drm/amd/display: Fix pageflip event race condition for DCN.

Commit '16f17eda8bad ("drm/amd/display: Send vblank and user
events at vsartup for DCN")' introduces a new way of pageflip
completion handling for DCN, and some trouble.

The current implementation introduces a race condition, which
can cause pageflip completion events to be sent out one vblank
too early, thereby confusing userspace and causing flicker:

prepare_flip_isr():

1. Pageflip programming takes the ddev->event_lock.
2. Sets acrtc->pflip_status == AMDGPU_FLIP_SUBMITTED
3. Releases ddev->event_lock.

--> Deadline for surface address regs double-buffering passes on
    target pipe.

4. dc_commit_updates_for_stream() MMIO programs the new pageflip
   into hw, but too late for current vblank.

=> pflip_status == AMDGPU_FLIP_SUBMITTED, but flip won't complete
   in current vblank due to missing the double-buffering deadline
   by a tiny bit.

5. VSTARTUP trigger point in vblank is reached, VSTARTUP irq fires,
   dm_dcn_crtc_high_irq() gets called.

6. Detects pflip_status == AMDGPU_FLIP_SUBMITTED and assumes the
   pageflip has been completed/will complete in this vblank and
   sends out pageflip completion event to userspace and resets
   pflip_status = AMDGPU_FLIP_NONE.

=> Flip completion event sent out one vblank too early.

This behaviour has been observed during my testing with measurement
hardware a couple of time.

The commit message says that the extra flip event code was added to
dm_dcn_crtc_high_irq() to prevent missing to send out pageflip events
in case the pflip irq doesn't fire, because the "DCH HUBP" component
is clock gated and doesn't fire pflip irqs in that state. Also that
this clock gating may happen if no planes are active. This suggests
that the problem addressed by that commit can't happen if planes
are active.

The proposed solution is therefore to only execute the extra pflip
completion code iff the count of active planes is zero and otherwise
leave pflip completion handling to the pflip irq handler, for a
more race-free experience.

Note that i don't know if this fixes the problem the original commit
tried to address, as i don't know what the test scenario was. It
does fix the observed too early pageflip events though and points
out the problem introduced.

Fixes: 16f17eda8bad ("drm/amd/display: Send vblank and user events at vsartup for DCN")
Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/[radeon|amdgpu]: Remove HAINAN board from max_sclk override check

Works stable without the overrides.

Signed-off-by: Yassine Oudjana <y.oudjana@protonmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Revert "drm/amdgpu: add CAP fw loading"

This reverts commit 29e2501f8a64fa2fa8f6fe4be53cce5a5a4fe79f.

Signed-off-by: Zhigang Luo <zhigang.luo@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/sriov : Don't resume RLCG for SRIOV guest

RLCG is enabled by host driver, no need to enable it in guest for none-PSP load path

Signed-off-by: shaoyunl <shaoyun.liu@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: protect RAS sysfs during GPU reset

MMHub EDC becomes dirty after BACO reset

EDC registers should be cleared early on in reset phase

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>

Merge tag 'mediatek-drm-next-5.7' of https://github.com/ckhu-mediatek/linux.git-tags into drm-next

Mediatek DRM Next for Linux 5.7

This include MT8183 DPI support.

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: CK Hu <ck.hu@mediatek.com>
Link: https://patchwork.freedesktop.org/patch/msgid/1584580683.29614.5.camel@mtksdaap41

Merge tag 'amd-drm-next-5.7-2020-03-19' of git://people.freedesktop.org/~agd5f/linux into drm-next

amd-drm-next-5.7-2020-03-19:

amdgpu:
- SR-IOV fixes
- RAS fixes
- Fallthrough cleanups
- Kconfig fix for ACP
- Fix load balancing with VCN
- DC fixes
- GPU reset fixes
- Various cleanups

scheduler:
- Revert job distribution optimization
- Add a helper to pick the least loaded scheduler

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Alex Deucher <alexdeucher@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200319175418.4237-1-alexander.deucher@amd.com

drm: amd: fix spelling mistake "shoudn't" -> "shouldn't"

There are spelling mistakes in pr_err messages and a comment. Fix these.

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Remove unnecessary variable shadow in gfx_v9_0_rlcg_wreg

clang warns:

drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c:754:6: warning: variable 'shadow'
is used uninitialized whenever 'if' condition is
false [-Wsometimes-uninitialized]
        if (offset == grbm_cntl || offset == grbm_idx)
            ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c:757:6: note: uninitialized use
occurs here
        if (shadow) {
            ^~~~~~
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c:754:2: note: remove the 'if' if
its condition is always true
        if (offset == grbm_cntl || offset == grbm_idx)
        ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c:738:13: note: initialize the
variable 'shadow' to silence this warning
        bool shadow;
                   ^
                    = 0
1 warning generated.

shadow is only assigned in one condition and used as the condition for
another if statement; combine the two if statements and remove shadow
to make the code cleaner and resolve this warning.

Fixes: 2e0cc4d48b91 ("drm/amdgpu: revise RLCG access path")
Link: https://github.com/ClangBuiltLinux/linux/issues/936
Suggested-by: Joe Perches <joe@perches.com>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: fix typo for vcn2.5/jpeg2.5 idle check

fix typo for vcn2.5/jpeg2.5 idle check

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

drm/amdgpu: fix typo for vcn2/jpeg2 idle check

fix typo for vcn2/jpeg2 idle check

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

drm/amdgpu: fix typo for vcn1 idle check

fix typo for vcn1 idle check

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

drm/amdgpu: add CAP fw loading

The CAP fw is for enabling driver compatibility. Currently, it only
enabled for vega10 VF.

Signed-off-by: Zhigang Luo <zhigang.luo@amd.com>
Reviewed-by: Shaoyun Liu <Shaoyun.Liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

amd/powerplay: arcturus baco reset disable all features

issue smu cmd to disable all features upon baco entry for arcturus

to mitigate potential dirty I2C controller on boot

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

drm/amdgpu: miss PRT case when bo update

Originally, only the PTE valid is taken in consider.
The PRT case is missied when bo update which raise problem.
We need add condition for PRT case.

v2: add PRT condition for amdgpu_vm_bo_update_mapping, too
v3: fix one typo error

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/amd/display: add on demand pipe merge logic for dcn2+

Adds logic that will determine if pipes need merging during validation.

