Skip to contentSkip to conversation

Add Arcturus support

PATCH New Alex Deucher 1 participant 89 messages Jul 15, 2019 1 message not in the archive

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

cover Add Arcturus support

Arcturus is a new GPU from AMD. This patch set contains support for the base functionality on most IPs: GC, SDMA, VCN, IH, etc.

I'm not sending out the new register headers due to size, but you can view the entire commit history in git:
https://cgit.freedesktop.org/~agd5f/linux/log/?h=amd-staging-drm-next-arcturus

Alex Deucher (2):
drm/amdgpu: drop unused function definitions drm/amdgpu: flag arcturus as experimental for now

Chengming Gui (3):
drm/amdgpu/powerplay: add arcturus ppt functions drm/amdgpu/powerplay: add smu11 driver interface for arcturus. (v2) drm/amd/powerplay: get smc firmware and pptable

Feifei Xu (1):
drm/amdgpu: add pci DID for Arcturus GL-XL.

Hawking Zhang (6):
drm/amdgpu: add arct sdma golden settings drm/amdgpu: add arct gc golden settings drm/amdgpu: init arct external rev id drm/amdgpu: keep stolen memory for arct drm/amdgpu: init gds config for arct drm/amdgpu: skip gfx 9 common golden settings for arct

James Zhu (10):
drm/amdgpu: Clear build undefined warning
drm/amdgpu/: add clientID for 2nd vcn instance drm/amdgpu/: add ucodeID for 2nd vcn instance drm/amdgpu/: add doorbell assignment for 2nd vcn instance drm/amdgpu/: increase AMDGPU_MAX_RINGS to add 2nd vcn instance drm/amdgpu: add vcn nbio doorbell range setting for 2nd vcn instance drm/amdgpu: modify amdgpu_vcn to support multiple instances drm/amdgpu: add multiple instances support for Arcturus drm/amdgpu: add harvest support for Arcturus drm/amdgpu:add all VCN rings into schedule request queue

Jay Cornwall (1):
drm/amdkfd: Merge gfx9/arcturus trap handlers, add ACC VGPR save

Le Ma (47):
drm/amdgpu: add mmhub 9.4.1 header files for Acrturus drm/amdgpu: add sdma 4.2.2 header files for Arcturus drm/amdgpu: add Arcturus ip_offset header (v3)
drm/amdgpu: add Arcturus asic type
drm/amdgpu: add gmc basic support for Arcturus
drm/amdgpu: rename AMDGPU_GFXHUB/MMHUB macro with hub number drm/amdgpu: add new member in amdgpu_device for vmhub counts per asic
chip
drm/amdgpu: add one more mmhub instance for Arcturus (v2) drm/amdgpu: add mmhub v9.4.1 block for Arcturus (v2) drm/amdgpu: use new mmhub interfaces for Arcturus drm/amdgpu: add SDMA 2~7 interrupt client id for Arcturus drm/amdgpu: add SDMA 2~7 ip block type
drm/amdgpu: increase max number of ip base instances to 8 drm/amdgpu: dynamically initialize IP offset for Arcturus drm/amdgpu: add VMC1 interrupt client id for Arcturus drm/amdgpu: update vmc interrupt routine to support 3 vmhubs drm/amdgpu: reorganize sdma v4 code to support more instances drm/amdgpu: specify sdma instance 5~7 with second mmhub type drm/amdgpu: support hdp flush for more sdma instances drm/amdgpu/soc15: add Arcturus common ip blocks drm/amdgpu: add to set Arcturus ip blocks
drm/amdgpu: set Arcturus fw load type as direct drm/amdgpu/dce_virtual: add Arcturus virtual display support drm/amdgpu: add support for Arcturus firmware
drm/amdgpu: add gfx config for Arcturus
drm/amdgpu: add number of mec for Arcturus
drm/amdgpu: add to set rlc funcs for Arcturus
drm/amdgpu: skip to get 3D engine clockgating state for Arcturus drm/amdgpu: skip pasid mapping for second mmhub on Arcturus drm/amdgpu: add Arcturus gpu info firmware
drm/amdgpu: optimize gfx9 init_microcode function drm/amdgpu: skip load cp gfx firmware for Arcturus drm/amdgpu: skip all gfx ring settings for Arcturus drm/amdgpu: support sdma 2~7 doorbell range register offset drm/amdgpu: correct Arcturus SDMA address space base index drm/amdgpu: enable 8 SDMA instances for Arcturus drm/amdgpu: add Arcturus chip_name for init sdma microcode drm/amdgpu: correct programming of ih_chicken for Arcturus drm/amdgpu: add paging queue support for 8 SDMA instances on Arcturus drm/amdgpu: declare sdma firmware binary files for Arcturus drm/amdgpu: skip get/update xgmi topology info when no psp exists drm/amdgpu: set system aperture to cover whole FB region in mmhub v9.4 drm/amdgpu: correct ip for mmHDP_READ_CACHE_INVALIDATE register access drm/amdgpu: assign fb_start/end in mmhub v9.4 interface drm/amdgpu: clean up nonexistent firmware declaration for Arcturus drm/amdgpu: limit sdma instances to 2 for Arcturus in BU phase drm/amdgpu: enable all 8 sdma instances for Arcturus silicon

Leo Liu (13):
drm/amdgpu: add VCN2.5 headers
drm/amdgpu/VCN2: put IB internal registers offset to structure drm/amdgpu/VCN2: expose rings functions
drm/amdgpu: add VCN2.5 basic supports
drm/amdgpu: add VCN2.5 VCPU start and stop drm/amdgpu: add Arcturus to the VCN family drm/amdgpu/VCN2.5: set decode ring functions drm/amdgpu/VCN2.5: set encode ring functions drm/amdgpu: add JPEG2.5 HW start and stop drm/amdgpu/VCN2.5: set JPEG decode ring functions drm/amdgpu: enable VCN2.5 on Arcturus
drm/amdgpu: add vcn doorbell range function to nbio7.4 (v2) drm/amdgpu: enable the Doorbell support for VCN2.5

Oak Zeng (12):
drm/amdgpu: Initialize asic functions for Arcturus drm/amdkfd: Extend PM4 packets to support 8 SDMA drm/amdkfd: Support bigger gds size
drm/amdkfd: Change arcturus sdma engines number drm/amdkfd: Fix sdma_bitmap overflow issue
drm/amdkfd: Implement kfd2kgd_calls for Arcturus drm/amdgpu: Hack xgmi topology info when there is no psp fw drm/amdgpu: Enable xgmi support for Arcturus drm/amdkfd: Set number of xgmi optimized SDMA engines for arcturus drm/amdkfd: Add arcturus CWSR trap handler
drm/amdkfd: Add device id for real asics
drm/amdkfd: Increase vcrat size for GPU

Yong Zhao (7):
amd/amdkfd: Add ASIC ARCTURUS to kfd
drm/amdkfd: Expose function mmhub_v9_4_setup_vm_pt_regs() for kfd to
use
drm/amdkfd: Support two MMHUBs when setting up page table base in KFD drm/amdgpu: Set VM_L2_CNTL.PDE_FAULT_CLASSIFICATION to 0 for MMHUB 9.4 drm/amdkfd: Support MMHUB1 in kfd interrupt path amd/powerplay: No SW XGMI dpm for Arcturus rev 2 drm/amdgpu: Add more detail to the VM fault printing

 drivers/gpu/drm/amd/amdgpu/Makefile           |     9 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu.h           |    17 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c    |     8 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h    |     1 +
 .../drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c   |   324 +
 .../gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c |   169 +-
 .../gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h |    69 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c       |    26 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c    |     6 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h  |     9 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c       |     4 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c       |    29 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h      |     2 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h      |     8 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c     |     2 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h     |     1 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c       |   182 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h       |    34 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c        |     4 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h        |     7 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c      |    76 +-
 drivers/gpu/drm/amd/amdgpu/arct_reg_init.c    |    57 +
 drivers/gpu/drm/amd/amdgpu/dce_virtual.c      |     1 +
 drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c        |     8 +-
 drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c         |   375 +-
 drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c      |     2 +-
 drivers/gpu/drm/amd/amdgpu/gfxhub_v2_0.c      |     2 +-
 drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c        |    31 +-
 drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c         |   112 +-
 drivers/gpu/drm/amd/amdgpu/gmc_v9_0.h         |     7 +
 drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c       |     2 +-
 drivers/gpu/drm/amd/amdgpu/mmhub_v2_0.c       |     2 +-
 drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c       |   517 +
 drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.h       |    33 +
 drivers/gpu/drm/amd/amdgpu/nbio_v2_3.c        |     2 +-
 drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c        |    72 +-
 drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c        |   314 +-
 drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c        |     2 +-
 drivers/gpu/drm/amd/amdgpu/soc15.c            |    29 +-
 drivers/gpu/drm/amd/amdgpu/soc15.h            |     1 +
 drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c         |     4 +-
 drivers/gpu/drm/amd/amdgpu/vce_v4_0.c         |     2 +-
 drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c         |   116 +-
 drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c         |   265 +-
 drivers/gpu/drm/amd/amdgpu/vcn_v2_0.h         |    38 +
 drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c         |  1417 +
 drivers/gpu/drm/amd/amdgpu/vcn_v2_5.h         |    29 +
 drivers/gpu/drm/amd/amdgpu/vega10_ih.c        |    11 +-
 drivers/gpu/drm/amd/amdgpu/vega20_reg_init.c  |     4 +
 .../gpu/drm/amd/amdkfd/cwsr_trap_handler.h    |   485 +
 .../drm/amd/amdkfd/cwsr_trap_handler_gfx9.asm |    83 +-
 drivers/gpu/drm/amd/amdkfd/kfd_crat.c         |     3 +-
 drivers/gpu/drm/amd/amdkfd/kfd_device.c       |    25 +-
 .../drm/amd/amdkfd/kfd_device_queue_manager.c |     9 +-
 drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c  |     1 +
 .../gpu/drm/amd/amdkfd/kfd_int_process_v9.c   |     2 +
 drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.c |     1 +
 .../gpu/drm/amd/amdkfd/kfd_kernel_queue_v9.c  |    59 +-
 .../gpu/drm/amd/amdkfd/kfd_packet_manager.c   |     1 +
 .../gpu/drm/amd/amdkfd/kfd_pm4_headers_ai.h   |    24 +-
 drivers/gpu/drm/amd/amdkfd/kfd_topology.c     |     1 +
 drivers/gpu/drm/amd/include/arct_ip_offset.h  |  1654 +
 .../asic_reg/mmhub/mmhub_9_4_1_default.h      |  3933 ++
 .../asic_reg/mmhub/mmhub_9_4_1_offset.h       |  7753 +++
 .../asic_reg/mmhub/mmhub_9_4_1_sh_mask.h      | 44884 ++++++++++++++++
 .../asic_reg/sdma0/sdma0_4_2_2_offset.h       |  1051 +
 .../asic_reg/sdma0/sdma0_4_2_2_sh_mask.h      |  3002 ++
 .../asic_reg/sdma1/sdma1_4_2_2_offset.h       |  1043 +
 .../asic_reg/sdma1/sdma1_4_2_2_sh_mask.h      |  2956 +
 .../asic_reg/sdma2/sdma2_4_2_2_offset.h       |  1043 +
 .../asic_reg/sdma2/sdma2_4_2_2_sh_mask.h      |  2956 +
 .../asic_reg/sdma3/sdma3_4_2_2_offset.h       |  1043 +
 .../asic_reg/sdma3/sdma3_4_2_2_sh_mask.h      |  2956 +
 .../asic_reg/sdma4/sdma4_4_2_2_offset.h       |  1043 +
 .../asic_reg/sdma4/sdma4_4_2_2_sh_mask.h      |  2956 +
 .../asic_reg/sdma5/sdma5_4_2_2_offset.h       |  1043 +
 .../asic_reg/sdma5/sdma5_4_2_2_sh_mask.h      |  2956 +
 .../asic_reg/sdma6/sdma6_4_2_2_offset.h       |  1043 +
 .../asic_reg/sdma6/sdma6_4_2_2_sh_mask.h      |  2956 +
 .../asic_reg/sdma7/sdma7_4_2_2_offset.h       |  1043 +
 .../asic_reg/sdma7/sdma7_4_2_2_sh_mask.h      |  2956 +
 .../amd/include/asic_reg/vcn/vcn_2_5_offset.h |   979 +
 .../include/asic_reg/vcn/vcn_2_5_sh_mask.h    |  3609 ++
 .../gpu/drm/amd/include/soc15_ih_clientid.h   |    11 +-
 drivers/gpu/drm/amd/powerplay/Makefile        |     2 +-
 drivers/gpu/drm/amd/powerplay/amdgpu_smu.c    |    11 +
 drivers/gpu/drm/amd/powerplay/arcturus_ppt.c  |   124 +
 drivers/gpu/drm/amd/powerplay/arcturus_ppt.h  |    28 +
 .../gpu/drm/amd/powerplay/inc/amdgpu_smu.h    |    11 +
 .../drm/amd/powerplay/inc/arcturus_ppsmc.h    |   120 +
 .../powerplay/inc/smu11_driver_if_arcturus.h  |   878 +
 drivers/gpu/drm/amd/powerplay/smu_v11_0.c     |    15 +-
 include/drm/amd_asic_type.h                   |     1 +
 93 files changed, 100491 insertions(+), 704 deletions(-)
 create mode 100644 drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c
 create mode 100644 drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h
 create mode 100644 drivers/gpu/drm/amd/amdgpu/arct_reg_init.c
 create mode 100644 drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c
 create mode 100644 drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.h
 create mode 100644 drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c
 create mode 100644 drivers/gpu/drm/amd/amdgpu/vcn_v2_5.h
 create mode 100644 drivers/gpu/drm/amd/include/arct_ip_offset.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/mmhub/mmhub_9_4_1_default.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/mmhub/mmhub_9_4_1_offset.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/mmhub/mmhub_9_4_1_sh_mask.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/sdma0/sdma0_4_2_2_offset.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/sdma0/sdma0_4_2_2_sh_mask.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/sdma1/sdma1_4_2_2_offset.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/sdma1/sdma1_4_2_2_sh_mask.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/sdma2/sdma2_4_2_2_offset.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/sdma2/sdma2_4_2_2_sh_mask.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/sdma3/sdma3_4_2_2_offset.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/sdma3/sdma3_4_2_2_sh_mask.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/sdma4/sdma4_4_2_2_offset.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/sdma4/sdma4_4_2_2_sh_mask.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/sdma5/sdma5_4_2_2_offset.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/sdma5/sdma5_4_2_2_sh_mask.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/sdma6/sdma6_4_2_2_offset.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/sdma6/sdma6_4_2_2_sh_mask.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/sdma7/sdma7_4_2_2_offset.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/sdma7/sdma7_4_2_2_sh_mask.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/vcn/vcn_2_5_offset.h
 create mode 100644 drivers/gpu/drm/amd/include/asic_reg/vcn/vcn_2_5_sh_mask.h
 create mode 100644 drivers/gpu/drm/amd/powerplay/arcturus_ppt.c
 create mode 100644 drivers/gpu/drm/amd/powerplay/arcturus_ppt.h
 create mode 100644 drivers/gpu/drm/amd/powerplay/inc/arcturus_ppsmc.h
 create mode 100644 drivers/gpu/drm/amd/powerplay/inc/smu11_driver_if_arcturus.h
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

A message that is not in the archive (sent off-list or not yet received): 20190715212437.31793-1-alexander.deucher-5C7GfCeVMHo@public.gmane.org

#2 lore raw

PATCH 4/102 drm/amdgpu: add Arcturus ip_offset header (v3)

From: Le Ma <[email protected]>

Provides the absolute offsets of the IP register
blocks.

v2: update chip name in source code
v3: squash in MP offset updates (Alex)

