diff options
Diffstat (limited to 'drivers/gpu/drm/msm/adreno')
| -rw-r--r-- | drivers/gpu/drm/msm/adreno/a3xx_gpu.c | 8 | ||||
| -rw-r--r-- | drivers/gpu/drm/msm/adreno/a4xx_gpu.c | 8 | ||||
| -rw-r--r-- | drivers/gpu/drm/msm/adreno/a5xx_gpu.c | 11 | ||||
| -rw-r--r-- | drivers/gpu/drm/msm/adreno/a6xx.xml.h | 52 | ||||
| -rw-r--r-- | drivers/gpu/drm/msm/adreno/a6xx_gmu.c | 69 | ||||
| -rw-r--r-- | drivers/gpu/drm/msm/adreno/a6xx_gmu.h | 3 | ||||
| -rw-r--r-- | drivers/gpu/drm/msm/adreno/a6xx_gpu.c | 36 | ||||
| -rw-r--r-- | drivers/gpu/drm/msm/adreno/a6xx_gpu.h | 9 | ||||
| -rw-r--r-- | drivers/gpu/drm/msm/adreno/a6xx_gpu_state.c | 52 | ||||
| -rw-r--r-- | drivers/gpu/drm/msm/adreno/a6xx_gpu_state.h | 16 | ||||
| -rw-r--r-- | drivers/gpu/drm/msm/adreno/a6xx_hfi.c | 85 | ||||
| -rw-r--r-- | drivers/gpu/drm/msm/adreno/adreno_device.c | 11 | ||||
| -rw-r--r-- | drivers/gpu/drm/msm/adreno/adreno_gpu.c | 66 | ||||
| -rw-r--r-- | drivers/gpu/drm/msm/adreno/adreno_gpu.h | 17 |
14 files changed, 359 insertions, 84 deletions
diff --git a/drivers/gpu/drm/msm/adreno/a3xx_gpu.c b/drivers/gpu/drm/msm/adreno/a3xx_gpu.c index 7ad14937fcdf..b67f88872726 100644 --- a/drivers/gpu/drm/msm/adreno/a3xx_gpu.c +++ b/drivers/gpu/drm/msm/adreno/a3xx_gpu.c @@ -506,6 +506,14 @@ struct msm_gpu *a3xx_gpu_init(struct drm_device *dev) goto fail; } + /* + * Set the ICC path to maximum speed for now by multiplying the fastest + * frequency by the bus width (8). We'll want to scale this later on to + * improve battery life. + */ + icc_set_bw(gpu->icc_path, 0, Bps_to_icc(gpu->fast_rate) * 8); + icc_set_bw(gpu->ocmem_icc_path, 0, Bps_to_icc(gpu->fast_rate) * 8); + return gpu; fail: diff --git a/drivers/gpu/drm/msm/adreno/a4xx_gpu.c b/drivers/gpu/drm/msm/adreno/a4xx_gpu.c index b01388a9e89e..253d8d85daad 100644 --- a/drivers/gpu/drm/msm/adreno/a4xx_gpu.c +++ b/drivers/gpu/drm/msm/adreno/a4xx_gpu.c @@ -591,6 +591,14 @@ struct msm_gpu *a4xx_gpu_init(struct drm_device *dev) goto fail; } + /* + * Set the ICC path to maximum speed for now by multiplying the fastest + * frequency by the bus width (8). We'll want to scale this later on to + * improve battery life. + */ + icc_set_bw(gpu->icc_path, 0, Bps_to_icc(gpu->fast_rate) * 8); + icc_set_bw(gpu->ocmem_icc_path, 0, Bps_to_icc(gpu->fast_rate) * 8); + return gpu; fail: diff --git a/drivers/gpu/drm/msm/adreno/a5xx_gpu.c b/drivers/gpu/drm/msm/adreno/a5xx_gpu.c index b02e2042547f..7d9e63e20ded 100644 --- a/drivers/gpu/drm/msm/adreno/a5xx_gpu.c +++ b/drivers/gpu/drm/msm/adreno/a5xx_gpu.c @@ -753,11 +753,18 @@ static int a5xx_hw_init(struct msm_gpu *gpu) gpu->funcs->flush(gpu, gpu->rb[0]); if (!a5xx_idle(gpu, gpu->rb[0])) return -EINVAL; - } else { - /* Print a warning so if we die, we know why */ + } else if (ret == -ENODEV) { + /* + * This device does not use zap shader (but print a warning + * just in case someone got their dt wrong.. hopefully they + * have a debug UART to realize the error of their ways... + * if you mess this up you are about to crash horribly) + */ dev_warn_once(gpu->dev->dev, "Zap shader not enabled - using SECVID_TRUST_CNTL instead\n"); gpu_write(gpu, REG_A5XX_RBBM_SECVID_TRUST_CNTL, 0x0); + } else { + return ret; } /* Last step - yield the ringbuffer */ diff --git a/drivers/gpu/drm/msm/adreno/a6xx.xml.h b/drivers/gpu/drm/msm/adreno/a6xx.xml.h index f44553ec3193..ed78fee2a262 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx.xml.h +++ b/drivers/gpu/drm/msm/adreno/a6xx.xml.h @@ -16,11 +16,11 @@ The rules-ng-ng source files this header was generated from are: - /home/robclark/src/envytools/rnndb/adreno/a3xx.xml ( 83840 bytes, from 2018-07-03 19:37:13) - /home/robclark/src/envytools/rnndb/adreno/a4xx.xml ( 112086 bytes, from 2018-07-03 