Signed-off-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Reviewed-by: Chris Park <Chris.Park@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Allocate scratch space for DMUB CW7

[Why]
The scratch space can be used to pass data between x86 and DMCUB.  DMCUB
will manage the actually mapping of CW7 internally, driver does not
program the window.

[How]
Allocate extra space within the DMUB service's framebuffer for this
scratch space and expose them from the service for use in DC.

Signed-off-by: Wyatt Wood <wyatt.wood@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/amd/display: DPP DTO isn't update properly.

[Why]
before update dpp DTO, we check dppclks in context to determine it is
changed or not, but dppclks in context will be updated anyways after
flip is done, so compare dppclks in context will always get an equal
result.

[How]
Add pipe dpp clks in dccg and compare values between dccg and context.

Signed-off-by: Yongqiang Sun <yongqiang.sun@amd.com>
Reviewed-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Remove connect DIG FE to its BE during timing programming

[why]
Causes regression with MST DSC displays not lighting up after DPMS

[how]
Revert commit 8cc426d79be1c3 ("drm/amd/display: Program DSC during timing programming")

Signed-off-by: Nikola Cornij <nikola.cornij@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: combine watermark change and clock change for update clocks.

[Why]
underflow happened when playing video on 1366x768 + 4K clone mode due to
incorrect handle watermark change flag and lower down clocks to early.

[How]
Check watermark change flag when decide doing optimized, and check
optimized required flag to do clock update.

Signed-off-by: Yongqiang Sun <yongqiang.sun@amd.com>
Reviewed-by: Eric Yang <eric.yang2@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: guard DPPHY_Internal_ctrl

[why]
this register not exist in some asic, based on request remove this from
dc.

[how]
add guard for sanization.

Signed-off-by: Charlene Liu <Charlene.Liu@amd.com>
Reviewed-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: differentiate vsc sdp colorimetry use criteria between MST and SST

[Why]
We should check MST BU support capability on output port before building
vsc info packet.

[How]
Add a new definition for port and sink capability check.

Signed-off-by: Martin Tsai <martin.tsai@amd.com>
Reviewed-by: Wenjing Liu <Wenjing.Liu@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Program self refresh control register on boot

[WHY]
In headless boot cases, self refresh control registers are not
programmed on boot. In certain hybrid graphics cases this may cause
cstate entering to get blocked causing a hang.

[HOW]
Program self refresh control register on boot.

Signed-off-by: Sung Lee <sung.lee@amd.com>
Reviewed-by: Anthony Koo <Anthony.Koo@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: remove magic numbers in hdcp_ddc

[why]
DP doesn't have message id as the first byte of an hdcp message,
current hdcp psp unifies HDMI and DP message so that it is required
when reading DP HDCP messages in hdcp_ddc, a message id needs to be
added as the first byte of the HDCP message.
The id is currently assigned as a magic number which is not a good
coding practice.

[how]
Replace magic numbers with macro defined in hdcp headers.

Signed-off-by: Wenjing Liu <Wenjing.Liu@amd.com>
Reviewed-by: Ashley Thomas <Ashley.Thomas2@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fallback to dmcub for psr when dmcu is disabled

[Why]
We want to be able to enable/disable psr on dmcub and fallback to dmcu
when necessary.

[How]
Use dc config option to do so.

Signed-off-by: Wyatt Wood <wyatt.wood@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/amd/display: Set disable_dmcu flag properly per asic

[Why]
The default value for disable_dmcu is true, even for asics that require
dmcu.

[How]
Set flag properly per asic.

Signed-off-by: Wyatt Wood <wyatt.wood@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/amd/display: Remove PSR dependency on swizzle mode

[Why]
The PSR enablement was dependent on swizzle as a workaround for
non-pageflipping fb console. It's no longer required.

[How]
Remove PSR-enable dependency on swizzle mode.

Signed-off-by: Roman Li <roman.li@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/amd/display: Explicitly disable triplebuffer flips

[Why]
This is enabled by default on Renoir but there's userspace/API support
to actually make use of this.

Since we're not passing this down through surface updates, let's
explicitly disable this for now.

This fixes "dcn20_program_front_end_for_ctx" warnings associated with
incorrect/unexpected programming sequences performed while this is
enabled.

[How]
Disable it at the topmost level in DM in case anyone tries to flip this
to enabled for any of the other ASICs like Navi10/14.

Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Hersen Wu <hersenxs.wu@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Pass triplebuffer surface flip flags down to plane state

[Why]
A "dcn20_program_front_end_for_ctx" warning is observed on Renoir.

Since the resource definition doesn't explicitly disable triplebuffer
flips like Navi10 DC actually attempts to go and setup triplebuffering
even when we pass in false to the plane state.

If we hit a full update after triplebuffering has been setup we see the
assertion since we don't expect full updates while performing
triplebuffer flips.

Normally this would get reset back to false whne we pass in the new
plane state, but since we never actually copy the flag when doing
surface updates this doesn't happen.

[How]
Copy the flag onto the plane update based on the requested surface
update state.

Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Hersen Wu <hersenxs.wu@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: fix split threshold w/a to work with mpo

Right now only stream count is used to avoid split.  This change updates
the W/A to check plane count instead.

Signed-off-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Reviewed-by: Wesley Chalmers <Wesley.Chalmers@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: workaround for HDMI hotplug in DPMSOFF state

[Why]
When hotplug a HDMI monitor during entering S0i3 or DPMSOFF state due to
entering infinite loop when calling vbios to program pixel clocks. In
this scenario, pll is enabled but phy is not, and there is not a
programing guide for this case.

[How]
Before we having the proper programing guide, before disable pll, doing
a phy enable and disable to avoid the issue.

