Robin Murphy [Tue, 5 Apr 2022 14:17:48 +0000 (15:17 +0100)]
drm/msm: Stop using iommu_present()
Even if some IOMMU has registered itself on the platform "bus", that
doesn't necessarily mean it provides translation for the device we
care about. Replace iommu_present() with a more appropriate check.
Xiaoke Wang [Thu, 7 Apr 2022 02:31:51 +0000 (10:31 +0800)]
drm/msm/mdp5: check the return of kzalloc()
kzalloc() is a memory allocation function which can return NULL when
some internal memory errors happen. So it is better to check it to
prevent potential wrong memory access.
Besides, since mdp5_plane_reset() is void type, so we should better
set `plane-state` to NULL after releasing it.
Signed-off-by: Xiaoke Wang <xkernel.wang@foxmail.com> Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/481055/ Link: https://lore.kernel.org/r/tencent_8E2A1C78140EE1784AB2FF4B2088CC0AB908@qq.com Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org> Signed-off-by: Rob Clark <robdclark@chromium.org>
Rob Clark [Thu, 7 Apr 2022 20:28:33 +0000 (13:28 -0700)]
drm/msm: Fix range size vs end confusion
The fourth param is size, rather than range_end.
Note that we could increase the address space size if we had a way to
prevent buffers from spanning a 4G split, mostly just to avoid fw bugs
with 64b math.
Marijn Suijten [Sat, 26 Feb 2022 19:46:32 +0000 (20:46 +0100)]
drm/msm/dpu: Use indexed array initializer to prevent mismatches
While there's a comment pointing from dpu_intr_set to dpu_hw_intr_reg
and vice-versa, an array initializer using indices makes it so that the
indices between the enum and array cannot possibly get out of sync even
if they're accidentially ordered wrongly. It is still useful to keep
the comment to be made aware where the register offset mapping resides
while looking at dpu_hw_intr_reg.
Xiaoke Wang [Thu, 24 Mar 2022 09:15:08 +0000 (17:15 +0800)]
drm/msm/disp: check the return value of kzalloc()
kzalloc() is a memory allocation function which can return NULL when
some internal memory errors happen. So it is better to check it to
prevent potential wrong memory access.
Dmitry Baryshkov [Thu, 24 Mar 2022 11:55:36 +0000 (14:55 +0300)]
dt-bindings: display/msm: another fix for the dpu-qcm2290 example
Make dpu-qcm2290 example really follow the defined schema:
- Drop qcom,mdss compatible. It's only used for MDP5 devices.
- Change display controller name to display-controller as specified in
the yaml
Reported-by: Rob Herring <robh@kernel.org> Cc: Loic Poulain <loic.poulain@linaro.org> Fixes: 80ee6f777960 ("dt-bindings: msm: disp: add yaml schemas for QCM2290 DPU bindings") Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org> Reviewed-by: Krzysztof Kozlowski <krzk@kernel.org> Link: https://lore.kernel.org/r/20220324115536.2090818-1-dmitry.baryshkov@linaro.org Signed-off-by: Rob Clark <robdclark@chromium.org>
Rob Clark [Thu, 10 Mar 2022 23:46:06 +0000 (15:46 -0800)]
drm/msm/gpu: Remove mutex from wait_event condition
The mutex wasn't really protecting anything before. Before the previous
patch we could still be racing with the scheduler's kthread, as that is
not necessarily frozen yet. Now that we've parked the sched threads,
the only race is with jobs retiring, and that is harmless, ie.
Rob Clark [Thu, 10 Mar 2022 23:46:05 +0000 (15:46 -0800)]
drm/msm/gpu: Park scheduler threads for system suspend
In the system suspend path, we don't want to be racing with the
scheduler kthreads pushing additional queued up jobs to the hw
queue (ringbuffer). So park them first. While we are at it,
move the wait for active jobs to complete into the new system-
suspend path.
Rob Clark [Tue, 8 Mar 2022 18:48:44 +0000 (10:48 -0800)]
drm/msm/gpu: Fix crash on devices without devfreq support (v2)
Avoid going down devfreq paths on devices where devfreq is not
initialized.
v2: Change has_devfreq() logic [Dmitry]
Reported-by: Linux Kernel Functional Testing <lkft@linaro.org> Reported-by: Anders Roxell <anders.roxell@linaro.org> Signed-off-by: Rob Clark <robdclark@chromium.org> Fixes: 714332d909e1 ("drm/msm/gpu: Cancel idle/boost work on suspend") Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org> Link: https://lore.kernel.org/r/20220308184844.1121029-1-robdclark@gmail.com
Dan Carpenter [Mon, 7 Mar 2022 13:31:05 +0000 (16:31 +0300)]
drm/msm/adreno: fix cast in adreno_get_param()
These casts need to happen before the shift. The only time it would
matter would be if "rev.core" is >= 128. In that case the sign bit
would be extended and we do not want that.
Fixes: 9c31d0f7efdb ("drm/msm/adreno: Expose speedbin to userspace") Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Reviewed-by: Akhil P Oommen <quic_akhilpo@quicinc.com> Link: https://lore.kernel.org/r/20220307133105.GA17534@kili Signed-off-by: Rob Clark <robdclark@chromium.org>
Rob Clark [Fri, 4 Mar 2022 20:21:45 +0000 (12:21 -0800)]
drm/msm: Fix dirtyfb refcounting
refcount_t complains about 0->1 transitions, which isn't *quite* what we
wanted. So use dirtyfb==1 to mean that the fb is not connected to any
output that requires dirtyfb flushing, so that we can keep the underflow
and overflow checking.
