A pasid might be bound to a page table from a VM guest via the iommu
ops.sva_bind_gpasid. In this case, when a DMA page fault is detected
on the physical IOMMU, we need to inject the page fault request into
the guest. After the guest completes handling the page fault, a page
response need to be sent back via the iommu ops.page_response().
This adds support to report a page request fault. Any external module
which is interested in handling this fault should regiester a notifier
with iommu_register_device_fault_handler().
Co-developed-by: Jacob Pan <jacob.jun.pan@linux.intel.com>
Co-developed-by: Liu Yi L <yi.l.liu@intel.com>
Signed-off-by: Jacob Pan <jacob.jun.pan@linux.intel.com>
Signed-off-by: Liu Yi L <yi.l.liu@intel.com>
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Link: https://lore.kernel.org/r/20200724014925.15523-11-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
For the unlikely use case where multiple aux domains from the same pdev
are attached to a single guest and then bound to a single process
(thus same PASID) within that guest, we cannot easily support this case
by refcounting the number of users. As there is only one SL page table
per PASID while we have multiple aux domains thus multiple SL page tables
for the same PASID.
Extra unbinding guest PASID can happen due to race between normal and
exception cases. Termination of one aux domain may affect others unless
we actively track and switch aux domains to ensure the validity of SL
page tables and TLB states in the shared PASID entry.
Support for sharing second level PGDs across domains can reduce the
complexity but this is not available due to the limitations on VFIO
container architecture. We can revisit this decision once sharing PGDs
are available.
Overall, the complexity and potential glitch do not warrant this unlikely
use case thereby removed by this patch.
Fixes: 56722a4398 ("iommu/vt-d: Add bind guest PASID support")
Signed-off-by: Liu Yi L <yi.l.liu@intel.com>
Signed-off-by: Jacob Pan <jacob.jun.pan@linux.intel.com>
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Cc: Kevin Tian <kevin.tian@intel.com>
Link: https://lore.kernel.org/r/20200724014925.15523-8-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
For guest requested IOTLB invalidation, address and mask are provided as
part of the invalidation data. VT-d HW silently ignores any address bits
below the mask. SW shall also allow such case but give warning if
address does not align with the mask. This patch relax the fault
handling from error to warning and proceed with invalidation request
with the given mask.
Fixes: 6ee1b77ba3 ("iommu/vt-d: Add svm/sva invalidate function")
Signed-off-by: Jacob Pan <jacob.jun.pan@linux.intel.com>
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Link: https://lore.kernel.org/r/20200724014925.15523-7-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
For guest SVA usage, in order to optimize for less VMEXIT, guest request
of IOTLB flush also includes device TLB.
On the host side, IOMMU driver performs IOTLB and implicit devTLB
invalidation. When PASID-selective granularity is requested by the guest
we need to derive the equivalent address range for devTLB instead of
using the address information in the UAPI data. The reason for that is,
unlike IOTLB flush, devTLB flush does not support PASID-selective
granularity. This is to say, we need to set the following in the PASID
based devTLB invalidation descriptor:
- entire 64 bit range in address ~(0x1 << 63)
- S bit = 1 (VT-d CH 6.5.2.6).
Without this fix, device TLB flush range is not set properly for PASID
selective granularity. This patch also merged devTLB flush code for both
implicit and explicit cases.
Fixes: 6ee1b77ba3 ("iommu/vt-d: Add svm/sva invalidate function")
Signed-off-by: Liu Yi L <yi.l.liu@intel.com>
Signed-off-by: Jacob Pan <jacob.jun.pan@linux.intel.com>
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Link: https://lore.kernel.org/r/20200724014925.15523-6-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Address information for device TLB invalidation comes from userspace
when device is directly assigned to a guest with vIOMMU support.
VT-d requires page aligned address. This patch checks and enforce
address to be page aligned, otherwise reserved bits can be set in the
invalidation descriptor. Unrecoverable fault will be reported due to
non-zero value in the reserved bits.
Fixes: 61a06a16e3 ("iommu/vt-d: Support flushing more translation cache types")
Signed-off-by: Liu Yi L <yi.l.liu@intel.com>
Signed-off-by: Jacob Pan <jacob.jun.pan@linux.intel.com>
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Link: https://lore.kernel.org/r/20200724014925.15523-5-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
DevTLB flush can be used for both DMA request with and without PASIDs.
The former uses PASID#0 (RID2PASID), latter uses non-zero PASID for SVA
usage.
