Request by Ben Widawsky in his review of a patch touching this code.

v2: Clarify the disdinction between evicting vmas (to free up virtual
address space) and evicting objects (to free up actual system memory).
Suggested by Ben.

Cc: Ben Widawsky <benjamin.widawsky@intel.com>
Acked-by: NBen Widawsky <benjamin.widawsky@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

During eviction, we are only considering how to free up space within the
current address space and not concerned with freeing up physical memory.
As such we need only skip nodes that pinned in the current VM and not
globally.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Ben Widawsky <benjamin.widawsky@intel.com>
Reviewed-by: NBen Widawsky <ben@bwidawsk.net>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

The initial jiffies value can be non-0, so set the inital panel power
sequencer timestamps accordingly. This didn't cause a problem on 64 bit
machines but on 32 bit jiffies is initially -300*HZ, so if the panel
power is initally off in the call from edp_panel_vdd_on()->
wait_panel_power_cycle() we'd wait up to ~300 sec more than needed.
Signed-off-by: NImre Deak <imre.deak@intel.com>
Reviewed-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

schedule_timeout_uninterruptible() takes jiffies not ms.

v2:
- ignore the overflow issue, the practical part of that should
  be solved instead in the caller (Chris)

Note that this issue was introduced in

commit dce56b3c
Author: Paulo Zanoni <paulo.r.zanoni@intel.com>
Date:   Thu Dec 19 14:29:40 2013 -0200

drm/i915: save some time when waiting the eDP timings

I've accidentally merged the broken v4 version of the patch (where
Jani noticed the issue [1]) instead of the v5, which was fixed [2].

[1] http://mid.gmane.org/87fvpnkgyg.fsf@intel.com
[2] http://mid.gmane.org/1388778311-2020-1-git-send-email-przanoni@gmail.comSigned-off-by: NImre Deak <imre.deak@intel.com>
Reviewed-by: NChris Wilson <chris@chris-wilson.co.uk>
[danvet: Add admission of incompetence in the form of a note.]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

When current delay is already at max delay, Let's disable the PM UP
THRESHOLD INTRRUPTS, so that we will not get further interrupts until
current delay is less than max delay, Also request for the PM DOWN
THRESHOLD INTRRUPTS to indicate the decrease in clock freq. and
viceversa for PM DOWN THRESHOLD INTRRUPTS.

v2: Use bool variables (Daniel)

v3: Fix Interrupt masking bit (Deepak)

v4: Use existing symbolic constants in i915_reg.h (Daniel)

v5: Add pm interrupt mask after new_delay calculation (Ville)
Signed-off-by: NDeepak S <deepak.s@intel.com>
[danvet: Pass new_delay by value as suggested by Ville. Also appease
checkpatch.]
Reviewed-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

We're disabling a boatload of clock gating features on VLV. Maybe these
days we don't need to do that. At least I'm not aware of any workarounds
with this level of paranoia.

This reverts commit 4e8c84a5.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NRodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

WaDisable4x2SubspanOptimization isn't listed for VLV in the workaround
database, but BSpec says that the relevant bit must be set. Add a
comment to remind people of this.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NRodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Call gen7_setup_fixed_func_scheduler() on VLV as well.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NRodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

BSpec states that the thread override values set by
gen7_setup_fixed_func_scheduler() are invalid for HSW. So let's not
muck around with them.

Since gen7_setup_fixed_func_scheduler() now has two totally independent
parts, one for IVB and one for HSW, move the HSW part directly into
haswell_init_clock_gating().

Note tht there's another workaround by the name of
WaHSWVSRefCountFullforceMissDisable which basically claims that later
steppings don't need the fix, but since WaVSRefCountFullforceMissDisable
is listed to be needed for all steppings play it safe and keep applying
the workaround.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NRodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

The current comments indicate that this function implements
WaVSRefCountFullforceMissDisable, which is only true for HSW.
The original purpose of the function is to implement
WaVSThreadDispatchOverride (and a bit more). Fix up the comments
to match reality.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NRodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

WaDisableTDLUnitClockGating is only relevant for early steppings of VLV.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NRodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