---
 drivers/gpu/drm/amd/include/arct_ip_offset.h | 1654 ++++++++++++++++++
 1 file changed, 1654 insertions(+)
 create mode 100644 drivers/gpu/drm/amd/include/arct_ip_offset.h
drivers/gpu/drm/amd/include/arct_ip_offset.h new file +1654 -0
new file mode 100644
index 000000000000..3211b3a96d68
@@ -0,0 +1,1654 @@
+/*
+ * Copyright (C) 2018  Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+#ifndef _arct_ip_offset_HEADER
+#define _arct_ip_offset_HEADER
+
+#define MAX_INSTANCE                                       8
+#define MAX_SEGMENT                                         6
+
+
+struct IP_BASE_INSTANCE
+{
+    unsigned int segment[MAX_SEGMENT];
+};
+
+struct IP_BASE
+{
+    struct IP_BASE_INSTANCE instance[MAX_INSTANCE];
+};
+
+
+static const struct IP_BASE ATHUB_BASE            ={ { { { 0x00000C20, 0x00012460, 0x00408C00, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } } } };
+static const struct IP_BASE CLK_BASE            ={ { { { 0x000120C0, 0x00016C00, 0x00401800, 0, 0, 0 } },
+                                        { { 0x000120E0, 0x00016E00, 0x00401C00, 0, 0, 0 } },
+                                        { { 0x00012100, 0x00017000, 0x00402000, 0, 0, 0 } },
+                                        { { 0x00012120, 0x00017200, 0x00402400, 0, 0, 0 } },
+                                        { { 0x000136C0, 0x0001B000, 0x0042D800, 0, 0, 0 } },
+                                        { { 0x00013720, 0x0001B200, 0x0042E400, 0, 0, 0 } },
+                                        { { 0x000125E0, 0x00017E00, 0x0040BC00, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } } } };
+static const struct IP_BASE DF_BASE            ={ { { { 0x00007000, 0x000125C0, 0x0040B800, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } } } };
+static const struct IP_BASE FUSE_BASE            ={ { { { 0x000120A0, 0x00017400, 0x00401400, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } } } };
+static const struct IP_BASE GC_BASE            ={ { { { 0x00002000, 0x0000A000, 0x00012160, 0x00402C00, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } } } };
+static const struct IP_BASE HDP_BASE            ={ { { { 0x00000F20, 0x00012520, 0x0040A400, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } } } };
+static const struct IP_BASE MMHUB_BASE            ={ { { { 0x00012440, 0x0001A000, 0x00408800, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } } } };
+static const struct IP_BASE MP0_BASE            ={ { { { 0x00016000, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } } } };
+static const struct IP_BASE MP1_BASE            ={ { { { 0x00016000, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } } } };
+static const struct IP_BASE NBIF0_BASE            ={ { { { 0x00000000, 0x00000014, 0x00000D20, 0x00010400, 0x00012D80, 0x0041B000 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } } } };
+static const struct IP_BASE OSSSYS_BASE            ={ { { { 0x000010A0, 0x00012500, 0x0040A000, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } } } };
+static const struct IP_BASE PCIE0_BASE            ={ { { { 0x000128C0, 0x00411800, 0x04440000, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } } } };
+static const struct IP_BASE SDMA0_BASE            ={ { { { 0x00001260, 0x00012540, 0x0040A800, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } } } };
+static const struct IP_BASE SDMA1_BASE            ={ { { { 0x00001860, 0x00012560, 0x0040AC00, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } } } };
+static const struct IP_BASE SDMA2_BASE            ={ { { { 0x00013760, 0x0001E000, 0x0042EC00, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } } } };
+static const struct IP_BASE SDMA3_BASE            ={ { { { 0x00013780, 0x0001E400, 0x0042F000, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } } } };
+static const struct IP_BASE SDMA4_BASE            ={ { { { 0x000137A0, 0x0001E800, 0x0042F400, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } } } };
+static const struct IP_BASE SDMA5_BASE            ={ { { { 0x000137C0, 0x0001EC00, 0x0042F800, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } } } };
+static const struct IP_BASE SDMA6_BASE            ={ { { { 0x000137E0, 0x0001F000, 0x0042FC00, 0, 0, 0 } },
+                                       { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } } } };
+static const struct IP_BASE SDMA7_BASE            ={ { { { 0x00013800, 0x0001F400, 0x00430000, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } } } };
+static const struct IP_BASE SMUIO_BASE            ={ { { { 0x00012080, 0x00016800, 0x00016A00, 0x00401000, 0x00440000, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } } } };
+static const struct IP_BASE THM_BASE            ={ { { { 0x00012060, 0x00016600, 0x00400C00, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } } } };
+static const struct IP_BASE UMC_BASE            ={ { { { 0x000132C0, 0x00014000, 0x00425800, 0, 0, 0 } },
+                                        { { 0x000132E0, 0x00054000, 0x00425C00, 0, 0, 0 } },
+                                        { { 0x00013300, 0x00094000, 0x00426000, 0, 0, 0 } },
+                                        { { 0x00013320, 0x000D4000, 0x00426400, 0, 0, 0 } },
+                                        { { 0x00013340, 0x00114000, 0x00426800, 0, 0, 0 } },
+                                        { { 0x00013360, 0x00154000, 0x00426C00, 0, 0, 0 } },
+                                        { { 0x00013380, 0x00194000, 0x00427000, 0, 0, 0 } },
+                                        { { 0x000133A0, 0x001D4000, 0x00427400, 0, 0, 0 } } } };
+static const struct IP_BASE UVD_BASE            ={ { { { 0x00007800, 0x00007E00, 0x00012180, 0x00403000, 0, 0 } },
+                                        { { 0x00007A00, 0x00009000, 0x000136E0, 0x0042DC00, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } } } };
+static const struct IP_BASE DBGU_IO_BASE            ={ { { { 0x000001E0, 0x000125A0, 0x0040B400, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } } } };
+static const struct IP_BASE RSMU_BASE            ={ { { { 0x00012000, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } },
+                                        { { 0, 0, 0, 0, 0, 0 } } } };
+
+
+
+#define ATHUB_BASE__INST0_SEG0                     0x00000C20
+#define ATHUB_BASE__INST0_SEG1                     0x00012460
+#define ATHUB_BASE__INST0_SEG2                     0x00408C00
+#define ATHUB_BASE__INST0_SEG3                     0
+#define ATHUB_BASE__INST0_SEG4                     0
+#define ATHUB_BASE__INST0_SEG5                     0
+
+#define ATHUB_BASE__INST1_SEG0                     0
+#define ATHUB_BASE__INST1_SEG1                     0
+#define ATHUB_BASE__INST1_SEG2                     0
+#define ATHUB_BASE__INST1_SEG3                     0
+#define ATHUB_BASE__INST1_SEG4                     0
+#define ATHUB_BASE__INST1_SEG5                     0
+
+#define ATHUB_BASE__INST2_SEG0                     0
+#define ATHUB_BASE__INST2_SEG1                     0
+#define ATHUB_BASE__INST2_SEG2                     0
+#define ATHUB_BASE__INST2_SEG3                     0
+#define ATHUB_BASE__INST2_SEG4                     0
+#define ATHUB_BASE__INST2_SEG5                     0
+
+#define ATHUB_BASE__INST3_SEG0                     0
+#define ATHUB_BASE__INST3_SEG1                     0
+#define ATHUB_BASE__INST3_SEG2                     0
+#define ATHUB_BASE__INST3_SEG3                     0
+#define ATHUB_BASE__INST3_SEG4                     0
+#define ATHUB_BASE__INST3_SEG5                     0
+
+#define ATHUB_BASE__INST4_SEG0                     0
+#define ATHUB_BASE__INST4_SEG1                     0
+#define ATHUB_BASE__INST4_SEG2                     0
+#define ATHUB_BASE__INST4_SEG3                     0
+#define ATHUB_BASE__INST4_SEG4                     0
+#define ATHUB_BASE__INST4_SEG5                     0
+
+#define ATHUB_BASE__INST5_SEG0                     0
+#define ATHUB_BASE__INST5_SEG1                     0
+#define ATHUB_BASE__INST5_SEG2                     0
+#define ATHUB_BASE__INST5_SEG3                     0
+#define ATHUB_BASE__INST5_SEG4                     0
+#define ATHUB_BASE__INST5_SEG5                     0
+
+#define ATHUB_BASE__INST6_SEG0                     0
+#define ATHUB_BASE__INST6_SEG1                     0
+#define ATHUB_BASE__INST6_SEG2                     0
+#define ATHUB_BASE__INST6_SEG3                     0
+#define ATHUB_BASE__INST6_SEG4                     0
+#define ATHUB_BASE__INST6_SEG5                     0
+
+#define ATHUB_BASE__INST7_SEG0                     0
+#define ATHUB_BASE__INST7_SEG1                     0
+#define ATHUB_BASE__INST7_SEG2                     0
+#define ATHUB_BASE__INST7_SEG3                     0
+#define ATHUB_BASE__INST7_SEG4                     0
+#define ATHUB_BASE__INST7_SEG5                     0
+
+#define CLK_BASE__INST0_SEG0                       0x000120C0
+#define CLK_BASE__INST0_SEG1                       0x00016C00
+#define CLK_BASE__INST0_SEG2                       0x00401800
+#define CLK_BASE__INST0_SEG3                       0
+#define CLK_BASE__INST0_SEG4                       0
+#define CLK_BASE__INST0_SEG5                       0
+
+#define CLK_BASE__INST1_SEG0                       0x000120E0
+#define CLK_BASE__INST1_SEG1                       0x00016E00
+#define CLK_BASE__INST1_SEG2                       0x00401C00
+#define CLK_BASE__INST1_SEG3                       0
+#define CLK_BASE__INST1_SEG4                       0
+#define CLK_BASE__INST1_SEG5                       0
+
+#define CLK_BASE__INST2_SEG0                       0x00012100
+#define CLK_BASE__INST2_SEG1                       0x00017000
+#define CLK_BASE__INST2_SEG2                       0x00402000
+#define CLK_BASE__INST2_SEG3                       0
+#define CLK_BASE__INST2_SEG4                       0
+#define CLK_BASE__INST2_SEG5                       0
+
+#define CLK_BASE__INST3_SEG0                       0x00012120
+#define CLK_BASE__INST3_SEG1                       0x00017200
+#define CLK_BASE__INST3_SEG2                       0x00402400
+#define CLK_BASE__INST3_SEG3                       0
+#define CLK_BASE__INST3_SEG4                       0
+#define CLK_BASE__INST3_SEG5                       0
+
+#define CLK_BASE__INST4_SEG0                       0x000136C0
+#define CLK_BASE__INST4_SEG1                       0x0001B000
+#define CLK_BASE__INST4_SEG2                       0x0042D800
+#define CLK_BASE__INST4_SEG3                       0
+#define CLK_BASE__INST4_SEG4                       0
+#define CLK_BASE__INST4_SEG5                       0
+
+#define CLK_BASE__INST5_SEG0                       0x00013720
+#define CLK_BASE__INST5_SEG1                       0x0001B200
+#define CLK_BASE__INST5_SEG2                       0x0042E400
+#define CLK_BASE__INST5_SEG3                       0
+#define CLK_BASE__INST5_SEG4                       0
+#define CLK_BASE__INST5_SEG5                       0
+
+#define CLK_BASE__INST6_SEG0                       0x000125E0
+#define CLK_BASE__INST6_SEG1                       0x00017E00
+#define CLK_BASE__INST6_SEG2                       0x0040BC00
+#define CLK_BASE__INST6_SEG3                       0
+#define CLK_BASE__INST6_SEG4                       0
+#define CLK_BASE__INST6_SEG5                       0
+
+#define CLK_BASE__INST7_SEG0                       0
+#define CLK_BASE__INST7_SEG1                       0
+#define CLK_BASE__INST7_SEG2                       0
+#define CLK_BASE__INST7_SEG3                       0
+#define CLK_BASE__INST7_SEG4                       0
+#define CLK_BASE__INST7_SEG5                       0
+
+#define DF_BASE__INST0_SEG0                        0x00007000
+#define DF_BASE__INST0_SEG1                        0x000125C0
+#define DF_BASE__INST0_SEG2                        0x0040B800
+#define DF_BASE__INST0_SEG3                        0
+#define DF_BASE__INST0_SEG4                        0
+#define DF_BASE__INST0_SEG5                        0
+
+#define DF_BASE__INST1_SEG0                        0
+#define DF_BASE__INST1_SEG1                        0
+#define DF_BASE__INST1_SEG2                        0
+#define DF_BASE__INST1_SEG3                        0
+#define DF_BASE__INST1_SEG4                        0
+#define DF_BASE__INST1_SEG5                        0
+
+#define DF_BASE__INST2_SEG0                        0
+#define DF_BASE__INST2_SEG1                        0
+#define DF_BASE__INST2_SEG2                        0
+#define DF_BASE__INST2_SEG3                        0
+#define DF_BASE__INST2_SEG4                        0
+#define DF_BASE__INST2_SEG5                        0
+
+#define DF_BASE__INST3_SEG0                        0
+#define DF_BASE__INST3_SEG1                        0
+#define DF_BASE__INST3_SEG2                        0
+#define DF_BASE__INST3_SEG3                        0
+#define DF_BASE__INST3_SEG4                        0
+#define DF_BASE__INST3_SEG5                        0
+
+#define DF_BASE__INST4_SEG0                        0
+#define DF_BASE__INST4_SEG1                        0
+#define DF_BASE__INST4_SEG2                        0
+#define DF_BASE__INST4_SEG3                        0
+#define DF_BASE__INST4_SEG4                        0
+#define DF_BASE__INST4_SEG5                        0
+
+#define DF_BASE__INST5_SEG0                        0
+#define DF_BASE__INST5_SEG1                        0
+#define DF_BASE__INST5_SEG2                        0
+#define DF_BASE__INST5_SEG3                        0
+#define DF_BASE__INST5_SEG4                        0
+#define DF_BASE__INST5_SEG5                        0
+
+#define DF_BASE__INST6_SEG0                        0
+#define DF_BASE__INST6_SEG1                        0
+#define DF_BASE__INST6_SEG2                        0
+#define DF_BASE__INST6_SEG3                        0
+#define DF_BASE__INST6_SEG4                        0
+#define DF_BASE__INST6_SEG5                        0
+
+#define DF_BASE__INST7_SEG0                        0
+#define DF_BASE__INST7_SEG1                        0
+#define DF_BASE__INST7_SEG2                        0
+#define DF_BASE__INST7_SEG3                        0
+#define DF_BASE__INST7_SEG4                        0
+#define DF_BASE__INST7_SEG5                        0
+
+#define FUSE_BASE__INST0_SEG0                      0x000120A0
+#define FUSE_BASE__INST0_SEG1                      0x00017400
+#define FUSE_BASE__INST0_SEG2                      0x00401400
+#define FUSE_BASE__INST0_SEG3                      0
+#define FUSE_BASE__INST0_SEG4                      0
+#define FUSE_BASE__INST0_SEG5                      0
+
+#define FUSE_BASE__INST1_SEG0                      0
+#define FUSE_BASE__INST1_SEG1                      0
+#define FUSE_BASE__INST1_SEG2                      0
+#define FUSE_BASE__INST1_SEG3                      0
+#define FUSE_BASE__INST1_SEG4                      0
+#define FUSE_BASE__INST1_SEG5                      0
+
+#define FUSE_BASE__INST2_SEG0                      0
+#define FUSE_BASE__INST2_SEG1                      0
+#define FUSE_BASE__INST2_SEG2                      0
+#define FUSE_BASE__INST2_SEG3                      0
+#define FUSE_BASE__INST2_SEG4                      0
+#define FUSE_BASE__INST2_SEG5                      0
+
+#define FUSE_BASE__INST3_SEG0                      0
+#define FUSE_BASE__INST3_SEG1                      0
+#define FUSE_BASE__INST3_SEG2                      0
+#define FUSE_BASE__INST3_SEG3                      0
+#define FUSE_BASE__INST3_SEG4                      0
+#define FUSE_BASE__INST3_SEG5                      0
+
+#define FUSE_BASE__INST4_SEG0                      0
+#define FUSE_BASE__INST4_SEG1                      0
+#define FUSE_BASE__INST4_SEG2                      0
+#define FUSE_BASE__INST4_SEG3                      0
+#define FUSE_BASE__INST4_SEG4                      0
+#define FUSE_BASE__INST4_SEG5                      0
+
+#define FUSE_BASE__INST5_SEG0                      0
+#define FUSE_BASE__INST5_SEG1                      0
+#define FUSE_BASE__INST5_SEG2                      0
+#define FUSE_BASE__INST5_SEG3                      0
+#define FUSE_BASE__INST5_SEG4                      0
+#define FUSE_BASE__INST5_SEG5                      0
+
+#define FUSE_BASE__INST6_SEG0                      0
+#define FUSE_BASE__INST6_SEG1                      0
+#define FUSE_BASE__INST6_SEG2                      0
+#define FUSE_BASE__INST6_SEG3                      0
+#define FUSE_BASE__INST6_SEG4                      0
+#define FUSE_BASE__INST6_SEG5                      0
+
+#define FUSE_BASE__INST7_SEG0                      0
+#define FUSE_BASE__INST7_SEG1                      0
+#define FUSE_BASE__INST7_SEG2                      0
+#define FUSE_BASE__INST7_SEG3                      0
+#define FUSE_BASE__INST7_SEG4                      0
+#define FUSE_BASE__INST7_SEG5                      0
+
+#define GC_BASE__INST0_SEG0                        0x00002000
+#define GC_BASE__INST0_SEG1                        0x0000A000
+#define GC_BASE__INST0_SEG2                        0x00012160
+#define GC_BASE__INST0_SEG3                        0x00402C00
+#define GC_BASE__INST0_SEG4                        0
+#define GC_BASE__INST0_SEG5                        0
+
+#define GC_BASE__INST1_SEG0                        0
+#define GC_BASE__INST1_SEG1                        0
+#define GC_BASE__INST1_SEG2                        0
+#define GC_BASE__INST1_SEG3                        0
+#define GC_BASE__INST1_SEG4                        0
+#define GC_BASE__INST1_SEG5                        0
+
+#define GC_BASE__INST2_SEG0                        0
+#define GC_BASE__INST2_SEG1                        0
+#define GC_BASE__INST2_SEG2                        0
+#define GC_BASE__INST2_SEG3                        0
+#define GC_BASE__INST2_SEG4                        0
+#define GC_BASE__INST2_SEG5                        0
+
+#define GC_BASE__INST3_SEG0                        0
+#define GC_BASE__INST3_SEG1                        0
+#define GC_BASE__INST3_SEG2                        0
+#define GC_BASE__INST3_SEG3                        0
+#define GC_BASE__INST3_SEG4                        0
+#define GC_BASE__INST3_SEG5                        0
+
+#define GC_BASE__INST4_SEG0                        0
+#define GC_BASE__INST4_SEG1                        0
+#define GC_BASE__INST4_SEG2                        0
+#define GC_BASE__INST4_SEG3                        0
+#define GC_BASE__INST4_SEG4                        0
+#define GC_BASE__INST4_SEG5                        0
+
+#define GC_BASE__INST5_SEG0                        0
+#define GC_BASE__INST5_SEG1                        0
+#define GC_BASE__INST5_SEG2                        0
+#define GC_BASE__INST5_SEG3                        0
+#define GC_BASE__INST5_SEG4                        0
+#define GC_BASE__INST5_SEG5                        0
+
+#define GC_BASE__INST6_SEG0                        0
+#define GC_BASE__INST6_SEG1                        0
+#define GC_BASE__INST6_SEG2                        0
+#define GC_BASE__INST6_SEG3                        0
+#define GC_BASE__INST6_SEG4                        0
+#define GC_BASE__INST6_SEG5                        0
+
+#define GC_BASE__INST7_SEG0                        0
+#define GC_BASE__INST7_SEG1                        0
+#define GC_BASE__INST7_SEG2                        0
+#define GC_BASE__INST7_SEG3                        0
+#define GC_BASE__INST7_SEG4                        0
+#define GC_BASE__INST7_SEG5                        0
+
+#define HDP_BASE__INST0_SEG0                       0x00000F20
+#define HDP_BASE__INST0_SEG1                       0x00012520
+#define HDP_BASE__INST0_SEG2                       0x0040A400
+#define HDP_BASE__INST0_SEG3                       0
+#define HDP_BASE__INST0_SEG4                       0
+#define HDP_BASE__INST0_SEG5                       0
+
+#define HDP_BASE__INST1_SEG0                       0
+#define HDP_BASE__INST1_SEG1                       0
+#define HDP_BASE__INST1_SEG2                       0
+#define HDP_BASE__INST1_SEG3                       0
+#define HDP_BASE__INST1_SEG4                       0
+#define HDP_BASE__INST1_SEG5                       0
+
+#define HDP_BASE__INST2_SEG0                       0
+#define HDP_BASE__INST2_SEG1                       0
+#define HDP_BASE__INST2_SEG2                       0
+#define HDP_BASE__INST2_SEG3                       0
+#define HDP_BASE__INST2_SEG4                       0
+#define HDP_BASE__INST2_SEG5                       0
+
+#define HDP_BASE__INST3_SEG0                       0
+#define HDP_BASE__INST3_SEG1                       0
+#define HDP_BASE__INST3_SEG2                       0
+#define HDP_BASE__INST3_SEG3                       0
+#define HDP_BASE__INST3_SEG4                       0
+#define HDP_BASE__INST3_SEG5                       0
+
+#define HDP_BASE__INST4_SEG0                       0
+#define HDP_BASE__INST4_SEG1                       0
+#define HDP_BASE__INST4_SEG2                       0
+#define HDP_BASE__INST4_SEG3                       0
+#define HDP_BASE__INST4_SEG4                       0
+#define HDP_BASE__INST4_SEG5                       0
+
+#define HDP_BASE__INST5_SEG0                       0
+#define HDP_BASE__INST5_SEG1                       0
+#define HDP_BASE__INST5_SEG2                       0
+#define HDP_BASE__INST5_SEG3                       0
+#define HDP_BASE__INST5_SEG4                       0
+#define HDP_BASE__INST5_SEG5                       0
+
+#define HDP_BASE__INST6_SEG0                       0
+#define HDP_BASE__INST6_SEG1                       0
+#define HDP_BASE__INST6_SEG2                       0
+#define HDP_BASE__INST6_SEG3                       0
+#define HDP_BASE__INST6_SEG4                       0
+#define HDP_BASE__INST6_SEG5                       0
+
+#define HDP_BASE__INST7_SEG0                       0
+#define HDP_BASE__INST7_SEG1                       0
+#define HDP_BASE__INST7_SEG2                       0
+#define HDP_BASE__INST7_SEG3                       0
+#define HDP_BASE__INST7_SEG4                       0
+#define HDP_BASE__INST7_SEG5                       0
+
+#define MMHUB_BASE__INST0_SEG0                     0x00012440
+#define MMHUB_BASE__INST0_SEG1                     0x0001A000
+#define MMHUB_BASE__INST0_SEG2                     0x00408800
+#define MMHUB_BASE__INST0_SEG3                     0
+#define MMHUB_BASE__INST0_SEG4                     0
+#define MMHUB_BASE__INST0_SEG5                     0
+
+#define MMHUB_BASE__INST1_SEG0                     0
+#define MMHUB_BASE__INST1_SEG1                     0
+#define MMHUB_BASE__INST1_SEG2                     0
+#define MMHUB_BASE__INST1_SEG3                     0
+#define MMHUB_BASE__INST1_SEG4                     0
+#define MMHUB_BASE__INST1_SEG5                     0
+
+#define MMHUB_BASE__INST2_SEG0                     0
+#define MMHUB_BASE__INST2_SEG1                     0
+#define MMHUB_BASE__INST2_SEG2                     0
+#define MMHUB_BASE__INST2_SEG3                     0
+#define MMHUB_BASE__INST2_SEG4                     0
+#define MMHUB_BASE__INST2_SEG5                     0
+
+#define MMHUB_BASE__INST3_SEG0                     0
+#define MMHUB_BASE__INST3_SEG1                     0
+#define MMHUB_BASE__INST3_SEG2                     0
+#define MMHUB_BASE__INST3_SEG3                     0
+#define MMHUB_BASE__INST3_SEG4                     0
+#define MMHUB_BASE__INST3_SEG5                     0
+
+#define MMHUB_BASE__INST4_SEG0                     0
+#define MMHUB_BASE__INST4_SEG1                     0
+#define MMHUB_BASE__INST4_SEG2                     0
+#define MMHUB_BASE__INST4_SEG3                     0
+#define MMHUB_BASE__INST4_SEG4                     0
+#define MMHUB_BASE__INST4_SEG5                     0
+
+#define MMHUB_BASE__INST5_SEG0                     0
+#define MMHUB_BASE__INST5_SEG1                     0
+#define MMHUB_BASE__INST5_SEG2                     0
+#define MMHUB_BASE__INST5_SEG3                     0
+#define MMHUB_BASE__INST5_SEG4                     0
+#define MMHUB_BASE__INST5_SEG5                     0
+
+#define MMHUB_BASE__INST6_SEG0                     0
+#define MMHUB_BASE__INST6_SEG1                     0
+#define MMHUB_BASE__INST6_SEG2                     0
+#define MMHUB_BASE__INST6_SEG3                     0
+#define MMHUB_BASE__INST6_SEG4                     0
+#define MMHUB_BASE__INST6_SEG5                     0
+
+#define MMHUB_BASE__INST7_SEG0                     0
+#define MMHUB_BASE__INST7_SEG1                     0
+#define MMHUB_BASE__INST7_SEG2                     0
+#define MMHUB_BASE__INST7_SEG3                     0
+#define MMHUB_BASE__INST7_SEG4                     0
+#define MMHUB_BASE__INST7_SEG5                     0
+
+#define MP0_BASE__INST0_SEG0                       0x00013FE0
+#define MP0_BASE__INST0_SEG1                       0x00016000
+#define MP0_BASE__INST0_SEG2                       0x0043FC00
+#define MP0_BASE__INST0_SEG3                       0x00DC0000
+#define MP0_BASE__INST0_SEG4                       0x00E00000
+#define MP0_BASE__INST0_SEG5                       0x00E40000
+
+#define MP0_BASE__INST1_SEG0                       0
+#define MP0_BASE__INST1_SEG1                       0
+#define MP0_BASE__INST1_SEG2                       0
+#define MP0_BASE__INST1_SEG3                       0
+#define MP0_BASE__INST1_SEG4                       0
+#define MP0_BASE__INST1_SEG5                       0
+
+#define MP0_BASE__INST2_SEG0                       0
+#define MP0_BASE__INST2_SEG1                       0
+#define MP0_BASE__INST2_SEG2                       0
+#define MP0_BASE__INST2_SEG3                       0
+#define MP0_BASE__INST2_SEG4                       0
+#define MP0_BASE__INST2_SEG5                       0
+
+#define MP0_BASE__INST3_SEG0                       0
+#define MP0_BASE__INST3_SEG1                       0
+#define MP0_BASE__INST3_SEG2                       0
+#define MP0_BASE__INST3_SEG3                       0
+#define MP0_BASE__INST3_SEG4                       0
+#define MP0_BASE__INST3_SEG5                       0
+
+#define MP0_BASE__INST4_SEG0                       0
+#define MP0_BASE__INST4_SEG1                       0
+#define MP0_BASE__INST4_SEG2                       0
+#define MP0_BASE__INST4_SEG3                       0
+#define MP0_BASE__INST4_SEG4                       0
+#define MP0_BASE__INST4_SEG5                       0
+
+#define MP0_BASE__INST5_SEG0                       0
+#define MP0_BASE__INST5_SEG1                       0
+#define MP0_BASE__INST5_SEG2                       0
+#define MP0_BASE__INST5_SEG3                       0
+#define MP0_BASE__INST5_SEG4                       0
+#define MP0_BASE__INST5_SEG5                       0
+
+#define MP0_BASE__INST6_SEG0                       0
+#define MP0_BASE__INST6_SEG1                       0
+#define MP0_BASE__INST6_SEG2                       0
+#define MP0_BASE__INST6_SEG3                       0
+#define MP0_BASE__INST6_SEG4                       0
+#define MP0_BASE__INST6_SEG5                       0
+
+#define MP0_BASE__INST7_SEG0                       0
+#define MP0_BASE__INST7_SEG1                       0
+#define MP0_BASE__INST7_SEG2                       0
+#define MP0_BASE__INST7_SEG3                       0
+#define MP0_BASE__INST7_SEG4                       0
+#define MP0_BASE__INST7_SEG5                       0
+
+#define MP1_BASE__INST0_SEG0                       0x00012020
+#define MP1_BASE__INST0_SEG1                       0x00016200
+#define MP1_BASE__INST0_SEG2                       0x00400400
+#define MP1_BASE__INST0_SEG3                       0x00E80000
+#define MP1_BASE__INST0_SEG4                       0x00EC0000
+#define MP1_BASE__INST0_SEG5                       0x00F00000
+
+#define MP1_BASE__INST1_SEG0                       0
+#define MP1_BASE__INST1_SEG1                       0
+#define MP1_BASE__INST1_SEG2                       0
+#define MP1_BASE__INST1_SEG3                       0
+#define MP1_BASE__INST1_SEG4                       0
+#define MP1_BASE__INST1_SEG5                       0
+
+#define MP1_BASE__INST2_SEG0                       0
+#define MP1_BASE__INST2_SEG1                       0
+#define MP1_BASE__INST2_SEG2                       0
+#define MP1_BASE__INST2_SEG3                       0
+#define MP1_BASE__INST2_SEG4                       0
+#define MP1_BASE__INST2_SEG5                       0
+
+#define MP1_BASE__INST3_SEG0                       0
+#define MP1_BASE__INST3_SEG1                       0
+#define MP1_BASE__INST3_SEG2                       0
+#define MP1_BASE__INST3_SEG3                       0
+#define MP1_BASE__INST3_SEG4                       0
+#define MP1_BASE__INST3_SEG5                       0
+
+#define MP1_BASE__INST4_SEG0                       0
+#define MP1_BASE__INST4_SEG1                       0
+#define MP1_BASE__INST4_SEG2                       0
+#define MP1_BASE__INST4_SEG3                       0
+#define MP1_BASE__INST4_SEG4                       0
+#define MP1_BASE__INST4_SEG5                       0
+
+#define MP1_BASE__INST5_SEG0                       0
+#define MP1_BASE__INST5_SEG1                       0
+#define MP1_BASE__INST5_SEG2                       0
+#define MP1_BASE__INST5_SEG3                       0
+#define MP1_BASE__INST5_SEG4                       0
+#define MP1_BASE__INST5_SEG5                       0
+
+#define MP1_BASE__INST6_SEG0                       0
+#define MP1_BASE__INST6_SEG1                       0
+#define MP1_BASE__INST6_SEG2                       0
+#define MP1_BASE__INST6_SEG3                       0
+#define MP1_BASE__INST6_SEG4                       0
+#define MP1_BASE__INST6_SEG5                       0
+
+#define MP1_BASE__INST7_SEG0                       0
+#define MP1_BASE__INST7_SEG1                       0
+#define MP1_BASE__INST7_SEG2                       0
+#define MP1_BASE__INST7_SEG3                       0
+#define MP1_BASE__INST7_SEG4                       0
+#define MP1_BASE__INST7_SEG5                       0
+
+#define NBIF0_BASE__INST0_SEG0                     0x00000000
+#define NBIF0_BASE__INST0_SEG1                     0x00000014
+#define NBIF0_BASE__INST0_SEG2                     0x00000D20
+#define NBIF0_BASE__INST0_SEG3                     0x00010400
+#define NBIF0_BASE__INST0_SEG4                     0x00012D80
+#define NBIF0_BASE__INST0_SEG5                     0x0041B000
+
+#define NBIF0_BASE__INST1_SEG0                     0
+#define NBIF0_BASE__INST1_SEG1                     0
+#define NBIF0_BASE__INST1_SEG2                     0
+#define NBIF0_BASE__INST1_SEG3                     0
+#define NBIF0_BASE__INST1_SEG4                     0
+#define NBIF0_BASE__INST1_SEG5                     0
+
+#define NBIF0_BASE__INST2_SEG0                     0
+#define NBIF0_BASE__INST2_SEG1                     0
+#define NBIF0_BASE__INST2_SEG2                     0
+#define NBIF0_BASE__INST2_SEG3                     0
+#define NBIF0_BASE__INST2_SEG4                     0
+#define NBIF0_BASE__INST2_SEG5                     0
+
+#define NBIF0_BASE__INST3_SEG0                     0
+#define NBIF0_BASE__INST3_SEG1                     0
+#define NBIF0_BASE__INST3_SEG2                     0
+#define NBIF0_BASE__INST3_SEG3                     0
+#define NBIF0_BASE__INST3_SEG4                     0
+#define NBIF0_BASE__INST3_SEG5                     0
+
+#define NBIF0_BASE__INST4_SEG0                     0
+#define NBIF0_BASE__INST4_SEG1                     0
+#define NBIF0_BASE__INST4_SEG2                     0
+#define NBIF0_BASE__INST4_SEG3                     0
+#define NBIF0_BASE__INST4_SEG4                     0
+#define NBIF0_BASE__INST4_SEG5                     0
+
+#define NBIF0_BASE__INST5_SEG0                     0
+#define NBIF0_BASE__INST5_SEG1                     0
+#define NBIF0_BASE__INST5_SEG2                     0
+#define NBIF0_BASE__INST5_SEG3                     0
+#define NBIF0_BASE__INST5_SEG4                     0
+#define NBIF0_BASE__INST5_SEG5                     0
+
+#define NBIF0_BASE__INST6_SEG0                     0
+#define NBIF0_BASE__INST6_SEG1                     0
+#define NBIF0_BASE__INST6_SEG2                     0
+#define NBIF0_BASE__INST6_SEG3                     0
+#define NBIF0_BASE__INST6_SEG4                     0
+#define NBIF0_BASE__INST6_SEG5                     0
+
+#define NBIF0_BASE__INST7_SEG0                     0
+#define NBIF0_BASE__INST7_SEG1                     0
+#define NBIF0_BASE__INST7_SEG2                     0
+#define NBIF0_BASE__INST7_SEG3                     0
+#define NBIF0_BASE__INST7_SEG4                     0
+#define NBIF0_BASE__INST7_SEG5                     0
+
+#define OSSSYS_BASE__INST0_SEG0                    0x000010A0
+#define OSSSYS_BASE__INST0_SEG1                    0x00012500
+#define OSSSYS_BASE__INST0_SEG2                    0x0040A000
+#define OSSSYS_BASE__INST0_SEG3                    0
+#define OSSSYS_BASE__INST0_SEG4                    0
+#define OSSSYS_BASE__INST0_SEG5                    0
+
+#define OSSSYS_BASE__INST1_SEG0                    0
+#define OSSSYS_BASE__INST1_SEG1                    0
+#define OSSSYS_BASE__INST1_SEG2                    0
+#define OSSSYS_BASE__INST1_SEG3                    0
+#define OSSSYS_BASE__INST1_SEG4                    0
+#define OSSSYS_BASE__INST1_SEG5                    0
+
+#define OSSSYS_BASE__INST2_SEG0                    0
+#define OSSSYS_BASE__INST2_SEG1                    0
+#define OSSSYS_BASE__INST2_SEG2                    0
+#define OSSSYS_BASE__INST2_SEG3                    0
+#define OSSSYS_BASE__INST2_SEG4                    0
+#define OSSSYS_BASE__INST2_SEG5                    0
+
+#define OSSSYS_BASE__INST3_SEG0                    0
+#define OSSSYS_BASE__INST3_SEG1                    0
+#define OSSSYS_BASE__INST3_SEG2                    0
+#define OSSSYS_BASE__INST3_SEG3                    0
+#define OSSSYS_BASE__INST3_SEG4                    0
+#define OSSSYS_BASE__INST3_SEG5                    0
+
+#define OSSSYS_BASE__INST4_SEG0                    0
+#define OSSSYS_BASE__INST4_SEG1                    0
+#define OSSSYS_BASE__INST4_SEG2                    0
+#define OSSSYS_BASE__INST4_SEG3                    0
+#define OSSSYS_BASE__INST4_SEG4                    0
+#define OSSSYS_BASE__INST4_SEG5                    0
+
+#define OSSSYS_BASE__INST5_SEG0                    0
+#define OSSSYS_BASE__INST5_SEG1                    0
+#define OSSSYS_BASE__INST5_SEG2                    0
+#define OSSSYS_BASE__INST5_SEG3                    0
+#define OSSSYS_BASE__INST5_SEG4                    0
+#define OSSSYS_BASE__INST5_SEG5                    0
+
+#define OSSSYS_BASE__INST6_SEG0                    0
+#define OSSSYS_BASE__INST6_SEG1                    0
+#define OSSSYS_BASE__INST6_SEG2                    0
+#define OSSSYS_BASE__INST6_SEG3                    0
+#define OSSSYS_BASE__INST6_SEG4                    0
+#define OSSSYS_BASE__INST6_SEG5                    0
+
+#define OSSSYS_BASE__INST7_SEG0                    0
+#define OSSSYS_BASE__INST7_SEG1                    0
+#define OSSSYS_BASE__INST7_SEG2                    0
+#define OSSSYS_BASE__INST7_SEG3                    0
+#define OSSSYS_BASE__INST7_SEG4                    0
+#define OSSSYS_BASE__INST7_SEG5                    0
+
+#define PCIE0_BASE__INST0_SEG0                     0x000128C0
+#define PCIE0_BASE__INST0_SEG1                     0x00411800
+#define PCIE0_BASE__INST0_SEG2                     0x04440000
+#define PCIE0_BASE__INST0_SEG3                     0
+#define PCIE0_BASE__INST0_SEG4                     0
+#define PCIE0_BASE__INST0_SEG5                     0
+
+#define PCIE0_BASE__INST1_SEG0                     0
+#define PCIE0_BASE__INST1_SEG1                     0
+#define PCIE0_BASE__INST1_SEG2                     0
+#define PCIE0_BASE__INST1_SEG3                     0
+#define PCIE0_BASE__INST1_SEG4                     0
+#define PCIE0_BASE__INST1_SEG5                     0
+
+#define PCIE0_BASE__INST2_SEG0                     0
+#define PCIE0_BASE__INST2_SEG1                     0
+#define PCIE0_BASE__INST2_SEG2                     0
+#define PCIE0_BASE__INST2_SEG3                     0
+#define PCIE0_BASE__INST2_SEG4                     0
+#define PCIE0_BASE__INST2_SEG5                     0
+
+#define PCIE0_BASE__INST3_SEG0                     0
+#define PCIE0_BASE__INST3_SEG1                     0
+#define PCIE0_BASE__INST3_SEG2                     0
+#define PCIE0_BASE__INST3_SEG3                     0
+#define PCIE0_BASE__INST3_SEG4                     0
+#define PCIE0_BASE__INST3_SEG5                     0
+
+#define PCIE0_BASE__INST4_SEG0                     0
+#define PCIE0_BASE__INST4_SEG1                     0
+#define PCIE0_BASE__INST4_SEG2                     0
+#define PCIE0_BASE__INST4_SEG3                     0
+#define PCIE0_BASE__INST4_SEG4                     0
+#define PCIE0_BASE__INST4_SEG5                     0
+
+#define PCIE0_BASE__INST5_SEG0                     0
+#define PCIE0_BASE__INST5_SEG1                     0
+#define PCIE0_BASE__INST5_SEG2                     0
+#define PCIE0_BASE__INST5_SEG3                     0
+#define PCIE0_BASE__INST5_SEG4                     0
+#define PCIE0_BASE__INST5_SEG5                     0
+
+#define PCIE0_BASE__INST6_SEG0                     0
+#define PCIE0_BASE__INST6_SEG1                     0
+#define PCIE0_BASE__INST6_SEG2                     0
+#define PCIE0_BASE__INST6_SEG3                     0
+#define PCIE0_BASE__INST6_SEG4                     0
+#define PCIE0_BASE__INST6_SEG5                     0
+
+#define PCIE0_BASE__INST7_SEG0                     0
+#define PCIE0_BASE__INST7_SEG1                     0
+#define PCIE0_BASE__INST7_SEG2                     0
+#define PCIE0_BASE__INST7_SEG3                     0
+#define PCIE0_BASE__INST7_SEG4                     0
+#define PCIE0_BASE__INST7_SEG5                     0
+
+#define SDMA0_BASE__INST0_SEG0                     0x00001260
+#define SDMA0_BASE__INST0_SEG1                     0x00012540
+#define SDMA0_BASE__INST0_SEG2                     0x0040A800
+#define SDMA0_BASE__INST0_SEG3                     0
+#define SDMA0_BASE__INST0_SEG4                     0
+#define SDMA0_BASE__INST0_SEG5                     0
+
+#define SDMA0_BASE__INST1_SEG0                     0
+#define SDMA0_BASE__INST1_SEG1                     0
+#define SDMA0_BASE__INST1_SEG2                     0
+#define SDMA0_BASE__INST1_SEG3                     0
+#define SDMA0_BASE__INST1_SEG4                     0
+#define SDMA0_BASE__INST1_SEG5                     0
+
+#define SDMA0_BASE__INST2_SEG0                     0
+#define SDMA0_BASE__INST2_SEG1                     0
+#define SDMA0_BASE__INST2_SEG2                     0
+#define SDMA0_BASE__INST2_SEG3                     0
+#define SDMA0_BASE__INST2_SEG4                     0
+#define SDMA0_BASE__INST2_SEG5                     0
+
+#define SDMA0_BASE__INST3_SEG0                     0
+#define SDMA0_BASE__INST3_SEG1                     0
+#define SDMA0_BASE__INST3_SEG2                     0
+#define SDMA0_BASE__INST3_SEG3                     0
+#define SDMA0_BASE__INST3_SEG4                     0
+#define SDMA0_BASE__INST3_SEG5                     0
+
+#define SDMA0_BASE__INST4_SEG0                     0
+#define SDMA0_BASE__INST4_SEG1                     0
+#define SDMA0_BASE__INST4_SEG2                     0
+#define SDMA0_BASE__INST4_SEG3                     0
+#define SDMA0_BASE__INST4_SEG4                     0
+#define SDMA0_BASE__INST4_SEG5                     0
+
+#define SDMA0_BASE__INST5_SEG0                     0
+#define SDMA0_BASE__INST5_SEG1                     0
+#define SDMA0_BASE__INST5_SEG2                     0
+#define SDMA0_BASE__INST5_SEG3                     0
+#define SDMA0_BASE__INST5_SEG4                     0
+#define SDMA0_BASE__INST5_SEG5                     0
+
+#define SDMA0_BASE__INST6_SEG0                     0
+#define SDMA0_BASE__INST6_SEG1                     0
+#define SDMA0_BASE__INST6_SEG2                     0
+#define SDMA0_BASE__INST6_SEG3                     0
+#define SDMA0_BASE__INST6_SEG4                     0
+#define SDMA0_BASE__INST6_SEG5                     0
+
+#define SDMA1_BASE__INST0_SEG0                     0x00001860
+#define SDMA1_BASE__INST0_SEG1                     0x00012560
+#define SDMA1_BASE__INST0_SEG2                     0x0040AC00
+#define SDMA1_BASE__INST0_SEG3                     0
+#define SDMA1_BASE__INST0_SEG4                     0
+#define SDMA1_BASE__INST0_SEG5                     0
+
+#define SDMA1_BASE__INST1_SEG0                     0
+#define SDMA1_BASE__INST1_SEG1                     0
+#define SDMA1_BASE__INST1_SEG2                     0
+#define SDMA1_BASE__INST1_SEG3                     0
+#define SDMA1_BASE__INST1_SEG4                     0
+#define SDMA1_BASE__INST1_SEG5                     0
+
+#define SDMA1_BASE__INST2_SEG0                     0
+#define SDMA1_BASE__INST2_SEG1                     0
+#define SDMA1_BASE__INST2_SEG2                     0
+#define SDMA1_BASE__INST2_SEG3                     0
+#define SDMA1_BASE__INST2_SEG4                     0
+#define SDMA1_BASE__INST2_SEG5                     0
+
+#define SDMA1_BASE__INST3_SEG0                     0
+#define SDMA1_BASE__INST3_SEG1                     0
+#define SDMA1_BASE__INST3_SEG2                     0
+#define SDMA1_BASE__INST3_SEG3                     0
+#define SDMA1_BASE__INST3_SEG4                     0
+#define SDMA1_BASE__INST3_SEG5                     0
+
+#define SDMA1_BASE__INST4_SEG0                     0
+#define SDMA1_BASE__INST4_SEG1                     0
+#define SDMA1_BASE__INST4_SEG2                     0
+#define SDMA1_BASE__INST4_SEG3                     0
+#define SDMA1_BASE__INST4_SEG4                     0
+#define SDMA1_BASE__INST4_SEG5                     0
+
+#define SDMA1_BASE__INST5_SEG0                     0
+#define SDMA1_BASE__INST5_SEG1                     0
+#define SDMA1_BASE__INST5_SEG2                     0
+#define SDMA1_BASE__INST5_SEG3                     0
+#define SDMA1_BASE__INST5_SEG4                     0
+#define SDMA1_BASE__INST5_SEG5                     0
+
+
+#define SDMA1_BASE__INST6_SEG0                     0
+#define SDMA1_BASE__INST6_SEG1                     0
+#define SDMA1_BASE__INST6_SEG2                     0
+#define SDMA1_BASE__INST6_SEG3                     0
+#define SDMA1_BASE__INST6_SEG4                     0
+#define SDMA1_BASE__INST6_SEG5                     0
+
+
+#define SDMA2_BASE__INST0_SEG0                     0x00013760
+#define SDMA2_BASE__INST0_SEG1                     0x0001E000
+#define SDMA2_BASE__INST0_SEG2                     0x0042EC00
+#define SDMA2_BASE__INST0_SEG3                     0
+#define SDMA2_BASE__INST0_SEG4                     0
+#define SDMA2_BASE__INST0_SEG5                     0
+
+
+#define SDMA2_BASE__INST1_SEG0                     0
+#define SDMA2_BASE__INST1_SEG1                     0
+#define SDMA2_BASE__INST1_SEG2                     0
+#define SDMA2_BASE__INST1_SEG3                     0
+#define SDMA2_BASE__INST1_SEG4                     0
+#define SDMA2_BASE__INST1_SEG5                     0
+
+#define SDMA2_BASE__INST2_SEG0                     0
+#define SDMA2_BASE__INST2_SEG1                     0
+#define SDMA2_BASE__INST2_SEG2                     0
+#define SDMA2_BASE__INST2_SEG3                     0
+#define SDMA2_BASE__INST2_SEG4                     0
+#define SDMA2_BASE__INST2_SEG5                     0
+
+#define SDMA2_BASE__INST3_SEG0                     0
+#define SDMA2_BASE__INST3_SEG1                     0
+#define SDMA2_BASE__INST3_SEG2                     0
+#define SDMA2_BASE__INST3_SEG3                     0
+#define SDMA2_BASE__INST3_SEG4                     0
+#define SDMA2_BASE__INST3_SEG5                     0
+
+#define SDMA2_BASE__INST4_SEG0                     0
+#define SDMA2_BASE__INST4_SEG1                     0
+#define SDMA2_BASE__INST4_SEG2                     0
+#define SDMA2_BASE__INST4_SEG3                     0
+#define SDMA2_BASE__INST4_SEG4                     0
+#define SDMA2_BASE__INST4_SEG5                     0
+
+#define SDMA2_BASE__INST5_SEG0                     0
+#define SDMA2_BASE__INST5_SEG1                     0
+#define SDMA2_BASE__INST5_SEG2                     0
+#define SDMA2_BASE__INST5_SEG3                     0
+#define SDMA2_BASE__INST5_SEG4                     0
+#define SDMA2_BASE__INST5_SEG5                     0
+
+#define SDMA2_BASE__INST6_SEG0                     0
+#define SDMA2_BASE__INST6_SEG1                     0
+#define SDMA2_BASE__INST6_SEG2                     0
+#define SDMA2_BASE__INST6_SEG3                     0
+#define SDMA2_BASE__INST6_SEG4                     0
+#define SDMA2_BASE__INST6_SEG5                     0
+
+#define SDMA3_BASE__INST0_SEG0                     0x00013780
+#define SDMA3_BASE__INST0_SEG1                     0x0001E400
+#define SDMA3_BASE__INST0_SEG2                     0x0042F000
+#define SDMA3_BASE__INST0_SEG3                     0
+#define SDMA3_BASE__INST0_SEG4                     0
+#define SDMA3_BASE__INST0_SEG5                     0
+
+#define SDMA3_BASE__INST1_SEG0                     0
+#define SDMA3_BASE__INST1_SEG1                     0
+#define SDMA3_BASE__INST1_SEG2                     0
+#define SDMA3_BASE__INST1_SEG3                     0
+#define SDMA3_BASE__INST1_SEG4                     0
+#define SDMA3_BASE__INST1_SEG5                     0
+
+#define SDMA3_BASE__INST2_SEG0                     0
+#define SDMA3_BASE__INST2_SEG1                     0
+#define SDMA3_BASE__INST2_SEG2                     0
+#define SDMA3_BASE__INST2_SEG3                     0
+#define SDMA3_BASE__INST2_SEG4                     0
+#define SDMA3_BASE__INST2_SEG5                     0
+
+#define SDMA3_BASE__INST3_SEG0                     0
+#define SDMA3_BASE__INST3_SEG1                     0
+#define SDMA3_BASE__INST3_SEG2                     0
+#define SDMA3_BASE__INST3_SEG3                     0
+#define SDMA3_BASE__INST3_SEG4                     0
+#define SDMA3_BASE__INST3_SEG5                     0
+
+#define SDMA3_BASE__INST4_SEG0                     0
+#define SDMA3_BASE__INST4_SEG1                     0
+#define SDMA3_BASE__INST4_SEG2                     0
+#define SDMA3_BASE__INST4_SEG3                     0
+#define SDMA3_BASE__INST4_SEG4                     0
+#define SDMA3_BASE__INST4_SEG5                     0
+
+#define SDMA3_BASE__INST5_SEG0                     0
+#define SDMA3_BASE__INST5_SEG1                     0
+#define SDMA3_BASE__INST5_SEG2                     0
+#define SDMA3_BASE__INST5_SEG3                     0
+#define SDMA3_BASE__INST5_SEG4                     0
+#define SDMA3_BASE__INST5_SEG5                     0
+
+#define SDMA3_BASE__INST6_SEG0                     0
+#define SDMA3_BASE__INST6_SEG1                     0
+#define SDMA3_BASE__INST6_SEG2                     0
+#define SDMA3_BASE__INST6_SEG3                     0
+#define SDMA3_BASE__INST6_SEG4                     0
+#define SDMA3_BASE__INST6_SEG5                     0
+
+#define SDMA4_BASE__INST0_SEG0                     0x000137A0
+#define SDMA4_BASE__INST0_SEG1                     0x0001E800
+#define SDMA4_BASE__INST0_SEG2                     0x0042F400
+#define SDMA4_BASE__INST0_SEG3                     0
+#define SDMA4_BASE__INST0_SEG4                     0
+#define SDMA4_BASE__INST0_SEG5                     0
+
+#define SDMA4_BASE__INST1_SEG0                     0
+#define SDMA4_BASE__INST1_SEG1                     0
+#define SDMA4_BASE__INST1_SEG2                     0
+#define SDMA4_BASE__INST1_SEG3                     0
+#define SDMA4_BASE__INST1_SEG4                     0
+#define SDMA4_BASE__INST1_SEG5                     0
+
+#define SDMA4_BASE__INST2_SEG0                     0
+#define SDMA4_BASE__INST2_SEG1                     0
+#define SDMA4_BASE__INST2_SEG2                     0
+#define SDMA4_BASE__INST2_SEG3                     0
+#define SDMA4_BASE__INST2_SEG4                     0
+#define SDMA4_BASE__INST2_SEG5                     0
+
+#define SDMA4_BASE__INST3_SEG0                     0
+#define SDMA4_BASE__INST3_SEG1                     0
+#define SDMA4_BASE__INST3_SEG2                     0
+#define SDMA4_BASE__INST3_SEG3                     0
+#define SDMA4_BASE__INST3_SEG4                     0
+#define SDMA4_BASE__INST3_SEG5                     0
+
+#define SDMA4_BASE__INST4_SEG0                     0
+#define SDMA4_BASE__INST4_SEG1                     0
+#define SDMA4_BASE__INST4_SEG2                     0
+#define SDMA4_BASE__INST4_SEG3                     0
+#define SDMA4_BASE__INST4_SEG4                     0
+#define SDMA4_BASE__INST4_SEG5                     0
+
+#define SDMA4_BASE__INST5_SEG0                     0
+#define SDMA4_BASE__INST5_SEG1                     0
+#define SDMA4_BASE__INST5_SEG2                     0
+#define SDMA4_BASE__INST5_SEG3                     0
+#define SDMA4_BASE__INST5_SEG4                     0
+#define SDMA4_BASE__INST5_SEG5                     0
+
+#define SDMA4_BASE__INST6_SEG0                     0
+#define SDMA4_BASE__INST6_SEG1                     0
+#define SDMA4_BASE__INST6_SEG2                     0
+#define SDMA4_BASE__INST6_SEG3                     0
+#define SDMA4_BASE__INST6_SEG4                     0
+#define SDMA4_BASE__INST6_SEG5                     0
+
+#define SDMA5_BASE__INST0_SEG0                     0x000137C0
+#define SDMA5_BASE__INST0_SEG1                     0x0001EC00
+#define SDMA5_BASE__INST0_SEG2                     0x0042F800
+#define SDMA5_BASE__INST0_SEG3                     0
+#define SDMA5_BASE__INST0_SEG4                     0
+#define SDMA5_BASE__INST0_SEG5                     0
+
+#define SDMA5_BASE__INST1_SEG0                     0
+#define SDMA5_BASE__INST1_SEG1                     0
+#define SDMA5_BASE__INST1_SEG2                     0
+#define SDMA5_BASE__INST1_SEG3                     0
+#define SDMA5_BASE__INST1_SEG4                     0
+#define SDMA5_BASE__INST1_SEG5                     0
+
+#define SDMA5_BASE__INST2_SEG0                     0
+#define SDMA5_BASE__INST2_SEG1                     0
+#define SDMA5_BASE__INST2_SEG2                     0
+#define SDMA5_BASE__INST2_SEG3                     0
+#define SDMA5_BASE__INST2_SEG4                     0
+#define SDMA5_BASE__INST2_SEG5                     0
+
+#define SDMA5_BASE__INST3_SEG0                     0
+#define SDMA5_BASE__INST3_SEG1                     0
+#define SDMA5_BASE__INST3_SEG2                     0
+#define SDMA5_BASE__INST3_SEG3                     0
+#define SDMA5_BASE__INST3_SEG4                     0
+#define SDMA5_BASE__INST3_SEG5                     0
+
+#define SDMA5_BASE__INST4_SEG0                     0
+#define SDMA5_BASE__INST4_SEG1                     0
+#define SDMA5_BASE__INST4_SEG2                     0
+#define SDMA5_BASE__INST4_SEG3                     0
+#define SDMA5_BASE__INST4_SEG4                     0
+#define SDMA5_BASE__INST4_SEG5                     0
+
+#define SDMA5_BASE__INST5_SEG0                     0
+#define SDMA5_BASE__INST5_SEG1                     0
+#define SDMA5_BASE__INST5_SEG2                     0
+#define SDMA5_BASE__INST5_SEG3                     0
+#define SDMA5_BASE__INST5_SEG4                     0
+#define SDMA5_BASE__INST5_SEG5                     0
+
+#define SDMA5_BASE__INST6_SEG0                     0
+#define SDMA5_BASE__INST6_SEG1                     0
+#define SDMA5_BASE__INST6_SEG2                     0
+#define SDMA5_BASE__INST6_SEG3                     0
+#define SDMA5_BASE__INST6_SEG4                     0
+#define SDMA5_BASE__INST6_SEG5                     0
+
+#define SDMA6_BASE__INST0_SEG0                     0x000137E0
+#define SDMA6_BASE__INST0_SEG1                     0x0001F000
+#define SDMA6_BASE__INST0_SEG2                     0x0042FC00
+#define SDMA6_BASE__INST0_SEG3                     0
+#define SDMA6_BASE__INST0_SEG4                     0
+#define SDMA6_BASE__INST0_SEG5                     0
+
+#define SDMA6_BASE__INST1_SEG0                     0
+#define SDMA6_BASE__INST1_SEG1                     0
+#define SDMA6_BASE__INST1_SEG2                     0
+#define SDMA6_BASE__INST1_SEG3                     0
+#define SDMA6_BASE__INST1_SEG4                     0
+#define SDMA6_BASE__INST1_SEG5                     0
+
+#define SDMA6_BASE__INST2_SEG0                     0
+#define SDMA6_BASE__INST2_SEG1                     0
+#define SDMA6_BASE__INST2_SEG2                     0
+#define SDMA6_BASE__INST2_SEG3                     0
+#define SDMA6_BASE__INST2_SEG4                     0
+#define SDMA6_BASE__INST2_SEG5                     0
+
+#define SDMA6_BASE__INST3_SEG0                     0
+#define SDMA6_BASE__INST3_SEG1                     0
+#define SDMA6_BASE__INST3_SEG2                     0
+#define SDMA6_BASE__INST3_SEG3                     0
+#define SDMA6_BASE__INST3_SEG4                     0
+#define SDMA6_BASE__INST3_SEG5                     0
+
+#define SDMA6_BASE__INST4_SEG0                     0
+#define SDMA6_BASE__INST4_SEG1                     0
+#define SDMA6_BASE__INST4_SEG2                     0
+#define SDMA6_BASE__INST4_SEG3                     0
+#define SDMA6_BASE__INST4_SEG4                     0
+#define SDMA6_BASE__INST4_SEG5                     0
+
+#define SDMA6_BASE__INST5_SEG0                     0
+#define SDMA6_BASE__INST5_SEG1                     0
+#define SDMA6_BASE__INST5_SEG2                     0
+#define SDMA6_BASE__INST5_SEG3                     0
+#define SDMA6_BASE__INST5_SEG4                     0
+#define SDMA6_BASE__INST5_SEG5                     0
+
+#define SDMA6_BASE__INST6_SEG0                     0
+#define SDMA6_BASE__INST6_SEG1                     0
+#define SDMA6_BASE__INST6_SEG2                     0
+#define SDMA6_BASE__INST6_SEG3                     0
+#define SDMA6_BASE__INST6_SEG4                     0
+#define SDMA6_BASE__INST6_SEG5                     0
+
+#define SDMA7_BASE__INST0_SEG0                     0x00013800
+#define SDMA7_BASE__INST0_SEG1                     0x0001F400
+#define SDMA7_BASE__INST0_SEG2                     0x00430000
+#define SDMA7_BASE__INST0_SEG3                     0
+#define SDMA7_BASE__INST0_SEG4                     0
+#define SDMA7_BASE__INST0_SEG5                     0
+
+#define SDMA7_BASE__INST1_SEG0                     0
+#define SDMA7_BASE__INST1_SEG1                     0
+#define SDMA7_BASE__INST1_SEG2                     0
+#define SDMA7_BASE__INST1_SEG3                     0
+#define SDMA7_BASE__INST1_SEG4                     0
+#define SDMA7_BASE__INST1_SEG5                     0
+
+#define SDMA7_BASE__INST2_SEG0                     0
+#define SDMA7_BASE__INST2_SEG1                     0
+#define SDMA7_BASE__INST2_SEG2                     0
+#define SDMA7_BASE__INST2_SEG3                     0
+#define SDMA7_BASE__INST2_SEG4                     0
+#define SDMA7_BASE__INST2_SEG5                     0
+
+#define SDMA7_BASE__INST3_SEG0                     0
+#define SDMA7_BASE__INST3_SEG1                     0
+#define SDMA7_BASE__INST3_SEG2                     0
+#define SDMA7_BASE__INST3_SEG3                     0
+#define SDMA7_BASE__INST3_SEG4                     0
+#define SDMA7_BASE__INST3_SEG5                     0
+
+#define SDMA7_BASE__INST4_SEG0                     0
+#define SDMA7_BASE__INST4_SEG1                     0
+#define SDMA7_BASE__INST4_SEG2                     0
+#define SDMA7_BASE__INST4_SEG3                     0
+#define SDMA7_BASE__INST4_SEG4                     0
+#define SDMA7_BASE__INST4_SEG5                     0
+
+#define SDMA7_BASE__INST5_SEG0                     0
+#define SDMA7_BASE__INST5_SEG1                     0
+#define SDMA7_BASE__INST5_SEG2                     0
+#define SDMA7_BASE__INST5_SEG3                     0
+#define SDMA7_BASE__INST5_SEG4                     0
+#define SDMA7_BASE__INST5_SEG5                     0
+
+#define SDMA7_BASE__INST6_SEG0                     0
+#define SDMA7_BASE__INST6_SEG1                     0
+#define SDMA7_BASE__INST6_SEG2                     0
+#define SDMA7_BASE__INST6_SEG3                     0
+#define SDMA7_BASE__INST6_SEG4                     0
+#define SDMA7_BASE__INST6_SEG5                     0
+
+#define SMUIO_BASE__INST0_SEG0                     0x00012080
+#define SMUIO_BASE__INST0_SEG1                     0x00016800
+#define SMUIO_BASE__INST0_SEG2                     0x00016A00
+#define SMUIO_BASE__INST0_SEG3                     0x00401000
+#define SMUIO_BASE__INST0_SEG4                     0x00440000
+#define SMUIO_BASE__INST0_SEG5                     0
+
+#define SMUIO_BASE__INST1_SEG0                     0
+#define SMUIO_BASE__INST1_SEG1                     0
+#define SMUIO_BASE__INST1_SEG2                     0
+#define SMUIO_BASE__INST1_SEG3                     0
+#define SMUIO_BASE__INST1_SEG4                     0
+#define SMUIO_BASE__INST1_SEG5                     0
+
+#define SMUIO_BASE__INST2_SEG0                     0
+#define SMUIO_BASE__INST2_SEG1                     0
+#define SMUIO_BASE__INST2_SEG2                     0
+#define SMUIO_BASE__INST2_SEG3                     0
+#define SMUIO_BASE__INST2_SEG4                     0
+#define SMUIO_BASE__INST2_SEG5                     0
+
+#define SMUIO_BASE__INST3_SEG0                     0
+#define SMUIO_BASE__INST3_SEG1                     0
+#define SMUIO_BASE__INST3_SEG2                     0
+#define SMUIO_BASE__INST3_SEG3                     0
+#define SMUIO_BASE__INST3_SEG4                     0
+#define SMUIO_BASE__INST3_SEG5                     0
+
+#define SMUIO_BASE__INST4_SEG0                     0
+#define SMUIO_BASE__INST4_SEG1                     0
+#define SMUIO_BASE__INST4_SEG2                     0
+#define SMUIO_BASE__INST4_SEG3                     0
+#define SMUIO_BASE__INST4_SEG4                     0
+#define SMUIO_BASE__INST4_SEG5                     0
+
+#define SMUIO_BASE__INST5_SEG0                     0
+#define SMUIO_BASE__INST5_SEG1                     0
+#define SMUIO_BASE__INST5_SEG2                     0
+#define SMUIO_BASE__INST5_SEG3                     0
+#define SMUIO_BASE__INST5_SEG4                     0
+#define SMUIO_BASE__INST5_SEG5                     0
+
+#define SMUIO_BASE__INST6_SEG0                     0
+#define SMUIO_BASE__INST6_SEG1                     0
+#define SMUIO_BASE__INST6_SEG2                     0
+#define SMUIO_BASE__INST6_SEG3                     0
+#define SMUIO_BASE__INST6_SEG4                     0
+#define SMUIO_BASE__INST6_SEG5                     0
+
+#define SMUIO_BASE__INST7_SEG0                     0
+#define SMUIO_BASE__INST7_SEG1                     0
+#define SMUIO_BASE__INST7_SEG2                     0
+#define SMUIO_BASE__INST7_SEG3                     0
+#define SMUIO_BASE__INST7_SEG4                     0
+#define SMUIO_BASE__INST7_SEG5                     0
+
+#define THM_BASE__INST0_SEG0                       0x00012060
+#define THM_BASE__INST0_SEG1                       0x00016600
+#define THM_BASE__INST0_SEG2                       0x00400C00
+#define THM_BASE__INST0_SEG3                       0
+#define THM_BASE__INST0_SEG4                       0
+#define THM_BASE__INST0_SEG5                       0
+
+#define THM_BASE__INST1_SEG0                       0
+#define THM_BASE__INST1_SEG1                       0
+#define THM_BASE__INST1_SEG2                       0
+#define THM_BASE__INST1_SEG3                       0
+#define THM_BASE__INST1_SEG4                       0
+#define THM_BASE__INST1_SEG5                       0
+
+#define THM_BASE__INST2_SEG0                       0
+#define THM_BASE__INST2_SEG1                       0
+#define THM_BASE__INST2_SEG2                       0
+#define THM_BASE__INST2_SEG3                       0
+#define THM_BASE__INST2_SEG4                       0
+#define THM_BASE__INST2_SEG5                       0
+
+#define THM_BASE__INST3_SEG0                       0
+#define THM_BASE__INST3_SEG1                       0
+#define THM_BASE__INST3_SEG2                       0
+#define THM_BASE__INST3_SEG3                       0
+#define THM_BASE__INST3_SEG4                       0
+#define THM_BASE__INST3_SEG5                       0
+
+#define THM_BASE__INST4_SEG0                       0
+#define THM_BASE__INST4_SEG1                       0
+#define THM_BASE__INST4_SEG2                       0
+#define THM_BASE__INST4_SEG3                       0
+#define THM_BASE__INST4_SEG4                       0
+#define THM_BASE__INST4_SEG5                       0
+
+#define THM_BASE__INST5_SEG0                       0
+#define THM_BASE__INST5_SEG1                       0
+#define THM_BASE__INST5_SEG2                       0
+#define THM_BASE__INST5_SEG3                       0
+#define THM_BASE__INST5_SEG4                       0
+#define THM_BASE__INST5_SEG5                       0
+
+#define THM_BASE__INST6_SEG0                       0
+#define THM_BASE__INST6_SEG1                       0
+#define THM_BASE__INST6_SEG2                       0
+#define THM_BASE__INST6_SEG3                       0
+#define THM_BASE__INST6_SEG4                       0
+#define THM_BASE__INST6_SEG5                       0
+
+#define THM_BASE__INST7_SEG0                       0
+#define THM_BASE__INST7_SEG1                       0
+#define THM_BASE__INST7_SEG2                       0
+#define THM_BASE__INST7_SEG3                       0
+#define THM_BASE__INST7_SEG4                       0
+#define THM_BASE__INST7_SEG5                       0
+
+#define UMC_BASE__INST0_SEG0                       0x000132C0
+#define UMC_BASE__INST0_SEG1                       0x00014000
+#define UMC_BASE__INST0_SEG2                       0x00425800
+#define UMC_BASE__INST0_SEG3                       0
+#define UMC_BASE__INST0_SEG4                       0
+#define UMC_BASE__INST0_SEG5                       0
+
+#define UMC_BASE__INST1_SEG0                       0x000132E0
+#define UMC_BASE__INST1_SEG1                       0x00054000
+#define UMC_BASE__INST1_SEG2                       0x00425C00
+#define UMC_BASE__INST1_SEG3                       0
+#define UMC_BASE__INST1_SEG4                       0
+#define UMC_BASE__INST1_SEG5                       0
+
+#define UMC_BASE__INST2_SEG0                       0x00013300
+#define UMC_BASE__INST2_SEG1                       0x00094000
+#define UMC_BASE__INST2_SEG2                       0x00426000
+#define UMC_BASE__INST2_SEG3                       0
+#define UMC_BASE__INST2_SEG4                       0
+#define UMC_BASE__INST2_SEG5                       0
+
+#define UMC_BASE__INST3_SEG0                       0x00013320
+#define UMC_BASE__INST3_SEG1                       0x000D4000
+#define UMC_BASE__INST3_SEG2                       0x00426400
+#define UMC_BASE__INST3_SEG3                       0
+#define UMC_BASE__INST3_SEG4                       0
+#define UMC_BASE__INST3_SEG5                       0
+
+#define UMC_BASE__INST4_SEG0                       0x00013340
+#define UMC_BASE__INST4_SEG1                       0x00114000
+#define UMC_BASE__INST4_SEG2                       0x00426800
+#define UMC_BASE__INST4_SEG3                       0
+#define UMC_BASE__INST4_SEG4                       0
+#define UMC_BASE__INST4_SEG5                       0
+
+#define UMC_BASE__INST5_SEG0                       0x00013360
+#define UMC_BASE__INST5_SEG1                       0x00154000
+#define UMC_BASE__INST5_SEG2                       0x00426C00
+#define UMC_BASE__INST5_SEG3                       0
+#define UMC_BASE__INST5_SEG4                       0
+#define UMC_BASE__INST5_SEG5                       0
+
+#define UMC_BASE__INST6_SEG0                       0x00013380
+#define UMC_BASE__INST6_SEG1                       0x00194000
+#define UMC_BASE__INST6_SEG2                       0x00427000
+#define UMC_BASE__INST6_SEG3                       0
+#define UMC_BASE__INST6_SEG4                       0
+#define UMC_BASE__INST6_SEG5                       0
+
+#define UMC_BASE__INST7_SEG0                       0x000133A0
+#define UMC_BASE__INST7_SEG1                       0x001D4000
+#define UMC_BASE__INST7_SEG2                       0x00427400
+#define UMC_BASE__INST7_SEG3                       0
+#define UMC_BASE__INST7_SEG4                       0
+#define UMC_BASE__INST7_SEG5                       0
+
+#define UVD_BASE__INST0_SEG0                       0x00007800
+#define UVD_BASE__INST0_SEG1                       0x00007E00
+#define UVD_BASE__INST0_SEG2                       0x00012180
+#define UVD_BASE__INST0_SEG3                       0x00403000
+#define UVD_BASE__INST0_SEG4                       0
+#define UVD_BASE__INST0_SEG5                       0
+
+#define UVD_BASE__INST1_SEG0                       0x00007A00
+#define UVD_BASE__INST1_SEG1                       0x00009000
+#define UVD_BASE__INST1_SEG2                       0x000136E0
+#define UVD_BASE__INST1_SEG3                       0x0042DC00
+#define UVD_BASE__INST1_SEG4                       0
+#define UVD_BASE__INST1_SEG5                       0
+
+#define UVD_BASE__INST2_SEG0                       0
+#define UVD_BASE__INST2_SEG1                       0
+#define UVD_BASE__INST2_SEG2                       0
+#define UVD_BASE__INST2_SEG3                       0
+#define UVD_BASE__INST2_SEG4                       0
+#define UVD_BASE__INST2_SEG5                       0
+
+#define UVD_BASE__INST3_SEG0                       0
+#define UVD_BASE__INST3_SEG1                       0
+#define UVD_BASE__INST3_SEG2                       0
+#define UVD_BASE__INST3_SEG3                       0
+#define UVD_BASE__INST3_SEG4                       0
+#define UVD_BASE__INST3_SEG5                       0
+
+#define UVD_BASE__INST4_SEG0                       0
+#define UVD_BASE__INST4_SEG1                       0
+#define UVD_BASE__INST4_SEG2                       0
+#define UVD_BASE__INST4_SEG3                       0
+#define UVD_BASE__INST4_SEG4                       0
+#define UVD_BASE__INST4_SEG5                       0
+
+#define UVD_BASE__INST5_SEG0                       0
+#define UVD_BASE__INST5_SEG1                       0
+#define UVD_BASE__INST5_SEG2                       0
+#define UVD_BASE__INST5_SEG3                       0
+#define UVD_BASE__INST5_SEG4                       0
+#define UVD_BASE__INST5_SEG5                       0
+
+#define UVD_BASE__INST6_SEG0                       0
+#define UVD_BASE__INST6_SEG1                       0
+#define UVD_BASE__INST6_SEG2                       0
+#define UVD_BASE__INST6_SEG3                       0
+#define UVD_BASE__INST6_SEG4                       0
+#define UVD_BASE__INST6_SEG5                       0
+
+#define UVD_BASE__INST7_SEG0                       0
+#define UVD_BASE__INST7_SEG1                       0
+#define UVD_BASE__INST7_SEG2                       0
+#define UVD_BASE__INST7_SEG3                       0
+#define UVD_BASE__INST7_SEG4                       0
+#define UVD_BASE__INST7_SEG5                       0
+
+#define DBGU_IO_BASE__INST0_SEG0                   0x000001E0
+#define DBGU_IO_BASE__INST0_SEG1                   0x000125A0
+#define DBGU_IO_BASE__INST0_SEG2                   0x0040B400
+#define DBGU_IO_BASE__INST0_SEG3                   0
+#define DBGU_IO_BASE__INST0_SEG4                   0
+#define DBGU_IO_BASE__INST0_SEG5                   0
+
+#define DBGU_IO_BASE__INST1_SEG0                   0
+#define DBGU_IO_BASE__INST1_SEG1                   0
+#define DBGU_IO_BASE__INST1_SEG2                   0
+#define DBGU_IO_BASE__INST1_SEG3                   0
+#define DBGU_IO_BASE__INST1_SEG4                   0
+#define DBGU_IO_BASE__INST1_SEG5                   0
+
+#define DBGU_IO_BASE__INST2_SEG0                   0
+#define DBGU_IO_BASE__INST2_SEG1                   0
+#define DBGU_IO_BASE__INST2_SEG2                   0
+#define DBGU_IO_BASE__INST2_SEG3                   0
+#define DBGU_IO_BASE__INST2_SEG4                   0
+#define DBGU_IO_BASE__INST2_SEG5                   0
+
+#define DBGU_IO_BASE__INST3_SEG0                   0
+#define DBGU_IO_BASE__INST3_SEG1                   0
+#define DBGU_IO_BASE__INST3_SEG2                   0
+#define DBGU_IO_BASE__INST3_SEG3                   0
+#define DBGU_IO_BASE__INST3_SEG4                   0
+#define DBGU_IO_BASE__INST3_SEG5                   0
+
+#define DBGU_IO_BASE__INST4_SEG0                   0
+#define DBGU_IO_BASE__INST4_SEG1                   0
+#define DBGU_IO_BASE__INST4_SEG2                   0
+#define DBGU_IO_BASE__INST4_SEG3                   0
+#define DBGU_IO_BASE__INST4_SEG4                   0
+#define DBGU_IO_BASE__INST4_SEG5                   0
+
+#define DBGU_IO_BASE__INST5_SEG0                   0
+#define DBGU_IO_BASE__INST5_SEG1                   0
+#define DBGU_IO_BASE__INST5_SEG2                   0
+#define DBGU_IO_BASE__INST5_SEG3                   0
+#define DBGU_IO_BASE__INST5_SEG4                   0
+#define DBGU_IO_BASE__INST5_SEG5                   0
+
+#define DBGU_IO_BASE__INST6_SEG0                   0
+#define DBGU_IO_BASE__INST6_SEG1                   0
+#define DBGU_IO_BASE__INST6_SEG2                   0
+#define DBGU_IO_BASE__INST6_SEG3                   0
+#define DBGU_IO_BASE__INST6_SEG4                   0
+#define DBGU_IO_BASE__INST6_SEG5                   0
+
+#define DBGU_IO_BASE__INST7_SEG0                   0
+#define DBGU_IO_BASE__INST7_SEG1                   0
+#define DBGU_IO_BASE__INST7_SEG2                   0
+#define DBGU_IO_BASE__INST7_SEG3                   0
+#define DBGU_IO_BASE__INST7_SEG4                   0
+#define DBGU_IO_BASE__INST7_SEG5                   0
+
+#define RSMU_BASE__INST0_SEG0                   0x00012000
+#define RSMU_BASE__INST0_SEG1                   0
+#define RSMU_BASE__INST0_SEG2                   0
+#define RSMU_BASE__INST0_SEG3                   0
+#define RSMU_BASE__INST0_SEG4                   0
+#define RSMU_BASE__INST0_SEG5                   0
+
+#define RSMU_BASE__INST1_SEG0                   0
+#define RSMU_BASE__INST1_SEG1                   0
+#define RSMU_BASE__INST1_SEG2                   0
+#define RSMU_BASE__INST1_SEG3                   0
+#define RSMU_BASE__INST1_SEG4                   0
+#define RSMU_BASE__INST1_SEG5                   0
+
+#define RSMU_BASE__INST2_SEG0                   0
+#define RSMU_BASE__INST2_SEG1                   0
+#define RSMU_BASE__INST2_SEG2                   0
+#define RSMU_BASE__INST2_SEG3                   0
+#define RSMU_BASE__INST2_SEG4                   0
+#define RSMU_BASE__INST2_SEG5                   0
+
+#define RSMU_BASE__INST3_SEG0                   0
+#define RSMU_BASE__INST3_SEG1                   0
+#define RSMU_BASE__INST3_SEG2                   0
+#define RSMU_BASE__INST3_SEG3                   0
+#define RSMU_BASE__INST3_SEG4                   0
+#define RSMU_BASE__INST3_SEG5                   0
+
+#define RSMU_BASE__INST4_SEG0                   0
+#define RSMU_BASE__INST4_SEG1                   0
+#define RSMU_BASE__INST4_SEG2                   0
+#define RSMU_BASE__INST4_SEG3                   0
+#define RSMU_BASE__INST4_SEG4                   0
+#define RSMU_BASE__INST4_SEG5                   0
+
+#define RSMU_BASE__INST5_SEG0                   0
+#define RSMU_BASE__INST5_SEG1                   0
+#define RSMU_BASE__INST5_SEG2                   0
+#define RSMU_BASE__INST5_SEG3                   0
+#define RSMU_BASE__INST5_SEG4                   0
+#define RSMU_BASE__INST5_SEG5                   0
+
+#define RSMU_BASE__INST6_SEG0                   0
+#define RSMU_BASE__INST6_SEG1                   0
+#define RSMU_BASE__INST6_SEG2                   0
+#define RSMU_BASE__INST6_SEG3                   0
+#define RSMU_BASE__INST6_SEG4                   0
+#define RSMU_BASE__INST6_SEG5                   0
+
+#define RSMU_BASE__INST7_SEG0                   0
+#define RSMU_BASE__INST7_SEG1                   0
+#define RSMU_BASE__INST7_SEG2                   0
+#define RSMU_BASE__INST7_SEG3                   0
+#define RSMU_BASE__INST7_SEG4                   0
+#define RSMU_BASE__INST7_SEG5                   0
+
+
+#endif
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#3 lore raw

PATCH 5/102 drm/amdgpu: add Arcturus asic type

In reply to a message not in the archive

From: Le Ma <[email protected]>

Add asic type for Arcturus.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 1 +
 include/drm/amd_asic_type.h                | 1 +
 2 files changed, 2 insertions(+)
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +1 -0
index 086d2f3b0db3..f8250a776b93 100644
@@ -98,6 +98,7 @@ static const char *amdgpu_asic_name[] = {
 	"VEGA12",
 	"VEGA20",
 	"RAVEN",
+	"ARCTURUS",
 	"NAVI10",
 	"NAVI14",
 	"LAST",
include/drm/amd_asic_type.h +1 -0
index 0c4766af04af..0f5a12a99948 100644
@@ -49,6 +49,7 @@ enum amd_asic_type {
 	CHIP_VEGA12,
 	CHIP_VEGA20,
 	CHIP_RAVEN,
+	CHIP_ARCTURUS,
 	CHIP_NAVI10,
 	CHIP_NAVI14,
 	CHIP_LAST,
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#4 lore raw

PATCH 6/102 drm/amdgpu: add gmc basic support for Arcturus

In reply to a message not in the archive

From: Le Ma <[email protected]>

Add initial GMC support for Arcturus

---
 drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c | 5 +++++
 1 file changed, 5 insertions(+)
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c +5 -0
index bd5d36944481..e78be96008e6 100644
@@ -883,6 +883,7 @@ static int gmc_v9_0_mc_init(struct amdgpu_device *adev)
 		case CHIP_VEGA10:  /* all engines support GPUVM */
 		case CHIP_VEGA12:  /* all engines support GPUVM */
 		case CHIP_VEGA20:
+		case CHIP_ARCTURUS:
 		default:
 			adev->gmc.gart_size = 512ULL << 20;
 			break;
@@ -998,6 +999,10 @@ static int gmc_v9_0_sw_init(void *handle)
 		else
 			amdgpu_vm_adjust_size(adev, 256 * 1024, 9, 3, 48);
 		break;
+	case CHIP_ARCTURUS:
+		/* Keep the vm size same with Vega20 */
+		amdgpu_vm_adjust_size(adev, 256 * 1024, 9, 3, 48);
+		break;
 	default:
 		break;
 	}
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#5 lore raw

PATCH 7/102 drm/amdgpu: rename AMDGPU_GFXHUB/MMHUB macro with hub number

In reply to a message not in the archive

From: Le Ma <[email protected]>

The number of GFXHUB/MMHUB may be expanded in later ASICs.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c   |  4 ++--
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h   |  4 ++--
 drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c   |  8 +++----
 drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c    |  8 +++----
 drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c |  2 +-
 drivers/gpu/drm/amd/amdgpu/gfxhub_v2_0.c |  2 +-
 drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c   | 30 ++++++++++++------------
 drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c    |  6 ++---
 drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c  |  2 +-
 drivers/gpu/drm/amd/amdgpu/mmhub_v2_0.c  |  2 +-
 drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c   |  2 +-
 drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c   |  2 +-
 drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c    |  4 ++--
 drivers/gpu/drm/amd/amdgpu/vce_v4_0.c    |  2 +-
 drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c    |  6 ++---
 drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c    |  6 ++---
 16 files changed, 45 insertions(+), 45 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +2 -2
index 1951f2abbdbc..735e835a2df8 100644
@@ -3060,12 +3060,12 @@ int amdgpu_vm_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 	switch (args->in.op) {
 	case AMDGPU_VM_OP_RESERVE_VMID:
 		/* current, we only have requirement to reserve vmid from gfxhub */
-		r = amdgpu_vmid_alloc_reserved(adev, &fpriv->vm, AMDGPU_GFXHUB);
+		r = amdgpu_vmid_alloc_reserved(adev, &fpriv->vm, AMDGPU_GFXHUB_0);
 		if (r)
 			return r;
 		break;
 	case AMDGPU_VM_OP_UNRESERVE_VMID:
-		amdgpu_vmid_free_reserved(adev, &fpriv->vm, AMDGPU_GFXHUB);
+		amdgpu_vmid_free_reserved(adev, &fpriv->vm, AMDGPU_GFXHUB_0);
 		break;
 	default:
 		return -EINVAL;
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h +2 -2
index 489a162ca620..8e78b81d0a05 100644
@@ -101,8 +101,8 @@ struct amdgpu_bo_list_entry;
 