19:37:13) - /home/robclark/src/envytools/rnndb/adreno/a5xx.xml ( 147240 bytes, from 2018-12-02 17:29:54) -- /home/robclark/src/envytools/rnndb/adreno/a6xx.xml ( 140790 bytes, from 2018-12-02 17:29:54) +- /home/smasetty/playarea/envytools/rnndb/adreno/a6xx.xml ( 161969 bytes, from 2019-11-29 07:18:16) - /home/robclark/src/envytools/rnndb/adreno/a6xx_gmu.xml ( 10431 bytes, from 2018-09-14 13:03:07) - /home/robclark/src/envytools/rnndb/adreno/ocmem.xml ( 1773 bytes, from 2018-07-03 19:37:13) -Copyright (C) 2013-2018 by the following authors: +Copyright (C) 2013-2019 by the following authors: - Rob Clark <robdclark@gmail.com> (robclark) - Ilia Mirkin <imirkin@alum.mit.edu> (imirkin) @@ -2519,6 +2519,54 @@ static inline uint32_t A6XX_VBIF_TEST_BUS2_CTRL1_DATA_SEL(uint32_t val) #define REG_A6XX_VBIF_PERF_PWR_CNT_HIGH2 0x0000311a +#define REG_A6XX_GBIF_SCACHE_CNTL1 0x00003c02 + +#define REG_A6XX_GBIF_QSB_SIDE0 0x00003c03 + +#define REG_A6XX_GBIF_QSB_SIDE1 0x00003c04 + +#define REG_A6XX_GBIF_QSB_SIDE2 0x00003c05 + +#define REG_A6XX_GBIF_QSB_SIDE3 0x00003c06 + +#define REG_A6XX_GBIF_HALT 0x00003c45 + +#define REG_A6XX_GBIF_HALT_ACK 0x00003c46 + +#define REG_A6XX_GBIF_PERF_PWR_CNT_EN 0x00003cc0 + +#define REG_A6XX_GBIF_PERF_CNT_SEL 0x00003cc2 + +#define REG_A6XX_GBIF_PERF_PWR_CNT_SEL 0x00003cc3 + +#define REG_A6XX_GBIF_PERF_CNT_LOW0 0x00003cc4 + +#define REG_A6XX_GBIF_PERF_CNT_LOW1 0x00003cc5 + +#define REG_A6XX_GBIF_PERF_CNT_LOW2 0x00003cc6 + +#define REG_A6XX_GBIF_PERF_CNT_LOW3 0x00003cc7 + +#define REG_A6XX_GBIF_PERF_CNT_HIGH0 0x00003cc8 + +#define REG_A6XX_GBIF_PERF_CNT_HIGH1 0x00003cc9 + +#define REG_A6XX_GBIF_PERF_CNT_HIGH2 0x00003cca + +#define REG_A6XX_GBIF_PERF_CNT_HIGH3 0x00003ccb + +#define REG_A6XX_GBIF_PWR_CNT_LOW0 0x00003ccc + +#define REG_A6XX_GBIF_PWR_CNT_LOW1 0x00003ccd + +#define REG_A6XX_GBIF_PWR_CNT_LOW2 0x00003cce + +#define REG_A6XX_GBIF_PWR_CNT_HIGH0 0x00003ccf + +#define REG_A6XX_GBIF_PWR_CNT_HIGH1 0x00003cd0 + +#define REG_A6XX_GBIF_PWR_CNT_HIGH2 0x00003cd1 + #define REG_A6XX_RB_WINDOW_OFFSET2 0x000088d4 #define A6XX_RB_WINDOW_OFFSET2_WINDOW_OFFSET_DISABLE 0x80000000 #define A6XX_RB_WINDOW_OFFSET2_X__MASK 0x00007fff diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gmu.c b/drivers/gpu/drm/msm/adreno/a6xx_gmu.c index 85f14feafdec..748cd379065f 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_gmu.c +++ b/drivers/gpu/drm/msm/adreno/a6xx_gmu.c @@ -1,5 +1,5 @@ // SPDX-License-Identifier: GPL-2.0 -/* Copyright (c) 2017-2018 The Linux Foundation. All rights reserved. */ +/* Copyright (c) 2017-2019 The Linux Foundation. All rights reserved. */ #include <linux/clk.h> #include <linux/interconnect.h> @@ -149,6 +149,8 @@ void a6xx_gmu_set_freq(struct msm_gpu *gpu, unsigned long freq) if (freq == gmu->gpu_freqs[perf_index]) break; + gmu->current_perf_index = perf_index; + __a6xx_gmu_set_freq(gmu, perf_index); } @@ -433,6 +435,8 @@ static void __iomem *a6xx_gmu_get_mmio(struct platform_device *pdev, static void a6xx_gmu_rpmh_init(struct a6xx_gmu *gmu) { + struct a6xx_gpu *a6xx_gpu = container_of(gmu, struct a6xx_gpu, gmu); + struct adreno_gpu *adreno_gpu = &a6xx_gpu->base; struct platform_device *pdev = to_platform_device(gmu->dev); void __iomem *pdcptr = a6xx_gmu_get_mmio(pdev, "gmu_pdc"); void __iomem *seqptr = a6xx_gmu_get_mmio(pdev, "gmu_pdc_seq"); @@ -480,20 +484,34 @@ static void a6xx_gmu_rpmh_init(struct a6xx_gmu *gmu) pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD0_MSGID + 4, 0x10108); pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD0_ADDR + 4, 0x30000); pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD0_DATA + 4, 0x0); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD0_MSGID + 8, 0x10108); - pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD0_ADDR + 8, 0x30080); + if (adreno_is_a618(adreno_gpu)) + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD0_ADDR + 8, 0x30090); + else + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD0_ADDR + 8, 0x30080); pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS1_CMD0_DATA + 8, 0x0); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD_ENABLE_BANK, 7); pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD_WAIT_FOR_CMPL_BANK, 0); pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CONTROL, 0); pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_MSGID, 0x10108); pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_ADDR, 0x30010); pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_DATA, 2); + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_MSGID + 4, 0x10108); pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_ADDR + 4, 0x30000); - pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_DATA + 4, 0x3); + if (adreno_is_a618(adreno_gpu)) + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_DATA + 4, 0x2); + else + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_DATA + 4, 0x3); + + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_MSGID + 8, 0x10108); - pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_ADDR + 8, 0x30080); + if (adreno_is_a618(adreno_gpu)) + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_ADDR + 8, 0x30090); + else + pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_ADDR + 8, 0x30080); pdc_write(pdcptr, REG_A6XX_PDC_GPU_TCS3_CMD0_DATA + 8, 0x3); /* Setup GPU PDC */ @@ -741,8 +759,8 @@ int a6xx_gmu_resume(struct a6xx_gpu *a6xx_gpu) gmu_write(gmu, REG_A6XX_GMU_GMU2HOST_INTR_MASK, ~A6XX_HFI_IRQ_MASK); enable_irq(gmu->hfi_irq); - /* Set the GPU to the highest power frequency */ - __a6xx_gmu_set_freq(gmu, gmu->nr_gpu_freqs - 1); + /* Set the GPU to the current freq */ + __a6xx_gmu_set_freq(gmu, gmu->current_perf_index); /* * "enable" the GX power domain which won't actually do anything but it @@ -778,12 +796,41 @@ bool a6xx_gmu_isidle(struct a6xx_gmu *gmu) return true; } +#define GBIF_CLIENT_HALT_MASK BIT(0) +#define GBIF_ARB_HALT_MASK BIT(1) + +static void a6xx_bus_clear_pending_transactions(struct adreno_gpu *adreno_gpu) +{ + struct msm_gpu *gpu = &adreno_gpu->base; + + if (!a6xx_has_gbif(adreno_gpu)) { + gpu_write(gpu, REG_A6XX_VBIF_XIN_HALT_CTRL0, 0xf); + spin_until((gpu_read(gpu, REG_A6XX_VBIF_XIN_HALT_CTRL1) & + 0xf) == 0xf); + gpu_write(gpu, REG_A6XX_VBIF_XIN_HALT_CTRL0, 0); + + return; + } + + /* Halt new client requests on GBIF */ + gpu_write(gpu, REG_A6XX_GBIF_HALT, GBIF_CLIENT_HALT_MASK); + spin_until((gpu_read(gpu, REG_A6XX_GBIF_HALT_ACK) & + (GBIF_CLIENT_HALT_MASK)) == GBIF_CLIENT_HALT_MASK); + + /* Halt all AXI requests on GBIF */ + gpu_write(gpu, REG_A6XX_GBIF_HALT, GBIF_ARB_HALT_MASK); + spin_until((gpu_read(gpu, REG_A6XX_GBIF_HALT_ACK) & + (GBIF_ARB_HALT_MASK)) == GBIF_ARB_HALT_MASK); + + /* The GBIF halt needs to be explicitly cleared */ + gpu_write(gpu, REG_A6XX_GBIF_HALT, 0x0); +} + /* Gracefully try to shut down the GMU and by extension the GPU */ static void a6xx_gmu_shutdown(struct a6xx_gmu *gmu) { struct a6xx_gpu *a6xx_gpu = container_of(gmu, struct a6xx_gpu, gmu); struct adreno_gpu *adreno_gpu = &a6xx_gpu->base; - struct msm_gpu *gpu = &adreno_gpu->base; u32 val; /* @@ -801,11 +848,7 @@ static void a6xx_gmu_shutdown(struct a6xx_gmu *gmu) return; } - /* Clear the VBIF pipe before shutting down */ - gpu_write(gpu, REG_A6XX_VBIF_XIN_HALT_CTRL0, 0xf); - spin_until((gpu_read(gpu, REG_A6XX_VBIF_XIN_HALT_CTRL1) & 0xf) - == 0xf); - gpu_write(gpu, REG_A6XX_VBIF_XIN_HALT_CTRL0, 0); + a6xx_bus_clear_pending_transactions(adreno_gpu); /* tell the GMU we want to slumber */ a6xx_gmu_notify_slumber(gmu); @@ -1166,6 +1209,8 @@ static int a6xx_gmu_pwrlevels_probe(struct