firefly-linux-kernel-4.4.55.git
12 years agodrm/i915: rip out intel_dpio_write again
Daniel Vetter [Wed, 2 May 2012 09:06:23 +0000 (11:06 +0200)]
drm/i915: rip out intel_dpio_write again

Unfortunately it looks like further vlv patches are still stalled due
to fried hw, and too many people are a bit annoyed about the unused
function warning.

So let's just rip it out, we can easily put it back in again.

Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: extract some common olr+wedge code
Ben Widawsky [Thu, 26 Apr 2012 03:50:12 +0000 (20:50 -0700)]
drm/i915: extract some common olr+wedge code

The new wait_rendering ioctl also needs to check for an oustanding
lazy request, and we already duplicate that logic at three places. So
extract it.

While at it, also extract the code to check the gpu wedging state to
improve code flow.

v2: Don't use seqno as an outparam (Chris)

v3 by danvet: Kill stale comment and pimp commit message

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: move pnv|ilk_gem_mem_freq to intel_pm.c
Daniel Vetter [Thu, 26 Apr 2012 21:28:17 +0000 (23:28 +0200)]
drm/i915: move pnv|ilk_gem_mem_freq to intel_pm.c

Because this is the place where we actually use the results of
them.

Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: fixup __iomem mixups in ringbuffer.c
Daniel Vetter [Thu, 26 Apr 2012 21:28:16 +0000 (23:28 +0200)]
drm/i915: fixup __iomem mixups in ringbuffer.c

Two things:
- ring->virtual start is an __iomem pointer, treat it accordingly.
- dev_priv->status_page.page_addr is now always a cpu addr, no pointer
  casting needed for that.

Take the opportunity to remove the unnecessary drm indirection when
setting up the ringbuffer iomapping.

v2: Add a compiler barrier before reading the hw status page.

Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: rework legacy GFX HWS handling
Daniel Vetter [Thu, 26 Apr 2012 21:28:15 +0000 (23:28 +0200)]
drm/i915: rework legacy GFX HWS handling

To get the fun stuff out of the way, the legacy hws is allocated by
userspace when the gpu needs a gfx hws. And there's no reference-counting
going on, so userspace can simply screw everyone over.

At least it's not as horrible as i810, where the ringbuffer is allocated
by userspace ...

We can't fix this disaster, but we can at least tidy up the code a
bit to make things clearer:
- Drop the drm ioremap indirection.
- Add a new new read_legacy_status_page to paper over the differences
  between the legacy gfx hws and the physical hws shared with the
  new ringbuffer code.
- Add a pointer in dev_priv->dri1 for the cpu addresses - that one is
  an iomem remapping as opposed to all other hw status pages. This is
  just prep work to make sparse happy.

Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: kill pointless clearing of dev_priv->hws_map
Daniel Vetter [Thu, 26 Apr 2012 21:28:14 +0000 (23:28 +0200)]
drm/i915: kill pointless clearing of dev_priv->hws_map

We kzalloc dev_priv, and we never use hws_map in intel_ringbuffer.c.

Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: move rps/emon function declarations
Daniel Vetter [Thu, 26 Apr 2012 21:28:13 +0000 (23:28 +0200)]
drm/i915: move rps/emon function declarations

They're now in intel_pm.c, so group them a bit better.

Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: move the ips code to intel_pm.c
Daniel Vetter [Thu, 26 Apr 2012 21:28:12 +0000 (23:28 +0200)]
drm/i915: move the ips code to intel_pm.c

We now have a nice home for power management code, so let's use it!

v2: Resolve conflict agains "Only enable IPS polling for gen5"

Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: disallow clip rects on gen5+
Daniel Vetter [Thu, 26 Apr 2012 21:28:11 +0000 (23:28 +0200)]
drm/i915: disallow clip rects on gen5+

Unfortunately there has been dri1 userspace that used gem to manage
the gtt and hence also needed cliprects in the execbuf ioctl. So
we can't ever remove that code without breaking the ioctl abi.

But at least we can disable it on gen5+, because these horrible
versions of mesa have not supported these chips.

Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: move LP_RING&friends to i915_dma.c
Daniel Vetter [Thu, 26 Apr 2012 21:28:10 +0000 (23:28 +0200)]
drm/i915: move LP_RING&friends to i915_dma.c

Wohoo!

Now we only need to move all the gem/kms stuff that accidentally
landed in i915_dma.c out of it, and this will be our legacy dri1
grave-yard.

Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: extract dri1 breadcrumb update from irq handler
Daniel Vetter [Thu, 26 Apr 2012 21:28:09 +0000 (23:28 +0200)]
drm/i915: extract dri1 breadcrumb update from irq handler

... and hide it in i915_dma.c.

This way all the legacy stuff dealing with READ_BREADCRUMB and
LP_RING and friends is in i915_dma.c.

v2: Rebase on top of Chris Wilson's rework irq handling code.

Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: move dri1 irq ioctl code to i915_dma.c
Daniel Vetter [Thu, 26 Apr 2012 21:28:08 +0000 (23:28 +0200)]
drm/i915: move dri1 irq ioctl code to i915_dma.c

Let's just get this out of the way.

v2: Rebase against ENODEV changes.

Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: rip out dri1 breadcrumb updates from gen5+ irq handlers
Daniel Vetter [Thu, 26 Apr 2012 21:28:07 +0000 (23:28 +0200)]
drm/i915: rip out dri1 breadcrumb updates from gen5+ irq handlers

We never supported dri1 on gen5+.

VLV never had that code, so no need to remove it.

Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: kill intel_clear_scanline_wait
Daniel Vetter [Thu, 26 Apr 2012 21:28:06 +0000 (23:28 +0200)]
drm/i915: kill intel_clear_scanline_wait

This is a pretty racy way to close these races, and we have
much better means to cope with these races meanwhile: For
non-broken userspace we correctly wait for any outstanding
rendering, for broken userspace the hangcheck will save the
day.

Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: remove LP_RING&friends from modeset code
Daniel Vetter [Thu, 26 Apr 2012 21:28:05 +0000 (23:28 +0200)]
drm/i915: remove LP_RING&friends from modeset code

The LP refers to 'low priority' as opposed to the high priority
ring on gen2/3. So lets constrain its use to the code of that era.

Unfortunately we can't yet completely remove the associated
macros from common headers and shove them into i915_dma.c to
the other dri1 legacy support code, a few cleanups are still
missing for that.

Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: rip out dev_priv->tex_lru_log_granularity
Daniel Vetter [Thu, 26 Apr 2012 21:28:04 +0000 (23:28 +0200)]
drm/i915: rip out dev_priv->tex_lru_log_granularity

Assigned in setparam, used never.

I didn't bother to dig through the archives to figure out what
this was supposed to do.

Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915 disallow physical batchbuffers for KMS
Daniel Vetter [Thu, 26 Apr 2012 21:28:03 +0000 (23:28 +0200)]
drm/i915 disallow physical batchbuffers for KMS

Even the horrible gen3 XvMC code has learned to do this
right by the time xf86-video-intel releases learned to do
kernel modesetting. So we can just disallow this.

Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: create dev_priv->dri1 dragon dungeon^W^W sub-struct
Daniel Vetter [Wed, 2 May 2012 09:49:32 +0000 (11:49 +0200)]
drm/i915: create dev_priv->dri1 dragon dungeon^W^W sub-struct

... and shove allow_batchbuffer in there. More dragons will
follow suit.

There's the curious case that we allow this for KMS ...

Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: move dri1 vblank stubs to i915_dma.c
Daniel Vetter [Thu, 26 Apr 2012 21:28:01 +0000 (23:28 +0200)]
drm/i915: move dri1 vblank stubs to i915_dma.c

i915_dma.c contains most of the old dri1 horror-show, so move
the remaining bits there, too. The code has been removed and
the only thing left are some stubs to ensure that userspace
doesn't try to use this stuff. vblank_pipe_set only returns 0
without any side-effects, so we can even stub it out with
the canonical drm_noop.

v2: Rebase against ENODEV changes.

Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: Remove unused dev_priv->vblank_pipe
Chris Wilson [Mon, 30 Apr 2012 18:17:35 +0000 (19:17 +0100)]
drm/i915: Remove unused dev_priv->vblank_pipe

vblank_pipe was intended to be used for tracking DRI1 state. However,
the vblank_pipe reported to DRI1 is fixed to umask both pipes, and the
dev_priv->vblank_pipe unused and superfluous.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: Only enable IPS polling for gen5
Chris Wilson [Mon, 30 Apr 2012 18:35:02 +0000 (19:35 +0100)]
drm/i915: Only enable IPS polling for gen5

On SandyBridge IPS was entirely implemented in hardware and not reliant
on the driver monitoring power consumption and feeding back desired run
states, so the hardware is able to adapt quicker and more flexibly. Which
is a huge relief for us as we no longer have to carry empirically
derived magic algorithms.

Yet despite the advance in technology, the driver was still doing its
IPS polling on all machines. Restrict it to the only supported hardware,
Clarkdale/Arrandale.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: Remove unused ring->irq_seqno
Chris Wilson [Sat, 28 Apr 2012 08:00:03 +0000 (09:00 +0100)]
drm/i915: Remove unused ring->irq_seqno

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: use __wait_seqno for ring throttle
Ben Widawsky [Thu, 26 Apr 2012 23:03:05 +0000 (16:03 -0700)]
drm/i915: use __wait_seqno for ring throttle

It turns out throttle had an almost identical  bit of code to do the
wait. Now we can call the new helper directly.  This is just a bonus,
and not needed for the overall series.

v2: remove irq_get/put which is now in __wait_seqno (Ben)

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: remove polled wait from throttle
Ben Widawsky [Thu, 26 Apr 2012 23:03:04 +0000 (16:03 -0700)]
drm/i915: remove polled wait from throttle

It's about to go away anyway. Just here to help bisection.

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: extract __wait_seqno from i915_wait_request
Ben Widawsky [Thu, 26 Apr 2012 23:03:03 +0000 (16:03 -0700)]
drm/i915: extract __wait_seqno from i915_wait_request

i915_wait_request is actually a fairly large function encapsulating
quite a few different operations. Because being able to wait on seqnos
in various conditions is useful, extracting that bit of code to a helper
function seems useful

v2: pull the irq_get/put as well (Ben)

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: drop polled waits from i915_wait_request
Ben Widawsky [Thu, 26 Apr 2012 23:03:02 +0000 (16:03 -0700)]
drm/i915: drop polled waits from i915_wait_request

The only time irq_get should fail is during unload or suspend. Both of
these points should try to quiesce the GPU before disabling interrupts
and so the atomic polling should never occur.

This was recommended by Chris Wilson as a way of reducing added
complexity to the polled wait which I introduced in an RFC patch.

09:57 < ickle_> it's only there as a fudge for waiting after irqs
after uninstalled during s&r, we aren't actually meant to hit it
09:57 < ickle_> so maybe we should just kill the code there and fix the breakage

v2: return -ENODEV instead of -EBUSY when irq_get fails

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: kill waiting_seqno
Ben Widawsky [Thu, 26 Apr 2012 23:03:01 +0000 (16:03 -0700)]
drm/i915: kill waiting_seqno

The waiting_seqno is not terribly useful, and as such we can remove it
so that we'll be able to extract lockless code.

v2: Keep the information for error_state (Chris)
Check if ring is initialized in hangcheck (Chris)
Capture the waiting ring (Chris)

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
[danvet: add some bikeshed to clarify a comment.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: move vbetool invoked ier stuff
Ben Widawsky [Thu, 26 Apr 2012 23:03:00 +0000 (16:03 -0700)]
drm/i915: move vbetool invoked ier stuff

This extra bit of interrupt enabling code doesn't belong in the wait
seqno function. If anything we should pull it out to a helper so the
throttle code can also use it. The history is a bit vague, but I am
going to attempt to just dump it, unless someone can argue otherwise.

