In the PCI hotplug path of the Intel IOMMU driver, replace
the usage of the BUS_NOTIFY_DEL_DEVICE notifier, which is
executed before the driver is unbound from the device, with
BUS_NOTIFY_REMOVED_DEVICE, which runs after that.
This fixes a kernel BUG being triggered in the VT-d code
when the device driver tries to unmap DMA buffers and the
VT-d driver already destroyed all mappings.
Reported-by: Stefani Seibold <stefani@seibold.net>
Cc: stable@vger.kernel.org # v4.3+
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Detach the device that is about to be removed from its
domain (if it has one) to clear any related state like DTE
entry and device's ATS state.
Reported-by: Kelly Zytaruk <Kelly.Zytaruk@amd.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
The AMD Family 15h Models 30h-3Fh (Kaveri) BIOS and Kernel Developer's
Guide omitted part of the BIOS IOMMU L2 register setup specification.
Without this setup the IOMMU L2 does not fully respect write permissions
when handling an ATS translation request.
The IOMMU L2 will set PTE dirty bit when handling an ATS translation with
write permission request, even when PTE RW bit is clear. This may occur by
direct translation (which would cause a PPR) or by prefetch request from
the ATC.
This is observed in practice when the IOMMU L2 modifies a PTE which maps a
pagecache page. The ext4 filesystem driver BUGs when asked to writeback
these (non-modified) pages.
Enable ATS write permission check in the Kaveri IOMMU L2 if BIOS has not.
Signed-off-by: Jay Cornwall <jay@jcornwall.me>
Cc: <stable@vger.kernel.org> # v3.19+
Signed-off-by: Joerg Roedel <jroedel@suse.de>
The setup code for the performance counters in the AMD IOMMU driver
tests whether the counters can be written. It tests to setup a counter
for device 00:00.0, which fails on systems where this particular device
is not covered by the IOMMU.
Fix this by not relying on device 00:00.0 but only on the IOMMU being
present.
Cc: stable@vger.kernel.org
Signed-off-by: Suravee Suthikulpanit <Suravee.Suthikulpanit@amd.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Minor register size and interrupt acknowledgement fixes which only showed
up in testing on newer hardware, but mostly a fix to the MM refcount
handling to prevent a recursive refcount issue when mmap() is used on
the file descriptor associated with a bound PASID.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iEYEABECAAYFAlbC/gAACgkQdwG7hYl686OY8QCfUPH+IB0zou9/MH3JNMz1ujot
I6wAoK0R4KiOFXvjNeNPy+XroZ9xKqv/
=RM+0
-----END PGP SIGNATURE-----
Merge tag 'for-linus-20160216' of git://git.infradead.org/intel-iommu
Pull IOMMU SVM fixes from David Woodhouse:
"Minor register size and interrupt acknowledgement fixes which only
showed up in testing on newer hardware, but mostly a fix to the MM
refcount handling to prevent a recursive refcount issue when mmap() is
used on the file descriptor associated with a bound PASID"
* tag 'for-linus-20160216' of git://git.infradead.org/intel-iommu:
iommu/vt-d: Clear PPR bit to ensure we get more page request interrupts
iommu/vt-d: Fix 64-bit accesses to 32-bit DMAR_GSTS_REG
iommu/vt-d: Fix mm refcounting to hold mm_count not mm_users
According to the VT-d specification we need to clear the PPR bit in
the Page Request Status register when handling page requests, or the
hardware won't generate any more interrupts.
This wasn't actually necessary on SKL/KBL (which may well be the
subject of a hardware erratum, although it's harmless enough). But
other implementations do appear to get it right, and we only ever get
one interrupt unless we clear the PPR bit.
