Oded Gabbay [Mon, 31 Oct 2022 13:28:35 +0000 (15:28 +0200)]
drivers/accel: define kconfig and register a new major
Add a new Kconfig for the accel subsystem. The Kconfig currently
contains only the basic CONFIG_DRM_ACCEL option that will be used to
decide whether to compile the accel registration code. Therefore, the
kconfig option is defined as bool.
The accel code will be compiled as part of drm.ko and will be called
directly from the DRM core code. The reason we compile it as part of
drm.ko and not as a separate module is because of cyclic dependency
between drm.ko and the separate module (if it would have existed).
This is due to the fact that DRM core code calls accel functions and
vice-versa.
The accelerator devices will be exposed to the user space with a new,
dedicated major number - 261.
The accel init function registers the new major number as a char device
and create corresponding sysfs and debugfs root entries, similar to
what is done in DRM init function.
I added a new header called drm_accel.h to include/drm/, that will hold
the prototypes of the drm_accel.c functions. In case CONFIG_DRM_ACCEL
is set to 'N', that header will contain empty inline implementations of
those functions, to allow DRM core code to compile successfully
without dependency on CONFIG_DRM_ACCEL.
I Updated the MAINTAINERS file accordingly with the newly added folder
and I have taken the liberty to appropriate the dri-devel mailing list
and the dri-devel IRC channel for the accel subsystem.
Dave Airlie [Tue, 22 Nov 2022 03:41:11 +0000 (13:41 +1000)]
Merge tag 'amd-drm-next-6.2-2022-11-18' of https://gitlab.freedesktop.org/agd5f/linux into drm-next
amd-drm-next-6.2-2022-11-18:
amdgpu:
- SR-IOV fixes
- Clean up DC checks
- DCN 3.2.x fixes
- DCN 3.1.x fixes
- Don't enable degamma on asics which don't support it
- IP discovery fixes
- BACO fixes
- Fix vbios allocation handling when vkms is enabled
- Drop buggy tdr advanced mode GPU reset handling
- Fix the build when DCN is not set in kconfig
- MST DSC fixes
- Userptr fixes
- FRU and RAS EEPROM fixes
- VCN 4.x RAS support
- Aldrebaran CU occupancy reporting fix
- PSP ring cleanup
amdgpu:
- SMU 13.x updates
- GPUVM TLB race fix
- DCN 3.1.4 updates
- DCN 3.2.x updates
- PSR fixes
- Kerneldoc fix
- Vega10 fan fix
- GPUVM locking fixes in error pathes
- BACO fix for Beige Goby
- EEPROM I2C address cleanup
- GFXOFF fix
- Fix DC memory leak in error pathes
- Flexible array updates
- Mtype fix for GPUVM PTEs
- Move Kconfig into amdgpu directory
- SR-IOV updates
- Fix possible memory leak in CS IOCTL error path
amdkfd:
- Fix possible memory overrun
- CRIU fixes
radeon:
- ACPI ref count fix
- HDA audio notifier support
- Move Kconfig into radeon directory
UAPI:
- Add new GEM_CREATE flags to help to transition more KFD functionality to the DRM UAPI.
These are used internally in the driver to align location based memory coherency
requirements from memory allocated in the KFD with how we manage GPUVM PTEs. They
are currently blocked in the GEM_CREATE IOCTL as we don't have a user right now.
They are just used internally in the kernel driver for now for existing KFD memory
allocations. So a change to the UAPI header, but no functional change in the UAPI.
Dave Airlie [Tue, 22 Nov 2022 02:29:47 +0000 (12:29 +1000)]
Merge branch 'etnaviv/next' of https://git.pengutronix.de/git/lst/linux into drm-next
please pull the following etnaviv changes for the next merge window.
Mostly some small workarounds to get new hardware support going. But
also more fixes to the softpin MMU handling and a nice addition from
Christian to make the kernel logs on hang detection more useful.
