linux-stable.git/drivers/vfio, branch v5.9.2

vfio iommu type1: Fix memory leak in vfio_iommu_type1_pin_pages

2020-10-29T09:11:57+00:00

[ Upstream commit 2e6cfd496f5b57034cf2aec738799571b5a52124 ]

pfn is not added to pfn_list when vfio_add_to_pfn_list fails.
vfio_unpin_page_external will exit directly without calling
vfio_iova_put_vfio_pfn.  This will lead to a memory leak.

Fixes: a54eb55045ae ("vfio iommu type1: Add support for mediated devices")
Signed-off-by: Xiaoyang Xu 
[aw: simplified logic, add Fixes]
Signed-off-by: Alex Williamson 
Signed-off-by: Sasha Levin

vfio/pci: Clear token on bypass registration failure

2020-10-29T09:11:57+00:00

[ Upstream commit 852b1beecb6ff9326f7ca4bc0fe69ae860ebdb9e ]

The eventfd context is used as our irqbypass token, therefore if an
eventfd is re-used, our token is the same.  The irqbypass code will
return an -EBUSY in this case, but we'll still attempt to unregister
the producer, where if that duplicate token still exists, results in
removing the wrong object.  Clear the token of failed producers so
that they harmlessly fall out when unregistered.

Fixes: 6d7425f109d2 ("vfio: Register/unregister irq_bypass_producer")
Reported-by: guomin chen 
Tested-by: guomin chen 
Signed-off-by: Alex Williamson 
Signed-off-by: Sasha Levin

vfio/type1: fix dirty bitmap calculation in vfio_dma_rw

2020-10-29T09:11:54+00:00

[ Upstream commit 2c5af98592f65517170c7bcc714566590d3f7397 ]

The count of dirtied pages is not only determined by count of copied
pages, but also by the start offset.

e.g. if offset = PAGE_SIZE - 1, and *copied=2, the dirty pages count
is 2, instead of 1 or 0.

Fixes: d6a4c185660c ("vfio iommu: Implementation of ioctl for dirty pages tracking")
Signed-off-by: Yan Zhao 
Signed-off-by: Alex Williamson 
Signed-off-by: Sasha Levin

vfio: fix a missed vfio group put in vfio_pin_pages

2020-10-29T09:11:54+00:00

[ Upstream commit 28b130244061863cf0437b7af1625fb45ec1a71e ]

When error occurs, need to put vfio group after a successful get.

Fixes: 95fc87b44104 ("vfio: Selective dirty page tracking if IOMMU backed device pins pages")
Signed-off-by: Yan Zhao 
Reviewed-by: Cornelia Huck 
Signed-off-by: Alex Williamson 
Signed-off-by: Sasha Levin

vfio/pci: Decouple PCI_COMMAND_MEMORY bit checks from is_virtfn

2020-10-29T09:11:54+00:00

[ Upstream commit 515ecd5368f1510152fa4f9b9ce55b66ac56c334 ]

While it is true that devices with is_virtfn=1 will have a Memory Space
Enable bit that is hard-wired to 0, this is not the only case where we
see this behavior -- For example some bare-metal hypervisors lack
Memory Space Enable bit emulation for devices not setting is_virtfn
(s390). Fix this by instead checking for the newly-added
no_command_memory bit which directly denotes the need for
PCI_COMMAND_MEMORY emulation in vfio.

Fixes: abafbc551fdd ("vfio-pci: Invalidate mmaps and block MMIO access on disabled memory")
Signed-off-by: Matthew Rosato 
Reviewed-by: Niklas Schnelle 
Reviewed-by: Pierre Morel 
Signed-off-by: Alex Williamson 
Signed-off-by: Sasha Levin

vfio: add a singleton check for vfio_group_pin_pages

2020-10-29T09:11:53+00:00

[ Upstream commit 7ef32e52368f62a4e041a4f0abefb4fb64e7fd4a ]

Page pinning is used both to translate and pin device mappings for DMA
purpose, as well as to indicate to the IOMMU backend to limit the dirty
page scope to those pages that have been pinned, in the case of an IOMMU
backed device.
To support this, the vfio_pin_pages() interface limits itself to only
singleton groups such that the IOMMU backend can consider dirty page
scope only at the group level.  Implement the same requirement for the
vfio_group_pin_pages() interface.

