netdev - Re: [PATCH V1 net 2/4] net: ena: set initial DMA width to avoid intel iommu issue

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]

Date:   Wed, 18 Nov 2020 23:35:28 +0100
From:   Heiner Kallweit <hkallweit1@...il.com>
To:     Shay Agroskin <shayagr@...zon.com>, kuba@...nel.org,
        netdev@...r.kernel.org
Cc:     dwmw@...zon.com, zorik@...zon.com, matua@...zon.com,
        saeedb@...zon.com, msw@...zon.com, aliguori@...zon.com,
        nafea@...zon.com, gtzalik@...zon.com, netanel@...zon.com,
        alisaidi@...zon.com, benh@...zon.com, akiyano@...zon.com,
        sameehj@...zon.com, ndagan@...zon.com,
        Mike Cui <mikecui@...zon.com>
Subject: Re: [PATCH V1 net 2/4] net: ena: set initial DMA width to avoid intel
 iommu issue

Am 18.11.2020 um 22:59 schrieb Shay Agroskin:
> The ENA driver uses the readless mechanism, which uses DMA, to find
> out what the DMA mask is supposed to be.
> 
> If DMA is used without setting the dma_mask first, it causes the
> Intel IOMMU driver to think that ENA is a 32-bit device and therefore
> disables IOMMU passthrough permanently.
> 
> This patch sets the dma_mask to be ENA_MAX_PHYS_ADDR_SIZE_BITS=48
> before readless initialization in
> ena_device_init()->ena_com_mmio_reg_read_request_init(),
> which is large enough to workaround the intel_iommu issue.
> 
> DMA mask is set again to the correct value after it's received from the
> device after readless is initialized.
> 
> Fixes: 1738cd3ed342 ("net: ena: Add a driver for Amazon Elastic Network Adapters (ENA)")
> Signed-off-by: Mike Cui <mikecui@...zon.com>
> Signed-off-by: Arthur Kiyanovski <akiyano@...zon.com>
> Signed-off-by: Shay Agroskin <shayagr@...zon.com>
> ---
>  drivers/net/ethernet/amazon/ena/ena_netdev.c | 13 +++++++++++++
>  1 file changed, 13 insertions(+)
> 
> diff --git a/drivers/net/ethernet/amazon/ena/ena_netdev.c b/drivers/net/ethernet/amazon/ena/ena_netdev.c
> index 574c2b5ba21e..854a22e692bf 100644
> --- a/drivers/net/ethernet/amazon/ena/ena_netdev.c
> +++ b/drivers/net/ethernet/amazon/ena/ena_netdev.c
> @@ -4146,6 +4146,19 @@ static int ena_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
>  		return rc;
>  	}
>  
> +	rc = pci_set_dma_mask(pdev, DMA_BIT_MASK(ENA_MAX_PHYS_ADDR_SIZE_BITS));
> +	if (rc) {
> +		dev_err(&pdev->dev, "pci_set_dma_mask failed %d\n", rc);
> +		goto err_disable_device;
> +	}
> +
> +	rc = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(ENA_MAX_PHYS_ADDR_SIZE_BITS));
> +	if (rc) {
> +		dev_err(&pdev->dev, "err_pci_set_consistent_dma_mask failed %d\n",
> +			rc);
> +		goto err_disable_device;
> +	}
> +
>  	pci_set_master(pdev);
>  
>  	ena_dev = vzalloc(sizeof(*ena_dev));
> 

The old pci_ dma wrappers are being phased out and shouldn't be used in
new code. See e.g. e059c6f340f6 ("tulip: switch from 'pci_' to 'dma_' API").
So better use:
dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(ENA_MAX_PHYS_ADDR_SIZE_BITS));