I changed 945's self refresh to work without the need for the driver to
enable/disable self refresh manually based on the idle state of the gpu.
This is much better than enabling/disabling self refresh for various
reasons, including staying in a lower power state for more time and
avoiding the need for cpu cycles.

This was originally done manually to workaround issues with the hardware
hanging. However, since 94400120: drm/i915: enable low power render
writes on GEN3 hardware, automatic CxSR seems stable.
Signed-off-by: NAlexander Lam <lambchop468@gmail.com>
Acked-by : Li Peng <peng.li@linux.intel.com>
[ickle: play safe with the ordering and disable CxSR before tweaking any
watermark and enable afterwards.]
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

945 class hardware has an interesting quirk in which the vblank
interrupt is not raised if the CPU is in a low power state. (We also
suspect that the memory bus is clocked to the CPU/c-state and not the
GPU so there are secondary starvation issues.) In order to prevent the
most obvious issue of the low of the vblank interrupt (stuttering
compositing that only updates when the mouse is moving) is to install a
PM QoS request to prevent low c-states whilst the GPU is active.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

eDP on the CPU doesn't need the PCH set up at all, it can in fact cause
problems.  So avoid FDI training and PCH PLL enabling in that case.
Signed-off-by: NJesse Barnes <jbarnes@virtuousgeek.org>
Tested-by: NYuanhan Liu <yuanhan.liu@linux.intel.com>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

We need to unlock the phase sync pointer enable bit before we can
actually enable the phase sync pointer workaround on Ironlake.
Signed-off-by: NJesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

Factor out the FDI disable function (make it a mirror of
ironlake_fdi_enable) and add some FDI related assertions to the FDI
training code (we need an active pipe & plane before we start
transmitting bits).
Signed-off-by: NJesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

Along with assertion checks for the FDI transmitters and receivers
(including PLLs).  Modify the pipe enable function to check for FDI PLL
status as well, when driving PCH ports.
Signed-off-by: NJesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

Otherwise our writes will be silently ignored.
Signed-off-by: NJesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

With assertions to check transcoder and reference clock state.
Signed-off-by: NJesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

For pre-ILK only.  Saves some code in the CRTC enable/disable functions
and allows us to check for pipe and panel status at enable/disable time.
Signed-off-by: NJesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

When PLLs or timing regs are changed, we need to make sure the panel
lock will allow it.
Signed-off-by: NJesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

Add plane enable/disable functions to prevent duplicated code and allow
us to easily check for plane enable/disable requirements (such as pipe
enable, plane status, pll status etc).
Signed-off-by: NJesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

On Ironlake+ we need to enable these in a specific order.
Signed-off-by: NJesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

Admittedly, trusting ACPI or the BIOS at all to be correct is littered
with numerous examples where it is wrong. Maybe, just maybe, we will
have better luck using the ACPI OpRegion lid status...
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

First, we were calling mc_stop() at the top of the function
which turns off all MC (memory controller) clients,
then checking if the GPU is idle.  If it was idle we
returned without re-enabling the MC clients which would
lead to a blank screen, etc.  This patch checks if the
GPU is idle before calling mc_stop().

Second, if the reset failed, we were returning without
re-enabling the MC clients.  This patch re-enables
the MC clients before returning regardless of whether
the reset was successful or not.
Signed-off-by: NAlex Deucher <alexdeucher@gmail.com>
Cc: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Use the common one for all asics.
Signed-off-by: NAlex Deucher <alexdeucher@gmail.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Switching to pcie gen2 causes problems on some
boards.  Add a module option to turn it on/off.

There are gen2 compatability issues with some
motherboards it seems.

Fixes:
https://bugs.freedesktop.org/show_bug.cgi?id=33027Signed-off-by: NAlex Deucher <alexdeucher@gmail.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Signed-off-by: NAlex Deucher <alexdeucher@gmail.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

The switch to separate BAR and channel address spaces made the fbcon memory
address calculation incorrect on NV50+ boards, this commit fixes that.
Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

Reviewed-by: NFrancisco Jerez <currojerez@riseup.net>
Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

Somehow missed this in the original merge of the nvc0 code.
Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

Signed-off-by: NBen Skeggs <bskeggs@redhat.com>

This reverts commit dfe63bb0.

