From: "Zhang, Jerry (Junwei)" <Jerry.Zhang-5C7GfCeVMHo@public.gmane.org>
To: Alex Deucher <alexdeucher-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
Cc: Alex Deucher <alexander.deucher-5C7GfCeVMHo@public.gmane.org>,
amd-gfx list
<amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org>,
Alex Xie <AlexBin.Xie-5C7GfCeVMHo@public.gmane.org>
Subject: Re: [PATCH 046/100] drm/amdgpu: Add GMC 9.0 support
Date: Thu, 23 Mar 2017 10:54:46 +0800 [thread overview]
Message-ID: <58D338F6.1090607@amd.com> (raw)
In-Reply-To: <CADnq5_N_8r-pesBn1NwDrHJcOx8jnryhtcWow01Cbndj8B9N6w-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
On 03/23/2017 10:53 AM, Alex Deucher wrote:
> On Wed, Mar 22, 2017 at 10:42 PM, Zhang, Jerry (Junwei)
> <Jerry.Zhang@amd.com> wrote:
>> Hi Alex,
>>
>> I remember we had a patch to remove the FB location programming in
>> gmc/vmhub.
>> I saw it's not in gmc v9 in this patch, but pre-gmcv9 still program FB
>> register.
>>
>> Is that any missing for sync here?
>> Or it's only supported for gmc v9 now.
>
> I never landed the patches for older asics because I couldn't get vce
> to work properly with dal disabled without reprogramming the fb
> location.
Thanks, got the latest info.
Jerry
>
> Alex
>
>
>>
>> Jerry
>>
>> On 03/21/2017 04:29 AM, Alex Deucher wrote:
>>>
>>> From: Alex Xie <AlexBin.Xie@amd.com>
>>>
>>> On SOC-15 parts, the GMC (Graphics Memory Controller) consists
>>> of two hubs: GFX (graphics and compute) and MM (sdma, uvd, vce).
>>>
>>> Signed-off-by: Alex Xie <AlexBin.Xie@amd.com>
>>> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
>>> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
>>> ---
>>> drivers/gpu/drm/amd/amdgpu/Makefile | 6 +-
>>> drivers/gpu/drm/amd/amdgpu/amdgpu.h | 30 ++
>>> drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 28 +-
>>> drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c | 447 +++++++++++++++++
>>> drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.h | 35 ++
>>> drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c | 826
>>> +++++++++++++++++++++++++++++++
>>> drivers/gpu/drm/amd/amdgpu/gmc_v9_0.h | 30 ++
>>> drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c | 585 ++++++++++++++++++++++
>>> drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.h | 35 ++
>>> drivers/gpu/drm/amd/include/amd_shared.h | 2 +
>>> 10 files changed, 2016 insertions(+), 8 deletions(-)
>>> create mode 100644 drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c
>>> create mode 100644 drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.h
>>> create mode 100644 drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
>>> create mode 100644 drivers/gpu/drm/amd/amdgpu/gmc_v9_0.h
>>> create mode 100644 drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c
>>> create mode 100644 drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.h
>>>
>>> diff --git a/drivers/gpu/drm/amd/amdgpu/Makefile
>>> b/drivers/gpu/drm/amd/amdgpu/Makefile
>>> index 69823e8..b5046fd 100644
>>> --- a/drivers/gpu/drm/amd/amdgpu/Makefile
>>> +++ b/drivers/gpu/drm/amd/amdgpu/Makefile
>>> @@ -45,7 +45,8 @@ amdgpu-y += \
>>> # add GMC block
>>> amdgpu-y += \
>>> gmc_v7_0.o \
>>> - gmc_v8_0.o
>>> + gmc_v8_0.o \
>>> + gfxhub_v1_0.o mmhub_v1_0.o gmc_v9_0.o
>>>
>>> # add IH block
>>> amdgpu-y += \
>>> @@ -74,7 +75,8 @@ amdgpu-y += \
>>> # add async DMA block
>>> amdgpu-y += \
>>> sdma_v2_4.o \
>>> - sdma_v3_0.o
>>> + sdma_v3_0.o \
>>> + sdma_v4_0.o
>>>
>>> # add UVD block
>>> amdgpu-y += \
>>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
>>> b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
>>> index aaded8d..d7257b6 100644
>>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
>>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
>>> @@ -123,6 +123,11 @@ extern int amdgpu_param_buf_per_se;
>>> /* max number of IP instances */
>>> #define AMDGPU_MAX_SDMA_INSTANCES 2
>>>
>>> +/* max number of VMHUB */
>>> +#define AMDGPU_MAX_VMHUBS 2
>>> +#define AMDGPU_MMHUB 0
>>> +#define AMDGPU_GFXHUB 1
>>> +
>>> /* hardcode that limit for now */
>>> #define AMDGPU_VA_RESERVED_SIZE (8 << 20)
>>>
>>> @@ -310,6 +315,12 @@ struct amdgpu_gart_funcs {
>>> uint32_t flags);
>>> };
>>>
>>> +/* provided by the mc block */
>>> +struct amdgpu_mc_funcs {
>>> + /* adjust mc addr in fb for APU case */
>>> + u64 (*adjust_mc_addr)(struct amdgpu_device *adev, u64 addr);
>>> +};
>>> +
>>> /* provided by the ih block */
>>> struct amdgpu_ih_funcs {
>>> /* ring read/write ptr handling, called from interrupt context */
>>> @@ -559,6 +570,21 @@ int amdgpu_gart_bind(struct amdgpu_device *adev,
>>> uint64_t offset,
>>> int amdgpu_ttm_recover_gart(struct amdgpu_device *adev);
>>>
>>> /*
>>> + * VMHUB structures, functions & helpers
>>> + */
>>> +struct amdgpu_vmhub {
>>> + uint32_t ctx0_ptb_addr_lo32;
>>> + uint32_t ctx0_ptb_addr_hi32;
>>> + uint32_t vm_inv_eng0_req;
>>> + uint32_t vm_inv_eng0_ack;
>>> + uint32_t vm_context0_cntl;
>>> + uint32_t vm_l2_pro_fault_status;
>>> + uint32_t vm_l2_pro_fault_cntl;
>>> + uint32_t (*get_invalidate_req)(unsigned int vm_id);
>>> + uint32_t (*get_vm_protection_bits)(void);
>>> +};
>>> +
>>> +/*
>>> * GPU MC structures, functions & helpers
>>> */
>>> struct amdgpu_mc {
>>> @@ -591,6 +617,9 @@ struct amdgpu_mc {
>>> u64 shared_aperture_end;
>>> u64 private_aperture_start;
>>> u64 private_aperture_end;
>>> + /* protects concurrent invalidation */
>>> + spinlock_t invalidate_lock;
>>> + const struct amdgpu_mc_funcs *mc_funcs;
>>> };
>>>
>>> /*
>>> @@ -1479,6 +1508,7 @@ struct amdgpu_device {
>>> struct amdgpu_gart gart;
>>> struct amdgpu_dummy_page dummy_page;
>>> struct amdgpu_vm_manager vm_manager;
>>> + struct amdgpu_vmhub vmhub[AMDGPU_MAX_VMHUBS];
>>>
>>> /* memory management */
>>> struct amdgpu_mman mman;
>>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
>>> b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
>>> index df615d7..47a8080 100644
>>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
>>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
>>> @@ -375,6 +375,16 @@ static bool amdgpu_vm_ring_has_compute_vm_bug(struct
>>> amdgpu_ring *ring)
>>> return false;
>>> }
>>>
>>> +static u64 amdgpu_vm_adjust_mc_addr(struct amdgpu_device *adev, u64
>>> mc_addr)
>>> +{
>>> + u64 addr = mc_addr;
>>> +
>>> + if (adev->mc.mc_funcs && adev->mc.mc_funcs->adjust_mc_addr)
>>> + addr = adev->mc.mc_funcs->adjust_mc_addr(adev, addr);
>>> +
>>> + return addr;
>>> +}
>>> +
>>> /**
>>> * amdgpu_vm_flush - hardware flush the vm
>>> *
>>> @@ -405,9 +415,10 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct
>>> amdgpu_job *job)
>>> if (ring->funcs->emit_vm_flush && (job->vm_needs_flush ||
>>> amdgpu_vm_is_gpu_reset(adev, id))) {
>>> struct fence *fence;
>>> + u64 pd_addr = amdgpu_vm_adjust_mc_addr(adev,
>>> job->vm_pd_addr);
>>>
>>> - trace_amdgpu_vm_flush(job->vm_pd_addr, ring->idx,
>>> job->vm_id);
>>> - amdgpu_ring_emit_vm_flush(ring, job->vm_id,
>>> job->vm_pd_addr);
>>> + trace_amdgpu_vm_flush(pd_addr, ring->idx, job->vm_id);
>>> + amdgpu_ring_emit_vm_flush(ring, job->vm_id, pd_addr);
>>>
>>> r = amdgpu_fence_emit(ring, &fence);
>>> if (r)
>>> @@ -643,15 +654,18 @@ int amdgpu_vm_update_page_directory(struct
>>> amdgpu_device *adev,
>>> (count == AMDGPU_VM_MAX_UPDATE_SIZE)) {
>>>
>>> if (count) {
>>> + uint64_t pt_addr =
>>> + amdgpu_vm_adjust_mc_addr(adev,
>>> last_pt);
>>> +
>>> if (shadow)
>>> amdgpu_vm_do_set_ptes(¶ms,
>>> last_shadow,
>>> - last_pt,
>>> count,
>>> + pt_addr,
>>> count,
>>> incr,
>>>
>>> AMDGPU_PTE_VALID);
>>>
>>> amdgpu_vm_do_set_ptes(¶ms, last_pde,
>>> - last_pt, count,
>>> incr,
>>> + pt_addr, count,
>>> incr,
>>> AMDGPU_PTE_VALID);
>>> }
>>>
>>> @@ -665,11 +679,13 @@ int amdgpu_vm_update_page_directory(struct
>>> amdgpu_device *adev,
>>> }
>>>
>>> if (count) {
>>> + uint64_t pt_addr = amdgpu_vm_adjust_mc_addr(adev,
>>> last_pt);
>>> +
>>> if (vm->page_directory->shadow)
>>> - amdgpu_vm_do_set_ptes(¶ms, last_shadow,
>>> last_pt,
>>> + amdgpu_vm_do_set_ptes(¶ms, last_shadow,
>>> pt_addr,
>>> count, incr,
>>> AMDGPU_PTE_VALID);
>>>
>>> - amdgpu_vm_do_set_ptes(¶ms, last_pde, last_pt,
>>> + amdgpu_vm_do_set_ptes(¶ms, last_pde, pt_addr,
>>> count, incr, AMDGPU_PTE_VALID);
>>> }
>>>
>>> diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c
>>> b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c
>>> new file mode 100644
>>> index 0000000..1ff019c
>>> --- /dev/null
>>> +++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c
>>> @@ -0,0 +1,447 @@
>>> +/*
>>> + * Copyright 2016 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.
