upgrade the callchain to drm_dbg() and drm_dev_dbg(); add a struct
_ddebug ptr parameter to them, and supply that additional param by
replacing the '_no_desc' flavor of dyndbg Factory macro currently used
with the flavor that supplies the descriptor.
NOTES:
The descriptor gives these fns access to the decorator flags, but they
do none of the dynamic-prefixing done by dynamic_emit_prefix(), which
is currently static.
DRM already has conventions for logging/messaging; just tossing
optional decorations on top probably wouldn't help. Instead, existing
flags (or new ones, perhaps 'sd' ala lspci) can be used to make
current message conventions optional. This suggests a new
drmdbg_prefix_emit() to handle prefixing locally.
For CONFIG_DRM_USE_DYNAMIC_DEBUG=N, just pass null descriptor.
desc->class_id is redundant with category parameter, but its
availability is dependent on desc.
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Link: https://lore.kernel.org/r/20220912052852.1123868-10-jim.cromie@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drm_print.c calls pr_debug() just once, from __drm_printfn_debug(),
which is a generic/service fn. The callsite is compile-time enabled
by DEBUG in both DYNAMIC_DEBUG=y/n builds.
For dyndbg builds, reverting this callsite back to bare printk is
correcting a few anti-features:
1- callsite is generic, serves multiple drm users.
it is soft-wired on currently by #define DEBUG
could accidentally: #> echo -p > /proc/dynamic_debug/control
2- optional "decorations" by dyndbg are unhelpful/misleading here,
they describe only the generic site, not end users
IOW, 1,2 are unhelpful at best, and possibly confusing.
reverting yields a nominal data and text shrink:
text data bss dec hex filename
462583 36604 54592 553779 87333 /kernel/drivers/gpu/drm/drm.ko
462515 36532 54592 553639 872a7 -dirty/kernel/drivers/gpu/drm/drm.ko
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Link: https://lore.kernel.org/r/20220912052852.1123868-9-jim.cromie@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When CONFIG_DRM_USE_DYNAMIC_DEBUG=y, the drm.debug API (a macro stack,
calling _+drm_*dbg() eventually) invokes a dyndbg Factory macro to
create a descriptor for each callsite, thus making them individually
>control-able.
In this case, the calls to _drm_*dbg are unreachable unless the
callsite is enabled. So those calls can short-circuit their early
do-nothing returns. Provide and use __drm_debug_enabled(), to do this
when config'd, or the _raw flags-check otherwize.
And since dyndbg is in use, lets also instrument the remaining users
of drm_debug_enabled, by wrapping the _raw in a macro with a:
pr_debug("todo: is this frequent enough to optimize ?\n");
For CONFIG_DRM_USE_DYNAMIC_DEBUG=n, do no site instrumenting at all,
since JUMP_LABEL might be off, and we don't want to make work.
With drm, amdgpu, i915, nouveau loaded, heres remaining uses of
drm_debug_enabled(), which costs ~1.5kb data to control the
pr_debug("todo:..")s.
Some of those uses might be ok to use __drm_debug_enabled() by
inspection, others might warrant conversion to use dyndbg Factory
macros, and that would want callrate data to estimate the savings
possible. TBH, any remaining savings are probably small; drm.debug
covers the vast bulk of the uses. Maybe "vblank" is the exception.
