- fbc improvements when stolen memory is tight (Ben)
- cdclk handling improvements for vlv/chv (Ville)
- proper fix for stuck primary planes on gmch platforms with cxsr (Imre&Ebgert
Eich)
- gen8 hw semaphore support (Ben)
- more execlist prep work from Oscar Mateo
- locking fixes for primary planes (Matt Roper)
- code rework to support runtime pm for dpms on hsw/bdw (Paulo, Imre & me), but
not yet enabled because some fixes from Paulo haven't made the cut
- more gpu boost tuning from Chris
- as usual piles of little things all over
* tag 'drm-intel-next-2014-07-11' of git://anongit.freedesktop.org/drm-intel: (93 commits)
drm/i915: Make the RPS interrupt generation mask handle the vlv wa
drm/i915: Move RPS evaluation interval counters to i915->rps
drm/i915: Don't cast a pointer to void* unnecessarily
drm/i915: don't read LVDS regs at compute_config time
drm/i915: check the power domains in intel_lvds_get_hw_state()
drm/i915: check the power domains in ironlake_get_pipe_config()
drm/i915: don't skip shared DPLL assertion on LPT
drm/i915: Only touch WRPLL hw state in enable/disable hooks
drm/i915: Switch to common shared dpll framework for WRPLLs
drm/i915: ->enable hook for WRPLLs
drm/i915: ->disable hook for WRPLLs
drm/i915: State readout support for WRPLLs
drm/i915: add POWER_DOMAIN_PLLS
drm/i915: Document that the pll->mode_set hook is optional
drm/i915: Basic shared dpll support for WRPLLs
drm/i915: Precompute static ddi_pll_sel values in encoders
drm/i915: BDW also has special-purpose DP DDI clocks
drm/i915: State readout and cross-checking for ddi_pll_sel
drm/i915: Move ddi_pll_sel into the pipe config
drm/i915: Add a debugfs file for the shared dpll state
...
Just like we already do in haswell_get_pipe_config(). This should
prevent some WARNs when we run pm_rpm on SNB.
Testcase: igt/pm_rpm
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=80463
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>
Since we now have support for shared DPLLS.
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>
To be able to do this we need to separately keep track of how many
crtcs need a given WRPLL and how many actually actively use it. The
common shared dpll framework already has all this, including massive
state readout and cross checking. Which allows us to do this switch in
a fairly small patch.
Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Mostly this patch is one big excersize in deleting code and asserts
which are no longer needed. Note that we still abuse the shared dpll
framework a bit since we call the enable/disable functions from the
crtc mode_set and off hooks. But changing the actual hardware sequence
will be done in the next step.
Note that besides the massive amount of changes in this patch the
places and order in which the low-level WRPLL code is called is
absolutely unchanged.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
[imre: rebased on patchset version w/o pch/crt/fdi refactoring]
Signed-off-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Currently still with a redudant WARN_ON in there, the common shared
dpll code will take care of this in the future.
Also we need to flip the switch for the transitional hack now to make
sure that we disable the right pll.
Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Still tacked onto the side, but slowly getting there.
v2: Don't forget the debugfs file.
v3 (from Paulo): Don't forget to check the power domains.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
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>
And get/put it when needed. The special thing about this commit is
that it will now return false in ibx_pch_dpll_get_hw_state() in case
the power domain is not enabled. This will fix some WARNs we have when
we run pm_rpm on SNB.
Testcase: igt/pm_rpm
Bugzilla:https://bugs.freedesktop.org/show_bug.cgi?id=80463
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>
Just filing in names and ids, but not yet officially registering them
so that the hw state cross checker doesn't completely freak out about
them. Still since we do already read out and cross check
config->shared_dpll the basics are now there to flesh out the wrpll
shared dpll implementation.
The idea is now to roll out all the callbacks step-by-step and then at
the end switch to the shared dpll framework. This way hw and sw
changes are clearly separated.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
[imre: added const to hsw_ddi_pll_names (Damien)]
Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
To make things a bit more manageable extract a new function for
reading out common ddi port state. This means a bit of duplication
between encoders and the core since both look at the same registers,
but doesn't seem worth to make a fuzz about.
We can also remove the state readout code in intel_ddi_setup_hw_pll_state.