Dave Airlie [Mon, 21 Nov 2022 23:21:10 +0000 (09:21 +1000)]
Merge tag 'drm-misc-next-2022-11-17' of git://anongit.freedesktop.org/drm/drm-misc into drm-next
drm-misc-next for 6.2:
UAPI Changes:
Cross-subsystem Changes:
- fbdev: Add support for the nomodeset kernel parameter
Core Changes:
- client: Add kunit tests for drm_connector_pick_cmdline_mode()
- dma-buf: Move dma_buf_mmap_internal() to new locking specification
- edid: Dump EDID on drm_edid_get_panel_id() failure, Stop using a
temporary device to load the EDID through the firmware mechanism
- fb-helper: Remove damage worker
- gem-vram: Fix deadlock in drm_gem_vram_vmap()
- modes: Named mode parsing improvements
- tests: Add Kunit helpers to create a DRM device
Driver Changes:
- hisilicon: convert to drm_mode_init()
- malidp: Use drm-managed resources
- msm: convert to drm_mode_init() and drm_mode_copy()
- mtk: convert to drm_mode_init()
- nouveau: Support backlight control for nva3
- rockchip: convert to drm_mode_copy()
- sti: convert to drm_mode_copy()
- v3d: Switch to drm-managed resources
- vc4: Fix potential NULL pointer dereference
Randy Dunlap [Thu, 17 Nov 2022 17:20:09 +0000 (09:20 -0800)]
drm/amd/display: fix kernel-doc issues in dc.h
Fix these kernel-doc complaints:
drivers/gpu/drm/amd/display/dc/dc.h:505: warning: cannot understand function prototype: 'struct dc_clocks '
dc.h:472: warning: Enum value 'MPC_SPLIT_AVOID' not described in enum 'pipe_split_policy'
dc.h:472: warning: Enum value 'MPC_SPLIT_AVOID_MULT_DISP' not described in enum 'pipe_split_policy'
dc.h:532: warning: Incorrect use of kernel-doc format: * @fw_based_mclk_switching
Ramesh Errabolu [Wed, 16 Nov 2022 16:46:08 +0000 (10:46 -0600)]
drm/amdgpu: Enable Aldebaran devices to report CU Occupancy
Allow user to know number of compute units (CU) that are in use at any
given moment. Enable access to the method kgd_gfx_v9_get_cu_occupancy
that computes CU occupancy.
Luben Tuikov [Tue, 15 Nov 2022 05:23:14 +0000 (00:23 -0500)]
drm/amdgpu: Interpret IPMI data for product information (v2)
Don't assume FRU MCU memory locations for the FRU data fields, or their sizes,
instead read and interpret the IPMI data, as stipulated in the IPMI spec
version 1.0 rev 1.2.
Extract the Product Name, Product Part/Model Number, and the Product Serial
Number by interpreting the IPMI data.
Check the checksums of the stored IPMI data to make sure we don't read and
give corrupted data back the the user.
Eliminate small I2C reads, and instead read the whole Product Info Area in one
go, and then extract the information we're seeking from it.
Eliminates a whole function, making this file smaller.
Luben Tuikov [Tue, 15 Nov 2022 02:21:55 +0000 (21:21 -0500)]
drm/amdgpu: Bug-fix: Reading I2C FRU data on newer ASICs
Set the new correct default FRU MCU I2C address for newer ASICs, so that we
can correctly read the Product Name, Product Part/Model Number and Serial
Number.
On newer ASICs, the FRU MCU was moved to I2C address 0x58.
Luben Tuikov [Tue, 15 Nov 2022 01:48:40 +0000 (20:48 -0500)]
drm/amdgpu: Allow non-standard EEPROM I2C address
Allow non-standard EEPROM I2C address of 0x58, where the Device Type
Identifier is 1011b, where we form 1011000b = 0x58 I2C address, as on some
ASICs the FRU data lives there.
Lyude Paul [Mon, 14 Nov 2022 22:17:55 +0000 (17:17 -0500)]
drm/amdgpu/dm/dp_mst: Don't grab mst_mgr->lock when computing DSC state
Now that we've fixed the issue with using the incorrect topology manager,
we're actually grabbing the topology manager's lock - and consequently
deadlocking. Luckily for us though, there's actually nothing in AMD's DSC
state computation code that really should need this lock. The one exception
is the mutex_lock() in dm_dp_mst_is_port_support_mode(), however we grab no
locks beneath &mgr->lock there so that should be fine to leave be.