Signed-off-by: Yongqiang Sun <yongqiang.sun@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Remove redundant hdcp display state

[Why]
Due to previous code changes displays which are in active state
immediately transition to the active and added state. This makes the two
states redundant and unnecessary.

[How]
Instead of updating the device state to active and added after
successful addition, change state to inactive if addition failed. Also,
change references to active and added state to just added state.

Signed-off-by: Isabel Zhang <isabel.zhang@amd.com>
Reviewed-by: Wenjing Liu <Wenjing.Liu@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fix test pattern color space inconsistency for Linux

[why]
When reprogram MSA with updated color space, the test color space shows
inconsistency. Linux has separate routine to set up test pattern color
space, but it fails to configure RGB.

[How]
Add RGB to test pattern.

Fixes: 43563bc2e6a769 ("drm/amd/display: update MSA and VSC SDP on video test pattern request")
Signed-off-by: Jerry (Fangzhi) Zuo <Jerry.Zuo@amd.com>
Reviewed-by: Hersen Wu <hersenxs.wu@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: update connector->display_info after read edid

[Why]
drm_connector->display_info is not passed to amdgpu_dm right way after
read edid.

[How]
display_info is parsed from edid and saved into drm_connector by
drm_connector_update_edid_proerty which is called within
amdgpu_dm_update_connector_after_detect. call this function after read
edid to update drm_connector->display_info

Signed-off-by: Hersen Wu <hersenxs.wu@amd.com>
Reviewed-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Revert "DCN2.x Do not program DPPCLK if same value"

[WHY]
Not programming dto with same values causes test failures in DCN2 diags
DPP tests.

[HOW]
This reverts commit 1b53e733238c0f7faa4744ec7c8c6f193649f168.

Signed-off-by: Sung Lee <sung.lee@amd.com>
Reviewed-by: Yongqiang Sun <yongqiang.sun@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Merge tag 'exynos-drm-next-for-v5.7' of git://git./linux/kernel/git/daeinki/drm-exynos into drm-next

Two cleanups
. Replace the hand rolled encoder bitmask thing with drm_encoder_mask()
. Use mode->clock instead of reverse calculating it from the vrefresh

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Inki Dae <inki.dae@samsung.com>
Link: https://patchwork.freedesktop.org/patch/msgid/1584320957-9442-1-git-send-email-inki.dae@samsung.com

Merge tag 'drm-misc-next-2020-03-17' of git://anongit.freedesktop.org/drm/drm-misc into drm-next

drm-misc-next for 5.7:

UAPI Changes:

Cross-subsystem Changes:

Core Changes:
  - dp-mst: Remove register_connector callback, add drm_dp_destroy_connector
  - Changes to scnprintf on multiple instances

Driver Changes:
  - meson: Support for YUV420
  - panel: Support Ortustech COM37H3M, idk-1110wr and idk-2121wr,
           multiple dotclock fixes

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Maxime Ripard <maxime@cerno.tech>
Link: https://patchwork.freedesktop.org/patch/msgid/20200317082858.lubmvlmvoprn2tuh@gilmour.lan

Merge branch 'vmwgfx-next' of git://people.freedesktop.org/~thomash/linux into drm-next

- Disable DMA when using SEV encryption
- An -RT fix
- Code cleanups

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: "Thomas Hellstrom (VMware)" <thomas_os@shipmail.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200316105212.26504-1-thomas_os@shipmail.org

Merge tag 'drm-intel-next-2020-03-13' of git://anongit.freedesktop.org/drm/drm-intel into drm-next

UAPI Changes:

On i915 we have a new UAPI to allow userspace to specify CS ring buffer size on
construction (I915_CONTEXT_PARAM_RINGSIZE) and also new sysfs entries exposing
various engine properties

GVT Changes:

VFIO edid getting expanded to all platforms and a big cleanup around attr
group, unused vblank complete, kvmgt, Intel engine and dev_priv usages.

i915 Changes:

- new UAPI to allow userspace to specify CS ring buffer size on construction
  (I915_CONTEXT_PARAM_RINGSIZE) -  (Chris)
- New sysfs entries exposing various engine properties (Chris)
- Tiger Lake is out of require_force_probe protection (Jose)
- Changes in many places around active requests, reset and heartbeat (Chris)
- Stop assigning drm-dev_private pointer (Jani)
- Many code refactor in many places, including intel_modeset_init,
  increasing use of intel_uncore_*, vgpu, and gvt stuff (Jani)
- Fixes around display pipe iterators (Anshuman)
- Tigerlake enabling work (Matt Ropper, Matt Atwood, Ville, Lucas, Daniele,
  Jose, Anusha, Vivek, Swathi, Caz. Kai)
- Code clean-up like reducing use of drm/i915_drv.h, removing unused
  registers, removing garbage warns, and some other code polishing (Jani, Lucas,
  Ville)
- Selftests fixes, improvements and additions (Chris, Dan, Aditya, Matt Auld)
- Fix plane possible_crtcs bit mask (Anshuman)
- Fixes and cleanup on GLK pre production identification and w/a (Ville)
- Fix display orientation on few cases (Hans, Ville)
- dbuf clean-up and improvements for slice arrays handling (Ville)
- Improvement around min cdclk calculation (Stanislav)
- Fixes and refactor around display PLLs (Imre)
- Other execlists and perf fixes (Chris)
- Documentation fixes (Jani, Chris)
- Fix build issue (Anshuman)
- Many more fixes around the locking mechanisms (Chris)
- Other fixes and debugability info around preemption (Chris, Tvrtko)
- Add mechanism to submit a context WA on ring submission (Mika)
- Clear all Eu/L3 resitual context (Prathap)
- More changes around local memory (Abdiel, Matt, Chris)
- Fix RPS (Chris)
- DP MST fix (Lyude)
- Display FBC fixes (Jose, RK)
- debugfs cleanup (Tvrtko)
- More convertion towards drm_debive based loggin (Wambui, Ram)
- Avoid potential buffer overflow (Takashi)
- Ice Lake and Elkhart Lake workarounds (Matt Roper)

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200314001535.GA2969344@intel.com

Merge tag 'drm/tegra/for-5.7-rc1' of git://anongit.freedesktop.org/tegra/linux into drm-next

drm/tegra: Changes for v5.7-rc1

This contains some minor cleanups, nothing too exciting.

