- 12 Feb, 2024 6 commits
-
-
Likun Gao authored
Support to set fw_load_type=3 to use backdoor rlc autoload. Signed-off-by:
Likun Gao <Likun.Gao@amd.com> Reviewed-by:
Hawking Zhang <Hawking.Zhang@amd.com> Signed-off-by:
Alex Deucher <alexander.deucher@amd.com>
-
Likun Gao authored
Skip ucode BO reservation for backdoor RLC autoload. Signed-off-by:
-
Likun Gao authored
Only enable VCN/JPEG dpm when VCN/JPEG PG flag was set when smu set dpm table. Signed-off-by:
Kenneth Feng <kenneth.feng@amd.com> Signed-off-by:
-
Kent Russell authored
Its currently incorrectly multiplied by number of XCCs in the partition Fixes: be457b22 ("drm/amdkfd: Update cache info for GFX 9.4.3") Signed-off-by:
Kent Russell <kent.russell@amd.com> Reviewed-by:
Mukul Joshi <mukul.joshi@amd.com> Signed-off-by:
-
Laurent Morichetti authored
Call the 2nd level trap handler if the cwsr handler is entered with any one of wave_start, wave_end, or trap_after_inst exceptions. Signed-off-by:
Laurent Morichetti <laurent.morichetti@amd.com> Tested-by:
Lancelot Six <lancelot.six@amd.com> Reviewed-by:
Jay Cornwall <jay.cornwall@amd.com> Acked-by:
-
Jonathan Kim authored
The debugger requires the control stack header to be filled in to update_waves. Reviewed-by:
Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by:
Jonathan Kim <jonathan.kim@amd.com> Signed-off-by:
-
- 07 Feb, 2024 34 commits
-
-
Lijo Lazar authored
HDP flush remapping is not done for VFs. Keep the original offsets in VF environment. Signed-off-by:
Lijo Lazar <lijo.lazar@amd.com> Reviewed-by:
-
Srinivasan Shanmugam authored
'stream_enc_regs' array is an array of dcn10_stream_enc_registers structures. The array is initialized with four elements, corresponding to the four calls to stream_enc_regs() in the array initializer. This means that valid indices for this array are 0, 1, 2, and 3. The error message 'stream_enc_regs' 4 <= 5 below, is indicating that there is an attempt to access this array with an index of 5, which is out of bounds. This could lead to undefined behavior Here, eng_id is used as an index to access the stream_enc_regs array. If eng_id is 5, this would result in an out-of-bounds access on the stream_enc_regs array. Thus fixing Buffer overflow error in dcn301_stream_encoder_create reported by Smatch: drivers/gpu/drm/amd/amdgpu/../display/dc/resource/dcn301/dcn301_resource.c:1011 dcn301_stream_encoder_create() error: buffer overflow 'stream_enc_regs' 4 <= 5 Fixes: 3a83e4e6 ("drm/amd/display: Add dcn3.01 support to DC (v2)") Cc: Roman Li <Roman.Li@amd.com> Cc: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com> Cc: Aurabindo Pillai <aurabindo.pillai@amd.com> Signed-off-by:
Srinivasan Shanmugam <srinivasan.shanmugam@amd.com> Reviewed-by:
Roman Li <roman.li@amd.com> Signed-off-by:
-
Nathan Chancellor authored
After a recent change in LLVM, allmodconfig (which has CONFIG_KCSAN=y and CONFIG_WERROR=y enabled) has a few new instances of -Wframe-larger-than for the mode support and system configuration functions: drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dcn20/display_mode_vba_20v2.c:3393:6: error: stack frame size (2144) exceeds limit (2048) in 'dml20v2_ModeSupportAndSystemConfigurationFull' [-Werror,-Wframe-larger-than] 3393 | void dml20v2_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_lib) | ^ 1 error generated. drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dcn21/display_mode_vba_21.c:3520:6: error: stack frame size (2192) exceeds limit (2048) in 'dml21_ModeSupportAndSystemConfigurationFull' [-Werror,-Wframe-larger-than] 3520 | void dml21_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_lib) | ^ 1 error generated. drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dcn20/display_mode_vba_20.c:3286:6: error: stack frame size (2128) exceeds limit (2048) in 'dml20_ModeSupportAndSystemConfigurationFull' [-Werror,-Wframe-larger-than] 3286 | void