Lyude Paul [Mon, 14 Nov 2022 22:17:54 +0000 (17:17 -0500)]
drm/amdgpu/dm/mst: Use the correct topology mgr pointer in amdgpu_dm_connector
This bug hurt me. Basically, it appears that we've been grabbing the
entirely wrong mutex in the MST DSC computation code for amdgpu! While
we've been grabbing:
amdgpu_dm_connector->mst_mgr
That's zero-initialized memory, because the only connectors we'll ever
actually be doing DSC computations for are MST ports. Which have mst_mgr
zero-initialized, and instead have the correct topology mgr pointer located
at:
amdgpu_dm_connector->mst_port->mgr;
I'm a bit impressed that until now, this code has managed not to crash
anyone's systems! It does seem to cause a warning in LOCKDEP though:
This was causing the problems that appeared to have been introduced by:
commit 4d07b0bc4034 ("drm/display/dp_mst: Move all payload info into the atomic state")
This wasn't actually where they came from though. Presumably, before the
only thing we were doing with the topology mgr pointer was attempting to
grab mst_mgr->lock. Since the above commit however, we grab much more
information from mst_mgr including the atomic MST state and respective
modesetting locks.
This patch also implies that up until now, it's quite likely we could be
susceptible to race conditions when going through the MST topology state
for DSC computations since we technically will not have grabbed any lock
when going through it.
So, let's fix this by adjusting all the respective code paths to look at
the right pointer and skip things that aren't actual MST connectors from a
topology.
Looks like that we're accidentally dropping a pretty important return code
here. For some reason, we just return -EINVAL if we fail to get the MST
topology state. This is wrong: error codes are important and should never
be squashed without being handled, which here seems to have the potential
to cause a deadlock.
Lyude Paul [Mon, 14 Nov 2022 22:17:52 +0000 (17:17 -0500)]
drm/amdgpu/mst: Stop ignoring error codes and deadlocking
It appears that amdgpu makes the mistake of completely ignoring the return
values from the DP MST helpers, and instead just returns a simple
true/false. In this case, it seems to have come back to bite us because as
a result of simply returning false from
compute_mst_dsc_configs_for_state(), amdgpu had no way of telling when a
deadlock happened from these helpers. This could definitely result in some
kernel splats.
V2:
* Address Wayne's comments (fix another bunch of spots where we weren't
passing down return codes)
Roman Li [Mon, 14 Nov 2022 20:50:27 +0000 (15:50 -0500)]
drm/amd/display: Align dcn314_smu logging with other DCNs
[Why]
Assert on non-OK response from SMU is unnecessary.
It was replaced with respective log message on other asics
in the past with commit:
"drm/amd/display: Removing assert statements for Linux"
[How]
Remove assert and add dbg logging as on other DCNs.
Lucas Stach [Fri, 9 Sep 2022 09:20:44 +0000 (11:20 +0200)]
drm/etnaviv: switch to PFN mappings
There is no reason to use page based mappings, as the established
mappings are special driver mappings anyways and should not be
handled like normal pages.
Be consistent with what other drivers do and use raw PFN based
mappings.
Jani Nikula [Mon, 14 Nov 2022 11:17:09 +0000 (13:17 +0200)]
drm/edid/firmware: stop using a throwaway platform device
We've used a temporary platform device for firmware EDID loading since
it was introduced in commit da0df92b5731 ("drm: allow loading an EDID as
firmware to override broken monitor"), but there's no explanation why.
Using a temporary device does not play well with CONFIG_FW_CACHE=y,
which caches firmware images (e.g. on suspend) so that drivers can
request firmware when the system is not ready for it, and return the
images from the cache (e.g. during resume). This works automatically for
regular devices, but obviously not for a temporarily created device.
Stop using the throwaway platform device, and use the drm device
instead.
Note that this may still be problematic for cases where the display was
plugged in during suspend, and the firmware wasn't loaded and therefore
not cached before suspend.
fbdev: Add support for the nomodeset kernel parameter
Support the kernel's nomodeset parameter for all PCI-based fbdev
drivers that use aperture helpers to remove other, hardware-agnostic
graphics drivers.
The parameter is a simple way of using the firmware-provided scanout
buffer if the hardware's native driver is broken. The same effect
could be achieved with per-driver options, but the importance of the
graphics output for many users makes a single, unified approach
worthwhile.
With nomodeset specified, the fbdev driver module will not load. This
unifies behavior with similar DRM drivers. In DRM helpers, modules
first check the nomodeset parameter before registering the PCI
driver. As fbdev has no such module helpers, we have to modify each
driver individually.
