From: Leon Romanovsky <leon@kernel.org>
To: Jens Axboe <axboe@kernel.dk>, Jason Gunthorpe <jgg@ziepe.ca>,
Robin Murphy <robin.murphy@arm.com>,
Joerg Roedel <joro@8bytes.org>, Will Deacon <will@kernel.org>,
Keith Busch <kbusch@kernel.org>, Christoph Hellwig <hch@lst.de>,
"Zeng, Oak" <oak.zeng@intel.com>,
Chaitanya Kulkarni <kch@nvidia.com>
Cc: "Leon Romanovsky" <leonro@nvidia.com>,
"Sagi Grimberg" <sagi@grimberg.me>,
"Bjorn Helgaas" <bhelgaas@google.com>,
"Logan Gunthorpe" <logang@deltatee.com>,
"Yishai Hadas" <yishaih@nvidia.com>,
"Shameer Kolothum" <shameerali.kolothum.thodi@huawei.com>,
"Kevin Tian" <kevin.tian@intel.com>,
"Alex Williamson" <alex.williamson@redhat.com>,
"Marek Szyprowski" <m.szyprowski@samsung.com>,
"Jérôme Glisse" <jglisse@redhat.com>,
"Andrew Morton" <akpm@linux-foundation.org>,
linux-block@vger.kernel.org, linux-kernel@vger.kernel.org,
linux-rdma@vger.kernel.org, iommu@lists.linux.dev,
linux-nvme@lists.infradead.org, linux-pci@vger.kernel.org,
kvm@vger.kernel.org, linux-mm@kvack.org
Subject: [RFC v2 01/21] iommu/dma: Provide an interface to allow preallocate IOVA
Date: Thu, 12 Sep 2024 14:15:36 +0300 [thread overview]
Message-ID: <8ae3944565cd7b140625a71b8c7e74ca466bd3ec.1726138681.git.leon@kernel.org> (raw)
In-Reply-To: <cover.1726138681.git.leon@kernel.org>
From: Leon Romanovsky <leonro@nvidia.com>
Separate IOVA allocation to dedicated callback so it will allow
cache of IOVA and reuse it in fast paths for devices which support
ODP (on-demand-paging) mechanism.
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
---
drivers/iommu/dma-iommu.c | 57 ++++++++++++++++++++++++++++++---------
include/linux/iommu-dma.h | 11 ++++++++
2 files changed, 56 insertions(+), 12 deletions(-)
diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c
index 65a38b5695f9..09deea2fc86b 100644
--- a/drivers/iommu/dma-iommu.c
+++ b/drivers/iommu/dma-iommu.c
@@ -358,7 +358,7 @@ int iommu_dma_init_fq(struct iommu_domain *domain)
atomic_set(&cookie->fq_timer_on, 0);
/*
* Prevent incomplete fq state being observable. Pairs with path from
- * __iommu_dma_unmap() through iommu_dma_free_iova() to queue_iova()
+ * __iommu_dma_unmap() through __iommu_dma_free_iova() to queue_iova()
*/
smp_wmb();
WRITE_ONCE(cookie->fq_domain, domain);
@@ -759,7 +759,7 @@ static int dma_info_to_prot(enum dma_data_direction dir, bool coherent,
}
}
-static dma_addr_t iommu_dma_alloc_iova(struct iommu_domain *domain,
+static dma_addr_t __iommu_dma_alloc_iova(struct iommu_domain *domain,
size_t size, u64 dma_limit, struct device *dev)
{
struct iommu_dma_cookie *cookie = domain->iova_cookie;
@@ -805,7 +805,7 @@ static dma_addr_t iommu_dma_alloc_iova(struct iommu_domain *domain,
return (dma_addr_t)iova << shift;
}
-static void iommu_dma_free_iova(struct iommu_dma_cookie *cookie,
+static void __iommu_dma_free_iova(struct iommu_dma_cookie *cookie,
dma_addr_t iova, size_t size, struct iommu_iotlb_gather *gather)
{
struct iova_domain *iovad = &cookie->iovad;
@@ -842,7 +842,7 @@ static void __iommu_dma_unmap(struct device *dev, dma_addr_t dma_addr,
if (!iotlb_gather.queued)
iommu_iotlb_sync(domain, &iotlb_gather);
- iommu_dma_free_iova(cookie, dma_addr, size, &iotlb_gather);
+ __iommu_dma_free_iova(cookie, dma_addr, size, &iotlb_gather);
}
static dma_addr_t __iommu_dma_map(struct device *dev, phys_addr_t phys,
@@ -865,12 +865,12 @@ static dma_addr_t __iommu_dma_map(struct device *dev, phys_addr_t phys,
size = iova_align(iovad, size + iova_off);
- iova = iommu_dma_alloc_iova(domain, size, dma_mask, dev);
+ iova = __iommu_dma_alloc_iova(domain, size, dma_mask, dev);
if (!iova)
return DMA_MAPPING_ERROR;
if (iommu_map(domain, iova, phys - iova_off, size, prot, GFP_ATOMIC)) {
- iommu_dma_free_iova(cookie, iova, size, NULL);