This patch adds a check for PASID value such that devTLB flush with
PASID is used for SVA case. This is more efficient in that multiple
PASIDs can be used by a single device, when tearing down a PASID entry
we shall flush only the devTLB specific to a PASID.
Fixes: 6f7db75e1c ("iommu/vt-d: Add second level page table")
Signed-off-by: Jacob Pan <jacob.jun.pan@linux.intel.com>
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Link: https://lore.kernel.org/r/20200724014925.15523-4-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
The Scalable-mode Page-walk Coherency (SMPWC) field in the VT-d extended
capability register indicates the hardware coherency behavior on paging
structures accessed through the pasid table entry. This is ignored in
current code and using ECAP.C instead which is only valid in legacy mode.
Fix this so that paging structure updates could be manually flushed from
the cache line if hardware page walking is not snooped.
Fixes: 765b6a98c1 ("iommu/vt-d: Enumerate the scalable mode capability")
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Cc: Ashok Raj <ashok.raj@intel.com>
Cc: Kevin Tian <kevin.tian@intel.com>
Cc: Jacob Pan <jacob.jun.pan@linux.intel.com>
Link: https://lore.kernel.org/r/20200622231345.29722-6-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
PCI ACS is disabled if Intel IOMMU is off by default or intel_iommu=off
is used in command line. Unfortunately, Intel IOMMU will be forced on if
there're devices sitting on an external facing PCI port that is marked
as untrusted (for example, thunderbolt peripherals). That means, PCI ACS
is disabled while Intel IOMMU is forced on to isolate those devices. As
the result, the devices of an MFD will be grouped by a single group even
the ACS is supported on device.
[ 0.691263] pci 0000:00:07.1: Adding to iommu group 3
[ 0.691277] pci 0000:00:07.2: Adding to iommu group 3
[ 0.691292] pci 0000:00:07.3: Adding to iommu group 3
Fix it by requesting PCI ACS when Intel IOMMU is detected with platform
opt in hint.
Fixes: 89a6079df7 ("iommu/vt-d: Force IOMMU on for platform opt in hint")
Co-developed-by: Lalithambika Krishnakumar <lalithambika.krishnakumar@intel.com>
Signed-off-by: Lalithambika Krishnakumar <lalithambika.krishnakumar@intel.com>
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Reviewed-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Cc: Mika Westerberg <mika.westerberg@linux.intel.com>
Cc: Ashok Raj <ashok.raj@intel.com>
Link: https://lore.kernel.org/r/20200622231345.29722-5-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
When using first-level translation for IOVA, currently the U/S bit in the
page table is cleared which implies DMA requests with user privilege are
blocked. As the result, following error messages might be observed when
passing through a device to user level:
DMAR: DRHD: handling fault status reg 3
DMAR: [DMA Read] Request device [41:00.0] PASID 1 fault addr 7ecdcd000
[fault reason 129] SM: U/S set 0 for first-level translation
with user privilege
This fixes it by setting U/S bit in the first level page table and makes
IOVA over first level compatible with previous second-level translation.
Fixes: b802d070a5 ("iommu/vt-d: Use iova over first level")
Reported-by: Xin Zeng <xin.zeng@intel.com>
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Link: https://lore.kernel.org/r/20200622231345.29722-3-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Current Intel SVM is designed by setting the pgd_t of the processor page
table to FLPTR field of the PASID entry. The first level translation only
supports 4 and 5 level paging structures, hence it's infeasible for the
IOMMU to share a processor's page table when it's running in 32-bit mode.
Let's disable 32bit support for now and claim support only when all the
missing pieces are ready in the future.
Fixes: 1c4f88b7f1 ("iommu/vt-d: Shared virtual address in scalable mode")
Suggested-by: Joerg Roedel <jroedel@suse.de>
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Link: https://lore.kernel.org/r/20200622231345.29722-2-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Pull more Kbuild updates from Masahiro Yamada:
- fix build rules in binderfs sample
- fix build errors when Kbuild recurses to the top Makefile
- covert '---help---' in Kconfig to 'help'
* tag 'kbuild-v5.8-2' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild:
treewide: replace '---help---' in Kconfig files with 'help'
kbuild: fix broken builds because of GZIP,BZIP2,LZOP variables
samples: binderfs: really compile this sample and fix build issues
Since commit 84af7a6194 ("checkpatch: kconfig: prefer 'help' over
'---help---'"), the number of '---help---' has been gradually
decreasing, but there are still more than 2400 instances.