>>> + *
>>> + */
>>> +#include "amdgpu.h"
>>> +#include "gfxhub_v1_0.h"
>>> +
>>> +#include "vega10/soc15ip.h"
>>> +#include "vega10/GC/gc_9_0_offset.h"
>>> +#include "vega10/GC/gc_9_0_sh_mask.h"
>>> +#include "vega10/GC/gc_9_0_default.h"
>>> +#include "vega10/vega10_enum.h"
>>> +
>>> +#include "soc15_common.h"
>>> +
>>> +int gfxhub_v1_0_gart_enable(struct amdgpu_device *adev)
>>> +{
>>> + u32 tmp;
>>> + u64 value;
>>> + u32 i;
>>> +
>>> + /* Program MC. */
>>> + /* Update configuration */
>>> + WREG32(SOC15_REG_OFFSET(GC, 0, mmMC_VM_SYSTEM_APERTURE_LOW_ADDR),
>>> + adev->mc.vram_start >> 18);
>>> + WREG32(SOC15_REG_OFFSET(GC, 0, mmMC_VM_SYSTEM_APERTURE_HIGH_ADDR),
>>> + adev->mc.vram_end >> 18);
>>> +
>>> + value = adev->vram_scratch.gpu_addr - adev->mc.vram_start
>>> + + adev->vm_manager.vram_base_offset;
>>> + WREG32(SOC15_REG_OFFSET(GC, 0,
>>> + mmMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB),
>>> + (u32)(value >> 12));
>>> + WREG32(SOC15_REG_OFFSET(GC, 0,
>>> + mmMC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB),
>>> + (u32)(value >> 44));
>>> +
>>> + /* Disable AGP. */
>>> + WREG32(SOC15_REG_OFFSET(GC, 0, mmMC_VM_AGP_BASE), 0);
>>> + WREG32(SOC15_REG_OFFSET(GC, 0, mmMC_VM_AGP_TOP), 0);
>>> + WREG32(SOC15_REG_OFFSET(GC, 0, mmMC_VM_AGP_BOT), 0xFFFFFFFF);
>>> +
>>> + /* GART Enable. */
>>> +
>>> + /* Setup TLB control */
>>> + tmp = RREG32(SOC15_REG_OFFSET(GC, 0, mmMC_VM_MX_L1_TLB_CNTL));
>>> + tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 1);
>>> + tmp = REG_SET_FIELD(tmp,
>>> + MC_VM_MX_L1_TLB_CNTL,
>>> + SYSTEM_ACCESS_MODE,
>>> + 3);
>>> + tmp = REG_SET_FIELD(tmp,
>>> + MC_VM_MX_L1_TLB_CNTL,
>>> + ENABLE_ADVANCED_DRIVER_MODEL,
>>> + 1);
>>> + tmp = REG_SET_FIELD(tmp,
>>> + MC_VM_MX_L1_TLB_CNTL,
>>> + SYSTEM_APERTURE_UNMAPPED_ACCESS,
>>> + 0);
>>> + tmp = REG_SET_FIELD(tmp,
>>> + MC_VM_MX_L1_TLB_CNTL,
>>> + ECO_BITS,
>>> + 0);
>>> + tmp = REG_SET_FIELD(tmp,
>>> + MC_VM_MX_L1_TLB_CNTL,
>>> + MTYPE,
>>> + MTYPE_UC);/* XXX for emulation. */
>>> + tmp = REG_SET_FIELD(tmp,
>>> + MC_VM_MX_L1_TLB_CNTL,
>>> + ATC_EN,
>>> + 1);
>>> + WREG32(SOC15_REG_OFFSET(GC, 0, mmMC_VM_MX_L1_TLB_CNTL), tmp);
>>> +
>>> + /* Setup L2 cache */
>>> + tmp = RREG32(SOC15_REG_OFFSET(GC, 0, mmVM_L2_CNTL));
>>> + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, ENABLE_L2_CACHE, 1);
>>> + tmp = REG_SET_FIELD(tmp,
>>> + VM_L2_CNTL,
>>> + ENABLE_L2_FRAGMENT_PROCESSING,
>>> + 0);
>>> + tmp = REG_SET_FIELD(tmp,
>>> + VM_L2_CNTL,
>>> + L2_PDE0_CACHE_TAG_GENERATION_MODE,
>>> + 0);/* XXX for emulation, Refer to closed
>>> source code.*/
>>> + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, PDE_FAULT_CLASSIFICATION, 1);
>>> + tmp = REG_SET_FIELD(tmp,
>>> + VM_L2_CNTL,
>>> + CONTEXT1_IDENTITY_ACCESS_MODE,
>>> + 1);
>>> + tmp = REG_SET_FIELD(tmp,
>>> + VM_L2_CNTL,
>>> + IDENTITY_MODE_FRAGMENT_SIZE,
>>> + 0);
>>> + WREG32(SOC15_REG_OFFSET(GC, 0, mmVM_L2_CNTL), tmp);
>>> +
>>> + tmp = RREG32(SOC15_REG_OFFSET(GC, 0, mmVM_L2_CNTL2));
>>> + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS, 1);
>>> + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL2, INVALIDATE_L2_CACHE, 1);
>>> + WREG32(SOC15_REG_OFFSET(GC, 0, mmVM_L2_CNTL2), tmp);
>>> +
>>> + tmp = mmVM_L2_CNTL3_DEFAULT;
>>> + WREG32(SOC15_REG_OFFSET(GC, 0, mmVM_L2_CNTL3), tmp);
>>> +
>>> + tmp = RREG32(SOC15_REG_OFFSET(GC, 0, mmVM_L2_CNTL4));
>>> + tmp = REG_SET_FIELD(tmp,
>>> + VM_L2_CNTL4,
>>> + VMC_TAP_PDE_REQUEST_PHYSICAL,
>>> + 0);
>>> + tmp = REG_SET_FIELD(tmp,
>>> + VM_L2_CNTL4,
>>> + VMC_TAP_PTE_REQUEST_PHYSICAL,
>>> + 0);
>>> + WREG32(SOC15_REG_OFFSET(GC, 0, mmVM_L2_CNTL4), tmp);
>>> +
>>> + /* setup context0 */
>>> + WREG32(SOC15_REG_OFFSET(GC, 0,
>>> + mmVM_CONTEXT0_PAGE_TABLE_START_ADDR_LO32),
>>> + (u32)(adev->mc.gtt_start >> 12));
>>> + WREG32(SOC15_REG_OFFSET(GC, 0,
>>> + mmVM_CONTEXT0_PAGE_TABLE_START_ADDR_HI32),
>>> + (u32)(adev->mc.gtt_start >> 44));
>>> +
>>> + WREG32(SOC15_REG_OFFSET(GC, 0,
>>> + mmVM_CONTEXT0_PAGE_TABLE_END_ADDR_LO32),
>>> + (u32)(adev->mc.gtt_end >> 12));
>>> + WREG32(SOC15_REG_OFFSET(GC, 0,
>>> + mmVM_CONTEXT0_PAGE_TABLE_END_ADDR_HI32),
>>> + (u32)(adev->mc.gtt_end >> 44));
>>> +
>>> + BUG_ON(adev->gart.table_addr & (~0x0000FFFFFFFFF000ULL));
>>> + value = adev->gart.table_addr - adev->mc.vram_start
>>> + + adev->vm_manager.vram_base_offset;
>>> + value &= 0x0000FFFFFFFFF000ULL;
>>> + value |= 0x1; /*valid bit*/
>>> +
>>> + WREG32(SOC15_REG_OFFSET(GC, 0,
>>> + mmVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32),
>>> + (u32)value);
>>> + WREG32(SOC15_REG_OFFSET(GC, 0,
>>> + mmVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32),
>>> + (u32)(value >> 32));
>>> +
>>> + WREG32(SOC15_REG_OFFSET(GC, 0,
>>> +
>>> mmVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_LO32),
>>> + (u32)(adev->dummy_page.addr >> 12));
>>> + WREG32(SOC15_REG_OFFSET(GC, 0,
>>> +
>>> mmVM_L2_PROTECTION_FAULT_DEFAULT_ADDR_HI32),
>>> + (u32)(adev->dummy_page.addr >> 44));
>>> +
>>> + tmp = RREG32(SOC15_REG_OFFSET(GC, 0,
>>> mmVM_L2_PROTECTION_FAULT_CNTL2));
>>> + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL2,
>>> + ACTIVE_PAGE_MIGRATION_PTE_READ_RETRY,
>>> + 1);
>>> + WREG32(SOC15_REG_OFFSET(GC, 0, mmVM_L2_PROTECTION_FAULT_CNTL2),
>>> tmp);
>>> +
>>> + tmp = RREG32(SOC15_REG_OFFSET(GC, 0, mmVM_CONTEXT0_CNTL));
>>> + tmp = REG_SET_FIELD(tmp, VM_CONTEXT0_CNTL, ENABLE_CONTEXT, 1);
>>> + tmp = REG_SET_FIELD(tmp, VM_CONTEXT0_CNTL, PAGE_TABLE_DEPTH, 0);
>>> + WREG32(SOC15_REG_OFFSET(GC, 0, mmVM_CONTEXT0_CNTL), tmp);
>>> +
>>> + /* Disable identity aperture.*/
>>> + WREG32(SOC15_REG_OFFSET(GC, 0,
>>> + mmVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_LO32),
>>> 0XFFFFFFFF);
>>> + WREG32(SOC15_REG_OFFSET(GC, 0,
>>> + mmVM_L2_CONTEXT1_IDENTITY_APERTURE_LOW_ADDR_HI32),
>>> 0x0000000F);
>>> +
>>> + WREG32(SOC15_REG_OFFSET(GC, 0,
>>> + mmVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_LO32), 0);
>>> + WREG32(SOC15_REG_OFFSET(GC, 0,
>>> + mmVM_L2_CONTEXT1_IDENTITY_APERTURE_HIGH_ADDR_HI32), 0);
>>> +
>>> + WREG32(SOC15_REG_OFFSET(GC, 0,
>>> + mmVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_LO32), 0);
>>> + WREG32(SOC15_REG_OFFSET(GC, 0,
>>> + mmVM_L2_CONTEXT_IDENTITY_PHYSICAL_OFFSET_HI32), 0);
>>> +
>>> + for (i = 0; i <= 14; i++) {
>>> + tmp = RREG32(SOC15_REG_OFFSET(GC, 0, mmVM_CONTEXT1_CNTL) +
>>> i);
>>> + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL, ENABLE_CONTEXT,
>>> 1);
>>> + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
>>> PAGE_TABLE_DEPTH, 1);
>>> + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
>>> + RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
>>> + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
>>> +
>>> DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
>>> + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
>>> + PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
>>> + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
>>> + VALID_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
>>> + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
>>> + READ_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
>>> + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
>>> + WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, 1);
>>> + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
>>> + EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT,
>>> 1);
>>> + tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
>>> + PAGE_TABLE_BLOCK_SIZE,
>>> + amdgpu_vm_block_size - 9);
>>> + WREG32(SOC15_REG_OFFSET(GC, 0, mmVM_CONTEXT1_CNTL) + i,
>>> tmp);
>>> + WREG32(SOC15_REG_OFFSET(GC, 0,
>>> mmVM_CONTEXT1_PAGE_TABLE_START_ADDR_LO32) + i*2, 0);
>>> + WREG32(SOC15_REG_OFFSET(GC, 0,
>>> mmVM_CONTEXT1_PAGE_TABLE_START_ADDR_HI32) + i*2, 0);
>>> + WREG32(SOC15_REG_OFFSET(GC, 0,
>>> mmVM_CONTEXT1_PAGE_TABLE_END_ADDR_LO32) + i*2,
>>> + adev->vm_manager.max_pfn - 1);
>>> + WREG32(SOC15_REG_OFFSET(GC, 0,
>>> mmVM_CONTEXT1_PAGE_TABLE_END_ADDR_HI32) + i*2, 0);
>>> + }
>>> +
>>> +
>>> + return 0;
>>> +}
>>> +
>>> +void gfxhub_v1_0_gart_disable(struct amdgpu_device *adev)
>>> +{
>>> + u32 tmp;
>>> + u32 i;
>>> +
>>> + /* Disable all tables */
>>> + for (i = 0; i < 16; i++)
>>> + WREG32(SOC15_REG_OFFSET(GC, 0, mmVM_CONTEXT0_CNTL) + i,
>>> 0);
>>> +
>>> + /* Setup TLB control */
>>> + tmp = RREG32(SOC15_REG_OFFSET(GC, 0, mmMC_VM_MX_L1_TLB_CNTL));
>>> + tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB, 0);
>>> + tmp = REG_SET_FIELD(tmp,
