Daniel Vetter [Thu, 20 Jul 2017 17:57:51 +0000 (19:57 +0200)]
drm/i915: Rip out legacy page_flip completion/irq handling
All these races and things are now solved through the vblank evasion
trick, plus event handling is done using normal vblank even processing
and drm_crtc_arm_vblank_event. We can get rid of all this complexity.
Daniel Vetter [Wed, 19 Jul 2017 12:54:55 +0000 (14:54 +0200)]
drm/i915: Unbreak gpu reset vs. modeset locking
Taking the modeset locks unconditionally isn't the greatest idea,
because atm that part is still broken and times out (and then atomic
keels over). And there's really no reason to do so, the old code
didn't do that either.
To make the patch a bit simpler let's also nuke 2 cases that are only
around for the old mmioflip paths. Atomic nonblocking workers will not
die (minus bugs) when a gpu reset happens.
And of course this doesn't fix any of the gpu reset vs. modeset
deadlock fun, but it at least stop modern CI machines from keeling
over all over the place for no reason at all.
And we still have the explicit testcases to run the fake gpu reset, so
coverage isn't that much worse.
v2: Split out additional changes on top, restrict this to purely reducing
the critical section of modeset locks.
v2: Review from Maarten
- update comments
- don't oops when state is NULL in intel_finish_reset, but try to at
least still drop locks properly. The hw is going to be toast anyway.
Fixes: 739748939974 ("drm/i915: Fix modeset handling during gpu reset, v5.") Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20170719125502.25696-3-daniel.vetter@ffwll.ch
drm/i915: Pass enum pipe to intel_set_pch_fifo_underrun_reporting()
Commit a21960339c8c ("drm/i915: Consistently use enum pipe for PCH
transcoders") misses some pieces, due to a problem with the patch
format, this patch adds the remaining bits.
Fixes: a21960339c8c ("drm/i915: Consistently use enum pipe for PCH
transcoders")
Chris Wilson [Wed, 19 Jul 2017 13:59:57 +0000 (14:59 +0100)]
drm/i915/selftests: Mark contexts as lost during freeing of mock device
We need to unpin the last retired context early in the shutdown sequence
so that its RCU free is done before we try to free the context ida. I
included this in a later patch ("drm/i915: Keep a recent cache of freed
contexts objects for reuse") and so missed that the selftests were broken
in the meantime.
Reported-by: Matthew Auld <matthew.auld@intel.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101627 Fixes: 5f09a9c8ab6b ("drm/i915: Allow contexts to be unreferenced locklessly") Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: Mika Kuoppala <mika.kuoppala@intel.com> Cc: Matthew Auld <matthew.auld@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20170719135957.14603-1-chris@chris-wilson.co.uk Tested-by: Matthew Auld <matthew.auld@intel.com> Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Daniel Vetter [Fri, 14 Jul 2017 22:46:55 +0000 (00:46 +0200)]
drm/i915: Fix fbdev unload sequence
First thing we need to do is unregister the fbdev instance, but we
can't just go ahead and kfree it. That must wait until the hotplug and
polling work are stopped, since they can race with the with the
teardown. That means we need to split up the fbdev teardown into the
unregister part and the cleanup part.
I originally suspected that this was broken in one of the unload
shuffles, but on closer inspection the oldest sequence I've dug out
also gets this wrong. Just not quite so badly.
I've run drv_module_reload a few hundred times and it's rock solid
compared to insta-death beforehand. This bug seems to have been
uncovered by
But the effect of that seems to only be to increase the race window
enough to make it blow up easier. I'm not exactly clear on what's
going on there ...
v2: Fix whitespace and use fetch_and_zero (Chris).
Daniel Vetter [Sat, 15 Jul 2017 09:31:06 +0000 (11:31 +0200)]
drm/atomic-helper: Fix leak in disable_all
The legacy plane->fb pointer is refcounted by calling
drm_atomic_clean_old_fb().
In practice this isn't a real problem because:
- The caller in the i915 gpu reset code restores the original state
again, which means the plane->fb pointer won't change, hence can't
leak.
- Drivers using drm_atomic_helper_shutdown call the fbdev cleanup
first, and that usually cleans up the fb through
drm_remove_framebuffer, which does this correctly.
- Without fbdev the only framebuffers are from userspace, and those
get cleaned up (again using drm_remove_framebuffer) befor the driver
can even be unloaded.
But in i915 I've switched the cleanup sequence around so that the
_shutdown() calls happens after the drm_remove_framebuffer(), which is
how I discovered this issue.
v2: My analysis why the current code was ok for gpu reset and
suspend/resume was correct, but then I totally failed to realize that
we better keep this symmetric. Thanksfully CI noticed that for
balance, a refcounting bug must exist at 2 places if previously there
was no issue ...
v3: Don't be lazy and compute the plane_mask in
commit_duplicated_state properly too, instead of just using ~0U.
