- 25 Oct, 2019 40 commits
-
-
Tianci.Yin authored
update registers: mmCGTT_SPI_CLK_CTRL Reviewed-by:
Feifei Xu <Feifei.Xu@amd.com> Signed-off-by:
Tianci.Yin <tianci.yin@amd.com> Signed-off-by:
Alex Deucher <alexander.deucher@amd.com>
-
Tianci.Yin authored
update registers: mmCGTT_SPI_CLK_CTRL Reviewed-by:
-
Tianci.Yin authored
update registers: mmCGTT_SPI_CLK_CTRL Reviewed-by:
-
Evan Quan authored
For dpm disabled case, it's assumed the only one support clock level is always current clock level. Signed-off-by:
Evan Quan <evan.quan@amd.com> Acked-by:
-
Evan Quan authored
For Arcturus, clock limit settings on uclk/socclk/fclk domains are not supported. V2: simplify the code to support both SGPU and MGPU cases Signed-off-by:
-
Marek Olšák authored
Allow userspace to read the same status registers for every family. Based on commit c7890fea, added any of these registers if defined in the include files of each architecture. Signed-off-by:
Marek Olšák <marek.olsak@amd.com> Reviewed-by:
-
Andrey Grodzovsky authored
For Arcturus the I2C traffic is done through SMU tables and so we must postpone RAS recovery init to after they are ready which is in amdgpu_device_ip_hw_init_phase2. Signed-off-by:
Andrey Grodzovsky <andrey.grodzovsky@amd.com> Reviewed-by:
Guchun Chen <guchun.chen@amd.com> Signed-off-by:
-
Andrey Grodzovsky authored
Add Arcturus EEPROM/I2C support in generic EEPROM code. Signed-off-by:
-
Andrey Grodzovsky authored
The communication is done through SMU table and hence the code is in powerplay. Signed-off-by:
-
Andrey Grodzovsky authored
Will be used by Arcturus support for RAS page retirement. Signed-off-by:
-
Leo Li authored
[Why] Some LED panel drivers might not like fractional PWM. In such cases, backlight flickering may be observed. [How] Add a DC feature mask to disable fractional PWM, and associate it with the preexisting dc_config flag. The flag is only plumbed through the dmcu firmware, so plumb it through the driver path as well. To disable, add the following to the linux cmdline: amdgpu.dcfeaturemask=0x4 Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=204957Signed-off-by:
Leo Li <sunpeng.li@amd.com> Reviewed-by:
Anthony Koo <anthony.koo@amd.com> Tested-by:
Lukáš Krejčí <lskrejci@gmail.com> Signed-off-by:
-
Nirmoy Das authored
Signed-off-by:
Nirmoy Das <nirmoy.das@amd.com> Acked-by:
Christian König <christian.koenig@amd.com> Signed-off-by:
-
James Zhu authored
After VCN2.5 firmware (Version ENC: 1.1 Revision: 11), VCN2.5 encoding can work properly. Signed-off-by:
James Zhu <James.Zhu@amd.com> Reviewed-by:
Leo Liu <leo.liu@amd.com> Reviewed-by:
-
Pelloux-prayer, Pierre-eric authored
This seems to help with https://bugs.freedesktop.org/show_bug.cgi?id=111481. v2: insert a NOP instead of skipping all 0-sized IBs to avoid breaking older hw Signed-off-by:
Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com> Reviewed-by:
-
chen gong authored
VKexample test hang during Occlusion/SDMA/Varia runs. Clear XNACK_WATERMK in reg SDMA0_UTCL1_WATERMK to fix this issue. Signed-off-by:
chen gong <curry.gong@amd.com> Reviewed-by:
Aaron Liu <aaron.liu@amd.com> Signed-off-by:
-
Guchun Chen authored
Easy for maintainance. Signed-off-by:
-
Guchun Chen authored