 /* max number of VMHUB */
 #define AMDGPU_MAX_VMHUBS			2
-#define AMDGPU_GFXHUB				0
-#define AMDGPU_MMHUB				1
+#define AMDGPU_GFXHUB_0				0
+#define AMDGPU_MMHUB_0				1
 
 /* hardcode that limit for now */
 #define AMDGPU_VA_RESERVED_SIZE			(1ULL << 20)
drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c +4 -4
index 7d5207bbe382..a9b7ddaa2d6a 100644
@@ -1601,7 +1601,7 @@ static void gfx_v10_0_constants_init(struct amdgpu_device *adev)
 	/* XXX SH_MEM regs */
 	/* where to put LDS, scratch, GPUVM in FSA64 space */
 	mutex_lock(&adev->srbm_mutex);
-	for (i = 0; i < adev->vm_manager.id_mgr[AMDGPU_GFXHUB].num_ids; i++) {
+	for (i = 0; i < adev->vm_manager.id_mgr[AMDGPU_GFXHUB_0].num_ids; i++) {
 		nv_grbm_select(adev, 0, 0, 0, i);
 		/* CP and shaders */
 		if (i == 0) {
@@ -5012,7 +5012,7 @@ static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_gfx = {
 	.align_mask = 0xff,
 	.nop = PACKET3(PACKET3_NOP, 0x3FFF),
 	.support_64bit_ptrs = true,
-	.vmhub = AMDGPU_GFXHUB,
+	.vmhub = AMDGPU_GFXHUB_0,
 	.get_rptr = gfx_v10_0_ring_get_rptr_gfx,
 	.get_wptr = gfx_v10_0_ring_get_wptr_gfx,
 	.set_wptr = gfx_v10_0_ring_set_wptr_gfx,
@@ -5063,7 +5063,7 @@ static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_compute = {
 	.align_mask = 0xff,
 	.nop = PACKET3(PACKET3_NOP, 0x3FFF),
 	.support_64bit_ptrs = true,
-	.vmhub = AMDGPU_GFXHUB,
+	.vmhub = AMDGPU_GFXHUB_0,
 	.get_rptr = gfx_v10_0_ring_get_rptr_compute,
 	.get_wptr = gfx_v10_0_ring_get_wptr_compute,
 	.set_wptr = gfx_v10_0_ring_set_wptr_compute,
@@ -5096,7 +5096,7 @@ static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_kiq = {
 	.align_mask = 0xff,
 	.nop = PACKET3(PACKET3_NOP, 0x3FFF),
 	.support_64bit_ptrs = true,
-	.vmhub = AMDGPU_GFXHUB,
+	.vmhub = AMDGPU_GFXHUB_0,
 	.get_rptr = gfx_v10_0_ring_get_rptr_compute,
 	.get_wptr = gfx_v10_0_ring_get_wptr_compute,
 	.set_wptr = gfx_v10_0_ring_set_wptr_compute,
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +4 -4
index 8e04e86cfd97..c67bb7fc51c0 100644
@@ -1932,7 +1932,7 @@ static void gfx_v9_0_constants_init(struct amdgpu_device *adev)
 	/* XXX SH_MEM regs */
 	/* where to put LDS, scratch, GPUVM in FSA64 space */
 	mutex_lock(&adev->srbm_mutex);
-	for (i = 0; i < adev->vm_manager.id_mgr[AMDGPU_GFXHUB].num_ids; i++) {
+	for (i = 0; i < adev->vm_manager.id_mgr[AMDGPU_GFXHUB_0].num_ids; i++) {
 		soc15_grbm_select(adev, 0, 0, 0, i);
 		/* CP and shaders */
 		if (i == 0) {
@@ -5170,7 +5170,7 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_gfx = {
 	.align_mask = 0xff,
 	.nop = PACKET3(PACKET3_NOP, 0x3FFF),
 	.support_64bit_ptrs = true,
-	.vmhub = AMDGPU_GFXHUB,
+	.vmhub = AMDGPU_GFXHUB_0,
 	.get_rptr = gfx_v9_0_ring_get_rptr_gfx,
 	.get_wptr = gfx_v9_0_ring_get_wptr_gfx,
 	.set_wptr = gfx_v9_0_ring_set_wptr_gfx,
@@ -5221,7 +5221,7 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_compute = {
 	.align_mask = 0xff,
 	.nop = PACKET3(PACKET3_NOP, 0x3FFF),
 	.support_64bit_ptrs = true,
-	.vmhub = AMDGPU_GFXHUB,
+	.vmhub = AMDGPU_GFXHUB_0,
 	.get_rptr = gfx_v9_0_ring_get_rptr_compute,
 	.get_wptr = gfx_v9_0_ring_get_wptr_compute,
 	.set_wptr = gfx_v9_0_ring_set_wptr_compute,
@@ -5256,7 +5256,7 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_kiq = {
 	.align_mask = 0xff,
 	.nop = PACKET3(PACKET3_NOP, 0x3FFF),
 	.support_64bit_ptrs = true,
-	.vmhub = AMDGPU_GFXHUB,
+	.vmhub = AMDGPU_GFXHUB_0,
 	.get_rptr = gfx_v9_0_ring_get_rptr_compute,
 	.get_wptr = gfx_v9_0_ring_get_wptr_compute,
 	.set_wptr = gfx_v9_0_ring_set_wptr_compute,
drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c +1 -1
index 15986748f59f..6ce37ce77d14 100644
@@ -357,7 +357,7 @@ void gfxhub_v1_0_set_fault_enable_default(struct amdgpu_device *adev,
 
 void gfxhub_v1_0_init(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0];
 
 	hub->ctx0_ptb_addr_lo32 =
 		SOC15_REG_OFFSET(GC, 0,
drivers/gpu/drm/amd/amdgpu/gfxhub_v2_0.c +1 -1
index d605b4963f8a..8ce5bf5feb45 100644
@@ -333,7 +333,7 @@ void gfxhub_v2_0_set_fault_enable_default(struct amdgpu_device *adev,
 
 void gfxhub_v2_0_init(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB_0];
 
 	hub->ctx0_ptb_addr_lo32 =
 		SOC15_REG_OFFSET(GC, 0,
drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c +15 -15
index c377dc22e68c..8cfc3aa32391 100644
@@ -61,7 +61,7 @@ gmc_v10_0_vm_fault_interrupt_state(struct amdgpu_device *adev,
 	struct amdgpu_vmhub *hub;
 	u32 tmp, reg, bits[AMDGPU_MAX_VMHUBS], i;
 
-	bits[AMDGPU_GFXHUB] = GCVM_CONTEXT1_CNTL__RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+	bits[AMDGPU_GFXHUB_0] = GCVM_CONTEXT1_CNTL__RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
 		GCVM_CONTEXT1_CNTL__DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
 		GCVM_CONTEXT1_CNTL__PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
 		GCVM_CONTEXT1_CNTL__VALID_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
@@ -69,7 +69,7 @@ gmc_v10_0_vm_fault_interrupt_state(struct amdgpu_device *adev,
 		GCVM_CONTEXT1_CNTL__WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
 		GCVM_CONTEXT1_CNTL__EXECUTE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK;
 
-	bits[AMDGPU_MMHUB] = MMVM_CONTEXT1_CNTL__RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
+	bits[AMDGPU_MMHUB_0] = MMVM_CONTEXT1_CNTL__RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
 		MMVM_CONTEXT1_CNTL__DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
 		MMVM_CONTEXT1_CNTL__PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
 		MMVM_CONTEXT1_CNTL__VALID_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
@@ -80,39 +80,39 @@ gmc_v10_0_vm_fault_interrupt_state(struct amdgpu_device *adev,
 	switch (state) {
 	case AMDGPU_IRQ_STATE_DISABLE:
 		/* MM HUB */
-		hub = &adev->vmhub[AMDGPU_MMHUB];
+		hub = &adev->vmhub[AMDGPU_MMHUB_0];
 		for (i = 0; i< 16; i++) {
 			reg = hub->vm_context0_cntl + i;
 			tmp = RREG32(reg);
-			tmp &= ~bits[AMDGPU_MMHUB];
+			tmp &= ~bits[AMDGPU_MMHUB_0];
 			WREG32(reg, tmp);
 		}
 
 		/* GFX HUB */
-		hub = &adev->vmhub[AMDGPU_GFXHUB];
+		hub = &adev->vmhub[AMDGPU_GFXHUB_0];
 		for (i = 0; i < 16; i++) {
 			reg = hub->vm_context0_cntl + i;
 			tmp = RREG32(reg);
-			tmp &= ~bits[AMDGPU_GFXHUB];
+			tmp &= ~bits[AMDGPU_GFXHUB_0];
 			WREG32(reg, tmp);
 		}
 		break;
 	case AMDGPU_IRQ_STATE_ENABLE:
 		/* MM HUB */
-		hub = &adev->vmhub[AMDGPU_MMHUB];
+		hub = &adev->vmhub[AMDGPU_MMHUB_0];
 		for (i = 0; i< 16; i++) {
 			reg = hub->vm_context0_cntl + i;
 			tmp = RREG32(reg);
-			tmp |= bits[AMDGPU_MMHUB];
+			tmp |= bits[AMDGPU_MMHUB_0];
 			WREG32(reg, tmp);
 		}
 
 		/* GFX HUB */
-		hub = &adev->vmhub[AMDGPU_GFXHUB];
+		hub = &adev->vmhub[AMDGPU_GFXHUB_0];
 		for (i = 0; i < 16; i++) {
 			reg = hub->vm_context0_cntl + i;
 			tmp = RREG32(reg);
-			tmp |= bits[AMDGPU_GFXHUB];
+			tmp |= bits[AMDGPU_GFXHUB_0];
 			WREG32(reg, tmp);
 		}
 		break;
@@ -243,12 +243,12 @@ static void gmc_v10_0_flush_gpu_tlb(struct amdgpu_device *adev,
 
 	mutex_lock(&adev->mman.gtt_window_lock);
 
-	gmc_v10_0_flush_vm_hub(adev, vmid, AMDGPU_MMHUB, 0);
+	gmc_v10_0_flush_vm_hub(adev, vmid, AMDGPU_MMHUB_0, 0);
 	if (!adev->mman.buffer_funcs_enabled ||
 	    !adev->ib_pool_ready ||
 	    adev->asic_type > CHIP_NAVI14 ||
 	    adev->in_gpu_reset) {
-		gmc_v10_0_flush_vm_hub(adev, vmid, AMDGPU_GFXHUB, 0);
+		gmc_v10_0_flush_vm_hub(adev, vmid, AMDGPU_GFXHUB_0, 0);
 		mutex_unlock(&adev->mman.gtt_window_lock);
 		return;
 	}
@@ -313,7 +313,7 @@ static void gmc_v10_0_emit_pasid_mapping(struct amdgpu_ring *ring, unsigned vmid
 	struct amdgpu_device *adev = ring->adev;
 	uint32_t reg;
 
-	if (ring->funcs->vmhub == AMDGPU_GFXHUB)
+	if (ring->funcs->vmhub == AMDGPU_GFXHUB_0)
 		reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_VMID_0_LUT) + vmid;
 	else
 		reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_VMID_0_LUT_MM) + vmid;
@@ -682,8 +682,8 @@ static int gmc_v10_0_sw_init(void *handle)
 	 * amdgpu graphics/compute will use VMIDs 1-7
 	 * amdkfd will use VMIDs 8-15
 	 */
-	adev->vm_manager.id_mgr[AMDGPU_GFXHUB].num_ids = AMDGPU_NUM_OF_VMIDS;
-	adev->vm_manager.id_mgr[AMDGPU_MMHUB].num_ids = AMDGPU_NUM_OF_VMIDS;
+	adev->vm_manager.id_mgr[AMDGPU_GFXHUB_0].num_ids = AMDGPU_NUM_OF_VMIDS;
+	adev->vm_manager.id_mgr[AMDGPU_MMHUB_0].num_ids = AMDGPU_NUM_OF_VMIDS;
 
 	amdgpu_vm_manager_init(adev);
 
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c +3 -3
index e78be96008e6..af2a4305d6bc 100644
@@ -476,7 +476,7 @@ static void gmc_v9_0_emit_pasid_mapping(struct amdgpu_ring *ring, unsigned vmid,
 	struct amdgpu_device *adev = ring->adev;
 	uint32_t reg;
 
-	if (ring->funcs->vmhub == AMDGPU_GFXHUB)
+	if (ring->funcs->vmhub == AMDGPU_GFXHUB_0)
 		reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_VMID_0_LUT) + vmid;
 	else
 		reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_VMID_0_LUT_MM) + vmid;
@@ -1078,8 +1078,8 @@ static int gmc_v9_0_sw_init(void *handle)
 	 * amdgpu graphics/compute will use VMIDs 1-7
 	 * amdkfd will use VMIDs 8-15
 	 */
-	adev->vm_manager.id_mgr[AMDGPU_GFXHUB].num_ids = AMDGPU_NUM_OF_VMIDS;
-	adev->vm_manager.id_mgr[AMDGPU_MMHUB].num_ids = AMDGPU_NUM_OF_VMIDS;
+	adev->vm_manager.id_mgr[AMDGPU_GFXHUB_0].num_ids = AMDGPU_NUM_OF_VMIDS;
+	adev->vm_manager.id_mgr[AMDGPU_MMHUB_0].num_ids = AMDGPU_NUM_OF_VMIDS;
 
 	amdgpu_vm_manager_init(adev);
 
drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c +1 -1
index dc5ce03034d3..3abd02bd5222 100644
@@ -407,7 +407,7 @@ void mmhub_v1_0_set_fault_enable_default(struct amdgpu_device *adev, bool value)
 
 void mmhub_v1_0_init(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
 
 	hub->ctx0_ptb_addr_lo32 =
 		SOC15_REG_OFFSET(MMHUB, 0,
drivers/gpu/drm/amd/amdgpu/mmhub_v2_0.c +1 -1
index a5c7ed1f37eb..d2f4775299c7 100644
@@ -324,7 +324,7 @@ void mmhub_v2_0_set_fault_enable_default(struct amdgpu_device *adev, bool value)
 
 void mmhub_v2_0_init(struct amdgpu_device *adev)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB];
+	struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_MMHUB_0];
 
 	hub->ctx0_ptb_addr_lo32 =
 		SOC15_REG_OFFSET(MMHUB, 0,
drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c +1 -1
index 9e571ca283c3..8d9efbd88a49 100644
@@ -2130,7 +2130,7 @@ static const struct amdgpu_ring_funcs sdma_v4_0_ring_funcs = {
 	.align_mask = 0xf,
 	.nop = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP),
 	.support_64bit_ptrs = true,
-	.vmhub = AMDGPU_MMHUB,
+	.vmhub = AMDGPU_MMHUB_0,
 	.get_rptr = sdma_v4_0_ring_get_rptr,
 	.get_wptr = sdma_v4_0_ring_get_wptr,
 	.set_wptr = sdma_v4_0_ring_set_wptr,
drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c +1 -1
index 89893261f145..3e536140bfd6 100644
@@ -1554,7 +1554,7 @@ static const struct amdgpu_ring_funcs sdma_v5_0_ring_funcs = {
 	.align_mask = 0xf,
 	.nop = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP),
 	.support_64bit_ptrs = true,
-	.vmhub = AMDGPU_GFXHUB,
+	.vmhub = AMDGPU_GFXHUB_0,
 	.get_rptr = sdma_v5_0_ring_get_rptr,
 	.get_wptr = sdma_v5_0_ring_get_wptr,
 	.set_wptr = sdma_v5_0_ring_set_wptr,
drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c +2 -2
index 74811b2aece1..2f3d4e8032d5 100644
@@ -1763,7 +1763,7 @@ static const struct amdgpu_ring_funcs uvd_v7_0_ring_vm_funcs = {
 	.align_mask = 0xf,
 	.support_64bit_ptrs = false,
 	.no_user_fence = true,
-	.vmhub = AMDGPU_MMHUB,
+	.vmhub = AMDGPU_MMHUB_0,
 	.get_rptr = uvd_v7_0_ring_get_rptr,
 	.get_wptr = uvd_v7_0_ring_get_wptr,
 	.set_wptr = uvd_v7_0_ring_set_wptr,
@@ -1796,7 +1796,7 @@ static const struct amdgpu_ring_funcs uvd_v7_0_enc_ring_vm_funcs = {
 	.nop = HEVC_ENC_CMD_NO_OP,
 	.support_64bit_ptrs = false,
 	.no_user_fence = true,
-	.vmhub = AMDGPU_MMHUB,
+	.vmhub = AMDGPU_MMHUB_0,
 	.get_rptr = uvd_v7_0_enc_ring_get_rptr,
 	.get_wptr = uvd_v7_0_enc_ring_get_wptr,
 	.set_wptr = uvd_v7_0_enc_ring_set_wptr,
drivers/gpu/drm/amd/amdgpu/vce_v4_0.c +1 -1
index e267b073f525..93d80ef17685 100644
@@ -1070,7 +1070,7 @@ static const struct amdgpu_ring_funcs vce_v4_0_ring_vm_funcs = {
 	.nop = VCE_CMD_NO_OP,
 	.support_64bit_ptrs = false,
 	.no_user_fence = true,
-	.vmhub = AMDGPU_MMHUB,
+	.vmhub = AMDGPU_MMHUB_0,
 	.get_rptr = vce_v4_0_ring_get_rptr,
 	.get_wptr = vce_v4_0_ring_get_wptr,
 	.set_wptr = vce_v4_0_ring_set_wptr,
drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c +3 -3
index 855b1f9609e4..09dc9c87ebd1 100644
@@ -2198,7 +2198,7 @@ static const struct amdgpu_ring_funcs vcn_v1_0_dec_ring_vm_funcs = {
 	.align_mask = 0xf,
 	.support_64bit_ptrs = false,
 	.no_user_fence = true,
-	.vmhub = AMDGPU_MMHUB,
+	.vmhub = AMDGPU_MMHUB_0,
 	.get_rptr = vcn_v1_0_dec_ring_get_rptr,
 	.get_wptr = vcn_v1_0_dec_ring_get_wptr,
 	.set_wptr = vcn_v1_0_dec_ring_set_wptr,
@@ -2232,7 +2232,7 @@ static const struct amdgpu_ring_funcs vcn_v1_0_enc_ring_vm_funcs = {
 	.nop = VCN_ENC_CMD_NO_OP,
 	.support_64bit_ptrs = false,
 	.no_user_fence = true,
-	.vmhub = AMDGPU_MMHUB,
+	.vmhub = AMDGPU_MMHUB_0,
 	.get_rptr = vcn_v1_0_enc_ring_get_rptr,
 	.get_wptr = vcn_v1_0_enc_ring_get_wptr,
 	.set_wptr = vcn_v1_0_enc_ring_set_wptr,
@@ -2264,7 +2264,7 @@ static const struct amdgpu_ring_funcs vcn_v1_0_jpeg_ring_vm_funcs = {
 	.nop = PACKET0(0x81ff, 0),
 	.support_64bit_ptrs = false,
 	.no_user_fence = true,
-	.vmhub = AMDGPU_MMHUB,
+	.vmhub = AMDGPU_MMHUB_0,
 	.extra_dw = 64,
 	.get_rptr = vcn_v1_0_jpeg_ring_get_rptr,
 	.get_wptr = vcn_v1_0_jpeg_ring_get_wptr,
drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c +3 -3
index 70c071d52149..84e6ee48670c 100644
@@ -2131,7 +2131,7 @@ static const struct amd_ip_funcs vcn_v2_0_ip_funcs = {
 static const struct amdgpu_ring_funcs vcn_v2_0_dec_ring_vm_funcs = {
 	.type = AMDGPU_RING_TYPE_VCN_DEC,
 	.align_mask = 0xf,
-	.vmhub = AMDGPU_MMHUB,
+	.vmhub = AMDGPU_MMHUB_0,
 	.get_rptr = vcn_v2_0_dec_ring_get_rptr,
 	.get_wptr = vcn_v2_0_dec_ring_get_wptr,
 	.set_wptr = vcn_v2_0_dec_ring_set_wptr,
@@ -2162,7 +2162,7 @@ static const struct amdgpu_ring_funcs vcn_v2_0_enc_ring_vm_funcs = {
 	.type = AMDGPU_RING_TYPE_VCN_ENC,
 	.align_mask = 0x3f,
 	.nop = VCN_ENC_CMD_NO_OP,
-	.vmhub = AMDGPU_MMHUB,
+	.vmhub = AMDGPU_MMHUB_0,
 	.get_rptr = vcn_v2_0_enc_ring_get_rptr,
 	.get_wptr = vcn_v2_0_enc_ring_get_wptr,
 	.set_wptr = vcn_v2_0_enc_ring_set_wptr,
@@ -2191,7 +2191,7 @@ static const struct amdgpu_ring_funcs vcn_v2_0_enc_ring_vm_funcs = {
 static const struct amdgpu_ring_funcs vcn_v2_0_jpeg_ring_vm_funcs = {
 	.type = AMDGPU_RING_TYPE_VCN_JPEG,
 	.align_mask = 0xf,
-	.vmhub = AMDGPU_MMHUB,
+	.vmhub = AMDGPU_MMHUB_0,
 	.get_rptr = vcn_v2_0_jpeg_ring_get_rptr,
 	.get_wptr = vcn_v2_0_jpeg_ring_get_wptr,
 	.set_wptr = vcn_v2_0_jpeg_ring_set_wptr,
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#6 lore raw

PATCH 8/102 drm/amdgpu: add new member in amdgpu_device for vmhub counts per asic chip

In reply to a message not in the archive

From: Le Ma <[email protected]>

It aims to replace AMDGPU_MAX_VMHUBS in for loop to initialize registers.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h    |  1 +
 drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c |  1 +
 drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c  | 10 +++++++---
 3 files changed, 9 insertions(+), 3 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu.h +1 -0
index 2cdf46bbeddf..aecb4f7f9c43 100644
@@ -839,6 +839,7 @@ struct amdgpu_device {
 	dma_addr_t			dummy_page_addr;
 	struct amdgpu_vm_manager	vm_manager;
 	struct amdgpu_vmhub             vmhub[AMDGPU_MAX_VMHUBS];
+	unsigned			num_vmhubs;
 
 	/* memory management */
 	struct amdgpu_mman		mman;
drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c +1 -0
index 8cfc3aa32391..a0bd14e9b8fe 100644
@@ -603,6 +603,7 @@ static int gmc_v10_0_sw_init(void *handle)
 	switch (adev->asic_type) {
 	case CHIP_NAVI10:
 	case CHIP_NAVI14:
+		adev->num_vmhubs = 2;
 		/*
 		 * To fulfill 4-level page support,
 		 * vm size is 256TB (48bit), maximum size of Navi10/Navi14,
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c +7 -3
index af2a4305d6bc..6d4212406c1c 100644
@@ -280,7 +280,7 @@ static int gmc_v9_0_vm_fault_interrupt_state(struct amdgpu_device *adev,
 
 	switch (state) {
 	case AMDGPU_IRQ_STATE_DISABLE:
-		for (j = 0; j < AMDGPU_MAX_VMHUBS; j++) {
+		for (j = 0; j < adev->num_vmhubs; j++) {
 			hub = &adev->vmhub[j];
 			for (i = 0; i < 16; i++) {
 				reg = hub->vm_context0_cntl + i;
@@ -291,7 +291,7 @@ static int gmc_v9_0_vm_fault_interrupt_state(struct amdgpu_device *adev,
 		}
 		break;
 	case AMDGPU_IRQ_STATE_ENABLE:
-		for (j = 0; j < AMDGPU_MAX_VMHUBS; j++) {
+		for (j = 0; j < adev->num_vmhubs; j++) {
 			hub = &adev->vmhub[j];
 			for (i = 0; i < 16; i++) {
 				reg = hub->vm_context0_cntl + i;
@@ -415,7 +415,7 @@ static void gmc_v9_0_flush_gpu_tlb(struct amdgpu_device *adev,
 	const unsigned eng = 17;
 	unsigned i, j;
 
-	for (i = 0; i < AMDGPU_MAX_VMHUBS; ++i) {
+	for (i = 0; i < adev->num_vmhubs; ++i) {
 		struct amdgpu_vmhub *hub = &adev->vmhub[i];
 		u32 tmp = gmc_v9_0_get_invalidate_req(vmid, flush_type);
 
@@ -976,6 +976,8 @@ static int gmc_v9_0_sw_init(void *handle)
 	adev->gmc.vram_type = amdgpu_atomfirmware_get_vram_type(adev);
 	switch (adev->asic_type) {
 	case CHIP_RAVEN:
+		adev->num_vmhubs = 2;
+
 		if (adev->rev_id == 0x0 || adev->rev_id == 0x1) {
 			amdgpu_vm_adjust_size(adev, 256 * 1024, 9, 3, 48);
 		} else {
@@ -988,6 +990,8 @@ static int gmc_v9_0_sw_init(void *handle)
 	case CHIP_VEGA10:
 	case CHIP_VEGA12:
 	case CHIP_VEGA20:
+		adev->num_vmhubs = 2;
+
 		/*
 		 * To fulfill 4-level page support,
 		 * vm size is 256TB (48bit), maximum size of Vega10,
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#7 lore raw

PATCH 9/102 drm/amdgpu: add one more mmhub instance for Arcturus (v2)

In reply to a message not in the archive

From: Le Ma <[email protected]>

v2: set mmhub num under CHIP_ARCTURUS switch case and add one more mmhub id_mgr

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h | 3 ++-
 drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c  | 6 +++++-
 2 files changed, 7 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h +2 -1
index 8e78b81d0a05..42daf8877c3a 100644
@@ -100,9 +100,10 @@ struct amdgpu_bo_list_entry;
 #define AMDGPU_VM_FAULT_STOP_ALWAYS	2
 
 /* max number of VMHUB */
-#define AMDGPU_MAX_VMHUBS			2
+#define AMDGPU_MAX_VMHUBS			3
 #define AMDGPU_GFXHUB_0				0
 #define AMDGPU_MMHUB_0				1
+#define AMDGPU_MMHUB_1				2
 
 /* hardcode that limit for now */
 #define AMDGPU_VA_RESERVED_SIZE			(1ULL << 20)
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c +5 -1
index 6d4212406c1c..fd22eaba80e2 100644
@@ -637,7 +637,8 @@ static int gmc_v9_0_allocate_vm_inv_eng(struct amdgpu_device *adev)
 {
 	struct amdgpu_ring *ring;
 	unsigned vm_inv_engs[AMDGPU_MAX_VMHUBS] =
-		{GFXHUB_FREE_VM_INV_ENGS_BITMAP, MMHUB_FREE_VM_INV_ENGS_BITMAP};
+		{GFXHUB_FREE_VM_INV_ENGS_BITMAP, MMHUB_FREE_VM_INV_ENGS_BITMAP,
+		GFXHUB_FREE_VM_INV_ENGS_BITMAP};
 	unsigned i;
 	unsigned vmhub, inv_eng;
 
@@ -1004,6 +1005,8 @@ static int gmc_v9_0_sw_init(void *handle)
 			amdgpu_vm_adjust_size(adev, 256 * 1024, 9, 3, 48);
 		break;
 	case CHIP_ARCTURUS:
+		adev->num_vmhubs = 3;
+
 		/* Keep the vm size same with Vega20 */
 		amdgpu_vm_adjust_size(adev, 256 * 1024, 9, 3, 48);
 		break;
@@ -1084,6 +1087,7 @@ static int gmc_v9_0_sw_init(void *handle)
 	 */
 	adev->vm_manager.id_mgr[AMDGPU_GFXHUB_0].num_ids = AMDGPU_NUM_OF_VMIDS;
 	adev->vm_manager.id_mgr[AMDGPU_MMHUB_0].num_ids = AMDGPU_NUM_OF_VMIDS;
+	adev->vm_manager.id_mgr[AMDGPU_MMHUB_1].num_ids = AMDGPU_NUM_OF_VMIDS;
 
 	amdgpu_vm_manager_init(adev);
 
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#8 lore raw

PATCH 10/102 drm/amdgpu: add mmhub v9.4.1 block for Arcturus (v2)

In reply to a message not in the archive

From: Le Ma <[email protected]>

Arcturus as an updated mmhub block. mmhub is the memory controller hub used for sdma and multimedia.

v2: squash in AGP BAR programming (Alex)

---
 drivers/gpu/drm/amd/amdgpu/Makefile     |   2 +-
 drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c | 504 ++++++++++++++++++++++++
 drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.h |  33 ++
 3 files changed, 538 insertions(+), 1 deletion(-)
 create mode 100644 drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c
 create mode 100644 drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.h
drivers/gpu/drm/amd/amdgpu/Makefile +1 -1
index 3f5329906fce..7b6fb0e9ee76 100644
@@ -77,7 +77,7 @@ amdgpu-y += \
 amdgpu-y += \
 	gmc_v7_0.o \
 	gmc_v8_0.o \
-	gfxhub_v1_0.o mmhub_v1_0.o gmc_v9_0.o gfxhub_v1_1.o \
+	gfxhub_v1_0.o mmhub_v1_0.o gmc_v9_0.o gfxhub_v1_1.o mmhub_v9_4.o \
 	gfxhub_v2_0.o mmhub_v2_0.o gmc_v10_0.o
 
 # add IH block
drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c new file +504 -0

504 lines not on this page · raw message

drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.h new file +33 -0
new file mode 100644
index 000000000000..9ba3dd808826
@@ -0,0 +1,33 @@
+/*
+ * Copyright 2018 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __MMHUB_V9_4_H__
+#define __MMHUB_V9_4_H__
+
+u64 mmhub_v9_4_get_fb_location(struct amdgpu_device *adev);
+int mmhub_v9_4_gart_enable(struct amdgpu_device *adev);
+void mmhub_v9_4_gart_disable(struct amdgpu_device *adev);
+void mmhub_v9_4_set_fault_enable_default(struct amdgpu_device *adev,
+					 bool value);
+void mmhub_v9_4_init(struct amdgpu_device *adev);
+
+#endif
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#9 lore raw

PATCH 11/102 drm/amdgpu: use new mmhub interfaces for Arcturus

In reply to a message not in the archive

From: Le Ma <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c | 35 ++++++++++++++++++++++-----
 1 file changed, 29 insertions(+), 6 deletions(-)
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c +29 -6
index fd22eaba80e2..a0fdfcb5d334 100644
@@ -44,6 +44,7 @@
 #include "gfxhub_v1_0.h"
 #include "mmhub_v1_0.h"
 #include "gfxhub_v1_1.h"
+#include "mmhub_v9_4.h"
 
 #include "ivsrcid/vmc/irqsrcs_vmc_1_0.h"
 
@@ -803,8 +804,12 @@ static void gmc_v9_0_vram_gtt_location(struct amdgpu_device *adev,
 					struct amdgpu_gmc *mc)
 {
 	u64 base = 0;
-	if (!amdgpu_sriov_vf(adev))
-		base = mmhub_v1_0_get_fb_location(adev);
+	if (!amdgpu_sriov_vf(adev)) {
+		if (adev->asic_type == CHIP_ARCTURUS)
+			base = mmhub_v9_4_get_fb_location(adev);
+		else
+			base = mmhub_v1_0_get_fb_location(adev);
+	}
 	/* add the xgmi offset of the physical node */
 	base += adev->gmc.xgmi.physical_node_id * adev->gmc.xgmi.node_segment_size;
 	amdgpu_gmc_vram_location(adev, mc, base);
@@ -970,7 +975,10 @@ static int gmc_v9_0_sw_init(void *handle)
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
 	gfxhub_v1_0_init(adev);
-	mmhub_v1_0_init(adev);
+	if (adev->asic_type == CHIP_ARCTURUS)
+		mmhub_v9_4_init(adev);
+	else
+		mmhub_v1_0_init(adev);
 
 	spin_lock_init(&adev->gmc.invalidate_lock);
 
@@ -1190,7 +1198,10 @@ static int gmc_v9_0_gart_enable(struct amdgpu_device *adev)
 	if (r)
 		return r;
 
-	r = mmhub_v1_0_gart_enable(adev);
+	if (adev->asic_type == CHIP_ARCTURUS)
+		r = mmhub_v9_4_gart_enable(adev);
+	else
+		r = mmhub_v1_0_gart_enable(adev);
 	if (r)
 		return r;
 
@@ -1211,7 +1222,10 @@ static int gmc_v9_0_gart_enable(struct amdgpu_device *adev)
 		value = true;
 
 	gfxhub_v1_0_set_fault_enable_default(adev, value);
-	mmhub_v1_0_set_fault_enable_default(adev, value);
+	if (adev->asic_type == CHIP_ARCTURUS)
+		mmhub_v9_4_set_fault_enable_default(adev, value);
+	else
+		mmhub_v1_0_set_fault_enable_default(adev, value);
 	gmc_v9_0_flush_gpu_tlb(adev, 0, 0);
 
 	DRM_INFO("PCIE GART of %uM enabled (table at 0x%016llX).\n",
@@ -1252,7 +1266,10 @@ static int gmc_v9_0_hw_init(void *handle)
 static void gmc_v9_0_gart_disable(struct amdgpu_device *adev)
 {
 	gfxhub_v1_0_gart_disable(adev);
-	mmhub_v1_0_gart_disable(adev);
+	if (adev->asic_type == CHIP_ARCTURUS)
+		mmhub_v9_4_gart_disable(adev);
+	else
+		mmhub_v1_0_gart_disable(adev);
 	amdgpu_gart_table_vram_unpin(adev);
 }
 
@@ -1317,6 +1334,9 @@ static int gmc_v9_0_set_clockgating_state(void *handle,
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
+	if (adev->asic_type == CHIP_ARCTURUS)
+		return 0;
+
 	return mmhub_v1_0_set_clockgating(adev, state);
 }
 
@@ -1324,6 +1344,9 @@ static void gmc_v9_0_get_clockgating_state(void *handle, u32 *flags)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
+	if (adev->asic_type == CHIP_ARCTURUS)
+		return;
+
 	mmhub_v1_0_get_clockgating(adev, flags);
 }
 
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#10 lore raw

PATCH 12/102 drm/amdgpu: add SDMA 2~7 interrupt client id for Arcturus

In reply to a message not in the archive

From: Le Ma <[email protected]>

---
 drivers/gpu/drm/amd/include/soc15_ih_clientid.h | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/include/soc15_ih_clientid.h +7 -1
index 12e196c15bbe..5b40b9d1b788 100644
@@ -63,7 +63,13 @@ enum soc15_ih_clientid {
 
 	SOC15_IH_CLIENTID_MAX,
 
-	SOC15_IH_CLIENTID_VCN		= SOC15_IH_CLIENTID_UVD
+	SOC15_IH_CLIENTID_VCN		= SOC15_IH_CLIENTID_UVD,
+	SOC15_IH_CLIENTID_SDMA2		= SOC15_IH_CLIENTID_ACP,
+	SOC15_IH_CLIENTID_SDMA3		= SOC15_IH_CLIENTID_DCE,
+	SOC15_IH_CLIENTID_SDMA4		= SOC15_IH_CLIENTID_ISP,
+	SOC15_IH_CLIENTID_SDMA5		= SOC15_IH_CLIENTID_VCE0,
+	SOC15_IH_CLIENTID_SDMA6		= SOC15_IH_CLIENTID_XDMA,
+	SOC15_IH_CLIENTID_SDMA7		= SOC15_IH_CLIENTID_VCE1,
 };
 
 #endif
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#11 lore raw

PATCH 13/102 drm/amdgpu: add SDMA 2~7 ip block type

In reply to a message not in the archive

From: Le Ma <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h | 6 ++++++
 1 file changed, 6 insertions(+)
drivers/gpu/drm/amd/amdgpu/amdgpu.h +6 -0
index aecb4f7f9c43..d881657fb874 100644
@@ -715,6 +715,12 @@ enum amd_hw_ip_block_type {
 	HDP_HWIP,
 	SDMA0_HWIP,
 	SDMA1_HWIP,
+	SDMA2_HWIP,
+	SDMA3_HWIP,
+	SDMA4_HWIP,
+	SDMA5_HWIP,
+	SDMA6_HWIP,
+	SDMA7_HWIP,
 	MMHUB_HWIP,
 	ATHUB_HWIP,
 	NBIO_HWIP,
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#12 lore raw

PATCH 14/102 drm/amdgpu: increase max number of ip base instances to 8

In reply to a message not in the archive

From: Le Ma <[email protected]>

For Arcturus, the number of IP base instances is 8.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/amdgpu.h +1 -1
index d881657fb874..8575ed6713c3 100644
@@ -740,7 +740,7 @@ enum amd_hw_ip_block_type {
 	MAX_HWIP
 };
 
-#define HWIP_MAX_INSTANCE	7
+#define HWIP_MAX_INSTANCE	8
 
 struct amd_powerplay {
 	void *pp_handle;
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#13 lore raw

PATCH 15/102 drm/amdgpu: dynamically initialize IP offset for Arcturus

In reply to a message not in the archive

From: Le Ma <[email protected]>

Add support for the IP offsets on Arcturus.

---
 drivers/gpu/drm/amd/amdgpu/Makefile        |  3 +-
 drivers/gpu/drm/amd/amdgpu/arct_reg_init.c | 57 ++++++++++++++++++++++
 drivers/gpu/drm/amd/amdgpu/soc15.c         |  3 ++
 drivers/gpu/drm/amd/amdgpu/soc15.h         |  1 +
 4 files changed, 63 insertions(+), 1 deletion(-)
 create mode 100644 drivers/gpu/drm/amd/amdgpu/arct_reg_init.c
drivers/gpu/drm/amd/amdgpu/Makefile +2 -1
index 7b6fb0e9ee76..97671df6a5f0 100644
@@ -66,7 +66,8 @@ amdgpu-$(CONFIG_DRM_AMDGPU_SI)+= si.o gmc_v6_0.o gfx_v6_0.o si_ih.o si_dma.o dce
 
 amdgpu-y += \
 	vi.o mxgpu_vi.o nbio_v6_1.o soc15.o emu_soc.o mxgpu_ai.o nbio_v7_0.o vega10_reg_init.o \
-	vega20_reg_init.o nbio_v7_4.o nbio_v2_3.o nv.o navi10_reg_init.o navi14_reg_init.o
+	vega20_reg_init.o nbio_v7_4.o nbio_v2_3.o nv.o navi10_reg_init.o navi14_reg_init.o \
+	arct_reg_init.o
 
 # add DF block
 amdgpu-y += \
drivers/gpu/drm/amd/amdgpu/arct_reg_init.c new file +57 -0

57 lines not on this page · raw message

drivers/gpu/drm/amd/amdgpu/soc15.c +3 -0
index 3b5223fca9a0..800dd76dcf92 100644
@@ -585,6 +585,9 @@ int soc15_set_ip_blocks(struct amdgpu_device *adev)
 	case CHIP_VEGA20:
 		vega20_reg_base_init(adev);
 		break;
+	case CHIP_ARCTURUS:
+		arct_reg_base_init(adev);
+		break;
 	default:
 		return -EINVAL;
 	}
drivers/gpu/drm/amd/amdgpu/soc15.h +1 -0
index 7a6b2cc6d9f5..a3dde0c31f57 100644
@@ -77,6 +77,7 @@ void soc15_program_register_sequence(struct amdgpu_device *adev,
 
 int vega10_reg_base_init(struct amdgpu_device *adev);
 int vega20_reg_base_init(struct amdgpu_device *adev);
+int arct_reg_base_init(struct amdgpu_device *adev);
 
 void vega10_doorbell_index_init(struct amdgpu_device *adev);
 void vega20_doorbell_index_init(struct amdgpu_device *adev);
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#14 lore raw

PATCH 16/102 drm/amdgpu: add VMC1 interrupt client id for Arcturus

In reply to a message not in the archive

From: Le Ma <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c           | 7 +++++++
 drivers/gpu/drm/amd/include/soc15_ih_clientid.h | 1 +
 2 files changed, 8 insertions(+)
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c +7 -0
index a0fdfcb5d334..697319b77d74 100644
@@ -1028,6 +1028,13 @@ static int gmc_v9_0_sw_init(void *handle)
 	if (r)
 		return r;
 
+	if (adev->asic_type == CHIP_ARCTURUS) {
+		r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VMC1, VMC_1_0__SRCID__VM_FAULT,
+					&adev->gmc.vm_fault);
+		if (r)
+			return r;
+	}
+
 	r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_UTCL2, UTCL2_1_0__SRCID__FAULT,
 				&adev->gmc.vm_fault);
 
drivers/gpu/drm/amd/include/soc15_ih_clientid.h +1 -0
index 5b40b9d1b788..0f386b2e1f4f 100644
@@ -70,6 +70,7 @@ enum soc15_ih_clientid {
 	SOC15_IH_CLIENTID_SDMA5		= SOC15_IH_CLIENTID_VCE0,
 	SOC15_IH_CLIENTID_SDMA6		= SOC15_IH_CLIENTID_XDMA,
 	SOC15_IH_CLIENTID_SDMA7		= SOC15_IH_CLIENTID_VCE1,
+	SOC15_IH_CLIENTID_VMC1		= SOC15_IH_CLIENTID_PCIE0,
 };
 
 #endif
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#15 lore raw

PATCH 17/102 drm/amdgpu: update vmc interrupt routine to support 3 vmhubs

In reply to a message not in the archive

From: Le Ma <[email protected]>

There is one more vmc interrupt and mmhub on Arcturus.

---
 drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c | 17 ++++++++++++++---
 1 file changed, 14 insertions(+), 3 deletions(-)
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c +14 -3
index 697319b77d74..d6aa290d567c 100644
@@ -312,10 +312,11 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev,
 				struct amdgpu_irq_src *source,
 				struct amdgpu_iv_entry *entry)
 {
-	struct amdgpu_vmhub *hub = &adev->vmhub[entry->vmid_src];
+	struct amdgpu_vmhub *hub;
 	bool retry_fault = !!(entry->src_data[1] & 0x80);
 	uint32_t status = 0;
 	u64 addr;
+	char hub_name[10];
 
 	addr = (u64)entry->src_data[0] << 12;
 	addr |= ((u64)entry->src_data[1] & 0xf) << 44;
@@ -324,6 +325,17 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev,
 						    entry->timestamp))
 		return 1; /* This also prevents sending it to KFD */
 
+	if (entry->client_id == SOC15_IH_CLIENTID_VMC) {
+		snprintf(hub_name, sizeof(hub_name), "mmhub0");
+		hub = &adev->vmhub[AMDGPU_MMHUB_0];
+	} else if (entry->client_id == SOC15_IH_CLIENTID_VMC1) {
+		snprintf(hub_name, sizeof(hub_name), "mmhub1");
+		hub = &adev->vmhub[AMDGPU_MMHUB_1];
+	} else {
+		snprintf(hub_name, sizeof(hub_name), "gfxhub0");
+		hub = &adev->vmhub[AMDGPU_GFXHUB_0];
+	}
+
 	/* If it's the first fault for this address, process it normally */
 	if (!amdgpu_sriov_vf(adev)) {
 		status = RREG32(hub->vm_l2_pro_fault_status);
@@ -339,8 +351,7 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev,
 		dev_err(adev->dev,
 			"[%s] %s page fault (src_id:%u ring:%u vmid:%u "
 			"pasid:%u, for process %s pid %d thread %s pid %d)\n",
-			entry->vmid_src ? "mmhub" : "gfxhub",
-			retry_fault ? "retry" : "no-retry",
+			hub_name, retry_fault ? "retry" : "no-retry",
 			entry->src_id, entry->ring_id, entry->vmid,
 			entry->pasid, task_info.process_name, task_info.tgid,
 			task_info.task_name, task_info.pid);
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#16 lore raw

PATCH 18/102 drm/amdgpu: reorganize sdma v4 code to support more instances

In reply to a message not in the archive

From: Le Ma <[email protected]>

This change is needed for Arcturus which has 8 sdma instances. The CG/PG part is not covered for now.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h |   8 +-
 drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c   | 144 ++++++++++++++++-------
 2 files changed, 107 insertions(+), 45 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h +7 -1
index 35dd152f9d5c..a9ae0d8a0589 100644
@@ -25,11 +25,17 @@
 #define __AMDGPU_SDMA_H__
 
 /* max number of IP instances */
-#define AMDGPU_MAX_SDMA_INSTANCES		2
+#define AMDGPU_MAX_SDMA_INSTANCES		8
 
 enum amdgpu_sdma_irq {
 	AMDGPU_SDMA_IRQ_INSTANCE0  = 0,
 	AMDGPU_SDMA_IRQ_INSTANCE1,
+	AMDGPU_SDMA_IRQ_INSTANCE2,
+	AMDGPU_SDMA_IRQ_INSTANCE3,
+	AMDGPU_SDMA_IRQ_INSTANCE4,
+	AMDGPU_SDMA_IRQ_INSTANCE5,
+	AMDGPU_SDMA_IRQ_INSTANCE6,
+	AMDGPU_SDMA_IRQ_INSTANCE7,
 	AMDGPU_SDMA_IRQ_LAST
 };
 
drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c +100 -44

200 lines not on this page · raw message

Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#17 lore raw

PATCH 19/102 drm/amdgpu: specify sdma instance 5~7 with second mmhub type

In reply to a message not in the archive

From: Le Ma <[email protected]>

On Arcturus, sdma instance 5~7 is connected to the second mmhub. The vmhub type in amdgpu_ring_funcs is constant, so we create an individual amdgpu_ring_funcs with different vmhub type(AMDGPU_MMHUB_1) for these sdma instances.

