Monk Liu [Fri, 26 Aug 2016 06:15:03 +0000 (14:15 +0800)]
drm/amdgpu: new method to sync ce&de
sync switch buffer scheme with windows kmd for gfx v8,
step2:
Insert 128NOP after&before VM flush to prevent CE vm fault.
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Maruthi Srinivas Bayyavarapu [Tue, 30 Aug 2016 09:46:36 +0000 (15:16 +0530)]
drm/amdgpu: add VCE VM mode support
This adds VCE VM mode support from Stoney onwards. Session tracking
is an open issue, yet to be supported.
v2: Fixed warnings from checkpatch.pl
Signed-off-by: Maruthi Bayyavarapu <maruthi.bayyavarapu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Tom St Denis [Fri, 2 Sep 2016 12:19:15 +0000 (08:19 -0400)]
drm/amd/amdgpu: Merge get_wptr functions in gfx6
Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Tom St Denis [Fri, 2 Sep 2016 12:15:42 +0000 (08:15 -0400)]
drm/amd/amdgpu: Merge get_rptr functions in gfx8
Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Tom St Denis [Fri, 2 Sep 2016 12:14:42 +0000 (08:14 -0400)]
drm/amd/amdgpu: Merge get_rptr functions in gfx7
Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Tom St Denis [Fri, 2 Sep 2016 12:13:48 +0000 (08:13 -0400)]
drm/amd/amdgpu: Merge get_rptr functions in gfx6
Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Tom St Denis [Thu, 1 Sep 2016 17:39:37 +0000 (13:39 -0400)]
drm/amd/amdgpu: Various tidy ups for gfx6
Various whitespace and logical simplifications for gfx6.
Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Tom St Denis [Thu, 1 Sep 2016 17:31:35 +0000 (13:31 -0400)]
drm/amd/amdgpu: Simplify mask creation in gfx6
Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Tom St Denis [Thu, 1 Sep 2016 17:29:49 +0000 (13:29 -0400)]
drm/amd/amdgpu: Add GRBM lock to various SI functions
Add missing lock around SE/SH/INSTANCE selections.
Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Rex Zhu [Wed, 31 Aug 2016 18:30:42 +0000 (02:30 +0800)]
drm/amd/powerplay: original power state table should not be changed.
power state table was set based on vbios and should not be changed.
when client need to change power state, just make a copy and send to
smu.
Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Rex Zhu [Wed, 31 Aug 2016 10:13:08 +0000 (18:13 +0800)]
drm/amd/powerplay: refine struct name for coding style reason.
delete useless member.
Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Rex Zhu [Wed, 31 Aug 2016 09:27:11 +0000 (17:27 +0800)]
drm/amd/powerplay: add vce state tables initialize for ppt v1.
Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Rex Zhu [Wed, 31 Aug 2016 09:25:49 +0000 (17:25 +0800)]
drm/amd/powerplay: refine struct name for code style reason.
Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Christian König [Thu, 1 Sep 2016 10:13:18 +0000 (12:13 +0200)]
drm/amdgpu: prevent command submission failures under memory pressure v2
As last resort try to evict BOs from the current working set into other
memory domains. This effectively prevents command submission failures when
VM page tables have been swapped out.
v2: fix typos
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Christian König [Wed, 31 Aug 2016 15:28:11 +0000 (17:28 +0200)]
drm/amdgpu: only try again if we actually run into -ENOMEM
All other errors can't be fixed by using a different memory domain.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Mon, 22 Aug 2016 18:29:44 +0000 (14:29 -0400)]
drm/radeon: wire up a pci shutdown callback
Normally on shutdown or reboot we don't care about necessarily
making sure the hw is in a good state because the system is about
to be powered down or reset. However, after a shutdown or reboot
in a VM, it's best to tear down the hw properly otherwise there
can be problems with the next VM use.
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Mon, 22 Aug 2016 17:50:22 +0000 (13:50 -0400)]
drm/amdgpu: wire up a pci shutdown callback
Normally on shutdown or reboot we don't care about necessarily
making sure the hw is in a good state because the system is about
to be powered down or reset. However, after a shutdown or reboot
in a VM, it's best to tear down the hw properly otherwise there
can be problems with the next VM use.
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Wed, 31 Aug 2016 21:23:31 +0000 (17:23 -0400)]
drm/radeon: handle runtime pm in drm pre/post close
Close was not handled correctly.
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Sat, 27 Aug 2016 16:27:24 +0000 (12:27 -0400)]
drm/amdgpu: handle runtime pm in drm pre/post close
Close was not handled correctly.