Reported-by: CQ Tang <cq.tang@intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Cc: stable@vger.kernel.org
In below commit alias DTE is set when its peripheral is
setting DTE. However there's a code bug here to wrongly
set the alias DTE, correct it in this patch.
commit e25bfb56ea
Author: Joerg Roedel <jroedel@suse.de>
Date: Tue Oct 20 17:33:38 2015 +0200
iommu/amd: Set alias DTE in do_attach/do_detach
Signed-off-by: Baoquan He <bhe@redhat.com>
Tested-by: Mark Hounschell <markh@compro.net>
Cc: stable@vger.kernel.org # v4.4
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Trying to build a kernel for ARC with both options CONFIG_COMPILE_TEST
and CONFIG_IOMMU_IO_PGTABLE_LPAE enabled (e.g. as a result of "make
allyesconfig") results in the following build failure:
| CC drivers/iommu/io-pgtable-arm.o
| linux/drivers/iommu/io-pgtable-arm.c: In
| function ‘__arm_lpae_alloc_pages’:
| linux/drivers/iommu/io-pgtable-arm.c:221:3:
| error: implicit declaration of function ‘dma_map_single’
| [-Werror=implicit-function-declaration]
| dma = dma_map_single(dev, pages, size, DMA_TO_DEVICE);
| ^
| linux/drivers/iommu/io-pgtable-arm.c:221:42:
| error: ‘DMA_TO_DEVICE’ undeclared (first use in this function)
| dma = dma_map_single(dev, pages, size, DMA_TO_DEVICE);
| ^
Since IOMMU_IO_PGTABLE_LPAE depends on DMA API, io-pgtable-arm.c should
include linux/dma-mapping.h. This fixes the reported failure.
Cc: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Vineet Gupta <vgupta@synopsys.com>
Cc: Joerg Roedel <joro@8bytes.org>
Signed-off-by: Lada Trimasova <ltrimas@synopsys.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
The updates include:
* Small code cleanups in the AMD IOMMUv2 driver
* Scalability improvements for the DMA-API implementation of the
AMD IOMMU driver. This is just a starting point, but already
showed some good improvements in my tests.
* Removal of the unused Renesas IPMMU/IPMMUI driver
* Updates for ARM-SMMU include:
* Some fixes to get the driver working nicely on
Broadcom hardware
* A change to the io-pgtable API to indicate the unit in
which to flush (all callers converted, with Ack from
Laurent)
* Use of devm_* for allocating/freeing the SMMUv3
buffers
* Some other small fixes and improvements for other drivers
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (GNU/Linux)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=bm5E
-----END PGP SIGNATURE-----
Merge tag 'iommu-updates-v4.5' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu
Pull IOMMU updates from Joerg Roedel:
"The updates include:
- Small code cleanups in the AMD IOMMUv2 driver
- Scalability improvements for the DMA-API implementation of the AMD
IOMMU driver. This is just a starting point, but already showed
some good improvements in my tests.
- Removal of the unused Renesas IPMMU/IPMMUI driver
- Updates for ARM-SMMU include:
* Some fixes to get the driver working nicely on Broadcom hardware
* A change to the io-pgtable API to indicate the unit in which to
flush (all callers converted, with Ack from Laurent)
* Use of devm_* for allocating/freeing the SMMUv3 buffers
- Some other small fixes and improvements for other drivers"
* tag 'iommu-updates-v4.5' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu: (46 commits)
iommu/vt-d: Fix up error handling in alloc_iommu
iommu/vt-d: Check the return value of iommu_device_create()
iommu/amd: Remove an unneeded condition
iommu/amd: Preallocate dma_ops apertures based on dma_mask
iommu/amd: Use trylock to aquire bitmap_lock
iommu/amd: Make dma_ops_domain->next_index percpu
iommu/amd: Relax locking in dma_ops path
iommu/amd: Initialize new aperture range before making it visible
iommu/amd: Build io page-tables with cmpxchg64
iommu/amd: Allocate new aperture ranges in dma_ops_alloc_addresses
iommu/amd: Optimize dma_ops_free_addresses
iommu/amd: Remove need_flush from struct dma_ops_domain
iommu/amd: Iterate over all aperture ranges in dma_ops_area_alloc
iommu/amd: Flush iommu tlb in dma_ops_free_addresses
iommu/amd: Rename dma_ops_domain->next_address to next_index
iommu/amd: Remove 'start' parameter from dma_ops_area_alloc
iommu/amd: Flush iommu tlb in dma_ops_aperture_alloc()
iommu/amd: Retry address allocation within one aperture
iommu/amd: Move aperture_range.offset to another cache-line
iommu/amd: Add dma_ops_aperture_alloc() function
...