---
 drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c | 43 +++++++++++++++++++++++++-
 1 file changed, 42 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c +42 -1
index b6509525f4fc..1dce974e0c58 100644
@@ -2213,6 +2213,42 @@ static const struct amdgpu_ring_funcs sdma_v4_0_ring_funcs = {
 	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
 };
 
+/*
+ * On Arcturus, SDMA instance 5~7 has a different vmhub type(AMDGPU_MMHUB_1).
+ * So create a individual constant ring_funcs for those instances.
+ */
+static const struct amdgpu_ring_funcs sdma_v4_0_ring_funcs_2nd_mmhub = {
+	.type = AMDGPU_RING_TYPE_SDMA,
+	.align_mask = 0xf,
+	.nop = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP),
+	.support_64bit_ptrs = true,
+	.vmhub = AMDGPU_MMHUB_1,
+	.get_rptr = sdma_v4_0_ring_get_rptr,
+	.get_wptr = sdma_v4_0_ring_get_wptr,
+	.set_wptr = sdma_v4_0_ring_set_wptr,
+	.emit_frame_size =
+		6 + /* sdma_v4_0_ring_emit_hdp_flush */
+		3 + /* hdp invalidate */
+		6 + /* sdma_v4_0_ring_emit_pipeline_sync */
+		/* sdma_v4_0_ring_emit_vm_flush */
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 6 +
+		10 + 10 + 10, /* sdma_v4_0_ring_emit_fence x3 for user fence, vm fence */
+	.emit_ib_size = 7 + 6, /* sdma_v4_0_ring_emit_ib */
+	.emit_ib = sdma_v4_0_ring_emit_ib,
+	.emit_fence = sdma_v4_0_ring_emit_fence,
+	.emit_pipeline_sync = sdma_v4_0_ring_emit_pipeline_sync,
+	.emit_vm_flush = sdma_v4_0_ring_emit_vm_flush,
+	.emit_hdp_flush = sdma_v4_0_ring_emit_hdp_flush,
+	.test_ring = sdma_v4_0_ring_test_ring,
+	.test_ib = sdma_v4_0_ring_test_ib,
+	.insert_nop = sdma_v4_0_ring_insert_nop,
+	.pad_ib = sdma_v4_0_ring_pad_ib,
+	.emit_wreg = sdma_v4_0_ring_emit_wreg,
+	.emit_reg_wait = sdma_v4_0_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+};
+
 static const struct amdgpu_ring_funcs sdma_v4_0_page_ring_funcs = {
 	.type = AMDGPU_RING_TYPE_SDMA,
 	.align_mask = 0xf,
@@ -2250,7 +2286,12 @@ static void sdma_v4_0_set_ring_funcs(struct amdgpu_device *adev)
 	int i;
 
 	for (i = 0; i < adev->sdma.num_instances; i++) {
-		adev->sdma.instance[i].ring.funcs = &sdma_v4_0_ring_funcs;
+		if (adev->asic_type == CHIP_ARCTURUS && i >=5)
+			adev->sdma.instance[i].ring.funcs =
+					&sdma_v4_0_ring_funcs_2nd_mmhub;
+		else
+			adev->sdma.instance[i].ring.funcs =
+					&sdma_v4_0_ring_funcs;
 		adev->sdma.instance[i].ring.me = i;
 		if (adev->sdma.has_page_queue) {
 			adev->sdma.instance[i].page.funcs = &sdma_v4_0_page_ring_funcs;
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#18 lore raw

PATCH 20/102 drm/amdgpu: support hdp flush for more sdma instances

In reply to a message not in the archive

From: Le Ma <[email protected]>

The bit RSVD_ENG0 to RSVD_ENG5 in GPU_HDP_FLUSH_REQ/GPU_HDP_FLUSH_DONE can be leveraged for sdma instance 2~7 to poll register/memory.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h    |  6 ++++++
 drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c | 17 +++++++++++++++++
 drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c |  5 +----
 3 files changed, 24 insertions(+), 4 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu.h +6 -0
index 8575ed6713c3..e9c386cec2ad 100644
@@ -651,6 +651,12 @@ struct nbio_hdp_flush_reg {
 	u32 ref_and_mask_cp9;
 	u32 ref_and_mask_sdma0;
 	u32 ref_and_mask_sdma1;
+	u32 ref_and_mask_sdma2;
+	u32 ref_and_mask_sdma3;
+	u32 ref_and_mask_sdma4;
+	u32 ref_and_mask_sdma5;
+	u32 ref_and_mask_sdma6;
+	u32 ref_and_mask_sdma7;
 };
 
 struct amdgpu_mmio_remap {
drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c +17 -0
index bfaaa327ae3c..2e1098c5c32b 100644
@@ -31,6 +31,17 @@
 
 #define smnNBIF_MGCG_CTRL_LCLK	0x1013a21c
 
+/*
+ * These are nbio v7_4_1 registers mask. Temporarily define these here since
+ * nbio v7_4_1 header is incomplete.
+ */
+#define GPU_HDP_FLUSH_DONE__RSVD_ENG0_MASK	0x00001000L
+#define GPU_HDP_FLUSH_DONE__RSVD_ENG1_MASK	0x00002000L
+#define GPU_HDP_FLUSH_DONE__RSVD_ENG2_MASK	0x00004000L
+#define GPU_HDP_FLUSH_DONE__RSVD_ENG3_MASK	0x00008000L
+#define GPU_HDP_FLUSH_DONE__RSVD_ENG4_MASK	0x00010000L
+#define GPU_HDP_FLUSH_DONE__RSVD_ENG5_MASK	0x00020000L
+
 static void nbio_v7_4_remap_hdp_registers(struct amdgpu_device *adev)
 {
 	WREG32_SOC15(NBIO, 0, mmREMAP_HDP_MEM_FLUSH_CNTL,
@@ -220,6 +231,12 @@ static const struct nbio_hdp_flush_reg nbio_v7_4_hdp_flush_reg = {
 	.ref_and_mask_cp9 = GPU_HDP_FLUSH_DONE__CP9_MASK,
 	.ref_and_mask_sdma0 = GPU_HDP_FLUSH_DONE__SDMA0_MASK,
 	.ref_and_mask_sdma1 = GPU_HDP_FLUSH_DONE__SDMA1_MASK,
+	.ref_and_mask_sdma2 = GPU_HDP_FLUSH_DONE__RSVD_ENG0_MASK,
+	.ref_and_mask_sdma3 = GPU_HDP_FLUSH_DONE__RSVD_ENG1_MASK,
+	.ref_and_mask_sdma4 = GPU_HDP_FLUSH_DONE__RSVD_ENG2_MASK,
+	.ref_and_mask_sdma5 = GPU_HDP_FLUSH_DONE__RSVD_ENG3_MASK,
+	.ref_and_mask_sdma6 = GPU_HDP_FLUSH_DONE__RSVD_ENG4_MASK,
+	.ref_and_mask_sdma7 = GPU_HDP_FLUSH_DONE__RSVD_ENG5_MASK,
 };
 
 static void nbio_v7_4_detect_hw_virt(struct amdgpu_device *adev)
drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c +1 -4
index 1dce974e0c58..fc01502baf00 100644
@@ -627,10 +627,7 @@ static void sdma_v4_0_ring_emit_hdp_flush(struct amdgpu_ring *ring)
 	u32 ref_and_mask = 0;
 	const struct nbio_hdp_flush_reg *nbio_hf_reg = adev->nbio_funcs->hdp_flush_reg;
 
-	if (ring->me == 0)
-		ref_and_mask = nbio_hf_reg->ref_and_mask_sdma0;
-	else
-		ref_and_mask = nbio_hf_reg->ref_and_mask_sdma1;
+	ref_and_mask = nbio_hf_reg->ref_and_mask_sdma0 << ring->me;
 
 	sdma_v4_0_wait_reg_mem(ring, 0, 1,
 			       adev->nbio_funcs->get_hdp_flush_done_offset(adev),
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#19 lore raw

PATCH 21/102 drm/amdgpu/soc15: add Arcturus common ip blocks

In reply to a message not in the archive

From: Le Ma <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/soc15.c | 18 ++++++++++++++++--
 1 file changed, 16 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/amdgpu/soc15.c +16 -2
index 800dd76dcf92..bb6aed206c8e 100644
@@ -597,12 +597,13 @@ int soc15_set_ip_blocks(struct amdgpu_device *adev)
 
 	if (adev->flags & AMD_IS_APU)
 		adev->nbio_funcs = &nbio_v7_0_funcs;
-	else if (adev->asic_type == CHIP_VEGA20)
+	else if (adev->asic_type == CHIP_VEGA20 ||
+		adev->asic_type == CHIP_ARCTURUS)
 		adev->nbio_funcs = &nbio_v7_4_funcs;
 	else
 		adev->nbio_funcs = &nbio_v6_1_funcs;
 
-	if (adev->asic_type == CHIP_VEGA20)
+	if (adev->asic_type == CHIP_VEGA20 || adev->asic_type == CHIP_ARCTURUS)
 		adev->df_funcs = &df_v3_6_funcs;
 	else
 		adev->df_funcs = &df_v1_7_funcs;
@@ -674,6 +675,15 @@ int soc15_set_ip_blocks(struct amdgpu_device *adev)
 #endif
 		amdgpu_device_ip_block_add(adev, &vcn_v1_0_ip_block);
 		break;
+	case CHIP_ARCTURUS:
+		amdgpu_device_ip_block_add(adev, &vega10_common_ip_block);
+		amdgpu_device_ip_block_add(adev, &gmc_v9_0_ip_block);
+		amdgpu_device_ip_block_add(adev, &vega10_ih_ip_block);
+		if (adev->enable_virtual_display || amdgpu_sriov_vf(adev))
+			amdgpu_device_ip_block_add(adev, &dce_virtual_ip_block);
+		amdgpu_device_ip_block_add(adev, &gfx_v9_0_ip_block);
+		amdgpu_device_ip_block_add(adev, &sdma_v4_0_ip_block);
+		break;
 	default:
 		return -EINVAL;
 	}
@@ -1000,6 +1010,10 @@ static int soc15_common_early_init(void *handle)
 				AMD_PG_SUPPORT_CP |
 				AMD_PG_SUPPORT_RLC_SMU_HS;
 		break;
+	case CHIP_ARCTURUS:
+		adev->cg_flags = 0;
+		adev->pg_flags = 0;
+		break;
 	default:
 		/* FIXME: not supported yet */
 		return -EINVAL;
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#20 lore raw

PATCH 22/102 drm/amdgpu: add to set Arcturus ip blocks

In reply to a message not in the archive

From: Le Ma <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 1 +
 1 file changed, 1 insertion(+)
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +1 -0
index f8250a776b93..017bbd541f00 100644
@@ -1534,6 +1534,7 @@ static int amdgpu_device_ip_early_init(struct amdgpu_device *adev)
 	case CHIP_VEGA12:
 	case CHIP_VEGA20:
 	case CHIP_RAVEN:
+	case CHIP_ARCTURUS:
 		if (adev->asic_type == CHIP_RAVEN)
 			adev->family = AMDGPU_FAMILY_RV;
 		else
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#21 lore raw

PATCH 23/102 drm/amdgpu: set Arcturus fw load type as direct

In reply to a message not in the archive

From: Le Ma <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c +2 -0
index 2042b6059127..1ec696858d78 100644
@@ -356,6 +356,8 @@ amdgpu_ucode_get_load_type(struct amdgpu_device *adev, int load_type)
 			return AMDGPU_FW_LOAD_DIRECT;
 		else
 			return AMDGPU_FW_LOAD_PSP;
+	case CHIP_ARCTURUS:
+		return AMDGPU_FW_LOAD_DIRECT;
 	default:
 		DRM_ERROR("Unknown firmware load type\n");
 	}
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#22 lore raw

PATCH 24/102 drm/amdgpu/dce_virtual: add Arcturus virtual display support

In reply to a message not in the archive

From: Le Ma <[email protected]>

Virtual dce is a sw only display driver for emulation and virtualization and cases where we want to use a virtual display subsystem.

---
 drivers/gpu/drm/amd/amdgpu/dce_virtual.c | 1 +
 1 file changed, 1 insertion(+)
drivers/gpu/drm/amd/amdgpu/dce_virtual.c +1 -0
index 4195b1172bdd..9b45f6e12c7c 100644
@@ -455,6 +455,7 @@ static int dce_virtual_hw_init(void *handle)
 	case CHIP_VEGA10:
 	case CHIP_VEGA12:
 	case CHIP_VEGA20:
+	case CHIP_ARCTURUS:
 	case CHIP_NAVI10:
 	case CHIP_NAVI14:
 		break;
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#23 lore raw

PATCH 25/102 drm/amdgpu: add support for Arcturus firmware

In reply to a message not in the archive

From: Le Ma <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 10 ++++++++++
 1 file changed, 10 insertions(+)
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +10 -0
index c67bb7fc51c0..626070e0d1e2 100644
@@ -100,6 +100,13 @@ MODULE_FIRMWARE("amdgpu/raven2_mec2.bin");
 MODULE_FIRMWARE("amdgpu/raven2_rlc.bin");
 MODULE_FIRMWARE("amdgpu/raven_kicker_rlc.bin");
 
+MODULE_FIRMWARE("amdgpu/arcturus_ce.bin");
+MODULE_FIRMWARE("amdgpu/arcturus_pfp.bin");
+MODULE_FIRMWARE("amdgpu/arcturus_me.bin");
+MODULE_FIRMWARE("amdgpu/arcturus_mec.bin");
+MODULE_FIRMWARE("amdgpu/arcturus_mec2.bin");
+MODULE_FIRMWARE("amdgpu/arcturus_rlc.bin");
+
 static const struct soc15_reg_golden golden_settings_gc_9_0[] =
 {
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_DEBUG2, 0xf00fffff, 0x00000400),
@@ -641,6 +648,9 @@ static int gfx_v9_0_init_microcode(struct amdgpu_device *adev)
 		else
 			chip_name = "raven";
 		break;
+	case CHIP_ARCTURUS:
+		chip_name = "arcturus";
+		break;
 	default:
 		BUG();
 	}
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#24 lore raw

PATCH 26/102 drm/amdgpu: add gfx config for Arcturus

In reply to a message not in the archive

From: Le Ma <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 10 ++++++++++
 1 file changed, 10 insertions(+)
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +10 -0
index 626070e0d1e2..9d57b4ae49cf 100644
@@ -1383,6 +1383,16 @@ static int gfx_v9_0_gpu_early_init(struct amdgpu_device *adev)
 		else
 			gb_addr_config = RAVEN_GB_ADDR_CONFIG_GOLDEN;
 		break;
+	case CHIP_ARCTURUS:
+		adev->gfx.config.max_hw_contexts = 8;
+		adev->gfx.config.sc_prim_fifo_size_frontend = 0x20;
+		adev->gfx.config.sc_prim_fifo_size_backend = 0x100;
+		adev->gfx.config.sc_hiz_tile_fifo_size = 0x30;
+		adev->gfx.config.sc_earlyz_tile_fifo_size = 0x4C0;
+		gb_addr_config = RREG32_SOC15(GC, 0, mmGB_ADDR_CONFIG);
+		gb_addr_config &= ~0xf3e777ff;
+		gb_addr_config |= 0x22014042;
+		break;
 	default:
 		BUG();
 		break;
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#25 lore raw

PATCH 27/102 drm/amdgpu: add number of mec for Arcturus

In reply to a message not in the archive

From: Le Ma <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 1 +
 1 file changed, 1 insertion(+)
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +1 -0
index 9d57b4ae49cf..e9d992dc04d6 100644
@@ -1669,6 +1669,7 @@ static int gfx_v9_0_sw_init(void *handle)
 	case CHIP_VEGA12:
 	case CHIP_VEGA20:
 	case CHIP_RAVEN:
+	case CHIP_ARCTURUS:
 		adev->gfx.mec.num_mec = 2;
 		break;
 	default:
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#26 lore raw

PATCH 28/102 drm/amdgpu: add to set rlc funcs for Arcturus

In reply to a message not in the archive

From: Le Ma <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 1 +
 1 file changed, 1 insertion(+)
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +1 -0
index e9d992dc04d6..c84419f73536 100644
@@ -5357,6 +5357,7 @@ static void gfx_v9_0_set_rlc_funcs(struct amdgpu_device *adev)
 	case CHIP_VEGA12:
 	case CHIP_VEGA20:
 	case CHIP_RAVEN:
+	case CHIP_ARCTURUS:
 		adev->gfx.rlc.funcs = &gfx_v9_0_rlc_funcs;
 		break;
 	default:
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#27 lore raw

PATCH 29/102 drm/amdgpu: skip to get 3D engine clockgating state for Arcturus

In reply to a message not in the archive

From: Le Ma <[email protected]>

It's because Arcturus has not 3D engine.

---
 drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 18 ++++++++++--------
 1 file changed, 10 insertions(+), 8 deletions(-)
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +10 -8
index c84419f73536..e0db0ecc334e 100644
@@ -4338,14 +4338,16 @@ static void gfx_v9_0_get_clockgating_state(void *handle, u32 *flags)
 	if (data & CP_MEM_SLP_CNTL__CP_MEM_LS_EN_MASK)
 		*flags |= AMD_CG_SUPPORT_GFX_CP_LS | AMD_CG_SUPPORT_GFX_MGLS;
 
-	/* AMD_CG_SUPPORT_GFX_3D_CGCG */
-	data = RREG32_SOC15(GC, 0, mmRLC_CGCG_CGLS_CTRL_3D);
-	if (data & RLC_CGCG_CGLS_CTRL_3D__CGCG_EN_MASK)
-		*flags |= AMD_CG_SUPPORT_GFX_3D_CGCG;
-
-	/* AMD_CG_SUPPORT_GFX_3D_CGLS */
-	if (data & RLC_CGCG_CGLS_CTRL_3D__CGLS_EN_MASK)
-		*flags |= AMD_CG_SUPPORT_GFX_3D_CGLS;
+	if (adev->asic_type != CHIP_ARCTURUS) {
+		/* AMD_CG_SUPPORT_GFX_3D_CGCG */
+		data = RREG32_SOC15(GC, 0, mmRLC_CGCG_CGLS_CTRL_3D);
+		if (data & RLC_CGCG_CGLS_CTRL_3D__CGCG_EN_MASK)
+			*flags |= AMD_CG_SUPPORT_GFX_3D_CGCG;
+
+		/* AMD_CG_SUPPORT_GFX_3D_CGLS */
+		if (data & RLC_CGCG_CGLS_CTRL_3D__CGLS_EN_MASK)
+			*flags |= AMD_CG_SUPPORT_GFX_3D_CGLS;
+	}
 }
 
 static u64 gfx_v9_0_ring_get_rptr_gfx(struct amdgpu_ring *ring)
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#28 lore raw

PATCH 30/102 drm/amdgpu: skip pasid mapping for second mmhub on Arcturus

In reply to a message not in the archive

From: Le Ma <[email protected]>

There's no LUT register for second mmhub to convert pasid since it has no ATC.

---
 drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c +4 -0
index d6aa290d567c..11968fbe0d60 100644
@@ -488,6 +488,10 @@ static void gmc_v9_0_emit_pasid_mapping(struct amdgpu_ring *ring, unsigned vmid,
 	struct amdgpu_device *adev = ring->adev;
 	uint32_t reg;
 
+	/* Do nothing because there's no lut register for mmhub1. */
+	if (ring->funcs->vmhub == AMDGPU_MMHUB_1)
+		return;
+
 	if (ring->funcs->vmhub == AMDGPU_GFXHUB_0)
 		reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_VMID_0_LUT) + vmid;
 	else
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#29 lore raw

PATCH 31/102 drm/amdgpu: add Arcturus gpu info firmware

In reply to a message not in the archive

From: Le Ma <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +4 -0
index 017bbd541f00..4425ff06ecc4 100644
@@ -69,6 +69,7 @@ MODULE_FIRMWARE("amdgpu/vega12_gpu_info.bin");
 MODULE_FIRMWARE("amdgpu/raven_gpu_info.bin");
 MODULE_FIRMWARE("amdgpu/picasso_gpu_info.bin");
 MODULE_FIRMWARE("amdgpu/raven2_gpu_info.bin");
+MODULE_FIRMWARE("amdgpu/arcturus_gpu_info.bin");
 MODULE_FIRMWARE("amdgpu/navi10_gpu_info.bin");
 MODULE_FIRMWARE("amdgpu/navi14_gpu_info.bin");
 
@@ -1386,6 +1387,9 @@ static int amdgpu_device_parse_gpu_info_fw(struct amdgpu_device *adev)
 		else
 			chip_name = "raven";
 		break;
+	case CHIP_ARCTURUS:
+		chip_name = "arcturus";
+		break;
 	case CHIP_NAVI10:
 		chip_name = "navi10";
 		break;
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#30 lore raw

PATCH 32/102 drm/amdgpu: optimize gfx9 init_microcode function

In reply to a message not in the archive

From: Le Ma <[email protected]>

Split each type of firmware into single function for easy to maintain.

---
 drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 247 ++++++++++++++++----------
 1 file changed, 153 insertions(+), 94 deletions(-)
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +153 -94
index e0db0ecc334e..edd3224720dd 100644
@@ -613,47 +613,14 @@ static void gfx_v9_0_check_if_need_gfxoff(struct amdgpu_device *adev)
 	}
 }
 
-static int gfx_v9_0_init_microcode(struct amdgpu_device *adev)
+static int gfx_v9_0_init_cp_gfx_microcode(struct amdgpu_device *adev,
+					  const char *chip_name)
 {
-	const char *chip_name;
 	char fw_name[30];
 	int err;
 	struct amdgpu_firmware_info *info = NULL;
 	const struct common_firmware_header *header = NULL;
 	const struct gfx_firmware_header_v1_0 *cp_hdr;
-	const struct rlc_firmware_header_v2_0 *rlc_hdr;
-	unsigned int *tmp = NULL;
-	unsigned int i = 0;
-	uint16_t version_major;
-	uint16_t version_minor;
-	uint32_t smu_version;
-
-	DRM_DEBUG("\n");
-
-	switch (adev->asic_type) {
-	case CHIP_VEGA10:
-		chip_name = "vega10";
-		break;
-	case CHIP_VEGA12:
-		chip_name = "vega12";
-		break;
-	case CHIP_VEGA20:
-		chip_name = "vega20";
-		break;
-	case CHIP_RAVEN:
-		if (adev->rev_id >= 8)
-			chip_name = "raven2";
-		else if (adev->pdev->device == 0x15d8)
-			chip_name = "picasso";
-		else
-			chip_name = "raven";
-		break;
-	case CHIP_ARCTURUS:
-		chip_name = "arcturus";
-		break;
-	default:
-		BUG();
-	}
 
 	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_pfp.bin", chip_name);
 	err = request_firmware(&adev->gfx.pfp_fw, fw_name, adev->dev);
@@ -688,6 +655,58 @@ static int gfx_v9_0_init_microcode(struct amdgpu_device *adev)
 	adev->gfx.ce_fw_version = le32_to_cpu(cp_hdr->header.ucode_version);
 	adev->gfx.ce_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version);
 
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_PFP];
+		info->ucode_id = AMDGPU_UCODE_ID_CP_PFP;
+		info->fw = adev->gfx.pfp_fw;
+		header = (const struct common_firmware_header *)info->fw->data;
+		adev->firmware.fw_size +=
+			ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
+
+		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_ME];
+		info->ucode_id = AMDGPU_UCODE_ID_CP_ME;
+		info->fw = adev->gfx.me_fw;
+		header = (const struct common_firmware_header *)info->fw->data;
+		adev->firmware.fw_size +=
+			ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
+
+		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_CE];
+		info->ucode_id = AMDGPU_UCODE_ID_CP_CE;
+		info->fw = adev->gfx.ce_fw;
+		header = (const struct common_firmware_header *)info->fw->data;
+		adev->firmware.fw_size +=
+			ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
+	}
+
+out:
+	if (err) {
+		dev_err(adev->dev,
+			"gfx9: Failed to load firmware \"%s\"\n",
+			fw_name);
+		release_firmware(adev->gfx.pfp_fw);
+		adev->gfx.pfp_fw = NULL;
+		release_firmware(adev->gfx.me_fw);
+		adev->gfx.me_fw = NULL;
+		release_firmware(adev->gfx.ce_fw);
+		adev->gfx.ce_fw = NULL;
+	}
+	return err;
+}
+
+static int gfx_v9_0_init_rlc_microcode(struct amdgpu_device *adev,
+					  const char *chip_name)
+{
+	char fw_name[30];
+	int err;
+	struct amdgpu_firmware_info *info = NULL;
+	const struct common_firmware_header *header = NULL;
+	const struct rlc_firmware_header_v2_0 *rlc_hdr;
+	unsigned int *tmp = NULL;
+	unsigned int i = 0;
+	uint16_t version_major;
+	uint16_t version_minor;
+	uint32_t smu_version;
+
 	/*
 	 * For Picasso && AM4 SOCKET board, we use picasso_rlc_am4.bin
 	 * instead of picasso_rlc.bin.
@@ -762,57 +781,7 @@ static int gfx_v9_0_init_microcode(struct amdgpu_device *adev)
 	if (adev->gfx.rlc.is_rlc_v2_1)
 		gfx_v9_0_init_rlc_ext_microcode(adev);
 
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec.bin", chip_name);
-	err = request_firmware(&adev->gfx.mec_fw, fw_name, adev->dev);
-	if (err)
-		goto out;
-	err = amdgpu_ucode_validate(adev->gfx.mec_fw);
-	if (err)
-		goto out;
-	cp_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.mec_fw->data;
-	adev->gfx.mec_fw_version = le32_to_cpu(cp_hdr->header.ucode_version);
-	adev->gfx.mec_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version);
-
-
-	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec2.bin", chip_name);
-	err = request_firmware(&adev->gfx.mec2_fw, fw_name, adev->dev);
-	if (!err) {
-		err = amdgpu_ucode_validate(adev->gfx.mec2_fw);
-		if (err)
-			goto out;
-		cp_hdr = (const struct gfx_firmware_header_v1_0 *)
-		adev->gfx.mec2_fw->data;
-		adev->gfx.mec2_fw_version =
-		le32_to_cpu(cp_hdr->header.ucode_version);
-		adev->gfx.mec2_feature_version =
-		le32_to_cpu(cp_hdr->ucode_feature_version);
-	} else {
-		err = 0;
-		adev->gfx.mec2_fw = NULL;
-	}
-
 	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
-		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_PFP];
-		info->ucode_id = AMDGPU_UCODE_ID_CP_PFP;
-		info->fw = adev->gfx.pfp_fw;
-		header = (const struct common_firmware_header *)info->fw->data;
-		adev->firmware.fw_size +=
-			ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
-
-		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_ME];
-		info->ucode_id = AMDGPU_UCODE_ID_CP_ME;
-		info->fw = adev->gfx.me_fw;
-		header = (const struct common_firmware_header *)info->fw->data;
-		adev->firmware.fw_size +=
-			ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
-
-		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_CE];
-		info->ucode_id = AMDGPU_UCODE_ID_CP_CE;
-		info->fw = adev->gfx.ce_fw;
-		header = (const struct common_firmware_header *)info->fw->data;
-		adev->firmware.fw_size +=
-			ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
-
 		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_G];
 		info->ucode_id = AMDGPU_UCODE_ID_RLC_G;
 		info->fw = adev->gfx.rlc_fw;
@@ -842,7 +811,58 @@ static int gfx_v9_0_init_microcode(struct amdgpu_device *adev)
 			adev->firmware.fw_size +=
 				ALIGN(adev->gfx.rlc.save_restore_list_srm_size_bytes, PAGE_SIZE);
 		}
+	}
+
+out:
+	if (err) {
+		dev_err(adev->dev,
+			"gfx9: Failed to load firmware \"%s\"\n",
+			fw_name);
+		release_firmware(adev->gfx.rlc_fw);
+		adev->gfx.rlc_fw = NULL;
+	}
+	return err;
+}
+
+static int gfx_v9_0_init_cp_compute_microcode(struct amdgpu_device *adev,
+					  const char *chip_name)
+{
+	char fw_name[30];
+	int err;
+	struct amdgpu_firmware_info *info = NULL;
+	const struct common_firmware_header *header = NULL;
+	const struct gfx_firmware_header_v1_0 *cp_hdr;
+
+	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec.bin", chip_name);
+	err = request_firmware(&adev->gfx.mec_fw, fw_name, adev->dev);
+	if (err)
+		goto out;
+	err = amdgpu_ucode_validate(adev->gfx.mec_fw);
+	if (err)
+		goto out;
+	cp_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.mec_fw->data;
+	adev->gfx.mec_fw_version = le32_to_cpu(cp_hdr->header.ucode_version);
+	adev->gfx.mec_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version);
+
 
+	snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec2.bin", chip_name);
+	err = request_firmware(&adev->gfx.mec2_fw, fw_name, adev->dev);
+	if (!err) {
+		err = amdgpu_ucode_validate(adev->gfx.mec2_fw);
+		if (err)
+			goto out;
+		cp_hdr = (const struct gfx_firmware_header_v1_0 *)
+		adev->gfx.mec2_fw->data;
+		adev->gfx.mec2_fw_version =
+		le32_to_cpu(cp_hdr->header.ucode_version);
+		adev->gfx.mec2_feature_version =
+		le32_to_cpu(cp_hdr->ucode_feature_version);
+	} else {
+		err = 0;
+		adev->gfx.mec2_fw = NULL;
+	}
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
 		info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_MEC1];
 		info->ucode_id = AMDGPU_UCODE_ID_CP_MEC1;
 		info->fw = adev->gfx.mec_fw;
@@ -871,7 +891,6 @@ static int gfx_v9_0_init_microcode(struct amdgpu_device *adev)
 			adev->firmware.fw_size +=
 				ALIGN(le32_to_cpu(cp_hdr->jt_size) * 4, PAGE_SIZE);
 		}
-
 	}
 
 out:
@@ -881,14 +900,6 @@ static int gfx_v9_0_init_microcode(struct amdgpu_device *adev)
 		dev_err(adev->dev,
 			"gfx9: Failed to load firmware \"%s\"\n",
 			fw_name);
-		release_firmware(adev->gfx.pfp_fw);
-		adev->gfx.pfp_fw = NULL;
-		release_firmware(adev->gfx.me_fw);
-		adev->gfx.me_fw = NULL;
-		release_firmware(adev->gfx.ce_fw);
-		adev->gfx.ce_fw = NULL;
-		release_firmware(adev->gfx.rlc_fw);
-		adev->gfx.rlc_fw = NULL;
 		release_firmware(adev->gfx.mec_fw);
 		adev->gfx.mec_fw = NULL;
 		release_firmware(adev->gfx.mec2_fw);
@@ -897,6 +908,54 @@ static int gfx_v9_0_init_microcode(struct amdgpu_device *adev)
 	return err;
 }
 
+static int gfx_v9_0_init_microcode(struct amdgpu_device *adev)
+{
+	const char *chip_name;
+	int r;
+
+	DRM_DEBUG("\n");
+
+	switch (adev->asic_type) {
+	case CHIP_VEGA10:
+		chip_name = "vega10";
+		break;
+	case CHIP_VEGA12:
+		chip_name = "vega12";
+		break;
+	case CHIP_VEGA20:
+		chip_name = "vega20";
+		break;
+	case CHIP_RAVEN:
+		if (adev->rev_id >= 8)
+			chip_name = "raven2";
+		else if (adev->pdev->device == 0x15d8)
+			chip_name = "picasso";
+		else
+			chip_name = "raven";
+		break;
+		break;
+	case CHIP_ARCTURUS:
+		chip_name = "arcturus";
+		break;
+	default:
+		BUG();
+	}
+
+	r = gfx_v9_0_init_cp_gfx_microcode(adev, chip_name);
+	if (r)
+		return r;
+
+	r = gfx_v9_0_init_rlc_microcode(adev, chip_name);
+	if (r)
+		return r;
+
+	r = gfx_v9_0_init_cp_compute_microcode(adev, chip_name);
+	if (r)
+		return r;
+
+	return r;
+}
+
 static u32 gfx_v9_0_get_csb_size(struct amdgpu_device *adev)
 {
 	u32 count = 0;
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#31 lore raw

PATCH 33/102 drm/amdgpu: skip load cp gfx firmware for Arcturus

In reply to a message not in the archive

From: Le Ma <[email protected]>

Arcturus has no CPG component any more.

---
 drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 19 ++++++++++++-------
 1 file changed, 12 insertions(+), 7 deletions(-)
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +12 -7
index edd3224720dd..0676ae38ec1e 100644
@@ -941,9 +941,12 @@ static int gfx_v9_0_init_microcode(struct amdgpu_device *adev)
 		BUG();
 	}
 
-	r = gfx_v9_0_init_cp_gfx_microcode(adev, chip_name);
-	if (r)
-		return r;
+	/* No CPG in Arcturus */
+	if (adev->asic_type != CHIP_ARCTURUS) {
+		r = gfx_v9_0_init_cp_gfx_microcode(adev, chip_name);
+		if (r)
+			return r;
+	}
 
 	r = gfx_v9_0_init_rlc_microcode(adev, chip_name);
 	if (r)
@@ -3319,10 +3322,12 @@ static int gfx_v9_0_cp_resume(struct amdgpu_device *adev)
 		gfx_v9_0_enable_gui_idle_interrupt(adev, false);
 
 	if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) {
-		/* legacy firmware loading */
-		r = gfx_v9_0_cp_gfx_load_microcode(adev);
-		if (r)
-			return r;
+		if (adev->asic_type != CHIP_ARCTURUS) {
+			/* legacy firmware loading */
+			r = gfx_v9_0_cp_gfx_load_microcode(adev);
+			if (r)
+				return r;
+		}
 
 		r = gfx_v9_0_cp_compute_load_microcode(adev);
 		if (r)
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#32 lore raw

PATCH 34/102 drm/amdgpu: skip all gfx ring settings for Arcturus

In reply to a message not in the archive

From: Le Ma <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 39 +++++++++++++++++----------
 1 file changed, 25 insertions(+), 14 deletions(-)
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +25 -14
index 0676ae38ec1e..3ac3ee7fcb1a 100644
@@ -3338,18 +3338,22 @@ static int gfx_v9_0_cp_resume(struct amdgpu_device *adev)
 	if (r)
 		return r;
 
-	r = gfx_v9_0_cp_gfx_resume(adev);
-	if (r)
-		return r;
+	if (adev->asic_type != CHIP_ARCTURUS) {
+		r = gfx_v9_0_cp_gfx_resume(adev);
+		if (r)
+			return r;
+	}
 
 	r = gfx_v9_0_kcq_resume(adev);
 	if (r)
 		return r;
 
-	ring = &adev->gfx.gfx_ring[0];
-	r = amdgpu_ring_test_helper(ring);
-	if (r)
-		return r;
+	if (adev->asic_type != CHIP_ARCTURUS) {
+		ring = &adev->gfx.gfx_ring[0];
+		r = amdgpu_ring_test_helper(ring);
+		if (r)
+			return r;
+	}
 
 	for (i = 0; i < adev->gfx.num_compute_rings; i++) {
 		ring = &adev->gfx.compute_ring[i];
@@ -3363,7 +3367,8 @@ static int gfx_v9_0_cp_resume(struct amdgpu_device *adev)
 
 static void gfx_v9_0_cp_enable(struct amdgpu_device *adev, bool enable)
 {
-	gfx_v9_0_cp_gfx_enable(adev, enable);
+	if (adev->asic_type != CHIP_ARCTURUS)
+		gfx_v9_0_cp_gfx_enable(adev, enable);
 	gfx_v9_0_cp_compute_enable(adev, enable);
 }
 
@@ -3388,9 +3393,11 @@ static int gfx_v9_0_hw_init(void *handle)
 	if (r)
 		return r;
 
-	r = gfx_v9_0_ngg_en(adev);
-	if (r)
-		return r;
+	if (adev->asic_type != CHIP_ARCTURUS) {
+		r = gfx_v9_0_ngg_en(adev);
+		if (r)
+			return r;
+	}
 
 	return r;
 }
@@ -3538,8 +3545,9 @@ static int gfx_v9_0_soft_reset(void *handle)
 		/* stop the rlc */
 		adev->gfx.rlc.funcs->stop(adev);
 
-		/* Disable GFX parsing/prefetching */
-		gfx_v9_0_cp_gfx_enable(adev, false);
+		if (adev->asic_type != CHIP_ARCTURUS)
+			/* Disable GFX parsing/prefetching */
+			gfx_v9_0_cp_gfx_enable(adev, false);
 
 		/* Disable MEC parsing/prefetching */
 		gfx_v9_0_cp_compute_enable(adev, false);
@@ -3882,7 +3890,10 @@ static int gfx_v9_0_early_init(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	adev->gfx.num_gfx_rings = GFX9_NUM_GFX_RINGS;
+	if (adev->asic_type == CHIP_ARCTURUS)
+		adev->gfx.num_gfx_rings = 0;
+	else
+		adev->gfx.num_gfx_rings = GFX9_NUM_GFX_RINGS;
 	adev->gfx.num_compute_rings = AMDGPU_MAX_COMPUTE_RINGS;
 	gfx_v9_0_set_ring_funcs(adev);
 	gfx_v9_0_set_irq_funcs(adev);
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#33 lore raw

PATCH 35/102 drm/amdgpu: support sdma 2~7 doorbell range register offset

In reply to a message not in the archive

From: Le Ma <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c | 20 +++++++++++++++++---
 1 file changed, 17 insertions(+), 3 deletions(-)
drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c +17 -3
index 2e1098c5c32b..fc45eaeaba6e 100644
@@ -86,10 +86,24 @@ static u32 nbio_v7_4_get_memsize(struct amdgpu_device *adev)
 static void nbio_v7_4_sdma_doorbell_range(struct amdgpu_device *adev, int instance,
 			bool use_doorbell, int doorbell_index, int doorbell_size)
 {
-	u32 reg = instance == 0 ? SOC15_REG_OFFSET(NBIO, 0, mmBIF_SDMA0_DOORBELL_RANGE) :
-			SOC15_REG_OFFSET(NBIO, 0, mmBIF_SDMA1_DOORBELL_RANGE);
+	u32 reg, doorbell_range;
 
-	u32 doorbell_range = RREG32(reg);
+	if (instance < 2)
+		reg = instance +
+			SOC15_REG_OFFSET(NBIO, 0, mmBIF_SDMA0_DOORBELL_RANGE);
+	else
+		/*
+		 * These registers address of SDMA2~7 is not consecutive
+		 * from SDMA0~1. Need plus 4 dwords offset.
+		 *
+		 *   BIF_SDMA0_DOORBELL_RANGE:  0x3bc0
+		 *   BIF_SDMA1_DOORBELL_RANGE:  0x3bc4
+		 *   BIF_SDMA2_DOORBELL_RANGE:  0x3bd8
+		 */
+		reg = instance + 0x4 +
+			SOC15_REG_OFFSET(NBIO, 0, mmBIF_SDMA0_DOORBELL_RANGE);
+
+	doorbell_range = RREG32(reg);
 
 	if (use_doorbell) {
 		doorbell_range = REG_SET_FIELD(doorbell_range, BIF_SDMA0_DOORBELL_RANGE, OFFSET, doorbell_index);
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#34 lore raw

PATCH 36/102 drm/amdgpu: correct Arcturus SDMA address space base index

In reply to a message not in the archive

From: Le Ma <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)
drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c +6 -6
index fc01502baf00..974228b918d2 100644
@@ -208,17 +208,17 @@ static u32 sdma_v4_0_get_reg_offset(struct amdgpu_device *adev,
 	case 1:
 		return (adev->reg_offset[SDMA1_HWIP][0][0] + offset);
 	case 2:
-		return (adev->reg_offset[SDMA2_HWIP][0][0] + offset);
+		return (adev->reg_offset[SDMA2_HWIP][0][1] + offset);
 	case 3:
-		return (adev->reg_offset[SDMA3_HWIP][0][0] + offset);
+		return (adev->reg_offset[SDMA3_HWIP][0][1] + offset);
 	case 4:
-		return (adev->reg_offset[SDMA4_HWIP][0][0] + offset);
+		return (adev->reg_offset[SDMA4_HWIP][0][1] + offset);
 	case 5:
-		return (adev->reg_offset[SDMA5_HWIP][0][0] + offset);
+		return (adev->reg_offset[SDMA5_HWIP][0][1] + offset);
 	case 6:
-		return (adev->reg_offset[SDMA6_HWIP][0][0] + offset);
+		return (adev->reg_offset[SDMA6_HWIP][0][1] + offset);
 	case 7:
-		return (adev->reg_offset[SDMA7_HWIP][0][0] + offset);
+		return (adev->reg_offset[SDMA7_HWIP][0][1] + offset);
 	default:
 		break;
 	}
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#35 lore raw

PATCH 37/102 drm/amdgpu: enable 8 SDMA instances for Arcturus

In reply to a message not in the archive

From: Le Ma <[email protected]>

All the 8 SDMA instances work fine on the latest Gopher build model.

---
 drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c +2 -0
index 974228b918d2..0ca41a8a7e5b 100644
@@ -1539,6 +1539,8 @@ static int sdma_v4_0_early_init(void *handle)
 
 	if (adev->asic_type == CHIP_RAVEN)
 		adev->sdma.num_instances = 1;
+	else if (adev->asic_type == CHIP_ARCTURUS)
+		adev->sdma.num_instances = 8;
 	else
 		adev->sdma.num_instances = 2;
 
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#36 lore raw

PATCH 38/102 drm/amdgpu: add Arcturus chip_name for init sdma microcode

In reply to a message not in the archive

From: Le Ma <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c +3 -0
index 0ca41a8a7e5b..cf064caa160c 100644
@@ -366,6 +366,9 @@ static int sdma_v4_0_init_microcode(struct amdgpu_device *adev)
 		else
 			chip_name = "raven";
 		break;
+	case CHIP_ARCTURUS:
+		chip_name = "arcturus";
+		break;
 	default:
 		BUG();
 	}
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#37 lore raw

PATCH 39/102 drm/amdgpu: correct programming of ih_chicken for Arcturus

In reply to a message not in the archive

From: Le Ma <[email protected]>

ih_chicken is a register that is not allowed to access by driver in the L0 security policy.
psp bl need to enable field to allow driver to use physical bus address for ih ring on secure part.

---
 drivers/gpu/drm/amd/amdgpu/vega10_ih.c | 11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/vega10_ih.c +10 -1
index 5f54acc70fec..ee9cd8579038 100644
@@ -217,7 +217,7 @@ static uint32_t vega10_ih_doorbell_rptr(struct amdgpu_ih_ring *ih)
 static int vega10_ih_irq_init(struct amdgpu_device *adev)
 {
 	struct amdgpu_ih_ring *ih;
-	u32 ih_rb_cntl;
+	u32 ih_rb_cntl, ih_chicken;
 	int ret = 0;
 	u32 tmp;
 
@@ -245,6 +245,15 @@ static int vega10_ih_irq_init(struct amdgpu_device *adev)
 		WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL, ih_rb_cntl);
 	}
 
+	if (adev->asic_type == CHIP_ARCTURUS &&
+		adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
+		if (adev->irq.ih.use_bus_addr) {
+			ih_chicken = RREG32_SOC15(OSSSYS, 0, mmIH_CHICKEN);
+			ih_chicken |= 0x00000010;
+			WREG32_SOC15(OSSSYS, 0, mmIH_CHICKEN, ih_chicken);
+		}
+	}
+
 	/* set the writeback address whether it's enabled or not */
 	WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR_ADDR_LO,
 		     lower_32_bits(ih->wptr_addr));
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#38 lore raw

PATCH 40/102 drm/amdgpu: add paging queue support for 8 SDMA instances on Arcturus

In reply to a message not in the archive

From: Le Ma <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c | 62 +++++++++++++++++++++-----
 1 file changed, 51 insertions(+), 11 deletions(-)
drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c +51 -11
index cf064caa160c..594750a9f2fc 100644
@@ -730,16 +730,20 @@ static void sdma_v4_0_rlc_stop(struct amdgpu_device *adev)
  */
 static void sdma_v4_0_page_stop(struct amdgpu_device *adev)
 {
-	struct amdgpu_ring *sdma0 = &adev->sdma.instance[0].page;
-	struct amdgpu_ring *sdma1 = &adev->sdma.instance[1].page;
+	struct amdgpu_ring *sdma[AMDGPU_MAX_SDMA_INSTANCES];
 	u32 rb_cntl, ib_cntl;
 	int i;
-
-	if ((adev->mman.buffer_funcs_ring == sdma0) ||
-	    (adev->mman.buffer_funcs_ring == sdma1))
-		amdgpu_ttm_set_buffer_funcs_status(adev, false);
+	bool unset = false;
 
 	for (i = 0; i < adev->sdma.num_instances; i++) {
+		sdma[i] = &adev->sdma.instance[i].page;
+
+		if ((adev->mman.buffer_funcs_ring == sdma[i]) &&
+			(unset == false)) {
+			amdgpu_ttm_set_buffer_funcs_status(adev, false);
+			unset = true;
+		}
+
 		rb_cntl = RREG32_SDMA(i, mmSDMA0_PAGE_RB_CNTL);
 		rb_cntl = REG_SET_FIELD(rb_cntl, SDMA0_PAGE_RB_CNTL,
 					RB_ENABLE, 0);
@@ -748,10 +752,9 @@ static void sdma_v4_0_page_stop(struct amdgpu_device *adev)
 		ib_cntl = REG_SET_FIELD(ib_cntl, SDMA0_PAGE_IB_CNTL,
 					IB_ENABLE, 0);
 		WREG32_SDMA(i, mmSDMA0_PAGE_IB_CNTL, ib_cntl);
-	}
 
-	sdma0->sched.ready = false;
-	sdma1->sched.ready = false;
+		sdma[i]->sched.ready = false;
+	}
 }
 
 /**
@@ -2256,7 +2259,39 @@ static const struct amdgpu_ring_funcs sdma_v4_0_page_ring_funcs = {
 	.align_mask = 0xf,
 	.nop = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP),
 	.support_64bit_ptrs = true,
-	.vmhub = AMDGPU_MMHUB,
+	.vmhub = AMDGPU_MMHUB_0,
+	.get_rptr = sdma_v4_0_ring_get_rptr,
+	.get_wptr = sdma_v4_0_page_ring_get_wptr,
+	.set_wptr = sdma_v4_0_page_ring_set_wptr,
+	.emit_frame_size =
+		6 + /* sdma_v4_0_ring_emit_hdp_flush */
+		3 + /* hdp invalidate */
+		6 + /* sdma_v4_0_ring_emit_pipeline_sync */
+		/* sdma_v4_0_ring_emit_vm_flush */
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 6 +
+		10 + 10 + 10, /* sdma_v4_0_ring_emit_fence x3 for user fence, vm fence */
+	.emit_ib_size = 7 + 6, /* sdma_v4_0_ring_emit_ib */
+	.emit_ib = sdma_v4_0_ring_emit_ib,
+	.emit_fence = sdma_v4_0_ring_emit_fence,
+	.emit_pipeline_sync = sdma_v4_0_ring_emit_pipeline_sync,
+	.emit_vm_flush = sdma_v4_0_ring_emit_vm_flush,
+	.emit_hdp_flush = sdma_v4_0_ring_emit_hdp_flush,
+	.test_ring = sdma_v4_0_ring_test_ring,
+	.test_ib = sdma_v4_0_ring_test_ib,
+	.insert_nop = sdma_v4_0_ring_insert_nop,
+	.pad_ib = sdma_v4_0_ring_pad_ib,
+	.emit_wreg = sdma_v4_0_ring_emit_wreg,
+	.emit_reg_wait = sdma_v4_0_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+};
+
+static const struct amdgpu_ring_funcs sdma_v4_0_page_ring_funcs_2nd_mmhub = {
+	.type = AMDGPU_RING_TYPE_SDMA,
+	.align_mask = 0xf,
+	.nop = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP),
+	.support_64bit_ptrs = true,
+	.vmhub = AMDGPU_MMHUB_1,
 	.get_rptr = sdma_v4_0_ring_get_rptr,
 	.get_wptr = sdma_v4_0_page_ring_get_wptr,
 	.set_wptr = sdma_v4_0_page_ring_set_wptr,
@@ -2296,7 +2331,12 @@ static void sdma_v4_0_set_ring_funcs(struct amdgpu_device *adev)
 					&sdma_v4_0_ring_funcs;
 		adev->sdma.instance[i].ring.me = i;
 		if (adev->sdma.has_page_queue) {
-			adev->sdma.instance[i].page.funcs = &sdma_v4_0_page_ring_funcs;
+			if (adev->asic_type == CHIP_ARCTURUS && i >=5)
+				adev->sdma.instance[i].page.funcs =
+					&sdma_v4_0_page_ring_funcs_2nd_mmhub;
+			else
+				adev->sdma.instance[i].page.funcs =
+					&sdma_v4_0_page_ring_funcs;
 			adev->sdma.instance[i].page.me = i;
 		}
 	}
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#39 lore raw

PATCH 41/102 drm/amdgpu: Initialize asic functions for Arcturus

In reply to a message not in the archive

From: Oak Zeng <[email protected]>

After cherry-picking doorbell rework changes from drm-next branch, Arcturus asic functions pointer need to be initialized to init doorbell index for Arcturus.

---
 drivers/gpu/drm/amd/amdgpu/soc15.c | 1 +
 1 file changed, 1 insertion(+)
drivers/gpu/drm/amd/amdgpu/soc15.c +1 -0
index bb6aed206c8e..8c881697d29b 100644
@@ -1011,6 +1011,7 @@ static int soc15_common_early_init(void *handle)
 				AMD_PG_SUPPORT_RLC_SMU_HS;
 		break;
 	case CHIP_ARCTURUS:
+		adev->asic_funcs = &vega20_asic_funcs;
 		adev->cg_flags = 0;
 		adev->pg_flags = 0;
 		break;
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#40 lore raw

PATCH 42/102 drm/amdkfd: Extend PM4 packets to support 8 SDMA

In reply to a message not in the archive

From: Oak Zeng <[email protected]>

Extend map_queue and unmap_queue PM4 packets to support 8 SDMA engines. The new format is backward compatible.

