[ Upstream commit 6d9870b7e5 ]
Add the IOMMU callback for DMA mapping API dma_opt_mapping_size(), which
allows the drivers to know the optimal mapping limit and thus limit the
requested IOVA lengths.
This value is based on the IOVA rcache range limit, as IOVAs allocated
above this limit must always be newly allocated, which may be quite slow.
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Acked-by: Robin Murphy <robin.murphy@arm.com>
Acked-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Stable-dep-of: afc5aa46ed ("iommu/dma: Force swiotlb_max_mapping_size on an untrusted device")
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit dcdb3ba7e2 upstream.
In __alloc_and_insert_iova_range, there is an issue that retry_pfn
overflows. The value of iovad->anchor.pfn_hi is ~0UL, then when
iovad->cached_node is iovad->anchor, curr_iova->pfn_hi + 1 will
overflow. As a result, if the retry logic is executed, low_pfn is
updated to 0, and then new_pfn < low_pfn returns false to make the
allocation successful.
This issue occurs in the following two situations:
1. The first iova size exceeds the domain size. When initializing
iova domain, iovad->cached_node is assigned as iovad->anchor. For
example, the iova domain size is 10M, start_pfn is 0x1_F000_0000,
and the iova size allocated for the first time is 11M. The
following is the log information, new->pfn_lo is smaller than
iovad->cached_node.
Example log as follows:
[ 223.798112][T1705487] sh: [name:iova&]__alloc_and_insert_iova_range
start_pfn:0x1f0000,retry_pfn:0x0,size:0xb00,limit_pfn:0x1f0a00
[ 223.799590][T1705487] sh: [name:iova&]__alloc_and_insert_iova_range
success start_pfn:0x1f0000,new->pfn_lo:0x1efe00,new->pfn_hi:0x1f08ff
2. The node with the largest iova->pfn_lo value in the iova domain
is deleted, iovad->cached_node will be updated to iovad->anchor,
and then the alloc iova size exceeds the maximum iova size that can
be allocated in the domain.
After judging that retry_pfn is less than limit_pfn, call retry_pfn+1
to fix the overflow issue.
Signed-off-by: jianjiao zeng <jianjiao.zeng@mediatek.com>
Signed-off-by: Yunfei Wang <yf.wang@mediatek.com>
Cc: <stable@vger.kernel.org> # 5.15.*
Fixes: 4e89dce725 ("iommu/iova: Retry from last rb tree node if iova search fails")
Acked-by: Robin Murphy <robin.murphy@arm.com>
Link: https://lore.kernel.org/r/20230111063801.25107-1-yf.wang@mediatek.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 5b61343b50 upstream.
For various reasons based on the allocator behaviour and typical
use-cases at the time, when the max32_alloc_size optimisation was
introduced it seemed reasonable to couple the reset of the tracked
size to the update of cached32_node upon freeing a relevant IOVA.
However, since subsequent optimisations focused on helping genuine
32-bit devices make best use of even more limited address spaces, it
is now a lot more likely for cached32_node to be anywhere in a "full"
32-bit address space, and as such more likely for space to become
available from IOVAs below that node being freed.
At this point, the short-cut in __cached_rbnode_delete_update() really
doesn't hold up any more, and we need to fix the logic to reliably
provide the expected behaviour. We still want cached32_node to only move
upwards, but we should reset the allocation size if *any* 32-bit space
has become available.
Reported-by: Yunfei Wang <yf.wang@mediatek.com>
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Reviewed-by: Miles Chen <miles.chen@mediatek.com>
Link: https://lore.kernel.org/r/033815732d83ca73b13c11485ac39336f15c3b40.1646318408.git.robin.murphy@arm.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Cc: Miles Chen <miles.chen@mediatek.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit d7061627d7 ]
It turns out to be possible for hotplugging out a device to reach the
stage of tearing down the device's group and default domain before the
domain's flush queue has drained naturally. At this point, it is then
possible for the timeout to expire just before the del_timer() call
in free_iova_flush_queue(), such that we then proceed to free the FQ
resources while fq_flush_timeout() is still accessing them on another
CPU. Crashes due to this have been observed in the wild while removing
NVMe devices.
Close the race window by using del_timer_sync() to safely wait for any
active timeout handler to finish before we start to free things. We
already avoid any locking in free_iova_flush_queue() since the FQ is
supposed to be inactive anyway, so the potential deadlock scenario does
not apply.
Fixes: 9a005a800a ("iommu/iova: Add flush timer")
Reviewed-by: John Garry <john.garry@huawei.com>
Signed-off-by: Xiongfeng Wang <wangxiongfeng2@huawei.com>
[ rm: rewrite commit message ]
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Link: https://lore.kernel.org/r/0a365e5b07f14b7344677ad6a9a734966a8422ce.1639753638.git.robin.murphy@arm.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>
vduse driver supporting blk
virtio-vsock support for end of record with SEQPACKET
vdpa: mac and mq support for ifcvf and mlx5
vdpa: management netlink for ifcvf
virtio-i2c, gpio dt bindings
misc fixes, cleanups
NB: when merging this with
b542e383d8 ("eventfd: Make signal recursion protection a task bit")
from Linus' tree, replace eventfd_signal_count with
eventfd_signal_allowed, and drop the export of eventfd_wake_count from
("eventfd: Export eventfd_wake_count to modules").
