linux-kernel - Re: [PATCH RFC] iommu/mediatek: Flush IOTLB completely only if domain has been attached

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 for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <cc053ab0-1ed3-0d2c-204d-a5ae85b759a5@collabora.com>
Date:   Mon, 29 May 2023 09:55:42 +0200
From:   AngeloGioacchino Del Regno 
        <angelogioacchino.delregno@...labora.com>
To:     Chen-Yu Tsai <wenst@...omium.org>, Yong Wu <yong.wu@...iatek.com>,
        Joerg Roedel <joro@...tes.org>, Will Deacon <will@...nel.org>,
        Robin Murphy <robin.murphy@....com>,
        Matthias Brugger <matthias.bgg@...il.com>
Cc:     iommu@...ts.linux.dev, linux-mediatek@...ts.infradead.org,
        linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH RFC] iommu/mediatek: Flush IOTLB completely only if domain
 has been attached

Il 26/05/23 10:53, Chen-Yu Tsai ha scritto:
> If an IOMMU domain was never attached, it lacks any linkage to the
> actual IOMMU hardware. Attempting to do flush_iotlb_all() on it will
> result in a NULL pointer dereference. This seems to happen after the
> recent IOMMU core rework in v6.4-rc1.
> 
>      Unable to handle kernel read from unreadable memory at virtual address 0000000000000018
>      Call trace:
>       mtk_iommu_flush_iotlb_all+0x20/0x80
>       iommu_create_device_direct_mappings.part.0+0x13c/0x230
>       iommu_setup_default_domain+0x29c/0x4d0
>       iommu_probe_device+0x12c/0x190
>       of_iommu_configure+0x140/0x208
>       of_dma_configure_id+0x19c/0x3c0
>       platform_dma_configure+0x38/0x88
>       really_probe+0x78/0x2c0
> 
> Check if the "bank" field has been filled in before actually attempting
> the IOTLB flush to avoid it. The IOTLB is also flushed when the device
> comes out of runtime suspend, so it should have a clean initial state.
> 
> Fixes: 08500c43d4f7 ("iommu/mediatek: Adjust the structure")
> Signed-off-by: Chen-Yu Tsai <wenst@...omium.org>

Not only ARM SMMU does this, others are doing the same, some in a different
form (walking a list)... So I agree with this being a valid fix.

Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@...labora.com>

> ---
> 
> I think this is a valid fix, but I'm not very familiar with the hardware
> or the design of the driver. The ARM SMMU drivers seem to do this as well.
> 
>   drivers/iommu/mtk_iommu.c | 3 ++-
>   1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/iommu/mtk_iommu.c b/drivers/iommu/mtk_iommu.c
> index aecc7d154f28..e93906d6e112 100644
> --- a/drivers/iommu/mtk_iommu.c
> +++ b/drivers/iommu/mtk_iommu.c
> @@ -781,7 +781,8 @@ static void mtk_iommu_flush_iotlb_all(struct iommu_domain *domain)
>   {
>   	struct mtk_iommu_domain *dom = to_mtk_domain(domain);
>   
> -	mtk_iommu_tlb_flush_all(dom->bank->parent_data);
> +	if (dom->bank)
> +		mtk_iommu_tlb_flush_all(dom->bank->parent_data);
>   }
>   
>   static void mtk_iommu_iotlb_sync(struct iommu_domain *domain,