drm/amd/display: fix undefined struct member reference

An initialization was added for two optional struct members.  One of
these is always present in the dcn20_resource file, but the other one
depends on CONFIG_DRM_AMD_DC_DSC_SUPPORT and causes a build failure if
that is missing:

drivers/gpu/drm/amd/amdgpu/../display/dc/dcn20/dcn20_resource.c:926:14: error: excess elements in struct initializer [-Werror]
   .num_dsc = 5,

Add another #ifdef around the assignment.

Fixes: c3d03c5a196f ("drm/amd/display: Include num_vmid and num_dsc within NV14's resource caps")
Reviewed-by: Zhan Liu <zhan.liu@amd.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Compare clock state member to determine optimization.

[Why]
It seems always request passive flip on RN due to incorrect compare
clock state to determine optization.

[How]
Instead of calling memcmp, compare clock state member to determine the
condition.

Signed-off-by: Yongqiang Sun <yongqiang.sun@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/display: add fallthrough comment

To avoid a compiler warning.

Reviewed-by: Zhan Liu <zhan.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Increase the number of retries after AUX DEFER

[Why]
When a timeout occurs after a DEFER, some devices require more retries
than in the case of a regular timeout.

[How]
In a timeout occurrence, check whether a DEFER has occurred before the
timeout and retry MAX_DEFER_RETRIES retries times instead of
MAX_TIMEOUT_RETRIES.

Signed-off-by: George Shen <george.shen@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Abdoulaye Berthe <Abdoulaye.Berthe@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: update dispclk and dppclk vco frequency

Value obtained from DV is not allowing 8k60 CTA mode with DSC to
pass, after checking real value being used in hw, find out that
correct value is 3600, which will allow that mode.

Signed-off-by: Eric Yang <Eric.Yang2@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fix screen tearing on vrr tests

[Why]
Screen tearing is present in tests when setting the frame rate to
certain fps

[How]
Revert previous optimizations for low frame rates.

Signed-off-by: Amanda Liu <amanda.liu@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Reset steer fifo before unblanking the stream

[why]
During mode transition steer fifo could overflow. Quite often it
recovers by itself, but sometimes it doesn't.

[how]
Add steer fifo reset before unblanking the stream. Also add a short
delay when resetting dig resync fifo to make sure register writes
don't end up back-to-back, in which case the HW might miss the reset
request.

Signed-off-by: Nikola Cornij <nikola.cornij@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: fixed that I2C over AUX didn't read data issue

[Why]
The variable mismatch assignment error.

[How]
To use uint32_t replace it.

Signed-off-by: Brandon Syu <Brandon.Syu@amd.com>
Reviewed-by: Charlene Liu <Charlene.Liu@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Change the delay time before enabling FEC

[why]
DP spec requires 1000 symbols delay between the end of link training
and enabling FEC in the stream. Currently we are using 1 miliseconds
delay which is not accurate.

[how]
One lane RBR should have the maximum time for transmitting 1000 LL
codes which is 6.173 us. So using 7 microseconds delay instead of
1 miliseconds.

Signed-off-by: Leo (Hanghong) Ma <hanghong.ma@amd.com>
Reviewed-by: Harry Wentland <Harry.Wentland@amd.com>
Reviewed-by: Nikola Cornij <Nikola.Cornij@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: have two different sr and pstate latency tables for renoir

[Why]
new sr and pstate latencies are optimized for the case when we are not
using lpddr4 memory

[How]
have two different wm tables, one for the lpddr case and one for
non lpddr case

Signed-off-by: Joseph Gravenor <joseph.gravenor@amd.com>
Reviewed-by: Eric Yang <eric.yang2@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fixed kernel panic when booting with DP-to-HDMI dongle

[Why]
In dc_link_is_dp_sink_present, if dal_ddc_open fails, then
dal_gpio_destroy_ddc is called, destroying pin_data and pin_clock. They
are created only on dc_construct, and next aux access will cause a panic.

[How]
Instead of calling dal_gpio_destroy_ddc, call dal_ddc_close.

Signed-off-by: David Galiffi <David.Galiffi@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: populate bios integrated info for renoir

[Why]
When video_memory_type bw_params->vram_type
is assigned, wedistinguish between Ddr4MemType and LpDdr4MemType.
Because of this we will never report that we are using
LpDdr4MemType and never re-purpose WM set D

[How]
populate bios integrated info for renoir by adding the
revision number for renoir and use that integrated info
table instead of of asic_id to get the vram type

Signed-off-by: Joseph Gravenor <joseph.gravenor@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: fix DalDramClockChangeLatencyNs override

[why]
pstate_latency_us never gets updated from the hard coded value
in rn_clk_mgr.c

[how]
update the wm table's values before we do calculations with them

Signed-off-by: Joseph Gravenor <joseph.gravenor@amd.com>
Reviewed-by: Eric Yang <eric.yang2@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Map DSC resources 1-to-1 if numbers of OPPs and DSCs are equal

[why]
On ASICs where number of DSCs is the same as OPPs there's no need
for DSC resource management. Mappping 1-to-1 fixes mode-set- or S3-
-related issues for such platforms.

[how]
Map DSC resources 1-to-1 to pipes only if number of OPPs is the same
as number of DSCs. This will still keep other ASICs working.
A follow-up patch to fix mode-set issues on those ASICs will be
required if testing shows issues with mode set.

Signed-off-by: Nikola Cornij <nikola.cornij@amd.com>
Reviewed-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: update sr and pstate latencies for Renoir

[Why]
DF team has produced more optimized latency numbers.

[How]
Add sr latencies to the wm table, use different latencies
for different wm sets.
Also fix bb override from registery key for these latencies.

Signed-off-by: Eric Yang <Eric.Yang2@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add cache flush workaround to gfx8 emit_fence

The same workaround is used for gfx7.
Both PAL and Mesa use it for gfx8 too, so port this commit to
gfx_v8_0_ring_emit_fence_gfx.