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
-----BEGIN PGP SIGNATURE-----
iQFDBAABCAAtFiEEXQn9CHHI+FuUyooNKB8NuNKNVGkFAmE1+awPHG1zdEByZWRo
YXQuY29tAAoJECgfDbjSjVRpt6EIAJy0qrc62lktNA0IiIVJSLbUbTMmFj8MzkGR
8UxZdhpjWqBPJPyaOuNeksAqTGm/UAPEYx3C2c95Jhej7anFpy7dbCtIXcPHLJME
DjcJg+EDrlNCj8m0FcsHpHWsFzPMERJpyEZNxgB5WazQbv+yWhGrg2FN5DCnF0Ro
ZFYeKSVty148pQ0nHl8X0JM2XMtqit+O+LvKN2HQZ+fubh7BCzMxzkHY0QLHIzUS
UeZqd3Qm8YcbqnlX38P5D6k+NPiTEgknmxaBLkPxg6H3XxDAmaIRFb8Ldd1rsgy1
zTLGDiSGpVDIpawRnuEAzqJThV3Y5/MVJ1WD+mDYQ96tmhfp+KY=
=DBH/
-----END PGP SIGNATURE-----
Merge tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost
Pull virtio updates from Michael Tsirkin:
- vduse driver ("vDPA Device in Userspace") supporting emulated virtio
block devices
- virtio-vsock support for end of record with SEQPACKET
- vdpa: mac and mq support for ifcvf and mlx5
- vdpa: management netlink for ifcvf
- virtio-i2c, gpio dt bindings
- misc fixes and cleanups
* tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost: (39 commits)
Documentation: Add documentation for VDUSE
vduse: Introduce VDUSE - vDPA Device in Userspace
vduse: Implement an MMU-based software IOTLB
vdpa: Support transferring virtual addressing during DMA mapping
vdpa: factor out vhost_vdpa_pa_map() and vhost_vdpa_pa_unmap()
vdpa: Add an opaque pointer for vdpa_config_ops.dma_map()
vhost-iotlb: Add an opaque pointer for vhost IOTLB
vhost-vdpa: Handle the failure of vdpa_reset()
vdpa: Add reset callback in vdpa_config_ops
vdpa: Fix some coding style issues
file: Export receive_fd() to modules
eventfd: Export eventfd_wake_count to modules
iova: Export alloc_iova_fast() and free_iova_fast()
virtio-blk: remove unneeded "likely" statements
virtio-balloon: Use virtio_find_vqs() helper
vdpa: Make use of PFN_PHYS/PFN_UP/PFN_DOWN helper macro
vsock_test: update message bounds test for MSG_EOR
af_vsock: rename variables in receive loop
virtio/vsock: support MSG_EOR bit processing
vhost/vsock: support MSG_EOR bit processing
...
Export alloc_iova_fast() and free_iova_fast() so that
some modules can make use of the per-CPU cache to get
rid of rbtree spinlock in alloc_iova() and free_iova()
during IOVA allocation.
Signed-off-by: Xie Yongji <xieyongji@bytedance.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Acked-by: Will Deacon <will@kernel.org>
Link: https://lore.kernel.org/r/20210831103634.33-2-xieyongji@bytedance.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Allocating and enabling a flush queue is in fact something we can
reasonably do while a DMA domain is active, without having to rebuild it
from scratch. Thus we can allow a strict -> non-strict transition from
sysfs without requiring to unbind the device's driver, which is of
particular interest to users who want to make selective relaxations to
critical devices like the one serving their root filesystem.
Disabling and draining a queue also seems technically possible to
achieve without rebuilding the whole domain, but would certainly be more
involved. Furthermore there's not such a clear use-case for tightening
up security *after* the device may already have done whatever it is that
you don't trust it not to do, so we only consider the relaxation case.
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Link: https://lore.kernel.org/r/d652966348c78457c38bf18daf369272a4ebc2c9.1628682049.git.robin.murphy@arm.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Since iommu_iotlb_gather exists to help drivers optimise flushing for a
given unmap request, it is also the logical place to indicate whether
the unmap is strict or not, and thus help them further optimise for
whether to expect a sync or a flush_all subsequently. As part of that,
it also seems fair to make the flush queue code take responsibility for
enforcing the really subtle ordering requirement it brings, so that we
don't need to worry about forgetting that if new drivers want to add
flush queue support, and can consolidate the existing versions.