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Thierry Reding <thierry.reding@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200313171042.2924890-1-thierry.reding@gmail.com

drm/exynos: Use mode->clock instead of reverse calculating it from the vrefresh

htotal*vtotal*vrefresh ~= clock. So just use say "clock" when we mean it.

Cc: Inki Dae <inki.dae@samsung.com>
Cc: Joonyoung Shim <jy0922.shim@samsung.com>
Cc: Seung-Woo Kim <sw0312.kim@samsung.com>
Cc: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos: Use drm_encoder_mask()

Replace the hand rolled encoder bitmask thing with drm_encoder_mask()

Cc: Inki Dae <inki.dae@samsung.com>
Cc: Joonyoung Shim <jy0922.shim@samsung.com>
Cc: Seung-Woo Kim <sw0312.kim@samsung.com>
Cc: Kyungmin Park <kyungmin.park@samsung.com>
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/amdgpu: disable gpu_sched load balancer for vcn jobs

VCN HW doesn't support dynamic load balance on multiple instances
for a context. This patch initializes VNC entities with only one
drm_gpu_scheduler picked by drm_sched_pick_best(). Picking a
drm_gpu_scheduler using drm_sched_pick_best() ensures that we
do load balance among multiple contexts but not among multiple
jobs in a context.

Signed-off-by: Nirmoy Das <nirmoy.das@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/sched: implement and export drm_sched_pick_best

Remove drm_sched_entity_get_free_sched() and use the logic of picking
the least loaded drm scheduler from a drm scheduler list to implement
drm_sched_pick_best(). This patch also exports drm_sched_pick_best() so
that it can be utilized by other drm drivers.

Signed-off-by: Nirmoy Das <nirmoy.das@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Revert "drm/scheduler: improve job distribution with multiple queues"

It needs to revert this patch to avoid amdgpu_test compute hang problem
on picasso.

This reverts commit 56822db194232c089601728d68ed078dccb97f8b.

Signed-off-by: changzhu <Changfeng.Zhu@amd.com>
Reviewed-by: Feifei Xu <Feifei.Xu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Move EEPROM I2C adapter to amdgpu_device

Puts the i2c adapter in common place for sharing by RAS
and upcoming data read from FRU EEPROM feature.

v2:
Move i2c adapter to amdgpu_pm and rename it.

v3: Move i2c adapter init to ASIC specific code and get rid
of the switch case in amdgpu_device

Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm_amdgpu: Add job fence to resv conditionally

Job fence on page table should be a shared one, so add it to the root
page talbe bo resv.
last_delayed field is not needed anymore. so remove it.

Cc: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Felix Kuehling <Felix.Kuehling@amd.com>
Suggested-by: Christian König <christian.koenig@amd.com>
Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: fix switch-case indentation

Fix switch-case indentation in amdgpu_ctx_init_entity()

Signed-off-by: Nirmoy Das <nirmoy.das@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: revise RLCG access path

what changed:
1)provide new implementation interface for the rlcg access path
2)put SQ_CMD/SQ_IND_INDEX to GFX9 RLCG path to let debugfs's reg_op
function can access reg that need RLCG path help

now even debugfs's reg_op can used to dump wave.

tested-by: Monk Liu <monk.liu@amd.com>
tested-by: Zhou pengju <pengju.zhou@amd.com>
Signed-off-by: Zhou pengju <pengju.zhou@amd.com>
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Emily Deng <Emily.Deng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm: Mark up racy check of drm_gem_object.handle_count

[ 1715.899800] BUG: KCSAN: data-race in drm_gem_handle_create_tail / drm_gem_object_handle_put_unlocked
[ 1715.899838]
[ 1715.899861] write to 0xffff8881830f3604 of 4 bytes by task 7834 on cpu 1:
[ 1715.899896]  drm_gem_handle_create_tail+0x62/0x250
[ 1715.899927]  drm_gem_open_ioctl+0xc1/0x160
[ 1715.899956]  drm_ioctl_kernel+0xe4/0x120
[ 1715.899981]  drm_ioctl+0x297/0x4c7
[ 1715.900003]  ksys_ioctl+0x89/0xb0
[ 1715.900027]  __x64_sys_ioctl+0x42/0x60
[ 1715.900052]  do_syscall_64+0x6e/0x2c0
[ 1715.900079]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
[ 1715.900100]
[ 1715.900119] read to 0xffff8881830f3604 of 4 bytes by task 8137 on cpu 0:
[ 1715.900149]  drm_gem_object_handle_put_unlocked+0x31/0x130
[ 1715.900180]  drm_gem_object_release_handle+0x93/0xe0
[ 1715.900208]  drm_gem_handle_delete+0x7b/0xe0
[ 1715.900235]  drm_gem_close_ioctl+0x61/0x80
[ 1715.900264]  drm_ioctl_kernel+0xe4/0x120
[ 1715.900291]  drm_ioctl+0x297/0x4c7
[ 1715.900316]  ksys_ioctl+0x89/0xb0
[ 1715.900340]  __x64_sys_ioctl+0x42/0x60
[ 1715.900363]  do_syscall_64+0x6e/0x2c0
[ 1715.900388]  entry_SYSCALL_64_after_hwframe+0x44/0xa9

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20200309120151.7675-1-chris@chris-wilson.co.uk

drm/mm: Allow drm_mm_initialized() to be used outside of the locks

Mark up the potential racy read in drm_mm_initialized(), as we want a
cheap and cheerful check:

[  121.098731] BUG: KCSAN: data-race in _i915_gem_object_create_stolen [i915] / rm_hole
[  121.098766]
[  121.098789] write (marked) to 0xffff8881f01ed330 of 8 bytes by task 3568 on cpu 3:
[  121.098831]  rm_hole+0x64/0x140
[  121.098860]  drm_mm_insert_node_in_range+0x3d3/0x6c0
[  121.099254]  i915_gem_stolen_insert_node_in_range+0x91/0xe0 [i915]
[  121.099646]  _i915_gem_object_create_stolen+0x9d/0x100 [i915]
[  121.100047]  i915_gem_object_create_region+0x7a/0xa0 [i915]
[  121.100451]  i915_gem_object_create_stolen+0x33/0x50 [i915]
[  121.100849]  intel_engine_create_ring+0x1af/0x280 [i915]
[  121.101242]  __execlists_context_alloc+0xce/0x3d0 [i915]
[  121.101635]  execlists_context_alloc+0x25/0x40 [i915]
[  121.102030]  intel_context_alloc_state+0xb6/0xf0 [i915]
[  121.102420]  __intel_context_do_pin+0x1ff/0x220 [i915]
[  121.102815]  i915_gem_do_execbuffer+0x46b4/0x4c20 [i915]
[  121.103211]  i915_gem_execbuffer2_ioctl+0x2c3/0x580 [i915]
[  121.103244]  drm_ioctl_kernel+0xe4/0x120
[  121.103269]  drm_ioctl+0x297/0x4c7
[  121.103296]  ksys_ioctl+0x89/0xb0
[  121.103321]  __x64_sys_ioctl+0x42/0x60
[  121.103349]  do_syscall_64+0x6e/0x2c0
[  121.103377]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
[  121.103403]
[  121.103426] read to 0xffff8881f01ed330 of 8 bytes by task 3109 on cpu 1:
[  121.103819]  _i915_gem_object_create_stolen+0x30/0x100 [i915]
[  121.104228]  i915_gem_object_create_region+0x7a/0xa0 [i915]
[  121.104631]  i915_gem_object_create_stolen+0x33/0x50 [i915]
[  121.105025]  intel_engine_create_ring+0x1af/0x280 [i915]
[  121.105420]  __execlists_context_alloc+0xce/0x3d0 [i915]
[  121.105818]  execlists_context_alloc+0x25/0x40 [i915]
[  121.106202]  intel_context_alloc_state+0xb6/0xf0 [i915]
[  121.106595]  __intel_context_do_pin+0x1ff/0x220 [i915]
[  121.106985]  i915_gem_do_execbuffer+0x46b4/0x4c20 [i915]
[  121.107375]  i915_gem_execbuffer2_ioctl+0x2c3/0x580 [i915]
[  121.107409]  drm_ioctl_kernel+0xe4/0x120
[  121.107437]  drm_ioctl+0x297/0x4c7
[  121.107464]  ksys_ioctl+0x89/0xb0
[  121.107489]  __x64_sys_ioctl+0x42/0x60
[  121.107511]  do_syscall_64+0x6e/0x2c0
[  121.107535]  entry_SYSCALL_64_after_hwframe+0x44/0xa9

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20200309121529.16497-1-chris@chris-wilson.co.uk

drm/edid: Distribute switch variables for initialization

Variables declared in a switch statement before any case statements
cannot be automatically initialized with compiler instrumentation (as
they are not part of any execution flow). With GCC's proposed automatic
stack variable initialization feature, this triggers a warning (and they
don't get initialized). Clang's automatic stack variable initialization
(via CONFIG_INIT_STACK_ALL=y) doesn't throw a warning, but it also
doesn't initialize such variables[1]. Note that these warnings (or silent
skipping) happen before the dead-store elimination optimization phase,
so even when the automatic initializations are later elided in favor of
direct initializations, the warnings remain.

To avoid these problems, lift such variables up into the next code
block.

drivers/gpu/drm/drm_edid.c: In function ‘drm_edid_to_eld’:
drivers/gpu/drm/drm_edid.c:4395:9: warning: statement will never be
executed [-Wswitch-unreachable]
 4395 |     int sad_count;
      |         ^~~~~~~~~

[1] https://bugs.llvm.org/show_bug.cgi?id=44916

v2: move into function block instead being switch-local (Ville Syrjälä)

Signed-off-by: Kees Cook <keescook@chromium.org>
[danvet: keep the changelog]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/202003060930.DDCCB6659@keescook

drm/vmwgfx: Replace zero-length array with flexible-array member

The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:

struct foo {
        int stuff;
        struct boo array[];
};

By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.

Also, notice that, dynamic memory allocations won't be affected by
this change:

"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]

This issue was found with the help of Coccinelle.

[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit 76497732932f ("cxgb3/l2t: Fix undefined behaviour")

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>

drm: lock: Clean up documentation

Fix kernel doc comments to avoid warnings when compiling with W=1.

Signed-off-by: Benjamin Gaignard <benjamin.gaignard@st.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20200306102937.4932-2-benjamin.gaignard@st.com

drm: bufs: Clean up documentation

Fix kernel doc comments to avoid warnings when compiling with W=1.

Signed-off-by: Benjamin Gaignard <benjamin.gaignard@st.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20200306102937.4932-3-benjamin.gaignard@st.com

drm: vm: Clean up documentation

Fix kernel doc comments to avoid warnings when compiling with W=1.

Signed-off-by: Benjamin Gaignard <benjamin.gaignard@st.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20200306102937.4932-1-benjamin.gaignard@st.com

drm: context: Clean up documentation

Fix kernel doc comments to avoid warnings when compiling with W=1.

Signed-off-by: Benjamin Gaignard <benjamin.gaignard@st.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20200306102937.4932-4-benjamin.gaignard@st.com

dt-bindings: display: Add idk-1110wr binding

Add binding for the idk-1110wr LVDS panel from Advantech.

