Hayes wrote:
> Modify device_shutdown() so that supplier devices do not wait for
> consumer devices to be shut down first when the devlink is sync state
> only, since the consumer is not dependent on the supplier in this case.
> 
> Without this change, a circular dependency could hang the system.
> 
> Fixes: 8064952c6504 ("driver core: shut down devices asynchronously")
> 
> Signed-off-by: Stuart Hayes <stuart.w.hayes@...il.com>

All of my virtual testing seems happy with this change.

Tested-by: Nathan Chancellor <nathan@...nel.org>

> ---
> The patch this fixes is in driver-core-next and linux-next.
> 
> Please let me know if this needs to be a V2 or if it needs anything
> else... it is the identical patch I sent in yesterday, except I added
> a "Fixes:" tag and the comments.  Thank you for the help!
> 
>  drivers/base/core.c | 10 +++++++++-
>  1 file changed, 9 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/base/core.c b/drivers/base/core.c
> index b69b82da8837..76513e360496 100644
> --- a/drivers/base/core.c
> +++ b/drivers/base/core.c
> @@ -4898,8 +4898,16 @@ void device_shutdown(void)
>  
>  		idx = device_links_read_lock();
>  		list_for_each_entry_rcu(link, &dev->links.suppliers, c_node,
> -				device_links_read_lock_held())
> +				device_links_read_lock_held()) {
> +			/*
> +			 * sync_state_only suppliers don't need to wait,
> +			 * aren't reordered on devices_kset, so making them
> +			 * wait could result in a hang
> +			 */
> +			if (device_link_flag_is_sync_state_only(link->flags))
> +				continue;
>  			link->supplier->p->shutdown_after = cookie;
> +		}
>  		device_links_read_unlock(idx);
>  		put_device(dev);
>  
> -- 
> 2.39.3
>