:#> grep todo /proc/dynamic_debug/control | wc
21 168 2357
:#> grep todo /proc/dynamic_debug/control
drivers/gpu/drm/drm_edid_load.c:178 [drm]edid_load =_ "todo: maybe avoid via dyndbg\n"
drivers/gpu/drm/drm_vblank.c:410 [drm]drm_crtc_accurate_vblank_count =_ "todo: maybe avoid via dyndbg\n"
drivers/gpu/drm/drm_vblank.c:787 [drm]drm_crtc_vblank_helper_get_vblank_timestamp_internal =_ "todo: maybe avoid via dyndbg\n"
drivers/gpu/drm/drm_vblank.c:1491 [drm]drm_vblank_restore =_ "todo: maybe avoid via dyndbg\n"
drivers/gpu/drm/drm_vblank.c:1433 [drm]drm_vblank_enable =_ "todo: maybe avoid via dyndbg\n"
drivers/gpu/drm/drm_plane.c:2168 [drm]drm_mode_setplane =_ "todo: maybe avoid via dyndbg\n"
drivers/gpu/drm/display/drm_dp_mst_topology.c:1359 [drm_display_helper]drm_dp_mst_wait_tx_reply =_ "todo: maybe avoid via dyndbg\n"
drivers/gpu/drm/display/drm_dp_mst_topology.c:2864 [drm_display_helper]process_single_tx_qlock =_ "todo: maybe avoid via dyndbg\n"
drivers/gpu/drm/display/drm_dp_mst_topology.c:2909 [drm_display_helper]drm_dp_queue_down_tx =_ "todo: maybe avoid via dyndbg\n"
drivers/gpu/drm/display/drm_dp_mst_topology.c:1686 [drm_display_helper]drm_dp_mst_update_slots =_ "todo: maybe avoid via dyndbg\n"
drivers/gpu/drm/i915/display/intel_dp.c:1111 [i915]intel_dp_print_rates =_ "todo: maybe avoid via dyndbg\n"
drivers/gpu/drm/i915/display/intel_backlight.c:5434 [i915]cnp_enable_backlight =_ "todo: maybe avoid via dyndbg\n"
drivers/gpu/drm/i915/display/intel_backlight.c:5459 [i915]intel_backlight_device_register =_ "todo: maybe avoid via dyndbg\n"
drivers/gpu/drm/i915/display/intel_opregion.c:43 [i915]intel_opregion_notify_encoder =_ "todo: maybe avoid via dyndbg\n"
drivers/gpu/drm/i915/display/intel_opregion.c:53 [i915]asle_set_backlight =_ "todo: maybe avoid via dyndbg\n"
drivers/gpu/drm/i915/display/intel_bios.c:1088 [i915]intel_bios_is_dsi_present =_ "todo: maybe avoid via dyndbg\n"
drivers/gpu/drm/i915/display/intel_display_debugfs.c:6153 [i915]i915_drrs_ctl_set =_ "todo: maybe avoid via dyndbg\n"
drivers/gpu/drm/i915/intel_pcode.c:26 [i915]snb_pcode_read =_ "todo: maybe avoid via dyndbg\n"
drivers/gpu/drm/i915/i915_getparam.c:785 [i915]i915_getparam_ioctl =_ "todo: maybe avoid via dyndbg\n"
drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c:282 [amdgpu]vcn_v2_5_process_interrupt =_ "todo: maybe avoid via dyndbg\n"
drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c:433 [amdgpu]vcn_v2_0_process_interrupt =_ "todo: maybe avoid via dyndbg\n"
:#>
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Link: https://lore.kernel.org/r/20220912052852.1123868-8-jim.cromie@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
For CONFIG_DRM_USE_DYNAMIC_DEBUG=y, wrap __drm_dbg() & __drm_dev_dbg()
in one of dyndbg's Factory macros: _dynamic_func_call_no_desc().
This adds the callsite descriptor into the code, and an entry for each
into /proc/dynamic_debug/control.
#> echo class DRM_UT_ATOMIC +p > /proc/dynamic_debug/control
CONFIG_DRM_USE_DYNAMIC_DEBUG=y/n is configurable because of the .data
footprint cost of per-callsite control; 56 bytes/site * ~2k for i915,
~4k callsites for amdgpu. This is large enough that a kernel builder
might not want it.
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Link: https://lore.kernel.org/r/20220912052852.1123868-5-jim.cromie@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
change drm_dev_dbg & drm_dbg to macros, which forward to the renamed
functions (with __ prefix added).
Those functions sit below the categorized layer of macros implementing
the DRM debug.category API, and implement most of it. These are good
places to insert dynamic-debug jump-label mechanics, which will allow
DRM to avoid the runtime cost of drm_debug_enabled().
no functional changes.
memory cost baseline: (unchanged)
bash-5.1# drms_load
[ 9.220389] dyndbg: 1 debug prints in module drm
[ 9.224426] ACPI: bus type drm_connector registered
[ 9.302192] dyndbg: 2 debug prints in module ttm
[ 9.305033] dyndbg: 8 debug prints in module video
[ 9.627563] dyndbg: 127 debug prints in module i915
[ 9.721505] AMD-Vi: AMD IOMMUv2 functionality not available on this system - This is not a bug.
[ 10.091345] dyndbg: 2196 debug prints in module amdgpu
[ 10.106589] [drm] amdgpu kernel modesetting enabled.