That code is only called from the hardware take over and not the cross
check code, and only after the crtc state is reconstructed. So we can
rely on an accurate value of crtc->config.ddi_pll_sel already.
Compared to the old code also trust the hw state more and don't
special-case port A - we want to cross-check the actual-state, not
bake in our own assumptions about how this is supposed to all be
linked up.
v2: Make use of the read-out ddi_pll_sel in intel_ddi_clock_get.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
[imre: rebased on patchset version w/o pch/crt/fdi refactoring]
Signed-off-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
This is needed by an upcoming patch that moves the PCH/CRT PLL disabling
into the post_disable hook, after which we want to keep the modeset
sequence at its current state. At this point this won't have an effect
since the PCH/CRT post_disable hook is atm a NOP.
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
This is needed by an upcoming patch that moves the PCH/CRT PLL enabling
into the pre_enable hook, after which we want to keep the modeset
sequence at its current state. At this point this won't have an effect
since the PCH/CRT pre_enable hook is atm a NOP.
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
All the other checks also check hw state, so checking our software
refcounts for the plls looks a bit odd. Also this will simplify the
conversion over to the shared dpll framework, which itself has massive
amounts of checks to make sure that we never leave a display pll
enabled when we shouldn't.
So after that conversion we should stil have a good enough coverage of
asserts for entering pc8/runtime pm on hsw/bdw.
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Add !mutex_is_locked() checks to intel_pin_and_fence_fb_obj() and
intel_unpin_fb_obj() to help catch failures to grab struct_mutex when
operating on fb objects.
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
intel_primary_plane_{setplane,disable} were lacking struct_mutex locking
around their GEM operations.
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reported-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
On HSW, the D_COMP register can be accessed through the mailbox (read
and write) or through MMIO on a MCHBAR offset (read only). On BDW, the
access should be done through MMIO on another address. So to account
for all these cases, create hsw_read_dcomp() with the correct
implementation for reading, and also fix hsw_write_dcomp() to do the
correct thing on BDW.
With this patch, we can now get back from the PC8+ state on BDW. We
were previously getting a black screen and lots of dmesg errors.
Please notice that the bug only happens when you actually reach the
PC8+ states, not when you only allow it.
Testcase: igt/pm_rpm/rte
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>
That function can be used to write anything on D_COMP, not just
disable it, so print a more appropriate message.
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>
This should hopefully simplify the display code slightly and also
solves at least one mistake in intel_pipe_set_base() where
to_intel_framebuffer(fb)->obj is referenced during local variable
initialization, before 'if (!fb)' gets checked.
Potential uses of this macro were identified via the following
Coccinelle patch:
@@
expression E;
@@
* to_intel_framebuffer(E)->obj
@@
expression E;
identifier I;
@@
I = to_intel_framebuffer(E);
...
* I->obj
v2: Rewrite some NULL tests in terms of the obj rather than the fb.
Also add a WARN() if trying to pageflip with a disabled primary
plane. [Suggested by Chris Wilson]
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
- Accurate frontbuffer tracking and frontbuffer rendering invalidate, flush and
flip events. This is prep work for proper PSR support and should also be
useful for DRRS&fbc.
- Runtime suspend hardware on system suspend to support the new SOix sleep
states, from Jesse.
- PSR updates for broadwell (Rodrigo)
- Universal plane support for cursors (Matt Roper), including core drm patches.
