| Age | Commit message (Collapse) | Author |
|---|
|
git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull irq updates from Thomas Gleixner:
"Updates for the interrupt core and driver subsystem:
The bulk is the rework of the MSI subsystem to support per device MSI
interrupt domains. This solves conceptual problems of the current
PCI/MSI design which are in the way of providing support for
PCI/MSI[-X] and the upcoming PCI/IMS mechanism on the same device.
IMS (Interrupt Message Store] is a new specification which allows
device manufactures to provide implementation defined storage for MSI
messages (as opposed to PCI/MSI and PCI/MSI-X that has a specified
message store which is uniform accross all devices). The PCI/MSI[-X]
uniformity allowed us to get away with "global" PCI/MSI domains.
IMS not only allows to overcome the size limitations of the MSI-X
table, but also gives the device manufacturer the freedom to store the
message in arbitrary places, even in host memory which is shared with
the device.
There have been several attempts to glue this into the current MSI
code, but after lengthy discussions it turned out that there is a
fundamental design problem in the current PCI/MSI-X implementation.
This needs some historical background.
When PCI/MSI[-X] support was added around 2003, interrupt management
was completely different from what we have today in the actively
developed architectures. Interrupt management was completely
architecture specific and while there were attempts to create common
infrastructure the commonalities were rudimentary and just providing
shared data structures and interfaces so that drivers could be written
in an architecture agnostic way.
The initial PCI/MSI[-X] support obviously plugged into this model
which resulted in some basic shared infrastructure in the PCI core
code for setting up MSI descriptors, which are a pure software
construct for holding data relevant for a particular MSI interrupt,
but the actual association to Linux interrupts was completely
architecture specific. This model is still supported today to keep
museum architectures and notorious stragglers alive.
In 2013 Intel tried to add support for hot-pluggable IO/APICs to the
kernel, which was creating yet another architecture specific mechanism
and resulted in an unholy mess on top of the existing horrors of x86
interrupt handling. The x86 interrupt management code was already an
incomprehensible maze of indirections between the CPU vector
management, interrupt remapping and the actual IO/APIC and PCI/MSI[-X]
implementation.
At roughly the same time ARM struggled with the ever growing SoC
specific extensions which were glued on top of the architected GIC
interrupt controller.
This resulted in a fundamental redesign of interrupt management and
provided the today prevailing concept of hierarchical interrupt
domains. This allowed to disentangle the interactions between x86
vector domain and interrupt remapping and also allowed ARM to handle
the zoo of SoC specific interrupt components in a sane way.
The concept of hierarchical interrupt domains aims to encapsulate the
functionality of particular IP blocks which are involved in interrupt
delivery so that they become extensible and pluggable. The X86
encapsulation looks like this:
|--- device 1
[Vector]---[Remapping]---[PCI/MSI]--|...
|--- device N
where the remapping domain is an optional component and in case that
it is not available the PCI/MSI[-X] domains have the vector domain as
their parent. This reduced the required interaction between the
domains pretty much to the initialization phase where it is obviously
required to establish the proper parent relation ship in the
components of the hierarchy.
While in most cases the model is strictly representing the chain of IP
blocks and abstracting them so they can be plugged together to form a
hierarchy, the design stopped short on PCI/MSI[-X]. Looking at the
hardware it's clear that the actual PCI/MSI[-X] interrupt controller
is not a global entity, but strict a per PCI device entity.
Here we took a short cut on the hierarchical model and went for the
easy solution of providing "global" PCI/MSI domains which was possible
because the PCI/MSI[-X] handling is uniform across the devices. This
also allowed to keep the existing PCI/MSI[-X] infrastructure mostly
unchanged which in turn made it simple to keep the existing
architecture specific management alive.
A similar problem was created in the ARM world with support for IP
block specific message storage. Instead of going all the way to stack
a IP block specific domain on top of the generic MSI domain this ended
in a construct which provides a "global" platform MSI domain which
allows overriding the irq_write_msi_msg() callback per allocation.
In course of the lengthy discussions we identified other abuse of the
MSI infrastructure in wireless drivers, NTB etc. where support for
implementation specific message storage was just mindlessly glued into
the existing infrastructure. Some of this just works by chance on
particular platforms but will fail in hard to diagnose ways when the
driver is used on platforms where the underlying MSI interrupt
management code does not expect the creative abuse.
Another shortcoming of today's PCI/MSI-X support is the inability to
allocate or free individual vectors after the initial enablement of
MSI-X. This results in an works by chance implementation of VFIO (PCI
pass-through) where interrupts on the host side are not set up upfront
to avoid resource exhaustion. They are expanded at run-time when the
guest actually tries to use them. The way how this is implemented is
that the host disables MSI-X and then re-enables it with a larger
number of vectors again. That works by chance because most device
drivers set up all interrupts before the device actually will utilize
them. But that's not universally true because some drivers allocate a
large enough number of vectors but do not utilize them until it's
actually required, e.g. for acceleration support. But at that point
other interrupts of the device might be in active use and the MSI-X
disable/enable dance can just result in losing interrupts and
therefore hard to diagnose subtle problems.