Signed-off-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add header line for power profile on Arcturus

So the output is consistent with other asics.

Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add check before enabling/disabling broadcast mode

When security violation from new vbios happens, data fabric is
risky to stop working. So prevent the direct access to DF
mmFabricConfigAccessControl from the new vbios and onwards.

Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Merge tag 'drm-misc-next-fixes-2019-12-04' of git://anongit.freedesktop.org/drm/drm-misc into drm-next

mgag200- Fix hw with broken 'startadd' support (Thomas)
mst- Avoid skipping payloads in payload deletion loop (Wayne)
omap- Fix dma_addr refcounting (Tomi)

Cc: Wayne Lin <Wayne.Lin@amd.com>
Cc: Tomi Valkeinen <tomi.valkeinen@ti.com>
Cc: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Sean Paul <sean@poorly.run>
Link: https://patchwork.freedesktop.org/patch/msgid/20191204212255.GA256395@art_vandelay

Merge tag 'drm/tegra/for-5.5-rc1-fixes' of git://anongit.freedesktop.org/tegra/linux into drm-next

drm/tegra: Fixes for v5.5-rc1

This is a set of small fixes, mostly for regressions introduced with the
DMA API and DisplayPort support in the main pull request for v5.5-rc1.

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Thierry Reding <thierry.reding@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191204124316.3534855-1-thierry.reding@gmail.com

drm/dp_mst: Correct the bug in drm_dp_update_payload_part1()

[Why]
If the payload_state is DP_PAYLOAD_DELETE_LOCAL in series, current
code doesn't delete the payload at current index and just move the
index to next one after shuffling payloads.

[How]
Drop the i++ increasing part in for loop head and decide whether
to increase the index or not according to payload_state of current
payload.

Changes since v1:
* Refine the code to have it easy reading
* Amend the commit message to meet the way code is modified now.

Signed-off-by: Wayne Lin <Wayne.Lin@amd.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
Fixes: 706246c761dd ("drm/dp_mst: Refactor drm_dp_update_payload_part1()")
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Juston Li <juston.li@intel.com>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Maxime Ripard <mripard@kernel.org>
Cc: Sean Paul <sean@poorly.run>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: dri-devel@lists.freedesktop.org
Cc: <stable@vger.kernel.org> # v5.1+
[Added cc for stable]
Signed-off-by: Lyude Paul <lyude@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191203042423.5961-1-Wayne.Lin@amd.com

drm/omap: fix dma_addr refcounting

cec4fa7511ef7a73eb635834e9d85b25a5b47a98 ("drm/omap: use refcount API to
track the number of users of dma_addr") changed omap_gem.c to use
refcounting API to track dma_addr uses.  However, the driver only tracks
the refcounts for non-contiguous buffers, and the patch didn't fully
take this in account.

After the patch, the driver always decreased refcount in omap_gem_unpin,
instead of decreasing the refcount only for non-contiguous buffers. This
leads to refcounting mismatch.

As for the contiguous cases the refcount is never increased, fix this
issue by returning from omap_gem_unpin if the buffer being unpinned is
contiguous.

Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191114080343.30704-1-tomi.valkeinen@ti.com
Fixes: cec4fa7511ef ("drm/omap: use refcount API to track the number of users of dma_addr")
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

drm/tegra: Run hub cleanup on ->remove()

The call to tegra_display_hub_cleanup() that takes care of disabling the
window groups is missing from the driver's ->remove() callback. Call it
to make sure the runtime PM reference counts for the display controllers
are balanced.

Signed-off-by: Thierry Reding <treding@nvidia.com>

drm/tegra: sor: Make the +5V HDMI supply optional

The SOR supports multiple display modes, but only when driving an HDMI
monitor does it make sense to control the +5V power supply. eDP and DP
don't need this, so make it optional.

This fixes a crash observed during system suspend/resume.

Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Thierry Reding <treding@nvidia.com>

drm/tegra: Silence expected errors on IOMMU attach

Subdevices may not be hooked up to an IOMMU via device tree. Detect such
situations and avoid confusing users by not emitting an error message.

Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Thierry Reding <treding@nvidia.com>

drm/tegra: vic: Export module device table

Export the module device table to ensure the VIC compatible strings are
listed in the module's aliases table. This in turn causes the driver to
be automatically loaded on boot if VIC is the only enabled subdevice of
the logical host1x DRM device.

Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Thierry Reding <treding@nvidia.com>

drm/tegra: sor: Implement system suspend/resume

Upon system suspend, make sure the +5V HDMI regulator is disabled. This
avoids potentially leaking current to the HDMI connector. This also
makes sure that upon resume the regulator is enabled again, which in
some cases is necessary to properly restore the state of the supply on
resume.

Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Thierry Reding <treding@nvidia.com>

drm/tegra: Use proper IOVA address for cursor image

The IOVA address for the cursor is the result of mapping the buffer
object for the given display controller. Make sure to use the proper
IOVA address as stored in the cursor's plane state.

Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Thierry Reding <treding@nvidia.com>

drm/tegra: gem: Remove premature import restrictions

All the display related blocks on Tegra require contiguous memory. Using
the DMA API, there is no knowing at import time which device will end up
using the buffer, so it's not known whether or not an IOMMU will be used
to map the buffer.

Move the check for non-contiguous buffers/mappings to the tegra_dc_pin()
function which is now the earliest point where it is known if a DMA BUF
can be used by the given device or not.

v2: add check for contiguous buffer/mapping in tegra_dc_pin()

Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Thierry Reding <treding@nvidia.com>

drm/tegra: gem: Properly pin imported buffers

Buffers that are imported from a DMA-BUF don't have pages allocated with
them. At the same time an SG table for them can't be derived using the
DMA API helpers because the necessary information doesn't exist. However
there's already an SG table that was created during import, so this can
simply be duplicated.