- Prefault gtt mappings (Chris)
- baytrail write-enable pte bit support (Akash Goel)
- mmio based flips (Sourab Gupta) instead of blitter ring flips
- interrupt handling race fixes (Oscar Mateo)
And old, not yet merged features from the previous round:
- rps/turbo support for chv (Deepak)
- some other straggling chv patches (Ville)
- proper universal plane conversion for the primary plane (Matt Roper)
- ppgtt on vlv from Jesse
- pile of cleanups, little fixes for insane corner cases and improved debug
support all over
* tag 'drm-intel-next-2014-06-20' of git://anongit.freedesktop.org/drm-intel: (99 commits)
drm/i915: Update DRIVER_DATE to 20140620
drivers/i915: Fix unnoticed failure of init_ring_common()
drm/i915: Track frontbuffer invalidation/flushing
drm/i915: Use new frontbuffer bits to increase pll clock
drm/i915: don't take runtime PM reference around freeze/thaw
drm/i915: use runtime irq suspend/resume in freeze/thaw
drm/i915: Properly track domain of the fbcon fb
drm/i915: Print obj->frontbuffer_bits in debugfs output
drm/i915: Introduce accurate frontbuffer tracking
drm/i915: Drop schedule_back from psr_exit
drm/i915: Ditch intel_edp_psr_update
drm/i915: Drop unecessary complexity from psr_inactivate
drm/i915: Remove ctx->last_ring
drm/i915/chv: Ack interrupts before handling them (CHV)
drm/i915/bdw: Ack interrupts before handling them (GEN8)
drm/i915/vlv: Ack interrupts before handling them (VLV)
drm/i915: Ack interrupts before handling them (GEN5 - GEN7)
drm/i915: Don't BUG_ON in i915_gem_obj_offset
drm/i915: Grab dev->struct_mutex in i915_gem_pageflip_info
drm/i915: Add some L3 registers to the parser whitelist
...
Conflicts:
drivers/gpu/drm/i915/i915_drv.c
For whatever reason, MI_DISPLAY_FLIP fails to change tiling mode on
Baytrail, so just use CPU driven mmio flips instead.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=76176
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
We currently see random GPU hangs when using RCS flips with multiple
pipes on Ivybridge. Now that we have mmio flips, we can fairly cheaply
fallback to using CPU driven flips instead.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=77104
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
misc core patches picked up by Daniel and Jani.
* tag 'topic/core-stuff-2014-06-30' of git://anongit.freedesktop.org/drm-intel:
drm/fb-helper: Remove unnecessary list empty check in drm_fb_helper_debug_enter()
drm/fb-helper: Redundant info->fix.type_aux setting in drm_fb_helper_fill_fix()
drm/debugfs: add an "edid_override" file per connector
drm/debugfs: add a "force" file per connector
drm: add register and unregister functions for connectors
drm: fix uninitialized acquire_ctx fields (v2)
drm: Driver-specific ioctls range from 0x40 to 0x9f
drm: Don't export internal module variables
The always-on power well pixel path on haswell is routed such that it
bypasses the panel fitter when we use is. Which means the pfit CRC
source won't work in that configuration.
Add a new disallow-bypass flags to the pfit pipe config state and set
it when we want to use the pf CRC. Results in a bit of flicker, but
should get the job done. We'll also undo do it afterwards to make sure
other tests arent' negatively affected.
Totally untested due to lack of hsw laptops around here.
v2: s/disallow_bypass/force_power_well_on/ to avoid a double negative
(Damien).
v3: force_thru because roadsigns.
v4: Don't forget the power wells! Also note that until the runtime pm
for DPMS series is fully merged the simple disable/enable trick won't
work since the ->crtc_mode_set callback is still required to do nasty
things. This stuff is tricky, but I think by both fixing up
get_crtc_power_domains and the debugfs wa code we should always
grab/drop the additional power well correctly.
v5: Wrap in () as suggested by Damien to avoid setting reserved values
for the edp transcoder path on bdw+
References: https://bugs.freedesktop.org/show_bug.cgi?id=72864
Cc: Damien Lespiau <damien.lespiau@intel.com>
Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
Tested-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Since the root cause is understood now and with the fix
commit 564ed191f5
Author: Imre Deak <imre.deak@intel.com>
Date: Fri Jun 13 14:54:21 2014 +0300
drm/i915: gmch: fix stuck primary plane due to memory self-refresh mode
in place the magic for G4x chipsets introduced with commit
commit 61bc95c1fb
Author: Egbert Eich <eich@suse.com>
Date: Mon Mar 4 09:24:38 2013 -0500
DRM/i915: On G45 enable cursor plane briefly after enabling the display plane.
to avoided occasional screen blanking on mode changes can finally
be removed.
It's been verified that Imre's fix also resolves the said issue.