>>> + MC_VM_MX_L1_TLB_CNTL,
>>> + ENABLE_ADVANCED_DRIVER_MODEL,
>>> + 0);
>>> + WREG32(SOC15_REG_OFFSET(GC, 0, mmMC_VM_MX_L1_TLB_CNTL), tmp);
>>> +
>>> + /* Setup L2 cache */
>>> + tmp = RREG32(SOC15_REG_OFFSET(GC, 0, mmVM_L2_CNTL));
>>> + tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, ENABLE_L2_CACHE, 0);
>>> + WREG32(SOC15_REG_OFFSET(GC, 0, mmVM_L2_CNTL), tmp);
>>> + WREG32(SOC15_REG_OFFSET(GC, 0, mmVM_L2_CNTL3), 0);
>>> +}
>>> +
>>> +/**
>>> + * gfxhub_v1_0_set_fault_enable_default - update GART/VM fault handling
>>> + *
>>> + * @adev: amdgpu_device pointer
>>> + * @value: true redirects VM faults to the default page
>>> + */
>>> +void gfxhub_v1_0_set_fault_enable_default(struct amdgpu_device *adev,
>>> + bool value)
>>> +{
>>> + u32 tmp;
>>> + tmp = RREG32(SOC15_REG_OFFSET(GC, 0,
>>> mmVM_L2_PROTECTION_FAULT_CNTL));
>>> + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
>>> + RANGE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
>>> + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
>>> + PDE0_PROTECTION_FAULT_ENABLE_DEFAULT, value);
>>> + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
>>> + PDE1_PROTECTION_FAULT_ENABLE_DEFAULT, value);
>>> + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
>>> + PDE2_PROTECTION_FAULT_ENABLE_DEFAULT, value);
>>> + tmp = REG_SET_FIELD(tmp,
>>> + VM_L2_PROTECTION_FAULT_CNTL,
>>> + TRANSLATE_FURTHER_PROTECTION_FAULT_ENABLE_DEFAULT,
>>> + value);
>>> + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
>>> + NACK_PROTECTION_FAULT_ENABLE_DEFAULT, value);
>>> + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
>>> + DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT,
>>> value);
>>> + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
>>> + VALID_PROTECTION_FAULT_ENABLE_DEFAULT, value);
>>> + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
>>> + READ_PROTECTION_FAULT_ENABLE_DEFAULT, value);
>>> + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
>>> + WRITE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
>>> + tmp = REG_SET_FIELD(tmp, VM_L2_PROTECTION_FAULT_CNTL,
>>> + EXECUTE_PROTECTION_FAULT_ENABLE_DEFAULT, value);
>>> + WREG32(SOC15_REG_OFFSET(GC, 0, mmVM_L2_PROTECTION_FAULT_CNTL),
>>> tmp);
>>> +}
>>> +
>>> +static uint32_t gfxhub_v1_0_get_invalidate_req(unsigned int vm_id)
>>> +{
>>> + u32 req = 0;
>>> +
>>> + /* invalidate using legacy mode on vm_id*/
>>> + req = REG_SET_FIELD(req, VM_INVALIDATE_ENG0_REQ,
>>> + PER_VMID_INVALIDATE_REQ, 1 << vm_id);
>>> + req = REG_SET_FIELD(req, VM_INVALIDATE_ENG0_REQ, FLUSH_TYPE, 0);
>>> + req = REG_SET_FIELD(req, VM_INVALIDATE_ENG0_REQ,
>>> INVALIDATE_L2_PTES, 1);
>>> + req = REG_SET_FIELD(req, VM_INVALIDATE_ENG0_REQ,
>>> INVALIDATE_L2_PDE0, 1);
>>> + req = REG_SET_FIELD(req, VM_INVALIDATE_ENG0_REQ,
>>> INVALIDATE_L2_PDE1, 1);
>>> + req = REG_SET_FIELD(req, VM_INVALIDATE_ENG0_REQ,
>>> INVALIDATE_L2_PDE2, 1);
>>> + req = REG_SET_FIELD(req, VM_INVALIDATE_ENG0_REQ,
>>> INVALIDATE_L1_PTES, 1);
>>> + req = REG_SET_FIELD(req, VM_INVALIDATE_ENG0_REQ,
>>> + CLEAR_PROTECTION_FAULT_STATUS_ADDR, 0);
>>> +
>>> + return req;
>>> +}
>>> +
>>> +static uint32_t gfxhub_v1_0_get_vm_protection_bits(void)
>>> +{
>>> + return
>>> (VM_CONTEXT1_CNTL__RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
>>> +
>>> VM_CONTEXT1_CNTL__DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
>>> +
>>> VM_CONTEXT1_CNTL__PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
>>> +
>>> VM_CONTEXT1_CNTL__VALID_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
>>> +
>>> VM_CONTEXT1_CNTL__READ_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
>>> +
>>> VM_CONTEXT1_CNTL__WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK |
>>> +
>>> VM_CONTEXT1_CNTL__EXECUTE_PROTECTION_FAULT_ENABLE_INTERRUPT_MASK);
>>> +}
>>> +
>>> +static int gfxhub_v1_0_early_init(void *handle)
>>> +{
>>> + return 0;
>>> +}
>>> +
>>> +static int gfxhub_v1_0_late_init(void *handle)
>>> +{
>>> + return 0;
>>> +}
>>> +
>>> +static int gfxhub_v1_0_sw_init(void *handle)
>>> +{
>>> + struct amdgpu_device *adev = (struct amdgpu_device *)handle;
>>> + struct amdgpu_vmhub *hub = &adev->vmhub[AMDGPU_GFXHUB];
>>> +
>>> + hub->ctx0_ptb_addr_lo32 =
>>> + SOC15_REG_OFFSET(GC, 0,
>>> + mmVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32);
>>> + hub->ctx0_ptb_addr_hi32 =
>>> + SOC15_REG_OFFSET(GC, 0,
>>> + mmVM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32);
>>> + hub->vm_inv_eng0_req =
>>> + SOC15_REG_OFFSET(GC, 0, mmVM_INVALIDATE_ENG0_REQ);
>>> + hub->vm_inv_eng0_ack =
>>> + SOC15_REG_OFFSET(GC, 0, mmVM_INVALIDATE_ENG0_ACK);
>>> + hub->vm_context0_cntl =
>>> + SOC15_REG_OFFSET(GC, 0, mmVM_CONTEXT0_CNTL);
>>> + hub->vm_l2_pro_fault_status =
>>> + SOC15_REG_OFFSET(GC, 0, mmVM_L2_PROTECTION_FAULT_STATUS);
>>> + hub->vm_l2_pro_fault_cntl =
>>> + SOC15_REG_OFFSET(GC, 0, mmVM_L2_PROTECTION_FAULT_CNTL);
>>> +
>>> + hub->get_invalidate_req = gfxhub_v1_0_get_invalidate_req;
>>> + hub->get_vm_protection_bits = gfxhub_v1_0_get_vm_protection_bits;
>>> +
>>> + return 0;
>>> +}
>>> +
>>> +static int gfxhub_v1_0_sw_fini(void *handle)
>>> +{
>>> + return 0;
>>> +}
>>> +
>>> +static int gfxhub_v1_0_hw_init(void *handle)
>>> +{
>>> + struct amdgpu_device *adev = (struct amdgpu_device *)handle;
>>> + unsigned i;
>>> +
>>> + for (i = 0 ; i < 18; ++i) {
>>> + WREG32(SOC15_REG_OFFSET(GC, 0,
>>> +
>>> mmVM_INVALIDATE_ENG0_ADDR_RANGE_LO32) +
>>> + 2 * i, 0xffffffff);
>>> + WREG32(SOC15_REG_OFFSET(GC, 0,
>>> +
>>> mmVM_INVALIDATE_ENG0_ADDR_RANGE_HI32) +
>>> + 2 * i, 0x1f);
>>> + }
>>> +
>>> + return 0;
>>> +}
>>> +
>>> +static int gfxhub_v1_0_hw_fini(void *handle)
>>> +{
>>> + return 0;
>>> +}
>>> +
>>> +static int gfxhub_v1_0_suspend(void *handle)
>>> +{
>>> + return 0;
>>> +}
>>> +
>>> +static int gfxhub_v1_0_resume(void *handle)
>>> +{
>>> + return 0;
>>> +}
>>> +
>>> +static bool gfxhub_v1_0_is_idle(void *handle)
>>> +{
>>> + return true;
>>> +}
>>> +
>>> +static int gfxhub_v1_0_wait_for_idle(void *handle)
>>> +{
>>> + return 0;
>>> +}
>>> +
>>> +static int gfxhub_v1_0_soft_reset(void *handle)
>>> +{
>>> + return 0;
>>> +}
>>> +
>>> +static int gfxhub_v1_0_set_clockgating_state(void *handle,
>>> + enum amd_clockgating_state
>>> state)
>>> +{
>>> + return 0;
>>> +}
>>> +
>>> +static int gfxhub_v1_0_set_powergating_state(void *handle,
>>> + enum amd_powergating_state
>>> state)
>>> +{
>>> + return 0;
>>> +}
>>> +
>>> +const struct amd_ip_funcs gfxhub_v1_0_ip_funcs = {
>>> + .name = "gfxhub_v1_0",
>>> + .early_init = gfxhub_v1_0_early_init,
>>> + .late_init = gfxhub_v1_0_late_init,
>>> + .sw_init = gfxhub_v1_0_sw_init,
>>> + .sw_fini = gfxhub_v1_0_sw_fini,
>>> + .hw_init = gfxhub_v1_0_hw_init,
>>> + .hw_fini = gfxhub_v1_0_hw_fini,
>>> + .suspend = gfxhub_v1_0_suspend,
>>> + .resume = gfxhub_v1_0_resume,
>>> + .is_idle = gfxhub_v1_0_is_idle,
>>> + .wait_for_idle = gfxhub_v1_0_wait_for_idle,
>>> + .soft_reset = gfxhub_v1_0_soft_reset,
>>> + .set_clockgating_state = gfxhub_v1_0_set_clockgating_state,
>>> + .set_powergating_state = gfxhub_v1_0_set_powergating_state,
>>> +};
>>> +
>>> +const struct amdgpu_ip_block_version gfxhub_v1_0_ip_block =
>>> +{
>>> + .type = AMD_IP_BLOCK_TYPE_GFXHUB,
>>> + .major = 1,
>>> + .minor = 0,
>>> + .rev = 0,
>>> + .funcs = &gfxhub_v1_0_ip_funcs,
>>> +};
>>> diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.h
>>> b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.h
>>> new file mode 100644
>>> index 0000000..5129a8f
>>> --- /dev/null
>>> +++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.h
>>> @@ -0,0 +1,35 @@
>>> +/*
>>> + * Copyright 2016 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 __GFXHUB_V1_0_H__
>>> +#define __GFXHUB_V1_0_H__
>>> +
>>> +int gfxhub_v1_0_gart_enable(struct amdgpu_device *adev);
>>> +void gfxhub_v1_0_gart_disable(struct amdgpu_device *adev);
>>> +void gfxhub_v1_0_set_fault_enable_default(struct amdgpu_device *adev,
>>> + bool value);
>>> +
>>> +extern const struct amd_ip_funcs gfxhub_v1_0_ip_funcs;
>>> +extern const struct amdgpu_ip_block_version gfxhub_v1_0_ip_block;
>>> +
>>> +#endif
>>> diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
>>> b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
>>> new file mode 100644
>>> index 0000000..5cf0fc3
>>> --- /dev/null
>>> +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
>>> @@ -0,0 +1,826 @@
>>> +/*
>>> + * Copyright 2016 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.