This is a 32-bit register. Apparently harmless on real hardware, but
causing justified warnings in simulation.
Signed-off-by: CQ Tang <cq.tang@intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Cc: stable@vger.kernel.org
Holding mm_users works OK for graphics, which was the first user of SVM
with VT-d. However, it works less well for other devices, where we actually
do a mmap() from the file descriptor to which the SVM PASID state is tied.
In this case on process exit we end up with a recursive reference count:
- The MM remains alive until the file is closed and the driver's release()
call ends up unbinding the PASID.
- The VMA corresponding to the mmap() remains intact until the MM is
destroyed.
- Thus the file isn't closed, even when exit_files() runs, because the
VMA is still holding a reference to it. And the MM remains alive…
To address this issue, we *stop* holding mm_users while the PASID is bound.
We already hold mm_count by virtue of the MMU notifier, and that can be
made to be sufficient.
It means that for a period during process exit, the fun part of mmput()
has happened and exit_mmap() has been called so the MM is basically
defunct. But the PGD still exists and the PASID is still bound to it.
During this period, we have to be very careful — exit_mmap() doesn't use
mm->mmap_sem because it doesn't expect anyone else to be touching the MM
(quite reasonably, since mm_users is zero). So we also need to fix the
fault handler to just report failure if mm_users is already zero, and to
temporarily bump mm_users while handling any faults.
Additionally, exit_mmap() calls mmu_notifier_release() *before* it tears
down the page tables, which is too early for us to flush the IOTLB for
this PASID. And __mmu_notifier_release() removes every notifier from the
list, so when exit_mmap() finally *does* tear down the mappings and
clear the page tables, we don't get notified. So we work around this by
clearing the PASID table entry in our MMU notifier release() callback.
That way, the hardware *can't* get any pages back from the page tables
before they get cleared.
Hardware designers have confirmed that the resulting 'PASID not present'
faults should be handled just as gracefully as 'page not present' faults,
the important criterion being that they don't perturb the operation for
any *other* PASID in the system.
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Cc: stable@vger.kernel.org
Only check for error when iommu->iommu_dev has been assigned
and only assign drhd->iommu when the function can't fail
anymore.
Signed-off-by: Joerg Roedel <jroedel@suse.de>
This adds the proper check to alloc_iommu to make sure that
the call to iommu_device_create has completed successfully
and if not return the error code to the caller after freeing
up resources allocated previously.
Signed-off-by: Nicholas Krause <xerofoify@gmail.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
When mapping a non-page-aligned scatterlist entry, we copy the original
offset to the output DMA address before aligning it to hand off to
iommu_map_sg(), then later adding the IOVA page address portion to get
the final mapped address. However, when the IOVA page size is smaller
than the CPU page size, it is the offset within the IOVA page we want,
not that within the CPU page, which can easily be larger than an IOVA
page and thus result in an incorrect final address.
Fix the bug by taking only the IOVA-aligned part of the offset as the
basis of the DMA address, not the whole thing.
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
get_device_id() returns an unsigned short device id. It never fails and
it never returns a negative so we can remove this condition.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Preallocate between 4 and 8 apertures when a device gets it
dma_mask. With more apertures we reduce the lock contention
of the domain lock significantly.
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Make this pointer percpu so that we start searching for new
addresses in the range we last stopped and which is has a
higher probability of being still in the cache.
Signed-off-by: Joerg Roedel <jroedel@suse.de>
This allows to build up the page-tables without holding any
locks. As a consequence it removes the need to pre-populate
dma_ops page-tables.
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Don't flush the iommu tlb when we free something behind the
current next_bit pointer. Update the next_bit pointer
instead and let the flush happen on the next wraparound in
the allocation path.
Signed-off-by: Joerg Roedel <jroedel@suse.de>
The flushing of iommu tlbs is now done on a per-range basis.
So there is no need anymore for domain-wide flush tracking.
Signed-off-by: Joerg Roedel <jroedel@suse.de>
It points to the next aperture index to allocate from. We
don't need the full address anymore because this is now
tracked in struct aperture_range.
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Moving it before the pte_pages array puts in into the same
cache-line as the spin-lock and the bitmap array pointer.
This should safe a cache-miss.