While we're adding to the kerneldoc, also fill in some info for
@freelist which was overlooked previously.
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Link: https://lore.kernel.org/r/bf5f8e2ad84e48c712ccbf80fa8c610594c7595f.1628682049.git.robin.murphy@arm.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
It is not necessary to put free_iova_mem() inside of spinlock/unlock
iova_rbtree_lock which only leads to more completion for the spinlock.
It has a small promote on the performance after the change. And also
rename private_free_iova() as remove_iova() because the function will not
free iova after that change.
Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Reviewed-by: John Garry <john.garry@huawei.com>
Acked-by: Robin Murphy <robin.murphy@arm.com>
Link: https://lore.kernel.org/r/1620647582-194621-1-git-send-email-chenxiang66@hisilicon.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Function free_iova_fast() is only referenced by dma-iommu.c, which can
only be in-built, so stop exporting it.
This was missed in an earlier tidy-up patch.
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Link: https://lore.kernel.org/r/1616675401-151997-5-git-send-email-john.garry@huawei.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Function iommu_dma_free_cpu_cached_iovas() no longer has any caller, so
delete it.
With that, function free_cpu_cached_iovas() may be made static.
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Link: https://lore.kernel.org/r/1616675401-151997-4-git-send-email-john.garry@huawei.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Like the Intel IOMMU driver already does, flush the per-IOVA domain
CPU rcache when a CPU goes offline - there's no point in keeping it.
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Link: https://lore.kernel.org/r/1616675401-151997-2-git-send-email-john.garry@huawei.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
When restarting after searching below the cached node fails, resetting
the start point to the anchor node is often overly pessimistic. If
allocations are made with mixed limits - particularly in the case of the
opportunistic 32-bit allocation for PCI devices - this could mean
significant time wasted walking through the whole populated upper range
just to reach the initial limit. We can improve on that by implementing
a proper tree traversal to find the first node above the relevant limit,
and set the exact start point.
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Link: https://lore.kernel.org/r/076b3484d1e5057b95d8c387c894bd6ad2514043.1614962123.git.robin.murphy@arm.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
The following functions are not referenced outside dma-iommu.c (and
iova.c), which can only be built-in:
- init_iova_flush_queue()
- free_iova_fast()
- queue_iova()
- alloc_iova_fast()
So stop exporting them.
Signed-off-by: John Garry <john.garry@huawei.com>
Acked-by: Will Deacon <will@kernel.org>
Link: https://lore.kernel.org/r/1609940111-28563-4-git-send-email-john.garry@huawei.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Since commit c588072bba ("iommu/vt-d: Convert intel iommu driver to the
iommu ops"), function copy_reserved_iova() is not referenced, so delete
it.
Signed-off-by: John Garry <john.garry@huawei.com>
Acked-by: Will Deacon <will@kernel.org>
Link: https://lore.kernel.org/r/1609940111-28563-3-git-send-email-john.garry@huawei.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Both find_iova() and __free_iova() take iova_rbtree_lock,
there is no reason to take and release it twice inside
free_iova().
Fold them into one critical section by calling the unlock
versions instead.
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Link: https://lore.kernel.org/r/1605608734-84416-5-git-send-email-john.garry@huawei.com
Signed-off-by: Will Deacon <will@kernel.org>
When ever an iova alloc request fails we free the iova
ranges present in the percpu iova rcaches and then retry
but the global iova rcache is not freed as a result we could
still see iova alloc failure even after retry as global
rcache is holding the iova's which can cause fragmentation.
So, free the global iova rcache as well and then go for the
retry.
Signed-off-by: Vijayanand Jitta <vjitta@codeaurora.org>
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Acked-by: John Garry <john.garry@huaqwei.com>
Link: https://lore.kernel.org/r/1601451864-5956-2-git-send-email-vjitta@codeaurora.org
Signed-off-by: Will Deacon <will@kernel.org>
When ever a new iova alloc request comes iova is always searched
from the cached node and the nodes which are previous to cached
node. So, even if there is free iova space available in the nodes
which are next to the cached node iova allocation can still fail
because of this approach.
Consider the following sequence of iova alloc and frees on
1GB of iova space
1) alloc - 500MB
2) alloc - 12MB
3) alloc - 499MB
4) free - 12MB which was allocated in step 2
5) alloc - 13MB
After the above sequence we will have 12MB of free iova space and
cached node will be pointing to the iova pfn of last alloc of 13MB
which will be the lowest iova pfn of that iova space. Now if we get an
alloc request of 2MB we just search from cached node and then look
for lower iova pfn's for free iova and as they aren't any, iova alloc
fails though there is 12MB of free iova space.
To avoid such iova search failures do a retry from the last rb tree node
when iova search fails, this will search the entire tree and get an iova
if its available.