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Wed, 31 Aug 2016 21:27:03 +0000 (17:27 -0400)]
drm/radeon: handle runtime pm correctly in amdgpu_driver_open_kms
Need to fix the error paths.
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Sat, 27 Aug 2016 16:30:25 +0000 (12:30 -0400)]
drm/amdgpu: handle runtime pm correctly in amdgpu_driver_open_kms
Need to fix the error paths.
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Wed, 31 Aug 2016 21:31:30 +0000 (17:31 -0400)]
drm/radeon: skip suspend/resume on DRM_SWITCH_POWER_DYNAMIC_OFF
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>
Alex Deucher [Wed, 31 Aug 2016 21:34:23 +0000 (17:34 -0400)]
drm/radeon: set runtime pm state to active on resume
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>
Alex Deucher [Sat, 27 Aug 2016 16:57:20 +0000 (12:57 -0400)]
drm/amdgpu: skip suspend/resume on DRM_SWITCH_POWER_DYNAMIC_OFF
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>
Alex Deucher [Sat, 27 Aug 2016 18:53:08 +0000 (14:53 -0400)]
drm/amdgpu: set runtime pm state to active on resume
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>
Alex Deucher [Thu, 1 Sep 2016 18:10:46 +0000 (14:10 -0400)]
drm/amdgpu/gfx6: drop some dead code
The mqd is only used on CI and newer.
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Chunming Zhou [Tue, 30 Aug 2016 08:36:25 +0000 (16:36 +0800)]
drm/amdgpu: add gart recovery by gtt list V2
V2:
a. gart recovery should be ahead of ring test.
b. rename to amdgpu_ttm_recover_gart
Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Chunming Zhou [Tue, 30 Aug 2016 08:13:10 +0000 (16:13 +0800)]
drm/amdgpu: link all gtt when binding them V2
V2:
spin lock instead of mutex for gtt list
Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Ken Wang [Thu, 21 Jan 2016 09:33:00 +0000 (17:33 +0800)]
drm/amdgpu: add si pciids v2
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Ken Wang <Qingqing.Wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Mon, 1 Aug 2016 20:05:47 +0000 (16:05 -0400)]
drm/amdgpu/si: Add updated smc firmware for SI kickers
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Ken Wang [Tue, 24 May 2016 13:02:53 +0000 (21:02 +0800)]
drm/amdgpu: Add SI Family information
Signed-off-by: Ken Wang <Qingqing.Wang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Maruthi Srinivas Bayyavarapu [Tue, 26 Apr 2016 15:05:36 +0000 (20:35 +0530)]
drm/amdgpu: enable SI DPM
Signed-off-by: Maruthi Bayyavarapu <maruthi.bayyavarapu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Huang Rui [Wed, 31 Aug 2016 05:23:18 +0000 (13:23 +0800)]
drm/amdgpu: use DRM print instead of printk
Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Maruthi Bayyavarapu [Mon, 1 Aug 2016 16:42:32 +0000 (12:42 -0400)]
drm/amdgpu: add SI DPM support (v4)
v2: corrected register offset shift
v3: rebase fixes
v4: fix firmware paths
add SI smc firmware versions for sysfs dump
remove unused function forward define
fix the tahiti specific value of DEEP_SLEEP_CLK_SEL field
fix to miss adding thermal controller
use vram_type instead of checking mem_gddr5 flag
fix incorrect index of CG_FFCT_0 register
fix incorrect reading method at si_get_current_pcie_speed
Signed-off-by: Maruthi Bayyavarapu <maruthi.bayyavarapu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Maruthi Srinivas Bayyavarapu [Tue, 26 Apr 2016 15:01:54 +0000 (20:31 +0530)]
drm/amdgpu: add SI SMC support
Signed-off-by: Maruthi Bayyavarapu <maruthi.bayyavarapu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Maruthi Srinivas Bayyavarapu [Tue, 26 Apr 2016 14:54:38 +0000 (20:24 +0530)]
drm/amdgpu: add si dpm support in amdgpu_atombios
v2: renamed _atom_ to _atombios_ for consistency
added ulClockParams to _COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V3 and
_COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V5 to avoid build break
Signed-off-by: Maruthi Bayyavarapu <maruthi.bayyavarapu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Ken Wang [Thu, 21 Jan 2016 11:08:55 +0000 (19:08 +0800)]
drm/amdgpu: add si specific logic into the device initialize function v3
v3: guard doorbell_fini as well
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Ken Wang <Qingqing.Wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Ken Wang [Thu, 21 Jan 2016 09:29:41 +0000 (17:29 +0800)]
drm/amdgpu: add si ip blocks setup v3
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Ken Wang <Qingqing.Wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Ken Wang [Tue, 19 Jan 2016 06:19:53 +0000 (14:19 +0800)]
drm/amdgpu: add all the components for si into Makefile/kconfig v3
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Ken Wang <Qingqing.Wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Huang Rui [Wed, 31 Aug 2016 05:23:25 +0000 (13:23 +0800)]
drm/amdgpu: introduce pcie port read/write entry
This patch adds pcie port read/write entry, because it will be also
used on si dpm part.
Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Ken Wang [Tue, 19 Jan 2016 06:08:49 +0000 (14:08 +0800)]
drm/amdgpu: add si implementation v10
v5: rebase fixes
v6: add mgcg arrays
v7: rebase fixes
v8: rebase fixes
v9: add get_disabled_bios(), make get_xclk static
v10: fix oland and hainan asic specific handle at si_program_aspm
Signed-off-by: Ken Wang <Qingqing.Wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Ken Wang [Tue, 19 Jan 2016 06:05:23 +0000 (14:05 +0800)]
drm/amdgpu: add DMA implementation for si v8
v4: rebase fixes
v5: use the generic nop fill
v6: rebase fixes
v7: rebase fixes
copy count fixes from Jonathan
general cleanup
add fill buffer implementation
v8: adapt write_pte and copy_pte to latest changes
Signed-off-by: Ken Wang <Qingqing.Wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Ken Wang [Tue, 19 Jan 2016 06:04:28 +0000 (14:04 +0800)]
drm/amdgpu: add graphic pipeline implementation for si v8
v5: rebase fixes
v6: rebase fixes
v7: rebase fixes
fix tile reg offset as noticed by Jonathan
Drop some debugging remnants
v8: add gfx v6 firmware versions for sysfs dump
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Ken Wang <Qingqing.Wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Ken Wang [Tue, 19 Jan 2016 07:07:29 +0000 (15:07 +0800)]
drm/amdgpu: atombios change for dce6 to work v3
v3: white space fixes
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Ken Wang <Qingqing.Wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Ken Wang [Tue, 19 Jan 2016 06:03:24 +0000 (14:03 +0800)]
drm/amdgpu: add display controller implementation for si v10
v4: rebase fixups
v5: more fixes based on dce8 code
v6: squash in dmif offset fix
v7: rebase fixups
v8: rebase fixups, drop some debugging remnants
v9: fix BE build
v10: include Marek's tiling fixes, add support for
page_flip_target, set MASTER_UDPATE_MODE=0,
fix cursor
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Ken Wang <Qingqing.Wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Ken Wang [Tue, 19 Jan 2016 06:02:14 +0000 (14:02 +0800)]
drm/amdgpu: add interupt handler implementation for si v3
v3: rebase fixups
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Ken Wang <Qingqing.Wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Ken Wang [Tue, 19 Jan 2016 06:01:19 +0000 (14:01 +0800)]
drm/amdgpu: add graphic memory controller implementation for si v7
v4: rebase fixups
v5: rebase fixups
v5: rebase fixups
v6: rebase fixups for gart size changes
v7: add gmc v6 firmware versions for sysfs dump
Signed-off-by: Ken Wang <Qingqing.Wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Ken Wang [Tue, 19 Jan 2016 05:53:10 +0000 (13:53 +0800)]
drm/amdgpu: add si header files v4
v4: drop unused DCE6 macro
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Ken Wang <Qingqing.Wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Ken Wang [Thu, 21 Jan 2016 09:00:06 +0000 (17:00 +0800)]
drm/amdgpu: add SI asics types v2
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Ken Wang <Qingqing.Wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Monk Liu [Fri, 26 Aug 2016 06:12:37 +0000 (14:12 +0800)]
drm/amdgpu:add switch buffer to end of CS (v2)
sync switch buffer scheme with windows kmd for gfx v8,
step1: append a switch_buffer to the end of CS
v2:rebase on latest staging
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Tom St Denis [Mon, 29 Aug 2016 12:39:29 +0000 (08:39 -0400)]
drm/amd/amdgpu: debugfs SMC addresses are byte addresses
Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Jordan Lazare [Fri, 26 Aug 2016 21:10:28 +0000 (17:10 -0400)]
drm/amd/powerplay: Remove unused variable causing compile warning
If treating warnings as errors this causes a build error
Signed-off-by: Jordan Lazare <Jordan.Lazare@amd.com>
Reviewed-by: Alex Deucher <Alexander.Deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Tom St Denis [Fri, 26 Aug 2016 16:44:01 +0000 (12:44 -0400)]
drm/amd/powerplay: Only load SDMA0/MEC firmware once on Stoney (v2)
Only load the SDMA0/MEC1 firmware once in the Carrizo SMU manager
driver.