This commit finishes the conversion. While I touched the lines,
I also fixed the indentation.
There are a variety of indentation styles found.
a) 4 spaces + '---help---'
b) 7 spaces + '---help---'
c) 8 spaces + '---help---'
d) 1 space + 1 tab + '---help---'
e) 1 tab + '---help---' (correct indentation)
f) 1 tab + 1 space + '---help---'
g) 1 tab + 2 spaces + '---help---'
In order to convert all of them to 1 tab + 'help', I ran the
following commend:
$ find . -name 'Kconfig*' | xargs sed -i 's/^[[:space:]]*---help---/\thelp/'
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Pull iommu driver directory structure cleanup from Joerg Roedel:
"Move the Intel and AMD IOMMU drivers into their own subdirectory.
Both drivers consist of several files by now and giving them their own
directory unclutters the IOMMU top-level directory a bit"
* tag 'iommu-drivers-move-v5.8' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu:
iommu/vt-d: Move Intel IOMMU driver into subdirectory
iommu/amd: Move AMD IOMMU driver into subdirectory
Pull s390 updates from Vasily Gorbik:
- Add support for multi-function devices in pci code.
- Enable PF-VF linking for architectures using the pdev->no_vf_scan
flag (currently just s390).
- Add reipl from NVMe support.
- Get rid of critical section cleanup in entry.S.
- Refactor PNSO CHSC (perform network subchannel operation) in cio and
qeth.
- QDIO interrupts and error handling fixes and improvements, more
refactoring changes.
- Align ioremap() with generic code.
- Accept requests without the prefetch bit set in vfio-ccw.
- Enable path handling via two new regions in vfio-ccw.
- Other small fixes and improvements all over the code.
* tag 's390-5.8-1' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux: (52 commits)
vfio-ccw: make vfio_ccw_regops variables declarations static
vfio-ccw: Add trace for CRW event
vfio-ccw: Wire up the CRW irq and CRW region
vfio-ccw: Introduce a new CRW region
vfio-ccw: Refactor IRQ handlers
vfio-ccw: Introduce a new schib region
vfio-ccw: Refactor the unregister of the async regions
vfio-ccw: Register a chp_event callback for vfio-ccw
vfio-ccw: Introduce new helper functions to free/destroy regions
vfio-ccw: document possible errors
vfio-ccw: Enable transparent CCW IPL from DASD
s390/pci: Log new handle in clp_disable_fh()
s390/cio, s390/qeth: cleanup PNSO CHSC
s390/qdio: remove q->first_to_kick
s390/qdio: fix up qdio_start_irq() kerneldoc
s390: remove critical section cleanup from entry.S
s390: add machine check SIGP
s390/pci: ioremap() align with generic code
s390/ap: introduce new ap function ap_get_qdev()
Documentation/s390: Update / remove developerWorks web links
...
Pull iommu updates from Joerg Roedel:
"A big part of this is a change in how devices get connected to IOMMUs
in the core code. It contains the change from the old add_device() /
remove_device() to the new probe_device() / release_device()
call-backs.
As a result functionality that was previously in the IOMMU drivers has
been moved to the IOMMU core code, including IOMMU group allocation
for each device. The reason for this change was to get more robust
allocation of default domains for the iommu groups.
A couple of fixes were necessary after this was merged into the IOMMU
tree, but there are no known bugs left. The last fix is applied on-top
of the merge commit for the topic branches.
Other than that change, we have:
- Removal of the driver private domain handling in the Intel VT-d
driver. This was fragile code and I am glad it is gone now.
- More Intel VT-d updates from Lu Baolu:
- Nested Shared Virtual Addressing (SVA) support to the Intel VT-d
driver
- Replacement of the Intel SVM interfaces to the common IOMMU SVA
API
- SVA Page Request draining support
- ARM-SMMU Updates from Will:
- Avoid mapping reserved MMIO space on SMMUv3, so that it can be
claimed by the PMU driver
- Use xarray to manage ASIDs on SMMUv3
- Reword confusing shutdown message
- DT compatible string updates
- Allow implementations to override the default domain type
- A new IOMMU driver for the Allwinner Sun50i platform
- Support for ATS gets disabled for untrusted devices (like
Thunderbolt devices). This includes a PCI patch, acked by Bjorn.
- Some cleanups to the AMD IOMMU driver to make more use of IOMMU
core features.
