Missed this case previously. No need to do anything if the
device is already off. runtime pm will handle it.
Acked-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The sbios always powers up the dGPU on resume.
Acked-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Replace packet2's with packet0 writes to UVD_NO_OP. The
value written to UVD_NO_OP does not matter.
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Writes to this register are the preferred way to do NOPs.
Bump the driver version as well.
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
In an if block for (running == 0) running cannot be non-zero.
v2: agd: remove unused variable
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
In an if block for (running == 0) running cannot be non-zero.
v2: agd: remove unused variable
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
With the previous change, it's safe to let page flips take effect
anytime during a vertical blank period.
This can avoid delaying a flip by a frame in some cases where we get to
radeon_flip_work_func -> adev->mode_info.funcs->page_flip during a
vertical blank period.
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Now we can program a flip during a vertical blank period, if it's the
one targeted by the flip (or a later one). This allows simplifying
radeon_flip_work_func considerably.
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
When this code was written, we didn't retry DP aux transactions on any
error, which required retrying important transactions like this in
individual drivers. Since that's no longer the case, retrying here is
not necessary. As well, we retry any aux transaction on any error 32
times. 7 * 32 = 224, which means this loop causes us to retry grabbing
the dpcd 224 times. This is definitely far more then we actually need to
do.
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Lyude <cpaul@redhat.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Since it's normal for DRM to retry our aux transaction helpers multiple
times in a row, up to 32 times for each attempted transaction, we're
making a lot of noise that is no longer necessary now that DRM will just
print the return code we give it.
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Lyude <cpaul@redhat.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Fixes hangs under memory pressure, e.g. running the piglit test
tex3d-maxsize concurrently with other tests.
Fixes: 17d33bc9d6 ("drm/ttm: drop waiting for idle in ttm_bo_evict.")
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
alloc_workqueue replaces deprecated create_singlethread_workqueue().
Each hardware CRTC has a single flip work queue.
When a radeon_flip_work_func item is queued, it needs to be executed
ASAP because even a slight delay may cause the flip to be delayed by
one refresh cycle.
Hence, a dedicated workqueue with WQ_HIGHPRI set, has been used here
since a delay can cause the outcome to miss the refresh cycle.
Since there are only a fixed number of work items, explicit concurrency
limit is unnecessary here.
Reviewed-by: Christian König <christian.koenig@amd.com>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Bhaktipriya Shridhar <bhaktipriya96@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The drm_gem_object_unreference_unlocked() function tests whether
its argument is NULL and then returns immediately.
Thus the test around the call is not needed.
This issue was detected by using the Coccinelle software.
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The message is harmless and confusing. On PX systems,
there is one ATIF method, but potentially multiple GPUs
leading to an error on the GPU with no backlight control.
Bug: https://bugzilla.kernel.org/show_bug.cgi?id=115011
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: Hans de Goede <jwrdegoede@fedoraproject.org>
commit d967be9b80 ("drm/radeon/ci: disable needless sclk changes")
introduces an unreachable if(C != C) conditional code section
flagged by coccinelle script bad_conditional.cocci:
Add a comment to make it clear that this is intentional.
Fixes: d967be9b80 ("drm/radeon/ci: disable needless sclk changes")
Signed-off-by: Nicholas Mc Guire <hofrat@osadl.org>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
If we do it at enable time, it's too late for the feature
checks.
v2: drop .init setting as per Peter's comments
bug: https://bugzilla.kernel.org/show_bug.cgi?id=115321
Reviewed-by: Peter Wu <peter@lekensteyn.nl>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: Peter Wu <peter@lekensteyn.nl>
A few more patches for 4.8. Mostly bug fixes and some prep work
for iceland powerplay support. I have a couple polaris patches and
Edward's misc cleanups that require a merge with Linus'. I don't know
if you are planning a merge anytime soon.
[airlied: fixed up endian vs 32-bit change in ppatomctrl]
* 'drm-next-4.8' of git://people.freedesktop.org/~agd5f/linux: (26 commits)
drm/amdgpu: comment out unused defaults_bonaire_pro static const structures to fix the build
drm/amdgpu: temporary comment out unused static const structures to fix the build
drm/amdgpu: S3 resume fail on Polaris10
drm/amd/powerplay: add pp_tables_get_response_times function in process pptables
drm/amd/powerplay: fix the incorrect return value
drm/amd/powerplay: add atomctrl_get_voltage_evv function in ppatomctrl
drm/amdgpu: add new definitions into ppsmc.h for iceland
drm/amd/powerplay: add SMU register macro for future use
drm/amdgpu: add ucode_start_address into cgs_firmware_info
drm/amdgpu: no need load microcode at sdma if powerplay is enabled
drm/amdgpu: rename smumgr to smum for dpm
drm/amdgpu: disable GFX PG on CZ/BR/ST
drivers: gpu: drm: amd: powerplay: hwmgr: Remove unused variable
drm/amdgpu: return -ENOSPC when running out of UVD handles
drm/amdgpu: trace need_flush in grab_vm as well
drm/amdgpu: always signal all fences
drm/amdgpu: check flush fence context instead of same ring v2
drm/radeon: support backlight control for UNIPHY3
drm/amdgpu: support backlight control for UNIPHY3
drm/amdgpu: remove usec timeout loop from IB tests
...
