dma-mapping: add benchmark support for streaming DMA APIs

Nowadays, there are increasing requirements to benchmark the performance
of dma_map and dma_unmap particually while the device is attached to an
IOMMU.

This patch enables the support. Users can run specified number of threads
to do dma_map_page and dma_unmap_page on a specific NUMA node with the
specified duration. Then dma_map_benchmark will calculate the average
latency for map and unmap.

A difficulity for this benchmark is that dma_map/unmap APIs must run on
a particular device. Each device might have different backend of IOMMU or
non-IOMMU.

So we use the driver_override to bind dma_map_benchmark to a particual
device by:
For platform devices:
echo dma_map_benchmark > /sys/bus/platform/devices/xxx/driver_override
echo xxx > /sys/bus/platform/drivers/xxx/unbind
echo xxx > /sys/bus/platform/drivers/dma_map_benchmark/bind

For PCI devices:
echo dma_map_benchmark > /sys/bus/pci/devices/0000:00:01.0/driver_override
echo 0000:00:01.0 > /sys/bus/pci/drivers/xxx/unbind
echo 0000:00:01.0 > /sys/bus/pci/drivers/dma_map_benchmark/bind

Cc: Will Deacon <will@kernel.org>
Cc: Shuah Khan <shuah@kernel.org>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Marek Szyprowski <m.szyprowski@samsung.com>
Cc: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Barry Song <song.bao.hua@hisilicon.com>
[hch: folded in two fixes from Colin Ian King <colin.king@canonical.com>]
Signed-off-by: Christoph Hellwig <hch@lst.de>

3 files changed, 371 insertions, 0 deletions

diff --git a/kernel/dma/Kconfig b/kernel/dma/Kconfig
index 43d106598e82..ba50d4588bc5 100644
--- a/kernel/dma/Kconfig
+++ b/kernel/dma/Kconfig
@@ -229,3 +229,12 @@ config DMA_API_DEBUG_SG
 	  is technically out-of-spec.
 