a6xx_gmu *gmu) gmu->nr_gpu_freqs = a6xx_gmu_build_freq_table(&gpu->pdev->dev, gmu->gpu_freqs, ARRAY_SIZE(gmu->gpu_freqs)); + gmu->current_perf_index = gmu->nr_gpu_freqs - 1; + /* Build the list of RPMh votes that we'll send to the GMU */ return a6xx_gmu_rpmh_votes_init(gmu); } diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gmu.h b/drivers/gpu/drm/msm/adreno/a6xx_gmu.h index 39a26dd63674..2af91ed7ed0c 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_gmu.h +++ b/drivers/gpu/drm/msm/adreno/a6xx_gmu.h @@ -63,6 +63,9 @@ struct a6xx_gmu { struct clk_bulk_data *clocks; struct clk *core_clk; + /* current performance index set externally */ + int current_perf_index; + int nr_gpu_freqs; unsigned long gpu_freqs[16]; u32 gx_arc_votes[16]; diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c index dc8ec2c94301..68af24150de5 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c +++ b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c @@ -1,5 +1,5 @@ // SPDX-License-Identifier: GPL-2.0 -/* Copyright (c) 2017-2018 The Linux Foundation. All rights reserved. */ +/* Copyright (c) 2017-2019 The Linux Foundation. All rights reserved. */ #include "msm_gem.h" @@ -406,12 +406,17 @@ static int a6xx_hw_init(struct msm_gpu *gpu) gpu_write(gpu, REG_A6XX_TPL1_ADDR_MODE_CNTL, 0x1); gpu_write(gpu, REG_A6XX_RBBM_SECVID_TSB_ADDR_MODE_CNTL, 0x1); - /* enable hardware clockgating */ - a6xx_set_hwcg(gpu, true); + /* + * enable hardware clockgating + * For now enable clock gating only for a630 + */ + if (adreno_is_a630(adreno_gpu)) + a6xx_set_hwcg(gpu, true); - /* VBIF start */ - gpu_write(gpu, REG_A6XX_VBIF_GATE_OFF_WRREQ_EN, 0x00000009); + /* VBIF/GBIF start*/ gpu_write(gpu, REG_A6XX_RBBM_VBIF_CLIENT_QOS_CNTL, 0x3); + if (adreno_is_a630(adreno_gpu)) + gpu_write(gpu, REG_A6XX_VBIF_GATE_OFF_WRREQ_EN, 0x00000009); /* Make all blocks contribute to the GPU BUSY perf counter */ gpu_write(gpu, REG_A6XX_RBBM_PERFCTR_GPU_BUSY_MASKED, 0xffffffff); @@ -453,10 +458,12 @@ static int a6xx_hw_init(struct msm_gpu *gpu) /* Select CP0 to always count cycles */ gpu_write(gpu, REG_A6XX_CP_PERFCTR_CP_SEL_0, PERF_CP_ALWAYS_COUNT); - gpu_write(gpu, REG_A6XX_RB_NC_MODE_CNTL, 2 << 1); - gpu_write(gpu, REG_A6XX_TPL1_NC_MODE_CNTL, 2 << 1); - gpu_write(gpu, REG_A6XX_SP_NC_MODE_CNTL, 2 << 1); - gpu_write(gpu, REG_A6XX_UCHE_MODE_CNTL, 2 << 21); + if (adreno_is_a630(adreno_gpu)) { + gpu_write(gpu, REG_A6XX_RB_NC_MODE_CNTL, 2 << 1); + gpu_write(gpu, REG_A6XX_TPL1_NC_MODE_CNTL, 2 << 1); + gpu_write(gpu, REG_A6XX_SP_NC_MODE_CNTL, 2 << 1); + gpu_write(gpu, REG_A6XX_UCHE_MODE_CNTL, 2 << 21); + } /* Enable fault detection */ gpu_write(gpu, REG_A6XX_RBBM_INTERFACE_HANG_INT_CNTL, @@ -537,12 +544,19 @@ static int a6xx_hw_init(struct msm_gpu *gpu) a6xx_flush(gpu, gpu->rb[0]); if (!a6xx_idle(gpu, gpu->rb[0])) return -EINVAL; - } else { - /* Print a warning so if we die, we know why */ + } else if (ret == -ENODEV) { + /* + * This device does not use zap shader (but print a warning + * just in case someone got their dt wrong.. hopefully they + * have a debug UART to realize the error of their ways... + * if you mess this up you are about to crash horribly) + */ dev_warn_once(gpu->dev->dev, "Zap shader not enabled - using SECVID_TRUST_CNTL instead\n"); gpu_write(gpu, REG_A6XX_RBBM_SECVID_TRUST_CNTL, 0x0); ret = 0; + } else { + return ret; } out: diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gpu.h b/drivers/gpu/drm/msm/adreno/a6xx_gpu.h index 64399554f2dd..7239b8b60939 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_gpu.h +++ b/drivers/gpu/drm/msm/adreno/a6xx_gpu.h @@ -1,5 +1,5 @@ /* SPDX-License-Identifier: GPL-2.0 */ -/* Copyright (c) 2017 The Linux Foundation. All rights reserved. */ +/* Copyright (c) 2017, 2019 The Linux Foundation. All