Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Wed, 9 Aug 2023 16:55:20 +0300 (EEST)
From:   =?ISO-8859-15?Q?Ilpo_J=E4rvinen?= <ilpo.jarvinen@linux.intel.com>
To:     Sui Jingfeng <sui.jingfeng@linux.dev>
cc:     Bjorn Helgaas <bhelgaas@google.com>,
        Dave Airlie <airlied@redhat.com>,
        Daniel Vetter <daniel@ffwll.ch>, linux-pci@vger.kernel.org,
        dri-devel@lists.freedesktop.org,
        LKML <linux-kernel@vger.kernel.org>,
        Sui Jingfeng <suijingfeng@loongson.cn>
Subject: Re: [PATCH v2 05/11] PCI/VGA: Move the new_state assignment out of
 the loop
In-Reply-To: <20230808223412.1743176-6-sui.jingfeng@linux.dev>
Message-ID: <ce60a772-4a6d-e31d-a195-c6a7e36ff26@linux.intel.com>
References: <20230808223412.1743176-1-sui.jingfeng@linux.dev> <20230808223412.1743176-6-sui.jingfeng@linux.dev>
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Precedence: bulk

On Wed, 9 Aug 2023, Sui Jingfeng wrote:

> From: Sui Jingfeng <suijingfeng@loongson.cn>
> 
> In the vga_arbiter_notify_clients() function, the value of the 'new_state'
> variable will be 'false' on systems that have more than one VGA device.
> The value will be 'true' if there is only one VGA device or no VGA device
> at all. Hence, its value is not relevant to the iteration of the loop.
> 
> So move the assignment clause out of the loop. For a system with multiple
> video cards, this patch saves unnecessary assignment.
> 
> Signed-off-by: Sui Jingfeng <suijingfeng@loongson.cn>
> ---
>  drivers/pci/vgaarb.c | 16 +++++++---------
>  1 file changed, 7 insertions(+), 9 deletions(-)
> 
> diff --git a/drivers/pci/vgaarb.c b/drivers/pci/vgaarb.c
> index dc10a262fb5e..6883067a802a 100644
> --- a/drivers/pci/vgaarb.c
> +++ b/drivers/pci/vgaarb.c
> @@ -1468,22 +1468,20 @@ static void vga_arbiter_notify_clients(void)
>  {
>  	struct vga_device *vgadev;
>  	unsigned long flags;
> -	uint32_t new_decodes;
> -	bool new_state;
> +	bool state;
>  
>  	if (!vga_arbiter_used)
>  		return;
>  
> +	state = (vga_count > 1) ? false : true;
> +

Is it safe to move this outside of the lock?

This would be enough (no need for ?: construct):

state = vga_count <= 1;

Or if you want to keep it as > 1:

state = !(vga_count > 1);

>  	spin_lock_irqsave(&vga_lock, flags);
>  	list_for_each_entry(vgadev, &vga_list, list) {
> -		if (vga_count > 1)
> -			new_state = false;
> -		else
> -			new_state = true;
>  		if (vgadev->set_decode) {
> -			new_decodes = vgadev->set_decode(vgadev->pdev,
> -							 new_state);
> -			vga_update_device_decodes(vgadev, new_decodes);
> +			unsigned int decodes;
> +
> +			decodes = vgadev->set_decode(vgadev->pdev, state);
> +			vga_update_device_decodes(vgadev, decodes);
>  		}
>  	}
>  	spin_unlock_irqrestore(&vga_lock, flags);
> 


-- 
 i.