 	  If unsure, say N.
+
+config DMA_MAP_BENCHMARK
+	bool "Enable benchmarking of streaming DMA mapping"
+	depends on DEBUG_FS
+	help
+	  Provides /sys/kernel/debug/dma_map_benchmark that helps with testing
+	  performance of dma_(un)map_page.
+
+	  See tools/testing/selftests/dma/dma_map_benchmark.c
diff --git a/kernel/dma/Makefile b/kernel/dma/Makefile
index dc755ab68aab..7aa6b26b1348 100644
--- a/kernel/dma/Makefile
+++ b/kernel/dma/Makefile
@@ -10,3 +10,4 @@ obj-$(CONFIG_DMA_API_DEBUG)		+= debug.o
 obj-$(CONFIG_SWIOTLB)			+= swiotlb.o
 obj-$(CONFIG_DMA_COHERENT_POOL)		+= pool.o
 obj-$(CONFIG_DMA_REMAP)			+= remap.o
+obj-$(CONFIG_DMA_MAP_BENCHMARK)		+= map_benchmark.o
diff --git a/kernel/dma/map_benchmark.c b/kernel/dma/map_benchmark.c
new file mode 100644
index 000000000000..b1496e744c68
--- /dev/null
+++ b/kernel/dma/map_benchmark.c
@@ -0,0 +1,361 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2020 Hisilicon Limited.
+ */
+
+#define pr_fmt(fmt)	KBUILD_MODNAME ": " fmt
+
+#include <linux/debugfs.h>
+#include <linux/delay.h>
+#include <linux/device.h>
+#include <linux/dma-mapping.h>
+#include <linux/kernel.h>
+#include <linux/kthread.h>
+#include <linux/math64.h>
+#include <linux/module.h>
+#include <linux/pci.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+#include <linux/timekeeping.h>
+
+#define DMA_MAP_BENCHMARK	_IOWR('d', 1, struct map_benchmark)
+#define DMA_MAP_MAX_THREADS	1024
+#define DMA_MAP_MAX_SECONDS	300
+
+#define DMA_MAP_BIDIRECTIONAL	0
+#define DMA_MAP_TO_DEVICE	1
+#define DMA_MAP_FROM_DEVICE	2
+
+struct map_benchmark {
+	__u64 avg_map_100ns; /* average map latency in 100ns */
+	__u64 map_stddev; /* standard deviation of map latency */
+	__u64 avg_unmap_100ns; /* as above */
+	__u64 unmap_stddev;
+	__u32 threads; /* how many threads will do map/unmap in parallel */
+	__u32 seconds; /* how long the test will last */
+	__s32 node; /* which numa node this benchmark will run on */
+	__u32 dma_bits; /* DMA addressing capability */
+	__u32 dma_dir; /* DMA data direction */
+	__u64 expansion[10];	/* For future use */
+};
+
+struct map_benchmark_data {
+	struct map_benchmark bparam;
+	struct device *dev;
+	struct dentry  *debugfs;
+	enum dma_data_direction dir;
+	atomic64_t sum_map_100ns;
+	atomic64_t sum_unmap_100ns;
+	atomic64_t sum_sq_map;
+	atomic64_t sum_sq_unmap;
+	atomic64_t loops;
+};
+
+static int map_benchmark_thread(void *data)
+{
+	void *buf;
+	dma_addr_t dma_addr;
+	struct map_benchmark_data *map = data;
+	int ret = 0;
+
+	buf = (void *)__get_free_page(GFP_KERNEL);
+	if (!buf)
+		return -ENOMEM;
+
+	while (!kthread_should_stop())  {
+		u64 map_100ns, unmap_100ns, map_sq, unmap_sq;
+		ktime_t map_stime, map_etime, unmap_stime, unmap_etime;
+		ktime_t map_delta, unmap_delta;
+
+		/*
+		 * for a non-coherent device, if we don't stain them in the
+		 * cache, this will give an underestimate of the real-world
+		 * overhead of BIDIRECTIONAL or TO_DEVICE mappings;
+		 * 66 means evertything goes well! 66 is lucky.
+		 */
+		if (map->dir != DMA_FROM_DEVICE)
+			memset(buf, 0x66, PAGE_SIZE);
+
+		map_stime = ktime_get();
+		dma_addr = dma_map_single(map->dev, buf, PAGE_SIZE, map->dir);
+		if (unlikely(dma_mapping_error(map->dev, dma_addr))) {
+			pr_err("dma_map_single failed on %s\n",
+				dev_name(map->dev));
+			ret = -ENOMEM;
+			goto out;
+		}
+		map_etime = ktime_get();
+		map_delta = ktime_sub(map_etime, map_stime);
+