[ 10.107270] amdgpu: CRAT table not found
[ 10.107926] amdgpu: Virtual CRAT table created for CPU
[ 10.108398] amdgpu: Topology: Add CPU node
[ 10.168507] dyndbg: 3 debug prints in module wmi
[ 10.329587] dyndbg: 3 debug prints in module nouveau
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Link: https://lore.kernel.org/r/20220912052852.1123868-4-jim.cromie@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Use DECLARE_DYNDBG_CLASSMAP across DRM:
- in .c files, since macro defines/initializes a record
- in drivers, $mod_{drv,drm,param}.c
ie where param setup is done, since a classmap is param related
- in drm/drm_print.c
since existing __drm_debug param is defined there,
and we ifdef it, and provide an elaborated alternative.
- in drm_*_helper modules:
dp/drm_dp - 1st item in makefile target
drivers/gpu/drm/drm_crtc_helper.c - random pick iirc.
Since these modules all use identical CLASSMAP declarations (ie: names
and .class_id's) they will all respond together to "class DRM_UT_*"
query-commands:
:#> echo class DRM_UT_KMS +p > /proc/dynamic_debug/control
NOTES:
This changes __drm_debug from int to ulong, so BIT() is usable on it.
DRM's enum drm_debug_category values need to sync with the index of
their respective class-names here. Then .class_id == category, and
dyndbg's class FOO mechanisms will enable drm_dbg(DRM_UT_KMS, ...).
Though DRM needs consistent categories across all modules, thats not
generally needed; modules X and Y could define FOO differently (ie a
different NAME => class_id mapping), changes are made according to
each module's private class-map.
No callsites are actually selected by this patch, since none are
class'd yet.
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Link: https://lore.kernel.org/r/20220912052852.1123868-3-jim.cromie@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This reverts commit 211f276ed3.
For quite some time, core DRM helpers already ensure that any relevant
connectors/CRTCs/etc. are disabled, as well as their associated
components (e.g., bridges) when suspending the system. Thus,
analogix_dp_bridge_{enable,disable}() already get called, which in turn
call drm_panel_{prepare,unprepare}(). This makes these drm_panel_*()
calls redundant.
Besides redundancy, there are a few problems with this handling:
(1) drm_panel_{prepare,unprepare}() are *not* reference-counted APIs and
are not in general designed to be handled by multiple callers --
although some panel drivers have a coarse 'prepared' flag that mitigates
some damage, at least. So at a minimum this is redundant and confusing,
but in some cases, this could be actively harmful.
(2) The error-handling is a bit non-standard. We ignored errors in
suspend(), but handled errors in resume(). And recently, people noticed
that the clk handling is unbalanced in error paths, and getting *that*
right is not actually trivial, given the current way errors are mostly
ignored.
(3) In the particular way analogix_dp_{suspend,resume}() get used (e.g.,
in rockchip_dp_*(), as a late/early callback), we don't necessarily have
a proper PM relationship between the DP/bridge device and the panel
device. So while the DP bridge gets resumed, the panel's parent device
(e.g., platform_device) may still be suspended, and so any prepare()
calls may fail.
So remove the superfluous, possibly-harmful suspend()/resume() handling
of panel state.
Fixes: 211f276ed3 ("drm: bridge: analogix/dp: add panel prepare/unprepare in suspend/resume time")
Link: https://lore.kernel.org/all/Yv2CPBD3Picg%2FgVe@google.com/
Signed-off-by: Brian Norris <briannorris@chromium.org>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20220822180729.1.I8ac5abe3a4c1c6fd5c061686c6e883c22f69022c@changeid
The ->ring_idx_mask variable is a u64 so static checkers, Smatch in
this case, complain if the BIT() is not also a u64.
drivers/gpu/drm/virtio/virtgpu_ioctl.c:50 virtio_gpu_fence_event_create()
warn: should '(1 << ring_idx)' be a 64 bit type?
Fixes: cd7f5ca335 ("drm/virtio: implement context init: add virtio_gpu_fence_event")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Link: http://patchwork.freedesktop.org/patch/msgid/YygN7jY0GdUSQSy0@kili
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Existing code is causing a race condition where dirt_needed value is
already set by the host and gets overwritten with default value. Remove
this default setting of dirt_needed, to avoid overwriting the value
received in the channel callback set by vmbus_open. Removing this
setting also means the default value for dirt_needed is changed to false
as it's allocated by kzalloc which is similar to legacy hyperv_fb driver.
