This reverts commit 009a82f643.
The ability to optimise here relies on compiler being able to optimise
away tail calls to avoid stack overflows. Unfortunately, we are seeing
reports of problems, so let's just revert.
Reported-by: Daniel Mack <[email protected]>
Signed-off-by: Trond Myklebust <[email protected]>
According to the NFSv4.2 spec if the input and output file is the
same file, operation should fail with EINVAL. However, linux
copy_file_range() system call has no such restrictions. Therefore,
in such case let's return EOPNOTSUPP and allow VFS to fallback
to doing do_splice_direct(). Also when copy_file_range is called
on an NFSv4.0 or 4.1 mount (ie., a server that doesn't support
COPY functionality), we also need to return EOPNOTSUPP and
fallback to a regular copy.
Fixes xfstest generic/075, generic/091, generic/112, generic/263
for all NFSv4.x versions.
Signed-off-by: Olga Kornievskaia <[email protected]>
Signed-off-by: Trond Myklebust <[email protected]>
We want to drain only the RQ first. Otherwise the transport can
deadlock on ->close if there are outstanding Send completions.
Fixes: 6d2d0ee27c ("xprtrdma: Replace rpcrdma_receive_wq ... ")
Signed-off-by: Chuck Lever <[email protected]>
Cc: [email protected] # v5.0+
Signed-off-by: Trond Myklebust <[email protected]>
NFSv4 GETACL and FS_LOCATIONS requests stopped working in v5.1-rc.
These two need the extra padding to be added directly to the reply
length.
Reported-by: Olga Kornievskaia <[email protected]>
Fixes: 02ef04e432 ("NFS: Account for XDR pad of buf->pages")
Signed-off-by: Chuck Lever <[email protected]>
Tested-by: Olga Kornievskaia <[email protected]>
Signed-off-by: Trond Myklebust <[email protected]>
syzbot is reporting uninitialized value at rpc_sockaddr2uaddr() [1]. This
is because syzbot is setting AF_INET6 to "struct sockaddr_in"->sin_family
(which is embedded into user-visible "struct nfs_mount_data" structure)
despite nfs23_validate_mount_data() cannot pass sizeof(struct sockaddr_in6)
bytes of AF_INET6 address to rpc_sockaddr2uaddr().
Since "struct nfs_mount_data" structure is user-visible, we can't change
"struct nfs_mount_data" to use "struct sockaddr_storage". Therefore,
assuming that everybody is using AF_INET family when passing address via
"struct nfs_mount_data"->addr, reject if its sin_family is not AF_INET.
[1] https://syzkaller.appspot.com/bug?id=599993614e7cbbf66bc2656a919ab2a95fb5d75c
Reported-by: syzbot <[email protected]>
Signed-off-by: Tetsuo Handa <[email protected]>
Signed-off-by: Trond Myklebust <[email protected]>
With skip set to 1, I get a traceback like this:
[ 106.867637] DMA-API: Mapped at:
[ 106.870784] afu_dma_map_region+0x2cd/0x4f0 [dfl_afu]
[ 106.875839] afu_ioctl+0x258/0x380 [dfl_afu]
[ 106.880108] do_vfs_ioctl+0xa9/0x720
[ 106.883688] ksys_ioctl+0x60/0x90
[ 106.887007] __x64_sys_ioctl+0x16/0x20
With the previous value of 2, afu_dma_map_region was being omitted. I
suspect that the code paths have simply changed since the value of 2 was
chosen a decade ago, but it's also possible that it varies based on which
mapping function was used, compiler inlining choices, etc. In any case,
it's best to err on the side of skipping less.
Signed-off-by: Scott Wood <[email protected]>
Signed-off-by: Christoph Hellwig <[email protected]>
It's used by probe and that isn't an init function. Drop this so that we
don't get a section mismatch.
Reported-by: kbuild test robot <[email protected]>
Cc: David Müller <[email protected]>
Cc: Hans de Goede <[email protected]>
Cc: Andy Shevchenko <[email protected]>
Fixes: 7c2e071300 ("clk: x86: Add system specific quirk to mark clocks as critical")
Signed-off-by: Stephen Boyd <[email protected]>
Pull NVMe fixes from Christoph:
"Two nvme fixes for 5.1 - fixing the initial CSN for nvme-fc, and handle
log page offsets properly in the target."