Ras reboot debugfs node allows user one easy control to avoid gpu recovery hang problem and directly reboot system per card basis, after ras uncorrectable error happens. However, it is one common entry, which should get rid of ras_ctrl node and remove ip dependence when inputting by user. So add one new auto_reboot node in ras debugfs dir to achieve this. v2: in commit mssage, add justification why ras reboot debugfs node is needed. v3: use debugfs_create_bool to create debugfs file for boolean value Signed-off-by:
-
Philip Yang authored
If device reset/suspend/resume failed for some reason, dqm lock is hold forever and this causes deadlock. Below is a kernel backtrace when application open kfd after suspend/resume failed. Instead of holding dqm lock in pre_reset and releasing dqm lock in post_reset, add dqm->sched_running flag which is modified in dqm->ops.start and dqm->ops.stop. The flag doesn't need lock protection because write/read are all inside dqm lock. For HWS case, map_queues_cpsch and unmap_queues_cpsch checks sched_running flag before sending the updated runlist. v2: For no-HWS case, when device is stopped, don't call load/destroy_mqd for eviction, restore and create queue, and avoid debugfs dump hdqs. Backtrace of dqm lock deadlock: [Thu Oct 17 16:43:37 2019] INFO: task rocminfo:3024 blocked for more than 120 seconds. [Thu Oct 17 16:43:37 2019] Not tainted 5.0.0-rc1-kfd-compute-rocm-dkms-no-npi-1131 #1 [Thu Oct 17 16:43:37 2019] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message. [Thu Oct 17 16:43:37 2019] rocminfo D 0 3024 2947 0x80000000 [Thu Oct 17 16:43:37 2019] Call Trace: [Thu Oct 17 16:43:37 2019] ? __schedule+0x3d9/0x8a0 [Thu Oct 17 16:43:37 2019] schedule+0x32/0x70 [Thu Oct 17 16:43:37 2019] schedule_preempt_disabled+0xa/0x10 [Thu Oct 17 16:43:37 2019] __mutex_lock.isra.9+0x1e3/0x4e0 [Thu Oct 17 16:43:37 2019] ? __call_srcu+0x264/0x3b0 [Thu Oct 17 16:43:37 2019] ? process_termination_cpsch+0x24/0x2f0 [amdgpu] [Thu Oct 17 16:43:37 2019] process_termination_cpsch+0x24/0x2f0 [amdgpu] [Thu Oct 17 16:43:37 2019] kfd_process_dequeue_from_all_devices+0x42/0x60 [amdgpu] [Thu Oct 17 16:43:37 2019] kfd_process_notifier_release+0x1be/0x220 [amdgpu] [Thu Oct 17 16:43:37 2019] __mmu_notifier_release+0x3e/0xc0 [Thu Oct 17 16:43:37 2019] exit_mmap+0x160/0x1a0 [Thu Oct 17 16:43:37 2019] ? __handle_mm_fault+0xba3/0x1200 [Thu Oct 17 16:43:37 2019] ? exit_robust_list+0x5a/0x110 [Thu Oct 17 16:43:37 2019] mmput+0x4a/0x120 [Thu Oct 17 16:43:37 2019] do_exit+0x284/0xb20 [Thu Oct 17 16:43:37 2019] ? handle_mm_fault+0xfa/0x200 [Thu Oct 17 16:43:37 2019] do_group_exit+0x3a/0xa0 [Thu Oct 17 16:43:37 2019] __x64_sys_exit_group+0x14/0x20 [Thu Oct 17 16:43:37 2019] do_syscall_64+0x4f/0x100 [Thu Oct 17 16:43:37 2019] entry_SYSCALL_64_after_hwframe+0x44/0xa9 Suggested-by:
Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by:
Philip Yang <Philip.Yang@amd.com> Reviewed-by:
-
Zhan liu authored
[Why] DWB (Display Writeback) flag needs to be enabled as 1, or system will throw out a few warnings when creating dcn20 resource pool. Also, Navi14's dwb setting needs to match Navi10's, which has already been set to 1. [How] Change value of num_dwb from 0 to 1. Signed-off-by:
Zhan Liu <zhan.liu@amd.com> Reviewed-by:
Nicholas Kazlauskas <nicholas.kazlauskas@amd.com> Signed-off-by:
-
Roman Li authored
[Why] Adding psr mask to dc features allows selectively disable/enable psr. Current psr implementation may not work with non-pageflipping application. Until resolved it should be disabled by default. [How] Add dcfeaturemask for psr enablement. Disable by default. To enable set amdgpu.dcfeaturemask=0x8 in grub kernel command line. Signed-off-by:
