Update the self-refresh watermark for display plane/cursor and enable
the memory self-refresh on Ironlake. The watermark is also updated for
the active display plane.
More than 1W idle power is saved on one Ironlake laptop after enabling
memory self-refresh.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Previously, after setting up the Pineview CxSR state, i9xx_update_wm would
get called and overwrite our state.
BTW: We will disable the self-refresh and never enable it any more if we
can't find the appropriate the latency on pineview plaftorm. In such case
the update_wm callback will be NULL.
The bitmask macro is also defined to access the corresponding fifo
watermark register.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
SDVO on Ibexpeak PCH with Ironlake is multiplexed with
HDMIB port, and only has SDVOB port.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
The existing code handling the DPMS ON event is much more careful to
ensure that these registers are enabled according to strict sequencing
requirements. Enabling these early in mode_set simply defeats that.
Signed-off-by: Carl Worth <cworth@cworth.org>
Signed-off-by: Eric Anholt <eric@anholt.net>
As all display drivers have been converted, remove the left reference
for connector object in old structure.
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
DP on Cougarpoint has new training pattern definitions, and
new transcoder DP control register is used to determine the mapping
for transcoder and DP digital output. And eDP for Sandybridge has
new voltage and pre-emphasis level definitions.
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
Cougarpoint is the new PCH for Sandybridge CPU. This one resolves the
chipset change for display pipeline compared to previous Ibexpeak PCH.
Sandybridge/Cougarpoint has different FDI training parameters, so this also
makes seperate FDI training functions for IBX and CPT. Other change includes
new transcoder DPLL select function to set which DPLL for transcoder to pick
up.
And with another new transcoder C introduced in Cougarpoint, each connector
has new transcoder select bits. This one adds that change to light up VGA.
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
In load detection, connector's encoder assignment must be kept
consistent for proper mode setting, and this makes connector as
explicit parameter for load detect function to not require single
data structure to hold both encoder and connector reference, ease
the transition for splitted encoder/connector model.
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
For introducing splitted encoder/connector structure, this helper will return
connector's attached encoder when needed.
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
What we really want is encoder info instead of connector, so change
some more list walk in pipeline setup functions from connector_list
to encoder_list.
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
The intel_output naming is inherited from the UMS code, which had a
structure of screen -> CRTC -> output. The DRM code has an additional
notion of encoder/connector, so the structure is screen -> CRTC ->
encoder -> connector. This is a useful structure for SDVO encoders
which can support multiple connectors (each of which requires
different programming in the one encoder and could be connected to
different CRTCs), or for DVI-I, where multiple encoders feed into the
connector for whether it's used for digital or analog. Most of our
code is encoder-related, so transition it to talking about encoders
before we start trying to distinguish connectors.
This patch is produced by sed s/intel_output/intel_encoder/ over the
driver.
Signed-off-by: Eric Anholt <eric@anholt.net>
This is a purely cosmetic change to make changes in this area easier.
And hey, it's not only clearer and typechecked, but actually shorter,
too!
[anholt: To clarify, this is a change to let us later make
drm_i915_gem_object subclass drm_gem_object, instead of having
drm_gem_object have a pointer to i915's private data]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Acked-by: Dave Airlie <airlied@gmail.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
It is causing hangs after a suspend/resume cycle with the default
powersave=1 module option on these chipsets since 2.6.32-rc.