Last but not least the "global" PCI/MSI-X domain approach prevents to
utilize PCI/MSI[-X] and PCI/IMS on the same device due to the fact
that IMS is not longer providing a uniform storage and configuration
model.
The solution to this is to implement the missing step and switch from
global PCI/MSI domains to per device PCI/MSI domains. The resulting
hierarchy then looks like this:
|--- [PCI/MSI] device 1
[Vector]---[Remapping]---|...
|--- [PCI/MSI] device N
which in turn allows to provide support for multiple domains per
device:
|--- [PCI/MSI] device 1
|--- [PCI/IMS] device 1
[Vector]---[Remapping]---|...
|--- [PCI/MSI] device N
|--- [PCI/IMS] device N
This work converts the MSI and PCI/MSI core and the x86 interrupt
domains to the new model, provides new interfaces for post-enable
allocation/free of MSI-X interrupts and the base framework for
PCI/IMS. PCI/IMS has been verified with the work in progress IDXD
driver.
There is work in progress to convert ARM over which will replace the
platform MSI train-wreck. The cleanup of VFIO, NTB and other creative
"solutions" are in the works as well.
Drivers:
- Updates for the LoongArch interrupt chip drivers
- Support for MTK CIRQv2
- The usual small fixes and updates all over the place"
* tag 'irq-core-2022-12-10' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (134 commits)
irqchip/ti-sci-inta: Fix kernel doc
irqchip/gic-v2m: Mark a few functions __init
irqchip/gic-v2m: Include arm-gic-common.h
irqchip/irq-mvebu-icu: Fix works by chance pointer assignment
iommu/amd: Enable PCI/IMS
iommu/vt-d: Enable PCI/IMS
x86/apic/msi: Enable PCI/IMS
PCI/MSI: Provide pci_ims_alloc/free_irq()
PCI/MSI: Provide IMS (Interrupt Message Store) support
genirq/msi: Provide constants for PCI/IMS support
x86/apic/msi: Enable MSI_FLAG_PCI_MSIX_ALLOC_DYN
PCI/MSI: Provide post-enable dynamic allocation interfaces for MSI-X
PCI/MSI: Provide prepare_desc() MSI domain op
PCI/MSI: Split MSI-X descriptor setup
genirq/msi: Provide MSI_FLAG_MSIX_ALLOC_DYN
genirq/msi: Provide msi_domain_alloc_irq_at()
genirq/msi: Provide msi_domain_ops:: Prepare_desc()
genirq/msi: Provide msi_desc:: Msi_data
genirq/msi: Provide struct msi_map
x86/apic/msi: Remove arch_create_remap_msi_irq_domain()
...
|
|
PCI/IMS works like PCI/MSI-X in the remapping. Just add the feature flag,
but only when on real hardware.
Virtualized IOMMUs need additional support, e.g. for PASID.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Acked-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20221124232327.081482253@linutronix.de
|
|
Remove the global PCI/MSI irqdomain implementation and provide the required
MSI parent ops so the PCI/MSI code can detect the new parent and setup per
device domains.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Acked-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20221124232326.151226317@linutronix.de
|
|
Enable MSI parent domain support in the x86 vector domain and fixup the
checks in the iommu implementations to check whether device::msi::domain is
the default MSI parent domain. That keeps the existing logic to protect
e.g. devices behind VMD working.
The interrupt remap PCI/MSI code still works because the underlying vector
domain still provides the same functionality.
None of the other x86 PCI/MSI, e.g. XEN and HyperV, implementations are
affected either. They still work the same way both at the low level and the
PCI/MSI implementations they provide.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Acked-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20221124232326.034672592@linutronix.de
|
|
Impacted QAT device IDs that need extra dtlb flush quirk is ranging
from 0x4940 to 0x4943. After bitwise AND device ID with 0xfffc the
result should be 0x4940 instead of 0x494c to identify these devices.
Fixes: e65a6897be5e ("iommu/vt-d: Add a fix for devices need extra dtlb flush")
Reported-by: Raghunathan Srinivasan <raghunathan.srinivasan@intel.com>
Signed-off-by: Ashok Raj <ashok.raj@intel.com>
Signed-off-by: Jacob Pan <jacob.jun.pan@linux.intel.com>
Link: https://lore.kernel.org/r/20221203005610.2927487-1-jacob.jun.pan@linux.intel.com
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
for_each_pci_dev() is implemented by pci_get_device(). The comment of
pci_get_device() says that it will increase the reference count for the
returned pci_dev and also decrease the reference count for the input
pci_dev @from if it is not NULL.
If we break for_each_pci_dev() loop with pdev not NULL, we need to call
pci_dev_put() to decrease the reference count. Add the missing
pci_dev_put() for the error path to avoid reference count leak.
Fixes: 2e4552893038 ("iommu/vt-d: Unify the way to process DMAR device scope array")
Signed-off-by: Xiongfeng Wang <wangxiongfeng2@huawei.com>
Link: https://lore.kernel.org/r/20221121113649.190393-3-wangxiongfeng2@huawei.com
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
for_each_pci_dev() is implemented by pci_get_device(). The comment of
pci_get_device() says that it will increase the reference count for the
returned pci_dev and also decrease the reference count for the input
pci_dev @from if it is not NULL.
