- 13 Mar, 2020 23 commits
-
-
Lucas Stach authored
1db8c142 (drm/scheduler: Add drm_sched_suspend/resume_timeout()) made the job_list_lock IRQ safe in as the suspend/resume calls were expected to be called from IRQ context. This usage never materialized in upstream. Instead amdgpu started locking the job_list_lock in an IRQ unsafe way in amdgpu_ib_preempt_mark_partial_job() and amdgpu_ib_preempt_job_recovery(), which leads to potential deadlock if one would actually start to call the drm_sched_suspend/resume_timeout functions from IRQ context. As no current user needs the locking to be IRQ safe, the local IRQ disable/enable is pure overhead. Fix the inconsistent locking by changing all uses of job_list_lock to use the IRQ unsafe locking primitives. Reviewed-by:
Christian König <christian.koenig@amd.com> Signed-off-by:
Lucas Stach <l.stach@pengutronix.de> Signed-off-by:
Alex Deucher <alexander.deucher@amd.com>
-
Robert Beckett authored
Add a new trace event to show when jobs are run on the HW. Acked-by:
Robert Beckett <bob.beckett@collabora.com> Signed-off-by:
-
Kevin Wang authored
clean up unused header in swsmu driver stack: 1. pp_debug.h 2. amd_pcie.h 3. soc15_common.h Signed-off-by:
Kevin Wang <kevin1.wang@amd.com> Reviewed-by:
-
Dennis Li authored
AccVGPRs are newly added in arcturus. Before reading these registers, they should be initialized. Otherwise edc error happens, when RAS is enabled. v2: reuse the existing logical to calculate register size Signed-off-by:
Dennis Li <Dennis.Li@amd.com> Reviewed-by:
Hawking Zhang <Hawking.Zhang@amd.com> Signed-off-by:
-
Mario Kleiner authored
This fixes a problem found on the MacBookPro 2017 Retina panel: The panel reports 10 bpc color depth in its EDID, and the firmware chooses link settings at boot which support enough bandwidth for 10 bpc (324000 kbit/sec aka LINK_RATE_RBR2 aka 0xc), but the DP_MAX_LINK_RATE dpcd register only reports 2.7 Gbps (multiplier value 0xa) as possible, in direct contradiction of what the firmware successfully set up. This restricts the panel to 8 bpc, not providing the full color depth of the panel on Linux <= 5.5. Additionally, commit '4a8ca46b ("drm/amd/display: Default max bpc to 16 for eDP")' introduced into Linux 5.6-rc1 will unclamp panel depth to its full 10 bpc, thereby requiring a eDP bandwidth for all modes that exceeds the bandwidth available and causes all modes to fail validation -> No modes for the laptop panel -> failure to set any mode -> Panel goes dark. This patch adds a quirk specific to the MBP 2017 15" Retina panel to override reported max link rate to the correct maximum of 0xc = LINK_RATE_RBR2 to fix the darkness and reduced display precision. Please apply for Linux 5.6+ to avoid regressing Apple MBP panel support. Signed-off-by:
Mario Kleiner <mario.kleiner.de@gmail.com> Signed-off-by:
-
Lyude Paul authored
These are some very loud debug statements that get printed on every vblank when driver level debug printing is enabled in DRM, and doesn't really tell us anything that isn't related to vblanks. So let's move this over to the proper debug flag to be a little less spammy with our debug output. Reviewed-by:
Harry Wentland <harry.wentland@amd.com> Signed-off-by:
Lyude Paul <lyude@redhat.com> Signed-off-by:
-
Dan Carpenter authored
If "speed" is zero then we use it as a divisor to find "prescale". It's better to move the check for zero to the very start of the function. Fixes: 9eeec26a ("drm/amd/display: Refine i2c frequency calculating sequence") Signed-off-by:
Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by:
-
Dan Carpenter authored
These lines were accidentally indented 4 spaces more than they should be. Signed-off-by:
-
Dan Carpenter authored
We can remove the NULL check for "res_ctx" and "res_ctx->pipe_ctx[i].stream->link". Also it's nicer to align the conditions using spaces so I re-indented a bit. Longer explanation: The "res_ctx" pointer points to an address in the middle of a struct so it can't be NULL. For "res_ctx->pipe_ctx[i].stream->link" we know that it is equal to "link" and "link" is non-NULL. Signed-off-by:
-
Randy Dunlap authored
Fix the Kconfig dependencies so that the menu is presented correctly by adding a dependency on DRM_AMDGPU to the "menu" Kconfig statement. This makes a continuous dependency on DRM_AMDGPU in the DRM AMD menus and eliminates a broken menu structure. Fixes: a8fe58ce ("drm/amd: add ACP driver support") Signed-off-by:
