The 64KB Translation Granule Supplement to the SMMUv1 architecture
allows an SMMUv1 implementation to support 64KB pages for stage 2
translations, using a constrained VMSAv8 descriptor format limited
to 40-bit addresses. Now that we can freely mix and match context
formats, we can actually handle having 4KB pages via an AArch32
context but 64KB pages via an AArch64 context, so plumb it in.
It is assumed that any implementations will have hardware capabilities
matching the format constraints, thus obviating the need for excessive
sanity-checking; this is the case for MMU-401, the only ARM Ltd.
implementation.
CC: Eric Auger <eric.auger@linaro.org>
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
The way the driver currently forces an AArch32 or AArch64 context format
based on the kernel config and SMMU architecture version is suboptimal,
in that it makes it very hard to support oddball mix-and-match cases
like the SMMUv1 64KB supplement, or situations where the reduced table
depth of an AArch32 short descriptor context may be desirable under an
AArch64 kernel. It also only happens to work on current implementations
which do support all the relevant formats.
Introduce an explicit notion of context format, so we can manage that
independently and get rid of the inflexible #ifdeffery.
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
With {read,write}q_relaxed now able to fall back to the common
nonatomic-hi-lo helper, make use of that so that we don't have to
open-code our own. In the process, also convert the other remaining
split accesses, and repurpose the custom accessor to smooth out the
couple of troublesome instances where we really want to avoid
nonatomic writes (and a 64-bit access is unnecessary in the 32-bit
context formats we would use on a 32-bit CPU).
This paves the way for getting rid of some of the assumptions currently
baked into the driver which make it really awkward to use 32-bit context
formats with SMMUv2 under a 64-bit kernel.
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
MMU-500 erratum #841119 is tickled by a particular set of circumstances
interacting with the next-page prefetcher. Since said prefetcher is
quite dumb and actually detrimental to performance in some cases (by
causing unwanted TLB evictions for non-sequential access patterns), we
lose very little by turning it off, and what we gain is a guarantee that
the erratum is never hit.
As a bonus, the same workaround will also prevent erratum #826419 once
v7 short descriptor support is implemented.
CC: Catalin Marinas <catalin.marinas@arm.com>
CC: Will Deacon <will.deacon@arm.com>
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
With a framework for implementation-specific funtionality in place, the
currently-FDT-dependent ThunderX workaround gets to be the first user.
Acked-by: Tirumalesh Chalamarla <tchalamarla@caviumnetworks.com>
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
As the inevitable reality of implementation-specific errata workarounds
begin to accrue alongside our integration quirk handling, it's about
time the driver had a decent way of keeping track. Extend the per-SMMU
data so we can identify specific implementations in an efficient and
firmware-agnostic manner.
Acked-by: Tirumalesh Chalamarla <tchalamarla@caviumnetworks.com>
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Due to erratum #27704, the CN88xx SMMUv2 implementation supports only
shared ASID and VMID numberspaces.
This patch ensures that ASID and VMIDs are unique across all SMMU
instances on affected Cavium systems.
Signed-off-by: Tirumalesh Chalamarla <tchalamarla@caviumnetworks.com>
Signed-off-by: Akula Geethasowjanya <Geethasowjanya.Akula@caviumnetworks.com>
[will: commit message, comments and formatting]
Signed-off-by: Will Deacon <will.deacon@arm.com>
This patch adds support for 16-bit VMIDs on implementations of SMMUv2
that support it.
Signed-off-by: Tirumalesh Chalamarla <tchalamarla@caviumnetworks.com>
[will: commit messsage and comments]
Signed-off-by: Will Deacon <will.deacon@arm.com>
Until all upstream devices have their DMA ops swizzled to point at the
SMMU, we need to treat the IOMMU_DOMAIN_DMA domain as bypass to avoid
putting devices into an empty address space when detaching from VFIO.
Signed-off-by: Will Deacon <will.deacon@arm.com>
The ARM SMMU attach_dev implementations returns -EEXIST if the device
being attached is already attached to a domain. This doesn't play nicely
with the default domain, resulting in splats such as:
WARNING: at drivers/iommu/iommu.c:1257
Modules linked in:
CPU: 3 PID: 1939 Comm: virtio-net-tx Tainted: G S 4.5.0-rc4+ #1
Hardware name: FVP Base (DT)
task: ffffffc87a9d0000 ti: ffffffc07a278000 task.ti: ffffffc07a278000
PC is at __iommu_detach_group+0x68/0xe8
LR is at __iommu_detach_group+0x48/0xe8
This patch fixes the problem by forcefully detaching the device from
its old domain, if present, when attaching to a new one. The unused
->detach_dev callback is also removed the iommu_ops structures.
Signed-off-by: Will Deacon <will.deacon@arm.com>
Borrow the disable_bypass parameter from the SMMUv3 driver as a handy
debugging/security feature so that unmatched stream IDs (i.e. devices
not attached to an IOMMU domain) may be configured to fault.
Rather than introduce unsightly inconsistency, or repeat the existing
unnecessary use of module_param_named(), fix that as well in passing.
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
With DMA mapping ops provided by the iommu-dma code, only a minimal
contribution from the IOMMU driver is needed to create a suitable
DMA-API domain for them to use. Implement this for the ARM SMMUs.
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
The IOMMU API has no concept of privilege so assumes all devices and
mappings are equal, and indeed most non-CPU master devices on an AMBA
interconnect make little use of the attribute bits on the bus thus by
default perform unprivileged data accesses.
Some devices, however, believe themselves more equal than others, such
as programmable DMA controllers whose 'master' thread issues bus
transactions marked as privileged instruction fetches, while the data
accesses of its channel threads (under the control of Linux, at least)
are marked as unprivileged. This poses a problem for implementing the
DMA API on an IOMMU conforming to ARM VMSAv8, under which a page that is
unprivileged-writeable is also implicitly privileged-execute-never.
Given that, there is no one set of attributes with which iommu_map() can
implement, say, dma_alloc_coherent() that will allow every possible type
of access without something running into unexecepted permission faults.
Fortunately the SMMU architecture provides a means to mitigate such
issues by overriding the incoming attributes of a transaction; make use
of that to strip the privileged/unprivileged status off incoming
transactions, leaving just the instruction/data dichotomy which the
IOMMU API does at least understand; Four states good, two states better.
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
When invalidating an IOVA range potentially spanning multiple pages,
such as when removing an entire intermediate-level table, we currently
only issue an invalidation for the first IOVA of that range. Since the
architecture specifies that address-based TLB maintenance operations
target a single entry, an SMMU could feasibly retain live entries for
subsequent pages within that unmapped range, which is not good.
Make sure we hit every possible entry by iterating over the whole range
at the granularity provided by the pagetable implementation.
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
[will: added missing semicolons...]
Signed-off-by: Will Deacon <will.deacon@arm.com>
IOMMU hardware with range-based TLB maintenance commands can work
happily with the iova and size arguments passed via the tlb_add_flush
callback, but for IOMMUs which require separate commands per entry in
the range, it is not straightforward to infer the necessary granularity
when it comes to issuing the actual commands.
Add an additional argument indicating the granularity for the benefit
of drivers needing to know, and update the ARM LPAE code appropriately
(for non-leaf invalidations we currently just assume the worst-case
page granularity rather than walking the table to check).
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
The basic flow for add a device:
arm_smmu_add_device
|->iommu_group_get_for_dev
|->iommu_group_get
return group; (1)
|->ops->device_group : Init/increase reference count to/by 1.
|->iommu_group_add_device : Increase reference count by 1.
return group (2)
|->return 0;
Since we are adding one device, the flow is (2) and the group reference
count will be increased by 2. So, we need to add iommu_group_put at the
end of arm_smmu_add_device to decrease the count by 1.
Also take the failure path into consideration when fail to add a device.
Signed-off-by: Peng Fan <van.freenix@gmail.com>
Cc: Will Deacon <will.deacon@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
The free_io_pgtable_ops() function tests whether its argument is NULL
and then returns immediately. Thus the test around the call is not needed.
This issue was detected by using the Coccinelle software.
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Signed-off-by: Will Deacon <will.deacon@arm.com>
This converts the ARM SMMU and the SMMUv3 driver to use the
new device_group call-back.
Cc: Will Deacon <will.deacon@arm.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Since commit 1463fe44fd ("iommu/arm-smmu: Don't use VMIDs for stage-1
translations"), we don't need the GR0 base address when initialising a
context bank, so remove the useless local variable and its init code.
Signed-off-by: Will Deacon <will.deacon@arm.com>
The SMMU architecture defines two different behaviors when 64-bit
registers are written with 32-bit writes. The first behavior causes
zero extension into the upper 32-bits. The second behavior splits a
64-bit register into "normal" 32-bit register pairs.
On some buggy implementations, registers incorrectly zero extended
when they should instead behave as normal 32-bit register pairs.
Signed-off-by: Tirumalesh Chalamarla <tchalamarla@caviumnetworks.com>
[will: removed redundant macro parameters]
Signed-off-by: Will Deacon <will.deacon@arm.com>
'%pad' automatically prints with '0x', so remove the explicit '0x'
annotation.
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
With the io-pgtable code now enforcing its own appropriate sync points,
the vestigial flush_pgtable callback becomes entirely redundant, so
remove it altogether.
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
With the correct DMA API calls now integrated into the io-pgtable code,
let that handle the flushing of non-coherent page table updates.
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Currently, we detect whether the SMMU has coherent page table walk
capability from the IDR0.CTTW field, and base our cache maintenance
decisions on that. In preparation for fixing the bogus DMA API usage,
however, we need to ensure that the DMA API agrees about this, which
necessitates deferring to the dma-coherent property in the device tree
for the final say.
As an added bonus, since systems exist where an external CTTW signal
has been tied off incorrectly at integration, allowing DT to override
it offers a neat workaround for coherency issues with such SMMUs.
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Commit 83a60ed8f0 ("iommu/arm-smmu: fix ARM_SMMU_FEAT_TRANS_OPS
condition") accidentally negated the ID0_ATOSNS predicate in the ATOS
feature check, causing the driver to attempt ATOS requests on SMMUv2
hardware without the ATOS feature implemented.
This patch restores the predicate to the correct value.
Cc: <stable@vger.kernel.org> # 4.0+
Reported-by: Varun Sethi <varun.sethi@freescale.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Changing force_stage dynamically isn't supported by the driver and it
also doesn't make a whole lot of sense to change it once the SMMU is up
and running.
This patch makes the sysfs entry for the parameter read-only.
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
The ATS1* address translation registers only support being written
atomically - in SMMUv2 where they are 64 bits wide, 32-bit writes to
the lower half are automatically zero-extended, whilst 32-bit writes
to the upper half are ignored. Thus, the current logic of performing
64-bit writes as two 32-bit accesses is wrong.
Since we already limit IOVAs to 32 bits on 32-bit ARM, the lack of a
suitable writeq() implementation there is not an issue, and we only
need a little preprocessor ugliness to safely hide the 64-bit case.
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Stage 1 translation is controlled by two sets of page tables (TTBR0 and
TTBR1) which grow up and down from zero respectively in the ARMv8
translation regime. For the SMMU, we only care about TTBR0 and, in the
case of a 48-bit virtual space, we expect to map virtual addresses 0x0
through to 0xffff_ffff_ffff.
Given that some masters may be incapable of emitting virtual addresses
targetting TTBR1 (e.g. because they sit on a 48-bit bus), the SMMU
architecture allows bit 47 to be sign-extended, halving the virtual
range of TTBR0 but allowing TTBR1 to be used. This is controlled by the
SEP field in TTBCR2.
The SMMU driver incorrectly enables this sign-extension feature, which
causes problems when userspace addresses are programmed into a master
device with the SMMU expecting to map the incoming transactions via
TTBR0; if the top bit of address is set, we will instead get a
translation fault since TTBR1 walks are disabled in the TTBCR.
This patch fixes the issue by disabling sign-extension of a fixed
virtual address bit and instead basing the behaviour on the upstream bus
size: the incoming address is zero extended unless the upstream bus is
only 49 bits wide, in which case bit 48 is used as the sign bit and is
replicated to the upper bits.
Cc: <stable@vger.kernel.org> # v4.0+
Reported-by: Varun Sethi <varun.sethi@freescale.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Implement domain_alloc and domain_free iommu-ops as a
replacement for domain_init/domain_destroy.
Acked-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
IOMMU groups for PCI devices can correspond to multiple DMA aliases due
to things like ACS and PCI quirks.
This patch extends the ARM SMMU ->add_device callback so that we
consider all of the DMA aliases for a PCI IOMMU group, rather than
creating a separate group for each Requester ID.
Signed-off-by: Will Deacon <will.deacon@arm.com>
Since we use dma_map_page() as an architecture-independent means of
making page table updates visible to non-coherent SMMUs, we need to
have a suitable DMA mask set to discourage the DMA mapping layer from
creating bounce buffers and flushing those instead, if said page tables
happen to lie outside the default 32-bit mask.
Tested-by: Mitchel Humpherys <mitchelh@codeaurora.org>
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
[will: added error checking]
Signed-off-by: Will Deacon <will.deacon@arm.com>
The VMID16 (8.1) extension to SMMUv2 added a 16-bit VMID16 field to the
CBA2R registers. Unfortunately, if software writes this field as zero
after setting an 8-bit VMID in a stage-2 CBAR, then the VMID may also be
overwritten with zero on some early implementations (the architecture
was later updated to fix this issue).
This patch ensures that we initialise CBA2R before CBAR, therefore
ensuring that the VMID is set correctly.
Tested-by: Manish Jaggi <mjaggi@caviumnetworks.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
This patch is a fix to "iommu/arm-smmu: add support for iova_to_phys
through ATS1PR".
According to ARM documentation, translation registers are optional even
in SMMUv1, so ID0_S1TS needs to be checked to verify their presence.
Also, we check that the domain is a stage-1 domain.
Signed-off-by: Baptiste Reynal <b.reynal@virtualopensystems.com>
Acked-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
ARM allmodconfig gained a new warning when dma_addr_t is 32-bit wide:
drivers/iommu/arm-smmu.c: In function 'arm_smmu_iova_to_phys_hard':
drivers/iommu/arm-smmu.c:1255:3: warning: right shift count >= width of type
This changes the calculation so that the effective type is always
64-bit.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Fixes: 859a732e4f ("iommu/arm-smmu: add support for iova_to_phys through ATS1PR")
Acked-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Currently, we provide the iommu_ops.iova_to_phys service by doing a
table walk in software to translate IO virtual addresses to physical
addresses. On SMMUs that support it, it can be useful to ask the SMMU
itself to do the translation. This can be used to warm the TLBs for an
SMMU. It can also be useful for testing and hardware validation.
Since the address translation registers are optional on SMMUv2, only
enable hardware translations when using SMMUv1 or when SMMU_IDR0.S1TS=1
and SMMU_IDR0.ATOSNS=0, as described in the ARM SMMU v1-v2 spec.
Signed-off-by: Mitchel Humpherys <mitchelh@codeaurora.org>
[will: reworked on top of generic iopgtbl changes]
Signed-off-by: Will Deacon <will.deacon@arm.com>
Currently we do a STLBIALL when we initialize the SMMU. However, in
some configurations that register is not supposed to be touched and is
marked as "Secure only" in the spec. Rip it out.
Signed-off-by: Mitchel Humpherys <mitchelh@codeaurora.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Here's the set of driver core patches for 3.19-rc1.
They are dominated by the removal of the .owner field in platform
drivers. They touch a lot of files, but they are "simple" changes, just
removing a line in a structure.
Other than that, a few minor driver core and debugfs changes. There are
some ath9k patches coming in through this tree that have been acked by
the wireless maintainers as they relied on the debugfs changes.
Everything has been in linux-next for a while.
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
iEYEABECAAYFAlSOD20ACgkQMUfUDdst+ylLPACg2QrW1oHhdTMT9WI8jihlHVRM
53kAoLeteByQ3iVwWurwwseRPiWa8+MI
=OVRS
-----END PGP SIGNATURE-----
Merge tag 'driver-core-3.19-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core
Pull driver core update from Greg KH:
"Here's the set of driver core patches for 3.19-rc1.
They are dominated by the removal of the .owner field in platform
drivers. They touch a lot of files, but they are "simple" changes,
just removing a line in a structure.
Other than that, a few minor driver core and debugfs changes. There
are some ath9k patches coming in through this tree that have been
acked by the wireless maintainers as they relied on the debugfs
changes.
Everything has been in linux-next for a while"
* tag 'driver-core-3.19-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core: (324 commits)
Revert "ath: ath9k: use debugfs_create_devm_seqfile() helper for seq_file entries"
fs: debugfs: add forward declaration for struct device type
firmware class: Deletion of an unnecessary check before the function call "vunmap"
firmware loader: fix hung task warning dump
devcoredump: provide a one-way disable function
device: Add dev_<level>_once variants
ath: ath9k: use debugfs_create_devm_seqfile() helper for seq_file entries
ath: use seq_file api for ath9k debugfs files
debugfs: add helper function to create device related seq_file
drivers/base: cacheinfo: remove noisy error boot message
Revert "core: platform: add warning if driver has no owner"
drivers: base: support cpu cache information interface to userspace via sysfs
drivers: base: add cpu_device_create to support per-cpu devices
topology: replace custom attribute macros with standard DEVICE_ATTR*
cpumask: factor out show_cpumap into separate helper function
driver core: Fix unbalanced device reference in drivers_probe
driver core: fix race with userland in device_add()
sysfs/kernfs: make read requests on pre-alloc files use the buffer.
sysfs/kernfs: allow attributes to request write buffer be pre-allocated.
fs: sysfs: return EGBIG on write if offset is larger than file size
...
When domains are set with the DOMAIN_ATTR_NESTING flag, we must ensure
that we allocate them to stage-2 context banks if the hardware permits
it.
This patch adds support for the attribute to the ARM SMMU driver, with
the actual stage being determined depending on the features supported
by the hardware.
Signed-off-by: Will Deacon <will.deacon@arm.com>
Currently the driver registers IOMMU bus operations for all busses even
if no ARM SMMU is present on a system. Depending on the driver probing
order this prevents the driver for the real IOMMU to register itself as
the bus-wide IOMMU.
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
The ARM SMMU supports the IOMMU_NOEXEC protection flag. Add the
corresponding IOMMU capability.
Signed-off-by: Antonios Motakis <a.motakis@virtualopensystems.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Exposing the XN flag of the SMMU driver as IOMMU_NOEXEC instead of
IOMMU_EXEC makes it enforceable, since for IOMMUs that don't support
the XN flag pages will always be executable.
Signed-off-by: Antonios Motakis <a.motakis@virtualopensystems.com>
Acked-by: Joerg Roedel <jroedel@suse.de>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Mapping and unmapping are more often than not in the critical path.
map_sg allows IOMMU driver implementations to optimize the process
of mapping buffers into the IOMMU page tables.
Instead of mapping a buffer one page at a time and requiring potentially
expensive TLB operations for each page, this function allows the driver
to map all pages in one go and defer TLB maintenance until after all
pages have been mapped.
Additionally, the mapping operation would be faster in general since
clients does not have to keep calling map API over and over again for
each physically contiguous chunk of memory that needs to be mapped to a
virtually contiguous region.
Signed-off-by: Olav Haugan <ohaugan@codeaurora.org>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Make of_device_id array const, because all OF functions handle it as const.
Signed-off-by: Kiran Padwal <kiran.padwal@smartplayin.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Robin Murphy