Navid Emamdoost [Wed, 2 Oct 2019 03:46:07 +0000 (22:46 -0500)]
drm/amdgpu: fix multiple memory leaks in acp_hw_init
In acp_hw_init there are some allocations that needs to be released in
case of failure:
1- adev->acp.acp_genpd should be released if any allocation attemp for
adev->acp.acp_cell, adev->acp.acp_res or i2s_pdata fails.
2- all of those allocations should be released if
mfd_add_hotplug_devices or pm_genpd_add_device fail.
3- Release is needed in case of time out values expire.
Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Navid Emamdoost <navid.emamdoost@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Tue, 1 Oct 2019 21:45:27 +0000 (16:45 -0500)]
drm/amdgpu: don't increment vram lost if we are in hibernation
We reset the GPU as part of our hibernation sequence so we need
to make sure we don't mark vram as lost in that case.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=111879 Acked-by: Christian König <christian.koenig@amd.com> Reviewed-by: Huang Rui <ray.huang@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
shaoyunl [Tue, 1 Oct 2019 19:52:31 +0000 (15:52 -0400)]
drm/amdgpu : enable msix for amdgpu driver
We might used out of the msi resources in some cloud project
which have a lot gpu devices(including PF and VF), msix can
provide enough resources from system level view
Signed-off-by: shaoyunl <shaoyun.liu@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Jack Zhang [Thu, 26 Sep 2019 07:24:55 +0000 (15:24 +0800)]
drm/amd/amdgpu/sriov ip block setting of Arcturus
Add ip block setting for Arcturus SRIOV
1.PSP need to be initialized before IH.
2.SMU doesn't need to be initialized at kmd driver.
3.Arcturus doesn't support DCE hardware,it needs to skip
register access to DCE.
Signed-off-by: Jack Zhang <Jack.Zhang1@amd.com> Reviewed-by: Le Ma <Le.Ma@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Marek Olšák [Tue, 24 Sep 2019 21:53:25 +0000 (17:53 -0400)]
drm/amdgpu: return tcc_disabled_mask to userspace
UMDs need this for correct programming of harvested chips.
Signed-off-by: Marek Olšák <marek.olsak@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Lyude Paul [Thu, 26 Sep 2019 22:51:05 +0000 (18:51 -0400)]
drm/amdgpu/dm/mst: Use ->atomic_best_encoder
We are supposed to be atomic after all. We'll need this in a moment for
the next commit.
Acked-by: Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Harry Wentland <harry.wentland@amd.com> Signed-off-by: Lyude Paul <lyude@redhat.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Lyude Paul [Thu, 26 Sep 2019 22:51:03 +0000 (18:51 -0400)]
drm/amdgpu/dm/mst: Don't create MST topology managers for eDP ports
Reviewed-by: Harry Wentland <harry.wentland@amd.com> Signed-off-by: Lyude Paul <lyude@redhat.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Lyude Paul [Thu, 26 Sep 2019 22:51:04 +0000 (18:51 -0400)]
drm/amdgpu/dm/mst: Remove unnecessary NULL check
kfree() checks this automatically.
Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Lyude Paul <lyude@redhat.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Lyude Paul [Tue, 3 Sep 2019 20:46:01 +0000 (16:46 -0400)]
drm/amdgpu: Iterate through DRM connectors correctly
Currently, every single piece of code in amdgpu that loops through
connectors does it incorrectly and doesn't use the proper list iteration
helpers, drm_connector_list_iter_begin() and
drm_connector_list_iter_end(). Yeesh.
So, do that.
Cc: Juston Li <juston.li@intel.com> Cc: Imre Deak <imre.deak@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Harry Wentland <hwentlan@amd.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Lyude Paul <lyude@redhat.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Lyude Paul [Wed, 25 Sep 2019 21:52:48 +0000 (17:52 -0400)]
drm/amdgpu/dm: Resume short HPD IRQs before resuming MST topology
Since we're going to be reprobing the entire topology state on resume
now using sideband transactions, we need to ensure that we actually have
short HPD irqs enabled before calling drm_dp_mst_topology_mgr_resume().
So, do that.
Changes since v4:
* Fix typo in comments
Cc: Juston Li <juston.li@intel.com> Cc: Imre Deak <imre.deak@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Harry Wentland <hwentlan@amd.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Lyude Paul <lyude@redhat.com> Acked-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The patch c670707 drm/amd: Pass drm_device to kfd introduced this issue and
fix the following compiler error.