Signed-off-by: Saurabh Sengar <ssengar@linux.microsoft.com>
Reviewed-by: Dexuan Cui <decui@microsoft.com>
Reviewed-by: Michael Kelley <mikelley@microsoft.com>
Link: https://lore.kernel.org/r/1662996766-19304-1-git-send-email-ssengar@linux.microsoft.com
Signed-off-by: Wei Liu <wei.liu@kernel.org>
As the comment right before the mtk_dsi_stop() call advises,
mtk_dsi_stop() should only be called after
mtk_drm_crtc_atomic_disable(). That's because that function calls
drm_crtc_wait_one_vblank(), which requires the vblank irq to be enabled.
Previously mtk_dsi_stop(), being in mtk_dsi_poweroff() and guarded by a
refcount, would only be called at the end of
mtk_drm_crtc_atomic_disable(), through the call to mtk_crtc_ddp_hw_fini().
Commit cde7e2e35c ("drm/mediatek: Separate poweron/poweroff from
enable/disable and define new funcs") moved the mtk_dsi_stop() call to
mtk_output_dsi_disable(), causing it to be called before
mtk_drm_crtc_atomic_disable(), and consequently generating vblank
timeout warnings during suspend.
Move the mtk_dsi_stop() call back to mtk_dsi_poweroff() so that we have
a working vblank irq during mtk_drm_crtc_atomic_disable() and stop
getting vblank timeout warnings.
Fixes: cde7e2e35c ("drm/mediatek: Separate poweron/poweroff from enable/disable and define new funcs")
Signed-off-by: Nícolas F. R. A. Prado <nfraprado@collabora.com>
Tested-by: Hsin-Yi Wang <hsinyi@chromium.org>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Tested-by: Allen-KH Cheng <allen-kh.cheng@mediatek.com>
Link: http://lists.infradead.org/pipermail/linux-mediatek/2022-August/046713.html
Signed-off-by: Chun-Kuang Hu <chunkuang.hu@kernel.org>
Most of the arguments are identical between the two call sites and they
can be accessed through the 'struct vba_vars_st' pointer. This reduces
the total amount of stack space that
dml314_ModeSupportAndSystemConfigurationFull() uses by 112 bytes with
LLVM 16 (1976 -> 1864), helping clear up the following clang warning:
drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dcn314/display_mode_vba_314.c:4020:6: error: stack frame size (2216) exceeds limit (2048) in 'dml314_ModeSupportAndSystemConfigurationFull' [-Werror,-Wframe-larger-than]
void dml314_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_lib)
^
1 error generated.
Link: https://github.com/ClangBuiltLinux/linux/issues/1710
Reported-by: "kernelci.org bot" <bot@kernelci.org>
Tested-by: Maíra Canal <mairacanal@riseup.net>
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Most of the arguments are identical between the two call sites and they
can be accessed through the 'struct vba_vars_st' pointer. This reduces
the total amount of stack space that
dml314_ModeSupportAndSystemConfigurationFull() uses by 240 bytes with
LLVM 16 (2216 -> 1976), helping clear up the following clang warning:
drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dcn314/display_mode_vba_314.c:4020:6: error: stack frame size (2216) exceeds limit (2048) in 'dml314_ModeSupportAndSystemConfigurationFull' [-Werror,-Wframe-larger-than]
void dml314_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_lib)
^
1 error generated.
Link: https://github.com/ClangBuiltLinux/linux/issues/1710
Reported-by: "kernelci.org bot" <bot@kernelci.org>
Tested-by: Maíra Canal <mairacanal@riseup.net>
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
This shouldn't trigger during tiled display hotplug/unplug but it does
because one of the tiles can end up with a NULL plane state.
This also doesn't guard against the hang that it was originally trying
to resolve, and can instead cause DIO corruption due to OTG sync
being lost.
[How]
This was reverted at one point out of DCN31 so revert it here too.
Reviewed-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Acked-by: Wayne Lin <wayne.lin@amd.com>
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
When USB4 DP link training failed and fell back to lower link rate,
the time slot calculation uses the verified_link_cap.
And the verified_link_cap was not updated to the new one.
It caused the wrong VC payload time-slot was allocated.
[How]
Updated verified_link_cap with the new one from cur_link_settings
after the LT completes successfully.