Signed-off-by: Vijayanand Jitta <vjitta@codeaurora.org>
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Link: https://lore.kernel.org/r/1601451864-5956-1-git-send-email-vjitta@codeaurora.org
Signed-off-by: Will Deacon <will@kernel.org>
The performance of the atomic_xchg is better than atomic_cmpxchg because
no comparison is required. While the value of @fq_timer_on can only be 0
or 1. Let's use atomic_xchg instead of atomic_cmpxchg here because we
only need to check that the value changes from 0 to 1 or from 1 to 1.
Signed-off-by: Yuqi Jin <jinyuqi@huawei.com>
Signed-off-by: Shaokun Zhang <zhangshaokun@hisilicon.com>
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Cc: Joerg Roedel <joro@8bytes.org>
Link: https://lore.kernel.org/r/1598517834-30275-1-git-send-email-zhangshaokun@hisilicon.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Unlike the other instances which represent a complete loss of
consistency within the rcache mechanism itself, or a fundamental
and obvious misconfiguration by an IOMMU driver, the BUG_ON() in
iova_magazine_free_pfns() can be provoked at more or less any time
in a "spooky action-at-a-distance" manner by any old device driver
passing nonsense to dma_unmap_*() which then propagates through to
queue_iova().
Not only is this well outside the IOVA layer's control, it's also
nowhere near fatal enough to justify panicking anyway - all that
really achieves is to make debugging the offending driver more
difficult. Let's simply WARN and otherwise ignore bogus PFNs.
Reported-by: Prakash Gupta <guptap@codeaurora.org>
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Reviewed-by: Prakash Gupta <guptap@codeaurora.org>
Link: https://lore.kernel.org/r/acbd2d092b42738a03a21b417ce64e27f8c91c86.1591103298.git.robin.murphy@arm.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
In commit 14bd9a607f ("iommu/iova: Separate atomic variables
to improve performance") Jinyu Qi identified that the atomic_cmpxchg()
in queue_iova() was causing a performance loss and moved critical fields
so that the false sharing would not impact them.
However, avoiding the false sharing in the first place seems easy.
We should attempt the atomic_cmpxchg() no more than 100 times
per second. Adding an atomic_read() will keep the cache
line mostly shared.
This false sharing came with commit 9a005a800a
("iommu/iova: Add flush timer").
Signed-off-by: Eric Dumazet <edumazet@google.com>
Fixes: 9a005a800a ('iommu/iova: Add flush timer')
Cc: Jinyu Qi <jinyuqi@huawei.com>
Cc: Joerg Roedel <jroedel@suse.de>
Acked-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Since the cached32_node is allowed to be advanced above dma_32bit_pfn
(to provide a shortcut into the limited range), we need to be careful to
remove the to be freed node if it is the cached32_node.
[ 48.477773] BUG: KASAN: use-after-free in __cached_rbnode_delete_update+0x68/0x110
[ 48.477812] Read of size 8 at addr ffff88870fc19020 by task kworker/u8:1/37
[ 48.477843]
[ 48.477879] CPU: 1 PID: 37 Comm: kworker/u8:1 Tainted: G U 5.2.0+ #735
[ 48.477915] Hardware name: Intel Corporation NUC7i5BNK/NUC7i5BNB, BIOS BNKBL357.86A.0052.2017.0918.1346 09/18/2017
[ 48.478047] Workqueue: i915 __i915_gem_free_work [i915]
[ 48.478075] Call Trace:
[ 48.478111] dump_stack+0x5b/0x90
[ 48.478137] print_address_description+0x67/0x237
[ 48.478178] ? __cached_rbnode_delete_update+0x68/0x110
[ 48.478212] __kasan_report.cold.3+0x1c/0x38
[ 48.478240] ? __cached_rbnode_delete_update+0x68/0x110
[ 48.478280] ? __cached_rbnode_delete_update+0x68/0x110
[ 48.478308] __cached_rbnode_delete_update+0x68/0x110
[ 48.478344] private_free_iova+0x2b/0x60
[ 48.478378] iova_magazine_free_pfns+0x46/0xa0
[ 48.478403] free_iova_fast+0x277/0x340
[ 48.478443] fq_ring_free+0x15a/0x1a0
[ 48.478473] queue_iova+0x19c/0x1f0
[ 48.478597] cleanup_page_dma.isra.64+0x62/0xb0 [i915]
[ 48.478712] __gen8_ppgtt_cleanup+0x63/0x80 [i915]
[ 48.478826] __gen8_ppgtt_cleanup+0x42/0x80 [i915]
[ 48.478940] __gen8_ppgtt_clear+0x433/0x4b0 [i915]
[ 48.479053] __gen8_ppgtt_clear+0x462/0x4b0 [i915]
[ 48.479081] ? __sg_free_table+0x9e/0xf0
[ 48.479116] ? kfree+0x7f/0x150
[ 48.479234] i915_vma_unbind+0x1e2/0x240 [i915]
[ 48.479352] i915_vma_destroy+0x3a/0x280 [i915]
[ 48.479465] __i915_gem_free_objects+0xf0/0x2d0 [i915]
[ 48.479579] __i915_gem_free_work+0x41/0xa0 [i915]