Signed-off-by: Thierry Reding <treding@nvidia.com>

drm/tegra: hub: Remove bogus connection mutex check

The Tegra DRM never actually took that lock, so the driver was broken
until generic locking was added in commit:

commit b962a12050a387e4bbf3a48745afe1d29d396b0d
Author: Rob Clark <robdclark@gmail.com>
Date:   Mon Oct 22 14:31:22 2018 +0200

    drm/atomic: integrate modeset lock with private objects

It's now no longer necessary to take that lock, so drop the check.

v2: add rationale for why it is now safe to drop the check (Daniel)

Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Thierry Reding <treding@nvidia.com>

ia64: agp: Replace empty define with do while

It's dangerous to use empty code define.
Furthermore it lead to the following warning:
drivers/char/agp/generic.c: In function « agp_generic_destroy_page »:
drivers/char/agp/generic.c:1266:28: attention : suggest braces around empty body in an « if » statement [-Wempty-body]

So let's replace emptyness by "do {} while(0)"

Signed-off-by: Corentin Labbe <clabbe@baylibre.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/1574324085-4338-6-git-send-email-clabbe@baylibre.com

agp: Add bridge parameter documentation

This patch add documentation about the bridge parameter in several
function.

This will fix the following build warning:
drivers/char/agp/generic.c:220: warning: No description found for parameter 'bridge'
drivers/char/agp/generic.c:364: warning: No description found for parameter 'bridge'
drivers/char/agp/generic.c:1283: warning: No description found for parameter 'bridge'

Signed-off-by: Corentin Labbe <clabbe@baylibre.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/1574324085-4338-5-git-send-email-clabbe@baylibre.com

agp: remove unused variable num_segments

This patch fix the following build warning:
warning: variable 'num_segments' set but not used [-Wunused-but-set-variable]

Signed-off-by: Corentin Labbe <clabbe@baylibre.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/1574324085-4338-4-git-send-email-clabbe@baylibre.com

agp: move AGPGART_MINOR to include/linux/miscdevice.h

This patch move the define for AGPGART_MINOR to include/linux/miscdevice.h.
It is better that all minor number definitions are in the same place.

Signed-off-by: Corentin Labbe <clabbe@baylibre.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/1574324085-4338-3-git-send-email-clabbe@baylibre.com

agp: remove unused variable size in agp_generic_create_gatt_table

This patch fix the following warning:
drivers/char/agp/generic.c:853:6: attention : variable ‘size’ set but not used [-Wunused-but-set-variable]
by removing the unused variable size in agp_generic_create_gatt_table

Signed-off-by: Corentin Labbe <clabbe@baylibre.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/1574324085-4338-2-git-send-email-clabbe@baylibre.com

Merge tag 'drm-next-5.5-2019-12-03' of git://people.freedesktop.org/~agd5f/linux into drm-next

drm-next-5.5-2019-12-03:

amdgpu:
- Fix vram lost handling with BACO on VI/CI asics
- DC fixes for Navi14
- Misc gfx10 fixes
- SR-IOV fixes
- Fix driver unload
- Fix XGMI limits on Arcturus

amdkfd:
- Enable KFD on PPC
- Optimize KFD page table reservations

radeon:
- Fix register checker for r1xx/r2xx

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Alex Deucher <alexdeucher@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191203204135.5437-1-alexander.deucher@amd.com

drm/dp_mst: Fix build on systems with STACKTRACE_SUPPORT=n

On systems with STACKTRACE_SUPPORT=n, we get:

WARNING: unmet direct dependencies detected for STACKTRACE
  Depends on [n]: STACKTRACE_SUPPORT
  Selected by [y]:
  - STACKDEPOT [=y]

and build errors such as:

m68k-linux-ld: kernel/stacktrace.o: in function `stack_trace_save':
(.text+0x11c): undefined reference to `save_stack_trace'

Add the missing deendency on STACKTRACE_SUPPORT.

Fixes: 12a280c72868 ("drm/dp_mst: Add topology ref history tracking for debugging")
Cc: Lyude Paul <lyude@redhat.com>
Cc: Sean Paul <sean@poorly.run>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Lyude Paul <lyude@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191202133650.11964-1-linux@roeck-us.net

drm/radeon: fix r1xx/r2xx register checker for POT textures

Shift and mask were reversed.  Noticed by chance.

Tested-by: Meelis Roos <mroos@linux.ee>
Reviewed-by: Michel Dänzer <mdaenzer@redhat.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

drm/amdgpu: fix GFX10 missing CSIB set(v3)

still need to init csb even for SRIOV

v2:
drop init_pg() for gfx10 at all since
PG and GFX off feature will be fully controled
by RLC and SMU fw for gfx10

v3:
drop the flush_gpu_tlb lines since we consider
it is only usefull in emulation

Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: should stop GFX ring in hw_fini

To align with the scheme from gfx9

disabling GFX ring after VM shutdown could avoid
garbage data be fetched to GFX RB which may lead
to unnecessary screw up on GFX

Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: do autoload right after MEC loaded for SRIOV VF

since we don't have RLCG ucode loading and no SRlist as well

Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: skip rlc ucode loading for SRIOV gfx10

Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: fix calltrace during kmd unload(v3)

issue:
kernel would report a warning from a double unpin
during the driver unloading on the CSB bo

why:
we unpin it during hw_fini, and there will be another
unpin in sw_fini on CSB bo.

fix:
actually we don't need to pin/unpin it during
hw_init/fini since it is created with kernel pinned,
we only need to fullfill the CSB again during hw_init
to prevent CSB/VRAM lost after S3

v2:
get_csb in init_rlc so hw_init() will make CSIB content
back even after reset or s3

v3:
use bo_create_kernel instead of bo_create_reserved for CSB
otherwise the bo_free_kernel() on CSB is not aligned and
would lead to its internal reserve pending there forever

take care of gfx7/8 as well

Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Xiaojie Yuan <xiaojie.yuan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/gfx10: unlock srbm_mutex after queue programming finish

srbm_mutex is to guarantee atomicity for r/w of gfx indexed registers

Signed-off-by: Xiaojie Yuan <xiaojie.yuan@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Added ASIC specific checks in gfxhub V1.1 get XGMI info