If we break for_each_pci_dev() loop with pdev not NULL, we need to call
pci_dev_put() to decrease the reference count. Add the missing
pci_dev_put() before 'return true' to avoid reference count leak.
Fixes: 89a6079df791 ("iommu/vt-d: Force IOMMU on for platform opt in hint")
Signed-off-by: Xiongfeng Wang <wangxiongfeng2@huawei.com>
Link: https://lore.kernel.org/r/20221121113649.190393-2-wangxiongfeng2@huawei.com
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
As comment of pci_get_domain_bus_and_slot() says, it returns a pci device
with refcount increment, when finish using it, the caller must decrease
the reference count by calling pci_dev_put(). So call pci_dev_put() after
using the 'pdev' to avoid refcount leak.
Besides, if the 'pdev' is null or intel_svm_prq_report() returns error,
there is no need to trace this fault.
Fixes: 06f4b8d09dba ("iommu/vt-d: Remove unnecessary SVA data accesses in page fault path")
Suggested-by: Lu Baolu <baolu.lu@linux.intel.com>
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Link: https://lore.kernel.org/r/20221119144028.2452731-1-yangyingliang@huawei.com
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
QAT devices on Intel Sapphire Rapids and Emerald Rapids have a defect in
address translation service (ATS). These devices may inadvertently issue
ATS invalidation completion before posted writes initiated with
translated address that utilized translations matching the invalidation
address range, violating the invalidation completion ordering.
This patch adds an extra device TLB invalidation for the affected devices,
it is needed to ensure no more posted writes with translated address
following the invalidation completion. Therefore, the ordering is
preserved and data-corruption is prevented.
Device TLBs are invalidated under the following six conditions:
1. Device driver does DMA API unmap IOVA
2. Device driver unbind a PASID from a process, sva_unbind_device()
3. PASID is torn down, after PASID cache is flushed. e.g. process
exit_mmap() due to crash
4. Under SVA usage, called by mmu_notifier.invalidate_range() where
VM has to free pages that were unmapped
5. userspace driver unmaps a DMA buffer
6. Cache invalidation in vSVA usage (upcoming)
For #1 and #2, device drivers are responsible for stopping DMA traffic
before unmap/unbind. For #3, iommu driver gets mmu_notifier to
invalidate TLB the same way as normal user unmap which will do an extra
invalidation. The dTLB invalidation after PASID cache flush does not
need an extra invalidation.
Therefore, we only need to deal with #4 and #5 in this patch. #1 is also
covered by this patch due to common code path with #5.
Tested-by: Yuzhang Luo <yuzhang.luo@intel.com>
Reviewed-by: Ashok Raj <ashok.raj@intel.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Signed-off-by: Jacob Pan <jacob.jun.pan@linux.intel.com>
Link: https://lore.kernel.org/r/20221130062449.1360063-1-jacob.jun.pan@linux.intel.com
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
SRS cap is the hardware cap telling if the hardware IOMMU can support
requests seeking supervisor privilege or not. SRE bit in scalable-mode
PASID table entry is treated as Reserved(0) for implementation not
supporting SRS cap.
Checking SRS cap before setting SRE bit can avoid the non-recoverable
fault of "Non-zero reserved field set in PASID Table Entry" caused by
setting SRE bit while there is no SRS cap support. The fault messages
look like below:
DMAR: DRHD: handling fault status reg 2
DMAR: [DMA Read NO_PASID] Request device [00:0d.0] fault addr 0x1154e1000
[fault reason 0x5a]
SM: Non-zero reserved field set in PASID Table Entry
Fixes: 6f7db75e1c46 ("iommu/vt-d: Add second level page table interface")
Cc: stable@vger.kernel.org
Signed-off-by: Tina Zhang <tina.zhang@intel.com>
Link: https://lore.kernel.org/r/20221115070346.1112273-1-tina.zhang@intel.com
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Link: https://lore.kernel.org/r/20221116051544.26540-3-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
The A/D bits are preseted for IOVA over first level(FL) usage for both
kernel DMA (i.e, domain typs is IOMMU_DOMAIN_DMA) and user space DMA
usage (i.e., domain type is IOMMU_DOMAIN_UNMANAGED).
Presetting A bit in FL requires to preset the bit in every related paging
entries, including the non-leaf ones. Otherwise, hardware may treat this
as an error. For example, in a case of ECAP_REG.SMPWC==0, DMA faults might
occur with below DMAR fault messages (wrapped for line length) dumped.
DMAR: DRHD: handling fault status reg 2
DMAR: [DMA Read NO_PASID] Request device [aa:00.0] fault addr 0x10c3a6000
[fault reason 0x90]
SM: A/D bit update needed in first-level entry when set up in no snoop
Fixes: 289b3b005cb9 ("iommu/vt-d: Preset A/D bits for user space DMA usage")
Cc: stable@vger.kernel.org
Signed-off-by: Tina Zhang <tina.zhang@intel.com>
Link: https://lore.kernel.org/r/20221113010324.1094483-1-tina.zhang@intel.com
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Link: https://lore.kernel.org/r/20221116051544.26540-2-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
Now that the PCI/MSI core code does early checking for multi-MSI support
X86_IRQ_ALLOC_CONTIGUOUS_VECTORS is not required anymore.