Fixes: 95fc87b44104 ("vfio: Selective dirty page tracking if IOMMU backed device pins pages")
Signed-off-by: Yan Zhao 
Signed-off-by: Alex Williamson 
Signed-off-by: Sasha Levin

treewide: Use fallthrough pseudo-keyword

2020-08-23T22:36:59+00:00

Replace the existing /* fall through */ comments and its variants with
the new pseudo-keyword macro fallthrough[1]. Also, remove unnecessary
fall-through markings when it is the case.

[1] https://www.kernel.org/doc/html/v5.7/process/deprecated.html?highlight=fallthrough#implicit-switch-case-fall-through

Signed-off-by: Gustavo A. R. Silva

vfio/type1: Add proper error unwind for vfio_iommu_replay()

2020-08-17T17:09:13+00:00

The vfio_iommu_replay() function does not currently unwind on error,
yet it does pin pages, perform IOMMU mapping, and modify the vfio_dma
structure to indicate IOMMU mapping.  The IOMMU mappings are torn down
when the domain is destroyed, but the other actions go on to cause
trouble later.  For example, the iommu->domain_list can be empty if we
only have a non-IOMMU backed mdev attached.  We don't currently check
if the list is empty before getting the first entry in the list, which
leads to a bogus domain pointer.  If a vfio_dma entry is erroneously
marked as iommu_mapped, we'll attempt to use that bogus pointer to
retrieve the existing physical page addresses.

This is the scenario that uncovered this issue, attempting to hot-add
a vfio-pci device to a container with an existing mdev device and DMA
mappings, one of which could not be pinned, causing a failure adding
the new group to the existing container and setting the conditions
for a subsequent attempt to explode.

To resolve this, we can first check if the domain_list is empty so
that we can reject replay of a bogus domain, should we ever encounter
this inconsistent state again in the future.  The real fix though is
to add the necessary unwind support, which means cleaning up the
current pinning if an IOMMU mapping fails, then walking back through
the r-b tree of DMA entries, reading from the IOMMU which ranges are
mapped, and unmapping and unpinning those ranges.  To be able to do
this, we also defer marking the DMA entry as IOMMU mapped until all
entries are processed, in order to allow the unwind to know the
disposition of each entry.

Fixes: a54eb55045ae ("vfio iommu type1: Add support for mediated devices")
Reported-by: Zhiyi Guo 
Tested-by: Zhiyi Guo 
Reviewed-by: Cornelia Huck 
Signed-off-by: Alex Williamson

vfio-pci: Avoid recursive read-lock usage

2020-08-17T17:08:18+00:00

A down_read on memory_lock is held when performing read/write accesses
to MMIO BAR space, including across the copy_to/from_user() callouts
which may fault.  If the user buffer for these copies resides in an
mmap of device MMIO space, the mmap fault handler will acquire a
recursive read-lock on memory_lock.  Avoid this by reducing the lock
granularity.  Sequential accesses requiring multiple ioread/iowrite
cycles are expected to be rare, therefore typical accesses should not
see additional overhead.

VGA MMIO accesses are expected to be non-fatal regardless of the PCI
memory enable bit to allow legacy probing, this behavior remains with
a comment added.  ioeventfds are now included in memory access testing,
with writes dropped while memory space is disabled.

Fixes: abafbc551fdd ("vfio-pci: Invalidate mmaps and block MMIO access on disabled memory")
Reported-by: Zhiyi Guo 
Tested-by: Zhiyi Guo 
Reviewed-by: Cornelia Huck 
Signed-off-by: Alex Williamson

Merge tag 'vfio-v5.9-rc1' of git://github.com/awilliam/linux-vfio

2020-08-12T19:09:36+00:00

Pull VFIO updates from Alex Williamson:

 - Inclusive naming updates (Alex Williamson)

 - Intel X550 INTx quirk (Alex Williamson)

 - Error path resched between unmaps (Xiang Zheng)

 - SPAPR IOMMU pin_user_pages() conversion (John Hubbard)

 - Trivial mutex simplification (Alex Williamson)

 - QAT device denylist (Giovanni Cabiddu)

 - type1 IOMMU ioctl refactor (Liu Yi L)

* tag 'vfio-v5.9-rc1' of git://github.com/awilliam/linux-vfio:
  vfio/type1: Refactor vfio_iommu_type1_ioctl()
  vfio/pci: Add QAT devices to denylist
  vfio/pci: Add device denylist
  PCI: Add Intel QuickAssist device IDs
  vfio/pci: Hold igate across releasing eventfd contexts
  vfio/spapr_tce: convert get_user_pages() --> pin_user_pages()
  vfio/type1: Add conditional rescheduling after iommu map failed
  vfio/pci: Add Intel X550 to hidden INTx devices
  vfio: Cleanup allowed driver naming