Rob Clark [Thu, 24 Feb 2022 22:23:21 +0000 (14:23 -0800)]
drm/msm: Add MSM_SUBMIT_FENCE_SN_IN
Add a way for userspace to specify the sequence number fence used to
track completion of the submit. As the seqno fence is simply an
incrementing counter which is local to the submitqueue, it is easy for
userspace to know the next value.
This is useful for native userspace drivers in a vm guest, as the guest
to host roundtrip can have high latency. Assigning the fence seqno in
the guest userspace allows the guest to continue without waiting for
response from the host.
Rob Clark [Fri, 4 Mar 2022 00:52:17 +0000 (16:52 -0800)]
drm/msm/a6xx: Zap counters across context switch
Any app controlled perfcntr collection (GL_AMD_performance_monitor, etc)
does not require counters to maintain state across context switches. So
clear them if systemwide profiling is not active.
Rob Clark [Fri, 4 Mar 2022 00:52:16 +0000 (16:52 -0800)]
drm/msm: Add SYSPROF param (v2)
Add a SYSPROF param for system profiling tools like Mesa's pps-producer
(perfetto) to control behavior related to system-wide performance
counter collection. In particular, for profiling, one wants to ensure
that GPU context switches do not effect perfcounter state, and might
want to suppress suspend (which would cause counters to lose state).
v2: Swap the order in msm_file_private_set_sysprof() [sboyd] and
initialize the sysprof_active refcount to one (because the under/
overflow checking in refcount_t doesn't expect a 0->1 transition)
meaning that values greater than 1 means sysprof is active.
Rob Clark [Fri, 4 Mar 2022 00:52:15 +0000 (16:52 -0800)]
drm/msm: Add SET_PARAM ioctl
It was always expected to have a use for this some day, so we left a
placeholder. Now we do. (And I expect another use in the not too
distant future when we start allowing userspace to allocate GPU iova.)
Rob Clark [Wed, 23 Feb 2022 19:11:08 +0000 (11:11 -0800)]
drm/msm: Avoid dirtyfb stalls on video mode displays (v2)
Someone on IRC once asked an innocent enough sounding question: Why
with xf86-video-modesetting is es2gears limited at 120fps.
So I broke out the perfetto tracing mesa MR and took a look. It turns
out the problem was drm_atomic_helper_dirtyfb(), which would end up
waiting for vblank.. es2gears would rapidly push two frames to Xorg,
which would blit them to screen and in idle hook (I assume) call the
DIRTYFB ioctl. Which in turn would do an atomic update to flush the
dirty rects, which would stall until the next vblank. And then the
whole process would repeat.
But this is a bit silly, we only need dirtyfb for command mode DSI
panels. So track in plane state whether dirtyfb is required, and
track in the fb how many attached planes require dirtyfb so that we
can skip it when not required. (Note, mdp4 does not actually have
cmd mode support.)
Dmitry Baryshkov [Tue, 22 Feb 2022 03:19:48 +0000 (06:19 +0300)]
drm/msm/dpu: wire up MSM8998's DSPP blocks
The commit adding msm8998 support didn't added msm8998's DSPP blocks
configuration, but did not use them in msm8998_cfg_init(). Wire them up
to be used for display post processing.
Reported-by: kernel test robot <lkp@intel.com> Fixes: 1a87bbb7786b ("drm/msm/dpu1: Add MSM8998 to hw catalog") Cc: AngeloGioacchino Del Regno <angelogioacchino.delregno@somainline.org> Cc: Jami Kettunen <jami.kettunen@somainline.org> Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/475544/ Link: https://lore.kernel.org/r/20220222031948.228727-1-dmitry.baryshkov@linaro.org Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Loic Poulain [Mon, 14 Feb 2022 15:37:39 +0000 (16:37 +0100)]
drm/msm/dsi: Add support for qcm2290 dsi controller
QCM2290 MDSS includes a Qualcomm DSI controller v2.4.1. Since this
controller version is not SoC specific, and already assigned to sc7180
for auto configuration, we rely on DSI block specific compatible
string "qcom,dsi-ctrl-6g-qcm2290", and use the device's data to point
to the right dsi config handler.
Loic Poulain [Mon, 14 Feb 2022 15:37:38 +0000 (16:37 +0100)]
drm/msm/dsi: Allow to specify dsi config as pdata
Config autodetect based on DSI controller version is quite limited
since several qcom SoCs can integrate a DSI controller with the same
version, but with different config (io_offset, supplies, etc).
This change allows to specify dsi config via device data pointer.
config autodetect is still used in case data pointer is NULL.
Rob Clark [Tue, 1 Feb 2022 16:16:12 +0000 (08:16 -0800)]
drm/msm/gpu: Track global faults per address-space
Other processes don't need to know about faults that they are isolated
from by virtue of address space isolation. They are only interested in
whether some of their state might have been corrupted.
