aboutsummaryrefslogtreecommitdiff
path: root/drivers/gpu
AgeCommit message (Collapse)Author
2014-03-26drm/gf104/gr: rename gf104 (nvc4), it came before gf106 (nvc3)Ben Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nvc0-/graph: fix gpccs fuc stack setupBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nouveau/bios: add more checks to PRAMIN image fetchingBen Skeggs
Prevents an attempt to access VRAM on an un-posted board, which, on a particular system with a GRID K1 installed, causes a MCE and chokes the entire system. Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26support for platform devicesAlexandre Courbot
Upcoming mobile Kepler GPUs (such as GK20A) use the platform bus instead of PCI to which Nouveau is tightly dependent. This patch allows Nouveau to handle platform devices by: - abstracting PCI-dependent functions that were typically used for resource querying and page mapping, - introducing a nv_device_is_pci() function that allows to make PCI-dependent code conditional, - providing a nouveau_drm_platform_probe() function that takes a GPU platform device to be probed. Core code as well as engine/subdev drivers are updated wherever possible to make use of these functions. Some older drivers are too dependent on PCI to be properly updated, but all newer code on which future chips may depend should at least be runnable with platform devices. Signed-off-by: Alexandre Courbot <acourbot@nvidia.com> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nouveau: support modesetting on GM107Ben Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/gm100/device: recognise GM107Ben Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/gm107/disp: initial implementationBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/gm107/ltcg: initial implementationBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nouveau/bios: add HDMI-C (mini) connector typeBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/gm107/fb: initial implementationBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/gk20a/timer: initial implementationBen Skeggs
A bit different from NVIDIA's RFC patch, but I want this now for GM107. Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/gm100/bios: hw disable register has movedBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/gm107/devinit: initial implementationBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nve0/fifo: bind intrBen Skeggs
Error code names from Android GK20A driver. Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nve0/fifo: attempt to recover from engine ctxsw timeoutsBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nve0/fifo: attempt to recover engines from mmu faultsBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nve0/fifo: allow copy engine channel to be looked up by instanceBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nve0/fifo: use runlist event instead of pollingBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nve0/fifo: allow channels to be marked as unrunnableBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nve0/fifo: single printk for sched error dataBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nve0/fifo: single printk for mmu fault dataBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nve0/fifo: ack pb intr individually after handling each unitBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nve0/fifo: runlist intrBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nve0/fifo: engine intrBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nve0/fifo: mask unhandled intr bits when seen, rather than all intrsBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nvc0/fifo: attempt to recover from engine ctxsw timeoutsBen Skeggs
My test cases don't seem to trigger this on all Fermi boards, not sure if they're broken tests or it didn't work until later versions. GF119 definitely works. Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nvc0/fifo: attempt to recover engines from mmu faultsBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nvc0/fifo: use subdev identifiers for bar/ifb fault recovery casesBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nvc0/fifo: use runlist event instead of pollingBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nvc0/fifo: allow channels to be marked as unrunnableBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nvc0/fifo: sched intrBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nvc0/fifo: single printk for mmu fault dataBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nvc0/fifo: ack pb intr individually after handling each unitBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nvc0/fifo: runlist intrBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nvc0/fifo: engine intrBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nvc0/fifo: mask unhandled intr bits when seen, rather than all intrsBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nvc0/fifo: rename a couple of unitsBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nvd0/disp: add debugging to show which supervisor actions are takenBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nv50-/disp: dump channel state when update method failsBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nv50-/disp: allow dumping core channel state at first supervisor intrBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nv50-/disp: add method descriptions for debuggingBen Skeggs
Lists of known methods for the DMA channel classes, and mappings to their priv register addresses (where known). Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nv50/disp: decode the known error codes to human readable formBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nv50/disp: preparation for storing static class dataBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nouveau/core: extend width of engine mask for namedbBen Skeggs
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nouveau/bios: fix INDEX_ADDRESS_LATCHED trace printoutIlia Mirkin
Having a \n in the middle of a format string means that the next line doesn't get the prefixes unlike every other line printed by the trace. Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nv50/gr: decode texture trap status codeIlia Mirkin
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nouveau: use nv_debug for NV_DEBUG, make DRM a separate subflagIlia Mirkin
It's really confusing for NV_DEBUG's printing to be controlled via drm.debug while everything else is controlled via nouveau.debug. These messages can be turned on with nouveau.debug=DRM=debug. Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nouveau: make hdmi device finding failure prints debug levelIlia Mirkin
The hdmi device is required for runtime pm. However it is not available on many esp older devices, which were all seeing these error messages. Take this opportunity to also convert to nv_debug instead of the DRM_* messages, like the rest of nouveau does. Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nve0/fifo: allocate usermem as neededAlexandre Courbot
Memory was always allocated for 4096 channels. Change this to allocate what we actually need according to the number of channels we use. Signed-off-by: Alexandre Courbot <acourbot@nvidia.com> Reviewed-by: Thierry Reding <treding@nvidia.com> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2014-03-26drm/nouveau: handle -EACCES runtime PM return codeAlexandre Courbot
pm_runtime_get*() may return -EACCES to indicate a device does not have runtime PM enabled. This is currently the case with platform devices on Nouveau, and is not an error in that context. Handle this case without failure. Signed-off-by: Alexandre Courbot <acourbot@nvidia.com> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2026-01-19 10:06:06 +0000