Chris Wilson [Tue, 18 Jul 2017 17:30:28 +0000 (18:30 +0100)]
drm/i915/selftests: Attach a stub pm_domain
Supply a pm_domain and its ops for our mock GEM device so that
device runtime pm doesn't complain even though we only want to mark it
permanently active!
Chris Wilson [Tue, 18 Jul 2017 13:41:24 +0000 (14:41 +0100)]
drm/i915: Drain the device workqueue on unload
Workers on the i915->wq may rearm themselves so for completeness we need
to replace our flush_workqueue() with a call to drain_workqueue() before
unloading the device.
v2: Reinforce the drain_workqueue with an preceding rcu_barrier() as a
few of the tasks that need to be drained may first be armed by RCU.
References: https://bugs.freedesktop.org/show_bug.cgi?id=101627 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Matthew Auld <matthew.auld@intel.com> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20170718134124.14832-1-chris@chris-wilson.co.uk Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Daniel Vetter [Wed, 19 Jul 2017 10:00:43 +0000 (12:00 +0200)]
drm/i915: More stolen quirking
I've found a bios with an off-by-one at the other end. There's a pnp
reservation for 0xc5400000-0xc7fffffe and we want stolen in 0xc6000000
through 0xc8000000.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=99872
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=98683 Cc: Martin Peres <martin.peres@linux.intel.com> Reviewed-by: Martin Peres <martin.peres@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20170719100043.30851-1-daniel.vetter@ffwll.ch
drm/i915: Explicit the connector name for DP link training result
This adds the connector name when printing a debug message about the DP
link training result. It is useful to figure out what connector is
failing when multiple DP connectors are used.
Ville Syrjälä [Fri, 14 Jul 2017 15:52:27 +0000 (18:52 +0300)]
drm/i915: Fix cursor updates on some platforms
Turns out that just writing CURPOS isn't sufficient to move the cursor
on some platforms. My 830 works just fine, but eg. 945 and PNV don't.
On those platforms we need to arm even the CURPOS update with a
CURBASE write.
Even worse, a write to any of the cursor register apart from CURBASE
will cancel an already pending cursor update. So if we have armed a
CURCNTR/CURBASE update, a subsequent CURPOS write prior to vblank
would cancel that armed update. Thus we're left with a cursor that
doesn't appear to move, or even change shape.
Fix the problem by always performing the CURBASE write after a
CURPOS write. Bspec is somewhat unclear which platforms actually
require this CURBASE write and which don't. So to keep it simple
and to make sure we really fix the problem across all supported
devices, let's just perform the CURBASE write unconditionally.
Cc: Paul Menzel <pmenzel@molgen.mpg.de>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101790 Fixes: 75343a44c901 ("drm/i915: Drop useless posting reads from cursor commit") Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Tested-by: Paul Menzel <paulepanter@users.sourceforge.net> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: https://patchwork.freedesktop.org/patch/msgid/20170714155227.6089-1-ville.syrjala@linux.intel.com
drm/i915: Consistently use enum pipe for PCH transcoders
The current code uses in some instances enum transcoder for PCH
transcoders and enum pipe in others. This is error prone and clang
raises warnings like this:
drivers/gpu/drm/i915/intel_dp.c:3546:51: warning: implicit conversion
from enumeration type 'enum pipe' to different enumeration type
'enum transcoder' [-Wenum-conversion]
intel_set_pch_fifo_underrun_reporting(dev_priv, PIPE_A, false);
Consistently use the type enum pipe for PCH transcoders.
Imre Deak [Fri, 14 Jul 2017 15:12:42 +0000 (18:12 +0300)]
drm/i915: Fix user ptr check size in eb_relocate_vma()
Fix the sizeof(ptr) vs. sizeof(*ptr) typo.
Fixes: 2889caa92321 ("drm/i915: Eliminate lots of iterations over the execobjects array") Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/20170714151242.517-2-imre.deak@intel.com
Imre Deak [Fri, 14 Jul 2017 15:12:41 +0000 (18:12 +0300)]
drm/i915: Fix error checking/locking in perf/lookup_context()
1acfc104cdf8 missed to convert this one caller to be lockless. The side
effect of that was that the error check in lookup_context() became
incorrect. Convert now this caller too.
Fixes: 1acfc104cdf ("drm/i915: Enable rcu-only context lookups") Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Signed-off-by: Imre Deak <imre.deak@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20170714151242.517-1-imre.deak@intel.com
Daniel Vetter [Thu, 6 Jul 2017 13:00:20 +0000 (15:00 +0200)]
drm/i915: Protect against deferred fbdev setup
We could probably hit this already with our current async fbdev init,
but it's much easier to hit this with the new deferred fbdev setup
that I'm working on polishing.