Randy Dunlap <rdunlap@infradead.org> Cc: Alex Deucher <alexander.deucher@amd.com> Cc: Christian König <christian.koenig@amd.com> Cc: David (ChunMing) Zhou <David1.Zhou@amd.com> Cc: Maruthi Bayyavarapu <maruthi.bayyavarapu@amd.com> Cc: amd-gfx@lists.freedesktop.org Signed-off-by:
-
Joe Perches authored
Convert the various uses of fallthrough comments to fallthrough; Done via script Link: https://lore.kernel.org/lkml/b56602fcf79f849e733e7b521bb0e17895d390fa.1582230379.git.joe@perches.com/Acked-by:
Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by:
Joe Perches <joe@perches.com> Signed-off-by:
-
Joe Perches authored
The automated conversion of /* fallthrough */ comments converted a comment outside of an #ifdef/#endif case block that should be inside the block. Move the fallthrough inside the block to silence the warning. Signed-off-by:
-
Joe Perches authored
Convert the various uses of fallthrough comments to fallthrough; Done via script Link: https://lore.kernel.org/lkml/b56602fcf79f849e733e7b521bb0e17895d390fa.1582230379.git.joe@perches.com/Signed-off-by:
-
Joe Perches authored
Convert the various uses of fallthrough comments to fallthrough; Done via script Link: https://lore.kernel.org/lkml/b56602fcf79f849e733e7b521bb0e17895d390fa.1582230379.git.joe@perches.com/Signed-off-by:
-
Stanley.Yang authored
Fix the warning "warn: variable dereferenced before check 'obj' (see line 1131)" by removing unnecessary checks as amdgpu_ras_debugfs_create_all() is only called from amdgpu_debugfs_init() where obj member in con->head list is not NULL. Use list_for_each_entry() instead list_for_each_entry_safe() as obj do not to be freeing or removing from list during this process. Signed-off-by:
Stanley.Yang <Stanley.Yang@amd.com> Reviewed-by:
-
Evan Quan authored
This can fix the baco reset failure seen on Navi10. And this should be a low risk fix as the same sequence is already used for system suspend/resume. Signed-off-by:
Evan Quan <evan.quan@amd.com> Reviewed-by:
-
Guchun Chen authored
RAS support capability needs to be updated on top of different memeory ECC enablement, and remove redundant memory ecc check in gmc module for vega20 and arcturus. v2: check HBM ECC enablement and set ras mask accordingly. v3: avoid to invoke atomfirmware interface to query twice. Suggested-by:
Guchun Chen <guchun.chen@amd.com> Reviewed-by:
-
Tom St Denis authored
The offset into the array was specified in bytes but should be in terms of 32-bit words. Also prevent large reads that would also cause a buffer overread. v2: Read from correct offset from internal storage buffer. Signed-off-by:
Tom St Denis <tom.stdenis@amd.com> Acked-by:
-
Stanley.Yang authored
In dcn20_funcs and dcn21_funcs struct, the member ".dsc_pg_control = NULL" should be removed due to .dsc_pg_control be assigned to dcn20_dsc_pg_control. Signed-off-by:
Anthony Koo <Anthony.Koo@amd.com> Signed-off-by:
-
Stanley.Yang authored
include amdgpu_ras.h head file instead of use extern ras_debugfs_create_all function Signed-off-by:
Tao Zhou <tao.zhou1@amd.com> Reviewed-by:
-
Hawking Zhang authored
disallow the logical to be enabled on platforms that don't support gfx ras at this stage, like sriov skus, dgpu with legacy ras.etc Signed-off-by:
Monk Liu <monk.liu@amd.com> Signed-off-by:
-
John Clements authored
invoking an error injection successfully will cause an at_event intterrupt that will occur before the invoke sequence can complete causing an invalid error Reviewed-by:
John Clements <john.clements@amd.com> Signed-off-by:
-
Jack Zhang authored
refine the assignment for vcn.num_vcn_inst, vcn.harvest_config, vcn.num_enc_rings in VF Signed-off-by:
Jack Zhang <Jack.Zhang1@amd.com> Reviewed-by:
-
- 12 Mar, 2020 2 commits
-
-
git://people.freedesktop.org/~agd5f/linuxDave Airlie authored
amd-drm-next-5.7-2020-03-10: amdgpu: - SR-IOV fixes - Fix up fallout from drm load/unload callback removal - Navi, renoir power management watermark fixes - Refactor smu parameter handling - Display FEC fixes - Display DCC fixes - HDCP fixes - Add support for USB-C PD firmware updates - Pollock detection fix - Rework compute ring priority handling - RAS fixes - Misc cleanups amdkfd: - Consolidate more gfx config details in amdgpu - Consolidate bo alloc flags - Improve code comments - SDMA MQD fixes - Misc cleanups gpu scheduler: - Add suport for modifying the sched list uapi: - Clarify comments about GEM_CREATE flags that are not used by userspace. The kernel driver has always prevented userspace from using these. They are only used internally in the kernel driver. Signed-off-by:
