From 32f0a6ddc8c98a1aade2bf3d07c79d5d2c6ceb9a Mon Sep 17 00:00:00 2001 From: Stanimir Varbanov Date: Tue, 13 Aug 2019 12:25:08 -0300 Subject: [PATCH] media: venus: Use on-chip interconnect API This aims to add a requests for bandwidth scaling depending on the resolution and framerate (macroblocks per second). The exact value of the requested bandwidth is get from a pre-calculated tables for encoder and decoder. Acked-by: Georgi Djakov Signed-off-by: Stanimir Varbanov Signed-off-by: Mauro Carvalho Chehab --- drivers/media/platform/Kconfig | 1 + drivers/media/platform/qcom/venus/core.c | 34 +++++++++++ drivers/media/platform/qcom/venus/core.h | 14 +++++ drivers/media/platform/qcom/venus/helpers.c | 67 ++++++++++++++++++++- 4 files changed, 115 insertions(+), 1 deletion(-) diff --git a/drivers/media/platform/Kconfig b/drivers/media/platform/Kconfig index 1175f122bb1d..997de1a83ff9 100644 --- a/drivers/media/platform/Kconfig +++ b/drivers/media/platform/Kconfig @@ -483,6 +483,7 @@ config VIDEO_QCOM_VENUS tristate "Qualcomm Venus V4L2 encoder/decoder driver" depends on VIDEO_DEV && VIDEO_V4L2 depends on (ARCH_QCOM && IOMMU_DMA) || COMPILE_TEST + depends on INTERCONNECT || !INTERCONNECT select QCOM_MDT_LOADER if ARCH_QCOM select QCOM_SCM if ARCH_QCOM select VIDEOBUF2_DMA_SG diff --git a/drivers/media/platform/qcom/venus/core.c b/drivers/media/platform/qcom/venus/core.c index 84e982f259a0..8d662c8f3769 100644 --- a/drivers/media/platform/qcom/venus/core.c +++ b/drivers/media/platform/qcom/venus/core.c @@ -5,6 +5,7 @@ */ #include #include +#include #include #include #include @@ -239,6 +240,14 @@ static int venus_probe(struct platform_device *pdev) if (IS_ERR(core->base)) return PTR_ERR(core->base); + core->video_path = of_icc_get(dev, "video-mem"); + if (IS_ERR(core->video_path)) + return PTR_ERR(core->video_path); + + core->cpucfg_path = of_icc_get(dev, "cpu-cfg"); + if (IS_ERR(core->cpucfg_path)) + return PTR_ERR(core->cpucfg_path); + core->irq = platform_get_irq(pdev, 0); if (core->irq < 0) return core->irq; @@ -273,6 +282,10 @@ static int venus_probe(struct platform_device *pdev) if (ret) return ret; + ret = icc_set_bw(core->cpucfg_path, 0, kbps_to_icc(1000)); + if (ret) + return ret; + ret = hfi_create(core, &venus_core_ops); if (ret) return ret; @@ -355,6 +368,9 @@ static int venus_remove(struct platform_device *pdev) pm_runtime_put_sync(dev); pm_runtime_disable(dev); + icc_put(core->video_path); + icc_put(core->cpucfg_path); + v4l2_device_unregister(&core->v4l2_dev); return ret; @@ -465,9 +481,27 @@ static const struct freq_tbl sdm845_freq_table[] = { { 244800, 100000000 }, /* 1920x1080@30 */ }; +static const struct bw_tbl sdm845_bw_table_enc[] = { + { 1944000, 1612000, 0, 2416000, 0 }, /* 3840x2160@60 */ + { 972000, 951000, 0, 1434000, 0 }, /* 3840x2160@30 */ + { 489600, 723000, 0, 973000, 0 }, /* 1920x1080@60 */ + { 244800, 370000, 0, 495000, 0 }, /* 1920x1080@30 */ +}; + +static const struct bw_tbl sdm845_bw_table_dec[] = { + { 2073600, 3929000, 0, 5551000, 0 }, /* 4096x2160@60 */ + { 1036800, 1987000, 0, 2797000, 0 }, /* 4096x2160@30 */ + { 489600, 1040000, 0, 1298000, 0 }, /* 1920x1080@60 */ + { 244800, 530000, 0, 659000, 0 }, /* 1920x1080@30 */ +}; + static const struct venus_resources sdm845_res = { .freq_tbl = sdm845_freq_table, .freq_tbl_size = ARRAY_SIZE(sdm845_freq_table), + .bw_tbl_enc = sdm845_bw_table_enc, + .bw_tbl_enc_size = ARRAY_SIZE(sdm845_bw_table_enc), + .bw_tbl_dec = sdm845_bw_table_dec, + .bw_tbl_dec_size = ARRAY_SIZE(sdm845_bw_table_dec), .clks = {"core", "iface", "bus" }, .clks_num = 3, .max_load = 3110400, /* 4096x2160@90 */ diff --git a/drivers/media/platform/qcom/venus/core.h b/drivers/media/platform/qcom/venus/core.h index 922cb7e64bfa..661eaa7b81ec 100644 --- a/drivers/media/platform/qcom/venus/core.h +++ b/drivers/media/platform/qcom/venus/core.h @@ -26,10 +26,22 @@ struct reg_val { u32 value; }; +struct bw_tbl { + u32 mbs_per_sec; + u32 avg; + u32 peak; + u32 avg_10bit; + u32 peak_10bit; +}; + struct venus_resources { u64 dma_mask; const struct freq_tbl *freq_tbl; unsigned int freq_tbl_size; + const struct bw_tbl *bw_tbl_enc; + unsigned int bw_tbl_enc_size; + const struct bw_tbl *bw_tbl_dec; + unsigned int bw_tbl_dec_size; const struct reg_val *reg_tbl; unsigned int reg_tbl_size; const char * const clks[VIDC_CLKS_NUM_MAX]; @@ -115,6 +127,8 @@ struct venus_core { struct clk *core1_clk; struct clk *core0_bus_clk; struct clk *core1_bus_clk; + struct icc_path *video_path; + struct icc_path *cpucfg_path; struct video_device *vdev_dec; struct video_device *vdev_enc; struct v4l2_device v4l2_dev; diff --git a/drivers/media/platform/qcom/venus/helpers.c b/drivers/media/platform/qcom/venus/helpers.c index 1ad96c25ab09..f18458921f5d 100644 --- a/drivers/media/platform/qcom/venus/helpers.c +++ b/drivers/media/platform/qcom/venus/helpers.c @@ -5,6 +5,7 @@ */ #include #include +#include #include #include #include @@ -388,6 +389,65 @@ static u32 load_per_type(struct venus_core *core, u32 session_type) return mbs_per_sec; } +static void mbs_to_bw(struct venus_inst *inst, u32 mbs, u32 *avg, u32 *peak) +{ + const struct venus_resources *res = inst->core->res; + const struct bw_tbl *bw_tbl; + unsigned int num_rows, i; + + *avg = 0; + *peak = 0; + + if (mbs == 0) + return; + + if (inst->session_type == VIDC_SESSION_TYPE_ENC) { + num_rows = res->bw_tbl_enc_size; + bw_tbl = res->bw_tbl_enc; + } else if (inst->session_type == VIDC_SESSION_TYPE_DEC) { + num_rows = res->bw_tbl_dec_size; + bw_tbl = res->bw_tbl_dec; + } else { + return; + } + + if (!bw_tbl || num_rows == 0) + return; + + for (i = 0; i < num_rows; i++) { + if (mbs > bw_tbl[i].mbs_per_sec) + break; + + if (inst->dpb_fmt & HFI_COLOR_FORMAT_10_BIT_BASE) { + *avg = bw_tbl[i].avg_10bit; + *peak = bw_tbl[i].peak_10bit; + } else { + *avg = bw_tbl[i].avg; + *peak = bw_tbl[i].peak; + } + } +} + +static int load_scale_bw(struct venus_core *core) +{ + struct venus_inst *inst = NULL; + u32 mbs_per_sec, avg, peak, total_avg = 0, total_peak = 0; + + mutex_lock(&core->lock); + list_for_each_entry(inst, &core->instances, list) { + mbs_per_sec = load_per_instance(inst); + mbs_to_bw(inst, mbs_per_sec, &avg, &peak); + total_avg += avg; + total_peak += peak; + } + mutex_unlock(&core->lock); + + dev_dbg(core->dev, "total: avg_bw: %u, peak_bw: %u\n", + total_avg, total_peak); + + return icc_set_bw(core->video_path, total_avg, total_peak); +} + int venus_helper_load_scale_clocks(struct venus_core *core) { const struct freq_tbl *table = core->res->freq_tbl; @@ -431,10 +491,15 @@ int venus_helper_load_scale_clocks(struct venus_core *core) if (ret) goto err; + ret = load_scale_bw(core); + if (ret) + goto err; + return 0; err: - dev_err(dev, "failed to set clock rate %lu (%d)\n", freq, ret); + dev_err(dev, "failed to set clock rate %lu or bandwidth (%d)\n", + freq, ret); return ret; } EXPORT_SYMBOL_GPL(venus_helper_load_scale_clocks);