[ 48.479607] process_one_work+0x495/0x710
[ 48.479642] worker_thread+0x4c7/0x6f0
[ 48.479687] ? process_one_work+0x710/0x710
[ 48.479724] kthread+0x1b2/0x1d0
[ 48.479774] ? kthread_create_worker_on_cpu+0xa0/0xa0
[ 48.479820] ret_from_fork+0x1f/0x30
[ 48.479864]
[ 48.479907] Allocated by task 631:
[ 48.479944] save_stack+0x19/0x80
[ 48.479994] __kasan_kmalloc.constprop.6+0xc1/0xd0
[ 48.480038] kmem_cache_alloc+0x91/0xf0
[ 48.480082] alloc_iova+0x2b/0x1e0
[ 48.480125] alloc_iova_fast+0x58/0x376
[ 48.480166] intel_alloc_iova+0x90/0xc0
[ 48.480214] intel_map_sg+0xde/0x1f0
[ 48.480343] i915_gem_gtt_prepare_pages+0xb8/0x170 [i915]
[ 48.480465] huge_get_pages+0x232/0x2b0 [i915]
[ 48.480590] ____i915_gem_object_get_pages+0x40/0xb0 [i915]
[ 48.480712] __i915_gem_object_get_pages+0x90/0xa0 [i915]
[ 48.480834] i915_gem_object_prepare_write+0x2d6/0x330 [i915]
[ 48.480955] create_test_object.isra.54+0x1a9/0x3e0 [i915]
[ 48.481075] igt_shared_ctx_exec+0x365/0x3c0 [i915]
[ 48.481210] __i915_subtests.cold.4+0x30/0x92 [i915]
[ 48.481341] __run_selftests.cold.3+0xa9/0x119 [i915]
[ 48.481466] i915_live_selftests+0x3c/0x70 [i915]
[ 48.481583] i915_pci_probe+0xe7/0x220 [i915]
[ 48.481620] pci_device_probe+0xe0/0x180
[ 48.481665] really_probe+0x163/0x4e0
[ 48.481710] device_driver_attach+0x85/0x90
[ 48.481750] __driver_attach+0xa5/0x180
[ 48.481796] bus_for_each_dev+0xda/0x130
[ 48.481831] bus_add_driver+0x205/0x2e0
[ 48.481882] driver_register+0xca/0x140
[ 48.481927] do_one_initcall+0x6c/0x1af
[ 48.481970] do_init_module+0x106/0x350
[ 48.482010] load_module+0x3d2c/0x3ea0
[ 48.482058] __do_sys_finit_module+0x110/0x180
[ 48.482102] do_syscall_64+0x62/0x1f0
[ 48.482147] entry_SYSCALL_64_after_hwframe+0x44/0xa9
[ 48.482190]
[ 48.482224] Freed by task 37:
[ 48.482273] save_stack+0x19/0x80
[ 48.482318] __kasan_slab_free+0x12e/0x180
[ 48.482363] kmem_cache_free+0x70/0x140
[ 48.482406] __free_iova+0x1d/0x30
[ 48.482445] fq_ring_free+0x15a/0x1a0
[ 48.482490] queue_iova+0x19c/0x1f0
[ 48.482624] cleanup_page_dma.isra.64+0x62/0xb0 [i915]
[ 48.482749] __gen8_ppgtt_cleanup+0x63/0x80 [i915]
[ 48.482873] __gen8_ppgtt_cleanup+0x42/0x80 [i915]
[ 48.482999] __gen8_ppgtt_clear+0x433/0x4b0 [i915]
[ 48.483123] __gen8_ppgtt_clear+0x462/0x4b0 [i915]
[ 48.483250] i915_vma_unbind+0x1e2/0x240 [i915]
[ 48.483378] i915_vma_destroy+0x3a/0x280 [i915]
[ 48.483500] __i915_gem_free_objects+0xf0/0x2d0 [i915]
[ 48.483622] __i915_gem_free_work+0x41/0xa0 [i915]
[ 48.483659] process_one_work+0x495/0x710
[ 48.483704] worker_thread+0x4c7/0x6f0
[ 48.483748] kthread+0x1b2/0x1d0
[ 48.483787] ret_from_fork+0x1f/0x30
[ 48.483831]
[ 48.483868] The buggy address belongs to the object at ffff88870fc19000
[ 48.483868] which belongs to the cache iommu_iova of size 40
[ 48.483920] The buggy address is located 32 bytes inside of
[ 48.483920] 40-byte region [ffff88870fc19000, ffff88870fc19028)
[ 48.483964] The buggy address belongs to the page:
[ 48.484006] page:ffffea001c3f0600 refcount:1 mapcount:0 mapping:ffff8888181a91c0 index:0x0 compound_mapcount: 0
[ 48.484045] flags: 0x8000000000010200(slab|head)
[ 48.484096] raw: 8000000000010200 ffffea001c421a08 ffffea001c447e88 ffff8888181a91c0
[ 48.484141] raw: 0000000000000000 0000000000120012 00000001ffffffff 0000000000000000
[ 48.484188] page dumped because: kasan: bad access detected
[ 48.484230]
[ 48.484265] Memory state around the buggy address:
[ 48.484314] ffff88870fc18f00: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
[ 48.484361] ffff88870fc18f80: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
[ 48.484406] >ffff88870fc19000: fb fb fb fb fb fc fc fc fc fc fc fc fc fc fc fc
[ 48.484451] ^
[ 48.484494] ffff88870fc19080: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
[ 48.484530] ffff88870fc19100: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=108602
Fixes: e60aa7b538 ("iommu/iova: Extend rbtree node caching")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Robin Murphy <robin.murphy@arm.com>
Cc: Joerg Roedel <jroedel@suse.de>
Cc: Joerg Roedel <joro@8bytes.org>
Cc: <stable@vger.kernel.org> # v4.15+
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Intel VT-d driver was reworked to use common deferred flushing
implementation. Previously there was one global per-cpu flush queue,
afterwards - one per domain.
