[git pull] drm - fixes + radeon KMS (part 2)

Discussion:

(too old to reply)

Dave Airlie

2009-06-15 02:10:12 UTC

Hi Linus,

Please pull the 'drm-linus' branch from
ssh://master.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6.git drm-linus

This is big. It contains the initial TTM memory manager + ATI radeon KMS
support. Currently the KMS code is part of the DRM radeon driver however
it is very clearly separated internally from the old codepaths. We've
elected to keep the radeon KMS Kconfig build/enable under staging for now
since we may have some ABI tweaks to sort out in this release cycle,
however the code is all in the drm. I don't think this enables crap
tainting, but at least no-one will find kms by accident.

The plan is to remove this from staging area in the next kernel.

This also contains one common memory manager code bug fix, and one pci ids
fixup.

Thanks to Thomas and Jerome for doing all the cleanups.

Dave.

drivers/gpu/drm/Kconfig | 13 +
drivers/gpu/drm/Makefile | 2 +-
drivers/gpu/drm/drm_mm.c | 20 +-
drivers/gpu/drm/radeon/Kconfig | 34 +
drivers/gpu/drm/radeon/Makefile | 12 +-
drivers/gpu/drm/radeon/ObjectID.h | 578 +++
drivers/gpu/drm/radeon/atom-bits.h | 48 +
drivers/gpu/drm/radeon/atom-names.h | 100 +
drivers/gpu/drm/radeon/atom-types.h | 42 +
drivers/gpu/drm/radeon/atom.c | 1215 ++++++
drivers/gpu/drm/radeon/atom.h | 149 +
drivers/gpu/drm/radeon/atombios.h | 4785 +++++++++++++++++++++++
drivers/gpu/drm/radeon/atombios_crtc.c | 695 ++++
drivers/gpu/drm/radeon/r100.c | 1524 ++++++++
drivers/gpu/drm/radeon/r300.c | 1116 ++++++
drivers/gpu/drm/radeon/r300_reg.h | 52 +-
drivers/gpu/drm/radeon/r420.c | 223 ++
drivers/gpu/drm/radeon/r500_reg.h | 749 ++++
drivers/gpu/drm/radeon/r520.c | 234 ++
drivers/gpu/drm/radeon/r600.c | 169 +
drivers/gpu/drm/radeon/r600_reg.h | 114 +
drivers/gpu/drm/radeon/radeon.h | 793 ++++
drivers/gpu/drm/radeon/radeon_agp.c | 249 ++
drivers/gpu/drm/radeon/radeon_asic.h | 405 ++
drivers/gpu/drm/radeon/radeon_atombios.c | 1298 ++++++
drivers/gpu/drm/radeon/radeon_benchmark.c | 133 +
drivers/gpu/drm/radeon/radeon_bios.c | 390 ++
drivers/gpu/drm/radeon/radeon_clocks.c | 833 ++++
drivers/gpu/drm/radeon/radeon_combios.c | 2481 ++++++++++++
drivers/gpu/drm/radeon/radeon_connectors.c | 603 +++
drivers/gpu/drm/radeon/radeon_cs.c | 249 ++
drivers/gpu/drm/radeon/radeon_cursor.c | 252 ++
drivers/gpu/drm/radeon/radeon_device.c | 813 ++++
drivers/gpu/drm/radeon/radeon_display.c | 692 ++++
drivers/gpu/drm/radeon/radeon_drv.c | 215 +-
drivers/gpu/drm/radeon/radeon_encoders.c | 1708 ++++++++
drivers/gpu/drm/radeon/radeon_fb.c | 825 ++++
drivers/gpu/drm/radeon/radeon_fence.c | 387 ++
drivers/gpu/drm/radeon/radeon_fixed.h | 50 +
drivers/gpu/drm/radeon/radeon_gart.c | 233 ++
drivers/gpu/drm/radeon/radeon_gem.c | 287 ++
drivers/gpu/drm/radeon/radeon_i2c.c | 209 +
drivers/gpu/drm/radeon/radeon_irq_kms.c | 158 +
drivers/gpu/drm/radeon/radeon_kms.c | 295 ++
drivers/gpu/drm/radeon/radeon_legacy_crtc.c | 1276 ++++++
drivers/gpu/drm/radeon/radeon_legacy_encoders.c | 1288 ++++++
drivers/gpu/drm/radeon/radeon_mode.h | 398 ++
drivers/gpu/drm/radeon/radeon_object.c | 511 +++
drivers/gpu/drm/radeon/radeon_object.h | 45 +
drivers/gpu/drm/radeon/radeon_reg.h | 3570 +++++++++++++++++
drivers/gpu/drm/radeon/radeon_ring.c | 485 +++
drivers/gpu/drm/radeon/radeon_ttm.c | 653 +++
drivers/gpu/drm/radeon/rs400.c | 411 ++
drivers/gpu/drm/radeon/rs600.c | 324 ++
drivers/gpu/drm/radeon/rs690.c | 181 +
drivers/gpu/drm/radeon/rs780.c | 102 +
drivers/gpu/drm/radeon/rv515.c | 504 +++
drivers/gpu/drm/radeon/rv770.c | 124 +
drivers/gpu/drm/ttm/Makefile | 8 +
drivers/gpu/drm/ttm/ttm_agp_backend.c | 150 +
drivers/gpu/drm/ttm/ttm_bo.c | 1698 ++++++++
drivers/gpu/drm/ttm/ttm_bo_util.c | 561 +++
drivers/gpu/drm/ttm/ttm_bo_vm.c | 454 +++
drivers/gpu/drm/ttm/ttm_global.c | 114 +
drivers/gpu/drm/ttm/ttm_memory.c | 234 ++
drivers/gpu/drm/ttm/ttm_module.c | 50 +
drivers/gpu/drm/ttm/ttm_tt.c | 635 +++
drivers/staging/Kconfig | 2 +
include/drm/drm_pciids.h | 8 +-
include/drm/radeon_drm.h | 130 +
include/drm/ttm/ttm_bo_api.h | 618 +++
include/drm/ttm/ttm_bo_driver.h | 867 ++++
include/drm/ttm/ttm_memory.h | 153 +
include/drm/ttm/ttm_module.h | 58 +
include/drm/ttm/ttm_placement.h | 92 +
75 files changed, 41091 insertions(+), 45 deletions(-)
create mode 100644 drivers/gpu/drm/radeon/Kconfig
create mode 100644 drivers/gpu/drm/radeon/ObjectID.h
create mode 100644 drivers/gpu/drm/radeon/atom-bits.h
create mode 100644 drivers/gpu/drm/radeon/atom-names.h
create mode 100644 drivers/gpu/drm/radeon/atom-types.h
create mode 100644 drivers/gpu/drm/radeon/atom.c
create mode 100644 drivers/gpu/drm/radeon/atom.h
create mode 100644 drivers/gpu/drm/radeon/atombios.h
create mode 100644 drivers/gpu/drm/radeon/atombios_crtc.c
create mode 100644 drivers/gpu/drm/radeon/r100.c
create mode 100644 drivers/gpu/drm/radeon/r300.c
create mode 100644 drivers/gpu/drm/radeon/r420.c
create mode 100644 drivers/gpu/drm/radeon/r500_reg.h
create mode 100644 drivers/gpu/drm/radeon/r520.c
create mode 100644 drivers/gpu/drm/radeon/r600.c
create mode 100644 drivers/gpu/drm/radeon/r600_reg.h
create mode 100644 drivers/gpu/drm/radeon/radeon.h
create mode 100644 drivers/gpu/drm/radeon/radeon_agp.c
create mode 100644 drivers/gpu/drm/radeon/radeon_asic.h
create mode 100644 drivers/gpu/drm/radeon/radeon_atombios.c
create mode 100644 drivers/gpu/drm/radeon/radeon_benchmark.c
create mode 100644 drivers/gpu/drm/radeon/radeon_bios.c
create mode 100644 drivers/gpu/drm/radeon/radeon_clocks.c
create mode 100644 drivers/gpu/drm/radeon/radeon_combios.c
create mode 100644 drivers/gpu/drm/radeon/radeon_connectors.c
create mode 100644 drivers/gpu/drm/radeon/radeon_cs.c
create mode 100644 drivers/gpu/drm/radeon/radeon_cursor.c
create mode 100644 drivers/gpu/drm/radeon/radeon_device.c
create mode 100644 drivers/gpu/drm/radeon/radeon_display.c
create mode 100644 drivers/gpu/drm/radeon/radeon_encoders.c
create mode 100644 drivers/gpu/drm/radeon/radeon_fb.c
create mode 100644 drivers/gpu/drm/radeon/radeon_fence.c
create mode 100644 drivers/gpu/drm/radeon/radeon_fixed.h
create mode 100644 drivers/gpu/drm/radeon/radeon_gart.c
create mode 100644 drivers/gpu/drm/radeon/radeon_gem.c
create mode 100644 drivers/gpu/drm/radeon/radeon_i2c.c
create mode 100644 drivers/gpu/drm/radeon/radeon_irq_kms.c
create mode 100644 drivers/gpu/drm/radeon/radeon_kms.c
create mode 100644 drivers/gpu/drm/radeon/radeon_legacy_crtc.c
create mode 100644 drivers/gpu/drm/radeon/radeon_legacy_encoders.c
create mode 100644 drivers/gpu/drm/radeon/radeon_mode.h
create mode 100644 drivers/gpu/drm/radeon/radeon_object.c
create mode 100644 drivers/gpu/drm/radeon/radeon_object.h
create mode 100644 drivers/gpu/drm/radeon/radeon_reg.h
create mode 100644 drivers/gpu/drm/radeon/radeon_ring.c
create mode 100644 drivers/gpu/drm/radeon/radeon_ttm.c
create mode 100644 drivers/gpu/drm/radeon/rs400.c
create mode 100644 drivers/gpu/drm/radeon/rs600.c
create mode 100644 drivers/gpu/drm/radeon/rs690.c
create mode 100644 drivers/gpu/drm/radeon/rs780.c
create mode 100644 drivers/gpu/drm/radeon/rv515.c
create mode 100644 drivers/gpu/drm/radeon/rv770.c
create mode 100644 drivers/gpu/drm/ttm/Makefile
create mode 100644 drivers/gpu/drm/ttm/ttm_agp_backend.c
create mode 100644 drivers/gpu/drm/ttm/ttm_bo.c
create mode 100644 drivers/gpu/drm/ttm/ttm_bo_util.c
create mode 100644 drivers/gpu/drm/ttm/ttm_bo_vm.c
create mode 100644 drivers/gpu/drm/ttm/ttm_global.c
create mode 100644 drivers/gpu/drm/ttm/ttm_memory.c
create mode 100644 drivers/gpu/drm/ttm/ttm_module.c
create mode 100644 drivers/gpu/drm/ttm/ttm_tt.c
create mode 100644 include/drm/ttm/ttm_bo_api.h
create mode 100644 include/drm/ttm/ttm_bo_driver.h
create mode 100644 include/drm/ttm/ttm_memory.h
create mode 100644 include/drm/ttm/ttm_module.h
create mode 100644 include/drm/ttm/ttm_placement.h

commit 6fe7ac3f5b544703581f3829c8c950dc721d976e
Author: Alex Deucher <***@gmail.com>
Date: Fri Jun 12 17:26:08 2009 +0000

radeon legacy chips: tv dac bg/dac adj updates

COMBIOS - fallback to table values if there are no tv dac or lvds bios tables
ATOMBIOS - add support for looking up these values from the bios table

Signed-off-by: Alex Deucher <***@gmail.com>
Signed-off-by: Dave Airlie <***@redhat.com>

commit 771fe6b912fca54f03e8a72eb63058b582775362
Author: Jerome Glisse <***@redhat.com>
Date: Fri Jun 5 14:42:42 2009 +0200

drm/radeon: introduce kernel modesetting for radeon hardware

Add kernel modesetting support to radeon driver, use the ttm memory
manager to manage memory and DRM/GEM to provide userspace API.
In order to avoid backward compatibility issue and to allow clean
design and code the radeon kernel modesetting use different code path
than old radeon/drm driver.

When kernel modesetting is enabled the IOCTL of radeon/drm
driver are considered as invalid and an error message is printed
in the log and they return failure.

KMS enabled userspace will use new API to talk with the radeon/drm
driver. The new API provide functions to create/destroy/share/mmap
buffer object which are then managed by the kernel memory manager
(here TTM). In order to submit command to the GPU the userspace
provide a buffer holding the command stream, along this buffer
userspace have to provide a list of buffer object used by the
command stream. The kernel radeon driver will then place buffer
in GPU accessible memory and will update command stream to reflect
the position of the different buffers.

The kernel will also perform security check on command stream
provided by the user, we want to catch and forbid any illegal use
of the GPU such as DMA into random system memory or into memory
not owned by the process supplying the command stream. This part
of the code is still incomplete and this why we propose that patch
as a staging driver addition, future security might forbid current
experimental userspace to run.

This code support the following hardware : R1XX,R2XX,R3XX,R4XX,R5XX
(radeon up to X1950). Works is underway to provide support for R6XX,
R7XX and newer hardware (radeon from HD2XXX to HD4XXX).

Authors:
Jerome Glisse <***@redhat.com>
Dave Airlie <***@redhat.com>
Alex Deucher <***@gmail.com>

Signed-off-by: Jerome Glisse <***@redhat.com>
Signed-off-by: Dave Airlie <***@redhat.com>
Signed-off-by: Alex Deucher <***@gmail.com>
Signed-off-by: Dave Airlie <***@redhat.com>

commit ba4e7d973dd09b66912ac4c0856add8b0703a997
Author: Thomas Hellstrom <***@vmware.com>
Date: Wed Jun 10 15:20:19 2009 +0200

drm: Add the TTM GPU memory manager subsystem.

TTM is a GPU memory manager subsystem designed for use with GPU
devices with various memory types (On-card VRAM, AGP,
PCI apertures etc.). It's essentially a helper library that assists
the DRM driver in creating and managing persistent buffer objects.

TTM manages placement of data and CPU map setup and teardown on
data movement. It can also optionally manage synchronization of
data on a per-buffer-object level.

TTM takes care to provide an always valid virtual user-space address
to a buffer object which makes user-space sub-allocation of
big buffer objects feasible.

TTM uses a fine-grained per buffer-object locking scheme, taking
care to release all relevant locks when waiting for the GPU.
Although this implies some locking overhead, it's probably a big
win for devices with multiple command submission mechanisms, since
the lock contention will be minimal.

TTM can be used with whatever user-space interface the driver
chooses, including GEM. It's used by the upcoming Radeon KMS DRM driver
and is also the GPU memory management core of various new experimental
DRM drivers.

Signed-off-by: Thomas Hellstrom <***@vmware.com>
Signed-off-by: Jerome Glisse <***@redhat.com>
Signed-off-by: Dave Airlie <***@redhat.com>

commit e6c03c5b40314d787f7053f631594d6b1bd609e8
Author: Chris Wilson <***@chris-wilson.co.uk>
Date: Fri May 22 14:14:22 2009 +0100

drm: Memory fragmentation from lost alignment blocks

If the block needs an alignment but otherwise fits exactly into the tail,
then the split-off block from the start would remain marked as non-free.

Signed-off-by: Chris Wilson <***@chris-wilson.co.uk>
Acked-by: Thomas Hellstrom <***@vmware.com>
Signed-off-by: Dave Airlie <***@redhat.com>

commit 602c11a8ee62d49cddbc5972e5edb876dd415113
Author: Dave Airlie <***@redhat.com>
Date: Mon Jun 15 09:24:41 2009 +1000

drm/radeon: fix mobility flags on new PCI IDs.

These aren't used that much on r600, but may be needed in the future,
so get them correct now.

Signed-off-by: Dave Airlie <***@redhat.com>

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Greg KH

2009-06-15 02:30:18 UTC

Permalink

Post by Dave Airlie
Hi Linus,
Please pull the 'drm-linus' branch from
ssh://master.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6.git drm-linus
This is big. It contains the initial TTM memory manager + ATI radeon KMS
support. Currently the KMS code is part of the DRM radeon driver however
it is very clearly separated internally from the old codepaths. We've
elected to keep the radeon KMS Kconfig build/enable under staging for now
since we may have some ABI tweaks to sort out in this release cycle,
however the code is all in the drm. I don't think this enables crap
tainting, but at least no-one will find kms by accident.

No, the module loader looks for stuff in drivers/staging/ to cause a
"taint".

But why not just keep the Kconfig stuff in your own directory, and
depend on CONFIG_STAGING if you want to not have it show up for "normal"
users? It seems odd to put anything in drivers/staging/Kconfig for
something that is not in drivers/staging.

I'm guessing this Kconfig change was not in linux-next? Or had it been
there and I just missed it somehow?

thanks,

greg k-h
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Dave Airlie

2009-06-15 02:50:08 UTC

Permalink

Post by Greg KH

No since I was on holidays until the just before the merge window
opened, the patches
did get posted to lkml but missed your cc.

Well we'd like to make sure people go via the staging menus to get at
the kconfig option
for now, granted it probably doesn't matter whether it goes in staging
menus or in drm depends
on CONFIG_STAGING. Up to you I can post a patch after this merge to
move it to drm.

Since its only temporary I'm not hugely pushed about taint.

Dave.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Ryan Hope

2009-06-16 16:00:20 UTC

Permalink

Dave, you mentioned that you were going to pull in the TTM Kconfig fix
that I submitted (the one that allows TTM to be built into the kernel
instead of just a module). What are you going to do about the hardlock
that occurs when drm/radeon/fbcon are all built into the kernel. It
seems like fbcon does not get initialized before radeon and as a
result, there is a hardlock. My fix was to init radeon later which
works fine for me. Maybe it would be better to init fbcon earlier if
possible? Either way, I am not sure if you missed that patch so I am
going to resubmit it here as a reminder that this bug exists if you
build everything into the kernel..

diff --git a/85adafcf5d18e58c60d9fdbb718abe9149661736:drivers/gpu/drm/radeon/radeon_drv.c
b/093f1cb1a1850951f86244ada018672885ec4ace:drivers/gpu/drm/radeon/radeon_drv.c
index f70c351bb43e5770b9d18f76ad2a6fc1a2c7d1d4..ff4e052f997deaaa87081069d17d2026d925eabb
100644
--- a/85adafcf5d18e58c60d9fdbb718abe9149661736:drivers/gpu/drm/radeon/radeon_drv.c
+++ b/093f1cb1a1850951f86244ada018672885ec4ace:drivers/gpu/drm/radeon/radeon_drv.c
@@ -345,7 +345,11 @@ static void __exit radeon_exit(void)
drm_exit(driver);
}

+#ifdef MODULE
module_init(radeon_init);
+#else
+late_initcall(radeon_init);
+#endif
module_exit(radeon_exit);

MODULE_AUTHOR(DRIVER_AUTHOR);

Post by Dave Airlie

Post by Greg KH

No since I was on holidays until the just before the merge window
opened, the patches
did get posted to lkml but missed your cc.
Well we'd like to make sure people go via the staging menus to get at
the kconfig option
for now, granted it probably doesn't matter whether it goes in staging
menus or in drm depends
on CONFIG_STAGING. Up to you I can post a patch after this merge to
move it to drm.
Since its only temporary I'm not hugely pushed about taint.
Dave.
------------------------------------------------------------------------------
Crystal Reports - New Free Runtime and 30 Day Trial
Check out the new simplified licensing option that enables unlimited
royalty-free distribution of the report engine for externally facing
server and web deployment.
http://p.sf.net/sfu/businessobjects
--
_______________________________________________
Dri-devel mailing list
https://lists.sourceforge.net/lists/listinfo/dri-devel

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Linus Torvalds

2009-06-16 18:30:12 UTC

Permalink

Post by Ryan Hope
+#ifdef MODULE
module_init(radeon_init);
+#else
+late_initcall(radeon_init);
+#endif

You should never need something like that.

Just do

late_initcall(radeon_init);

and if it's a module (which doesn't have "early" vs "late" etc), it will
just be a normal module_init.

Linus
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Dave Airlie

2009-06-16 20:50:10 UTC

Permalink

On Wed, Jun 17, 2009 at 4:20 AM, Linus

Post by Linus Torvalds

Post by Ryan Hope
+#ifdef MODULE
module_init(radeon_init);
+#else
+late_initcall(radeon_init);
+#endif

You should never need something like that.
Just do
late_initcall(radeon_init);

I actually don't think we need either of them, the ordering shuoldn't
matter, its just a bug we need to fix
in the radeon code during staging I expect, in theory you should be
able to load fbcon before/after radeon.

I don't think the intel code suffers from this issue.

Linus can you pull this tree?

Dave.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Linus Torvalds

2009-06-16 21:00:18 UTC

Permalink

Post by Dave Airlie
Linus can you pull this tree?

I hate pulling trees when I know there are _known_ bugs.

Even during the merge window. The rest of the -rc series is for fixing up
bugs, yes, but not bugs that were introduced on purpose.

Linus
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Dave Airlie

2009-06-16 21:00:29 UTC

Permalink

On Wed, Jun 17, 2009 at 6:52 AM, Linus

Post by Linus Torvalds

Post by Dave Airlie
Linus can you pull this tree?

I hate pulling trees when I know there are _known_ bugs.
Even during the merge window. The rest of the -rc series is for fixing up
bugs, yes, but not bugs that were introduced on purpose.

It only happens if the user enable staging drivers, and I suspect staging
has much worse bugs, but okay I'll put the initcall fix into the tree
and send you
another pull request.

Dave.

Post by Linus Torvalds
Linus

Greg KH

2009-06-16 21:30:15 UTC

Permalink

Post by Dave Airlie
On Wed, Jun 17, 2009 at 6:52 AM, Linus

Post by Linus Torvalds

Post by Dave Airlie
Linus can you pull this tree?

I hate pulling trees when I know there are _known_ bugs.
Even during the merge window. The rest of the -rc series is for fixing up
bugs, yes, but not bugs that were introduced on purpose.

It only happens if the user enable staging drivers, and I suspect staging
has much worse bugs, but okay I'll put the initcall fix into the tree
and send you another pull request.

Note, distros enable the staging tree to pick up drivers they were
previously including anyway. So that is quite a large userbase :)

thanks,

greg k-h
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Greg KH

2009-06-16 19:00:15 UTC

Permalink

Post by Dave Airlie

Post by Greg KH

Ok, I'll defer to your judgement here.

thanks,

greg k-h
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/