Removing this allows for a shared lock free wait seqno function. To keep
tabs on this issue though, the IER value is stored on error capture
(recommended by Chris Wilson)

v2: fixed typo EIR->IER (Ben)
Fix some white space (Ben)
Move IER capture to globally instead of per ring (Ben)

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
[danvet: ier is a 16 bit reg on gen2!]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: remove do_retire from i915_wait_request
Ben Widawsky [Thu, 26 Apr 2012 23:02:58 +0000 (16:02 -0700)]
drm/i915: remove do_retire from i915_wait_request

This originates from a hack by me to quickly fix a bug in an earlier
patch where we needed control over whether or not waiting on a seqno
actually did any retire list processing. Since the two operations aren't
clearly related, we should pull the parameter out of the wait function,
and make the caller responsible for retiring if the action is desired.

The only function call site which did not get an explicit retire_request call
(on purpose) is i915_gem_inactive_shrink(). That code was already calling
retire_request a second time.

v2: don't modify any behavior excepit i915_gem_inactive_shrink(Daniel)

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: use the new masked bit macro some more
Daniel Vetter [Thu, 26 Apr 2012 20:02:54 +0000 (22:02 +0200)]
drm/i915: use the new masked bit macro some more

I've missed this one.

v2: Chris Wilson noticed another register.
v3: Color choice improvements.

Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: Remove unused and unloved vblank macros
Chris Wilson [Tue, 24 Apr 2012 21:59:52 +0000 (22:59 +0100)]
drm/i915: Remove unused and unloved vblank macros

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: Inline I915_INTERRUPT_ENABLE_FIX
Chris Wilson [Tue, 24 Apr 2012 21:59:51 +0000 (22:59 +0100)]
drm/i915: Inline I915_INTERRUPT_ENABLE_FIX

Since there is only one remaining user of I915_INTERRUPT_ENABLE_FIX,
expand it at the callsite. Quoting Jesse Barnes:

"I'd really like to get rid of these defines at the top of i915_irq.c.
Some are unused and the others just make you check for the right bits
everytime your read the code."

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
[danvet: Add bikeshed suggested by Jesse.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: Handle PendingFlip on gen3 robustly
Chris Wilson [Tue, 24 Apr 2012 21:59:50 +0000 (22:59 +0100)]
drm/i915: Handle PendingFlip on gen3 robustly

We appear to allow too many pending pageflips as evidenced by an
apparent pin-leak. So borrow the pageflip completion logic from i8xx for
handling PendingFlip in a robust manner.

v2: Address Jesse's reminders about the nuances of gen3 IRQ handling.

References: https://bugzilla.kernel.org/show_bug.cgi?id=41882
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: Cleanup gen3 irq uninstall
Chris Wilson [Tue, 24 Apr 2012 21:59:49 +0000 (22:59 +0100)]
drm/i915: Cleanup gen3 irq uninstall

Bring the for-each-pipe loops together so that the code is easier on the
eyes.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: HWSTAM is only 16-bit on gen3
Chris Wilson [Tue, 24 Apr 2012 21:59:48 +0000 (22:59 +0100)]
drm/i915: HWSTAM is only 16-bit on gen3

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: Remove gen4 irq code from gen3 irq routine
Chris Wilson [Tue, 24 Apr 2012 21:59:47 +0000 (22:59 +0100)]
drm/i915: Remove gen4 irq code from gen3 irq routine

A couple of miscellaneous cleanups as well to move per-loop condition
variables within the scope of the loop and the update of the DRI1
breadcrumb to the tail of the function.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: Remove gen3 irq code from gen4 irq routine
Chris Wilson [Tue, 24 Apr 2012 21:59:46 +0000 (22:59 +0100)]
drm/i915: Remove gen3 irq code from gen4 irq routine

And a couple of miscellaneous cleanups to the main body of the IRQ loop;
move per-loop condition variables within the scope of the loop and move
the old DRI1 breadcrumb to the tail of the function and so only execute
it once.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: Clear FlipDone semantics change for pageflipping on gen3
Chris Wilson [Tue, 24 Apr 2012 21:59:45 +0000 (22:59 +0100)]
drm/i915: Clear FlipDone semantics change for pageflipping on gen3

On later gen3, you are able to select the meaning of the FlipPending
status bit in IIR and change it to FlipDone. This was sometimes done by
the BIOS leading to confusion on just how pageflipping worked on gen3.
Simplify the implementation by using the legacy meaning for all gen3
machines.

Note: this makes all gen3 machines equally broken...

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: Duplicate and split the gen3/4 irq handler
Chris Wilson [Tue, 24 Apr 2012 21:59:44 +0000 (22:59 +0100)]
drm/i915: Duplicate and split the gen3/4 irq handler

In preparation for rewriting the gen3 irq handler.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: pending_flip_is_done is gen3, name it so
Chris Wilson [Tue, 24 Apr 2012 21:59:43 +0000 (22:59 +0100)]
drm/i915: pending_flip_is_done is gen3, name it so

And remove the cargo-culted copy from the valleyview irq handler.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: Remove redundant initialisation of per-ring IRQ waitqueues
Chris Wilson [Tue, 24 Apr 2012 21:59:42 +0000 (22:59 +0100)]
drm/i915: Remove redundant initialisation of per-ring IRQ waitqueues

The waitqueues are already initialised during ring initialisation so
kill the redundant and duplicated code to do so in each generations IRQ
installer.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: Unconditionally initialise the interrupt workers
Chris Wilson [Tue, 24 Apr 2012 21:59:41 +0000 (22:59 +0100)]
drm/i915: Unconditionally initialise the interrupt workers

Rather than duplicate similar code across the IRQ installers, perform
the initialisation of the workers upfront. This will lead to simpler
teardown and quiescent code as we can assume that the workers have
been initialised.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: check for kms in dri1 ioctls
Daniel Vetter [Tue, 24 Apr 2012 06:29:42 +0000 (08:29 +0200)]
drm/i915: check for kms in dri1 ioctls

Calling these when gem assumes full control of the hw won't end
in anything else than tears. So be a bit more paranoid here.