Before deferring a flush, the queue should be allocated and initialized.
Currently only domains with IOMMU_DOMAIN_DMA type initialize their flush
queue. It's probably worth to init it for static or unmanaged domains
too, but it may be arguable - I'm leaving it to iommu folks.
Prevent queuing an iova flush if the domain doesn't have a queue.
The defensive check seems to be worth to keep even if queue would be
initialized for all kinds of domains. And is easy backportable.
On 4.19.43 stable kernel it has a user-visible effect: previously for
devices in si domain there were crashes, on sata devices:
BUG: spinlock bad magic on CPU#6, swapper/0/1
lock: 0xffff88844f582008, .magic: 00000000, .owner: <none>/-1, .owner_cpu: 0
CPU: 6 PID: 1 Comm: swapper/0 Not tainted 4.19.43 #1
Call Trace:
<IRQ>
dump_stack+0x61/0x7e
spin_bug+0x9d/0xa3
do_raw_spin_lock+0x22/0x8e
_raw_spin_lock_irqsave+0x32/0x3a
queue_iova+0x45/0x115
intel_unmap+0x107/0x113
intel_unmap_sg+0x6b/0x76
__ata_qc_complete+0x7f/0x103
ata_qc_complete+0x9b/0x26a
ata_qc_complete_multiple+0xd0/0xe3
ahci_handle_port_interrupt+0x3ee/0x48a
ahci_handle_port_intr+0x73/0xa9
ahci_single_level_irq_intr+0x40/0x60
__handle_irq_event_percpu+0x7f/0x19a
handle_irq_event_percpu+0x32/0x72
handle_irq_event+0x38/0x56
handle_edge_irq+0x102/0x121
handle_irq+0x147/0x15c
do_IRQ+0x66/0xf2
common_interrupt+0xf/0xf
RIP: 0010:__do_softirq+0x8c/0x2df
The same for usb devices that use ehci-pci:
BUG: spinlock bad magic on CPU#0, swapper/0/1
lock: 0xffff88844f402008, .magic: 00000000, .owner: <none>/-1, .owner_cpu: 0
CPU: 0 PID: 1 Comm: swapper/0 Not tainted 4.19.43 #4
Call Trace:
<IRQ>
dump_stack+0x61/0x7e
spin_bug+0x9d/0xa3
do_raw_spin_lock+0x22/0x8e
_raw_spin_lock_irqsave+0x32/0x3a
queue_iova+0x77/0x145
intel_unmap+0x107/0x113
intel_unmap_page+0xe/0x10
usb_hcd_unmap_urb_setup_for_dma+0x53/0x9d
usb_hcd_unmap_urb_for_dma+0x17/0x100
unmap_urb_for_dma+0x22/0x24
__usb_hcd_giveback_urb+0x51/0xc3
usb_giveback_urb_bh+0x97/0xde
tasklet_action_common.isra.4+0x5f/0xa1
tasklet_action+0x2d/0x30
__do_softirq+0x138/0x2df
irq_exit+0x7d/0x8b
smp_apic_timer_interrupt+0x10f/0x151
apic_timer_interrupt+0xf/0x20
</IRQ>
RIP: 0010:_raw_spin_unlock_irqrestore+0x17/0x39
Cc: David Woodhouse <dwmw2@infradead.org>
Cc: Joerg Roedel <joro@8bytes.org>
Cc: Lu Baolu <baolu.lu@linux.intel.com>
Cc: iommu@lists.linux-foundation.org
Cc: <stable@vger.kernel.org> # 4.14+
Fixes: 13cf017446 ("iommu/vt-d: Make use of iova deferred flushing")
Signed-off-by: Dmitry Safonov <dima@arista.com>
Reviewed-by: Lu Baolu <baolu.lu@linux.intel.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Based on 1 normalized pattern(s):
this program is free software you can redistribute it and or modify
it under the terms and conditions of the gnu general public license
version 2 as published by the free software foundation this program
is distributed in the hope it will be useful but without any
warranty without even the implied warranty of merchantability or
fitness for a particular purpose see the gnu general public license
for more details you should have received a copy of the gnu general
public license along with this program if not write to the free
software foundation inc 59 temple place suite 330 boston ma 02111
1307 usa
extracted by the scancode license scanner the SPDX license identifier
GPL-2.0-only
has been chosen to replace the boilerplate/reference in 33 file(s).