But to be safe, also track unattributed faults. This case should really
never happen unless there is a kernel bug (and that would never happen,
right?)
v2: Instead of adding a new param, just change the behavior of the
existing param to match what userspace actually wants [anholt]
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/5934 Signed-off-by: Rob Clark <robdclark@chromium.org> Link: https://lore.kernel.org/r/20220201161618.778455-3-robdclark@gmail.com Reviewed-by: Emma Anholt <emma@anholt.net> Signed-off-by: Rob Clark <robdclark@chromium.org>
Dmitry Baryshkov [Fri, 11 Feb 2022 22:40:03 +0000 (01:40 +0300)]
drm/msm/dp: support attaching bridges to the DP encoder
Currently DP driver will allocate panel bridge for eDP panels. This
supports only the following topology:
- eDP encoder ⇒ eDP panel (wrapped using panel-bridge)
Simplify this code to just check if there is any next bridge in the
chain (be it a panel bridge or regular bridge). Rename panel_bridge
field to next_bridge accordingly.
This allows one to use e.g. one of the following display topologies:
Dmitry Baryshkov [Fri, 11 Feb 2022 22:40:02 +0000 (01:40 +0300)]
drm/msm/dp: fix panel bridge attachment
In commit 3fc9f2bcab38 ("drm/msm/dp: employ bridge mechanism for display
enable and disable") the DP driver received a drm_bridge instance, which
is always attached to the encoder as a root bridge. However it conflicts
with the panel_bridge support for eDP panels. The panel bridge attaches
to the encoder before the "dp" bridge (DP driver's drm_bridge instance
created in msm_dp_bridge_init()) has a chance to do so. Change
panel bridge attachment to come after the "dp" bridge attachment (and to
use it as a previous bridge).
Fixes: 3fc9f2bcab38 ("drm/msm/dp: employ bridge mechanism for display enable and disable") Cc: Kuogee Hsieh <quic_khsieh@quicinc.com> Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org> Tested-by: Kuogee Hsieh <quic_khsieh@quicinc.com> Reviewed-by: Stephen Boyd <swboyd@chromium.org> Link: https://lore.kernel.org/r/20220211224006.1797846-2-dmitry.baryshkov@linaro.org
[db: fixed commit message according to Stephen's suggestions] Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
drm/msm/dpu: Fix timeout issues on command mode panels
In function dpu_encoder_phys_cmd_wait_for_commit_done we are always
checking if the relative CTL is started by waiting for an interrupt
to fire: it is fine to do that, but then sometimes we call this
function while the CTL is up and has never been put down, but that
interrupt gets raised only when the CTL gets a state change from
0 to 1 (disabled to enabled), so we're going to wait for something
that will never happen on its own.
Solving this while avoiding to restart the CTL is actually possible
and can be done by just checking if it is already up and running
when the wait_for_commit_done function is called: in this case, so,
if the CTL was already running, we can say that the commit is done
if the command transmission is complete (in other terms, if the
interface has been flushed).
Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@somainline.org> Reviewed-by: Marijn Suijten <marijn.suijten@somainline.org> Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org> Link: https://lore.kernel.org/r/20210911163919.47173-2-angelogioacchino.delregno@somainline.org Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Dmitry Baryshkov [Thu, 17 Feb 2022 05:55:25 +0000 (08:55 +0300)]
drm/msm/dpu: simplify clocks handling
DPU driver contains code to parse clock items from device tree into
special data struct and then enable/disable/set rate for the clocks
using that data struct. However the DPU driver itself uses only parsing
and enabling/disabling part (the rate setting is used by DP driver).
Move this implementation to the DP driver (which actually uses rate
setting) and replace hand-coded enable/disable/get loops in the DPU
with the respective clk_bulk operations. Put operation is removed
completely because, it is handled using devres instead.
Dmitry Baryshkov [Thu, 17 Feb 2022 03:53:58 +0000 (06:53 +0300)]
drm/msm/dpu: pull connector from dpu_encoder_phys to dpu_encoder_virt
All physical encoders used by virtual encoder share the same connector,
so pull the connector field from dpu_encoder_phys into dpu_encoder_virt
structure. Otherwise code suggests that different phys_encs
can have different connectors.
Dmitry Baryshkov [Thu, 17 Feb 2022 03:53:57 +0000 (06:53 +0300)]
drm/msm/dpu: switch dpu_encoder to use atomic_mode_set
Make dpu_encoder use atomic_mode_set to receive connector and CRTC
states as arguments rather than finding connector and CRTC by manually
looping through the respective lists.
Dmitry Baryshkov [Thu, 17 Feb 2022 03:53:52 +0000 (06:53 +0300)]
drm/msm/dpu: fix dp audio condition
DP audio enablement code which is comparing intf_type,
DRM_MODE_ENCODER_TMDS (= 2) with DRM_MODE_CONNECTOR_DisplayPort (= 10).
Which would never succeed. Fix it to check for DRM_MODE_ENCODER_TMDS.
Fixes: 80d99f9671f8 ("drm/msm/dp: add audio support for Display Port on MSM") Reviewed-by: Abhinav Kumar <quic_abhinavk@quicinc.com> Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org> Reviewed-by: Stephen Boyd <swboyd@chromium.org> Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org> Link: https://lore.kernel.org/r/20220217035358.465904-2-dmitry.baryshkov@linaro.org Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Bjorn Andersson [Sat, 12 Feb 2022 00:38:11 +0000 (03:38 +0300)]
drm/msm/dpu: Only create debugfs for PRIMARY minor
dpu_kms_debugfs_init() is invoked for each minor being registered. Most
of the files created are unrelated to the minor, so there's no reason to
present them per minor.
The exception to this is the DisplayPort code, which ends up invoking
dp_debug_get() for each minor, each time associate the allocated object
with dp->debug.