WaDisableVDSUtnitClockGating was only relevant for early steepings of
VLV.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NRodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Add a few new debugfs files which allow changing the watermark memory
latency values during runtime. This can be used to determine the if the
original BIOS provided latency values are no good.

v2: Drop superfluous plane name from output
    Take modeset locks around the latency value read/write
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NDamien Lespiau <damien.lespiau@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Only early VLV steppings needed thist. Should no longer be relevant.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NRodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

WaApplyL3ControlAndL3ChickenMode is only relevant to early HSW
steppings..
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NRodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

WaDisableRCZUnitClockGating was needed with early HSW steppings only.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NRodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Someone copy pasted the comment from the SNB code w/o reading it.
We never actually implemented the workaround to disable RCPB unit
clock gating on IVB. It would have been needed for early steppings,
but we don't care about those anymore, so just remove the stale
comment.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NRodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Can't find WaDisableRHWOOptimizationForRenderHang listed for VLV.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NRodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Can't find WaDisableRHWOOptimizationForRenderHang listed for HSW.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NRodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

WaDisableRCCUnitClockGating is only relevant for SNB.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NRodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

WaDisableRCCUnitClockGating is only relevant for SNB.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NRodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Can't find any mention of WaDisableVDSUnitClockGating ever being
relevant for SNB. Remove it.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NRodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NRodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

There are cases where we want to know if there is a full, or aliased
PPGTT. Currently, in fact the only distinction we ever need to make is
when we're using full PPGTT.

This patch is simply to promote readability and clarify for the
confusing existing usage where "aliasing" meant aliasing and full.

v2: Remove USES_ALIASING_PPGTT since there are currently no cases where
we need to check if we're using aliasing, but not full PPGTT. (Daniel)

Cc: Daniel Vetter <daniel@ffwll.ch>
Signed-off-by: NBen Widawsky <ben@bwidawsk.net>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Having to use i915.i915_foo is inconsistent and a bit on the verbose
side. Drop the prefix per Daniel's request, who also says this is not
ABI we need to maintain.
Signed-off-by: NJani Nikula <jani.nikula@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

At least I couldn't find it in the Haswell Bspec any more and we've
tried to test-boot a Haswell machine with num_pipes forced to 0 (i.e.
hit the PCH_NOP path) and the unclaimed register logic complained.

So restrict this dance to just ivb platforms.

v2: Art pointed out that the bits simply moved on hsw+

v3: Buy code terseneness with a notch of sublety as suggested by
Chris.

v4: Frob the right bit, spotted by Art.

Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Arthur Ranyan <arthur.j.runyan@intel.com>
Cc: Dave Airlie <airlied@gmail.com>
Reviewed-by: NArt Runyan <arthur.j.runyan@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

A lot of the WM functions are only reading from that structure and are
already using const. While converting the code to use dev_priv instead
of dev, I noticed a few places where we can give that hint.
Signed-off-by: NDamien Lespiau <damien.lespiau@intel.com>
Reviewed-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

With 20+ module parameters, I think referring to them via a struct
improves clarity over just having a bunch of globals. While at it, move
the parameter initialization and definitions into a new file
i915_params.c to reduce clutter in i915_drv.c.

Apart from the ill-named i915_enable_rc6, i915_enable_fbc and
i915_enable_ppgtt parameters, for which we lose the "i915_" prefix
internally, the module parameters now look the same both on the kernel
command line and in code. For example, "i915.modeset".

The downsides of the change are losing static on a couple of variables
and not having the initialization and module_param_named() right next to
each other. On the other hand, all module parameters are now defined in
one place at i915_params.c. Plus you can do this to find all module
parameter references:

$ git grep "i915\." -- drivers/gpu/drm/i915

v2:
- move the definitions into a new file
- s/i915_params/i915/
- make i915_try_reset i915.reset, for consistency
Signed-off-by: NJani Nikula <jani.nikula@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

WaMiSetContext_Hang tells us that a MI_NOOP must follow MI_SET_CONTEXT.