>>> + *
>>> + */
>>> +#include <linux/firmware.h>
>>> +#include "amdgpu.h"
>>> +#include "gmc_v9_0.h"
>>> +
>>> +#include "vega10/soc15ip.h"
>>> +#include "vega10/HDP/hdp_4_0_offset.h"
>>> +#include "vega10/HDP/hdp_4_0_sh_mask.h"
>>> +#include "vega10/GC/gc_9_0_sh_mask.h"
>>> +#include "vega10/vega10_enum.h"
>>> +
>>> +#include "soc15_common.h"
>>> +
>>> +#include "nbio_v6_1.h"
>>> +#include "gfxhub_v1_0.h"
>>> +#include "mmhub_v1_0.h"
>>> +
>>> +#define mmDF_CS_AON0_DramBaseAddress0
>>> 0x0044
>>> +#define mmDF_CS_AON0_DramBaseAddress0_BASE_IDX
>>> 0
>>> +//DF_CS_AON0_DramBaseAddress0
>>> +#define DF_CS_AON0_DramBaseAddress0__AddrRngVal__SHIFT
>>> 0x0
>>> +#define DF_CS_AON0_DramBaseAddress0__LgcyMmioHoleEn__SHIFT
>>> 0x1
>>> +#define DF_CS_AON0_DramBaseAddress0__IntLvNumChan__SHIFT
>>> 0x4
>>> +#define DF_CS_AON0_DramBaseAddress0__IntLvAddrSel__SHIFT
>>> 0x8
>>> +#define DF_CS_AON0_DramBaseAddress0__DramBaseAddr__SHIFT
>>> 0xc
>>> +#define DF_CS_AON0_DramBaseAddress0__AddrRngVal_MASK
>>> 0x00000001L
>>> +#define DF_CS_AON0_DramBaseAddress0__LgcyMmioHoleEn_MASK
>>> 0x00000002L
>>> +#define DF_CS_AON0_DramBaseAddress0__IntLvNumChan_MASK
>>> 0x000000F0L
>>> +#define DF_CS_AON0_DramBaseAddress0__IntLvAddrSel_MASK
>>> 0x00000700L
>>> +#define DF_CS_AON0_DramBaseAddress0__DramBaseAddr_MASK
>>> 0xFFFFF000L
>>> +
>>> +/* XXX Move this macro to VEGA10 header file, which is like vid.h for
>>> VI.*/
>>> +#define AMDGPU_NUM_OF_VMIDS 8
>>> +
>>> +static const u32 golden_settings_vega10_hdp[] =
>>> +{
>>> + 0xf64, 0x0fffffff, 0x00000000,
>>> + 0xf65, 0x0fffffff, 0x00000000,
>>> + 0xf66, 0x0fffffff, 0x00000000,
>>> + 0xf67, 0x0fffffff, 0x00000000,
>>> + 0xf68, 0x0fffffff, 0x00000000,
>>> + 0xf6a, 0x0fffffff, 0x00000000,
>>> + 0xf6b, 0x0fffffff, 0x00000000,
>>> + 0xf6c, 0x0fffffff, 0x00000000,
>>> + 0xf6d, 0x0fffffff, 0x00000000,
>>> + 0xf6e, 0x0fffffff, 0x00000000,
>>> +};
>>> +
>>> +static int gmc_v9_0_vm_fault_interrupt_state(struct amdgpu_device *adev,
>>> + struct amdgpu_irq_src *src,
>>> + unsigned type,
>>> + enum amdgpu_interrupt_state state)
>>> +{
>>> + struct amdgpu_vmhub *hub;
>>> + u32 tmp, reg, bits, i;
>>> +
>>> + switch (state) {
>>> + case AMDGPU_IRQ_STATE_DISABLE:
>>> + /* MM HUB */
>>> + hub = &adev->vmhub[AMDGPU_MMHUB];
>>> + bits = hub->get_vm_protection_bits();
>>> + for (i = 0; i< 16; i++) {
>>> + reg = hub->vm_context0_cntl + i;
>>> + tmp = RREG32(reg);
>>> + tmp &= ~bits;
>>> + WREG32(reg, tmp);
>>> + }
>>> +
>>> + /* GFX HUB */
>>> + hub = &adev->vmhub[AMDGPU_GFXHUB];
>>> + bits = hub->get_vm_protection_bits();
>>> + for (i = 0; i < 16; i++) {
>>> + reg = hub->vm_context0_cntl + i;
>>> + tmp = RREG32(reg);
>>> + tmp &= ~bits;
>>> + WREG32(reg, tmp);
>>> + }
>>> + break;
>>> + case AMDGPU_IRQ_STATE_ENABLE:
>>> + /* MM HUB */
>>> + hub = &adev->vmhub[AMDGPU_MMHUB];
>>> + bits = hub->get_vm_protection_bits();
>>> + for (i = 0; i< 16; i++) {
>>> + reg = hub->vm_context0_cntl + i;
>>> + tmp = RREG32(reg);
>>> + tmp |= bits;
>>> + WREG32(reg, tmp);
>>> + }
>>> +
>>> + /* GFX HUB */
>>> + hub = &adev->vmhub[AMDGPU_GFXHUB];
>>> + bits = hub->get_vm_protection_bits();
>>> + for (i = 0; i < 16; i++) {
>>> + reg = hub->vm_context0_cntl + i;
>>> + tmp = RREG32(reg);
>>> + tmp |= bits;
>>> + WREG32(reg, tmp);
>>> + }
>>> + break;
>>> + default:
>>> + break;
>>> + }
>>> +
>>> + return 0;
>>> + return 0;
>>> +}
>>> +
>>> +static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev,
>>> + struct amdgpu_irq_src *source,
>>> + struct amdgpu_iv_entry *entry)
>>> +{
>>> + struct amdgpu_vmhub *gfxhub = &adev->vmhub[AMDGPU_GFXHUB];
>>> + struct amdgpu_vmhub *mmhub = &adev->vmhub[AMDGPU_MMHUB];
>>> + uint32_t status;
>>> + u64 addr;
>>> +
>>> + addr = (u64)entry->src_data[0] << 12;
>>> + addr |= ((u64)entry->src_data[1] & 0xf) << 44;
>>> +
>>> + if (entry->vm_id_src) {
>>> + status = RREG32(mmhub->vm_l2_pro_fault_status);
>>> + WREG32_P(mmhub->vm_l2_pro_fault_cntl, 1, ~1);
>>> + } else {
>>> + status = RREG32(gfxhub->vm_l2_pro_fault_status);
>>> + WREG32_P(gfxhub->vm_l2_pro_fault_cntl, 1, ~1);
>>> + }
>>> +
>>> + DRM_ERROR("[%s]VMC page fault (src_id:%u ring:%u vm_id:%u
>>> pas_id:%u) "
>>> + "at page 0x%016llx from %d\n"
>>> + "VM_L2_PROTECTION_FAULT_STATUS:0x%08X\n",
>>> + entry->vm_id_src ? "mmhub" : "gfxhub",
>>> + entry->src_id, entry->ring_id, entry->vm_id,
>>> entry->pas_id,
>>> + addr, entry->client_id, status);
>>> +
>>> + return 0;
>>> +}
>>> +
>>> +static const struct amdgpu_irq_src_funcs gmc_v9_0_irq_funcs = {
>>> + .set = gmc_v9_0_vm_fault_interrupt_state,
>>> + .process = gmc_v9_0_process_interrupt,
>>> +};
>>> +
>>> +static void gmc_v9_0_set_irq_funcs(struct amdgpu_device *adev)
>>> +{
>>> + adev->mc.vm_fault.num_types = 1;
>>> + adev->mc.vm_fault.funcs = &gmc_v9_0_irq_funcs;
>>> +}
>>> +
>>> +/*
>>> + * GART
>>> + * VMID 0 is the physical GPU addresses as used by the kernel.