Remove the flag and rely on MSI_FLAG_MULTI_PCI_MSI.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Link: https://lore.kernel.org/r/20221111122015.865042356@linutronix.de
|
|
PCI/Multi-MSI is MSI specific and not supported for MSI-X.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Ashok Raj <ashok.raj@intel.com>
Link: https://lore.kernel.org/r/20221111122013.713848846@linutronix.de
|
|
A splat from kmem_cache_destroy() was seen with a kernel prior to
commit ee2653bbe89d ("iommu/vt-d: Remove domain and devinfo mempool")
when there was a failure in init_dmars(), because the iommu_domain
cache still had objects. While the mempool code is now gone, there
still is a leak of the si_domain memory if init_dmars() fails. So
clean up si_domain in the init_dmars() error path.
Cc: Lu Baolu <baolu.lu@linux.intel.com>
Cc: Joerg Roedel <joro@8bytes.org>
Cc: Will Deacon <will@kernel.org>
Cc: Robin Murphy <robin.murphy@arm.com>
Fixes: 86080ccc223a ("iommu/vt-d: Allocate si_domain in init_dmars()")
Signed-off-by: Jerry Snitselaar <jsnitsel@redhat.com>
Link: https://lore.kernel.org/r/20221010144842.308890-1-jsnitsel@redhat.com
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
Commit 5f64ce5411b46 ("iommu/vt-d: Duplicate iommu_resv_region objects
per device list") converted rcu_lock in get_resv_regions to
dmar_global_lock to allow sleeping in iommu_alloc_resv_region(). This
introduced possible recursive locking if get_resv_regions is called from
within a section where intel_iommu_init() already holds dmar_global_lock.
Especially, after commit 57365a04c921 ("iommu: Move bus setup to IOMMU
device registration"), below lockdep splats could always be seen.
============================================
WARNING: possible recursive locking detected
6.0.0-rc4+ #325 Tainted: G I
--------------------------------------------
swapper/0/1 is trying to acquire lock:
ffffffffa8a18c90 (dmar_global_lock){++++}-{3:3}, at:
intel_iommu_get_resv_regions+0x25/0x270
but task is already holding lock:
ffffffffa8a18c90 (dmar_global_lock){++++}-{3:3}, at:
intel_iommu_init+0x36d/0x6ea
...
Call Trace:
<TASK>
dump_stack_lvl+0x48/0x5f
__lock_acquire.cold.73+0xad/0x2bb
lock_acquire+0xc2/0x2e0
? intel_iommu_get_resv_regions+0x25/0x270
? lock_is_held_type+0x9d/0x110
down_read+0x42/0x150
? intel_iommu_get_resv_regions+0x25/0x270
intel_iommu_get_resv_regions+0x25/0x270
iommu_create_device_direct_mappings.isra.28+0x8d/0x1c0
? iommu_get_dma_cookie+0x6d/0x90
bus_iommu_probe+0x19f/0x2e0
iommu_device_register+0xd4/0x130
intel_iommu_init+0x3e1/0x6ea
? iommu_setup+0x289/0x289
? rdinit_setup+0x34/0x34
pci_iommu_init+0x12/0x3a
do_one_initcall+0x65/0x320
? rdinit_setup+0x34/0x34
? rcu_read_lock_sched_held+0x5a/0x80
kernel_init_freeable+0x28a/0x2f3
? rest_init+0x1b0/0x1b0
kernel_init+0x1a/0x130
ret_from_fork+0x1f/0x30
</TASK>
This rolls back dmar_global_lock to rcu_lock in get_resv_regions to avoid
the lockdep splat.
Fixes: 57365a04c921 ("iommu: Move bus setup to IOMMU device registration")
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Tested-by: Alex Williamson <alex.williamson@redhat.com>
Link: https://lore.kernel.org/r/20220927053109.4053662-3-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
Add gfp parameter to iommu_alloc_resv_region() for the callers to specify
the memory allocation behavior. Thus iommu_alloc_resv_region() could also
be available in critical contexts.
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Tested-by: Alex Williamson <alex.williamson@redhat.com>
Link: https://lore.kernel.org/r/20220927053109.4053662-2-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
'virtio', 'x86/vt-d', 'x86/amd' and 'core' into next
|
|
Some VT-d hardware implementations invalidate all DMA remapping hardware
translation caches as part of SRTP flow. The VT-d spec adds a ESRTPS
(Enhanced Set Root Table Pointer Support, section 11.4.2 in VT-d spec)
capability bit to indicate this. With this bit set, software has no need
to issue the global invalidation request.
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Reviewed-by: Jerry Snitselaar <jsnitsel@redhat.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Link: https://lore.kernel.org/r/20220919062523.3438951-3-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
Some VT-d hardware implementations invalidate all interrupt remapping
hardware translation caches as part of SIRTP flow. The VT-d spec adds
a ESIRTPS (Enhanced Set Interrupt Remap Table Pointer Support, section
11.4.2 in VT-d spec) capability bit to indicate this.