Added max hive/node info checks per supported ASIC

Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: John Clements <john.clements@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/powerplay: unify smu send message function

Drop smu_send_smc_msg function from ASIC specify structure.
Reuse smu_send_smc_msg_with_param function for smu_send_smc_msg.
Set paramer to 0 for smu_send_msg function, otherwise it will send
with previous paramer value (Not a certain value).
Materialize msg type for smu send message function definition.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Kevin Wang <kevin1.wang@amd.com>
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: re-enable wait in pipelock, but add timeout

Removing this causes hangs in some games, so re-add it, but add
a timeout so we don't hang while switching flip types.

Bug: https://bugzilla.kernel.org/show_bug.cgi?id=205169
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=112266
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

drm/amd/display: Get NV14 specific ip params as needed

[Why]
NV14 is using its own ip params that's different from other
DCN2.0 ASICs.

[How]
Add ASIC revision check to make sure NV14 gets correct
ip params.

Signed-off-by: Zhan Liu <zhan.liu@amd.com>
Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Adding NV14 IP Parameters

[Why]
NV14 IP Parameters are missing.

[How]
Add IP Parameters in.

Signed-off-by: Zhan liu <zhan.liu@amd.com>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Include num_vmid and num_dsc within NV14's resource caps

[Why]
"num_vmid" and "num_dsc" are missing within NV14's resource caps structure.

[How]
Add the missing parts.

Signed-off-by: Zhan Liu <zhan.liu@amd.com>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: use CPU to flush vmhub if sched stopped

otherwse the flush_gpu_tlb will hang if we unload the
KMD becuase the schedulers already stopped

Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Merge tag 'drm-intel-next-fixes-2019-11-28' of git://anongit.freedesktop.org/drm/drm-intel into drm-next

- Important fix to uAPI alignment on query IOCTL
- Fixes for the power regression introduced by the previous security patches
- Avoid regressing super heavy benchmarks by increasing the default request pre-emption timeout from 100 ms to 640 ms to
- Resulting set of smaller fixes done while problem was inspected
- Display fixes for EHL voltage level programming and TGL DKL PHY vswing for HDMI

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191128141524.GA11992@jlahtine-desk.ger.corp.intel.com

Merge tag 'drm-msm-next-2019-11-05' of https://gitlab.freedesktop.org/drm/msm into drm-next

+ OCMEM support to enable the couple generations that had shared OCMEM
  rather than GMEM exclusively for the GPU (late a3xx and I think basically
  all of a4xx).  Bjorn and Brian decided to land this through the drm
  tree to avoid having to coordinate merge requests.
+ a510 support, and various associated display support
+ the usual misc cleanups and fixes

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Rob Clark <robdclark@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/

drm/i915: Reduce nested prepare_remote_context() to a trylock

On context retiring, we may invoke the kernel_context to unpin this
context. Elsewhere, we may use the kernel_context to modify this
context. This currently leads to an AB-BA lock inversion, so we need to
back-off from the contended lock, and repeat.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=111732
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Fixes: a9877da2d629 ("drm/i915/oa: Reconfigure contexts on the fly")
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191126065521.2331017-1-chris@chris-wilson.co.uk
(cherry picked from commit 58b4c1a07ada7fb91ea757bdb9bd47df02207357)
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915: Default to a more lenient forced preemption timeout

Based on a sampling of a number of benchmarks across platforms, by
default opt for a much more lenient timeout so that we should not
adversely affect existing "good" clients.

640ms ought to be enough for anyone.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=112169
Fixes: 3a7a92aba8fb ("drm/i915/execlists: Force preemption")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Eero Tamminen <eero.t.tamminen@intel.com>
Cc: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191125162737.2161069-1-chris@chris-wilson.co.uk
(cherry picked from commit 5766a5ffc6a69595903865518c43636bde0e4ac4)
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/mgag200: Add workaround for HW that does not support 'startadd'

There's at least one system that does not interpret the value of
the device's 'startadd' field correctly, which leads to incorrectly
displayed scanout buffers. Always placing the active scanout buffer
at offset 0 works around the problem.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reported-by: John Donnelly <john.p.donnelly@oracle.com>
Tested-by: John Donnelly <john.p.donnelly@oracle.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Fixes: 81da87f63a1e ("drm: Replace drm_gem_vram_push_to_system() with kunmap + unpin")
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Dave Airlie <airlied@redhat.com>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Maxime Ripard <mripard@kernel.org>
Cc: David Airlie <airlied@linux.ie>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: "Y.C. Chen" <yc_chen@aspeedtech.com>
Cc: Neil Armstrong <narmstrong@baylibre.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "José Roberto de Souza" <jose.souza@intel.com>
Cc: Andrzej Pietrasiewicz <andrzej.p@collabora.com>
Cc: dri-devel@lists.freedesktop.org
Cc: <stable@vger.kernel.org> # v5.3+
Link: https://gitlab.freedesktop.org/drm/misc/issues/7
Link: https://patchwork.freedesktop.org/patch/msgid/20191126101529.20356-4-tzimmermann@suse.de