The name 'nomodeset' is slightly misleading, but has been chosen for
historical reasons. Several drivers implemented it before it became a
general option for DRM. So keeping the existing name was preferred over
introducing a new one.
v2:
* print a warning if a driver does not init (Helge)
* wrap video_firmware_drivers_only() in helper
drm: Move nomodeset kernel parameter to drivers/video
Move the nomodeset kernel parameter to drivers/video to make it
available to non-DRM drivers. Adapt the interface, but keep the DRM
interface drm_firmware_drivers_only() to avoid churn within DRM. The
function should later be inlined into callers.
The parameter disables any DRM graphics driver that would replace a
driver for firmware-provided scanout buffers. It is an option to easily
fallback to basic graphics output if the hardware's native driver is
broken. Moving it to a more prominent location wil make it available
to fbdev as well.
v2:
* clarify the meaning of the nomodeset parameter (Javier)
drm/fb-helper: Schedule deferred-I/O worker after writing to framebuffer
Schedule the deferred-I/O worker instead of the damage worker after
writing to the fbdev framebuffer. The deferred-I/O worker then performs
the dirty-fb update. The fbdev emulation will initialize deferred I/O
for all drivers that require damage updates. It is therefore a valid
assumption that the deferred-I/O worker is present.
It would be possible to perform the damage handling directly from within
the write operation. But doing this could increase the overhead of the
write or interfere with a concurrently scheduled deferred-I/O worker.
Instead, scheduling the deferred-I/O worker with its regular delay of
50 ms removes load off the write operation and allows the deferred-I/O
worker to handle multiple write operations that arrived during the delay
time window.
drm/fb-helper: Perform damage handling in deferred-I/O helper
Call fb_dirty directly from drm_fb_helper_deferred_io() to avoid the
latency of running the damage worker.
The deferred-I/O helper drm_fb_helper_deferred_io() runs in a worker
thread at regular intervals as part of writing to mmaped framebuffer
memory. It used to schedule the fbdev damage worker to flush the
framebuffer. Changing this to flushing the framebuffer directly avoids
the latency introduced by the damage worker.
v2:
* remove fb_dirty from defio in separate patch (Daniel)
drm/fb-helper: Remove test for fb_dirty callback from deferred-I/O helper
The helper for processing deferred I/O on pages has no dependency on
the fb_dirty damge-handling callback; so remove the test. In practice,
deferred I/O is only used with damage handling and the damage worker
already guarantees the presence of the fb_dirty callback.
drm/fb-helper: Move dirty-fb update into helper function
Move the dirty-fb update from the damage-worker callback into the
new helper drm_fb_helper_fb_dirty(), so that it can run outside the
damage worker. This change will help to remove the damage worker
entirely. No functional changes.
Set the damage area in the new helper drm_fb_helper_add_damage_clip().
It can now be updated without scheduling the damage worker. This change
will help to remove the damage worker entirely. No functional changes.
Danilo Krummrich [Wed, 26 Oct 2022 15:59:34 +0000 (17:59 +0200)]
drm/arm/malidp: remove calls to drm_mode_config_cleanup()
drm_mode_config_init() simply calls drmm_mode_config_init(), hence
cleanup is automatically handled through registering
drm_mode_config_cleanup() with drmm_add_action_or_reset().
While at it, get rid of the deprecated drm_mode_config_init() and
replace it with drmm_mode_config_init() directly.
Danilo Krummrich [Wed, 26 Oct 2022 15:59:31 +0000 (17:59 +0200)]
drm/arm/malidp: replace drm->dev_private with drm_to_malidp()
Using drm_device->dev_private is deprecated. Since we've switched to
devm_drm_dev_alloc(), struct drm_device is now embedded in struct
malidp_drm, hence we can use container_of() to get the struct drm_device
instance instead.
Christian König [Wed, 26 Oct 2022 10:57:44 +0000 (12:57 +0200)]
drm/amdgpu: stop resubmitting jobs for GPU reset v2
Re-submitting IBs by the kernel has many problems because pre-
requisite state is not automatically re-created as well. In
other words neither binary semaphores nor things like ring
buffer pointers are in the state they should be when the
hardware starts to work on the IBs again.
Additional to that even after more than 5 years of
developing this feature it is still not stable and we have
massively problems getting the reference counts right.
As discussed with user space developers this behavior is not
helpful in the first place. For graphics and multimedia
workloads it makes much more sense to either completely
re-create the context or at least re-submitting the IBs
from userspace.