* 'nvme-5.1' of git://git.infradead.org/nvme:
nvmet: fix discover log page when offsets are used
nvme-fc: correct csn initialization and increments on error
The nvme target hadn't been taking the Get Log Page offset parameter
into consideration, and so has been returning corrupted log pages when
offsets are used. Since many tools, including nvme-cli, split the log
request to 4k, we've been breaking discovery log responses when more
than 3 subsystems exist.
Fix the returned data by internally generating the entire discovery
log page and copying only the requested bytes into the user buffer. The
command log page offset type has been modified to a native __le64 to
make it easier to extract the value from a command.
Signed-off-by: Keith Busch <[email protected]>
Tested-by: Minwoo Im <[email protected]>
Reviewed-by: Chaitanya Kulkarni <[email protected]>
Reviewed-by: Hannes Reinecke <[email protected]>
Reviewed-by: James Smart <[email protected]>
Signed-off-by: Christoph Hellwig <[email protected]>
This patch fixes a long-standing bug that initialized the FC-NVME
cmnd iu CSN value to 1. Early FC-NVME specs had the connection starting
with CSN=1. By the time the spec reached approval, the language had
changed to state a connection should start with CSN=0. This patch
corrects the initialization value for FC-NVME connections.
Additionally, in reviewing the transport, the CSN value is assigned to
the new IU early in the start routine. It's possible that a later dma
map request may fail, causing the command to never be sent to the
controller. Change the location of the assignment so that it is
immediately prior to calling the lldd. Add a comment block to explain
the impacts if the lldd were to additionally fail sending the command.
Signed-off-by: Dick Kennedy <[email protected]>
Signed-off-by: James Smart <[email protected]>
Reviewed-by: Ewan D. Milne <[email protected]>
Signed-off-by: Christoph Hellwig <[email protected]>
Due to copy/paste error, the fbdev format was changed to 32bpp = XRGB8888
which is an emulated format for the RGB565 drivers. Revert to to using the
fallback which is dev->mode_config.preferred_depth for the drivers that
set it or 32bpp for those that don't (repaper, st7586).
Fixes: 3eba392281 ("drm/tinydrm: Drop using tinydrm_device")
Signed-off-by: Noralf Trønnes <[email protected]>
Reviewed-by: Neil Armstrong <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
A few core fixes along with the driver specific ones, mainly fixing
small issues that only affect x86 platforms for various reasons (their
unusual machine enumeration mechanisms mainly, plus a fix for error
handling in topology).
There's some of the driver fixes that look larger than they are, like
the hdmi-codec changes which resulted in an indentation change, and most
of the other large changes are for new drivers like the STM32 changes.
-----BEGIN PGP SIGNATURE-----
iQFHBAABCgAxFiEEreZoqmdXGLWf4p/qJNaLcl1Uh9AFAlyvIEoTHGJyb29uaWVA
a2VybmVsLm9yZwAKCRAk1otyXVSH0N5OB/4uphscKPebkxqa9jGkXDoatGfBjQLE
KfLJzbuwqdg8gHNcWWXmZ1v+5UkJN2lbC9rKoOmGhAsGYNSftpWV1cUOemW2ryo7
OAiW14h/GjfI5JDwveVcyZJQrtBw2KBH42A0tE7VGOYu8lUWSFgLAr8wJRevXw1h
uYhgFwz5drn7yAg9mked0uGCvHN1mXpwQdRExKytXNLCnAm1TmQck0cwXn8oia3h
NKzRk1ofnKFolesjllc1rKJP2vhv8+SRKj1mi6FB5Zo2zlBAOdu4f+e/cf+2VQyS
7ltdK3tcWSb1laIATk47lNMIFoanSn9DN2vPzg91po5mUh6BGF/UUho/
=Zijb
-----END PGP SIGNATURE-----
Merge tag 'asoc-fix-v5.1-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound into for-linus
ASoC: Fixes for v5.1
A few core fixes along with the driver specific ones, mainly fixing
small issues that only affect x86 platforms for various reasons (their
unusual machine enumeration mechanisms mainly, plus a fix for error
handling in topology).