Some panel-specific documentation can be found here:
https://buy.advantech.eu/Displays/Embedded-LCD-Kits-LCD-Kit-Modules/model-IDK-1110WR-55WSA1E.htm

Signed-off-by: Fabrizio Castro <fabrizio.castro@bp.renesas.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/1583957020-16359-2-git-send-email-prabhakar.mahadev-lad.rj@bp.renesas.com

drm/tiny: fix sparse warning: incorrect type in assignment (different base types)

This fixes the following sparse warning:

drivers/gpu/drm/tiny/ili9486.c:61:16: sparse: sparse: incorrect type in assignment (different base types)
drivers/gpu/drm/tiny/ili9486.c:61:16: sparse:    expected unsigned short [usertype]
drivers/gpu/drm/tiny/ili9486.c:61:16: sparse:    got restricted __be16 [usertype]
drivers/gpu/drm/tiny/ili9486.c:71:32: sparse: sparse: incorrect type in assignment (different base types)
drivers/gpu/drm/tiny/ili9486.c:71:32: sparse:    expected unsigned short [usertype]
drivers/gpu/drm/tiny/ili9486.c:71:32: sparse:    got restricted __be16 [usertype]

Reported-by: kbuild test robot <lkp@intel.com>
Signed-off-by: Kamlesh Gurudasani <kamlesh.gurudasani@gmail.com>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/1583684084-4694-1-git-send-email-kamlesh.gurudasani@gmail.com

drm/i915: Update DRIVER_DATE to 20200313

Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

dt-bindings: display: fix panel warnings

Fix following type af warnings in the panel bindings:

Warning (unit_address_vs_reg): /example-0/dsi/panel: node has a reg or ranges property, but no unit name
Warning (unit_address_vs_reg): /example-0/dsi@ff450000: node has a unit name, but no reg property

Removing the "@xxx" from the node name fixed first warning.
Adding a missing reg property fixed the second warning

v2:
  - renamed mdss_dsi to dsi in panel-simple-dsi.yaml (Rob)

Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Benjamin Gaignard <benjamin.gaignard@st.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Cc: Thierry Reding <thierry.reding@gmail.com>
Cc: Linus Walleij <linus.walleij@linaro.org>
Cc: Rob Herring <robh@kernel.org>
Cc: Heiko Stuebner <heiko.stuebner@theobroma-systems.com>
Cc: Maxime Ripard <mripard@kernel.org>
Cc: Benjamin Gaignard <benjamin.gaignard@st.com>
Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200308115017.18563-2-sam@ravnborg.org

drm/i915/tgl: Remove require_force_probe protection

We have a few TGL machines in our CI and it is mostly green with
failures in tests that will not impact future Linux installations.
Also there is no warnings, errors, flickering or any visual defects
while doing ordinary tasks like browsing and editing documents in a
dual monitor setup.

As a reminder i915.require_force_probe was created to protect
future Linux installation's iso images that might contain a
kernel from the enabling time of the new platform. Without this
protection most of linux installation was recommending
nomodeset option during installation that was getting stick
there after installation.

Reference: https://intel-gfx-ci.01.org/tree/drm-tip/fi-tgl-u.html
Reference: https://intel-gfx-ci.01.org/tree/drm-tip/shard-tglb.html
Cc: James Ausmus <james.ausmus@intel.com>
Cc: Jani Saarinen <jani.saarinen@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Jani Nikula <jani.nikula@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Acked-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200218230822.66801-1-jose.souza@intel.com

drm/dp_mst: Convert drm_dp_mst_topology_mgr.is_waiting_for_dwn_reply to bitfield

Small nitpick that I noticed a second ago - we can save some space in
the struct by making this a bitfield and sticking it with the rest of
the bitfields. Also, some small cleanup to the kdocs for this member.

There should be no functional changes in this patch.

Signed-off-by: Lyude Paul <lyude@redhat.com>
Cc: Wayne Lin <Wayne.Lin@amd.com>
Reviewed-by: Wayne Lin <Wayne.Lin@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200122194846.16025-1-lyude@redhat.com

drm/tegra: hdmi: Silence deferred-probe error

Driver fails to probe with -EPROBE_DEFER, which produces a bit noisy error
message in KMSG during kernel's boot up. This happens because voltage
regulators tend to be probed later than the DRM driver.

Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>

drm/tegra: dc: Silence RGB output deferred-probe error

Driver fails to probe with -EPROBE_DEFER if display output isn't ready
yet. This produces a bit noisy error message in KMSG during kernel's boot
up on Tegra20 and Tegra30 because RGB output tends to be probed earlier
than a corresponding voltage regulator driver.

Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>

drm/i915: Add Wa_1605460711 / Wa_1408767742 to ICL and EHL

This workaround appears under two different numbers (and with somewhat
confused stepping applicability on ICL).  Ultimately it appears we
should just implement this for all stepping of ICL and EHL.

Note that this is identical to Wa_1407928979:tgl that already exists in
our driver too...yet another number referencing the same actual
workaround.

Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200311162300.1838847-7-matthew.d.roper@intel.com
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>

drm/i915: Apply Wa_1406680159:icl,ehl as an engine workaround

The register this workaround updates is a render engine register in the
MCR range, so we should initialize this in rcs_engine_wa_init() rather
than gt_wa_init().

Closes: https://gitlab.freedesktop.org/drm/intel/issues/1222
Fixes: 36204d80bacb ("drm/i915/icl: Wa_1406680159")
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200311162300.1838847-6-matthew.d.roper@intel.com
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>

drm/i915: Add Wa_1406306137:icl,ehl

v2:
 - Move to context workarounds.  ROW_CHICKEN4 is part of the context
   image on gen11 (although it isn't on gen12).

Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200311162300.1838847-5-matthew.d.roper@intel.com
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>

drm/i915: Add Wa_1604278689:icl,ehl

The bspec description for this workaround tells us to program
0xFFFF_FFFF into both FBC_RT_BASE_ADDR_REGISTER_* registers, but we've
previously found that this leads to failures in CI.  Our suspicion is
that the failures are caused by this valid turning on the "address valid
bit" even though we're intentionally supplying an invalid address.
Experimentation has shown that setting all bits _except_ for the
RT_VALID bit seems to avoid these failures.

v2:
 - Mask off the RT_VALID bit.  Experimentation with CI trybot indicates
   that this is necessary to avoid reset failures on BCS.

v3:
 - Program RT_BASE before RT_BASE_UPPER so that the valid bit is turned
   off by the first write.  (Chris)

Bspec: 11388
Bspec: 33451
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20200311162300.1838847-4-matthew.d.roper@intel.com
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>

drm/i915: Add Wa_1209644611:icl,ehl

On gen11 the XY_FAST_COPY_BLT command has some size restrictions on its
usage.  Although this instruction is mainly used by userspace, i915 also
uses it to copy object contents during some selftests, so let's ensure
the restrictions are followed.

