GitHub/MotorolaMobilityLLC/kernel-slsi.git
13 years agodrm/nv50/pm: rewrite clock management, and switch to the new pm hooks
Ben Skeggs [Tue, 25 Oct 2011 23:11:02 +0000 (09:11 +1000)]
drm/nv50/pm: rewrite clock management, and switch to the new pm hooks

This area is horrifically complicated on these chipsets, and it's likely we
will need at least a few more tweaks yet.

Oh yes, and it's completely disabled on IGPs for the moment.  From traces,
things look potentially different there yet again.  Sigh...

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nv50/pm: s/PLL_UNK05/PLL_VDEC/
Martin Peres [Thu, 6 Oct 2011 21:47:58 +0000 (23:47 +0200)]
drm/nv50/pm: s/PLL_UNK05/PLL_VDEC/

Following to "drm/nv50/pm: s/unk05/vdec/", let's rename the PLL to PLL_VDEC

PLL names are purely indicative and are based on the most important engine
it clocks.

Signed-off-by: Martin Peres <martin.peres@labri.fr>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nouveau/pm: make clocks_set return an error code clocks_set can fail.
Martin Peres [Sat, 9 Jul 2011 22:08:41 +0000 (00:08 +0200)]
drm/nouveau/pm: make clocks_set return an error code clocks_set can fail.

Reporting an error is better than silently refusing to reclock.

V2: Use the same logic on nv40

Signed-off-by: Martin Peres <martin.peres@labri.fr>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nvd0: read temperature as we did on nv84+ boards
Martin Peres [Fri, 21 Oct 2011 23:40:40 +0000 (01:40 +0200)]
drm/nvd0: read temperature as we did on nv84+ boards

Signed-off-by: Martin Peres <martin.peres@labri.fr>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nv50/disp: fix scaling of doublescan modes
Ben Skeggs [Tue, 25 Oct 2011 06:29:13 +0000 (16:29 +1000)]
drm/nv50/disp: fix scaling of doublescan modes

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nv50/disp: rewrite crtc timing calculation, with proper names and fixes
Ben Skeggs [Thu, 20 Oct 2011 05:00:22 +0000 (15:00 +1000)]
drm/nv50/disp: rewrite crtc timing calculation, with proper names and fixes

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nouveau/disp: kill off nouveau_crtc.mode
Ben Skeggs [Thu, 20 Oct 2011 04:25:53 +0000 (14:25 +1000)]
drm/nouveau/disp: kill off nouveau_crtc.mode

This hasn't been necessary for a long time now..

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nv50/disp: allow interlaced and doublescan modes on digital outputs
Ben Skeggs [Wed, 19 Oct 2011 03:06:48 +0000 (13:06 +1000)]
drm/nv50/disp: allow interlaced and doublescan modes on digital outputs

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nouveau: don't pretend to support the DVI-I 'select subconnector' prop
Ben Skeggs [Fri, 14 Oct 2011 13:55:47 +0000 (23:55 +1000)]
drm/nouveau: don't pretend to support the DVI-I 'select subconnector' prop

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nouveau: improve dithering properties, and implement proper auto mode
Ben Skeggs [Mon, 17 Oct 2011 02:23:41 +0000 (12:23 +1000)]
drm/nouveau: improve dithering properties, and implement proper auto mode

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nouveau: no need to pass parameters into set_scale/dither
Ben Skeggs [Mon, 17 Oct 2011 00:38:10 +0000 (10:38 +1000)]
drm/nouveau: no need to pass parameters into set_scale/dither

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nouveau: determine a value for display_info.bpc if edid doesn't
Ben Skeggs [Mon, 17 Oct 2011 00:24:49 +0000 (10:24 +1000)]
drm/nouveau: determine a value for display_info.bpc if edid doesn't

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nv50/disp: wait for encoder disconnect to complete before link training
Ben Skeggs [Sun, 16 Oct 2011 22:56:09 +0000 (08:56 +1000)]
drm/nv50/disp: wait for encoder disconnect to complete before link training

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nv50/disp: disconnect encoders before reprogramming them
Ben Skeggs [Fri, 14 Oct 2011 04:43:20 +0000 (14:43 +1000)]
drm/nv50/disp: disconnect encoders before reprogramming them

Fixes a case where we don't get separate supervisor interrupt sequences for
disconnect and modeset events.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nv50/disp: completely reset disp if master evo channel active at init
Ben Skeggs [Fri, 14 Oct 2011 06:19:42 +0000 (16:19 +1000)]
drm/nv50/disp: completely reset disp if master evo channel active at init

Should fix issues with kexec, and as a nice side bonus, the code to avoid
having PDISP disappear will also fix hibernate on those effected systems.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nv50/disp: synchronise display right after init
Ben Skeggs [Fri, 14 Oct 2011 06:13:10 +0000 (16:13 +1000)]
drm/nv50/disp: synchronise display right after init

This has the effect of ensuring the encoders which were active before we
loaded get disconnected properly before we start reprogramming them.