dml20_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_lib) | ^ 1 error generated. Without the sanitizers enabled, there are no warnings. This was the catalyst for commit 6740ec97 ("drm/amd/display: Increase frame warning limit with KASAN or KCSAN in dml2") and that same change was made to dml in commit 5b750b22 ("drm/amd/display: Increase frame warning limit with KASAN or KCSAN in dml") but the frame_warn_flag variable was not applied to all files. Do so now to clear up the warnings and make all these files consistent. Cc: stable@vger.kernel.org Closes: https://github.com/ClangBuiltLinux/linux/issue/1990Signed-off-by:Nathan Chancellor <nathan@kernel.org> Signed-off-by:
-
Mario Limonciello authored
When dc_state_destruct() was refactored the new phantom_stream_count and phantom_plane_count members weren't cleared. Fixes: 012a04b1 ("drm/amd/display: Refactor phantom resource allocation") Acked-by:
Mario Limonciello <mario.limonciello@amd.com> Signed-off-by:
-
Jiapeng Chong authored
./drivers/gpu/drm/amd/display/dc/dml2/dml2_utils.c:236:49-51: WARNING !A || A && B is equivalent to !A || B. Reported-by:
Abaci Robot <abaci@linux.alibaba.com> Closes: https://bugzilla.openanolis.cn/show_bug.cgi?id=8169Signed-off-by:
Jiapeng Chong <jiapeng.chong@linux.alibaba.com> Signed-off-by:
-
Lijo Lazar authored
The present way to fetch VRAM vendor information turns out to be not reliable on GFX 9.4.3 dGPUs as well. Avoid using the data. Signed-off-by:
-
Aric Cyr authored
This version brings along the following: - DCN35 fixes - DMUB fixes - Link training fixes - Misc code style fixes - MST fixes - ODM fixes - SubVP fixes Acked-by:
Hamza Mahfooz <hamza.mahfooz@amd.com> Signed-off-by:
Aric Cyr <aric.cyr@amd.com> Signed-off-by:
-
Rodrigo Siqueira authored
Some of the CONFIG_DRM_AMD_DC_FP was added in some non-related FPU code, which may cause confusion. This commit dropped some of the unnecessary guards. Acked-by:
Rodrigo Siqueira <Rodrigo.Siqueira@amd.com> Signed-off-by:
-
Rodrigo Siqueira authored
Just adjust the code indentation in the header and add a simple comment in the dm_cp_psp file. Acked-by:
-
Rodrigo Siqueira authored
Just ensure that ODM optimization is disabled by default. Acked-by:
-
Rodrigo Siqueira authored
Display code keeps getting improvements, and because of that, some legacy code is left behind. This commit drops some of those unused codes. Acked-by:
-
Alvin Lee authored
Previously we would call apply_ctx_to_hw to enable and disable phantom pipes. However, apply_ctx_to_hw can potentially update non-phantom pipes as well which is undesired. Instead of calling apply_ctx_to_hw as a whole, call the relevant helpers for each phantom pipe when enabling / disabling which will avoid us modifying hardware state for non-phantom pipes unknowingly. The use case is for an FRL display where FRL_Update is requested by the display. In this case link_state_valid flag is cleared in a passive callback thread and should be handled in the next stream / link update. However, due to the call to apply_ctx_to_hw for the phantom pipes during a flip, the main pipes were modified outside of the desired sequence (driver does not handle link_state_valid = 0 on flips). Cc: stable@vger.kernel.org # 6.6+ Reviewed-by:
Samson Tam <samson.tam@amd.com> Acked-by:
Alvin Lee <alvin.lee2@amd.com> Signed-off-by:
-
Alvin Lee authored
This reverts commit 6b2b782a. Since, it was causing regression for some DRR scenarios. Reviewed-by:
Nevenko Stupar <nevenko.stupar@amd.com> Acked-by:
-
Ethan Bitnun authored
Update the p_state type before update_clocks is called to ensure accurate values are used for logging. Reviewed-by:
Ethan Bitnun <etbitnun@amd.com> Signed-off-by:
-
Ethan Bitnun authored
Add a small delay before reading clks from hw, to ensure correct values are used for logging. Reviewed-by:
-
Fangzhi Zuo authored