Just serves as documentation.

v2: Bail out with ENODEV as suggested by Chris Wilson.

Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: rip out dev_priv->has_gem
Daniel Vetter [Mon, 23 Apr 2012 14:50:51 +0000 (16:50 +0200)]
drm/i915: rip out dev_priv->has_gem

Always true these days. It has been added originally to work
around some issues with the agp layer in 2.6.29:

commit ac5c4e76180a74c7f922f6fa71ace0cef45fa433
Author: Dave Airlie <airlied@redhat.com>
Date:   Fri Dec 19 15:38:34 2008 +1000

    drm/i915: GEM on PAE has problems - disable it for now.

Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: rip out GEM drm feature checks
Daniel Vetter [Mon, 23 Apr 2012 14:50:50 +0000 (16:50 +0200)]
drm/i915: rip out GEM drm feature checks

We always set it so there's no point in checking. We could
instead add a bit that tells us whether gem is actually
initialized (i.e. either kms or gem_init_ioctl called), but
that's imho not worth it.

So just rip it out.

There's a little change in the wait_ring timeout, but we've never
run with anything else than the 60 second timeout, even on dri1
userspace.

Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: disallow gem ums init ioctl for kms
Daniel Vetter [Tue, 24 Apr 2012 06:22:52 +0000 (08:22 +0200)]
drm/i915: disallow gem ums init ioctl for kms

This ioctl used in a kms driver is only useful to create massive
havoc.

v2: Bail out with -ENODEV as suggested by Chris Wilson.

Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: properly check for MODESET for kms driver ioctls
Daniel Vetter [Tue, 24 Apr 2012 07:55:08 +0000 (09:55 +0200)]
drm/i915: properly check for MODESET for kms driver ioctls

Also ditch the cargo-culted dev_priv checks - either we have a
giant hole in our setup code or this is useless. Plainly bogus
to check for it in either case.

v2: Chris Wilson noticed that I've missed one bogus dev_priv check.

v3: The check in the overlay code is redundant (Chris)

Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: CR clock gating is recommend to be set on PineView
Chris Wilson [Tue, 24 Apr 2012 13:51:43 +0000 (14:51 +0100)]
drm/i915: CR clock gating is recommend to be set on PineView

The specs recommend that this bit be set on PineView. No reason is
given, but it sounds like a powersaving bit that we should expect the
BIOS to be setting...

v2: Rebase on top of _MASKED_ENABLE_BIT

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: Move GEM initialisation from i915_dma.c to i915_gem.c
Chris Wilson [Tue, 24 Apr 2012 14:47:41 +0000 (15:47 +0100)]
drm/i915: Move GEM initialisation from i915_dma.c to i915_gem.c

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: Split the stolen handling for GEM out of i915_dma.c
Chris Wilson [Tue, 24 Apr 2012 14:47:39 +0000 (15:47 +0100)]
drm/i915: Split the stolen handling for GEM out of i915_dma.c

We slightly modify the initialisation sequence to move the
initialisation of the memory managers earlier and in particular before
probing outputs and detecting any existing output configuration. This is
essential if we wish to track preallocated objects and preserve them
whilst initialising GEM.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: Remove the deferred-free list
Chris Wilson [Tue, 24 Apr 2012 14:47:31 +0000 (15:47 +0100)]
drm/i915: Remove the deferred-free list

The use of the mm_list by deferred-free breaks the following patches to
extend the range of objects tracked. We can simplify things if we just
make the unbind during free uninterrutible.

Note that unbinding should never fail, because we hold an additional
reference on every active object. Only the ilk vt-d workaround breaks
this, but already takes care of not failing by waiting for the gpu to
quiescent non-interruptible. But the existence of the deferred free
list casted some doubts on this theory, hence WARN if the unbind fails
and only then retry non-interruptible.

We can kill this additional code after a release in case the theory is
indeed right and no one has hit that WARN.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: Remove the list of pinned inactive objects
Chris Wilson [Tue, 24 Apr 2012 14:47:30 +0000 (15:47 +0100)]
drm/i915: Remove the list of pinned inactive objects

Simplify object tracking by removing the inactive but pinned list. The
only place where this was used is for counting the available memory,
which is just as easy performed by checking all objects on the rare
occasions it is required (application startup). For ease of debugging,
we keep the reporting of pinned objects through the error-state and
debugfs.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: Remove i915_gem_evict_inactive()
Chris Wilson [Tue, 24 Apr 2012 17:22:52 +0000 (18:22 +0100)]
drm/i915: Remove i915_gem_evict_inactive()

This was only used by one external caller who would just be as happy
with evict-everything, so perform the replacement and make the function
private.

In the process we note that unbinding the inactive list should not fail,
and make it a warning instead.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: Bump the inactive LRU on set-to-GTT-domain
Chris Wilson [Tue, 24 Apr 2012 14:52:35 +0000 (15:52 +0100)]
drm/i915: Bump the inactive LRU on set-to-GTT-domain

Currently, we only bump the inactive LRU of an object when we bind
into the GTT for a page-fault. As the object may be used many times
before its mapping is zapped, we do not mark it as active as
frequently as we should. Userspace should be calling set-to-GTT-domain
before each pointer deference (for synchronous access) and so is a
good place to mark the buffer as active.