>>> + * VMIDs 1-15 are used for userspace clients and are handled
>>> + * by the amdgpu vm/hsa code.
>>> + */
>>> +
>>> +/**
>>> + * gmc_v9_0_gart_flush_gpu_tlb - gart tlb flush callback
>>> + *
>>> + * @adev: amdgpu_device pointer
>>> + * @vmid: vm instance to flush
>>> + *
>>> + * Flush the TLB for the requested page table.
>>> + */
>>> +static void gmc_v9_0_gart_flush_gpu_tlb(struct amdgpu_device *adev,
>>> + uint32_t vmid)
>>> +{
>>> + /* Use register 17 for GART */
>>> + const unsigned eng = 17;
>>> + unsigned i, j;
>>> +
>>> + /* flush hdp cache */
>>> + nbio_v6_1_hdp_flush(adev);
>>> +
>>> + spin_lock(&adev->mc.invalidate_lock);
>>> +
>>> + for (i = 0; i < AMDGPU_MAX_VMHUBS; ++i) {
>>> + struct amdgpu_vmhub *hub = &adev->vmhub[i];
>>> + u32 tmp = hub->get_invalidate_req(vmid);
>>> +
>>> + WREG32(hub->vm_inv_eng0_req + eng, tmp);
>>> +
>>> + /* Busy wait for ACK.*/
>>> + for (j = 0; j < 100; j++) {
>>> + tmp = RREG32(hub->vm_inv_eng0_ack + eng);
>>> + tmp &= 1 << vmid;
>>> + if (tmp)
>>> + break;
>>> + cpu_relax();
>>> + }
>>> + if (j < 100)
>>> + continue;
>>> +
>>> + /* Wait for ACK with a delay.*/
>>> + for (j = 0; j < adev->usec_timeout; j++) {
>>> + tmp = RREG32(hub->vm_inv_eng0_ack + eng);
>>> + tmp &= 1 << vmid;
>>> + if (tmp)
>>> + break;
>>> + udelay(1);
>>> + }
>>> + if (j < adev->usec_timeout)
>>> + continue;
>>> +
>>> + DRM_ERROR("Timeout waiting for VM flush ACK!\n");
>>> + }
>>> +
>>> + spin_unlock(&adev->mc.invalidate_lock);
>>> +}
>>> +
>>> +/**
>>> + * gmc_v9_0_gart_set_pte_pde - update the page tables using MMIO
>>> + *
>>> + * @adev: amdgpu_device pointer
>>> + * @cpu_pt_addr: cpu address of the page table
>>> + * @gpu_page_idx: entry in the page table to update
>>> + * @addr: dst addr to write into pte/pde
>>> + * @flags: access flags
>>> + *
>>> + * Update the page tables using the CPU.
>>> + */
>>> +static int gmc_v9_0_gart_set_pte_pde(struct amdgpu_device *adev,
>>> + void *cpu_pt_addr,
>>> + uint32_t gpu_page_idx,
>>> + uint64_t addr,
>>> + uint64_t flags)
>>> +{
>>> + void __iomem *ptr = (void *)cpu_pt_addr;
>>> + uint64_t value;
>>> +
>>> + /*
>>> + * PTE format on VEGA 10:
>>> + * 63:59 reserved
>>> + * 58:57 mtype
>>> + * 56 F
>>> + * 55 L
>>> + * 54 P
>>> + * 53 SW
>>> + * 52 T
>>> + * 50:48 reserved
>>> + * 47:12 4k physical page base address
>>> + * 11:7 fragment
>>> + * 6 write
>>> + * 5 read
>>> + * 4 exe
>>> + * 3 Z
>>> + * 2 snooped
>>> + * 1 system
>>> + * 0 valid
>>> + *
>>> + * PDE format on VEGA 10:
>>> + * 63:59 block fragment size
>>> + * 58:55 reserved
>>> + * 54 P
>>> + * 53:48 reserved
>>> + * 47:6 physical base address of PD or PTE
>>> + * 5:3 reserved
>>> + * 2 C
>>> + * 1 system
>>> + * 0 valid
>>> + */
>>> +
>>> + /*
>>> + * The following is for PTE only. GART does not have PDEs.
>>> + */
>>> + value = addr & 0x0000FFFFFFFFF000ULL;
>>> + value |= flags;
>>> + writeq(value, ptr + (gpu_page_idx * 8));
>>> + return 0;
>>> +}
>>> +
>>> +static uint64_t gmc_v9_0_get_vm_pte_flags(struct amdgpu_device *adev,
>>> + uint32_t flags)
>>> +
>>> +{
>>> + uint64_t pte_flag = 0;
>>> +
>>> + if (flags & AMDGPU_VM_PAGE_EXECUTABLE)
>>> + pte_flag |= AMDGPU_PTE_EXECUTABLE;
>>> + if (flags & AMDGPU_VM_PAGE_READABLE)
>>> + pte_flag |= AMDGPU_PTE_READABLE;
>>> + if (flags & AMDGPU_VM_PAGE_WRITEABLE)
>>> + pte_flag |= AMDGPU_PTE_WRITEABLE;
>>> +
>>> + switch (flags & AMDGPU_VM_MTYPE_MASK) {
>>> + case AMDGPU_VM_MTYPE_DEFAULT:
>>> + pte_flag |= AMDGPU_PTE_MTYPE(MTYPE_NC);
>>> + break;
>>> + case AMDGPU_VM_MTYPE_NC:
>>> + pte_flag |= AMDGPU_PTE_MTYPE(MTYPE_NC);
>>> + break;
>>> + case AMDGPU_VM_MTYPE_WC:
>>> + pte_flag |= AMDGPU_PTE_MTYPE(MTYPE_WC);
>>> + break;
>>> + case AMDGPU_VM_MTYPE_CC:
>>> + pte_flag |= AMDGPU_PTE_MTYPE(MTYPE_CC);
>>> + break;
>>> + case AMDGPU_VM_MTYPE_UC:
>>> + pte_flag |= AMDGPU_PTE_MTYPE(MTYPE_UC);
>>> + break;
>>> + default:
>>> + pte_flag |= AMDGPU_PTE_MTYPE(MTYPE_NC);
>>> + break;
>>> + }
>>> +
>>> + if (flags & AMDGPU_VM_PAGE_PRT)
>>> + pte_flag |= AMDGPU_PTE_PRT;
>>> +
>>> + return pte_flag;
>>> +}
>>> +
>>> +static const struct amdgpu_gart_funcs gmc_v9_0_gart_funcs = {
>>> + .flush_gpu_tlb = gmc_v9_0_gart_flush_gpu_tlb,
>>> + .set_pte_pde = gmc_v9_0_gart_set_pte_pde,
>>> + .get_vm_pte_flags = gmc_v9_0_get_vm_pte_flags
>>> +};
>>> +
>>> +static void gmc_v9_0_set_gart_funcs(struct amdgpu_device *adev)
>>> +{
>>> + if (adev->gart.gart_funcs == NULL)
>>> + adev->gart.gart_funcs = &gmc_v9_0_gart_funcs;
>>> +}
>>> +
>>> +static u64 gmc_v9_0_adjust_mc_addr(struct amdgpu_device *adev, u64
>>> mc_addr)
>>> +{
>>> + return adev->vm_manager.vram_base_offset + mc_addr -
>>> adev->mc.vram_start;
>>> +}
>>> +
>>> +static const struct amdgpu_mc_funcs gmc_v9_0_mc_funcs = {
>>> + .adjust_mc_addr = gmc_v9_0_adjust_mc_addr,
>>> +};
>>> +
>>> +static void gmc_v9_0_set_mc_funcs(struct amdgpu_device *adev)
>>> +{
>>> + adev->mc.mc_funcs = &gmc_v9_0_mc_funcs;
>>> +}
>>> +
>>> +static int gmc_v9_0_early_init(void *handle)
>>> +{
>>> + struct amdgpu_device *adev = (struct amdgpu_device *)handle;
>>> +
>>> + gmc_v9_0_set_gart_funcs(adev);
>>> + gmc_v9_0_set_mc_funcs(adev);
>>> + gmc_v9_0_set_irq_funcs(adev);
>>> +
>>> + return 0;
>>> +}
>>> +
>>> +static int gmc_v9_0_late_init(void *handle)
>>> +{
>>> + struct amdgpu_device *adev = (struct amdgpu_device *)handle;
>>> + return amdgpu_irq_get(adev, &adev->mc.vm_fault, 0);
>>> +}
>>> +
>>> +static void gmc_v9_0_vram_gtt_location(struct amdgpu_device *adev,
>>> + struct amdgpu_mc *mc)
>>> +{
>>> + u64 base = mmhub_v1_0_get_fb_location(adev);
>>> + amdgpu_vram_location(adev, &adev->mc, base);
>>> + adev->mc.gtt_base_align = 0;
>>> + amdgpu_gtt_location(adev, mc);
>>> +}
>>> +
>>> +/**
>>> + * gmc_v9_0_mc_init - initialize the memory controller driver params
>>> + *
>>> + * @adev: amdgpu_device pointer
>>> + *
>>> + * Look up the amount of vram, vram width, and decide how to place
>>> + * vram and gart within the GPU's physical address space.
>>> + * Returns 0 for success.
>>> + */
>>> +static int gmc_v9_0_mc_init(struct amdgpu_device *adev)
>>> +{
>>> + u32 tmp;
>>> + int chansize, numchan;
>>> +
>>> + /* hbm memory channel size */
>>> + chansize = 128;
>>> +
>>> + tmp = RREG32(SOC15_REG_OFFSET(DF, 0,
>>> mmDF_CS_AON0_DramBaseAddress0));
>>> + tmp &= DF_CS_AON0_DramBaseAddress0__IntLvNumChan_MASK;
>>> + tmp >>= DF_CS_AON0_DramBaseAddress0__IntLvNumChan__SHIFT;
>>> + switch (tmp) {
>>> + case 0:
>>> + default:
>>> + numchan = 1;
>>> + break;
>>> + case 1:
>>> + numchan = 2;
>>> + break;
>>> + case 2:
>>> + numchan = 0;
>>> + break;
>>> + case 3:
>>> + numchan = 4;
>>> + break;
>>> + case 4:
>>> + numchan = 0;
>>> + break;
>>> + case 5:
>>> + numchan = 8;
>>> + break;
>>> + case 6:
>>> + numchan = 0;
>>> + break;
>>> + case 7:
>>> + numchan = 16;
>>> + break;
>>> + case 8:
>>> + numchan = 2;
>>> + break;
>>> + }
>>> + adev->mc.vram_width = numchan * chansize;
>>> +
>>> + /* Could aper size report 0 ? */
>>> + adev->mc.aper_base = pci_resource_start(adev->pdev, 0);
>>> + adev->mc.aper_size = pci_resource_len(adev->pdev, 0);
>>> + /* size in MB on si */
>>> + adev->mc.mc_vram_size =
>>> + nbio_v6_1_get_memsize(adev) * 1024ULL * 1024ULL;
>>> + adev->mc.real_vram_size = adev->mc.mc_vram_size;
>>> + adev->mc.visible_vram_size = adev->mc.aper_size;
>>> +
>>> + /* In case the PCI BAR is larger than the actual amount of vram */
>>> + if (adev->mc.visible_vram_size > adev->mc.real_vram_size)
>>> + adev->mc.visible_vram_size = adev->mc.real_vram_size;
>>> +
>>> + /* unless the user had overridden it, set the gart
>>> + * size equal to the 1024 or vram, whichever is larger.