Also removing a bit of cargo-cult from the initial evo pushbuf.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nv50/disp: move sync routine to where it can be used by other modules
Ben Skeggs [Fri, 14 Oct 2011 04:35:19 +0000 (14:35 +1000)]
drm/nv50/disp: move sync routine to where it can be used by other modules

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nouveau: restore cursors after restoring mode
Maxim Levitsky [Sun, 9 Oct 2011 20:58:34 +0000 (22:58 +0200)]
drm/nouveau: restore cursors after restoring mode

PDISP doesn't like it when disabled CRTCs are poked.

Fixes external output not coming to life when it has cursor on.
https://bugs.freedesktop.org/show_bug.cgi?id=41608

Signed-off-by: Maxim Levitsky <maximlevitsky@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nouveau: restore performance mode a bit later.
Maxim Levitsky [Sun, 9 Oct 2011 20:58:35 +0000 (22:58 +0200)]
drm/nouveau: restore performance mode a bit later.

Otherwice code that responsible for idling the card can't work.
BIOS init tables are supposed to init the clocks to correct values,
so that shouldn't cause any problems (we don't reclock by default anyway)

Signed-off-by: Maxim Levitsky <maximlevitsky@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nouveau: disable output polling through suspend.
Maxim Levitsky [Sun, 9 Oct 2011 20:58:33 +0000 (22:58 +0200)]
drm/nouveau: disable output polling through suspend.

Because doing polling while hardware is disabled is a bad idea...

Signed-off-by: Maxim Levitsky <maximlevitsky@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nv50: also report errors in MP1/MP2 when they happen.
Maxim Levitsky [Sun, 9 Oct 2011 20:58:32 +0000 (22:58 +0200)]
drm/nv50: also report errors in MP1/MP2 when they happen.

Signed-off-by: Maxim Levitsky <maximlevitsky@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nouveau: add overscan compensation connector properties
Ben Skeggs [Thu, 6 Oct 2011 03:29:05 +0000 (13:29 +1000)]
drm/nouveau: add overscan compensation connector properties

Exposes the same connector properties as the Radeon implementation, however
their behaviour isn't exactly the same.  The primary difference being that
unless both hborder/vborder have been defined by the user, the driver will
keep the aspect ratio of the overscanned area the same as the mode the
display is programmed for.

Enabled for digital outputs on GeForce 8 and up, excluding GF119.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nouveau: move master modesetting init to nouveau_display
Ben Skeggs [Thu, 6 Oct 2011 02:46:40 +0000 (12:46 +1000)]
drm/nouveau: move master modesetting init to nouveau_display

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nv50/crtc: disable flip overlay around scaling mode changes
Ben Skeggs [Thu, 6 Oct 2011 01:51:45 +0000 (11:51 +1000)]
drm/nv50/crtc: disable flip overlay around scaling mode changes

Prevents EVO getting all angry at us.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nouveau/hdmi: enable sending of avi/audio infoframes
Ben Skeggs [Wed, 5 Oct 2011 04:59:14 +0000 (14:59 +1000)]
drm/nouveau/hdmi: enable sending of avi/audio infoframes

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nouveau/hdmi: add hdmi register accessors to handle hdmi block move
Ben Skeggs [Wed, 5 Oct 2011 04:52:02 +0000 (14:52 +1000)]
drm/nouveau/hdmi: add hdmi register accessors to handle hdmi block move

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nouveau/hdmi: build ELD from EDID, notify audio driver of its presence
Ben Skeggs [Wed, 5 Oct 2011 01:05:07 +0000 (11:05 +1000)]
drm/nouveau/hdmi: build ELD from EDID, notify audio driver of its presence

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nv50/pm: s/unk05/vdec/
Ben Skeggs [Tue, 27 Sep 2011 00:31:36 +0000 (10:31 +1000)]
drm/nv50/pm: s/unk05/vdec/

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nouveau/pm: remove defunct fanspeed_set/get from pm table
Ben Skeggs [Fri, 16 Sep 2011 20:38:56 +0000 (06:38 +1000)]
drm/nouveau/pm: remove defunct fanspeed_set/get from pm table

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nv40/pm: convert to new pwm hooks, also fixing pwm type detection
Ben Skeggs [Fri, 16 Sep 2011 16:11:39 +0000 (02:11 +1000)]
drm/nv40/pm: convert to new pwm hooks, also fixing pwm type detection

A NV49 appeared a while back that was using the "nv41 style" pwm registers,
rather than the "nv40 style" ones my board is using.  This disproves the
previous theory that the pwm controller choice is chipset-specific.

So, after looking at a bunch of vbios images it appears that the next viable
theory is that we should select the pwm controller to use based on the gpio
line the fan is tied to, just like we do on nv50.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nv50/pm: convert to new fanspeed pwm controller hooks
Ben Skeggs [Fri, 16 Sep 2011 16:01:24 +0000 (02:01 +1000)]
drm/nv50/pm: convert to new fanspeed pwm controller hooks

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nouveau/pm: introduce generic handler for on-chip fan controller
Ben Skeggs [Fri, 16 Sep 2011 15:42:12 +0000 (01:42 +1000)]
drm/nouveau/pm: introduce generic handler for on-chip fan controller

The handling of the internal pwm fan controller is similar enough between
current chipsets that it makes sense to share the logic, and bugfixes :)

