On 29/04/2019 15:44, Julien Grall wrote:
> On RT, iommu_dma_map_msi_msg() may be called from non-preemptible
> context. This will lead to a splat with CONFIG_DEBUG_ATOMIC_SLEEP as
> the function is using spin_lock (they can sleep on RT).
>
> iommu_dma_map_msi_msg() is used to map the MSI page in the IOMMU PT
> and update the MSI message with the IOVA.
>
> Only the part to lookup for the MSI page requires to be called in
> preemptible context. As the MSI page cannot change over the lifecycle
> of the MSI interrupt, the lookup can be cached and re-used later on.
>
> iomma_dma_map_msi_msg() is now split in two functions:
> - iommu_dma_prepare_msi(): This function will prepare the mapping
> in the IOMMU and store the cookie in the structure msi_desc. This
> function should be called in preemptible context.
> - iommu_dma_compose_msi_msg(): This function will update the MSI
> message with the IOVA when the device is behind an IOMMU.
>
> Signed-off-by: Julien Grall <julien.grall@xxxxxxx>
>
> ---
> Changes in v2:
> - Rework the commit message to use imperative mood
> - Use the MSI accessor to get/set the iommu cookie
> - Don't use ternary on return
> - Select CONFIG_IRQ_MSI_IOMMU
> - Pass an msi_desc rather than the irq number
> ---
> drivers/iommu/Kconfig | 1 +
> drivers/iommu/dma-iommu.c | 47 ++++++++++++++++++++++++++++++++++++++---------
> include/linux/dma-iommu.h | 23 +++++++++++++++++++++++
> 3 files changed, 62 insertions(+), 9 deletions(-)
>
> diff --git a/drivers/iommu/Kconfig b/drivers/iommu/Kconfig
> index 6f07f3b21816..eb1c8cd243f9 100644
> --- a/drivers/iommu/Kconfig
> +++ b/drivers/iommu/Kconfig
> @@ -94,6 +94,7 @@ config IOMMU_DMA
> bool
> select IOMMU_API
> select IOMMU_IOVA
> + select IRQ_MSI_IOMMU
> select NEED_SG_DMA_LENGTH
>
> config FSL_PAMU
> diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c
> index 77aabe637a60..2309f59cefa4 100644
> --- a/drivers/iommu/dma-iommu.c
> +++ b/drivers/iommu/dma-iommu.c
> @@ -888,17 +888,18 @@ static struct iommu_dma_msi_page *iommu_dma_get_msi_page(struct device *dev,
> return NULL;
> }
>
> -void iommu_dma_map_msi_msg(int irq, struct msi_msg *msg)
> +int iommu_dma_prepare_msi(struct msi_desc *desc, phys_addr_t msi_addr)
> {
> - struct device *dev = msi_desc_to_dev(irq_get_msi_desc(irq));
> + struct device *dev = msi_desc_to_dev(desc);
> struct iommu_domain *domain = iommu_get_domain_for_dev(dev);
> struct iommu_dma_cookie *cookie;
> struct iommu_dma_msi_page *msi_page;
> - phys_addr_t msi_addr = (u64)msg->address_hi << 32 | msg->address_lo;
> unsigned long flags;
>
> - if (!domain || !domain->iova_cookie)
> - return;
> + if (!domain || !domain->iova_cookie) {
> + desc->iommu_cookie = NULL;
nit: This could be
msi_desc_set_iommu_cookie(desc, NULL);
now that you have introduced the relevant accessors.