The change try to fix below error specific to RV platform: BUG: kernel NULL pointer dereference, address: 0000000000000008 PGD 0 P4D 0 Oops: 0000 [#1] PREEMPT SMP NOPTI CPU: 4 PID: 917 Comm: sway Not tainted 6.3.9-arch1-1 #1 124dc55df4f5272ccb409f39ef4872fc2b3376a2 Hardware name: LENOVO 20NKS01Y00/20NKS01Y00, BIOS R12ET61W(1.31 ) 07/28/2022 RIP: 0010:drm_dp_atomic_find_time_slots+0x5e/0x260 [drm_display_helper] Code: 01 00 00 48 8b 85 60 05 00 00 48 63 80 88 00 00 00 3b 43 28 0f 8d 2e 01 00 00 48 8b 53 30 48 8d 04 80 48 8d 04 c2 48 8b 40 18 <48> 8> RSP: 0018:ffff960cc2df77d8 EFLAGS: 00010293 RAX: 0000000000000000 RBX: ffff8afb87e81280 RCX: 0000000000000224 RDX: ffff8afb9ee37c00 RSI: ffff8afb8da1a578 RDI: ffff8afb87e81280 RBP: ffff8afb83d67000 R08: 0000000000000001 R09: ffff8afb9652f850 R10: ffff960cc2df7908 R11: 0000000000000002 R12: 0000000000000000 R13: ffff8afb8d7688a0 R14: ffff8afb8da1a578 R15: 0000000000000224 FS: 00007f4dac35ce00(0000) GS:ffff8afe30b00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000000000008 CR3: 000000010ddc6000 CR4: 00000000003506e0 Call Trace: <TASK> ? __die+0x23/0x70 ? page_fault_oops+0x171/0x4e0 ? plist_add+0xbe/0x100 ? exc_page_fault+0x7c/0x180 ? asm_exc_page_fault+0x26/0x30 ? drm_dp_atomic_find_time_slots+0x5e/0x260 [drm_display_helper 0e67723696438d8e02b741593dd50d80b44c2026] ? drm_dp_atomic_find_time_slots+0x28/0x260 [drm_display_helper 0e67723696438d8e02b741593dd50d80b44c2026] compute_mst_dsc_configs_for_link+0x2ff/0xa40 [amdgpu 62e600d2a75e9158e1cd0a243bdc8e6da040c054] ? fill_plane_buffer_attributes+0x419/0x510 [amdgpu 62e600d2a75e9158e1cd0a243bdc8e6da040c054] compute_mst_dsc_configs_for_state+0x1e1/0x250 [amdgpu 62e600d2a75e9158e1cd0a243bdc8e6da040c054] amdgpu_dm_atomic_check+0xecd/0x1190 [amdgpu 62e600d2a75e9158e1cd0a243bdc8e6da040c054] drm_atomic_check_only+0x5c5/0xa40 drm_mode_atomic_ioctl+0x76e/0xbc0 ? _copy_to_user+0x25/0x30 ? drm_ioctl+0x296/0x4b0 ? __pfx_drm_mode_atomic_ioctl+0x10/0x10 drm_ioctl_kernel+0xcd/0x170 drm_ioctl+0x26d/0x4b0 ? __pfx_drm_mode_atomic_ioctl+0x10/0x10 amdgpu_drm_ioctl+0x4e/0x90 [amdgpu 62e600d2a75e9158e1cd0a243bdc8e6da040c054] __x64_sys_ioctl+0x94/0xd0 do_syscall_64+0x60/0x90 ? do_syscall_64+0x6c/0x90 entry_SYSCALL_64_after_hwframe+0x72/0xdc RIP: 0033:0x7f4dad17f76f Code: 00 48 89 44 24 18 31 c0 48 8d 44 24 60 c7 04 24 10 00 00 00 48 89 44 24 08 48 8d 44 24 20 48 89 44 24 10 b8 10 00 00 00 0f 05 <89> c> RSP: 002b:00007ffd9ae859f0 EFLAGS: 00000246 ORIG_RAX: 0000000000000010 RAX: ffffffffffffffda RBX: 000055e255a55900 RCX: 00007f4dad17f76f RDX: 00007ffd9ae85a90 RSI: 00000000c03864bc RDI: 000000000000000b RBP: 00007ffd9ae85a90 R08: 0000000000000003 R09: 0000000000000003 R10: 0000000000000000 R11: 0000000000000246 R12: 00000000c03864bc R13: 000000000000000b R14: 000055e255a7fc60 R15: 000055e255a01eb0 </TASK> Modules linked in: rfcomm snd_seq_dummy snd_hrtimer snd_seq snd_seq_device ccm cmac algif_hash algif_skcipher af_alg joydev mousedev bnep > typec libphy k10temp ipmi_msghandler roles i2c_scmi acpi_cpufreq mac_hid nft_reject_inet nf_reject_ipv4 nf_reject_ipv6 nft_reject nft_mas> CR2: 0000000000000008 ---[ end trace 0000000000000000 ]--- RIP: 0010:drm_dp_atomic_find_time_slots+0x5e/0x260 [drm_display_helper] Code: 01 00 00 48 8b 85 60 05 00 00 48 63 80 88 00 00 00 3b 43 28 0f 8d 2e 01 00 00 48 8b 53 30 48 8d 04 80 48 8d 04 c2 48 8b 40 18 <48> 8> RSP: 0018:ffff960cc2df77d8 EFLAGS: 00010293 RAX: 0000000000000000 RBX: ffff8afb87e81280 RCX: 0000000000000224 RDX: ffff8afb9ee37c00 RSI: ffff8afb8da1a578 RDI: ffff8afb87e81280 RBP: ffff8afb83d67000 R08: 0000000000000001 R09: ffff8afb9652f850 R10: ffff960cc2df7908 R11: 0000000000000002 R12: 0000000000000000 R13: ffff8afb8d7688a0 R14: ffff8afb8da1a578 R15: 0000000000000224 FS: 00007f4dac35ce00(0000) GS:ffff8afe30b00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000000000008 CR3: 000000010ddc6000 CR4: 00000000003506e0 With a second DP monitor connected, drm_atomic_state in