CC [M] drivers/gpu/drm/amd/amdgpu//../powerplay/smumgr/fiji_smumgr.o
drivers/gpu/drm/amd/amdgpu//amdgpu_amdkfd.c:746:6: error: conflicting types for ‘kgd2kfd_device_init’
bool kgd2kfd_device_init(struct kfd_dev *kfd,
^
In file included from drivers/gpu/drm/amd/amdgpu//amdgpu_amdkfd.c:23:0:
drivers/gpu/drm/amd/amdgpu//amdgpu_amdkfd.h:253:6: note: previous declaration of ‘kgd2kfd_device_init’ was here
bool kgd2kfd_device_init(struct kfd_dev *kfd,
^
scripts/Makefile.build:273: recipe for target 'drivers/gpu/drm/amd/amdgpu//amdgpu_amdkfd.o' failed
make[1]: *** [drivers/gpu/drm/amd/amdgpu//amdgpu_amdkfd.o] Error 1
Signed-off-by: Prike Liang <Prike.Liang@amd.com> Acked-by: Evan Quan <evan.quan@amd.com> Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amd/powerplay: bypass dpm_context null pointer check guard for some smu series
For now APU has no smu_dpm_context structure for containing default/current related dpm tables,
thus will needn't initialize smu_dpm_context to aviod APU null pointer issue.
Signed-off-by: Prike Liang <Prike.Liang@amd.com> Reviewed-by: Kevin Wang <kevin1.wang@amd.com> Reviewed-by: Evan Quan <evan.quan@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amd/amdgpu: add IH cg support on soc15 project
enable/disable IH clock gating on soc15 projects.
Signed-off-by: Kenneth Feng <kenneth.feng@amd.com> Reviewed-by: Kevin Wang <kevin1.wang@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Kevin Wang [Thu, 26 Sep 2019 08:22:13 +0000 (16:22 +0800)]
drm/amd/powerplay: add sensor lock support for smu
when multithreading access sysfs of amdgpu_pm_info at the sametime.
the swsmu driver cause smu firmware hang.
eg:
single thread access:
Message A + Param A ==> right
Message B + Param B ==> right
Message C + Param C ==> right
multithreading access:
Message A + Param B ==> error
Message B + Param A ==> error
Message C + Param C ==> right
the patch will add sensor lock(mutex) to avoid this error.
Signed-off-by: Kevin Wang <kevin1.wang@amd.com> Reviewed-by: Kenneth Feng <kenneth.feng@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Kevin Wang [Thu, 26 Sep 2019 08:16:41 +0000 (16:16 +0800)]
drm/amd/powerplay: change metrics update period from 1ms to 100ms
v2:
change period from 10ms to 100ms (typo error)
too high frequence to update mertrics table will cause smu firmware
error,so change mertrics table update period from 1ms to 100ms
(navi10, 12, 14)
Signed-off-by: Kevin Wang <kevin1.wang@amd.com> Reviewed-by: Kenneth Feng <kenneth.feng@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
get_atc_vmid_pasid_mapping_valid() is very similar to
get_atc_vmid_pasid_mapping_pasid(), so they can be merged into a new
function get_atc_vmid_pasid_mapping_info() to reduce register access
times. More importantly, getting the PASID and the valid bit atomically
with a single read fixes some potential race conditions where the
mapping changes between the two reads.
Signed-off-by: Yong Zhao <Yong.Zhao@amd.com> Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amdkfd: Query vmid pasid mapping through stored info for non HWS
Because we record the mapping under non HWS mode in the software,
we can query pasid through vmid using the stored mapping instead of
reading from ATC registers.
This also prepares for the defeatured ATC block in future ASICs.
Signed-off-by: Yong Zhao <Yong.Zhao@amd.com> Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Since KFD pasid starts from 0x8000 (32768 in decimal), it is better
perceived as a hex number. Meanwhile, change the pasid type from
unsigned int to uint16_t to be consistent throughout the code.
Signed-off-by: Yong Zhao <Yong.Zhao@amd.com> Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Kent Russell [Mon, 23 Sep 2019 01:20:14 +0000 (21:20 -0400)]
drm/amdgpu: Add SMUIO values for other I2C controller v2
These are the offsets for CKSVII2C1, and match up with the values
already added for CKSVII2C
v2: Don't remove some of the CSKVII2C values
Signed-off-by: Kent Russell <kent.russell@amd.com> Reviewed-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amdkfd: Use better name for sdma queue non HWS path
The old name is prone to confusion. The register offset is for a RLC queue
rather than a SDMA engine. The value is not a base address, but a
register offset.
Signed-off-by: Yong Zhao <Yong.Zhao@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Christian König [Thu, 19 Sep 2019 13:16:49 +0000 (15:16 +0200)]
drm/amdgpu: restrict hotplug error message
We should print the error only when we are hotplugged and crash
basically all userspace applications.