> + return 0;
> + }
>
> cookie = domain->iova_cookie;
>
> @@ -911,7 +912,37 @@ void iommu_dma_map_msi_msg(int irq, struct msi_msg *msg)
> msi_page = iommu_dma_get_msi_page(dev, msi_addr, domain);
> spin_unlock_irqrestore(&cookie->msi_lock, flags);
>
> - if (WARN_ON(!msi_page)) {
> + msi_desc_set_iommu_cookie(desc, msi_page);
> +
> + if (!msi_page)
> + return -ENOMEM;
> + else
> + return 0;
> +}
> +
> +void iommu_dma_compose_msi_msg(struct msi_desc *desc,
> + struct msi_msg *msg)
> +{
> + struct device *dev = msi_desc_to_dev(desc);
> + const struct iommu_domain *domain = iommu_get_domain_for_dev(dev);
> + const struct iommu_dma_msi_page *msi_page;
> +
> + msi_page = msi_desc_get_iommu_cookie(desc);
> +
> + if (!domain || !domain->iova_cookie || WARN_ON(!msi_page))
> + return;
> +
> + msg->address_hi = upper_32_bits(msi_page->iova);
> + msg->address_lo &= cookie_msi_granule(domain->iova_cookie) - 1;
> + msg->address_lo += lower_32_bits(msi_page->iova);
> +}
> +
> +void iommu_dma_map_msi_msg(int irq, struct msi_msg *msg)
> +{
> + struct msi_desc *desc = irq_get_msi_desc(irq);
> + phys_addr_t msi_addr = (u64)msg->address_hi << 32 | msg->address_lo;
> +
> + if (WARN_ON(iommu_dma_prepare_msi(desc, msi_addr))) {
> /*
> * We're called from a void callback, so the best we can do is
> * 'fail' by filling the message with obviously bogus values.
> @@ -922,8 +953,6 @@ void iommu_dma_map_msi_msg(int irq, struct msi_msg *msg)
> msg->address_lo = ~0U;
> msg->data = ~0U;
> } else {
> - msg->address_hi = upper_32_bits(msi_page->iova);
> - msg->address_lo &= cookie_msi_granule(cookie) - 1;
> - msg->address_lo += lower_32_bits(msi_page->iova);
> + iommu_dma_compose_msi_msg(desc, msg);
> }
> }
> diff --git a/include/linux/dma-iommu.h b/include/linux/dma-iommu.h
> index e760dc5d1fa8..3fc48fbd6f63 100644
> --- a/include/linux/dma-iommu.h
> +++ b/include/linux/dma-iommu.h
> @@ -71,12 +71,24 @@ void iommu_dma_unmap_resource(struct device *dev, dma_addr_t handle,
> size_t size, enum dma_data_direction dir, unsigned long attrs);
>
> /* The DMA API isn't _quite_ the whole story, though... */
> +/*
> + * Map the MSI page in the IOMMU device and store it in @desc
> + *
> + * Return 0 if succeeded other an error if the preparation has failed.
s/other/or/
This could be a proper kerneldoc thing, while you're at it.
> + */
> +int iommu_dma_prepare_msi(struct msi_desc *desc, phys_addr_t msi_addr);
> +
> +/* Update the MSI message if required. */
> +void iommu_dma_compose_msi_msg(struct msi_desc *desc,
> + struct msi_msg *msg);
> +
> void iommu_dma_map_msi_msg(int irq, struct msi_msg *msg);
> void iommu_dma_get_resv_regions(struct device *dev, struct list_head *list);
>
> #else
>
> struct iommu_domain;
> +struct msi_desc;
> struct msi_msg;
> struct device;
>
> @@ -99,6 +111,17 @@ static inline void iommu_put_dma_cookie(struct iommu_domain *domain)
> {
> }
>
> +static inline int iommu_dma_prepare_msi(struct msi_desc *desc,
> + phys_addr_t msi_addr)
> +{
> + return 0;
> +}
> +
> +static inline void iommu_dma_compose_msi_msg(struct msi_desc *desc,
> + struct msi_msg *msg)
> +{
> +}
> +
> static inline void iommu_dma_map_msi_msg(int irq, struct msi_msg *msg)
> {
> }
>
Thanks,
M.
--
Jazz is not dead. It just smells funny...