For compute use cases re-submitting is also not very
helpful since userspace must rely on the accuracy of
the result.
Because of this we stop this practice and instead just
properly note that the fence submission was canceled. The
only use case we keep the re-submission for now is SRIOV
and function level resets.
v2: as suggested by Sshaoyun stop resubmitting jobs even for SRIOV
This feature basically re-submits one job after another to
figure out which one was the one causing a hang.
This is obviously incompatible with gang-submit which requires
that multiple jobs run at the same time. It's also absolutely
not helpful to crash the hardware multiple times if a clean
recovery is desired.
For testing and debugging environments we should rather disable
recovery alltogether to be able to inspect the state with a hw
debugger.
Additional to that the sw implementation is clearly buggy and causes
reference count issues for the hardware fence.
Dan Carpenter [Tue, 15 Nov 2022 12:56:57 +0000 (15:56 +0300)]
amdgpu/pm: prevent array underflow in vega20_odn_edit_dpm_table()
In the PP_OD_EDIT_VDDC_CURVE case the "input_index" variable is capped at
2 but not checked for negative values so it results in an out of bounds
read. This value comes from the user via sysfs.
Fixes: d5bf26539494 ("drm/amd/powerplay: added vega20 overdrive support V3") Signed-off-by: Dan Carpenter <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
drm/amdgpu: Replace one-elements array with flex-array members
One-element arrays are deprecated, and we are replacing them with
flexible array members instead. So, replace one-element array with
flexible-array member in structs ATOM_I2C_VOLTAGE_OBJECT_V3,
ATOM_ASIC_INTERNAL_SS_INFO_V2, ATOM_ASIC_INTERNAL_SS_INFO_V3,
and refactor the rest of the code accordingly.
Important to mention is that doing a build before/after this patch
results in no functional binary output differences.
This helps with the ongoing efforts to tighten the FORTIFY_SOURCE
routines on memcpy() and help us make progress towards globally
enabling -fstrict-flex-arrays=3 [1].
YiPeng Chai [Fri, 4 Nov 2022 09:50:12 +0000 (17:50 +0800)]
drm/amdgpu: Fixed the problem that ras error can't be queried after gpu recovery is completed
Amdgpu_ras_set_error_query_ready is called at the start of
amdgpu_device_gpu_recover to disable query ras error, but the
code behind only enables query ras error in full reset path,
but not in soft reset path, emergency restart path and skip
the hardware reset path.
Alex Deucher [Fri, 11 Nov 2022 17:50:38 +0000 (12:50 -0500)]
drm/amdgpu: there is no vbios fb on devices with no display hw (v2)
If we enable virtual display functionality on parts with
no display hardware we can end up trying to check for and
reserve the vbios FB area on devices where it doesn't exist.
Check if display hardware is actually present on the hardware
before trying to reserve the memory.
Eric Huang [Mon, 14 Nov 2022 16:43:48 +0000 (11:43 -0500)]
drm/amdkfd: Fix a memory limit issue
It is to resolve a regression, which fails to allocate
VRAM due to no free memory in application, the reason
is we add check of vram_pin_size for memory limit, and
application is pinning the memory for Peerdirect, KFD
should not count it in memory limit. So removing
vram_pin_size will resolve it.
Perry Yuan [Mon, 10 Oct 2022 16:31:47 +0000 (00:31 +0800)]
drm/amdgpu: remove the DID of Vangogh from pciidlist
change the vangogh family to use IP discovery path to initialize IP
list, this needs to remove the DID from the PCI ID list to allow the IP
discovery path to set all the IP versions correctly.
Melissa Wen [Thu, 3 Nov 2022 18:45:00 +0000 (17:45 -0100)]
drm/amd/display: don't enable DRM CRTC degamma property for DCE
DM maps DRM CRTC degamma to DPP (pre-blending) degamma block, but DCE doesn't
support programmable degamma curve anywhere. Currently, a custom degamma is
accepted by DM but just ignored by DCE driver and degamma correction isn't
actually applied. There is no way to map custom degamma in DCE, therefore, DRM
CRTC degamma property shouldn't be enabled for DCE drivers.
