Commits · 0b33c936599d75b8d8ff6868fa0cbd5676d88b89 · nexedi / linux

23 Jun, 2011 40 commits
- drm/nvc0: push prunk140 irq messages to debug loglevel · 0b33c936
  Ben Skeggs authored Jun 19, 2011
```
We know they happen, we don't know why.  They're annoying, so hide them
from users for the moment.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  0b33c936
- drm/nouveau: un-blacklist nvce accel · 63305de7
  Ben Skeggs authored Jun 17, 2011
```
Reported working on IRC.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  63305de7
- drm/nouveau: fix null pointer deref on pre-nv50 chipsets · bf08bcc6
  Ben Skeggs authored Jun 13, 2011
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  bf08bcc6
- drm/nouveau: rework vram init/fini ordering a little · 24f246ac
  Ben Skeggs authored Jun 10, 2011
```
Commit "drm/nouveau: add some debug output if nouveau_mm busy at destroy time"
revealed an issue where vram mm takedown would actually fail due to there
still being nodes present, causing nouveau to leak a small amount of memory
on module unload.

This splits TTM/nouveau_mm a bit more cleanly and ensures nouveau_mm fini
isn't done until all gpuobjs are also destroyed.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  24f246ac
- drm/nouveau: shut lockdep up if last vm ref needs to destroy pgd · 15ba79ad
  Ben Skeggs authored Jun 08, 2011
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  15ba79ad
- drm/nouveau: fix display takedown order to match reverse init order · 06b75e35
  Ben Skeggs authored Jun 08, 2011
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  06b75e35
- drm/nvc0: enable per-client address spaces · 5de8037a
  Ben Skeggs authored Jun 08, 2011
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  5de8037a
- drm/nouveau: add some debug output if nouveau_mm busy at destroy time · ad9ac437
  Ben Skeggs authored Jun 07, 2011
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  ad9ac437
- drm/nv50: enable use of per-client gpu address space · e41f26e7
  Ben Skeggs authored Jun 07, 2011
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  e41f26e7
- drm/nouveau: remove implicit mapping of every bo into chan_vm · 2fd3db6f
  Ben Skeggs authored Jun 07, 2011
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  2fd3db6f
- drm/nouveau: remove 'chan' argument from nouveau_bo_new · 7375c95b
  Ben Skeggs authored Jun 07, 2011
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  7375c95b
- drm/nouveau: fixup gem_info ioctl to return client-specific bo virtual · e758a311
  Ben Skeggs authored Jun 07, 2011
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  e758a311
- drm/nvc0: explicitly map PDISP semaphore buffer into each channel's vm · 3d483d57
  Ben Skeggs authored Jun 07, 2011
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  3d483d57
- drm/nv50-nvc0: lookup pushbuf virtual address on dma_push · 9f9f51fc
  Ben Skeggs authored Jun 07, 2011
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  9f9f51fc
- drm/nv84-nvc0: explicitly map semaphore buffer into channel vm · d02836b4
  Ben Skeggs authored Jun 07, 2011
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  d02836b4
- drm/nv50-nvc0: explicitly map pushbuf bo into channel vm · ce163f69
  Ben Skeggs authored Jun 07, 2011
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  ce163f69
- drm/nv50-nvc0: explicitly map notifier bo into channel vm · 0b718733
  Ben Skeggs authored Jun 07, 2011
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  0b718733
- drm/nv50-nvc0: explicitly map fbcon fb into channel vm · 45143cb5
  Ben Skeggs authored Jun 07, 2011
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  45143cb5
- drm/nouveau: initial changes to support multiple VMAs per buffer object · fd2871af
  Ben Skeggs authored Jun 06, 2011
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  fd2871af
- drm/nv50-nvc0: completely disable relocs · a3fcd0a9
  Ben Skeggs authored Jun 07, 2011
```
GPU virtual addresses are constant now so this should never be getting hit
anyway and userspace shouldn't break from them being ignored.

This is being done in preference to teaching the code how to deal with BOs
that exist at different virtual addresses within separate VMs.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  a3fcd0a9
- drm/nouveau: convert bo.mem.start usage to bo.offset · 180cc306
  Ben Skeggs authored Jun 07, 2011
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  180cc306
- drm/nouveau: convert some bo.offset use to vma.offset · 07533ea5
  Ben Skeggs authored Jun 07, 2011
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  07533ea5
- drm/nouveau: create temp vmas for both src and dst of bo moves · d2f96666
  Ben Skeggs authored Jun 06, 2011
```
Greatly simplifies a number of things, particularly once per-client GPU
address spaces are involved.

May add this back later once I know what things'll look like.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  d2f96666
- drm/nouveau: store bo's page size in nouveau_bo · f91bac5b
  Ben Skeggs authored Jun 06, 2011
```
Was previously assuming a page size of 4KiB unless a VMA was present to
override it.  Eventually, a buffer won't necessarily have a VMA at all at
some stages of its life, so we need to store this info elsewhere.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  f91bac5b
- drm/nouveau: skip move_notify() if bo does not have a vma attached · 111af5c1
  Ben Skeggs authored Jun 03, 2011
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  111af5c1
- drm/nouveau: will need to specify channel for vm-ful gpuobj allocations · 6e32fedc
  Ben Skeggs authored Jun 03, 2011
```
Abuses existing gpuobj_new() chan argument for this, which in turn forces
all NVOBJ_FLAG_VM allocations to be done from the global heap, not
suballocated from the channel's private heap.  Not a problem though in
practise.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  6e32fedc
- drm/nouveau: initialise any vm for a channel before pushbuf/ntfy · dd6a46cc
  Ben Skeggs authored Jun 03, 2011
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  dd6a46cc
- drm/nouveau: modify gpuobj/ntfy takedown ordering · b7cb6c01
  Ben Skeggs authored Jun 06, 2011
```
gpuobj really needs splitting into channel/gpuobj code instead...
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  b7cb6c01
- drm/nouveau/gem: implement stub hooks for GEM object open/close · 639212d0
  Ben Skeggs authored Jun 03, 2011
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  639212d0
- drm/nouveau: remove 'chan' argument from nouveau_gem_new · f6d4e621
  Ben Skeggs authored Jun 07, 2011
```
Userspace hasn't passed us a channel_hint for a long long time now, and
there isn't actually a need to do so anymore anyway.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  f6d4e621
- drm/nv50-nvc0/chan: inherit vm from fpriv, rather than chan_vm directly · 0320d791
  Ben Skeggs authored Jun 03, 2011
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  0320d791
- drm/nv50-nvc0/vm: take client reference on shared channel vm · fe32b16e
  Ben Skeggs authored Jun 03, 2011
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  fe32b16e
- drm/nv50-nvc0/vm: don't touch chan_vm · b79181cb
  Ben Skeggs authored Jun 03, 2011
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  b79181cb
- drm/nouveau: no need to update bo.offset from vma after validate · d359d51c
  Ben Skeggs authored Jun 03, 2011
```
On chipsets using nouveau_vm, the virtual address stays constant, so
the value set at bo creation time is fine.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  d359d51c
- drm/nouveau: store a per-client channel list · e8a863c1
  Ben Skeggs authored Jun 01, 2011
```
Removes the need to disable IRQs to lookup channel struct on every pushbuf
ioctl, among others.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  e8a863c1
- drm/nouveau: use NULL file_priv for DRM-created channels · f8656f0b
  Ben Skeggs authored May 31, 2011
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  f8656f0b
- drm/nouveau: allocate structure to store per-client data · 3f0a68d8
  Ben Skeggs authored May 31, 2011
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  3f0a68d8
- drm/nouveau: silence error for missing dac loadval table · 1562ffde
  Ben Skeggs authored Jun 01, 2011
```
There's lots of boards (all recent ones) that don't have this anymore, so
punt the message to debug loglevel.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  1562ffde
- drm/nv50: DCB table quirks for another busted XFX board · e540afc3
  Ben Skeggs authored May 30, 2011
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  e540afc3
- drm/nouveau: log if accel is disabled by default on a chipset · ad830d23
  Ben Skeggs authored May 27, 2011
```
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
```
  ad830d23