BugLink: http://bugs.launchpad.net/bugs/492392
Signed-off-by: Robert Hooker <sarvatt@ubuntu.com>
Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Eric Anholt <eric@anholt.net>
* anholt/drm-intel-next:
drm/i915: Record batch buffer following GPU error
drm/i915: give up on 8xx lid status
drm/i915: reduce some of the duplication of tiling checking
drm/i915: blow away userspace mappings before fence change
drm/i915: move a gtt flush to the correct place
agp/intel: official names for Pineview and Ironlake
drm/i915: overlay: drop superflous gpu flushes
drm/i915: overlay: nuke readback to flush wc caches
drm/i915: provide self-refresh status in debugfs
drm/i915: provide FBC status in debugfs
drm/i915: fix drps disable so unload & re-load works
drm/i915: Fix OGLC performance regression on 945
drm/i915: Deobfuscate the render p-state obfuscation
drm/i915: add dynamic performance control support for Ironlake
drm/i915: enable memory self refresh on 9xx
drm/i915: Don't reserve compatibility fence regs in KMS mode.
drm/i915: Keep MCHBAR always enabled
drm/i915: Replace open-coded eviction in i915_gem_idle()
Tools like powertop want to check the current FBC status and report it
to the user. So add a debugfs file indicating whether FBC is enabled,
and if not, why.
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Eric Anholt <eric@anholt.net>
At unload time, we need to disable DRPS, but we need to do it correctly
or the GPU will hang and we won't be able to load the module again. So
set the SFCAVM bit so we can properly restore the DRPS config at unload.
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Eric Anholt <eric@anholt.net>
He Shuang reported an OGLC performance regression introduced in the patch
"enable memory self refresh on 9xx", In that patch, SR on 945 is disabled
everytime when calling intel_mark_busy(), while too much of such operation
will impact performance. Actually disable SR is necessary only when GPU and
Crtc changing from idle to busy. This patch make such optimization.
It fixes upstream bug
http://bugs.freedesktop.org/show_bug.cgi?id=26422
Signed-off-by: Li Peng <peng.li@intel.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
The ironlake render p-state support includes some rather odd variable
names. Clean them up in order to improve the readability of the code.
Signed-off-by: Matthew Garrett <mjg@redhat.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
Ironlake (and 965GM, which this patch doesn't support) supports a
hardware performance and power management feature that allows it to
adjust to changes in GPU load over time with software help. The goal
if this is to maximize performance/power for a given workload.
This patch enables that feature, which is also a requirement for
supporting Intelligent Power Sharing, a feature which allows for
dynamic budgeting of power between the CPU and GPU in Arrandale
platforms.
Tested-by: ykzhao <yakui.zhao@intel.com>
[anholt: Resolved against the irq handler loop removal]
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Eric Anholt <eric@anholt.net>
Enabling memory self refresh (SR) on 9xx needs to set additional
register bits. On 945, we need bit 31 of FW_BLC_SELF to enable the
write to self refresh bit and bit 16 to enable the write of self
refresh watermark. On 915, bit 12 of INSTPM is used to enable SR.
SR will take effect when CPU enters C3+ state and its entry/exit
should be automatically controlled by H/W, driver only needs to set
SR enable bits in wm update. But this isn't safe in my test on 945
because GPU is hung. So this patch explicitly enables SR when GPU
is idle, and disables SR when it is busy. In my test on a netbook of
945GSE chipset, it saves about 0.8W idle power.
Signed-off-by: Li Peng <peng.li@intel.com>
[anholt: rebased against 33c5fd121eabbccc9103daf6cda36941eb3c349f
by adding disable of INSTPM SR bit on 915GM for two pipe setup]
Signed-off-by: Eric Anholt <eric@anholt.net>
Mostly obvious simplifications.
The i915 pread/pwrite ioctls, intel_overlay_put_image and
nouveau_gem_new were incorrectly using the locked versions
without locking: this is also fixed in this patch.
Signed-off-by: Luca Barbieri <luca@luca-barbieri.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This will prevent things from falling over if the user frees the flip
buffer before we complete the flip, since we'll hold an internal
reference.
Reported-by: Kristian Høgsberg <krh@bitplanet.net>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Eric Anholt <eric@anholt.net>
The MI_DISPLAY_FLIP command needs to be set the same pipe
source image like in pipe source register, e.g source image
size minus one. This fixes screen corrupt issue on Ironlake.
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
When a new page flip is requested, we need to both queue an unpin for
the current framebuffer, and also increment the flip pending count on
the newly submitted buffer.
