I pulled the wrong version of the patch from Daniel Vetter which was
missing the read barriers -- and the one that was causing all the trouble
was from i915_gem_object_put_fence_reg(), leading to GPU hangs on gen3.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Chris Wilson [Mon, 13 Sep 2010 22:44:34 +0000 (23:44 +0100)]
drm/i915: Reduce hangcheck frequency
By reducing the hangcheck frequency we check less often, conserving
resources, and still detect a lock up quickly. On a fast machine with a
slow GPU (like a Core2 paired with a 945G) it is easy for the hangcheck to
misfire as we check too fast.
Also once hung and if we fail to completely reset the chip, we have a
nasty habit of proclaming a hang many times a second and generating a
strobe-like display.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Chris Wilson [Mon, 13 Sep 2010 15:58:39 +0000 (16:58 +0100)]
drm/i915: Initialize intel_crtc->active
Fix a regression in the previous regression fix...
In order to turn off the pipes entirely upon the first modeset, we
pretend that BIOS (or earlier module incarnation) left them active.
The first task performed by setup_initial_configuration() is to disable
all pipes and so to avoid skipping that step and so to ensure a known
configuration we need to mark all the crtcs as active.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
When separating out the prepare/commit into its own separate functions
we overlooked that the intel_crtc->dpms_mode was being used elsewhere to
check on the actual status of the pipe.
Track that bit of logic separately from the actual dpms mode, so there
is no confusion should we be able to handle multiple dpms modes, nor
any semantic conflict between prepare/commit and dpms.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Chris Wilson [Mon, 13 Sep 2010 12:54:26 +0000 (13:54 +0100)]
drm/i915: Share crtc setup and teardown between dpms and disable/enable
This closes a couple of corner cases where we introduced and forgot
about a couple of routines that need to be called when disabling the
crtc and then re-enabling it. The code needs to be moved again so that
the common bits are shared across generations.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Chris Wilson [Mon, 13 Sep 2010 00:16:10 +0000 (01:16 +0100)]
drm/i915/lvds: Ensure panel is unlocked for Ironlake or the panel fitter
Commit 77d07fd9d73ef28689737c0952dbd5d6a5017743 introduced a regression
where by not waiting for the panel to be turned off, left the panel and
PLL registers locked across the modeset. Thus the panel remaining blank.
As pointed out by Daniel Vetter, when testing LVDS it helps to open the
laptop and look at the actual panel you are purporting to test.
A second issue with the patch was that in order to modify the panel
fitter before gen5, the pipe and the panel must have be completely
powered down. So we wait.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Chris Wilson [Thu, 9 Sep 2010 18:06:13 +0000 (19:06 +0100)]
drm/i915/sdvo: Poll command status 5 times without delay on read
The documentation says that an SDVO command takes a maximum of 15us to be
processed by the device, and that it is sufficient to read the status byte
3 times (whilst the command is still in the PENDING state) for the driver
to be confident that sufficient time has elapsed.
We err on the safe side and try 5 times before giving up.
The only question that remains: was the old behaviour derived by
experiments with real hardware?
A look into the murky history of UMS, implies that the behaviour was
accidental and the current retry mechanism was solely designed to catch
the status byte indicating PENDING with no reference to hardware
behaviour. (commit ac9181c014638dbeb334b40b4029d0ccb2b7a0fc in
xf86-video-intel)
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Thermal reporting may not be enabled by default on some machines, so
enable the appropriate bits to allow IPS to get the data it needs from
the CPU thermal device.
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
drm/i915: enable PCH PLL, FDI training and transcoder even for eDP
eDP panels require these to be set up prior to panel power sequencing,
or they'll fail to power on due to an "asset not ready" check. And of
course, eDP panels attached to anything other than DP_A need them
enabled regardless, since they'll be driven from the CPU through FDI out
to the PCH.
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
drm/i915: use i915 and Ironlake CRTC enable/disable functions in prepare/commit
This will allow us to optimize our prepare/commit paths a bit better.
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
[ickle: minor tweak to handle the cursor across pipe resizing] Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Chris Wilson [Thu, 9 Sep 2010 11:46:34 +0000 (12:46 +0100)]
drm/i915: Adapt workqueue to new alloc_workqueue interface
create_singlethreaded_workqueue() is being phased out for a new
concurrency managed task infrastructure.
Adapt our workqueue constructor to explicitly create a domain that only
allows the execution of a single task at any time. All the tasks are
expected to require the dev->struct_mutex, so would block concurrency of
other tasks if we allow more than a single i915 task to be run at once.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
We don't know how to enable it safely, especially as outputs turn on and
off. When disabling LP1 we also need to make sure LP2 and 3 are already
disabled.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=29173
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=29082 Reported-by: Chris Lord <chris@linux.intel.com> Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org> Tested-by: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: stable@kernel.org Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Chris Wilson [Thu, 9 Sep 2010 15:20:55 +0000 (16:20 +0100)]
drm/i915: Make the connector->encoder relationship explicit
Currently we have a exact mapping of a connector onto an encoder for its
whole lifetime. Make this an explicit property of the structure and so
simplify the code.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Chris Wilson [Wed, 8 Sep 2010 23:02:03 +0000 (00:02 +0100)]
drm/i915: Double check that the wait_request is not pending before warning
If we are busy, then we may have woken up the wait_request handler but
not yet serviced it before the hang check fires. So in hang check,
double check that the i915_gem_do_wait_request() is still pending the
wake-up before declaring all hope lost.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=30073 Reported-and-tested-by: Sitsofe Wheeler <sitsofe@yahoo.com> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
drm/i915: split DP link training across panel power sequencing
Mode set sequence requires that we start training, then enable the
panel, then complete training. So split the DP training function into
two parts; the first enables the DP port and sets training pattern 1 and
the second completes the training.
As part of this, remove some redundant function args from the various DP
handling functions and use the intel_dp fields everywhere we can.
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
[ickle: removed first ironlake_edp_backlight_on() on advice of jbarnes] Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
drm/i915: use VDD AUX for panel power around detection and in prepare
Mode setting sequence specifies that we use VDD AUX for configuration
and detection, and early in the mode set sequence. Only later (after
DP_A has started training) should we actually enable panel power.
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
[ickle: checkpatch.pl complaining about whitespace] Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Dan Carpenter [Wed, 8 Sep 2010 19:44:47 +0000 (21:44 +0200)]
i915: snprintf returns large values
snprintf() returns the number of bytes which would have been used if
there was enough space. It can be larger than the size of the buffer.
Obviously in this case the buffer is large enough but everyone just
copy and pastes this code so it's better to limit it and set a good
example.
Signed-off-by: Dan Carpenter <error27@gmail.com> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Daniel Vetter [Mon, 6 Sep 2010 18:08:44 +0000 (20:08 +0200)]
drm/i915: die, i915_probe_agp, die
Use the detection from intel-gtt.ko instead. Hooray!
Also move the stolen mem allocator to the other gtt stuff in dev_prv->mem.
v2: Chris Wilson noted that my error handling was crap. Fix it. He also
said that this fixes a problem on his i845. Indeed, i915_probe_agp
misses a special case for i830/i845 stolen mem detection.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=25476 Cc: stable@kernel.org Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Daniel Vetter [Sun, 29 Aug 2010 12:18:49 +0000 (14:18 +0200)]
intel-gtt: call init_gtt_init in probe function
This way create_gatt_table become dummy glue functions for the fake
agp driver - rename them accordingly (and kill the now unnecessary
i9xx copy).
With this change, the gtt initialization code is almost independant
from the agp stuff. Two things are still missing:
- the scratch page is created by the generic agp code.
- filling the whole gtt with scratch_page ptes is not yet consolidated -
this needs abstracted pte handling, first.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Daniel Vetter [Sun, 29 Aug 2010 15:35:30 +0000 (17:35 +0200)]
intel-gtt: consolidate i9xx setup
The only difference between i915 and i965 was the calculation of the
gtt address. So merge these two paths into one. Otherwise the same
changes as in the i830 setup consolidation.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Daniel Vetter [Sun, 29 Aug 2010 15:29:50 +0000 (17:29 +0200)]
intel-gtt: consolidate i830 setup
Slighlty reordered sequence was necessary. Also don't set
agp_bridge->gatt_bus_addr anymore. Only used by generic agp helper
functions, hence unnecessary for the intel fake agp driver.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Daniel Vetter [Wed, 8 Sep 2010 19:18:53 +0000 (21:18 +0200)]
intel-gtt: introduce intel_gtt_driver
Same idea as INTEL_INFO from drm/i915. This
- reduces the dependancy on agp_driver
- stops the what-does-IS_I965G-mean confusion (here it's just gen4, in
drm/i915 it's gen >=4)
- further prepares the separation of the fake agp driver from the rest.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Daniel Vetter [Sat, 28 Aug 2010 09:04:32 +0000 (11:04 +0200)]
intel-gtt: fix gtt_total_entries detection
In commit f1befe71 Chris Wilson added some code to clear the full gtt
on g33/pineview instead of just the mappable part. The code looks like
it was copy-pasted from agp/intel-gtt.c, at least an identical piece
of code is still there (in intel_i830_init_gtt_entries). This lead to
a regression in 2.6.35 which was supposedly fixed in commit e7b96f28
Now this commit makes absolutely no sense to me. It seems to be
slightly confused about chipset generations - it references docs for
4th gen but the regression concerns 3rd gen g33. Luckily the the g33
gmch docs are available with the GMCH Graphics Control pci config
register definitions. The other (bigger problem) is that the new
check in there uses the i830 stolen mem bits (.5M, 1M or 8M of stolen
mem). They are different since the i855GM.
The most likely case is that it hits the 512M fallback, which was
probably the right thing for the boxes this was tested on.
So the original approach by Chris Wilson seems to be wrong and the
current code is definitely wrong. There is a third approach by Jesse
Barnes from his RFC patch "Who wants a bigger GTT mapping range?"
where he simply shoves g33 in the same clause like later chipset
generations.
I've asked him and Jesse confirmed that this should work. So implement
it.
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=16891$ Tested-by: Anisse Astier <anisse@astier.eu> Cc: stable@kernel.org Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Daniel Vetter [Fri, 27 Aug 2010 15:12:41 +0000 (17:12 +0200)]
intel-gtt: adjust overhead entries in intel_gtt_stolen_entries
agp/intel_gtt.c and drm/i915/i915_dma.c don't calculate this the same
way: The intel-gtt code seems to use the actual gtt size, the drm
module just the mappable. Go with the logic from the drm module because
that's the more conservative choice.
But conserve the original code in intel_gtt_total_size for later use.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Daniel Vetter [Fri, 27 Aug 2010 14:13:52 +0000 (16:13 +0200)]
intel-gtt: sane variable names for intel_gtt_stolen_entries
This somewhat aligns it with the version in drm/i915/i915_dma.c.
Changes:
- s/gtt_entries/stolen_size
- track overhead entries in a seperate var (the effective gtt size
calculation will be extracted later on).
- subtract the overhead at the end instead of in each clause.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Daniel Vetter [Wed, 8 Sep 2010 19:01:04 +0000 (21:01 +0200)]
intel-gtt: new function intel_gtt_mappable_entries
This implementation is stolen from drm/i915, but is equivalent to
the code sprinkled over intel-gtt.c in the various fetch_size functions.
It's not yet used anywhere, though.
Also introduce intel_gtt_init which only calls intel_gtt_stolen_entries.
Over the course of the next patches, this will grow untill it contains
the complete init sequence starting from the call to gtt_mappable_entries.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
First simple step towards a more generic initialization. This
is needed to disentangle the agp stuff from the stuff that is
actually needed by drm/i915.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Daniel Vetter [Tue, 24 Aug 2010 20:18:41 +0000 (22:18 +0200)]
intel-gtt: introduce drm/intel-gtt.h
Add a few definitions to it that are already shared and that will
be shared in the future (like the number of stolen entries).
No functional changes in here.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Daniel Vetter [Wed, 8 Sep 2010 15:29:51 +0000 (17:29 +0200)]
agp/intel: make intel-gtt.c into a real source file
Now that the disentangling is complete, stop including intel-gtt.c
from intel-agp.c.
The linux build system _really_ doesn't allow .c source files with the
same name as the module. It fails with the following message when trying
to build such a bugger:
Instead of renameing intel-agp.c I've simply created a new module out
of intel-gtt.c. Renaming intel-agp.ko to something else is not an option
for it will surely kill someones boot process.
This also paves the way to use the gtt code without loading the agp
driver.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Daniel Vetter [Tue, 24 Aug 2010 17:39:48 +0000 (19:39 +0200)]
agp/intel: split out gmch/gtt probe, part 2
This just splits the device list into two and moves the gtt related stuff
to intel-gtt.c. The two new devices lists also lose the not longer needed
fields. There where only about 5 cases anyway with both a gmch and a
possible agp port, so the duplication of entries is rather small.
Additionally kill 2 out of the three Ironlake mobile entries that
only differed in host bridge pci id.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Matthew Garrett [Tue, 18 May 2010 17:53:16 +0000 (13:53 -0400)]
drm/i915: Don't disable panel for modesetting if pfit hasn't changed
It seems to be possible to program a new mode without disabling the panel
if the panel fitter setup doesn't change. Add support for that.
Signed-off-by: Matthew Garrett <mjg@redhat.com> Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
We really need a macro to test whether a given connector has a panel
attached rather than sprinkling HAS_PCH_SPLIT/IS_eDP/has_edp_encoder
etc all over. In the meantime, fix the bug...
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
[ickle: tidy up the duplicity in the conditionals] Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Chris Wilson [Sun, 8 Aug 2010 19:38:12 +0000 (20:38 +0100)]
drm/i915: Add ringbuffer wait reset to hangcheck
The GPU records whether it is currently waiting for a completion of a
WAIT_FOR_EVENT in the RB_WAIT bit in the ringbuffer control registers.
On third generation chipsets and later, a write of 1 to this bit breaks
the hang and returns the GPU to arbitration, i.e. the GPU should
continue executing the reminder of the batchbuffer and return to normal
operations.
By adding this to hangcheck we can avoid a full GPU reset under these
conditions.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Chris Wilson [Sun, 8 Aug 2010 10:53:53 +0000 (11:53 +0100)]
drm/i915: Clear scanline waits after disabling the pipe.
If we disable the pipe and the GPU is currently waiting on a scanline
WAIT_FOR_EVENT, the GPU will hang. Fortunately, there is a magic bit
which we can write on i915+ to break this wait after disabling the
pipe.
References:
Bug 29252 - [Arrandale] Hung WAIT_FOR_EVENT when running rss-glx-skyrocket
https://bugs.freedesktop.org/show_bug.cgi?id=29252
Bug 28964 - [i965gm] GPU infinite MI_WAIT_FOR_EVENT while watching video in Totem
https://bugs.freedesktop.org/show_bug.cgi?id=28964
and many others.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Chris Wilson [Sat, 3 Jul 2010 06:58:38 +0000 (07:58 +0100)]
drm/i915: Kill the active list spinlock
This spinlock only served debugging purposes in a time when we could not
be sure of the mutex ever being released upon a GPU hang. As we now
should be able rely on hangcheck to do the job for us (and that error
reporting should not itself require the struct mutex) we can kill the
incomplete attempt at protection.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Chris Wilson [Thu, 19 Aug 2010 07:19:30 +0000 (08:19 +0100)]
drm/i915: Compile out error state without DEBUG_FS
Alexander reported that the compilation of intel_overlay.c was failing
due to an inclusion that was only valid with CONFIG_DEBUG_FS. As the
whole error reporting is only useful with debugfs enabled, remove all
the redundant error state collection code when compiling without
CONFIG_DEBUG_FS.
Reported-by: Alexander Lam <lambchop468@gmail.com> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
On i830, there exists a bug where an overlay on pipe B requires the mode
clock on pipe A in order to activate. So workaround this by activating
pipe A when trying to enable the overlay on pipe B.
References:
[Bug 29007] GPU hang on video playback with overlay
https://bugs.freedesktop.org/show_bug.cgi?id=29007
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Chris Wilson [Thu, 12 Aug 2010 11:07:32 +0000 (12:07 +0100)]
drm/i915/overlay: Combine SWITCH_OFF into a single step
We can program the h/w to first wait on the flip and then switch off
without relying on s/w intervention. This removes the need for a double
step switch off, bringing much rejoicing.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
projects / linux-beck.git / log