This commit was causing nouveau not to work properly, for -rc1 I'd
prefer it worked and we can look if this is useful for 2.6.39.

Cc: James Simmons <jsimmons@infradead.org>
Signed-off-by: NDave Airlie <airlied@redhat.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Signed-off-by: NKnut Petersen <knut_petersen@t-online.de>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: stable@kernel.org

Hopefully, this is a temporary measure whilst the root cause is
understood. At the moment, we experience a hard hang whilst looping
urbanterror that has been identified as a result of the use of
semaphores, but so far only on SNB mobile.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=32752
Tested-by: mengmeng.meng@intel.com
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

After reordering the sequence of relocating objects, commit 6fe4f140,
we can no longer rely on seeing all reloc targets prior to performing
the relocation. As a result we were ignoring the need to flush objects
from the render cache and invalidate the sampler caches, resulting in
rendering glitches. So we need to clear the relocation domains earlier.
Reported-by: NLinus Torvalds <torvalds@linux-foundation.org>
Tested-by: NLinus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

On the fault path, commit 6fe4f140 introduction a regression whereby it
changed the sequence of the objects but continued to use the original
ordering of relocation entries. The result was that incorrect GTT offsets
were being fed into the execbuffer causing lots of misrendering and
potential hangs.
Reported-by: NLinus Torvalds <torvalds@linux-foundation.org>
Tested-by: NLinus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

Whilst we had no older batches on the active list, everything was fine.
However, if the GPU is free running and the requests are only being
reaped by the periodic retirer, than the current seqno may not be at the
start of the list. In this case we need to select the first batch after
the last seqno written by the gpu and not inclusive of the seqno.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

In order to workaround the issue with LVDS not working on the Lenovo
U160 apparently due to using the wrong SSC frequency, add an option to
disable SSC.
Suggested-by: NLukács, Árpád <lukacs.arpad@gmail.com>
Bugzillla: https://bugs.freedesktop.org/show_bug.cgi?id=32748Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: stable@kernel.org

... and not if the maximum is non-zero. This fixes the typo introduced
in 47356eb6 and preserves the backlight value from boot.

[ickle: My thanks also to Indan Zupancic for diagnosing the original
        regression and suggesting the appropriate fix.]
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: stable@kernel.org # after 47356eb6

Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

As the mappable portion of the aperture is always a small subset at the
start of the GTT, it is allocated preferentially by drm_mm. This is
useful in case we ever need to map an object later. However, if you have
a large object that can consume the entire mappable region of the
GTT this prevents the batchbuffer from fitting and so causing an error.
Instead allocate all those that require a mapping up front in order to
improve the likelihood of finding sufficient space to bind them.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

Rather than evicting an object at random, which is unlikely to alleviate
the memory pressure sufficient to allow us to continue, zap the entire
aperture. That should give the system long enough to recover and reap
some pages from the evicted objects, forestalling the allocation error
for the new object.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

... and not leave the objects in a inconsistent state.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: stable@kernel.org

Before releasing the lock in order to copy the relocation list from user
pages, we need to drop all the object references as another thread may
usurp and execute another batchbuffer before we reacquire the lock.
However, the code was buggy and failed to clear the list...
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: stable@kernel.org

Useful for determining the layout.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

In order to retire active buffers whilst no client is active, we need to
insert our own flush requests onto the ring.

This is useful for servers that queue up some rendering and then go to
sleep as it allows us to the complete processing of those requests,
potentially making that memory available again much earlier.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

Dave Airlie spotted that his ILK laptop with DMAR enabled was generating
the occasional DMAR warning.

"The ordering in the previous code was to rewrite the GTT table before
unmapping the pages and that makes sense to me."

This is his stable patch ported to d-i-n.
Reported-by: NDave Airlie <airlied@redhat.com>
Original-patch-by: NDave Airlie <airlied@redhat.com>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>