At flip finish time, we need to unpin the old fb and decrement the flip
pending count on the new buffer.
The old code was conflating the two, and led to hangs when new direct
rendered apps were started, replacing the existing frame buffer. This
patch splits out the buffers and prevents the hangs.
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Signed-off-by: Eric Anholt <eric@anholt.net>
On 945, we need to avoid entering self-refresh if the compressor is
busy, or we may cause display FIFO underruns leading to ugly flicker.
Fixes fdo bug #24314, kernel bug #15043.
Tested-by: Alexander Lam <lambchop468@gmail.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Tested-by: Julien Cristau <jcristau@debian.org> (fd.o #25371)
Cc: stable@kernel.org
Signed-off-by: Eric Anholt <eric@anholt.net>
Got Ironlake DPLL parameter table, which reflects the hardware
optimized values. So this one trys to list DPLL parameters for
different output types, should potential fix clock issue seen
on new Arrandale CPUs.
This fixes DPLL setting failure on one 1920x1080 dual channel
LVDS for Ironlake. Test has also been made on LVDS panels with
smaller size and CRT/HDMI/DP ports for different monitors on
their all supported modes.
Update:
- Change name of double LVDS to dual LVDS.
- Fix SSC 120M reference clock to use the right range.
Cc: CSJ <changsijay@gmail.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
Self Refresh should be disabled on dual plane configs. Otherwise, as
the SR watermark is not calculated for such configs, switching to non
VGA mode causes FIFO underrun and display flicker.
This fixes Korg Bug #14897.
Signed-off-by: David John <davidjon@xenontk.org>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Cc: stable@kernel.org
Signed-off-by: Eric Anholt <eric@anholt.net>
The first page flip queued will replace the current front buffer, which
should have a 0 pending flip count. So at finish time we need to handle
that case (i.e. if the flip count is 0 *or* dec_and_test is 0 we need to
wake the waiters).
Also fix up an error path in the queue function and add some debug
output (only enabled with driver debugging).
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Eric Anholt <eric@anholt.net>
so far vblank interrupt on ironlake is disabled, this would cause
bad gfx performance if userspace calls drm_wait_vblank. This patch
enables vblank interrupt on ironlake and follows vblank get/put
model.
Signed-off-by: Li Peng <peng.li@intel.com>
Acked-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
Original DP mode_valid check didn't take pixel color depth into account,
which made one 1600x900 eDP panel's mode check invalid because of overclock,
but actually this 6bpc panel does can work with x1 lane at 2.7G. This one
trys to take bpp value properly both in mode validation and mode setting.
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
When we setup buffer for display plane, we'll check any pending
required GPU flush and possible make interruptible wait for flush
complete. But that wait would be most possibly to fail in case of
signals received for X process, which will then fail modeset process
and put display engine in unconsistent state. The result could be
blank screen or CPU hang, and DDX driver would always turn on outputs
DPMS after whatever modeset fails or not.
So this one creates new helper for setup display plane buffer, and
when needing flush using uninterruptible wait for that.
This one should fix bug like https://bugs.freedesktop.org/show_bug.cgi?id=24009.
Also fixing mode switch stress test on Ironlake.
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
Without this fix, some modes couldn't find appropriate clocks.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
Tested-by: Matthew Garrett <mjg@redhat.com>
For any given clock we can use the find_pll to get the corresponding DPLL
setting. It is unnecessary to use the find_reduce_pll callback function
to calculate the DPLL parameter for LVDS downclock in order to get the same
divider factor(P) for the normal and downclock.
In theory when the LVDS downclock is supported by LVDS panel, we should get the
same DPLL divider factor(P) for the normal clock and reduced downclock.
If we get the diferent divider factor(P) for normal clock and reduced downclock,
it means that the found downclock is incorrect and should be discarded.
So we should use find_pll callback to calculate the DPLL parameter for the
LVDS reduced downclock as for the normal clock. Then we can do the cleanup
about find_reduced_pll.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
cc: Jesse Barnes <jbarnes@virtuousgeek.org>
cc: Matthew Garrett <mjg@redhat.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
For some clocks, the old Ironlake DPLL calculator wold give m/n/p
combinations that didn't match the spreadsheet of what HW validation
tests. Instead, use the G4X DPLL calculator, which does a better job
at it.
So we use the intel_g4x_find_best_pll to calculate the DPLL for CRT/HDMI/LVDS
on ironlake. At the same time to consider the dpll setting for display port, we
add the display port DPLL limit on ironlake, which will directly use the
function of intel_find_pll_ironlake_dp to get the corresponding dpll setting.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
Select the correct BPC for LVDS on Ironlake. If it is 18-bit LVDS panel,
the BPC will be 6. When it is 24-bit LVDS panel, the BPC will 8.
At the same time the BPC will be 8 when the output device is CRT/HDMI/DP.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
Make the BPC in FDI rx/transcoder be consistent with that in pipeconf on Ironlake.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
Enable/disable the dithering for LVDS based on VBT setting. On the 965/g4x
platform the dithering flag is defined in LVDS register. And on the ironlake
the dithering flag is defined in pipeconf register.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
Hugh found an error path where we were attempting to unref a bo without
holding the struct mutex:
[drm:intel_init_clock_gating] *ERROR* failed to pin power context: -16
------------[ cut here ]------------
WARNING: at drivers/gpu/drm/drm_gem.c:438 drm_gem_object_free+0x20/0x5e()
Hardware name: ESPRIMO Mobile V5505
Modules linked in: snd_pcm_oss snd_mixer_oss snd_seq snd_seq_device
Pid: 3793, comm: s2ram Not tainted 2.6.33-rc2 #4
Call Trace:
[<7815298e>] warn_slowpath_common+0x59/0x6b
[<781529b3>] warn_slowpath_null+0x13/0x18
[<78317c1a>] ? drm_gem_object_free+0x20/0x5e
[<78317c1a>] drm_gem_object_free+0x20/0x5e
[<78317bfa>] ? drm_gem_object_free+0x0/0x5e
[<7829df11>] kref_put+0x38/0x45
[<7833a5f0>] intel_init_clock_gating+0x232/0x271
[<78317bfa>] ? drm_gem_object_free+0x0/0x5e
[<7832c307>] i915_restore_state+0x21a/0x2b3
[<7832379d>] i915_resume+0x3c/0xbb
[<78174fe5>] ? trace_hardirqs_on_caller+0xfc/0x123
[<7831c756>] ? drm_class_resume+0x0/0x3e
[<7831c78d>] drm_class_resume+0x37/0x3e
[<78351e0a>] legacy_resume+0x1e/0x51
[<78351ece>] device_resume+0x91/0xab
[<7831c756>] ? drm_class_resume+0x0/0x3e
[<78352226>] dpm_resume+0x58/0x10f
[<783522fb>] dpm_resume_end+0x1e/0x2c
[<78180f80>] suspend_devices_and_enter+0x61/0x84
[<78180ff8>] enter_state+0x55/0x83
[<7818091c>] state_store+0x94/0xaa
[<7829d09e>] kobj_attr_store+0x1e/0x23
[<782098e0>] sysfs_write_file+0x66/0x99
[<781cd2f0>] vfs_write+0x8a/0x108
[<781cd408>] sys_write+0x3c/0x63
[<78125c10>] sysenter_do_call+0x12/0x36
---[ end trace a343537f29950fda ]---
It is in fact slightly more insiduous that first appears since we are
attempting to not just free the object without the lock, but are trying
to do the whole bo manipulation without holding the lock.
Reported-by: Hugh Dickins <hugh.dickins@tiscali.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: stable@kernel.org
Signed-off-by: Eric Anholt <eric@anholt.net>
