This is a note to let you know that I've just added the patch titled drm/msm: simplify gpu_busy callback to the 5.18-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary The filename of the patch is: drm-msm-simplify-gpu_busy-callback.patch and it can be found in the queue-5.18 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable@xxxxxxxxxxxxxxx> know about it. commit 09cc842722fc60c24487ffa8b8ebaa5c69b20ecf Author: Chia-I Wu <olvaffe@xxxxxxxxx> Date: Fri Apr 15 17:33:13 2022 -0700 drm/msm: simplify gpu_busy callback [ Upstream commit 15c411980bacddf294452fd1cf7308b14f3f8c63 ] Move tracking and busy time calculation to msm_devfreq_get_dev_status. Signed-off-by: Chia-I Wu <olvaffe@xxxxxxxxx> Cc: Rob Clark <robdclark@xxxxxxxxxxxx> Link: https://lore.kernel.org/r/20220416003314.59211-2-olvaffe@xxxxxxxxx Signed-off-by: Rob Clark <robdclark@xxxxxxxxxxxx> Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx> diff --git a/drivers/gpu/drm/msm/adreno/a5xx_gpu.c b/drivers/gpu/drm/msm/adreno/a5xx_gpu.c index 407f50a15faa..217615e0e850 100644 --- a/drivers/gpu/drm/msm/adreno/a5xx_gpu.c +++ b/drivers/gpu/drm/msm/adreno/a5xx_gpu.c @@ -1662,28 +1662,23 @@ static struct msm_ringbuffer *a5xx_active_ring(struct msm_gpu *gpu) return a5xx_gpu->cur_ring; } -static unsigned long a5xx_gpu_busy(struct msm_gpu *gpu) +static u64 a5xx_gpu_busy(struct msm_gpu *gpu, unsigned long *out_sample_rate) { - u64 busy_cycles, busy_time; + u64 busy_cycles; /* Only read the gpu busy if the hardware is already active */ - if (pm_runtime_get_if_in_use(&gpu->pdev->dev) == 0) + if (pm_runtime_get_if_in_use(&gpu->pdev->dev) == 0) { + *out_sample_rate = 1; return 0; + } busy_cycles = gpu_read64(gpu, REG_A5XX_RBBM_PERFCTR_RBBM_0_LO, REG_A5XX_RBBM_PERFCTR_RBBM_0_HI); - - busy_time = busy_cycles - gpu->devfreq.busy_cycles; - do_div(busy_time, clk_get_rate(gpu->core_clk) / 1000000); - - gpu->devfreq.busy_cycles = busy_cycles; + *out_sample_rate = clk_get_rate(gpu->core_clk); pm_runtime_put(&gpu->pdev->dev); - if (WARN_ON(busy_time > ~0LU)) - return ~0LU; - - return (unsigned long)busy_time; + return busy_cycles; } static uint32_t a5xx_get_rptr(struct msm_gpu *gpu, struct msm_ringbuffer *ring) diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c index a8f6d73197b1..40fb92becc78 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c +++ b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c @@ -1649,12 +1649,14 @@ static void a6xx_destroy(struct msm_gpu *gpu) kfree(a6xx_gpu); } -static unsigned long a6xx_gpu_busy(struct msm_gpu *gpu) +static u64 a6xx_gpu_busy(struct msm_gpu *gpu, unsigned long *out_sample_rate) { struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu); struct a6xx_gpu *a6xx_gpu = to_a6xx_gpu(adreno_gpu); - u64 busy_cycles, busy_time; + u64 busy_cycles; + /* 19.2MHz */ + *out_sample_rate = 19200000; /* Only read the gpu busy if the hardware is already active */ if (pm_runtime_get_if_in_use(a6xx_gpu->gmu.dev) == 