In order to handle differences in pte encoding between architectures it
is desirable to have one helper function, pte_encode, do it all for us.
As such, this commit moves the code around so we're in good shape to do
that.

Luckily the ppgtt pte and the ggtt pte look very similar.
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

HSW will change the PTE encoding, and laying this out now will be
helpful when we're ready to implement that. More importantly, GGTT and
PPGTT PTE encoding is quite similar, so moving this out into a helper
function will enable us to lance the AGP layer.
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

This will make the calculations of size easier to read instead of just
assuming uint32_t everywhere.
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

Some subsequent commits will need to know what generation we're running
on to do different pte encoding for the ppgtt. Since it's not much
hassle or overhead to store it in the ppgtt structure, do that.
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

The mid-level cache or as it's more commonly referred to now as L3, is
not setup this way on HSW.
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

Make intel_render_ring_init_dri and intel_init_ring_buffer symmetrical
with regards of workaround introduced by:

commit 27c1cbd0
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Mon Apr 9 13:59:46 2012 +0100

    drm/i915/ringbuffer: Exclude last 2 cachlines of ring on 845g
Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

Now intel_ddi_init is just like intel_hdmi_init and intel_dp_init: it
inits the encoder and then calls the proper init_connector functions.
Notice that for non-eDP ports we call both HDMI and DP connector init,
so we have 2 connectors attached to each DDI encoder.

After this change, intel_hdmi_init and intel_dp_init are only called
by Ivy Bridge and earlier, while hardware containing DDI outputs
should call intel_ddi_init.

Also added/removed quite a few "static" keywords due to the fact that
some function pointers were moved from intel_dp.c and intel_hdmi.c to
intel_ddi.c.

DP finally works on Haswell now! \o/
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

We need this since now on DDI we will have 2 connectors on each
encoder.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

Both "intel_dp" and "intel_hdmi" structs had a "port" field, which
always had the same value. It makes more sense to move this to
intel_digital_port, so we can know the port independently of the
connector type.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

When intel_hdmi_detect detects a monitor, set intel_encoder->type with
INTEL_OUTPUT_HDMI. Same for DP.

This should not break the current code because these variables never
change. This will be used after we create the DDI encoder because it
will have both DP and HDMI connectors.

We won't support eDP+HDMI on the same port, so if an encoder is eDP we
should expect it to always remain eDP and never change.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

Same reason as the previous HDMI commit: the DDI code will have its
own encoder init function but still use the DP and HDMI connectors.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
[danvet: kill the unnecessarily added line that Damien spotted in
review.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

We want to split the HDMI connector and encoder initialization because
in the future the DDI code will have its own "encoder init" function,
but it will still call intel_hdmi_init_connector. The DDI encoder will
actually have two connectors attached to it: HDMI and DP.

The best way to look at this patch is to imagine that we're renaming
intel_hdmi_init to intel_hdmi_init_connector and removing the
encoder-specific pieces and placing them into intel_hdmi_init.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

The goal is to have one single encoder capable of controlling both DP
and HDMI outputs. This patch just adds the initial infrastructure, no
functional changes.

Previously, both intel_dp and intel_hdmi were intel_encoders. Now,
these 2 structs do not have intel_encoder as members anymore. The new
struct intel_digital_port has intel_encoder as a member, and it also
includes intel_dp and intel_hdmi as members. In other words: see the
changes inside intel_drv.h: it's the most important change, everything
else is only to make it compile and work.

For now, each intel_digital_port is still only able to control one of
HDMI or DP, but not both together.

In the future we should also try to merge the common fields from
intel_dp and intel_hdmi (e.g., port).
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
[danvet: Add the missing ' ' spotted by Damien Lespiau.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

When we add struct intel_digital_port, there will be no direct way of
going from intel_{dp,hdmi} to drm_device: we will need to call
container_of().

This patch adds functions to go from intel_{dp,hdmi} to drm_device.
The main goal here is to greatly reduce the size of the next patch,
where we will change the implementation of the functions we just
added here (among other things).
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Damien Lespiau <damien.lespiau@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

 - Replace container_of with enc_to_intel_dp.
 - Walk through less structures when making assignments.
 - Rename some variables to keep our naming standards.

