linux-kernel - Re: [PATCH v2 01/10] KVM: arm/arm64: vgic-its: fix return value for restore

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: <87030c8b-2464-8602-5672-e9ad29a9db90@arm.com>
Date:   Fri, 6 Oct 2017 15:37:15 +0100
From:   Andre Przywara <andre.przywara@....com>
To:     Eric Auger <eric.auger@...hat.com>, eric.auger.pro@...il.com,
        linux-kernel@...r.kernel.org, kvm@...r.kernel.org,
        marc.zyngier@....com, cdall@...aro.org, peter.maydell@...aro.org,
        wanghaibin.wang@...wei.com
Cc:     wu.wubin@...wei.com
Subject: Re: [PATCH v2 01/10] KVM: arm/arm64: vgic-its: fix return value for
 restore

Hi,

On 27/09/17 14:28, Eric Auger wrote:
> From: wanghaibin <wanghaibin.wang@...wei.com>
> 
> This patch fix the migrate restore tables failure.
> 
> The same scene, at the destination, the restore tables
> interface traversal guest memory, and check the dte/ite
> is valid or not.  If all dtes/ites are invalid, we will do
> try next one, and the last it will take the 1 return value,
> but currently, it be treated as error. That's not correct.
> 
> This patch try to fix this problem.
> 
> Signed-off-by: wanghaibin <wanghaibin.wang@...wei.com>

Looks right to me. But I wonder if we actually should go over the file
and unify the return value semantics or at least document them.
It's a bit puzzling to have functions which return negative errors and 0
*or 1* on success, and then functions which go with the traditional C
convention. That would help explaining the second hunk.

Also this return value handling is a bit weird in cases, like in
handle_l1_dte():

	if (ret <= 0)
		return ret;
	return 1;

which looks like a glorified "return ret;" in that case to me.

But actually this is just nitpicking and the actual patch seems correct.

Cheers,
Andre.

> ---
> 
> need to CC stable
> 
> v1 -> v2:
> - if (ret > 0) ret = 0
> ---
>  virt/kvm/arm/vgic/vgic-its.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/virt/kvm/arm/vgic/vgic-its.c b/virt/kvm/arm/vgic/vgic-its.c
> index f51c1e1..fbbc97b 100644
> --- a/virt/kvm/arm/vgic/vgic-its.c
> +++ b/virt/kvm/arm/vgic/vgic-its.c
> @@ -2018,7 +2018,7 @@ static int vgic_its_restore_dte(struct vgic_its *its, u32 id,
>  		return PTR_ERR(dev);
>  
>  	ret = vgic_its_restore_itt(its, dev);
> -	if (ret) {
> +	if (ret < 0) {
>  		vgic_its_free_device(its->dev->kvm, dev);
>  		return ret;
>  	}
> @@ -2141,7 +2141,7 @@ static int vgic_its_restore_device_tables(struct vgic_its *its)
>  	}
>  
>  	if (ret > 0)
> -		ret = -EINVAL;
> +		ret = 0;
>  
>  	return ret;
>  }
>