- 11 Feb, 2021 40 commits
-
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
None of the chipsets we use this on have instanced engines, so this is fine. Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
This is easier to deal with in some situations than the existing accessor functions. Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
MC/FIFO will need this info as they're switched over. Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Rather than having to add new engines / engine instances to multiple places, define everything in include/nvkm/core/layout.h and use macros to generate the required plumbing. Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
This switches to using the subdev list for lookup, and otherwise should be a no-op aside from switching the function signatures. Callers will be transitioned to split type+inst individually. Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
We use subdev id bitmasks (as a u64) in a number of places, and GA100 adds enough new engine instances that we run out of bits. We could alias IDs of engines that no longer exist, but it's cleaner for a number of reasons to just split the subdev index into a subdev type, and instance ID instead. Just a lot more painful to do. This magics up the values for old-style subdev constructors, and provides a way to incrementally transition each subdev to the new style. Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
Much easier to store this to avoid having to reconstruct a string for a specific subdev, taking into account whether it's instanced or not. Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
This is somewhat nicer to read. Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
There's not really any nice way to assign the lock classes when we split subdev indices into type+inst, and saves a few bytes in the structs when a subdev has no need for it. Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-
Ben Skeggs authored
nvkm_subdev.mutex is going away. Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com>
-