---
 .../gpu/drm/amd/amdkfd/kfd_kernel_queue_v9.c  | 26 ++++++++++++++++---
 .../gpu/drm/amd/amdkfd/kfd_pm4_headers_ai.h   | 14 ++++++++--
 2 files changed, 34 insertions(+), 6 deletions(-)
drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue_v9.c +22 -4
index 2d5ddf199bd0..91da72d0d405 100644
@@ -161,6 +161,8 @@ static int pm_map_queues_v9(struct packet_manager *pm, uint32_t *buffer,
 	packet->bitfields2.engine_sel =
 		engine_sel__mes_map_queues__compute_vi;
 	packet->bitfields2.gws_control_queue = q->gws ? 1 : 0;
+	packet->bitfields2.extended_engine_sel =
+		extended_engine_sel__mes_map_queues__legacy_engine_sel;
 	packet->bitfields2.queue_type =
 		queue_type__mes_map_queues__normal_compute_vi;
 
@@ -176,9 +178,15 @@ static int pm_map_queues_v9(struct packet_manager *pm, uint32_t *buffer,
 		break;
 	case KFD_QUEUE_TYPE_SDMA:
 	case KFD_QUEUE_TYPE_SDMA_XGMI:
-		packet->bitfields2.engine_sel = q->properties.sdma_engine_id +
-				engine_sel__mes_map_queues__sdma0_vi;
 		use_static = false; /* no static queues under SDMA */
+		if (q->properties.sdma_engine_id < 2)
+			packet->bitfields2.engine_sel = q->properties.sdma_engine_id +
+				engine_sel__mes_map_queues__sdma0_vi;
+		else {
+			packet->bitfields2.extended_engine_sel =
+				extended_engine_sel__mes_map_queues__sdma0_to_7_sel;
+			packet->bitfields2.engine_sel = q->properties.sdma_engine_id;
+		}
 		break;
 	default:
 		WARN(1, "queue type %d", q->properties.type);
@@ -218,13 +226,23 @@ static int pm_unmap_queues_v9(struct packet_manager *pm, uint32_t *buffer,
 	switch (type) {
 	case KFD_QUEUE_TYPE_COMPUTE:
 	case KFD_QUEUE_TYPE_DIQ:
+		packet->bitfields2.extended_engine_sel =
+			extended_engine_sel__mes_unmap_queues__legacy_engine_sel;
 		packet->bitfields2.engine_sel =
 			engine_sel__mes_unmap_queues__compute;
 		break;
 	case KFD_QUEUE_TYPE_SDMA:
 	case KFD_QUEUE_TYPE_SDMA_XGMI:
-		packet->bitfields2.engine_sel =
-			engine_sel__mes_unmap_queues__sdma0 + sdma_engine;
+		if (sdma_engine < 2) {
+			packet->bitfields2.extended_engine_sel =
+				extended_engine_sel__mes_unmap_queues__legacy_engine_sel;
+			packet->bitfields2.engine_sel =
+				engine_sel__mes_unmap_queues__sdma0 + sdma_engine;
+		} else {
+			packet->bitfields2.extended_engine_sel =
+				extended_engine_sel__mes_unmap_queues__sdma0_to_7_sel;
+			packet->bitfields2.engine_sel = sdma_engine;
+		}
 		break;
 	default:
 		WARN(1, "queue type %d", type);
drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers_ai.h +12 -2
index e3e21404cfa0..44ed94239513 100644
@@ -260,6 +260,10 @@ enum mes_map_queues_engine_sel_enum {
 	engine_sel__mes_map_queues__sdma1_vi = 3
 };
 
+enum mes_map_queues_extended_engine_sel_enum {
+	extended_engine_sel__mes_map_queues__legacy_engine_sel = 0,
+	extended_engine_sel__mes_map_queues__sdma0_to_7_sel = 1
+};
 
 struct pm4_mes_map_queues {
 	union {
@@ -269,7 +273,8 @@ struct pm4_mes_map_queues {
 
 	union {
 		struct {
-			uint32_t reserved1:4;
+			uint32_t reserved1:2;
+			enum mes_map_queues_extended_engine_sel_enum extended_engine_sel:2;
 			enum mes_map_queues_queue_sel_enum queue_sel:2;
 			uint32_t reserved5:6;
 			uint32_t gws_control_queue:1;
@@ -382,6 +387,11 @@ enum mes_unmap_queues_engine_sel_enum {
 	engine_sel__mes_unmap_queues__sdmal = 3
 };
 
+enum mes_unmap_queues_extended_engine_sel_enum {
+	extended_engine_sel__mes_unmap_queues__legacy_engine_sel = 0,
+	extended_engine_sel__mes_unmap_queues__sdma0_to_7_sel = 1
+};
+
 struct pm4_mes_unmap_queues {
 	union {
 		union PM4_MES_TYPE_3_HEADER   header;            /* header */
@@ -391,7 +401,7 @@ struct pm4_mes_unmap_queues {
 	union {
 		struct {
 			enum mes_unmap_queues_action_enum action:2;
-			uint32_t reserved1:2;
+			enum mes_unmap_queues_extended_engine_sel_enum extended_engine_sel:2;
 			enum mes_unmap_queues_queue_sel_enum queue_sel:2;
 			uint32_t reserved2:20;
 			enum mes_unmap_queues_engine_sel_enum engine_sel:3;
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#41 lore raw

PATCH 43/102 drm/amdkfd: Support bigger gds size

In reply to a message not in the archive

From: Oak Zeng <[email protected]>

Extend map_process and set_resources pm4 packet to support bigger gds size for arcturus.

v2: Only make the change for v9

---
 .../gpu/drm/amd/amdkfd/kfd_kernel_queue_v9.c  | 33 +++++++++++++++++--
 .../gpu/drm/amd/amdkfd/kfd_pm4_headers_ai.h   | 10 +++---
 2 files changed, 36 insertions(+), 7 deletions(-)
drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue_v9.c +31 -2
index 91da72d0d405..9a4bafb2e175 100644
@@ -81,7 +81,8 @@ static int pm_map_process_v9(struct packet_manager *pm,
 	packet->bitfields2.diq_enable = (qpd->is_debug) ? 1 : 0;
 	packet->bitfields2.process_quantum = 1;
 	packet->bitfields2.pasid = qpd->pqm->process->pasid;
-	packet->bitfields14.gds_size = qpd->gds_size;
+	packet->bitfields14.gds_size = qpd->gds_size & 0x3F;
+	packet->bitfields14.gds_size_hi = (qpd->gds_size >> 6) & 0xF;
 	packet->bitfields14.num_gws = qpd->num_gws;
 	packet->bitfields14.num_oac = qpd->num_oac;
 	packet->bitfields14.sdma_enable = 1;
@@ -143,6 +144,34 @@ static int pm_runlist_v9(struct packet_manager *pm, uint32_t *buffer,
 	return 0;
 }
 
+static int pm_set_resources_v9(struct packet_manager *pm, uint32_t *buffer,
+				struct scheduling_resources *res)
+{
+	struct pm4_mes_set_resources *packet;
+
+	packet = (struct pm4_mes_set_resources *)buffer;
+	memset(buffer, 0, sizeof(struct pm4_mes_set_resources));
+
+	packet->header.u32All = pm_build_pm4_header(IT_SET_RESOURCES,
+					sizeof(struct pm4_mes_set_resources));
+
+	packet->bitfields2.queue_type =
+			queue_type__mes_set_resources__hsa_interface_queue_hiq;
+	packet->bitfields2.vmid_mask = res->vmid_mask;
+	packet->bitfields2.unmap_latency = KFD_UNMAP_LATENCY_MS / 100;
+	packet->bitfields7.oac_mask = res->oac_mask;
+	packet->bitfields8.gds_heap_base = res->gds_heap_base;
+	packet->bitfields8.gds_heap_size = res->gds_heap_size;
+
+	packet->gws_mask_lo = lower_32_bits(res->gws_mask);
+	packet->gws_mask_hi = upper_32_bits(res->gws_mask);
+
+	packet->queue_mask_lo = lower_32_bits(res->queue_mask);
+	packet->queue_mask_hi = upper_32_bits(res->queue_mask);
+
+	return 0;
+}
+
 static int pm_map_queues_v9(struct packet_manager *pm, uint32_t *buffer,
 		struct queue *q, bool is_static)
 {
@@ -344,7 +373,7 @@ static int pm_release_mem_v9(uint64_t gpu_addr, uint32_t *buffer)
 const struct packet_manager_funcs kfd_v9_pm_funcs = {
 	.map_process		= pm_map_process_v9,
 	.runlist		= pm_runlist_v9,
-	.set_resources		= pm_set_resources_vi,
+	.set_resources		= pm_set_resources_v9,
 	.map_queues		= pm_map_queues_v9,
 	.unmap_queues		= pm_unmap_queues_v9,
 	.query_status		= pm_query_status_v9,
drivers/gpu/drm/amd/amdkfd/kfd_pm4_headers_ai.h +5 -5
index 44ed94239513..4d7add843746 100644
@@ -83,10 +83,10 @@ struct pm4_mes_set_resources {
 
 	union {
 		struct {
-		uint32_t gds_heap_base:6;
-		uint32_t reserved3:5;
-		uint32_t gds_heap_size:6;
-		uint32_t reserved4:15;
+		uint32_t gds_heap_base:10;
+		uint32_t reserved3:1;
+		uint32_t gds_heap_size:10;
+		uint32_t reserved4:11;
 		} bitfields8;
 		uint32_t ordinal8;
 	};
@@ -179,7 +179,7 @@ struct pm4_mes_map_process {
 			uint32_t num_gws:7;
 			uint32_t sdma_enable:1;
 			uint32_t num_oac:4;
-			uint32_t reserved8:4;
+			uint32_t gds_size_hi:4;
 			uint32_t gds_size:6;
 			uint32_t num_queues:10;
 		} bitfields14;
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#42 lore raw

PATCH 44/102 amd/amdkfd: Add ASIC ARCTURUS to kfd

In reply to a message not in the archive

From: Yong Zhao <[email protected]>

Add initial support for ARCTURUS to kfd.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c       |  1 +
 drivers/gpu/drm/amd/amdkfd/kfd_crat.c            |  1 +
 drivers/gpu/drm/amd/amdkfd/kfd_device.c          | 16 ++++++++++++++++
 .../drm/amd/amdkfd/kfd_device_queue_manager.c    |  1 +
 drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c     |  1 +
 drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.c    |  1 +
 drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c  |  1 +
 drivers/gpu/drm/amd/amdkfd/kfd_topology.c        |  1 +
 8 files changed, 23 insertions(+)
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c +1 -0
index 658b1654e4b7..d695e0762510 100644
@@ -85,6 +85,7 @@ void amdgpu_amdkfd_device_probe(struct amdgpu_device *adev)
 	case CHIP_VEGA12:
 	case CHIP_VEGA20:
 	case CHIP_RAVEN:
+	case CHIP_ARCTURUS:
 		kfd2kgd = amdgpu_amdkfd_gfx_9_0_get_functions();
 		break;
 	case CHIP_NAVI10:
drivers/gpu/drm/amd/amdkfd/kfd_crat.c +1 -0
index 792371442195..b30ccbfeb648 100644
@@ -662,6 +662,7 @@ static int kfd_fill_gpu_cache_info(struct kfd_dev *kdev,
 	case CHIP_VEGA10:
 	case CHIP_VEGA12:
 	case CHIP_VEGA20:
+	case CHIP_ARCTURUS:
 		pcache_info = vega10_cache_info;
 		num_of_cache_types = ARRAY_SIZE(vega10_cache_info);
 		break;
drivers/gpu/drm/amd/amdkfd/kfd_device.c +16 -0
index 3322a443dfb2..81602fb89d87 100644
@@ -317,6 +317,22 @@ static const struct kfd_device_info vega20_device_info = {
 	.num_sdma_queues_per_engine = 8,
 };
 
+static const struct kfd_device_info arcturus_device_info = {
+	.asic_family = CHIP_ARCTURUS,
+	.max_pasid_bits = 16,
+	.max_no_of_hqd	= 24,
+	.doorbell_size	= 8,
+	.ih_ring_entry_size = 8 * sizeof(uint32_t),
+	.event_interrupt_class = &event_interrupt_class_v9,
+	.num_of_watch_points = 4,
+	.mqd_size_aligned = MQD_SIZE_ALIGNED,
+	.supports_cwsr = true,
+	.needs_iommu_device = false,
+	.needs_pci_atomics = false,
+	.num_sdma_engines = 2,
+	.num_sdma_queues_per_engine = 8,
+};
+
 static const struct kfd_device_info navi10_device_info = {
 	.asic_family = CHIP_NAVI10,
 	.max_pasid_bits = 16,
drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c +1 -0
index f23e17bc5222..587362bb912b 100644
@@ -1786,6 +1786,7 @@ struct device_queue_manager *device_queue_manager_init(struct kfd_dev *dev)
 	case CHIP_VEGA12:
 	case CHIP_VEGA20:
 	case CHIP_RAVEN:
+	case CHIP_ARCTURUS:
 		device_queue_manager_init_v9(&dqm->asic_ops);
 		break;
 	case CHIP_NAVI10:
drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c +1 -0
index 60521366dd31..9dc4bff8085e 100644
@@ -405,6 +405,7 @@ int kfd_init_apertures(struct kfd_process *process)
 			case CHIP_VEGA12:
 			case CHIP_VEGA20:
 			case CHIP_RAVEN:
+			case CHIP_ARCTURUS:
 			case CHIP_NAVI10:
 				kfd_init_apertures_v9(pdd, id);
 				break;
drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.c +1 -0
index 29c0bd2d7a5c..8b4564f71a7a 100644
@@ -330,6 +330,7 @@ struct kernel_queue *kernel_queue_init(struct kfd_dev *dev,
 	case CHIP_VEGA12:
 	case CHIP_VEGA20:
 	case CHIP_RAVEN:
+	case CHIP_ARCTURUS:
 		kernel_queue_init_v9(&kq->ops_asic_specific);
 		break;
 	case CHIP_NAVI10:
drivers/gpu/drm/amd/amdkfd/kfd_packet_manager.c +1 -0
index ccf6b2310316..2c8624c5b42c 100644
@@ -239,6 +239,7 @@ int pm_init(struct packet_manager *pm, struct device_queue_manager *dqm)
 	case CHIP_VEGA12:
 	case CHIP_VEGA20:
 	case CHIP_RAVEN:
+	case CHIP_ARCTURUS:
 		pm->pmf = &kfd_v9_pm_funcs;
 		break;
 	case CHIP_NAVI10:
drivers/gpu/drm/amd/amdkfd/kfd_topology.c +1 -0
index c2e6e47abaf2..36fa98fe858b 100644
@@ -1321,6 +1321,7 @@ int kfd_topology_add_device(struct kfd_dev *gpu)
 	case CHIP_VEGA12:
 	case CHIP_VEGA20:
 	case CHIP_RAVEN:
+	case CHIP_ARCTURUS:
 	case CHIP_NAVI10:
 		dev->node_props.capability |= ((HSA_CAP_DOORBELL_TYPE_2_0 <<
 			HSA_CAP_DOORBELL_TYPE_TOTALBITS_SHIFT) &
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#43 lore raw

PATCH 45/102 drm/amdkfd: Expose function mmhub_v9_4_setup_vm_pt_regs() for kfd to use

In reply to a message not in the archive

From: Yong Zhao <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/gmc_v9_0.h   |  7 +++++++
 drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c | 17 ++++++++++++-----
 2 files changed, 19 insertions(+), 5 deletions(-)
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.h +7 -0
index 5c8deac65580..971c0840358f 100644
@@ -37,4 +37,11 @@
 extern const struct amd_ip_funcs gmc_v9_0_ip_funcs;
 extern const struct amdgpu_ip_block_version gmc_v9_0_ip_block;
 
+/* amdgpu_amdkfd*.c */
+void gfxhub_v1_0_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid,
+				uint64_t value);
+void mmhub_v1_0_setup_vm_pt_regs(struct amdgpu_device *adev, uint32_t vmid,
+				uint64_t value);
+void mmhub_v9_4_setup_vm_pt_regs(struct amdgpu_device *adev, int hubid,
+				uint32_t vmid, uint64_t value);
 #endif
drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c +12 -5
index aa9b43b6ba6b..7f79fa4b6ad1 100644
@@ -46,18 +46,23 @@ u64 mmhub_v9_4_get_fb_location(struct amdgpu_device *adev)
 	return base;
 }
 
-static void mmhub_v9_4_init_gart_pt_regs(struct amdgpu_device *adev, int hubid)
+void mmhub_v9_4_setup_vm_pt_regs(struct amdgpu_device *adev, int hubid,
+				uint32_t vmid, uint64_t value)
 {
-	uint64_t value = amdgpu_gmc_pd_addr(adev->gart.bo);
+	/* two registers distance between mmVML2VC0_VM_CONTEXT0_* to
+	 * mmVML2VC0_VM_CONTEXT1_*
+	 */
+	int dist = mmVML2VC0_VM_CONTEXT1_PAGE_TABLE_BASE_ADDR_LO32
+			- mmVML2VC0_VM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32;
 
 	WREG32_SOC15_OFFSET(MMHUB, 0,
 			    mmVML2VC0_VM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32,
-			    hubid * MMHUB_INSTANCE_REGISTER_OFFSET,
+			    dist * vmid + hubid * MMHUB_INSTANCE_REGISTER_OFFSET,
 			    lower_32_bits(value));
 
 	WREG32_SOC15_OFFSET(MMHUB, 0,
 			    mmVML2VC0_VM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32,
-			    hubid * MMHUB_INSTANCE_REGISTER_OFFSET,
+			    dist * vmid + hubid * MMHUB_INSTANCE_REGISTER_OFFSET,
 			    upper_32_bits(value));
 
 }
@@ -65,7 +70,9 @@ static void mmhub_v9_4_init_gart_pt_regs(struct amdgpu_device *adev, int hubid)
 static void mmhub_v9_4_init_gart_aperture_regs(struct amdgpu_device *adev,
 					       int hubid)
 {
-	mmhub_v9_4_init_gart_pt_regs(adev, hubid);
+	uint64_t pt_base = amdgpu_gmc_pd_addr(adev->gart.bo);
+
+	mmhub_v9_4_setup_vm_pt_regs(adev, hubid, 0, pt_base);
 
 	WREG32_SOC15_OFFSET(MMHUB, 0,
 			    mmVML2VC0_VM_CONTEXT0_PAGE_TABLE_START_ADDR_LO32,
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#44 lore raw

PATCH 46/102 drm/amdkfd: Support two MMHUBs when setting up page table base in KFD

In reply to a message not in the archive

From: Yong Zhao <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c +7 -1
index d5af41143d12..f53140223c77 100644
@@ -47,6 +47,7 @@
 #include "soc15d.h"
 #include "mmhub_v1_0.h"
 #include "gfxhub_v1_0.h"
+#include "gmc_v9_0.h"
 
 
 #define V9_PIPE_PER_MEC		(4)
@@ -884,7 +885,12 @@ static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid,
 	 * now, all processes share the same address space size, like
 	 * on GFX8 and older.
 	 */
-	mmhub_v1_0_setup_vm_pt_regs(adev, vmid, page_table_base);
+	if (adev->asic_type == CHIP_ARCTURUS) {
+		/* Two MMHUBs */
+		mmhub_v9_4_setup_vm_pt_regs(adev, 0, vmid, page_table_base);
+		mmhub_v9_4_setup_vm_pt_regs(adev, 1, vmid, page_table_base);
+	} else
+		mmhub_v1_0_setup_vm_pt_regs(adev, vmid, page_table_base);
 
 	gfxhub_v1_0_setup_vm_pt_regs(adev, vmid, page_table_base);
 }
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#45 lore raw

PATCH 47/102 drm/amdkfd: Change arcturus sdma engines number

In reply to a message not in the archive

From: Oak Zeng <[email protected]>

Arcturus has 8 sdma engines

---
 drivers/gpu/drm/amd/amdkfd/kfd_device.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/amdkfd/kfd_device.c +1 -1
index 81602fb89d87..511bc2523f2f 100644
@@ -329,7 +329,7 @@ static const struct kfd_device_info arcturus_device_info = {
 	.supports_cwsr = true,
 	.needs_iommu_device = false,
 	.needs_pci_atomics = false,
-	.num_sdma_engines = 2,
+	.num_sdma_engines = 8,
 	.num_sdma_queues_per_engine = 8,
 };
 
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#46 lore raw

PATCH 48/102 drm/amdkfd: Fix sdma_bitmap overflow issue

In reply to a message not in the archive

From: Oak Zeng <[email protected]>

In the original formula, when sdma queue number is 64, the left shift overflows. Use an equivalence that won't overflow.

---
 drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)
drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c +4 -4
index 587362bb912b..97ef55282c35 100644
@@ -880,8 +880,8 @@ static int initialize_nocpsch(struct device_queue_manager *dqm)
 	}
 
 	dqm->vmid_bitmap = (1 << dqm->dev->vm_info.vmid_num_kfd) - 1;
-	dqm->sdma_bitmap = (1ULL << get_num_sdma_queues(dqm)) - 1;
-	dqm->xgmi_sdma_bitmap = (1ULL << get_num_xgmi_sdma_queues(dqm)) - 1;
+	dqm->sdma_bitmap = ~0ULL >> (64 - get_num_sdma_queues(dqm));
+	dqm->xgmi_sdma_bitmap = ~0ULL >> (64 - get_num_xgmi_sdma_queues(dqm));
 
 	return 0;
 }
@@ -1019,8 +1019,8 @@ static int initialize_cpsch(struct device_queue_manager *dqm)
 	dqm->sdma_queue_count = 0;
 	dqm->xgmi_sdma_queue_count = 0;
 	dqm->active_runlist = false;
-	dqm->sdma_bitmap = (1ULL << get_num_sdma_queues(dqm)) - 1;
-	dqm->xgmi_sdma_bitmap = (1ULL << get_num_xgmi_sdma_queues(dqm)) - 1;
+	dqm->sdma_bitmap = ~0ULL >> (64 - get_num_sdma_queues(dqm));
+	dqm->xgmi_sdma_bitmap = ~0ULL >> (64 - get_num_xgmi_sdma_queues(dqm));
 
 	INIT_WORK(&dqm->hw_exception_work, kfd_process_hw_exception);
 
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#47 lore raw

PATCH 49/102 drm/amdkfd: Implement kfd2kgd_calls for Arcturus

In reply to a message not in the archive

From: Oak Zeng <[email protected]>

Arcturus shares most of the kfd2kgd_calls with gfx9. But due to SDMA register address change, it can't share SDMA related functions. Export gfx9 kfd2kgd_calls and implement SDMA related functions for Arcturus.

---
 drivers/gpu/drm/amd/amdgpu/Makefile           |   1 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c    |   4 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h    |   1 +
 .../drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c   | 324 ++++++++++++++++++
 .../gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c | 161 +++------
 .../gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h |  83 +++++
 6 files changed, 465 insertions(+), 109 deletions(-)
 create mode 100644 drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c
 create mode 100644 drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h
drivers/gpu/drm/amd/amdgpu/Makefile +1 -0
index 97671df6a5f0..dc22569e9732 100644
@@ -163,6 +163,7 @@ amdgpu-y += \
 	 amdgpu_amdkfd_gpuvm.o \
 	 amdgpu_amdkfd_gfx_v8.o \
 	 amdgpu_amdkfd_gfx_v9.o \
+	 amdgpu_amdkfd_arcturus.o \
 	 amdgpu_amdkfd_gfx_v10.o
 
 ifneq ($(CONFIG_DRM_AMDGPU_CIK),)
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c +3 -1
index d695e0762510..702d05ce1fa5 100644
@@ -85,9 +85,11 @@ void amdgpu_amdkfd_device_probe(struct amdgpu_device *adev)
 	case CHIP_VEGA12:
 	case CHIP_VEGA20:
 	case CHIP_RAVEN:
-	case CHIP_ARCTURUS:
 		kfd2kgd = amdgpu_amdkfd_gfx_9_0_get_functions();
 		break;
+	case CHIP_ARCTURUS:
+		kfd2kgd = amdgpu_amdkfd_arcturus_get_functions();
+		break;
 	case CHIP_NAVI10:
 		kfd2kgd = amdgpu_amdkfd_gfx_10_0_get_functions();
 		break;
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h +1 -0
index b6076d19e442..e519df3fd2b6 100644
@@ -140,6 +140,7 @@ bool amdgpu_amdkfd_have_atomics_support(struct kgd_dev *kgd);
 struct kfd2kgd_calls *amdgpu_amdkfd_gfx_7_get_functions(void);
 struct kfd2kgd_calls *amdgpu_amdkfd_gfx_8_0_get_functions(void);
 struct kfd2kgd_calls *amdgpu_amdkfd_gfx_9_0_get_functions(void);
+struct kfd2kgd_calls *amdgpu_amdkfd_arcturus_get_functions(void);
 struct kfd2kgd_calls *amdgpu_amdkfd_gfx_10_0_get_functions(void);
 
 bool amdgpu_amdkfd_is_kfd_vmid(struct amdgpu_device *adev, u32 vmid);
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c new file +324 -0

324 lines not on this page · raw message

drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c +53 -108

286 lines not on this page · raw message

drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.h new file +83 -0

83 lines not on this page · raw message

Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#48 lore raw

PATCH 50/102 drm/amdgpu: Set VM_L2_CNTL.PDE_FAULT_CLASSIFICATION to 0 for MMHUB 9.4

In reply to a message not in the archive

From: Yong Zhao <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c +1 -1
index 7f79fa4b6ad1..6b7cdaadbd70 100644
@@ -195,7 +195,7 @@ static void mmhub_v9_4_init_cache_regs(struct amdgpu_device *adev, int hubid)
 	tmp = REG_SET_FIELD(tmp, VML2PF0_VM_L2_CNTL,
 			    L2_PDE0_CACHE_TAG_GENERATION_MODE, 0);
 	tmp = REG_SET_FIELD(tmp, VML2PF0_VM_L2_CNTL,
-			    PDE_FAULT_CLASSIFICATION, 1);
+			    PDE_FAULT_CLASSIFICATION, 0);
 	tmp = REG_SET_FIELD(tmp, VML2PF0_VM_L2_CNTL,
 			    CONTEXT1_IDENTITY_ACCESS_MODE, 1);
 	tmp = REG_SET_FIELD(tmp, VML2PF0_VM_L2_CNTL,
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#49 lore raw

PATCH 51/102 drm/amdkfd: Support MMHUB1 in kfd interrupt path

In reply to a message not in the archive

From: Yong Zhao <[email protected]>

---
 drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c +2 -0
index a85904ad0d5f..3ef67d2e0d9f 100644
@@ -80,6 +80,7 @@ static bool event_interrupt_isr_v9(struct kfd_dev *dev,
 		source_id == SOC15_INTSRC_SQ_INTERRUPT_MSG ||
 		source_id == SOC15_INTSRC_CP_BAD_OPCODE ||
 		client_id == SOC15_IH_CLIENTID_VMC ||
+		client_id == SOC15_IH_CLIENTID_VMC1 ||
 		client_id == SOC15_IH_CLIENTID_UTCL2;
 }
 
@@ -104,6 +105,7 @@ static void event_interrupt_wq_v9(struct kfd_dev *dev,
 	else if (source_id == SOC15_INTSRC_CP_BAD_OPCODE)
 		kfd_signal_hw_exception_event(pasid);
 	else if (client_id == SOC15_IH_CLIENTID_VMC ||
+		client_id == SOC15_IH_CLIENTID_VMC1 ||
 		 client_id == SOC15_IH_CLIENTID_UTCL2) {
 		struct kfd_vm_fault_info info = {0};
 		uint16_t ring_id = SOC15_RING_ID_FROM_IH_ENTRY(ih_ring_entry);
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#50 lore raw

PATCH 52/102 drm/amdgpu: Hack xgmi topology info when there is no psp fw

In reply to a message not in the archive

From: Oak Zeng <[email protected]>

This is only needed on emulation platform where psp fw might not be available, to hack xgmi topology info such as hive id and node id.

v2: Add offset to hacked hive/node id
v3: Don't use introduce new module parameter.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c | 27 ++++++++++++++----------
 1 file changed, 16 insertions(+), 11 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c +16 -11
index d11eba09eadd..b024070ad1cf 100644
@@ -301,18 +301,23 @@ int amdgpu_xgmi_add_device(struct amdgpu_device *adev)
 	if (!adev->gmc.xgmi.supported)
 		return 0;
 
-	ret = psp_xgmi_get_node_id(&adev->psp, &adev->gmc.xgmi.node_id);
-	if (ret) {
-		dev_err(adev->dev,
-			"XGMI: Failed to get node id\n");
-		return ret;
-	}
+	if (amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_PSP)) {
+		ret = psp_xgmi_get_hive_id(&adev->psp, &adev->gmc.xgmi.hive_id);
+		if (ret) {
+			dev_err(adev->dev,
+				"XGMI: Failed to get hive id\n");
+			return ret;
+		}
 
-	ret = psp_xgmi_get_hive_id(&adev->psp, &adev->gmc.xgmi.hive_id);
-	if (ret) {
-		dev_err(adev->dev,
-			"XGMI: Failed to get hive id\n");
-		return ret;
+		ret = psp_xgmi_get_node_id(&adev->psp, &adev->gmc.xgmi.node_id);
+		if (ret) {
+			dev_err(adev->dev,
+				"XGMI: Failed to get node id\n");
+			return ret;
+		}
+	} else {
+		adev->gmc.xgmi.hive_id = 16;
+		adev->gmc.xgmi.node_id = adev->gmc.xgmi.physical_node_id + 16;
 	}
 
 	hive = amdgpu_get_xgmi_hive(adev, 1);
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#51 lore raw

PATCH 53/102 drm/amdgpu: Enable xgmi support for Arcturus

In reply to a message not in the archive

From: Oak Zeng <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/soc15.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/soc15.c +1 -1
index 8c881697d29b..fb4c7aac8d9e 100644
@@ -592,7 +592,7 @@ int soc15_set_ip_blocks(struct amdgpu_device *adev)
 		return -EINVAL;
 	}
 
-	if (adev->asic_type == CHIP_VEGA20)
+	if (adev->asic_type == CHIP_VEGA20 || adev->asic_type == CHIP_ARCTURUS)
 		adev->gmc.xgmi.supported = true;
 
 	if (adev->flags & AMD_IS_APU)
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#52 lore raw

PATCH 54/102 drm/amdgpu: Clear build undefined warning

In reply to a message not in the archive

From: James Zhu <[email protected]>

Add amdgpu_amdkfd_arcturus_get_functions stub when CONFIG_HSA_AMD is undefinded.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c | 5 +++++
 1 file changed, 5 insertions(+)
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c +5 -0
index 702d05ce1fa5..f052c70e4659 100644
@@ -718,6 +718,11 @@ struct kfd2kgd_calls *amdgpu_amdkfd_gfx_9_0_get_functions(void)
 	return NULL;
 }
 
+struct kfd2kgd_calls *amdgpu_amdkfd_arcturus_get_functions(void)
+{
+	return NULL;
+}
+
 struct kfd2kgd_calls *amdgpu_amdkfd_gfx_10_0_get_functions(void)
 {
 	return NULL;
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#53 lore raw

PATCH 55/102 drm/amdgpu: declare sdma firmware binary files for Arcturus

In reply to a message not in the archive

From: Le Ma <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c | 8 ++++++++
 1 file changed, 8 insertions(+)
drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c +8 -0
index 594750a9f2fc..cd142e4304a8 100644
@@ -52,6 +52,14 @@ MODULE_FIRMWARE("amdgpu/vega20_sdma1.bin");
 MODULE_FIRMWARE("amdgpu/raven_sdma.bin");
 MODULE_FIRMWARE("amdgpu/picasso_sdma.bin");
 MODULE_FIRMWARE("amdgpu/raven2_sdma.bin");
+MODULE_FIRMWARE("amdgpu/arcturus_sdma.bin");
+MODULE_FIRMWARE("amdgpu/arcturus_sdma1.bin");
+MODULE_FIRMWARE("amdgpu/arcturus_sdma2.bin");
+MODULE_FIRMWARE("amdgpu/arcturus_sdma3.bin");
+MODULE_FIRMWARE("amdgpu/arcturus_sdma4.bin");
+MODULE_FIRMWARE("amdgpu/arcturus_sdma5.bin");
+MODULE_FIRMWARE("amdgpu/arcturus_sdma6.bin");
+MODULE_FIRMWARE("amdgpu/arcturus_sdma7.bin");
 
 #define SDMA0_POWER_CNTL__ON_OFF_CONDITION_HOLD_TIME_MASK  0x000000F8L
 #define SDMA0_POWER_CNTL__ON_OFF_STATUS_DURATION_TIME_MASK 0xFC000000L
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#54 lore raw

PATCH 56/102 drm/amdgpu/VCN2: put IB internal registers offset to structure

In reply to a message not in the archive

From: Leo Liu <[email protected]>

So the ring functions can be shared with different VCN versions with different internal registers offsets

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h |  6 +++
 drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c   | 64 ++++++++++++++++---------
 2 files changed, 47 insertions(+), 23 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h +6 -0
index 99f14fcc1460..bfd8c3cea13a 100644
@@ -145,6 +145,12 @@ struct amdgpu_vcn_reg{
 	unsigned	data1;
 	unsigned	cmd;
 	unsigned	nop;
+	unsigned	context_id;
+	unsigned	ib_vmid;
+	unsigned	ib_bar_low;
+	unsigned	ib_bar_high;
+	unsigned	ib_size;
+	unsigned	gp_scratch8;
 	unsigned	scratch9;
 	unsigned	jpeg_pitch;
 };
drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c +41 -23

147 lines not on this page · raw message

Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#55 lore raw

PATCH 57/102 drm/amdgpu/VCN2: expose rings functions

In reply to a message not in the archive

From: Leo Liu <[email protected]>

They can be reused by VCN2.x family

---
 drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c | 87 +++++++++++++--------------
 drivers/gpu/drm/amd/amdgpu/vcn_v2_0.h | 38 ++++++++++++
 2 files changed, 79 insertions(+), 46 deletions(-)
drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c +41 -46
index c801b69f0b75..52a5e3ac9f55 100644
@@ -1490,7 +1490,7 @@ static void vcn_v2_0_dec_ring_set_wptr(struct amdgpu_ring *ring)
  *
  * Write a start command to the ring.
  */
-static void vcn_v2_0_dec_ring_insert_start(struct amdgpu_ring *ring)
+void vcn_v2_0_dec_ring_insert_start(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
 
@@ -1507,7 +1507,7 @@ static void vcn_v2_0_dec_ring_insert_start(struct amdgpu_ring *ring)
  *
  * Write a end command to the ring.
  */
-static void vcn_v2_0_dec_ring_insert_end(struct amdgpu_ring *ring)
+void vcn_v2_0_dec_ring_insert_end(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
 
@@ -1522,7 +1522,7 @@ static void vcn_v2_0_dec_ring_insert_end(struct amdgpu_ring *ring)
  *
  * Write a nop command to the ring.
  */
-static void vcn_v2_0_dec_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count)
+void vcn_v2_0_dec_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count)
 {
 	struct amdgpu_device *adev = ring->adev;
 	int i;
@@ -1543,8 +1543,8 @@ static void vcn_v2_0_dec_ring_insert_nop(struct amdgpu_ring *ring, uint32_t coun
  *
  * Write a fence and a trap command to the ring.
  */
-static void vcn_v2_0_dec_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq,
-				     unsigned flags)
+void vcn_v2_0_dec_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq,
+				unsigned flags)
 {
 	struct amdgpu_device *adev = ring->adev;
 
@@ -1580,10 +1580,10 @@ static void vcn_v2_0_dec_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64
  *
  * Write ring commands to execute the indirect buffer
  */
-static void vcn_v2_0_dec_ring_emit_ib(struct amdgpu_ring *ring,
-				      struct amdgpu_job *job,
-				      struct amdgpu_ib *ib,
-				      uint32_t flags)
+void vcn_v2_0_dec_ring_emit_ib(struct amdgpu_ring *ring,
+			       struct amdgpu_job *job,
+			       struct amdgpu_ib *ib,
+			       uint32_t flags)
 {
 	struct amdgpu_device *adev = ring->adev;
 	unsigned vmid = AMDGPU_JOB_GET_VMID(job);
@@ -1599,9 +1599,8 @@ static void vcn_v2_0_dec_ring_emit_ib(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, ib->length_dw);
 }
 
-static void vcn_v2_0_dec_ring_emit_reg_wait(struct amdgpu_ring *ring,
-					    uint32_t reg, uint32_t val,
-					    uint32_t mask)
+void vcn_v2_0_dec_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
+				uint32_t val, uint32_t mask)
 {
 	struct amdgpu_device *adev = ring->adev;
 
@@ -1619,8 +1618,8 @@ static void vcn_v2_0_dec_ring_emit_reg_wait(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, VCN_DEC_CMD_REG_READ_COND_WAIT << 1);
 }
 
-static void vcn_v2_0_dec_ring_emit_vm_flush(struct amdgpu_ring *ring,
-					    unsigned vmid, uint64_t pd_addr)
+void vcn_v2_0_dec_ring_emit_vm_flush(struct amdgpu_ring *ring,
+				unsigned vmid, uint64_t pd_addr)
 {
 	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
 	uint32_t data0, data1, mask;
@@ -1634,8 +1633,8 @@ static void vcn_v2_0_dec_ring_emit_vm_flush(struct amdgpu_ring *ring,
 	vcn_v2_0_dec_ring_emit_reg_wait(ring, data0, data1, mask);
 }
 
-static void vcn_v2_0_dec_ring_emit_wreg(struct amdgpu_ring *ring,
-					uint32_t reg, uint32_t val)
+void vcn_v2_0_dec_ring_emit_wreg(struct amdgpu_ring *ring,
+				uint32_t reg, uint32_t val)
 {
 	struct amdgpu_device *adev = ring->adev;
 
@@ -1727,8 +1726,8 @@ static void vcn_v2_0_enc_ring_set_wptr(struct amdgpu_ring *ring)
  *
  * Write enc a fence and a trap command to the ring.
  */
-static void vcn_v2_0_enc_ring_emit_fence(struct amdgpu_ring *ring, u64 addr,
-			u64 seq, unsigned flags)
+void vcn_v2_0_enc_ring_emit_fence(struct amdgpu_ring *ring, u64 addr,
+				u64 seq, unsigned flags)
 {
 	WARN_ON(flags & AMDGPU_FENCE_FLAG_64BIT);
 
@@ -1739,7 +1738,7 @@ static void vcn_v2_0_enc_ring_emit_fence(struct amdgpu_ring *ring, u64 addr,
 	amdgpu_ring_write(ring, VCN_ENC_CMD_TRAP);
 }
 
-static void vcn_v2_0_enc_ring_insert_end(struct amdgpu_ring *ring)
+void vcn_v2_0_enc_ring_insert_end(struct amdgpu_ring *ring)
 {
 	amdgpu_ring_write(ring, VCN_ENC_CMD_END);
 }
@@ -1752,10 +1751,10 @@ static void vcn_v2_0_enc_ring_insert_end(struct amdgpu_ring *ring)
  *
  * Write enc ring commands to execute the indirect buffer
  */
-static void vcn_v2_0_enc_ring_emit_ib(struct amdgpu_ring *ring,
-				      struct amdgpu_job *job,
-				      struct amdgpu_ib *ib,
-				      uint32_t flags)
+void vcn_v2_0_enc_ring_emit_ib(struct amdgpu_ring *ring,
+			       struct amdgpu_job *job,
+			       struct amdgpu_ib *ib,
+			       uint32_t flags)
 {
 	unsigned vmid = AMDGPU_JOB_GET_VMID(job);
 
@@ -1766,9 +1765,8 @@ static void vcn_v2_0_enc_ring_emit_ib(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, ib->length_dw);
 }
 
-static void vcn_v2_0_enc_ring_emit_reg_wait(struct amdgpu_ring *ring,
-					    uint32_t reg, uint32_t val,
-					    uint32_t mask)
+void vcn_v2_0_enc_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
+				uint32_t val, uint32_t mask)
 {
 	amdgpu_ring_write(ring, VCN_ENC_CMD_REG_WAIT);
 	amdgpu_ring_write(ring, reg << 2);
@@ -1776,8 +1774,8 @@ static void vcn_v2_0_enc_ring_emit_reg_wait(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, val);
 }
 
-static void vcn_v2_0_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
-					    unsigned int vmid, uint64_t pd_addr)
+void vcn_v2_0_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
+				unsigned int vmid, uint64_t pd_addr)
 {
 	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
 
@@ -1788,8 +1786,7 @@ static void vcn_v2_0_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
 					lower_32_bits(pd_addr), 0xffffffff);
 }
 
-static void vcn_v2_0_enc_ring_emit_wreg(struct amdgpu_ring *ring,
-					uint32_t reg, uint32_t val)
+void vcn_v2_0_enc_ring_emit_wreg(struct amdgpu_ring *ring, uint32_t reg, uint32_t val)
 {
 	amdgpu_ring_write(ring, VCN_ENC_CMD_REG_WRITE);
 	amdgpu_ring_write(ring,	reg << 2);
@@ -1853,7 +1850,7 @@ static void vcn_v2_0_jpeg_ring_set_wptr(struct amdgpu_ring *ring)
  *
  * Write a start command to the ring.
  */
-static void vcn_v2_0_jpeg_ring_insert_start(struct amdgpu_ring *ring)
+void vcn_v2_0_jpeg_ring_insert_start(struct amdgpu_ring *ring)
 {
 	amdgpu_ring_write(ring, PACKETJ(mmUVD_JRBC_EXTERNAL_REG_INTERNAL_OFFSET,
 		0, 0, PACKETJ_TYPE0));
@@ -1871,7 +1868,7 @@ static void vcn_v2_0_jpeg_ring_insert_start(struct amdgpu_ring *ring)
  *
  * Write a end command to the ring.
  */
-static void vcn_v2_0_jpeg_ring_insert_end(struct amdgpu_ring *ring)
+void vcn_v2_0_jpeg_ring_insert_end(struct amdgpu_ring *ring)
 {
 	amdgpu_ring_write(ring, PACKETJ(mmUVD_JRBC_EXTERNAL_REG_INTERNAL_OFFSET,
 		0, 0, PACKETJ_TYPE0));
@@ -1890,8 +1887,8 @@ static void vcn_v2_0_jpeg_ring_insert_end(struct amdgpu_ring *ring)
  *
  * Write a fence and a trap command to the ring.
  */
-static void vcn_v2_0_jpeg_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq,
-				     unsigned flags)
+void vcn_v2_0_jpeg_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq,
+				unsigned flags)
 {
 	WARN_ON(flags & AMDGPU_FENCE_FLAG_64BIT);
 
@@ -1939,10 +1936,10 @@ static void vcn_v2_0_jpeg_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u6
  *
  * Write ring commands to execute the indirect buffer.
  */
-static void vcn_v2_0_jpeg_ring_emit_ib(struct amdgpu_ring *ring,
-				       struct amdgpu_job *job,
-				       struct amdgpu_ib *ib,
-				       uint32_t flags)
+void vcn_v2_0_jpeg_ring_emit_ib(struct amdgpu_ring *ring,
+				struct amdgpu_job *job,
+				struct amdgpu_ib *ib,
+				uint32_t flags)
 {
 	unsigned vmid = AMDGPU_JOB_GET_VMID(job);
 
@@ -1990,9 +1987,8 @@ static void vcn_v2_0_jpeg_ring_emit_ib(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, 0x2);
 }
 
-static void vcn_v2_0_jpeg_ring_emit_reg_wait(struct amdgpu_ring *ring,
-					    uint32_t reg, uint32_t val,
-					    uint32_t mask)
+void vcn_v2_0_jpeg_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
+				uint32_t val, uint32_t mask)
 {
 	uint32_t reg_offset = (reg << 2);
 
@@ -2018,8 +2014,8 @@ static void vcn_v2_0_jpeg_ring_emit_reg_wait(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, mask);
 }
 
-static void vcn_v2_0_jpeg_ring_emit_vm_flush(struct amdgpu_ring *ring,
-		unsigned vmid, uint64_t pd_addr)
+void vcn_v2_0_jpeg_ring_emit_vm_flush(struct amdgpu_ring *ring,
+				unsigned vmid, uint64_t pd_addr)
 {
 	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
 	uint32_t data0, data1, mask;
@@ -2033,8 +2029,7 @@ static void vcn_v2_0_jpeg_ring_emit_vm_flush(struct amdgpu_ring *ring,
 	vcn_v2_0_jpeg_ring_emit_reg_wait(ring, data0, data1, mask);
 }
 
-static void vcn_v2_0_jpeg_ring_emit_wreg(struct amdgpu_ring *ring,
-					uint32_t reg, uint32_t val)
+void vcn_v2_0_jpeg_ring_emit_wreg(struct amdgpu_ring *ring, uint32_t reg, uint32_t val)
 {
 	uint32_t reg_offset = (reg << 2);
 
@@ -2052,7 +2047,7 @@ static void vcn_v2_0_jpeg_ring_emit_wreg(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, val);
 }
 
-static void vcn_v2_0_jpeg_ring_nop(struct amdgpu_ring *ring, uint32_t count)
+void vcn_v2_0_jpeg_ring_nop(struct amdgpu_ring *ring, uint32_t count)
 {
 	int i;
 
drivers/gpu/drm/amd/amdgpu/vcn_v2_0.h +38 -0
index a74227f4663b..8467292f32e5 100644
@@ -24,6 +24,44 @@
 #ifndef __VCN_V2_0_H__
 #define __VCN_V2_0_H__
 
+extern void vcn_v2_0_dec_ring_insert_start(struct amdgpu_ring *ring);
+extern void vcn_v2_0_dec_ring_insert_end(struct amdgpu_ring *ring);
+extern void vcn_v2_0_dec_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count);
+extern void vcn_v2_0_dec_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq,
+				unsigned flags);
+extern void vcn_v2_0_dec_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_job *job,
+				struct amdgpu_ib *ib, uint32_t flags);
+extern void vcn_v2_0_dec_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
+				uint32_t val, uint32_t mask);
+extern void vcn_v2_0_dec_ring_emit_vm_flush(struct amdgpu_ring *ring,
+				unsigned vmid, uint64_t pd_addr);
+extern void vcn_v2_0_dec_ring_emit_wreg(struct amdgpu_ring *ring,
+				uint32_t reg, uint32_t val);
+
+extern void vcn_v2_0_enc_ring_insert_end(struct amdgpu_ring *ring);
+extern void vcn_v2_0_enc_ring_emit_fence(struct amdgpu_ring *ring, u64 addr,
+				u64 seq, unsigned flags);
+extern void vcn_v2_0_enc_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_job *job,
+				struct amdgpu_ib *ib, uint32_t flags);
+extern void vcn_v2_0_enc_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
+				uint32_t val, uint32_t mask);
+extern void vcn_v2_0_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
+				unsigned int vmid, uint64_t pd_addr);
+extern void vcn_v2_0_enc_ring_emit_wreg(struct amdgpu_ring *ring, uint32_t reg, uint32_t val);
+
+extern void vcn_v2_0_jpeg_ring_insert_start(struct amdgpu_ring *ring);
+extern void vcn_v2_0_jpeg_ring_insert_end(struct amdgpu_ring *ring);
+extern void vcn_v2_0_jpeg_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq,
+				unsigned flags);
+extern void vcn_v2_0_jpeg_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_job *job,
+				struct amdgpu_ib *ib, uint32_t flags);
+extern void vcn_v2_0_jpeg_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
+				uint32_t val, uint32_t mask);
+extern void vcn_v2_0_jpeg_ring_emit_vm_flush(struct amdgpu_ring *ring,
+				unsigned vmid, uint64_t pd_addr);
+extern void vcn_v2_0_jpeg_ring_emit_wreg(struct amdgpu_ring *ring, uint32_t reg, uint32_t val);
+extern void vcn_v2_0_jpeg_ring_nop(struct amdgpu_ring *ring, uint32_t count);
+
 extern const struct amdgpu_ip_block_version vcn_v2_0_ip_block;
 
 #endif /* __VCN_V2_0_H__ */
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#56 lore raw

PATCH 58/102 drm/amdgpu: add VCN2.5 basic supports

In reply to a message not in the archive

From: Leo Liu <[email protected]>

i.e. basic VCN IP SW structures

---
 drivers/gpu/drm/amd/amdgpu/Makefile   |   3 +-
 drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c | 414 ++++++++++++++++++++++++++
 2 files changed, 416 insertions(+), 1 deletion(-)
 create mode 100644 drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c
drivers/gpu/drm/amd/amdgpu/Makefile +2 -1
index dc22569e9732..7a1a78c7b329 100644
@@ -145,7 +145,8 @@ amdgpu-y += \
 amdgpu-y += \
 	amdgpu_vcn.o \
 	vcn_v1_0.o \
-	vcn_v2_0.o
+	vcn_v2_0.o \
+	vcn_v2_5.o
 
 # add ATHUB block
 amdgpu-y += \
drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c new file +414 -0

414 lines not on this page · raw message

Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#57 lore raw

PATCH 59/102 drm/amdgpu: add VCN2.5 VCPU start and stop

In reply to a message not in the archive

From: Leo Liu <[email protected]>

HW engine initialization and suspend sequences.