drm/mgag200: Store flags from PCI driver data in device structure

The flags field in struct mga_device has been unused so far. We now
use it to store flag bits from the PCI driver.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Fixes: 81da87f63a1e ("drm: Replace drm_gem_vram_push_to_system() with kunmap + unpin")
Cc: John Donnelly <john.p.donnelly@oracle.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Dave Airlie <airlied@redhat.com>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Maxime Ripard <mripard@kernel.org>
Cc: David Airlie <airlied@linux.ie>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: "Y.C. Chen" <yc_chen@aspeedtech.com>
Cc: Neil Armstrong <narmstrong@baylibre.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "José Roberto de Souza" <jose.souza@intel.com>
Cc: Andrzej Pietrasiewicz <andrzej.p@collabora.com>
Cc: dri-devel@lists.freedesktop.org
Cc: <stable@vger.kernel.org> # v5.3+
Link: https://patchwork.freedesktop.org/patch/msgid/20191126101529.20356-3-tzimmermann@suse.de

drm/mgag200: Extract device type from flags

Adds a conversion function that extracts the device type from the
PCI id-table flags. Allows for storing additional information in the
other flag bits.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Fixes: 81da87f63a1e ("drm: Replace drm_gem_vram_push_to_system() with kunmap + unpin")
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: John Donnelly <john.p.donnelly@oracle.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Dave Airlie <airlied@redhat.com>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Maxime Ripard <mripard@kernel.org>
Cc: David Airlie <airlied@linux.ie>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Emil Velikov <emil.velikov@collabora.com>
Cc: "Y.C. Chen" <yc_chen@aspeedtech.com>
Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Cc: "José Roberto de Souza" <jose.souza@intel.com>
Cc: Andrzej Pietrasiewicz <andrzej.p@collabora.com>
Cc: dri-devel@lists.freedesktop.org
Cc: <stable@vger.kernel.org> # v5.3+
Link: https://patchwork.freedesktop.org/patch/msgid/20191126101529.20356-2-tzimmermann@suse.de

amdgpu: Enable KFD on POWER systems

KFD has been verified to function on POWER systems (Talos II / Vega 64).
It should be available as a kernel configuration option on these systems.

Signed-off-by: Timothy Pearson <tpearson@raptorengineering.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Optimize KFD page table reservation

Be less pessimistic about estimated page table use for KFD. Most
allocations use 2MB pages and therefore need less VRAM for page
tables. This allows more VRAM to be used for applications especially
on large systems with many GPUs and hundreds of GB of system memory.

Example: 8 GPUs with 32GB VRAM each + 256GB system memory = 512GB
Old page table reservation per GPU:  1GB
New page table reservation per GPU: 32MB

Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: xinhui pan <xinhui.pan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: flag vram lost on baco reset for VI/CIK

VI/CIK BACO was inflight when this fix landed for SOC15/NV.
Add the fix to VI/CIK as well.

Acked-by: Evan Quan <evan.quan@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

MAINTAINERS: Drop Rex Zhu for amdgpu powerplay

No longer works on the driver.

Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Resolved offchip EEPROM I/O issue

Updated target I2C address

Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: John Clements <john.clements@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: add default clocks if not able to fetch them

dm_pp_get_clock_levels_by_type needs to add the default clocks
to the powerplay case as well.  This was accidently dropped.

Fixes: b3ea88fef321de ("drm/amd/powerplay: add get_clock_by_type interface for display")
Bug: https://gitlab.freedesktop.org/drm/amd/issues/906
Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

drm/i915/gt: Make intel_ring_unpin() safe for concurrent pint

In order to avoid some nasty mutex inversions, commit 09c5ab384f6f
("drm/i915: Keep rings pinned while the context is active") allowed the
intel_ring unpinning to be run concurrently with the next context
pinning it. Thus each step in intel_ring_unpin() needed to be atomic and
ordered in a nice onion with intel_ring_pin() so that the lifetimes
overlapped and were always safe.

Sadly, a few steps in intel_ring_unpin() were overlooked, such as
closing the read/write pointers of the ring and discarding the
intel_ring.vaddr, as these steps were not serialised with
intel_ring_pin() and so could leave the ring in disarray.

Fixes: 09c5ab384f6f ("drm/i915: Keep rings pinned while the context is active")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191118230254.2615942-6-chris@chris-wilson.co.uk
(cherry picked from commit a266bf42006004306dd48a9082c35dfbff153307)
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

Merge tag 'drm-next-5.5-2019-11-22' of git://people.freedesktop.org/~agd5f/linux into drm-next

drm-next-5.5-2019-11-22:

amdgpu:
- Fix bad DMA on some PPC platforms
- MMHUB fix for powergating
- BACO fix for Navi
- Misc raven fixes
- Enable vbios fetch directly from rom on navi
- debugfs fix for DC
- SR-IOV fixes for arcturus
- Misc power fixes

radeon:
- Fix bad DMA on some PPC platforms

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Alex Deucher <alexdeucher@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191122203025.3787-1-alexander.deucher@amd.com

Merge tag 'drm-intel-next-fixes-2019-11-22' of git://anongit.freedesktop.org/drm/drm-intel into drm-next

- Reverts a patch to avoid spinning forever when context's timeline
  is active but has no requests

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191122155523.GA20167@jlahtine-desk.ger.corp.intel.com

drm/i915/gt: Schedule request retirement when timeline idles

The major drawback of commit 7e34f4e4aad3 ("drm/i915/gen8+: Add RC6 CTX
corruption WA") is that it disables RC6 while Skylake (and friends) is
active, and we do not consider the GPU idle until all outstanding
requests have been retired and the engine switched over to the kernel
context. If userspace is idle, this task falls onto our background idle
worker, which only runs roughly once a second, meaning that userspace has
to have been idle for a couple of seconds before we enable RC6 again.
Naturally, this causes us to consume considerably more energy than
before as powersaving is effectively disabled while a display server
(here's looking at you Xorg) is running.