As such dp_debug will create debugfs files in both the PRIMARY and the
RENDER minor's debugfs directory, but only the last reference will be
remembered.
The only use of this reference today is in the cleanup path in
dp_display_deinit_sub_modules() and the dp_debug_private object does
outlive the debugfs entries in either case, so there doesn't seem to be
any adverse effects of this, but per the code the current behavior is
unexpected, so change it to only create debugfs files for the PRIMARY
minor.
Rob Clark [Thu, 17 Feb 2022 01:21:55 +0000 (17:21 -0800)]
drm/msm/dpu: Add SC8180x to hw catalog
Add SC8180x to the hardware catalog, for initial support for the
platform. Due to limitations in the DP driver only one of the four DP
interfaces is left enabled.
The SC8180x platform supports the newly added DPU_INTF_WIDEBUS flag and
the Windows-on-Snapdragon bootloader leaves the widebus bit set, so this
is flagged appropriately to ensure widebus is disabled - for now.
Signed-off-by: Rob Clark <robdclark@chromium.org>
[bjorn: Reworked intf and irq definitions] Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org> Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Dmitry Baryshkov [Fri, 21 Jan 2022 21:06:17 +0000 (00:06 +0300)]
drm/msm/dpu: fix error handling in dpu_rm_init
Using IS_ERR_OR_NULL() together with PTR_ERR() is a typical mistake. If
the value is NULL, then the function will return 0 instead of a proper
return code. Moreover none of dpu_hw_*_init() functions can return NULL.
So, replace all dpu_rm_init()'s IS_ERR_OR_NULL() calls with IS_ERR().
Dmitry Baryshkov [Fri, 21 Jan 2022 21:06:16 +0000 (00:06 +0300)]
drm/msm/dpu: stop embedding dpu_hw_blk into dpu_hw_intf
Now as dpu_hw_intf is not hanled by dpu_rm_get_assigned_resources, there
is no point in embedding the (empty) struct dpu_hw_blk into dpu_hw_intf
(and using typecasts between dpu_hw_blk and dpu_hw_intf). Drop it and
use dpu_hw_intf directly.
Dmitry Baryshkov [Fri, 21 Jan 2022 21:06:15 +0000 (00:06 +0300)]
drm/msm/dpu: get INTF blocks directly rather than through RM
INTF blocks are not really handled by resource manager, they are
assigned at dpu_encoder_setup_display using dpu_encoder_get_intf().
Then this allocation is passed to RM and then returned to then
dpu_encoder.
So allocate them outside of RM and use them directly.
Dmitry Baryshkov [Fri, 21 Jan 2022 21:06:13 +0000 (00:06 +0300)]
drm/msm/dpu: drop unused lm_max_width from RM
No code uses lm_max_width from resource manager, so drop it. Instead of
calculating the lm_max_width, code can use max_mixer_width field from
the hw catalog.
Kuogee Hsieh [Mon, 14 Feb 2022 21:46:54 +0000 (13:46 -0800)]
drm/msm: populate intf_audio_select() base on hardware capability
intf_audio_select() callback function use to configure
HDMI_DP_CORE_SELECT to decide audio output routes to HDMI or DP
interface. HDMI is obsoleted at newer chipset. To keep supporting
legacy hdmi application, intf_audio_select call back function have
to be populated base on hardware chip capability where legacy
chipsets have has_audio_select flag set to true.
Changes in V2:
-- remove has_audio_select flag
-- add BIT(DPU_MDP_AUDIO_SELECT) into dpu_mdp_cfg
Vinod Koul [Thu, 10 Feb 2022 11:41:06 +0000 (17:11 +0530)]
drm/msm/dpu: Update function parameter documentation
dpu_core_irq_callback_handler() function comments seem to have become
stale and emit a warning:
drivers/gpu/drm/msm/disp/dpu1/dpu_hw_interrupts.c:147:
warning: Function parameter or member 'dpu_kms' not described in 'dpu_core_irq_callback_handler'
drivers/gpu/drm/msm/disp/dpu1/dpu_hw_interrupts.c:147:
warning: Excess function parameter 'arg' description in 'dpu_core_irq_callback_handler'
Vinod Koul [Thu, 10 Feb 2022 11:41:05 +0000 (17:11 +0530)]
drm/msm/dpu: Update the comment style
The multi line comment style is wrongly used as kernel-doc comment. This
gives a warning:
drivers/gpu/drm/msm/disp/dpu1/dpu_hw_interrupts.c:17:
warning: This comment starts with '/**', but isn't a kernel-doc comment.
Refer Documentation/doc-guide/kernel-doc.rst
Vinod Koul [Thu, 10 Feb 2022 11:41:04 +0000 (17:11 +0530)]
drm/msm/dpu: Remove set but unused variables
We get warning:
In function ‘dpu_encoder_virt_enable’: drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c:1145:33:
warning: variable ‘priv’ set but not used [-Wunused-but-set-variable]
1145 | struct msm_drm_private *priv;
In function ‘dpu_encoder_virt_disable’: drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c:1182:33:
warning: variable ‘priv’ set but not used [-Wunused-but-set-variable]
1182 | struct msm_drm_private *priv;
dt-bindings: display: msm: Add binding for msm8998 dpu
Add yaml binding for msm8998 dpu1 support.
Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@somainline.org> Signed-off-by: Jami Kettunen <jami.kettunen@somainline.org> Reviewed-by: Rob Herring <robh@kernel.org> Link: https://lore.kernel.org/r/20220113145111.29984-4-jami.kettunen@somainline.org Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
drm/msm/dpu1: Add DMA2, DMA3 clock control to enum
The enum dpu_clk_ctrl_type misses DPU_CLK_CTRL_DMA{2,3} even though
this driver does actually handle both, if present: add the two in
preparation for adding support for SoCs having them.
Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@somainline.org> Signed-off-by: Jami Kettunen <jami.kettunen@somainline.org> Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org> Link: https://lore.kernel.org/r/20220113145111.29984-2-jami.kettunen@somainline.org Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Yury Norov [Thu, 10 Feb 2022 22:48:52 +0000 (14:48 -0800)]
drm: replace bitmap_weight with bitmap_empty where appropriate
smp_request_block() in drivers/gpu/drm/msm/disp/mdp5/mdp5_smp.c calls
bitmap_weight() to check if any bit of a given bitmap is set. It's
better to use bitmap_empty() in that case because bitmap_empty() stops
traversing the bitmap as soon as it finds first set bit, while
bitmap_weight() counts all bits unconditionally.
Daniel Thompson [Tue, 1 Feb 2022 17:47:33 +0000 (17:47 +0000)]
drm/msm/hdmi: Remove spurious IRQF_ONESHOT flag
Quoting the header comments, IRQF_ONESHOT is "Used by threaded interrupts
which need to keep the irq line disabled until the threaded handler has
been run.". When applied to an interrupt that doesn't request a threaded
irq then IRQF_ONESHOT has a lesser known (undocumented?) side effect,
which it to disable the forced threading of irqs. For "normal" kernels
if there is no thread_fn then IRQF_ONESHOT is a nop.
In this case disabling forced threading is not appropriate because the
driver calls wake_up_all() (via msm_hdmi_i2c_irq) and also directly uses
the regular spinlock API for locking (in msm_hdmi_hdcp_irq() ). Neither
of these APIs can be called from no-thread interrupt handlers on
PREEMPT_RT systems.
Dmitry Baryshkov [Thu, 17 Feb 2022 00:08:37 +0000 (03:08 +0300)]
drm/msm/dsi/phy: fix 7nm v4.0 settings for C-PHY mode
The dsi_7nm_phy_enable() disagrees with downstream for
glbl_str_swi_cal_sel_ctrl and glbl_hstx_str_ctrl_0 values. Update
programmed settings to match downstream driver. To remove the
possibility for such errors in future drop less_than_1500_mhz
assignment and specify settings explicitly.
drm/msm/dsi: move DSI host powerup to modeset time
The DSI subsystem does not fully fall into the pre-enable/enable system
of callbacks, since typically DSI device bridge drivers expect to be
able to communicate with DSI devices at the pre-enable() callback. The
reason is that for some DSI hosts enabling the video stream would
prevent other drivers from sending DSI commands. For example see the
panel-bridge driver, which does drm_panel_prepare() from the
pre_enable() callback (which would be called before our pre_enable()
callback, resulting in panel preparation failures as the link is not yet
ready).
Therewere several attempts to solve this issue, but currently the best
approach is to power up the DSI link from the mode_set() callback,
allowing next bridge/panel to use DSI transfers in the pre_enable()
time. Follow this approach.
Daniel Thompson [Tue, 1 Feb 2022 17:47:32 +0000 (17:47 +0000)]
drm/msm/dsi: Remove spurious IRQF_ONESHOT flag
Quoting the header comments, IRQF_ONESHOT is "Used by threaded interrupts
which need to keep the irq line disabled until the threaded handler has
been run.". When applied to an interrupt that doesn't request a threaded
irq then IRQF_ONESHOT has a lesser known (undocumented?) side effect,
which it to disable the forced threading of irqs (and for "normal" kernels
it is a nop). In this case I can find no evidence that suppressing forced
threading is intentional. Had it been intentional then a driver must adopt
the raw_spinlock API in order to avoid deadlocks on PREEMPT_RT kernels
(and avoid calling any kernel API that uses regular spinlocks).
Fix this by removing the spurious additional flag.
This change is required for my Snapdragon 7cx Gen2 tablet to boot-to-GUI
with PREEMPT_RT enabled.
Rajeev Nandan [Sun, 30 Jan 2022 20:15:19 +0000 (01:45 +0530)]
drm/msm/dsi: Add 10nm dsi phy tuning configuration support
The clock and data lanes of the DSI PHY have a calibration circuitry
feature. As per the MSM DSI PHY tuning guidelines, the drive strength
tuning can be done by adjusting rescode offset for hstop/hsbot, and
the drive level tuning can be done by adjusting the LDO output level
for the HSTX drive.
Rajeev Nandan [Sun, 30 Jan 2022 20:15:18 +0000 (01:45 +0530)]
drm/msm/dsi: Add dsi phy tuning configuration support
Add support for MSM DSI PHY tuning configuration. Current design is
to support drive strength and drive level/amplitude tuning for
10nm PHY version, but this can be extended to other PHY versions.
In most cases, the default values of DSI PHY tuning registers should be
sufficient as they are fully optimized. However, in some cases where
extreme board parasitics cause the eye shape to degrade, the override
bits can be used to improve the signal quality.
The general guidelines for DSI PHY tuning include:
- High and moderate data rates may benefit from the drive strength and
drive level tuning.
