Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Subject: Re: [PATCH 1/2] KVM: arm64: vgic: Forbid invalid userspace
 Redistributor accesses
To:     Marc Zyngier <maz@kernel.org>, Zenghui Yu <yuzenghui@huawei.com>
Cc:     suzuki.poulose@arm.com, linux-kernel@vger.kernel.org,
        james.morse@arm.com, linux-arm-kernel@lists.infradead.org,
        wanghaibin.wang@huawei.com, Keqian Zhu <zhukeqian1@huawei.com>,
        kvmarm@lists.cs.columbia.edu, julien.thierry.kdev@gmail.com
References: <20201113142801.1659-1-yuzenghui@huawei.com>
 <20201113142801.1659-2-yuzenghui@huawei.com>
 <724c43702b52aac0d3c9beb9604d1bfb@kernel.org>
 <584b7ff1-ecf2-b0ec-cea3-ccc29902f43a@huawei.com>
 <cc45285fe491aff5c28a24f94c124508@kernel.org>
 <7e58200c-814e-3598-155a-9a7e6cc24374@huawei.com>
 <c20865a267e44d1e2c0d52ce4e012263@kernel.org>
From:   Auger Eric <eric.auger@redhat.com>
Message-ID: <5ba4a98e-276b-2462-0580-fe0e007e9b38@redhat.com>
Date:   Tue, 17 Nov 2020 10:59:27 +0100
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101
 Thunderbird/68.5.0
MIME-Version: 1.0
In-Reply-To: <c20865a267e44d1e2c0d52ce4e012263@kernel.org>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 8bit
Precedence: bulk

Hi Marc,

On 11/17/20 9:49 AM, Marc Zyngier wrote:
> Hi Zenghui,
> 
> On 2020-11-16 14:57, Zenghui Yu wrote:
>> Hi Marc,
>>
>> On 2020/11/16 22:10, Marc Zyngier wrote:
>>>> My take is that only if the "[Re]Distributor base address" is specified
>>>> in the system memory map, will the user-provided kvm_device_attr.offset
>>>> make sense. And we can then handle the access to the register which is
>>>> defined by "base address + offset".
>>>
>>> I'd tend to agree, but it is just that this is a large change at -rc4.
>>> I'd rather have a quick fix for 5.10, and a more invasive change for
>>> 5.11,
>>> spanning all the possible vgic devices.
>>
>> So you prefer fixing it by "return a value that doesn't have the Last
>> bit set" for v5.10? I'm ok with it and can send v2 for it.
> 
> Cool. Thanks for that.
> 
>> Btw, looking again at the way we handle the user-reading of GICR_TYPER
>>
>>     vgic_mmio_read_v3r_typer(vcpu, addr, len)
>>
>> it seems that @addr is actually the *offset* of GICR_TYPER (0x0008) and
>> @addr is unlikely to be equal to last_rdist_typer, which is the *GPA* of
>> the last RD. Looks like the user-reading of GICR_TYPER.Last is always
>> broken?
> 
> I think you are right. Somehow, we don't seem to track the index of
> the RD in the region, so we can never compute the address of the RD
> even if the base address is set.
> 
> Let's drop the reporting of Last for userspace for now, as it never
> worked. If you post a patch addressing that quickly, I'll get it to
> Paolo by the end of the week (there's another fix that needs merging).
> 
> Eric: do we have any test covering the userspace API?

So as this issue seems related to the changes made when implementing the
multiple RDIST regions, I volunteer to write those KVM selftests :-)

Thanks

Eric

> 
> Thanks,
> 
>         M.