The point is to share more code between the PFB/PGRAPH tile region
hooks, and give the hardware specific functions a chance to allocate
per-region resources.
Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
nouveau_bo_move_m2mf() needs to lock the kernel channel, and it may be
called from the pushbuf IOCTL with an user channel already locked. Use
a separate subclass for the kernel channel mutex because this is
legitimate mutex nesting.
Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
In a multihead setup vblank interrupts may end up enabled in both
heads. In that case we want to ignore the vblank interrupts coming
from the wrong CRTC to avoid tearing and unbalanced calls to
drm_vblank_get/put (fdo bug 31074).
Reported-by: Felix Leimbach <felix.leimbach@gmx.net>
Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
nv0x-nv4x should be mostly fine, nv50 doesn't work yet.
Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
nouveau_fence_* functions are not type safe, which could lead to bugs.
Additionally every use of nouveau_fence_unref had to cast struct
nouveau_fence to void **.
Fix it by renaming old functions and creating static inline functions with
new prototypes. We still need old functions, because we pass function
pointers to ttm.
As we are wrapping functions, drop unused "void *arg" parameter where possible.
Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Only supported on NV50+ so far, and disabled by default currently. The
module parameter "msi=1" will enable it.
There's a kernel bug which will cause this to fail if the module (or the
NVIDIA binary driver) has ever been loaded before loading nouveau with
MSI enabled. As such, this is only safe to enable if you have nouveau
load on boot, and don't wish to ever reload it.
The workaround is to "echo 0 > /sys/bus/pci/devices/<device>/enable"
until the enable count reads 0. Then you should be able to load nouveau
with MSI enabled.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Not an issue right now, we're forced to 64k size/alignment by the BO
allocator anyway. This won't be the case soon.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
This really needs cleaning up somehow, and probably investigate what's
needed to do this on earlier generations. NVIDIA do something similar
there too.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
We previously added all the available classes for the entire generation,
even though the objects wouldn't work on the hardware.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
The structs themselves, as well as the non-sw object creation function are
probably very misnamed now. That's a problem for later :)
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Without it there's a potential race with nouveau_fence_update().
Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
It needs a "strong" channel reference because it actually writes to
the channel pushbuf, otherwise the corresponding FIFO context could
get kicked off in the middle of nouveau_fence_sync().
Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Fences didn't increment the channel reference count, and the fenced
channel could go away at any time. Fixes a potential race in
nouveau_fence_update().
Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
nouveau_channel_ref() takes a "weak" channel reference that doesn't
prevent the hardware channel resources from being released, it just
keeps the channel data structure alive.
Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
nouveau_channel_put() can be executed after the 'refcount == 0' check
in nouveau_channel_get() and before the channel reference count is
incremented. In that case CPU0 will take the context down while CPU1
thinks it owns the channel and 'refcount == 1'.
Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
The destroy_context() engine hooks call gpuobj management functions to
release the channel resources, these functions use HARDIRQ-unsafe locks
whereas destroy_context() is called with the HARDIRQ-safe
context_switch_lock held, that's a lock ordering violation.
Push the engine-specific channel destruction logic into destroy_context()
and let the hardware-specific code lock and unlock when it's actually
needed. Change the engine destruction order to avoid a race in the small
gap between pgraph and pfifo context uninitialization.
Reported-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
The pushbuf ioctl syncs after validation, no need for this anymore.
Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
No other driver uses this, and userspace should be responsible for handling
locking between them if they share BOs.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
This fixes a race condition between fbcon acceleration and TTM buffer
moves. To reproduce:
- start X
- switch to vt and "while (true); do dmesg; done"
- switch to another vt and "sleep 2 && cat /path/to/debugfs/dri/0/evict_vram"
- switch back to vt running dmesg
We don't make use of this on any other channel yet, they're currently
protected by drm_global_mutex. This will change in the near future.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
A future commit will add locking to the DRM's channel, and there's numerous
problems that come up if we allow printk from an interrupt context to be
accelerated. It seems saner to just disallow it completely.
As a nice side-effect, all the "to accel or not to accel" logic gets moved
out of the chipset-specific code.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
* 'drm-radeon-fusion' of ../drm-radeon-next:
drm/radeon/kms: add Ontario APU ucode loading support
drm/radeon/kms: add Ontario Fusion APU pci ids
drm/radeon/kms: enable MSIs on fusion APUs
drm/radeon/kms: add power table parsing support for Ontario fusion APUs
drm/radeon/kms: refactor atombios power state fetching
drm/radeon/kms: add bo blit support for Ontario fusion APUs
drm/radeon/kms: add thermal sensor support for fusion APUs
drm/radeon/kms: fill in GPU init for AMD Ontario Fusion APUs
drm/radeon/kms: add radeon_asic struct for AMD Ontario fusion APUs
drm/radeon/kms: evergreen.c updates for fusion
drm/radeon/kms: MC setup changes for fusion APUs
drm/radeon/kms: move r7xx/evergreen to its own vram_gtt setup function
drm/radeon/kms: add support for ss overrides on Fusion APUs
drm/radeon/kms: Add support for external encoders on fusion APUs
drm/radeon/kms: atom changes for DCE4.1 devices
drm/radeon/kms: add new family id for AMD Ontario APUs
drm/radeon/kms: upstream power table updates
drm/radeon/kms: upstream atombios.h updates
drm/radeon/kms: upstream ObjectID.h updates
drm/radeon/kms: setup mc chremap properly on r7xx/evergreen
* 'drm-radeon-next' of ../drm-radeon-next:
drm/radeon/kms: improve pflip precision on r1xx-r4xx
drm/kms/radeon: Use high precision timestamps for pageflip completion events.
drm/kms/radeon: Reorder vblank and pageflip interrupt handling.
drm/radeon/kms: add pageflip ioctl support (v3)
drm/kms/radeon: Add support for precise vblank timestamping.
This workaround only applies to Ironlake.
Signed-off-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: stable@kernel.org
The pipe is always set to 8BPC, but here we were leaving whatever
previous bits were set by the BIOS in place.
Signed-off-by: Eric Anholt <eric@anholt.net>
Tested-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: stable@kernel.org
As the tracepoint is now decoupled from when the actual register is
assigned and was never complemented by detailing when the object lost
its fence, it has outlived its limited usefulness. Profiling the actual
stalls is a far more profitable venture anyway.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
As the userspace mappings are torn down on every GPU write, we prefer to
track when the buffer is activated (via a fresh i915_gem_fault). This
makes the LRU conceptually simpler. With coherent mappings, the
remaining use-case for set_domain_ioctl is GPU synchronisation.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
With this change, every batchbuffer can use all available fences (save
pinned and scanout, of course) without ever stalling the gpu!
In theory. Currently the actual pipelined update of the register is
disabled due to some stability issues. However, just the deferred update
is a significant win.
Based on a series of patches by Daniel Vetter.
The premise is that before every access to a buffer through the GTT we
have to declare whether we need a register or not. If the access is by
the GPU, a pipelined update to the register is made via the ringbuffer,
and we track the last seqno of the batches that access it. If by the
CPU we wait for the last GPU access and update the register (either
to clear or to set it for the current buffer).
One advantage of being able to pipeline changes is that we can defer the
actual updating of the fence register until we first need to access the
object through the GTT, i.e. we can eliminate the stall on set_tiling.
This is important as the userspace bo cache does not track the tiling
status of active buffers which generate frequent stalls on gen3 when
enabling tiling for an already bound buffer.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
6xx/7xx was hitting the wrong BUS_CNTL reg and bits.
Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Cc: stable@kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>
* 'intel/drm-intel-fixes' of /ssd/git/drm-next:
Revert "drm/i915/dp: use VBT provided eDP params if available"
drm/i915: Clear pfit registers when not used by any outputs
drm/i915: fix regression due to ba3d8d749b
This makes the various rings more consistent by removing the anomalous
handing of the rendering ring execbuffer dispatch.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
This reverts commit 869184a675.
This is required for the Sony Vaio Jesse was working on at the time, but
breaks most other eDP machines - machines that were working in earlier
kernels.
Reported-and-tested-by: Dave Airlie <airlied@redhat.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=31188
Tested-by: Zhao Jian <jian.j.zhao@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
... otherwise the panel-fitter may be left enabled with random settings
and cause unintended filtering (i.e. blurring of native modes on external
panels).
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=31942
Reported-and-tested-by: Ben Kohler <bkohler@gmail.com>
Tested-by: Ciprian Docan <docan@eden.rutgers.edu>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
In order to correctly report monitor connected status changes, the
previous monitor status must be recorded in the connector->status
value instead of being discarded.
Signed-off-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
When setting a new crtc configuration, force the DPMS state of all
connectors to ON. Otherwise, they'll be left at OFF and a future mode set
that disables the specified connector will not turn the connector off.
Signed-off-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
... so that upon first use after resume we will reacquire the fence reg.
Reported-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
As the error occurred on the current object, it means that its state was
not changed and so it should be excluded from the unwind.
Reported-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
We don't track gpu flush request in any special way. So even with
obj->write_domain == 0, a gpu flush might be outstanding but no
yet executed. Even worse, the latest request might use the object
only for reading. So and unconditional call to object_wait_rendering
is needed for !pipelined.
Hence revert that patch fully and untangle the flushing from the
synchronization again.
Reported-by: Keith Packard <keithp@keithp.com>
Tested-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
* 'intel/drm-intel-fixes' of /ssd/git/drm-next:
drm/i915/sdvo: Always add a 30ms delay to make SDVO TV detection reliable
MAINTAINERS: INTEL DRM DRIVERS list (intel-gfx) is subscribers-only
drm/i915/sdvo: Always fallback to querying the shared DDC line
drm/i915: Handle pagefaults in execbuffer user relocations
drm/i915/sdvo: Only enable HDMI encodings only if the commandset is supported
drm/i915: Only save/restore cursor regs if !KMS
drm/i915: Prevent integer overflow when validating the execbuffer
The update pending bit has a separate enable bit.
Cc: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This reverts commit d33ef52d9d.
This change seems to expose a bug in the 3D driver tiggered by
certain apps, so revert it to keep userspace working.
Reported-by: Rafael J. Wysocki <rjw@sisk.pl>
Cc: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Besides the minimal improvement in reducing the execbuffer overhead, the
real benefit is clarifying a few routines.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
A number of dragons have been seen lurking within the execbuffer code.
The first step is then to isolate them from the rest and begin to
scrutinise them in depth. Suggested by Daniel Vetter.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Simply remove our accounting of objects inside the aperture, keeping
only track of what is in the aperture and its current usage. This
removes the over-complication of BUGs that were attempting to keep the
accounting correct and also removes the overhead of the accounting on
the hot-paths.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
With KMS, we can simply relinquish the fence when we idle the GPU and
reassign it upon first use.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Commit d09c23de intended to add a 30ms delay to give the ADD time to
detect any TVs connected. However, it used the sdvo->is_tv flag to do so
which is dependent upon the previous detection result and not whether the
output supports TVs.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: stable@kernel.org
Avoid evicting buffers that will be used later in the batch in order to
make room for the initial buffers by pinning all bound buffers in a
single pass before binding (and evicting for) fresh buffer.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
On a few devices, like the Mac Mini, the CRT DDC pins are shared between
the analog connector and the digital connector. In this scenario, rely
on the EDID to determine if a digital panel is connected to the digital
connector.
Reported-and-tested-by: Tino Keitel <tino.keitel@tikei.de>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
This used to check the precondition that all fences were to be located
in a mappable area, redundant now as those two parameters are combined
into one.
After pinning, we assert that the buffer is bound into the desired
region.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
The pipe control object is allocated by the device for the sole use of the
render ringbuffer. Move this detail from the general code to the render
ring buffer initialisation.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Having seen the effects of erroneous fencing on the batchbuffer, a
useful sanity check is to record the fence registers at the time of an
error.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Combining map_and_fenceable revealed a bug in
i915_gem_object_gtt_size() in that it always computed the appropriate
fence size for the object regardless of tiling state which caused us to
over-allocate linear buffers when binding to the GTT.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
This still uses the agp functions to actually reinstate the mappings
(with a gross hack to make agp cooperate), but it wires everything
up correctly for the switchover.
The call to agp_rebind_memory can be dropped because all non-kms drivers
do all their rebinding on EnterVT.
v2: Be more paranoid and flush the chipset cache after restoring gtt
mappings.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
This is required to restore gtt mappings on resume when agp is gone.
The right way to do this would be to make sturct drm_mm_node embeddable
and use the allocation list maintained by the drm memory manager. But
that's a bigger project. Getting rid of the per bo agp_mem will save
more memory than this wastes, anyway.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
No longer used.
Cc: Dave Airlie <airlied@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Currently if we hit a pagefault when applying a user relocation for the
execbuffer, we bail and return EFAULT to the application. Instead, we
need to unwind, drop the dev->struct_mutex, copy all the relocation
entries to a vmalloc array (to avoid any potential circular deadlocks
when resolving the pagefault), retake the mutex and then apply the
relocations. Afterwards, we need to again drop the lock and copy the
vmalloc array back to userspace.
v2: Incorporate feedback from Daniel Vetter.
Reported-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
The GATT is a write-only set of registers, reading from them in the
manner of i915_gtt_to_phys() is supposed to be undefined. However a
simple solution exists as we allocate linear memory from the stolen
area, we can simply add the block offset to the base register. As a
side-effect we recover all the unused stolen GTT entries and so enlarge
our aperture.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
After a GPU reset, the backlight controller registers may be also reset
to 0. In that case we should restore those to the original values
programmed by the BIOS. Note that we still lack the code to handle the
case where the BIOS failed to program those registers at all...
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
The vbios power tables on my inagua board seem a bit funky...
Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
The function was getting too large. Rework it to share
more state better handle new power table formats.
Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
- CONFIG_MEMSIZE is in bytes on fusion.
- FB_BASE and FB_TOP are finer grained.
Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
MC_VM_FB_LOCATION is at a different offset between r6xx and r7xx/evergreen.
The location is needed for vram setup on fusion chips.
Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
System specific spread spectrum overrides can be specified in
the integrated system info table for Fusion APUs. This adds
support for using those overrides.
Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Should improve performance slightly and possibly fix some
issues.
Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
As we conflated intel_sdvo->is_hdmi with both having HDMI support on the
ADD along with having HDMI support on the monitor, we would attempt to
use HDMI encodings even if the interface did not support those commands.
Reported-by: Simon Farnsworth <simon.farnsworth@onelan.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Tested-by: Simon Farnsworth <simon.farnsworth@onelan.co.uk>
Reviewed-by: Simon Farnsworth <simon.farnsworth@onelan.co.uk>
We were reading our 64-bit value in I915_READ64 and returning 32 bits
of it. The restoration of fence regs at resume then had a zero end
value, and the fence had no effect.
Version 2: Split register access functions into per-size versions
Sharing code between different sizes seemed reasonable when we only
needed a single copy, but as 64-bit access requires its own version,
it makes sense to just split them out for each size.
Reported-by: Peter Clifton <pcjc2@cam.ac.uk>
Signed-off-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
[ickle: use a macro to create the various read/write routines]
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
This has proven sufficient to recover from a hang of the GPU using the
gem_bad_blit test while at the KMS console then starting X. When
attempting the same during an X session, the timer doesn't appear to
trigger.
Signed-off-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
It isn't used for the hangcheck, which does its work right from the
timer trigger, but hangcheck can lead to error state recording, which
is run off of the workqueue.
Signed-off-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
When trying to diagnose mysterious errors on resume, capture the
display register contents as well.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
The pinned buffers are useful for diagnosing errors in setting up state
for the chipset, which may not necessarily be 'active' at the time of
the error, e.g. the cursor buffer object.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Rather than re-implementing in the Radeon driver,
Use the execbuf / cs / pushbuf utilities that comes with TTM.
This comes with an even greater benefit now that many spinlocks have been
optimized away...
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This patch attempts to fix up shortcomings with the current calling
sequences.
1) There's a fastpath where no locking occurs and only io_mem_reserved is
called to obtain needed info for mapping. The fastpath is set per
memory type manager.
2) If the fastpath is disabled, io_mem_reserve and io_mem_free will be exactly
balanced and not called recursively for the same struct ttm_mem_reg.
3) Optionally the driver can choose to enable a per memory type manager LRU
eviction mechanism that, when io_mem_reserve returns -EAGAIN will attempt
to kill user-space mappings of memory in that manager to free up needed
resources
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Ben Skeggs <bskeggs@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Rather than having the driver supply the validation sequence, leave that
responsibility to TTM. This saves some confusion and a function argument.
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Drastically reduce the number of spin lock / unlock operations by performing
unreserving and fencing under global locks.
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Jerome Glisse <j.glisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
The bo lock used only to protect the bo sync object members, and since it
is a per bo lock, fencing a buffer list will see a lot of locks and unlocks.
Replace it with a per-device lock that protects the sync object members on
*all* bos. Reading and setting these members will always be very quick, so
the risc of heavy lock contention is microscopic. Note that waiting for
sync objects will always take place outside of this lock.
The bo device fence lock will eventually be replaced with a seqlock /
rcu mechanism so we can determine that a bo is idle under a
rcu / read seqlock.
However this change will allow us to batch fencing and unreserving of
buffers with a minimal amount of locking.
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Jerome Glisse <j.glisse@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Add an aid for the driver to detect deadlocks on multi-bo reservations
Update documentation.
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Jerome Glisse <j.glisse@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Avoid the ttm_bo_unreserve() spinlocks by calling
ttm_eu_backoff_reservation_locked under the lru spinlock.
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Jerome Glisse <j.glisse@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Makes it possible to reserve a list of buffer objects with a single
spin lock / unlock if there is no contention.
Should improve cpu usage on SMP kernels.
v2: Initialize private list members on reserve and don't call
ttm_bo_list_ref_sub() with zero put_count.
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
I had removed this when I switched the atom indirect io methods
to use the io bar rather than the mmio bar, but it appears it's
still needed.
Reported-by: Mark Lord <kernel@teksavvy.com>
Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Cc: stable@kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>