Signed-off-by: Christian König <christian.koenig@amd.com> Reviewed-by: Emily Deng <Emily.Deng@amd.com> Acked-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Christian König [Tue, 24 Sep 2019 11:29:27 +0000 (13:29 +0200)]
drm/amdgpu: once more fix amdgpu_bo_create_kernel_at
When CPU access is needed we should tell that to
amdgpu_bo_create_reserved() or otherwise the access is denied later on.
Signed-off-by: Christian König <christian.koenig@amd.com> Reviewed-by: Emily Deng <Emily.Deng@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Tao Zhou [Thu, 12 Sep 2019 05:38:44 +0000 (13:38 +0800)]
drm/amdgpu: update parameter of ras_ih_cb
change struct ras_err_data *err_data to void *err_data, align with
umc code and the callback's declaration in each ras block could
pay no attention to the structure type
Signed-off-by: Tao Zhou <tao.zhou1@amd.com> Reviewed-by: Guchun Chen <guchun.chen@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Monk Liu [Tue, 24 Sep 2019 08:08:00 +0000 (16:08 +0800)]
drm/amdgpu: fix an UMC hw arbitrator bug(v3)
issue:
the UMC6 h/w bug is that when MCLK is doing the switch
in the middle of a page access being preempted by high
priority client (e.g. DISPLAY) then UMC and the mclk switch
would stuck there due to deadlock
how:
fixed by disabling auto PreChg for UMC to avoid high
priority client preempting other client's access on
the same page, thus the deadlock could be avoided
v2:
put the patch in callback of UMC6
v3:
rename the callback to "init_registers"
Signed-off-by: Monk Liu <Monk.Liu@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Hawking Zhang <hawking.zhang@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Marek Olšák [Fri, 20 Sep 2019 02:04:43 +0000 (22:04 -0400)]
drm/amdgpu: remove gfx9 NGG
Never used.
Signed-off-by: Marek Olšák <marek.olsak@amd.com> Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Fri, 20 Sep 2019 19:43:44 +0000 (14:43 -0500)]
drm/amdgpu/atomfirmware: use proper index for querying vram type (v3)
The index is stored in scratch register 4 after asic init. Use
that index. No functional change since all asics in a family
use the same type of vram (G5, G6, HBM) and that is all we use
at the monent, but if we ever need to query other info, we will
now have the proper index.
v2: module array is variable sized, handle that.
v3: fix off by one in array handling
Reviewed-by: Xiaojie Yuan <xiaojie.yuan@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Shirish S [Tue, 24 Sep 2019 09:15:54 +0000 (14:45 +0530)]
drm/amdgpu/psp: silence response status warning
log the response status related error to the driver's
debug log since psp response status is not 0 even though
there was no problem while the command was submitted.
This warning misleads, hence this change.
Signed-off-by: Shirish S <shirish.s@amd.com> Acked-by: Huang Rui <ray.huang@amd.com> Reviewed-by: Feifei Xu <Feifei.Xu@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amd/amdgpu:Fix compute ring unable to detect hang.
When compute fence did not signal, compute ring cannot detect hardware hang
because its timeout value is set to be infinite by default.
In SR-IOV and passthrough mode, if user does not declare custome timeout
value for compute ring, then use gfx ring timeout value as default. So
that when there is a ture hardware hang, compute ring can detect it.
Signed-off-by: Jesse Zhang <zhexi.zhang@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amdkfd: Fix NULL pointer dereference for set_scratch_backing_va()
Currently this function pointer is missing for GFX10. Considering it is
a void function since GFX9, fix it by checking the function pointer
before dereferencing it.
Signed-off-by: Yong Zhao <Yong.Zhao@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Yong Zhao [Wed, 23 Jan 2019 01:09:17 +0000 (20:09 -0500)]
drm/amdkfd: Remove unnecessary pm_init() for non HWS mode
The packet manager is not needed for non HWS mode except Hawaii, so only
initialize it for Hawaii under non HWS mode. This will simplify debugging
under non HWS mode for all new asics, because it eliminates one variable
out of the equation in non HWS mode
Signed-off-by: Yong Zhao <Yong.Zhao@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Allen Pais <allen.pais@oracle.com> Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Thu, 19 Sep 2019 19:24:11 +0000 (14:24 -0500)]
drm/amdgpu/mn: fix documentation for amdgpu_mn_read_lock
Document the new parameter.
Fixes: 93065ac753e4 ("mm, oom: distinguish blockable mode for mmu notifiers") Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>