- Drive strength tuning will affect the output impedance and may be used
for matching optimization.
- Drive level tuning will affect the output levels without affecting the
impedance.
The clock and data lanes have a calibration circuitry feature. The drive
strength tuning can be done by adjusting rescode offset for hstop/hsbot,
and the drive level tuning can be done by adjusting the LDO output level
for the HSTX drive.
drm/msm/dsi: Use "ref" fw clock instead of global name for VCO parent
All DSI PHY/PLL drivers were referencing their VCO parent clock by a
global name, most of which don't exist or have been renamed. These
clock drivers seem to function fine without that except the 14nm driver
for sdm6xx [1].
At the same time all DTs provide a "ref" clock as per the requirements
of dsi-phy-common.yaml, but the clock is never used. This patchset puts
that clock to use without relying on a global clock name, so that all
dependencies are explicitly defined in DT (the firmware) in the end.
Note that this patch intentionally breaks older firmware (DT) that
relies on the clock to be found globally instead. The only affected
platform is msm8974 [2] for whose dsi_phy_28nm a .name="xo" fallback is
left in place to accommodate a more graceful transition period. All
other platforms had the "ref" clock added to their phy node since its
inception, or in a followup patch some time after. These patches
wrongly assumed that the "ref" clock was actively used and have hence
been listed as "Fixes:" below.
Furthermore apq8064 was providing the wrong 19.2MHz cxo instead of
27MHz pxo clock, which has been addressed in [3].
It is expected that both [2] and [3] are applied to the tree well in
advance of this patch such that any actual breakage is extremely
unlikely, but might still occur if kernel upgrades are performed without
the DT to match. After some time the fallback for msm8974 can be
removed again as well.
Fixes: d6cdf194b4ff ("arm64: dts: qcom: msm8916: Set 'xo_board' as ref clock of the DSI PHY") Fixes: 8b18982e1291 ("ARM: dts: qcom-apq8064: Set 'cxo_board' as ref clock of the DSI PHY") Fixes: b1a4d80c825a ("arm64: dts: sdm845: Set 'bi_tcxo' as ref clock of the DSI PHYs") Signed-off-by: Marijn Suijten <marijn.suijten@somainline.org> Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@somainline.org> Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org> Link: https://lore.kernel.org/r/20210911131922.387964-2-marijn.suijten@somainline.org Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Kuogee Hsieh [Wed, 2 Feb 2022 18:56:38 +0000 (10:56 -0800)]
drm/msm/dp: enhance debug info related to dp phy
DP phy should be initialized and exited symmetrically to avoid
clock being stucked at either on or off error. Add debug info
to cover all DP phy to identify clock issues easily.
Kuogee Hsieh [Wed, 2 Feb 2022 18:56:37 +0000 (10:56 -0800)]
drm/msm/dp: add connector type to enhance debug messages
DP driver is a generic driver which supports both eDP and DP.
For debugging purpose it is required to have capabilities to
differentiate message are generated from eDP or DP. This patch
add connector type into debug messages for this purpose.
Changes in v3:
-- replace original patch into 3 patches
Kuogee Hsieh [Mon, 24 Jan 2022 23:17:54 +0000 (15:17 -0800)]
drm/msm/dp: always add fail-safe mode into connector mode list
Some of DP link compliant test expects to return fail-safe mode
if prefer detailed timing mode can not be supported by mainlink's
lane and rate after link training. Therefore add fail-safe mode
into connector mode list as backup mode. This patch fixes test
case 4.2.2.1.
Changes in v2:
-- add Fixes text string
Fixes: 4c34f33bbdae ( "drm/msm/dp: reduce link rate if failed at link training 1") Signed-off-by: Kuogee Hsieh <quic_khsieh@quicinc.com> Reviewed-by: Stephen Boyd <swboyd@chromium.org> Link: https://lore.kernel.org/r/1643066274-25814-1-git-send-email-quic_khsieh@quicinc.com Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Kuogee Hsieh [Tue, 18 Jan 2022 18:47:28 +0000 (10:47 -0800)]
drm/msm/dp: stop link training after link training 2 failed
Each DP link training contains link training 1 followed by link
training 2. There is maximum of 5 retries of DP link training
before declared link training failed. It is required to stop link
training at end of link training 2 if it is failed so that next
link training 1 can start freshly. This patch fixes link compliance
test case 4.3.1.13 (Source Device Link Training EQ Fallback Test).
Changes in v10:
-- group into one series
Changes in v11:
-- drop drm/msm/dp: dp_link_parse_sink_count() return immediately if aux read
Fixes: bb7f27b18d7c ("drm/msm/dp: do not end dp link training until video is ready") Signed-off-by: Kuogee Hsieh <quic_khsieh@quicinc.com> Reviewed-by: Stephen Boyd <swboyd@chromium.org> Link: https://lore.kernel.org/r/1642531648-8448-5-git-send-email-quic_khsieh@quicinc.com Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Kuogee Hsieh [Tue, 18 Jan 2022 18:47:27 +0000 (10:47 -0800)]
drm/msm/dp: add support of tps4 (training pattern 4) for HBR3
Some DP sinkers prefer to use tps4 instead of tps3 during training #2.
This patch will use tps4 to perform link training #2 if sinker's DPCD
supports it.