---
 drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c | 467 +++++++++++++++++++++++++-
 1 file changed, 466 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c +466 -1
index 0f553563ceb9..b6e72fff94f5 100644
@@ -307,6 +307,446 @@ static int vcn_v2_5_resume(void *handle)
 	return r;
 }
 
+/**
+ * vcn_v2_5_mc_resume - memory controller programming
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Let the VCN memory controller know it's offsets
+ */
+static void vcn_v2_5_mc_resume(struct amdgpu_device *adev)
+{
+	uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw->size + 4);
+	uint32_t offset;
+
+	/* cache window 0: fw */
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+		WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
+			(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN].tmr_mc_addr_lo));
+		WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
+			(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN].tmr_mc_addr_hi));
+		WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_OFFSET0, 0);
+		offset = 0;
+	} else {
+		WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
+			lower_32_bits(adev->vcn.gpu_addr));
+		WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
+			upper_32_bits(adev->vcn.gpu_addr));
+		offset = size;
+		/* No signed header for now from firmware
+		WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_OFFSET0,
+			AMDGPU_UVD_FIRMWARE_OFFSET >> 3);
+		*/
+		WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_OFFSET0, 0);
+	}
+	WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_SIZE0, size);
+
+	/* cache window 1: stack */
+	WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW,
+		lower_32_bits(adev->vcn.gpu_addr + offset));
+	WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH,
+		upper_32_bits(adev->vcn.gpu_addr + offset));
+	WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_OFFSET1, 0);
+	WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_SIZE1, AMDGPU_VCN_STACK_SIZE);
+
+	/* cache window 2: context */
+	WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW,
+		lower_32_bits(adev->vcn.gpu_addr + offset + AMDGPU_VCN_STACK_SIZE));
+	WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH,
+		upper_32_bits(adev->vcn.gpu_addr + offset + AMDGPU_VCN_STACK_SIZE));
+	WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_OFFSET2, 0);
+	WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_SIZE2, AMDGPU_VCN_CONTEXT_SIZE);
+}
+
+/**
+ * vcn_v2_5_disable_clock_gating - disable VCN clock gating
+ *
+ * @adev: amdgpu_device pointer
+ * @sw: enable SW clock gating
+ *
+ * Disable clock gating for VCN block
+ */
+static void vcn_v2_5_disable_clock_gating(struct amdgpu_device *adev)
+{
+	uint32_t data;
+	int ret = 0;
+
+	/* UVD disable CGC */
+	data = RREG32_SOC15(VCN, 0, mmUVD_CGC_CTRL);
+	if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG)
+		data |= 1 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+	else
+		data &= ~ UVD_CGC_CTRL__DYN_CLOCK_MODE_MASK;
+	data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
+	data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
+	WREG32_SOC15(VCN, 0, mmUVD_CGC_CTRL, data);
+
+	data = RREG32_SOC15(VCN, 0, mmUVD_CGC_GATE);
+	data &= ~(UVD_CGC_GATE__SYS_MASK
+		| UVD_CGC_GATE__UDEC_MASK
+		| UVD_CGC_GATE__MPEG2_MASK
+		| UVD_CGC_GATE__REGS_MASK
+		| UVD_CGC_GATE__RBC_MASK
+		| UVD_CGC_GATE__LMI_MC_MASK
+		| UVD_CGC_GATE__LMI_UMC_MASK
+		| UVD_CGC_GATE__IDCT_MASK
+		| UVD_CGC_GATE__MPRD_MASK
+		| UVD_CGC_GATE__MPC_MASK
+		| UVD_CGC_GATE__LBSI_MASK
+		| UVD_CGC_GATE__LRBBM_MASK
+		| UVD_CGC_GATE__UDEC_RE_MASK
+		| UVD_CGC_GATE__UDEC_CM_MASK
+		| UVD_CGC_GATE__UDEC_IT_MASK
+		| UVD_CGC_GATE__UDEC_DB_MASK
+		| UVD_CGC_GATE__UDEC_MP_MASK
+		| UVD_CGC_GATE__WCB_MASK
+		| UVD_CGC_GATE__VCPU_MASK
+		| UVD_CGC_GATE__MMSCH_MASK);
+
+	WREG32_SOC15(VCN, 0, mmUVD_CGC_GATE, data);
+
+	SOC15_WAIT_ON_RREG(VCN, 0, mmUVD_CGC_GATE, 0,  0xFFFFFFFF, ret);
+
+	data = RREG32_SOC15(VCN, 0, mmUVD_CGC_CTRL);
+	data &= ~(UVD_CGC_CTRL__UDEC_RE_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_CM_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_IT_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_DB_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_MP_MODE_MASK
+		| UVD_CGC_CTRL__SYS_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_MODE_MASK
+		| UVD_CGC_CTRL__MPEG2_MODE_MASK
+		| UVD_CGC_CTRL__REGS_MODE_MASK
+		| UVD_CGC_CTRL__RBC_MODE_MASK
+		| UVD_CGC_CTRL__LMI_MC_MODE_MASK
+		| UVD_CGC_CTRL__LMI_UMC_MODE_MASK
+		| UVD_CGC_CTRL__IDCT_MODE_MASK
+		| UVD_CGC_CTRL__MPRD_MODE_MASK
+		| UVD_CGC_CTRL__MPC_MODE_MASK
+		| UVD_CGC_CTRL__LBSI_MODE_MASK
+		| UVD_CGC_CTRL__LRBBM_MODE_MASK
+		| UVD_CGC_CTRL__WCB_MODE_MASK
+		| UVD_CGC_CTRL__VCPU_MODE_MASK
+		| UVD_CGC_CTRL__MMSCH_MODE_MASK);
+	WREG32_SOC15(VCN, 0, mmUVD_CGC_CTRL, data);
+
+	/* turn on */
+	data = RREG32_SOC15(VCN, 0, mmUVD_SUVD_CGC_GATE);
+	data |= (UVD_SUVD_CGC_GATE__SRE_MASK
+		| UVD_SUVD_CGC_GATE__SIT_MASK
+		| UVD_SUVD_CGC_GATE__SMP_MASK
+		| UVD_SUVD_CGC_GATE__SCM_MASK
+		| UVD_SUVD_CGC_GATE__SDB_MASK
+		| UVD_SUVD_CGC_GATE__SRE_H264_MASK
+		| UVD_SUVD_CGC_GATE__SRE_HEVC_MASK
+		| UVD_SUVD_CGC_GATE__SIT_H264_MASK
+		| UVD_SUVD_CGC_GATE__SIT_HEVC_MASK
+		| UVD_SUVD_CGC_GATE__SCM_H264_MASK
+		| UVD_SUVD_CGC_GATE__SCM_HEVC_MASK
+		| UVD_SUVD_CGC_GATE__SDB_H264_MASK
+		| UVD_SUVD_CGC_GATE__SDB_HEVC_MASK
+		| UVD_SUVD_CGC_GATE__SCLR_MASK
+		| UVD_SUVD_CGC_GATE__UVD_SC_MASK
+		| UVD_SUVD_CGC_GATE__ENT_MASK
+		| UVD_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK
+		| UVD_SUVD_CGC_GATE__SIT_HEVC_ENC_MASK
+		| UVD_SUVD_CGC_GATE__SITE_MASK
+		| UVD_SUVD_CGC_GATE__SRE_VP9_MASK
+		| UVD_SUVD_CGC_GATE__SCM_VP9_MASK
+		| UVD_SUVD_CGC_GATE__SIT_VP9_DEC_MASK
+		| UVD_SUVD_CGC_GATE__SDB_VP9_MASK
+		| UVD_SUVD_CGC_GATE__IME_HEVC_MASK);
+	WREG32_SOC15(VCN, 0, mmUVD_SUVD_CGC_GATE, data);
+
+	data = RREG32_SOC15(VCN, 0, mmUVD_SUVD_CGC_CTRL);
+	data &= ~(UVD_SUVD_CGC_CTRL__SRE_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SIT_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SMP_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SCM_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SDB_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SCLR_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__UVD_SC_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__ENT_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__IME_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SITE_MODE_MASK);
+	WREG32_SOC15(VCN, 0, mmUVD_SUVD_CGC_CTRL, data);
+}
+
+/**
+ * vcn_v2_5_enable_clock_gating - enable VCN clock gating
+ *
+ * @adev: amdgpu_device pointer
+ * @sw: enable SW clock gating
+ *
+ * Enable clock gating for VCN block
+ */
+static void vcn_v2_5_enable_clock_gating(struct amdgpu_device *adev)
+{
+	uint32_t data = 0;
+
+	/* enable UVD CGC */
+	data = RREG32_SOC15(VCN, 0, mmUVD_CGC_CTRL);
+	if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG)
+		data |= 1 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+	else
+		data |= 0 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+	data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
+	data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
+	WREG32_SOC15(VCN, 0, mmUVD_CGC_CTRL, data);
+
+	data = RREG32_SOC15(VCN, 0, mmUVD_CGC_CTRL);
+	data |= (UVD_CGC_CTRL__UDEC_RE_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_CM_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_IT_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_DB_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_MP_MODE_MASK
+		| UVD_CGC_CTRL__SYS_MODE_MASK
+		| UVD_CGC_CTRL__UDEC_MODE_MASK
+		| UVD_CGC_CTRL__MPEG2_MODE_MASK
+		| UVD_CGC_CTRL__REGS_MODE_MASK
+		| UVD_CGC_CTRL__RBC_MODE_MASK
+		| UVD_CGC_CTRL__LMI_MC_MODE_MASK
+		| UVD_CGC_CTRL__LMI_UMC_MODE_MASK
+		| UVD_CGC_CTRL__IDCT_MODE_MASK
+		| UVD_CGC_CTRL__MPRD_MODE_MASK
+		| UVD_CGC_CTRL__MPC_MODE_MASK
+		| UVD_CGC_CTRL__LBSI_MODE_MASK
+		| UVD_CGC_CTRL__LRBBM_MODE_MASK
+		| UVD_CGC_CTRL__WCB_MODE_MASK
+		| UVD_CGC_CTRL__VCPU_MODE_MASK);
+	WREG32_SOC15(VCN, 0, mmUVD_CGC_CTRL, data);
+
+	data = RREG32_SOC15(VCN, 0, mmUVD_SUVD_CGC_CTRL);
+	data |= (UVD_SUVD_CGC_CTRL__SRE_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SIT_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SMP_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SCM_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SDB_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SCLR_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__UVD_SC_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__ENT_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__IME_MODE_MASK
+		| UVD_SUVD_CGC_CTRL__SITE_MODE_MASK);
+	WREG32_SOC15(VCN, 0, mmUVD_SUVD_CGC_CTRL, data);
+}
+
+static int vcn_v2_5_start(struct amdgpu_device *adev)
+{
+	struct amdgpu_ring *ring = &adev->vcn.ring_dec;
+	uint32_t rb_bufsz, tmp;
+	int i, j, r;
+
+	/* disable register anti-hang mechanism */
+	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_POWER_STATUS), 0,
+		~UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
+
+	/* set uvd status busy */
+	tmp = RREG32_SOC15(UVD, 0, mmUVD_STATUS) | UVD_STATUS__UVD_BUSY;
+	WREG32_SOC15(UVD, 0, mmUVD_STATUS, tmp);
+
+	/*SW clock gating */
+	vcn_v2_5_disable_clock_gating(adev);
+
+	/* enable VCPU clock */
+	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_VCPU_CNTL),
+		UVD_VCPU_CNTL__CLK_EN_MASK, ~UVD_VCPU_CNTL__CLK_EN_MASK);
+
+	/* disable master interrupt */
+	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_MASTINT_EN), 0,
+		~UVD_MASTINT_EN__VCPU_EN_MASK);
+
+	/* setup mmUVD_LMI_CTRL */
+	tmp = RREG32_SOC15(UVD, 0, mmUVD_LMI_CTRL);
+	tmp &= ~0xff;
+	WREG32_SOC15(UVD, 0, mmUVD_LMI_CTRL, tmp | 0x8|
+		UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK	|
+		UVD_LMI_CTRL__MASK_MC_URGENT_MASK |
+		UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK |
+		UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK);
+
+	/* setup mmUVD_MPC_CNTL */
+	tmp = RREG32_SOC15(UVD, 0, mmUVD_MPC_CNTL);
+	tmp &= ~UVD_MPC_CNTL__REPLACEMENT_MODE_MASK;
+	tmp |= 0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT;
+	WREG32_SOC15(VCN, 0, mmUVD_MPC_CNTL, tmp);
+
+	/* setup UVD_MPC_SET_MUXA0 */
+	WREG32_SOC15(UVD, 0, mmUVD_MPC_SET_MUXA0,
+		((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) |
+		(0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) |
+		(0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) |
+		(0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT)));
+
+	/* setup UVD_MPC_SET_MUXB0 */
+	WREG32_SOC15(UVD, 0, mmUVD_MPC_SET_MUXB0,
+		((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) |
+		(0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) |
+		(0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) |
+		(0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT)));
+
+	/* setup mmUVD_MPC_SET_MUX */
+	WREG32_SOC15(UVD, 0, mmUVD_MPC_SET_MUX,
+		((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) |
+		(0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) |
+		(0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT)));
+
+	vcn_v2_5_mc_resume(adev);
+
+	/* VCN global tiling registers */
+	WREG32_SOC15(UVD, 0, mmUVD_GFX8_ADDR_CONFIG,
+		adev->gfx.config.gb_addr_config);
+	WREG32_SOC15(UVD, 0, mmUVD_GFX8_ADDR_CONFIG,
+		adev->gfx.config.gb_addr_config);
+
+	/* enable LMI MC and UMC channels */
+	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_CTRL2), 0,
+		~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK);
+
+	/* unblock VCPU register access */
+	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_RB_ARB_CTRL), 0,
+		~UVD_RB_ARB_CTRL__VCPU_DIS_MASK);
+
+	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_VCPU_CNTL), 0,
+		~UVD_VCPU_CNTL__BLK_RST_MASK);
+
+	for (i = 0; i < 10; ++i) {
+		uint32_t status;
+
+		for (j = 0; j < 100; ++j) {
+			status = RREG32_SOC15(UVD, 0, mmUVD_STATUS);
+			if (status & 2)
+				break;
+			if (amdgpu_emu_mode == 1)
+				msleep(500);
+			else
+				mdelay(10);
+		}
+		r = 0;
+		if (status & 2)
+			break;
+
+		DRM_ERROR("VCN decode not responding, trying to reset the VCPU!!!\n");
+		WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_VCPU_CNTL),
+			UVD_VCPU_CNTL__BLK_RST_MASK,
+			~UVD_VCPU_CNTL__BLK_RST_MASK);
+		mdelay(10);
+		WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_VCPU_CNTL), 0,
+			~UVD_VCPU_CNTL__BLK_RST_MASK);
+
+		mdelay(10);
+		r = -1;
+	}
+
+	if (r) {
+		DRM_ERROR("VCN decode not responding, giving up!!!\n");
+		return r;
+	}
+
+	/* enable master interrupt */
+	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_MASTINT_EN),
+		UVD_MASTINT_EN__VCPU_EN_MASK,
+		~UVD_MASTINT_EN__VCPU_EN_MASK);
+
+	/* clear the busy bit of VCN_STATUS */
+	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_STATUS), 0,
+		~(2 << UVD_STATUS__VCPU_REPORT__SHIFT));
+
+	WREG32_SOC15(UVD, 0, mmUVD_LMI_RBC_RB_VMID, 0);
+
+	/* force RBC into idle state */
+	rb_bufsz = order_base_2(ring->ring_size);
+	tmp = REG_SET_FIELD(0, UVD_RBC_RB_CNTL, RB_BUFSZ, rb_bufsz);
+	tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_BLKSZ, 1);
+	tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_FETCH, 1);
+	tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_UPDATE, 1);
+	tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_RPTR_WR_EN, 1);
+	WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_CNTL, tmp);
+
+	/* programm the RB_BASE for ring buffer */
+	WREG32_SOC15(UVD, 0, mmUVD_LMI_RBC_RB_64BIT_BAR_LOW,
+		lower_32_bits(ring->gpu_addr));
+	WREG32_SOC15(UVD, 0, mmUVD_LMI_RBC_RB_64BIT_BAR_HIGH,
+		upper_32_bits(ring->gpu_addr));
+
+	/* Initialize the ring buffer's read and write pointers */
+	WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_RPTR, 0);
+
+	ring->wptr = RREG32_SOC15(UVD, 0, mmUVD_RBC_RB_RPTR);
+	WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR,
+			lower_32_bits(ring->wptr));
+	ring = &adev->vcn.ring_enc[0];
+	WREG32_SOC15(UVD, 0, mmUVD_RB_RPTR, lower_32_bits(ring->wptr));
+	WREG32_SOC15(UVD, 0, mmUVD_RB_WPTR, lower_32_bits(ring->wptr));
+	WREG32_SOC15(UVD, 0, mmUVD_RB_BASE_LO, ring->gpu_addr);
+	WREG32_SOC15(UVD, 0, mmUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr));
+	WREG32_SOC15(UVD, 0, mmUVD_RB_SIZE, ring->ring_size / 4);
+
+	ring = &adev->vcn.ring_enc[1];
+	WREG32_SOC15(UVD, 0, mmUVD_RB_RPTR2, lower_32_bits(ring->wptr));
+	WREG32_SOC15(UVD, 0, mmUVD_RB_WPTR2, lower_32_bits(ring->wptr));
+	WREG32_SOC15(UVD, 0, mmUVD_RB_BASE_LO2, ring->gpu_addr);
+	WREG32_SOC15(UVD, 0, mmUVD_RB_BASE_HI2, upper_32_bits(ring->gpu_addr));
+	WREG32_SOC15(UVD, 0, mmUVD_RB_SIZE2, ring->ring_size / 4);
+
+	return r;
+}
+
+static int vcn_v2_5_stop(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+	int r;
+
+	/* wait for vcn idle */
+	SOC15_WAIT_ON_RREG(VCN, 0, mmUVD_STATUS, UVD_STATUS__IDLE, 0x7, r);
+	if (r)
+		return r;
+
+	tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK |
+		UVD_LMI_STATUS__READ_CLEAN_MASK |
+		UVD_LMI_STATUS__WRITE_CLEAN_MASK |
+		UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK;
+	SOC15_WAIT_ON_RREG(VCN, 0, mmUVD_LMI_STATUS, tmp, tmp, r);
+	if (r)
+		return r;
+
+	/* block LMI UMC channel */
+	tmp = RREG32_SOC15(VCN, 0, mmUVD_LMI_CTRL2);
+	tmp |= UVD_LMI_CTRL2__STALL_ARB_UMC_MASK;
+	WREG32_SOC15(VCN, 0, mmUVD_LMI_CTRL2, tmp);
+
+	tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK|
+		UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK;
+	SOC15_WAIT_ON_RREG(VCN, 0, mmUVD_LMI_STATUS, tmp, tmp, r);
+	if (r)
+		return r;
+
+	/* block VCPU register access */
+	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_RB_ARB_CTRL),
+		UVD_RB_ARB_CTRL__VCPU_DIS_MASK,
+		~UVD_RB_ARB_CTRL__VCPU_DIS_MASK);
+
+	/* reset VCPU */
+	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_VCPU_CNTL),
+		UVD_VCPU_CNTL__BLK_RST_MASK,
+		~UVD_VCPU_CNTL__BLK_RST_MASK);
+
+	/* disable VCPU clock */
+	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_VCPU_CNTL), 0,
+		~(UVD_VCPU_CNTL__CLK_EN_MASK));
+
+	/* clear status */
+	WREG32_SOC15(VCN, 0, mmUVD_STATUS, 0);
+
+	vcn_v2_5_enable_clock_gating(adev);
+
+	/* enable register anti-hang mechanism */
+	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_POWER_STATUS),
+		UVD_POWER_STATUS__UVD_POWER_STATUS_MASK,
+		~UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
+
+	return 0;
+}
+
 static bool vcn_v2_5_is_idle(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
@@ -328,13 +768,38 @@ static int vcn_v2_5_wait_for_idle(void *handle)
 static int vcn_v2_5_set_clockgating_state(void *handle,
 					  enum amd_clockgating_state state)
 {
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	bool enable = (state == AMD_CG_STATE_GATE) ? true : false;
+
+	if (enable) {
+		if (vcn_v2_5_is_idle(handle))
+			return -EBUSY;
+		vcn_v2_5_enable_clock_gating(adev);
+	} else {
+		vcn_v2_5_disable_clock_gating(adev);
+	}
+
 	return 0;
 }
 
 static int vcn_v2_5_set_powergating_state(void *handle,
 					  enum amd_powergating_state state)
 {
-	return 0;
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	int ret;
+
+	if(state == adev->vcn.cur_state)
+		return 0;
+
+	if (state == AMD_PG_STATE_GATE)
+		ret = vcn_v2_5_stop(adev);
+	else
+		ret = vcn_v2_5_start(adev);
+
+	if(!ret)
+		adev->vcn.cur_state = state;
+
+	return ret;
 }
 
 static int vcn_v2_5_set_interrupt_state(struct amdgpu_device *adev,
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#58 lore raw

PATCH 60/102 drm/amdgpu: add Arcturus to the VCN family

In reply to a message not in the archive

From: Leo Liu <[email protected]>

including firmware support etc.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c | 5 +++++
 1 file changed, 5 insertions(+)
drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c +5 -0
index 8a162bdad79e..bb0d1ef50c9c 100644
@@ -45,12 +45,14 @@
 #define FIRMWARE_RAVEN		"amdgpu/raven_vcn.bin"
 #define FIRMWARE_PICASSO	"amdgpu/picasso_vcn.bin"
 #define FIRMWARE_RAVEN2		"amdgpu/raven2_vcn.bin"
+#define FIRMWARE_ARCTURUS 	"amdgpu/arcturus_vcn.bin"
 #define FIRMWARE_NAVI10 	"amdgpu/navi10_vcn.bin"
 #define FIRMWARE_NAVI14 	"amdgpu/navi14_vcn.bin"
 
 MODULE_FIRMWARE(FIRMWARE_RAVEN);
 MODULE_FIRMWARE(FIRMWARE_PICASSO);
 MODULE_FIRMWARE(FIRMWARE_RAVEN2);
+MODULE_FIRMWARE(FIRMWARE_ARCTURUS);
 MODULE_FIRMWARE(FIRMWARE_NAVI10);
 MODULE_FIRMWARE(FIRMWARE_NAVI14);
 
@@ -75,6 +77,9 @@ int amdgpu_vcn_sw_init(struct amdgpu_device *adev)
 		else
 			fw_name = FIRMWARE_RAVEN;
 		break;
+	case CHIP_ARCTURUS:
+		fw_name = FIRMWARE_ARCTURUS;
+		break;
 	case CHIP_NAVI10:
 		fw_name = FIRMWARE_NAVI10;
 		if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) &&
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#59 lore raw

PATCH 61/102 drm/amdgpu/VCN2.5: set decode ring functions

In reply to a message not in the archive

From: Leo Liu <[email protected]>

Also reuse most of the VCN2.0 decode ring functions

---
 drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c | 78 +++++++++++++++++++++++++++
 1 file changed, 78 insertions(+)
drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c +78 -0
index b6e72fff94f5..f16a4f682e26 100644
@@ -747,6 +747,84 @@ static int vcn_v2_5_stop(struct amdgpu_device *adev)
 	return 0;
 }
 
+/**
+ * vcn_v2_5_dec_ring_get_rptr - get read pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware read pointer
+ */
+static uint64_t vcn_v2_5_dec_ring_get_rptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	return RREG32_SOC15(UVD, 0, mmUVD_RBC_RB_RPTR);
+}
+
+/**
+ * vcn_v2_5_dec_ring_get_wptr - get write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware write pointer
+ */
+static uint64_t vcn_v2_5_dec_ring_get_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	return RREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR);
+}
+
+/**
+ * vcn_v2_5_dec_ring_set_wptr - set write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Commits the write pointer to the hardware
+ */
+static void vcn_v2_5_dec_ring_set_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR, lower_32_bits(ring->wptr));
+}
+
+static const struct amdgpu_ring_funcs vcn_v2_5_dec_ring_vm_funcs = {
+	.type = AMDGPU_RING_TYPE_VCN_DEC,
+	.align_mask = 0xf,
+	.vmhub = AMDGPU_MMHUB_1,
+	.get_rptr = vcn_v2_5_dec_ring_get_rptr,
+	.get_wptr = vcn_v2_5_dec_ring_get_wptr,
+	.set_wptr = vcn_v2_5_dec_ring_set_wptr,
+	.emit_frame_size =
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 8 +
+		8 + /* vcn_v2_0_dec_ring_emit_vm_flush */
+		14 + 14 + /* vcn_v2_0_dec_ring_emit_fence x2 vm fence */
+		6,
+	.emit_ib_size = 8, /* vcn_v2_0_dec_ring_emit_ib */
+	.emit_ib = vcn_v2_0_dec_ring_emit_ib,
+	.emit_fence = vcn_v2_0_dec_ring_emit_fence,
+	.emit_vm_flush = vcn_v2_0_dec_ring_emit_vm_flush,
+	.test_ring = amdgpu_vcn_dec_ring_test_ring,
+	.test_ib = amdgpu_vcn_dec_ring_test_ib,
+	.insert_nop = vcn_v2_0_dec_ring_insert_nop,
+	.insert_start = vcn_v2_0_dec_ring_insert_start,
+	.insert_end = vcn_v2_0_dec_ring_insert_end,
+	.pad_ib = amdgpu_ring_generic_pad_ib,
+	.begin_use = amdgpu_vcn_ring_begin_use,
+	.end_use = amdgpu_vcn_ring_end_use,
+	.emit_wreg = vcn_v2_0_dec_ring_emit_wreg,
+	.emit_reg_wait = vcn_v2_0_dec_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+};
+
+static void vcn_v2_5_set_dec_ring_funcs(struct amdgpu_device *adev)
+{
+	adev->vcn.ring_dec.funcs = &vcn_v2_5_dec_ring_vm_funcs;
+	DRM_INFO("VCN decode is enabled in VM mode\n");
+}
+
 static bool vcn_v2_5_is_idle(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#60 lore raw

PATCH 62/102 drm/amdgpu/VCN2.5: set encode ring functions

In reply to a message not in the archive

From: Leo Liu <[email protected]>

Also reuse most of the VCN2.0 encode ring functions

---
 drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c | 91 +++++++++++++++++++++++++++
 1 file changed, 91 insertions(+)
drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c +91 -0
index f16a4f682e26..b42f6769ae06 100644
@@ -819,12 +819,103 @@ static const struct amdgpu_ring_funcs vcn_v2_5_dec_ring_vm_funcs = {
 	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
 };
 
+/**
+ * vcn_v2_5_enc_ring_get_rptr - get enc read pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware enc read pointer
+ */
+static uint64_t vcn_v2_5_enc_ring_get_rptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring == &adev->vcn.ring_enc[0])
+		return RREG32_SOC15(UVD, 0, mmUVD_RB_RPTR);
+	else
+		return RREG32_SOC15(UVD, 0, mmUVD_RB_RPTR2);
+}
+
+/**
+ * vcn_v2_5_enc_ring_get_wptr - get enc write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware enc write pointer
+ */
+static uint64_t vcn_v2_5_enc_ring_get_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring == &adev->vcn.ring_enc[0])
+		return RREG32_SOC15(UVD, 0, mmUVD_RB_WPTR);
+	else
+		return RREG32_SOC15(UVD, 0, mmUVD_RB_WPTR2);
+}
+
+/**
+ * vcn_v2_5_enc_ring_set_wptr - set enc write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Commits the enc write pointer to the hardware
+ */
+static void vcn_v2_5_enc_ring_set_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring == &adev->vcn.ring_enc[0])
+		WREG32_SOC15(UVD, 0, mmUVD_RB_WPTR, lower_32_bits(ring->wptr));
+	else
+		WREG32_SOC15(UVD, 0, mmUVD_RB_WPTR2, lower_32_bits(ring->wptr));
+}
+
+static const struct amdgpu_ring_funcs vcn_v2_5_enc_ring_vm_funcs = {
+	.type = AMDGPU_RING_TYPE_VCN_ENC,
+	.align_mask = 0x3f,
+	.nop = VCN_ENC_CMD_NO_OP,
+	.vmhub = AMDGPU_MMHUB_1,
+	.get_rptr = vcn_v2_5_enc_ring_get_rptr,
+	.get_wptr = vcn_v2_5_enc_ring_get_wptr,
+	.set_wptr = vcn_v2_5_enc_ring_set_wptr,
+	.emit_frame_size =
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 4 +
+		4 + /* vcn_v2_0_enc_ring_emit_vm_flush */
+		5 + 5 + /* vcn_v2_0_enc_ring_emit_fence x2 vm fence */
+		1, /* vcn_v2_0_enc_ring_insert_end */
+	.emit_ib_size = 5, /* vcn_v2_0_enc_ring_emit_ib */
+	.emit_ib = vcn_v2_0_enc_ring_emit_ib,
+	.emit_fence = vcn_v2_0_enc_ring_emit_fence,
+	.emit_vm_flush = vcn_v2_0_enc_ring_emit_vm_flush,
+	.test_ring = amdgpu_vcn_enc_ring_test_ring,
+	.test_ib = amdgpu_vcn_enc_ring_test_ib,
+	.insert_nop = amdgpu_ring_insert_nop,
+	.insert_end = vcn_v2_0_enc_ring_insert_end,
+	.pad_ib = amdgpu_ring_generic_pad_ib,
+	.begin_use = amdgpu_vcn_ring_begin_use,
+	.end_use = amdgpu_vcn_ring_end_use,
+	.emit_wreg = vcn_v2_0_enc_ring_emit_wreg,
+	.emit_reg_wait = vcn_v2_0_enc_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+};
+
 static void vcn_v2_5_set_dec_ring_funcs(struct amdgpu_device *adev)
 {
 	adev->vcn.ring_dec.funcs = &vcn_v2_5_dec_ring_vm_funcs;
 	DRM_INFO("VCN decode is enabled in VM mode\n");
 }
 
+static void vcn_v2_5_set_enc_ring_funcs(struct amdgpu_device *adev)
+{
+	int i;
+
+	for (i = 0; i < adev->vcn.num_enc_rings; ++i)
+		adev->vcn.ring_enc[i].funcs = &vcn_v2_5_enc_ring_vm_funcs;
+
+	DRM_INFO("VCN encode is enabled in VM mode\n");
+}
+
 static bool vcn_v2_5_is_idle(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#61 lore raw

PATCH 63/102 drm/amdgpu: add JPEG2.5 HW start and stop

In reply to a message not in the archive

From: Leo Liu <[email protected]>

JPEG engine initialization and suspend sequences

---
 drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c | 104 ++++++++++++++++++++++++++
 1 file changed, 104 insertions(+)
drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c +104 -0
index b42f6769ae06..82c9c40e9ae4 100644
@@ -530,6 +530,104 @@ static void vcn_v2_5_enable_clock_gating(struct amdgpu_device *adev)
 	WREG32_SOC15(VCN, 0, mmUVD_SUVD_CGC_CTRL, data);
 }
 
+/**
+ * jpeg_v2_5_start - start JPEG block
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Setup and start the JPEG block
+ */
+static int jpeg_v2_5_start(struct amdgpu_device *adev)
+{
+	struct amdgpu_ring *ring = &adev->vcn.ring_jpeg;
+	uint32_t tmp;
+
+	/* disable anti hang mechanism */
+	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_JPEG_POWER_STATUS), 0,
+		~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK);
+
+	/* JPEG disable CGC */
+	tmp = RREG32_SOC15(VCN, 0, mmJPEG_CGC_CTRL);
+	tmp |= 1 << JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+	tmp |= 1 << JPEG_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
+	tmp |= 4 << JPEG_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
+	WREG32_SOC15(VCN, 0, mmJPEG_CGC_CTRL, tmp);
+
+	tmp = RREG32_SOC15(VCN, 0, mmJPEG_CGC_GATE);
+	tmp &= ~(JPEG_CGC_GATE__JPEG_DEC_MASK
+		| JPEG_CGC_GATE__JPEG2_DEC_MASK
+		| JPEG_CGC_GATE__JMCIF_MASK
+		| JPEG_CGC_GATE__JRBBM_MASK);
+	WREG32_SOC15(VCN, 0, mmJPEG_CGC_GATE, tmp);
+
+	tmp = RREG32_SOC15(VCN, 0, mmJPEG_CGC_CTRL);
+	tmp &= ~(JPEG_CGC_CTRL__JPEG_DEC_MODE_MASK
+		| JPEG_CGC_CTRL__JPEG2_DEC_MODE_MASK
+		| JPEG_CGC_CTRL__JMCIF_MODE_MASK
+		| JPEG_CGC_CTRL__JRBBM_MODE_MASK);
+	WREG32_SOC15(VCN, 0, mmJPEG_CGC_CTRL, tmp);
+
+	/* MJPEG global tiling registers */
+	WREG32_SOC15(UVD, 0, mmJPEG_DEC_GFX8_ADDR_CONFIG,
+		adev->gfx.config.gb_addr_config);
+	WREG32_SOC15(UVD, 0, mmJPEG_DEC_GFX10_ADDR_CONFIG,
+		adev->gfx.config.gb_addr_config);
+
+	/* enable JMI channel */
+	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_JMI_CNTL), 0,
+		~UVD_JMI_CNTL__SOFT_RESET_MASK);
+
+	/* enable System Interrupt for JRBC */
+	WREG32_P(SOC15_REG_OFFSET(VCN, 0, mmJPEG_SYS_INT_EN),
+		JPEG_SYS_INT_EN__DJRBC_MASK,
+		~JPEG_SYS_INT_EN__DJRBC_MASK);
+
+	WREG32_SOC15(UVD, 0, mmUVD_LMI_JRBC_RB_VMID, 0);
+	WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_CNTL, (0x00000001L | 0x00000002L));
+	WREG32_SOC15(UVD, 0, mmUVD_LMI_JRBC_RB_64BIT_BAR_LOW,
+		lower_32_bits(ring->gpu_addr));
+	WREG32_SOC15(UVD, 0, mmUVD_LMI_JRBC_RB_64BIT_BAR_HIGH,
+		upper_32_bits(ring->gpu_addr));
+	WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_RPTR, 0);
+	WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_WPTR, 0);
+	WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_CNTL, 0x00000002L);
+	WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_SIZE, ring->ring_size / 4);
+	ring->wptr = RREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_WPTR);
+
+	return 0;
+}
+
+/**
+ * jpeg_v2_5_stop - stop JPEG block
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * stop the JPEG block
+ */
+static int jpeg_v2_5_stop(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+
+	/* reset JMI */
+	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_JMI_CNTL),
+		UVD_JMI_CNTL__SOFT_RESET_MASK,
+		~UVD_JMI_CNTL__SOFT_RESET_MASK);
+
+	tmp = RREG32_SOC15(VCN, 0, mmJPEG_CGC_GATE);
+	tmp |= (JPEG_CGC_GATE__JPEG_DEC_MASK
+		|JPEG_CGC_GATE__JPEG2_DEC_MASK
+		|JPEG_CGC_GATE__JMCIF_MASK
+		|JPEG_CGC_GATE__JRBBM_MASK);
+	WREG32_SOC15(VCN, 0, mmJPEG_CGC_GATE, tmp);
+
+	/* enable anti hang mechanism */
+	WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_JPEG_POWER_STATUS),
+		UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK,
+		~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK);
+
+	return 0;
+}
+
 static int vcn_v2_5_start(struct amdgpu_device *adev)
 {
 	struct amdgpu_ring *ring = &adev->vcn.ring_dec;
@@ -688,6 +786,8 @@ static int vcn_v2_5_start(struct amdgpu_device *adev)
 	WREG32_SOC15(UVD, 0, mmUVD_RB_BASE_HI2, upper_32_bits(ring->gpu_addr));
 	WREG32_SOC15(UVD, 0, mmUVD_RB_SIZE2, ring->ring_size / 4);
 
+	r = jpeg_v2_5_start(adev);
+
 	return r;
 }
 
@@ -696,6 +796,10 @@ static int vcn_v2_5_stop(struct amdgpu_device *adev)
 	uint32_t tmp;
 	int r;
 
+	r = jpeg_v2_5_stop(adev);
+	if (r)
+		return r;
+
 	/* wait for vcn idle */
 	SOC15_WAIT_ON_RREG(VCN, 0, mmUVD_STATUS, UVD_STATUS__IDLE, 0x7, r);
 	if (r)
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#62 lore raw

PATCH 64/102 drm/amdgpu/VCN2.5: set JPEG decode ring functions

In reply to a message not in the archive

From: Leo Liu <[email protected]>

Also reuse most of the JPEG2.0 decode ring functions

---
 drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c | 78 +++++++++++++++++++++++++++
 1 file changed, 78 insertions(+)
drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c +78 -0
index 82c9c40e9ae4..0ffc0d60fad8 100644
@@ -1004,6 +1004,78 @@ static const struct amdgpu_ring_funcs vcn_v2_5_enc_ring_vm_funcs = {
 	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
 };
 
+/**
+ * vcn_v2_5_jpeg_ring_get_rptr - get read pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware read pointer
+ */
+static uint64_t vcn_v2_5_jpeg_ring_get_rptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	return RREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_RPTR);
+}
+
+/**
+ * vcn_v2_5_jpeg_ring_get_wptr - get write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware write pointer
+ */
+static uint64_t vcn_v2_5_jpeg_ring_get_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	return RREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_WPTR);
+}
+
+/**
+ * vcn_v2_5_jpeg_ring_set_wptr - set write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Commits the write pointer to the hardware
+ */
+static void vcn_v2_5_jpeg_ring_set_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_WPTR, lower_32_bits(ring->wptr));
+}
+
+static const struct amdgpu_ring_funcs vcn_v2_5_jpeg_ring_vm_funcs = {
+	.type = AMDGPU_RING_TYPE_VCN_JPEG,
+	.align_mask = 0xf,
+	.vmhub = AMDGPU_MMHUB_1,
+	.get_rptr = vcn_v2_5_jpeg_ring_get_rptr,
+	.get_wptr = vcn_v2_5_jpeg_ring_get_wptr,
+	.set_wptr = vcn_v2_5_jpeg_ring_set_wptr,
+	.emit_frame_size =
+		SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 +
+		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 8 +
+		8 + /* vcn_v2_0_jpeg_ring_emit_vm_flush */
+		18 + 18 + /* vcn_v2_0_jpeg_ring_emit_fence x2 vm fence */
+		8 + 16,
+	.emit_ib_size = 22, /* vcn_v2_0_jpeg_ring_emit_ib */
+	.emit_ib = vcn_v2_0_jpeg_ring_emit_ib,
+	.emit_fence = vcn_v2_0_jpeg_ring_emit_fence,
+	.emit_vm_flush = vcn_v2_0_jpeg_ring_emit_vm_flush,
+	.test_ring = amdgpu_vcn_jpeg_ring_test_ring,
+	.test_ib = amdgpu_vcn_jpeg_ring_test_ib,
+	.insert_nop = vcn_v2_0_jpeg_ring_nop,
+	.insert_start = vcn_v2_0_jpeg_ring_insert_start,
+	.insert_end = vcn_v2_0_jpeg_ring_insert_end,
+	.pad_ib = amdgpu_ring_generic_pad_ib,
+	.begin_use = amdgpu_vcn_ring_begin_use,
+	.end_use = amdgpu_vcn_ring_end_use,
+	.emit_wreg = vcn_v2_0_jpeg_ring_emit_wreg,
+	.emit_reg_wait = vcn_v2_0_jpeg_ring_emit_reg_wait,
+	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+};
+
 static void vcn_v2_5_set_dec_ring_funcs(struct amdgpu_device *adev)
 {
 	adev->vcn.ring_dec.funcs = &vcn_v2_5_dec_ring_vm_funcs;
@@ -1020,6 +1092,12 @@ static void vcn_v2_5_set_enc_ring_funcs(struct amdgpu_device *adev)
 	DRM_INFO("VCN encode is enabled in VM mode\n");
 }
 
+static void vcn_v2_5_set_jpeg_ring_funcs(struct amdgpu_device *adev)
+{
+	adev->vcn.ring_jpeg.funcs = &vcn_v2_5_jpeg_ring_vm_funcs;
+	DRM_INFO("VCN jpeg decode is enabled in VM mode\n");
+}
+
 static bool vcn_v2_5_is_idle(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#63 lore raw

PATCH 65/102 drm/amdgpu: enable VCN2.5 on Arcturus

In reply to a message not in the archive

From: Leo Liu <[email protected]>

VCN is the video decode and encode engine on Arcturus

---
 drivers/gpu/drm/amd/amdgpu/soc15.c    |  2 ++
 drivers/gpu/drm/amd/amdgpu/vcn_v2_5.h | 29 +++++++++++++++++++++++++++
 2 files changed, 31 insertions(+)
 create mode 100644 drivers/gpu/drm/amd/amdgpu/vcn_v2_5.h
drivers/gpu/drm/amd/amdgpu/soc15.c +2 -0
index fb4c7aac8d9e..217afe23a2f6 100644
@@ -62,6 +62,7 @@
 #include "uvd_v7_0.h"
 #include "vce_v4_0.h"
 #include "vcn_v1_0.h"
+#include "vcn_v2_5.h"
 #include "dce_virtual.h"
 #include "mxgpu_ai.h"
 #include "amdgpu_smu.h"
@@ -683,6 +684,7 @@ int soc15_set_ip_blocks(struct amdgpu_device *adev)
 			amdgpu_device_ip_block_add(adev, &dce_virtual_ip_block);
 		amdgpu_device_ip_block_add(adev, &gfx_v9_0_ip_block);
 		amdgpu_device_ip_block_add(adev, &sdma_v4_0_ip_block);
+		amdgpu_device_ip_block_add(adev, &vcn_v2_5_ip_block);
 		break;
 	default:
 		return -EINVAL;
drivers/gpu/drm/amd/amdgpu/vcn_v2_5.h new file +29 -0
new file mode 100644
index 000000000000..8d9c0800b8e0
@@ -0,0 +1,29 @@
+/*
+ * Copyright 2019 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __VCN_V2_5_H__
+#define __VCN_V2_5_H__
+
+extern const struct amdgpu_ip_block_version vcn_v2_5_ip_block;
+
+#endif /* __VCN_V2_5_H__ */
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#64 lore raw

PATCH 66/102 drm/amdgpu: add vcn doorbell range function to nbio7.4 (v2)

In reply to a message not in the archive

From: Leo Liu <[email protected]>

To setup the aperture for VCN2.5

v2: setup vcn doorbells in vcn2.5 hw_init (Alex)

---
 drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c | 21 +++++++++++++++++++++
 drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c  |  3 +++
 2 files changed, 24 insertions(+)
drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c +21 -0
index fc45eaeaba6e..d8c9972a315b 100644
@@ -114,6 +114,26 @@ static void nbio_v7_4_sdma_doorbell_range(struct amdgpu_device *adev, int instan
 	WREG32(reg, doorbell_range);
 }
 
+static void nbio_v7_4_vcn_doorbell_range(struct amdgpu_device *adev, bool use_doorbell,
+					 int doorbell_index)
+{
+	u32 reg = SOC15_REG_OFFSET(NBIO, 0, mmBIF_MMSCH0_DOORBELL_RANGE);
+
+	u32 doorbell_range = RREG32(reg);
+
+	if (use_doorbell) {
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       BIF_MMSCH0_DOORBELL_RANGE, OFFSET,
+					       doorbell_index);
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       BIF_MMSCH0_DOORBELL_RANGE, SIZE, 8);
+	} else
+		doorbell_range = REG_SET_FIELD(doorbell_range,
+					       BIF_MMSCH0_DOORBELL_RANGE, SIZE, 0);
+
+	WREG32(reg, doorbell_range);
+}
+
 static void nbio_v7_4_enable_doorbell_aperture(struct amdgpu_device *adev,
 					       bool enable)
 {
@@ -292,6 +312,7 @@ const struct amdgpu_nbio_funcs nbio_v7_4_funcs = {
 	.hdp_flush = nbio_v7_4_hdp_flush,
 	.get_memsize = nbio_v7_4_get_memsize,
 	.sdma_doorbell_range = nbio_v7_4_sdma_doorbell_range,
+	.vcn_doorbell_range = nbio_v7_4_vcn_doorbell_range,
 	.enable_doorbell_aperture = nbio_v7_4_enable_doorbell_aperture,
 	.enable_doorbell_selfring_aperture = nbio_v7_4_enable_doorbell_selfring_aperture,
 	.ih_doorbell_range = nbio_v7_4_ih_doorbell_range,
drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c +3 -0
index 0ffc0d60fad8..f9d6819f0260 100644
@@ -205,6 +205,9 @@ static int vcn_v2_5_hw_init(void *handle)
 	struct amdgpu_ring *ring = &adev->vcn.ring_dec;
 	int i, r;
 
+	adev->nbio_funcs->vcn_doorbell_range(adev, ring->use_doorbell,
+		ring->doorbell_index);
+
 	r = amdgpu_ring_test_ring(ring);
 	if (r) {
 		ring->sched.ready = false;
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#65 lore raw

PATCH 67/102 drm/amdgpu: enable the Doorbell support for VCN2.5

In reply to a message not in the archive

From: Leo Liu <[email protected]>

Including decode, encode, and JPEG decode rings

---
 drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c        | 64 ++++++++++++++++----
 drivers/gpu/drm/amd/amdgpu/vega20_reg_init.c |  4 ++
 2 files changed, 56 insertions(+), 12 deletions(-)
drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c +52 -12
index f9d6819f0260..840737df19c0 100644
@@ -128,6 +128,8 @@ static int vcn_v2_5_sw_init(void *handle)
 		return r;
 
 	ring = &adev->vcn.ring_dec;
+	ring->use_doorbell = true;
+	ring->doorbell_index = adev->doorbell_index.vcn.vcn_ring0_1 << 1;
 	sprintf(ring->name, "vcn_dec");
 	r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.irq, 0);
 	if (r)
@@ -153,6 +155,8 @@ static int vcn_v2_5_sw_init(void *handle)
 
 	for (i = 0; i < adev->vcn.num_enc_rings; ++i) {
 		ring = &adev->vcn.ring_enc[i];
+		ring->use_doorbell = true;
+		ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 2 + i;
 		sprintf(ring->name, "vcn_enc%d", i);
 		r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.irq, 0);
 		if (r)
@@ -160,6 +164,8 @@ static int vcn_v2_5_sw_init(void *handle)
 	}
 
 	ring = &adev->vcn.ring_jpeg;
+	ring->use_doorbell = true;
+	ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 1;
 	sprintf(ring->name, "vcn_jpeg");
 	r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.irq, 0);
 	if (r)
@@ -879,7 +885,10 @@ static uint64_t vcn_v2_5_dec_ring_get_wptr(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
 
-	return RREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR);
+	if (ring->use_doorbell)
+		return adev->wb.wb[ring->wptr_offs];
+	else
+		return RREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR);
 }
 
 /**
@@ -893,7 +902,12 @@ static void vcn_v2_5_dec_ring_set_wptr(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
 
-	WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR, lower_32_bits(ring->wptr));
+	if (ring->use_doorbell) {
+		adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
+		WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
+	} else {
+		WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR, lower_32_bits(ring->wptr));
+	}
 }
 
 static const struct amdgpu_ring_funcs vcn_v2_5_dec_ring_vm_funcs = {
@@ -954,10 +968,17 @@ static uint64_t vcn_v2_5_enc_ring_get_wptr(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
 
-	if (ring == &adev->vcn.ring_enc[0])
-		return RREG32_SOC15(UVD, 0, mmUVD_RB_WPTR);
-	else
-		return RREG32_SOC15(UVD, 0, mmUVD_RB_WPTR2);
+	if (ring == &adev->vcn.ring_enc[0]) {
+		if (ring->use_doorbell)
+			return adev->wb.wb[ring->wptr_offs];
+		else
+			return RREG32_SOC15(UVD, 0, mmUVD_RB_WPTR);
+	} else {
+		if (ring->use_doorbell)
+			return adev->wb.wb[ring->wptr_offs];
+		else
+			return RREG32_SOC15(UVD, 0, mmUVD_RB_WPTR2);
+	}
 }
 
 /**
@@ -971,10 +992,21 @@ static void vcn_v2_5_enc_ring_set_wptr(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
 
-	if (ring == &adev->vcn.ring_enc[0])
-		WREG32_SOC15(UVD, 0, mmUVD_RB_WPTR, lower_32_bits(ring->wptr));
-	else
-		WREG32_SOC15(UVD, 0, mmUVD_RB_WPTR2, lower_32_bits(ring->wptr));
+	if (ring == &adev->vcn.ring_enc[0]) {
+		if (ring->use_doorbell) {
+			adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
+			WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
+		} else {
+			WREG32_SOC15(UVD, 0, mmUVD_RB_WPTR, lower_32_bits(ring->wptr));
+		}
+	} else {
+		if (ring->use_doorbell) {
+			adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
+			WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
+		} else {
+			WREG32_SOC15(UVD, 0, mmUVD_RB_WPTR2, lower_32_bits(ring->wptr));
+		}
+	}
 }
 
 static const struct amdgpu_ring_funcs vcn_v2_5_enc_ring_vm_funcs = {
@@ -1032,7 +1064,10 @@ static uint64_t vcn_v2_5_jpeg_ring_get_wptr(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
 
-	return RREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_WPTR);
+	if (ring->use_doorbell)
+		return adev->wb.wb[ring->wptr_offs];
+	else
+		return RREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_WPTR);
 }
 
 /**
@@ -1046,7 +1081,12 @@ static void vcn_v2_5_jpeg_ring_set_wptr(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
 
-	WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_WPTR, lower_32_bits(ring->wptr));
+	if (ring->use_doorbell) {
+		adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
+		WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
+	} else {
+		WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_WPTR, lower_32_bits(ring->wptr));
+	}
 }
 
 static const struct amdgpu_ring_funcs vcn_v2_5_jpeg_ring_vm_funcs = {
drivers/gpu/drm/amd/amdgpu/vega20_reg_init.c +4 -0
index 0db84386252a..79223188bd47 100644
@@ -85,6 +85,10 @@ void vega20_doorbell_index_init(struct amdgpu_device *adev)
 	adev->doorbell_index.uvd_vce.vce_ring2_3 = AMDGPU_VEGA20_DOORBELL64_VCE_RING2_3;
 	adev->doorbell_index.uvd_vce.vce_ring4_5 = AMDGPU_VEGA20_DOORBELL64_VCE_RING4_5;
 	adev->doorbell_index.uvd_vce.vce_ring6_7 = AMDGPU_VEGA20_DOORBELL64_VCE_RING6_7;
+	adev->doorbell_index.vcn.vcn_ring0_1 = AMDGPU_VEGA20_DOORBELL64_VCN0_1;
+	adev->doorbell_index.vcn.vcn_ring2_3 = AMDGPU_VEGA20_DOORBELL64_VCN2_3;
+	adev->doorbell_index.vcn.vcn_ring4_5 = AMDGPU_VEGA20_DOORBELL64_VCN4_5;
+	adev->doorbell_index.vcn.vcn_ring6_7 = AMDGPU_VEGA20_DOORBELL64_VCN6_7;
 
 	adev->doorbell_index.first_non_cp = AMDGPU_VEGA20_DOORBELL64_FIRST_NON_CP;
 	adev->doorbell_index.last_non_cp = AMDGPU_VEGA20_DOORBELL64_LAST_NON_CP;
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#66 lore raw

PATCH 68/102 drm/amdgpu/powerplay: add arcturus ppt functions

In reply to a message not in the archive

From: Chengming Gui <[email protected]>

add arcturus_ppsmc.h rcturus_ppt.c and arcturus_ppt.h files.

This is the initial power management support for Arcturus.