Bspec: 6544
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200311162300.1838847-3-matthew.d.roper@intel.com
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>

drm/i915: Handle all MCR ranges

The bspec documents multiple MCR ranges; make sure they're all captured
by the driver.

Bspec: 13991, 52079
Fixes: 592a7c5e082e ("drm/i915: Extend non readable mcr range")
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200311162300.1838847-2-matthew.d.roper@intel.com
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>

drm/scheduler: fix inconsistent locking of job_list_lock

1db8c142b6c5 (drm/scheduler: Add drm_sched_suspend/resume_timeout()) made
the job_list_lock IRQ safe in as the suspend/resume calls were expected to
be called from IRQ context. This usage never materialized in upstream.
Instead amdgpu started locking the job_list_lock in an IRQ unsafe way in
amdgpu_ib_preempt_mark_partial_job() and amdgpu_ib_preempt_job_recovery(),
which leads to potential deadlock if one would actually start to call the
drm_sched_suspend/resume_timeout functions from IRQ context.

As no current user needs the locking to be IRQ safe, the local IRQ
disable/enable is pure overhead. Fix the inconsistent locking by changing
all uses of job_list_lock to use the IRQ unsafe locking primitives.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/sched: add run job trace

Add a new trace event to show when jobs are run on the HW.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Robert Beckett <bob.beckett@collabora.com>
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/swsmu: clean up unused header in swsmu

clean up unused header in swsmu driver stack:
1. pp_debug.h
2. amd_pcie.h
3. soc15_common.h

Signed-off-by: Kevin Wang <kevin1.wang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add codes to clear AccVGPR for arcturus

AccVGPRs are newly added in arcturus. Before reading these
registers, they should be initialized. Otherwise edc error
happens, when RAS is enabled.

v2: reuse the existing logical to calculate register size

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

drm/amd/display: Add link_rate quirk for Apple 15" MBP 2017

This fixes a problem found on the MacBookPro 2017 Retina panel:

The panel reports 10 bpc color depth in its EDID, and the
firmware chooses link settings at boot which support enough
bandwidth for 10 bpc (324000 kbit/sec aka LINK_RATE_RBR2
aka 0xc), but the DP_MAX_LINK_RATE dpcd register only reports
2.7 Gbps (multiplier value 0xa) as possible, in direct
contradiction of what the firmware successfully set up.

This restricts the panel to 8 bpc, not providing the full
color depth of the panel on Linux <= 5.5. Additionally, commit
'4a8ca46bae8a ("drm/amd/display: Default max bpc to 16 for eDP")'
introduced into Linux 5.6-rc1 will unclamp panel depth to
its full 10 bpc, thereby requiring a eDP bandwidth for all
modes that exceeds the bandwidth available and causes all modes
to fail validation -> No modes for the laptop panel -> failure
to set any mode -> Panel goes dark.

This patch adds a quirk specific to the MBP 2017 15" Retina
panel to override reported max link rate to the correct maximum
of 0xc = LINK_RATE_RBR2 to fix the darkness and reduced display
precision.

Please apply for Linux 5.6+ to avoid regressing Apple MBP panel
support.

Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Stop using the DRIVER debugging flag for vblank debugging messages

These are some very loud debug statements that get printed on every
vblank when driver level debug printing is enabled in DRM, and doesn't
really tell us anything that isn't related to vblanks. So let's move
this over to the proper debug flag to be a little less spammy with our
debug output.

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Lyude Paul <lyude@redhat.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Possible divide by zero in set_speed()

If "speed" is zero then we use it as a divisor to find "prescale".  It's
better to move the check for zero to the very start of the function.

Fixes: 9eeec26a1339 ("drm/amd/display: Refine i2c frequency calculating sequence")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/display: clean up some indenting

These lines were accidentally indented 4 spaces more than they should
be.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: clean up a condition in dmub_psr_copy_settings()

We can remove the NULL check for "res_ctx" and
"res_ctx->pipe_ctx[i].stream->link".  Also it's nicer to align the
conditions using spaces so I re-indented a bit.

Longer explanation: The "res_ctx" pointer points to an address in the
middle of a struct so it can't be NULL.  For
"res_ctx->pipe_ctx[i].stream->link" we know that it is equal to "link"
and "link" is non-NULL.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm: amd/acp: fix broken menu structure

Fix the Kconfig dependencies so that the menu is presented
correctly by adding a dependency on DRM_AMDGPU to the "menu"
Kconfig statement.  This makes a continuous dependency on
DRM_AMDGPU in the DRM AMD menus and eliminates a broken menu
structure.

Fixes: a8fe58cec351 ("drm/amd: add ACP driver support")
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: David (ChunMing) Zhou <David1.Zhou@amd.com>
Cc: Maruthi Bayyavarapu <maruthi.bayyavarapu@amd.com>
Cc: amd-gfx@lists.freedesktop.org
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

AMD KFD: Use fallthrough;

Convert the various uses of fallthrough comments to fallthrough;

Done via script
Link: https://lore.kernel.org/lkml/b56602fcf79f849e733e7b521bb0e17895d390fa.1582230379.git.joe@perches.com/
Acked-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/powerplay: Move fallthrough; into containing #ifdef/#endif

The automated conversion of /* fallthrough */ comments converted
a comment outside of an #ifdef/#endif case block that should be
inside the block.

Move the fallthrough inside the block to silence the warning.

