Re-uses the implementation's accessor functions rather than requiring
and init/fini implementation for each chipset.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
There's really not a great deal this time due to me spending most of this window on Maxwell. But, here's the random bits and pieces that's currently queued.
* 'drm-nouveau-next' of git://anongit.freedesktop.org/git/nouveau/linux-2.6: (25 commits)
drm/gk208/gr: add missing registers to grctx init
drm/nouveau/kms/nv04-nv40: fix pageflip events via special case.
drm/nv50-/mc: fix kms pageflip events by reordering irq handling order.
drm/nouveau/disp/nv04-nv40: abort scanoutpos query on vga analog.
drm/nv50-/kms: wait for enough ring space in crtc_prepare()
drm/nouveau/disp/dp: support training pattern 3
drm/nouveau/disp/dp: support aux read interval during link training
drm/gk104/gpio: fix incorrect interrupt register usage
drm/nouveau/core: punt all object state change messages to trace level
drm/nouveau/clk: allow end-user reclocking for nv40, nvaa, and nve0 clock types
drm/nouveau/fb: default NvMemExec to on, turning it off is used for debugging only
drm/nouveau/bios: fix a potential NULL deref in the PROM shadowing function
drm/nouveau/i2c: bump the i2c delay for the adt7473
drm/nouveau/therm/fan/tach: default to 2 pulses per revolution
drm/nvf0/device: enable video decoding engines on gk110/gk208
drm/nvf1/device: add support for 0xf1 (gk110b)
drm/nouveau/device: support for probing GK20A
drm/nouveau/graph: add GK20A support
drm/nouveau/graph: pad firmware code at load time
drm/nouveau/graph: enable when using external fw
...
This fixes hangs on GK208 which happen instantaneously on trying to use a
geometry shader.
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Cc: stable@vger.kernel.org # v3.14+
Cards with nv04 display engine can't reliably use vblank
counts and timestamps computed via drm_handle_vblank(), as
the function gets invoked after sending the pageflip events.
Fix this by defaulting to the old crtcid = -1 fallback path
on <= NV-50 cards, and only using the precise path on NV-50
and later.
Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Cc: <stable@vger.kernel.org> # 3.13+
Whenever a single nouveau_mc_intr() main gpu irq-handler invocation was
responsible for calling both, the vblank-irq handler (display engine irq)
and kms-pageflip completion handler (from fifo irq), the order of
invocation was wrong. nouveau_finish_flip() was called before
drm_handle_vblank() for the vblank of pageflip completion, so the
emitted pageflip event contained stale vblank count and timestamp
from previous vblank. This caused failure in userspace to timestamp
properly.
Reorder order of invocation of engine irq handlers: Put
NVDEV_ENGINE_DISP always on top, and thereby before NVDEV_ENGINE_FIFO,
so that drm_handle_vblank() gets called to update vblank timestamps
and count before potential pageflip events make use of that
information.
This works on nv-50 and later, where kms-pageflip completion triggers
an irq either after a separate vblank irq, or both pageflip and vblank
trigger one common irq invocation, but never before vblank irqs.
v2 (Ben):
- removed mods for nv04-nv40, it doesn't help there anyway
- this is considered a hack, and a better solution should be found
Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Cc: <stable@vger.kernel.org> # 3.13+
nv04_disp_scanoutpos() must abort to trigger simple timestamping
fallback if vtotal/htotal regs return zero. This happens if the
output isn't a digital output, but a vga analog output, as the
regs don't get initialized in that case.
Fixes timestamping failure on nv-40 and earlier with vga output.
Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Cc: <stable@vger.kernel.org> # 3.14+
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Martin Peres <martin.peres@free.fr>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Some adt7473 can't manage the 20µs delay we use for the bitbanging, bumping
it to 40µs seem to do the trick.
Signed-off-by: Martin Peres <martin.peres@free.fr>
Tested-by: Marcel Dopita <mdop@seznam.cz>
I spent some time this weekend trying to find in the vbios the number of
pulses per revolutions in the vbios but couldn't find it. It would seem
all my cards have 2 pulses per revolution so let's stick to that until
further notice.
Thermal table's id 0x48 may indicate this information but it would seem
that changing the value results in the blob power or clock gating the
RPM counter... We should ask NVIDIA about that, should be trivial-enough
for them to answer.
Signed-off-by: Martin Peres <martin.peres@free.fr>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Set the correct subdev/engine classes when GK20A (0xea) is probed.
Signed-off-by: Alexandre Courbot <acourbot@nvidia.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Add a GR device for GK20A based on NVE4, with the correct classes
definitions (GK20A's 3D class is 0xa297).
Most of the NVE4 code can be used on GK20A, so make relevant bits of
NVE4 available to other chips as well.
Signed-off-by: Alexandre Courbot <acourbot@nvidia.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Pad the microcode to a multiple of 0x40 words, otherwise firmware will
fail to run from non-prepadded firmware files.
Signed-off-by: Alexandre Courbot <acourbot@nvidia.com>
Reviewed-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
nvc0_graph_ctor() would only let the graphics engine be enabled if its
oclass has a proper microcode linked to it. This prevents GR from being
enabled at all on chips that rely exclusively on external firmware, even
though such a use-case is valid.
Relax the conditions enabling the GR engine to also include the case
where an external firmware has also been loaded.
Also switch to external firmware if the graph class has no microcode
linked to it.
Signed-off-by: Alexandre Courbot <acourbot@nvidia.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
GK20A's FIFO is compatible with NVE0, but only features 128 channels and
1 runlist.
Signed-off-by: Alexandre Courbot <acourbot@nvidia.com>
Reviewed-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Add a simple FB device for GK20A, as well as a RAM implementation
suitable for chips that use system memory as video RAM.
Signed-off-by: Alexandre Courbot <acourbot@nvidia.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Add support for initializing the priv ring of GK20A. This is done by the
BIOS on desktop GPUs, but needs to be done by hand on Tegra.
Signed-off-by: Alexandre Courbot <acourbot@nvidia.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Adapt the NVC0 BAR driver to make it able to support chips that do not
expose a BAR3. When this happens, BAR1 is then used for USERD mapping
and the BAR alloc() functions is disabled, making GPU objects unable
to rely on BAR for data access and falling back to PRAMIN.
Signed-off-by: Alexandre Courbot <acourbot@nvidia.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Some chips that use system memory exclusively (e.g. GK20A) do not
expose 2 BAR regions. For them only BAR1 exists, and it should be used
for USERD mapping. Do not map BAR3 if its resource does not exist.
Signed-off-by: Alexandre Courbot <acourbot@nvidia.com>
Reviewed-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Some additional patches for radeon for 3.16 now that -fixes has been merged.
- Gart fix for all asics r6xx+
- Add some VM tuning parameters
- misc fixes
* 'drm-next-3.16' of git://people.freedesktop.org/~agd5f/linux:
drm/radeon: Move fb update from radeon_flip_work_func to radeon_crtc_page_flip
drm/radeon/dpm: powertune updates for SI
Revert "drm/radeon: use variable UVD clocks"
drm/radeon: add query for number of active CUs
drm/radeon: add debugfs file to trigger GPU reset
drm/radeon: make vm_block_size a module parameter
drm/radeon: make VM size a module parameter (v2)
drm/radeon: rename alt_domain to allowed_domains
drm/radeon: use the SDMA on for buffer moves on CIK again
drm/radeon: remove range check from *_gart_set_page
drm/radeon: stop poisoning the GART TLB
drm/radeon: hdmi deep color modes must obey clock limit of sink.
drm/edid: Store all supported hdmi deep color modes in drm_display_info
drm/radeon: add missing vce init case for hawaii
drm/radeon: use lower_32_bits where appropriate
Fixes WARN()s from the DRM core since the page flip rework.
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=77521
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This caused reduced performance for some users with advanced post
processing enabled. We need a better method to pick the
UVD state based on the amount of post processing required or tune
the advanced post processing to fit within the lower power state
envelope.
This reverts commit 14a9579ddb.
Cc: "3.15" <stable@vger.kernel.org>
And also domain to prefered_domains. That matches better
what those values represent.
Signed-off-by: Christian König <christian.koenig@amd.com>
Cc: Marek Olšák <maraeo@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The underlying reason for the crashes seems to be fixed now.
Signed-off-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
We never check the return value anyway and if the
index isn't valid would crash way before calling
the functions.
Signed-off-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
When we set the valid bit on invalid GART entries they are
loaded into the TLB when an adjacent entry is loaded. This
poisons the TLB with invalid entries which are sometimes
not correctly removed on TLB flush.
For stable inclusion the patch probably needs to be modified a bit.
Signed-off-by: Christian König <christian.koenig@amd.com>
Cc: stable@vger.kernel.org
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Make sure that a hdmi deep color mode can't exceed the max tmds
clock limit of a hdmi sink if such a limit is defined by edid.
If requested deep color bpc would exceed the limit given the mode
to be set, try to degrade gracefully to lower supported deep color
bpc or to standard 8 bpc if needed.
Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
HDMI deep color setup must know which modes are supported if
it needs to degrade gracefully, as only 12 bpc / dc_36 is
guaranteed, but 10 bpc / dc_30 is optional. The maximum bpc
is not sufficient for this.
Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Ben Skeggs