- Unification of some printk formats in the Intel and AMD IOMMU
drivers and in the IOVA code.
- Updates for DT bindings
- A number of smaller fixes and cleanups.
* tag 'iommu-updates-v5.8' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu: (109 commits)
iommu: Check for deferred attach in iommu_group_do_dma_attach()
iommu/amd: Remove redundant devid checks
iommu/amd: Store dev_data as device iommu private data
iommu/amd: Merge private header files
iommu/amd: Remove PD_DMA_OPS_MASK
iommu/amd: Consolidate domain allocation/freeing
iommu/amd: Free page-table in protection_domain_free()
iommu/amd: Allocate page-table in protection_domain_init()
iommu/amd: Let free_pagetable() not rely on domain->pt_root
iommu/amd: Unexport get_dev_data()
iommu/vt-d: Fix compile warning
iommu/vt-d: Remove real DMA lookup in find_domain
iommu/vt-d: Allocate domain info for real DMA sub-devices
iommu/vt-d: Only clear real DMA device's context entries
iommu: Remove iommu_sva_ops::mm_exit()
uacce: Remove mm_exit() op
iommu/sun50i: Constify sun50i_iommu_ops
iommu/hyper-v: Constify hyperv_ir_domain_ops
iommu/vt-d: Use pci_ats_supported()
iommu/arm-smmu-v3: Use pci_ats_supported()
...
Pull dma-mapping updates from Christoph Hellwig:
- enhance the dma pool to allow atomic allocation on x86 with AMD SEV
(David Rientjes)
- two small cleanups (Jason Yan and Peter Collingbourne)
* tag 'dma-mapping-5.8' of git://git.infradead.org/users/hch/dma-mapping:
dma-contiguous: fix comment for dma_release_from_contiguous
dma-pool: scale the default DMA coherent pool size with memory capacity
x86/mm: unencrypted non-blocking DMA allocations use coherent pools
dma-pool: add pool sizes to debugfs
dma-direct: atomic allocations must come from atomic coherent pools
dma-pool: dynamically expanding atomic pools
dma-pool: add additional coherent pools to map to gfp mask
dma-remap: separate DMA atomic pools from direct remap code
dma-debug: make __dma_entry_alloc_check_leak() static
Use 'struct domain_pgtable' instead to free_pagetable(). This solves
the problem that amd_iommu_domain_direct_map() needs to restore
domain->pt_root after the device table has been updated just to make
free_pagetable release the domain page-table.
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Reviewed-by: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
Link: https://lore.kernel.org/r/20200527115313.7426-4-joro@8bytes.org
This function is internal to the AMD IOMMU driver and only exported
because the amd_iommu_v2 modules calls it. But the reason it is called
from there could better be handled by amd_iommu_is_attach_deferred().
So unexport get_dev_data() and use amd_iommu_is_attach_deferred()
instead.
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Reviewed-by: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
Link: https://lore.kernel.org/r/20200527115313.7426-3-joro@8bytes.org
Sub-devices of a real DMA device might exist on a separate segment than
the real DMA device and its IOMMU. These devices should still have a
valid device_domain_info, but the current dma alias model won't
allocate info for the subdevice.
This patch adds a segment member to struct device_domain_info and uses
the sub-device's BDF so that these sub-devices won't alias to other
devices.
Fixes: 2b0140c696 ("iommu/vt-d: Use pci_real_dma_dev() for mapping")
Cc: stable@vger.kernel.org # v5.6+
Signed-off-by: Jon Derrick <jonathan.derrick@intel.com>
Acked-by: Lu Baolu <baolu.lu@linux.intel.com>
Link: https://lore.kernel.org/r/20200527165617.297470-3-jonathan.derrick@intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Domain context mapping can encounter issues with sub-devices of a real
DMA device. A sub-device cannot have a valid context entry due to it
potentially aliasing another device's 16-bit ID. It's expected that
sub-devices of the real DMA device uses the real DMA device's requester
when context mapping.
This is an issue when a sub-device is removed where the context entry is
cleared for all aliases. Other sub-devices are still valid, resulting in
those sub-devices being stranded without valid context entries.
The correct approach is to use the real DMA device when programming the
context entries. The insertion path is correct because device_to_iommu()
will return the bus and devfn of the real DMA device. The removal path
needs to only operate on the real DMA device, otherwise the entire
context entry would be cleared for all sub-devices of the real DMA
device.