As execlists will get a completion event as each context is completed,
we can use this interrupt to queue a retire worker bound to this engine
to cleanup idle timelines. We will then immediately notice the idle
engine (without userspace intervention or the aid of the background
retire worker) and start parking the GPU. Thus during light workloads,
we will do much more work to idle the GPU faster...  Hopefully with
commensurate power saving!

v2: Watch context completions and only look at those local to the engine
when retiring to reduce the amount of excess work we perform.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=112315
References: 7e34f4e4aad3 ("drm/i915/gen8+: Add RC6 CTX corruption WA")
References: 2248a28384fe ("drm/i915/gen8+: Add RC6 CTX corruption WA")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191125105858.1718307-3-chris@chris-wilson.co.uk
(cherry picked from commit 4f88f8747fa43c97c3b3712d8d87295ea757cc51)
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915/gt: Adapt engine_park synchronisation rules for engine_retire

In the next patch, we will introduce a new asynchronous retirement
worker, fed by execlists CS events. Here we may queue a retirement as
soon as a request is submitted to HW (and completes instantly), and we
also want to process that retirement as early as possible and cannot
afford to postpone (as there may not be another opportunity to retire it
for a few seconds). To allow the new async retirer to run in parallel
with our submission, pull the __i915_request_queue (that passes the
request to HW) inside the timelines spinlock so that the retirement
cannot release the timeline before we have completed the submission.

v2: Actually to play nicely with engine_retire, we have to raise the
timeline.active_lock before releasing the HW. intel_gt_retire_requsts()
is still serialised by the outer lock so they cannot see this
intermediate state, and engine_retire is serialised by HW submission.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191125105858.1718307-2-chris@chris-wilson.co.uk
(cherry picked from commit 88a4655e75ac8f55eea5e3f38b176cba9cf653b5)
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915/execlists: Fixup cancel_port_requests()

I rushed a last minute correction to cancel_port_requests() to prevent
the snooping of *execlists->active as the inflight array was being
updated, without noticing we iterated the inflight array starting from
active! Oops.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=112387
Fixes: 97f9af78f38d ("drm/i915/gt: Mark the execlists->active as the primary volatile access")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Andi Shyti <andi.shyti@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191125112520.1760492-1-chris@chris-wilson.co.uk
(cherry picked from commit da0ef77e1e0ccff703efee82406c629d5c4f4bbb)
[Joonas: Fixed Fixes: tag to match drm-intel-next-fixes]
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915/gt: Mark the execlists->active as the primary volatile access

Since we want to do a lockless read of the current active request, and
that request is written to by process_csb also without serialisation, we
need to instruct gcc to take care in reading the pointer itself.

Otherwise, we have observed execlists_active() to report 0x40.

[ 2400.760381] igt/para-4098    1..s. 2376479300us : process_csb: rcs0 cs-irq head=3, tail=4
[ 2400.760826] igt/para-4098    1..s. 2376479303us : process_csb: rcs0 csb[4]: status=0x00000001:0x00000000
[ 2400.761271] igt/para-4098    1..s. 2376479306us : trace_ports: rcs0: promote { b9c59:2622, b9c55:2624 }
[ 2400.761726] igt/para-4097    0d... 2376479311us : __i915_schedule: rcs0: -2147483648->3, inflight:0000000000000040, rq:ffff888208c1e940

which is impossible!

The answer is that as we keep the existing execlists->active pointing
into the array as we copy over that array, the unserialised read may see
a partial pointer value.

Fixes: df403069029d ("drm/i915/execlists: Lift process_csb() out of the irq-off spinlock")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191125094318.1630806-1-chris@chris-wilson.co.uk
(cherry picked from commit 331bf90591573dfe6c8e892239713ef9702f1396)
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915/gt: Unlock engine-pm after queuing the kernel context switch

In commit a79ca656b648 ("drm/i915: Push the wakeref->count deferral to
the backend"), I erroneously concluded that we last modify the engine
inside __i915_request_commit() meaning that we could enable concurrent
submission for userspace as we enqueued this request. However, this
falls into a trap with other users of the engine->kernel_context waking
up and submitting their request before the idle-switch is queued, with
the result that the kernel_context is executed out-of-sequence most
likely upsetting the GPU and certainly ourselves when we try to retire
the out-of-sequence requests.

As such we need to hold onto the effective engine->kernel_context mutex
lock (via the engine pm mutex proxy) until we have finish queuing the
request to the engine.

v2: Serialise against concurrent intel_gt_retire_requests()
v3: Describe the hairy locking scheme with intel_gt_retire_requests()
for future reference.
v4: Combine timeline->lock and engine pm release; it's hairy.

Fixes: a79ca656b648 ("drm/i915: Push the wakeref->count deferral to the backend")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191120165514.3955081-2-chris@chris-wilson.co.uk
(cherry picked from commit 5cba288466e9b229feb68295675246e7522fb5eb)
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915/gt: Close race between engine_park and intel_gt_retire_requests

The general concept was that intel_timeline.active_count was locked by
the intel_timeline.mutex. The exception was for power management, where
the engine->kernel_context->timeline could be manipulated under the
global wakeref.mutex.

This was quite solid, as we always manipulated the timeline only while
we held an engine wakeref.

And then we started retiring requests outside of struct_mutex, only
using the timelines.active_list and the timeline->mutex. There we
started manipulating intel_timeline.active_count outside of an engine
wakeref, and so introduced a race between __engine_park() and
intel_gt_retire_requests(), a race that could result in the
engine->kernel_context not being added to the active timelines and so
losing requests, which caused us to keep the system permanently powered
up [and unloadable].

The race would be easy to close if we could take the engine wakeref for
the timeline before we retire -- except timelines are not bound to any
engine and so we would need to keep all active engines awake. The
alternative is to guard intel_timeline_enter/intel_timeline_exit for use
outside of the timeline->mutex.