rights reserved. */ #ifndef __A6XX_GPU_H__ #define __A6XX_GPU_H__ @@ -42,6 +42,13 @@ struct a6xx_gpu { #define A6XX_PROTECT_RDONLY(_reg, _len) \ ((((_len) & 0x3FFF) << 18) | ((_reg) & 0x3FFFF)) +static inline bool a6xx_has_gbif(struct adreno_gpu *gpu) +{ + if(adreno_is_a630(gpu)) + return false; + + return true; +} int a6xx_gmu_resume(struct a6xx_gpu *gpu); int a6xx_gmu_stop(struct a6xx_gpu *gpu); diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gpu_state.c b/drivers/gpu/drm/msm/adreno/a6xx_gpu_state.c index 691c1a277d91..d6023ba8033c 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_gpu_state.c +++ b/drivers/gpu/drm/msm/adreno/a6xx_gpu_state.c @@ -1,5 +1,5 @@ // SPDX-License-Identifier: GPL-2.0 -/* Copyright (c) 2018 The Linux Foundation. All rights reserved. */ +/* Copyright (c) 2018-2019 The Linux Foundation. All rights reserved. */ #include <linux/ascii85.h> #include "msm_gem.h" @@ -320,6 +320,7 @@ static void a6xx_get_debugbus(struct msm_gpu *gpu, { struct resource *res; void __iomem *cxdbg = NULL; + int nr_debugbus_blocks; /* Set up the GX debug bus */ @@ -374,9 +375,11 @@ static void a6xx_get_debugbus(struct msm_gpu *gpu, cxdbg_write(cxdbg, REG_A6XX_CX_DBGC_CFG_DBGBUS_MASKL_3, 0); } - a6xx_state->debugbus = state_kcalloc(a6xx_state, - ARRAY_SIZE(a6xx_debugbus_blocks), - sizeof(*a6xx_state->debugbus)); + nr_debugbus_blocks = ARRAY_SIZE(a6xx_debugbus_blocks) + + (a6xx_has_gbif(to_adreno_gpu(gpu)) ? 1 : 0); + + a6xx_state->debugbus = state_kcalloc(a6xx_state, nr_debugbus_blocks, + sizeof(*a6xx_state->debugbus)); if (a6xx_state->debugbus) { int i; @@ -388,15 +391,31 @@ static void a6xx_get_debugbus(struct msm_gpu *gpu, &a6xx_state->debugbus[i]); a6xx_state->nr_debugbus = ARRAY_SIZE(a6xx_debugbus_blocks); + + /* + * GBIF has same debugbus as of other GPU blocks, fall back to + * default path if GPU uses GBIF, also GBIF uses exactly same + * ID as of VBIF. + */ + if (a6xx_has_gbif(to_adreno_gpu(gpu))) { + a6xx_get_debugbus_block(gpu, a6xx_state, + &a6xx_gbif_debugbus_block, + &a6xx_state->debugbus[i]); + + a6xx_state->nr_debugbus += 1; + } } - a6xx_state->vbif_debugbus = - state_kcalloc(a6xx_state, 1, - sizeof(*a6xx_state->vbif_debugbus)); + /* Dump the VBIF debugbus on applicable targets */ + if (!a6xx_has_gbif(to_adreno_gpu(gpu))) { + a6xx_state->vbif_debugbus = + state_kcalloc(a6xx_state, 1, + sizeof(*a6xx_state->vbif_debugbus)); - if (a6xx_state->vbif_debugbus) - a6xx_get_vbif_debugbus_block(gpu, a6xx_state, - a6xx_state->vbif_debugbus); + if (a6xx_state->vbif_debugbus) + a6xx_get_vbif_debugbus_block(gpu, a6xx_state, + a6xx_state->vbif_debugbus); + } if (cxdbg) { a6xx_state->cx_debugbus = @@ -770,14 +789,16 @@ static void a6xx_get_gmu_registers(struct msm_gpu *gpu, &a6xx_state->gmu_registers[1]); } +#define A6XX_GBIF_REGLIST_SIZE 1 static void a6xx_get_registers(struct msm_gpu *gpu, struct a6xx_gpu_state *a6xx_state, struct a6xx_crashdumper *dumper) { int i, count = ARRAY_SIZE(a6xx_ahb_reglist) + ARRAY_SIZE(a6xx_reglist) + - ARRAY_SIZE(a6xx_hlsq_reglist); + ARRAY_SIZE(a6xx_hlsq_reglist) + A6XX_GBIF_REGLIST_SIZE; int index = 0; + struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu); a6xx_state->registers = state_kcalloc(a6xx_state, count, sizeof(*a6xx_state->registers)); @@ -792,6 +813,15 @@ static void a6xx_get_registers(struct msm_gpu *gpu, a6xx_state, &a6xx_ahb_reglist[i], &a6xx_state->registers[index++]); + if (a6xx_has_gbif(adreno_gpu)) + a6xx_get_ahb_gpu_registers(gpu, + a6xx_state, &a6xx_gbif_reglist, + &a6xx_state->registers[index++]); + else + a6xx_get_ahb_gpu_registers(gpu, + a6xx_state, &a6xx_vbif_reglist, + &a6xx_state->registers[index++]); + for (i = 0; i < ARRAY_SIZE(a6xx_reglist); i++) a6xx_get_crashdumper_registers(gpu, a6xx_state, &a6xx_reglist[i], diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gpu_state.h