I recovered dri-devel backlog from my vacation, more misc stuff:
- of_put_node fixes from Peter Chen (not all yet)
- more patches from Gustavo to use kms-native drm_crtc_vblank_* funcs
- docs sphinxification from Lukas Wunner
- bunch of fixes all over from Dan Carpenter
- more follow up work from Chris register/unregister rework in various
places
- vgem dma-buf export (for writing testcases)
- small things all over from tons of different people
* tag 'topic/drm-misc-2016-07-14' of git://anongit.freedesktop.org/drm-intel: (52 commits)
drm: Don't overwrite user ioctl arg unless requested
dma-buf/sync_file: improve Kconfig description for Sync Files
MAINTAINERS: add entry for the Sync File Framework
drm: Resurrect atomic rmfb code
drm/vgem: Use PAGE_KERNEL in place of x86-specific PAGE_KERNEL_IO
qxl: silence uninitialized variable warning
qxl: check for kmap failures
vga_switcheroo: Sphinxify docs
drm: Restore double clflush on the last partial cacheline
gpu: drm: rockchip_drm_drv: add missing of_node_put after calling of_parse_phandle
gpu: drm: sti_vtg: add missing of_node_put after calling of_parse_phandle
gpu: drm: sti_hqvdp: add missing of_node_put after calling of_parse_phandle
gpu: drm: sti_vdo: add missing of_node_put after calling of_parse_phandle
gpu: drm: sti_compositor: add missing of_node_put after calling of_parse_phandle
drm/tilcdc: use drm_crtc_handle_vblank()
drm/rcar-du: use drm_crtc_handle_vblank()
drm/nouveau: use drm_crtc_handle_vblank()
drm/atmel: use drm_crtc_handle_vblank()
drm/armada: use drm_crtc_handle_vblank()
drm: make drm_vblank_count_and_time() static
...
DRM_CONNECTOR_POLL_CONNECT only enables polling for connections, not
disconnections. Because of this, we end up losing hotplug polling for
analog connectors once they get connected.
Easy way to reproduce:
- Grab a machine with a radeon GPU and a VGA port
- Plug a monitor into the VGA port, wait for it to update the connector
from disconnected to connected
- Disconnect the monitor on VGA, a hotplug event is never sent for the
removal of the connector.
Originally, only using DRM_CONNECTOR_POLL_CONNECT might have been a good
idea since doing VGA polling can sometimes result in having to mess with
the DAC voltages to figure out whether or not there's actually something
there since VGA doesn't have HPD. Doing this would have the potential of
showing visible artifacts on the screen every time we ran a poll while a
VGA display was connected. Luckily, radeon_vga_detect() only resorts to
this sort of polling if the poll is forced, and DRM's polling helper
doesn't force it's polls.
Additionally, this removes some assignments to connector->polled that
weren't actually doing anything.
Cc: stable@vger.kernel.org
Signed-off-by: Lyude <cpaul@redhat.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
It isn't used and not waiting for the GPU after scheduling a move is
actually quite dangerous.
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
When we want to pipeline accelerated moves we need to wait in the fallback path.
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Wait for idle before moving the BO in all drivers implementing
an accelerated move function.
This should keep the current behavior when removing the pre move wait.
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Align to the jump table offset. May fix hangs on some
asics with GFX PG enabled.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Acked-by: Tom St Denis <tom.stdenis@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Until Dave's patch to support the new hybrid gfx ACPI method goes
upstream, we can fallback to the old ATPX method which seems to
still work.
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The ATPX power control method does this for you.
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The platform d3 cold is used to power down the dGPU.
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
hybrid graphics in this case refers to systems which use the new
platform d3 cold ACPI methods as opposed to ATPX for dGPU power
control.
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Now that we handle this correctly, there is no need to force
it.
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
On PX systems without dGPU power control, use PCI_D3hot.
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The runtime pm sequence is different depending on whether or
not the platform supports ATPX dGPU power control.
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
ATPX dGPU power control requires a 200ms delay between
power off and on. This should fix dGPU failures on
resume from power off.