Marking the buffer as recently used places it at the end of the
inactive eviction queue, though still before anything with outstanding
rendering. This reduces the likelihood of evicting a buffer that is
going to be used again by the CPU in the near future. This way we can
hopefully avoid to kick out upload buffers right before we use them on
the gpu.

Note that we need to check that the object is not active or pinned,
for otherwise we create havoc on the active/pinned lists, which also
use obj->mm_list.

The active lists are sorted by and evicted in last GPU rendering
order, access by the CPU to a still active buffer therefore does not
affect its eviction ordering. Pinned objects are currently excluded
from eviction, therefore the only list that we need to bump for GTT
access by the CPU is the inactive list.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
[danvet: Added further explanations to the commit message as discussed
on irc.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: Remove too early plane enable on pre-PCH hardware
Chris Wilson [Tue, 24 Apr 2012 15:36:50 +0000 (16:36 +0100)]
drm/i915: Remove too early plane enable on pre-PCH hardware

Enabling the plane before we have assigned valid address means that it
will access random PTE (often with conflicting memory types) and cause
GPU lockups. However, enabling the plane too early appears to workaround
a number of bugs in our modesetting code.

Cc: Franz Melchior <melchior.franz@gmail.com>
References: https://bugs.freedesktop.org/show_bug.cgi?id=39947
References: https://bugs.freedesktop.org/show_bug.cgi?id=41091
References: https://bugs.freedesktop.org/show_bug.cgi?id=49041
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: Use a global lock for modifying global irq flags
Chris Wilson [Tue, 24 Apr 2012 20:48:47 +0000 (21:48 +0100)]
drm/i915: Use a global lock for modifying global irq flags

We were attempting to use a per-ring spinlock whilst modifying global
IRQ flags. A recipe for rare missed interrupts.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: implement Disable4x2SubspanOptimization w/a for ivb, too
Daniel Vetter [Tue, 24 Apr 2012 14:00:21 +0000 (16:00 +0200)]
drm/i915: implement Disable4x2SubspanOptimization w/a for ivb, too

Copy&pasted from the vlv setup code. According to docs, we need that
on ivb, too.

v2: Use new masked bit handling macros.

Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: create macros to handle masked bits
Daniel Vetter [Tue, 24 Apr 2012 12:04:12 +0000 (14:04 +0200)]
drm/i915: create macros to handle masked bits

... and put them to so good use.

Note that there's functional change in vlv clock gating code, we now
no longer spuriously read back the current value of the bit. According
to Bspec the high bits should always read zero, so ORing this in
should have no effect.

Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: manage PCH PLLs separately from pipes
Jesse Barnes [Fri, 20 Apr 2012 16:11:53 +0000 (17:11 +0100)]
drm/i915: manage PCH PLLs separately from pipes

PCH PLLs aren't required for outputs on the CPU, so we shouldn't just
treat them as part of the pipe.

So split the code out and manage PCH PLLs separately, allocating them
when needed or trying to re-use existing PCH PLL setups when the timings
match.

v2: add num_pch_pll field to dev_priv (Daniel)
    don't NULL the pch_pll pointer in disable or DPMS will fail (Jesse)
    put register offsets in pll struct (Chris)

v3: Decouple enable/disable of PLLs from get/put.
v4: Track temporary PLL disabling during modeset
v5: Tidy PLL initialisation by only checking for num_pch_pll == 0 (Eugeni)
v6: Avoid mishandling allocation failure by embedding the small array of
    PLLs into the device struct

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=44309
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org> (up to v2)
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> (v3+)
Reviewed-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Tested-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: i8xx interrupt handler
Chris Wilson [Sun, 22 Apr 2012 20:13:57 +0000 (21:13 +0100)]
drm/i915: i8xx interrupt handler

gen2 hardware has some significant differences from the other interrupt
routines that were glossed over and then forgotten about in the
transition to KMS. Such as

- 16bit IIR
- PendingFlip status bit

This patch reintroduces a handler specifically for gen2 for the purpose
of handling pageflips correctly, simplifying code in the process.

v2: Also fixup ring get/put irq to only access 16bit registers (Daniel)

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=24202
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=41793
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
[danvet: use posting_read16 in intel_ringbuffer.c and kill _driver
from the function names.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: Only the zap the VMA after updating the tiling parameters
Chris Wilson [Sat, 21 Apr 2012 15:23:24 +0000 (16:23 +0100)]
drm/i915: Only the zap the VMA after updating the tiling parameters

If we fail to unbind and so abort the change in tiling, we will have
removed the VMA for the object for no reason. The likelihood of unbind
failing is slim (other than ERESTARTSYS which will cause userspace to
try again), so the change is mostly for the principle.

Also improve the slightly stale comment.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: Clarify the semantics of tiling_changed
Chris Wilson [Sat, 21 Apr 2012 15:23:23 +0000 (16:23 +0100)]
drm/i915: Clarify the semantics of tiling_changed

Rename obj->tiling_changed to obj->fence_dirty so that it is clear that
it flags when the parameters for an active fence (including the
no-fence) register are changed.

Also, do not set this flag when the object does not have a fence
register allocated currently and the gpu does not depend upon the
unfence. This case works exactly like when a tiled object lost its
fence and hence does not need additional handling for the tiling
change in the code.

v2: Use fence_dirty to better express what the flag tracks and add a few
more details to the comments to serve as a reminder of how the GPU also
uses the unfenced register slot.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
[danvet: Add some bikeshed to the commit message about the stricter
use of fence_dirty.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: fixup tv load-detect on enabled but not active crtc
Daniel Vetter [Fri, 20 Apr 2012 19:25:04 +0000 (21:25 +0200)]
drm/i915: fixup tv load-detect on enabled but not active crtc

When fixing up the crt load detect code I've failed to notice the same
problem in the tv load detect code. Again, unconditionally use the
load detect pipe infrastructure, it gets things right.

Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: re-add static qualifier to get_cxsr_latency
Daniel Vetter [Sat, 21 Apr 2012 21:17:55 +0000 (23:17 +0200)]
drm/i915: re-add static qualifier to get_cxsr_latency

This got lost in the intel_pm.c move.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: [sparse] __iomem fixes for gem
Ben Widawsky [Mon, 16 Apr 2012 21:07:47 +0000 (14:07 -0700)]
drm/i915: [sparse] __iomem fixes for gem

As with one of the earlier patches in the series, we're forced to cast
for copy_[to|from]_user. Again because of the nature of the GEN x86
exclusivity, this should be safe.

Signed-off-by: Ben Widawsky <benjamin.widawsky@intel.com>
[danvet: Added some bikeshed.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: [sparse] __iomem fixes for debugfs
Daniel Vetter [Sat, 21 Apr 2012 20:49:10 +0000 (22:49 +0200)]
drm/i915: [sparse] __iomem fixes for debugfs

These were mostly straight forward. No forced casting needed.

Signed-off-by: Ben Widawsky <benjamin.widawsky@intel.com>
[danvet: fix conflict with ringbuffer_data removal and drop the hunk
about the status page - that needs more care to fix up.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: [sparse] __iomem fixes for overlay
Ben Widawsky [Mon, 16 Apr 2012 21:07:43 +0000 (14:07 -0700)]
drm/i915: [sparse] __iomem fixes for overlay

With the exception of a forced cast for phys_obj stuff (a problem in
other patches as well) all of these are fairly simple __iomem compliance
fixes.

As with other patches, yank/paste errors may exist.

Signed-off-by: Ben Widawsky <benjamin.widawsky@intel.com>
[danvet: Added comment to explain the __iomem cast.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: [sparse] __iomem fixes for opregion
Ben Widawsky [Mon, 16 Apr 2012 21:07:42 +0000 (14:07 -0700)]
drm/i915: [sparse] __iomem fixes for opregion

Almost all of the errors related __iomem problems.

Most of the changes here are trivial, however there is plenty of chance
for yank/paste errors.

Signed-off-by: Ben Widawsky <benjamin.widawsky@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: Remove i915_gem_ringbuffer_(data|info) from debugfs
Chris Wilson [Thu, 19 Apr 2012 16:55:23 +0000 (17:55 +0100)]
drm/i915: Remove i915_gem_ringbuffer_(data|info) from debugfs

This was originally used as an attempt to diagnose GPU hangs, but was
never very reliable and superseded by the i915_error_state capture on
hangcheck. It now lies languishing unused and unwanted.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915/sprite: Avoid incurring extra vblank stall when updating plane on IVB
Chris Wilson [Wed, 18 Apr 2012 16:12:26 +0000 (17:12 +0100)]
drm/i915/sprite: Avoid incurring extra vblank stall when updating plane on IVB

IvyBridge requires an extra frame between disabling the low power
watermarks and enabling scaling on the sprite plane. If the scaling
is already enabled, then we have already disabled the low power
watermarks and need not incur an extra wait.

Similarly, as we disable the scaling when turning off the sprite plane,
we can update the scaling enabled flag and restore the low power
watermarks.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agodrm/i915: print computed bpp in dp link configuration
Daniel Vetter [Fri, 20 Apr 2012 18:23:49 +0000 (20:23 +0200)]
drm/i915: print computed bpp in dp link configuration

Pretty useful to debug our DP bandwidth woes.

v2: Also print out the required and available link bandwidth,
suggested by Chris Wilson.

v3: Also print out the input parameters so that diagnosing failures to
find a valid dp link configuration is possible.

v4: s/Display port/DP/ to shorten the output.

Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agoMerge tag 'drm-intel-next-2012-04-23' of git://people.freedesktop.org/~danvet/drm...
Dave Airlie [Wed, 2 May 2012 08:21:50 +0000 (09:21 +0100)]
Merge tag 'drm-intel-next-2012-04-23' of git://people.freedesktop.org/~danvet/drm-intel into drm-core-next

Daniel Vetter writes:

A new drm-intel-next pull. Highlights:
- More gmbus patches from Daniel Kurtz, I think gmbus is now ready, all
 known issues fixed.
- Fencing cleanup and pipelined fencing removal from Chris.
- rc6 residency interface from Ben, useful for powertop.
- Cleanups and code reorg around the ringbuffer code (Ben&me).
- Use hw semaphores in the pageflip code from Ben.
- More vlv stuff from Jesse, unfortunately his vlv cpu is doa, so less
 merged than I've hoped for - we still have the unused function warning :(
- More hsw patches from Eugeni, again, not yet enabled fully.
- intel_pm.c refactoring from Eugeni.
- Ironlake sprite support from Chris.
- And various smaller improvements/fixes all over the place.

Note that this pull request also contains a backmerge of -rc3 to sort out
a few things in -next. I've also had to frob the shortlog a bit to exclude
anything that -rc3 brings in with this pull.

Regression wise we have a few strange bugs going on, but for all of them
closer inspection revealed that they've been pre-existing, just now
slightly more likely to be hit. And for most of them we have a patch
already. Otherwise QA has not reported any regressions, and I'm also not
aware of anything bad happening in 3.4.