dm atomic check sequence does not include the connector state for the old/existing/first DP monitor. In such case, dsc determination policy would hit a null ptr when it tries to iterate the old/existing stream that does not have a valid connector state attached to it. When that happens, dm atomic check should call drm_atomic_get_connector_state for a new connector state. Existing dm has already done that, except for RV due to it does not have official support of dsc where .num_dsc is not defined in dcn10 resource cap, that prevent from getting drm_atomic_get_connector_state called. So, skip dsc determination policy for ASICs that don't have DSC support. Cc: stable@vger.kernel.org # 6.1+ Link: https://gitlab.freedesktop.org/drm/amd/-/issues/2314Reviewed-by:
Wayne Lin <wayne.lin@amd.com> Acked-by:
Fangzhi Zuo <jerry.zuo@amd.com> Signed-off-by:
-
Camille Cho authored
0 nits is a valid default value for OLED panels. So, update the relevant comment to account for that fact. Reviewed-by:
Krunoslav Kovac <krunoslav.kovac@amd.com> Signed-off-by:
Camille Cho <camille.cho@amd.com> Signed-off-by:
-
George Shen authored
[Why] Currently 3-tap chroma subsampling is used for YCbCr422/420. When ODM pipesplit is used, pixels on the left edge of ODM slices need one extra pixel from the right edge of the previous slice to calculate the correct chroma value. Without this change, the chroma value is slightly different than expected. This is usually imperceptible visually, but it impacts test pattern CRCs for compliance test automation. [How] Update logic to use the register for adding extra left edge pixel for YCbCr422/420 ODM cases. Reviewed-by:
George Shen <george.shen@amd.com> Signed-off-by:
-
George Shen authored
[Why] Default driver behaviour is 3-tap subsampling, so we should keep it the same for test patterns as well. However, it is also useful to force 1-tap subsampling for testing purposes. Reviewed-by:
Michael Strauss <michael.strauss@amd.com> Acked-by:
-
Nicholas Kazlauskas authored
[Why] Workaroud for a race condition where DMCUB is in the process of committing to IPS1 during the handshake causing us to miss the transition into IPS2 and touch the INBOX1 RPTR causing a HW hang. [How] Disable the reallow to ensure that we have enough of a gap between entry and exit and we're not seeing back-to-back wake_and_executes. Reviewed-by:
Ovidiu Bunea <ovidiu.bunea@amd.com> Acked-by:
Nicholas Kazlauskas <nicholas.kazlauskas@amd.com> Signed-off-by:
-
Stanley.Yang authored
ta if invoke node buffer |-------- ta type ----------| |-------- ta id ----------| |-------- cmd id ----------| |------ shared buf len -----| |------ shared buffer ------| ta if invoke node buffer is as above, copy shared buffer data to correct location Signed-off-by:
Stanley.Yang <Stanley.Yang@amd.com> Reviewed-by:
-
Nicholas Kazlauskas authored
[Why] To match firmware measurements and avoid hanging when accessing HW that's in idle. [How] Increase the delays to what we've measured. Reviewed-by:
-
Nicholas Kazlauskas authored
[Why] We're still missing a few and we'd like to avoid continuining when a hang occurs for debug purposes. [How] Add the loop anywhere we try to wait on rptr == wptr in dc_dmub_srv. Reviewed-by:
-
Srinivasan Shanmugam authored
The issue arises when the array 'adev->vcn.vcn_config' is accessed before checking if the index 'adev->vcn.num_vcn_inst' is within the bounds of the array. The fix involves moving the bounds check before the array access. This ensures that 'adev->vcn.num_vcn_inst' is within the bounds of the array before it is used as an index. Fixes the below: drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c:1289 amdgpu_discovery_reg_base_init() error: testing array offset 'adev->vcn.num_vcn_inst' after use. Fixes: a0ccc717 ("drm/amdgpu/discovery: validate VCN and SDMA instances") Cc: Christian König <christian.koenig@amd.com> Cc: Alex Deucher <alexander.deucher@amd.com> Signed-off-by:
-
Candice Li authored
Instead of software managed counters. Signed-off-by:
Candice Li <candice.li@amd.com> Reviewed-by:
Tao Zhou <tao.zhou1@amd.com> Signed-off-by:
-
Nicholas Kazlauskas authored
[Why] Any interface that touches registers needs to wake up the system. [How] Add a new interface dc_exit_ips_for_hw_access that wraps the check for IPS support and insert it into the public DC interfaces that touch registers. We don't re-enter, since we expect that the enter/exit to have been done on the DM side. Cc: stable@vger.kernel.org # 6.1+ Reviewed-by:
-
Allen Pan authored
[Why] Hardware register definition changed Reviewed-by:
Charlene Liu <charlene.liu@amd.com> Acked-by:
Allen Pan <allen.pan@amd.com> Signed-off-by:
-
Li Ma authored
A supplement to commit: 615dd56a There is an irq warning of jpeg during resume in s2idle process. No irq enabled in jpeg 4.0.5 resume. Fixes: 615dd56a ("drm/amdgpu: remove asymmetrical irq disabling in vcn 4.0.5 suspend") Signed-off-by:
Li Ma <li.ma@amd.com> Acked-By:
Saleemkhan Jamadar <saleemkhan.jamadar@amd.com> Reviewed-by:
Yifan Zhang <yifan1.zhang@amd.com> Reviewed-by:
Veerabadhran Gopalakrishnan <Veerabadhran.Gopalakrishnan@amd.com> Signed-off-by:
-
Prike Liang authored
In the s3 suspend abort case some type of gfx9 power rail not turn off from FCH side and this will put the GPU in an unknown power status, so let's reset the gpu to a known good power state before reinitialize gpu device. Signed-off-by:
Prike Liang <Prike.Liang@amd.com> Acked-by:
-
Prike Liang authored
In the suspend abort cases, the gfx power rail doesn't turn off so some GFXDEC registers/CSB can't reset to default value and at this moment reinitialize GFXDEC/CSB will result in an unexpected error. So let skip those program sequence for the suspend abort case. Signed-off-by:
-
Wenjing Liu authored
[why] When populating dml pipes, odm combine policy should be assigned based on the pipe topology of the context passed in. DML pipes could be repopulated multiple times during single validate bandwidth attempt. We need to make sure that whenever we repopulate the dml pipes it is always aligned with the updated context. There is a case where DML pipes get repopulated during FPO optimization after ODM combine policy is changed. Since in the current code we reinitlaize ODM combine policy, even though the current context has ODM combine enabled, we overwrite it despite the pipes are already split. This causes DML to think that MPC combine is used so we mistakenly enable MPC combine because we apply pipe split with ODM combine policy reset. This issue doesn't impact non windowed MPO with ODM case because the legacy policy has restricted use cases. We don't encounter the case where both ODM and FPO optimizations are enabled together. So we decide to leave it as is because it is about to be replaced anyway. Cc: stable@vger.kernel.org # 6.6+ Reviewed-by:
Chaitanya Dhere <chaitanya.dhere@amd.com> Reviewed-by:
Wenjing Liu <wenjing.liu@amd.com> Signed-off-by:
-
Michael Strauss authored
[WHY] Only required if FIXED_VS retimer does not support DP2-capable. [HOW] Gate link rate toggle with DP 128b/132b LTTPR channel coding cap check. Reviewed-by:
-
Joseph Greathouse authored
The KFD topology includes cache line size, but we have not been filling that information out unless we are parsing a CRAT table. Fill in this information for the devices where we have cache information structs, and pipe this information to the topology sysfs files. v2: squash in fix from Joe (Alex) Signed-off-by:
Joseph Greathouse <Joseph.Greathouse@amd.com> Acked-by:
-
Michael Strauss authored
The New sequence has been in use in DCN314 with no regressions introduced. Therefore, it is safe to enable this sequence for all devices using FIXED_VS retimers. So, remove the legacy codepath and its associated config flag. Reviewed-by:
-
