| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
| |
These are copied directly from the mesa repository.
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Acked-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
rendercheck -f a8r8g8b8 passes as much as on a GK208, and xv appears to
work. Very lightly tested.
Instead of sticking coordinates into pushbufs, the vertex shader is
modified to read them from a constbuf, indexed by vertex id. This
approach could be used for all nvc0 generations, but I didn't want to
rock the boat.
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
|
|
|
|
|
|
|
|
| |
If you want glamor, just use modesetting instead. It's hooked up
incorrectly in nouveau, and there doesn't seem to be any point in fixing
it when the modesetting driver does it all properly.
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
|
|
|
|
|
|
| |
Until glamor grows its own implementation.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
|
|
|
|
| |
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
|
|
|
|
| |
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
|
|
|
|
| |
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
|
|
|
|
|
|
| |
As a side-effect, we now allow Kepler to use async copies too.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
|
|
|
|
|
|
| |
add all the nvf0 stuff.
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
|
|
|
|
|
|
|
| |
The DDX uses symbols from libdrm but currently does not link against
it, but rather relies on indirect linkage from libdrm_nouveau. This
could lead to insufficient dependencies in distribution packages, seen
in http://bugs.debian.org/633786.
Signed-off-by: Sven Joachim <svenjoac@gmx.de>
|
| |
|
|
|
|
| |
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
|
| |
|
|
|
|
| |
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
|
|
|
|
|
| |
Also removes A8 surface shader hack, and implements proper shaders for
the specific purpose.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
|
|
|
|
| |
calim: nv50,nvc0/xv: fix src_w/h width/height confusion
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Squashed commit of the following:
commit e8ce1ef351ee51bec51211dca6fda88fdbfbefee
Author: Ben Skeggs <bskeggs@redhat.com>
Date: Mon Jan 17 08:55:04 2011 +1000
fix hardcoding of nvc0 wfb path
commit b414b1f948da18d96309bad84c54948d877efaa6
Author: Ben Skeggs <bskeggs@redhat.com>
Date: Mon Jan 17 08:52:53 2011 +1000
general tidy-ups in preparation for merge to master
commit 73bf71d5de0cb990178a1d2a31e0ea1717c214a4
Merge: fb499a4 b795ca6
Author: Ben Skeggs <bskeggs@redhat.com>
Date: Mon Jan 17 08:47:40 2011 +1000
Merge branch 'master' into nvc0
commit fb499a4e9d95650dc89f4c1820b94d01344733f6
Author: Ben Skeggs <bskeggs@redhat.com>
Date: Mon Dec 20 12:54:43 2010 +1000
nvc0: switch to "standard" RING macros
commit 64f0e1d0e9e29c5437ad4e52d01ccfb0b03a2ae9
Merge: eba7e17 c2092a4
Author: Ben Skeggs <bskeggs@redhat.com>
Date: Mon Dec 20 11:58:42 2010 +1000
Merge branch 'master' into nvc0
Conflicts:
src/nv_dma.c
commit eba7e173a5006ff50b950d0e8b8154564471a9ce
Author: Ben Skeggs <bskeggs@redhat.com>
Date: Fri Dec 17 11:04:40 2010 +1000
nvc0/xv: use mad in nv12 shader, rather than mul+add
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
commit b025ceb40a5e94a3657f3c7f76044c49a1a05e8d
Author: Christoph Bumiller <e0425955@student.tuwien.ac.at>
Date: Thu Dec 16 23:34:29 2010 +0100
nvc0: fix offset of 17bc buffer
I had messed it up in 5e4da3451c3c0f645f98fea7fb0ce2629b1bfd2f.
commit 6ff5a9a7f3411ef218f17fa481c583596ffde2cc
Author: Christoph Bumiller <e0425955@student.tuwien.ac.at>
Date: Thu Dec 16 01:07:25 2010 +0100
nvc0/xv: bind textures for the FP, not the TEP
commit 5e4da3451c3c0f645f98fea7fb0ce2629b1bfd2f
Author: Christoph Bumiller <e0425955@student.tuwien.ac.at>
Date: Thu Dec 16 01:02:58 2010 +0100
nvc0: switch to rnn headers
commit 0839041e79dbb0e629326d0f240220b33921db1a
Author: Ben Skeggs <bskeggs@redhat.com>
Date: Wed Dec 15 16:42:41 2010 +1000
nvc0/exa: fix m2mf dfs
commit 88c1c43302331e1cb8b7035bfb4677a5e7da3090
Author: Ben Skeggs <bskeggs@redhat.com>
Date: Mon Dec 13 12:00:23 2010 +1000
nvc0: add Xv implementation
commit ad5b4cfe388a102a0c93f7281cb4bba532b1796b
Merge: 1942a81 b6cb21b
Author: Ben Skeggs <bskeggs@redhat.com>
Date: Wed Dec 15 14:57:04 2010 +1000
Merge branch 'master' into nvc0
commit 1942a81c0eb592972d4a3e8c46b80192ae8d1a59
Author: Ben Skeggs <bskeggs@redhat.com>
Date: Thu Dec 9 12:48:27 2010 +1000
nvc0: call grobj_alloc for all used object classes
Though the hardware has no concept of object handles any more, the nouveau
DRM still uses this call to know what engines are in use.
commit 0c670a6d406fe0a57373fa9e03005be6f28ecc3e
Author: Ben Skeggs <bskeggs@redhat.com>
Date: Thu Dec 9 12:44:41 2010 +1000
nvc0: don't try and create DmaNotifier0
commit 5d3d20d908a50ad7742567d58969c7f12dd944f7
Author: Christoph Bumiller <e0425955@student.tuwien.ac.at>
Date: Thu Dec 9 01:36:32 2010 +0100
nvc0: change tile_mode to contain 0xZYX instead of 0xZY
Warning: the kernel will not like that (EVO)
commit dd7fd8263e0e9f92412fb20086e4c8acf8cab099
Author: Christoph Bumiller <e0425955@student.tuwien.ac.at>
Date: Thu Dec 9 00:29:58 2010 +0100
exa: don't fall back to nv04 m2mf if nvc0 m2mf fails
commit b4cf5a8f2276392945c2530f0d6fe96b5ab4b5da
Author: Christoph Bumiller <e0425955@student.tuwien.ac.at>
Date: Thu Dec 9 00:07:59 2010 +0100
nvc0/exa: remove noisy debug messages
commit 5419ef6d2cfabb7329982cb8dbdb5bbc8e973fe3
Author: Christoph Bumiller <e0425955@student.tuwien.ac.at>
Date: Wed Dec 8 23:55:51 2010 +0100
nvc0/accel: remove unneeded scratch method 0x3420 init
It was intended be read by a MACRO method to get the address for
constants upload.
commit 178356391ff6831599eba257a5912079894641a1
Merge: 8573280 8bb8231
Author: Christoph Bumiller <e0425955@student.tuwien.ac.at>
Date: Thu Dec 9 01:31:40 2010 +0100
Merge remote branch 'origin/master' into nvc0
Conflicts:
src/drmmode_display.c
src/nouveau_exa.c
src/nv_driver.c
src/nv_type.h
commit 8573280871401e29f27fa24d0b7c6ac5e35b2cc1
Author: Christoph Bumiller <e0425955@student.tuwien.ac.at>
Date: Sat Jul 3 14:27:36 2010 +0200
nvc0: initial implementation
EXA should work, Xv not yet done.
|
|
|
|
|
|
|
| |
When a hotplug event is received from the kernel we should notify the
client side to reconfigure the display.
Ported to nouveau from radeon driver, based on work by ajax in intel driver.
|
| |
|
|
|
|
|
|
|
| |
- remove README.NV1 reference from Makefile.am
- add nv_rop.h to src/Makefile.am
Signed-off-by: Francisco Jerez <currojerez@riseup.net>
|
| |
|
|
|
|
| |
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
|
|
|
|
|
|
|
|
|
| |
It's likely there's a lot more that can be cleaned up/removed as a result
of this, we can handle that as we come across it.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Acked-by: Maarten Maathuis <madman2003@gmail.com>
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
New nv_crtc->state shortcut and some renaming included for convenience
|
|
|
|
|
|
| |
Pull in pll setting (common, no reason to bloat nv_bios with it), state
load/unload from nv_crtc, and useful bits of nv_hw.c (remnants in nv_hw.c
are only used on the non-randr12 path)
|
| |
|
|
|
|
| |
Make bios->chip_version public so the pll calcs can use it
|
|
|
|
| |
Also, no DRI1 when driver pixmaps enabled.
|
| |
|
| |
|
|
|
|
|
|
|
|
|
| |
It actually still worked just fine, but it was wrong. It's possible we'll
have more than 1 clipbox, in which case the check for enough space in the
pushbuf for a single frame is insufficient.
Check for enough space just before emitting quads, if there's not enough
flush and reemit state - then go again.
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The NV12 YCbCr->RGB shader is hand-written just like the composite shaders,
hopefully correct! I'll probably put a disasm next to each opcode at some
point so it doesn't look like just a bunch of magics.
You'll need to use a composite manager if you want to watch videos, we can't
render directly to the screen for the same reasons as mentioned in the
composite commit. Though, a simple workaround is possible in the Xv case,
so that'll probably come at some point. For now, PutImage() will fail with
BadMatch.
The Y and UV plane upload may not be 100% for what NV50 requires, I want to
rework all that at some point anyway. For now, at least there's some form
of accelerated video to work with.
Enjoy :)
|
| |
|
|
|
|
| |
- This will get a cleanup later on.
|
|
|
|
|
|
|
|
|
| |
- There is now interface independent code for most things.
- This structure, although not exclusive, is only used for NV50 atm.
- The change in approach is that the concept of a connector is added to the model (randr12 sucks in this aspect).
- I'm hacking around randr trying to create a single output for each connector, the indices are the bus numbers.
- Limitations: Hot swapping vga and dvi of same monitor will not make randr12 think it changed, so for the moment force a change.
- Concept is largely inspired by the radeonhd driver.
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
| |
- Only supports YV12.
- No attributes.
- Only the Y component gets bilinear filtering.
- You're welcome to improve it ;-)
|
|
|
|
|
|
|
|
|
|
| |
Implemented on top of the current MM interface, and currently used in the
exact same way NVAllocate/FreeMemory was used. Can transition to something
more ttm-friendly gradually.
The "old mm" bo backend is extremely primitive. Buffers can only be created
in mappable memory, forcibly pinned, and never moved once they're created.
Some of these things could be improved, but probably not much point.
|