+		unmap_stime = ktime_get();
+		dma_unmap_single(map->dev, dma_addr, PAGE_SIZE, map->dir);
+		unmap_etime = ktime_get();
+		unmap_delta = ktime_sub(unmap_etime, unmap_stime);
+
+		/* calculate sum and sum of squares */
+
+		map_100ns = div64_ul(map_delta,  100);
+		unmap_100ns = div64_ul(unmap_delta, 100);
+		map_sq = map_100ns * map_100ns;
+		unmap_sq = unmap_100ns * unmap_100ns;
+
+		atomic64_add(map_100ns, &map->sum_map_100ns);
+		atomic64_add(unmap_100ns, &map->sum_unmap_100ns);
+		atomic64_add(map_sq, &map->sum_sq_map);
+		atomic64_add(unmap_sq, &map->sum_sq_unmap);
+		atomic64_inc(&map->loops);
+	}
+
+out:
+	free_page((unsigned long)buf);
+	return ret;
+}
+
+static int do_map_benchmark(struct map_benchmark_data *map)
+{
+	struct task_struct **tsk;
+	int threads = map->bparam.threads;
+	int node = map->bparam.node;
+	const cpumask_t *cpu_mask = cpumask_of_node(node);
+	u64 loops;
+	int ret = 0;
+	int i;
+
+	tsk = kmalloc_array(threads, sizeof(*tsk), GFP_KERNEL);
+	if (!tsk)
+		return -ENOMEM;
+
+	get_device(map->dev);
+
+	for (i = 0; i < threads; i++) {
+		tsk[i] = kthread_create_on_node(map_benchmark_thread, map,
+				map->bparam.node, "dma-map-benchmark/%d", i);
+		if (IS_ERR(tsk[i])) {
+			pr_err("create dma_map thread failed\n");
+			ret = PTR_ERR(tsk[i]);
+			goto out;
+		}
+
+		if (node != NUMA_NO_NODE)
+			kthread_bind_mask(tsk[i], cpu_mask);
+	}
+
+	/* clear the old value in the previous benchmark */
+	atomic64_set(&map->sum_map_100ns, 0);
+	atomic64_set(&map->sum_unmap_100ns, 0);
+	atomic64_set(&map->sum_sq_map, 0);
+	atomic64_set(&map->sum_sq_unmap, 0);
+	atomic64_set(&map->loops, 0);
+
+	for (i = 0; i < threads; i++)
+		wake_up_process(tsk[i]);
+
+	msleep_interruptible(map->bparam.seconds * 1000);
+
+	/* wait for the completion of benchmark threads */
+	for (i = 0; i < threads; i++) {
+		ret = kthread_stop(tsk[i]);
+		if (ret)
+			goto out;
+	}
+
+	loops = atomic64_read(&map->loops);
+	if (likely(loops > 0)) {
+		u64 map_variance, unmap_variance;
+		u64 sum_map = atomic64_read(&map->sum_map_100ns);
+		u64 sum_unmap = atomic64_read(&map->sum_unmap_100ns);
+		u64 sum_sq_map = atomic64_read(&map->sum_sq_map);
+		u64 sum_sq_unmap = atomic64_read(&map->sum_sq_unmap);
+
+		/* average latency */
+		map->bparam.avg_map_100ns = div64_u64(sum_map, loops);
+		map->bparam.avg_unmap_100ns = div64_u64(sum_unmap, loops);
+
+		/* standard deviation of latency */
+		map_variance = div64_u64(sum_sq_map, loops) -
+				map->bparam.avg_map_100ns *
+				map->bparam.avg_map_100ns;
+		unmap_variance = div64_u64(sum_sq_unmap, loops) -
+				map->bparam.avg_unmap_100ns *
+				map->bparam.avg_unmap_100ns;
+		map->bparam.map_stddev = int_sqrt64(map_variance);
+		map->bparam.unmap_stddev = int_sqrt64(unmap_variance);
+	}
+
+out:
+	put_device(map->dev);
+	kfree(tsk);
+	return ret;
+}
+
+static long map_benchmark_ioctl(struct file *file, unsigned int cmd,
+		unsigned long arg)
+{
+	struct map_benchmark_data *map = file->private_data;
+	void __user *argp = (void __user *)arg;
+	u64 old_dma_mask;
+
+	int ret;
+
+	if (copy_from_user(&map->bparam, argp, sizeof(map->bparam)))
+		return -EFAULT;
+
+	switch (cmd) {
+	case DMA_MAP_BENCHMARK:
+		if (map->bparam.threads == 0 ||
+		    map->bparam.threads > DMA_MAP_MAX_THREADS) {
+			pr_err("invalid thread number\n");
+			return -EINVAL;
+		}
+
+		if (map->bparam.seconds == 0 ||
+		    map->bparam.seconds > DMA_MAP_MAX_SECONDS) {