Fixes: e5dadff4b093 ("drm/i915: Protect request retirement with timeline->mutex")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Matthew Auld <matthew.auld@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191120165514.3955081-1-chris@chris-wilson.co.uk
(cherry picked from commit a6edbca74b305adc165e67065d7ee766006e6a48)
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915: Mark up the calling context for intel_wakeref_put()

Previously, we assumed we could use mutex_trylock() within an atomic
context, falling back to a worker if contended. However, such trickery
is illegal inside interrupt context, and so we need to always use a
worker under such circumstances. As we normally are in process context,
we can typically use a plain mutex, and only defer to a work when we
know we are being called from an interrupt path.

Fixes: 51fbd8de87dc ("drm/i915/pmu: Atomically acquire the gt_pm wakeref")
References: a0855d24fc22d ("locking/mutex: Complain upon mutex API misuse in IRQ contexts")
References: https://bugs.freedesktop.org/show_bug.cgi?id=111626
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191120125433.3767149-1-chris@chris-wilson.co.uk
(cherry picked from commit 07779a76ee1f93f930cf697b22be73d16e14f50c)
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915: Wait until the intel_wakeref idle callback is complete

When waiting for idle, serialise with any ongoing callback so that it
will have completed before completing the wait.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191118230254.2615942-12-chris@chris-wilson.co.uk
(cherry picked from commit f4ba0707c825d60f1d0f5ce7bd3d875e68f3e204)
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915/gt: Fixup config ifdeffery for pm_suspend_target_state

pm_suspend_target_state is declared under CONFIG_PM_SLEEP but only
defined under CONFIG_SUSPEND. Play safe and only use the symbol if it is
both declared and defined.

Reported-by: kbuild-all@lists.01.org
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Fixes: a70a9e998e8e ("drm/i915: Defer rc6 shutdown to suspend_late")
Cc: Andi Shyti <andi.shyti@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191120182209.3967833-1-chris@chris-wilson.co.uk
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915/query: Align flavour of engine data lookup

Commit 750e76b4f9f6 ("drm/i915/gt: Move the [class][inst] lookup for
engines onto the GT") changed the engine query to iterate over uabi
engines but left the buffer size calculation look at the physical engine
count. Difference has no practical consequence but it is nicer to align
both queries.

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Fixes: 750e76b4f9f6 ("drm/i915/gt: Move the [class][inst] lookup for engines onto the GT")
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20191122104115.29610-1-tvrtko.ursulin@linux.intel.com
(cherry picked from commit 9acc99d8f278e3da398e927774431bd3e947ab2e)
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915/tgl: Add DKL PHY vswing table for HDMI

The bspec initially provided a single DKL PHY vswing table for both HDMI
and DP, but was recently updated to include an independent table for
HDMI.

Bspec: 49292
Fixes: 978c3e539be2 ("drm/i915/tgl: Add dkl phy programming sequences")
Cc: Clinton A Taylor <clinton.a.taylor@intel.com>
Cc: Lucas De Marchi <lucas.demarchi@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191118180219.9309-1-matthew.d.roper@intel.com
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
(cherry picked from commit 362bfb995b78394aefe61f7cc0511ef7c50bb11e)
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915/ehl: Update voltage level checks

The bspec was recently updated with new cdclk -> voltage level tables to
accommodate the new 324/326.4 cdclk values.

Bspec: 21809
Fixes: 63c9dae71dc5 ("drm/i915/ehl: Add voltage level requirement table")
Cc: José Roberto de Souza <jose.souza@intel.com>
Cc: Vivek Kasireddy <vivek.kasireddy@intel.com>
Cc: Bob Paauwe <bob.j.paauwe@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191118164412.26216-1-matthew.d.roper@intel.com
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
(cherry picked from commit d147483884ed08ee6bb618ef610ee0329a27fda7)
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/amdgpu: invalidate mmhub semaphore workaround in gmc9/gmc10

It may lose gpuvm invalidate acknowldege state across power-gating off
cycle. To avoid this issue in gmc9/gmc10 invalidation, add semaphore acquire
before invalidation and semaphore release after invalidation.

After adding semaphore acquire before invalidation, the semaphore
register become read-only if another process try to acquire semaphore.
Then it will not be able to release this semaphore. Then it may cause
deadlock problem. If this deadlock problem happens, it needs a semaphore
firmware fix.

Signed-off-by: changzhu <Changfeng.Zhu@amd.com>
Acked-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

drm/amdgpu: initialize vm_inv_eng0_sem for gfxhub and mmhub

SW must acquire/release one of the vm_invalidate_eng*_sem around the
invalidation req/ack. Through this way,it can avoid losing invalidate
acknowledge state across power-gating off cycle.
To use vm_invalidate_eng*_sem, it needs to initialize
vm_invalidate_eng*_sem firstly.

Signed-off-by: changzhu <Changfeng.Zhu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

drm/amd/amdgpu/sriov skip RLCG s/r list for arcturus VF.

After rlcg fw 2.1, kmd driver starts to load extra fw for
LIST_CNTL,GPM_MEM,SRM_MEM. We needs to skip the three fw
because all rlcg related fw have been loaded by host driver.
Guest driver would load the three fw fail without this change.

Signed-off-by: Jack Zhang <Jack.Zhang1@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/amdgpu/sriov temporarily skip ras,dtm,hdcp for arcturus VF

Temporarily skip ras,dtm,hdcp initialize and terminate for arcturus VF
Currently the three features haven't been enabled at SRIOV, it would
trigger guest driver load fail with the bare-metal path of the three
features.

Signed-off-by: Jack Zhang <Jack.Zhang1@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/gfx10: re-init clear state buffer after gpu reset

This patch fixes 2nd baco reset failure with gfxoff enabled on navi1x.

clear state buffer (resides in vram) is corrupted after 1st baco reset,
upon gfxoff exit, CPF gets garbage header in CSIB and hangs.

