linux-kernel - Re: [PATCH] iommu/iova: Replace cmpxchg with xchg in queue

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: <3afcc7b2-0bfb-b79c-513f-1beb66c5f164@arm.com>
Date:   Thu, 27 Aug 2020 20:41:36 +0100
From:   Robin Murphy <robin.murphy@....com>
To:     Shaokun Zhang <zhangshaokun@...ilicon.com>,
        iommu@...ts.linux-foundation.org, linux-kernel@...r.kernel.org
Cc:     Yuqi Jin <jinyuqi@...wei.com>
Subject: Re: [PATCH] iommu/iova: Replace cmpxchg with xchg in queue_iova

On 2020-08-27 09:43, Shaokun Zhang wrote:
> From: Yuqi Jin <jinyuqi@...wei.com>
> 
> The performance of the atomic_xchg is better than atomic_cmpxchg because
> no comparison is required. While the value of @fq_timer_on can only be 0
> or 1. Let's use atomic_xchg instead of atomic_cmpxchg here because we
> only need to check that the value changes from 0 to 1 or from 1 to 1.

Looks reasonable to me - the "compare" part is already covered by the 
separate atomic_read(), and as you say there's no harm if the "exchange" 
part races as long as it's still atomic.

Reviewed-by: Robin Murphy <robin.murphy@....com>

> Cc: Joerg Roedel <joro@...tes.org>
> Signed-off-by: Yuqi Jin <jinyuqi@...wei.com>
> Signed-off-by: Shaokun Zhang <zhangshaokun@...ilicon.com>
> ---
>   drivers/iommu/iova.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/iommu/iova.c b/drivers/iommu/iova.c
> index 45a251da5453..30d969a4c5fd 100644
> --- a/drivers/iommu/iova.c
> +++ b/drivers/iommu/iova.c
> @@ -579,7 +579,7 @@ void queue_iova(struct iova_domain *iovad,
>   
>   	/* Avoid false sharing as much as possible. */
>   	if (!atomic_read(&iovad->fq_timer_on) &&
> -	    !atomic_cmpxchg(&iovad->fq_timer_on, 0, 1))
> +	    !atomic_xchg(&iovad->fq_timer_on, 1))
>   		mod_timer(&iovad->fq_timer,
>   			  jiffies + msecs_to_jiffies(IOVA_FQ_TIMEOUT));
>   }
>