Changes in V2:
-- replace dp_catalog_ctrl_set_pattern() with dp_catalog_ctrl_set_pattern_state_bit()
Changes in V3:
-- change state_ctrl_bits type to u32 and pattern type to u8
Changes in V4:
-- align } else if { and } else {
Changes in v10:
-- group into one series
Changes in v11:
-- drop drm/msm/dp: dp_link_parse_sink_count() return immediately if aux read
Kuogee Hsieh [Tue, 18 Jan 2022 18:47:26 +0000 (10:47 -0800)]
drm/msm/dp: populate connector of struct dp_panel
DP CTS test case 4.2.2.6 has valid edid with bad checksum on purpose
and expect DP source return correct checksum. During drm edid read,
correct edid checksum is calculated and stored at
connector::real_edid_checksum.
The problem is struct dp_panel::connector never be assigned, instead the
connector is stored in struct msm_dp::connector. When we run compliance
testing test case 4.2.2.6 dp_panel_handle_sink_request() won't have a valid
edid set in struct dp_panel::edid so we'll try to use the connectors
real_edid_checksum and hit a NULL pointer dereference error because the
connector pointer is never assigned.
Changes in V2:
-- populate panel connector at msm_dp_modeset_init() instead of at dp_panel_read_sink_caps()
Changes in V3:
-- remove unhelpful kernel crash trace commit text
-- remove renaming dp_display parameter to dp
Changes in V4:
-- add more details to commit text
Changes in v10:
-- group into one series
Changes in v11:
-- drop drm/msm/dp: dp_link_parse_sink_count() return immediately if aux read
Kuogee Hsieh [Tue, 18 Jan 2022 18:47:25 +0000 (10:47 -0800)]
drm/msm/dp: do not initialize phy until plugin interrupt received
Current DP drivers have regulators, clocks, irq and phy are grouped
together within a function and executed not in a symmetric manner.
This increase difficulty of code maintenance and limited code scalability.
This patch divides the driver life cycle of operation into four states,
resume (including booting up), dongle plugin, dongle unplugged and suspend.
Regulators, core clocks and irq are grouped together and enabled at resume
(or booting up) so that the DP controller is armed and ready to receive HPD
plugin interrupts. HPD plugin interrupt is generated when a dongle plugs
into DUT (device under test). Once HPD plugin interrupt is received, DP
controller will initialize phy so that dpcd read/write will function and
following link training can be proceeded successfully. DP phy will be
disabled after main link is teared down at end of unplugged HPD interrupt
handle triggered by dongle unplugged out of DUT. Finally regulators, code
clocks and irq are disabled at corresponding suspension.
Changes in V2:
-- removed unnecessary dp_ctrl NULL check
-- removed unnecessary phy init_count and power_count DRM_DEBUG_DP logs
-- remove flip parameter out of dp_ctrl_irq_enable()
-- add fixes tag
Changes in V3:
-- call dp_display_host_phy_init() instead of dp_ctrl_phy_init() at
dp_display_host_init() for eDP
Changes in V4:
-- rewording commit text to match this commit changes
Changes in V5:
-- rebase on top of msm-next branch
Changes in V6:
-- delete flip variable
Changes in V7:
-- dp_ctrl_irq_enable/disabe() merged into dp_ctrl_reset_irq_ctrl()
Changes in V8:
-- add more detail comment regrading dp phy at dp_display_host_init()
Changes in V9:
-- remove set phy_initialized to false when -ECONNRESET detected
Changes in v10:
-- group into one series
Changes in v11:
-- drop drm/msm/dp: dp_link_parse_sink_count() return immediately
if aux read
Changes in v12:
-- move dp_display_host_phy_exit() after dp_display_host_deinit()
Changes in v13:
-- do not execute phy_init until plugged_in interrupt for edp, same as DP.
Changes in v14:
-- remove redundant dp->core_initialized = false form dp_pm_suspend.
Changes in v15:
-- remove core_initialized flag check at both host_init and host_deinit
Changes in v16:
-- remove dp_display_host_phy_exit core_initialized=false at dp_pm_suspend
Changes in v17:
-- remove core_initialized checking before execute attention_cb()
Changes in v18:
-- remove core_initialized checking at dp_pm_suspend
Bjorn Andersson [Fri, 15 Oct 2021 23:17:02 +0000 (16:17 -0700)]
drm/msm/dp: Move debugfs files into subdirectory
In the cleanup path of the MSM DP driver the DP driver's debugfs files
are destroyed by invoking debugfs_remove_recursive() on debug->root,
which during initialization has been set to minor->debugfs_root.
To allow cleaning up the DP driver's debugfs files either each dentry
needs to be kept track of or the files needs to be put in a subdirectory
which can be removed in one go.
By choosing to put the debugfs files in a subdirectory, based on the
name of the associated connector this also solves the problem that these
names would collide as support for multiple DP instances are introduced.
One alternative solution to the problem with colliding file names would
have been to put keep track of the individual files and put them under
the connector's debugfs directory. But while the drm_connector has been
allocated, its associated debugfs directory has not been created at the
time of initialization of the dp_debug.
Rob Clark [Fri, 14 Jan 2022 18:57:42 +0000 (10:57 -0800)]
drm/msm/rd: Add chip-id
For newer devices which deprecate gpu-id and do matching based on
chip-id, we need this information in cmdstream dumps so that the
decoding tools know how to decode them.
The round_pixclk() callback returns different rate only on MDP4 in HDMI
(DTV) case. Stop using this callback in other cases to simplify
mode_valid callbacks.