Signed-off-by: Egbert Eich <eich@suse.de>
Tested-by: Stefan Dirsch <sndirsch@suse.de>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Blanking/unblanking the console in a loop on an Asus T100 sometimes
leaves the console blank. After some digging I found that applying
commit 61bc95c1fb
Author: Egbert Eich <eich@suse.com>
Date: Mon Mar 4 09:24:38 2013 -0500
DRM/i915: On G45 enable cursor plane briefly after enabling the display plane.
fixed VLV too.
In my case the problem seemed to happen already during the previous crtc
disabling and went away if I disabled self-refresh mode before disabling
the primary plane.
The root cause for this is that updates from the shadow to live plane
control register are blocked at vblank time if the memory self-refresh
mode (aka max-fifo mode on VLV) is active at that moment. The controller
checks at frame start time if the CPU is in C0 and the self-refresh mode
enable bit is set and if so activates self-reresh mode, otherwise
deactivates it. So to make sure that the plane truly gets disabled before
pipe-off we have to:
1. disable memory self-refresh mode
2. disable plane
3. wait for vblank
4. disable pipe
5. wait for pipe-off
v2:
- add explanation for the root cause from HW team (Cesar Mancini et al)
- remove note about the CPU C7S state, in my latest tests disabling it
alone didn't make a difference
- add vblank between disabling plane and pipe (Ville)
- apply the same workaround for all gmch platforms (Ville)
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Vijay Purushothaman <vijay.a.purushothaman@intel.com>
Reviewed-by: Deepak S<deepak.s@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Now that the CMNRESET deassert is part of the cmnlane power well,
intel_reset_dpio() is called too late to make any difference. We've
deasserted CMNRESET by that time, and so the off+on toggle w/a will
never kick in.
Move the workaround to intel_power_domains_init_hw() where it gets
called before we enable the init power domain.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
We have a slightly different way of readoing out the cdclk in
gmbus_set_freq(). Kill that and just call .get_display_clock_speed().
Also need to remove the GMBUSFREQ update from intel_i2c_reset() since
that gets called way too early. Let's do it in intel_modeset_init_hw()
instead, and also pull the initial vlv_cdclk_freq update there from
init_clock gating.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
If someone is interested in the current cdclk frquency it should
be stable and not in process of changing frquency. Warn if the current
and requested cdclk don't match in .get_display_clock_spee() on vlv.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
VLV Punit doesn't support the 400MHz cdclk option, so we bypass the
Punit and poke at CCK directly. However we forgot to wait for the
frequeency change to complete. Poll the CCK clock status to make sure
the clock has changed before we fire up any pipes.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Drop the cdclk frequency to 200MHz on vlv when all pipes are off. In
theory we should be able to use 200MHz also when the pixel clock is at
most 90% of 200MHz. However in practice all we seem to get is a solid
color picture or an otherwise corrupted display.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Depending on the HPLL frequency one of the supported cdclk frquencies is
either 320MHz or 333MHz. Figure out which one it is to accurately pick
the minimal required cdclk. This would also avoid a warning from the
cdclk code where it compares the actual cdclk read out from the hardware
with a value that was calculated using valleyview_calc_cdclk().
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
We have a standard hook for reading out the current cdclk. Move the VLV
code from valleyview_cur_cdclk() to .get_display_clock_speed().
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Avoid using magic values for CCK frequency bits. Also the mask we were
using for the requested frequency was one bit too short. Fix it up.
Note: This also fixes the #define for a mask (spotted by Jesse in his
review).
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
[danvet: Add note about mask change.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Use kHz units in vlv cdclk code since that's more customary.
Also replace the precomputed 90% values with *9/10 computation
for extra clarity.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iQEcBAABAgAGBQJTuaWZAAoJEHm+PkMAQRiGfkIH/2Hhwrg51GWazUYIXVxz5zLU
kPMlaws3vankbhka9HCg02eS3tkzr6shO3F/qlBba+5GUkUDKCcCisIsvk4hgZZg
7YqepTvcaupNxIp4TmTGm1FYVK1GpaWFdJVgg2PDdGFahw3HSlfZoTkBzirNCwga
p/jfeRzathbUixpz9OAC1AEn2gP1AxNRpSt1wShL5rexBb1YRXCPuCEt9B0UsVoR
mzKf5xEsuaZnpCuvWK4S60fjfVhTe8UJ/xGPPfdLyIXU0rvhaKzfeVQO6F5nIQBy
Xvrar1f7oOPZaJRdlmPvAimS7iS8lq/YctuHu7ia1NdJSihtA5sRPf7cWAw2d7s=
=4PrL
-----END PGP SIGNATURE-----
Merge tag 'v3.16-rc4' into drm-intel-next-queued
Due to Dave's vacation drm-next hasn't opened yet for 3.17 so I
couldn't move my drm-intel-next queue forward yet like I usually do.