The other thing WaMiSetContext_Hang seems to say is that URB_FENCE isn't
allowed to straddle two cachelines. But we don't issue those from the
kernel so we don't care.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NRodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NRodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

WaApplyL3ControlAndL3ChickenMode is only listed for IVB and HSW in
W/A database and BSpec.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NRodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NRodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

The w/a database lists both WaPsdDispatchEnable and
WaDisablePSDDualDispatchEnable for VLV. They appear to be the same
thing, so list both names.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NRodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NRodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NRodrigo Vivi <rodrigo.vivi@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Many times in the past we have concluded that the cause of the GPU hang
has been that the hw status page was stale, usually because the GPU and
CPU disagreed over the address of the page. Having stumbled across yet
another issue that seems to be related to the HWSP, it is time to
include that information in the GPU error dump.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Through a twisty and circuituous path it is possible to currently trick
the code into creating a default context and forgetting to pin it
immediately into the GGTT. (This requires a system using contexts without
an aliasing ppgtt, which is currently restricted to Baytrails machines
manually specifying a module parameter to force enable contexts, or
on Sandybridge and later that manually disable the aliasing ppgtt.) The
consequence is that during module unload we attempt to unpin the default
context twice and encounter a BUG remonstrating that we attempt to unpin
an unbound object.

[  161.002869] Kernel BUG at f84861f8 [verbose debug info unavailable]
[  161.002875] invalid opcode: 0000 [#1] SMP
[  161.002882] Modules linked in: coretemp kvm_intel kvm crc32_pclmul aesni_intel aes_i586 xts lrw gf128mul ablk_helper cryptd hid_sensor_accel_3d hid_sensor_gyro_3d hid_sensor_magn_3d hid_sensor_trigger industrialio_triggered_buffer kfifo_buf industrialio hid_sensor_iio_common snd_hda_codec_hdmi snd_hda_codec_realtek snd_hda_intel snd_hda_codec snd_hwdep snd_pcm snd_page_alloc snd_seq_midi snd_seq_midi_event dm_multipath scsi_dh asix ppdev usbnet snd_rawmidi mii hid_sensor_hub microcode snd_seq rfcomm bnep snd_seq_device bluetooth snd_timer snd parport_pc binfmt_misc soundcore dw_dmac_pci dw_dmac_core mac_hid lp parport dm_mirror dm_region_hash dm_log hid_generic usbhid hid i915(O-) drm_kms_helper(O) igb dca ptp pps_core i2c_algo_bit drm(O) ahci libahci video
[  161.002991] CPU: 0 PID: 2114 Comm: rmmod Tainted: G        W  O 3.13.0-rc8+ #2
[  161.002997] Hardware name: NEXCOM VTC1010/Aptio CRB, BIOS 5.6.5 09/24/2013
[  161.003004] task: dbdd6800 ti: dbe0e000 task.ti: dbe0e000
[  161.003010] EIP: 0060:[<f84861f8>] EFLAGS: 00010246 CPU: 0
[  161.003044] EIP is at i915_gem_object_ggtt_unpin+0x88/0x90 [i915]
[  161.003050] EAX: dfce3840 EBX: 00000000 ECX: dfafd690 EDX: dfce3874
[  161.003056] ESI: c0086b40 EDI: df962e00 EBP: dbe0fe1c ESP: dbe0fe0c
[  161.003062]  DS: 007b ES: 007b FS: 00d8 GS: 00e0 SS: 0068
[  161.003068] CR0: 8005003b CR2: b7718000 CR3: 1bec0000 CR4: 001007f0
[  161.003076] Stack:
[  161.003081]  00afc014 00000004 c0086b40 dfafc000 dbe0fe38 f8487e5a dfaa5400 c0086b40
[  161.003099]  dfafc000 dfaa5400 dfaa5414 dbe0fe58 f84741aa 00000000 f89c34b9 dfaa5414
[  161.003117]  dfaa5400 dfaa5400 f644b000 dbe0fe6c f89a5443 dfaa5400 f8505000 f644b000
[  161.003134] Call Trace:
[  161.003169]  [<f8487e5a>] i915_gem_context_fini+0xba/0x1c0 [i915]
[  161.003202]  [<f84741aa>] i915_driver_unload+0x1fa/0x2f0 [i915]
[  161.003232]  [<f89a5443>] drm_dev_unregister+0x23/0x90 [drm]
[  161.003259]  [<f89a54ed>] drm_put_dev+0x3d/0x70 [drm]
[  161.003294]  [<f8470615>] i915_pci_remove+0x15/0x20 [i915]
[  161.003306]  [<c1338a6f>] pci_device_remove+0x2f/0xa0
[  161.003317]  [<c140c871>] __device_release_driver+0x61/0xc0
[  161.003328]  [<c140d12f>] driver_detach+0x8f/0xa0
[  161.003341]  [<c140c54f>] bus_remove_driver+0x4f/0xc0
[  161.003353]  [<c140d708>] driver_unregister+0x28/0x60
[  161.003362]  [<c10cee42>] ? stop_cpus+0x32/0x40
[  161.003372]  [<c10bd510>] ? module_refcount+0x90/0x90
[  161.003383]  [<c13378c5>] pci_unregister_driver+0x15/0x60
[  161.003413]  [<f89a739f>] drm_pci_exit+0x9f/0xb0 [drm]
[  161.003458]  [<f84e624a>] i915_exit+0x1b/0x1d [i915]
[  161.003468]  [<c10bf8a8>] SyS_delete_module+0x158/0x1f0
[  161.003480]  [<c1173d5d>] ? ____fput+0xd/0x10
[  161.003488]  [<c106f0fe>] ? task_work_run+0x7e/0xb0
[  161.003499]  [<c165a68d>] sysenter_do_call+0x12/0x28
[  161.003505] Code: 0f b6 4d f3 8d 51 0f 83 e1 f0 83 e2 0f 09 d1 84 d2 88 48 54 75 07 80 a7 91 00 00 00 7f 83 c4 04 5b 5e 5f 5d c3 8d b6 00 00 00 00 <0f> 0b 8d b6 00 00 00 00 55 89 e5 57 56 53 83 ec 64 3e 8d 74 26
[  161.003586] EIP: [<f84861f8>] i915_gem_object_ggtt_unpin+0x88/0x90 [i915] SS:ESP 0068:dbe0fe0c

v2: Rename the local variable (is_default_ctx) to avoid confusion with
the function is_default_ctx(). And correct Jesse's email address.
Reported-by: NJesse Barnes <jbarnes@virtuousgeek.org>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=73985Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Cc: Ben Widawsky <benjamin.widawsky@intel.com>
Tested-by: NJesse Barnes <jbarnes@virtuousgeek.org>
Reviewed-by: NBen Widawsky <benjamin.widawsky@intel.com>
[danvet: Fix up the rebase fail from my first attempt, thankfully
pointed out by Ville.]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

This debugfs interface will allow intel-gpu-tools test case
to verify if screen has been updated properly on cases like PSR.

v2: Accepted all Daniel's suggestions:
    * grab modeset lock
    * loop over connector and check DPMS on
    * return errors
    * use _eDP1 suffix for easy future extension
    * don't cache crc_supported neither latest crc
    * return crc as a full array and read it at once with aux.
    * use 0 to turn TEST_SINK off.
    * split the drm_helpers definitions in another patch.

v3: Accepted 2 Damien's suggestion: remove h from printf hexa
    and return ENODEV when eDP not present instead of EAGAIN.

v4: Accepted 2 Jani' s suggestion: 1 path for unlock and remove
    _retry from aux read.

v5: removing last missing useless _retry (by Damien)

Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Damien Lespiau <damien.lespiau@intel.com>
Cc: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: NRodrigo Vivi <rodrigo.vivi@gmail.com>
Reviewed-by: NDamien Lespiau <damien.lespiau@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

This address will be used to verify panel CRC for test and
validation purposes.
Signed-off-by: NRodrigo Vivi <rodrigo.vivi@gmail.com>
[danvet: Fix whitespace fail.]
Acked-by: NDave Airlie <airlied@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Having a 4 byte register at 0x321b seems unlikely as that's not
4 byte aligned. Since later platforms have more or less the same FBC
registers with new names, assume that FBC_FENCE_OFF is at 0x3218 just
like DPFC_FENCE_YOFF.

This feels like a simple typo in BSpec. 321Bh looks a lot like 3218h
after all.

Should still be tested on real hardware of course. But I don't have
any mobile gen4 systems.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Acked-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>