Daniel Vetter [Thu, 6 Jul 2017 13:00:19 +0000 (15:00 +0200)]
drm/i915/fbdev: Always forward hotplug events
With deferred fbdev setup we always need to forward hotplug events,
even if fbdev isn't fully set up yet. Otherwise the deferred setup
will neer happen.
Originally this check was added in
commit c45eb4fed12d278d3619f1904885bd0d7bcbf036 (tag: drm-intel-next-fixes-2016-08-05)
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date: Wed Jul 13 18:34:45 2016 +0100
drm/i915/fbdev: Check for the framebuffer before use
But the specific case of the hotplug function blowing up was fixed in
height of plane was require to swap width/height in case of 90/270
rotation. Now src structure contains already swapped values, So we
don't have to calculate height of the plane.
drm/i915: Addition wrapper for fixed16.16 operation
This patch introduce addition wrapper for fixed point 16.16 operations.
Which will be used by later patches to avoid direct member variables
access of fixed_16_16_t structure.
This patch make naming of fixed-point wrappers consistent
operation_<any_post_operation>_<1st operand>_<2nd operand>
also shorten the name for fixed_16_16 to fixed16
drm/i915: Always perform internal fixed16 division in 64 bits
This patch combines fixed_16_16_div & fixed_16_16_div_u64 wrappers.
And new fixed_16_16_div wrapper always performs division operation in
u64 internally, to avoid any data loss which was happening in earlier
version of wrapper.
earlier wrapper was converting u32 to fixed16 in 32 bit so we were
losing 16-MSB data.
Paulo had noticed that inside cnl_ddi_vswing_program
the case was handling voltage but with no indication
of type where a missing type could also take us to that
path. So my first attempt was to add a message to
let clear who trigger that path.
However DK had a better idea that is to handle the
missed type sooner before it might take to that path.
So this v2 here uses his approach.
Daniel Vetter [Mon, 10 Jul 2017 19:56:39 +0000 (21:56 +0200)]
Merge tag 'drm-for-v4.13' into drm-intel-next-queued
Resync with the main drm-next pull request for 4.13. What we really
need is to fully resync with pending drm-misc, but that's not yet
possible due to the still ongoing merge window.
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Commit 9a148a96fc3a ("drm/i915/debugfs: add dp mst info") adds support
for DP-MST to intel_connector_info, but forgot to remove the early
return for DP-MST.
drm/i915/cnl: Don't trust VBT's alternate pin for port D for now.
Cannon Lake's VBT that is currently available for B0 stepping
states that port D uses alternate pin 3 messing up with the
default pin-port mapping table. Using that information we cannot
get HDMI working properly. So for now we don't relly on VBT for
this information.
drm/i915: Fix the kernel panic when using aliasing ppgtt
The ppgtt should be get directly from i915_address_space *vm instead of
vma->vm.
v2:
- add one more fix for bxt. (Chris)
Fixes: 4a234c5fae16 ("drm/i915: pass the vma to insert_entries")
Bugzilla:https://bugs.freedesktop.org/show_bug.cgi?id=101713 Signed-off-by: Chuanxiao Dong <chuanxiao.dong@intel.com> Reviewed-by: Matthew Auld <matthew.auld@intel.com> v1 Cc: Matthew Auld <matthew.auld@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Zhenyu Wang <zhenyuw@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1499421059-18262-1-git-send-email-chuanxiao.dong@intel.com Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Imre Deak [Thu, 29 Jun 2017 15:37:01 +0000 (18:37 +0300)]
drm/i915/gen9+: Don't remove secondary power well requests
So far in an attempt to make sure all power wells get disabled during
display uninitialization the driver removed any secondary request bits
(BIOS, KVMR, DEBUG) that were set for a given power well. The known
source for these requests was DMC's request on power well 1 and the misc
IO power well. Since DMC is inactive (DC states are disabled) at the
point we disable these power wells, there shouldn't be any reason to
leave them on. However there are two problems with the above
assumption: Bspec requires that the misc IO power well stays enabled
(without providing a reason) and there can be KVMR requests that we
can't remove anyway (the KVMR request register is R/O). Atm, a KVMR
request can trigger a timeout WARN when trying to disable power wells.
To make the code aligned to Bspec and to get rid of the KVMR WARN, don't
try to remove the secondary requests, only detect them and stop polling
for the power well disabled state when any one is set.
Also add a comment about the timeout values required by Bspec when
enabling power wells and the fact that waiting for them to get disabled
is not required by Bspec.
Imre Deak [Thu, 29 Jun 2017 15:37:00 +0000 (18:37 +0300)]
drm/i915/bxt, glk: Fix assert on conditions for DC9 enabling
What we want to assert based on the conditions required by Bspec is that
power well 2 is disabled, so no need to check for other power wells.