---
 drivers/gpu/drm/amd/powerplay/Makefile        |   2 +-
 drivers/gpu/drm/amd/powerplay/arcturus_ppt.c  | 124 ++++++++++++++++++
 drivers/gpu/drm/amd/powerplay/arcturus_ppt.h  |  28 ++++
 .../gpu/drm/amd/powerplay/inc/amdgpu_smu.h    |  10 ++
 .../drm/amd/powerplay/inc/arcturus_ppsmc.h    | 120 +++++++++++++++++
 drivers/gpu/drm/amd/powerplay/smu_v11_0.c     |   4 +
 6 files changed, 287 insertions(+), 1 deletion(-)
 create mode 100644 drivers/gpu/drm/amd/powerplay/arcturus_ppt.c
 create mode 100644 drivers/gpu/drm/amd/powerplay/arcturus_ppt.h
 create mode 100644 drivers/gpu/drm/amd/powerplay/inc/arcturus_ppsmc.h
drivers/gpu/drm/amd/powerplay/Makefile +1 -1
index 727c5cff231c..e05a7e3d6d8d 100644
@@ -35,7 +35,7 @@ AMD_POWERPLAY = $(addsuffix /Makefile,$(addprefix $(FULL_AMD_PATH)/powerplay/,$(
 
 include $(AMD_POWERPLAY)
 
-POWER_MGR = amd_powerplay.o amdgpu_smu.o smu_v11_0.o vega20_ppt.o navi10_ppt.o
+POWER_MGR = amd_powerplay.o amdgpu_smu.o smu_v11_0.o vega20_ppt.o arcturus_ppt.o navi10_ppt.o
 
 AMD_PP_POWER = $(addprefix $(AMD_PP_PATH)/,$(POWER_MGR))
 
drivers/gpu/drm/amd/powerplay/arcturus_ppt.c new file +124 -0

124 lines not on this page · raw message

drivers/gpu/drm/amd/powerplay/arcturus_ppt.h new file +28 -0
new file mode 100644
index 000000000000..7b808d091b31
@@ -0,0 +1,28 @@
+/*
+ * Copyright 2019 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __ARCTURUS_PPT_H__
+#define __ARCTURUS_PPT_H__
+
+extern void arcturus_set_ppt_funcs(struct smu_context *smu);
+
+#endif
drivers/gpu/drm/amd/powerplay/inc/amdgpu_smu.h +10 -0
index 135a3236f7ed..79e34097eb0f 100644
@@ -242,6 +242,16 @@ enum smu_message_type
 	SMU_MSG_PowerDownJpeg,
 	SMU_MSG_BacoAudioD3PME,
 	SMU_MSG_ArmD3,
+	SMU_MSG_RunGfxDcBtc,
+	SMU_MSG_RunSocDcBtc,
+	SMU_MSG_SetMemoryChannelEnable,
+	SMU_MSG_SetDfSwitchType,
+	SMU_MSG_GetVoltageByDpm,
+	SMU_MSG_GetVoltageByDpmOverdrive,
+	SMU_MSG_PowerUpVcn0,
+	SMU_MSG_PowerDownVcn01,
+	SMU_MSG_PowerUpVcn1,
+	SMU_MSG_PowerDownVcn1,
 	SMU_MSG_MAX_COUNT,
 };
 
drivers/gpu/drm/amd/powerplay/inc/arcturus_ppsmc.h new file +120 -0

120 lines not on this page · raw message

drivers/gpu/drm/amd/powerplay/smu_v11_0.c +4 -0
index c60899de88bb..d0764ba2705c 100644
@@ -30,6 +30,7 @@
 #include "soc15_common.h"
 #include "atom.h"
 #include "vega20_ppt.h"
+#include "arcturus_ppt.h"
 #include "navi10_ppt.h"
 
 #include "asic_reg/thm/thm_11_0_2_offset.h"
@@ -1793,6 +1794,9 @@ void smu_v11_0_set_smu_funcs(struct smu_context *smu)
 	case CHIP_VEGA20:
 		vega20_set_ppt_funcs(smu);
 		break;
+	case CHIP_ARCTURUS:
+		arcturus_set_ppt_funcs(smu);
+		break;
 	case CHIP_NAVI10:
 	case CHIP_NAVI14:
 		navi10_set_ppt_funcs(smu);
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#67 lore raw

PATCH 69/102 drm/amdgpu/powerplay: add smu11 driver interface for arcturus. (v2)

In reply to a message not in the archive

From: Chengming Gui <[email protected]>

add smu11_driver_if_arcturus.h file.

v2: add license, fix header guard (Alex)

---
 .../powerplay/inc/smu11_driver_if_arcturus.h  | 878 ++++++++++++++++++
 1 file changed, 878 insertions(+)
 create mode 100644 drivers/gpu/drm/amd/powerplay/inc/smu11_driver_if_arcturus.h
drivers/gpu/drm/amd/powerplay/inc/smu11_driver_if_arcturus.h new file +878 -0
new file mode 100644
index 000000000000..7a9969e075d4
@@ -0,0 +1,878 @@
+/*
+ * Copyright 2019 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+
+#ifndef SMU11_DRIVER_IF_ARCTURUS_H
+#define SMU11_DRIVER_IF_ARCTURUS_H
+
+// *** IMPORTANT ***
+// SMU TEAM: Always increment the interface version if
+// any structure is changed in this file
+#define SMU11_DRIVER_IF_VERSION 0x06
+
+#define PPTABLE_ARCTURUS_SMU_VERSION 3
+
+#define NUM_GFXCLK_DPM_LEVELS  16
+#define NUM_VCLK_DPM_LEVELS    8
+#define NUM_DCLK_DPM_LEVELS    8
+#define NUM_MP0CLK_DPM_LEVELS  2
+#define NUM_SOCCLK_DPM_LEVELS  8
+#define NUM_UCLK_DPM_LEVELS    4
+#define NUM_FCLK_DPM_LEVELS    8
+#define NUM_XGMI_LEVELS        2
+
+#define MAX_GFXCLK_DPM_LEVEL  (NUM_GFXCLK_DPM_LEVELS  - 1)
+#define MAX_VCLK_DPM_LEVEL    (NUM_VCLK_DPM_LEVELS    - 1)
+#define MAX_DCLK_DPM_LEVEL    (NUM_DCLK_DPM_LEVELS    - 1)
+#define MAX_MP0CLK_DPM_LEVEL  (NUM_MP0CLK_DPM_LEVELS  - 1)
+#define MAX_SOCCLK_DPM_LEVEL  (NUM_SOCCLK_DPM_LEVELS  - 1)
+#define MAX_UCLK_DPM_LEVEL    (NUM_UCLK_DPM_LEVELS    - 1)
+#define MAX_FCLK_DPM_LEVEL    (NUM_FCLK_DPM_LEVELS    - 1)
+#define MAX_XGMI_LEVEL        (NUM_XGMI_LEVELS        - 1)
+
+// Feature Control Defines
+// DPM
+#define FEATURE_DPM_PREFETCHER_BIT      0
+#define FEATURE_DPM_GFXCLK_BIT          1
+#define FEATURE_DPM_UCLK_BIT            2
+#define FEATURE_DPM_SOCCLK_BIT          3
+#define FEATURE_DPM_FCLK_BIT            4
+#define FEATURE_DPM_MP0CLK_BIT          5
+#define FEATURE_DPM_XGMI_BIT            6
+// Idle
+#define FEATURE_DS_GFXCLK_BIT           7
+#define FEATURE_DS_SOCCLK_BIT           8
+#define FEATURE_DS_LCLK_BIT             9
+#define FEATURE_DS_FCLK_BIT             10
+#define FEATURE_DS_UCLK_BIT             11
+#define FEATURE_GFX_ULV_BIT             12
+#define FEATURE_DPM_VCN_BIT             13
+#define FEATURE_RSMU_SMN_CG_BIT         14
+#define FEATURE_WAFL_CG_BIT             15
+// Throttler/Response
+#define FEATURE_PPT_BIT                 16
+#define FEATURE_TDC_BIT                 17
+#define FEATURE_APCC_PLUS_BIT           18
+#define FEATURE_VR0HOT_BIT              19
+#define FEATURE_VR1HOT_BIT              20
+#define FEATURE_FW_CTF_BIT              21
+#define FEATURE_FAN_CONTROL_BIT         22
+#define FEATURE_THERMAL_BIT             23
+// Other
+#define FEATURE_OUT_OF_BAND_MONITOR_BIT 24
+#define FEATURE_TEMP_DEPENDENT_VMIN_BIT 25
+
+#define FEATURE_SPARE_26_BIT            26
+#define FEATURE_SPARE_27_BIT            27
+#define FEATURE_SPARE_28_BIT            28
+#define FEATURE_SPARE_29_BIT            29
+#define FEATURE_SPARE_30_BIT            30
+#define FEATURE_SPARE_31_BIT            31
+#define FEATURE_SPARE_32_BIT            32
+#define FEATURE_SPARE_33_BIT            33
+#define FEATURE_SPARE_34_BIT            34
+#define FEATURE_SPARE_35_BIT            35
+#define FEATURE_SPARE_36_BIT            36
+#define FEATURE_SPARE_37_BIT            37
+#define FEATURE_SPARE_38_BIT            38
+#define FEATURE_SPARE_39_BIT            39
+#define FEATURE_SPARE_40_BIT            40
+#define FEATURE_SPARE_41_BIT            41
+#define FEATURE_SPARE_42_BIT            42
+#define FEATURE_SPARE_43_BIT            43
+#define FEATURE_SPARE_44_BIT            44
+#define FEATURE_SPARE_45_BIT            45
+#define FEATURE_SPARE_46_BIT            46
+#define FEATURE_SPARE_47_BIT            47
+#define FEATURE_SPARE_48_BIT            48
+#define FEATURE_SPARE_49_BIT            49
+#define FEATURE_SPARE_50_BIT            50
+#define FEATURE_SPARE_51_BIT            51
+#define FEATURE_SPARE_52_BIT            52
+#define FEATURE_SPARE_53_BIT            53
+#define FEATURE_SPARE_54_BIT            54
+#define FEATURE_SPARE_55_BIT            55
+#define FEATURE_SPARE_56_BIT            56
+#define FEATURE_SPARE_57_BIT            57
+#define FEATURE_SPARE_58_BIT            58
+#define FEATURE_SPARE_59_BIT            59
+#define FEATURE_SPARE_60_BIT            60
+#define FEATURE_SPARE_61_BIT            61
+#define FEATURE_SPARE_62_BIT            62
+#define FEATURE_SPARE_63_BIT            63
+
+#define NUM_FEATURES                    64
+
+
+#define FEATURE_DPM_PREFETCHER_MASK       (1 << FEATURE_DPM_PREFETCHER_BIT       )
+#define FEATURE_DPM_GFXCLK_MASK           (1 << FEATURE_DPM_GFXCLK_BIT           )
+#define FEATURE_DPM_UCLK_MASK             (1 << FEATURE_DPM_UCLK_BIT             )
+#define FEATURE_DPM_SOCCLK_MASK           (1 << FEATURE_DPM_SOCCLK_BIT           )
+#define FEATURE_DPM_FCLK_MASK             (1 << FEATURE_DPM_FCLK_BIT             )
+#define FEATURE_DPM_MP0CLK_MASK           (1 << FEATURE_DPM_MP0CLK_BIT           )
+#define FEATURE_DPM_XGMI_MASK             (1 << FEATURE_DPM_XGMI_BIT             )
+
+#define FEATURE_DS_GFXCLK_MASK            (1 << FEATURE_DS_GFXCLK_BIT            )
+#define FEATURE_DS_SOCCLK_MASK            (1 << FEATURE_DS_SOCCLK_BIT            )
+#define FEATURE_DS_LCLK_MASK              (1 << FEATURE_DS_LCLK_BIT              )
+#define FEATURE_DS_FCLK_MASK              (1 << FEATURE_DS_FCLK_BIT              )
+#define FEATURE_DS_LCLK_MASK              (1 << FEATURE_DS_LCLK_BIT              )
+#define FEATURE_GFX_ULV_MASK              (1 << FEATURE_GFX_ULV_BIT              )
+#define FEATURE_VCN_PG_MASK               (1 << FEATURE_VCN_PG_BIT               )
+#define FEATURE_RSMU_SMN_CG_MASK          (1 << FEATURE_RSMU_SMN_CG_BIT          )
+#define FEATURE_WAFL_CG_MASK              (1 << FEATURE_WAFL_CG_BIT              )
+
+#define FEATURE_PPT_MASK                  (1 << FEATURE_PPT_BIT                  )
+#define FEATURE_TDC_MASK                  (1 << FEATURE_TDC_BIT                  )
+#define FEATURE_APCC_MASK                 (1 << FEATURE_APCC_BIT                 )
+#define FEATURE_VR0HOT_MASK               (1 << FEATURE_VR0HOT_BIT               )
+#define FEATURE_VR1HOT_MASK               (1 << FEATURE_VR1HOT_BIT               )
+#define FEATURE_FW_CTF_MASK               (1 << FEATURE_FW_CTF_BIT               )
+#define FEATURE_FAN_CONTROL_MASK          (1 << FEATURE_FAN_CONTROL_BIT          )
+#define FEATURE_THERMAL_MASK              (1 << FEATURE_THERMAL_BIT              )
+
+#define FEATURE_OUT_OF_BAND_MONITOR_MASK  (1 << EATURE_OUT_OF_BAND_MONITOR_BIT   )
+#define FEATURE_TEMP_DEPENDENT_VMIN_MASK  (1 << FEATURE_TEMP_DEPENDENT_VMIN_MASK )
+
+
+//FIXME need updating
+// Debug Overrides Bitmask
+#define DPM_OVERRIDE_DISABLE_UCLK_PID               0x00000001
+#define DPM_OVERRIDE_ENABLE_VOLT_LINK_VCN_FCLK      0x00000002
+
+// I2C Config Bit Defines
+#define I2C_CONTROLLER_ENABLED           1
+#define I2C_CONTROLLER_DISABLED          0
+
+// VR Mapping Bit Defines
+#define VR_MAPPING_VR_SELECT_MASK  0x01
+#define VR_MAPPING_VR_SELECT_SHIFT 0x00
+
+#define VR_MAPPING_PLANE_SELECT_MASK  0x02
+#define VR_MAPPING_PLANE_SELECT_SHIFT 0x01
+
+// PSI Bit Defines
+#define PSI_SEL_VR0_PLANE0_PSI0  0x01
+#define PSI_SEL_VR0_PLANE0_PSI1  0x02
+#define PSI_SEL_VR0_PLANE1_PSI0  0x04
+#define PSI_SEL_VR0_PLANE1_PSI1  0x08
+#define PSI_SEL_VR1_PLANE0_PSI0  0x10
+#define PSI_SEL_VR1_PLANE0_PSI1  0x20
+#define PSI_SEL_VR1_PLANE1_PSI0  0x40
+#define PSI_SEL_VR1_PLANE1_PSI1  0x80
+
+// Throttler Control/Status Bits
+#define THROTTLER_PADDING_BIT      0
+#define THROTTLER_TEMP_EDGE_BIT    1
+#define THROTTLER_TEMP_HOTSPOT_BIT 2
+#define THROTTLER_TEMP_MEM_BIT     3
+#define THROTTLER_TEMP_VR_GFX_BIT  4
+#define THROTTLER_TEMP_VR_MEM_BIT  5
+#define THROTTLER_TEMP_VR_SOC_BIT  6
+#define THROTTLER_TDC_GFX_BIT      7
+#define THROTTLER_TDC_SOC_BIT      8
+#define THROTTLER_PPT0_BIT         9
+#define THROTTLER_PPT1_BIT         10
+#define THROTTLER_PPT2_BIT         11
+#define THROTTLER_PPT3_BIT         12
+#define THROTTLER_PPM_BIT          13
+#define THROTTLER_FIT_BIT          14
+#define THROTTLER_APCC_BIT         15
+
+// Table transfer status
+#define TABLE_TRANSFER_OK         0x0
+#define TABLE_TRANSFER_FAILED     0xFF
+#define TABLE_TRANSFER_PENDING    0xAB
+
+// Workload bits
+#define WORKLOAD_PPLIB_DEFAULT_BIT        0
+#define WORKLOAD_PPLIB_POWER_SAVING_BIT   1
+#define WORKLOAD_PPLIB_VIDEO_BIT          2
+#define WORKLOAD_PPLIB_COMPUTE_BIT        3
+#define WORKLOAD_PPLIB_CUSTOM_BIT         4
+#define WORKLOAD_PPLIB_COUNT              5
+
+//XGMI performance states
+#define XGMI_STATE_D0                      1
+#define XGMI_STATE_D3                      0
+
+#define NUM_I2C_CONTROLLERS                8
+
+#define I2C_CONTROLLER_ENABLED             1
+#define I2C_CONTROLLER_DISABLED            0
+
+#define MAX_SW_I2C_COMMANDS                8
+
+typedef enum {
+  I2C_CONTROLLER_PORT_0 = 0,  //CKSVII2C0
+  I2C_CONTROLLER_PORT_1 = 1,  //CKSVII2C1
+  I2C_CONTROLLER_PORT_COUNT,
+} I2cControllerPort_e;
+
+typedef enum {
+  I2C_CONTROLLER_NAME_VR_GFX = 0,
+  I2C_CONTROLLER_NAME_VR_SOC,
+  I2C_CONTROLLER_NAME_VR_MEM,
+  I2C_CONTROLLER_NAME_SPARE,
+  I2C_CONTROLLER_NAME_COUNT,
+} I2cControllerName_e;
+
+typedef enum {
+  I2C_CONTROLLER_THROTTLER_TYPE_NONE = 0,
+  I2C_CONTROLLER_THROTTLER_VR_GFX,
+  I2C_CONTROLLER_THROTTLER_VR_SOC,
+  I2C_CONTROLLER_THROTTLER_VR_MEM,
+  I2C_CONTROLLER_THROTTLER_COUNT,
+} I2cControllerThrottler_e;
+
+typedef enum {
+  I2C_CONTROLLER_PROTOCOL_VR_0,
+  I2C_CONTROLLER_PROTOCOL_VR_1,
+  I2C_CONTROLLER_PROTOCOL_TMP_0,
+  I2C_CONTROLLER_PROTOCOL_TMP_1,
+  I2C_CONTROLLER_PROTOCOL_SPARE_0,
+  I2C_CONTROLLER_PROTOCOL_SPARE_1,
+  I2C_CONTROLLER_PROTOCOL_COUNT,
+} I2cControllerProtocol_e;
+
+typedef struct {
+  uint8_t   Enabled;
+  uint8_t   Speed;
+  uint8_t   Padding[2];
+  uint32_t  SlaveAddress;
+  uint8_t   ControllerPort;
+  uint8_t   ControllerName;
+  uint8_t   ThermalThrotter;
+  uint8_t   I2cProtocol;
+} I2cControllerConfig_t;
+
+typedef enum {
+  I2C_PORT_SVD_SCL = 0,
+  I2C_PORT_GPIO,
+} I2cPort_e;
+
+typedef enum {
+  I2C_SPEED_FAST_50K = 0,      //50  Kbits/s
+  I2C_SPEED_FAST_100K,         //100 Kbits/s
+  I2C_SPEED_FAST_400K,         //400 Kbits/s
+  I2C_SPEED_FAST_PLUS_1M,      //1   Mbits/s (in fast mode)
+  I2C_SPEED_HIGH_1M,           //1   Mbits/s (in high speed mode)
+  I2C_SPEED_HIGH_2M,           //2.3 Mbits/s
+  I2C_SPEED_COUNT,
+} I2cSpeed_e;
+
+typedef enum {
+  I2C_CMD_READ = 0,
+  I2C_CMD_WRITE,
+  I2C_CMD_COUNT,
+} I2cCmdType_e;
+
+#define CMDCONFIG_STOP_BIT      0
+#define CMDCONFIG_RESTART_BIT   1
+
+#define CMDCONFIG_STOP_MASK     (1 << CMDCONFIG_STOP_BIT)
+#define CMDCONFIG_RESTART_MASK  (1 << CMDCONFIG_RESTART_BIT)
+
+typedef struct {
+  uint8_t RegisterAddr; ////only valid for write, ignored for read
+  uint8_t Cmd;  //Read(0) or Write(1)
+  uint8_t Data;  //Return data for read. Data to send for write
+  uint8_t CmdConfig; //Includes whether associated command should have a stop or restart command
+} SwI2cCmd_t; //SW I2C Command Table
+
+typedef struct {
+  uint8_t     I2CcontrollerPort; //CKSVII2C0(0) or //CKSVII2C1(1)
+  uint8_t     I2CSpeed;          //Slow(0) or Fast(1)
+  uint16_t    SlaveAddress;
+  uint8_t     NumCmds;           //Number of commands
+  uint8_t     Padding[3];
+
+  SwI2cCmd_t  SwI2cCmds[MAX_SW_I2C_COMMANDS];
+
+  uint32_t     MmHubPadding[8]; // SMU internal use
+
+} SwI2cRequest_t; // SW I2C Request Table
+
+//D3HOT sequences
+//sequence codes from spec: atlvp4p01.amd.com:1677@//gpu/doc/soc_arch/spec/feature/BACO/Navi/Navi2x/
+typedef enum {
+  BACO_SEQUENCE,
+  MSR_SEQUENCE,
+  BAMACO_SEQUENCE,
+  ULPS_SEQUENCE,
+  D3HOT_SEQUENCE_COUNT,
+}D3HOTSequence_e;
+
+//THis is aligned with RSMU PGFSM Register Mapping
+typedef enum {
+  PG_DYNAMIC_MODE = 0,
+  PG_STATIC_MODE,
+} PowerGatingMode_e;
+
+//This is aligned with RSMU PGFSM Register Mapping
+typedef enum {
+  PG_POWER_DOWN = 0,
+  PG_POWER_UP,
+} PowerGatingSettings_e;
+
+typedef struct {
+  uint32_t a;  // store in IEEE float format in this variable
+  uint32_t b;  // store in IEEE float format in this variable
+  uint32_t c;  // store in IEEE float format in this variable
+} QuadraticInt_t;
+
+typedef struct {
+  uint32_t m;  // store in IEEE float format in this variable
+  uint32_t b;  // store in IEEE float format in this variable
+} LinearInt_t;
+
+typedef struct {
+  uint32_t a;  // store in IEEE float format in this variable
+  uint32_t b;  // store in IEEE float format in this variable
+  uint32_t c;  // store in IEEE float format in this variable
+} DroopInt_t;
+
+typedef enum {
+  GFXCLK_SOURCE_PLL = 0,
+  GFXCLK_SOURCE_AFLL,
+  GFXCLK_SOURCE_COUNT,
+} GfxclkSrc_e;
+
+typedef enum {
+  PPCLK_GFXCLK,
+  PPCLK_VCLK,
+  PPCLK_DCLK,
+  PPCLK_SOCCLK,
+  PPCLK_UCLK,
+  PPCLK_FCLK,
+  PPCLK_COUNT,
+} PPCLK_e;
+
+typedef enum {
+  TEMP_EDGE,
+  TEMP_HOTSPOT,
+  TEMP_MEM,
+  TEMP_VR_GFX,
+  TEMP_VR_SOC,
+  TEMP_VR_MEM,
+  TEMP_COUNT
+} TEMP_TYPE_e;
+
+typedef enum  {
+  PPT_THROTTLER_PPT0,
+  PPT_THROTTLER_PPT1,
+  PPT_THROTTLER_PPT2,
+  PPT_THROTTLER_PPT3,
+  PPT_THROTTLER_COUNT
+} PPT_THROTTLER_e;
+
+typedef enum {
+  VOLTAGE_MODE_AVFS = 0,
+  VOLTAGE_MODE_AVFS_SS,
+  VOLTAGE_MODE_SS,
+  VOLTAGE_MODE_COUNT,
+} VOLTAGE_MODE_e;
+
+typedef enum {
+  AVFS_VOLTAGE_GFX = 0,
+  AVFS_VOLTAGE_SOC,
+  AVFS_VOLTAGE_COUNT,
+} AVFS_VOLTAGE_TYPE_e;
+
+typedef enum {
+  GPIO_INT_POLARITY_ACTIVE_LOW = 0,
+  GPIO_INT_POLARITY_ACTIVE_HIGH,
+} GpioIntPolarity_e;
+
+typedef enum {
+  MEMORY_TYPE_GDDR6 = 0,
+  MEMORY_TYPE_HBM,
+} MemoryType_e;
+
+typedef enum {
+  PWR_CONFIG_TDP = 0,
+  PWR_CONFIG_TGP,
+  PWR_CONFIG_TCP_ESTIMATED,
+  PWR_CONFIG_TCP_MEASURED,
+} PwrConfig_e;
+
+typedef enum {
+  XGMI_LINK_RATE_12 = 0,  // 12Gbps
+  XGMI_LINK_RATE_16,      // 16Gbps
+  XGMI_LINK_RATE_22,      // 22Gbps
+  XGMI_LINK_RATE_25,      // 25Gbps
+  XGMI_LINK_RATE_COUNT
+} XGMI_LINK_RATE_e;
+
+typedef enum {
+  XGMI_LINK_WIDTH_2 = 0, // x2
+  XGMI_LINK_WIDTH_4,     // x4
+  XGMI_LINK_WIDTH_8,     // x8
+  XGMI_LINK_WIDTH_16,    // x16
+  XGMI_LINK_WIDTH_COUNT
+} XGMI_LINK_WIDTH_e;
+
+typedef struct {
+  uint8_t        VoltageMode;         // 0 - AVFS only, 1- min(AVFS,SS), 2-SS only
+  uint8_t        SnapToDiscrete;      // 0 - Fine grained DPM, 1 - Discrete DPM
+  uint8_t        NumDiscreteLevels;   // Set to 2 (Fmin, Fmax) when using fine grained DPM, otherwise set to # discrete levels used
+  uint8_t        padding;
+  LinearInt_t    ConversionToAvfsClk; // Transfer function to AVFS Clock (GHz->GHz)
+  QuadraticInt_t SsCurve;             // Slow-slow curve (GHz->V)
+  uint16_t       SsFmin;              // Fmin for SS curve. If SS curve is selected, will use V@SSFmin for F <= Fmin
+  uint16_t       Padding16;
+} DpmDescriptor_t;
+
+typedef struct {
+  uint32_t Version;
+
+  // SECTION: Feature Enablement
+  uint32_t FeaturesToRun[2];
+
+  // SECTION: Infrastructure Limits
+  uint16_t SocketPowerLimitAc[PPT_THROTTLER_COUNT];
+  uint16_t SocketPowerLimitAcTau[PPT_THROTTLER_COUNT];
+  uint16_t TdcLimitSoc;             // Amps
+  uint16_t TdcLimitSocTau;          // Time constant of LPF in ms
+  uint16_t TdcLimitGfx;             // Amps
+  uint16_t TdcLimitGfxTau;          // Time constant of LPF in ms
+
+  uint16_t TedgeLimit;              // Celcius
+  uint16_t ThotspotLimit;           // Celcius
+  uint16_t TmemLimit;               // Celcius
+  uint16_t Tvr_gfxLimit;            // Celcius
+  uint16_t Tvr_memLimit;            // Celcius
+  uint16_t Tvr_socLimit;            // Celcius
+  uint32_t FitLimit;                // Failures in time (failures per million parts over the defined lifetime)
+
+  uint16_t PpmPowerLimit;           // Switch this this power limit when temperature is above PpmTempThreshold
+  uint16_t PpmTemperatureThreshold;
+
+  // SECTION: Throttler settings
+  uint32_t ThrottlerControlMask;   // See Throtter masks defines
+
+  // SECTION: ULV Settings
+  uint16_t  UlvVoltageOffsetGfx; // In mV(Q2)
+  uint16_t  UlvPadding;          // Padding
+
+  uint8_t  UlvGfxclkBypass;  // 1 to turn off/bypass Gfxclk during ULV, 0 to leave Gfxclk on during ULV
+  uint8_t  Padding234[3];
+
+  // SECTION: Voltage Control Parameters
+  uint16_t     MinVoltageGfx;     // In mV(Q2) Minimum Voltage ("Vmin") of VDD_GFX
+  uint16_t     MinVoltageSoc;     // In mV(Q2) Minimum Voltage ("Vmin") of VDD_SOC
+  uint16_t     MaxVoltageGfx;     // In mV(Q2) Maximum Voltage allowable of VDD_GFX
+  uint16_t     MaxVoltageSoc;     // In mV(Q2) Maximum Voltage allowable of VDD_SOC
+
+  uint16_t     LoadLineResistanceGfx;   // In mOhms with 8 fractional bits
+  uint16_t     LoadLineResistanceSoc;   // In mOhms with 8 fractional bits
+
+  //SECTION: DPM Config 1
+  DpmDescriptor_t DpmDescriptor[PPCLK_COUNT];
+
+  uint16_t       FreqTableGfx      [NUM_GFXCLK_DPM_LEVELS  ];     // In MHz
+  uint16_t       FreqTableVclk     [NUM_VCLK_DPM_LEVELS    ];     // In MHz
+  uint16_t       FreqTableDclk     [NUM_DCLK_DPM_LEVELS    ];     // In MHz
+  uint16_t       FreqTableSocclk   [NUM_SOCCLK_DPM_LEVELS  ];     // In MHz
+  uint16_t       FreqTableUclk     [NUM_UCLK_DPM_LEVELS    ];     // In MHz
+  uint16_t       FreqTableFclk     [NUM_FCLK_DPM_LEVELS    ];     // In MHz
+
+  uint32_t       Paddingclks[16];
+
+  // SECTION: DPM Config 2
+  uint16_t       Mp0clkFreq        [NUM_MP0CLK_DPM_LEVELS];       // in MHz
+  uint16_t       Mp0DpmVoltage     [NUM_MP0CLK_DPM_LEVELS];       // mV(Q2)
+
+  // GFXCLK DPM
+  uint16_t        GfxclkFidle;          // In MHz
+  uint16_t        GfxclkSlewRate;       // for PLL babystepping???
+  uint8_t         Padding567[4];
+  uint16_t        GfxclkDsMaxFreq;      // In MHz
+  uint8_t         GfxclkSource;         // 0 = PLL, 1 = AFLL
+  uint8_t         Padding456;
+
+  // GFXCLK Thermal DPM (formerly 'Boost' Settings)
+  uint16_t     EnableTdpm;
+  uint16_t     TdpmHighHystTemperature;
+  uint16_t     TdpmLowHystTemperature;
+  uint16_t     GfxclkFreqHighTempLimit; // High limit on GFXCLK when temperature is high, for reliability.
+
+  // SECTION: Fan Control
+  uint16_t     FanStopTemp;          //Celcius
+  uint16_t     FanStartTemp;         //Celcius
+
+  uint16_t     FanGainEdge;
+  uint16_t     FanGainHotspot;
+  uint16_t     FanGainVrGfx;
+  uint16_t     FanGainVrSoc;
+  uint16_t     FanGainVrMem;
+  uint16_t     FanGainHbm;
+  uint16_t     FanPwmMin;
+  uint16_t     FanAcousticLimitRpm;
+  uint16_t     FanThrottlingRpm;
+  uint16_t     FanMaximumRpm;
+  uint16_t     FanTargetTemperature;
+  uint16_t     FanTargetGfxclk;
+  uint8_t      FanZeroRpmEnable;
+  uint8_t      FanTachEdgePerRev;
+  uint8_t      FanTempInputSelect;
+  uint8_t      padding8_Fan;
+
+  // The following are AFC override parameters. Leave at 0 to use FW defaults.
+  int16_t      FuzzyFan_ErrorSetDelta;
+  int16_t      FuzzyFan_ErrorRateSetDelta;
+  int16_t      FuzzyFan_PwmSetDelta;
+  uint16_t     FuzzyFan_Reserved;
+
+
+  // SECTION: AVFS
+  // Overrides
+  uint8_t           OverrideAvfsGb[AVFS_VOLTAGE_COUNT];
+  uint8_t           Padding8_Avfs[2];
+
+  QuadraticInt_t    qAvfsGb[AVFS_VOLTAGE_COUNT];              // GHz->V Override of fused curve
+  DroopInt_t        dBtcGbGfxPll;       // GHz->V BtcGb
+  DroopInt_t        dBtcGbGfxAfll;        // GHz->V BtcGb
+  DroopInt_t        dBtcGbSoc;            // GHz->V BtcGb
+  LinearInt_t       qAgingGb[AVFS_VOLTAGE_COUNT];          // GHz->V
+
+  QuadraticInt_t    qStaticVoltageOffset[AVFS_VOLTAGE_COUNT]; // GHz->V
+
+  uint16_t          DcTol[AVFS_VOLTAGE_COUNT];            // mV Q2
+
+  uint8_t           DcBtcEnabled[AVFS_VOLTAGE_COUNT];
+  uint8_t           Padding8_GfxBtc[2];
+
+  uint16_t          DcBtcMin[AVFS_VOLTAGE_COUNT];       // mV Q2
+  uint16_t          DcBtcMax[AVFS_VOLTAGE_COUNT];       // mV Q2
+
+  uint16_t          DcBtcGb[AVFS_VOLTAGE_COUNT];        // mV Q2
+
+  uint16_t          SsFmin[10]; // PPtable value to function similar to VFTFmin for SS Curve; Size is PPCLK_COUNT rounded to nearest multiple of 2
+
+  // SECTION: XGMI
+  uint8_t           XgmiLinkSpeed   [NUM_XGMI_LEVELS];
+  uint8_t           XgmiLinkWidth   [NUM_XGMI_LEVELS];
+
+  uint16_t          XgmiFclkFreq    [NUM_XGMI_LEVELS];
+  uint16_t          XgmiSocVoltage  [NUM_XGMI_LEVELS];
+
+  // Temperature Dependent Vmin
+  uint16_t     VDDGFX_TVmin;       //Celcius
+  uint16_t     VDDSOC_TVmin;       //Celcius
+  uint16_t     VDDGFX_Vmin_HiTemp; // mV Q2
+  uint16_t     VDDGFX_Vmin_LoTemp; // mV Q2
+  uint16_t     VDDSOC_Vmin_HiTemp; // mV Q2
+  uint16_t     VDDSOC_Vmin_LoTemp; // mV Q2
+
+  uint16_t     VDDGFX_TVminHystersis; // Celcius
+  uint16_t     VDDSOC_TVminHystersis; // Celcius
+
+
+  // SECTION: Advanced Options
+  uint32_t          DebugOverrides;
+  QuadraticInt_t    ReservedEquation0;
+  QuadraticInt_t    ReservedEquation1;
+  QuadraticInt_t    ReservedEquation2;
+  QuadraticInt_t    ReservedEquation3;
+
+  uint16_t     MinVoltageUlvGfx; // In mV(Q2)  Minimum Voltage ("Vmin") of VDD_GFX in ULV mode
+  uint16_t     PaddingUlv;       // Padding
+
+  // Total Power configuration, use defines from PwrConfig_e
+  uint8_t      TotalPowerConfig;    //0-TDP, 1-TGP, 2-TCP Estimated, 3-TCP Measured
+  uint8_t      TotalPowerSpare1;
+  uint16_t     TotalPowerSpare2;
+
+  // APCC Settings
+  uint16_t     PccThresholdLow;
+  uint16_t     PccThresholdHigh;
+  uint32_t     PaddingAPCC[6];  //FIXME pending SPEC
+
+  // SECTION: Reserved
+  uint32_t     Reserved[11];
+
+  // SECTION: BOARD PARAMETERS
+
+  // SVI2 Board Parameters
+  uint16_t     MaxVoltageStepGfx; // In mV(Q2) Max voltage step that SMU will request. Multiple steps are taken if voltage change exceeds this value.
+  uint16_t     MaxVoltageStepSoc; // In mV(Q2) Max voltage step that SMU will request. Multiple steps are taken if voltage change exceeds this value.
+
+  uint8_t      VddGfxVrMapping;     // Use VR_MAPPING* bitfields
+  uint8_t      VddSocVrMapping;     // Use VR_MAPPING* bitfields
+  uint8_t      VddMemVrMapping;     // Use VR_MAPPING* bitfields
+  uint8_t      BoardVrMapping;      // Use VR_MAPPING* bitfields
+
+  uint8_t      GfxUlvPhaseSheddingMask; // set this to 1 to set PSI0/1 to 1 in ULV mode
+  uint8_t      ExternalSensorPresent; // External RDI connected to TMON (aka TEMP IN)
+  uint8_t      Padding8_V[2];
+
+  // Telemetry Settings
+  uint16_t     GfxMaxCurrent;   // in Amps
+  int8_t       GfxOffset;       // in Amps
+  uint8_t      Padding_TelemetryGfx;
+
+  uint16_t     SocMaxCurrent;   // in Amps
+  int8_t       SocOffset;       // in Amps
+  uint8_t      Padding_TelemetrySoc;
+
+  uint16_t     MemMaxCurrent;   // in Amps
+  int8_t       MemOffset;       // in Amps
+  uint8_t      Padding_TelemetryMem;
+
+  uint16_t     BoardMaxCurrent;   // in Amps
+  int8_t       BoardOffset;       // in Amps
+  uint8_t      Padding_TelemetryBoardInput;
+
+  // GPIO Settings
+  uint8_t      VR0HotGpio;      // GPIO pin configured for VR0 HOT event
+  uint8_t      VR0HotPolarity;  // GPIO polarity for VR0 HOT event
+  uint8_t      VR1HotGpio;      // GPIO pin configured for VR1 HOT event
+  uint8_t      VR1HotPolarity;  // GPIO polarity for VR1 HOT event
+
+  // GFXCLK PLL Spread Spectrum
+  uint8_t      PllGfxclkSpreadEnabled;   // on or off
+  uint8_t      PllGfxclkSpreadPercent;   // Q4.4
+  uint16_t     PllGfxclkSpreadFreq;      // kHz
+
+  // UCLK Spread Spectrum
+  uint8_t      UclkSpreadEnabled;   // on or off
+  uint8_t      UclkSpreadPercent;   // Q4.4
+  uint16_t     UclkSpreadFreq;      // kHz
+
+  // FCLK Spread Spectrum
+  uint8_t      FclkSpreadEnabled;   // on or off
+  uint8_t      FclkSpreadPercent;   // Q4.4
+  uint16_t     FclkSpreadFreq;      // kHz
+
+  // GFXCLK Fll Spread Spectrum
+  uint8_t      FllGfxclkSpreadEnabled;   // on or off
+  uint8_t      FllGfxclkSpreadPercent;   // Q4.4
+  uint16_t     FllGfxclkSpreadFreq;      // kHz
+
+  // I2C Controller Structure
+  I2cControllerConfig_t  I2cControllers[NUM_I2C_CONTROLLERS];
+
+  // Memory section
+  uint32_t     MemoryChannelEnabled; // For DRAM use only, Max 32 channels enabled bit mask.
+
+  uint8_t      DramBitWidth; // For DRAM use only.  See Dram Bit width type defines
+  uint8_t      PaddingMem[3];
+
+  // Total board power
+  uint16_t     TotalBoardPower;     //Only needed for TCP Estimated case, where TCP = TGP+Total Board Power
+  uint16_t     BoardPadding;
+
+  uint32_t     BoardReserved[10];
+
+  // Padding for MMHUB - do not modify this
+  uint32_t     MmHubPadding[8]; // SMU internal use
+
+} PPTable_t;
+
+typedef struct {
+  // Time constant parameters for clock averages in ms
+  uint16_t     GfxclkAverageLpfTau;
+  uint16_t     SocclkAverageLpfTau;
+  uint16_t     UclkAverageLpfTau;
+  uint16_t     GfxActivityLpfTau;
+  uint16_t     UclkActivityLpfTau;
+
+  uint16_t     Padding;
+
+  // Padding - ignore
+  uint32_t     MmHubPadding[8]; // SMU internal use
+} DriverSmuConfig_t;
+
+typedef struct {
+  uint16_t CurrClock[PPCLK_COUNT];
+  uint16_t AverageGfxclkFrequency;
+  uint16_t AverageSocclkFrequency;
+  uint16_t AverageUclkFrequency  ;
+  uint16_t AverageGfxActivity    ;
+  uint16_t AverageUclkActivity   ;
+  uint8_t  CurrSocVoltageOffset  ;
+  uint8_t  CurrGfxVoltageOffset  ;
+  uint8_t  CurrMemVidOffset      ;
+  uint8_t  Padding8              ;
+  uint16_t CurrSocketPower       ;
+  uint16_t TemperatureEdge       ;
+  uint16_t TemperatureHotspot    ;
+  uint16_t TemperatureHBM        ;
+  uint16_t TemperatureVrGfx      ;
+  uint16_t TemperatureVrSoc      ;
+  uint16_t TemperatureVrMem      ;
+  uint32_t ThrottlerStatus       ;
+
+  // Padding - ignore
+  uint32_t     MmHubPadding[7]; // SMU internal use
+} SmuMetrics_t;
+
+
+typedef struct {
+  uint16_t avgPsmCount[45];
+  uint16_t minPsmCount[45];
+  float    avgPsmVoltage[45];
+  float    minPsmVoltage[45];
+
+  uint16_t avgScsPsmCount;
+  uint16_t minScsPsmCount;
+  float    avgScsPsmVoltage;
+  float    minScsPsmVoltage;
+
+  uint32_t MmHubPadding[6]; // SMU internal use
+} AvfsDebugTable_t;
+
+typedef struct {
+  uint8_t  AvfsVersion;
+  uint8_t  Padding;
+  uint8_t  AvfsEn[AVFS_VOLTAGE_COUNT];
+
+  uint8_t  OverrideVFT[AVFS_VOLTAGE_COUNT];
+  uint8_t  OverrideAvfsGb[AVFS_VOLTAGE_COUNT];
+
+  uint8_t  OverrideTemperatures[AVFS_VOLTAGE_COUNT];
+  uint8_t  OverrideVInversion[AVFS_VOLTAGE_COUNT];
+  uint8_t  OverrideP2V[AVFS_VOLTAGE_COUNT];
+  uint8_t  OverrideP2VCharzFreq[AVFS_VOLTAGE_COUNT];
+
+  int32_t VFT0_m1[AVFS_VOLTAGE_COUNT]; // Q8.24
+  int32_t VFT0_m2[AVFS_VOLTAGE_COUNT]; // Q12.12
+  int32_t VFT0_b[AVFS_VOLTAGE_COUNT];  // Q32
+
+  int32_t VFT1_m1[AVFS_VOLTAGE_COUNT]; // Q8.16
+  int32_t VFT1_m2[AVFS_VOLTAGE_COUNT]; // Q12.12
+  int32_t VFT1_b[AVFS_VOLTAGE_COUNT];  // Q32
+
+  int32_t VFT2_m1[AVFS_VOLTAGE_COUNT]; // Q8.16
+  int32_t VFT2_m2[AVFS_VOLTAGE_COUNT]; // Q12.12
+  int32_t VFT2_b[AVFS_VOLTAGE_COUNT];  // Q32
+
+  int32_t AvfsGb0_m1[AVFS_VOLTAGE_COUNT]; // Q8.24
+  int32_t AvfsGb0_m2[AVFS_VOLTAGE_COUNT]; // Q12.12
+  int32_t AvfsGb0_b[AVFS_VOLTAGE_COUNT];  // Q32
+
+  int32_t AcBtcGb_m1[AVFS_VOLTAGE_COUNT]; // Q8.24
+  int32_t AcBtcGb_m2[AVFS_VOLTAGE_COUNT]; // Q12.12
+  int32_t AcBtcGb_b[AVFS_VOLTAGE_COUNT];  // Q32
+
+  uint32_t AvfsTempCold[AVFS_VOLTAGE_COUNT];
+  uint32_t AvfsTempMid[AVFS_VOLTAGE_COUNT];
+  uint32_t AvfsTempHot[AVFS_VOLTAGE_COUNT];
+
+  uint32_t VInversion[AVFS_VOLTAGE_COUNT]; // in mV with 2 fractional bits
+
+
+  int32_t P2V_m1[AVFS_VOLTAGE_COUNT]; // Q8.24
+  int32_t P2V_m2[AVFS_VOLTAGE_COUNT]; // Q12.12
+  int32_t P2V_b[AVFS_VOLTAGE_COUNT];  // Q32
+
+  uint32_t P2VCharzFreq[AVFS_VOLTAGE_COUNT]; // in 10KHz units
+
+  uint32_t EnabledAvfsModules;
+
+  uint32_t MmHubPadding[7]; // SMU internal use
+} AvfsFuseOverride_t;
+
+/* NOT CURRENTLY USED
+typedef struct {
+  uint8_t   Gfx_ActiveHystLimit;
+  uint8_t   Gfx_IdleHystLimit;
+  uint8_t   Gfx_FPS;
+  uint8_t   Gfx_MinActiveFreqType;
+  uint8_t   Gfx_BoosterFreqType;
+  uint8_t   Gfx_MinFreqStep;                // Minimum delta between current and target frequeny in order for FW to change clock.
+  uint8_t   Gfx_UseRlcBusy;
+  uint8_t   PaddingGfx[3];
+  uint16_t  Gfx_MinActiveFreq;              // MHz
+  uint16_t  Gfx_BoosterFreq;                // MHz
+  uint16_t  Gfx_PD_Data_time_constant;      // Time constant of PD controller in ms
+  uint32_t  Gfx_PD_Data_limit_a;            // Q16
+  uint32_t  Gfx_PD_Data_limit_b;            // Q16
+  uint32_t  Gfx_PD_Data_limit_c;            // Q16
+  uint32_t  Gfx_PD_Data_error_coeff;        // Q16
+  uint32_t  Gfx_PD_Data_error_rate_coeff;   // Q16
+
+  uint8_t   Mem_ActiveHystLimit;
+  uint8_t   Mem_IdleHystLimit;
+  uint8_t   Mem_FPS;
+  uint8_t   Mem_MinActiveFreqType;
+  uint8_t   Mem_BoosterFreqType;
+  uint8_t   Mem_MinFreqStep;                // Minimum delta between current and target frequeny in order for FW to change clock.
+  uint8_t   Mem_UseRlcBusy;
+  uint8_t   PaddingMem[3];
+  uint16_t  Mem_MinActiveFreq;              // MHz
+  uint16_t  Mem_BoosterFreq;                // MHz
+  uint16_t  Mem_PD_Data_time_constant;      // Time constant of PD controller in ms
+  uint32_t  Mem_PD_Data_limit_a;            // Q16
+  uint32_t  Mem_PD_Data_limit_b;            // Q16
+  uint32_t  Mem_PD_Data_limit_c;            // Q16
+  uint32_t  Mem_PD_Data_error_coeff;        // Q16
+  uint32_t  Mem_PD_Data_error_rate_coeff;   // Q16
+
+  uint32_t  Mem_UpThreshold_Limit;          // Q16
+  uint8_t   Mem_UpHystLimit;
+  uint8_t   Mem_DownHystLimit;
+  uint16_t  Mem_Fps;
+
+  uint32_t  MmHubPadding[8]; // SMU internal use
+} DpmActivityMonitorCoeffInt_t;
+*/
+
+// These defines are used with the following messages:
+// SMC_MSG_TransferTableDram2Smu
+// SMC_MSG_TransferTableSmu2Dram
+#define TABLE_PPTABLE                 0
+#define TABLE_AVFS                    1
+#define TABLE_AVFS_PSM_DEBUG          2
+#define TABLE_AVFS_FUSE_OVERRIDE      3
+#define TABLE_PMSTATUSLOG             4
+#define TABLE_SMU_METRICS             5
+#define TABLE_DRIVER_SMU_CONFIG       6
+//#define TABLE_ACTIVITY_MONITOR_COEFF  7
+#define TABLE_OVERDRIVE               7
+#define TABLE_WAFL_XGMI_TOPOLOGY      8
+#define TABLE_COUNT                   9
+
+// These defines are used with the SMC_MSG_SetUclkFastSwitch message.
+typedef enum {
+  DF_SWITCH_TYPE_FAST = 0,
+  DF_SWITCH_TYPE_SLOW,
+  DF_SWITCH_TYPE_COUNT,
+} DF_SWITCH_TYPE_e;
+
+typedef enum {
+  DRAM_BIT_WIDTH_DISABLED = 0,
+  DRAM_BIT_WIDTH_X_8,
+  DRAM_BIT_WIDTH_X_16,
+  DRAM_BIT_WIDTH_X_32,
+  DRAM_BIT_WIDTH_X_64, // NOT USED.
+  DRAM_BIT_WIDTH_X_128,
+  DRAM_BIT_WIDTH_COUNT,
+} DRAM_BIT_WIDTH_TYPE_e;
+
+#define REMOVE_FMAX_MARGIN_BIT     0x0
+#define REMOVE_DCTOL_MARGIN_BIT    0x1
+#define REMOVE_PLATFORM_MARGIN_BIT 0x2
+
+#endif
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#68 lore raw

PATCH 70/102 drm/amd/powerplay: get smc firmware and pptable

In reply to a message not in the archive

From: Chengming Gui <[email protected]>

get smc firmware and pptable for arcturus

---
 drivers/gpu/drm/amd/powerplay/smu_v11_0.c | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)
drivers/gpu/drm/amd/powerplay/smu_v11_0.c +8 -3
index d0764ba2705c..7d723a5e4135 100644
@@ -43,6 +43,7 @@
 #include "asic_reg/smuio/smuio_11_0_0_sh_mask.h"
 
 MODULE_FIRMWARE("amdgpu/vega20_smc.bin");
+MODULE_FIRMWARE("amdgpu/arcturus_smc.bin");
 MODULE_FIRMWARE("amdgpu/navi10_smc.bin");
 MODULE_FIRMWARE("amdgpu/navi14_smc.bin");
 