b/drivers/gpu/drm/msm/adreno/a6xx_gpu_state.h index 68cccfa2870a..e67c20c415af 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_gpu_state.h +++ b/drivers/gpu/drm/msm/adreno/a6xx_gpu_state.h @@ -1,5 +1,5 @@ // SPDX-License-Identifier: GPL-2.0 -/* Copyright (c) 2018 The Linux Foundation. All rights reserved. */ +/* Copyright (c) 2018-2019 The Linux Foundation. All rights reserved. */ #ifndef _A6XX_CRASH_DUMP_H_ #define _A6XX_CRASH_DUMP_H_ @@ -307,11 +307,20 @@ static const u32 a6xx_vbif_registers[] = { 0x3410, 0x3410, 0x3800, 0x3801, }; +static const u32 a6xx_gbif_registers[] = { + 0x3C00, 0X3C0B, 0X3C40, 0X3C47, 0X3CC0, 0X3CD1, 0xE3A, 0xE3A, +}; + static const struct a6xx_registers a6xx_ahb_reglist[] = { REGS(a6xx_ahb_registers, 0, 0), - REGS(a6xx_vbif_registers, 0, 0), }; +static const struct a6xx_registers a6xx_vbif_reglist = + REGS(a6xx_vbif_registers, 0, 0); + +static const struct a6xx_registers a6xx_gbif_reglist = + REGS(a6xx_gbif_registers, 0, 0); + static const u32 a6xx_gmu_gx_registers[] = { /* GMU GX */ 0x0000, 0x0000, 0x0010, 0x0013, 0x0016, 0x0016, 0x0018, 0x001b, @@ -422,6 +431,9 @@ static const struct a6xx_debugbus_block { DEBUGBUS(A6XX_DBGBUS_TPL1_3, 0x100), }; +static const struct a6xx_debugbus_block a6xx_gbif_debugbus_block = + DEBUGBUS(A6XX_DBGBUS_VBIF, 0x100); + static const struct a6xx_debugbus_block a6xx_cx_debugbus_blocks[] = { DEBUGBUS(A6XX_DBGBUS_GMU_CX, 0x100), DEBUGBUS(A6XX_DBGBUS_CX, 0x100), diff --git a/drivers/gpu/drm/msm/adreno/a6xx_hfi.c b/drivers/gpu/drm/msm/adreno/a6xx_hfi.c index eda11abc5f01..e450e0b97211 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_hfi.c +++ b/drivers/gpu/drm/msm/adreno/a6xx_hfi.c @@ -7,6 +7,7 @@ #include "a6xx_gmu.h" #include "a6xx_gmu.xml.h" +#include "a6xx_gpu.h" #define HFI_MSG_ID(val) [val] = #val @@ -216,48 +217,82 @@ static int a6xx_hfi_send_perf_table(struct a6xx_gmu *gmu) NULL, 0); } -static int a6xx_hfi_send_bw_table(struct a6xx_gmu *gmu) +static void a618_build_bw_table(struct a6xx_hfi_msg_bw_table *msg) { - struct a6xx_hfi_msg_bw_table msg = { 0 }; + /* Send a single "off" entry since the 618 GMU doesn't do bus scaling */ + msg->bw_level_num = 1; + + msg->ddr_cmds_num = 3; + msg->ddr_wait_bitmask = 0x01; + + msg->ddr_cmds_addrs[0] = 0x50000; + msg->ddr_cmds_addrs[1] = 0x5003c; + msg->ddr_cmds_addrs[2] = 0x5000c; + + msg->ddr_cmds_data[0][0] = 0x40000000; + msg->ddr_cmds_data[0][1] = 0x40000000; + msg->ddr_cmds_data[0][2] = 0x40000000; /* - * The sdm845 GMU doesn't do bus frequency scaling on its own but it - * does need at least one entry in the list because it might be accessed - * when the GMU is shutting down. Send a single "off" entry. + * These are the CX (CNOC) votes - these are used by the GMU but the + * votes are known and fixed for the target */ + msg->cnoc_cmds_num = 1; + msg->cnoc_wait_bitmask = 0x01; + + msg->cnoc_cmds_addrs[0] = 0x5007c; + msg->cnoc_cmds_data[0][0] = 0x40000000; + msg->cnoc_cmds_data[1][0] = 0x60000001; +} - msg.bw_level_num = 1; +static void a6xx_build_bw_table(struct a6xx_hfi_msg_bw_table *msg) +{ + /* Send a single "off" entry since the 630 GMU doesn't do bus scaling */ + msg->bw_level_num = 1; - msg.ddr_cmds_num = 3; - msg.ddr_wait_bitmask = 0x07; + msg->ddr_cmds_num = 3; + msg->ddr_wait_bitmask = 0x07; - msg.ddr_cmds_addrs[0] = 0x50000; - msg.ddr_cmds_addrs[1] = 0x5005c; - msg.ddr_cmds_addrs[2] = 0x5000c; + msg->ddr_cmds_addrs[0] = 0x50000; + msg->ddr_cmds_addrs[1] = 0x5005c; + msg->ddr_cmds_addrs[2] = 0x5000c; - msg.ddr_cmds_data[0][0] = 0x40000000; - msg.ddr_cmds_data[0][1] = 0x40000000; - msg.ddr_cmds_data[0][2] = 0x40000000; + msg->ddr_cmds_data[0][0] = 0x40000000; + msg->ddr_cmds_data[0][1] = 0x40000000; + msg->ddr_cmds_data[0][2] = 