Just pull in the latest upstream -rc to unblock patch merging - I
don't want to needlessly rebase my current patch pile really and void
all the testing we've done already.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
use mm.h definition
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: intel-gfx@lists.freedesktop.org
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Fabian Frederick <fabf@skynet.be>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Apparently we can't trust this field on other platforms and need to find
some other way.
This fixes a regression introduced in
commit 27da3bdfcf
Author: Jesse Barnes <jbarnes@virtuousgeek.org>
Date: Fri Apr 4 16:12:07 2014 -0700
drm/i915: use VBT to determine whether to enumerate the VGA port
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Tested-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
BDW signals the flip done interrupt immediately after the DSPSURF write
when the plane is disabled. This is true even if we've already armed
DSPCNTR to enable the plane at the next vblank. This causes major
problems for our page flip code which relies on the flip done interrupts
happening at vblank time.
So what happens is that we enable the plane, and immediately allow
userspace to submit a page flip. If the plane is still in the process
of being enabled when the page flip is issued, the flip done gets
signalled immediately. Our DSPSURFLIVE check catches this to prevent
premature flip completion, but it also means that we don't get a flip
done interrupt when the plane actually gets enabled, and so the page
flip is never completed.
Work around this by re-introducing blocking vblank waits on BDW
whenever we enable the primary plane.
I removed some of the vblank waits here:
commit 6304cd91e7
Author: Ville Syrjälä <ville.syrjala@linux.intel.com>
Date: Fri Apr 25 13:30:12 2014 +0300
drm/i915: Drop the excessive vblank waits from modeset codepaths
To avoid these blocking vblank waits we should start using the vblank
interrupt instead of the flip done interrupt to complete page flips.
But that's material for another patch.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=79354
Tested-by: Guo Jinxian <jinxianx.guo@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
crtc->primary->fb may be NULL upon entry to intel_pipe_set_base() if the
primary plane has previously been disabled via the universal plane
interface. We need to check for NULL before trying to reference
old_fb's obj.
This fixes a regression introduced in
commit a071fa0064
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Wed Jun 18 23:28:09 2014 +0200
drm/i915: Introduce accurate frontbuffer tracking
Testcase: igt/kms_universal_plane
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
For MIPI, DSI PLL is configured separately in vlv_configure_dsi_pll
during the DSI enable sequence
Causing WARN dump otherwise in dpio_reads
v2: Add IS_CHERRYVIEW check as suggested by Ville
Signed-off-by: Shobhit Kumar <shobhit.kumar@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Jesse noticed that the punit communication needed to query the VLV power
well status can cause substantial delays. Since we can query the state
frequently, for example during I2C transfers, maintain a cached version
of the HW state to get rid of this delay.
This fixes at least one reported regression where boot time increased by
~4 seconds due to frequent power well state queries on VLV during eDP
EDID read.
This regression has been introduced in
commit bb4932c4f1
Author: Imre Deak <imre.deak@intel.com>
Date: Mon Apr 14 20:24:33 2014 +0300
drm/i915: vlv: check port power domain instead of only D0 for eDP VDD on
Reported-by: Jesse Barnes <jesse.barnes@intel.com>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
So these are the guts of the new beast. This tracks when a frontbuffer
gets invalidated (due to frontbuffer rendering) and hence should be
constantly scaned out, and when it's flushed again and can be
compressed/one-shot-upload.
Rules for flushing are simple: The frontbuffer needs one more full
upload starting from the next vblank. Which means that the flushing
can _only_ be called once the frontbuffer update has been latched.
But this poses a problem for pageflips: We can't just delay the
flushing until the pageflip is latched, since that would pose the risk
that we override frontbuffer rendering that has been scheduled
in-between the pageflip ioctl and the actual latching.
