From 88a1c40a04de534353fe7389654557aa886bc445 Mon Sep 17 00:00:00 2001 From: Leo Liu Date: Fri, 8 Nov 2019 10:00:24 -0500 Subject: drm/amdgpu: add JPEG HW IP and SW structures It will be used for JPEG IP 1.0, 2.0, 2.5 and later. Signed-off-by: Leo Liu Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h | 46 ++++++++++++++++++++++++++++++++ 1 file changed, 46 insertions(+) create mode 100644 drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h new file mode 100644 index 000000000000..36e2b7340c97 --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h @@ -0,0 +1,46 @@ +/* + * 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 __AMDGPU_JPEG_H__ +#define __AMDGPU_JPEG_H__ + +#define AMDGPU_MAX_JPEG_INSTANCES 2 + +struct amdgpu_jpeg_reg{ + unsigned jpeg_pitch; +}; + +struct amdgpu_jpeg_inst { + struct amdgpu_ring ring_dec; + struct amdgpu_irq_src irq; + struct amdgpu_jpeg_reg external; +}; + +struct amdgpu_jpeg { + uint8_t num_jpeg_inst; + struct amdgpu_jpeg_inst inst[AMDGPU_MAX_JPEG_INSTANCES]; + struct amdgpu_jpeg_reg internal; + unsigned harvest_config; +}; + +#endif /*__AMDGPU_JPEG_H__*/ -- cgit v1.2.3 From 9d9cc9b8fe85e7c7cdfc58b61830ca38f770cab1 Mon Sep 17 00:00:00 2001 From: Leo Liu Date: Fri, 8 Nov 2019 10:07:56 -0500 Subject: drm/amdgpu: add amdgpu_jpeg and JPEG tests It will be used for all versions of JPEG eventually. Previous JPEG tests will be removed later since they are still used by JPEG2.x. Signed-off-by: Leo Liu Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/Makefile | 5 +- drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c | 135 +++++++++++++++++++++++++++++++ drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h | 3 + 3 files changed, 141 insertions(+), 2 deletions(-) create mode 100644 drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h') diff --git a/drivers/gpu/drm/amd/amdgpu/Makefile b/drivers/gpu/drm/amd/amdgpu/Makefile index ca0e435559d5..b6c9d54266dc 100644 --- a/drivers/gpu/drm/amd/amdgpu/Makefile +++ b/drivers/gpu/drm/amd/amdgpu/Makefile @@ -147,12 +147,13 @@ amdgpu-y += \ vce_v3_0.o \ vce_v4_0.o -# add VCN block +# add VCN and JPEG block amdgpu-y += \ amdgpu_vcn.o \ vcn_v1_0.o \ vcn_v2_0.o \ - vcn_v2_5.o + vcn_v2_5.o \ + amdgpu_jpeg.o # add ATHUB block amdgpu-y += \ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c new file mode 100644 index 000000000000..d9a547d4d3b2 --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c @@ -0,0 +1,135 @@ +/* + * Copyright 2019 Advanced Micro Devices, Inc. + * All Rights Reserved. + * + * 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, sub license, 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 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 NON-INFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS 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. + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + */ + +#include "amdgpu.h" +#include "amdgpu_jpeg.h" +#include "soc15d.h" +#include "soc15_common.h" + +int amdgpu_jpeg_dec_ring_test_ring(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + uint32_t tmp = 0; + unsigned i; + int r; + + WREG32(adev->jpeg.inst[ring->me].external.jpeg_pitch, 0xCAFEDEAD); + r = amdgpu_ring_alloc(ring, 3); + if (r) + return r; + + amdgpu_ring_write(ring, PACKET0(adev->jpeg.internal.jpeg_pitch, 0)); + amdgpu_ring_write(ring, 0xDEADBEEF); + amdgpu_ring_commit(ring); + + for (i = 0; i < adev->usec_timeout; i++) { + tmp = RREG32(adev->jpeg.inst[ring->me].external.jpeg_pitch); + if (tmp == 0xDEADBEEF) + break; + udelay(1); + } + + if (i >= adev->usec_timeout) + r = -ETIMEDOUT; + + return r; +} + +static int amdgpu_jpeg_dec_set_reg(struct amdgpu_ring *ring, uint32_t handle, + struct dma_fence **fence) +{ + struct amdgpu_device *adev = ring->adev; + struct amdgpu_job *job; + struct amdgpu_ib *ib; + struct dma_fence *f = NULL; + const unsigned ib_size_dw = 16; + int i, r; + + r = amdgpu_job_alloc_with_ib(ring->adev, ib_size_dw * 4, &job); + if (r) + return r; + + ib = &job->ibs[0]; + + ib->ptr[0] = PACKETJ(adev->jpeg.internal.jpeg_pitch, 0, 0, PACKETJ_TYPE0); + ib->ptr[1] = 0xDEADBEEF; + for (i = 2; i < 16; i += 2) { + ib->ptr[i] = PACKETJ(0, 0, 0, PACKETJ_TYPE6); + ib->ptr[i+1] = 0; + } + ib->length_dw = 16; + + r = amdgpu_job_submit_direct(job, ring, &f); + if (r) + goto err; + + if (fence) + *fence = dma_fence_get(f); + dma_fence_put(f); + + return 0; + +err: + amdgpu_job_free(job); + return r; +} + +int amdgpu_jpeg_dec_ring_test_ib(struct amdgpu_ring *ring, long timeout) +{ + struct amdgpu_device *adev = ring->adev; + uint32_t tmp = 0; + unsigned i; + struct dma_fence *fence = NULL; + long r = 0; + + r = amdgpu_jpeg_dec_set_reg(ring, 1, &fence); + if (r) + goto error; + + r = dma_fence_wait_timeout(fence, false, timeout); + if (r == 0) { + r = -ETIMEDOUT; + goto error; + } else if (r < 0) { + goto error; + } else { + r = 0; + } + + for (i = 0; i < adev->usec_timeout; i++) { + tmp = RREG32(adev->jpeg.inst[ring->me].external.jpeg_pitch); + if (tmp == 0xDEADBEEF) + break; + udelay(1); + } + + if (i >= adev->usec_timeout) + r = -ETIMEDOUT; + + dma_fence_put(fence); +error: + return r; +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h index 36e2b7340c97..a8d988c25f45 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h @@ -43,4 +43,7 @@ struct amdgpu_jpeg { unsigned harvest_config; }; +int amdgpu_jpeg_dec_ring_test_ring(struct amdgpu_ring *ring); +int amdgpu_jpeg_dec_ring_test_ib(struct amdgpu_ring *ring, long timeout); + #endif /*__AMDGPU_JPEG_H__*/ -- cgit v1.2.3 From 2eb167293f5738c7d6b89ca154b5fddd04774db3 Mon Sep 17 00:00:00 2001 From: Leo Liu Date: Fri, 8 Nov 2019 13:12:05 -0500 Subject: drm/amdgpu: add JPEG common functions to amdgpu_jpeg They will be used for JPEG2.0 and later. Signed-off-by: Leo Liu Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c | 76 ++++++++++++++++++++++++++++++++ drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h | 10 +++++ 2 files changed, 86 insertions(+) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c index d9a547d4d3b2..5727f00afc8e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c @@ -26,9 +26,85 @@ #include "amdgpu.h" #include "amdgpu_jpeg.h" +#include "amdgpu_pm.h" #include "soc15d.h" #include "soc15_common.h" +#define JPEG_IDLE_TIMEOUT msecs_to_jiffies(1000) + +static void amdgpu_jpeg_idle_work_handler(struct work_struct *work); + +int amdgpu_jpeg_sw_init(struct amdgpu_device *adev) +{ + INIT_DELAYED_WORK(&adev->jpeg.idle_work, amdgpu_jpeg_idle_work_handler); + + return 0; +} + +int amdgpu_jpeg_sw_fini(struct amdgpu_device *adev) +{ + int i; + + cancel_delayed_work_sync(&adev->jpeg.idle_work); + + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { + if (adev->jpeg.harvest_config & (1 << i)) + continue; + + amdgpu_ring_fini(&adev->jpeg.inst[i].ring_dec); + } + + return 0; +} + +int amdgpu_jpeg_suspend(struct amdgpu_device *adev) +{ + cancel_delayed_work_sync(&adev->jpeg.idle_work); + + return 0; +} + +int amdgpu_jpeg_resume(struct amdgpu_device *adev) +{ + return 0; +} + +static void amdgpu_jpeg_idle_work_handler(struct work_struct *work) +{ + struct amdgpu_device *adev = + container_of(work, struct amdgpu_device, jpeg.idle_work.work); + unsigned int fences = 0; + unsigned int i; + + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { + if (adev->jpeg.harvest_config & (1 << i)) + continue; + + fences += amdgpu_fence_count_emitted(&adev->jpeg.inst[i].ring_dec); + } + + if (fences == 0) + amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_JPEG, + AMD_PG_STATE_GATE); + else + schedule_delayed_work(&adev->jpeg.idle_work, JPEG_IDLE_TIMEOUT); +} + +void amdgpu_jpeg_ring_begin_use(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + bool set_clocks = !cancel_delayed_work_sync(&adev->jpeg.idle_work); + + if (set_clocks) + amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_JPEG, + AMD_PG_STATE_UNGATE); +} + +void amdgpu_jpeg_ring_end_use(struct amdgpu_ring *ring) +{ + schedule_delayed_work(&ring->adev->jpeg.idle_work, JPEG_IDLE_TIMEOUT); +} + int amdgpu_jpeg_dec_ring_test_ring(struct amdgpu_ring *ring) { struct amdgpu_device *adev = ring->adev; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h index a8d988c25f45..5e2e06ec13df 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h @@ -41,8 +41,18 @@ struct amdgpu_jpeg { struct amdgpu_jpeg_inst inst[AMDGPU_MAX_JPEG_INSTANCES]; struct amdgpu_jpeg_reg internal; unsigned harvest_config; + struct delayed_work idle_work; + enum amd_powergating_state cur_state; }; +int amdgpu_jpeg_sw_init(struct amdgpu_device *adev); +int amdgpu_jpeg_sw_fini(struct amdgpu_device *adev); +int amdgpu_jpeg_suspend(struct amdgpu_device *adev); +int amdgpu_jpeg_resume(struct amdgpu_device *adev); + +void amdgpu_jpeg_ring_begin_use(struct amdgpu_ring *ring); +void amdgpu_jpeg_ring_end_use(struct amdgpu_ring *ring); + int amdgpu_jpeg_dec_ring_test_ring(struct amdgpu_ring *ring); int amdgpu_jpeg_dec_ring_test_ib(struct amdgpu_ring *ring, long timeout); -- cgit v1.2.3 From 14f43e8f88c57bbeaed3d05f13efab733f5e2338 Mon Sep 17 00:00:00 2001 From: Leo Liu Date: Mon, 11 Nov 2019 09:56:32 -0500 Subject: drm/amdgpu: move JPEG2.5 out from VCN2.5 And clean up the duplicated stuff Signed-off-by: Leo Liu Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/Makefile | 3 +- drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h | 3 + drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c | 105 ----- drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h | 5 - drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c | 641 +++++++++++++++++++++++++++++++ drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.h | 29 ++ drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c | 236 ------------ drivers/gpu/drm/amd/amdgpu/vcn_v2_0.h | 13 - drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c | 246 +----------- 9 files changed, 679 insertions(+), 602 deletions(-) create mode 100644 drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c create mode 100644 drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.h (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h') diff --git a/drivers/gpu/drm/amd/amdgpu/Makefile b/drivers/gpu/drm/amd/amdgpu/Makefile index c48b3bd73dfe..7ae3b22c5628 100644 --- a/drivers/gpu/drm/amd/amdgpu/Makefile +++ b/drivers/gpu/drm/amd/amdgpu/Makefile @@ -155,7 +155,8 @@ amdgpu-y += \ vcn_v2_5.o \ amdgpu_jpeg.o \ jpeg_v1_0.o \ - jpeg_v2_0.o + jpeg_v2_0.o \ + jpeg_v2_5.o # add ATHUB block amdgpu-y += \ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h index 5e2e06ec13df..5131a0a1bc8a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h @@ -26,6 +26,9 @@ #define AMDGPU_MAX_JPEG_INSTANCES 2 +#define AMDGPU_JPEG_HARVEST_JPEG0 (1 << 0) +#define AMDGPU_JPEG_HARVEST_JPEG1 (1 << 1) + struct amdgpu_jpeg_reg{ unsigned jpeg_pitch; }; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c index 4715115c8f06..428cfd58b37d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c @@ -706,108 +706,3 @@ error: amdgpu_bo_unref(&bo); return r; } - -int amdgpu_vcn_jpeg_ring_test_ring(struct amdgpu_ring *ring) -{ - struct amdgpu_device *adev = ring->adev; - uint32_t tmp = 0; - unsigned i; - int r; - - WREG32(adev->vcn.inst[ring->me].external.jpeg_pitch, 0xCAFEDEAD); - r = amdgpu_ring_alloc(ring, 3); - if (r) - return r; - - amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.jpeg_pitch, 0)); - amdgpu_ring_write(ring, 0xDEADBEEF); - amdgpu_ring_commit(ring); - - for (i = 0; i < adev->usec_timeout; i++) { - tmp = RREG32(adev->vcn.inst[ring->me].external.jpeg_pitch); - if (tmp == 0xDEADBEEF) - break; - udelay(1); - } - - if (i >= adev->usec_timeout) - r = -ETIMEDOUT; - - return r; -} - -static int amdgpu_vcn_jpeg_set_reg(struct amdgpu_ring *ring, uint32_t handle, - struct dma_fence **fence) -{ - struct amdgpu_device *adev = ring->adev; - struct amdgpu_job *job; - struct amdgpu_ib *ib; - struct dma_fence *f = NULL; - const unsigned ib_size_dw = 16; - int i, r; - - r = amdgpu_job_alloc_with_ib(ring->adev, ib_size_dw * 4, &job); - if (r) - return r; - - ib = &job->ibs[0]; - - ib->ptr[0] = PACKETJ(adev->vcn.internal.jpeg_pitch, 0, 0, PACKETJ_TYPE0); - ib->ptr[1] = 0xDEADBEEF; - for (i = 2; i < 16; i += 2) { - ib->ptr[i] = PACKETJ(0, 0, 0, PACKETJ_TYPE6); - ib->ptr[i+1] = 0; - } - ib->length_dw = 16; - - r = amdgpu_job_submit_direct(job, ring, &f); - if (r) - goto err; - - if (fence) - *fence = dma_fence_get(f); - dma_fence_put(f); - - return 0; - -err: - amdgpu_job_free(job); - return r; -} - -int amdgpu_vcn_jpeg_ring_test_ib(struct amdgpu_ring *ring, long timeout) -{ - struct amdgpu_device *adev = ring->adev; - uint32_t tmp = 0; - unsigned i; - struct dma_fence *fence = NULL; - long r = 0; - - r = amdgpu_vcn_jpeg_set_reg(ring, 1, &fence); - if (r) - goto error; - - r = dma_fence_wait_timeout(fence, false, timeout); - if (r == 0) { - r = -ETIMEDOUT; - goto error; - } else if (r < 0) { - goto error; - } else { - r = 0; - } - - for (i = 0; i < adev->usec_timeout; i++) { - tmp = RREG32(adev->vcn.inst[ring->me].external.jpeg_pitch); - if (tmp == 0xDEADBEEF) - break; - udelay(1); - } - - if (i >= adev->usec_timeout) - r = -ETIMEDOUT; - - dma_fence_put(fence); -error: - return r; -} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h index dface275c81a..402a5046b985 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h @@ -158,7 +158,6 @@ struct