Signed-off-by: Xiaojie Yuan <xiaojie.yuan@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

merge fix for "ftrace: Rework event_create_dir()"

Reviewed-by: Kevin Wang <kevin1.wang@amd.com>
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Update Arcturus golden registers

Signed-off-by: Jay Cornwall <jay.cornwall@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/gfx10: fix out-of-bound mqd_backup array access

Fixes: 0900a9efdb7909 ("drm/amdgpu/gfx10: fix mqd backup/restore for gfx rings (v2)")
Signed-off-by: Xiaojie Yuan <xiaojie.yuan@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/gfx10: explicitly wait for cp idle after halt/unhalt

50us is not enough to wait for cp ready after gpu reset on some navi asics.

Signed-off-by: Xiaojie Yuan <xiaojie.yuan@amd.com>
Suggested-by: Jack Xiao <Jack.Xiao@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

Revert "drm/amd/display: enable S/G for RAVEN chip"

This reverts commit 1c4259159132ae4ceaf7c6db37a6cf76417f73d9.

S/G display is not stable with the IOMMU enabled on some
platforms.

Bug: https://bugzilla.kernel.org/show_bug.cgi?id=205523
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: disable gfxoff on original raven

There are still combinations of sbios and firmware that
are not stable.

Bug: https://bugzilla.kernel.org/show_bug.cgi?id=204689
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: remove experimental flag for Navi14

5.4 and newer works fine with navi14.

Reviewed-by: Xiaojie Yuan <xiaojie.yuan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: disable gfxoff when using register read interface

When gfxoff is enabled, accessing gfx registers via MMIO
can lead to a hang.

Bug: https://bugzilla.kernel.org/show_bug.cgi?id=205497
Acked-by: Xiaojie Yuan <xiaojie.yuan@amd.com>
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/powerplay: properly set PP_GFXOFF_MASK (v2)

So that the setting reflects what the hw supports. This will
be used in a subsequent patch so needs to be correct.

v2: squash in fix from Colin Ian King

Bug: https://bugzilla.kernel.org/show_bug.cgi?id=205497
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: fix bad DMA from INTERRUPT_CNTL2

The INTERRUPT_CNTL2 register expects a valid DMA address, but is
currently set with a GPU MC address.  This can cause problems on
systems that detect the resulting DMA read from an invalid address
(found on a Power8 guest).

Instead, use the DMA address of the dummy page because it will always
be safe.

Fixes: 27ae10641e9c ("drm/amdgpu: add interupt handler implementation for si v3")
Signed-off-by: Sam Bobroff <sbobroff@linux.ibm.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon: fix bad DMA from INTERRUPT_CNTL2

The INTERRUPT_CNTL2 register expects a valid DMA address, but is
currently set with a GPU MC address.  This can cause problems on
systems that detect the resulting DMA read from an invalid address
(found on a Power8 guest).

Instead, use the DMA address of the dummy page because it will always
be safe.

Fixes: d8f60cfc9345 ("drm/radeon/kms: Add support for interrupts on r6xx/r7xx chips (v3)")
Fixes: 25a857fbe973 ("drm/radeon/kms: add support for interrupts on SI")
Fixes: a59781bbe528 ("drm/radeon: add support for interrupts on CIK (v5)")
Signed-off-by: Sam Bobroff <sbobroff@linux.ibm.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fix debugfs on MST connectors

[why]
Previous patch allowed to initialize debugfs entries on both MST
and SST connectors, but MST connectors get registered much later
which exposed an issue of debugfs entries being initialized in the
same folder.

[how]
Return SST debugfs entries' initialization back to where it was.
For MST connectors we should initialize debugfs entries in connector
register function after the connector is registered.

Signed-off-by: Mikita Lipski <mikita.lipski@amd.com>
Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/nv: add asic func for fetching vbios from rom directly

Needed as a fallback if the vbios can't be fetched by other means.

Reviewed-by: Xiaojie Yuan <xiaojie.yuan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: put flush_delayed_work at first

There is one regression from 042f3d7b745cd76aa
To put flush_delayed_work after adev->shutdown = true
which will make amdgpu_ih_process not response the irq
At last, all ib ring tests will be failed just like below

[drm] amdgpu: finishing device.
[drm] Fence fallback timer expired on ring gfx
[drm] Fence fallback timer expired on ring comp_1.0.0
[drm] Fence fallback timer expired on ring comp_1.1.0
[drm] Fence fallback timer expired on ring comp_1.2.0
[drm] Fence fallback timer expired on ring comp_1.3.0
[drm] Fence fallback timer expired on ring comp_1.0.1
amdgpu 0000:00:07.0: [drm:amdgpu_ib_ring_tests [amdgpu]] *ERROR* IB test failed on comp_1.1.1 (-110).
amdgpu 0000:00:07.0: [drm:amdgpu_ib_ring_tests [amdgpu]] *ERROR* IB test failed on comp_1.2.1 (-110).
amdgpu 0000:00:07.0: [drm:amdgpu_ib_ring_tests [amdgpu]] *ERROR* IB test failed on comp_1.3.1 (-110).
amdgpu 0000:00:07.0: [drm:amdgpu_ib_ring_tests [amdgpu]] *ERROR* IB test failed on sdma0 (-110).
amdgpu 0000:00:07.0: [drm:amdgpu_ib_ring_tests [amdgpu]] *ERROR* IB test failed on sdma1 (-110).
amdgpu 0000:00:07.0: [drm:amdgpu_ib_ring_tests [amdgpu]] *ERROR* IB test failed on uvd_enc_0.0 (-110).
amdgpu 0000:00:07.0: [drm:amdgpu_ib_ring_tests [amdgpu]] *ERROR* IB test failed on vce0 (-110).
[drm:amdgpu_device_delayed_init_work_handler [amdgpu]] *ERROR* ib ring test failed (-110).

v2: replace cancel_delayed_work_sync() with flush_delayed_work()

Signed-off-by: Yintian Tao <yttao@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/vcn2.5: fix the enc loop with hw fini

Signed-off-by: Leo Liu <leo.liu@amd.com>
Reviewed-by: James Zhu <James.Zhu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>