Dave Airlie <airlied@redhat.com> From: Alex Deucher <alexdeucher@gmail.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200310212748.4519-1-alexander.deucher@amd.com
-
git://anongit.freedesktop.org/drm/drm-miscDave Airlie authored
drm-misc-next for 5.7: UAPI Changes: Cross-subsystem Changes: Core Changes: Driver Changes: - fb-helper: Remove drm_fb_helper_{add,add_all,remove}_one_connector - fbdev: some cleanups and dead-code removal - Conversions to simple-encoder - zero-length array removal - Panel: panel-dpi support in panel-simple, Novatek NT35510, Elida KD35T133, Signed-off-by:
-
- 10 Mar, 2020 9 commits
-
-
Dave Airlie authored
Requested my mripard for some misc patches that need this as a base. Signed-off-by: -
Feifei Xu authored
Some framework test will fail if enable runpm on Vega10. Disable it untill issue fixed. Signed-off-by:
Feifei Xu <Feifei.Xu@amd.com> Tested-by:
Kyle Chen <Kyle.Chen@amd.com> Signed-off-by:
-
Tao Zhou authored
and remove each ras IP's own debugfs creation this is required to fix ras when the driver does not use the drm load and unload callbacks due to ordering issues with the drm device node. Signed-off-by:
-
Tao Zhou authored
centralize all debugfs creation in one place for ras this is required to fix ras when the driver does not use the drm load and unload callbacks due to ordering issues with the drm device node. Signed-off-by:
-
Prike Liang authored
There will be a coverity warning because min and max are both unsigned. Signed-off-by:
Prike Liang <Prike.Liang@amd.com> Reviewed-by:
-
xinhui pan authored
Only kernel bo has kfd eviction fence. This warning is to give a notice that kfd only remove eviction fence on individual bos. Tested-by:
Nicholas Johnson <nicholas.johnson-opensource@outlook.com.au> Reviewed-by:
xinhui pan <xinhui.pan@amd.com> Signed-off-by:
-
Yong Zhao authored
ALLOC_MEM_FLAGS_* used are the same as the KFD_IOC_ALLOC_MEM_FLAGS_*, but they are interweavedly used in kernel driver, resulting in bad readability. For example, KFD_IOC_ALLOC_MEM_FLAGS_COHERENT is not referenced in kernel, and it functions implicitly in kernel through ALLOC_MEM_FLAGS_COHERENT, causing unnecessary confusion. Replace all occurrences of ALLOC_MEM_FLAGS_* with KFD_IOC_ALLOC_MEM_FLAGS_* to solve the problem. Signed-off-by:
Yong Zhao <Yong.Zhao@amd.com> Reviewed-by:
-
Yong Zhao authored
People are inclined to think of the previous pr_warn message as an error, so use pre_debug instead. Signed-off-by:
-
Nirmoy Das authored
If there are no high priority compute queues available then set normal priority sched array to compute_prio_sched[AMDGPU_GFX_PIPE_PRIO_HIGH] Signed-off-by:
Nirmoy Das <nirmoy.das@amd.com> Reviewed-by:
-
- 09 Mar, 2020 6 commits
-
-
Nirmoy Das authored
AMDGPU statically sets priority for compute queues at initialization so remove all the functions responsible for changing compute queue priority dynamically. Signed-off-by:
-
Nirmoy Das authored
Switch to appropriate sched list for an entity on priority override. Signed-off-by:
-
Nirmoy Das authored
Implement drm_sched_entity_modify_sched() which modifies existing sched_list with a different one. This is going to be helpful when userspace changes priority of a ctx/entity then the driver can switch to the corresponding HW scheduler list for that priority. Signed-off-by:
-
Nirmoy Das authored
We were changing compute ring priority while rings were being used before every job submission which is not recommended. This patch sets compute queue priority at mqd initialization for gfx8, gfx9 and gfx10. Policy: make queue 0 of each pipe as high priority compute queue High/normal priority compute sched lists are generated from set of high/normal priority compute queues. At context creation, entity of compute queue get a sched list from high or normal priority depending on ctx->priority Signed-off-by:
-
Xiaojie Yuan authored
Fix version for navi12. Signed-off-by:
Xiaojie Yuan <xiaojie.yuan@amd.com> Reviewed-by:
-
Aly-Tawfik authored
Problem Description: Currently we are checking internal fused rev id with pci rev id. However, fused internal rev id is the same on all raven2 parts (in which Dali and Pollock were based on too), thus Pollock detection fails Fix: use the pci rev to preform the detection for bandwidth calculations. Reviewed-by:
Aly-Tawfik <altawfik@amd.com> Signed-off-by:
-