(v2) Avoid loading SDMA0 twice too.
Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Tom St Denis [Fri, 26 Aug 2016 16:43:14 +0000 (12:43 -0400)]
drm/amd/powerplay: Fix up return codes in cz SMU manager
Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Tom St Denis [Fri, 26 Aug 2016 16:42:36 +0000 (12:42 -0400)]
drm/amd/powerplay: Tidy up cz SMU manager
Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Tom St Denis [Fri, 26 Aug 2016 14:29:19 +0000 (10:29 -0400)]
drm/amd/powerplay: Fix CZ SMU firmware load check (v4)
Remove an errant return in the middle of the check
function as well as check for success in the start
function.
(v2) Add return check to smu_load_fw()
(v3) Don't return early if SMU load check fails
(v4) No returns! :-)
Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Marek Olšák [Wed, 17 Aug 2016 21:49:27 +0000 (23:49 +0200)]
drm/amdgpu: throttle buffer migrations at CS using a fixed MBps limit (v2)
The old mechanism used a per-submission limit that didn't take previous
submissions within the same time frame into account. It also filled VRAM
slowly when VRAM usage dropped due to a big eviction or buffer deallocation.
This new method establishes a configurable MBps limit that is obeyed when
VRAM usage is very high. When VRAM usage is not very high, it gives
the driver the freedom to fill it quickly. The result is more consistent
performance.
It can't keep the BO move rate low if lots of evictions are happening due
to VRAM fragmentation, or if a big buffer is being migrated.
The amdgpu.moverate parameter can be used to set a non-default limit.
Measurements can be done to find out which amdgpu.moverate setting gives
the best results.
Mainly APUs and cards with small VRAM will benefit from this. For F1 2015,
anything with 2 GB VRAM or less will benefit.
Some benchmark results - F1 2015 (Tonga 2GB):
Limit MinFPS AvgFPS
Old code: 14 32.6
128 MB/s: 28 41
64 MB/s: 15.5 43
32 MB/s: 28.7 43.4
8 MB/s: 27.8 44.4
8 MB/s: 21.9 42.8 (different run)
Random drops in Min FPS can still occur (due to fragmented VRAM?), but
the average FPS is much better. 8 MB/s is probably a good limit for this
game & the current VRAM management. The random FPS drops are still to be
tackled.
v2: use a spinlock
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>
Tom St Denis [Thu, 25 Aug 2016 17:02:06 +0000 (13:02 -0400)]
drm/amd/amdgpu: Tidy up cz_dpm.c
Various minor formatting changes.
Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Tom St Denis [Thu, 25 Aug 2016 16:41:15 +0000 (12:41 -0400)]
drm/amd/amdgpu: Clean up memory leak in cz_dpm_init().
If init fails free up any allocated memory.
Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Tom St Denis [Thu, 25 Aug 2016 16:16:24 +0000 (12:16 -0400)]
drm/amd/amdgpu: Fix memleak in cz_parse_power_table()
If one of the entries fails to be allocated then free
all of the previous entries before freeing the array which
holds their pointers.
Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Rex Zhu [Wed, 24 Aug 2016 11:39:06 +0000 (19:39 +0800)]
drm/amdgpu: refine uvd gate logic for CI.
uvd dpm will be controlled by uvd.
dpm just disable uvd dpm in case of suspend when play video.
due to the new logic of uvd_begin_use/end_use,
if disable uvd dpm in late init, will have no chance to
enable uvd dpm after resume until play video again.
Change-Id: I70e3d7efe63edad37f26e6c5ea089da1135c0ab5
Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Rex Zhu [Mon, 22 Aug 2016 12:47:28 +0000 (20:47 +0800)]
drm/amd/powerplay: simplify struct amd_pp_init.
delete the members not needed when amd_powerplay_init.
Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Rex Zhu [Mon, 22 Aug 2016 12:47:01 +0000 (20:47 +0800)]
drm/amd/powerplay: get system info by cgs interface.
Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Rex Zhu [Mon, 22 Aug 2016 12:48:13 +0000 (20:48 +0800)]
drm/amdgpu: add support for getting sub_device id and
sub_vendor_id in cgs interface.
Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Rex Zhu [Thu, 18 Aug 2016 10:27:14 +0000 (18:27 +0800)]
drm/amdgpu: add vce bypass mode for tonga.
fix issue that encode test failed on the second time when
vce dpm enabled on tonga.
Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Rex Zhu [Thu, 18 Aug 2016 05:02:46 +0000 (13:02 +0800)]
drm/amd/powerplay: add bypass mode for vce 2.0.
fix issue after vce encode, the eclk stay high.
Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Rex Zhu [Thu, 18 Aug 2016 04:37:02 +0000 (12:37 +0800)]
drm/amdgpu: refine function name for consistency
Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Wed, 24 Aug 2016 21:15:33 +0000 (17:15 -0400)]
drm/amdgpu/vce3: add support for third vce ring
Not of much use at the moment (we don't really use
the second ring either), but may be useful later.
Reviewed-by: JimQu <Jim.Qu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Wed, 24 Aug 2016 20:56:21 +0000 (16:56 -0400)]
drm/amdgpu: track the number of vce rings
Rather than using a hardcoded value. This allows
different versions to expose more or less rings.
No functional change.
Reviewed-by: JimQu <Jim.Qu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Tue, 23 Aug 2016 17:25:49 +0000 (13:25 -0400)]
drm/amdgpu: rename suspend_kms and resume_kms
The old names were dragged over from radeon. The new ones
better match the naming conventions used in the driver.
No functional change.
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Christian König [Tue, 23 Aug 2016 09:18:59 +0000 (11:18 +0200)]
drm/amdgpu: use memcpy_toio for VCE firmware upload
Try to be clean here, even when it's a noop on x86.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Christian König [Tue, 23 Aug 2016 09:00:17 +0000 (11:00 +0200)]
drm/amdgpu: use memcpy_to/fromio for UVD fw upload
Also use the firmware size from the header instead of calculating it.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Rex Zhu [Mon, 22 Aug 2016 12:49:36 +0000 (20:49 +0800)]
drm/amd/powerplay: delete useless code in iceland_hwmgr.c.
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Tue, 23 Aug 2016 14:07:28 +0000 (10:07 -0400)]
drm/radeon: switch UVD code to use UVD_NO_OP for padding
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>
Alex Deucher [Tue, 23 Aug 2016 13:12:21 +0000 (09:12 -0400)]
drm/amdgpu: switch UVD code to use UVD_NO_OP for padding
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>
Alex Deucher [Mon, 22 Aug 2016 22:03:22 +0000 (18:03 -0400)]
drm/radeon: add support for UVD_NO_OP register
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>
Alex Deucher [Mon, 22 Aug 2016 21:58:14 +0000 (17:58 -0400)]
drm/amdgpu: add support for UVD_NO_OP register
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>
Alex Deucher [Tue, 23 Aug 2016 14:44:16 +0000 (10:44 -0400)]
drm/amdgpu: fix VCE ib alignment value
The VCE rings only require single dword alignment.
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Mon, 22 Aug 2016 20:31:36 +0000 (16:31 -0400)]
drm/amdgpu: fix IB alignment for UVD
According to the hw team, it should be 16, not 8.
Cc: Peter Fang <peter.fang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Tom St Denis [Mon, 22 Aug 2016 14:54:28 +0000 (10:54 -0400)]
drm/amd/amdgpu: Print ring name in amdgpu_ib_schedule()
If the ring isn't ready lets print out which ring name
to help debugging.
Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Heinrich Schuchardt [Sun, 21 Aug 2016 20:36:29 +0000 (22:36 +0200)]
drm/radeon: remove dead code, si_mc_load_microcode (v2)
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>
Heinrich Schuchardt [Sun, 21 Aug 2016 19:41:49 +0000 (21:41 +0200)]
drm/radeon/cik: remove dead code (v2)
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>
Heinrich Schuchardt [Sun, 21 Aug 2016 18:27:02 +0000 (20:27 +0200)]
drm/amd/powerplay: avoid NULL dereference, cz_hwmgr.c
if (a == NULL || a->b == NULL)
leads to a NULL pointer dereference if a == NULL.
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>
Heinrich Schuchardt [Sun, 21 Aug 2016 18:21:27 +0000 (20:21 +0200)]
drm/amd/powerplay: avoid NULL pointer dereference
if (a == NULL || a->b == NULL)
leads to a NULL pointer dereference if a == NULL.