+			pr_err("invalid duration seconds\n");
+			return -EINVAL;
+		}
+
+		if (map->bparam.node != NUMA_NO_NODE &&
+		    !node_possible(map->bparam.node)) {
+			pr_err("invalid numa node\n");
+			return -EINVAL;
+		}
+
+		switch (map->bparam.dma_dir) {
+		case DMA_MAP_BIDIRECTIONAL:
+			map->dir = DMA_BIDIRECTIONAL;
+			break;
+		case DMA_MAP_FROM_DEVICE:
+			map->dir = DMA_FROM_DEVICE;
+			break;
+		case DMA_MAP_TO_DEVICE:
+			map->dir = DMA_TO_DEVICE;
+			break;
+		default:
+			pr_err("invalid DMA direction\n");
+			return -EINVAL;
+		}
+
+		old_dma_mask = dma_get_mask(map->dev);
+
+		ret = dma_set_mask(map->dev,
+				   DMA_BIT_MASK(map->bparam.dma_bits));
+		if (ret) {
+			pr_err("failed to set dma_mask on device %s\n",
+				dev_name(map->dev));
+			return -EINVAL;
+		}
+
+		ret = do_map_benchmark(map);
+
+		/*
+		 * restore the original dma_mask as many devices' dma_mask are
+		 * set by architectures, acpi, busses. When we bind them back
+		 * to their original drivers, those drivers shouldn't see
+		 * dma_mask changed by benchmark
+		 */
+		dma_set_mask(map->dev, old_dma_mask);
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	if (copy_to_user(argp, &map->bparam, sizeof(map->bparam)))
+		return -EFAULT;
+
+	return ret;
+}
+
+static const struct file_operations map_benchmark_fops = {
+	.open			= simple_open,
+	.unlocked_ioctl		= map_benchmark_ioctl,
+};
+
+static void map_benchmark_remove_debugfs(void *data)
+{
+	struct map_benchmark_data *map = (struct map_benchmark_data *)data;
+
+	debugfs_remove(map->debugfs);
+}
+
+static int __map_benchmark_probe(struct device *dev)
+{
+	struct dentry *entry;
+	struct map_benchmark_data *map;
+	int ret;
+
+	map = devm_kzalloc(dev, sizeof(*map), GFP_KERNEL);
+	if (!map)
+		return -ENOMEM;
+	map->dev = dev;
+
+	ret = devm_add_action(dev, map_benchmark_remove_debugfs, map);
+	if (ret) {
+		pr_err("Can't add debugfs remove action\n");
+		return ret;
+	}
+
+	/*
+	 * we only permit a device bound with this driver, 2nd probe
+	 * will fail
+	 */
+	entry = debugfs_create_file("dma_map_benchmark", 0600, NULL, map,
+			&map_benchmark_fops);
+	if (IS_ERR(entry))
+		return PTR_ERR(entry);
+	map->debugfs = entry;
+
+	return 0;
+}
+
+static int map_benchmark_platform_probe(struct platform_device *pdev)
+{
+	return __map_benchmark_probe(&pdev->dev);
+}
+
+static struct platform_driver map_benchmark_platform_driver = {
+	.driver		= {
+		.name	= "dma_map_benchmark",
+	},
+	.probe = map_benchmark_platform_probe,
+};
+
+static int
+map_benchmark_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id)
+{
+	return __map_benchmark_probe(&pdev->dev);
+}
+
+static struct pci_driver map_benchmark_pci_driver = {
+	.name	= "dma_map_benchmark",
+	.probe	= map_benchmark_pci_probe,
+};
+
+static int __init map_benchmark_init(void)
+{
+	int ret;
+
+	ret = pci_register_driver(&map_benchmark_pci_driver);
+	if (ret)
+		return ret;
+
+	ret = platform_driver_register(&map_benchmark_platform_driver);
+	if (ret) {
+		pci_unregister_driver(&map_benchmark_pci_driver);
+		return ret;
+	}
+
+	return 0;
+}
+
+static void __exit map_benchmark_cleanup(void)
+{
+	platform_driver_unregister(&map_benchmark_platform_driver);
+	pci_unregister_driver(&map_benchmark_pci_driver);
+}
+
+module_init(map_benchmark_init);
+module_exit(map_benchmark_cleanup);
+
+MODULE_AUTHOR("Barry Song <song.bao.hua@hisilicon.com>");
+MODULE_DESCRIPTION("dma_map benchmark driver");
+MODULE_LICENSE("GPL");