>>> + */
>>> + if (amdgpu_gart_size == -1)
>>> + adev->mc.gtt_size = max((1024ULL << 20),
>>> adev->mc.mc_vram_size);
>>> + else
>>> + adev->mc.gtt_size = (uint64_t)amdgpu_gart_size << 20;
>>> +
>>> + gmc_v9_0_vram_gtt_location(adev, &adev->mc);
>>> +
>>> + return 0;
>>> +}
>>> +
>>> +static int gmc_v9_0_gart_init(struct amdgpu_device *adev)
>>> +{
>>> + int r;
>>> +
>>> + if (adev->gart.robj) {
>>> + WARN(1, "VEGA10 PCIE GART already initialized\n");
>>> + return 0;
>>> + }
>>> + /* Initialize common gart structure */
>>> + r = amdgpu_gart_init(adev);
>>> + if (r)
>>> + return r;
>>> + adev->gart.table_size = adev->gart.num_gpu_pages * 8;
>>> + adev->gart.gart_pte_flags = AMDGPU_PTE_MTYPE(MTYPE_UC) |
>>> + AMDGPU_PTE_EXECUTABLE;
>>> + return amdgpu_gart_table_vram_alloc(adev);
>>> +}
>>> +
>>> +/*
>>> + * vm
>>> + * VMID 0 is the physical GPU addresses as used by the kernel.
>>> + * VMIDs 1-15 are used for userspace clients and are handled
>>> + * by the amdgpu vm/hsa code.
>>> + */
>>> +/**
>>> + * gmc_v9_0_vm_init - vm init callback
>>> + *
>>> + * @adev: amdgpu_device pointer
>>> + *
>>> + * Inits vega10 specific vm parameters (number of VMs, base of vram for
>>> + * VMIDs 1-15) (vega10).
>>> + * Returns 0 for success.
>>> + */
>>> +static int gmc_v9_0_vm_init(struct amdgpu_device *adev)
>>> +{
>>> + /*
>>> + * number of VMs
>>> + * VMID 0 is reserved for System
>>> + * amdgpu graphics/compute will use VMIDs 1-7
>>> + * amdkfd will use VMIDs 8-15
>>> + */
>>> + adev->vm_manager.num_ids = AMDGPU_NUM_OF_VMIDS;
>>> + amdgpu_vm_manager_init(adev);
>>> +
>>> + /* base offset of vram pages */
>>> + /*XXX This value is not zero for APU*/
>>> + adev->vm_manager.vram_base_offset = 0;
>>> +
>>> + return 0;
>>> +}
>>> +
>>> +/**
>>> + * gmc_v9_0_vm_fini - vm fini callback
>>> + *
>>> + * @adev: amdgpu_device pointer
>>> + *
>>> + * Tear down any asic specific VM setup.
>>> + */
>>> +static void gmc_v9_0_vm_fini(struct amdgpu_device *adev)
>>> +{
>>> + return;
>>> +}
>>> +
>>> +static int gmc_v9_0_sw_init(void *handle)
>>> +{
>>> + int r;
>>> + int dma_bits;
>>> + struct amdgpu_device *adev = (struct amdgpu_device *)handle;
>>> +
>>> + spin_lock_init(&adev->mc.invalidate_lock);
>>> +
>>> + if (adev->flags & AMD_IS_APU) {
>>> + adev->mc.vram_type = AMDGPU_VRAM_TYPE_UNKNOWN;
>>> + } else {
>>> + /* XXX Don't know how to get VRAM type yet. */
>>> + adev->mc.vram_type = AMDGPU_VRAM_TYPE_HBM;
>>> + }
>>> +
>>> + /* This interrupt is VMC page fault.*/
>>> + r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_VMC, 0,
>>> + &adev->mc.vm_fault);
>>> +
>>> + if (r)
>>> + return r;
>>> +
>>> + /* Adjust VM size here.
>>> + * Currently default to 64GB ((16 << 20) 4k pages).
>>> + * Max GPUVM size is 48 bits.
>>> + */
>>> + adev->vm_manager.max_pfn = amdgpu_vm_size << 18;
>>> +
>>> + /* Set the internal MC address mask
>>> + * This is the max address of the GPU's
>>> + * internal address space.
>>> + */
>>> + adev->mc.mc_mask = 0xffffffffffffULL; /* 48 bit MC */
>>> +
>>> + /* set DMA mask + need_dma32 flags.
>>> + * PCIE - can handle 44-bits.
>>> + * IGP - can handle 44-bits
>>> + * PCI - dma32 for legacy pci gart, 44 bits on vega10
>>> + */
>>> + adev->need_dma32 = false;
>>> + dma_bits = adev->need_dma32 ? 32 : 44;
>>> + r = pci_set_dma_mask(adev->pdev, DMA_BIT_MASK(dma_bits));
>>> + if (r) {
>>> + adev->need_dma32 = true;
>>> + dma_bits = 32;
>>> + printk(KERN_WARNING "amdgpu: No suitable DMA
>>> available.\n");
>>> + }
>>> + r = pci_set_consistent_dma_mask(adev->pdev,
>>> DMA_BIT_MASK(dma_bits));
>>> + if (r) {
>>> + pci_set_consistent_dma_mask(adev->pdev, DMA_BIT_MASK(32));
>>> + printk(KERN_WARNING "amdgpu: No coherent DMA
>>> available.\n");
>>> + }
>>> +
>>> + r = gmc_v9_0_mc_init(adev);
>>> + if (r)
>>> + return r;
>>> +
>>> + /* Memory manager */
>>> + r = amdgpu_bo_init(adev);
>>> + if (r)
>>> + return r;
>>> +
>>> + r = gmc_v9_0_gart_init(adev);
>>> + if (r)
>>> + return r;
>>> +
>>> + if (!adev->vm_manager.enabled) {
>>> + r = gmc_v9_0_vm_init(adev);
>>> + if (r) {
>>> + dev_err(adev->dev, "vm manager initialization
>>> failed (%d).\n", r);
>>> + return r;
>>> + }
>>> + adev->vm_manager.enabled = true;
>>> + }
>>> + return r;
>>> +}
>>> +
>>> +/**
>>> + * gmc_v8_0_gart_fini - vm fini callback
>>> + *
>>> + * @adev: amdgpu_device pointer
>>> + *
>>> + * Tears down the driver GART/VM setup (CIK).
>>> + */
>>> +static void gmc_v9_0_gart_fini(struct amdgpu_device *adev)
>>> +{
>>> + amdgpu_gart_table_vram_free(adev);
>>> + amdgpu_gart_fini(adev);
>>> +}
>>> +
>>> +static int gmc_v9_0_sw_fini(void *handle)
>>> +{
>>> + struct amdgpu_device *adev = (struct amdgpu_device *)handle;
>>> +
>>> + if (adev->vm_manager.enabled) {
>>> + amdgpu_vm_manager_fini(adev);
>>> + gmc_v9_0_vm_fini(adev);
>>> + adev->vm_manager.enabled = false;
>>> + }
>>> + gmc_v9_0_gart_fini(adev);
>>> + amdgpu_gem_force_release(adev);
>>> + amdgpu_bo_fini(adev);
>>> +
>>> + return 0;
>>> +}
>>> +
>>> +static void gmc_v9_0_init_golden_registers(struct amdgpu_device *adev)
>>> +{
>>> + switch (adev->asic_type) {
>>> + case CHIP_VEGA10:
>>> + break;
>>> + default:
>>> + break;
>>> + }
>>> +}
>>> +
>>> +/**
>>> + * gmc_v9_0_gart_enable - gart enable
>>> + *
>>> + * @adev: amdgpu_device pointer
>>> + */
>>> +static int gmc_v9_0_gart_enable(struct amdgpu_device *adev)
>>> +{
>>> + int r;
>>> + bool value;
>>> + u32 tmp;
>>> +
>>> + amdgpu_program_register_sequence(adev,
>>> + golden_settings_vega10_hdp,
>>> + (const u32)ARRAY_SIZE(golden_settings_vega10_hdp));
>>> +
>>> + if (adev->gart.robj == NULL) {
>>> + dev_err(adev->dev, "No VRAM object for PCIE GART.\n");
>>> + return -EINVAL;
>>> + }
>>> + r = amdgpu_gart_table_vram_pin(adev);
>>> + if (r)
>>> + return r;
>>> +
>>> + /* After HDP is initialized, flush HDP.*/
>>> + nbio_v6_1_hdp_flush(adev);
>>> +
>>> + r = gfxhub_v1_0_gart_enable(adev);
>>> + if (r)
>>> + return r;
>>> +
>>> + r = mmhub_v1_0_gart_enable(adev);
>>> + if (r)
>>> + return r;
>>> +
>>> + tmp = RREG32(SOC15_REG_OFFSET(HDP, 0, mmHDP_MISC_CNTL));
>>> + tmp |= HDP_MISC_CNTL__FLUSH_INVALIDATE_CACHE_MASK;
>>> + WREG32(SOC15_REG_OFFSET(HDP, 0, mmHDP_MISC_CNTL), tmp);
>>> +
>>> + tmp = RREG32(SOC15_REG_OFFSET(HDP, 0, mmHDP_HOST_PATH_CNTL));
>>> + WREG32(SOC15_REG_OFFSET(HDP, 0, mmHDP_HOST_PATH_CNTL), tmp);
>>> +
>>> +
>>> + if (amdgpu_vm_fault_stop == AMDGPU_VM_FAULT_STOP_ALWAYS)
>>> + value = false;
>>> + else
>>> + value = true;
>>> +
>>> + gfxhub_v1_0_set_fault_enable_default(adev, value);
>>> + mmhub_v1_0_set_fault_enable_default(adev, value);
>>> +
>>> + gmc_v9_0_gart_flush_gpu_tlb(adev, 0);
>>> +
>>> + DRM_INFO("PCIE GART of %uM enabled (table at 0x%016llX).\n",
>>> + (unsigned)(adev->mc.gtt_size >> 20),
>>> + (unsigned long long)adev->gart.table_addr);
>>> + adev->gart.ready = true;
>>> + return 0;
>>> +}
>>> +
>>> +static int gmc_v9_0_hw_init(void *handle)
>>> +{
>>> + int r;
>>> + struct amdgpu_device *adev = (struct amdgpu_device *)handle;
>>> +
>>> + /* The sequence of these two function calls matters.*/
>>> + gmc_v9_0_init_golden_registers(adev);
>>> +
>>> + r = gmc_v9_0_gart_enable(adev);
>>> +
>>> + return r;
>>> +}
>>> +
>>> +/**
>>> + * gmc_v9_0_gart_disable - gart disable
>>> + *
>>> + * @adev: amdgpu_device pointer
>>> + *
>>> + * This disables all VM page table.