+ __iommu_dma_free_iova(cookie, iova, size, NULL);
return DMA_MAPPING_ERROR;
}
return iova + iova_off;
@@ -973,7 +973,7 @@ static struct page **__iommu_dma_alloc_noncontiguous(struct device *dev,
return NULL;
size = iova_align(iovad, size);
- iova = iommu_dma_alloc_iova(domain, size, dev->coherent_dma_mask, dev);
+ iova = __iommu_dma_alloc_iova(domain, size, dev->coherent_dma_mask, dev);
if (!iova)
goto out_free_pages;
@@ -1007,7 +1007,7 @@ static struct page **__iommu_dma_alloc_noncontiguous(struct device *dev,
out_free_sg:
sg_free_table(sgt);
out_free_iova:
- iommu_dma_free_iova(cookie, iova, size, NULL);
+ __iommu_dma_free_iova(cookie, iova, size, NULL);
out_free_pages:
__iommu_dma_free_pages(pages, count);
return NULL;
@@ -1434,7 +1434,7 @@ int iommu_dma_map_sg(struct device *dev, struct scatterlist *sg, int nents,
if (!iova_len)
return __finalise_sg(dev, sg, nents, 0);
- iova = iommu_dma_alloc_iova(domain, iova_len, dma_get_mask(dev), dev);
+ iova = __iommu_dma_alloc_iova(domain, iova_len, dma_get_mask(dev), dev);
if (!iova) {
ret = -ENOMEM;
goto out_restore_sg;
@@ -1451,7 +1451,7 @@ int iommu_dma_map_sg(struct device *dev, struct scatterlist *sg, int nents,
return __finalise_sg(dev, sg, nents, iova);
out_free_iova:
- iommu_dma_free_iova(cookie, iova, iova_len, NULL);
+ __iommu_dma_free_iova(cookie, iova, iova_len, NULL);
out_restore_sg:
__invalidate_sg(sg, nents);
out:
@@ -1710,6 +1710,39 @@ size_t iommu_dma_max_mapping_size(struct device *dev)
return SIZE_MAX;
}
+int iommu_dma_alloc_iova(struct dma_iova_state *state, phys_addr_t phys,
+ size_t size)
+{
+ struct iommu_domain *domain = iommu_get_dma_domain(state->dev);
+ struct iommu_dma_cookie *cookie = domain->iova_cookie;
+ struct iova_domain *iovad = &cookie->iovad;
+ dma_addr_t addr;
+
+ size = iova_align(iovad, size + iova_offset(iovad, phys));
+ addr = __iommu_dma_alloc_iova(domain, size, dma_get_mask(state->dev),
+ state->dev);
+ if (addr == DMA_MAPPING_ERROR)
+ return -EINVAL;
+
+ state->addr = addr;
+ state->size = size;
+ return 0;
+}
+
+void iommu_dma_free_iova(struct dma_iova_state *state)
+{
+ struct iommu_domain *domain = iommu_get_dma_domain(state->dev);
+ struct iommu_dma_cookie *cookie = domain->iova_cookie;
+ struct iova_domain *iovad = &cookie->iovad;
+ size_t iova_off = iova_offset(iovad, state->addr);
+ struct iommu_iotlb_gather iotlb_gather;
+
+ iommu_iotlb_gather_init(&iotlb_gather);
+ __iommu_dma_free_iova(cookie, state->addr - iova_off,
+ iova_align(iovad, state->size + iova_off),
+ &iotlb_gather);
+}
+
void iommu_setup_dma_ops(struct device *dev)
{
struct iommu_domain *domain = iommu_get_domain_for_dev(dev);
@@ -1746,7 +1779,7 @@ static struct iommu_dma_msi_page *iommu_dma_get_msi_page(struct device *dev,
if (!msi_page)
return NULL;
- iova = iommu_dma_alloc_iova(domain, size, dma_get_mask(dev), dev);
+ iova = __iommu_dma_alloc_iova(domain, size, dma_get_mask(dev), dev);
if (!iova)
goto out_free_page;
@@ -1760,7 +1793,7 @@ static struct iommu_dma_msi_page *iommu_dma_get_msi_page(struct device *dev,
return msi_page;
out_free_iova:
- iommu_dma_free_iova(cookie, iova, size, NULL);
+ __iommu_dma_free_iova(cookie, iova, size, NULL);
out_free_page:
kfree(msi_page);
return NULL;
diff --git a/include/linux/iommu-dma.h b/include/linux/iommu-dma.h
index 13874f95d77f..698df67b152a 100644
--- a/include/linux/iommu-dma.h
+++ b/include/linux/iommu-dma.h
@@ -57,6 +57,9 @@ void iommu_dma_sync_sg_for_cpu(struct device *dev, struct scatterlist *sgl,
int nelems, enum dma_data_direction dir);
void iommu_dma_sync_sg_for_device(struct device *dev, struct scatterlist *sgl,
int nelems, enum dma_data_direction dir);
+int iommu_dma_alloc_iova(struct dma_iova_state *state, phys_addr_t phys,
+ size_t size);
+void iommu_dma_free_iova(struct dma_iova_state *state);
#else
static inline bool use_dma_iommu(struct device *dev)
{
@@ -173,5 +176,13 @@ static inline void iommu_dma_sync_sg_for_device(struct device *dev,
enum dma_data_direction dir)
{
}
+static inline int iommu_dma_alloc_iova(struct dma_iova_state *state,
+ phys_addr_t phys, size_t size)
+{
+ return -EOPNOTSUPP;
+}
+static inline void iommu_dma_free_iova(struct dma_iova_state *state)
+{
+}
#endif /* CONFIG_IOMMU_DMA */
#endif /* _LINUX_IOMMU_DMA_H */
--
2.46.0
next prev parent reply other threads:[~2024-09-12 11:16 UTC|newest]
Thread overview: 22+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-09-12 11:15 [RFC v2 00/21] Provide a new two step DMA API mapping API Leon Romanovsky
2024-09-12 11:15 ` Leon Romanovsky [this message]
2024-09-12 11:15 ` [RFC v2 02/21] iommu/dma: Implement link/unlink ranges callbacks Leon Romanovsky
2024-09-12 11:15 ` [RFC v2 03/21] iommu/dma: Add check if IOVA can be used Leon Romanovsky
2024-09-12 11:15 ` [RFC v2 04/21] dma-mapping: initialize IOVA state struct Leon Romanovsky
2024-09-12 11:15 ` [RFC v2 05/21] dma-mapping: provide an interface to allocate IOVA Leon Romanovsky
2024-09-12 11:15 ` [RFC v2 06/21] dma-mapping: set and query DMA IOVA state Leon Romanovsky
2024-09-12 11:15 ` [RFC v2 07/21] dma-mapping: implement link range API Leon Romanovsky
2024-09-12 11:15 ` [RFC v2 08/21] mm/hmm: let users to tag specific PFN with DMA mapped bit Leon Romanovsky
2024-09-12 11:15 ` [RFC v2 09/21] dma-mapping: provide callbacks to link/unlink HMM PFNs to specific IOVA Leon Romanovsky
2024-09-12 11:15 ` [RFC v2 10/21] RDMA/umem: Preallocate and cache IOVA for UMEM ODP Leon Romanovsky
2024-09-12 11:15 ` [RFC v2 11/21] RDMA/umem: Store ODP access mask information in PFN Leon Romanovsky
2024-09-12 11:15 ` [RFC v2 12/21] RDMA/core: Separate DMA mapping to caching IOVA and page linkage Leon Romanovsky
2024-09-12 11:15 ` [RFC v2 13/21] RDMA/umem: Prevent UMEM ODP creation with SWIOTLB Leon Romanovsky
2024-09-12 11:15 ` [RFC v2 14/21] vfio/mlx5: Explicitly use number of pages instead of allocated length Leon Romanovsky
2024-09-12 11:15 ` [RFC v2 15/21] vfio/mlx5: Rewrite create mkey flow to allow better code reuse Leon Romanovsky
2024-09-12 11:15 ` [RFC v2 16/21] vfio/mlx5: Explicitly store page list Leon Romanovsky
2024-09-12 11:15 ` [RFC v2 17/21] vfio/mlx5: Convert vfio to use DMA link API Leon Romanovsky
2024-09-12 11:15 ` [RFC v2 18/21] nvme-pci: remove optimizations for single DMA entry Leon Romanovsky
2024-09-12 11:15 ` [RFC v2 19/21] nvme-pci: precalculate number of DMA entries for each command Leon Romanovsky
2024-09-12 11:15 ` [RFC v2 20/21] nvme-pci: use new dma API Leon Romanovsky
2024-09-12 11:15 ` [RFC v2 21/21] nvme-pci: don't allow mapping of bvecs with offset Leon Romanovsky
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=8ae3944565cd7b140625a71b8c7e74ca466bd3ec.1726138681.git.leon@kernel.org \
--to=leon@kernel.org \
--cc=akpm@linux-foundation.org \
--cc=alex.williamson@redhat.com \
--cc=axboe@kernel.dk \
--cc=bhelgaas@google.com \
--cc=hch@lst.de \
--cc=iommu@lists.linux.dev \
--cc=jgg@ziepe.ca \
--cc=jglisse@redhat.com \
--cc=joro@8bytes.org \
--cc=kbusch@kernel.org \
--cc=kch@nvidia.com \
--cc=kevin.tian@intel.com \
--cc=kvm@vger.kernel.org \
--cc=leonro@nvidia.com \
--cc=linux-block@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linux-nvme@lists.infradead.org \
--cc=linux-pci@vger.kernel.org \
--cc=linux-rdma@vger.kernel.org \
--cc=logang@deltatee.com \
--cc=m.szyprowski@samsung.com \
--cc=oak.zeng@intel.com \
--cc=robin.murphy@arm.com \
--cc=sagi@grimberg.me \
--cc=shameerali.kolothum.thodi@huawei.com \
--cc=will@kernel.org \
--cc=yishaih@nvidia.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox