[PATCH v2 08/25] iommu: Allow an IDENTITY domain as the default_domain in ARM32
Robin Murphy
robin.murphy at arm.com
Thu Jun 1 11:57:25 PDT 2023
On 2023-05-16 01:00, Jason Gunthorpe wrote:
> Even though dma-iommu.c and CONFIG_ARM_DMA_USE_IOMMU do approximately the
> same stuff, the way they relate to the IOMMU core is quiet different.
>
> dma-iommu.c expects the core code to setup an UNMANAGED domain (of type
> IOMMU_DOMAIN_DMA) and then configures itself to use that domain. This
> becomes the default_domain for the group.
>
> ARM_DMA_USE_IOMMU does not use the default_domain, instead it directly
> allocates an UNMANAGED domain and operates it just like an external
> driver. In this case group->default_domain is NULL.
>
> If the driver provides a global static identity_domain then automatically
> use it as the default_domain when in ARM_DMA_USE_IOMMU mode.
>
> This allows drivers that implemented default_domain == NULL as an IDENTITY
> translation to trivially get a properly labeled non-NULL default_domain on
> ARM32 configs.
>
> With this arrangment when ARM_DMA_USE_IOMMU wants to disconnect from the
> device the normal detach_domain flow will restore the IDENTITY domain as
> the default domain. Overall this makes attach_dev() of the IDENTITY domain
> called in the same places as detach_dev().
>
> This effectively migrates these drivers to default_domain mode. For
> drivers that support ARM64 they will gain support for the IDENTITY
> translation mode for the dma_api and behave in a uniform way.
>
> Drivers use this by setting ops->identity_domain to a static singleton
> iommu_domain that implements the identity attach. If the core detects
> ARM_DMA_USE_IOMMU mode then it automatically attaches the IDENTITY domain
> during probe.
>
> Drivers can continue to prevent the use of DMA translation by returning
> IOMMU_DOMAIN_IDENTITY from def_domain_type, this will completely prevent
> IOMMU_DMA from running but will not impact ARM_DMA_USE_IOMMU.
>
> This allows removing the set_platform_dma_ops() from every remaining
> driver.
>
> Remove the set_platform_dma_ops from rockchip and mkt_v1 as all it does
> is set an existing global static identity domain. mkt_v1 does not support
> IOMMU_DOMAIN_DMA and it does not compile on ARM64 so this transformation
> is safe.
>
> Signed-off-by: Jason Gunthorpe <jgg at nvidia.com>
> ---
> drivers/iommu/iommu.c | 40 +++++++++++++++++++++++++++++-----
> drivers/iommu/mtk_iommu_v1.c | 12 ----------
> drivers/iommu/rockchip-iommu.c | 10 ---------
> 3 files changed, 35 insertions(+), 27 deletions(-)
>
> diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c
> index 8ba90571449cec..bed7cb6e5ee65b 100644
> --- a/drivers/iommu/iommu.c
> +++ b/drivers/iommu/iommu.c
> @@ -1757,18 +1757,48 @@ static int iommu_get_default_domain_type(struct iommu_group *group,
> int type;
>
> lockdep_assert_held(&group->mutex);
> +
> + /*
> + * ARM32 drivers supporting CONFIG_ARM_DMA_USE_IOMMU can declare an
> + * identity_domain and it will automatically become their default
> + * domain. Later on ARM_DMA_USE_IOMMU will install its UNMANAGED domain.
> + * Override the selection to IDENTITY if we are sure the driver supports
> + * it.
> + */
> + if (IS_ENABLED(CONFIG_ARM_DMA_USE_IOMMU) && ops->identity_domain) {
If I cared about arm-smmu on 32-bit, I'd bring that up again, but
honestly I'm not sure that I do... I think it might end up working after
patch #21, and it's currently still broken for lack of .set_platform_dma
anyway, so meh.
> + type = IOMMU_DOMAIN_IDENTITY;
> + if (best_type && type && best_type != type)
> + goto err;
> + best_type = target_type = IOMMU_DOMAIN_IDENTITY;
> + }
> +
> for_each_group_device(group, gdev) {
> type = best_type;
> if (ops->def_domain_type) {
> type = ops->def_domain_type(gdev->dev);
> - if (best_type && type && best_type != type)
> + if (best_type && type && best_type != type) {
> + /* Stick with the last driver override we saw */
> + best_type = type;
> goto err;
> + }
> }
>
> if (dev_is_pci(gdev->dev) && to_pci_dev(gdev->dev)->untrusted) {
> - type = IOMMU_DOMAIN_DMA;
> - if (best_type && type && best_type != type)
> - goto err;
> + /*
> + * We don't have any way for the iommu core code to
> + * force arm_iommu to activate so we can't enforce
> + * trusted. Log it and keep going with the IDENTITY
> + * default domain.