No hw backends converted yet, will automatically fall-through to the
"old" per-chipset fanspeed hooks for now.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nouveau/gpio: remove invert flag, use state[] everywhere
Ben Skeggs [Mon, 12 Sep 2011 19:32:43 +0000 (05:32 +1000)]
drm/nouveau/gpio: remove invert flag, use state[] everywhere

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nv50/pm: mostly nailed down fan pwm frequency selection
Ben Skeggs [Mon, 15 Aug 2011 06:13:34 +0000 (16:13 +1000)]
drm/nv50/pm: mostly nailed down fan pwm frequency selection

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nouveau/pm: manual pwm fanspeed management for nv40+ boards
Martin Peres [Mon, 15 Aug 2011 01:10:30 +0000 (11:10 +1000)]
drm/nouveau/pm: manual pwm fanspeed management for nv40+ boards

Exposes the following sysfs entries:
- fan0_input: read the rotational speed of the fan (poll a bit during 250ms)
- pwm0: set the pwm duty cycle
- pwm0_min/max: set the minimum/maximum pwm value

v2 (Ben Skeggs):
- nv50 pwm controller code removed in favour of other more complete code
- FAN_RPM -> FAN_SENSE
- merged FAN_SENSE readout into common code, not at all nv50-specific
- protected fanspeed changes with perflvl_wr
- formatting tidying
- added some comments where things are shaky

v3 (Martin Peres)
- ensure duty min/max from thermal table are sane

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Signed-off-by: Martin Peres <martin.peres@ensi-bourges.fr>
13 years agodrm/nv50/pm: add support for pwm fan control
Ben Skeggs [Sun, 14 Aug 2011 02:43:47 +0000 (12:43 +1000)]
drm/nv50/pm: add support for pwm fan control

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nouveau/vdec: implement stub modules for the known engines
Ben Skeggs [Thu, 11 Aug 2011 04:58:06 +0000 (14:58 +1000)]
drm/nouveau/vdec: implement stub modules for the known engines

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nouveau/pm: hook up fanspeed get/set if they're present
Ben Skeggs [Thu, 28 Jul 2011 01:01:21 +0000 (11:01 +1000)]
drm/nouveau/pm: hook up fanspeed get/set if they're present

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nv41/pm: implement a second type of fanspeed pwm
Ben Skeggs [Thu, 28 Jul 2011 00:52:13 +0000 (10:52 +1000)]
drm/nv41/pm: implement a second type of fanspeed pwm

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nv40/pm: implement first type of pwm fanspeed funcs
Ben Skeggs [Thu, 28 Jul 2011 00:40:48 +0000 (10:40 +1000)]
drm/nv40/pm: implement first type of pwm fanspeed funcs

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm/nv40/pm: parse fan pwm divisor from vbios tables
Ben Skeggs [Thu, 28 Jul 2011 00:17:40 +0000 (10:17 +1000)]
drm/nv40/pm: parse fan pwm divisor from vbios tables

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
13 years agodrm: Add drm_format_num_planes() utility function
Ville Syrjälä [Mon, 19 Dec 2011 22:33:24 +0000 (00:33 +0200)]
drm: Add drm_format_num_planes() utility function

This function returns the number of planes used by a specific pixel
format.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Rob Clark <rob.clark@linaro.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm: call connector dpms fxn, when setting config
Rob Clark [Thu, 15 Dec 2011 20:53:24 +0000 (14:53 -0600)]
drm: call connector dpms fxn, when setting config

Call connector->funcs->dpms(DPMS_ON) rather than just setting
connector->dpms = DPMS_ON.  This ensures that if the connector
has something to do to enable the output (rather than just using
drm_helper_connector_dpms helper directly), that this happens
at bootup.  This solves an issue with connectors not getting
enabled from fbcon_init() when the driver is loaded.

Signed-off-by: Rob Clark <rob@ti.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm/radeon/kms: don't use 0 bpc for adjusting hdmi clock
Dave Airlie [Tue, 20 Dec 2011 11:44:30 +0000 (11:44 +0000)]
drm/radeon/kms: don't use 0 bpc for adjusting hdmi clock

If the bpc is set from the connector is 0, we then use it later to adjust
in a special case the HDMI pixel clock, however if the bpc is 0, we end up
passing a 0 pixel clock into the code.

I'm not sure if this is the correct answer or if we should avoid the HDMI
clock adjustment for 0 values.

This fixes a divide by 0 on my Llano system with a HDMI monitor and hdmi
audio enabled.