>>> + */
>>> +static void gmc_v9_0_gart_disable(struct amdgpu_device *adev)
>>> +{
>>> + gfxhub_v1_0_gart_disable(adev);
>>> + mmhub_v1_0_gart_disable(adev);
>>> + amdgpu_gart_table_vram_unpin(adev);
>>> +}
>>> +
>>> +static int gmc_v9_0_hw_fini(void *handle)
>>> +{
>>> + struct amdgpu_device *adev = (struct amdgpu_device *)handle;
>>> +
>>> + amdgpu_irq_put(adev, &adev->mc.vm_fault, 0);
>>> + gmc_v9_0_gart_disable(adev);
>>> +
>>> + return 0;
>>> +}
>>> +
>>> +static int gmc_v9_0_suspend(void *handle)
>>> +{
>>> + struct amdgpu_device *adev = (struct amdgpu_device *)handle;
>>> +
>>> + if (adev->vm_manager.enabled) {
>>> + gmc_v9_0_vm_fini(adev);
>>> + adev->vm_manager.enabled = false;
>>> + }
>>> + gmc_v9_0_hw_fini(adev);
>>> +
>>> + return 0;
>>> +}
>>> +
>>> +static int gmc_v9_0_resume(void *handle)
>>> +{
>>> + int r;
>>> + struct amdgpu_device *adev = (struct amdgpu_device *)handle;
>>> +
>>> + r = gmc_v9_0_hw_init(adev);
>>> + if (r)
>>> + return r;
>>> +
>>> + if (!adev->vm_manager.enabled) {
>>> + r = gmc_v9_0_vm_init(adev);
>>> + if (r) {
>>> + dev_err(adev->dev,
>>> + "vm manager initialization failed
>>> (%d).\n", r);
>>> + return r;
>>> + }
>>> + adev->vm_manager.enabled = true;
>>> + }
>>> +
>>> + return r;
>>> +}
>>> +
>>> +static bool gmc_v9_0_is_idle(void *handle)
>>> +{
>>> + /* MC is always ready in GMC v9.*/
>>> + return true;
>>> +}
>>> +
>>> +static int gmc_v9_0_wait_for_idle(void *handle)
>>> +{
>>> + /* There is no need to wait for MC idle in GMC v9.*/
>>> + return 0;
>>> +}
>>> +
>>> +static int gmc_v9_0_soft_reset(void *handle)
>>> +{
>>> + /* XXX for emulation.*/
>>> + return 0;
>>> +}
>>> +
>>> +static int gmc_v9_0_set_clockgating_state(void *handle,
>>> + enum amd_clockgating_state state)
>>> +{
>>> + return 0;
>>> +}
>>> +
>>> +static int gmc_v9_0_set_powergating_state(void *handle,
>>> + enum amd_powergating_state state)
>>> +{
>>> + return 0;
>>> +}
>>> +
>>> +const struct amd_ip_funcs gmc_v9_0_ip_funcs = {
>>> + .name = "gmc_v9_0",
>>> + .early_init = gmc_v9_0_early_init,
>>> + .late_init = gmc_v9_0_late_init,
>>> + .sw_init = gmc_v9_0_sw_init,
>>> + .sw_fini = gmc_v9_0_sw_fini,
>>> + .hw_init = gmc_v9_0_hw_init,
>>> + .hw_fini = gmc_v9_0_hw_fini,
>>> + .suspend = gmc_v9_0_suspend,
>>> + .resume = gmc_v9_0_resume,
>>> + .is_idle = gmc_v9_0_is_idle,
>>> + .wait_for_idle = gmc_v9_0_wait_for_idle,
>>> + .soft_reset = gmc_v9_0_soft_reset,
>>> + .set_clockgating_state = gmc_v9_0_set_clockgating_state,
>>> + .set_powergating_state = gmc_v9_0_set_powergating_state,
>>> +};
>>> +
>>> +const struct amdgpu_ip_block_version gmc_v9_0_ip_block =
>>> +{
>>> + .type = AMD_IP_BLOCK_TYPE_GMC,
>>> + .major = 9,
>>> + .minor = 0,
>>> + .rev = 0,
>>> + .funcs = &gmc_v9_0_ip_funcs,
>>> +};
>>> diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.h
>>> b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.h
>>> new file mode 100644
>>> index 0000000..b030ca5
>>> --- /dev/null
>>> +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.h
>>> @@ -0,0 +1,30 @@
>>> +/*
>>> + * Copyright 2016 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 __GMC_V9_0_H__
>>> +#define __GMC_V9_0_H__
>>> +
>>> +extern const struct amd_ip_funcs gmc_v9_0_ip_funcs;
>>> +extern const struct amdgpu_ip_block_version gmc_v9_0_ip_block;
>>> +
>>> +#endif
>>> diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c
>>> b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c
>>> new file mode 100644
>>> index 0000000..b1e0e6b
>>> --- /dev/null
>>> +++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c
>>> @@ -0,0 +1,585 @@
>>> +/*
>>> + * Copyright 2016 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.
>>> + *
>>> + */
>>> +#include "amdgpu.h"
>>> +#include "mmhub_v1_0.h"
>>> +
>>> +#include "vega10/soc15ip.h"
>>> +#include "vega10/MMHUB/mmhub_1_0_offset.h"
>>> +#include "vega10/MMHUB/mmhub_1_0_sh_mask.h"
>>> +#include "vega10/MMHUB/mmhub_1_0_default.h"
>>> +#include "vega10/ATHUB/athub_1_0_offset.h"
>>> +#include "vega10/ATHUB/athub_1_0_sh_mask.h"
>>> +#include "vega10/ATHUB/athub_1_0_default.h"
>>> +#include "vega10/vega10_enum.h"
>>> +
>>> +#include "soc15_common.h"
>>> +
>>> +u64 mmhub_v1_0_get_fb_location(struct amdgpu_device *adev)
>>> +{
>>> + u64 base = RREG32(SOC15_REG_OFFSET(MMHUB, 0,
>>> mmMC_VM_FB_LOCATION_BASE));
>>> +
>>> + base &= MC_VM_FB_LOCATION_BASE__FB_BASE_MASK;
>>> + base <<= 24;
>>> +
>>> + return base;
>>> +}
>>> +
>>> +int mmhub_v1_0_gart_enable(struct amdgpu_device *adev)
>>> +{
>>> + u32 tmp;
>>> + u64 value;
>>> + uint64_t addr;
>>> + u32 i;
>>> +
>>> + /* Program MC. */
>>> + /* Update configuration */
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
next prev parent reply other threads:[~2017-03-23 2:54 UTC|newest]
Thread overview: 101+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-03-20 20:29 [PATCH 000/100] Add Vega10 Support Alex Deucher
[not found] ` <1490041835-11255-1-git-send-email-alexander.deucher-5C7GfCeVMHo@public.gmane.org>
2017-03-20 20:29 ` [PATCH 001/100] drm/amdgpu: add the new atomfirmware interface header Alex Deucher
2017-03-20 20:29 ` [PATCH 002/100] amdgpu: detect if we are using atomfirm or atombios for vbios (v2) Alex Deucher
2017-03-20 20:29 ` [PATCH 003/100] drm/amdgpu: move atom scratch setup into amdgpu_atombios.c Alex Deucher
2017-03-20 20:29 ` [PATCH 004/100] drm/amdgpu: add basic support for atomfirmware.h (v3) Alex Deucher
2017-03-20 20:29 ` [PATCH 005/100] drm/amdgpu: add soc15ip.h Alex Deucher
2017-03-20 20:29 ` [PATCH 021/100] drm/amd: Add MQD structs for GFX V9 Alex Deucher
2017-03-20 20:29 ` [PATCH 022/100] drm/amdgpu: add gfx9 clearstate header Alex Deucher
2017-03-20 20:29 ` [PATCH 023/100] drm/amdgpu: add SDMA 4.0 packet header Alex Deucher
2017-03-20 20:29 ` [PATCH 024/100] drm/amdgpu: add common soc15 headers Alex Deucher
2017-03-20 20:29 ` [PATCH 025/100] drm/amdgpu: add vega10 chip name Alex Deucher
2017-03-20 20:29 ` [PATCH 026/100] drm/amdgpu: add clinetid definition for vega10 Alex Deucher
2017-03-20 20:29 ` [PATCH 027/100] drm/amdgpu: use new flag to handle different firmware loading method Alex Deucher
2017-03-20 20:29 ` [PATCH 028/100] drm/amdgpu: gb_addr_config struct Alex Deucher
2017-03-20 20:29 ` [PATCH 029/100] drm/amdgpu: add 64bit doorbell assignments Alex Deucher
2017-03-20 20:29 ` [PATCH 030/100] drm/amdgpu: Add MTYPE flags to GPU VM IOCTL interface Alex Deucher
2017-03-20 20:29 ` [PATCH 031/100] drm/amdgpu: use atomfirmware interfaces for scratch reg save/restore Alex Deucher
2017-03-20 20:29 ` [PATCH 032/100] drm/amdgpu: update IH IV ring entry for soc-15 Alex Deucher
2017-03-20 20:29 ` [PATCH 033/100] drm/amdgpu: add IV trace point Alex Deucher
2017-03-20 20:29 ` [PATCH 034/100] drm/amdgpu: add PTE defines for MTYPE Alex Deucher
2017-03-20 20:29 ` [PATCH 035/100] drm/amdgpu: add NGG parameters Alex Deucher
2017-03-20 20:29 ` [PATCH 036/100] drm/amdgpu: Add asic family for vega10 Alex Deucher
2017-03-20 20:29 ` [PATCH 037/100] drm/amdgpu: add tiling flags for GFX9 Alex Deucher
2017-03-20 20:29 ` [PATCH 038/100] drm/amdgpu: don't validate TILE_SPLIT on GFX9 Alex Deucher
2017-03-20 20:29 ` [PATCH 039/100] drm/amdgpu: rework common ucode handling for vega10 Alex Deucher
2017-03-20 20:29 ` [PATCH 040/100] drm/amdgpu: add psp firmware header info Alex Deucher
2017-03-20 20:29 ` [PATCH 041/100] drm/amdgpu: get display info from DC when DC enabled Alex Deucher