Aric Cyr [Mon, 7 Nov 2022 01:37:34 +0000 (20:37 -0500)]
drm/amd/display: 3.2.212
DC version 3.2.212 brings along the following fixes:
- Fix DPIA AUX timeout issue
- Add HUBP surface flip interrupt handler
- Fix dpms_off issue when disabling bios mode
- Update SubVP Visual Confirm
- Program pipes for ODM when removing planes
- Use min transition for all SubVP plane add/remove
- Support parsing VRAM info v3.0 from VBIOS
- Enable dchub request limit for NV32
- Workaround to increase phantom pipe vactive in pipesplit
- Fix dcn3.1x mode validation on high bandwidth config
- Check context constructed for pipe split flags after pipesplit
- Align phantom pipe dst/src scaling with main for subvp
- Fix calculation for cursor CAB allocation
- Fix unbounded requesting for high pixel rate modes on dcn315
- Revert Disable DRR actions during state commit
- Fix array index out of bound error in bios parser
- Fix optc2_configure warning on dcn314
- Fix Subvp phantom pipe transition
- Fix prefetch calculations for dcn32
- Use uclk pstate latency for fw assisted mclk validation dcn32
- Set max for prefetch lines on dcn32
- Rework macros for DWB register access
- Adjust DP 8b10b LT exit behavior
Ilya Bakoulin [Wed, 2 Nov 2022 19:56:16 +0000 (15:56 -0400)]
drm/amd/display: Adjust DP 8b10b LT exit behavior
[Why]
Polling SINK_STATUS DPCD can cause some non-DP2.0 LTTPR hubs to hang and
no longer respond to AUX.
[How]
As a workaround for the problem hubs, and also to more closely match
DP spec 8b10b behavior, don't poll for intra-hop AUX reply indication
on 8b10b-encoded links.
drm/amd/display: rework macros for DWB register access
[Why]
A hack was used to access DWB register due to difference in the register
naming convention which was not compatible with existing SR/SRI* macros.
The additional macro needed were added to dwb ip specific header file
(dcnxx_dwb.h) instead of soc resource file (dcnxx_resource.c). Due to
this pattern, BASE macro had to be redefined in dcnxx_dwb.h, which in
turn needed us to undefine them in the resource file.
[How]
Add a separate macro for DWB access to the resource files that need it
instead of defining them in DWB ip header file. This will enable us to
reuse the BASE macro defined in the resource file.
Dillon Varone [Thu, 3 Nov 2022 22:33:38 +0000 (18:33 -0400)]
drm/amd/display: Fix prefetch calculations for dcn32
[Description]
Prefetch calculation loop was not exiting until utilizing all of vstartup if it
failed once. Locals need to be reset on each iteration of the loop.
[Description]
- When transitioning a pipe in use (non-phantom) to become
a phantom pipe, we must fully disable the pipe first before
doing any phantom pipe programming
- Move phantom OTG enablement to after "regular" front-end
programming sequence (including wait for vupdate)
- If a pipe is being transitioned into a phantom pipe,
fully disable it first
Roman Li [Thu, 3 Nov 2022 21:33:07 +0000 (17:33 -0400)]
drm/amd/display: Fix optc2_configure warning on dcn314
[Why]
dcn314 uses optc2_configure_crc() that wraps
optc1_configure_crc() + set additional registers
not applicable to dcn314.
It's not critical but when used leads to warning like:
WARNING: drivers/gpu/drm/amd/amdgpu/../display/dc/dc_helper.c
Call Trace:
<TASK>
generic_reg_set_ex+0x6d/0xe0 [amdgpu]
optc2_configure_crc+0x60/0x80 [amdgpu]
dc_stream_configure_crc+0x129/0x150 [amdgpu]
amdgpu_dm_crtc_configure_crc_source+0x5d/0xe0 [amdgpu]
Alvin Lee [Tue, 1 Nov 2022 20:10:49 +0000 (16:10 -0400)]
drm/amd/display: Don't check output BPP for phantom
[Description]
- If we're enabling phantom pipe for a high
link rate display we could
fail DML on the phantom pipe since it's
set to virtual signal
- Therefore don't consider Output BPP for
phantom pipe
George Shen [Wed, 2 Nov 2022 03:03:03 +0000 (23:03 -0400)]
drm/amd/display: Fix calculation for cursor CAB allocation
[Why]
The cursor size (in memory) is currently incorrectly calculated,
resulting not enough CAB being allocated for static screen cursor
in MALL refresh. This results in cursor image corruption.
[How]
Use cursor pitch instead of cursor width when calculating cursor size.