amdgpu_vcn_reg{ unsigned ib_size; unsigned gp_scratch8; unsigned scratch9; - unsigned jpeg_pitch; }; struct amdgpu_vcn_inst { @@ -168,7 +167,6 @@ struct amdgpu_vcn_inst { void *saved_bo; 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; }; @@ -209,7 +207,4 @@ int amdgpu_vcn_dec_ring_test_ib(struct amdgpu_ring *ring, long timeout); int amdgpu_vcn_enc_ring_test_ring(struct amdgpu_ring *ring); int amdgpu_vcn_enc_ring_test_ib(struct amdgpu_ring *ring, long timeout); -int amdgpu_vcn_jpeg_ring_test_ring(struct amdgpu_ring *ring); -int amdgpu_vcn_jpeg_ring_test_ib(struct amdgpu_ring *ring, long timeout); - #endif diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c new file mode 100644 index 000000000000..2c58939e6ad0 --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c @@ -0,0 +1,641 @@ +/* + * 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. + * + */ + +#include "amdgpu.h" +#include "amdgpu_jpeg.h" +#include "soc15.h" +#include "soc15d.h" +#include "jpeg_v2_0.h" + +#include "vcn/vcn_2_5_offset.h" +#include "vcn/vcn_2_5_sh_mask.h" +#include "ivsrcid/vcn/irqsrcs_vcn_2_0.h" + +#define mmUVD_JPEG_PITCH_INTERNAL_OFFSET 0x401f + +#define JPEG25_MAX_HW_INSTANCES_ARCTURUS 2 + +static void jpeg_v2_5_set_dec_ring_funcs(struct amdgpu_device *adev); +static void jpeg_v2_5_set_irq_funcs(struct amdgpu_device *adev); +static int jpeg_v2_5_set_powergating_state(void *handle, + enum amd_powergating_state state); + +static int amdgpu_ih_clientid_jpeg[] = { + SOC15_IH_CLIENTID_VCN, + SOC15_IH_CLIENTID_VCN1 +}; + +/** + * jpeg_v2_5_early_init - set function pointers + * + * @handle: amdgpu_device pointer + * + * Set ring and irq function pointers + */ +static int jpeg_v2_5_early_init(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + if (adev->asic_type == CHIP_ARCTURUS) { + u32 harvest; + int i; + + adev->jpeg.num_jpeg_inst = JPEG25_MAX_HW_INSTANCES_ARCTURUS; + for (i = 0; i < adev->jpeg.num_jpeg_inst; i++) { + harvest = RREG32_SOC15(JPEG, i, mmCC_UVD_HARVESTING); + if (harvest & CC_UVD_HARVESTING__UVD_DISABLE_MASK) + adev->jpeg.harvest_config |= 1 << i; + } + + if (adev->jpeg.harvest_config == (AMDGPU_JPEG_HARVEST_JPEG0 | + AMDGPU_JPEG_HARVEST_JPEG1)) + return -ENOENT; + } else + adev->jpeg.num_jpeg_inst = 1; + + jpeg_v2_5_set_dec_ring_funcs(adev); + jpeg_v2_5_set_irq_funcs(adev); + + return 0; +} + +/** + * jpeg_v2_5_sw_init - sw init for JPEG block + * + * @handle: amdgpu_device pointer + * + * Load firmware and sw initialization + */ +static int jpeg_v2_5_sw_init(void *handle) +{ + struct amdgpu_ring *ring; + int i, r; + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { + if (adev->jpeg.harvest_config & (1 << i)) + continue; + + /* JPEG TRAP */ + r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_jpeg[i], + VCN_2_0__SRCID__JPEG_DECODE, &adev->jpeg.inst[i].irq); + if (r) + return r; + } + + r = amdgpu_jpeg_sw_init(adev); + if (r) + return r; + + r = amdgpu_jpeg_resume(adev); + if (r) + return r; + + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { + if (adev->jpeg.harvest_config & (1 << i)) + continue; + + ring = &adev->jpeg.inst[i].ring_dec; + ring->use_doorbell = true; + ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 1 + 8 * i; + sprintf(ring->name, "jpeg_dec_%d", i); + r = amdgpu_ring_init(adev, ring, 512, &adev->jpeg.inst[i].irq, 0); + if (r) + return r; + + adev->jpeg.internal.jpeg_pitch = mmUVD_JPEG_PITCH_INTERNAL_OFFSET; + adev->jpeg.inst[i].external.jpeg_pitch = SOC15_REG_OFFSET(JPEG, i, mmUVD_JPEG_PITCH); + } + + return 0; +} + +/** + * jpeg_v2_5_sw_fini - sw fini for JPEG block + * + * @handle: amdgpu_device pointer + * + * JPEG suspend and free up sw allocation + */ +static int jpeg_v2_5_sw_fini(void *handle) +{ + int r; + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + r = amdgpu_jpeg_suspend(adev); + if (r) + return r; + + r = amdgpu_jpeg_sw_fini(adev); + + return r; +} + +/** + * jpeg_v2_5_hw_init - start and test JPEG block + * + * @handle: amdgpu_device pointer + * + */ +static int jpeg_v2_5_hw_init(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_ring *ring; + int i, r; + + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { + if (adev->jpeg.harvest_config & (1 << i)) + continue; + + ring = &adev->jpeg.inst[i].ring_dec; + adev->nbio.funcs->vcn_doorbell_range(adev, ring->use_doorbell, + (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 8 * i, i); + + r = amdgpu_ring_test_helper(ring); + if (r) + return r; + } + + DRM_INFO("JPEG decode initialized successfully.\n"); + + return 0; +} + +/** + * jpeg_v2_5_hw_fini - stop the hardware block + * + * @handle: amdgpu_device pointer + * + * Stop the JPEG block, mark ring as not ready any more + */ +static int jpeg_v2_5_hw_fini(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + struct amdgpu_ring *ring; + int i; + + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { + if (adev->jpeg.harvest_config & (1 << i)) + continue; + + ring = &adev->jpeg.inst[i].ring_dec; + if (adev->jpeg.cur_state != AMD_PG_STATE_GATE && + RREG32_SOC15(JPEG, i, mmUVD_JRBC_STATUS)) + jpeg_v2_5_set_powergating_state(adev, AMD_PG_STATE_GATE); + + ring->sched.ready = false; + } + + return 0; +} + +/** + * jpeg_v2_5_suspend - suspend JPEG block + * + * @handle: amdgpu_device pointer + * + * HW fini and suspend JPEG block + */ +static int jpeg_v2_5_suspend(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int r; + + r = jpeg_v2_5_hw_fini(adev); + if (r) + return r; + + r = amdgpu_jpeg_suspend(adev); + + return r; +} + +/** + * jpeg_v2_5_resume - resume JPEG block + * + * @handle: amdgpu_device pointer + * + * Resume firmware and hw init JPEG block + */ +static int jpeg_v2_5_resume(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int r; + + r = amdgpu_jpeg_resume(adev); + if (r) + return r; + + r = jpeg_v2_5_hw_init(adev); + + return r; +} + +static void jpeg_v2_5_disable_clock_gating(struct amdgpu_device* adev, int inst) +{ + uint32_t data; + + data = RREG32_SOC15(JPEG, inst, mmJPEG_CGC_CTRL); + if (adev->cg_flags & AMD_CG_SUPPORT_JPEG_MGCG) + data |= 1 << JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT; + else + data &= ~JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT; + + data |= 1 << JPEG_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT; + data |= 4 << JPEG_CGC_CTRL__CLK_OFF_DELAY__SHIFT; + WREG32_SOC15(JPEG, inst, mmJPEG_CGC_CTRL, data); + + data = RREG32_SOC15(JPEG, inst, mmJPEG_CGC_GATE); + data &= ~(JPEG_CGC_GATE__JPEG_DEC_MASK + | JPEG_CGC_GATE__JPEG2_DEC_MASK + | JPEG_CGC_GATE__JPEG_ENC_MASK + | JPEG_CGC_GATE__JMCIF_MASK + | JPEG_CGC_GATE__JRBBM_MASK); + WREG32_SOC15(JPEG, inst, mmJPEG_CGC_GATE, data); + + data = RREG32_SOC15(JPEG, inst, mmJPEG_CGC_CTRL); + data &= ~(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(JPEG, inst, mmJPEG_CGC_CTRL, data); +} + +static void jpeg_v2_5_enable_clock_gating(struct amdgpu_device* adev, int inst) +{ + uint32_t data; + + data = RREG32_SOC15(JPEG, inst, mmJPEG_CGC_GATE); + data |= (JPEG_CGC_GATE__JPEG_DEC_MASK + |JPEG_CGC_GATE__JPEG2_DEC_MASK + |JPEG_CGC_GATE__JPEG_ENC_MASK + |JPEG_CGC_GATE__JMCIF_MASK + |JPEG_CGC_GATE__JRBBM_MASK); + WREG32_SOC15(JPEG, inst, mmJPEG_CGC_GATE, 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; + int i; + + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { + if (adev->jpeg.harvest_config & (1 << i)) + continue; + + ring = &adev->jpeg.inst[i].ring_dec; + /* disable anti hang mechanism */ + WREG32_P(SOC15_REG_OFFSET(JPEG, i, mmUVD_JPEG_POWER_STATUS), 0, + ~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK); + + /* JPEG disable CGC */ + jpeg_v2_5_disable_clock_gating(adev, i); + + /* MJPEG global tiling registers */ + WREG32_SOC15(JPEG, i, mmJPEG_DEC_GFX8_ADDR_CONFIG, + adev->gfx.config.gb_addr_config); + WREG32_SOC15(JPEG, i, mmJPEG_DEC_GFX10_ADDR_CONFIG, + adev->gfx.config.gb_addr_config); + + /* enable JMI channel */ + WREG32_P(SOC15_REG_OFFSET(JPEG, i, mmUVD_JMI_CNTL), 0, + ~UVD_JMI_CNTL__SOFT_RESET_MASK); + + /* enable System Interrupt for JRBC */ + WREG32_P(SOC15_REG_OFFSET(JPEG, i, mmJPEG_SYS_INT_EN), + JPEG_SYS_INT_EN__DJRBC_MASK, + ~JPEG_SYS_INT_EN__DJRBC_MASK); + + WREG32_SOC15(JPEG, i, mmUVD_LMI_JRBC_RB_VMID, 0); + WREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_CNTL, (0x00000001L | 0x00000002L)); + WREG32_SOC15(JPEG, i, mmUVD_LMI_JRBC_RB_64BIT_BAR_LOW, + lower_32_bits(ring->gpu_addr)); + WREG32_SOC15(JPEG, i, mmUVD_LMI_JRBC_RB_64BIT_BAR_HIGH, + upper_32_bits(ring->gpu_addr)); + WREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_RPTR, 0); + WREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_WPTR, 0); + WREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_CNTL, 0x00000002L); + WREG32_SOC15(JPEG, i, mmUVD_JRBC_RB_SIZE, ring->ring_size / 4); + ring->wptr = RREG32_SOC15(JPEG, i, 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) +{ + int i; + + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { + if (adev->jpeg.harvest_config & (1 << i)) + continue; + + /* reset JMI */ + WREG32_P(SOC15_REG_OFFSET(JPEG, i, mmUVD_JMI_CNTL), + UVD_JMI_CNTL__SOFT_RESET_MASK, + ~UVD_JMI_CNTL__SOFT_RESET_MASK); + + jpeg_v2_5_enable_clock_gating(adev, i); + + /* enable anti hang mechanism */ + WREG32_P(SOC15_REG_OFFSET(JPEG, i, mmUVD_JPEG_POWER_STATUS), + UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK, + ~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK); + } + + return 0; +} + +/** + * jpeg_v2_5_dec_ring_get_rptr - get read pointer + * + * @ring: amdgpu_ring pointer + * + * Returns the current hardware read pointer + */ +static uint64_t jpeg_v2_5_dec_ring_get_rptr(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + + return RREG32_SOC15(JPEG, ring->me, mmUVD_JRBC_RB_RPTR); +} + +/** + * jpeg_v2_5_dec_ring_get_wptr - get write pointer + * + * @ring: amdgpu_ring pointer + * + * Returns the current hardware write pointer + */ +static uint64_t jpeg_v2_5_dec_ring_get_wptr(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + + if (ring->use_doorbell) + return adev->wb.wb[ring->wptr_offs]; + else + return RREG32_SOC15(JPEG, ring->me, mmUVD_JRBC_RB_WPTR); +} + +/** + * jpeg_v2_5_dec_ring_set_wptr - set write pointer + * + * @ring: amdgpu_ring pointer + * + * Commits the write pointer to the hardware + */ +static void jpeg_v2_5_dec_ring_set_wptr(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + + 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(JPEG, ring->me, mmUVD_JRBC_RB_WPTR, lower_32_bits(ring->wptr)); + } +} + +static bool jpeg_v2_5_is_idle(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int i, ret = 1; + + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { + if (adev->jpeg.harvest_config & (1 << i)) + continue; + + ret &= (((RREG32_SOC15(JPEG, i, mmUVD_JRBC_STATUS) & + UVD_JRBC_STATUS__RB_JOB_DONE_MASK) == + UVD_JRBC_STATUS__RB_JOB_DONE_MASK)); + } + + return ret; +} + +static int jpeg_v2_5_wait_for_idle(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int i, ret = 0; + + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { + if (adev->jpeg.harvest_config & (1 << i)) + continue; + + SOC15_WAIT_ON_RREG(JPEG, i, mmUVD_JRBC_STATUS, + UVD_JRBC_STATUS__RB_JOB_DONE_MASK, + UVD_JRBC_STATUS__RB_JOB_DONE_MASK, ret); + if (ret) + return ret; + } + + return ret; +} + +static int jpeg_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; + int i; + + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { + if (adev->jpeg.harvest_config & (1 << i)) + continue; + + if (enable) { + if (jpeg_v2_5_is_idle(handle)) + return -EBUSY; + jpeg_v2_5_enable_clock_gating(adev, i); + } else { + jpeg_v2_5_disable_clock_gating(adev, i); + } + } + + return 0; +} + +static int jpeg_v2_5_set_powergating_state(void *handle, + enum amd_powergating_state state) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int ret; + + if(state == adev->jpeg.cur_state) + return 0; + + if (state == AMD_PG_STATE_GATE) + ret = jpeg_v2_5_stop(adev); + else + ret = jpeg_v2_5_start(adev); + + if(!ret) + adev->jpeg.cur_state = state; + + return ret; +} + +static int jpeg_v2_5_set_interrupt_state(struct amdgpu_device *adev, + struct amdgpu_irq_src *source, + unsigned type, + enum amdgpu_interrupt_state state) +{ + return 0; +} + +static int jpeg_v2_5_process_interrupt(struct amdgpu_device *adev, + struct amdgpu_irq_src *source, + struct amdgpu_iv_entry *entry) +{ + uint32_t ip_instance; + + switch (entry->client_id) { + case SOC15_IH_CLIENTID_VCN: + ip_instance = 0; + break; + case SOC15_IH_CLIENTID_VCN1: + ip_instance = 1; + break; + default: + DRM_ERROR("Unhandled client id: %d\n", entry->client_id); + return 0; + } + + DRM_DEBUG("IH: JPEG TRAP\n"); + + switch (entry->src_id) { + case VCN_2_0__SRCID__JPEG_DECODE: + amdgpu_fence_process(&adev->jpeg.inst[ip_instance].ring_dec); + break; + default: + DRM_ERROR("Unhandled interrupt: %d %d\n", + entry->src_id, entry->src_data[0]); + break; + } + + return 0; +} + +static const struct amd_ip_funcs jpeg_v2_5_ip_funcs = { + .name = "jpeg_v2_5", + .early_init = jpeg_v2_5_early_init, + .late_init = NULL, + .sw_init = jpeg_v2_5_sw_init, + .sw_fini = jpeg_v2_5_sw_fini, + .hw_init = jpeg_v2_5_hw_init, + .hw_fini = jpeg_v2_5_hw_fini, + .suspend = jpeg_v2_5_suspend, + .resume = jpeg_v2_5_resume, + .is_idle = jpeg_v2_5_is_idle, + .wait_for_idle = jpeg_v2_5_wait_for_idle, + .check_soft_reset = NULL, + .pre_soft_reset = NULL, + .soft_reset = NULL, + .post_soft_reset = NULL, + .set_clockgating_state = jpeg_v2_5_set_clockgating_state, + .set_powergating_state = jpeg_v2_5_set_powergating_state, +}; + +static const struct amdgpu_ring_funcs jpeg_v2_5_dec_ring_vm_funcs = { + .type = AMDGPU_RING_TYPE_VCN_JPEG, + .align_mask = 0xf, + .vmhub = AMDGPU_MMHUB_1, + .get_rptr = jpeg_v2_5_dec_ring_get_rptr, + .get_wptr = jpeg_v2_5_dec_ring_get_wptr, + .set_wptr = jpeg_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 + /* jpeg_v2_5_dec_ring_emit_vm_flush */ + 18 + 18 + /* jpeg_v2_5_dec_ring_emit_fence x2 vm fence */ + 8 + 16, + .emit_ib_size = 22, /* jpeg_v2_5_dec_ring_emit_ib */ + .emit_ib = jpeg_v2_0_dec_ring_emit_ib, + .emit_fence = jpeg_v2_0_dec_ring_emit_fence, + .emit_vm_flush = jpeg_v2_0_dec_ring_emit_vm_flush, + .test_ring = amdgpu_jpeg_dec_ring_test_ring, + .test_ib = amdgpu_jpeg_dec_ring_test_ib, + .insert_nop = jpeg_v2_0_dec_ring_nop, + .insert_start = jpeg_v2_0_dec_ring_insert_start, + .insert_end = jpeg_v2_0_dec_ring_insert_end, + .pad_ib = amdgpu_ring_generic_pad_ib, + .begin_use = amdgpu_jpeg_ring_begin_use, + .end_use = amdgpu_jpeg_ring_end_use, + .emit_wreg = jpeg_v2_0_dec_ring_emit_wreg, + .emit_reg_wait = jpeg_v2_0_dec_ring_emit_reg_wait, + .emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper, +}; + +static void jpeg_v2_5_set_dec_ring_funcs(struct amdgpu_device *adev) +{ + int i; + + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { + if (adev->jpeg.harvest_config & (1 << i)) + continue; + + adev->jpeg.inst[i].ring_dec.funcs = &jpeg_v2_5_dec_ring_vm_funcs; + adev->jpeg.inst[i].ring_dec.me = i; + DRM_INFO("JPEG(%d) JPEG decode is enabled in VM mode\n", i); + } +} + +static const struct amdgpu_irq_src_funcs jpeg_v2_5_irq_funcs = { + .set = jpeg_v2_5_set_interrupt_state, + .process = jpeg_v2_5_process_interrupt, +}; + +static void jpeg_v2_5_set_irq_funcs(struct amdgpu_device *adev) +{ + int i; + + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { + if (adev->jpeg.harvest_config & (1 << i)) + continue; + + adev->jpeg.inst[i].irq.num_types = 1; + adev->jpeg.inst[i].irq.funcs = &jpeg_v2_5_irq_funcs; + } +} + +const struct amdgpu_ip_block_version jpeg_v2_5_ip_block = +{ + .type = AMD_IP_BLOCK_TYPE_JPEG, + .major = 2, + .minor = 5, + .rev = 0, + .funcs = &jpeg_v2_5_ip_funcs, +}; diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.h b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.h new file mode 100644 index 000000000000..2b4087c02620 --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.h @@ -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 __JPEG_V2_5_H__ +#define __JPEG_V2_5_H__ + +extern const struct amdgpu_ip_block_version jpeg_v2_5_ip_block; + +#endif /* __JPEG_V2_5_H__ */ diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c index ded0ab574f4e..5649190cb629 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c @@ -47,26 +47,6 @@ #define mmUVD_LMI_RBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET 0x5a7 #define mmUVD_RBC_IB_SIZE_INTERNAL_OFFSET 0x1e2 -#define mmUVD_JRBC_EXTERNAL_REG_INTERNAL_OFFSET 0x1bfff -#define mmUVD_JPEG_GPCOM_CMD_INTERNAL_OFFSET 0x4029 -#define mmUVD_JPEG_GPCOM_DATA0_INTERNAL_OFFSET 0x402a -#define mmUVD_JPEG_GPCOM_DATA1_INTERNAL_OFFSET 0x402b -#define mmUVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW_INTERNAL_OFFSET 0x40ea -#define mmUVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH_INTERNAL_OFFSET 0x40eb -#define mmUVD_LMI_JRBC_IB_VMID_INTERNAL_OFFSET 0x40cf -#define mmUVD_LMI_JPEG_VMID_INTERNAL_OFFSET 0x40d1 -#define mmUVD_LMI_JRBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET 0x40e8 -#define mmUVD_LMI_JRBC_IB_64BIT_BAR_HIGH_INTERNAL_OFFSET 0x40e9 -#define mmUVD_JRBC_IB_SIZE_INTERNAL_OFFSET 0x4082 -#define mmUVD_LMI_JRBC_RB_MEM_RD_64BIT_BAR_LOW_INTERNAL_OFFSET 0x40ec -#define mmUVD_LMI_JRBC_RB_MEM_RD_64BIT_BAR_HIGH_INTERNAL_OFFSET 0x40ed -#define mmUVD_JRBC_RB_COND_RD_TIMER_INTERNAL_OFFSET 0x4085 -#define mmUVD_JRBC_RB_REF_DATA_INTERNAL_OFFSET 0x4084 -#define mmUVD_JRBC_STATUS_INTERNAL_OFFSET 0x4089 -#define mmUVD_JPEG_PITCH_INTERNAL_OFFSET 0x401f - -#define JRBC_DEC_EXTERNAL_REG_WRITE_ADDR 0x18000 - #define mmUVD_RBC_XX_IB_REG_CHECK 0x026b #define mmUVD_RBC_XX_IB_REG_CHECK_BASE_IDX 1 #define mmUVD_REG_XX_MASK 0x026c @@ -1616,222 +1596,6 @@ void vcn_v2_0_enc_ring_emit_wreg(struct amdgpu_ring *ring, uint32_t reg, uint32_ amdgpu_ring_write(ring, val); } -/** - * vcn_v2_0_jpeg_ring_insert_start - insert a start command - * - * @ring: amdgpu_ring pointer - * - * Write a start command to the 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)); - amdgpu_ring_write(ring, 0x68e04); - - amdgpu_ring_write(ring, PACKETJ(JRBC_DEC_EXTERNAL_REG_WRITE_ADDR, - 0, 0, PACKETJ_TYPE0)); - amdgpu_ring_write(ring, 0x80010000); -} - -/** - * vcn_v2_0_jpeg_ring_insert_end - insert a end command - * - * @ring: amdgpu_ring pointer - * - * Write a end command to the 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)); - amdgpu_ring_write(ring, 0x68e04); - - amdgpu_ring_write(ring, PACKETJ(JRBC_DEC_EXTERNAL_REG_WRITE_ADDR, - 0, 0, PACKETJ_TYPE0)); - amdgpu_ring_write(ring, 0x00010000); -} - -/** - * vcn_v2_0_jpeg_ring_emit_fence - emit an fence & trap command - * - * @ring: amdgpu_ring pointer - * @fence: fence to emit - * - * Write a fence and a trap command to the ring. - */ -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); - - amdgpu_ring_write(ring, PACKETJ(mmUVD_JPEG_GPCOM_DATA0_INTERNAL_OFFSET, - 0, 0, PACKETJ_TYPE0)); - amdgpu_ring_write(ring, seq); - - amdgpu_ring_write(ring, PACKETJ(mmUVD_JPEG_GPCOM_DATA1_INTERNAL_OFFSET, - 0, 0, PACKETJ_TYPE0)); - amdgpu_ring_write(ring, seq); - - amdgpu_ring_write(ring, PACKETJ(mmUVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW_INTERNAL_OFFSET, - 0, 0, PACKETJ_TYPE0)); - amdgpu_ring_write(ring, lower_32_bits(addr)); - - amdgpu_ring_write(ring, PACKETJ(mmUVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH_INTERNAL_OFFSET, - 0, 0, PACKETJ_TYPE0)); - amdgpu_ring_write(ring, upper_32_bits(addr)); - - amdgpu_ring_write(ring, PACKETJ(mmUVD_JPEG_GPCOM_CMD_INTERNAL_OFFSET, - 0, 0, PACKETJ_TYPE0)); - amdgpu_ring_write(ring, 0x8); - - amdgpu_ring_write(ring, PACKETJ(mmUVD_JPEG_GPCOM_CMD_INTERNAL_OFFSET, - 0, PACKETJ_CONDITION_CHECK0, PACKETJ_TYPE4)); - amdgpu_ring_write(ring, 0); - - amdgpu_ring_write(ring, PACKETJ(mmUVD_JRBC_EXTERNAL_REG_INTERNAL_OFFSET, - 0, 0, PACKETJ_TYPE0)); - amdgpu_ring_write(ring, 0x3fbc); - - amdgpu_ring_write(ring, PACKETJ(JRBC_DEC_EXTERNAL_REG_WRITE_ADDR, - 0, 0, PACKETJ_TYPE0)); - amdgpu_ring_write(ring, 0x1); - - amdgpu_ring_write(ring, PACKETJ(0, 0, 0, PACKETJ_TYPE7)); - amdgpu_ring_write(ring, 0); -} - -/** - * vcn_v2_0_jpeg_ring_emit_ib - execute indirect buffer - * - * @ring: amdgpu_ring pointer - * @ib: indirect buffer to execute - * - * Write ring commands to execute the indirect buffer. - */ -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); - - amdgpu_ring_write(ring, PACKETJ(mmUVD_LMI_JRBC_IB_VMID_INTERNAL_OFFSET, - 0, 0, PACKETJ_TYPE0)); - amdgpu_ring_write(ring, (vmid | (vmid << 4))); - - amdgpu_ring_write(ring, PACKETJ(mmUVD_LMI_JPEG_VMID_INTERNAL_OFFSET, - 0, 0, PACKETJ_TYPE0)); - amdgpu_ring_write(ring, (vmid | (vmid << 4))); - - amdgpu_ring_write(ring, PACKETJ(mmUVD_LMI_JRBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET, - 0, 0, PACKETJ_TYPE0)); - amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr)); - - amdgpu_ring_write(ring, PACKETJ(mmUVD_LMI_JRBC_IB_64BIT_BAR_HIGH_INTERNAL_OFFSET, - 0, 0, PACKETJ_TYPE0)); - amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr)); - - amdgpu_ring_write(ring, PACKETJ(mmUVD_JRBC_IB_SIZE_INTERNAL_OFFSET, - 0, 0, PACKETJ_TYPE0)); - amdgpu_ring_write(ring, ib->length_dw); - - amdgpu_ring_write(ring, PACKETJ(mmUVD_LMI_JRBC_RB_MEM_RD_64BIT_BAR_LOW_INTERNAL_OFFSET, - 0, 0, PACKETJ_TYPE0)); - amdgpu_ring_write(ring, lower_32_bits(ring->gpu_addr)); - - amdgpu_ring_write(ring, PACKETJ(mmUVD_LMI_JRBC_RB_MEM_RD_64BIT_BAR_HIGH_INTERNAL_OFFSET, - 0, 0, PACKETJ_TYPE0)); - amdgpu_ring_write(ring, upper_32_bits(ring->gpu_addr)); - - amdgpu_ring_write(ring, PACKETJ(0, 0, PACKETJ_CONDITION_CHECK0, PACKETJ_TYPE2)); - amdgpu_ring_write(ring, 0); - - amdgpu_ring_write(ring, PACKETJ(mmUVD_JRBC_RB_COND_RD_TIMER_INTERNAL_OFFSET, - 0, 0, PACKETJ_TYPE0)); - amdgpu_ring_write(ring, 0x01400200); - - amdgpu_ring_write(ring, PACKETJ(mmUVD_JRBC_RB_REF_DATA_INTERNAL_OFFSET, - 0, 0, PACKETJ_TYPE0)); - amdgpu_ring_write(ring, 0x2); - - amdgpu_ring_write(ring, PACKETJ(mmUVD_JRBC_STATUS_INTERNAL_OFFSET, - 0, PACKETJ_CONDITION_CHECK3, PACKETJ_TYPE3)); - amdgpu_ring_write(ring, 0x2); -} - -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); - - amdgpu_ring_write(ring, PACKETJ(mmUVD_JRBC_RB_COND_RD_TIMER_INTERNAL_OFFSET, - 0, 0, PACKETJ_TYPE0)); - amdgpu_ring_write(ring, 0x01400200); - - amdgpu_ring_write(ring, PACKETJ(mmUVD_JRBC_RB_REF_DATA_INTERNAL_OFFSET, - 0, 0, PACKETJ_TYPE0)); - amdgpu_ring_write(ring, val); - - amdgpu_ring_write(ring, PACKETJ(mmUVD_JRBC_EXTERNAL_REG_INTERNAL_OFFSET, - 0, 0, PACKETJ_TYPE0)); - if (reg_offset >= 0x10000 && reg_offset <= 0x105ff) { - amdgpu_ring_write(ring, 0); - amdgpu_ring_write(ring, - PACKETJ((reg_offset >> 2), 0, 0, PACKETJ_TYPE3)); - } else { - amdgpu_ring_write(ring, reg_offset); - amdgpu_ring_write(ring, PACKETJ(JRBC_DEC_EXTERNAL_REG_WRITE_ADDR, - 0, 0, PACKETJ_TYPE3)); - } - amdgpu_ring_write(ring, mask); -} - -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; - - pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr); - - /* wait for register write */ - data0 = hub->ctx0_ptb_addr_lo32 + vmid * 2; - data1 = lower_32_bits(pd_addr); - mask = 0xffffffff; - vcn_v2_0_jpeg_ring_emit_reg_wait(ring, data0, data1, mask); -} - -void vcn_v2_0_jpeg_ring_emit_wreg(struct amdgpu_ring *ring, uint32_t reg, uint32_t val) -{ - uint32_t reg_offset = (reg << 2); - - amdgpu_ring_write(ring, PACKETJ(mmUVD_JRBC_EXTERNAL_REG_INTERNAL_OFFSET, - 0, 0, PACKETJ_TYPE0)); - if (reg_offset >= 0x10000 && reg_offset <= 0x105ff) { - amdgpu_ring_write(ring, 0); - amdgpu_ring_write(ring, - PACKETJ((reg_offset >> 2), 0, 0, PACKETJ_TYPE0)); - } else { - amdgpu_ring_write(ring, reg_offset); - amdgpu_ring_write(ring, PACKETJ(JRBC_DEC_EXTERNAL_REG_WRITE_ADDR, - 0, 0, PACKETJ_TYPE0)); - } - amdgpu_ring_write(ring, val); -} - -void vcn_v2_0_jpeg_ring_nop(struct amdgpu_ring *ring, uint32_t count) -{ - int i; - - WARN_ON(ring->wptr % 2 || count % 2); - - for (i = 0; i < count / 2; i++) { - amdgpu_ring_write(ring, PACKETJ(0, 0, 0, PACKETJ_TYPE6)); - amdgpu_ring_write(ring, 0); - } -} - static int vcn_v2_0_set_interrupt_state(struct amdgpu_device *adev, struct amdgpu_irq_src *source, unsigned type, diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.h b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.h index 8467292f32e5..ef749b02ded9 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.h +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.h @@ -49,19 +49,6 @@ 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__ */ diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c index 03083e5f731a..451dc814d845 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c @@ -47,13 +47,10 @@ #define mmUVD_LMI_RBC_IB_64BIT_BAR_HIGH_INTERNAL_OFFSET 0x3b5 #define mmUVD_RBC_IB_SIZE_INTERNAL_OFFSET 0x25c -#define mmUVD_JPEG_PITCH_INTERNAL_OFFSET 0x401f - -#define VCN25_MAX_HW_INSTANCES_ARCTURUS 2 +#define VCN25_MAX_HW_INSTANCES_ARCTURUS 2 static void vcn_v2_5_set_dec_ring_funcs(struct amdgpu_device *adev); static void vcn_v2_5_set_enc_ring_funcs(struct amdgpu_device *adev); -static void vcn_v2_5_set_jpeg_ring_funcs(struct amdgpu_device *adev); static void vcn_v2_5_set_irq_funcs(struct amdgpu_device *adev); static int vcn_v2_5_set_powergating_state(void *handle, enum amd_powergating_state state); @@ -95,7 +92,6 @@ static int vcn_v2_5_early_init(void *handle) vcn_v2_5_set_dec_ring_funcs(adev); vcn_v2_5_set_enc_ring_funcs(adev); - vcn_v2_5_set_jpeg_ring_funcs(adev); vcn_v2_5_set_irq_funcs(adev); return 0; @@ -130,12 +126,6 @@ static int vcn_v2_5_sw_init(void *handle) if (r) return r; } - - /* VCN JPEG TRAP */ - r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_vcns[j], - VCN_2_0__SRCID__JPEG_DECODE, &adev->vcn.inst[j].irq); - if (r) - return r; } r = amdgpu_vcn_sw_init(adev); @@ -184,9 +174,6 @@ static int vcn_v2_5_sw_init(void *handle) adev->vcn.internal.nop = mmUVD_NO_OP_INTERNAL_OFFSET; adev->vcn.inst[j].external.nop = SOC15_REG_OFFSET(UVD, j, mmUVD_NO_OP); - adev->vcn.internal.jpeg_pitch = mmUVD_JPEG_PITCH_INTERNAL_OFFSET; - adev->vcn.inst[j].external.jpeg_pitch = SOC15_REG_OFFSET(UVD, j, mmUVD_JPEG_PITCH); - ring = &adev->vcn.inst[j].ring_dec; ring->use_doorbell = true; ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 8*j; @@ -204,14 +191,6 @@ static int vcn_v2_5_sw_init(void *handle) if (r) return r; } - - ring = &adev->vcn.inst[j].ring_jpeg; - ring->use_doorbell = true; - ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 1 + 8*j; - sprintf(ring->name, "vcn_jpeg_%d", j); - r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.inst[j].irq, 0); - if (r) - return r; } return 0; @@ -269,12 +248,8 @@ static int vcn_v2_5_hw_init(void *handle) if (r) goto done; } - - ring = &adev->vcn.inst[j].ring_jpeg; - r = amdgpu_ring_test_helper(ring); - if (r) - goto done; } + done: if (!r) DRM_INFO("VCN decode and encode initialized successfully.