In addition we can only check if the driver's request is removed, the
actual state depends on whether the other request bits are set or not
(BIOS, KVMR, DEBUG). So check only the driver's request bit.
Chris Wilson [Wed, 5 Jul 2017 14:26:34 +0000 (15:26 +0100)]
drm/i915: Only free the oldest stale context before allocating
Currently, we move all unreferenced contexts to an RCU free list and
then onto a worker for eventual reaping. To compensate against this
growing into a long list with frequent allocations starving the system
of available memory, before we allocate a new context we reap all the
stale contexts. This puts all the cost of destroying the context into
the next allocator, which is presumably more sensitive to syscall
latency and unfair. We can limit the number of contexts being freed by
the new allocator to both keep the list trimmed and to allow the
allocator to be reasonably fast.
Chris Wilson [Wed, 5 Jul 2017 14:26:33 +0000 (15:26 +0100)]
drm/i915: Drop request retirement before reaping stale contexts
Before we create a new context, we try and reap all the stale contexts
(i.e. those that are freed but waiting for a worker to come and return
their allocations to the system). Before we do this, we retire all
requests so that we clear any inflight no longer used contexts (who are
only being kept alived by those inflght requests). However, any context
that is finally unreferenced by this retirement is put onto an RCU list
and not available for immediately reaping, we stall for no immediate
benefit.
Chris Wilson [Wed, 5 Jul 2017 14:26:32 +0000 (15:26 +0100)]
drm/i915: Move stale context reaping to common i915_gem_context_create
We need to reap the stale contexts for all new contexts, be they created
by user in i915_gem_context_ioctl or from opening a new file in
i915_gem_context_open. Both paths may be called very frequently
accumulating many stale contexts before any worker has a chance to run
and free their memory.
Xiong Zhang [Thu, 15 Jun 2017 03:11:45 +0000 (11:11 +0800)]
drm/i915: Setting pch_id for HSW/BDW in virtual environment
In a IGD passthrough environment, the real ISA bridge may doesn't exist.
then pch_id couldn't be correctly gotten from ISA bridge, but pch_id is
used to identify LPT_H and LPT_LP. Currently i915 treat all LPT pch as
LPT_H,then errors occur when i915 runs on LPT_LP machines with igd
passthrough.
This patch set pch_id for HSW/BDW according to IGD type and isn't fully
correct. But it solves such issue on HSW/BDW ult/ulx machines.
QA CI system is blocked by this issue for a long time, it's better that
we could merge it to unblock QA CI system.
We know the root cause is in device model of virtual passthrough, and
will resolve it in the future with several parts cooperation in kernel,
qemu and xen.
attribute_groups are not supposed to change at runtime. All functions
working with attribute_groups provided by <linux/sysfs.h> work with const
attribute_group. So mark the non-const structs as const.
File size before:
text data bss dec hex filename
4028 1088 0 5116 13fc drivers/gpu/drm/i915/i915_sysfs.o
File size After adding 'const':
text data bss dec hex filename
4196 928 0 5124 1404 drivers/gpu/drm/i915/i915_sysfs.o
Chris Wilson [Wed, 5 Jul 2017 20:12:44 +0000 (21:12 +0100)]
drm: Remove unused drm_file parameter to drm_syncobj_replace_fence()
the drm_file parameter is unused, so remove it.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Dave Airlie <airlied@redhat.com> Reviewed-by: Jason Ekstrand <jason@jlekstrand.net> Signed-off-by: Dave Airlie <airlied@redhat.com>
Dave Airlie [Thu, 6 Jul 2017 01:10:29 +0000 (11:10 +1000)]
Merge branch 'drm-next-4.13' of git://people.freedesktop.org/~agd5f/linux into drm-next
Fixes for 4.13:
- Various fixes for Raven
- Various fixes for Vega10
- Stability fixes for KIQ
- Fix reloading the driver
- Fix S3 on vega10
- Misc other fixes
* 'drm-next-4.13' of git://people.freedesktop.org/~agd5f/linux: (26 commits)
drm/amd/powerplay: fix bug fail to remove sysfs when rmmod amdgpu.