0) @@ -1664,17 +1666,10 @@ static unsigned long a6xx_gpu_busy(struct msm_gpu *gpu) REG_A6XX_GMU_CX_GMU_POWER_COUNTER_XOCLK_0_L, REG_A6XX_GMU_CX_GMU_POWER_COUNTER_XOCLK_0_H); - busy_time = (busy_cycles - gpu->devfreq.busy_cycles) * 10; - do_div(busy_time, 192); - - gpu->devfreq.busy_cycles = busy_cycles; pm_runtime_put(a6xx_gpu->gmu.dev); - if (WARN_ON(busy_time > ~0LU)) - return ~0LU; - - return (unsigned long)busy_time; + return busy_cycles; } static void a6xx_gpu_set_freq(struct msm_gpu *gpu, struct dev_pm_opp *opp) diff --git a/drivers/gpu/drm/msm/msm_gpu.h b/drivers/gpu/drm/msm/msm_gpu.h index 02419f2ca2bc..389c6dab751b 100644 --- a/drivers/gpu/drm/msm/msm_gpu.h +++ b/drivers/gpu/drm/msm/msm_gpu.h @@ -62,7 +62,7 @@ struct msm_gpu_funcs { /* for generation specific debugfs: */ void (*debugfs_init)(struct msm_gpu *gpu, struct drm_minor *minor); #endif - unsigned long (*gpu_busy)(struct msm_gpu *gpu); + u64 (*gpu_busy)(struct msm_gpu *gpu, unsigned long *out_sample_rate); struct msm_gpu_state *(*gpu_state_get)(struct msm_gpu *gpu); int (*gpu_state_put)(struct msm_gpu_state *state); unsigned long (*gpu_get_freq)(struct msm_gpu *gpu); @@ -106,11 +106,8 @@ struct msm_gpu_devfreq { struct dev_pm_qos_request boost_freq; /** - * busy_cycles: - * - * Used by implementation of gpu->gpu_busy() to track the last - * busy counter value, for calculating elapsed busy cycles since - * last sampling period. + * busy_cycles: Last busy counter value, for calculating elapsed busy + * cycles since last sampling period. */ u64 busy_cycles; diff --git a/drivers/gpu/drm/msm/msm_gpu_devfreq.c b/drivers/gpu/drm/msm/msm_gpu_devfreq.c index 12641616acd3..d2b4c646a0ae 100644 --- a/drivers/gpu/drm/msm/msm_gpu_devfreq.c +++ b/drivers/gpu/drm/msm/msm_gpu_devfreq.c @@ -49,18 +49,38 @@ static unsigned long get_freq(struct msm_gpu *gpu) return clk_get_rate(gpu->core_clk); } -static int msm_devfreq_get_dev_status(struct device *dev, +static void get_raw_dev_status(struct msm_gpu *gpu, struct devfreq_dev_status *status) { - struct msm_gpu *gpu = dev_to_gpu(dev); + struct msm_gpu_devfreq *df = &gpu->devfreq; + u64 busy_cycles, busy_time; + unsigned long sample_rate; ktime_t time; status->current_frequency = get_freq(gpu); - status->busy_time = gpu->funcs->gpu_busy(gpu); - + busy_cycles = gpu->funcs->gpu_busy(gpu, &sample_rate); time = ktime_get(); - status->total_time = ktime_us_delta(time, gpu->devfreq.time); - gpu->devfreq.time = time; + + busy_time = busy_cycles - df->busy_cycles; + status->total_time = ktime_us_delta(time, df->time); + + df->busy_cycles = busy_cycles; + df->time = time; + + busy_time *= USEC_PER_SEC; + do_div(busy_time, sample_rate); + if (WARN_ON(busy_time > ~0LU)) + busy_time = ~0LU; + + status->busy_time = busy_time; +} + +static int msm_devfreq_get_dev_status(struct device *dev, + struct devfreq_dev_status *status) +{ + struct msm_gpu *gpu = dev_to_gpu(dev); + + get_raw_dev_status(gpu, status); return 0; }