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

AMD POWERPLAY: Use fallthrough;

Convert the various uses of fallthrough comments to fallthrough;

Done via script
Link: https://lore.kernel.org/lkml/b56602fcf79f849e733e7b521bb0e17895d390fa.1582230379.git.joe@perches.com/
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

AMD DISPLAY CORE: Use fallthrough;

Convert the various uses of fallthrough comments to fallthrough;

Done via script
Link: https://lore.kernel.org/lkml/b56602fcf79f849e733e7b521bb0e17895d390fa.1582230379.git.joe@perches.com/
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: fix warning in ras_debugfs_create_all()

Fix the warning
"warn: variable dereferenced before check 'obj' (see line 1131)"
by removing unnecessary checks as amdgpu_ras_debugfs_create_all()
is only called from amdgpu_debugfs_init() where obj member in
con->head list is not NULL.
Use list_for_each_entry() instead list_for_each_entry_safe() as obj
do not to be freeing or removing from list during this process.

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

drm/amdgpu: add fbdev suspend/resume on gpu reset

This can fix the baco reset failure seen on Navi10.
And this should be a low risk fix as the same sequence
is already used for system suspend/resume.

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

drm/amdgpu: update ras capability's query based on mem ecc configuration

RAS support capability needs to be updated on top of different
memeory ECC enablement, and remove redundant memory ecc check
in gmc module for vega20 and arcturus.

v2: check HBM ECC enablement and set ras mask accordingly.
v3: avoid to invoke atomfirmware interface to query twice.

Suggested-by: Hawking Zhang <Hawking.Zhang@amd.com>
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>

drm/amd/amdgpu: Fix GPR read from debugfs (v2)

The offset into the array was specified in bytes but should
be in terms of 32-bit words.  Also prevent large reads that
would also cause a buffer overread.

v2:  Read from correct offset from internal storage buffer.

Signed-off-by: Tom St Denis <tom.stdenis@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>

drm/amd/display: fix typos for dcn20_funcs and dcn21_funcs struct

In dcn20_funcs and dcn21_funcs struct, the member ".dsc_pg_control = NULL"
should be removed due to .dsc_pg_control be assigned to dcn20_dsc_pg_control.

Signed-off-by: Stanley.Yang <Stanley.Yang@amd.com>
Reviewed-by: Anthony Koo <Anthony.Koo@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: use amdgpu_ras.h in amdgpu_debugfs.c

include amdgpu_ras.h head file instead of use extern
ras_debugfs_create_all function

Signed-off-by: Stanley.Yang <Stanley.Yang@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Guchun Chen <guchun.chen@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: check GFX RAS capability before reset counters

disallow the logical to be enabled on platforms that
don't support gfx ras at this stage, like sriov skus,
dgpu with legacy ras.etc

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

drm/amdgpu: resolve failed error inject msg

invoking an error injection successfully will cause an at_event intterrupt that

will occur before the invoke sequence can complete causing an invalid error

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/sriov refine vcn_v2_5_early_init func

refine the assignment for vcn.num_vcn_inst,
vcn.harvest_config, vcn.num_enc_rings in VF

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/i915/selftest: Add more poison patterns

Throw in the inverse patterns to create more examples of poison to use
against the LRC state.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200313102812.30173-1-chris@chris-wilson.co.uk

Merge tag 'amd-drm-next-5.7-2020-03-10' of git://people.freedesktop.org/~agd5f/linux into drm-next

amd-drm-next-5.7-2020-03-10:

amdgpu:
- SR-IOV fixes
- Fix up fallout from drm load/unload callback removal
- Navi, renoir power management watermark fixes
- Refactor smu parameter handling
- Display FEC fixes
- Display DCC fixes
- HDCP fixes
- Add support for USB-C PD firmware updates
- Pollock detection fix
- Rework compute ring priority handling
- RAS fixes
- Misc cleanups

amdkfd:
- Consolidate more gfx config details in amdgpu
- Consolidate bo alloc flags
- Improve code comments
- SDMA MQD fixes
- Misc cleanups

gpu scheduler:
- Add suport for modifying the sched list

uapi:
- Clarify comments about GEM_CREATE flags that are not used by userspace.
  The kernel driver has always prevented userspace from using these.
  They are only used internally in the kernel driver.

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Alex Deucher <alexdeucher@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200310212748.4519-1-alexander.deucher@amd.com

Revert "drm/i915/tgl: Add extra hdc flush workaround"

This reverts commit 36a6b5d964d995b536b1925ec42052ee40ba92c4.

The commit takes care Wa_1604544889 which was fixed on a0 stepping based on
a0 replan. So no SW workaround is required on any stepping now.

Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Signed-off-by: Caz Yokoyama <caz.yokoyama@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Fixes: 36a6b5d964d9 ("drm/i915/tgl: Add extra hdc flush workaround")
Link: https://patchwork.freedesktop.org/patch/msgid/1c751032ce79c80c5485cae315f1a9904ce07cac.1583359940.git.caz.yokoyama@intel.com

drm/i915/gt: Wait for RCUs frees before asserting idle on unload

During driver unload, we have many asserts that we have released our
bookkeeping structs and are idle. In some cases, these struct are
protected by RCU and we do not release them until after an RCU grace
period.

Reported-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Fixes: 130a95e9098e ("drm/i915/gem: Consolidate ctx->engines[] release")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200312115307.16460-1-chris@chris-wilson.co.uk

drm/i915/selftests: Use igt_random_offset()

Switch igt_vm_isolation() to using igt_random_offset().

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Matthew Auld <matthew.auld@intel.com>
Cc: Akeem G Abodunrin <akeem.g.abodunrin@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200312154708.1720-1-chris@chris-wilson.co.uk

drm/i915/gem: Drop relocation slowpath

Since the relocations are no longer performed under a global
struct_mutex, or any other lock, that is also held by pagefault handlers,
we can relax and allow our fast path to take a fault. As we no longer
need to abort the fast path for lock avoidance, we no longer need the
slow path handling at all.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200311160310.26711-1-chris@chris-wilson.co.uk