0x40000000; /* * These are the CX (CNOC) votes. This is used but the values for the * sdm845 GMU are known and fixed so we can hard code them. */ - msg.cnoc_cmds_num = 3; - msg.cnoc_wait_bitmask = 0x05; + msg->cnoc_cmds_num = 3; + msg->cnoc_wait_bitmask = 0x05; - msg.cnoc_cmds_addrs[0] = 0x50034; - msg.cnoc_cmds_addrs[1] = 0x5007c; - msg.cnoc_cmds_addrs[2] = 0x5004c; + msg->cnoc_cmds_addrs[0] = 0x50034; + msg->cnoc_cmds_addrs[1] = 0x5007c; + msg->cnoc_cmds_addrs[2] = 0x5004c; - msg.cnoc_cmds_data[0][0] = 0x40000000; - msg.cnoc_cmds_data[0][1] = 0x00000000; - msg.cnoc_cmds_data[0][2] = 0x40000000; + msg->cnoc_cmds_data[0][0] = 0x40000000; + msg->cnoc_cmds_data[0][1] = 0x00000000; + msg->cnoc_cmds_data[0][2] = 0x40000000; + + msg->cnoc_cmds_data[1][0] = 0x60000001; + msg->cnoc_cmds_data[1][1] = 0x20000001; + msg->cnoc_cmds_data[1][2] = 0x60000001; +} + + +static int a6xx_hfi_send_bw_table(struct a6xx_gmu *gmu) +{ + struct a6xx_hfi_msg_bw_table msg = { 0 }; + struct a6xx_gpu *a6xx_gpu = container_of(gmu, struct a6xx_gpu, gmu); + struct adreno_gpu *adreno_gpu = &a6xx_gpu->base; - msg.cnoc_cmds_data[1][0] = 0x60000001; - msg.cnoc_cmds_data[1][1] = 0x20000001; - msg.cnoc_cmds_data[1][2] = 0x60000001; + if (adreno_is_a618(adreno_gpu)) + a618_build_bw_table(&msg); + else + a6xx_build_bw_table(&msg); return a6xx_hfi_send_msg(gmu, HFI_H2F_MSG_BW_TABLE, &msg, sizeof(msg), NULL, 0); diff --git a/drivers/gpu/drm/msm/adreno/adreno_device.c b/drivers/gpu/drm/msm/adreno/adreno_device.c index fbbdf86504f5..cb3a6e597d76 100644 --- a/drivers/gpu/drm/msm/adreno/adreno_device.c +++ b/drivers/gpu/drm/msm/adreno/adreno_device.c @@ -167,6 +167,17 @@ static const struct adreno_info gpulist[] = { .init = a5xx_gpu_init, .zapfw = "a540_zap.mdt", }, { + .rev = ADRENO_REV(6, 1, 8, ANY_ID), + .revn = 618, + .name = "A618", + .fw = { + [ADRENO_FW_SQE] = "a630_sqe.fw", + [ADRENO_FW_GMU] = "a630_gmu.bin", + }, + .gmem = SZ_512K, + .inactive_period = DRM_MSM_INACTIVE_PERIOD, + .init = a6xx_gpu_init, + }, { .rev = ADRENO_REV(6, 3, 0, ANY_ID), .revn = 630, .name = "A630", diff --git a/drivers/gpu/drm/msm/adreno/adreno_gpu.c b/drivers/gpu/drm/msm/adreno/adreno_gpu.c index 0783e4b5486a..7fd29829b2fa 100644 --- a/drivers/gpu/drm/msm/adreno/adreno_gpu.c +++ b/drivers/gpu/drm/msm/adreno/adreno_gpu.c @@ -26,6 +26,7 @@ static int zap_shader_load_mdt(struct msm_gpu *gpu, const char *fwname, { struct device *dev = &gpu->pdev->dev; const struct firmware *fw; + const char *signed_fwname = NULL; struct device_node *np, *mem_np; struct resource r; phys_addr_t mem_phys; @@ -58,8 +59,43 @@ static int zap_shader_load_mdt(struct msm_gpu *gpu, const char *fwname, mem_phys = r.start; - /* Request the MDT file for the firmware */ - fw = adreno_request_fw(to_adreno_gpu(gpu), fwname); + /* + * Check for a firmware-name property. This is the new scheme + * to handle firmware that may be signed with device specific + * keys, allowing us to have a different zap fw path for different + * devices. + * + * If the firmware-name property is found, we bypass the + * adreno_request_fw() mechanism, because we don't need to handle + * the /lib/firmware/qcom/... vs /lib/firmware/... case. + * + * If the firmware-name property is not found, for backwards + * compatibility we fall back to the fwname from the gpulist + * table. + */ + of_property_read_string_index(np, "firmware-name", 0, &signed_fwname); + if (signed_fwname) { + fwname = signed_fwname; + ret = request_firmware_direct(&fw, fwname, gpu->dev->dev); + if (ret) + fw = ERR_PTR(ret); + } else if (fwname) { + /* Request the MDT file from the default location: */ + fw = adreno_request_fw(to_adreno_gpu(gpu), fwname); + } else { + /* + * For new targets, we require the firmware-name property, + * if a