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
The presence of the power control method should be determined
via the presence of the method in function 0. However, some
sbioses only set the appropriate bits in function 1 so use
then to override a missing power control function.
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Windows 10 (and some 8.1) systems use standardized
ACPI calls for hybrid laptops to control dGPU power.
Detect those cases and disable the AMD specific ATPX
power control.
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Edmondo Tommasina <edmondo.tommasina@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The power tables on some variants require different firmware.
This may fix stability issues on some newer CI parts.
bug:
https://bugs.freedesktop.org/show_bug.cgi?id=91880
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The power tables on some variants require different firmware.
This may fix stability issues on some newer SI parts.
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iQEcBAABAgAGBQJXcHi9AAoJEHm+PkMAQRiGSJ0H/2o4t9VWYmhyPC1sdIHoCExJ
P4tBrcZYBmKcsOmIfnJDa5g/+IdhouEUM0v0fHPogS2UUWT9eRuJWYD3sY+HpEQ+
heKTli8X73gsFB25odeIbIt0jAoSiiMYWDrWqLNsuUV1tjEYVA8rH0SM94FiOC/5
7WVWXLTuH+Rm7JHP18BnKxmMMbzrTFmwisLMqFKyfZRRSlS+/ix7iLUNO9AFa39B
YHxNPihLrZ0oONyCOAQoHTIXXrw0cQbxV2utg3vnMcCZdme2xOn+iXMntTSKfZ39
iC9/T0vsO3R6OrRo2aDZAnCPUAniXnMEIhrKG37WMyXpj6cucZ/2QiNXcXviGV4=
=iLte
-----END PGP SIGNATURE-----
Back-merge tag 'v4.7-rc5' into drm-next
Linux 4.7-rc5
The fsl-dcu pull needs -rc3 so go to -rc5 for now.
radeon leaks a runtime pm ref if at least one CRTC is enabled on unload.
The ref is taken by radeon_crtc_set_config() and held as long as a CRTC
is in use. Fix by turning off all CRTCs on unload.
Cc: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/f706b9db319e2bdaf153966a2b95a5d80f67e09b.1465392124.git.lukas@wunner.de
The PCI core calls pm_runtime_forbid() on device probe in pci_pm_init(),
making this the default state when radeon is loaded.
radeon_driver_load_kms() therefore calls pm_runtime_allow(), but there's
no pm_runtime_forbid() in radeon_driver_unload_kms() to balance it. Add
it so that we leave the device in the same state that we found it.
This isn't a bug, it's just good housekeeping. When radeon is first
loaded with runpm=1, then unloaded and loaded again with runpm=0,
pm_runtime_forbid() will be called from radeon_pmops_runtime_idle() or
radeon_pmops_runtime_suspend(), so the behaviour is correct. If there
ever is a third party driver for AMD cards, this commit avoids that it
has to clean up behind radeon.
Cc: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/25a3e20b786fd66b10f40fa24c61dd36c33270da.1465392124.git.lukas@wunner.de
radeon_device_init() returns an error if either of the two calls to
radeon_init() fail. One level up in the call stack,
radeon_driver_load_kms() will then skip runtime pm initialization and
call radeon_driver_unload_kms(), which acquires a runtime pm ref that
is leaked.
Balance by releasing a runtime pm ref in the error path of
radeon_device_init().
Fixes: 10ebc0bc09 ("drm/radeon: add runtime PM support (v2)")
Cc: Dave Airlie <airlied@redhat.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/fa5bb977c1fe00474acedae5b03232dbf0b49410.1465392124.git.lukas@wunner.de
radeon_driver_load_kms() calls pm_runtime_put_autosuspend() if
radeon_is_px(dev), but radeon_driver_unload_kms() calls
pm_runtime_get_sync() unconditionally. We therefore leak a runtime pm
ref whenever radeon is unloaded on a non-PX machine or if runpm=0. The
GPU will subsequently never runtime suspend after loading radeon again.
Fix by taking the runtime pm ref under the same condition that it was
released on driver load.
Fixes: 10ebc0bc09 ("drm/radeon: add runtime PM support (v2)")
Cc: Dave Airlie <airlied@redhat.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/aaf71106c042126817aeca8b8e54ed468ab61ef7.1465392124.git.lukas@wunner.de
When executing in a PCI passthrough based virtuzliation environment, the
hypervisor will usually attempt to send a PCIe bus reset signal to the
ASIC when the VM reboots. In this scenario, the card is not correctly
initialized, but we still consider it to be posted. Therefore, in a
passthrough based environemnt we should always post the card to guarantee
it is in a good state for driver initialization.
Ported from amdgpu commit:
amdgpu: fix asic initialization for virtualized environments
Cc: Andres Rodriguez <andres.rodriguez@amd.com>
Cc: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Alex Deucher