@@ -153,6 +154,9 @@ static int smu_v11_0_init_microcode(struct smu_context *smu)
 	case CHIP_VEGA20:
 		chip_name = "vega20";
 		break;
+	case CHIP_ARCTURUS:
+		chip_name = "arcturus";
+		break;
 	case CHIP_NAVI10:
 		chip_name = "navi10";
 		break;
@@ -204,7 +208,7 @@ static int smu_v11_0_load_microcode(struct smu_context *smu)
 	uint32_t i;
 	uint32_t mp1_fw_flags;
 
-	hdr = (const struct smc_firmware_header_v1_0 *)	adev->pm.fw->data;
+	hdr = (const struct smc_firmware_header_v1_0 *) adev->pm.fw->data;
 	src = (const uint32_t *)(adev->pm.fw->data +
 		le32_to_cpu(hdr->header.ucode_array_offset_bytes));
 
@@ -291,7 +295,8 @@ static int smu_v11_0_set_pptable_v2_0(struct smu_context *smu, void **table, uin
 	return 0;
 }
 
-static int smu_v11_0_set_pptable_v2_1(struct smu_context *smu, void **table, uint32_t *size, uint32_t pptable_id)
+static int smu_v11_0_set_pptable_v2_1(struct smu_context *smu, void **table,
+				      uint32_t *size, uint32_t pptable_id)
 {
 	struct amdgpu_device *adev = smu->adev;
 	const struct smc_firmware_header_v2_1 *v2_1;
@@ -1739,7 +1744,7 @@ static const struct smu_funcs smu_v11_0_funcs = {
 	.send_smc_msg = smu_v11_0_send_msg,
 	.send_smc_msg_with_param = smu_v11_0_send_msg_with_param,
 	.read_smc_arg = smu_v11_0_read_arg,
-	.setup_pptable= smu_v11_0_setup_pptable,
+	.setup_pptable = smu_v11_0_setup_pptable,
 	.init_smc_tables = smu_v11_0_init_smc_tables,
 	.fini_smc_tables = smu_v11_0_fini_smc_tables,
 	.init_power = smu_v11_0_init_power,
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#69 lore raw

PATCH 71/102 drm/amdgpu: skip get/update xgmi topology info when no psp exists

In reply to a message not in the archive

From: Le Ma <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c | 47 +++++++++++++-----------
 1 file changed, 25 insertions(+), 22 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c +25 -22
index b024070ad1cf..41e648d56eca 100644
@@ -296,7 +296,7 @@ int amdgpu_xgmi_add_device(struct amdgpu_device *adev)
 	struct amdgpu_xgmi	*entry;
 	struct amdgpu_device *tmp_adev = NULL;
 
-	int count = 0, ret = -EINVAL;
+	int count = 0, ret = 0;
 
 	if (!adev->gmc.xgmi.supported)
 		return 0;
@@ -337,29 +337,32 @@ int amdgpu_xgmi_add_device(struct amdgpu_device *adev)
 	top_info->num_nodes = count;
 	hive->number_devices = count;
 
-	list_for_each_entry(tmp_adev, &hive->device_list, gmc.xgmi.head) {
-		/* update node list for other device in the hive */
-		if (tmp_adev != adev) {
-			top_info = &tmp_adev->psp.xgmi_context.top_info;
-			top_info->nodes[count - 1].node_id = adev->gmc.xgmi.node_id;
-			top_info->num_nodes = count;
+	if (amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_PSP)) {
+		list_for_each_entry(tmp_adev, &hive->device_list, gmc.xgmi.head) {
+			/* update node list for other device in the hive */
+			if (tmp_adev != adev) {
+				top_info = &tmp_adev->psp.xgmi_context.top_info;
+				top_info->nodes[count - 1].node_id =
+					adev->gmc.xgmi.node_id;
+				top_info->num_nodes = count;
+			}
+			ret = amdgpu_xgmi_update_topology(hive, tmp_adev);
+			if (ret)
+				goto exit;
 		}
-		ret = amdgpu_xgmi_update_topology(hive, tmp_adev);
-		if (ret)
-			goto exit;
-	}
 
-	/* get latest topology info for each device from psp */
-	list_for_each_entry(tmp_adev, &hive->device_list, gmc.xgmi.head) {
-		ret = psp_xgmi_get_topology_info(&tmp_adev->psp, count,
-				&tmp_adev->psp.xgmi_context.top_info);
-		if (ret) {
-			dev_err(tmp_adev->dev,
-				"XGMI: Get topology failure on device %llx, hive %llx, ret %d",
-				tmp_adev->gmc.xgmi.node_id,
-				tmp_adev->gmc.xgmi.hive_id, ret);
-			/* To do : continue with some node failed or disable the whole hive */
-			goto exit;
+		/* get latest topology info for each device from psp */
+		list_for_each_entry(tmp_adev, &hive->device_list, gmc.xgmi.head) {
+			ret = psp_xgmi_get_topology_info(&tmp_adev->psp, count,
+					&tmp_adev->psp.xgmi_context.top_info);
+			if (ret) {
+				dev_err(tmp_adev->dev,
+					"XGMI: Get topology failure on device %llx, hive %llx, ret %d",
+					tmp_adev->gmc.xgmi.node_id,
+					tmp_adev->gmc.xgmi.hive_id, ret);
+				/* To do : continue with some node failed or disable the whole hive */
+				goto exit;
+			}
 		}
 	}
 
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#70 lore raw

PATCH 72/102 drm/amdgpu: set system aperture to cover whole FB region in mmhub v9.4

In reply to a message not in the archive

From: Le Ma <[email protected]>

In XGMI configuration, the FB region covers vram region from peer device, adjust system aperture to cover all of them

---
 drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)
drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c +2 -3
index 6b7cdaadbd70..c0eb8f0a2182 100644
@@ -114,12 +114,11 @@ static void mmhub_v9_4_init_system_aperture_regs(struct amdgpu_device *adev,
 	WREG32_SOC15_OFFSET(MMHUB, 0,
 			    mmVMSHAREDVC0_MC_VM_SYSTEM_APERTURE_LOW_ADDR,
 			    hubid * MMHUB_INSTANCE_REGISTER_OFFSET,
-			    min(adev->gmc.vram_start, adev->gmc.agp_start)
-			    >> 18);
+			    min(adev->gmc.fb_start, adev->gmc.agp_start) >> 18);
 	WREG32_SOC15_OFFSET(MMHUB, 0,
 			    mmVMSHAREDVC0_MC_VM_SYSTEM_APERTURE_HIGH_ADDR,
 			    hubid * MMHUB_INSTANCE_REGISTER_OFFSET,
-			    max(adev->gmc.vram_end, adev->gmc.agp_end) >> 18);
+			    max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18);
 
 	/* Set default page address. */
 	value = adev->vram_scratch.gpu_addr - adev->gmc.vram_start +
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#71 lore raw

PATCH 73/102 drm/amdgpu: correct ip for mmHDP_READ_CACHE_INVALIDATE register access

In reply to a message not in the archive

From: Le Ma <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/soc15.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/soc15.c +1 -1
index 217afe23a2f6..dc09469f77c5 100644
@@ -702,7 +702,7 @@ static void soc15_invalidate_hdp(struct amdgpu_device *adev,
 				 struct amdgpu_ring *ring)
 {
 	if (!ring || !ring->funcs->emit_wreg)
-		WREG32_SOC15_NO_KIQ(NBIO, 0, mmHDP_READ_CACHE_INVALIDATE, 1);
+		WREG32_SOC15_NO_KIQ(HDP, 0, mmHDP_READ_CACHE_INVALIDATE, 1);
 	else
 		amdgpu_ring_emit_wreg(ring, SOC15_REG_OFFSET(
 			HDP, 0, mmHDP_READ_CACHE_INVALIDATE), 1);
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#72 lore raw

PATCH 74/102 drm/amdkfd: Set number of xgmi optimized SDMA engines for arcturus

In reply to a message not in the archive

From: Oak Zeng <[email protected]>

---
 drivers/gpu/drm/amd/amdkfd/kfd_device.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/amdkfd/kfd_device.c +2 -1
index 511bc2523f2f..cc22289c92d1 100644
@@ -329,7 +329,8 @@ static const struct kfd_device_info arcturus_device_info = {
 	.supports_cwsr = true,
 	.needs_iommu_device = false,
 	.needs_pci_atomics = false,
-	.num_sdma_engines = 8,
+	.num_sdma_engines = 2,
+	.num_xgmi_sdma_engines = 6,
 	.num_sdma_queues_per_engine = 8,
 };
 
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#73 lore raw

PATCH 75/102 drm/amdgpu/: add clientID for 2nd vcn instance

In reply to a message not in the archive

From: James Zhu <[email protected]>

add clientID for 2nd vcn instance, remove unused SOC15_IH_CLIENTID_SYSHUB.

---
 drivers/gpu/drm/amd/include/soc15_ih_clientid.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/include/soc15_ih_clientid.h +1 -1
index 0f386b2e1f4f..1794ad1fc4fc 100644
@@ -42,7 +42,6 @@ enum soc15_ih_clientid {
 	SOC15_IH_CLIENTID_SE1SH		= 0x0b,
 	SOC15_IH_CLIENTID_SE2SH		= 0x0c,
 	SOC15_IH_CLIENTID_SE3SH		= 0x0d,
-	SOC15_IH_CLIENTID_SYSHUB	= 0x0e,
 	SOC15_IH_CLIENTID_UVD1		= 0x0e,
 	SOC15_IH_CLIENTID_THM		= 0x0f,
 	SOC15_IH_CLIENTID_UVD		= 0x10,
@@ -64,6 +63,7 @@ enum soc15_ih_clientid {
 	SOC15_IH_CLIENTID_MAX,
 
 	SOC15_IH_CLIENTID_VCN		= SOC15_IH_CLIENTID_UVD,
+	SOC15_IH_CLIENTID_VCN1		= SOC15_IH_CLIENTID_UVD1,
 	SOC15_IH_CLIENTID_SDMA2		= SOC15_IH_CLIENTID_ACP,
 	SOC15_IH_CLIENTID_SDMA3		= SOC15_IH_CLIENTID_DCE,
 	SOC15_IH_CLIENTID_SDMA4		= SOC15_IH_CLIENTID_ISP,
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#74 lore raw

PATCH 76/102 drm/amdgpu/: add ucodeID for 2nd vcn instance

In reply to a message not in the archive

From: James Zhu <[email protected]>

add ucodeID for 2nd vcn instance

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h | 1 +
 1 file changed, 1 insertion(+)
drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h +1 -0
index c1fb6dc86440..2be106e81eda 100644
@@ -281,6 +281,7 @@ enum AMDGPU_UCODE_ID {
 	AMDGPU_UCODE_ID_UVD1,
 	AMDGPU_UCODE_ID_VCE,
 	AMDGPU_UCODE_ID_VCN,
+	AMDGPU_UCODE_ID_VCN1,
 	AMDGPU_UCODE_ID_DMCU_ERAM,
 	AMDGPU_UCODE_ID_DMCU_INTV,
 	AMDGPU_UCODE_ID_VCN0_RAM,
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#75 lore raw

PATCH 77/102 drm/amdgpu/: add doorbell assignment for 2nd vcn instance

In reply to a message not in the archive

From: James Zhu <[email protected]>

add doorbell assignment for 2nd vcn instance

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h +7 -2
index 790263dcc064..3fa18003d4d6 100644
@@ -130,13 +130,18 @@ typedef enum _AMDGPU_VEGA20_DOORBELL_ASSIGNMENT
 	AMDGPU_VEGA20_DOORBELL_IH                      = 0x178,
 	/* MMSCH: 392~407
 	 * overlap the doorbell assignment with VCN as they are  mutually exclusive
-	 * VCE engine's doorbell is 32 bit and two VCE ring share one QWORD
+	 * VCN engine's doorbell is 32 bit and two VCN ring share one QWORD
 	 */
-	AMDGPU_VEGA20_DOORBELL64_VCN0_1                  = 0x188, /* lower 32 bits for VNC0 and upper 32 bits for VNC1 */
+	AMDGPU_VEGA20_DOORBELL64_VCN0_1                  = 0x188, /* VNC0 */
 	AMDGPU_VEGA20_DOORBELL64_VCN2_3                  = 0x189,
 	AMDGPU_VEGA20_DOORBELL64_VCN4_5                  = 0x18A,
 	AMDGPU_VEGA20_DOORBELL64_VCN6_7                  = 0x18B,
 
+	AMDGPU_VEGA20_DOORBELL64_VCN8_9                  = 0x18C, /* VNC1 */
+	AMDGPU_VEGA20_DOORBELL64_VCNa_b                  = 0x18D,
+	AMDGPU_VEGA20_DOORBELL64_VCNc_d                  = 0x18E,
+	AMDGPU_VEGA20_DOORBELL64_VCNe_f                  = 0x18F,
+
 	AMDGPU_VEGA20_DOORBELL64_UVD_RING0_1             = 0x188,
 	AMDGPU_VEGA20_DOORBELL64_UVD_RING2_3             = 0x189,
 	AMDGPU_VEGA20_DOORBELL64_UVD_RING4_5             = 0x18A,
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#76 lore raw

PATCH 78/102 drm/amdgpu/: increase AMDGPU_MAX_RINGS to add 2nd vcn instance

In reply to a message not in the archive

From: James Zhu <[email protected]>

increase AMDGPU_MAX_RINGS to add 2nd vcn instance

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h +1 -1
index 4410c97ac9b7..930316e60155 100644
@@ -29,7 +29,7 @@
 #include <drm/drm_print.h>
 
 /* max number of rings */
-#define AMDGPU_MAX_RINGS		24
+#define AMDGPU_MAX_RINGS		28
 #define AMDGPU_MAX_GFX_RINGS		2
 #define AMDGPU_MAX_COMPUTE_RINGS	8
 #define AMDGPU_MAX_VCE_RINGS		3
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#77 lore raw

PATCH 79/102 drm/amdgpu: add vcn nbio doorbell range setting for 2nd vcn instance

In reply to a message not in the archive

From: James Zhu <[email protected]>

add vcn nbio doorbell range setting for 2nd vcn instance

---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h    |  2 +-
 drivers/gpu/drm/amd/amdgpu/nbio_v2_3.c |  2 +-
 drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c | 20 +++++++++++++++++---
 drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c  |  2 +-
 drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c  |  2 +-
 5 files changed, 21 insertions(+), 7 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu.h +1 -1
index e9c386cec2ad..4b514a44184c 100644
@@ -677,7 +677,7 @@ struct amdgpu_nbio_funcs {
 	void (*sdma_doorbell_range)(struct amdgpu_device *adev, int instance,
 			bool use_doorbell, int doorbell_index, int doorbell_size);
 	void (*vcn_doorbell_range)(struct amdgpu_device *adev, bool use_doorbell,
-			int doorbell_index);
+				   int doorbell_index, int instance);
 	void (*enable_doorbell_aperture)(struct amdgpu_device *adev,
 					 bool enable);
 	void (*enable_doorbell_selfring_aperture)(struct amdgpu_device *adev,
drivers/gpu/drm/amd/amdgpu/nbio_v2_3.c +1 -1
index 72efe32f2406..f5611c479e28 100644
@@ -92,7 +92,7 @@ static void nbio_v2_3_sdma_doorbell_range(struct amdgpu_device *adev, int instan
 }
 
 static void nbio_v2_3_vcn_doorbell_range(struct amdgpu_device *adev, bool use_doorbell,
-					 int doorbell_index)
+					 int doorbell_index, int instance)
 {
 	u32 reg = SOC15_REG_OFFSET(NBIO, 0, mmBIF_MMSCH0_DOORBELL_RANGE);
 
drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c +17 -3
index d8c9972a315b..910fffced43b 100644
@@ -42,6 +42,14 @@
 #define GPU_HDP_FLUSH_DONE__RSVD_ENG4_MASK	0x00010000L
 #define GPU_HDP_FLUSH_DONE__RSVD_ENG5_MASK	0x00020000L
 
+#define mmBIF_MMSCH1_DOORBELL_RANGE                     0x01dc
+#define mmBIF_MMSCH1_DOORBELL_RANGE_BASE_IDX            2
+//BIF_MMSCH1_DOORBELL_RANGE
+#define BIF_MMSCH1_DOORBELL_RANGE__OFFSET__SHIFT        0x2
+#define BIF_MMSCH1_DOORBELL_RANGE__SIZE__SHIFT          0x10
+#define BIF_MMSCH1_DOORBELL_RANGE__OFFSET_MASK          0x00000FFCL
+#define BIF_MMSCH1_DOORBELL_RANGE__SIZE_MASK            0x001F0000L
+
 static void nbio_v7_4_remap_hdp_registers(struct amdgpu_device *adev)
 {
 	WREG32_SOC15(NBIO, 0, mmREMAP_HDP_MEM_FLUSH_CNTL,
@@ -115,11 +123,17 @@ static void nbio_v7_4_sdma_doorbell_range(struct amdgpu_device *adev, int instan
 }
 
 static void nbio_v7_4_vcn_doorbell_range(struct amdgpu_device *adev, bool use_doorbell,
-					 int doorbell_index)
+					 int doorbell_index, int instance)
 {
-	u32 reg = SOC15_REG_OFFSET(NBIO, 0, mmBIF_MMSCH0_DOORBELL_RANGE);
+	u32 reg;
+	u32 doorbell_range;
+
+	if (instance)
+		reg = SOC15_REG_OFFSET(NBIO, 0, mmBIF_MMSCH1_DOORBELL_RANGE);
+	else
+		reg = SOC15_REG_OFFSET(NBIO, 0, mmBIF_MMSCH0_DOORBELL_RANGE);
 
-	u32 doorbell_range = RREG32(reg);
+	doorbell_range = RREG32(reg);
 
 	if (use_doorbell) {
 		doorbell_range = REG_SET_FIELD(doorbell_range,
drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c +1 -1
index 52a5e3ac9f55..2b7bb6364e5a 100644
@@ -245,7 +245,7 @@ static int vcn_v2_0_hw_init(void *handle)
 	int i, r;
 
 	adev->nbio_funcs->vcn_doorbell_range(adev, ring->use_doorbell,
-		ring->doorbell_index);
+					     ring->doorbell_index, 0);
 
 	ring->sched.ready = true;
 	r = amdgpu_ring_test_ring(ring);
drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c +1 -1
index 840737df19c0..75fdb6881ac0 100644
@@ -212,7 +212,7 @@ static int vcn_v2_5_hw_init(void *handle)
 	int i, r;
 
 	adev->nbio_funcs->vcn_doorbell_range(adev, ring->use_doorbell,
-		ring->doorbell_index);
+					     ring->doorbell_index, 0);
 
 	r = amdgpu_ring_test_ring(ring);
 	if (r) {
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#78 lore raw

PATCH 80/102 drm/amdgpu: modify amdgpu_vcn to support multiple instances

In reply to a message not in the archive

From: James Zhu <[email protected]>

Arcturus has dual-VCN. Need Restruct amdgpu_device::vcn to support multiple vcns. There are no any logical changes here

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c |   6 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c |   6 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c |  68 +++++++--------
 drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h |  24 ++++--
 drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c   | 110 ++++++++++++------------
 drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c   | 106 +++++++++++------------
 drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c   |  87 ++++++++++---------
 7 files changed, 210 insertions(+), 197 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c +3 -3
index a28a3d722ba2..86cc3092a5e1 100644
@@ -155,15 +155,15 @@ static int amdgpu_ctx_init(struct amdgpu_device *adev,
 			num_rings = 1;
 			break;
 		case AMDGPU_HW_IP_VCN_DEC:
-			rings[0] = &adev->vcn.ring_dec;
+			rings[0] = &adev->vcn.inst[0].ring_dec;
 			num_rings = 1;
 			break;
 		case AMDGPU_HW_IP_VCN_ENC:
-			rings[0] = &adev->vcn.ring_enc[0];
+			rings[0] = &adev->vcn.inst[0].ring_enc[0];
 			num_rings = 1;
 			break;
 		case AMDGPU_HW_IP_VCN_JPEG:
-			rings[0] = &adev->vcn.ring_jpeg;
+			rings[0] = &adev->vcn.inst[0].ring_jpeg;
 			num_rings = 1;
 			break;
 		}
drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c +3 -3
index e2c9d8d31ed8..b56d0f15e457 100644
@@ -405,7 +405,7 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev,
 		break;
 	case AMDGPU_HW_IP_VCN_DEC:
 		type = AMD_IP_BLOCK_TYPE_VCN;
-		if (adev->vcn.ring_dec.sched.ready)
+		if (adev->vcn.inst[0].ring_dec.sched.ready)
 			++num_rings;
 		ib_start_alignment = 16;
 		ib_size_alignment = 16;
@@ -413,14 +413,14 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev,
 	case AMDGPU_HW_IP_VCN_ENC:
 		type = AMD_IP_BLOCK_TYPE_VCN;
 		for (i = 0; i < adev->vcn.num_enc_rings; i++)
-			if (adev->vcn.ring_enc[i].sched.ready)
+			if (adev->vcn.inst[0].ring_enc[i].sched.ready)
 				++num_rings;
 		ib_start_alignment = 64;
 		ib_size_alignment = 1;
 		break;
 	case AMDGPU_HW_IP_VCN_JPEG:
 		type = AMD_IP_BLOCK_TYPE_VCN;
-		if (adev->vcn.ring_jpeg.sched.ready)
+		if (adev->vcn.inst[0].ring_jpeg.sched.ready)
 			++num_rings;
 		ib_start_alignment = 16;
 		ib_size_alignment = 16;
drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c +34 -34

176 lines not on this page · raw message

drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h +17 -7
index bfd8c3cea13a..d2fc47a954ab 100644
@@ -30,6 +30,8 @@
 #define AMDGPU_VCN_FIRMWARE_OFFSET	256
 #define AMDGPU_VCN_MAX_ENC_RINGS	3
 
+#define AMDGPU_MAX_VCN_INSTANCES	2
+
 #define VCN_DEC_CMD_FENCE		0x00000000
 #define VCN_DEC_CMD_TRAP		0x00000001
 #define VCN_DEC_CMD_WRITE_REG		0x00000004
@@ -155,30 +157,38 @@ struct amdgpu_vcn_reg{
 	unsigned	jpeg_pitch;
 };
 
-struct amdgpu_vcn {
+struct amdgpu_vcn_inst {
 	struct amdgpu_bo	*vcpu_bo;
 	void			*cpu_addr;
 	uint64_t		gpu_addr;
-	unsigned		fw_version;
 	void			*saved_bo;
-	struct delayed_work	idle_work;
-	const struct firmware	*fw;	/* VCN firmware */
 	struct amdgpu_ring	ring_dec;
 	struct amdgpu_ring	ring_enc[AMDGPU_VCN_MAX_ENC_RINGS];
 	struct amdgpu_ring	ring_jpeg;
 	struct amdgpu_irq_src	irq;
+	struct amdgpu_vcn_reg	external;
+};
+
+struct amdgpu_vcn {
+	unsigned		fw_version;
+	struct delayed_work	idle_work;
+	const struct firmware	*fw;	/* VCN firmware */
 	unsigned		num_enc_rings;
 	enum amd_powergating_state cur_state;
 	struct dpg_pause_state pause_state;
-	struct amdgpu_vcn_reg	internal, external;
-	int (*pause_dpg_mode)(struct amdgpu_device *adev,
-		struct dpg_pause_state *new_state);
 
 	bool			indirect_sram;
 	struct amdgpu_bo	*dpg_sram_bo;
 	void			*dpg_sram_cpu_addr;
 	uint64_t		dpg_sram_gpu_addr;
 	uint32_t		*dpg_sram_curr_addr;
+
+	uint8_t	num_vcn_inst;
+	struct amdgpu_vcn_inst	inst[AMDGPU_MAX_VCN_INSTANCES];
+	struct amdgpu_vcn_reg	internal;
+
+	int (*pause_dpg_mode)(struct amdgpu_device *adev,
+		struct dpg_pause_state *new_state);
 };
 
 int amdgpu_vcn_sw_init(struct amdgpu_device *adev);
drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c +56 -54

346 lines not on this page · raw message

drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c +53 -53

342 lines not on this page · raw message

drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c +44 -43

277 lines not on this page · raw message

Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#79 lore raw

PATCH 81/102 drm/amdgpu: add multiple instances support for Arcturus

In reply to a message not in the archive

From: James Zhu <[email protected]>

Arcturus has dual-VCN. Need add multiple instances support for Arcturus.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c |   20 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c |  166 ++--
 drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c   | 1178 ++++++++++++-----------
 3 files changed, 737 insertions(+), 627 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c +13 -7
index b56d0f15e457..38f6c77a593a 100644
@@ -405,23 +405,29 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev,
 		break;
 	case AMDGPU_HW_IP_VCN_DEC:
 		type = AMD_IP_BLOCK_TYPE_VCN;
-		if (adev->vcn.inst[0].ring_dec.sched.ready)
-			++num_rings;
+		for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+			if (adev->vcn.inst[i].ring_dec.sched.ready)
+				++num_rings;
+		}
 		ib_start_alignment = 16;
 		ib_size_alignment = 16;
 		break;
 	case AMDGPU_HW_IP_VCN_ENC:
 		type = AMD_IP_BLOCK_TYPE_VCN;
-		for (i = 0; i < adev->vcn.num_enc_rings; i++)
-			if (adev->vcn.inst[0].ring_enc[i].sched.ready)
-				++num_rings;
+		for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+			for (j = 0; j < adev->vcn.num_enc_rings; j++)
+				if (adev->vcn.inst[i].ring_enc[j].sched.ready)
+					++num_rings;
+		}
 		ib_start_alignment = 64;
 		ib_size_alignment = 1;
 		break;
 	case AMDGPU_HW_IP_VCN_JPEG:
 		type = AMD_IP_BLOCK_TYPE_VCN;
-		if (adev->vcn.inst[0].ring_jpeg.sched.ready)
-			++num_rings;
+		for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+			if (adev->vcn.inst[i].ring_jpeg.sched.ready)
+				++num_rings;
+		}
 		ib_start_alignment = 16;
 		ib_size_alignment = 16;
 		break;
drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c +89 -77

267 lines not on this page · raw message

drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c +635 -543

1,459 lines not on this page · raw message

Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#80 lore raw

PATCH 82/102 drm/amdgpu: add harvest support for Arcturus

In reply to a message not in the archive

From: James Zhu <[email protected]>

Add harvest support for Arcturus

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c |  9 ++++
 drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c | 11 +++++
 drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h |  4 ++
 drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c   | 55 ++++++++++++++++++++++++-
 4 files changed, 77 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c +9 -0
index 38f6c77a593a..6cf61e01041f 100644
@@ -406,6 +406,9 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev,
 	case AMDGPU_HW_IP_VCN_DEC:
 		type = AMD_IP_BLOCK_TYPE_VCN;
 		for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+			if (adev->uvd.harvest_config & (1 << i))
+				continue;
+
 			if (adev->vcn.inst[i].ring_dec.sched.ready)
 				++num_rings;
 		}
@@ -415,6 +418,9 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev,
 	case AMDGPU_HW_IP_VCN_ENC:
 		type = AMD_IP_BLOCK_TYPE_VCN;
 		for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+			if (adev->uvd.harvest_config & (1 << i))
+				continue;
+
 			for (j = 0; j < adev->vcn.num_enc_rings; j++)
 				if (adev->vcn.inst[i].ring_enc[j].sched.ready)
 					++num_rings;
@@ -425,6 +431,9 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev,
 	case AMDGPU_HW_IP_VCN_JPEG:
 		type = AMD_IP_BLOCK_TYPE_VCN;
 		for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+			if (adev->uvd.harvest_config & (1 << i))
+				continue;
+
 			if (adev->vcn.inst[i].ring_jpeg.sched.ready)
 				++num_rings;
 		}
drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c +11 -0
index 5016fc570211..be34bdc47174 100644
@@ -147,6 +147,9 @@ int amdgpu_vcn_sw_init(struct amdgpu_device *adev)
 		bo_size += AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8);
 
 	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
+
 		r = amdgpu_bo_create_kernel(adev, bo_size, PAGE_SIZE,
 						AMDGPU_GEM_DOMAIN_VRAM, &adev->vcn.inst[i].vcpu_bo,
 						&adev->vcn.inst[i].gpu_addr, &adev->vcn.inst[i].cpu_addr);
@@ -180,6 +183,8 @@ int amdgpu_vcn_sw_fini(struct amdgpu_device *adev)
 	}
 
 	for (j = 0; j < adev->vcn.num_vcn_inst; ++j) {
+		if (adev->vcn.harvest_config & (1 << j))
+			continue;
 		kvfree(adev->vcn.inst[j].saved_bo);
 
 		amdgpu_bo_free_kernel(&adev->vcn.inst[j].vcpu_bo,
@@ -208,6 +213,8 @@ int amdgpu_vcn_suspend(struct amdgpu_device *adev)
 	cancel_delayed_work_sync(&adev->vcn.idle_work);
 
 	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
 		if (adev->vcn.inst[i].vcpu_bo == NULL)
 			return 0;
 
@@ -230,6 +237,8 @@ int amdgpu_vcn_resume(struct amdgpu_device *adev)
 	int i;
 
 	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+		if (adev->vcn.harvest_config & (1 << i))
+			continue;
 		if (adev->vcn.inst[i].vcpu_bo == NULL)
 			return -EINVAL;
 
@@ -266,6 +275,8 @@ static void amdgpu_vcn_idle_work_handler(struct work_struct *work)
 	unsigned int i, j;
 
 	for (j = 0; j < adev->vcn.num_vcn_inst; ++j) {
+		if (adev->vcn.harvest_config & (1 << j))
+			continue;
 		for (i = 0; i < adev->vcn.num_enc_rings; ++i) {
 			fence[j] += amdgpu_fence_count_emitted(&adev->vcn.inst[j].ring_enc[i]);
 		}
drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h +4 -0
index d2fc47a954ab..38f0d53a6381 100644
@@ -32,6 +32,9 @@
 
 #define AMDGPU_MAX_VCN_INSTANCES	2
 
+#define AMDGPU_VCN_HARVEST_VCN0 (1 << 0)
+#define AMDGPU_VCN_HARVEST_VCN1 (1 << 1)
+
 #define VCN_DEC_CMD_FENCE		0x00000000
 #define VCN_DEC_CMD_TRAP		0x00000001
 #define VCN_DEC_CMD_WRITE_REG		0x00000004
@@ -187,6 +190,7 @@ struct amdgpu_vcn {
 	struct amdgpu_vcn_inst	inst[AMDGPU_MAX_VCN_INSTANCES];
 	struct amdgpu_vcn_reg	internal;
 
+	unsigned	harvest_config;
 	int (*pause_dpg_mode)(struct amdgpu_device *adev,
 		struct dpg_pause_state *new_state);
 };
drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c +53 -2

178 lines not on this page · raw message

Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#81 lore raw

PATCH 83/102 drm/amdgpu: assign fb_start/end in mmhub v9.4 interface

In reply to a message not in the archive

From: Le Ma <[email protected]>

Align with mmhub v1.0.

---
 drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c | 7 +++++++
 1 file changed, 7 insertions(+)
drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c +7 -0
index c0eb8f0a2182..33b0de54a5da 100644
@@ -39,10 +39,17 @@ u64 mmhub_v9_4_get_fb_location(struct amdgpu_device *adev)
 {
 	/* The base should be same b/t 2 mmhubs on Acrturus. Read one here. */
 	u64 base = RREG32_SOC15(MMHUB, 0, mmVMSHAREDVC0_MC_VM_FB_LOCATION_BASE);
+	u64 top = RREG32_SOC15(MMHUB, 0, mmVMSHAREDVC0_MC_VM_FB_LOCATION_TOP);
 
 	base &= VMSHAREDVC0_MC_VM_FB_LOCATION_BASE__FB_BASE_MASK;
 	base <<= 24;
 
+	top &= VMSHAREDVC0_MC_VM_FB_LOCATION_TOP__FB_TOP_MASK;
+	top <<= 24;
+
+	adev->gmc.fb_start = base;
+	adev->gmc.fb_end = top;
+
 	return base;
 }
 
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#82 lore raw

PATCH 84/102 drm/amdgpu: add pci DID for Arcturus GL-XL.

In reply to a message not in the archive

From: Feifei Xu <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +4 -0
index 0e80a500aa22..fc8f82ac1382 100644
@@ -994,6 +994,10 @@ static const struct pci_device_id pciidlist[] = {
 	/* Raven */
 	{0x1002, 0x15dd, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_RAVEN|AMD_IS_APU},
 	{0x1002, 0x15d8, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_RAVEN|AMD_IS_APU},
+	/* Arcturus */
+	{0x1002, 0x738C, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ARCTURUS},
+	{0x1002, 0x7388, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ARCTURUS},
+	{0x1002, 0x738E, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ARCTURUS},
 	/* Navi10 */
 	{0x1002, 0x7310, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_NAVI10},
 	{0x1002, 0x7312, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_NAVI10},
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#83 lore raw

PATCH 85/102 drm/amdgpu: add arct sdma golden settings

In reply to a message not in the archive

From: Hawking Zhang <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c | 45 ++++++++++++++++++++++++++
 1 file changed, 45 insertions(+)
drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c +45 -0
index cd142e4304a8..166b1999dcbf 100644
@@ -31,6 +31,18 @@
 #include "sdma0/sdma0_4_2_sh_mask.h"
 #include "sdma1/sdma1_4_2_offset.h"
 #include "sdma1/sdma1_4_2_sh_mask.h"
+#include "sdma2/sdma2_4_2_2_offset.h"
+#include "sdma2/sdma2_4_2_2_sh_mask.h"
+#include "sdma3/sdma3_4_2_2_offset.h"
+#include "sdma3/sdma3_4_2_2_sh_mask.h"
+#include "sdma4/sdma4_4_2_2_offset.h"
+#include "sdma4/sdma4_4_2_2_sh_mask.h"
+#include "sdma5/sdma5_4_2_2_offset.h"
+#include "sdma5/sdma5_4_2_2_sh_mask.h"
+#include "sdma6/sdma6_4_2_2_offset.h"
+#include "sdma6/sdma6_4_2_2_sh_mask.h"
+#include "sdma7/sdma7_4_2_2_offset.h"
+#include "sdma7/sdma7_4_2_2_sh_mask.h"
 #include "hdp/hdp_4_0_offset.h"
 #include "sdma0/sdma0_4_1_default.h"
 
@@ -207,6 +219,34 @@ static const struct soc15_reg_golden golden_settings_sdma_rv2[] =
 	SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_GB_ADDR_CONFIG_READ, 0x0018773f, 0x00003001)
 };
 
+static const struct soc15_reg_golden golden_settings_sdma_arct[] =
+{
+	SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_CHICKEN_BITS, 0xfe931f07, 0x02831f07),
+	SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_GB_ADDR_CONFIG, 0x0000773f, 0x00004002),
+	SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_GB_ADDR_CONFIG_READ, 0x0000773f, 0x00004002),
+	SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_CHICKEN_BITS, 0xfe931f07, 0x02831f07),
+	SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_GB_ADDR_CONFIG, 0x0000773f, 0x00004002),
+	SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_GB_ADDR_CONFIG_READ, 0x0000773f, 0x00004002),
+	SOC15_REG_GOLDEN_VALUE(SDMA2, 0, mmSDMA2_CHICKEN_BITS, 0xfe931f07, 0x02831f07),
+	SOC15_REG_GOLDEN_VALUE(SDMA2, 0, mmSDMA2_GB_ADDR_CONFIG, 0x0000773f, 0x00004002),
+	SOC15_REG_GOLDEN_VALUE(SDMA2, 0, mmSDMA2_GB_ADDR_CONFIG_READ, 0x0000773f, 0x00004002),
+	SOC15_REG_GOLDEN_VALUE(SDMA3, 0, mmSDMA3_CHICKEN_BITS, 0xfe931f07, 0x02831f07),
+	SOC15_REG_GOLDEN_VALUE(SDMA3, 0, mmSDMA3_GB_ADDR_CONFIG, 0x0000773f, 0x00004002),
+	SOC15_REG_GOLDEN_VALUE(SDMA3, 0, mmSDMA3_GB_ADDR_CONFIG_READ, 0x0000773f, 0x00004002),
+	SOC15_REG_GOLDEN_VALUE(SDMA4, 0, mmSDMA4_CHICKEN_BITS, 0xfe931f07, 0x02831f07),
+	SOC15_REG_GOLDEN_VALUE(SDMA4, 0, mmSDMA4_GB_ADDR_CONFIG, 0x0000773f, 0x00004002),
+	SOC15_REG_GOLDEN_VALUE(SDMA4, 0, mmSDMA4_GB_ADDR_CONFIG_READ, 0x0000773f, 0x00004002),
+	SOC15_REG_GOLDEN_VALUE(SDMA5, 0, mmSDMA5_CHICKEN_BITS, 0xfe931f07, 0x02831f07),
+	SOC15_REG_GOLDEN_VALUE(SDMA5, 0, mmSDMA5_GB_ADDR_CONFIG, 0x0000773f, 0x00004002),
+	SOC15_REG_GOLDEN_VALUE(SDMA5, 0, mmSDMA5_GB_ADDR_CONFIG_READ, 0x0000773f, 0x00004002),
+	SOC15_REG_GOLDEN_VALUE(SDMA6, 0, mmSDMA6_CHICKEN_BITS, 0xfe931f07, 0x02831f07),
+	SOC15_REG_GOLDEN_VALUE(SDMA6, 0, mmSDMA6_GB_ADDR_CONFIG, 0x0000773f, 0x00004002),
+	SOC15_REG_GOLDEN_VALUE(SDMA6, 0, mmSDMA6_GB_ADDR_CONFIG_READ, 0x0000773f, 0x00004002),
+	SOC15_REG_GOLDEN_VALUE(SDMA7, 0, mmSDMA7_CHICKEN_BITS, 0xfe931f07, 0x02831f07),
+	SOC15_REG_GOLDEN_VALUE(SDMA7, 0, mmSDMA7_GB_ADDR_CONFIG, 0x0000773f, 0x00004002),
+	SOC15_REG_GOLDEN_VALUE(SDMA7, 0, mmSDMA7_GB_ADDR_CONFIG_READ, 0x0000773f, 0x00004002)
+};
+
 static u32 sdma_v4_0_get_reg_offset(struct amdgpu_device *adev,
 		u32 instance, u32 offset)
 {
@@ -315,6 +355,11 @@ static void sdma_v4_0_init_golden_registers(struct amdgpu_device *adev)
 						golden_settings_sdma1_4_2,
 						ARRAY_SIZE(golden_settings_sdma1_4_2));
 		break;
+	case CHIP_ARCTURUS:
+		soc15_program_register_sequence(adev,
+						golden_settings_sdma_arct,
+						ARRAY_SIZE(golden_settings_sdma_arct));
+		break;
 	case CHIP_RAVEN:
 		soc15_program_register_sequence(adev,
 						golden_settings_sdma_4_1,
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#84 lore raw

PATCH 86/102 drm/amdgpu: add arct gc golden settings

In reply to a message not in the archive

From: Hawking Zhang <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 30 +++++++++++++++++++++++++++
 1 file changed, 30 insertions(+)
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +30 -0
index 3ac3ee7fcb1a..560e607e7049 100644
@@ -107,6 +107,19 @@ MODULE_FIRMWARE("amdgpu/arcturus_mec.bin");
 MODULE_FIRMWARE("amdgpu/arcturus_mec2.bin");
 MODULE_FIRMWARE("amdgpu/arcturus_rlc.bin");
 
+#define mmTCP_CHAN_STEER_0_ARCT								0x0b03
+#define mmTCP_CHAN_STEER_0_ARCT_BASE_IDX							0
+#define mmTCP_CHAN_STEER_1_ARCT								0x0b04
+#define mmTCP_CHAN_STEER_1_ARCT_BASE_IDX							0
+#define mmTCP_CHAN_STEER_2_ARCT								0x0b09
+#define mmTCP_CHAN_STEER_2_ARCT_BASE_IDX							0
+#define mmTCP_CHAN_STEER_3_ARCT								0x0b0a
+#define mmTCP_CHAN_STEER_3_ARCT_BASE_IDX							0
+#define mmTCP_CHAN_STEER_4_ARCT								0x0b0b
+#define mmTCP_CHAN_STEER_4_ARCT_BASE_IDX							0
+#define mmTCP_CHAN_STEER_5_ARCT								0x0b0c
+#define mmTCP_CHAN_STEER_5_ARCT_BASE_IDX							0
+
 static const struct soc15_reg_golden golden_settings_gc_9_0[] =
 {
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_DEBUG2, 0xf00fffff, 0x00000400),
@@ -274,6 +287,18 @@ static const struct soc15_reg_golden golden_settings_gc_9_2_1_vg12[] =
 	SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_DEBUG, 0x00000000, 0x00008000)
 };
 
+static const struct soc15_reg_golden golden_settings_gc_9_4_1_arct[] =
+{
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmGB_ADDR_CONFIG, 0xffff77ff, 0x2a114042),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmTA_CNTL_AUX, 0xfffffeef, 0x10b0000),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmTCP_CHAN_STEER_0_ARCT, 0x3fffffff, 0x346f0a4e),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmTCP_CHAN_STEER_1_ARCT, 0x3fffffff, 0x1c642ca),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmTCP_CHAN_STEER_2_ARCT, 0x3fffffff, 0x26f45098),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmTCP_CHAN_STEER_3_ARCT, 0x3fffffff, 0x2ebd9fe3),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmTCP_CHAN_STEER_4_ARCT, 0x3fffffff, 0xb90f5b1),
+	SOC15_REG_GOLDEN_VALUE(GC, 0, mmTCP_CHAN_STEER_5_ARCT, 0x3ff, 0x135),
+};
+
 static const u32 GFX_RLC_SRM_INDEX_CNTL_ADDR_OFFSETS[] =
 {
 	mmRLC_SRM_INDEX_CNTL_ADDR_0 - mmRLC_SRM_INDEX_CNTL_ADDR_0,
@@ -343,6 +368,11 @@ static void gfx_v9_0_init_golden_registers(struct amdgpu_device *adev)
 						golden_settings_gc_9_0_vg20,
 						ARRAY_SIZE(golden_settings_gc_9_0_vg20));
 		break;
+	case CHIP_ARCTURUS:
+		soc15_program_register_sequence(adev,
+						golden_settings_gc_9_4_1_arct,
+						ARRAY_SIZE(golden_settings_gc_9_4_1_arct));
+		break;
 	case CHIP_RAVEN:
 		soc15_program_register_sequence(adev, golden_settings_gc_9_1,
 						ARRAY_SIZE(golden_settings_gc_9_1));
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#85 lore raw

PATCH 87/102 drm/amdgpu: init arct external rev id

In reply to a message not in the archive

From: Hawking Zhang <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/soc15.c | 1 +
 1 file changed, 1 insertion(+)
drivers/gpu/drm/amd/amdgpu/soc15.c +1 -0
index dc09469f77c5..f67ecf814c8c 100644
@@ -1016,6 +1016,7 @@ static int soc15_common_early_init(void *handle)
 		adev->asic_funcs = &vega20_asic_funcs;
 		adev->cg_flags = 0;
 		adev->pg_flags = 0;
+		adev->external_rev_id = adev->rev_id + 0x32;
 		break;
 	default:
 		/* FIXME: not supported yet */
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#86 lore raw

PATCH 88/102 drm/amdgpu: keep stolen memory for arct

In reply to a message not in the archive

From: Hawking Zhang <[email protected]>

Any dce register read back from arct is invalid. use hard code stolen memory for arct until we validate the s3.

---
 drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c +3 -1
index 11968fbe0d60..820cc6e0b51e 100644
@@ -641,6 +641,7 @@ static bool gmc_v9_0_keep_stolen_memory(struct amdgpu_device *adev)
 	switch (adev->asic_type) {
 	case CHIP_VEGA10:
 	case CHIP_RAVEN:
+	case CHIP_ARCTURUS:
 		return true;
 	case CHIP_VEGA12:
 	case CHIP_VEGA20:
@@ -941,7 +942,7 @@ static int gmc_v9_0_gart_init(struct amdgpu_device *adev)
 
 static unsigned gmc_v9_0_get_vbios_fb_size(struct amdgpu_device *adev)
 {
-	u32 d1vga_control = RREG32_SOC15(DCE, 0, mmD1VGA_CONTROL);
+	u32 d1vga_control;
 	unsigned size;
 
 	/*
@@ -951,6 +952,7 @@ static unsigned gmc_v9_0_get_vbios_fb_size(struct amdgpu_device *adev)
 	if (gmc_v9_0_keep_stolen_memory(adev))
 		return 9 * 1024 * 1024;
 
+	d1vga_control = RREG32_SOC15(DCE, 0, mmD1VGA_CONTROL);
 	if (REG_GET_FIELD(d1vga_control, D1VGA_CONTROL, D1VGA_MODE_ENABLE)) {
 		size = 9 * 1024 * 1024; /* reserve 8MB for vga emulator and 1 MB for FB */
 	} else {
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#87 lore raw

PATCH 89/102 drm/amdgpu: init gds config for arct

In reply to a message not in the archive

From: Hawking Zhang <[email protected]>

arct has 4KB gds (4 banks inside). there are 32 * 128 (CUs) so the max_wave_id should be 0xfff

---
 drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +4 -0
index 560e607e7049..2f553595a898 100644
@@ -5482,6 +5482,7 @@ static void gfx_v9_0_set_gds_init(struct amdgpu_device *adev)
 		adev->gds.gds_size = 0x10000;
 		break;
 	case CHIP_RAVEN:
+	case CHIP_ARCTURUS:
 		adev->gds.gds_size = 0x1000;
 		break;
 	default:
@@ -5503,6 +5504,9 @@ static void gfx_v9_0_set_gds_init(struct amdgpu_device *adev)
 		else
 			adev->gds.gds_compute_max_wave_id = 0x15f; /* raven1 */
 		break;
+	case CHIP_ARCTURUS:
+		adev->gds.gds_compute_max_wave_id = 0xfff;
+		break;
 	default:
 		/* this really depends on the chip */
 		adev->gds.gds_compute_max_wave_id = 0x7ff;
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#88 lore raw

PATCH 90/102 drm/amdgpu: clean up nonexistent firmware declaration for Arcturus

In reply to a message not in the archive

From: Le Ma <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 3 ---
 1 file changed, 3 deletions(-)
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +0 -3
index 2f553595a898..81f896b4aa6d 100644
@@ -100,9 +100,6 @@ MODULE_FIRMWARE("amdgpu/raven2_mec2.bin");
 MODULE_FIRMWARE("amdgpu/raven2_rlc.bin");
 MODULE_FIRMWARE("amdgpu/raven_kicker_rlc.bin");
 
-MODULE_FIRMWARE("amdgpu/arcturus_ce.bin");
-MODULE_FIRMWARE("amdgpu/arcturus_pfp.bin");
-MODULE_FIRMWARE("amdgpu/arcturus_me.bin");
 MODULE_FIRMWARE("amdgpu/arcturus_mec.bin");
 MODULE_FIRMWARE("amdgpu/arcturus_mec2.bin");
 MODULE_FIRMWARE("amdgpu/arcturus_rlc.bin");
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#89 lore raw

PATCH 91/102 amd/powerplay: No SW XGMI dpm for Arcturus rev 2

In reply to a message not in the archive

From: Yong Zhao <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c       |  2 +-
 drivers/gpu/drm/amd/powerplay/amdgpu_smu.c     | 11 +++++++++++
 drivers/gpu/drm/amd/powerplay/inc/amdgpu_smu.h |  1 +
 3 files changed, 13 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c +1 -1
index 41e648d56eca..4e8d60eec0fe 100644
@@ -248,7 +248,7 @@ int amdgpu_xgmi_set_pstate(struct amdgpu_device *adev, int pstate)
 
 	dev_dbg(adev->dev, "Set xgmi pstate %d.\n", pstate);
 
-	if (is_support_sw_smu(adev))
+	if (is_support_sw_smu_xgmi(adev))
 		ret = smu_set_xgmi_pstate(&adev->smu, pstate);
 	if (ret)
 		dev_err(adev->dev,
drivers/gpu/drm/amd/powerplay/amdgpu_smu.c +11 -0
index 7d1bd0997b59..342d57f2fc5f 100644
@@ -370,6 +370,17 @@ bool is_support_sw_smu(struct amdgpu_device *adev)
 		return false;
 }
 
+bool is_support_sw_smu_xgmi(struct amdgpu_device *adev)
+{
+	if (amdgpu_dpm != 1)
+		return false;
+
+	if (adev->asic_type == CHIP_VEGA20)
+		return true;
+
+	return false;
+}
+
 int smu_sys_get_pp_table(struct smu_context *smu, void **table)
 {
 	struct smu_table_context *smu_table = &smu->smu_table;
drivers/gpu/drm/amd/powerplay/inc/amdgpu_smu.h +1 -0
index 79e34097eb0f..c80077db6cf5 100644
@@ -951,6 +951,7 @@ int smu_update_table(struct smu_context *smu, uint32_t table_index,
 		     void *table_data, bool drv2smu);
 
 bool is_support_sw_smu(struct amdgpu_device *adev);
+bool is_support_sw_smu_xgmi(struct amdgpu_device *adev);
 int smu_reset(struct smu_context *smu);
 int smu_common_read_sensor(struct smu_context *smu, enum amd_pp_sensors sensor,
 			   void *data, uint32_t *size);
Signature · 6 lines
2.20.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx