[PATCH v5 0/3] iommu: Clean up cookie and sw_msi in struct iommu_domain

Nicolin Chen posted 3 patches 8 months, 4 weeks ago
drivers/iommu/iommufd/Kconfig           |   2 +-
drivers/iommu/dma-iommu.h               |  14 ++
drivers/iommu/iommu-priv.h              |  13 ++
drivers/iommu/iommufd/iommufd_private.h |   7 +-
include/linux/iommu.h                   |  35 ++--
drivers/iommu/dma-iommu.c               | 208 ++++++++++++------------
drivers/iommu/iommu-sva.c               |   1 +
drivers/iommu/iommu.c                   |  36 +++-
drivers/iommu/iommufd/device.c          | 131 +--------------
drivers/iommu/iommufd/driver.c          | 126 ++++++++++++++
drivers/iommu/iommufd/hw_pagetable.c    |   6 +-
11 files changed, 323 insertions(+), 256 deletions(-)
[PATCH v5 0/3] iommu: Clean up cookie and sw_msi in struct iommu_domain
Posted by Nicolin Chen 8 months, 4 weeks ago
A few things in iommu_domain structure (such as iova_cookie, iommufd_hwpt,
fault handler, and etc) are exclusive to certain types of iommu_domains.
So, all of them can be seen as domain cookies in different cookie types.

Furthermore, iova_cookie is shared with the msi_cookie use case. It could
be cleaner by decoupling these too.

Define an enum iommu_domain_cookie_type to fit all those cookies in.

Since there are only two sw_msi implementations, check domain->cookie_type
in iommu_dma_prepare_msi() to call the corresponding sw_msi function.

This is a clean-up series for the previous sw_msi Part-1 core series. It's
on github:
https://github.com/nicolinc/iommufd/commits/iommufd_msi_cleanup-v5

Changelog
v5
 * Add IMPORT line for IOMMUFD_INTERNAL and change IOMMUFD_DRIVER_CORE
   to a "bool" module, fixing build module errors reported by Arnd:
   https://lore.kernel.org/all/20250324210329.2809869-1-arnd@kernel.org/
 * In iommu_dma_prepare_msi(), let all IDENTITY DOMAINs pass through,
   fixing the regression with IDENTITY domains reported by Nathan:
   https://lore.kernel.org/linux-iommu/20250324162558.GA198799@ax162/
v4
 https://lore.kernel.org/all/cover.1741294235.git.nicolinc@nvidia.com/
 * Rebase on top of a bug fix for hwpt_iommufd cookie
   https://lore.kernel.org/all/20250305211800.229465-1-nicolinc@nvidia.com/
 * Add Reviewed-by from Robin and Kevin
 * Add missing "break;" in iommu_domain_free()
 * Drop "linux/msi.h" header from the iommufd/device code
 * Set iommufd_sw_msi_install's SYMBOL to "IOMMUFD_INTERNAL"
 * Update commit messages for typo fixing and size comparison
 * Add a local msi_page_list to cache cookie_msi_pages(domain) return
 * Add an inline iommufd_group_setup_msi() for !CONFIG_IRQ_MSI_IOMMU
v3
 https://lore.kernel.org/all/cover.1741034885.git.nicolinc@nvidia.com/
 * Integrate Robin's patch
   https://lore.kernel.org/linux-iommu/4e68d5820be06adc1b34fc0d1c9399481151daee.1740742271.git.robin.murphy@arm.com/
   replacing private_data_owner with the new cookie_type
 * Move iommu_sw_msi stub from the public header to iommu-priv.h
v2
 https://lore.kernel.org/all/cover.1740705776.git.nicolinc@nvidia.com/
 * Drop sw_msi function pointer
 * Add a new private_data_owner tag in iommu_domain
 * Let iommu core call the two sw_msi implementations directly
v1
 https://lore.kernel.org/all/cover.1740600272.git.nicolinc@nvidia.com/

Thanks
Nicolin

Nicolin Chen (2):
  iommufd: Move iommufd_sw_msi and related functions to driver.c
  iommu: Drop sw_msi from iommu_domain

Robin Murphy (1):
  iommu: Sort out domain user data

 drivers/iommu/iommufd/Kconfig           |   2 +-
 drivers/iommu/dma-iommu.h               |  14 ++
 drivers/iommu/iommu-priv.h              |  13 ++
 drivers/iommu/iommufd/iommufd_private.h |   7 +-
 include/linux/iommu.h                   |  35 ++--
 drivers/iommu/dma-iommu.c               | 208 ++++++++++++------------
 drivers/iommu/iommu-sva.c               |   1 +
 drivers/iommu/iommu.c                   |  36 +++-
 drivers/iommu/iommufd/device.c          | 131 +--------------
 drivers/iommu/iommufd/driver.c          | 126 ++++++++++++++
 drivers/iommu/iommufd/hw_pagetable.c    |   6 +-
 11 files changed, 323 insertions(+), 256 deletions(-)


base-commit: da0c56520e880441d0503d0cf0d6853dcfb5f1a4
-- 
2.43.0
Re: [PATCH v5 0/3] iommu: Clean up cookie and sw_msi in struct iommu_domain
Posted by Jason Gunthorpe 8 months, 3 weeks ago
On Mon, Mar 24, 2025 at 09:05:14PM -0700, Nicolin Chen wrote:

> Nicolin Chen (2):
>   iommufd: Move iommufd_sw_msi and related functions to driver.c
>   iommu: Drop sw_msi from iommu_domain
> 
> Robin Murphy (1):
>   iommu: Sort out domain user data

Refreshed, thanks

Jason
Re: [PATCH v5 0/3] iommu: Clean up cookie and sw_msi in struct iommu_domain
Posted by Nicolin Chen 8 months, 4 weeks ago
On Mon, Mar 24, 2025 at 09:05:14PM -0700, Nicolin Chen wrote:
> This is a clean-up series for the previous sw_msi Part-1 core series. It's
> on github:
> https://github.com/nicolinc/iommufd/commits/iommufd_msi_cleanup-v5
> 
> Changelog
> v5
>  * Add IMPORT line for IOMMUFD_INTERNAL and change IOMMUFD_DRIVER_CORE
>    to a "bool" module, fixing build module errors reported by Arnd:
>    https://lore.kernel.org/all/20250324210329.2809869-1-arnd@kernel.org/
>  * In iommu_dma_prepare_msi(), let all IDENTITY DOMAINs pass through,
>    fixing the regression with IDENTITY domains reported by Nathan:
>    https://lore.kernel.org/linux-iommu/20250324162558.GA198799@ax162/

> base-commit: da0c56520e880441d0503d0cf0d6853dcfb5f1a4

Hi Jason,

Assuming that you want to replace the commits in your for-next
tree with these patches, I rebased them on this commit:
da0c56520e88 iommu/arm-smmu-v3: Set MEV bit in nested STE for DoS mitigations

Again, sorry for the trouble.

Nicolin

P.S. Attaching git-diff against the current commit in your tree
"git diff -U1 e009e088d88e8..nicolinc/iommufd_msi_cleanup-v5":

diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c
index 0f4cc15ded1c..d96e6fabb4da 100644
--- a/drivers/iommu/iommu.c
+++ b/drivers/iommu/iommu.c
@@ -3653,3 +3653,4 @@ int iommu_dma_prepare_msi(struct msi_desc *desc, phys_addr_t msi_addr)
 	mutex_lock(&group->mutex);
-	if (group->domain) {
+	/* An IDENTITY domain must pass through */
+	if (group->domain && group->domain->type != IOMMU_DOMAIN_IDENTITY) {
 		switch (group->domain->cookie_type) {
diff --git a/drivers/iommu/iommufd/Kconfig b/drivers/iommu/iommufd/Kconfig
index 0a07f9449fd9..2beeb4f60ee5 100644
--- a/drivers/iommu/iommufd/Kconfig
+++ b/drivers/iommu/iommufd/Kconfig
@@ -2,3 +2,3 @@
 config IOMMUFD_DRIVER_CORE
-	tristate
+	bool
 	default (IOMMUFD_DRIVER || IOMMUFD) if IOMMUFD!=n
diff --git a/drivers/iommu/iommufd/driver.c b/drivers/iommu/iommufd/driver.c
index 352513974154..a08ff0f37fc6 100644
--- a/drivers/iommu/iommufd/driver.c
+++ b/drivers/iommu/iommufd/driver.c
@@ -249,2 +249,3 @@ EXPORT_SYMBOL_NS_GPL(iommufd_sw_msi, "IOMMUFD");
 MODULE_DESCRIPTION("iommufd code shared with builtin modules");
+MODULE_IMPORT_NS("IOMMUFD_INTERNAL");
 MODULE_LICENSE("GPL");
Re: [PATCH v5 0/3] iommu: Clean up cookie and sw_msi in struct iommu_domain
Posted by Jason Gunthorpe 8 months, 3 weeks ago
On Mon, Mar 24, 2025 at 09:14:23PM -0700, Nicolin Chen wrote:
> On Mon, Mar 24, 2025 at 09:05:14PM -0700, Nicolin Chen wrote:
> > This is a clean-up series for the previous sw_msi Part-1 core series. It's
> > on github:
> > https://github.com/nicolinc/iommufd/commits/iommufd_msi_cleanup-v5
> > 
> > Changelog
> > v5
> >  * Add IMPORT line for IOMMUFD_INTERNAL and change IOMMUFD_DRIVER_CORE
> >    to a "bool" module, fixing build module errors reported by Arnd:
> >    https://lore.kernel.org/all/20250324210329.2809869-1-arnd@kernel.org/
> >  * In iommu_dma_prepare_msi(), let all IDENTITY DOMAINs pass through,
> >    fixing the regression with IDENTITY domains reported by Nathan:
> >    https://lore.kernel.org/linux-iommu/20250324162558.GA198799@ax162/
> 
> > base-commit: da0c56520e880441d0503d0cf0d6853dcfb5f1a4
> 
> Hi Jason,
> 
> Assuming that you want to replace the commits in your for-next
> tree with these patches, I rebased them on this commit:
> da0c56520e88 iommu/arm-smmu-v3: Set MEV bit in nested STE for DoS mitigations
> 
> Again, sorry for the trouble.

Can you post a v6 please? That is a bit easier

Thanks,
Jason