Reviewed-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agoMerge branch 'drm-radeon-next' of ../drm-radeon-next into drm-core-next
Dave Airlie [Tue, 20 Dec 2011 19:53:44 +0000 (19:53 +0000)]
Merge branch 'drm-radeon-next' of ../drm-radeon-next into drm-core-next

* 'drm-radeon-next' of ../drm-radeon-next:
  drm/radeon: introduce a sub allocator and convert ib pool to it v4
  drm/radeon/kms: add support for per-ring fence interrupts
  drm/radeon/kms: add cayman specific fence_ring_emit
  drm/radeon/kms: add some new ring params to better handle other ring types
  drm/radeon: improve radeon_test_syncing function
  drm/radeon: precompute fence cpu/gpu addr once v3
  drm/radeon: move ring debugfs into radeon_ring.c
  drm/radeon: rename struct radeon_cp to radeon_ring
  drm/radeon: disable compute rings on cayman for now
  drm/radeon: add radeon_fence_count_emited function
  drm/radeon: make some asic pointers per ring
  drm/radeon: Add radeon_test_syncing function v2
  drm/radeon: make cp variable an array
  drm/radeon: make ring rptr and wptr register offsets variable
  drm/radeon: make all functions work with multiple rings.
  drm/radeon/kms: add support for semaphores v3
  drm/radeon/kms: add support for multiple fence queues v2
  drm/radeon: fix a spelling mistake
  drm/radeon: no need to check all relocs for duplicates
  drm/radeon: fix debugfs handling v3

13 years agodrm/radeon: introduce a sub allocator and convert ib pool to it v4
Jerome Glisse [Tue, 15 Nov 2011 16:48:34 +0000 (11:48 -0500)]
drm/radeon: introduce a sub allocator and convert ib pool to it v4

Somewhat specializaed sub-allocator designed to perform sub-allocation
for command buffer not only for current cs ioctl but for future command
submission ioctl as well. Patch also convert current ib pool to use
the sub allocator. Idea is that ib poll buffer can be share with other
command buffer submission not having 64K granularity.

v2 Harmonize pool handling and add suspend/resume callback to pin/unpin
sa bo (tested on rv280, rv370, r420, rv515, rv610, rv710, redwood, cayman,
rs480, rs690, rs880)
v3 Simplify allocator
v4 Fix radeon_ib_get error path to properly free fence

Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm/radeon/kms: add support for per-ring fence interrupts
Alex Deucher [Fri, 18 Nov 2011 01:13:28 +0000 (20:13 -0500)]
drm/radeon/kms: add support for per-ring fence interrupts

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Christian König <deathsimple@vodafone.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm/radeon/kms: add cayman specific fence_ring_emit
Alex Deucher [Thu, 17 Nov 2011 19:57:50 +0000 (14:57 -0500)]
drm/radeon/kms: add cayman specific fence_ring_emit

cayman is wb only and doesn't have a VC.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Christian König <deathsimple@vodafone.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm/radeon/kms: add some new ring params to better handle other ring types
Alex Deucher [Thu, 17 Nov 2011 19:25:56 +0000 (14:25 -0500)]
drm/radeon/kms: add some new ring params to better handle other ring types

Some rptr/wptrs fields have different offsets and not all rings are pm4
so add a new nop field.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Christian König <deathsimple@vodafone.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm/radeon: improve radeon_test_syncing function
Christian König [Thu, 17 Nov 2011 14:22:44 +0000 (15:22 +0100)]
drm/radeon: improve radeon_test_syncing function

Also test multiple waits on the same semaphore.

Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm/radeon: precompute fence cpu/gpu addr once v3
Jerome Glisse [Sun, 20 Nov 2011 20:45:34 +0000 (20:45 +0000)]
drm/radeon: precompute fence cpu/gpu addr once v3

Add a start fence driver helper function which will be call
once for each ring and will compute cpu/gpu addr for fence
depending on wether to use wb buffer or scratch reg.

This patch replace initialize fence driver separately which
was broken in regard of GPU lockup. The fence list for created,
emited, signaled must be initialize once and only from the
asic init callback not from the startup call back which is
call from the gpu reset.

v2: With this in place we no longer need to know the number of
    rings in fence_driver_init, also writing to the scratch reg
    before knowing its offset is a bad idea.

v3: rebase on top of change to previous patch in the serie

Signed-off-by: Christian König <deathsimple@vodafone.de>
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm/radeon: move ring debugfs into radeon_ring.c
Christian König [Mon, 24 Oct 2011 15:08:44 +0000 (17:08 +0200)]
drm/radeon: move ring debugfs into radeon_ring.c

Those debugfs files aren't r600 specific, so they
shouldn't be in r600.c. Move them to radeon_ring.c
and also add functionality to dump CP1 & CP2 ring
informations.

Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm/radeon: rename struct radeon_cp to radeon_ring
Christian König [Sun, 23 Oct 2011 10:56:27 +0000 (12:56 +0200)]
drm/radeon: rename struct radeon_cp to radeon_ring

That naming seems to make more sense, since we not
only want to run PM4 rings with it.

Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm/radeon: disable compute rings on cayman for now
Christian König [Thu, 20 Oct 2011 10:47:13 +0000 (12:47 +0200)]
drm/radeon: disable compute rings on cayman for now

Disable the additional compute rings on cayman
until their setup is fully implemented.

Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm/radeon: add radeon_fence_count_emited function
Christian König [Thu, 20 Oct 2011 10:38:09 +0000 (12:38 +0200)]
drm/radeon: add radeon_fence_count_emited function

Split counting of emited fences out of power
management into a seperate function.

Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm/radeon: make some asic pointers per ring
Christian König [Wed, 19 Oct 2011 17:02:21 +0000 (19:02 +0200)]
drm/radeon: make some asic pointers per ring

Emitting fences, semaphores and ib works differently
on different ring, so its is easier to maintain
separate functions for each ring.

Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm/radeon: Add radeon_test_syncing function v2
Christian König [Tue, 27 Sep 2011 10:31:00 +0000 (12:31 +0200)]
drm/radeon: Add radeon_test_syncing function v2

Tests syncing between all rings by using
semaphores and fences.

v2: use radeon_testing as a bit flag rather than on/off switch
    this allow to test for one thing at a time (bo_move or semaphore
    test). It kind of break the usage if user wheren't using 1
    for bo move test but as it's a test feature i believe it's ok.

Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm/radeon: make cp variable an array
Christian König [Thu, 13 Oct 2011 11:19:22 +0000 (13:19 +0200)]
drm/radeon: make cp variable an array

Replace cp, cp1 and cp2 members with just an array
of radeon_cp structs.

Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm/radeon: make ring rptr and wptr register offsets variable
Christian König [Thu, 13 Oct 2011 10:48:45 +0000 (12:48 +0200)]
drm/radeon: make ring rptr and wptr register offsets variable

Every ring seems to have the concept of read and
write pointers. Make the register offset variable
so we can use the functions for different types of rings.

Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm/radeon: make all functions work with multiple rings.
Christian König [Fri, 23 Sep 2011 13:11:23 +0000 (15:11 +0200)]
drm/radeon: make all functions work with multiple rings.

Give all asic and radeon_ring_* functions a
radeon_cp parameter, so they know the ring to work with.

Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm/radeon/kms: add support for semaphores v3
Christian König [Thu, 15 Sep 2011 17:02:22 +0000 (19:02 +0200)]
drm/radeon/kms: add support for semaphores v3

They are used to sync between rings, while fences
sync between a ring and the cpu.

v2 Fix radeon_semaphore_driver_fini when no semaphore were
allocated.

v3 Initialize list early on to avoid issue in case or early
error

Signed-off-by: Christian König <deathsimple@vodafone.de>
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm/radeon/kms: add support for multiple fence queues v2
Alex Deucher [Thu, 25 Aug 2011 17:39:48 +0000 (13:39 -0400)]
drm/radeon/kms: add support for multiple fence queues v2

For supporting multiple CP ring buffers, async DMA
engines and UVD.  We still need a way to synchronize
between engines.

v2 initialize unused fence driver ring to avoid issue in
   suspend/unload

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm/radeon: fix a spelling mistake
Christian König [Mon, 24 Oct 2011 13:05:29 +0000 (15:05 +0200)]
drm/radeon: fix a spelling mistake

Better fix it before this obvious typo spreads even more.

Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm/radeon: no need to check all relocs for duplicates
Christian König [Mon, 24 Oct 2011 12:59:17 +0000 (14:59 +0200)]
drm/radeon: no need to check all relocs for duplicates

Only check the previously checked relocs for
duplicates. Also leaving the handle uninitialized
isn't such a good idea.

Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm/radeon: fix debugfs handling v3
Christian König [Mon, 24 Oct 2011 12:54:54 +0000 (14:54 +0200)]
drm/radeon: fix debugfs handling v3

Having registered debugfs files globally causes
the files to not show up on the second, third
etc.. card in the system.

v2: fix crash on module unloading
v3: fix space indentation

Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agogma500/oaktrail: panel display quality fix
Alan Cox [Tue, 20 Dec 2011 15:58:44 +0000 (15:58 +0000)]
gma500/oaktrail: panel display quality fix

The GCT setup was used on Moorestown. The Oaktrail version uses a normal PC
interface. That means we must also honour the dither info from the BIOS
data.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agogma500: Oaktrail fixes
Alan Cox [Tue, 20 Dec 2011 15:58:29 +0000 (15:58 +0000)]
gma500: Oaktrail fixes

The move to connectors breaks Oaktrail again if we have memory poisoning
enabled. Fix this.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agoMerge tag 'v3.2-rc6' of /home/airlied/devel/kernel/linux-2.6 into drm-core-next
Dave Airlie [Tue, 20 Dec 2011 14:43:53 +0000 (14:43 +0000)]
Merge tag 'v3.2-rc6' of /home/airlied/devel/kernel/linux-2.6 into drm-core-next

Merge in the upstream tree to bring in the mainline fixes.

Conflicts:
drivers/gpu/drm/exynos/exynos_drm_fbdev.c
drivers/gpu/drm/nouveau/nouveau_sgdma.c

13 years agogma500: Final enables for Oaktrail
Alan Cox [Mon, 19 Dec 2011 21:06:35 +0000 (21:06 +0000)]
gma500: Final enables for Oaktrail

This switches the ifdef to match the Kconfig so that Oaktrail probing occurs
and adds some additional minor bulletproofing.

Tested on a Fujtisu Stylistic Q550 internal display. HDMI might work but that
remains to be seen.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agogma500: SDVO DDC bus guessing isn't working so hardcode it instead
Patrik Jakobsson [Mon, 19 Dec 2011 21:41:45 +0000 (21:41 +0000)]
gma500: SDVO DDC bus guessing isn't working so hardcode it instead

We currently don't have support for parsing SDVO mappings from BIOS so we're
guessing the bus switch parameter. This isn't working so hardcode it to a
configuration known to work on most poulsbo hardware.

Signed-off-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agogma500: Convert Oaktrail to work with new output handling
Patrik Jakobsson [Mon, 19 Dec 2011 21:41:33 +0000 (21:41 +0000)]
gma500: Convert Oaktrail to work with new output handling

Replace psb_intel_output with psb_intel_encoder and psb_intel_connector

Signed-off-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
[Changed Moorestown reference to Oaktrail]
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agogma500: Convert Cedarview to work with new output handling
Patrik Jakobsson [Mon, 19 Dec 2011 21:41:22 +0000 (21:41 +0000)]
gma500: Convert Cedarview to work with new output handling

Replace psb_intel_output with psb_intel_encoder and psb_intel_connector.
Things will need to be cleaned up and tested so consider this an initial
patch for Cedarview.

Signed-off-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agogma500: Replace SDVO code with slightly modified version from i915
Patrik Jakobsson [Mon, 19 Dec 2011 21:41:10 +0000 (21:41 +0000)]
gma500: Replace SDVO code with slightly modified version from i915

Our current SDVO implementation is not working properly, so replace it with
a modified version of the i915. Further testing and debugging is needed to make
sure we can handle the different SDVO setups and wiring.

Signed-off-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agogma500: Add support for Intel GMBUS
Patrik Jakobsson [Mon, 19 Dec 2011 21:40:58 +0000 (21:40 +0000)]
gma500: Add support for Intel GMBUS

Before we integrate the new SDVO code we need GMBUS support

Signed-off-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agogma500: Convert PSB LVDS to new output handling
Patrik Jakobsson [Mon, 19 Dec 2011 21:40:45 +0000 (21:40 +0000)]
gma500: Convert PSB LVDS to new output handling

LVDS for PSB now uses psb_intel_encoder and psb_intel_connectors instead of
psb_intel_output. i2c_bus and ddc_bus are moved to lvds_priv. There was also a
pointer to mode_dev (for no obvious reason) that we now get directly from
dev_priv.

Signed-off-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agogma500: Fix encoder type checking for connectors
Patrik Jakobsson [Mon, 19 Dec 2011 21:40:33 +0000 (21:40 +0000)]
gma500: Fix encoder type checking for connectors

Fix cases where we need to know what encoder type is behind a given connector.

Signed-off-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agogma500: Remove psb_intel_output from ddc_probe and ddc_get_modes
Patrik Jakobsson [Mon, 19 Dec 2011 21:40:05 +0000 (21:40 +0000)]
gma500: Remove psb_intel_output from ddc_probe and ddc_get_modes

Signed-off-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agogma500: Initial support for our encoder and connector structs
Patrik Jakobsson [Mon, 19 Dec 2011 21:39:53 +0000 (21:39 +0000)]
gma500: Initial support for our encoder and connector structs

First step towards adding i915 alike encoder and connector abstractions. This
will make life easier when adding i915 output code into our driver. It also
removes the old psb_intel_output struct.

Signed-off-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agostaging/gma500: fixup staging code to build following core changes.
Dave Airlie [Tue, 20 Dec 2011 10:22:18 +0000 (10:22 +0000)]
staging/gma500: fixup staging code to build following core changes.

This just fixes up the staging code to keep building.

Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm: Replace pitch with pitches[] in drm_framebuffer
Ville Syrjälä [Mon, 19 Dec 2011 22:06:49 +0000 (00:06 +0200)]
drm: Replace pitch with pitches[] in drm_framebuffer

Otherwise each driver would need to keep the information inside
their own framebuffer object structure. Also add offsets[]. BOs
on the other hand are driver specific, so those can be kept in
driver specific structures.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm: Check that the requested pixel format is valid
Ville Syrjälä [Mon, 19 Dec 2011 22:06:48 +0000 (00:06 +0200)]
drm: Check that the requested pixel format is valid

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm: plane: Check that the fb pixel format is supported by the plane
Ville Syrjälä [Mon, 19 Dec 2011 22:06:47 +0000 (00:06 +0200)]
drm: plane: Check that the fb pixel format is supported by the plane

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm: plane: Make 'formats' parameter to drm_plane_init() const
Ville Syrjälä [Mon, 19 Dec 2011 22:06:46 +0000 (00:06 +0200)]
drm: plane: Make 'formats' parameter to drm_plane_init() const

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm: plane: Check crtc coordinates against integer overflows in setplane ioctl
Ville Syrjälä [Mon, 19 Dec 2011 22:06:45 +0000 (00:06 +0200)]
drm: plane: Check crtc coordinates against integer overflows in setplane ioctl

Help drivers a little by guaranteeing that crtc_x+crtc_w and
crtc_y+crtc_h don't overflow.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm: plane: Check source coordinates
Ville Syrjälä [Mon, 19 Dec 2011 22:06:44 +0000 (00:06 +0200)]
drm: plane: Check source coordinates

Make sure the source coordinates stay within the buffer.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm: plane: Clear plane.crtc and plane.fb after disable_plane()
Ville Syrjälä [Mon, 19 Dec 2011 22:06:43 +0000 (00:06 +0200)]
drm: plane: Clear plane.crtc and plane.fb after disable_plane()

These are the only indication to user space that the plane was disabled.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm: Fix __user sparse warnings
Ville Syrjälä [Mon, 19 Dec 2011 22:06:42 +0000 (00:06 +0200)]
drm: Fix __user sparse warnings

Several pointers and casts were missing __user annotations.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm: plane: mutex_unlock() was missing
Ville Syrjälä [Mon, 19 Dec 2011 22:06:41 +0000 (00:06 +0200)]
drm: plane: mutex_unlock() was missing

Unlock the mode_config mutex if drm_plane_init() fails.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm: fourcc: Use __u32 instead of u32
Ville Syrjälä [Mon, 19 Dec 2011 22:06:40 +0000 (00:06 +0200)]
drm: fourcc: Use __u32 instead of u32

drm_fourcc.h can be included from user space so use the appropriate types.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm: Install drm_fourcc.h
Ville Syrjälä [Mon, 19 Dec 2011 22:06:39 +0000 (00:06 +0200)]
drm: Install drm_fourcc.h

Userspace needs this header.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm: Add a missing ')'
Ville Syrjälä [Mon, 19 Dec 2011 22:06:38 +0000 (00:06 +0200)]
drm: Add a missing ')'

The code happened to compile because the flag wasn't actually used yet.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm/intel: Fix initialization if startup happens in interlaced mode [v2]
Christian Schmidt [Mon, 19 Dec 2011 19:03:33 +0000 (20:03 +0100)]
drm/intel: Fix initialization if startup happens in interlaced mode [v2]

My EFI BIOS starts the graphics card up in my projector's preferred EDID
mode, 1080@60i. The Intel driver does not clear all the interlaced bits.

This patch introduces a new PIPECONF_INTERLACE_MASK define and uses it
to restore progressive mode.

Signed-of-by: Christian Schmidt <schmidt@digadd.de>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agoFix wrong assumptions in cea_for_each_detailed_block v2
Christian Schmidt [Mon, 19 Dec 2011 19:03:43 +0000 (20:03 +0100)]
Fix wrong assumptions in cea_for_each_detailed_block v2

The current logic misunderstands the spec about CEA 18byte descriptors.
First, the spec doesn't state "detailed timing descriptors" but "18 byte
descriptors", so any data record could be stored, mixed timings and
other data, just as in the standard EDID.
Second, the lower four bit of byte 3 of the CEA record do not contain
the number of descriptors, but "the total number of DTDs defining native
formats in the whole EDID [...], starting with the first DTD in the DTD
list (which starts in the base EDID block)." A device can of course
support non-native formats.

As such the number can't be used to determine n, and the existing code
will filter non-timing 18byte descriptors anyway.

Signed-off-by: Christian Schmidt <schmidt@digadd.de>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm_edid_to_eld: check for CEA data blocks only from structure revision 3 on
Christian Schmidt [Mon, 19 Dec 2011 19:03:38 +0000 (20:03 +0100)]
drm_edid_to_eld: check for CEA data blocks only from structure revision 3 on

CEA datablocks are only defined from revision 3 onwards. Only check for
them if the revision says so.

Signed-of-by: Christian Schmidt <schmidt@digadd.de>
Tested-by: James Cloos <cloos@jhcloos.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agodrm/edid: support CEA video modes.
Christian Schmidt [Mon, 19 Dec 2011 14:53:16 +0000 (14:53 +0000)]
drm/edid: support CEA video modes.

TFT/plasma televisions and projectors have become commonplace, and so
has the use of PCs to drive them. Add the video modes specified by an
EDID's CEA extension to the mode database for a connector.

Before:
[    1.158869] [drm:drm_mode_debug_printmodeline], Modeline
19:"1920x1080i" 0 74250 1920 2448 2492 2640 1080 1084 1094 1125 0x40 0x15
[    1.158875] [drm:drm_mode_debug_printmodeline], Modeline
18:"1920x1080i" 0 74250 1920 2008 2052 2200 1080 1084 1094 1125 0x48 0x15
[    1.158882] [drm:drm_mode_debug_printmodeline], Modeline
20:"1920x1080" 24 74250 1920 2558 2602 2750 1080 1084 1089 1125 0x40 0x5

After:
[    1.144175] [drm:drm_mode_debug_printmodeline], Modeline
22:"1920x1080" 0 74250 1920 2448 2492 2640 1080 1084 1094 1125 0x40 0x15
[    1.144179] [drm:drm_mode_debug_printmodeline], Modeline
21:"1920x1080" 0 74250 1920 2008 2052 2200 1080 1084 1094 1125 0x48 0x15
[    1.144187] [drm:drm_mode_debug_printmodeline], Modeline
30:"1920x1080" 50 148500 1920 2448 2492 2640 1080 1084 1089 1125 0x40 0x5
[    1.144190] [drm:drm_mode_debug_printmodeline], Modeline
29:"1920x1080" 60 148500 1920 2008 2052 2200 1080 1084 1089 1125 0x40 0x5
[    1.144192] [drm:drm_mode_debug_printmodeline], Modeline
25:"1920x1080" 24 74250 1920 2558 2602 2750 1080 1084 1089 1125 0x40 0x5
[    1.144195] [drm:drm_mode_debug_printmodeline], Modeline
24:"1280x720" 50 74250 1280 1720 1760 1980 720 725 730 750 0x40 0x5
[    1.144198] [drm:drm_mode_debug_printmodeline], Modeline
23:"1280x720" 60 74250 1280 1390 1430 1650 720 725 730 750 0x40 0x5
[    1.144201] [drm:drm_mode_debug_printmodeline], Modeline 27:"720x576"
50 27000 720 732 796 864 576 581 586 625 0x40 0xa
[    1.144203] [drm:drm_mode_debug_printmodeline], Modeline 26:"720x480"
60 27000 720 736 798 858 480 489 495 525 0x40 0xa
[    1.144206] [drm:drm_mode_debug_printmodeline], Modeline 28:"640x480"
60 25175 640 656 752 800 480 490 492 525 0x40 0xa

Signed-off-by: Christian Schmidt <schmidt@digadd.de>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
13 years agoLinux 3.2-rc6
Linus Torvalds [Sat, 17 Dec 2011 02:36:26 +0000 (18:36 -0800)]
Linux 3.2-rc6

13 years agoRevert "drm/i915: fix infinite recursion on unbind due to ilk vt-d w/a"
Linus Torvalds [Fri, 16 Dec 2011 20:58:39 +0000 (12:58 -0800)]
Revert "drm/i915: fix infinite recursion on unbind due to ilk vt-d w/a"

This reverts commit eb1711bb94991e93669c5a1b5f84f11be2d51ea1.

It blows up the i915 seqno tracking, resulting in the

BUG_ON(seqno == 0);

in i915_wait_request() triggering, which will cause lock-ups.

See for example
  https://bugs.launchpad.net/ubuntu/+source/linux/+bug/903010
  https://lkml.org/lkml/2011/12/14/395

Reported-requested-and-tested-by: Dirk Hohndel <dirk@hohndel.org>
Reported-by: Richard Eames <Richard.Eames@flinders.edu.au>
Reported-by: Rocko Requin <rockorequin@hotmail.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Dave Airlie <airlied@redhat.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Keith Packard <keithp@keithp.com>
Cc: Eric Anholt <eric@anholt.net>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc
Linus Torvalds [Fri, 16 Dec 2011 20:26:26 +0000 (12:26 -0800)]
Merge git://git./linux/kernel/git/davem/sparc

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc:
  sparc32: Be less strict in matching %lo part of relocation.
  sbus: convert drivers/sbus/char/* to use module_platform_driver()
  bbc_i2c: Remove unneeded err variable
  sparc: Use kmemdup rather than duplicating its implementation

13 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
Linus Torvalds [Fri, 16 Dec 2011 20:17:32 +0000 (12:17 -0800)]
Merge git://git./linux/kernel/git/davem/net

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net:
  ipv6: Check dest prefix length on original route not copied one in rt6_alloc_cow().
  sch_gred: should not use GFP_KERNEL while holding a spinlock
  ipip, sit: copy parms.name after register_netdevice
  ipv6: Fix for adding multicast route for loopback device automatically.
  ssb: fix init regression with SoCs
  rtl8192{ce,cu,de,se}: avoid problems because of possible ERFOFF -> ERFSLEEP transition
  mac80211: fix another race in aggregation start
  fsl_pq_mdio: Clean up tbi address configuration
  ppp: fix pptp double release_sock in pptp_bind()
  net/fec: fix the use of pdev->id
  ath9k: fix check for antenna diversity support
  batman-adv: delete global entry in case of roaming
  batman-adv: in case of roaming mark the client with TT_CLIENT_ROAM
  Bluetooth: Correct version check in hci_setup
  btusb: fix a memory leak in btusb_send_frame()
  Bluetooth: bnep: Fix module reference
  Bluetooth: cmtp: Fix module reference
  Bluetooth: btmrvl: support Marvell Bluetooth device SD8797

13 years agoMerge branches 'for-linus' and 'for-linus-3.2' of git://git.kernel.org/pub/scm/linux...
Linus Torvalds [Fri, 16 Dec 2011 20:15:50 +0000 (12:15 -0800)]
Merge branches 'for-linus' and 'for-linus-3.2' of git://git./linux/kernel/git/mason/linux-btrfs

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs:
  Btrfs: unplug every once and a while
  Btrfs: deal with NULL srv_rsv in the delalloc inode reservation code
  Btrfs: only set cache_generation if we setup the block group
  Btrfs: don't panic if orphan item already exists
  Btrfs: fix leaked space in truncate
  Btrfs: fix how we do delalloc reservations and how we free reservations on error
  Btrfs: deal with enospc from dirtying inodes properly
  Btrfs: fix num_workers_starting bug and other bugs in async thread
  BTRFS: Establish i_ops before calling d_instantiate
  Btrfs: add a cond_resched() into the worker loop
  Btrfs: fix ctime update of on-disk inode
  btrfs: keep orphans for subvolume deletion
  Btrfs: fix inaccurate available space on raid0 profile
  Btrfs: fix wrong disk space information of the files
  Btrfs: fix wrong i_size when truncating a file to a larger size
  Btrfs: fix btrfs_end_bio to deal with write errors to a single mirror

* 'for-linus-3.2' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs:
  btrfs: lower the dirty balance poll interval