> + */
> + if (IS_ENABLED(CONFIG_ARM_DMA_USE_IOMMU)) {
> + dev_warn(
> + gdev->dev,
> + "PCI device is untrusted but ARM32 does not support secure IOMMU operation, continuing anyway.\n");
To within experimental error, this is dead code. The ARM DMA ops don't
even understand groups, so already have the much bigger problem of being
broken for any non-trivial PCI setup anyway. That's if you could even
find a 32-bit SoC with both PCI(e) and a relevant IOMMU. None of those
will have Thunderbolt, and I expect even fewer would be using the
"external-facing" DT property (which is likely newer than they are) for
any other reason.
Thanks,
Robin.
> + } else {
> + type = IOMMU_DOMAIN_DMA;
> + if (best_type && type && best_type != type)
> + goto err;
> + }
> }
> best_type = type;
> last_dev = gdev->dev;
> @@ -1790,7 +1820,7 @@ static int iommu_get_default_domain_type(struct iommu_group *group,
> "Device needs domain type %s, but device %s in the same iommu group requires type %s - using default\n",
> iommu_domain_type_str(type), dev_name(last_dev),
> iommu_domain_type_str(best_type));
> - return 0;
> + return best_type;
> }
>
> static void iommu_group_do_probe_finalize(struct device *dev)
> diff --git a/drivers/iommu/mtk_iommu_v1.c b/drivers/iommu/mtk_iommu_v1.c
> index cc3e7d53d33ad9..7c0c1d50df5f75 100644
> --- a/drivers/iommu/mtk_iommu_v1.c
> +++ b/drivers/iommu/mtk_iommu_v1.c
> @@ -337,11 +337,6 @@ static struct iommu_domain mtk_iommu_v1_identity_domain = {
> .ops = &mtk_iommu_v1_identity_ops,
> };
>
> -static void mtk_iommu_v1_set_platform_dma(struct device *dev)
> -{
> - mtk_iommu_v1_identity_attach(&mtk_iommu_v1_identity_domain, dev);
> -}
> -
> static int mtk_iommu_v1_map(struct iommu_domain *domain, unsigned long iova,
> phys_addr_t paddr, size_t pgsize, size_t pgcount,
> int prot, gfp_t gfp, size_t *mapped)
> @@ -457,11 +452,6 @@ static int mtk_iommu_v1_create_mapping(struct device *dev, struct of_phandle_arg
> return 0;
> }
>
> -static int mtk_iommu_v1_def_domain_type(struct device *dev)
> -{
> - return IOMMU_DOMAIN_IDENTITY;
> -}
> -
> static struct iommu_device *mtk_iommu_v1_probe_device(struct device *dev)
> {
> struct iommu_fwspec *fwspec = dev_iommu_fwspec_get(dev);
> @@ -599,10 +589,8 @@ static const struct iommu_ops mtk_iommu_v1_ops = {
> .probe_device = mtk_iommu_v1_probe_device,
> .probe_finalize = mtk_iommu_v1_probe_finalize,
> .release_device = mtk_iommu_v1_release_device,
> - .def_domain_type = mtk_iommu_v1_def_domain_type,
> .device_group = generic_device_group,
> .pgsize_bitmap = MT2701_IOMMU_PAGE_SIZE,
> - .set_platform_dma_ops = mtk_iommu_v1_set_platform_dma,
> .owner = THIS_MODULE,
> .default_domain_ops = &(const struct iommu_domain_ops) {
> .attach_dev = mtk_iommu_v1_attach_device,
> diff --git a/drivers/iommu/rockchip-iommu.c b/drivers/iommu/rockchip-iommu.c
> index ebce56d6e9c634..9e1296a856ac4c 100644
> --- a/drivers/iommu/rockchip-iommu.c
> +++ b/drivers/iommu/rockchip-iommu.c
> @@ -1026,13 +1026,6 @@ static struct iommu_domain rk_identity_domain = {
> .ops = &rk_identity_ops,
> };
>
> -#ifdef CONFIG_ARM
> -static void rk_iommu_set_platform_dma(struct device *dev)
> -{
> - WARN_ON(rk_iommu_identity_attach(&rk_identity_domain, dev));
> -}
> -#endif
> -
> static int rk_iommu_attach_device(struct iommu_domain *domain,
> struct device *dev)
> {
> @@ -1211,9 +1204,6 @@ static const struct iommu_ops rk_iommu_ops = {
> .probe_device = rk_iommu_probe_device,
> .release_device = rk_iommu_release_device,
> .device_group = rk_iommu_device_group,
> -#ifdef CONFIG_ARM
> - .set_platform_dma_ops = rk_iommu_set_platform_dma,
> -#endif
> .pgsize_bitmap = RK_IOMMU_PGSIZE_BITMAP,
> .of_xlate = rk_iommu_of_xlate,
> .default_domain_ops = &(const struct iommu_domain_ops) {
More information about the Linux-rockchip
mailing list