amdgpu: Set cik/si_support to 1 by default if radeon isn't built
drm/amdgpu/gfx9: fix driver reload with KIQ
drm/amdgpu/gfx8: fix driver reload with KIQ
drm/amdgpu: Don't call amd_powerplay_destroy() if we don't have powerplay
drm/ttm: Fix use-after-free in ttm_bo_clean_mm
drm/amd/amdgpu: move get memory type function from early init to sw init
drm/amdgpu/cgs: always set reference clock in mode_info
drm/amdgpu: fix vblank_time when displays are off
drm/amd/powerplay: power value format change for Vega10
drm/amdgpu/gfx9: support the amdgpu.disable_cu option
drm/amd/powerplay: change PPSMC_MSG_GetCurrPkgPwr for Vega10
drm/amdgpu: Make amdgpu_cs_parser_init static (v2)
drm/amdgpu/cs: fix a typo in a comment
drm/amdgpu: Fix the exported always on CU bitmap
drm/amdgpu/gfx9: gfx_v9_0_enable_gfx_static_mg_power_gating() can be static
drm/amdgpu/psp: upper_32_bits/lower_32_bits for address setup
drm/amd/powerplay/cz: print message if smc message fails
drm/amdgpu: fix typo in amdgpu_debugfs_test_ib_init
drm/amdgpu: enable mmhub pg on raven
...
Rex Zhu [Tue, 4 Jul 2017 08:35:59 +0000 (16:35 +0800)]
drm/amd/powerplay: fix bug fail to remove sysfs when rmmod amdgpu.
when hw_fini, pp will disable dpm.so remove sysfs before
disable dpm.
Signed-off-by: Rex Zhu <Rex.Zhu@amd.com> Tested-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Tue, 4 Jul 2017 05:57:33 +0000 (01:57 -0400)]
drm/amdgpu/gfx9: fix driver reload with KIQ
Drop the KCQ disabling via KIQ. We disable the MEC shortly
after anyway, so there is no need to wait for all of this.
Doing so seems to leave the MEC in a bad way.
Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Tue, 4 Jul 2017 05:47:00 +0000 (01:47 -0400)]
drm/amdgpu/gfx8: fix driver reload with KIQ
Drop the deactivation in KIQ init and drop the KCQ disabling
via KIQ. We disable the MEC shortly after anyway, so there is
no need to wait for all of this. Doing so seems to leave the
MEC in a bad way.
Tested-by: Michel Dänzer <michel.daenzer@amd.com> Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Manasi Navare [Fri, 30 Jun 2017 16:33:48 +0000 (09:33 -0700)]
drm/i915/edp: Add a T12 panel delay quirk to fix DP AUX CH timeouts
This patch fixes the DP AUX CH timeouts observed during CI IGT
tests thus fixing the CI failures. This is done by adding a
quirk for a particular PCI device that requires the panel power
cycle delay (T12) to be set to 800ms which is 300msecs more than
the minimum value specified in the eDP spec. So a quirk is
implemented for that specific PCI device.
v4:
* Add Bugzilla links for FDO bugs in the commit message (Ville, Jani)
v3:
* Change some comments, specify the delay as 800 * 10 (Ville)
v2:
* Change the function and variable names to from PPS_T12_
to _T12 since it is a T12 delay (Clint)
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101144
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101154
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101167
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101515 Cc: Ville Syrjala <ville.syrjala@linux.intel.com> Cc: Clinton Taylor <clinton.a.taylor@intel.com> Signed-off-by: Manasi Navare <manasi.d.navare@intel.com> Reviewed-by: Clinton Taylor <clinton.a.taylor@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1498840428-23176-1-git-send-email-manasi.d.navare@intel.com Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Mahesh Kumar [Fri, 30 Jun 2017 12:11:00 +0000 (17:41 +0530)]
drm/i915/skl+: Scaling not supported in IF-ID Interlace mode
GEN9+ Interlace fetch mode doesn't support pipe/plane scaling,
This patch adds check to fail the flip if pipe/plane scaling is
requested in Interlace fetch mode.
Changes since V1:
- move check to skl_update_scaler (ville)
- mode to adjusted_mode (ville)
- combine pipe/plane scaling check
Changes since V2:
- Indentation fix
- Added TODO to handle/reject NV12 with interlace mode
Mahesh Kumar [Fri, 30 Jun 2017 12:10:59 +0000 (17:40 +0530)]
drm/i915/skl+: Check for supported plane configuration in Interlace mode
In Gen9 platform Interlaced fetch mode doesn't support following plane
configuration:
- Y/Yf tiling
- 90/270 rotation
- YUV420 hybrid planar source pixel formats.
This patch adds check to fail the flip if any of the above configuration
is requested.
Changes since V1:
- handle checks in intel_plane_atomic_check_with_state (ville)
- takeout plane scaler checks combile with pipe scaler in next patch
Changes since V2:
- No need to check for NV12 as it need scaling, so it will be rejected
by scaling check (ville)
Chris Wilson [Thu, 22 Jun 2017 16:02:11 +0000 (17:02 +0100)]
drm/i915/fbdev: Check for existence of ifbdev->vma before operations
Commit fabef825626d ("drm/i915: Drop struct_mutex around frontbuffer
flushes") adds a dependency to ifbdev->vma when flushing the framebufer,
but the checks are only against the existence of the ifbdev->fb and not
against ifbdev->vma. This leaves a window of opportunity where we may
try to operate on the fbdev prior to it being probed (thanks to
asynchronous booting).