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Allison Randal <allison@lohutok.net>
Reviewed-by: Kate Stewart <kstewart@linuxfoundation.org>
Reviewed-by: Alexios Zavras <alexios.zavras@intel.com>
Cc: linux-spdx@vger.kernel.org
Link: https://lkml.kernel.org/r/20190530000435.254582722@linutronix.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If a 32 bit allocation request is too big to possibly succeed, it
early exits with a failure and then should never update max32_alloc_
size. This patch fixes current code, now the size is only updated if
the slow path failed while walking the tree. Without the fix the
allocation may enter the slow path again even if there was a failure
before of a request with the same or a smaller size.
Cc: <stable@vger.kernel.org> # 4.20+
Fixes: bee60e94a1 ("iommu/iova: Optimise attempts to allocate iova from 32bit address range")
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Robert Richter <rrichter@marvell.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
As an optimisation for PCI devices, there is always first attempt
been made to allocate iova from SAC address range. This will lead
to unnecessary attempts, when there are no free ranges
available. Adding fix to track recently failed iova address size and
allow further attempts, only if requested size is lesser than a failed
size. The size is updated when any replenish happens.
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Ganapatrao Kulkarni <ganapatrao.kulkarni@cavium.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
get_cpu_ptr() disabled preemption and returns the ->fq object of the
current CPU. raw_cpu_ptr() does the same except that it not disable
preemption which means the scheduler can move it to another CPU after it
obtained the per-CPU object.
In this case this is not bad because the data structure itself is
protected with a spin_lock. This change shouldn't matter however on RT
it does because the sleeping lock can't be accessed with disabled
preemption.
Cc: Joerg Roedel <joro@8bytes.org>
Cc: iommu@lists.linux-foundation.org
Reported-by: vinadhy@gmail.com
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Since IOVA allocation failure is not unusual case we need to flush
CPUs' rcache in hope we will succeed in next round.
However, it is useful to decide whether we need rcache flush step because
of two reasons:
- Not scalability. On large system with ~100 CPUs iterating and flushing
rcache for each CPU becomes serious bottleneck so we may want to defer it.
- free_cpu_cached_iovas() does not care about max PFN we are interested in.
Thus we may flush our rcaches and still get no new IOVA like in the
commonly used scenario:
if (dma_limit > DMA_BIT_MASK(32) && dev_is_pci(dev))
iova = alloc_iova_fast(iovad, iova_len, DMA_BIT_MASK(32) >> shift);
if (!iova)
iova = alloc_iova_fast(iovad, iova_len, dma_limit >> shift);
1. First alloc_iova_fast() call is limited to DMA_BIT_MASK(32) to get
PCI devices a SAC address
2. alloc_iova() fails due to full 32-bit space
3. rcaches contain PFNs out of 32-bit space so free_cpu_cached_iovas()
throws entries away for nothing and alloc_iova() fails again
4. Next alloc_iova_fast() call cannot take advantage of rcache since we
have just defeated caches. In this case we pick the slowest option
to proceed.
This patch reworks flushed_rcache local flag to be additional function
argument instead and control rcache flush step. Also, it updates all users
to do the flush as the last chance.
Signed-off-by: Tomasz Nowicki <Tomasz.Nowicki@caviumnetworks.com>
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Tested-by: Nate Watterson <nwatters@codeaurora.org>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Anchor nodes are not reserved IOVAs in the way that copy_reserved_iova()
cares about - while the failure from reserve_iova() is benign since the
target domain will already have its own anchor, we still don't want to
be triggering spurious warnings.
Reported-by: kernel test robot <fengguang.wu@intel.com>
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Fixes: bb68b2fbfb ('iommu/iova: Add rbtree anchor node')
Signed-off-by: Joerg Roedel <jroedel@suse.de>
When devices with different DMA masks are using the same domain, or for
PCI devices where we usually try a speculative 32-bit allocation first,
there is a fair possibility that the top PFN of the rcache stack at any
given time may be unsuitable for the lower limit, prompting a fallback
to allocating anew from the rbtree. Consequently, we may end up
artifically increasing pressure on the 32-bit IOVA space as unused IOVAs
accumulate lower down in the rcache stacks, while callers with 32-bit
masks also impose unnecessary rbtree overhead.