2017-03-20 20:29 ` [PATCH 042/100] drm/amdgpu: gart fixes for vega10 Alex Deucher
2017-03-20 20:29 ` [PATCH 043/100] drm/amdgpu: handle PTE EXEC in amdgpu_vm_bo_split_mapping Alex Deucher
2017-03-20 20:29 ` [PATCH 044/100] drm/amdgpu: handle PTE MTYPE " Alex Deucher
2017-03-20 20:29 ` [PATCH 045/100] drm/amdgpu: add NBIO 6.1 driver Alex Deucher
2017-03-20 20:29 ` [PATCH 046/100] drm/amdgpu: Add GMC 9.0 support Alex Deucher
[not found] ` <1490041835-11255-32-git-send-email-alexander.deucher-5C7GfCeVMHo@public.gmane.org>
2017-03-21 8:49 ` Christian König
[not found] ` <003f0fba-4792-a32a-c982-73457dfbd1aa-ANTagKRnAhcb1SvskN2V4Q@public.gmane.org>
2017-03-21 15:09 ` Deucher, Alexander
[not found] ` <BN6PR12MB1652E0D9C22360FF77A4360AF73D0-/b2+HYfkarQqUD6E6FAiowdYzm3356FpvxpqHgZTriW3zl9H0oFU5g@public.gmane.org>
2017-03-22 16:51 ` Christian König
2017-03-22 19:41 ` Alex Deucher
2017-03-22 19:48 ` Dave Airlie
[not found] ` <CAPM=9tyv3RT0Q8i5zan_iaM7XxoTdb6nXuX=aT4C9vPPKjYfww-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-03-22 19:53 ` Deucher, Alexander
2017-03-23 2:42 ` Zhang, Jerry (Junwei)
[not found] ` <58D33609.3060304-5C7GfCeVMHo@public.gmane.org>
2017-03-23 2:53 ` Alex Deucher
[not found] ` <CADnq5_N_8r-pesBn1NwDrHJcOx8jnryhtcWow01Cbndj8B9N6w-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2017-03-23 2:54 ` Zhang, Jerry (Junwei) [this message]
2017-03-20 20:29 ` [PATCH 047/100] drm/amdgpu: add SDMA v4.0 implementation Alex Deucher
2017-03-20 20:29 ` [PATCH 048/100] drm/amdgpu: implement GFX 9.0 support Alex Deucher
2017-03-20 20:29 ` [PATCH 049/100] drm/amdgpu: add vega10 interrupt handler Alex Deucher
2017-03-20 20:29 ` [PATCH 050/100] drm/amdgpu: add initial uvd 7.0 support for vega10 Alex Deucher
2017-03-20 20:29 ` [PATCH 051/100] drm/amdgpu: add initial vce 4.0 " Alex Deucher
2017-03-20 20:29 ` [PATCH 052/100] drm/amdgpu: add PSP driver " Alex Deucher
2017-03-20 20:29 ` [PATCH 053/100] drm/amdgpu: add psp firmware info into info query and debugfs Alex Deucher
2017-03-20 20:29 ` [PATCH 054/100] drm/amdgpu: add SMC firmware into global ucode list for psp loading Alex Deucher
2017-03-20 20:29 ` [PATCH 055/100] drm/amd/powerplay: add smu9 header files for Vega10 Alex Deucher
2017-03-20 20:29 ` [PATCH 056/100] drm/amd/powerplay: add new Vega10's ppsmc header file Alex Deucher
2017-03-20 20:29 ` [PATCH 057/100] drm/amdgpu: add new atomfirmware based helpers for powerplay Alex Deucher
2017-03-20 20:29 ` [PATCH 058/100] drm/amd/powerplay: add global PowerPlay mutex Alex Deucher
2017-03-20 20:29 ` [PATCH 059/100] drm/amd/powerplay: add some new structures for Vega10 Alex Deucher
2017-03-20 20:29 ` [PATCH 060/100] drm/amd: add structures for display/powerplay interface Alex Deucher
2017-03-20 20:29 ` [PATCH 061/100] drm/amd/powerplay: add some display/powerplay interfaces Alex Deucher
2017-03-20 20:29 ` [PATCH 062/100] drm/amd/powerplay: add Vega10 powerplay support Alex Deucher
2017-03-20 20:29 ` [PATCH 063/100] drm/amd/display: Add DCE12 bios parser support Alex Deucher
[not found] ` <1490041835-11255-49-git-send-email-alexander.deucher-5C7GfCeVMHo@public.gmane.org>
2017-08-03 21:04 ` Mike Lothian
2017-03-20 20:29 ` [PATCH 064/100] drm/amd/display: Add DCE12 gpio support Alex Deucher
2017-03-20 20:30 ` [PATCH 065/100] drm/amd/display: Add DCE12 i2c/aux support Alex Deucher
2017-03-20 20:30 ` [PATCH 066/100] drm/amd/display: Add DCE12 irq support Alex Deucher
2017-03-20 20:30 ` [PATCH 067/100] drm/amd/display: Add DCE12 core support Alex Deucher
2017-03-20 20:30 ` [PATCH 068/100] drm/amd/display: Enable DCE12 support Alex Deucher
2017-03-20 20:30 ` [PATCH 069/100] drm/amd/display: need to handle DCE_Info table ver4.2 Alex Deucher
2017-03-20 20:30 ` [PATCH 070/100] drm/amd/display: Less log spam Alex Deucher
2017-03-20 20:30 ` [PATCH 071/100] drm/amdgpu: soc15 enable (v2) Alex Deucher
2017-03-20 20:30 ` [PATCH 072/100] drm/amdgpu: Set the IP blocks for vega10 Alex Deucher
2017-03-20 20:30 ` [PATCH 073/100] drm/amdgpu: add Vega10 Device IDs Alex Deucher
2017-03-20 20:30 ` [PATCH 074/100] drm/amdgpu/gfx9: programing wptr_poll_addr register Alex Deucher
2017-03-20 20:30 ` [PATCH 075/100] drm/amdgpu: impl sriov detection for vega10 Alex Deucher
2017-03-20 20:30 ` [PATCH 076/100] drm/amdgpu: add kiq ring for gfx9 Alex Deucher
2017-03-20 20:30 ` [PATCH 077/100] drm/amdgpu/gfx9: fullfill kiq funcs Alex Deucher
2017-03-20 20:30 ` [PATCH 078/100] drm/amdgpu/gfx9: fullfill kiq irq funcs Alex Deucher
2017-03-20 20:30 ` [PATCH 079/100] drm/amdgpu: init kiq and kcq for vega10 Alex Deucher
2017-03-20 20:30 ` [PATCH 080/100] drm/amdgpu:impl gfx9 cond_exec Alex Deucher
2017-03-20 20:30 ` [PATCH 081/100] drm/amdgpu/gfx9: impl gfx9 meta data emit Alex Deucher
2017-03-20 20:30 ` [PATCH 082/100] drm/amdgpu:bypass RLC init for SRIOV Alex Deucher
2017-03-20 20:30 ` [PATCH 083/100] drm/amdgpu/sdma4:re-org SDMA initial steps for sriov Alex Deucher
2017-03-20 20:30 ` [PATCH 084/100] drm/amdgpu/soc15: bypass PSP for VF Alex Deucher
2017-03-20 20:30 ` [PATCH 085/100] drm/amdgpu/gmc9: no need use kiq in vega10 tlb flush Alex Deucher
2017-03-20 20:30 ` [PATCH 086/100] drm/amdgpu/dce_virtual: bypass DPM for vf Alex Deucher
2017-03-20 20:30 ` [PATCH 087/100] drm/amdgpu/virt: impl mailbox for ai Alex Deucher
2017-03-20 20:30 ` [PATCH 088/100] drm/amdgpu/soc15: init virt ops for vf Alex Deucher
2017-03-20 20:30 ` [PATCH 089/100] drm/amdgpu/soc15: enable virtual dce " Alex Deucher
2017-03-20 20:30 ` [PATCH 090/100] drm/amdgpu/vega10:fix DOORBELL64 scheme Alex Deucher
2017-03-20 20:30 ` [PATCH 091/100] drm/amdgpu: Don't touch PG&CG for SRIOV MM Alex Deucher
2017-03-20 20:30 ` [PATCH 092/100] drm/amdgpu/vce4: enable doorbell for SRIOV Alex Deucher
2017-03-20 20:30 ` [PATCH 093/100] drm/amdgpu: disable uvd for sriov Alex Deucher
2017-03-20 20:30 ` [PATCH 094/100] drm/amdgpu/soc15: bypass pp block for vf Alex Deucher
2017-03-20 20:30 ` [PATCH 095/100] drm/amdgpu/virt: add structure for MM table Alex Deucher
2017-03-20 20:30 ` [PATCH 096/100] drm/amdgpu/vce4: alloc mm table for MM sriov Alex Deucher
2017-03-20 20:30 ` [PATCH 097/100] drm/amdgpu/vce4: Ignore vce ring/ib test temporarily Alex Deucher
2017-03-20 20:30 ` [PATCH 098/100] drm/amdgpu: add mmsch structures Alex Deucher
2017-03-20 20:30 ` [PATCH 099/100] drm/amdgpu/vce4: impl vce & mmsch sriov start Alex Deucher
2017-03-20 20:30 ` [PATCH 100/100] drm/amdgpu/gfx9: correct wptr pointer value Alex Deucher
2017-03-21 7:42 ` [PATCH 000/100] Add Vega10 Support Christian König
[not found] ` <50d03274-5a6e-fb77-9741-b6700a9949bd-ANTagKRnAhcb1SvskN2V4Q@public.gmane.org>
2017-03-21 11:51 ` Christian König
[not found] ` <b717602f-7573-6c20-ca68-491e3fe847c0-ANTagKRnAhcb1SvskN2V4Q@public.gmane.org>
2017-03-21 12:18 ` Christian König
[not found] ` <15b7d1b4-8ac7-d14b-40f6-aba529b301ea-ANTagKRnAhcb1SvskN2V4Q@public.gmane.org>
2017-03-21 15:54 ` Alex Deucher
2017-03-21 22:00 ` Alex Deucher
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=58D338F6.1090607@amd.com \
--to=jerry.zhang-5c7gfcevmho@public.gmane.org \
--cc=AlexBin.Xie-5C7GfCeVMHo@public.gmane.org \
--cc=alexander.deucher-5C7GfCeVMHo@public.gmane.org \
--cc=alexdeucher-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org \
--cc=amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.