* tag 'drm-intel-next-2012-04-23' of git://people.freedesktop.org/~danvet/drm-intel: (420 commits)
  drm/i915: rc6 residency (fix the fix)
  drm/i915/tv: fix open-coded ARRAY_SIZE.
  drm/i915: invalidate render cache on gen2
  drm/i915: Silence the change of LVDS sync polarity
  drm/i915: add generic power management initialization
  drm/i915: move clock gating functionality into intel_pm module
  drm/i915: move emon functionality into intel_pm module
  drm/i915: move drps, rps and rc6-related functions to intel_pm
  drm/i915: fix line breaks in intel_pm
  drm/i915: move watermarks settings into intel_pm module
  drm/i915: move fbc-related functionality into intel_pm module
  drm/i915: Refactor get_fence() to use the common fence writing routine
  drm/i915: Refactor fence clearing to use the common fence writing routine
  drm/i915: Refactor put_fence() to use the common fence writing routine
  drm/i915: Prepare to consolidate fence writing
  drm/i915: Remove the unsightly "optimisation" from flush_fence()
  drm/i915: Simplify fence finding
  drm/i915: Discard the unused obj->last_fenced_ring
  drm/i915: Remove unused ring->setup_seqno
  drm/i915: Remove fence pipelining
  ...

12 years agodrm/radeon/kms/hdmi: use relative offsets, official regs
Rafał Miłecki [Sat, 28 Apr 2012 21:35:24 +0000 (23:35 +0200)]
drm/radeon/kms/hdmi: use relative offsets, official regs

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Tested-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Christian König <deathsimple@vodafone.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agodrm/radeon/kms: keep HDMI state in separated variable
Rafał Miłecki [Sat, 28 Apr 2012 21:35:23 +0000 (23:35 +0200)]
drm/radeon/kms: keep HDMI state in separated variable

If we want hdmi_offset to be relative to the first block, zero value can
be used also for enabled block.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Tested-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Christian König <deathsimple@vodafone.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agodrm/radeon/kms: get rid of r600_hdmi_find_free_block
Rafał Miłecki [Sat, 28 Apr 2012 21:35:22 +0000 (23:35 +0200)]
drm/radeon/kms: get rid of r600_hdmi_find_free_block

R6xx has routable blocks, but there's nothing wrong in assignment based
on dig_encoder. We didn't really need that algorithm.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Tested-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Christian König <deathsimple@vodafone.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agodrm/radeon/kms: get rid of hdmi_config_offset
Rafał Miłecki [Sat, 28 Apr 2012 21:35:21 +0000 (23:35 +0200)]
drm/radeon/kms: get rid of hdmi_config_offset

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Tested-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Christian König <deathsimple@vodafone.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agodrm/radeon/kms: move audio params to separated struct
Rafał Miłecki [Sat, 28 Apr 2012 21:35:20 +0000 (23:35 +0200)]
drm/radeon/kms: move audio params to separated struct

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Tested-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Christian König <deathsimple@vodafone.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agodrm/edid: fix collision between two patches breaking build
Dave Airlie [Mon, 30 Apr 2012 06:26:16 +0000 (07:26 +0100)]
drm/edid: fix collision between two patches breaking build

this fixes a report that the new load code needed to be updated for
ajax's validity changes.

Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agodrm: fixed: Add dfixed_frac
Robert Morell [Wed, 25 Apr 2012 09:45:01 +0000 (11:45 +0200)]
drm: fixed: Add dfixed_frac

This helper macro retrieves the fractional part of a fixed20_12 20.12
fixed-point number.

Signed-off-by: Robert Morell <rmorell@nvidia.com>
Signed-off-by: Olof Johansson <olofj@chromium.org>
Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agogma500: Set the mapping mask
Alan Cox [Wed, 25 Apr 2012 13:38:47 +0000 (14:38 +0100)]
gma500: Set the mapping mask

Some boards such as the Intel D2700MUD allow you to have over 4GB of RAM.
The GTT on the PVR based devices is 32bit however. Hugh Dickins points out
that we should therefore be setting the mapping gfp mask.

This is not the whole fix for the problem. Some further shmem patches will
be needed to deal with the corner cases.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agogma500: Add the base elements of CDV hotplug support
Alan Cox [Wed, 25 Apr 2012 13:38:32 +0000 (14:38 +0100)]
gma500: Add the base elements of CDV hotplug support

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agogma500: Add ops for hotplug support.
Alan Cox [Wed, 25 Apr 2012 13:38:20 +0000 (14:38 +0100)]
gma500: Add ops for hotplug support.

This provides the needed callback hooks to add hotplug display support to
the GMA36x0 devices.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agocdv: continue synching up with updated reference code
Alan Cox [Wed, 25 Apr 2012 13:38:07 +0000 (14:38 +0100)]
cdv: continue synching up with updated reference code

In particular clean up the errata handling and correct the crtc masks. We do
this a bit differently using our device abstraction for neatness.

This doesn't address the ACPI opregion and hotplug plumbing, nor the IRQ related
changes that will need. It touches on backlight init but the full backlight
support is not in this change set.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agogma500: Clean up weirdness in the cdv mode test code
Alan Cox [Wed, 25 Apr 2012 13:37:53 +0000 (14:37 +0100)]
gma500: Clean up weirdness in the cdv mode test code

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agogma500: support 1080p
Alan Cox [Wed, 25 Apr 2012 13:37:40 +0000 (14:37 +0100)]
gma500: support 1080p

The problem in console mode is lack of linear memory. We can solve that by
dropping to 16bpp. The mode setting X server will allocate its own GEM
framebuffer in 32bpp and all will be well.