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Wayne Lin <wayne.lin@amd.com>
Signed-off-by: Cruise Hung <Cruise.Hung@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
During hot plug of specific 5K tiled display, sometimes both the tiles
are not synchronized resulting in distortion. The reason is that otgs of
both the tiles goes out of sync when otg workaround (dcnxxx_disable_otg_wa)
is applied for bandwidth optimization. The otg workaround reenables otg
but otg synchronization context is not reset and hence dc_trigger_sync()
does not resynchronize otg again.
[How]
Implement reset_sync_context_for_pipe() to reset the otg synchronization
context for the disabled pipe and its slave pipes when otg workaround is
applied.
Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Acked-by: Wayne Lin <wayne.lin@amd.com>
Signed-off-by: Meenakshikumar Somasundaram <meenakshikumar.somasundaram@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
Current DCN3.2 logic for finding the dummy P-state index uses the
DCN3.0 DML validation function instead of DCN3.2 DML.
This can result in either unexpected DML VBA values, or unexpected
dummy P-state index to be used.
[How]
Update the dummy P-state logic to use DCN3.2 DML validation function.
Reviewed-by: Alvin Lee <alvin.lee2@amd.com>
Reviewed-by: Nevenko Stupar <Nevenko.Stupar@amd.com>
Acked-by: Wayne Lin <wayne.lin@amd.com>
Signed-off-by: George Shen <george.shen@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[why]
We have minimal pipe split transition method to avoid pipe
allocation outage.However, this method will invoke audio setup
which cause audio output stuck once pipe reallocate.
[how]
skip audio setup for pipelines which audio stream has been enabled
Reviewed-by: Charlene Liu <Charlene.Liu@amd.com>
Acked-by: Wayne Lin <wayne.lin@amd.com>
Signed-off-by: zhikzhai <zhikai.zhai@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
The desktop plane and full-screen game plane may have different
gamut remap coefficients, if switching between desktop and
full-screen game without updating the gamut remap will cause
incorrect color.
[How]
Update gamut remap if planes change.
Reviewed-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Acked-by: Wayne Lin <wayne.lin@amd.com>
Signed-off-by: Hugo Hu <hugo.hu@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[WHY]
LTTPRs can in very rare instsances fail to increment DPCD LTTPR count.
This results in aux-i LTTPR requests to be sent to the wrong DPCD
address, which causes link training failure.
[HOW]
Override internal repeater count if fixed_vs flag is set for a given link
Reviewed-by: George Shen <George.Shen@amd.com>
Acked-by: Wayne Lin <wayne.lin@amd.com>
Signed-off-by: Michael Strauss <michael.strauss@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
DC makes use of layer_index (zpos) when picking the HW plane to enable
HW cursor on. However, some compositors will not attach zpos information
to each DRM plane. Consequently, in amdgpu, we default layer_index to 0
and do not update it.
This causes said DC logic to enable HW cursor on all planes of the same
layer_index, which manifests as a double cursor issue if one of the
planes is scaled (and hence scaling the cursor as well).
[How]
Use DRM core helpers to calculate a normalized_zpos value for each
drm_plane_state under each crtc, within the atomic state.
This helper will first consider existing zpos values, and if
identical/unset, fallback to plane ID ordering.
The normalized_zpos is then passed to dc_plane_info during atomic check
for later use by the cursor logic.
Reviewed-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Acked-by: Wayne Lin <wayne.lin@amd.com>
Signed-off-by: Leo Li <sunpeng.li@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
420 modes are limited by FMT buffer width of 4096
which requires multi-pipe support in form of ODM
combine. If 420 modes have greater HActive than
4096, the DML logic should accomodate whether
it should be rejected, or ODM combine 2:1 or 4:1
is triggered accordingly.
[How]
FMT Buffer limit of 4096 in DCN32. Force ODM
combine depending on HActive and FMT Buffer limit.
Reject modes if TMDS 420 and above 4096.
Acked-by: Wayne Lin <wayne.lin@amd.com>
Signed-off-by: Chris Park <chris.park@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This patch updates the PTE flags when translate further (TF) is
enabled:
- With translate_further enabled, invalid PTEs can be 0. Reading
consecutive invalid PTEs as 0 is considered a fault. To prevent
this, ensure invalid PTEs have at least 1 bit set.
- The current invalid PTE flags settings to translate a retry fault
into a no-retry fault, doesn't work with TF enabled. As a result,
update invalid PTE flags settings which works for both TF enabled
and disabled case.
Fixes: 352e683b72 ("drm/amdgpu: Enable translate_further to extend UTCL2 reach")
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Mukul Joshi <mukul.joshi@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
