Commits · 6dd123ba8a8be762c294cee293783345ed6dfec5 · Kirill Smelkov / linux

11 Feb, 2021 40 commits

drm/nouveau/mmu: switch to instanced constructor · 6dd123ba
Ben Skeggs authored Dec 04, 2020
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
```
6dd123ba
drm/nouveau/mc: use split type+inst when handling dev_top interrupts · 209ec1b8
Ben Skeggs authored Feb 06, 2021
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
```
209ec1b8
drm/nouveau/mc: use split type+inst in interrupt masking API · a35047ba
Ben Skeggs authored Feb 06, 2021
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
```
a35047ba
drm/nouveau/mc: use split type+inst in device reset APIs · 6997ea13
Ben Skeggs authored Feb 06, 2021
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
```
6997ea13
drm/nouveau/mc: lookup subdev interrupt handlers with split type+inst · c653ab83
Ben Skeggs authored Feb 04, 2021
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
```
c653ab83
drm/nouveau/mc: switch to instanced constructor · 1fc2fddf
Ben Skeggs authored Dec 04, 2020
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
```
1fc2fddf

drm/nouveau/top: expose parsed device info more directly · 3b9e93f7

Ben Skeggs authored Dec 07, 2020

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>

3b9e93f7

drm/nouveau/top: store device type and instance separately · 5e0d3dbc

Ben Skeggs authored Dec 07, 2020

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>

5e0d3dbc

drm/nouveau/ltc: switch to instanced constructor · 0afc1c4c
Ben Skeggs authored Dec 04, 2020
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
```
0afc1c4c
drm/nouveau/instmem: switch to instanced constructor · d9691a22
Ben Skeggs authored Dec 04, 2020
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
```
d9691a22
drm/nouveau/iccsense: switch to instanced constructor · 4dea1a96
Ben Skeggs authored Dec 04, 2020
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
```
4dea1a96
drm/nouveau/ibus: switch to instanced constructor · c5f38d67
Ben Skeggs authored Dec 04, 2020
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
```
c5f38d67
drm/nouveau/i2c: switch to instanced constructor · c6ce0861
Ben Skeggs authored Dec 04, 2020
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
```
c6ce0861
drm/nouveau/gsp: switch to instanced constructor · b240b212
Ben Skeggs authored Dec 04, 2020
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
```
b240b212
drm/nouveau/gpio: switch to instanced constructor · 01055c01
Ben Skeggs authored Dec 04, 2020
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
```
01055c01
drm/nouveau/fuse: switch to instanced constructor · 8d056d99
Ben Skeggs authored Dec 04, 2020
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
```
8d056d99
drm/nouveau/fb: switch to instanced constructor · b7a9369a
Ben Skeggs authored Dec 04, 2020
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
```
b7a9369a
drm/nouveau/fault: switch to instanced constructor · 77689f1b
Ben Skeggs authored Dec 04, 2020
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
```
77689f1b
drm/nouveau/devinit: switch to instanced constructor · 4a34fd0e
Ben Skeggs authored Dec 04, 2020
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
```
4a34fd0e
drm/nouveau/clk: switch to instanced constructor · 98fd7f83
Ben Skeggs authored Dec 04, 2020
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
```
98fd7f83
drm/nouveau/bus: switch to instanced constructor · d37766e5
Ben Skeggs authored Dec 04, 2020
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
```
d37766e5
drm/nouveau/bios: switch to instanced constructor · e07f50d3
Ben Skeggs authored Dec 04, 2020
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
```
e07f50d3
drm/nouveau/bar: switch to instanced constructor · 917b24a3
Ben Skeggs authored Dec 04, 2020
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
```
917b24a3
drm/nouveau/acr: switch to instanced constructor · c288b4de
Ben Skeggs authored Dec 02, 2020
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
```
c288b4de

drm/nouveau/nvkm: add macros for subdev layout · f483253f

Ben Skeggs authored Dec 06, 2020

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>

f483253f

drm/nouveau/device: pass instance id when looking up a subdev/engine · efe2a9ec

Ben Skeggs authored Feb 04, 2021

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>

efe2a9ec

drm/nouveau/subdev: track type+instance separately · 65a279c1

Ben Skeggs authored Jul 25, 2020

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>

65a279c1

drm/nouveau/subdev: store full subdev name in struct · 9c28abb7

Ben Skeggs authored Jul 25, 2020

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>

9c28abb7

drm/nouveau/subdev: store subdevs in list · 54d10db1

Ben Skeggs authored Dec 01, 2020

This is somewhat nicer to read.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>

54d10db1

drm/nouveau/subdev: remove nvkm_subdev.mutex · 149a23b0

Ben Skeggs authored Dec 02, 2020

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>

149a23b0

drm/nouveau/perfmon: use private spinlock to control exclusive access to perfmon · f8106922
Ben Skeggs authored Dec 03, 2020
```
nvkm_subdev.mutex is going away.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
```
f8106922

drm/nouveau/fifo: private mutex · a6419360

Ben Skeggs authored Dec 03, 2020

nvkm_subdev.mutex is going away.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>

a6419360

drm/nouveau/disp: use private spinlock to control exclusive access to disp · bfa7f6a6

Ben Skeggs authored Dec 03, 2020

nvkm_subdev.mutex is going away.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>

bfa7f6a6

drm/nouveau/pmu: serialise send() with private mutex · 5a479d45

Ben Skeggs authored Dec 03, 2020

nvkm_subdev.mutex is going away.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>

5a479d45

drm/nouveau/mmu: serialise mmu invalidations with private mutex · 5ec69c91

Ben Skeggs authored Dec 03, 2020

nvkm_subdev.mutex is going away.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>

5ec69c91

drm/nouveau/ltc: serialise cbc operations with private mutex · dbffdff7

Ben Skeggs authored Dec 02, 2020

nvkm_subdev.mutex is going away.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>

dbffdff7

drm/nouveau/instmem: protect mm/lru with private mutex · e5bf9a5c

Ben Skeggs authored Dec 02, 2020

nvkm_subdev.mutex is going away.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>

e5bf9a5c

drm/nouveau/fb: protect vram mm with private mutex · 0e65ec75

Ben Skeggs authored Dec 02, 2020

nvkm_subdev.mutex is going away.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>

0e65ec75

drm/nouveau/fb: protect comptags with private mutex · f5cfbd99

Ben Skeggs authored Dec 02, 2020

nvkm_subdev.mutex is going away.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>

f5cfbd99

drm/nouveau/engine: use refcount_t + private mutex · 71ccf2a0

Ben Skeggs authored Dec 02, 2020

nvkm_subdev.mutex is going away.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>

71ccf2a0