Roman Li <Roman.Li@amd.com> Reviewed-by:
-
Evan Quan authored
With this cleanup, the APIs from amdgpu_smu.c will map to ASIC specific ones directly. Those can be shared around all SMU V11/V12 ASICs will be put in smu_v11_0.c and smu_v12_0.c respectively. Signed-off-by:
-
Evan Quan authored
Those swSMU APIs used internally are moved to smu_internal.h while others are kept in amdgpu_smu.h. V2: give a better name smu_internal.h for the place to hold those internal APIs Signed-off-by: -
Evan Quan authored
This is a quick and low risk fix. Those APIs which are exposed to other IPs or to support sysfs/hwmon interfaces or DAL will have lock protection. Meanwhile no lock protection is enforced for swSMU internal used APIs. Future optimization is needed. V2: strip the lock protection for all swSMU internal APIs Signed-off-by:
-
Jane Jian authored
Add mmsch part registers Signed-off-by:
Jane Jian <jane.jian@amd.com> Reviewed-by:
-
Joshua Aberback authored
[Why] This workaround was put in dcn2 DMLv1, and now we need it in DMLv2. Signed-off-by:
Joshua Aberback <joshua.aberback@amd.com> Reviewed-by:
Jun Lei <Jun.Lei@amd.com> Acked-by:
-
Eric Yang authored
[Why] Fix the programming of DCHUBBUB_ARB_REFCYC_PER_TRIP_TO_MEMORY_A. Was not filled in. Signed-off-by:
Eric Yang <Eric.Yang2@amd.com> Reviewed-by:
Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com> Acked-by:
-
Dmytro Laktyushkin authored
We are currently incorrectly processing avoid split at highest voltage level. Signed-off-by:
Eric Bernstein <Eric.Bernstein@amd.com> Acked-by:
-
Sung Lee authored
[Why] In Hybrid Graphics, dcn2_1_soc struct stays alive through PnP. This causes an issue on dc init where dcn2_1_soc which has been updated by update_bw_bounding_box gets put into dml->soc. As update_bw_bounding_box is currently incorrect for dcn2.1, this makes dml calculations fail due to incorrect parameters, leading to a crash on PnP. [How] Comment out update_bw_bounding_box call for now. Signed-off-by:
Sung Lee <sung.lee@amd.com> Reviewed-by:
-
Reza Amini authored
[why] We want to use maximum space on display to show source [how] For Centered Mode: Replicate source as many times as possible to use maximum of display active space add borders. Signed-off-by:
Reza Amini <Reza.Amini@amd.com> Reviewed-by:
-
Lewis Huang authored
[Why] Signal is update to EDP when driver disable first encoder. The following encoder using SIGNAL_TYPE_EDP to handle other device. When encoder signal is HDMI, driver will detect it is dp and release phy. It cause hw hang. [How] Take signal type from link->connector_signal. Signed-off-by:
Lewis Huang <Lewis.Huang@amd.com> Reviewed-by:
-
Aric Cyr authored
Signed-off-by:
Aric Cyr <aric.cyr@amd.com> Acked-by:
-
Noah Abradjian authored
[Why] * Clk Mgr DTO update point did not cover all needed updates, as it included a check for plane_state which does not exist yet when the updater is called on driver startup * This resulted in another update path in the pipe programming sequence, based on a dppclk update flag * However, this alternate path allowed for stray DTO updates, some of which would occur in the wrong order during dppclk lowering and cause underflow [How] * Remove plane_state check and use of plane_res.dpp->inst, getting rid of sequence dependencies (this results in extra dto programming for unused pipes but that doesn't cause issues and is a small cost) * Allow DTOs to be updated even if global clock is equal, to account for edge case exposed by diags tests * Remove update_dpp_dto call in pipe programming sequence (leave update to dppclk_control there, as that update is necessary and shouldn't occur in clk mgr) * Remove call to optimize_bandwidth when committing state, as it is not needed and resulted in sporadic underflows even with other fixes in place Signed-off-by:
Noah Abradjian <noah.abradjian@amd.com> Reviewed-by:
-
Krunoslav Kovac authored
[Why&How] BT.2390 EETF is used for tone mapping/range reduction. Say display is 0.1 - 500 nits. The problematic case is when content is 0-400. We apply EETF because 0<0.1 so we need to reduce the range by 0.1. In the commit, we ignore the bottom range. Most displays map 0 to min and then have a ramp to 0.1, so sending 0.1 is actually >0.1. Furthermode, HW that uses 3D LUT also assumes min=0. Signed-off-by:
Krunoslav Kovac <Krunoslav.Kovac@amd.com> Reviewed-by:
-
Eric Yang authored
[Why] SMU does not keep the wm table across S3, S4, need to re-send the table. Also defer sending the cable to after DCN bave initialized [How] Send table at end of init hw Signed-off-by:
Yongqiang Sun <yongqiang.sun@amd.com> Acked-by:
-
Jun Lei authored
[why] A display that supports DRR can never really be considered "synchronized" with any other display because we can dynamically enable DRR (i.e. without modeset). this will cause their relative CRTC positions to drift and lose sync. this will disrupt features such as MCLK switching that assume and depend on their permanent alignment (that can only change with modeset) [how] check for ignore_msa in stream when considered synchronizability this ignore_msa is basically actually implemented as "supports drr" Signed-off-by:
-
Anthony Koo authored
[Why] When aux engine acquire fails, we missed populating the operation_result that describes the failure reason. [How] Set operation_result to new type: AUX_CHANNEL_OPERATION_FAILED_ENGINE_ACQUIRE in the case aux engine acquire has failed. Signed-off-by:
-
Michael Strauss authored
[WHY] force_single_disp_pipe_split is a debug flag for use on DCN1 but isn't necessary otherwise as DCN2+ splits by default Signed-off-by:
Michael Strauss <michael.strauss@amd.com> Reviewed-by:
Tony Cheng <Tony.Cheng@amd.com> Acked-by:
Charlene Liu <Charlene.Liu@amd.com> Acked-by:
-
Michael Strauss authored
[WHY] i2c_read is called to differentiate passive DP->HDMI and DP->DVI-D dongles The call is expected to fail in DVI-D case but pass in HDMI case Some HDMI dongles have a chance to fail as well, causing misdetection as DVI-D [HOW] Retry i2c_read to ensure failed result is valid Signed-off-by:
-
Yogesh Mohan Marimuthu authored
[Why] The enum value for TRANSMITTER_UNIPHY_G is 9. In resource dc_xx_resource file structure link_enc_regs[], the TRANSMITTER_UNIPHY_G registers are initialized at index 6. Due to this mismatch, if monitor is attached to port using TRANSMITTER_UNIPHY_G then the monitor blanks out. [How] add function map_transmitter_id_to_phy_instance() and use the function to map enum transmitter to link regs. Signed-off-by:
Yogesh Mohan Marimuthu <yogesh.mohanmarimuthu@amd.com> Reviewed-by:
-
Jun Lei authored
[why] SLT tests require that diag can drive eDP even if nothing is connected, this is not typical production use case, so we need to add flag [how] add flag, and this flag supercedes "should destroy" logic Signed-off-by:
-