Update num cache lines calculation to use the result of the cursor size
calculation instead of manually recalculating again.
Charlene Liu [Wed, 2 Nov 2022 16:13:43 +0000 (12:13 -0400)]
drm/amd/display: fix dcn3.1x mode validation on high bandwidth config
[why]
1. correct dram_channel_width (was hard coded to 4 for 32bit)
2. use dm's is_hvm_enable status flag for hostvm_en input for dml.
3. add a function to override to all dcn3.1x.
George Shen [Tue, 1 Nov 2022 04:16:56 +0000 (00:16 -0400)]
drm/amd/display: Workaround to increase phantom pipe vactive in pipesplit
[Why]
Certain high resolution displays exhibit DCC line corruption with SubVP
enabled. This is likely due to insufficient DCC meta data buffered
immediately after the mclk switch.
[How]
Add workaround to increase phantom pipe vactive height by
meta_row_height number of lines, thus increasing the amount of meta data
buffered immediately after mclk switch finishes.
Jun Lei [Wed, 2 Nov 2022 19:19:37 +0000 (15:19 -0400)]
drm/amd/display: enable dchub request limit for DCN 3.2.x
[why]
HW recommend we enable this for better DF QoS for DCN 3.2.x
[how]
Add interfaces to hubbub to program register
Add initialization as part of init_hw
Add default settings for DCN 3.2.x
Add registry override for debug/tuning
George Shen [Wed, 2 Nov 2022 19:06:48 +0000 (15:06 -0400)]
drm/amd/display: Support parsing VRAM info v3.0 from VBIOS
[Why]
For DCN3.2 and DCN3.21, VBIOS has switch to using v3.0 of the VRAM
info struct. We should read and override the VRAM info in driver with
values provided by VBIOS to support memory downbin cases.
Alvin Lee [Wed, 2 Nov 2022 16:38:12 +0000 (12:38 -0400)]
drm/amd/display: Use min transition for all SubVP plane add/remove
[Description]
- Whenever disabling a phantom pipe, we must run through the
minimal transition sequence
- In the case where SetVisibility = false for the main pipe,
we also need to run through the min transtion when disabling
the phantom pipes
Dillon Varone [Tue, 1 Nov 2022 20:38:20 +0000 (16:38 -0400)]
drm/amd/display: Program pipes for ODM when removing planes
[WHY?]
Software state expects pipes to be configured for ODM, but due to the transition
occurring on a plane disable, the ODM enablement code is not run.
[HOW?]
Update ODM when removing a plane, and dynamic ODM is active. Also acquire pipe
lock when removing a plane.
David Galiffi [Sat, 22 Oct 2022 02:22:46 +0000 (22:22 -0400)]
drm/amd/display: Update SubVP Visual Confirm
[Why]
The visual confirm bar is displayed on all displays, but is only
updated on a flip. If an update causes SubVP to be disabled on
all displays, but there are no updates on the other displays, it
will lead to stale information displayed on the non-active
displays. An example of this is playing a fullscreen video on a
rotated display.
[How]
Add a procedure to update visual confirm color for all pipes when
committing a plane.
Zhongwei Zhang [Fri, 28 Oct 2022 05:40:31 +0000 (13:40 +0800)]
drm/amd/display: fix dpms_off issue when disabling bios mode
[Why]
disable_vbios_mode_if_required() will set dpms_off to false
during boot when pixel clk dismatches with driver requires.
This will cause extra backlight on and off if OS call 2
times setmode.
[How]
Set dpms_off to true to keep power_off and
let OS control BL by display's powerState.
Stylon Wang [Wed, 26 Oct 2022 13:00:40 +0000 (21:00 +0800)]
drm/amd/display: Fix invalid DPIA AUX reply causing system hang
[Why]
Some DPIA AUX replies have incorrect data length from original request.
This could lead to overwriting of destination buffer if reply length is
larger, which could cause invalid access to stack since many destination
buffers are declared as local variables.
[How]
Check for invalid length from DPIA AUX replies and trigger a retry if
reply length is not the same as original request. A DRM_WARN() dmesg log
is also produced.
It is possible to see that the flip was submitted, but DC never replied
back, which generates time-out issues. This is an indication that the
HUBP surface flip is missing. This commit fixes this issue by adding
hubp1_set_flip_int to DCN31.
projects / linux.git / log