As a bonus, this will reduce the usage of "struct intel_dp", making
the future patch that introduces intel_digital_port smaller and easier
to review.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Damien Lespiau <damien.lespiau@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

Fix power well control state by reading real register offset.
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

We were writing DSP_ADDR and DSP_SURF unconditionally. This did not
trigger an unclaimed write before HSW as the address of DSP_ADDR has
been repurposed as DSP_LINOFF.

On HSW, though, DSP_LINOFF has been removed and then writting to it
triggers an unclaimed write.

This patch writes to DSP_ADDR or DSP_SURF to flush the display plane
configuration depending on the gen we're running on.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

Comment says for eaglelake/cantiga, but it's listed in the ilk table,
too. So apply it to both.
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

Or at least our best understanding of it.

v2: Fixup commit message and put the wa name into the comment block.
And actually update the commit, too.
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

Just like in:

commit c2c75131
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Thu Jul 5 12:17:30 2012 +0200

    drm/i915: adjust framebuffer base address on gen4+

but this time, for the sprite planes. This ensures that the
sprite offset are always inside the supported hardware limits since it
becomes the offset into a page and we adjust the base address to a page
boundary.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

HSW consolidates SPRTILEOFF and SPRLINOFF into a single SPROFFSET
register.

v2: Remove a useless level of indentation (Paulo Zanoni)
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com> (v1)
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

Haswell consolidates DSP_TILEOFF and DSP_LINOFF into DSP_OFFSET (aka
PRI_OFFSET).
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

v2: Use a switch for consistency (Chris Wilson)
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

Instead of writing to the DSP_ADDR ourselves. This will do the right
thing on gen >= 4 as well.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

And properly toggle the chicken bit in the pch to enable/disable fdi C
rx. If we don't set this bit correctly, the rx gets confused in link
training, which can result in an fdi link that silently fails to train
the link (since the corresponding register reports success). Note that
both fdi link B and C can suffer when this bit is not set correctly.

The code as-is has a few deficiencies:
- We presume all pipes use the pch which is not the case for cpu edp.
- We don't bother with disabling both pipes when we could make things
  work, e.g. when pipe B switched from 4 to 2 lanes due to a mode
  change, we don't bother updating the w/a bit.
- It's ugly.

All of these are because we compute ->fdi_lanes way too late, when
we're already setting up individual pipes. We need to have this
information in ->modeset_global_resources already, to set things up
correctly. But that is a much larger reorg of the code.

Note that we actually hit the 2 lanes limit in practice rather
quickly: Even though the 1920x1200 mode native mode of my screen fits
into 2 lanes, it needs 3 lanes for the 1920x1080 (since that somehow
has much more blanking ...). Not obeying this restriction seems to
results in cute-looking digital noise.

v2: Only ever clear the chicken bit when both pipes are off.

v3: Use the new ->modeset_global_resources callback.

v4: Move the WARNs to the right place. Oh how I hate hacks.

v5: Fix spelling, noticed by Paulo Zanoni.
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

After all relevant pipes are disabled and after we've updated all the
state with the staged state, but before we call the per-crtc
->mode_set functions there's a very natural point to set up any
shared/global resources like
- shared plls (obviously only the setup, the enabling needs to be
  separately handling with a separate refcount)
- global watermark state like the DSPARB on gmch platforms
- workaround bits that depend upon the exact global output
  configuration
- enabling the right set of refclocks
- enabling/disabling manual power wells.

Now for a lot of these things we can't move them into this function
yet, most often because we only compute the required information in
the per-crtc ->mode_set callback. Which is too late. But due to a
bunch of reasons (check-only atomic modeset, fastboot&hw state checks,
...) we need to separate the computation of that state from the actual
hw frobbery anyway. So we can move things into this new callback step-
by-step.

Others can't be moved here (or implemented at all) because our code
lacks the smarts to properly update them. E.g. the DSPARB can only be
updated when all pipes are disabled, so if we decide to change it's
value, we need to disable _all_ pipes. The infrastructure for that is
already in place (with the various pipe masks that driver the modeset
logic). But again we need to move a few things out of ->mode_set
first before we can even implement the correct decision making.

In any case, we need to start somewhere, so let's start with the
callback: Some small follow-up patches will make immediate good use of
it.
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

Since it is one. We need to move this code to encoder specific callbacks
eventually, to kill all that inversion of control ...
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

Atm we have a few funny issues where we enable/disable shared
pll clocks. To make it clear that we are not required to enable/
disable the pch plls together with the other pch resources (and
so should keep it running when it's used by another pipe in
a shared pll configuration) add a comment.

This note is lifted from "Graphics BSpec: vol4g North Display Engine
Registers [IVB], Display Mode Set Sequence", step 9.d. of the enable
sequence:

"Configure and enable PCH DPLL, wait for PCH DPLL warmup (Can be
done anytime before enabling PCH transcoder)."

Since fixing the pll sharing code to no longer disable shared plls
if they're still in use is more involved, let's just stick with the
comment for now.

v2: Make the comment in the code clearer, to address questions raised
by Paulo Zanoni in review.
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

My machine here has the correct ones already, but better safe
than sorry. IBX has different settings for that register, and
on IBX the device defaults match the recommended values. Hence
I did not add the respective writes for IBX.

LPT needs the same settings, but that has been done already

commit 4acf5186
Author: Eugeni Dodonov <eugeni.dodonov@intel.com>
Date:   Wed Jul 4 20:15:16 2012 -0300

    drm/i915: program FDI_RX TP and FDI delays

Cc: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

For reference, see "Graphics BSpec: vol4g North Display Engine
Registers [IVB], Display Mode Set Sequence", step 4 of the enabling
sequence:

a. "Enable PCH FDI Receiver PLL, wait for warmup plus DMI latency
b. "Switch from Rawclk to PCDclk in FDI Receiver
c. "Enable CPU FDI Transmitter PLL, wait for warmup"

Cc: Paulo Zanoni <przanoni@gmail.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

According to "Graphics BSpec: vol4g North Display Engine Registers [IVB],
Display Mode Set Sequence" We need to write the TU size register
of the fdi RX unit _before_ starting to train the link.

Note: The current code is actually correct as Paulo mentioned in
review, but it's a bit confusion since only the fdi rx/tx plls need to
be enabled before the cpu pipes/planes. Hence it's still a good idea
to move the TU_SIZE setting to the "right" spot in the sequence, to
better match Bspec.
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

The debug message is only relevant on CPT/PPT PCH ports, so move
it into the correct if clause.
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

Daniel's backmerge

commit c2fb7916
Merge: 29de6ce5 6f0c0580
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Mon Oct 22 14:34:51 2012 +0200

    Merge tag 'v3.7-rc2' into drm-intel-next-queued

to solve conflicts blew up (either git or Daniel was trying to be too
clever for their own good; it's usually convenient to blame tools ;) and
caused the changes of

commit 0c96c65b
Author: Jani Nikula <jani.nikula@intel.com>
Date:   Wed Sep 26 18:43:10 2012 +0300

    drm/i915: use adjusted_mode instead of mode for checking the 6bpc force flag

in ironlake_crtc_mode_set() to be dropped.

Fix the call in ironlake_crtc_mode_set() again, and while at it, also fix
the new, copy-pasted haswell_crtc_mode_set() to use adjusted_mode.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

Signed-off-by: Yuly Novikov <ynovikov@chromium.org>
[Jani: ripped this change separate from the scaling mode change support]
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Tested-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

LVDS allowed changing panel fitting scaling mode, while eDP didn't. Copied
relevant code from LVDS to eDP.
Signed-off-by: Yuly Novikov <ynovikov@chromium.org>
[Jani: use fitting mode in intel_panel, remove default mode change]
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Tested-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

Prepare for supporting scaling mode configuration also in eDP.

Includes a drive-by-removal of an outdated comment about fitting mode.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Tested-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

At some point the DPCD size was increased, but the debug print not. While
at it, switch to using hex dump.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

Just like HSW, VLV does not have a sprite scale. Set
intel_plane->can_scale accordingly.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

SDVOB may be multiplexed with HDMIB. If it's not SDVOB, the same i2c
adapter may be used for HDMIB, with the adjusted config (i.e. with GPIO
bit-banging instead of gmbus). Restore i2c adapter config before error
return from intel_sdvo_init(), letting HDMIB enjoy the joys of gmbus.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

commit 63abf3ed
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Wed Dec 8 16:48:21 2010 +0000

    drm/i915/sdvo: Only use the SDVO pin if it is in the valid range

added a default fallback if BIOS provides an invalid pin mapping, but
failed to force GPIO bit-banging on it. Finish the job, and also clean up
the function a bit. With bit-banging, setting the gmbus speed has no
effect, so drop it.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
[danvet: Extend comment about gmbus in the code a bit.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>