This patch also adds a helper to determine if a struct device is a
sub-device of a real DMA device.
Fixes: 2b0140c696 ("iommu/vt-d: Use pci_real_dma_dev() for mapping")
Cc: stable@vger.kernel.org # v5.6+
Signed-off-by: Jon Derrick <jonathan.derrick@intel.com>
Acked-by: Lu Baolu <baolu.lu@linux.intel.com>
Link: https://lore.kernel.org/r/20200527165617.297470-2-jonathan.derrick@intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
After binding a device to an mm, device drivers currently need to
register a mm_exit handler. This function is called when the mm exits,
to gracefully stop DMA targeting the address space and flush page faults
to the IOMMU.
This is deemed too complex for the MMU release() notifier, which may be
triggered by any mmput() invocation, from about 120 callsites [1]. The
upcoming SVA module has an example of such complexity: the I/O Page
Fault handler would need to call mmput_async() instead of mmput() after
handling an IOPF, to avoid triggering the release() notifier which would
in turn drain the IOPF queue and lock up.
Another concern is the DMA stop function taking too long, up to several
minutes [2]. For some mmput() callers this may disturb other users. For
example, if the OOM killer picks the mm bound to a device as the victim
and that mm's memory is locked, if the release() takes too long, it
might choose additional innocent victims to kill.
To simplify the MMU release notifier, don't forward the notification to
device drivers. Since they don't stop DMA on mm exit anymore, the PASID
lifetime is extended:
(1) The device driver calls bind(). A PASID is allocated.
Here any DMA fault is handled by mm, and on error we don't print
anything to dmesg. Userspace can easily trigger errors by issuing DMA
on unmapped buffers.
(2) exit_mmap(), for example the process took a SIGKILL. This step
doesn't happen during normal operations. Remove the pgd from the
PASID table, since the page tables are about to be freed. Invalidate
the IOTLBs.
Here the device may still perform DMA on the address space. Incoming
transactions are aborted but faults aren't printed out. ATS
Translation Requests return Successful Translation Completions with
R=W=0. PRI Page Requests return with Invalid Request.
(3) The device driver stops DMA, possibly following release of a fd, and
calls unbind(). PASID table is cleared, IOTLB invalidated if
necessary. The page fault queues are drained, and the PASID is
freed.
If DMA for that PASID is still running here, something went seriously
wrong and errors should be reported.
For now remove iommu_sva_ops entirely. We might need to re-introduce
them at some point, for example to notify device drivers of unhandled
IOPF.
[1] https://lore.kernel.org/linux-iommu/20200306174239.GM31668@ziepe.ca/
[2] https://lore.kernel.org/linux-iommu/4d68da96-0ad5-b412-5987-2f7a6aa796c3@amd.com/
Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
Acked-by: Jacob Pan <jacob.jun.pan@linux.intel.com>
Acked-by: Lu Baolu <baolu.lu@linux.intel.com>
Link: https://lore.kernel.org/r/20200423125329.782066-3-jean-philippe@linaro.org
Signed-off-by: Joerg Roedel <jroedel@suse.de>
The struct sun50i_iommu_ops is not modified and can be made const to
allow the compiler to put it in read-only memory.
Before:
text data bss dec hex filename
14358 2501 64 16923 421b drivers/iommu/sun50i-iommu.o
After:
text data bss dec hex filename
14726 2117 64 16907 420b drivers/iommu/sun50i-iommu.o
Signed-off-by: Rikard Falkeborn <rikard.falkeborn@gmail.com>
Acked-by: Maxime Ripard <mripard@kernel.org>
Link: https://lore.kernel.org/r/20200525214958.30015-3-rikard.falkeborn@gmail.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
The pci_ats_supported() function checks if a device supports ATS and is
allowed to use it. In addition to checking that the device has an ATS
capability and that the global pci=noats is not set
(pci_ats_disabled()), it also checks if a device is untrusted.
A device is untrusted if it is plugged into an external-facing port such
as Thunderbolt and could be spoofing an existing device to exploit
weaknesses in the IOMMU configuration. By calling pci_ats_supported() we
keep DTE[I]=0 for untrusted devices and abort transactions with
Pretranslated Addresses.
Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
Reviewed-by: Joerg Roedel <jroedel@suse.de>
Link: https://lore.kernel.org/r/20200520152201.3309416-3-jean-philippe@linaro.org
Signed-off-by: Joerg Roedel <jroedel@suse.de>