There's some of the driver fixes that look larger than they are, like
the hdmi-codec changes which resulted in an indentation change, and most
of the other large changes are for new drivers like the STM32 changes.
commit 5b0d62108b ("mmc: sdhci-omap: Add platform specific reset
callback") skips data resets during tuning operation. Because of this,
a data error or data finish interrupt might still arrive after a command
error has been handled and the mrq ended. This ends up with a "mmc0: Got
data interrupt 0x00000002 even though no data operation was in progress"
error message.
Fix this by adding a platform specific callback for sdhci_irq. Mark the
mrq as a failure but wait for a data interrupt instead of calling
finish_mrq().
Fixes: 5b0d62108b ("mmc: sdhci-omap: Add platform specific reset
callback")
Signed-off-by: Faiz Abbas <[email protected]>
Acked-by: Adrian Hunter <[email protected]>
Cc: [email protected]
Signed-off-by: Ulf Hansson <[email protected]>
A single, one-line fix for a build error introduced in v5.1-rc1.
-----BEGIN PGP SIGNATURE-----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=qL9y
-----END PGP SIGNATURE-----
Merge tag 'drm/tegra/for-5.1-rc5' of git://anongit.freedesktop.org/tegra/linux into drm-fixes
drm/tegra: Fixes for v5.1-rc5
A single, one-line fix for a build error introduced in v5.1-rc1.
Signed-off-by: Dave Airlie <[email protected]>
From: Thierry Reding <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
In case the IOMMU API is not available compiling host1x fails with
the following error:
In file included from drivers/gpu/host1x/hw/host1x06.c:27:
drivers/gpu/host1x/hw/channel_hw.c: In function ‘host1x_channel_set_streamid’:
drivers/gpu/host1x/hw/channel_hw.c:118:30: error: implicit declaration of function
‘dev_iommu_fwspec_get’; did you mean ‘iommu_fwspec_free’? [-Werror=implicit-function-declaration]
struct iommu_fwspec *spec = dev_iommu_fwspec_get(channel->dev->parent);
^~~~~~~~~~~~~~~~~~~~
iommu_fwspec_free
Fixes: de5469c21f ("gpu: host1x: Program the channel stream ID")
Signed-off-by: Stefan Agner <[email protected]>
Signed-off-by: Thierry Reding <[email protected]>
Consider two tasks that are running in parallel on a pair of engines
(vcs0, vcs1), but then must complete on a shared engine (rcs0). To
maximise throughput, we want to run the first ready task on rcs0 (i.e.
the first task that completes on either of vcs0 or vcs1). When using
semaphores, however, we will instead queue onto rcs in submission order.
To resolve this incorrect ordering, we want to re-evaluate the priority
queue when each of the request is ready. Normally this happens because
we only insert into the priority queue requests that are ready, but with
semaphores we are inserting ahead of their readiness and to compensate
we penalize those tasks with reduced priority (so that tasks that do not
need to busywait should naturally be run first). However, given a series
of tasks that each use semaphores, the queue degrades into submission
fifo rather than readiness fifo, and so to counter this we give a small
boost to semaphore users as their dependent tasks are completed (and so
we no longer require any busywait prior to running the user task as they
are then ready themselves).
v2: Fixup irqsave for schedule_lock (Tvrtko)
Testcase: igt/gem_exec_schedule/semaphore-codependency
Signed-off-by: Chris Wilson <[email protected]>
Cc: Tvrtko Ursulin <[email protected]>
Cc: Dmitry Rogozhkin <[email protected]>
Cc: Dmitry Ermilov <[email protected]>
Reviewed-by: Tvrtko Ursulin <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
When fb is tiled and fb->height isn't the multiple of tile's height,
the format fb->size = fb->stride * fb->height, will get a smaller size
than the actual size. As the memory height of tiled fb should be multiple
of tile's height.
Fixes: 7f1a93b1f1 ("drm/i915/gvt: Correct the calculation of plane size")
Reviewed-by: Zhenyu Wang <[email protected]>
Signed-off-by: Xiong Zhang <[email protected]>
Signed-off-by: Zhenyu Wang <[email protected]>
Since commit 648e921888 ("clk: x86: Stop marking clocks as
CLK_IS_CRITICAL"), the pmc_plt_clocks of the Bay Trail SoC are
unconditionally gated off. Unfortunately this will break systems where these
clocks are used for external purposes beyond the kernel's knowledge. Fix it
by implementing a system specific quirk to mark the necessary pmc_plt_clks as
critical.
Fixes: 648e921888 ("clk: x86: Stop marking clocks as CLK_IS_CRITICAL")
Signed-off-by: David Müller <[email protected]>
Signed-off-by: Hans de Goede <[email protected]>
Reviewed-by: Andy Shevchenko <[email protected]>
Signed-off-by: Stephen Boyd <[email protected]>
During a safe hot remove, the OS powers off the slot, which may cause a
Data Link Layer State Changed event. The slot has already been set to
OFF_STATE, so that event results in re-enabling the device, making it
impossible to safely remove it.
Clear out the Presence Detect Changed and Data Link Layer State Changed
events when the disabled slot has settled down.
It is still possible to re-enable the device if it remains in the slot
after pressing the Attention Button by pressing it again.
Fixes the problem that Micah reported below: an NVMe drive power button may
not actually turn off the drive.
Link: https://bugzilla.kernel.org/show_bug.cgi?id=203237
Reported-by: Micah Parrish <[email protected]>
Tested-by: Micah Parrish <[email protected]>
Signed-off-by: Sergey Miroshnichenko <[email protected]>
[bhelgaas: changelog, add bugzilla URL]
Signed-off-by: Bjorn Helgaas <[email protected]>
Reviewed-by: Lukas Wunner <[email protected]>
Cc: [email protected] # v4.19+
Finally have a reason for a backmerge other than "it's been a while"!
Backmerging drm-next to -misc-next to facilitate Rob Herring's work on
Panfrost.
Signed-off-by: Sean Paul <[email protected]>
Now that we allow drivers to always need to set larger than required
DMA masks we need to be a little more careful in the sun4v PCI iommu
driver to chose when to select the ATU support - a larger DMA mask
can be set even when the platform does not support ATU, so we always
have to check if it is avaiable before using it. Add a little helper
for that and use it in all the places where we make ATU usage decisions
based on the DMA mask.
Fixes: 24132a419c ("sparc64/pci_sun4v: allow large DMA masks")
Reported-by: Meelis Roos <[email protected]>
Signed-off-by: Christoph Hellwig <[email protected]>
Tested-by: Meelis Roos <[email protected]>
Acked-by: David S. Miller <[email protected]>
Several driver bug fixes posted in the last several weeks
- Several bug fixes for the hfi1 driver 'TID RDMA' functionality merged
into 5.1. Since TID RDMA is on by default these all seem to be
regressions.
- Wrong software permission checks on memory in mlx5
- Memory leak in vmw_pvrdma during driver remove
- Several bug fixes for hns driver features merged into 5.1
-----BEGIN PGP SIGNATURE-----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=Dq9o
-----END PGP SIGNATURE-----
Merge tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma
Pull rdma fixes from Jason Gunthorpe:
"Several driver bug fixes posted in the last several weeks
- Several bug fixes for the hfi1 driver 'TID RDMA' functionality
merged into 5.1. Since TID RDMA is on by default these all seem to
be regressions.
- Wrong software permission checks on memory in mlx5
- Memory leak in vmw_pvrdma during driver remove
- Several bug fixes for hns driver features merged into 5.1"
* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma:
IB/hfi1: Do not flush send queue in the TID RDMA second leg
RDMA/hns: Bugfix for SCC hem free
RDMA/hns: Fix bug that caused srq creation to fail
RDMA/vmw_pvrdma: Fix memory leak on pvrdma_pci_remove
IB/mlx5: Reset access mask when looping inside page fault handler
IB/hfi1: Fix the allocation of RSM table
IB/hfi1: Eliminate opcode tests on mr deref
IB/hfi1: Clear the IOWAIT pending bits when QP is put into error state
IB/hfi1: Failed to drain send queue when QP is put into error state
Not all archs have the __io_virt() macro, so cirrus can't simply convert
pointers that way. The drm format helpers have to use memcpy_toio()
instead.
This patch makes drm_fb_xrgb8888_to_rgb888_dstclip() accept a __iomem
dst pointer and use memcpy_toio() instead of memcpy(). The helper
function (drm_fb_xrgb8888_to_rgb888_line) has been changed to process a
single scanline.
Signed-off-by: Gerd Hoffmann <[email protected]>
Reviewed-by: Noralf Trønnes <[email protected]>
Link: http://patchwork.freedesktop.org/patch/msgid/[email protected]
Not all archs have the __io_virt() macro, so cirrus can't simply convert
pointers that way. The drm format helpers have to use memcpy_toio()
instead.
This patch makes drm_fb_xrgb8888_to_rgb565_dstclip() accept a __iomem
dst pointer and use memcpy_toio() instead of memcpy(). The helper
function (drm_fb_xrgb8888_to_rgb565_line) has been changed to process
a single scanline.
Signed-off-by: Gerd Hoffmann <[email protected]>
Reviewed-by: Noralf Trønnes <[email protected]>
Link: http://patchwork.freedesktop.org/patch/msgid/[email protected]
Not all archs have the __io_virt() macro, so cirrus can't simply convert
pointers that way. The drm format helpers have to use memcpy_toio()
instead.
This patch makes drm_fb_memcpy_dstclip() accept a __iomem dst pointer
and use memcpy_toio() instead of memcpy(). With that separating out the
memcpy loop into the drm_fb_memcpy_lines() helper isn't useful any more,
so move the code back into the calling functins.
Signed-off-by: Gerd Hoffmann <[email protected]>
Reviewed-by: Noralf Trønnes <[email protected]>
Link: http://patchwork.freedesktop.org/patch/msgid/[email protected]
[Why]
When deciding to add properties or expose formats on DRM planes we
should be querying the caps for the DC plane it's supposed to represent.
[How]
Pass plane caps down into plane initialization, refactoring overlay
plane initialization to have the overlay plane be represented by
the first overlay capable DC plane.
Signed-off-by: Nicholas Kazlauskas <[email protected]>
Reviewed-by: Leo Li <[email protected]>
Signed-off-by: Alex Deucher <[email protected]>
[Why]
dcn*_disable_plane() doesn't unlock the pipe anymore, making the extra
lock unnecessary.
In addition - during full plane updates - all necessary pipes should be
locked/unlocked together when modifying hubp to avoid tearing in
pipesplit setups.
[How]
Remove redundant locks, and add function to lock all pipes. If an
interdependent pipe update is required, lock down all pipes. Otherwise,
lock only the top pipe for the updated pipe tree.
Signed-off-by: Leo Li <[email protected]>
Reviewed-by: Tony Cheng <[email protected]>
Acked-by: Nicholas Kazlauskas <[email protected]>
Signed-off-by: Alex Deucher <[email protected]>
[Why]
When entering S3, amdgpu first calls DRM to cache the current atomic
state, then commit the 'all-disabled' state. This sets dc->current_state
to point at the private atomic object's dm_atomic_state->context, as
any regular atomic commit would.
Afterwards, amdgpu_dm calls dc_set_power_state() with S3 power state.
This invalidates dc->current_state by wiping it to 0, consequently
wiping dm_atomic_state->context.
During resume, the cached atomic state is restored. When getting the
private object however, the dm_atomic_state - containing the wiped
context - is duplicated into the atomic state. This causes DC validation
to fail during atomic check, as necessary function pointers in dc_state
are now NULL.
[How]
Recreate the private object's dm_atomic_state->context during resume,
restoring any static values such as function pointers.
A TODO item is added to move static read-only values out of dc_state -
they shouldn't be there anyways.
Signed-off-by: Leo Li <[email protected]>
Reviewed-by: Nicholas Kazlauskas <[email protected]>
Signed-off-by: Alex Deucher <[email protected]>