With the reglog removal, msm_readl/_writel became single line wrappers
around readl/writel. Move those two wrappers and msm_rmw to msm_drv.h to
remove need for extra function calls when doing register writes.
Register logging was used during early stages of msm driver development
to compare upstream and downstream register traces. However the tool was
never updated to work with mdp5 hardware. Later it was dropped
completely when Rob imported freedreno tools into mesa. All this makes
DRM_MSM_REGISTER_LOGGING irrelevant now, so it can be dropped.
Changcheng Deng [Tue, 21 Dec 2021 14:42:34 +0000 (14:42 +0000)]
drm/msm: replace DEFINE_SIMPLE_ATTRIBUTE with DEFINE_DEBUGFS_ATTRIBUTE
Fix the following coccicheck warning:
./drivers/gpu/drm/msm/msm_debugfs.c: 132: 0-23: WARNING: shrink_fops
should be defined with DEFINE_DEBUGFS_ATTRIBUTE
Use DEFINE_DEBUGFS_ATTRIBUTE rather than DEFINE_SIMPLE_ATTRIBUTE for
debugfs files.
Linus Torvalds [Sun, 30 Jan 2022 13:12:02 +0000 (15:12 +0200)]
Merge tag 'irq_urgent_for_v5.17_rc2_p2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull irq fixes from Borislav Petkov:
- Drop an unused private data field in the AIC driver
- Various fixes to the realtek-rtl driver
- Make the GICv3 ITS driver compile again in !SMP configurations
- Force reset of the GICv3 ITSs at probe time to avoid issues during kexec
- Yet another kfree/bitmap_free conversion
- Various DT updates (Renesas, SiFive)
* tag 'irq_urgent_for_v5.17_rc2_p2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
dt-bindings: interrupt-controller: sifive,plic: Group interrupt tuples
dt-bindings: interrupt-controller: sifive,plic: Fix number of interrupts
dt-bindings: irqchip: renesas-irqc: Add R-Car V3U support
irqchip/gic-v3-its: Reset each ITS's BASERn register before probe
irqchip/gic-v3-its: Fix build for !SMP
irqchip/loongson-pch-ms: Use bitmap_free() to free bitmap
irqchip/realtek-rtl: Service all pending interrupts
irqchip/realtek-rtl: Fix off-by-one in routing
irqchip/realtek-rtl: Map control data to virq
irqchip/apple-aic: Drop unused ipi_hwirq field
Linus Torvalds [Sun, 30 Jan 2022 13:02:32 +0000 (15:02 +0200)]
Merge tag 'perf_urgent_for_v5.17_rc2_p2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull perf fixes from Borislav Petkov:
- Prevent accesses to the per-CPU cgroup context list from another CPU
except the one it belongs to, to avoid list corruption
- Make sure parent events are always woken up to avoid indefinite hangs
in the traced workload
* tag 'perf_urgent_for_v5.17_rc2_p2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
perf/core: Fix cgroup event list management
perf: Always wake the parent event
Linus Torvalds [Sun, 30 Jan 2022 11:09:00 +0000 (13:09 +0200)]
Merge tag 'sched_urgent_for_v5.17_rc2_p2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull scheduler fix from Borislav Petkov:
"Make sure the membarrier-rseq fence commands are part of the reported
set when querying membarrier(2) commands through MEMBARRIER_CMD_QUERY"
* tag 'sched_urgent_for_v5.17_rc2_p2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
sched/membarrier: Fix membarrier-rseq fence command missing from query bitmask
Linus Torvalds [Sun, 30 Jan 2022 10:55:06 +0000 (12:55 +0200)]
Merge tag 'x86_urgent_for_v5.17_rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull x86 fixes from Borislav Petkov:
- Add another Intel CPU model to the list of CPUs supporting the
processor inventory unique number
- Allow writing to MCE thresholding sysfs files again - a previous
change had accidentally disabled it and no one noticed. Goes to show
how much is this stuff used
* tag 'x86_urgent_for_v5.17_rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/cpu: Add Xeon Icelake-D to list of CPUs that support PPIN
x86/MCE/AMD: Allow thresholding interface updates after init
Linus Torvalds [Sun, 30 Jan 2022 09:21:50 +0000 (11:21 +0200)]
Merge branch 'akpm' (patches from Andrew)
Merge misc fixes from Andrew Morton:
"12 patches.
Subsystems affected by this patch series: sysctl, binfmt, ia64, mm
(memory-failure, folios, kasan, and psi), selftests, and ocfs2"
* emailed patches from Andrew Morton <akpm@linux-foundation.org>:
ocfs2: fix a deadlock when commit trans
jbd2: export jbd2_journal_[grab|put]_journal_head
psi: fix "defined but not used" warnings when CONFIG_PROC_FS=n
psi: fix "no previous prototype" warnings when CONFIG_CGROUPS=n
mm, kasan: use compare-exchange operation to set KASAN page tag
kasan: test: fix compatibility with FORTIFY_SOURCE
tools/testing/scatterlist: add missing defines
mm: page->mapping folio->mapping should have the same offset
memory-failure: fetch compound_head after pgmap_pfn_valid()
ia64: make IA64_MCA_RECOVERY bool instead of tristate
binfmt_misc: fix crash when load/unload module
include/linux/sysctl.h: fix register_sysctl_mount_point() return type
projects / kernel.git / log