The spec also states in 11.4.4 that hardware also performs global
invalidation on all interrupt remapping caches as part of Interrupt
Remapping Disable operation if ESIRTPS capability bit is set.
This checks the ESIRTPS capability bit and skip software global cache
invalidation if it's set.
Signed-off-by: Jacob Pan <jacob.jun.pan@linux.intel.com>
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Reviewed-by: Jerry Snitselaar <jsnitsel@redhat.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Link: https://lore.kernel.org/r/20220921065741.3572495-1-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
This renaming better describes it is for first level page table (a.k.a
first stage page table since VT-d spec 3.4).
Signed-off-by: Yi Liu <yi.l.liu@intel.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Link: https://lore.kernel.org/r/20220916071326.2223901-1-yi.l.liu@intel.com
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
It is not used anywhere in the tree. Remove it to avoid dead code.
No functional change intended.
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Link: https://lore.kernel.org/r/20220915081645.1834555-1-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
Previously the PCI PASID and PRI capabilities are enabled in the path of
iommu device probe only if INTEL_IOMMU_SVM is configured and the device
supports ATS. As we've already decoupled the I/O page fault handler from
SVA, we could also decouple PASID and PRI enabling from it to make room
for growth of new features like kernel DMA with PASID, SIOV and nested
translation.
At the same time, the iommu_enable_dev_iotlb() helper is also called in
iommu_dev_enable_feature(dev, IOMMU_DEV_FEAT_SVA) path. It's unnecessary
and duplicate. This cleanups this helper to make the code neat.
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Link: https://lore.kernel.org/r/20220915085814.2261409-1-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
The existing I/O page fault handling code accesses the per-PASID SVA data
structures. This is unnecessary and makes the fault handling code only
suitable for SVA scenarios. This removes the SVA data accesses from the
I/O page fault reporting and responding code, so that the fault handling
code could be generic.
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Link: https://lore.kernel.org/r/20220914011821.400986-1-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
Check 5-level paging capability for 57 bits address width instead of
checking 1GB large page capability.
Fixes: 53fc7ad6edf2 ("iommu/vt-d: Correctly calculate sagaw value of IOMMU")
Cc: stable@vger.kernel.org
Reported-by: Raghunathan Srinivasan <raghunathan.srinivasan@intel.com>
Signed-off-by: Yi Liu <yi.l.liu@intel.com>
Reviewed-by: Jerry Snitselaar <jsnitsel@redhat.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Reviewed-by: Raghunathan Srinivasan <raghunathan.srinivasan@intel.com>
Link: https://lore.kernel.org/r/20220916071212.2223869-2-yi.l.liu@intel.com
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
This reverts commit 9cd4f1434479f1ac25c440c421fbf52069079914.
Some issues were reported on the original commit. Some thunderbolt devices
don't work anymore due to the following DMA fault.
DMAR: DRHD: handling fault status reg 2
DMAR: [INTR-REMAP] Request device [09:00.0] fault index 0x8080
[fault reason 0x25]
Blocked a compatibility format interrupt request
Bring it back for now to avoid functional regression.
Fixes: 9cd4f1434479f ("iommu/vt-d: Fix possible recursive locking in intel_iommu_init()")
Link: https://lore.kernel.org/linux-iommu/485A6EA5-6D58-42EA-B298-8571E97422DE@getmailspring.com/
Link: https://bugzilla.kernel.org/show_bug.cgi?id=216497
Cc: Mika Westerberg <mika.westerberg@linux.intel.com>
Cc: <stable@vger.kernel.org> # 5.19.x
Reported-and-tested-by: George Hilliard <thirtythreeforty@gmail.com>
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Link: https://lore.kernel.org/r/20220920081701.3453504-1-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
The global rwsem dmar_global_lock was introduced by commit 3a5670e8ac932
("iommu/vt-d: Introduce a rwsem to protect global data structures"). It
is used to protect DMAR related global data from DMAR hotplug operations.
The dmar_global_lock used in the intel_iommu_init() might cause recursive
locking issue, for example, intel_iommu_get_resv_regions() is taking the
dmar_global_lock from within a section where intel_iommu_init() already
holds it via probe_acpi_namespace_devices().
Using dmar_global_lock in intel_iommu_init() could be relaxed since it is
unlikely that any IO board must be hot added before the IOMMU subsystem is
initialized. This eliminates the possible recursive locking issue by moving
down DMAR hotplug support after the IOMMU is initialized and removing the
uses of dmar_global_lock in intel_iommu_init().
Fixes: d5692d4af08cd ("iommu/vt-d: Fix suspicious RCU usage in probe_acpi_namespace_devices()")
Reported-by: Robin Murphy <robin.murphy@arm.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/894db0ccae854b35c73814485569b634237b5538.1657034828.git.robin.murphy@arm.com
Link: https://lore.kernel.org/r/20220718235325.3952426-1-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
|
|
Now that dma-iommu.h only contains internal interfaces, make it
private to the IOMMU subsytem.
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Link: https://lore.kernel.org/r/b237e06c56a101f77af142a54b629b27aa179d22.1660668998.git.robin.murphy@arm.com
[ joro : re-add stub for iommu_dma_get_resv_regions ]
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
With CONFIG_INTEL_IOMMU_DEBUGFS enabled, below lockdep splat are seen
when an I/O fault occurs on a machine with an Intel IOMMU in it.
DMAR: DRHD: handling fault status reg 3
DMAR: [DMA Write NO_PASID] Request device [00:1a.0] fault addr 0x0
[fault reason 0x05] PTE Write access is not set
DMAR: Dump dmar0 table entries for IOVA 0x0
DMAR: root entry: 0x0000000127f42001
DMAR: context entry: hi 0x0000000000001502, low 0x000000012d8ab001
================================
WARNING: inconsistent lock state
5.20.0-0.rc0.20220812git7ebfc85e2cd7.10.fc38.x86_64 #1 Not tainted
--------------------------------
inconsistent {HARDIRQ-ON-W} -> {IN-HARDIRQ-W} usage.
rngd/1006 [HC1[1]:SC0[0]:HE0:SE1] takes:
ff177021416f2d78 (&k->k_lock){?.+.}-{2:2}, at: klist_next+0x1b/0x160
{HARDIRQ-ON-W} state was registered at:
lock_acquire+0xce/0x2d0
_raw_spin_lock+0x33/0x80
klist_add_tail+0x46/0x80
bus_add_device+0xee/0x150
device_add+0x39d/0x9a0
add_memory_block+0x108/0x1d0
memory_dev_init+0xe1/0x117
driver_init+0x43/0x4d
kernel_init_freeable+0x1c2/0x2cc
kernel_init+0x16/0x140
ret_from_fork+0x1f/0x30
irq event stamp: 7812
hardirqs last enabled at (7811): [<ffffffff85000e86>] asm_sysvec_apic_timer_interrupt+0x16/0x20
hardirqs last disabled at (7812): [<ffffffff84f16894>] irqentry_enter+0x54/0x60
softirqs last enabled at (7794): [<ffffffff840ff669>] __irq_exit_rcu+0xf9/0x170
softirqs last disabled at (7787): [<ffffffff840ff669>] __irq_exit_rcu+0xf9/0x170
The klist iterator functions using spin_*lock_irq*() but the klist
insertion functions using spin_*lock(), combined with the Intel DMAR
IOMMU driver iterating over klists from atomic (hardirq) context, where
pci_get_domain_bus_and_slot() calls into bus_find_device() which iterates
over klists.
As currently there's no plan to fix the klist to make it safe to use in
atomic context, this fixes the lockdep splat by avoid calling
pci_get_domain_bus_and_slot() in the hardirq context.
Fixes: 8ac0b64b9735 ("iommu/vt-d: Use pci_get_domain_bus_and_slot() in pgtable_walk()")
Reported-by: Lennert Buytenhek <buytenh@wantstofly.org>
Link: https://lore.kernel.org/linux-iommu/Yvo2dfpEh%2FWC+Wrr@wantstofly.org/
Link: https://lore.kernel.org/linux-iommu/YvyBdPwrTuHHbn5X@wantstofly.org/
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Link: https://lore.kernel.org/r/20220819015949.4795-1-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
The per domain spinlock is acquired in iommu_flush_dev_iotlb(), which
is possbile to be called in the interrupt context. For example, the
drm-intel's CI system got completely blocked with below error:
WARNING: inconsistent lock state
6.0.0-rc1-CI_DRM_11990-g6590d43d39b9+ #1 Not tainted
--------------------------------
inconsistent {SOFTIRQ-ON-W} -> {IN-SOFTIRQ-W} usage.
swapper/6/0 [HC0[0]:SC1[1]:HE1:SE0] takes:
ffff88810440d678 (&domain->lock){+.?.}-{2:2}, at: iommu_flush_dev_iotlb.part.61+0x23/0x80
{SOFTIRQ-ON-W} state was registered at:
lock_acquire+0xd3/0x310
_raw_spin_lock+0x2a/0x40
domain_update_iommu_cap+0x20b/0x2c0
intel_iommu_attach_device+0x5bd/0x860
__iommu_attach_device+0x18/0xe0
bus_iommu_probe+0x1f3/0x2d0
bus_set_iommu+0x82/0xd0
intel_iommu_init+0xe45/0x102a
pci_iommu_init+0x9/0x31
do_one_initcall+0x53/0x2f0
kernel_init_freeable+0x18f/0x1e1
kernel_init+0x11/0x120
ret_from_fork+0x1f/0x30
irq event stamp: 162354
hardirqs last enabled at (162354): [<ffffffff81b59274>] _raw_spin_unlock_irqrestore+0x54/0x70
hardirqs last disabled at (162353): [<ffffffff81b5901b>] _raw_spin_lock_irqsave+0x4b/0x50
softirqs last enabled at (162338): [<ffffffff81e00323>] __do_softirq+0x323/0x48e
softirqs last disabled at (162349): [<ffffffff810c1588>] irq_exit_rcu+0xb8/0xe0
other info that might help us debug this:
Possible unsafe locking scenario:
CPU0
----
lock(&domain->lock);
<Interrupt>
lock(&domain->lock);
*** DEADLOCK ***
1 lock held by swapper/6/0:
This coverts the spin_lock/unlock() into the irq save/restore varieties
to fix the recursive locking issues.
Fixes: ffd5869d93530 ("iommu/vt-d: Replace spin_lock_irqsave() with spin_lock()")
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Acked-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://lore.kernel.org/r/20220817025650.3253959-1-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
The Intel IOMMU driver possibly selects between the first-level and the
second-level translation tables for DMA address translation. However,
the levels of page-table walks for the 4KB base page size are calculated
from the SAGAW field of the capability register, which is only valid for
the second-level page table. This causes the IOMMU driver to stop working
if the hardware (or the emulated IOMMU) advertises only first-level
translation capability and reports the SAGAW field as 0.
This solves the above problem by considering both the first level and the
second level when calculating the supported page table levels.
Fixes: b802d070a52a1 ("iommu/vt-d: Use iova over first level")
Cc: stable@vger.kernel.org
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Link: https://lore.kernel.org/r/20220817023558.3253263-1-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
The translation table copying code for kdump kernels is currently based
on the extended root/context entry formats of ECS mode defined in older
VT-d v2.5, and doesn't handle the scalable mode formats. This causes
the kexec capture kernel boot failure with DMAR faults if the IOMMU was
enabled in scalable mode by the previous kernel.
The ECS mode has already been deprecated by the VT-d spec since v3.0 and
Intel IOMMU driver doesn't support this mode as there's no real hardware
implementation. Hence this converts ECS checking in copying table code
into scalable mode.
The existing copying code consumes a bit in the context entry as a mark
of copied entry. It needs to work for the old format as well as for the
extended context entries. As it's hard to find such a common bit for both
legacy and scalable mode context entries. This replaces it with a per-
IOMMU bitmap.
Fixes: 7373a8cc38197 ("iommu/vt-d: Setup context and enable RID2PASID support")
Cc: stable@vger.kernel.org
Reported-by: Jerry Snitselaar <jsnitsel@redhat.com>
Tested-by: Wen Jin <wen.jin@intel.com>
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Link: https://lore.kernel.org/r/20220817011035.3250131-1-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
Although iommu-dma is a per-architecture chonce, that is currently
implemented in a rather haphazard way. Selecting from the arch Kconfig
was the original logical approach, but is complicated by having to
manage dependencies; conversely, selecting from drivers ends up hiding
the architecture dependency *too* well. Instead, let's just have it
enable itself automatically when IOMMU API support is enabled for the
relevant architectures. It can't get much clearer than that.
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Link: https://lore.kernel.org/r/2e33c8bc2b1bb478157b7964bfed976cb7466139.1660668998.git.robin.murphy@arm.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
Clean up the remaining trivial bus_set_iommu() callsites along
with the implementation. Now drivers only have to know and care
about iommu_device instances, phew!
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Tested-by: Matthew Rosato <mjrosato@linux.ibm.com> # s390
Tested-by: Niklas Schnelle <schnelle@linux.ibm.com> # s390
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Link: https://lore.kernel.org/r/ea383d5f4d74ffe200ab61248e5de6e95846180a.1660572783.git.robin.murphy@arm.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
Currently we rely on registering all our instances before initially
allowing any .probe_device calls via bus_set_iommu(). In preparation for
phasing out the latter, make sure we won't inadvertently return success
for a device associated with a known but not yet registered instance,
otherwise we'll run straight into iommu_group_get_for_dev() trying to
use NULL ops.
That also highlights an issue with intel_iommu_get_resv_regions() taking
dmar_global_lock from within a section where intel_iommu_init() already
holds it, which already exists via probe_acpi_namespace_devices() when
an ANDD device is probed, but gets more obvious with the upcoming change
to iommu_device_register(). Since they are both read locks it manages
not to deadlock in practice, and a more in-depth rework of this locking
is underway, so no attempt is made to address it here.
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Link: https://lore.kernel.org/r/579f2692291bcbfc3ac64f7456fcff0d629af131.1660572783.git.robin.murphy@arm.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
With all callers now converted to the device-specific version, retire
the old bus-based interface, and give drivers the chance to indicate
accurate per-instance capabilities.