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101534 Fixes: fabef825626d ("drm/i915: Drop struct_mutex around frontbuffer flushes") Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Daniel Vetter <daniel.vetter@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20170622160211.783-1-chris@chris-wilson.co.uk Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: stable@vger.kernel.org
John Brooks [Mon, 3 Jul 2017 18:05:35 +0000 (14:05 -0400)]
drm/amdgpu: Don't call amd_powerplay_destroy() if we don't have powerplay
amd_powerplay_destroy() expects a handle pointing to a struct pp_instance.
On chips without PowerPlay, pp_handle points to a struct amdgpu_device. The
resulting attempt to kfree() fields of the wrong struct ends in fire:
Fixes: 1c8638024846 (drm/amd/powerplay: refine powerplay interface.) Signed-off-by: John Brooks <john@fastquake.com> Acked-by: Christian König <christian.koenig@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Cc: stable@vger.kernel.org
John Brooks [Mon, 3 Jul 2017 18:05:34 +0000 (14:05 -0400)]
drm/ttm: Fix use-after-free in ttm_bo_clean_mm
We unref the man->move fence in ttm_bo_clean_mm() and then call
ttm_bo_force_list_clean() which waits on it, except the refcount is now
zero so a warning is generated (or worse):
v2: Set man->move to NULL after dropping the last ref (Christian König)
Fixes: aff98ba1fdb8 (drm/ttm: wait for eviction in ttm_bo_force_list_clean) Signed-off-by: John Brooks <john@fastquake.com> Reviewed-by: Christian König <christian.koenig@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Cc: stable@vger.kernel.org
drm/i915: Prevent kernel panic when reading/writing compliance debugfs files, v2.
When reading all debugfs files on a system with DP-MST the kernel panics
on a null pointer dereference because intel_dp is null for a DP-MST
connector. Detect this case and skip those connectors.
Also fix the write for the DP compliance file in the same way.
Changes since v1:
- Fix i915_displayport_test_active_write too. (DK)
drm/i915: Hold RPM wakelock while initializing OA buffer
OA buffer initialization involves access to HW registers to set
the OA base, head and tail. Ensure device is awake while setting
these. With this, all oa.ops are covered under RPM and forcewake
wakelock.
Alex Deucher [Fri, 30 Jun 2017 13:58:34 +0000 (09:58 -0400)]
drm/amdgpu/cgs: always set reference clock in mode_info
It's relevent regardless of whether there are displays
enabled. Fixes garbage values for ref clock in powerplay
leading to incorrect fan speed reporting when displays
are disabled.
bug: https://bugs.freedesktop.org/show_bug.cgi?id=101653 Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Cc: stable@vger.kernel.org
Alex Deucher [Thu, 29 Jun 2017 20:08:49 +0000 (16:08 -0400)]
drm/amdgpu: fix vblank_time when displays are off
If the displays are off, set the vblank time to max to make
sure mclk switching is enabled. Avoid mclk getting set
to high when no displays are attached.
bug: https://bugs.freedesktop.org/show_bug.cgi?id=101528
fixes: 09be4a5219 (drm/amd/powerplay/smu7: add vblank check for mclk switching (v2)) Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Cc: stable@vger.kernel.org
Rodrigo Vivi [Mon, 19 Jun 2017 21:21:47 +0000 (14:21 -0700)]
drm/i915/cfl: Fix Workarounds.
During the review of Coffee Lake workarounds Mika pointed out
that WaDisableKillLogic and GEN9_DISABLE_OCL_OOB_SUPPRESS_LOGIC
should be removed from CFL and with that I should carry the rv-b.
However when doing the v2 I removed another Workaround that should
remain because although not mentioned by spec the history of hangs
around it advocates on its favor.
On some follow-up patches I continued operating on the wrong
workardound, but Ville noticed that, so here is the fix for the
current CFL code that is upstream already.
Fixes: 46c26662d2f ("drm/i915/cfl: Introduce Coffee Lake workarounds.") Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com> Cc: Mika Kuoppala <mika.kuoppala@intel.com> Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Reviewed-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Eric Huang [Tue, 27 Jun 2017 15:29:09 +0000 (11:29 -0400)]
drm/amd/powerplay: power value format change for Vega10
Power value is an integer on vega10.
Signed-off-by: Eric Huang <JinHuiEric.Huang@amd.com> Acked-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Eric Huang [Mon, 26 Jun 2017 15:41:33 +0000 (11:41 -0400)]
drm/amd/powerplay: change PPSMC_MSG_GetCurrPkgPwr for Vega10
To follow up SMU FW 28.35.0.
Signed-off-by: Eric Huang <JinHuiEric.Huang@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Xie [Tue, 20 Jun 2017 20:35:04 +0000 (16:35 -0400)]
drm/amdgpu: Make amdgpu_cs_parser_init static (v2)
The function is called only once inside the .c file.
v2: update the commit message (Michel)
Signed-off-by: Alex Xie <AlexBin.Xie@amd.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Xie [Tue, 20 Jun 2017 20:33:02 +0000 (16:33 -0400)]
drm/amdgpu/cs: fix a typo in a comment
Signed-off-by: Alex Xie <AlexBin.Xie@amd.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Flora Cui [Tue, 20 Jun 2017 03:08:35 +0000 (11:08 +0800)]
drm/amdgpu: Fix the exported always on CU bitmap
Newer asics with 4 SEs are not able to fit the entire bitmask in the
original field, use an array instead.
v2: keep cu_ao_mask for backward compatibility.
Signed-off-by: Flora Cui <Flora.Cui@amd.com> Acked-by: Christian König <christian.koenig@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Arnd Bergmann [Wed, 21 Jun 2017 21:51:02 +0000 (23:51 +0200)]
drm/amdgpu: fix typo in amdgpu_debugfs_test_ib_init
The debugfs interface has calls a function that was evidently
defined under the wrong name in some configurations:
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c:64:12: error: 'amdgpu_debugfs_test_ib_ring_init' used but never defined [-Werror]
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c:3803:12: error: 'amdgpu_debugfs_test_ib_init' defined but not used [-Werror=unused-function]
This fixes the function name.
Fixes: 4f0955fcc052 ("drm/amdgpu: export test ib debugfs interface") Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Rex Zhu [Mon, 19 Jun 2017 09:11:41 +0000 (17:11 +0800)]
drm/amd/powerplay: add support for ATOM GFXCLK table v2.
New vbios table format on some boards.
Signed-off-by: Rex Zhu <Rex.Zhu@amd.com> Reviewed-by: Ken Wang <Qingqing.Wang@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Chris Wilson [Thu, 29 Jun 2017 15:04:25 +0000 (16:04 +0100)]
drm/i915: Avoid undefined behaviour of "u32 >> 32"
When computing a hash for looking up relocation target handles in an
execbuf, we start with a large size for the hashtable and proceed to
halve it until the allocation succeeds. The final attempt is with an
order of 0 (i.e. a single element). This means that we then pass bits=0
to hash_32() which then computes "hash >> (32 - 0)" to lookup the single
element. Right shifting a value by the width of the operand is
undefined, so limit the smallest hash table we use to order 1.
v2: Keep the retry allocation flag for the final pass
Fixes: 4ff4b44cbb70 ("drm/i915: Store a direct lookup from object handle to vma") Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20170629150425.27508-1-chris@chris-wilson.co.uk
drm/i915: reintroduce VLV/CHV PFI programming power domain workaround
There are still cases on these platforms where an attempt is made to
configure the CDCLK while the power domain is off, like when coming back
from a suspend. So the workaround below is still needed.
This effectively reverts commit 63ff30442519 ("drm/i915: Nuke the
VLV/CHV PFI programming power domain workaround").
Cc: stable@vger.kernel.org
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101517 Suggested-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Gabriel Krisman Bertazi <krisman@collabora.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/20170628210605.4994-1-krisman@collabora.co.uk Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Chris Wilson [Wed, 28 Jun 2017 12:35:48 +0000 (13:35 +0100)]
drm/i915: Avoid keeping waitboost active for signaling threads
Once a client has requested a waitboost, we keep that waitboost active
until all clients are no longer waiting. This is because we don't
distinguish which waiter deserves the boost. However, with the advent of
fence signaling, the signaler threads appear as waiters to the RPS
interrupt handler. So instead of using a single boolean to track when to
keep the waitboost active, use a counter of all outstanding waitboosted
requests.
At this point, I have removed all vestiges of the rate limiting on
clients. Whilst this means that compositors should remain more fluid,
it also means that boosts are more prevalent. See commit b29c19b64528
("drm/i915: Boost RPS frequency for CPU stalls") for a longer discussion
on the pros and cons of both approaches.
A drawback of this implementation is that it requires constant request
submission to keep the waitboost trimmed (as it is now cancelled when the
request is completed). This will be fine for a busy system, but near
idle the boosts may be kept for longer than desired (effectively tens of
vblanks worstcase) and there is a reliance on rc6 instead.
Chris Wilson [Tue, 27 Jun 2017 17:37:31 +0000 (18:37 +0100)]
drm/i915: Drop flushing of the object free list/worker from i915_gem_suspend
i915_gem_suspend() is called from all of our finalization paths
(suspend, hibernate, unload). i915_gem_drain_freed_objects() adds an
arbitrary delay as it uses an rcu_barrier() to ensure that there are no
more freed objects in flight, and this delay causes a large amount of
variability in suspend timings. For S3 suspend, we do not need to free
pages as doing so does not impact at all upon the system in its
suspended state, unlike S4 hibernation where we do want the hibernation
image to be as small as possible. Therefore we can forgo waiting inside
i915_gem_suspend(), so long as we ensure that we do cleanup before
unload (see i915_gem_load_cleanup()) and prefer to reap our objects
prior to hibernation (see i915_gem_freeze()).
Removing the rcu_barrier() from i915_gem_suspend() improves S3 latency
by about 30ms on Skylake (ymmv).
Reported-by: David Weinehall <david.weinehall@linux.intel.com> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: David Weinehall <david.weinehall@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20170627173731.11566-1-chris@chris-wilson.co.uk Tested-by: David Weinehall <david.weinehall@linux.intel.com> Reviewed-by: David Weinehall <david.weinehall@linux.intel.com>
Chris Wilson [Tue, 27 Jun 2017 15:25:10 +0000 (16:25 +0100)]
drm/i915: Cancel pending execlists irq handler upon idling
Due to the slight asynchronicity in handling the execlists interrupts
(i.e. we defer the work to a handler that may consume more than one
interrupt event), when the engine is idle we may still have an irq
tasklet queued (especially when it has been deferred to a ksoftirqd).
At the beginning of the tasklet, we assert that we do hold a device
wakeref for the access we are about to perform. This assumes that when
we idle and release the GT wakeref, all execlists work has been
completed (since the elsp tracking says the hw is idle). However, there
may still be a tasklet queued, so as we mark the engine idle, also
cancel any pending tasklet.
Dave Airlie [Wed, 28 Jun 2017 07:11:12 +0000 (17:11 +1000)]
Merge tag 'drm-intel-next-fixes-2017-06-27' of git://anongit.freedesktop.org/git/drm-intel into drm-next
Just three minor fixups for stuff in -next.
* tag 'drm-intel-next-fixes-2017-06-27' of git://anongit.freedesktop.org/git/drm-intel:
drm/i915: Clear execbuf's vma backpointer upon release
drm/i915: Pass the right flags to i915_vma_move_to_active()
drm/i915/cnl: Fix RMW on ddi vswing sequence.
Dave Airlie [Wed, 28 Jun 2017 07:08:12 +0000 (17:08 +1000)]
Merge branch 'etnaviv/next' of https://git.pengutronix.de/git/lst/linux into drm-next
- a fix from Eric for synchronization with etnaviv exported dma-bufs
- thermal throttle support for newer GPU cores
- updated module clock gating to work around GPU errata
- a fix to restore userspace buffer cache performance
* 'etnaviv/next' of https://git.pengutronix.de/git/lst/linux:
drm/etnaviv: restore ETNA_PREP_NOSYNC behaviour
drm/etnaviv: implement cooling support for new GPU cores
drm/etnaviv: update MLCG disables with info from newer Vivante driver
drm/etnaviv: update common.xml.h
drm/etnaviv: Expose our reservation object when exporting a dmabuf.
Ville Syrjälä [Mon, 26 Jun 2017 20:30:51 +0000 (23:30 +0300)]
drm/i915: Disable MSI for all pre-gen5
We have pretty clear evidence that MSIs are getting lost on g4x and
somehow the interrupt logic doesn't seem to recover from that state
even if we try hard to clear the IIR.
Disabling IER around the normal IIR clearing in the irq handler isn't
sufficient to avoid this, so the problem really seems to be further
up the interrupt chain. This should guarantee that there's always
an edge if any IIR bits are set after the interrupt handler is done,
which should normally guarantee that the CPU interrupt is generated.
That approach seems to work perfectly on VLV/CHV, but apparently
not on g4x.
MSI is documented to be broken on 965gm at least. The chipset spec
says MSI is defeatured because interrupts can be delayed or lost,
which fits well with what we're seeing on g4x. Previously we've
already disabled GMBUS interrupts on g4x because somehow GMBUS
manages to raise legacy interrupts even when MSI is enabled.
Since there's such widespread MSI breakahge all over in the pre-gen5
land let's just give up on MSI on these platforms.
Seqno reporting might be negatively affected by this since the legcy
interrupts aren't guaranteed to be ordered with the seqno writes,
whereas MSI interrupts may be? But an occasioanlly missed seqno
seems like a small price to pay for generally working interrupts.
Cc: stable@vger.kernel.org Cc: Diego Viola <diego.viola@gmail.com> Tested-by: Diego Viola <diego.viola@gmail.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101261 Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20170626203051.28480-1-ville.syrjala@linux.intel.com Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
projects / mirror_ubuntu-bionic-kernel.git / log