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>
Heinrich Schuchardt [Sun, 21 Aug 2016 18:10:41 +0000 (20:10 +0200)]
drm/amdgpu/gmc8: remove dead code (v2)
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>
Heinrich Schuchardt [Sun, 21 Aug 2016 18:06:32 +0000 (20:06 +0200)]
drm/amdgpu/gmc7: remove dead code (v2)
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>
Alexandre Demers [Mon, 22 Aug 2016 00:38:26 +0000 (20:38 -0400)]
drm/amdgpu: Fix indentation in dce_v8_0_audio_write_sad_regs()
Fixed indentation for readability.
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Signed-off-by: Alexandre Demers <alexandre.f.demers@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alexandre Demers [Mon, 22 Aug 2016 00:38:41 +0000 (20:38 -0400)]
drm/amdgpu: Use correct mask in dce_v8_0_afmt_setmode() and fix comment typos.
We were using the same mask twice. Looking at radeon, it seems
we should be using HDMI_AVI_INFO_CONT instead as the second mask.
Being there, fix typos in comments and improved readability.
I haven't looked at other DCEs, the mask may also be wrong for them.
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alexandre Demers <alexandre.f.demers@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Christian König [Tue, 16 Aug 2016 12:43:17 +0000 (14:43 +0200)]
drm/amdgpu: cleanup amdgpu_vm_bo_update params
Make it more obvious what we are doing here.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Christian König [Mon, 15 Aug 2016 12:08:54 +0000 (14:08 +0200)]
drm/amdgpu: stop adding dummy entry in amdgpu_ttm_placement_init
AMDGPU_GEM_CREATE_NO_CPU_ACCESS and AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED are
obviously mutual exclusive. So stop adding a dummy entry without effect when
both are specified.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Christian König [Mon, 15 Aug 2016 12:06:50 +0000 (14:06 +0200)]
drm/amdgpu: cleanup amdgpu_ttm_placement_init
Make it more clear what this function does. No intendet functional change.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Christian König [Wed, 17 Aug 2016 11:44:20 +0000 (13:44 +0200)]
drm/amdgpu: fix lru size grouping v2
Adding a BO can make it the insertion point for larger sizes as well.
v2: add a comment about the guard structure.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Tom St Denis [Thu, 18 Aug 2016 14:25:52 +0000 (10:25 -0400)]
drm/amd/powerplay: Tidy up cz_hwmgr.c
Clean up whitespace and formatting.
Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Marek Olšák [Wed, 17 Aug 2016 21:58:58 +0000 (23:58 +0200)]
drm/amdgpu: add AMDGPU_INFO_NUM_EVICTIONS
For profiling.
v2: really bump the minor version
Signed-off-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Chunming Zhou [Thu, 21 Jul 2016 09:20:52 +0000 (17:20 +0800)]
drm/amdgpu: recover vram bo from shadow after gpu reset V2
V2:
1. don't directly submit to many jobs at the same time.
2. delete unrelated printk.
Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Chunming Zhou [Wed, 17 Aug 2016 03:41:30 +0000 (11:41 +0800)]
drm/amdgpu: link all shadow bo V2
V2:
1. use mutex instead of spinlock for shadow list, since its process could
sleep.
2. move list_del to bo destroy phase.
Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Chunming Zhou [Mon, 15 Aug 2016 03:46:21 +0000 (11:46 +0800)]
drm/amdgpu: update pt shadow while updating pt V2
V2:
move shadow parameter to amdgpu_pte_update_params.
Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Cc: minutemaidpark@hotmail.com
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Chunming Zhou [Mon, 15 Aug 2016 03:36:54 +0000 (11:36 +0800)]
drm/amdgpu: update pd shadow while updating pd V2
V2:
Checking if shadow is valid.
Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Chunming Zhou [Thu, 4 Aug 2016 08:51:18 +0000 (16:51 +0800)]
drm/amdgpu: sync bo and shadow V3
Use shadow flag to judge which direction to sync.
V2:
Don't need bo pin, so remove it.
V3:
1. Split to two functions, one is backup_to_shadow, another is
restore_from_shadow.
2. Clean up previous shadow direction difinitions.
Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Chunming Zhou [Mon, 15 Aug 2016 02:46:04 +0000 (10:46 +0800)]
drm/amdgpu: add direct submision option for copy_buffer
Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
projects / openwrt / staging / blogic.git / log