zap-shader is required, rather than falling back + * to a firmware name specified in gpulist. + * + * Because the firmware is signed with a (potentially) + * device specific key, having the name come from gpulist + * was a bad idea, and is only provided for backwards + * compatibility for older targets. + */ + return -ENODEV; + } + if (IS_ERR(fw)) { DRM_DEV_ERROR(dev, "Unable to load %s\n", fwname); return PTR_ERR(fw); @@ -95,7 +131,7 @@ static int zap_shader_load_mdt(struct msm_gpu *gpu, const char *fwname, * not. But since we've already gotten through adreno_request_fw() * we know which of the two cases it is: */ - if (to_adreno_gpu(gpu)->fwloc == FW_LOCATION_LEGACY) { + if (signed_fwname || (to_adreno_gpu(gpu)->fwloc == FW_LOCATION_LEGACY)) { ret = qcom_mdt_load(dev, fw, fwname, pasid, mem_region, mem_phys, mem_size, NULL); } else { @@ -146,14 +182,6 @@ int adreno_zap_shader_load(struct msm_gpu *gpu, u32 pasid) return -EPROBE_DEFER; } - /* Each GPU has a target specific zap shader firmware name to use */ - if (!adreno_gpu->info->zapfw) { - zap_available = false; - DRM_DEV_ERROR(&pdev->dev, - "Zap shader firmware file not specified for this target\n"); - return -ENODEV; - } - return zap_shader_load_mdt(gpu, adreno_gpu->info->zapfw, pasid); } @@ -826,7 +854,7 @@ static int adreno_get_legacy_pwrlevels(struct device *dev) node = of_get_compatible_child(dev->of_node, "qcom,gpu-pwrlevels"); if (!node) { - DRM_DEV_ERROR(dev, "Could not find the GPU powerlevels\n"); + DRM_DEV_DEBUG(dev, "Could not find the GPU powerlevels\n"); return -ENXIO; } @@ -887,10 +915,21 @@ static int adreno_get_pwrlevels(struct device *dev, DBG("fast_rate=%u, slow_rate=27000000", gpu->fast_rate); /* Check for an interconnect path for the bus */ - gpu->icc_path = of_icc_get(dev, NULL); + gpu->icc_path = of_icc_get(dev, "gfx-mem"); + if (!gpu->icc_path) { + /* + * Keep compatbility with device trees that don't have an + * interconnect-names property. + */ + gpu->icc_path = of_icc_get(dev, NULL); + } if (IS_ERR(gpu->icc_path)) gpu->icc_path = NULL; + gpu->ocmem_icc_path = of_icc_get(dev, "ocmem"); + if (IS_ERR(gpu->ocmem_icc_path)) + gpu->ocmem_icc_path = NULL; + return 0; } @@ -977,6 +1016,7 @@ void adreno_gpu_cleanup(struct adreno_gpu *adreno_gpu) release_firmware(adreno_gpu->fw[i]); icc_put(gpu->icc_path); + icc_put(gpu->ocmem_icc_path); msm_gpu_cleanup(&adreno_gpu->base); } diff --git a/drivers/gpu/drm/msm/adreno/adreno_gpu.h b/drivers/gpu/drm/msm/adreno/adreno_gpu.h index e71a7570ef72..9ff4e550e7bd 100644 --- a/drivers/gpu/drm/msm/adreno/adreno_gpu.h +++ b/drivers/gpu/drm/msm/adreno/adreno_gpu.h @@ -3,7 +3,7 @@ * Copyright (C) 2013 Red Hat * Author: Rob Clark <robdclark@gmail.com> * - * Copyright (c) 2014,2017 The Linux Foundation. All rights reserved. + * Copyright (c) 2014,2017, 2019 The Linux Foundation. All rights reserved. */ #ifndef __ADRENO_GPU_H__ @@ -227,6 +227,16 @@ static inline int adreno_is_a540(struct adreno_gpu *gpu) return gpu->revn == 540; } +static inline int adreno_is_a618(struct adreno_gpu *gpu) +{ + return gpu->revn == 618; +} + +static inline int adreno_is_a630(struct adreno_gpu *gpu) +{ + return gpu->revn == 630; +} + int adreno_get_param(struct msm_gpu *gpu, uint32_t param, uint64_t *value); const struct firmware *adreno_request_fw(struct adreno_gpu *adreno_gpu, const char *fwname); @@ -330,10 +340,7 @@ OUT_PKT7(struct msm_ringbuffer *ring, uint8_t opcode, uint16_t cnt) static inline bool adreno_reg_check(struct adreno_gpu *gpu, enum adreno_regs offset_name) { - if (offset_name >= REG_ADRENO_REGISTER_MAX || - !gpu->reg_offsets[offset_name]) { - BUG(); - } + BUG_ON(offset_name >= REG_ADRENO_REGISTER_MAX || !gpu->reg_offsets[offset_name]); /* * REG_SKIP is a special value that tell us that the register in |