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Reviewed-by: Lu Baolu <baolu.lu@linux.intel.com>
Link: https://lore.kernel.org/r/d8bd8777d06929ad8f49df7fc80e1b9af32a41b5.1660574547.git.robin.murphy@arm.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
Pull bitmap updates from Yury Norov:
- fix the duplicated comments on bitmap_to_arr64() (Qu Wenruo)
- optimize out non-atomic bitops on compile-time constants (Alexander
Lobakin)
- cleanup bitmap-related headers (Yury Norov)
- x86/olpc: fix 'logical not is only applied to the left hand side'
(Alexander Lobakin)
- lib/nodemask: inline wrappers around bitmap (Yury Norov)
* tag 'bitmap-6.0-rc1' of https://github.com/norov/linux: (26 commits)
lib/nodemask: inline next_node_in() and node_random()
powerpc: drop dependency on <asm/machdep.h> in archrandom.h
x86/olpc: fix 'logical not is only applied to the left hand side'
lib/cpumask: move some one-line wrappers to header file
headers/deps: mm: align MANITAINERS and Docs with new gfp.h structure
headers/deps: mm: Split <linux/gfp_types.h> out of <linux/gfp.h>
headers/deps: mm: Optimize <linux/gfp.h> header dependencies
lib/cpumask: move trivial wrappers around find_bit to the header
lib/cpumask: change return types to unsigned where appropriate
cpumask: change return types to bool where appropriate
lib/bitmap: change type of bitmap_weight to unsigned long
lib/bitmap: change return types to bool where appropriate
arm: align find_bit declarations with generic kernel
iommu/vt-d: avoid invalid memory access via node_online(NUMA_NO_NODE)
lib/test_bitmap: test the tail after bitmap_to_arr64()
lib/bitmap: fix off-by-one in bitmap_to_arr64()
lib: test_bitmap: add compile-time optimization/evaluations assertions
bitmap: don't assume compiler evaluates small mem*() builtins calls
net/ice: fix initializing the bitmap in the switch code
bitops: let optimize out non-atomic bitops on compile-time constants
...
|
|
'virtio', 'x86/vt-d', 'x86/amd' and 'core' into next
|
|
bitmap_weight() doesn't return negative values, so change it's type
to unsigned long. It may help compiler to generate better code and
catch bugs.
Signed-off-by: Yury Norov <yury.norov@gmail.com>
|
|
The g_iommus and g_num_of_iommus is not used anywhere. Remove them to
avoid dead code.
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Reviewed-by: Steve Wahl <steve.wahl@hpe.com>
Link: https://lore.kernel.org/r/20220702015610.2849494-6-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
The Intel IOMMU hot-add process starts from dmar_device_hotplug(). It
uses the global dmar_global_lock to synchronize all the hot-add and
hot-remove paths. In the hot-add path, the new IOMMU data structures
are allocated firstly by dmar_parse_one_drhd() and then initialized by
dmar_hp_add_drhd(). All the IOMMU units are allocated and initialized
in the same synchronized path. There is no case where any IOMMU unit
is created and then initialized for multiple times.
This removes the unnecessary check in intel_iommu_add() which is the
last reference place of the global IOMMU array.
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Reviewed-by: Steve Wahl <steve.wahl@hpe.com>
Link: https://lore.kernel.org/r/20220702015610.2849494-5-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
When a DMA domain is attached to a device, it needs to allocate a domain
ID from its IOMMU. Currently, the domain ID information is stored in two
static arrays embedded in the domain structure. This can lead to memory
waste when the driver is running on a small platform.
This optimizes these static arrays by replacing them with an xarray and
consuming memory on demand.
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Reviewed-by: Steve Wahl <steve.wahl@hpe.com>
Link: https://lore.kernel.org/r/20220702015610.2849494-4-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
Switch dmar unit sequence id allocation and release from bitmap to IDA
interface.
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Reviewed-by: Steve Wahl <steve.wahl@hpe.com>
Link: https://lore.kernel.org/r/20220702015610.2849494-3-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
It is not used anywhere. Remove it to avoid dead code.
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Reviewed-by: Steve Wahl <steve.wahl@hpe.com>
Link: https://lore.kernel.org/r/20220702015610.2849494-2-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
Using a global device_domain_lock spinlock to protect per-domain device
tracking lists is an inefficient way, especially considering this lock
is also needed in the hot paths. This optimizes the locking mechanism
by converting the global lock to per domain lock.
On the other hand, as the device tracking lists are never accessed in
any interrupt context, there is no need to disable interrupts while
spinning. Replace irqsave variant with spinlock calls.
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Link: https://lore.kernel.org/r/20220706025524.2904370-12-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
The device_domain_lock is used to protect the device tracking list of
a domain. Remove unnecessary spin_lock/unlock()'s and move the necessary
ones around the list access.
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Link: https://lore.kernel.org/r/20220706025524.2904370-11-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
Fold __dmar_remove_one_dev_info() into dmar_remove_one_dev_info() which
is its only caller. Make the spin lock critical range only cover the
device list change code and remove some unnecessary checks.
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Link: https://lore.kernel.org/r/20220706025524.2904370-10-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
When the IOMMU domain is about to be freed, it should not be set on any
device. Instead of silently dealing with some bug cases, it's better to
trigger a warning to report and fix any potential bugs at the first time.
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Link: https://lore.kernel.org/r/20220706025524.2904370-9-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
The iommu->lock is used to protect the per-IOMMU pasid directory table
and pasid table. Move the spinlock acquisition/release into the helpers
to make the code self-contained.
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Link: https://lore.kernel.org/r/20220706025524.2904370-8-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
The iommu->lock is used to protect the per-IOMMU domain ID resource.
Moving the lock into the ID alloc/free helpers makes the code more
compact. At the same time, the device_domain_lock is irrelevant to
the domain ID resource, remove its assertion as well.
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Link: https://lore.kernel.org/r/20220706025524.2904370-7-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