Signed-off-by: Joerg Roedel <jroedel@suse.de>
There have been present PTEs which in theory could have made
it to the IOMMU TLB. Flush the addresses out on the error
path to make sure no stale entries remain.
Signed-off-by: Joerg Roedel <jroedel@suse.de>
If CONFIG_PHYS_ADDR_T_64BIT=n:
drivers/iommu/ipmmu-vmsa.c: In function 'ipmmu_domain_init_context':
drivers/iommu/ipmmu-vmsa.c:434:2: warning: right shift count >= width of type
ipmmu_ctx_write(domain, IMTTUBR0, ttbr >> 32);
^
As io_pgtable_cfg.arm_lpae_s1_cfg.ttbr[] is an array of u64s, assigning
it to a phys_addr_t may truncates it. Make ttbr u64 to fix this.
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Doug reports that the equivalent page allocator on 32-bit ARM exhibits
particularly pathalogical behaviour under memory pressure when
fragmentation is high, where allocating a 4MB buffer takes tens of
seconds and the number of calls to alloc_pages() is over 9000![1]
We can drastically improve that situation without losing the other
benefits of high-order allocations when they would succeed, by assuming
memory pressure is relatively constant over the course of an allocation,
and not retrying allocations at orders we know to have failed before.
This way, the best-case behaviour remains unchanged, and in the worst
case we should see at most a dozen or so (MAX_ORDER - 1) failed attempts
before falling back to single pages for the remainder of the buffer.
[1]:http://lists.infradead.org/pipermail/linux-arm-kernel/2015-December/394660.html
Reported-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
dma-iommu.c was naughtily relying on an implicit transitive #include of
linux/vmalloc.h, which is apparently not present on some architectures.
Add that, plus a couple more headers for other functions which are used
similarly.
Reported-by: kbuild test robot <lkp@intel.com>
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Those are stupid and code should use static_cpu_has_safe() or
boot_cpu_has() instead. Kill the least used and unused ones.
The remaining ones need more careful inspection before a conversion can
happen. On the TODO.
Signed-off-by: Borislav Petkov <bp@suse.de>
Link: http://lkml.kernel.org/r/1449481182-27541-4-git-send-email-bp@alien8.de
Cc: David Sterba <dsterba@suse.com>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Matt Mackall <mpm@selenic.com>
Cc: Chris Mason <clm@fb.com>
Cc: Josef Bacik <jbacik@fb.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
* Two similar fixes for the Intel and AMD IOMMU drivers to add
proper access checks before calling handle_mm_fault.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (GNU/Linux)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=qrAB
-----END PGP SIGNATURE-----
Merge tag 'iommu-fixes-v4.4-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu
Pull IOMMU fixes from Joerg Roedel:
"Two similar fixes for the Intel and AMD IOMMU drivers to add proper
access checks before calling handle_mm_fault"
* tag 'iommu-fixes-v4.4-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu:
iommu/vt-d: Do access checks before calling handle_mm_fault()
iommu/amd: Do proper access checking before calling handle_mm_fault()
When tearing down page tables, we return early for the final level
since we know that we won't have any table pointers to follow.
Unfortunately, this also means that we forget to free the final level,
so we end up leaking memory.
Fix the issue by always freeing the current level, but just don't bother
to iterate over the ptes if we're at the final level.
Cc: <stable@vger.kernel.org>
Reported-by: Zhang Bo <zhangbo_a@xiaomi.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
It is ILLEGAL to set STE.S1STALLD to 1 if stage 1 is enabled and
either the stall or terminate models are not supported.
This patch fixes the STALLD check and ensures that we don't set STALLD
in the STE when it is not supported.
Signed-off-by: Prem Mallappa <pmallapp@broadcom.com>
[will: consistently use IDR0_STALL_MODEL_* prefix]
Signed-off-by: Will Deacon <will.deacon@arm.com>
When acknowledging global errors, the GERRORN register should be written
with the original GERROR value so that active errors are toggled.
This patch fixed the driver to write the original GERROR value to
GERRORN, instead of an active error mask.
Signed-off-by: Prem Mallappa <pmallapp@broadcom.com>
[will: reworked use of active bits and fixed commit log]
Signed-off-by: Will Deacon <will.deacon@arm.com>
Joerg Roedel