To handle this track asynchronous invalidations (and also pageflip)
state per-ring and delay any in-between flushing until the rendering
has completed. And also cancel any delayed flushing if we get a new
invalidation request (whether delayed or not).
Also call intel_mark_fb_busy in both cases in all cases to make sure
that we keep the screen at the highest refresh rate both on flips,
synchronous plane updates and for frontbuffer rendering.
v2: Lots of improvements
Suggestions from Chris:
- Move invalidate/flush in flush_*_domain and set_to_*_domain.
- Drop the flush in busy_ioctl since it's redundant. Was a leftover
from an earlier concept to track flips/delayed flushes.
- Don't forget about the initial modeset enable/final disable.
Suggested by Chris.
Track flips accurately, too. Since flips complete independently of
rendering we need to track pending flips in a separate mask. Again if
an invalidate happens we need to cancel the evenutal flush to avoid
races.
v3:
Provide correct header declarations for flip functions. Currently not
needed outside of intel_display.c, but part of the proper interface.
v4: Add proper domain management to fbcon so that the fbcon buffer is
also tracked correctly.
v5: Fixup locking around the fbcon set_to_gtt_domain call.
v6: More comments from Chris:
- Split out fbcon changes.
- Drop superflous checks for potential scanout before calling intel_fb
functions - we can micro-optimize this later.
- s/intel_fb_/intel_fb_obj_/ to make it clear that this deals in gem
object. We already have precedence for fb_obj in the pin_and_fence
functions.
v7: Clarify the semantics of the flip flush handling by renaming
things a bit:
- Don't go through a gem object but take the relevant frontbuffer bits
directly. These functions center on the plane, the actual object is
irrelevant - even a flip to the same object as already active should
cause a flush.
- Add a new intel_frontbuffer_flip for synchronous plane updates. It
currently just calls intel_frontbuffer_flush since the implemenation
differs.
This way we achieve a clear split between one-shot update events on
one side and frontbuffer rendering with potentially a very long delay
between the invalidate and flush.
Chris and I also had some discussions about mark_busy and whether it
is appropriate to call from flush. But mark busy is a state which
should be derived from the 3 events (invalidate, flush, flip) we now
have by the users, like psr does by tracking relevant information in
psr.busy_frontbuffer_bits. DRRS (the only real use of mark_busy for
frontbuffer) needs to have similar logic. With that the overall
mark_busy in the core could be removed.
v8: Only when retiring gpu buffers only flush frontbuffer bits we
actually invalidated in a batch. Just for safety since before any
additional usage/invalidate we should always retire current rendering.
Suggested by Chris Wilson.
v9: Actually use intel_frontbuffer_flip in all appropriate places.
Spotted by Chris.
v10: Address more comments from Chris:
- Don't call _flip in set_base when the crtc is inactive, avoids redunancy
in the modeset case with the initial enabling of all planes.
- Add comments explaining that the initial/final plane enable/disable
still has work left to do before it's fully generic.
v11: Only invalidate for gtt/cpu access when writing. Spotted by Chris.
v12: s/_flush/_flip/ in intel_overlay.c per Chris' comment.
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
The downclocking checks a few more things, so not that simple to
convert. Also, this should get unified with the drrs handling and also
use the locking of that. Otoh the drrs locking is about as hapzardous
as no locking, at least on first sight.
For easier conversion ditch the upclocking on unload - we'll turn off
everything anyway.
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
So from just a quick look we seem to have enough information to
accurately figure out whether a given gem bo is used as a frontbuffer
and where exactly: We have obj->pin_count as a first check with no
false negatives and only negligible false positives. And then we can
just walk the modeset objects and figure out where exactly a buffer is
used as scanout.
Except that we can't due to locking order: If we already hold
dev->struct_mutex we can't acquire any modeset locks, so could
potential chase freed pointers and other evil stuff.
So we need something else. For that introduce a new set of bits
obj->frontbuffer_bits to track where a buffer object is used. That we
can then chase without grabbing any modeset locks.
Of course the consumers of this (DRRS, PSR, FBC, ...) still need to be
able to do their magic both when called from modeset and from gem
code. But that can be easily achieved by adding locks for these
specific subsystems which always nest within either kms or gem
locking.