In such cases, let's try a bit harder to satisfy the allocation locally
first - scanning the whole stack should still be relatively inexpensive.
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
When popping a pfn from an rcache, we are currently checking it directly
against limit_pfn for viability. Since this represents iova->pfn_lo, it
is technically possible for the corresponding iova->pfn_hi to be greater
than limit_pfn. Although we generally get away with it in practice since
limit_pfn is typically a power-of-two boundary and the IOVAs are
size-aligned, it's pretty trivial to make the iova_rcache_get() path
take the allocation size into account for complete safety.
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
All put_iova_domain() should have to worry about is freeing memory - by
that point the domain must no longer be live, so the act of cleaning up
doesn't need to be concurrency-safe or maintain the rbtree in a
self-consistent state. There's no need to waste time with locking or
emptying the rcache magazines, and we can just use the postorder
traversal helper to clear out the remaining rbtree entries in-place.
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
The logic of __get_cached_rbnode() is a little obtuse, but then
__get_prev_node_of_cached_rbnode_or_last_node_and_update_limit_pfn()
wouldn't exactly roll off the tongue...
Now that we have the invariant that there is always a valid node to
start searching downwards from, everything gets a bit easier to follow
if we simplify that function to do what it says on the tin and return
the cached node (or anchor node as appropriate) directly. In turn, we
can then deduplicate the rb_prev() and limit_pfn logic into the main
loop itself, further reduce the amount of code under the lock, and
generally make the inner workings a bit less subtle.
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Add a permanent dummy IOVA reservation to the rbtree, such that we can
always access the top of the address space instantly. The immediate
benefit is that we remove the overhead of the rb_last() traversal when
not using the cached node, but it also paves the way for further
simplifications.
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Now that the cached node optimisation can apply to all allocations, the
couple of users which were playing tricks with dma_32bit_pfn in order to
benefit from it can stop doing so. Conversely, there is also no need for
all the other users to explicitly calculate a 'real' 32-bit PFN, when
init_iova_domain() can happily do that itself from the page granularity.
CC: Thierry Reding <thierry.reding@gmail.com>
CC: Jonathan Hunter <jonathanh@nvidia.com>
CC: David Airlie <airlied@linux.ie>
CC: Sudeep Dutt <sudeep.dutt@intel.com>
CC: Ashutosh Dixit <ashutosh.dixit@intel.com>
Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Tested-by: Zhen Lei <thunder.leizhen@huawei.com>
Tested-by: Nate Watterson <nwatters@codeaurora.org>
[rm: use iova_shift(), rewrote commit message]
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
The cached node mechanism provides a significant performance benefit for
allocations using a 32-bit DMA mask, but in the case of non-PCI devices
or where the 32-bit space is full, the loss of this benefit can be
significant - on large systems there can be many thousands of entries in
the tree, such that walking all the way down to find free space every
time becomes increasingly awful.
Maintain a similar cached node for the whole IOVA space as a superset of
the 32-bit space so that performance can remain much more consistent.
Inspired by work by Zhen Lei <thunder.leizhen@huawei.com>.
Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Tested-by: Zhen Lei <thunder.leizhen@huawei.com>
Tested-by: Nate Watterson <nwatters@codeaurora.org>
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
The mask for calculating the padding size doesn't change, so there's no
need to recalculate it every loop iteration. Furthermore, Once we've
done that, it becomes clear that we don't actually need to calculate a
padding size at all - by flipping the arithmetic around, we can just
combine the upper limit, size, and mask directly to check against the
lower limit.
For an arm64 build, this alone knocks 20% off the object code size of
the entire alloc_iova() function!
Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Tested-by: Zhen Lei <thunder.leizhen@huawei.com>
Tested-by: Nate Watterson <nwatters@codeaurora.org>
[rm: simplified more of the arithmetic, rewrote commit message]
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Checking the IOVA bounds separately before deciding which direction to
continue the search (if necessary) results in redundantly comparing both
pfns twice each. GCC can already determine that the final comparison op
is redundant and optimise it down to 3 in total, but we can go one
further with a little tweak of the ordering (which makes the intent of
the code that much cleaner as a bonus).
Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Tested-by: Zhen Lei <thunder.leizhen@huawei.com>
Tested-by: Nate Watterson <nwatters@codeaurora.org>
[rm: rewrote commit message to clarify]
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Add a timer to flush entries from the Flush-Queues every
10ms. This makes sure that no stale TLB entries remain for
too long after an IOVA has been unmapped.
Signed-off-by: Joerg Roedel <jroedel@suse.de>
The lock is taken from the same CPU most of the time. But
having it allows to flush the queue also from another CPU if
necessary.
This will be used by a timer to regularily flush any pending
IOVAs from the Flush-Queues.
Signed-off-by: Joerg Roedel <jroedel@suse.de>
John Garry