\n"); @@ -309,9 +284,6 @@ static int vcn_v2_5_hw_fini(void *handle) ring = &adev->vcn.inst[i].ring_enc[i]; ring->sched.ready = false; } - - ring = &adev->vcn.inst[i].ring_jpeg; - ring->sched.ready = false; } return 0; @@ -592,115 +564,6 @@ static void vcn_v2_5_enable_clock_gating(struct amdgpu_device *adev) } } -/** - * 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; - uint32_t tmp; - int i; - - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->vcn.harvest_config & (1 << i)) - continue; - ring = &adev->vcn.inst[i].ring_jpeg; - /* disable anti hang mechanism */ - WREG32_P(SOC15_REG_OFFSET(UVD, i, mmUVD_JPEG_POWER_STATUS), 0, - ~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK); - - /* JPEG disable CGC */ - tmp = RREG32_SOC15(VCN, i, 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, i, mmJPEG_CGC_CTRL, tmp); - - tmp = RREG32_SOC15(VCN, i, 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, i, mmJPEG_CGC_GATE, tmp); - - tmp = RREG32_SOC15(VCN, i, 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, i, mmJPEG_CGC_CTRL, tmp); - - /* MJPEG global tiling registers */ - WREG32_SOC15(UVD, i, mmJPEG_DEC_GFX8_ADDR_CONFIG, - adev->gfx.config.gb_addr_config); - WREG32_SOC15(UVD, i, mmJPEG_DEC_GFX10_ADDR_CONFIG, - adev->gfx.config.gb_addr_config); - - /* enable JMI channel */ - WREG32_P(SOC15_REG_OFFSET(UVD, i, mmUVD_JMI_CNTL), 0, - ~UVD_JMI_CNTL__SOFT_RESET_MASK); - - /* enable System Interrupt for JRBC */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, mmJPEG_SYS_INT_EN), - JPEG_SYS_INT_EN__DJRBC_MASK, - ~JPEG_SYS_INT_EN__DJRBC_MASK); - - WREG32_SOC15(UVD, i, mmUVD_LMI_JRBC_RB_VMID, 0); - WREG32_SOC15(UVD, i, mmUVD_JRBC_RB_CNTL, (0x00000001L | 0x00000002L)); - WREG32_SOC15(UVD, i, mmUVD_LMI_JRBC_RB_64BIT_BAR_LOW, - lower_32_bits(ring->gpu_addr)); - WREG32_SOC15(UVD, i, mmUVD_LMI_JRBC_RB_64BIT_BAR_HIGH, - upper_32_bits(ring->gpu_addr)); - WREG32_SOC15(UVD, i, mmUVD_JRBC_RB_RPTR, 0); - WREG32_SOC15(UVD, i, mmUVD_JRBC_RB_WPTR, 0); - WREG32_SOC15(UVD, i, mmUVD_JRBC_RB_CNTL, 0x00000002L); - WREG32_SOC15(UVD, i, mmUVD_JRBC_RB_SIZE, ring->ring_size / 4); - ring->wptr = RREG32_SOC15(UVD, i, 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; - int i; - - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->vcn.harvest_config & (1 << i)) - continue; - /* reset JMI */ - WREG32_P(SOC15_REG_OFFSET(UVD, i, mmUVD_JMI_CNTL), - UVD_JMI_CNTL__SOFT_RESET_MASK, - ~UVD_JMI_CNTL__SOFT_RESET_MASK); - - tmp = RREG32_SOC15(VCN, i, 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, i, mmJPEG_CGC_GATE, tmp); - - /* enable anti hang mechanism */ - WREG32_P(SOC15_REG_OFFSET(UVD, i, 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; @@ -874,19 +737,14 @@ static int vcn_v2_5_start(struct amdgpu_device *adev) WREG32_SOC15(UVD, i, mmUVD_RB_BASE_HI2, upper_32_bits(ring->gpu_addr)); WREG32_SOC15(UVD, i, mmUVD_RB_SIZE2, ring->ring_size / 4); } - r = jpeg_v2_5_start(adev); - return r; + return 0; } static int vcn_v2_5_stop(struct amdgpu_device *adev) { uint32_t tmp; - int i, r; - - r = jpeg_v2_5_stop(adev); - if (r) - return r; + int i, r = 0; for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { if (adev->vcn.harvest_config & (1 << i)) @@ -1125,86 +983,6 @@ 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, ring->me, 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; - - if (ring->use_doorbell) - return adev->wb.wb[ring->wptr_offs]; - else - return RREG32_SOC15(UVD, ring->me, 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; - - 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, ring->me, 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) { int i; @@ -1233,19 +1011,6 @@ static void vcn_v2_5_set_enc_ring_funcs(struct amdgpu_device *adev) } } -static void vcn_v2_5_set_jpeg_ring_funcs(struct amdgpu_device *adev) -{ - int i; - - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->vcn.harvest_config & (1 << i)) - continue; - adev->vcn.inst[i].ring_jpeg.funcs = &vcn_v2_5_jpeg_ring_vm_funcs; - adev->vcn.inst[i].ring_jpeg.me = i; - DRM_INFO("VCN(%d) jpeg decode is enabled in VM mode\n", i); - } -} - static bool vcn_v2_5_is_idle(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; @@ -1352,9 +1117,6 @@ static int vcn_v2_5_process_interrupt(struct amdgpu_device *adev, case VCN_2_0__SRCID__UVD_ENC_LOW_LATENCY: amdgpu_fence_process(&adev->vcn.inst[ip_instance].ring_enc[1]); break; - case VCN_2_0__SRCID__JPEG_DECODE: - amdgpu_fence_process(&adev->vcn.inst[ip_instance].ring_jpeg); - break; default: DRM_ERROR("Unhandled interrupt: %d %d\n", entry->src_id, entry->src_data[0]); -- cgit v1.2.3 From f880799d7fcf0a63ca2295d950cd12f5520251d9 Mon Sep 17 00:00:00 2001 From: Nirmoy Das Date: Mon, 16 Dec 2019 14:43:34 +0100 Subject: amd/amdgpu: add sched array to IPs with multiple run-queues MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This sched array can be passed on to entity creation routine instead of manually creating such sched array on every context creation. v2: squash in missing break fix Signed-off-by: Nirmoy Das Reviewed-by: Christian König Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c | 113 ++++++++++++++++++----------- drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.h | 3 + drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 2 + drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h | 4 + drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h | 2 + drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h | 2 + drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h | 9 ++- 7 files changed, 89 insertions(+), 46 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c index 38ec5c919bd9..63f6365312d5 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c @@ -74,7 +74,7 @@ static int amdgpu_ctx_init(struct amdgpu_device *adev, struct amdgpu_ctx *ctx) { unsigned num_entities = amdgpu_ctx_total_num_entities(); - unsigned i, j, k; + unsigned i, j; int r; if (priority < 0 || priority >= DRM_SCHED_PRIORITY_MAX) @@ -121,73 +121,56 @@ static int amdgpu_ctx_init(struct amdgpu_device *adev, ctx->override_priority = DRM_SCHED_PRIORITY_UNSET; for (i = 0; i < AMDGPU_HW_IP_NUM; ++i) { - struct amdgpu_ring *rings[AMDGPU_MAX_RINGS]; - struct drm_gpu_scheduler *sched_list[AMDGPU_MAX_RINGS]; - unsigned num_rings = 0; - unsigned num_rqs = 0; + struct drm_gpu_scheduler **scheds; + struct drm_gpu_scheduler *sched; + unsigned num_scheds = 0; switch (i) { case AMDGPU_HW_IP_GFX: - rings[0] = &adev->gfx.gfx_ring[0]; - num_rings = 1; + scheds = adev->gfx.gfx_sched; + num_scheds = 1; break; case AMDGPU_HW_IP_COMPUTE: - for (j = 0; j < adev->gfx.num_compute_rings; ++j) - rings[j] = &adev->gfx.compute_ring[j]; - num_rings = adev->gfx.num_compute_rings; + scheds = adev->gfx.compute_sched; + num_scheds = adev->gfx.num_compute_sched; break; case AMDGPU_HW_IP_DMA: - for (j = 0; j < adev->sdma.num_instances; ++j) - rings[j] = &adev->sdma.instance[j].ring; - num_rings = adev->sdma.num_instances; + scheds = adev->sdma.sdma_sched; + num_scheds = adev->sdma.num_sdma_sched; break; case AMDGPU_HW_IP_UVD: - rings[0] = &adev->uvd.inst[0].ring; - num_rings = 1; + sched = &adev->uvd.inst[0].ring.sched; + scheds = &sched; + num_scheds = 1; break; case AMDGPU_HW_IP_VCE: - rings[0] = &adev->vce.ring[0]; - num_rings = 1; + sched = &adev->vce.ring[0].sched; + scheds = &sched; + num_scheds = 1; break; case AMDGPU_HW_IP_UVD_ENC: - rings[0] = &adev->uvd.inst[0].ring_enc[0]; - num_rings = 1; + sched = &adev->uvd.inst[0].ring_enc[0].sched; + scheds = &sched; + num_scheds = 1; break; case AMDGPU_HW_IP_VCN_DEC: - for (j = 0; j < adev->vcn.num_vcn_inst; ++j) { - if (adev->vcn.harvest_config & (1 << j)) - continue; - rings[num_rings++] = &adev->vcn.inst[j].ring_dec; - } + scheds = adev->vcn.vcn_dec_sched; + num_scheds = adev->vcn.num_vcn_dec_sched; break; case AMDGPU_HW_IP_VCN_ENC: - for (j = 0; j < adev->vcn.num_vcn_inst; ++j) { - if (adev->vcn.harvest_config & (1 << j)) - continue; - for (k = 0; k < adev->vcn.num_enc_rings; ++k) - rings[num_rings++] = &adev->vcn.inst[j].ring_enc[k]; - } + scheds = adev->vcn.vcn_enc_sched; + num_scheds = adev->vcn.num_vcn_enc_sched; break; case AMDGPU_HW_IP_VCN_JPEG: - for (j = 0; j < adev->jpeg.num_jpeg_inst; ++j) { - if (adev->jpeg.harvest_config & (1 << j)) - continue; - rings[num_rings++] = &adev->jpeg.inst[j].ring_dec; - } + scheds = adev->jpeg.jpeg_sched; + num_scheds = adev->jpeg.num_jpeg_sched; break; } - for (j = 0; j < num_rings; ++j) { - if (!rings[j]->adev) - continue; - - sched_list[num_rqs++] = &rings[j]->sched; - } - for (j = 0; j < amdgpu_ctx_num_entities[i]; ++j) r = drm_sched_entity_init(&ctx->entities[i][j].entity, - priority, sched_list, - num_rqs, &ctx->guilty); + priority, scheds, + num_scheds, &ctx->guilty); if (r) goto error_cleanup_entities; } @@ -628,3 +611,45 @@ void amdgpu_ctx_mgr_fini(struct amdgpu_ctx_mgr *mgr) idr_destroy(&mgr->ctx_handles); mutex_destroy(&mgr->lock); } + +void amdgpu_ctx_init_sched(struct amdgpu_device *adev) +{ + int i, j; + + for (i = 0; i < adev->gfx.num_gfx_rings; i++) { + adev->gfx.gfx_sched[i] = &adev->gfx.gfx_ring[i].sched; + adev->gfx.num_gfx_sched++; + } + + for (i = 0; i < adev->gfx.num_compute_rings; i++) { + adev->gfx.compute_sched[i] = &adev->gfx.compute_ring[i].sched; + adev->gfx.num_compute_sched++; + } + + for (i = 0; i < adev->sdma.num_instances; i++) { + adev->sdma.sdma_sched[i] = &adev->sdma.instance[i].ring.sched; + adev->sdma.num_sdma_sched++; + } + + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { + if (adev->vcn.harvest_config & (1 << i)) + continue; + adev->vcn.vcn_dec_sched[adev->vcn.num_vcn_dec_sched++] = + &adev->vcn.inst[i].ring_dec.sched; + } + + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { + if (adev->vcn.harvest_config & (1 << i)) + continue; + for (j = 0; j < adev->vcn.num_enc_rings; ++j) + adev->vcn.vcn_enc_sched[adev->vcn.num_vcn_enc_sched++] = + &adev->vcn.inst[i].ring_enc[j].sched; + } + + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { + if (adev->jpeg.harvest_config & (1 << i)) + continue; + adev->jpeg.jpeg_sched[adev->jpeg.num_jpeg_sched++] = + &adev->jpeg.inst[i].ring_dec.sched; + } +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.h index da808633732b..4ad90a44dc3c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.h @@ -87,4 +87,7 @@ void amdgpu_ctx_mgr_entity_fini(struct amdgpu_ctx_mgr *mgr); long amdgpu_ctx_mgr_entity_flush(struct amdgpu_ctx_mgr *mgr, long timeout); void amdgpu_ctx_mgr_fini(struct amdgpu_ctx_mgr *mgr); +void amdgpu_ctx_init_sched(struct amdgpu_device *adev); + + #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index db91663df4f6..e1f8b715301a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -3036,6 +3036,8 @@ fence_driver_init: adev->gfx.config.max_cu_per_sh, adev->gfx.cu_info.number); + amdgpu_ctx_init_sched(adev); + adev->accel_working = true; amdgpu_vm_check_compute_bug(adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h index 0ae0a2715b0d..8e88e0411662 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h @@ -269,8 +269,12 @@ struct amdgpu_gfx { bool me_fw_write_wait; bool cp_fw_write_wait; struct amdgpu_ring gfx_ring[AMDGPU_MAX_GFX_RINGS]; + struct drm_gpu_scheduler *gfx_sched[AMDGPU_MAX_GFX_RINGS]; + uint32_t num_gfx_sched; unsigned num_gfx_rings; struct amdgpu_ring compute_ring[AMDGPU_MAX_COMPUTE_RINGS]; + struct drm_gpu_scheduler *compute_sched[AMDGPU_MAX_COMPUTE_RINGS]; + uint32_t num_compute_sched; unsigned num_compute_rings; struct amdgpu_irq_src eop_irq; struct amdgpu_irq_src priv_reg_irq; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h index 5131a0a1bc8a..bd9ef9cc86de 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h @@ -43,6 +43,8 @@ struct amdgpu_jpeg { uint8_t num_jpeg_inst; struct amdgpu_jpeg_inst inst[AMDGPU_MAX_JPEG_INSTANCES]; struct amdgpu_jpeg_reg internal; + struct drm_gpu_scheduler *jpeg_sched[AMDGPU_MAX_JPEG_INSTANCES]; + uint32_t num_jpeg_sched; unsigned harvest_config; struct delayed_work idle_work; enum amd_powergating_state cur_state; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h index 761ff8be6314..346dcb1f7146 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h @@ -52,6 +52,8 @@ struct amdgpu_sdma_instance { struct amdgpu_sdma { struct amdgpu_sdma_instance instance[AMDGPU_MAX_SDMA_INSTANCES]; + struct drm_gpu_scheduler *sdma_sched[AMDGPU_MAX_SDMA_INSTANCES]; + uint32_t num_sdma_sched; struct amdgpu_irq_src trap_irq; struct amdgpu_irq_src illegal_inst_irq; struct amdgpu_irq_src ecc_irq; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h index 3484ead62046..e6dee8224d33 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h @@ -31,6 +31,7 @@ #define AMDGPU_VCN_MAX_ENC_RINGS 3 #define AMDGPU_MAX_VCN_INSTANCES 2 +#define AMDGPU_MAX_VCN_ENC_RINGS AMDGPU_VCN_MAX_ENC_RINGS * AMDGPU_MAX_VCN_INSTANCES #define AMDGPU_VCN_HARVEST_VCN0 (1 << 0) #define AMDGPU_VCN_HARVEST_VCN1 (1 << 1) @@ -189,8 +190,12 @@ struct amdgpu_vcn { 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; + struct amdgpu_vcn_inst inst[AMDGPU_MAX_VCN_INSTANCES]; + struct amdgpu_vcn_reg internal; + struct drm_gpu_scheduler *vcn_enc_sched[AMDGPU_MAX_VCN_ENC_RINGS]; + struct drm_gpu_scheduler *vcn_dec_sched[AMDGPU_MAX_VCN_INSTANCES]; + uint32_t num_vcn_enc_sched; + uint32_t num_vcn_dec_sched; unsigned harvest_config; int (*pause_dpg_mode)(struct amdgpu_device *adev, -- cgit v1.2.3