This patch just adds the relevant update code to all places.
Note that if we ever support multi-planar scanout targets then we need
one frontbuffer tracking bit per attachment point that we expose to
userspace.
v2:
- Fix more oopsen. Oops.
- WARN if we leak obj->frontbuffer_bits when freeing a gem buffer. Fix
the bugs this brought to light.
- s/update_frontbuffer_bits/update_fb_bits/. More consistent with the
fb tracking functions (fb for gem object, frontbuffer for raw bits).
And the function name was way too long.
v3: Size obj->frontbuffer_bits correctly so that all pipes fit in.
v4: Don't update fb bits in set_base on failure. Noticed by Chris.
v5: s/i915_gem_update_fb_bits/i915_gem_track_fb/ Also remove a few
local enum pipe variables which are now no longer needed to make the
function arguments no drop over the 80 char limit.
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
It doesn't make sense to never again schedule the work, since by the
time we might want to re-enable psr the world might have changed and
we can do it again.
The only exception is when we shut down the pipe, but that's an
entirely different thing and needs to be handled in psr_disable.
Note that later patch will again split psr_exit into psr_invalidate
and psr_flush. But the split is different and this simplification
helps with the transition.
v2: Improve the commit message a bit.
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
We have _enable/_disable interfaces now for the modeset sequence and
intel_edp_psr_exit for workarounds.
The callsites in intel_display.c are all redundant with the modeset
sequence enable/disable calls in intel_ddi.c. The one in
intel_sprite.c is real and needs to be switched to psr_exit.
If this breaks anything then we need to augment the enable/disable
functions accordingly.
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Introduce generic functions to register and unregister connectors. This
provides a common place to add and remove associated user space
interfaces.
Signed-off-by: Thomas Wood <thomas.wood@intel.com>
Reviewed-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
This patch enables the framework for using MMIO based flip calls,
in contrast with the CS based flip calls which are being used currently.
MMIO based flip calls can be enabled on architectures where
Render and Blitter engines reside in different power wells. The
decision to use MMIO flips can be made based on workloads to give
100% residency for Media power well.
v2: The MMIO flips now use the interrupt driven mechanism for issuing the
flips when target seqno is reached. (Incorporating Ville's idea)
v3: Rebasing on latest code. Code restructuring after incorporating
Damien's comments
v4: Addressing Ville's review comments
-general cleanup
-updating only base addr instead of calling update_primary_plane
-extending patch for gen5+ platforms
v5: Addressed Ville's review comments
-Making mmio flip vs cs flip selection based on module parameter
-Adding check for DRIVER_MODESET feature in notify_ring before calling
notify mmio flip.
-Other changes mostly in function arguments
v6: -Having a seperate function to check condition for using mmio flips (Ville)
-propogating error code from i915_gem_check_olr (Ville)
v7: -Adding __must_check with i915_gem_check_olr (Chris)
-Renaming mmio_flip_data to mmio_flip (Chris)
-Rebasing on latest nightly
v8: -Rebasing on latest code
-squash 3rd patch in series(mmio setbase vs page flip race) with this patch
-Added new tiling mode update in intel_do_mmio_flip (Chris)
v9: -check for obj->last_write_seqno being 0 instead of obj->ring being NULL in
intel_postpone_flip, as this is a more restrictive condition (Chris)
v10: -Applied Chris's suggestions for squashing patches 2,3 into this patch.
These patches make the selection of CS vs MMIO flip at the page flip time, and
make the module parameter for using mmio flips as tristate, the states being
'force CS flips', 'force mmio flips', 'driver discretion'.
Changed the logic for driver discretion (Chris)
v11: Minor code cleanup(better readability, fixing whitespace errors, using
lockdep to check mutex locked status in postpone_flip, removal of __must_check
in function definition) (Chris)
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Sourab Gupta <sourab.gupta@intel.com>
Signed-off-by: Akash Goel <akash.goel@intel.com>
Tested-by: Chris Wilson <chris@chris-wilson.co.uk> # snb, ivb
[danvet: Fix up parameter alignement checkpatch spotted.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Spotted while crawling around in the area.
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