We could just do 16bpp anyway but that would be a regression on the lower
modes as many distributions don't yet ship the generic mode setting KMS
drivers.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agogma500: panel presence check
Alan Cox [Wed, 25 Apr 2012 13:37:27 +0000 (14:37 +0100)]
gma500: panel presence check

Introduce a panel presence check for Cedartrail. Non netbook devices don't
necessarily have a panel attached.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agogma500: intel_bios updates
Alan Cox [Wed, 25 Apr 2012 13:37:14 +0000 (14:37 +0100)]
gma500: intel_bios updates

Pull in various i915 bits that we will need to begin tackling the LVDS detect
and ACPI events. We try and drift towards the i915 version of the code with
the long term goal that at least some of it can one day be unified.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agogma500: mark framebuffer pages write combining
Alan Cox [Wed, 25 Apr 2012 13:37:00 +0000 (14:37 +0100)]
gma500: mark framebuffer pages write combining

We don't want them uncached, combining will do nicely and fixes the performance
problem with the generic modesetting X server.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agogma500: Update the Cedarview clock handling
Alan Cox [Wed, 25 Apr 2012 13:36:48 +0000 (14:36 +0100)]
gma500: Update the Cedarview clock handling

Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agogma500: read the PLL bits
Alan Cox [Wed, 25 Apr 2012 13:36:34 +0000 (14:36 +0100)]
gma500: read the PLL bits

We need to pull more stuff from the VBT in order to configure the clocking
correctly in all cases. Add the relevant bits from the other CDV driver work.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agogma500: Fix leak of uncached page
Alan Cox [Wed, 25 Apr 2012 13:36:13 +0000 (14:36 +0100)]
gma500: Fix leak of uncached page

This was reported a long time ago (and I apologize to whoever it was that
reported it as I've lost the original report).

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agodrm: Store vendor IDs directly in the EDID quirk structure
Ian Pilcher [Sun, 22 Apr 2012 16:40:26 +0000 (11:40 -0500)]
drm: Store vendor IDs directly in the EDID quirk structure

EDID vendor IDs are always 3 characters long (4 with the terminating
0).  It doesn't make any sense to have a (possibly 8-byte) pointer
to the ID string in the quirk structure.

Signed-off-by: Ian Pilcher <arequipeno@gmail.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agodrm/edid: Try harder to fix up base EDID blocks
Adam Jackson [Mon, 16 Apr 2012 14:40:08 +0000 (10:40 -0400)]
drm/edid: Try harder to fix up base EDID blocks

Requiring the first byte of the EDID base block header to be 0 means we
don't fix up as many transfer errors as we could.  Instead have the
callers specify whether it's meant to be block 0 or not, and
conditionally run header fixup based on that.

Bugzilla: https://bugzilla.redhat.com/812890
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Alex Deucher <alexdeucher@gmail.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agovga: fix build when fbdev is a module
Matthew Garrett [Tue, 24 Apr 2012 08:31:28 +0000 (09:31 +0100)]
vga: fix build when fbdev is a module

This fixes the build breakage reported by Stephen in -next
when merging the drm-next tree.

Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agodrm: Unify and fix idr error handling
Ville Syrjälä [Thu, 15 Mar 2012 17:58:31 +0000 (19:58 +0200)]
drm: Unify and fix idr error handling

The error handling code w.r.t. idr usage looks inconsistent.

In the case of drm_mode_object_get() and drm_ctxbitmap_next() the error
handling is also incomplete.

Unify the code to follow the same pattern always.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agovga-switcheroo: select VGA arbitration.
Dave Airlie [Mon, 16 Apr 2012 18:01:17 +0000 (19:01 +0100)]
vga-switcheroo: select VGA arbitration.

Since Matthew's changes we have to select arbitration.

Reported-by: devh on #radeon
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agoefifb: Implement vga_default_device() (v2)
Matthew Garrett [Mon, 16 Apr 2012 20:26:05 +0000 (16:26 -0400)]
efifb: Implement vga_default_device() (v2)

EFI doesn't typically make use of the legacy VGA ROM, but it may still be
configured to pass that through to a given video device. This may lead to
an inaccurate choice of default video device. Add support to efifb to pick
out the correct active video device.

v2: fix if->ifdef

Signed-off-by: Matthew Garrett <mjg@redhat.com>
Acked-by: hpa@zytor.com
Cc: matt.fleming@intel.com
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agox86: Use vga_default_device() when determining whether an fb is primary
Matthew Garrett [Mon, 16 Apr 2012 20:26:04 +0000 (16:26 -0400)]
x86: Use vga_default_device() when determining whether an fb is primary

IORESOURCE_ROM_SHADOW is not necessarily an indication that the hardware
is the primary device. Add support for using the vgaarb functions and
fall back if nothing's set them.

Signed-off-by: Matthew Garrett <mjg@redhat.com>
Cc: mingo@redhat.com
Acked-by: hpa@zytor.com
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agovga-switcheroo: Use vga_default_device()
Matthew Garrett [Mon, 16 Apr 2012 20:26:03 +0000 (16:26 -0400)]
vga-switcheroo: Use vga_default_device()

vga-switcheroo currently changes the default VGA device by fiddling with
the IORESOURCE_ROM_SHADOW flag on the device. This isn't strictly accurate,
since there's no guarantee that switching also changes the ROM decoding.
Switch over to using the vgaarb functions for this.

Signed-off-by: Matthew Garrett <mjg@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agovgaarb: Add support for setting the default video device (v2)
Matthew Garrett [Mon, 16 Apr 2012 20:26:02 +0000 (16:26 -0400)]
vgaarb: Add support for setting the default video device (v2)

The default VGA device is a somewhat fluid concept on platforms with
multiple GPUs. Add support for setting it so switching code can update
things appropriately, and make sure that the sysfs code returns the right
device if it's changed.

v2: Updated to fix builds when __ARCH_HAS_VGA_DEFAULT_DEVICE is false.

Signed-off-by: Matthew Garrett <mjg@redhat.com>
Acked-by: H. Peter Anvin <hpa@zytor.com>
Acked-by: benh@kernel.crashing.org
Cc: airlied@redhat.com
Signed-off-by: Dave Airlie <airlied@redhat.com>