summaryrefslogtreecommitdiff
path: root/drivers/gpu/drm/nouveau
AgeCommit message (Collapse)Author
2013-07-01drm/nouveau/vp/nv84: initial vp2 engine implementationIlia Mirkin
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nouveau/core: xtensa engine base class implementationIlia Mirkin
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nouveau/vdec: fork vp3 implementations from vp2Ilia Mirkin
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nouveau/core: move falcon class to engine/Ben Skeggs
Not really "core" per-se. About to merge Ilia's work adding another similar class for the VP2 xtensa engines, so, seems like a good time to move all these to engine/. Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nouveau/kms: don't fail if there's no dcb table entriesBen Skeggs
Fixes module not loading on Tesla K20. Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nouveau: remove limit on gartMaarten Lankhorst
Most graphics cards nowadays have a multiple of this limit as their vram, so limiting GART doesn't seem to make much sense. Signed-off-by: Maarten >Lnkhorst <maarten.lankhorst@canonical.com> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nouveau/vm: perform a bar flush when flushing vmMaarten Lankhorst
Appears to fix the regression from "drm/nvc0/vm: handle bar tlb flushes internally". nvidia always seems to do this flush after writing values. Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nvc0/gr: cleanup register lists, and add nvce/nvcf to switchesBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nvc8/gr: update initial register/context valuesBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nvc4/gr: update initial register/context valuesBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nvc1/gr: update initial register/context valuesBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nvc3/gr: update initial register/context valuesBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nvc0/gr: update initial register/context valuesBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nvd9/gr: update initial register/context valuesBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nve4/gr: update initial register/context valuesBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nvc0-/gr: bump maximum gpc/tpc limitsBen Skeggs
Needed for GK110, separate commit to catch any unexpected breaks to other parts of the code. Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nvf0/gr: initial register/context setupBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nve7/gr: update initial register/context valuesBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nve6/gr: update initial register/context valuesBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nouveau: delay busy bo vma removal until fence signalsBen Skeggs
As opposed to an explicit wait. Allows userspace to not stall waiting on buffer deletion. Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nouveau/vm: make each vma take a reference on its parent vmBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nouveau/core: remove nouveau_mm.mutex, no more usersBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nouveau/vm: take subdev mutex, not the mm, protects against race with ↵Ben Skeggs
vm/nvc0 nvc0_vm_flush() accesses the pgd list, which will soon be able to race with vm_unlink() during channel destruction. Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nvc0/vm: handle bar tlb flushes internallyBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nv50-/vm: take mutex rather than irqsave spinlockBen Skeggs
These operations can take quite some time, and we really don't want to have to hold a spinlock for too long. Now that the lock ordering for vm and the gr/nv84 hw bug workaround has been reversed, it's possible to use a mutex here. Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nv50/vm: remove explicit vm knowledge from enginesBen Skeggs
This reverses the lock ordering between VM and gr/nv84:nvc0. Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nv50/vm: handle bar tlb flushes internallyBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nvc0/gr: port mp trap handling from calim's kepler codeBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nve0/gr: attempt to resume after sm trapsBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nve0/gr: s/tp/tpc/Ben Skeggs
NVIDIA's name... Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nve0/fifo: create our playlists up-front, at startupBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nva3/clk: minor improvements to fractional N calculationBen Skeggs
Helps us to get identical numbers to the binary driver for (at least) Kepler memory PLLs, and fixes a rounding error. Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nouveau/fb: initialise vram controller as pfb sub-objectBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nouveau/clk: change init ordering, no longer needed by devinitBen Skeggs
And, will depend on FB/VOLT/DAEMON being ready when it gets initialised so that it can set/restore clocks. Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nouveau/devinit: move simple pll setting routines to devinitBen Skeggs
These are pretty much useless for reclocking purposes. Lets make it clearer what they're for and move them to DEVINIT to signify they're for the very simple PLL setting requirements of running the init tables. Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nouveau: pass generic subdev to calculation routinesBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nve0/ce: stub interrupt handlerBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nve0/ce: link ce2 to its engine, rather than from graphicsBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nve0/fifo: copy engine context stored in ramfc, not externallyBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nve0/ce: create engine object for ce2Ben Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-07-01drm/nouveau: pull in latest ucode builds from external treeBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2013-06-28drm/ttm: convert to the reservation apiMaarten Lankhorst
Now that the code is compatible in semantics, flip the switch. Use ww_mutex instead of the homegrown implementation. ww_mutex uses -EDEADLK to signal that the caller has to back off, and -EALREADY to indicate this buffer is already held by the caller. ttm used -EAGAIN and -EDEADLK for those, respectively. So some changes were needed to handle this correctly. Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Reviewed-by: Jerome Glisse <jglisse@redhat.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-06-28drm/nouveau: make flipping lockdep safeMaarten Lankhorst
cli->mutex was inverted with reservations, and multiple reservations were used without a ticket, fix both. This commit had to be done after the previous commit, because otherwise ttm_eu_* calls would use a different seqno counter.. Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Acked-by: Ben Skeggs <bskeggs@redhat.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-06-28drm/ttm: make ttm reservation calls behave like reservation callsMaarten Lankhorst
This commit converts the source of the val_seq counter to the ww_mutex api. The reservation objects are converted later, because there is still a lockdep splat in nouveau that has to resolved first. Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Reviewed-by: Jerome Glisse <jglisse@redhat.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-06-28drm/nouveau: always select ACPI_VIDEO if ACPI is enabled.Maarten Lankhorst
Having nouveau builtin would still allow ACPI_VIDEO to be used as external module if some of the deps for acpi_video have not been met, which would result in a linking failure. Solve this by selecting all dependencies as well. Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Acked-by: Ben Skeggs <bskeggs@redhat.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-06-28drm/nouveau: complain loudly if buffer is pinned during destructionMaarten Lankhorst
Shouldn't happen, and we invert the struct_mutex with reservation here, potentially leading to deadlocks. Once reservations become lockdep annotated, lockdep will go splat on this. Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Acked-by: Ben Skeggs <bskeggs@redhat.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-06-28drm/nouveau: fixup fbcon failure pathsMaarten Lankhorst
Add missing calls, and fix a leak from forgetting to call the unpin function. Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Acked-by: Ben Skeggs <bskeggs@redhat.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-06-28drm/nouveau: unpin notify object in chan_finiMaarten Lankhorst
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Acked-by: Ben Skeggs <bskeggs@redhat.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-06-28drm/nouveau: implement prime helper unpin functionMaarten Lankhorst
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Acked-by: Ben Skeggs <bskeggs@redhat.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-06-27Merge tag 'v3.10-rc7' into drm-nextDave Airlie
Linux 3.10-rc7 The sdvo lvds fix in this -fixes pull commit c3456fb3e4712d0448592af3c5d644c9472cd3c1 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Mon Jun 10 09:47:58 2013 +0200 drm/i915: prefer VBT modes for SVDO-LVDS over EDID has a silent functional conflict with commit 990256aec2f10800595dddf4d1c3441fcd6b2616 Author: Ville Syrjälä <ville.syrjala@linux.intel.com> Date: Fri May 31 12:17:07 2013 +0000 drm: Add probed modes in probe order in drm-next. W simply need to add the vbt modes before edid modes, i.e. the other way round than now. Conflicts: drivers/gpu/drm/drm_prime.c drivers/gpu/drm/i915/intel_sdvo.c
generated by cgit v0.10.2 at 2024-08-04 08:27:46 (GMT)